On the Regulation of Mitochondrial Fusion, Fission and Mitochondrial DNA

University of Calgary PRISM: University of Calgary's Digital Repository

Graduate Studies The Vault: Electronic Theses and Dissertations

2020-04-13 On the Regulation of Mitochondrial Fusion, Fission and Mitochondrial DNA

Sabouny, Rasha

Sabouny, R. (2020). On the Regulation of Mitochondrial Fusion, Fission and Mitochondrial DNA (Unpublished doctoral thesis). University of Calgary, Calgary, AB. http://hdl.handle.net/1880/111801 doctoral thesis

University of Calgary graduate students retain copyright ownership and moral rights for their thesis. You may use this material in any way that is permitted by the Copyright Act or through licensing that has been assigned to the document. For uses that are not allowable under copyright legislation or licensing, you are required to seek permission. Downloaded from PRISM: https://prism.ucalgary.ca UNIVERSITY OF CALGARY

On the Regulation of Mitochondrial Fusion, Fission and Mitochondrial DNA

Rasha Sabouny

A THESIS

SUBMITTED TO THE FACULTY OF GRADUATE STUDIES

IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE

DEGREE OF DOCTOR OF PHILOSOPHY

GRADUATE PROGRAM IN BIOCHEMISTRY AND MOLECULAR BIOLOGY

CALGARY, ALBERTA

APRIL, 2020

Abstract

Mitochondria are functionally and structurally fascinating organelles, well known for their role as the cellular powerhouse. Unlike other membrane bound organelles, mitochondria maintain their own genome (mtDNA), which is present in hundreds of copies per cell, packaged into nucleo-protein structures known as nucleoids. An important regulator of mitochondrial function is their dynamic nature, whereby ongoing fusion and fission events remodel mitochondrial network morphology and influence mitochondrial activity. Dynamic fusion and fission forces are also key for distributing mtDNA nucleoids throughout mitochondrial networks and the maintenance of mtDNA copy number. Notably, mutations in core mitochondrial fusion

(MFN2, OPA1) and fission (DRP1) proteins lead to enlarged nucleoids, mtDNA depletion and cause severe mitochondrial diseases. However, we do not completely understand how or why these processes are important for mtDNA. Additionally, there remains a lot to be learned about the molecular regulators mediating fusion and fission of mitochondrial networks.

This project set out to characterize novel mitochondrial fusion and fission factors and further understand how defective fusion and fission regulation influence mtDNA dynamics. The work outlined in this thesis showcases three nuclear-encoded mitochondrial disease genes

(FBXL4, MSTO1 & MYH14) implicated as regulators of mitochondrial morphology and shown to be important for mtDNA regulation. Firstly, this work characterizes an established mtDNA depletion syndrome gene, FBXL4 and provides the first evidence that FBXL4 protein is a mitochondrial fusion regulator. Secondly, MSTO1, a recently described cytosolic fusion regulator, is highlighted as perturbations in MSTO1 pro-fusion activity gives rise to mtDNA depletion and altered nucleoid distribution. Lastly, the largely uncharacterized non-muscle myosin protein, NMIIC, encoded by MYH14, is highlighted as novel component of the mitochondrial fission machinery. A pathogenic mutation in MYH14 causing peripheral neuropathy reduces fission and adversely affects the distribution of mtDNA nucleoids,

ii particularly at the cell periphery. Through genetic and pharmacological rescue approaches to restore mitochondrial network morphology in these models, this work contributes to our understanding on the interplay between fusion and fission dynamics and mtDNA maintenance.

iii

Preface

This thesis is written in a manuscript-based format, as such there may be some redundancy in the information included in introduction sections. In order to minimize repetition in materials & methods sections, a separate chapter (Chapter 2) was added compiling details on all methodology used to produce this work, and respective methods sections were removed from main chapters. A final bibliography section was added combining all referenced literature.

Chapter 1 includes sections from a literature review published as Sabouny R & Shutt TE.

“Mitochondrial Fission and Fusion are Reciprocally Regulated”. 2020. Trends in Biochemical

Sciences. (https://doi.org/10.1016/j.tibs.2020.03.009)

Chapter 3 of this thesis was published as Sabouny R, Wong R, Lee-Glover L, Greenway

SC, Sinasac D, Care4Rare Canada Consortium, Khan A, Shutt TE. “Characterization of the

C584R variant in the mtDNA depletion syndrome gene FBXL4, reveals a novel role for FBXL4 as a regulator of mitochondrial fusion”. 2019. BBA Molecular Basis of Disease;1865(11):165536

Chapter 4 has been published as Donkervoort S*, Sabouny R*, Yun P, Gauquelin L,

Chao KR, Hu Y, Al Khatib I, Töpf A, Mohassel P, Cummings BB, Kaur R, Saade D, Moore SA,

Waddell LB, Farrar MA, Goodrich JK, Uapinyoying P, Chan SHS, Javed A, Leach ME,

Karachunski P, Dalton J, Medne L, Harper A, Thompson C, Thiffault I, Specht S, Lamont RE,

Saunders C, Racher H, Bernier FP, Mowat D, Witting N, Vissing J, Hanson R, Coffman KA,

Hainlen M, Parboosingh JS, Carnevale A, Yoon G, Schnur RE; Care4Rare Canada Consortium,

Boycott KM, Mah JK, Straub V, Foley AR, Innes AM, Bönnemann CG, Shutt TE. “MSTO1 mutations cause mtDNA depletion, manifesting as muscular dystrophy with cerebellar involvement”. 2019. Acta Neuropathologica, doi: 10.1007/s00401-019-02059-z

(*co-first authors)

Chapter 5 includes data that were published as Almutawa W, Smith C, Sabouny R, Smit

RB, Zhao T, Wong R, Lee-Glover L, Desrochers-Goyette J, Ilamathi HS; Care4Rare Canada

Consortium, Suchowersky O, Germain M, Mains PE, Parboosingh JS, Pfeffer G, Innes AM,

Shutt TE. “The R941L mutation in MYH14 disrupts mitochondrial fission and associates with peripheral neuropathy”. 2019. EBioMedicine; 45:379-392.

Acknowledgements

Firstly, I would like to thank my supervisor Dr. Tim Shutt for introducing me to the world of mitochondria! Tim, I am very grateful to your unconditional support, your contagious enthusiasm for science and your kind spirit. Thank you for mentoring me to be a great scientist and a better person. I am very fortunate to have joined your lab and if time goes back, I would do it all over again in a heartbeat!

To my colleagues at the Shutt Lab, Iman Al Khatib, Tian Zhao, Erik Fraunberger, Dr.

Walaa Al Mutawa, Sandra Nishikawa, Dr. Govinda Sharma, Dr. Mezbah Uddin, Laurie Lee-

Glover, Alon Gilad, Rachel Wong, Liam Aleksiuk and Rafa Abbas. Thank you for making everyday fun and always finding a reason to ‘celebrate’. I will miss the Shutt Lab shenanigans!

Iman, thank you for reminding me that ‘finished is better than perfect!’ and for showing me that with determination it is possible to be Super-Mom-Super-Scientist! Tian, thank you for looking out for me and sharing lab snacks and free food location updates! Erik, Christmas carols still do not sound the same and I am sure this statement will be true 20 years from now. Thank you for bringing your lightheartedness and laughter to many lab mission-impossibles! And to my awesome summer students, Laurie, Alon, Rachel, Rafa and Liam, thank you for making this whole journey more worthwhile, I loved every moment of working with you, learning from you and teaching you!

Ms Lisa Mesluk, thank you for being an absolute star, helping me navigate administrative hurdles, find mysteriously missing lab orders and taking care of business!

My committee members, Dr. David Schriemer and Dr. David Sinasac, thank you for your support throughout my PhD, mentoring me to be a good researcher and always reminding me to keep the big picture in mind! Also, thanks to my internal examiner Dr. Jennifer Corcoran, and my external examiner Dr. Suzanne Hoppins. I am truly grateful to your support amidst the

COVID-19 global crisis. I am hoping that this will be an amusing memory to share down the road.

This work would not have been possible without our collaborations with Dr. Aneal Khan,

Dr. Michael Innes, Dr. Steve Greenway, Dr. Gerald Pfeffer, Dr. Jillian Parboosingh and the

Medical Genetics team at Alberta Children’s Hospital. Additionally, thanks to our collaborators at

NIH, Dr. Sandra Donkervoort and Dr. Carsten Bönnemann, and at SickKids, Dr. Grace Yoon, and their respective teams, for the opportunity to work together. Thank you all for providing us with patient fibroblasts and sharing your expertise to study many mitochondrial disease cases.

My friends, thank you for putting up with me despite the many times I cancelled our plans, dragged you to the lab on weekends or made you wait while I wrap up experiments.

Thank you for keeping me sane!

And finally, my family, thank you for constantly reminding me that there is life outside the lab and ‘trying’ not to ask how many more years till I graduate and when I would get a real job.

Thank you for loving and supporting me, I really could not have gotten this far without you.

vii

Dedication

To my parents,

Hala & Kadry,

my siblings,

Abdullah & Farah.

Thank you for believing in me.

“I solemnly swear that I am up to no good.” – Moony, Wormtail, Padfoot & Prongs

viii

Table of Contents Abstract...... ii Preface ...... iv Acknowledgements ...... vi Dedication ...... viii List of Tables ...... xiii List of Figures & Illustrations ...... xiv List of Symbols, Abbreviations, & Nomenclature ...... xxi INTRODUCTION ...... 26 1. Mitochondria ...... 27 1.1 Mitochondrial Dynamics ...... 28 1.2 Why do mitochondria fuse and divide? ...... 29 1.3 Mitochondrial Fusion Regulation ...... 34 1.3.1 Core Fusion GTPases ...... 34 1.3.2 Auxiliary Fusion Regulators ...... 36 1.4 Mitochondrial Fission Regulation ...... 37 1.4.1 Core Fission Machinery ...... 37 1.4.2 Auxiliary Fission Regulators ...... 39 1.5 Consequences of impaired fusion or fission ...... 39 1.5.1 Mitochondrial Dynamics and Mitochondrial Diseases ...... 40 1.5.2 mtDNA depletion syndrome ...... 41 1.6 Mitochondrial DNA and Links to Mitochondrial Dynamics ...... 43 1.6.1 MtDNA Nucleoid Organization ...... 43 1.6.2 MtDNA Replication ...... 45 1.6.3 MtDNA Copy Number...... 46 1.6.4 MtDNA Mutations & Quality Control ...... 47 1.7 Summary of Main Chapters ...... 49 1.7.1 Statement of Contribution ...... 51 MATERIALS & METHODS ...... 53 2.1 Ethics Statement ...... 54 2.2 Cell Culture ...... 54 2.3 Plasmids & Cloning ...... 54 2.4 Transfections ...... 55 2.5 Cell Sorting ...... 56 2.6 mtDNA Copy Number Analysis ...... 56

2.7 Long-range PCR ...... 56 2.8 Western Blot ...... 57 2.8.1 Densitometry Analysis: ...... 57 2.9 Mitochondrial Respiration ...... 58 2.10 Immunofluorescence Staining ...... 58 2.11 Live Cell Staining ...... 58 2.12 Microscopy ...... 59 2.13 Photo-activatable GFP fusion assay ...... 59 2.14 Phototoxicity-dependent fission assay ...... 60 2.15 Image analysis ...... 60 2.15.1 Mitochondrial Networks ...... 60 2.15.2 MtDNA Nucleoids ...... 60 2.15.3 Lysosomes ...... 61 2.15.4 Mitochondrial Fusion Analysis ...... 61 2.16 Mitochondrial Membrane Potential and Mitochondrial Mass Analyses ...... 62 2.17 In vitro Cell-Free Mitochondrial Fusion Assay ...... 63 2.17.1 Mitochondrial isolation and cytosol prep ...... 63 2.17.2 Mitochondrial Fusion Assay...... 63 2.18 Transmission electron microscopy ...... 64 Characterization of the C584R variant in the mtDNA depletion syndrome gene FBXL4, reveals a novel role for FBXL4 as a regulator of mitochondrial fusion...... 65 3.1 Introduction ...... 66 3.2 Results ...... 68 3.2.1 Clinical Description ...... 68 3.2.2 Dichloroacetate improved lactic acidosis and cardiac hypertrophy in patient 2 ...... 70 3.2.3 Genetic studies identify homozygous variants in FBXL4 ...... 70 3.2.4 Cells homozygous for the C584R variant in FBXL4 exhibit mtDNA depletion and bioenergetic defects ...... 71 3.2.5 DCA reduces lactate production, but does not correct mitochondrial function ...... 76 3.2.6 Impaired mitochondrial fusion in FBXL4 patient fibroblasts ...... 76 3.2.7 Overexpression of wildtype FBXL4, but not C584R mutant, promotes mitochondrial fusion ...... 79 3.2.8 Restoring mitochondrial morphology in FBXL4 fibroblasts rescues mtDNA depletion 80 3.3 Discussion ...... 83 MSTO1 mutations cause mtDNA depletion, manifesting as muscular dystrophy with cerebellar involvement...... 89

4.1 Introduction ...... 90 4.2 Results ...... 92 4.2.1 Human Subjects and Samples ...... 92 4.2.2 Clinical Characteristics ...... 92 4.2.3 Neuroimaging Characteristics ...... 95 4.2.4 Muscle Histopathology and Electron Microscopy...... 95 4.2.5 Molecular Results ...... 97 4.2.6 Characterization of Patient Fibroblasts ...... 98 4.3 Discussion ...... 106 The R941L mutation in MYH14 disrupts mitochondrial fission and alters the distribution of mtDNA nucleoids...... 113 5.1 Introduction ...... 114 5.2 Results ...... 115 5.2.1 Hyperconnected mitochondrial networks at the cell periphery in R941L patient fibroblasts ...... 115 5.2.2 Hyperfused mitochondria at the cell periphery are resistant to phototoxicity-induced fission ...... 117 5.2.3 Altered mtDNA nucleoids in R941L patient fibroblasts ...... 118 5.3 Discussion ...... 121 DISCUSSION ...... 124 Mitochondrial fusion, fission and the regulation of mtDNA ...... 125 How does restoring fused mitochondrial networks affect mtDNA? ...... 127 mtDNA copy number ...... 127 Enlarged mtDNA nucleoids ...... 129 mtDNA deletions in fusion-defective cells...... 131 MtDNA phenotypes in models of impaired fusion are discrepant – why? ...... 132 Limitations ...... 133 Future directions ...... 135 How does impaired fusion lead to mtDNA loss? ...... 135 Fused Mitochondrial Networks – are they all equal? ...... 136 How does FBXL4 promote mitochondrial fusion? ...... 137 Where is FBXL4 protein in the mitochondria? ...... 139 How does MSTO1 promote mitochondrial fusion? ...... 141 Does MSTO1 interact with MFN1 or MFN2? ...... 142 Does MSTO1 have a role in oxidative stress-induced mitochondrial hyperfusion? ...... 144

CONCLUSIONS ...... 146 BIBLIOGRAPHY ...... 148 Appendix A ...... 160 Appendix B – Copyright Permission (Chapter 1) ...... 163 Appendix C – Copyright Permission (Chapter 3) ...... 164 Appendix D – Copyright Permission (Chapter 4) ...... 165 Appendix E – Copyright Permission (Chapter 5) ...... 166

xii

List of Tables

Table 1.1. Pathogenic mutations in mitochondrial fusion and fission genes highlighting patient phenotypes and perturbations to mitochondrial genomes. This is a non-exhaustive list of pathogenic mutation; however, the aim here was to outline variants with relevance to mtDNA maintenance. CMT2A: Charcot Marie Tooth disease type 2A, DOA: Dominant Optic Atrophy, het: heterozygous, cPEO: chronic progressive external ophthalmoplegia, M: Muscle, F: Fibroblasts...... 42 Table 6.1. Summary of mitochondrial network morphology and mtDNA phenotypes in patient primary dermal fibroblasts harbouring pathogenic mutations in FBXL4, MSTO1 or NMIIC...... 127 Table 6.2. Summary of mitochondrial network morphology and mtDNA phenotypes following various approaches to rescue fused mitochondrial networks in patient primary dermal fibroblasts harbouring pathogenic mutations in FBXL4 or MSTO1...... 128

xiii

List of Figures & Illustrations

Figure 1.1. Representative confocal images of mitochondrial network morphology showing fragmented and hyperfused mitochondrial networks. Cells were fixed and stained with antibodies against the outer membrane protein, TOMM20...... 28 Figure 1.2. Functional Relevance of Mitochondrial Fusion and Fission. A) Constant cycling between fusion and fission events remodels mitochondrial network morphology. B) Fragmentation of mitochondria ensures their inheritance into daughter cells during cell division. C) Fission facilitates transporting mitochondrial puncta along neuronal axons. D) Fusion mediates buffering the effect of damaged mitochondrial proteins or mutant mtDNA and allows for functional complementation. E) Fission mediates segregation of dysfunctional mitochondria and their elimination by mitophagy. F) Fusion and fission of mitochondrial membranes ensures distributions of intra-mitochondrial components including mtDNA, proteins and lipids...... 30 Figure 1.3. Schematic of balanced mitochondrial network remodeling and coordinated mtDNA nucleoid distribution following fusion and fission (left). Impaired fusion regulation results in mitochondrial fragments devoid of mtDNA, enlarged mtDNA nucleoids and mtDNA loss. Meanwhile, impaired fission gives rise to hyperfused networks with enlarged mtDNA nucleoids (right)...... 32 Figure 1.4. Schematic of mitochondrial fusion machinery. A) Mitochondrial outer membrane fusion is mediated by the GTPases MFN1 and MFN2. The non-bilayer forming lipid, phosphatidic acid, promotes membrane curvature and drives membrane fusion. The cytosolic protein, MSTO1, has been proposed to participate in fusion regulation, however the mechanism remains unclear. B) Fusion of the inner mitochondrial membrane is mediated by the GTPase activity of the membrane associated L-OPA1. The soluble S-OPA1 isoform may participate in regulating IMM fusion however this role remains debatable. Notably, the IMM phospholipid, cardiolipin, can interact with OPA1 to promote membrane fusion...... 35 Figure 1.5. Schematic of mitochondrial fission machinery. A) ER-mediated constriction of mitochondria marks fission sites and actin-myosin complexes provide the necessary force for this pinching step. Notably, it has been proposed that mtDNA replication provides the initial signal for the assembly of the fission machinery. B) DRP1 is recruited to fission sites via OMM adaptor proteins. DRP1 oligomerization further constricts mitochondria at fission sites. C) Dynamin 2 may also play a role in constricting and severing mitochondria, though its role is debatable. GTP hydrolysis leads to conformational changes in DRP1 which enhances mitochondrial constriction and ultimately severs mitochondria. D) Fission ensures segregation of newly replicated mtDNA into daughter mitochondria...... 38 Figure 1.6. Representative confocal images of control dermal fibroblasts. Cells were stained with MitoTrackerRed (mitochondria, red) and PicoGreen (nDNA and mtDNA) dyes. Images demonstrate uniform size and even distribution of mtDNA nucleoid throughout reticular and fragmented networks in control cells...... 44 Figure 3.1. Clinical investigation of mitochondrial disease patients. A) Family pedigree of two mitochondrial disease patients (P1 and P2) from a consanguineous marriage, both of whom presented with encephalomyopathy, severe lactic acidosis and hypertrophic cardiomyopathy. Exome sequencing of P2 identified a homozygous recessive mutation in the mitochondrial protein, FBXL4 (c. 1750T>C; p.C584R). B) Increasing doses of dichloroacetate (DCA) administered to P2 around 6 weeks of age remarkably improved lactic acidosis. Dose range: 25- 50mg/kg/day. C) Interventricular septum thickness z-scores for P2 indicate thickening of the of

xiv the left ventricular walls which normalize after DCA administration. Z-scores were measured by the Boston Children’s Hospital z-score calculator. D) Sanger sequencing of exon 9 in FBXL4 confirms the homozygous variants in P1 and P2...... 69 Figure 3.2. Characterization of FBXL4 patient fibroblasts, and response to DCA treatment. A) QPCR analysis of mtDNA copy number from control and patient fibroblasts normalized to 18S. Data represents at least three independent biological replicates. B) Traces of oxygen consumption rates (OCR) over time analyzed in control and FBXL4 patient fibroblasts using the Seahorse XF24 extracellular flux analyzer. C) Traces of OCR of cells as in B following 24hrs treatment with 10 mM DCA. D) Basal respiration in control and patient cells calculated from B and C. Data is presented as % control from 3 independent biological replicates. E) Maximal respiratory capacity in control and patient fibroblasts calculated from B & C. Data is presented as % control from 3 independent biological replicates. F) Extracellular acidification rate (ECAR) was recorded for control and patient samples treated with DCA. Data represents ECAR recordings prior to drug injections. Bioenergetic profiles were assessed from at least three technical replicates per condition. P-values were determined by an unpaired student t-test, compared to the untreated control, unless otherwise indicated...... 72 Figure 3.3. Characterization of OXPHOS subunits in FBXL4 patient fibroblasts, and response to DCA treatment. A) Western analysis of subunits of oxidative phosphorylation complexes in control and FBXL4 patient fibroblasts treated as indicated. MtDNA-encoded COXII of complex IV is marked with an asterisk. All other subunits are nuclear-encoded. B-actin was used as a load control. B-F) Relative quantification of Complex V subunit (ATP5A) (B), Complex III subunit (UQCRC2) (C), Complex II subunit (SDHB) (D), the mtDNA-encoded subunit of Complex IV (COXII) (E) and Complex I subunit (NDUFB8) (F), in control and patient cells treated as indicated. Data represents three independent biological replicates and is presented as %control...... 73 Figure 3.4. A) Mitochondrial membrane potential in fibroblasts from patients and an unaffected control was assessed by flow cytometry following staining with a membrane-potential sensitive dye (TMRE). B) Mitochondrial mass was assessed in control and FBXL4 patient fibroblasts by flow cytometry following MitoTracker Green staining. P-values were determined by an unpaired student t-test, compared to the control...... 73 Figure 3.5. Mitochondrial network and nucleoid alterations in C584R FBXL4 patient fibroblasts. A) Representative confocal images of control and patient cells stained with MitoTracker Red and PicoGreen (dsDNA: nuclear and mtDNA). Arrowheads in zoomed images highlight mitochondrial fragments devoid of mtDNA. Scalebars: 10µm. B) Quantification of mitochondrial morphology from control and FBXL4 patient fibroblasts treated with DCA (10mM, 24hrs) or a vehicle control (DIH2O). Morphology analysis was performed on fixed cells stained with TOMM20 from three technical replicates, quantifying at least 50 cells per fibroblast line per replicate. C) Western Analysis of mitochondrial fusion proteins (MFN1, MFN2, OPA1) and the main fission GTPase (DRP1) in control and patient fibroblasts treated as indicated. HSP60 and β-tubulin were used as mitochondrial and cytosolic load controls, respectively. D) Quantification of mtDNA nucleoid counts in control and patient fibroblasts from cells stained as in A. E) Quantification of mtDNA nucleoid sizes in cells stained as in A. Data represents average nucleoid sizes and counts from 10 cells for each group. P-values were determined by an unpaired student t-test, compared to the control fibroblasts...... 75 Figure 3.6. Mitochondrial fusion is reduced in FBXL4 patient fibroblasts. A) Representative confocal images of live control and patient cells expressing a matrix-targeted PA-GFP and stained with TMRE. White boxes indicated region of photoactivation. Images were

xv acquired pre-photoactivation, 5s after photoactivation and then every 120s, up to 960s (16mins). Scalebars: 10µm. B) Photoactivated region of cells shown in A, zoomed in, and showing initial photoactivate frame (t=5s) and final frame at 960s. C) Quantification of area of photoactivated mitochondrial network at t=5s and the percentage of the network to which the GFP signal had spread at t=960s. Analysis was performed on binarized images, shown in Fig 3.7. Data represent analyses in 5 cells per fibroblast line. P-values were determined by an unpaired student t-test, compared to the control. D) Relative fluorescence intensity over time in the photoactivated region indicated in panel A, at t=0 (pre-photoactivation), t=5s (post- photoactivation) and 120s intervals. Data points represents relative fluorescence intensity in the representative images shown in panel A...... 78 Figure 3.7. Example of binarized confocal images of control and FBXL4 patient cells expressing a matrix-targeted PA-GFP and stained with TMRE as in Figure 3.6. Images shown were at t=5s post-photoactivation (A) and t=960s at the final acquisition frame (B). Images are representative of how PA-GFP signal diffusion was analyzed. The percentage of the network that was photoactivated at t=5s relative to the total network (TMRE signal) as well as the % network containing GFP signal at t=960s relative to the total network in TMRE channel were measured using ImageJ FIJI. Percentages displayed on respective images indicate measurements for these frames...... 79 Figure 3.8. FBXL4 overexpression shifts mitochondrial networks into a fused state. A) Representative confocal images of HEK cells transfected with an empty vector, wildtype FBXL4- HA, or C584R-HA, immunolabelled with antibodies against TOMM20 (red) and the epitope tag, HA (green). Scalebars: 10µm. B) Quantification of mitochondrial morphology from cells as in panel A. P-values were determined by an unpaired student t-test, compared to the percentage of cells with a fused morphology in the indicated lines...... 80 Figure 3.9. Restoration of fused mitochondrial networks in FBXL4 patient fibroblasts rescues mtDNA copy number. A) Representative confocal images of control and patient fibroblasts treated with mdivi-1 or DMSO for 7 days, fixed and immunolabeled with antibodies against TOMM20 (red) and DNA (green). Scalebars: 10µm. B) Mitochondrial morphology analysis of control and patient cells treated with mdivi-1 or a vehicle control for 7 days. Data represents two independent biological replicates. C) Relative mtDNA copy number in control and patient cells treated with mdivi-1 or a vehicle control for 7 days. D) Average mtDNA nucleoid counts from cells treated as indicated, fixed and immunolabeled with an anti-DNA antibody. E) Average mtDNA nucleoid sizes from cells as in D. P-values were determined by an unpaired student t-test, comparing each group to the vehicle control-treated control fibroblasts...... 82 Figure 3.10. Mdivi-1 treatment (25uM, 24hrs) in FBXL4 patient fibroblasts does not rescue mtDNA copy number. A) Relative mtDNA copy number in control and patient cells treated with mdivi-1 or a vehicle control for 24hrs. P-values were determined by an unpaired student t-test, comparing each group to the vehicle control-treated control fibroblasts...... 83 Figure 3.11. Proposed model for mtDNA depletion and altered nucleoids distribution in cells with impaired mitochondrial dynamics regulation. Under normal conditions, balanced fusion and fission events ensure even distribution of mtDNA nucleoids throughout reticular mitochondrial networks or fragmented puncta. Notably, mitochondrial fission at sites of mtDNA replication facilitates segregation of newly synthesized mtDNA, thus maintaining mtDNA copies throughout the network. An imbalance in mitochondrial dynamics arising from impaired fusion (e.g. in mutant FBXL4, OPA1 or MFN2) leads to excessive fragmentation of the mitochondrial network, clustering of mtDNA nucleoids and mtDNA depletion. Meanwhile, cells with defective

xvi fission (e.g. mutant DNM1L), mtDNA nucleoids form large aggregates likely due to lack of efficient segregation following replication. Impaired fission also leads to mtDNA depletion, however the exact mechanisms are poorly understood...... 88 Figure 4.1. Muscle and brain imaging in MSTO1 patients. (a) Lower extremity muscle MRI of patients P6 (p.(Asp236His); p.(Phe217Leu)), P7 (p.(Leu450Phe); deletion), P8 (p.(Phe217Leu); p.(Asp236His)) and P10 (p.(Asp236Gly); p.(Arg279His)) at ages 37 years, 9 years, 6 years and 16 years, respectively. Abnormal signaling of muscles such as the posterior gastrocnemius muscle in patient P6 (white arrow), reflects muscle breakdown with replacement with adipose tissue (b) Brain MRI completed in twelve patients consistently demonstrates moderate-to-severe cerebellar volume loss involving the vermis and both hemispheres. Repeat MRI images available in patients P4 (p.(Gly420ValfsX2); p.(Arg256Glu)), P8 (p.(Phe217Leu); p.(Asp236His)), P10 (p.(Asp236Gly); p.(Arg279His)) and P12 (p.(Arg345His); p.(Thr324Ile)) demonstrate mild (P12: p.(Arg345His); p.(Thr324Ile)) to no progression (P4: p.(Gly420ValfsX2); p.(Arg256Glu), P8: p.(Phe217Leu); p.(Asp236His) and P10: p.(Asp236Gly); p.(Arg279His)) of cerebellar volume loss over time...... 94 Figure 4.2. Muscle biopsy, MSTO1 pathogenic variants and pedigrees. (a) Histology findings from the vastus lateralis muscle biopsy of P7 (p.(Leu450Phe); deletion) at age 20 months include internalized nuclei on hematoxylin and eosin (H&E) staining (white arrow) (i) and variation in fiber size on nicotinamide dinucleotide (NADH) staining (ii) and whorled fibers evident on Gömöri trichrome (inset) (iii) and COX staining (white arrow) (iv). (b) Muscle biopsy electron microscopy (EM) findings are notable for aggregates of subsarcolemmal mitochondria in both P9 (p.(Tyr478Cys); missing)) (i and ii) and P10 (p.(Asp236Gly); p.(Arg279His)) (iii and iv) and non-specific mitochondrial morphologic abnormalities (variations in mitochondrial shape and size) (black boxes) in P10. (c) Schematic of new and reported human MSTO1 pathogenic variants. Shown in numbered light blue squares are cDNA exons (RefSeq isoform NM_018116.3 of MSTO1). Corresponding known protein domains are shown in orange (tubulin 3 domain) and beige (Misato segment II tublin-like domain). Variants written in black text are recessive; the single mutation in red has been previously reported to cause dominantly inherited MSTO1-related disease. The top half of the figure depicts novel variants reported in this publication; the bottom half of the figure depicts variants which have been previously reported. Bolded variants depict previously reported mutations that were also present in our cohort. The dotted line depicts a large deletion (exons 9-14). (d) Pedigree of two families consistent with recessive inheritance of MSTO1 pathogenic variants...... 96 Figure 4.3. Pathogenic variants lead to MSTO1 protein instability. (a) Western blot analysis of total cell lysates from control and patient fibroblast. As a control, total cell lysates from HeLa cells overexpressing MSTO1-V5 or empty vector were also included. Blots were probed with antibodies against endogenous MSTO1, VDAC1, HSP60 and V5. Black arrow corresponds with endogenous MSTO1 protein further verified in HeLa cell lysates, meanwhile bands underneath are nonspecific. (b) Western blot analysis of cell lysates as in (a). Blots were probed against fusion proteins (Mfn1, Mfn2 and Opa1) and loading controls...... 99 Figure 4.4. Characteristics of MSTO1 patient fibroblasts. (a) Representative confocal microscopy images of control and patient cells. Mitochondrial networks in MSTO1 patient cells are more fragmented and contain fewer but larger mtDNA nucleoids compared to the control cells. Live cells were stained with MitoTracker Red (Red, mitochondria) and PicoGreen (Green, nuclear and mitochondrial DNA). (b) Quantification of mitochondrial morphology from control and patient cells performed from three independent replicates. Statistical analysis was

xvii performed on the number of cells with partly fragmented mitochondrial morphology in control versus patient cells; Student T-test, *p < 0.05, **p < 0.001...... 100 Figure 4.5. Enlarged lysosomal vacuoles in MSTO1 patient fibroblasts. (a) Representative confocal images of control and patient cells fixed and stained with antibodies against TOMM20 (Red, mitochondria) and LAMP1 (Green, lysosomes). Compared to an unaffected control, patient cells contain distinct lysosomal clusters. (b) Quantification of cells containing enlarged lysosomes in control and patient fibroblasts performed from two independent replicates. Statistical analysis was performed; Student T-test, *p < 0.05...... 101 Figure 4.6. Pathogenic variants in MSTO1 are linked to mtDNA depletion. (a) Relative mtDNA copy number normalized to the nuclear-encoded 18S gene. Data represents at least three independent biological replicates. (b) Analysis of mtDNA nucleoid counts per cell from 35 cells for each group). (c) Quantification of nucleoid sizes in control and patient cells. Data represents average nucleoid sizes from the same cells as in (b). Average mtDNA nucleoid size is presented in a violin plot. K-S test was performed to determine statistical significance. (d) Frequency of nucleoids larger than 0.2µm2 in all 35 cells quantified per fibroblast line. Student T-test was performed as indicated for (a), (c) and (d). *p < 0.05, **p < 0.01, *** p < 0.0001. ... 103 Figure 4.7. Expression of wild-type MSTO1 rescues cellular phenotypes in MSTO1 patient fibroblasts. Control, P4 and P7 fibroblast cells were transfected with MSTO1-P2A-mCherry or the mCherry empty vector control. Representative images of fibroblasts transfected with MSTO1-P2A-mCherry, for (a) live cells stained with picogreen and MitoTracker Deep Red, or (b) fixed cells stained with antibodies against TOMM20 (Green, mitochondria) and LAMP1 (Blue, lysosomes). Scalebars: 10µm. Transfected fibroblasts, expressing cytosolic mCherry or MSTO1-P2A-mCherry, were characterized as described above for the following cellular phenotypes: (c) mitochondrial morphology, (d) lysosome morphology, (e) average mtDNA nucleoid size (f) mtDNA nucleoid counts, and (g) relative mtDNA copy number. Student T-test was performed as indicated for (c), (d), (f), and (g). K-S test was performed to determine statistical significance for (e). *p < 0.05, **p < 0.01, *** p < 0.0001...... 105 Figure 5.1. Hyperconnected mitochondrial networks at the cell periphery in R941L patient fibroblasts. A) Representative confocal images of control and patient fibroblast cells taken with an Olympus SD-OSR microscope. Live cells were stained with MitoTracker Red (Red, mitochondria) and PicoGreen (Green, nuclear and mitochondrial DNA). Scale bars indicate 10 μm. B) Quantification of control and patient cells containing hyperconnected mitochondrial networks at the cell periphery. At least 70 cells were quantified from two independent replicates. Error bars indicate standard deviations, and p-values (Student's t-test) were determined by comparison to the number of control cells with hyperfused mitochondria...... 116 Figure 5.2. Hyperconnected mitochondrial networks at the cell periphery in R941L patient fibroblasts are resistant to phototoxicity-induced fission. A) Representative confocal images of mitochondrial networks taken with an Olympus SD-OSR microscope. Control and patient cells stained with MitoTracker Red were imaged continuously over 5 min with high laser power to induce fission. Zoomed boxes represent regions with fragmented mitochondria (green hashed boxes) or resistant to fission (magenta hashed boxes) when imaging commenced, and at the end of 5 min. Signal intensity was enhanced for later frames to adjust for photobleaching. Scale bars indicate 10 μm...... 118 Figure 5.3. Altered mtDNA nucleoids in R941L patient fibroblasts. A) Representative confocal images of control and patient fibroblast cells taken with an Olympus SD-OSR microscope. Live cells were stained with MitoTracker Red (Red, mitochondria) and PicoGreen (Green, nuclear and mitochondrial DNA). Scale bars indicate 10 μm. B) Quantification of

xviii nucleoid size and C) number from 10 cells for each line. Error bars indicate standard deviations, and p-values (Student's t-test) were determined by comparison to control fibroblasts...... 120 Figure 5.4. Analysis of mtDNA copy number and mtDNA deletions in R941L patient fibroblasts. A) Copy number of mtDNA as determined by quantitative PCR. Error bars indicate standard deviations, and p-values (Student's t-test) were determined by comparison to control fibroblasts. B) Long range PCR of mtDNA in control and patient fibroblasts showing 16.3 kb amplicons and no mtDNA deletions...... 121 Figure 6.1. Proposed model for the contribution of fusion and fission to mtDNA nucleoid distribution and the formation of mtDNA clumps. A) Under steady state conditions, there are continuous, balanced fusion and fission events, and even distribution of mtDNA nucleoids. B) In cells with a fission defect, mitochondrial fission events are reduced, which is coupled with a reduction in the number of fusion events. As such, reduced fission impairs segregation of nucleoids giving rise to clumps while the reduction in fusion can impair mtDNA replication affecting total copy number. C) Cells with impaired fusion also exhibit a reduced number of fission events. Defective fusion may contribute to reduced synthesis (mtDNA depletion) while reduced fission impaired segregation of nucleoids. RESCUE EXPERIMENTS D) Restoring fused mitochondrial networks by inhibiting fragmentation results in a net increase in the number of fusion events relative to fission events, restoring network morphology and mtDNA copy number. However, the lack of fission events impaired nucleoid segregation. E) Restoring fused mitochondrial networks by transiently promoting fusion, rescues network morphology and because fission is active, mtDNA nucleoid clumps are resolved. KO: knockout, LOF: loss of function...... 130 Figure 6.2. The autophagy inhibitor chloroquine does not rescue mtDNA copy number in MFN1/2 double knockout cells or in FBXL4 patient fibroblasts. A) QPCR analysis of relative mtDNA copy number in control and MFN1/2 double knockout mouse embryonic fibroblasts. Cells were treated with 25uM chloroquine (CQ) for 24hrs. B) QPCR analysis of mtDNA copy number from control and patient fibroblasts under the same conditions as panel A...... 136 Figure 6.3. Targeting a constitutively active deubiquitilase to the mitochondrial intermembrane space disrupts mitochondrial networks and may phenocopy FBXL4 mutations. A) Representative confocal images of HEK cells transfected with an empty vector or IMS-targeted, GFP-tagged catalytic domain of USP8. Scalebars represent 5um. B) Quantification of mitochondrial morphology from cells as in panel A. Statistical analysis: no. of cells with fused mitochondrial morphology in control vs. IMS-USP8-overexpressing cells or in control vs. IMS-C784A overexpressing cells; Student T.test, *p<0.05. C) Western analyses of total cell lysates from cells overexpressing the USP8-GFP constructs or an empty vector. Blots were probed with antibodies against GFP or β-actin...... 138 Figure 6.4. Distinctive structural abnormalities of mitochondria and the endoplasmic reticulum in FBXL4 patient fibroblasts. A) Electron micrographs of control and FBXL4 patient cells showing mitochondria (top) and endoplasmic reticulum structures (bottom). Structural abnormalities in mitochondria are notable including swelling and abnormal cristae. As well, ER appears swollen in FBXL4 patient cells. B) Confocal images of control and FBXL4 patient fibroblasts stained with antibodies against TOMM20 (magenta) and Calnexin (green) to label mitochondrial and the endoplasmic reticulum, respectively. Abnormal ER networks in patient cells compared to control are notable. ER Tubules are indicated with an asterisk, and ER sheets with an arrowhead...... 140 Figure 6.5. MSTO1 protein promotes mitochondrial fusion. A) Overexpression of MSTO1 in human embryonic kidney (HEK) cells gives rise to hyperfused mitochondrial networks. HEK

xix cells were transfected with an empty vector control or MSTO1-V5 for 24hrs. Mitochondrial morphology was scored from at least 50 cells, from three independent replicates. B) Western Analyses from HEK cells overexpressing MSTO1-V5 or an empty vector control probed with antibodies against MSTO1, V5 and β-actin. Arrow indicates MSTO1 band. C) Cytosolic MSTO1 protein directly promotes mitochondrial fusion in a cell-free fusion assay. Isolated mitochondria were incubated with cytosol from control cells or cells overexpressing MSTO1 protein and fusion activity measured based on complementation of a split-luciferase protein and luciferase signal. Negative controls N-Luci, C-Luci refer to mitochondria containing only the N- or C-terminal end of the luciferase protein and Neg refers to fusion reactions with no mitochondria. P-values are based on Student’s t.tests...... 142 Figure 6.6. MSTO1 protein promotes fusion via interactions with MFN1 and MFN2. A) Model depicting mitofusin interactions to promote fusion and a putative role for MSTO1 in promoting homo- or heterotypic MFN interactions. B) Western analyses of total cell lysates from control, MFN1 or MFN2 knockout mouse embryonic fibroblasts overexpressing an empty vector or MSTO1-V5. Blots were probed with antibodies against endogenous MSTO1, V5, MFN1, MFN2 and β-actin. C) Mitochondrial morphology analysis from cells transfected as in B. At least 50 cells were scored for mitochondrial network morphology from three independent replicates. P-values are based on Student’s t.tests...... 143 Figure 6.7. MSTO1 may be a redox regulated protein and Cys222 is important for its pro- fusion activity. A) Western analyses of mitochondria enriched fractions from HEK cells overexpressing an empty vector or MSTO1-V5. Samples were prepared under non-reducing conditions (without β-mercaptoethanol) and blots were probed with antibodies against MSTO1 and VDAC1. Note MSTO1 bands appear at a significantly higher molecular weight than monomeric protein (monomeric MSTO1: ~62kDa). B) Sequence alignment of MSTO1 homologs in Homo sapiens NP_060586.2, Mus musculus NP_659147.2, Drosophila melanogaster NP_523435.1, Saccharomyces cerevisiae NP_013938.1, and Arabidopsis thaliana NP_195436.1. Protein sequences were aligned using COBALT [255]. A highly conserved cysteine residue is highlighted in red. C) Western analyses of total cell lysates from HEK cells expressing wildtype MSTO1-V5, C222A-V5 MSTO1 mutant or empty vector control. Blots were probes against endogenous MSTO1, V5 and β-actin. D) Mitochondrial morphology analysis from cells transfected as in C. At least 50 cells were scored for mitochondrial network morphology from three independent replicates. P-values are based on Student’s t.tests...... 145

List of Symbols, Abbreviations, & Nomenclature

Symbol/Abbreviation Definition 18S 18S rRNA gene ABAT 4-Aminobutyrate Aminotransferase ADP Adenosine diphosphate AGK Acylglycerol Kinase ARL2 ADP-ribosylation factor (ARF) like 2 ATP Adenosine triphosphate ATP5A ATP synthase F1 subunit alpha BAK BCL2 Antagonist/Killer 1 BAX BCL2 Associated X Protein BCA Bicinchoninic Acid assay cDNA Coding Deoxyribonucleic acid CL Cardiolipin C-Luci Carboxy-terminal split luciferase CMT2A Charcot Marie Tooth type 2A COX Cytochrome C Oxidase COXI Cytochrome c oxidase subunit I COXII Cytochrome c oxidase subunit II CQ Chloroquine CytB Cytochrome B DCA Dichloroacetate DGUOK Deoxyguanosine Kinase DMEM Dulbecco's Modified Eagle Medium DNA Deoxyribonucleic acid DNM1L Dynamin 1 Like dNTP Deoxynucleoside triphosphates DOA Dominant Optic Atrophy DRP1 dynamic related protein 1 DSB Double Strand Break dsDNA Double stranded DNA DYN2 Dynamin 2 ECAR Extracellular acidification rate ELMOD2 ELMO Domain Containing 2 ER Endoplasmic Reticulum FBS Fetal Bovine Serum FBXL4 F-Box and Leucine Rich Repeat Protein 4

xxi

FCCP Carbonyl cyanide-4-(trifluoromethoxy)phenylhydrazone FIS1 Fission, Mitochondrial 1 Fzo1 Fuzzy Onions 1 GFP Green Fluorescent Protein Gp78 Glycoprotein 78 GTP Guanosine-5'-triphosphate H&E Hematoxylin and eosin stain HA Hemagglutinin HEK Human Embryonic Kidney HeLa Henrietta Lacks (immortal cervical cancer cell line) HEPES 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid HSP60 Heat Shock Protein 60 HECT, UBA and WWE Domain Containing E3 Ubiquitin HUWE1 Protein Ligase 1 IMM Inner Mitochondrial Membrane IMS Inter Membrane Space INF2 Inverted formin-2 kb kilobase KD Knockdown KO Knockout LAMP1 Lysosomal-associated membrane protein 1 LOF Loss of Function L-OPA1 Long isoform of Optic Atrophy 1 LR PCR Long Range Polymerase Chain Reaction LSM Line Scanning Microscopy MAPL Mitochondrial-anchored protein ligase MARCHV Membrane-associated RING finger protein 5 mCherry Monomeric red fluorescent protein Mdivi-1 Mitochondrial division inhibitor 1 MEF Mouse Embryonic Fibroblasts MEM Minimum Essential Media MFF Mitochondrial fission factor MFN1 Mitofusin 1 MFN2 Mitofusin 2 Mgm1 Mitochondrial genome maintenance 1 MGME1 Mitochondrial Genome Maintenance Exonuclease 1 MGRN1 Mahogunin Ring Finger 1 MIB Mitochondrial Isolation Buffer MID49 Mitochondrial Dynamic Protein Of 49 kDa

xxii

MID51 Mitochondrial Dynamic Protein Of 49 kDa MIEF2 Mitochondrial Elongation Factor 2 MIGA1/2 Mitoguardin 1/2 MitoPLD Mitochondrial Phospholipase MPV17 Mitochondrial Inner Membrane Protein MPV17 Misato Mitochondrial Distribution And Morphology MSTO1 Regulator 1 MTCH2 Mitochondrial Carrier 2 mtDNA mitochondrial DNA MTDPS MtDNA depletion syndrome MTO1 Mitochondrial TRNA Translation Optimization 1 MTP18 Mitochondrial Protein 18 KDa MTR MitoTrackerRed mtSSBP Mitochondrial single stranded DNA-binding protein MYH10 Myosin Heavy Chain 10 MYH14 Myosin Heavy Chain 14 MYH9 Myosin Heavy Chain 9 NADH Nicotinamide adenine dinucleotide nDNA nuclear DNA NDUFB8 NADH:Ubiquinone Oxidoreductase Subunit B8 Neg Negative N-Luci Amino-terminal split luciferase NMIIA Non muscle myosin II A NMIIB Non muscle myosin II B NMIIC Non muscle myosin II C OCR Oxygen Consumption Rate OMA1 Overlapping with the M-AAA Protease 1 Homolog OMM Outer Mitochondrial Membrane OXPHOS Oxidative Phosphorylation P2A Self-cleaving peptide PA Phosphatidic acid PA-GFP Photo-activatable green fluorescent protein Parkin Parkinson Protein 2 E3 Ubiquitin Protein Ligase PBS Phosphate Buffered Saline PCR Polymerase Chain Reaction PINK1 PTEN-induced kinase 1 POLγ Polymerase gamma PTM Post translational modification qPCR quantitative Polymerase Chain Reaction

xxiii

RIPA Radioimmunoprecipitation assay buffer RNA Ribonucleic acid RNAi RNA interference ROS Reactive Oxygen Species Ribonucleotide Reductase Regulatory TP53 Inducible RRM2B Subunit M2B rRNA Ribosomal RNA SD Standard deviation SD OSR Spinning Disc Olympus Super Resolution Succinate Dehydrogenase Complex Iron Sulfur Subunit SDHB B Sodium Dodecyl Sulfate Poly-Acrylamide Gel SDS-PAGE Electrophoresis SIMH Stress-induced mitochondrial hyperfusion Solute Carrier Family 25 Member 4 (ADP/ATP SLC25A4 translocase 1) SLC25A46 Solute Carrier Family 25 Member 46 SLC25A50 Solute Carrier Family 25 Member 50 S-OPA1 Short isoform of Optic Atrophy 1 SUCLA2 Succinate-CoA Ligase ADP-Forming Subunit Beta SUCLG1 Succinate-CoA Ligase GDP/ADP-Forming Subunit Alpha TEM Transmission electron microscopy TFAM Transcription Factor A, Mitochondrial TFB2M Transcription Initiation Factor IIB TK2 Thymidine Kinase 2 TMRE Tetramethylrhodamine, ethyl ester TOMM20 Translocase of the Outer Mitochondrial Membrane tRNA Transfer RNA TWNK TWINKLE TYMP Thymidine Phosphorylase Ub ubiquitin UQCRC2 Ubiquinol-Cytochrome C Reductase Core Protein 2 USP8 Ubiquitin Specific Peptidase 8 V5 Short peptide derived from simian virus 5 VD Variable Domain VDAC1 Voltage Dependent Anion Channel 1 YME1L Yeast mitochondrial escape like 1 ATPase

xxiv

Nomenclature used for various genes/proteins (MFN2 is used as an example).

Organism Gene symbol Protein symbol Mouse Mfn2 Mfn2 Human MFN2 MFN2

xxv

INTRODUCTION

Some sections of this chapter are from the following literature review (Sabouny R and & Shutt TE. Mitochondrial Fission and Fusion are Reciprocally Regulated. Trends in Biochemical Sciences [1], with permission from Sabouny & Shutt, 2020 (see Appendix B).

1. Mitochondria

Mitochondria are functionally and structurally fascinating organelles. While oxidative phosphorylation (OXPHOS) and ATP production are mitochondria’s widely known functions, these organelles play a central role in maintaining metabolic homeostasis, lipid synthesis, redox balance, cell survival and innate immune responses [2]. As such, proper mitochondrial function is critical for virtually every cell type and tissue in the body. Nonetheless, it is evident from numerous mitochondrial diseases that mitochondrial dysfunction manifests most severely in high energy demanding tissues including neurons, skeletal and cardiac muscles [2, 3].

Notably, an important regulator of mitochondrial function is their dynamic nature [4].

Contrary to the bean-like structure depicted in most textbooks, mitochondria form intricate dynamic networks, the morphology of which can range between small fragments, elongated tubules or interconnected webs [5] (Fig 1.1). Ongoing movement, fusion and fission events dictate these architectural changes and collectively constitute what is known as mitochondrial dynamics [2]. Notably, fusion and fission events are important for maintaining mitochondrial constituents, including the 100-1000 copies of mitochondrial genomes (mtDNA) found in each cell [6]. As discussed in further detail below, mitochondrial network remodeling is instrumental for the regulation of mtDNA copy number and distribution throughout the network.

This thesis will explore consequences of impaired fusion or fission regulation in three mitochondrial disease cases presenting with a spectrum of neurological and muscular defects.

Through characterizing molecular dysfunction of mitochondria in patient cells, we begin to address important outstanding questions pertaining to the interplay between mitochondrial fusion, fission and mtDNA regulation.

1.1 Mitochondrial Dynamics

Mitochondrial dynamics describe ongoing changes in mitochondrial network shape, size, connectivity, trafficking and activity [3]. Mitochondrial network morphology, which can range from an interconnected reticulum to fragmented puncta, is determined by the balance between the opposing forces of fusion and fission [5]. As such, upregulated fusion or downregulated fission can both give rise to hyperfused mitochondrial networks. Conversely, fragmented networks can arise due a direct increase in fission activity or downregulated fusion. Rather than being phenomenological, remodeling mitochondrial networks influences mitochondrial functions in many ways. These include organelle transport and inheritance during cell division, buffering damage and quality control, and the maintenance of intra-mitochondrial components (e.g. mtDNA).

1.2 Why do mitochondria fuse and divide?

Mitochondrial trafficking and inheritance

Mitochondrial networks undergo dramatic remodeling throughout the cell cycle. This is observed in the G1-S phase transition during which mitochondria become tubular networks [7], followed by marked fragmentation during the G2-M phase transition [8, 9]. Given that mitochondria are not synthesized de novo, network fragmentation is believed to allow for equal partitioning of mitochondria to daughter cells after cell division (Fig 1.2). Furthermore, ongoing fusion and fission are also important in post-mitotic cells in order to mediate organelle transport

[3]. Mitochondrial fission facilitates organelle trafficking along the cellular cytoskeleton, and this is particularly important in polarized cells such as neurons which can have very long axons (Fig

1.2). It has been shown that fragmentation of mitochondria is required for their transport along neuronal axons and is important for proper synapse development and neuronal function [10,

11].

Figure 1.2. Functional Relevance of Mitochondrial Fusion and Fission. A) Constant cycling between fusion and fission events remodels mitochondrial network morphology. B) Fragmentation of mitochondria ensures their inheritance into daughter cells during cell division. C) Fission facilitates transporting mitochondrial puncta along neuronal axons. D) Fusion mediates buffering the effect of damaged mitochondrial proteins or mutant mtDNA and allows for functional complementation. E) Fission mediates segregation of dysfunctional mitochondria and their elimination by mitophagy. F) Fusion and fission of mitochondrial membranes ensures distributions of intra-mitochondrial components including mtDNA, proteins and lipids.

Content Mixing

An important feature afforded by fusion and fission dynamics is mitochondrial content exchange. Fusion mediates homogenization of mitochondrial contents, which is especially important if organelles accumulate damaged proteins or mtDNA mutations [12, 13]. Organelle fusion allows for functional complementation and helps buffer the effects of impaired proteins

(Fig 1.2). Moreover, large hyperfused mitochondrial networks provide a protective advantage to mitochondria as they are excluded from the autophagosome [14, 15].

Quality Control (Mitophagy)

If mitochondria accumulate high levels of irreparable damage, mitochondrial fission is activated to segregate dysfunctional organelles [16] (Fig 1.2). Damaged mitochondrial fragments are then targeted for degradation by the autophagosome in a process know as mitophagy (or mitochondrial autophagy) [17]. Notably, dysfunctional organelles can be identified based on mitochondrial membrane depolarization [16], which can arise from mutations in mtDNA or accumulation of dysfunctional mitochondrial proteins. Damaged organelles are specifically marked with poly-ubiquitin chains, a signal that mediates their capture by autophagosomes and subsequent elimination by the autophagosome-lysosome complex [18].

Fragmentation of damaged mitochondria is also relevant from a structural standpoint, as smaller fragments are engulfed more efficiently by the autophagosome [16]. Thus, mitochondrial dynamics also participate in mitochondrial quality control.

Maintenance of Mitochondrial DNA

Mitochondrial fusion and fission both play an important role in the maintenance of mitochondrial genomes. Mitochondrial DNA (mtDNA) is present in hundreds of copies per cell, each copy packaged into a nucleo-protein structure known as an mtDNA nucleoid [19]. Fusion

31 and fission participate in the maintenance of mtDNA copy number, organization of mtDNA into nucleoids and proper distribution of mtDNA throughout mitochondrial networks [3]. In the absence of mitochondrial fusion, cells display depletion in the amount of mtDNA [20-22].

Furthermore, loss of fusion also gives rise to altered mtDNA nucleoids distribution such that some mitochondrial fragments lack mtDNA while others have enlarged nucleoid aggregates (i.e. multiple copies of the genome in close proximity) [20, 22] (Fig 1.3). However, it is not entirely clear how or why fusion defects affect mtDNA. On the other hand, mitochondrial fission has been shown to be important for the segregation of mtDNA nucleoids following replication [23,

24]. In fission defective cells, mitochondrial networks hyperfuse and contain distinctly large mtDNA nucleoid aggregates due to a distribution defect [23, 24] (Fig 1.3).

Figure 1.3. Schematic of balanced mitochondrial network remodeling and coordinated mtDNA nucleoid distribution following fusion and fission (left). Impaired fusion regulation results in mitochondrial fragments devoid of mtDNA, enlarged mtDNA nucleoids and mtDNA loss. Meanwhile, impaired fission gives rise to hyperfused networks with enlarged mtDNA nucleoids (right).

Modulating Mitochondrial Function

The dynamic nature of mitochondria allows for the integration of their activity with key physiological functions such that mitochondria are responsive to changes in the cellular milieu.

For instance, acute oxidative stress and nutrient starvation are examples of cellular cues that have been shown to promote a stress-induced mitochondrial hyperfusion (SIMH) response [25-

27]. In these contexts, hyperfused mitochondrial networks offer cytoprotection and resistance to apoptosis while homeostasis is reinstated. Conversely, nutrient excess is a stimulus that leads to fragmented mitochondrial networks and reprograms mitochondrial function [28].

At the same time, mitochondrial network morphology can directly influence the bioenergetic output of mitochondria. Generally, fused networks have been associated with a higher efficiency for ATP production [27] and lower reactive oxygen species (ROS) [16], at least in part via remodeling inner mitochondrial membrane (IMM) cristae [29]. Meanwhile, several mitochondrial diseases, toxins and stress signals lead to fragmentation of the network and compromise OXPHOS activity [2, 3]. However, the underlying causal link between mitochondrial network shape and bioenergetic activity remains an open area for investigation.

Given that remodeling mitochondrial network morphology is intertwined with mitochondrial and cellular homeostasis, it is not surprising that a complex array of molecular regulators has been shown to govern these dynamic changes in mitochondrial networks. The following sections highlight what we know about core fusion and fission regulators as well as recent reports on accessory factors that participate in remodeling mitochondrial networks.

1.3 Mitochondrial Fusion Regulation

1.3.1 Core Fusion GTPases

Mitochondrial fusion is a multi-step process that begins with juxtaposition and tethering of adjacent mitochondria, followed sequentially by outer mitochondrial membrane (OMM) and inner mitochondrial membrane (IMM) fusion events [30] (Fig 1.4). Mechanistically, tethering and

OMM fusion are regulated by Mitofusin 1 and 2 (MFN1/2), homologous proteins that contain a

GTPase domain, two heptad repeats (coiled-coil domains) and a transmembrane domain [31,

32]. MFN proteins form homo- and hetero-oligomers in cis (i.e. between MFN proteins on the same mitochondria) to prime fusion events, as well as in trans (i.e. between adjacent mitochondria) in order to tether and increase surface contact between adjacent mitochondria

[32, 33]. Subsequent GTP hydrolysis drives conformational changes in MFN oligomers that mediate outer membrane fusion. Notably, while MFN1 and MFN2 have overlapping functions as fusion regulators, MFN2 has additional roles such as mediating mitochondria-endoplasmic reticulum (ER) contact sites [34], and lipid transfer to mitochondria [35].

Fusion of the IMM is regulated by the GTPase Optic Atrophy 1 (OPA1) [36]. Alternative splicing gives rise to long membrane bound OPA1 isoforms (L-OPA1), which can be further processed by the IMM proteases YME1L and OMA1 into short soluble (S-OPA1) isoforms [37,

38]. Maintaining a balance between L-OPA1 and S-OPA1 isoforms is critical for mitochondrial network morphology, as L-OPA1 is a pro-fusion isoform while S-OPA1 has been proposed to play a role in fission [38]. Like MFN proteins, OPA1 also forms oligomeric structures that undergo GTP hydrolysis-driven conformational changes to drive IMM fusion. In addition to regulating dynamic changes in mitochondrial networks, OPA1 protein has been shown to participate in maintaining cristae junctions [39]. Additionally, specific OPA1 isoforms have been proposed to bind with mtDNA nucleoids [40].

Figure 1.4. Schematic of mitochondrial fusion machinery. A) Mitochondrial outer membrane fusion is mediated by the GTPases MFN1 and MFN2. The non-bilayer forming lipid, phosphatidic acid, promotes membrane curvature and drives membrane fusion. The cytosolic protein, MSTO1, has been proposed to participate in fusion regulation, however the mechanism remains unclear. B) Fusion of the inner mitochondrial membrane is mediated by the GTPase activity of the membrane associated L-OPA1. The soluble S-OPA1 isoform may participate in regulating IMM fusion however this role remains debatable. Notably, the IMM phospholipid, cardiolipin, can interact with OPA1 to promote membrane fusion.

Notably, several post translational modifications (PTMs) regulate the activity and stability of fusion GTPases. Many of the reported PTMs have inhibitory effects on fusion, including acetylation of MFN1, MFN2 and OPA1 [41-44], MFN1 phosphorylation [45] and OPA1

OGlcNAcylation [46]. Moreover, ubiquitination can negatively regulate mitochondrial fusion by mediating the proteasomal degradation of MFN1 and MFN2 through the activity of several E3 ubiquitin ligases (e.g. MARCHV [47], Parkin [48, 49], MAPL (MUL1) [50], HUWE1 [51] and

Gp78 [52, 53]). Conversely, MFN1 and MFN2 ubiquitin modifications can also promote fusion

[54]. Notably, pro-fusion ubiquitination of MFN1 by the E3 ubiquitin ligase Mahogunin Ring

Finger-1 (MGRN1) was recently shown to stabilize MFN1 oligomers and drive fusion [55].

However, proteins mediating pro-fusion ubiquitination of MFN2 have yet to be identified. One potential candidate is the F-box containing protein FBXL4 [56, 57]. However, FBXL4 is thought to be localized primarily to the IMS [57], where very little is known about active ubiquitination events.

1.3.2 Auxiliary Fusion Regulators

There is mounting evidence for cytosolic and mitochondrial accessory proteins that participate in regulating fusion. For example, the MFN-binding protein (MIB) is a cytosolic protein that has been implicated as a negative regulator of MFN1 [58]. Meanwhile, BAX and

BAK (members of the pro-apoptotic BCL2 protein family) were shown to promote fusion in healthy cells in an MFN2 dependent manner [33, 59]. Additionally, a number of novel fusion factors were recently identified such as the cytosolic protein MSTO1 [60-62], the mitochondrial

F-box protein FBXL4 (discussed in this thesis) [63], the OMM mitochondrial carrier MTCH2

(SLC25A50) [64], the IMS GTPase ARL2 and its GTPase-activating protein ELMOD2 [65, 66].

The exact mechanism through which these proteins promote mitochondrial fusion remains elusive, emphasizing the fact that there are numerous gaps in our understanding of mitochondrial fusion.

Furthermore, recent work has shown that remodeling the phospholipid content of mitochondrial membranes can influence mitochondrial fusion [67, 68]. On the outer membrane, the phospholipase MitoPLD produces phosphatidic acid, a fusogenic phospholipid, to promote fusion [69]. Notably, the OMM proteins Mitoguardin-1 and -2 (MIGA1/2) have been implicated as fusion regulators by stabilizing MitoPLD [69]. Meanwhile, IMM cardiolipin (CL) has been shown to directly interact with OPA1 to mediate fusion [70, 71]. Finally, reduced levels of phosphatidylethanolamine in the IMM have been shown to activate IMM proteases, resulting in

36 increased OPA1 processing and fragmented mitochondrial networks [72]. SLC25A46, a member of the mitochondrial solute carrier family, was recently implicated in maintaining mitochondrial lipid homeostasis via contacts with ER and regulating mitochondrial dynamics

[73]. Notably, the yeast homolog of SLC25A46, Ugo1, coordinates mitochondrial fusion via interactions with mitofusin and OPA1 homologs in yeast, Fzo1 and Mgm1, respectively [74].

Meanwhile in humans, SLC25A46 has been suggested to negatively regulate mitochondrial fusion yet the exact mechanism of which remains unclear [73].

1.4 Mitochondrial Fission Regulation

1.4.1 Core Fission Machinery

Our mechanistic understanding of the molecular machinery governing mitochondrial fission is far more detailed compared to how little we know about fusion. The first step of mitochondrial fission involves a constriction event that is marked by mitochondrial contacts with the ER and is mediated by the actin-myosin cytoskeleton [75] (Fig 1.5). Several actin remodeling proteins have been implicated in promoting fission including Spire1c [76], an OMM- bound protein that nucleates actin filaments, and inverted formin protein 2 (INF2) [77], an ER- bound protein that regulates actin polymerization. Meanwhile, Cofilin1, an actin depolymerizing protein, negatively regulates fission [78]. Additionally, three non-muscle myosin isoforms

(NMIIA, NMIIB, NMIIC) provide the force required for mitochondrial constriction, however the functional overlap and relative contribution of each isoform remain unclear [79-81]. Moreover, the actin-based molecular motor protein Myosin Va interacts with Spire1C and DRP1 to promote fission [82].

The next step is recruitment of the GTPase DRP1 to fission sites, which can be mediated by OMM proteins Fis1, Mff, MiD51 and MiD49 [83-87]. DRP1 then assembles into oligomeric rings whereby GTP hydrolysis drives conformational changes that further constricts

37 these rings to drive mitochondrial scission [88, 89]. Finally, while Dynamin 2 (DNM2/DYN2) was proposed to be required for the final step of membrane scission [90], recent work suggests that it is not essential for fission [91, 92].

Figure 1.5. Schematic of mitochondrial fission machinery. A) ER-mediated constriction of mitochondria marks fission sites and actin-myosin complexes provide the necessary force for this pinching step. Notably, it has been proposed that mtDNA replication provides the initial signal for the assembly of the fission machinery. B) DRP1 is recruited to fission sites via OMM adaptor proteins. DRP1 oligomerization further constricts mitochondria at fission sites. C) Dynamin 2 may also play a role in constricting and severing mitochondria, though its role is debatable. GTP hydrolysis leads to conformational changes in DRP1 which enhances mitochondrial constriction and ultimately severs mitochondria. D) Fission ensures segregation of newly replicated mtDNA into daughter mitochondria.

Notably, DRP1 protein oligomerization and activity can be further modulated by PTMs.

Several activating modifications reside within the variable domain (VD) (also known as insert B) which is an intrinsically unstructured domain implicated in DRP1 stability and turnover [93]. Pro- fission DRP1 modifications include phosphorylation at S616, S-Nitrosylation, SUMOylation, acetylation and OGlcNAcylation [94-96]. Meanwhile, phosphorylation at DRP1 S637 is a well characterized inhibitory modification [94].

1.4.2 Auxiliary Fission Regulators

Regulation of IMM fission, which can occur independently of complete mitochondrial fission, is not well understood. Emerging evidence suggests that S-OPA1 isoforms and the IMM protein MTP18 (MTFP1) mediate IMM fission [38, 97]; however, further studies are required to identify bona fide IMM fission regulators. It is speculated that the activity of OMM fission machinery may be sufficient to drive IMM fission. Meanwhile, a recent study reported that targeting a human codon optimized version of FtsZ, the central component of bacterial division, to the mitochondrial matrix can induce inner membrane fission prior to DRP1 recruitment [98].

As such, it was speculated that there may be a matrix-localized IMM fission machinery yet to be identified [98].

Remodeling membrane lipids can also have a direct effect on the fission machinery [67].

Examples include the IMM lipid CL, which when localized to the OMM can interact with DRP1 and promote fission [99]. Conversely, creating a phosphatidic acid-rich region on the OMM was suggested to block mitochondrial division even after assembly of the fission machinery [100].

1.5 Consequences of impaired fusion or fission

The importance of mitochondrial fusion and fission regulation is most evident from various genetic ablation studies in murine models. Notably, loss of Mfn1, Mfn2 or Opa1 leads to

39 embryonic lethality in mice [31, 101]. Likewise, Drp1 knockout mice die in utero at embryonic day 11.5 [102], while mice lacking Mff die around 13 weeks due to dilated cardiomyopathy [103].

Thus, active fusion and fission processes are essential for proper cellular function, development and survival. Meanwhile, in humans, defects in fusion or fission regulation cause severe mitochondrial diseases.

1.5.1 Mitochondrial Dynamics and Mitochondrial Diseases

Mitochondrial diseases are a genetically and phenotypically heterogeneous group of diseases that can arise due to pathogenic mutations in the mitochondrial or nuclear genomes

[2]. The phenotypic spectrum of mitochondrial diseases is very broad and typically manifests in energetically demanding tissues (e.g. neuronal, skeletal, cardiac, hepatic tissue). However, mitochondrial diseases can affect any organ/tissue with varying degrees of severity, and disease manifestation can occur in infancy or in adulthood.

Given the well-established role of mitochondria in energy metabolism, mitochondrial diseases have been classically defined as defects in the respiratory chain complexes. However, our understanding of mitochondrial functions has significantly expanded, as has our appreciation for the numerous pathogenic mechanisms leading to mitochondrial dysfunction. In the context of mitochondrial dynamics, it is now well-established that fusion/fission imbalances, due to complete loss or dysfunction in fusion or fission factors, manifest as severe diseases in humans. Pathogenic mutations in MFN2 cause Charcot-Marie-Tooth disease type 2A, an autosomal dominant peripheral neuropathy characterized by early onset sensorimotor deficits in distal limbs [104]. Additionally, pathogenic OPA1 variants cause dominant optic atrophy and have been associated with peripheral neuropathy, deafness, ataxia and myopathy [105-107].

Moreover, heterozygous DRP1 variants have also been linked to global developmental delay, seizures, microcephaly and neonatal lethality [108-111]. Additionally, a number of newly recognized fusion or fission factors have also been associated with ataxia, myopathy (e.g.

MSTO1) [60-62, 112] (discussed in this thesis) and peripheral neuropathy (e.g. non-muscle myosin IIC) [80] (discussed in this thesis). Thus, balanced fusion and fission dynamics are crucial for proper mitochondrial function and overall homeostasis.

Notably, an interesting common feature in various diseases arising from mutations in mitochondrial fusion and fission regulators are disturbances to mtDNA. These include alterations in mtDNA abundance, nucleoid distribution, and mtDNA integrity (e.g. deletions)

(Table 1.1). In some cases, pathogenic variants in mitochondrial dynamics regulators have been reported to cause mtDNA depletion syndrome.

1.5.2 mtDNA depletion syndrome

MtDNA depletion syndromes are a severe class of mitochondrial diseases characterized by a significant reduction of mtDNA levels in affected tissues, typically manifesting in infancy as myopathic, encephalomyopathic, hepatocerebral or neurogastrointestinal disorders [113, 114].

The underlying cause of MTDPS are mutations in nuclear DNA encoded genes that regulate one of the following mechanisms: mtDNA replication, mitochondrial nucleotide pools or mitochondrial dynamics (e.g. OPA1, MFN2) [115-117]. Meanwhile, there are examples of mtDNA depletion syndrome genes with an unknown etiology (e.g. FBXL4, discussed in this thesis) [56, 57]. While it may be obvious that defects in mtDNA replication or alterations in mitochondrial nucleotide levels would lead to mtDNA loss, it is less clear why perturbations in the regulation of mitochondrial dynamics have profound effects on mtDNA levels.

The following sections outline what we know about mtDNA with an emphasis on how dynamic changes in mitochondrial networks influence mitochondrial genomes. It is worth noting that defects in fusion and fission regulation also perturb other mechanisms governed by mitochondrial dynamics regulators such as OXPHOS, cell division, apoptosis, metabolic reprograming and organellar contact sites. However, the focus of this thesis will be on understanding the interplay between fusion, fission and mtDNA maintenance.

Gene Mutation Patient Phenotype mtDNA abnormalities Ref (Protein)

p.Q74R, MtDNA Deletions (F, M) p.V226_S229del, CMT2A [116] ~ 2-fold Depletion (F, M) p.M376V, p.R707P

Optic Atrophy ‘Plus’ (Axonal Multiple mtDNA Deletions (M) MFN2 p.D210V Neuropathy, Mitochondrial [118] No Depletion (M) (MFN2) Myopathy)

Hypotonia, Ataxia, Axonal No mtDNA Deletions (M) p.D210Y Neuropathy, Optic Atrophy, [115] ~ 70% mtDNA Depletion (M) Hearing Loss

p.G439V, p.V910D, DOA ‘Plus’ (Sensorineural Multiple mtDNA Deletions (M) [105, p.R455H, p.S545R, deafness, Ataxia, No mtDNA Depletion (M) 106] OPA1 p.A357T Myopathy, cPEO) (OPA1) Encephalomyopathy, Optic No mtDNA Deletions (M) p.L534R Atrophy, Hypertrophic [117] ~ 80% mtDNA Depletion (M) Cardiomyopathy p.R345C, p.F376L; p.D236H, p.R279H; No MtDNA Deletions (F) MSTO1 Myopathy, Ataxia, Muscular [60- p.G420VfsX2, ~ 30-70% mtDNA Depletion (MSTO1) Dystrophy 62] p.R256Q (F, M) (compound het) # Infantile Lethality, Encephalopathy, Epilepsy, [109, DNM1L p.G362D, p.A395D, Optic Atrophy, Not Reported 110, (DRP1) p.R403C Microcephaly, Lactic 119] Acidosis Epileptic Encephalopathy, No mtDNA deletion (M) MFF [120, p.R298* Optic Atrophy, Peripheral No reports on mtDNA copy (MFF) 121] Neuropathy, Microcephaly number No mtDNA Deletions (M) MIEF2 p.Q92* Myopathy Elevated mtDNA copy number [122] (MID49) (M)

MYH14 No mtDNA deletions (F) [80, p.R941L Peripheral Neuropathy (NMIIC) No mtDNA depletion (F) 123]

Underlined mutations are homozygous, all other reported mutations were heterozygous. #: a comprehensive list of new MSTO1 mutations are summarized and discussed in Chapter 4.

1.6 Mitochondrial DNA and Links to Mitochondrial Dynamics

Mitochondria DNA (mtDNA) is a circular 16,569 bp genome, present within the mitochondrial matrix in a 100-1000 copies per cell. The mitochondrial genome contains 37 genes of which 2 encode ribosomal RNAs (rRNAs), 22 genes encode transfer RNAs (tRNAs) required for mitochondrial protein translation and the remaining 13 genes encode for subunits of the electron transport chain. As such, maintenance of mtDNA is key for supporting the bioenergetic function of mitochondria in a cell. Notably, mtDNA maintenance is completely dependent on nuclear-encoded factors imported or recruited to the mitochondria, discussed in more details below.

1.6.1 MtDNA Nucleoid Organization

MtDNA is spatially organized and tightly packaged into nucleo-protein structures known as nucleoids, mainly comprised of mtDNA and the mitochondrial transcription factor A (TFAM)

[124, 125]. Notably, mtDNA is a double-stranded circular DNA molecule with a contour length of approximately 5µm [126]. As such, the tight compaction of mtDNA by TFAM and other nucleoid associated proteins gives rise to nucleoids of about 100nm in diameter allowing them to fit within mitochondrial filaments (~250nm in diameter) [19, 126, 127]. Moreover, the proteinaceous components of nucleoids have also been proposed to tether nucleoids to the

IMM which facilitates nucleoid movement and distribution within mitochondria [6]. One example includes an alternatively spliced isoform of OPA1 (OPA1-exon4b), which is an IMM associated isoform. OPA1-exon4b has been shown to directly interact with nucleoids, mediate their even distribution throughout mitochondrial network and was further proposed to promote mtDNA replication via interacting with replisomes [40, 128].

Notably, visualizing mtDNA nucleoids by fluorescent microscopy in intact cells reveals an astonishing organization whereby uniformly sized nucleoids are distributed evenly throughout mitochondrial networks (Fig 1.6). This distribution pattern is observed in both reticular and

43 fragmented networks [22, 129]. The fact that each mitochondrial fragment typically contains at least one nucleoid is suggestive of a strictly controlled distribution mechanism. Indeed, mitochondrial fission has been shown to mediate segregation of nucleoids following replication which ensures redistribution of mtDNA to daughter mitochondria. Consistently, disrupting mitochondrial fission machinery (e.g. acute DRP1 knockdown) leads to significantly enlarged nucleoids (mito bulbs) due to a distribution defect [24].

Figure 1.6. Representative confocal images of control dermal fibroblasts. Cells were stained with MitoTrackerRed (mitochondria, red) and PicoGreen (nDNA and mtDNA) dyes. Images demonstrate uniform size and even distribution of mtDNA nucleoid throughout reticular and fragmented networks in control cells.

On the other hand, it is less clear how mitochondrial fusion contributes to nucleoid segregation. MFN1/2 double knockout and OPA1-exon4b RNAi fusion defective models demonstrate that reduced fusion gives rise to mtDNA nucleoid clumps and many mitochondrial fragments completely devoid of mtDNA [20, 22, 40]. Taken together, this strongly suggests that ongoing cycles of fission and fusion ensures even distribution of mtDNA nucleoids, though the

44 underlying mechanisms linking fusion defects to nucleoid clumps are unclear. Whether this phenotype is prevalent in other models of defective fusion is also an intriguing question.

1.6.2 MtDNA Replication

Replication of mitochondrial genomes is regulated by mitochondrial polymerase γ

(POLγ) which is composed of the catalytic subunit POLγA and the accessory subunit POLγB

[124]. In order to form a functional replisome POLγ works in concert with the mitochondrial single-stranded DNA binding protein (mtSSBP), mitochondrial DNA helicase (Twinkle), mitochondrial topoisomerase I [130, 131], mitochondrial DNA ligase III and mitochondrial genome maintenance exonuclease 1 (MGME1) [6]. The mitochondrial RNA polymerase

(mtRNAP), TFAM and mitochondrial transcription factor 2B (TFB2M), involved in mtDNA transcription, also generate an RNA primer required to initiate replication by POLγ [124].

Importantly, mitochondrial replisome components need to be in stoichiometrically balanced amounts for them to function properly. This is evident from a number of studies showing that overexpression or knockdown of replisome components can lead to mtDNA depletion and deletions [130, 132-134].

Proper replication of mitochondrial genomes also requires a balanced pool of intra- mitochondrial deoxyribonucleotide triphosphates (dNTPs) [135]. To this end, it is worth noting that nucleotides are either synthesized de novo or from pre-existing nucleosides by salvage pathways. While de novo synthesis of nucleotides does not occur in the mitochondria, the mitochondrial dNTP supply is based on import of cytosolic dNTPs to the matrix through specific transporters (e.g. SLC25A4, AGK and MPV17) or the activity of proteins in the mitochondrial salvage pathway (e.g. TK2, DGUOK, SUCLG1, SUCLA2, and ABAT). Additionally, cytosolic factors involved in nucleotide metabolism have also been implicated in balancing mitochondrial dNTP pools (e.g. RRM2B and TYMP) [135, 136].

1.6.3 MtDNA Copy Number

Mitochondrial genomes propagate constantly in order to maintain their high copy numbers in the cell. This occurs independent of nuclear genome replication and the cell cycle.

Furthermore, mtDNA replication is essential within dividing as well as postmitotic and terminally differentiated cells (e.g. neurons and myofibers) [137]. Notably, the absolute amount of mtDNA per cell varies in a tissue and cell type specific manner [138]. Maintenance of adequate levels of mtDNA copies per cell is important to support mitochondrial bioenergetic functions. However, the exact cellular mechanisms that sense and regulate mtDNA copy number are unclear.

Defects in maintenance of mtDNA copy number give rise to mtDNA depletion syndromes

[114]. As alluded to earlier, mtDNA depletion syndromes arise from mutations in nuclear genes that encode factors important for 1) mtDNA replication, 2) maintenance of mtDNA nucleotide pools or 3) mitochondrial dynamics. While it is expected that defects in mtDNA replication or imbalances in nucleotides would lead to mtDNA loss, how fusion-fission dynamics affect mtDNA levels is not fully understood.

Interestingly, the connection between mitochondrial network remodeling and mtDNA copy number dynamics was first noted in yeast models where disruption of OPA1 and MFN homologs, MGM1 and Fzo1p respectively, lead to complete loss of mtDNA [139-141]. Since then, this observation was extended to mammalian models of impaired mitochondrial fusion. For example, mouse embryonic fibroblasts (MEFs) lacking Mfn1/2 (double knockout) or Opa1 all display mtDNA depletion phenotypes [20, 22]. Likewise, conditional knockout mice also exhibit significant mtDNA loss in targeted tissues [20, 21].

Notably, because fusion is important for mitochondrial content mixing, impaired fusion gives rise to mitochondrial fragments with unbalanced proteomes [21]. In the context of mtDNA regulation, Ramos et al., demonstrated that in fusion defective MEFs as well as conditional

Mfn1/2 double knockout mice, the relative levels of mitochondrial replisome enzymes were significantly disrupted [20]. While mtSSBP1 and POLyA protein levels were reduced in all

46 models, the helicase TWINKLE was upregulated in Opa1 KO MEFs and conditional Mfn KO mice [20]. Clearly, mitochondrial fusion dynamics are important for maintaining the fine stoichiometric balance in mitochondrial replisome components. Nevertheless, whether this mechanism is valid in other models of impaired mitochondrial fusion with mtDNA depletion remains to be elucidated.

Furthermore, mitochondrial fission has also been shown to influence mtDNA copy number dynamics. For instance, ER-mediated constriction and DRP1 assembly at mitochondrial fission sites was shown to occur in close proximity to replicating nucleoids [23]. As such, fission is proposed to be important for the segregation of nascent mtDNA molecules after replication

[23]. Blocking fission has also been reported to lead to mtDNA loss in some experimental models (e.g. genetic ablation of Drp1 or NMIIB) [142, 143]. The functional link between fission and mtDNA copy number regulation is not completely understood, although ER-mediated mitochondrial constriction was proposed to prime mtDNA replication and coordinate subsequent distribution of mitochondrial genomes [23]. Exactly how cytosolic fission machinery can influence mtDNA in the matrix remains to be a puzzling question.

1.6.4 MtDNA Mutations & Quality Control

Given that mtDNA is present in hundreds of copies per cell, mtDNA can exist in a homoplasmic state, where all mtDNA sequences are identical. Additionally, mtDNA is susceptible to mutations, and a single cell can harbour a combination of wildtype and mutant mitochondrial genomes, a concept known as heteroplasmy. Typically, cells tolerate high levels of mutant mtDNA. However, when the proportion of mutant mtDNA surpasses a certain threshold and compromises mitochondrial functions, mtDNA mutations can lead to severe multisystemic diseases [144].

Fortunately, mitophagic clearance of dysfunctional mitochondria can select against the expansion of deleterious mtDNA mutations and shift heteroplasmy levels in favour of wildtype

47 mtDNA [137]. The most well-characterized mechanism regulating mitophagy involves the mitochondrial kinase PINK1 and the cytosolic E3 ubiquitin ligase Parkin which conjugates polyubiquitin chains to damaged mitochondria [17]. Notably, mutations in PINK1 and Parkin have been shown to shift mtDNA heteroplasmy in favour of mutant mtDNA in several C.elegans models [145-147]. In line with this observation, Parkin overexpression selects against deleterious mtDNA lesions and enriches mitochondria harbouring wildtype mtDNA [148, 149].

Furthermore, Parkin-mediated mitophagy was shown to reduce the accumulation of mtDNA mutations in PolγA mutant mice [150]. Thus, mitophagy plays a crucial role in preventing clonal expansion of deleterious mtDNA variants.

When it comes to clearing dysfunctional organelles, mitophagy is tightly coordinated with dynamic remodeling of mitochondrial networks. Fission has been proposed to segregate mutant mtDNA [151] and provide a necessary cue for initiating mitophagy [16]. However, having fragmented mitochondria is not sufficient for their autophagic clearance as it is also requisite that these fragments accumulate enough damage leading to membrane depolarization [16].

Notably, small mitochondrial fragments can be captured efficiently by the autophagosome while their depolarization also prevents refusion with the rest of the network [16, 152]. Thus, fission- mediated segregation of mutant mtDNA and mitophagic clearance provide an efficient mechanism for shifting mtDNA heteroplasmy in favour of wildtype genomes.

Taken together, dynamic remodeling of mitochondrial networks is crucial for various aspects of mtDNA maintenance. Meanwhile, perturbations in fusion and fission regulation and/or the maintenance of mitochondrial genomes give rise to debilitating mitochondrial diseases.

1.7 Summary of Main Chapters

Throughout the body of this thesis, I will highlight novel examples of mitochondrial or cytosolic proteins, pathogenic mutations in which are implicated in severe mitochondrial diseases. The common thread linking all these proteins is their function as regulators of dynamic changes in mitochondrial networks. Importantly, I will take a deep dive into understanding how these proteins participate in the maintenance of mitochondrial genomes and contribute to our understanding on the interplay between mitochondrial fusion, fission and mtDNA maintenance.

Chapter 3 highlights the mtDNA depletion syndrome gene, FBXL4. FBXL4 is a nuclear- encoded mitochondrial protein that has been proposed to localize to the mitochondrial intermembrane space [57]. Mutations in FBXL4 cause mtDNA depletion syndrome, a debilitating childhood onset disease presenting with mitochondrial encephalomyopathy, severe lactic acidosis and cardiac hypertrophy [56, 57, 153-158]. While FBXL4 mutations were shown to give rise to fragmented mitochondrial networks, altered nucleoid distribution and mtDNA loss, FBXL4 protein function was not known. Here, I studied a largely uncharacterized mutation in FBXL4

(C584R). I hypothesized that FBXL4 protein is a regulator of mitochondrial fusion and that the fusion defect in cells harbouring mutant FBXL4 was responsible for mtDNA loss and altered distribution of mitochondrial genomes. Additionally, I examined the consequences of restoring fused mitochondrial networks in cells with mutant FBXL4 protein, particularly examining mtDNA depletion and nucleoid organization phenotypes.

Chapter 4 discusses another mitochondrial disease gene, MSTO1. MSTO1 is a nuclear encoded cytosolic protein, mutations in which have been shown to lead to childhood onset muscular dystrophy, ataxia and cerebellar atrophy [61, 62, 112]. Notably, MSTO1 was recently characterized as a cytosolic regulator of mitochondrial fusion [61, 62]. While pathogenic mutations in MSTO1 protein were shown to impair mitochondrial fusion activity, it remained

49 unclear whether this fusion defect had an impact on mtDNA maintenance. Here, through a collaboration with multiple medical genetics centres, we compiled the largest patient cohort harbouring various mutations in MSTO1. Through functional characterization of mitochondrial dysfunction in fibroblasts from 7 patients, I asked whether these mutations were the underlying cause of disease. As well, I hypothesized that the fusion defect in cells harbouring mutant

MSTO1 protein leads to mtDNA depletion and abnormal nucleoid distribution. Finally, I explored the functional consequences of restoring wildtype MSTO1 protein on mitochondrial network morphology, mtDNA levels and nucleoid distribution in MSTO1 patient cells.

Chapter 5 investigates a member of the nonmuscle myosin protein family, nonmuscle myosin IIC (NMIIC). NMIIC is a cytosolic protein encoded by MYH14, mutations in which lead to progressive peripheral neuropathy and hearing loss [80]. Notably, work from our lab recently showed that NMIIC participates in the ER-mediated constriction of mitochondria at fission sites and mediates fragmentation of mitochondrial networks [80]. My contribution to this study was examining patient fibroblasts harbouring the pathogenic R941L mutation in MYH14. Given that the R941L variant impairs NMIIC pro-fission activity, I hypothesized that the fission defect is associated with abnormal mtDNA nucleoid distribution throughout mitochondrial networks.

Taken together, the work outlined in the following chapters implicates further components of the fusion and fission machinery in the maintenance of mtDNA.

1.7.1 Statement of Contribution

The introductory chapter of this thesis (Chapter 1) contains sections from a published literature review, written by myself and Dr. Shutt. [Sabouny R & Shutt TE. Mitochondrial Fission and Fusion are Reciprocally Regulated”. 2020. Trends in Biochemical Sciences.

(https://doi.org/10.1016/j.tibs.2020.03.009)]

Chapter 3 is based on the following published manuscript, written by myself and Dr.

Shutt, with contributions and comments from all the listed authors. [Sabouny R, Wong R, Lee-

Glover L, Greenway SC, Sinasac D, Care4Rare Canada Consortium, Khan A, Shutt TE.

Characterization of the C584R variant in the mtDNA depletion syndrome gene FBXL4, reveals a novel role for FBXL4 as a regulator of mitochondrial fusion. 2019. BBA Molecular Basis of

Disease;1865(11):165536].

In this study, I performed all the experiments, data collection and analyses outlined in this chapter. My colleagues Rachel Wong and Laurie Lee-Glover generated very valuable

FBXL4 and USP8 mammalian expression plasmids during their summer projects at the Shutt

Lab. Dr. Greenway provided clinical expertise on reporting patients’ cardiac anomalies. Dr.

Sinasac generously provided control and patient derived dermal fibroblasts and lead clinical biochemical genetics analyses. Dr. Khan was the clinician who cared for the patients, performed muscle biopsies and subsequent diagnoses. Dr. Shutt supervised this study.

Chapter 4 is based on the following published manuscript, as part of a large collaboration across many medical genetics clinics. The manuscript was written by myself, Dr.

Donkervoort and Dr. Shutt with contributions and comments from all the listed authors.

[Donkervoort S*, Sabouny R*, Yun P, Gauquelin L, Chao KR, Hu Y, Al Khatib I, Töpf A,

Mohassel P, Cummings BB, Kaur R, Saade D, Moore SA, Waddell LB, Farrar MA, Goodrich JK,

Uapinyoying P, Chan SHS, Javed A, Leach ME, Karachunski P, Dalton J, Medne L, Harper A,

Thompson C, Thiffault I, Specht S, Lamont RE, Saunders C, Racher H, Bernier FP, Mowat D,

Witting N, Vissing J, Hanson R, Coffman KA, Hainlen M, Parboosingh JS, Carnevale A, Yoon G,

Schnur RE; Care4Rare Canada Consortium, Boycott KM, Mah JK, Straub V, Foley AR, Innes

AM, Bönnemann CG, Shutt TE. MSTO1 mutations cause mtDNA depletion, manifesting as muscular dystrophy with cerebellar involvement. 2019. Acta Neuropathologica, doi:

10.1007/s00401-019-02059-z. (*co-first authors)].

In this study, I performed all the experiments, data collection and analyses on control and patient fibroblasts provided by medical genetics clinics at the Alberta Children’s Hospital

(Calgary, AB), SickKids Hospital (Toronto, ON) and National Institute of Health (USA). Iman Al

Khatib, my colleague at the Shutt lab, assisted in nucleoid analyses to complete revisions for this manuscript. Dr. Shutt supervised this study.

Chapter 5 is based on the following published manuscript, on which I was a co-author and have primarily investigated mtDNA copy number and nucleoid phenotypes in patient fibroblasts. [Almutawa W, Smith C, Sabouny R, Smit RB, Zhao T, Wong R, Lee-Glover L,

Desrochers-Goyette J, Ilamathi HS; Care4Rare Canada Consortium, Suchowersky O, Germain

M, Mains PE, Parboosingh JS, Pfeffer G, Innes AM, Shutt TE. The R941L mutation in MYH14 disrupts mitochondrial fission and associates with peripheral neuropathy. 2019. EBioMedicine;

45:379-392.]

In this chapter, I have included my work examining the effects of the R941L mutation in

MYH14 on mitochondrial morphology at the cell periphery as well as the consequences of defective fission regulation on mtDNA nucleoids. My colleagues Walaa Almutawa and Chris

Smith conducted the majority of experiments and analyses for this publication. Dr. Pfeffer and

Dr. Innes were the primary clinicians caring for the family in which the mutation was identified and Dr. Shutt supervised this study.

MATERIALS & METHODS

This chapter is based on materials and methods sections in published manuscripts with permission from Sabouny et al., 2019; Donkervoort et al., 2019; and Almutawa et al., 2019 (see Appendices C, D, E).

2.1 Ethics Statement

Informed consent for enrollment in the Care4Rare project and for publication was obtained according to standard procedures. This study was approved by the Conjoint Ethics

Board of the University of Calgary and work with human cell lines was conducted under the following certificate: REB16-1776.

2.2 Cell Culture

Control and patient primary fibroblasts (from skin or muscle biopsies, obtained with informed consent) were cultured in MEM media (Gibco, 11095080) containing L-Glutamine and supplemented with 10% fetal bovine serum (FBS) and 1mM sodium pyruvate. HEK cells, HeLa cells and mouse embryonic fibroblasts (MEFs) were grown in DMEM (Gibco, 11965092) supplemented with 10% FBS. Cells were maintained at 37°C and 5% CO2. Where indicated, cells were treated with 10mM DCA (Sigma Aldrich, 347795) for 24hrs, 25µM mdivi-1 (Sigma

Aldrich, M0199) for 24hrs or 1 week or 25µM Chloroquine (Sigma Aldrich, C6628) for 24hrs.

2.3 Plasmids & Cloning

The mito-PAGFP was a gift from Richard Youle (Addgene plasmid # 23348; http://n2t.net/addgene:23348; RRID: Addgene_23348) [159]. A mammalian expression vector containing C-terminal HA-tagged FBXL4 (a generous gift from Dr. Zeviani) was generated by cloning the FBXL4-HA open reading frame into a pcDNA3.1- backbone using InFusion cloning

(Takara, Clonetech). The FBXL4 C584R mutation was introduced using mutagenesis primers. A mammalian expression vector containing MSTO1-V5 was obtained from DNASU

(HsCD00440595). The MSTO1 C222A mutation was generated using mutagenesis primers. To generate the MSTO1-P2A-mCherry construct, the MSTO1 sequence was cloned into an

AmCyan-P2A-mCherry construct (AmCyan-P2A-mCherry was a gift from Ilpo Huhtaniemi,

Addgene plasmid # 45350 ; http://n2t.net/addgene:45350 ; RRID:Addgene_45350) [160],

54 replacing mCyan with MSTO1 sequence. An empty vector containing only mCherry was also generated.

IMS-GFP-USP8 construct was generated by InFusion cloning. The catalytic domain of

USP8 (residues 776-1110) [161] (DNASU, HsCD00436965) was used to generate the plasmid.

The IMS-targeting (RSVCSLFRYRQRFPVLANSKKRCFSELIKPWHKTVLTGFGMTLCAVPI) sequence was synthesized as a G-Block with an N-terminal methionine start codon (Integrated

DNA Technologies). GFP sequence was obtained by PCR amplification from another plasmid.

The three inserts were designed/PCR amplified to contain homologous flanking sequences and were cloned into a pcDNA3.1- backbone in this order (N- IMS-targeting sequence – GFP –

USP8 -C). The catalytically inactivating C786A mutation [162, 163] was introduced by mutagenesis primers.

2.4 Transfections

HEK and HeLa cells were transfected with the indicated plasmids using Lipofectamine

3000 (Thermo Fisher Scientific, L3000015) according to manufacturer’s instructions. Briefly, cells were seeded at 3.6X105 (HEK) or 4.5X105 (HeLa) in 6-well plates and allowed to grow overnight. The following day, 2ug of plasmid were used for transfections and cells were left to grow for 24 hours before harvesting/fixation.

Overexpression of PA-GFP in fibroblasts and MSTO1 genetic rescue experiments were mediated by electroporation using the Amaxa Nucleofector II system (Lonza). Fibroblasts were grown to 70-80% confluence, harvested and resuspended in OptiMEM media. Next, 1X106 cells in 100 µL media and 2 µg of plasmid DNA were transferred to a sterile 2mm electroporation cuvette (VWR 89047-208) and electroporated using the A-024 program. Cells were then plated onto 35mm glass bottom dishes or in 100mm plates and maintained at 37°C and 5% CO2 for 48 hours prior to further analysis.

2.5 Cell Sorting

Following electroporation with either mCherry empty vector or MSTO1-P2A-mCherry, approximately 4X106 fibroblast cells were sorted for red fluorescence (and MSTO1 expression) using a 130um nozzle on a BD FACSAria Fusion (FACSAriaIII) cytometer (BD Biosciences), supported by FACSDiva Version 8.0.1. Genomic DNA was subsequently purified from control and patient mCherry-positive cells as described below.

2.6 mtDNA Copy Number Analysis

Genomic DNA (nuclear and mitochondrial DNA) was isolated from control and patient fibroblasts (seeded at 5x105 cells) using the PureLink Genomic DNA Mini Kit (Thermo Fisher

Scientific, K182001) according to manufacturer’s instructions. The QuantStudio 6 Flex Real-

Time PCR system (Thermo Fisher Scientific) was used to assess relative mtDNA copy number.

Primer sequences to amplify mtDNA and the nuclear-encoded housekeeping gene 18S, and thermocycling conditions were exactly as described in [164]. Briefly, the 20µL quantitative PCR

(qPCR) reactions contained 10µL PowerUp SYBR Green Master Mix (Thermo Fisher Scientific,

A25742), 100ng gDNA and 500nM forward and 500nM reverse primers (final concentrations).

MtDNA copy number relative to 18S was analyzed using the delta delta Ct method and represented as percent control [165]. Reactions were performed in triplicate and mtDNA copy number analysis was performed on at least three independent biological replicates. Data is presented as mean ± SD and unpaired, 2-tailed Student’s t-tests were used to determine statistical significance.

2.7 Long-range PCR

To examine mtDNA deletions, the following primers were used to amplify nearly full length mtDNA (16.3 kb), (1482–1516 F: ACCGCCCGTCACCCTCCTCAAGTATACTTCAAAGG;

1180–1146 R: ACCGCCAGGTCCTTTGAGTTTTAAGCTGTGGCTCG) as reported previously

[166]. Long range PCR reactions were performed using the Takara LA Taq polymerase (Takara

Bio, RR002M), with 250 ng genomic DNA, 200 nM forward and reverse primers. The PCR cycling conditions were as follows: 94 °C for 1 min; 98 °C for 10 s and 68 °C for 11 min

(30 cycles); and a final extension cycle at 72 °C for 10 min. PCR products were visualized by electrophoresis on a 0.6% agarose gel, run for approximately 12 h at 20 V.

2.8 Western Blot

For Western analyses, 3-5x105 fibroblasts were seeded in 100mm plates, and allowed to grow for 2 days. Cells were harvested by trypsinization, pelleted, washed with 1X phosphate buffered saline (PBS) and lysed with RIPA buffer containing protease inhibitors. Total cell lysates (20-50ug) were resolved on SDS-PAGE gels and transferred onto PVDF membranes.

For analyses examining expression of OXPHOS subunits, 50µg total cell lysates were loaded, meanwhile for other Western analyses, 20µg protein was used. Blots were subsequently probed with the following antibodies (1:1000 dilution unless otherwise indicated): OXPHOS antibody cocktail (Abcam, ab110411), anti-MFN1 (Cell Signalling, 14739), anti-MFN2 (Abnova,

H00009927-M03), anti-OPA1 (BD Bioscience, 612606), anti-DRP1 (BD Bioscience, 611113), anti-β-actin (Sigma, A5316), anti-HSP60 (Cell Signalling, 12165), anti-tubulin (Santa Cruz

Biotechnology, D-10), anti-MSTO1 (Genetex, GTX105110) (1:500), anti-V5 (Millipore, AB3792), and the appropriate horseradish peroxidase (HRP)-conjugated secondary antibodies (1:3000).

Blots were incubated with Clarity ECL substrate (Biorad, 1705061) and imaged using an

Amersham Imager AI600.

2.8.1 Densitometry Analysis:

Densitometric analysis of protein band intensities were performed using ImageJ [167].

Band intensities were quantified, normalized to respective loading control (VDAC1, HSP60, or

β-actin) and presented as a percentage of vehicle control set to 100%.

2.9 Mitochondrial Respiration

The Seahorse XFe24 Extracellular Flux Analyzer (Agilent Technologies, Inc) was used to examine mitochondrial bioenergetic profiles in control and patient fibroblasts. Briefly, 3.75x104 cells/well were seeded in an XF24 microplate and incubated at 37°C, 5% CO2 for 24hrs. Next,

DCA (10mM) was added where indicated and the cells were allowed to grow for another 24hrs.

Subsequently, growth media was replaced with assay media supplemented with D-Glucose

(25mM), sodium pyruvate (2mM) and L-Glutamine (4mM). Oxygen consumption rates were measured following sequential injection of the following compounds into each well: oligomycin

(1µg/mL) (Enzo Life Sciences, BML-CM111), carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone (FCCP, 1µM) (Enzo life Sciences, BML-CM120) and Antimycin A (1µM)

(Sigma Aldrich, A8674). Upon completion of the assay, assay media was aspirated, wells were carefully washed with 1XPBS and 15µL of RIPA buffer were added. Protein concentrations from each well were measured by BCA assay (Thermo Fisher Scientific, 23225) and used to normalize data.

2.10 Immunofluorescence Staining

Fibroblasts were seeded on 12mm glass coverslips (no. 1.5) at 2X104 cells and incubated for 1-2 days. Subsequently, cells were fixed with 4% paraformaldehyde and stained with primary antibodies against TOMM20 (Santa Cruz Biotechnology, FL-145 or F-10), DNA

(Millipore, CBL186), HA (Santa Cruz Biotechnology, F-7), LAMP1 (Santa Cruz Biotechnology,

18821), Calnexin (Sigma Aldrich, MAB3126) and appropriate Alexafluor-conjugated secondary antibodies (Thermo Fisher Scientific) at 1:1000.

2.11 Live Cell Staining

In order to visualize mitochondrial DNA nucleoids, fibroblasts were stained with

PicoGreen (Thermo Fisher Scientific, P7581) as described previously [168]. Briefly, cells were

58 seeded on glass bottom dishes (Mattek, P35G-1.5-14-C) at 8X104 and incubated overnight.

Approximately one hour prior to imaging, cells were stained with PicoGreen at 3uL/mL for 30-45 minutes at 37°C. MitoTracker Red dye (50nM) (Thermo Fisher Scientific, M7512) or MitoTracker

Deep Red dye (50 nM) (ThermoFischer Scientific, M22426) were added simultaneously to the media to visualize mitochondrial networks. The media containing dyes was aspirated, cells were washed four times in pre-warmed 1XPBS, and fresh pre-warmed media was added to the cells.

Similarly, where indicated, cells were stained with the membrane potential-dependent dye

TMRE (Tetramethylrhodamine, ethyl ester) (50nM, 20 minutes) (Life Technologies, T669) as above, washed with 1XPBS and replenished with pre-warmed media and 10mM HEPES buffer.

2.12 Microscopy

Images from both fixed and live samples were acquired on an Olympus spinning disc confocal system (Olympus SD OSR) (UAPON 100XOTIRF / 1.49 oil objective) operated by

Metamorph software. A cell Vivo incubation module was used to maintain cells at 37°C and 5%

CO2 during live cell imaging. Where indicated, live and fixed cell imaging was performed on a line scanning confocal microscope (Zeiss LSM 700) (40X/1.4 oil objective) operated by Zen software. Live cells were maintained at 37ºC for the duration of imaging and media was buffered with HEPES.

2.13 Photo-activatable GFP fusion assay

Fibroblasts transfected with a matrix-targeted PA-GFP and stained with TMRE were imaged on the Zeiss LSM system to assess fusion rates. For photo-activation, the 405nm laser at 10% power was used to photoactivate a region of the mitochondrial network (approximately

5x5µm). An image was acquired pre-photoactivation, five seconds post-photoactivation and dynamic changes in mitochondrial networks were subsequently recorded at 120s intervals (up to 960s) using the 488nm and 561nm lasers.

2.14 Phototoxicity-dependent fission assay

For the phototoxic-dependent mitochondrial fission assay using the Olympus SD-OSR microscope, MitoTrackerRed-labelled fibroblasts were repeatedly exposed (100 ms at 1 frame per second intervals) to high levels of the 561 nm excitation laser (100 mW at 4% power), for a total of 5 min and images were acquired at these intervals.

2.15 Image analysis

2.15.1 Mitochondrial Networks

Mitochondrial network morphology was qualitatively analyzed by classifying networks into three or four categories, fragmented (predominantly small mitochondrial fragments), partly fragmented (cells containing short fragments and small mitochondrial puncta), Intermediate

(cells containing a mixture of small puncta and elongated networks) and fused (elongated, interconnected networks with few to no short fragments). For each fibroblast or transfected HEK line, at least 50 cells were scored. Morphology analyses were performed on 3 independent replicates, the results represent mean ± SD, and P values were based on unpaired, 2-tailed

Student’s t-tests.

2.15.2 MtDNA Nucleoids

Mitochondrial DNA nucleoid size and number were analyzed using the particle analysis tool in ImageJ FIJI [167]. Fibroblasts from fixed cells stained with an anti-DNA antibody (Z- stack), or live cells stained with PicoGreen (single plain) were imaged using the indicated acquisition parameters, and images were subsequently scaled and binarized. For each cell, a region of interest (ROI) encompassing the entire mitochondrial network was selected. Then, in binarized mtDNA nucleoid images, the particle analysis tool was used to measure surface area and total nucleoid counts within the selected ROI. Nuclear signal was excluded from the analysis. The analyses were performed on at least 10 fibroblasts for each patient and control

60 lines. Nucleoid counts represent mean ± SD and P values were based on unpaired, 2-tailed

Student’s t-tests.

In Chapter 3, nucleoid sizes are presented as the average size of all nucleoids per cell ±

SD and P values were based on unpaired, 2-tailed Student’s t-tests. In Chapter 4, nucleoid sizes are presented as the average size of all nucleoids per cell in a violin plot, highlighting the distribution of quantified mtDNA nucleoid sizes. The non-parametric Kolmogorov-Smirnov (K-S) test was used to determine statistical significance regarding the distribution of nucleoid sizes.

The relative frequencies of nucleoid sizes were assessed in each quantified cell and the percentage of large nucleoids (>0.2um2) in control and patient fibroblasts was plotted.

2.15.3 Lysosomes

The presence of enlarged lysosomal aggregates was quantified from confocal images of patient and control cells by scoring the number of cells with normal vs. enlarged lysosomes. At least 50 cells were analyzed per fibroblast line from two independent replicates. Results represent mean ± SD, and P values were obtained from unpaired, 2-tailed Student’s t-tests.

2.15.4 Mitochondrial Fusion Analysis

Mitochondrial fusion rates were assessed in control and patient fibroblasts expressing the matrix-targeted PA-GFP by two methods; examining 1) diffusion of the GFP signal and 2) mean fluorescence intensity of GFP signal overtime. Firstly, the diffusion of the GFP signal was assessed by comparing the percentage of the network that was photoactivated at 5s to the percentage of the network containing GFP signal at the final acquisition frame (t=960s). Images in the GFP and TMRE channels were binarized and the area of mitochondrial network in the

GFP channel relative to the total network in the TMRE channel was measured at each timepoint. Wider diffusion of the GFP signal throughout the network is indicative of higher fusion rates. Representative binarized images used for these analyses are depicted in (Fig 3.7).

Analyses were performed using ImageJ FIJI and data represent the average ± SD from 5 cells for each fibroblast line. Unpaired, Student’s t.tests were used to assess statistical significance.

Secondly, mean fluorescence intensity in the photo-activated region (5x5µm box) was recorded over the acquisition period using Zen Black software, as done previously [169]. A reduction in fluorescence intensity overtime is suggestive of higher fusion rates as the GFP signal spreads throughout the network. Notably, acquisition parameters were optimized to minimize photo-bleaching of the GFP and TMRE signals.

2.16 Mitochondrial Membrane Potential and Mitochondrial Mass Analyses

Control and patient fibroblasts were seeded in 6-well plates at 1.5X105 cells per well and incubated for 1-2 days. On the day of analysis, cells were stained with TMRE (50nM, 20 minutes) to examine mitochondrial membrane potential. The ionophore FCCP

(Carbonylcyanide-4-(trifluoromethoxy)-phenylhydrazone, 10uM) (Enzo Life Sciences, BML-

CM120) was used as a negative control. To evaluate mitochondrial mass, another set of cells were seeded as above and labelled with the membrane potential-independent dye MitoTracker

Green (50nM, 20 minutes) (Life technologies, M7514). After staining, cells were washed three times with pre-warmed PBS, trypsinized and pelleted. Cell pellets were resuspended in complete media and the BD LSR II flow cytometer supported by the BD FACSDiva software (BD

Biosciences) was used to measure signal intensity from the respective dyes. Mean fluorescence intensity was recorded for approximately 20,000 events in triplicates of each fibroblast line.

Results are presented as percent control and Unpaired Student’s t-tests were used to assess statistical significance.

2.17 In vitro Cell-Free Mitochondrial Fusion Assay

The in vitro mitochondrial fusion assay was performed as previously described with few modifications [170]. The following sections detail mitochondria and cytosol isolation steps and the in vitro fusion assay.

2.17.1 Mitochondrial isolation and cytosol prep

Mitochondria were isolated from two suspension HeLa cell lines (sHeLa) stably expressing a split YFP – split luciferase reporter construct in the mitochondrial matrix. One cell line contained the N-terminus of the reporter construct (N-Luci) and the other contained the C- terminus of the reporter (C-Luci). sHeLa cells were grown in suspension (2L flasks) and harvested by centrifugation at 4000xg, 4°C for 10 minutes. Cell pellets were washed and resuspended in cold mitochondria isolation buffer (MIB) (220 mM Mannitol, 68 mM Sucrose, 80 mM KCl, 0.5 mM EGTA, 2 mM MgAc2, and 10 mM HEPES, pH 7.4) and then mechanically broken using a Dounce homogenizer on ice. Mitochondria were subsequently isolated by differential centrifugation, first pelleting nuclei at 600g, 4°C for 10 minutes, then pelleting mitochondria from the post-nuclear supernatant at 10000g, 4°C for 15 minutes. Mitochondria- enriched pellets were then washed with MIB and resuspended in MIB containing 10% glycerol, snap frozen in liquid nitrogen and stored at -80°C. Meanwhile, the supernatant (cytosol) was aliquoted and stored at -80°C.

2.17.2 Mitochondrial Fusion Assay

The in vitro mitochondrial fusion assay was performed in 25uL reactions containing 50ug of each mitochondrial population (N-Luci and C-Luci), 0.5mM GTP, 2mM K2HPO4, 10mM

ATP(K+), 0.08 mM ADP, 5mM Na succinate and MIB. Where indicated, cytosolic extracts were added to the fusion reaction to a final concentration of 1ug/ul. Negative controls were prepared containing only N-Luci, only C-Luci mitochondria or no mitochondria. Reactions were prepared on ice, incubated at 37°C for 5 minutes. Luciferase activity, measuring mitochondrial fusion, was

63 determined using the Renilla Luciferase Assay System (Promega, WI, USA) and the

SpectraMax L luminometer (Molecular Devices) according to the manufacturer's protocol. In vitro mitochondrial fusion activity was measured in triplicates for each condition.

2.18 Transmission electron microscopy

For transmission electron microscopy (TEM), fibroblasts were grown in 24-well tissue culture plates and further processed for fixation, dehydration and embedding by the Microscopy

Imaging Facility at the University of Calgary. Briefly, the cells were prefixed with 1.6% paraformaldehyde and 2.5% glutaraldehyde in 0.1 M cacodylate buffer, pH 7.3, for 1 hour. After washing, cells were post-fixed with cacodylate-buffered 1% osmium tetroxide for 1 hour at room temperature, dehydrated through graded ethanol and embedded in an Epon resin. Ultrathin sections were obtained using an ultramicrotome and then stained with aqueous uranyl acetate and Reynolds’s lead citrate. Samples were visualized using a Hitachi H7650 TEM at 80kV and images were captured using an AMT600 digital camera.

Characterization of the C584R variant in the mtDNA depletion syndrome gene FBXL4, reveals a novel role for FBXL4 as a regulator of mitochondrial fusion.

This chapter is based on a published manuscript with permission from Sabouny et al., 2019 (see Appendix C).

3.1 Introduction

FBXL4 (F-Box and Leucine rich repeat protein 4) is a nuclear-encoded mitochondrial protein proposed to localize to the mitochondrial intermembrane space (IMS) [57]. Although mutations in FBXL4 cause a mitochondrial DNA (mtDNA) depletion syndrome with infant-onset encephalopathy and lactic acidosis (OMIM 615471, [171]) [56, 57, 153-158, 172-174], the exact function of the FBXL4 protein has remained elusive. Thus, we do not have a clear understanding of how FBXL4 impairment leads to disease. MtDNA depletion syndromes are a subclass of mitochondrial disease with genetic and phenotypic heterogeneity, which are characterized by significant reduction of mtDNA content in affected tissues [113, 114]. MtDNA depletion syndromes are typically caused by mutations in nuclear-encoded genes that support mtDNA replication (e.g., POLG and TWNK) or maintenance of mitochondrial deoxyribonucleotide triphosphate (dNTP) pools (e.g., TK2, DGUOK, RRM2B, TYMP, SUCLA2 and SUCLG1) [113, 114]. However, mounting evidence suggests that dysregulation of mitochondrial fusion can also cause mtDNA depletion in mitochondrial disease patients [105,

115-117, 175]. To date, FBXL4 is one of the few mtDNA depletion syndrome genes with an unknown pathogenesis mechanism [113, 114].

Fusion and fission events constantly remodel mitochondrial shape, which can range from connected reticular networks to fragmented puncta [4, 176]. Importantly, changes to mitochondrial morphology encompass redistribution of mitochondrial contents, including mtDNA nucleoids (nucleoprotein structures into which mtDNA is packaged) [6, 22, 24, 124].

Mitochondrial fission is mainly driven by the GTPase Dynamin Related protein 1 (DRP1) [5].

Meanwhile, fusion of mitochondrial membranes is also regulated by large GTPases namely,

Mitofusin 1 and 2 (MFN1/2) at the outer mitochondrial membrane, and Optic Atrophy 1 protein

(OPA1) at the inner mitochondrial membrane [5]. Cells lacking any of the core fusion proteins

66 contain not only fragmented mitochondrial networks, but also abnormal distribution of mtDNA nucleoids and mtDNA genome instability [22]. Moreover, pathogenic mutations in MFN2 and

OPA1 genes have been reported to cause mtDNA depletion in patients [105, 115-117, 175].

Nevertheless, we still do not have a complete understanding of how impaired fusion causes mtDNA depletion.

Early onset, multisystemic presentation is characteristic of mtDNA depletion syndromes, and in the absence of effective therapeutics, poor prognosis is projected if symptoms are left unmanaged. Lactic acidemia is a common metabolic problem among mitochondrial disease patients, especially in patients carrying pathogenic FBXL4 variants [56, 57, 153-158, 172-174].

One way to tackle elevated lactate levels is by administering dichloroacetate (DCA), a glycolysis inhibitor that functions by inhibiting pyruvate dehydrogenase kinase, leading to activation of the pyruvate dehydrogenase complex (PDHC) [177, 178]. Ultimately, activating PDHC prevents the conversion of pyruvate to lactate, and thus reduces lactate accumulation in blood and tissues

[177, 178].

Herein, we report two siblings harboring a previously uncharacterized homozygous mutation in FBXL4 (c.1750T>C; p.Cys584Arg, hereafter referred to as C584R), and exhibiting characteristic phenotypes of FBXL4 dysfunction. Notably, the second patient, who was treated with DCA, showed improved lactate levels and reversal of cardiac hypertrophy. We set out to characterize patient fibroblasts harbouring the C584R mutation and confirm the pathogenic nature of this variant. In addition, as no effective treatment have been reported for FBXL4 impairment, we examined whether DCA improved the underlying mitochondrial function. Finally, we wanted to test the hypothesis that FBXL4 promotes mitochondrial fusion, which would explain the mtDNA depletion in these patients.

3.2 Results

The sequencing component of this work was part of Care for Rare (http://care4rare.ca) and the tissue analysis was part of the MITO-FIND project (mitochondrial functional and integrative next generation diagnostics) sponsored by MitoCanada (http://mitocanada.org). For both projects, consent was obtained following standard procedures through the University of

Calgary Conjoint Research Ethics Board.

3.2.1 Clinical Description

We report two siblings (hereafter referred to as P1 and P2) from a consanguineous family of Arab descent with three healthy children and three miscarriages (Fig 3.1 A). The probands presented with encephalomyopathy, developmental delay, severe lactic acidosis and hypertrophic cardiomyopathy, suggestive of mitochondrial disease. P1, the older female sibling died at 7 months of age.

Figure 3.1. Clinical investigation of mitochondrial disease patients. A) Family pedigree of two mitochondrial disease patients (P1 and P2) from a consanguineous marriage, both of whom presented with encephalomyopathy, severe lactic acidosis and hypertrophic cardiomyopathy. Exome sequencing of P2 identified a homozygous recessive mutation in the mitochondrial protein, FBXL4 (c. 1750T>C; p.C584R). B) Increasing doses of dichloroacetate (DCA) administered to P2 around 6 weeks of age remarkably improved lactic acidosis. Dose range: 25- 50mg/kg/day. C) Interventricular septum thickness z-scores for P2 indicate thickening of the of the left ventricular walls which normalize after DCA administration. Z-scores were measured by the Boston Children’s Hospital z-score calculator. D) Sanger sequencing of exon 9 in FBXL4 confirms the homozygous variants in P1 and P2.

P2, the younger male sibling, was investigated more closely and showed features of polymicrogyria, cryptorchidism, chronic hydronephrosis, and cerebellar hypoplasia. Metabolic investigations in P2 showed an elevated plasma lactate:pyruvate ratio (lactate 9.5 mmol/L, pyruvate 0.23 mmol/L, L:P ratio 41; reference < 25), and elevated urinary 3-methylglutaconic acid of 18.2 mmol/mol creatinine (reference 1.9-9.1). In addition, electron transport chain activity was globally reduced in muscle tissue, and muscle pathology was abnormal as detailed previously, prior to exome sequencing, as Case 1 in Sarnat et al., [179]. P2 died at 2 years and

10 months of age after being transferred to palliative care owing to a respiratory infection that was secondary to the disease.

3.2.2 Dichloroacetate improved lactic acidosis and cardiac hypertrophy in patient 2

Based on the suspected mitochondrial disease in the first affected sibling (P1), P2 received vitamin D (600IU/day), vitamin B1 (10 mg/kg/day), CoQ10 (10 mg/kg/day) and propranolol starting at birth. In response to increasing blood lactate levels, administration of dichloroacetate (DCA; TCI America) was started at 6 weeks of age in the dose range between

25-50 mg/kg/day (Fig 3.1 B). The clinical experience showed doses above 25 mg/kg/day were needed to reduce blood lactate levels with dosing changes adjusted for increasing weight.

Remarkably, septal hypertrophy also normalized after DCA administration was initiated, with echocardiography showing no ventricular hypertrophy at 18 months of age (Fig 3.1 C). It is worth noting that DCA dosage was not adjusted after P2 was transferred to palliative care, hence the observed elevation in venous lactate later in life (Fig 3.1 B).

3.2.3 Genetic studies identify homozygous variants in FBXL4

Mitochondrial DNA (mtDNA) from muscle DNA was fully sequenced in P2, but no mutations or deletions were detected. Meanwhile, whole exome sequencing revealed a novel homozygous variant in the nuclear-encoded mitochondrial protein, FBXL4 (c.1750T>C; p.C584R) in P2. This variant was confirmed in both patients by Sanger sequencing (Fig 3.1 D).

Of note, while the C584R variant was reported previously, along with 10 other novel FBXL4 variants, no detailed molecular studies of the C584R variant were performed [173].

3.2.4 Cells homozygous for the C584R variant in FBXL4 exhibit mtDNA depletion and bioenergetic defects

Skin fibroblasts obtained from both patients were used to examine novel mitochondrial phenotypes, as well as phenotypes reported previously in fibroblasts from other FBXL4 patients

[56, 57, 155, 156]. Consistent with other reports, we observed 30-50% depletion of mtDNA in patient fibroblast lines (Fig 3.2 A). In line with these findings, we also examined mitochondrial respiration using the Seahorse analyzer. Basal respiration was reduced by 30-50% in patient fibroblasts (Fig 3.2 B, D). Similarly, a 30-60% decrease in maximal respiratory capacity was measured in FBXL4 patient fibroblasts (Fig 3.2 B, E). Of note, extracellular acidification rate

(ECAR), which can reflect lactate production, showed approximately 30% increase in FBXL4 patient cells compared to controls (Fig 3.2 F). When we examined the expression of several subunits of OXPHOS complexes, we observed noticeably reduced expression of several

OXPHOS subunits in patient cells (Fig 3.3). Finally, we observed approximately 40% reduction in mitochondrial membrane potential and a 20% decrease in mitochondrial mass in both FBXL4 patient fibroblast lines (Fig 3.4).

Figure 3.2. Characterization of FBXL4 patient fibroblasts, and response to DCA treatment. A) QPCR analysis of mtDNA copy number from control and patient fibroblasts normalized to 18S. Data represents at least three independent biological replicates. B) Traces of oxygen consumption rates (OCR) over time analyzed in control and FBXL4 patient fibroblasts using the Seahorse XF24 extracellular flux analyzer. C) Traces of OCR of cells as in B following 24hrs treatment with 10 mM DCA. D) Basal respiration in control and patient cells calculated from B and C. Data is presented as % control from 3 independent biological replicates. E) Maximal respiratory capacity in control and patient fibroblasts calculated from B & C. Data is presented as % control from 3 independent biological replicates. F) Extracellular acidification rate (ECAR) was recorded for control and patient samples treated with DCA. Data represents ECAR recordings prior to drug injections. Bioenergetic profiles were assessed from at least three technical replicates per condition. P-values were determined by an unpaired student t-test, compared to the untreated control, unless otherwise indicated.

Figure 3.3. Characterization of OXPHOS subunits in FBXL4 patient fibroblasts, and response to DCA treatment. A) Western analysis of subunits of oxidative phosphorylation complexes in control and FBXL4 patient fibroblasts treated as indicated. MtDNA-encoded COXII of complex IV is marked with an asterisk. All other subunits are nuclear-encoded. B-actin was used as a load control. B-F) Relative quantification of Complex V subunit (ATP5A) (B), Complex III subunit (UQCRC2) (C), Complex II subunit (SDHB) (D), the mtDNA-encoded subunit of Complex IV (COXII) (E) and Complex I subunit (NDUFB8) (F), in control and patient cells treated as indicated. Data represents three independent biological replicates and is presented as %control.

Figure 3.4. A) Mitochondrial membrane potential in fibroblasts from patients and an unaffected control was assessed by flow cytometry following staining with a membrane-potential sensitive dye (TMRE). B) Mitochondrial mass was assessed in control and FBXL4 patient fibroblasts by flow cytometry following MitoTracker Green staining. P-values were determined by an unpaired student t-test, compared to the control.

Next, we assessed mitochondrial network morphology in FBXL4 patient cells. Consistent with previous reports, cells harbouring the C584R FBXL4 mutation display noticeably fragmented mitochondrial networks (Fig 3.5 A, B). Additionally, we examined the levels of core mitochondrial fusion proteins (MFN1, MFN2, OPA1) in control and patient fibroblasts, but did not observe consistent changes in protein expression (Fig 3.5 C). However, we noted significant changes in the appearance and distribution of mtDNA nucleoids in patient cells, a feature that has been described previously in FBXL4 patient fibroblasts, but not quantified. We performed detailed quantitation of nucleoid morphology and found that on average nucleoids were ~40% larger in patient cells compared to control (Fig 3.5 D). Meanwhile the total number of nucleoids in patient cells was reduced by ~50% in patient cells (Fig 3.5 E). Qualitatively, mtDNA nucleoids showed a perinuclear distribution in FBXL4 fibroblasts, compared to a more evenly distributed nucleoids in control fibroblasts (Fig 3.5 A). In addition, patient cells contained many mitochondrial fragments devoid of mtDNA (Fig 3.5 A), which is not typical in control cells. Thus, our data indicate that the C584R variant in FBXL4 is pathogenic.

Figure 3.5. Mitochondrial network and nucleoid alterations in C584R FBXL4 patient fibroblasts. A) Representative confocal images of control and patient cells stained with MitoTracker Red and PicoGreen (dsDNA: nuclear and mtDNA). Arrowheads in zoomed images highlight mitochondrial fragments devoid of mtDNA. Scalebars: 10µm. B) Quantification of mitochondrial morphology from control and FBXL4 patient fibroblasts treated with DCA (10mM,

24hrs) or a vehicle control (DIH2O). Morphology analysis was performed on fixed cells stained with TOMM20 from three technical replicates, quantifying at least 50 cells per fibroblast line per replicate. C) Western Analysis of mitochondrial fusion proteins (MFN1, MFN2, OPA1) and the main fission GTPase (DRP1) in control and patient fibroblasts treated as indicated. HSP60 and β-tubulin were used as mitochondrial and cytosolic load controls, respectively. D) Quantification of mtDNA nucleoid counts in control and patient fibroblasts from cells stained as in A. E) Quantification of mtDNA nucleoid sizes in cells stained as in A. Data represents average nucleoid sizes and counts from 10 cells for each group. P-values were determined by an unpaired student t-test, compared to the control fibroblasts.

3.2.5 DCA reduces lactate production, but does not correct mitochondrial function

Given the reported improvements in patient symptoms following DCA administration, we examined whether DCA could reverse the mitochondrial dysfunction we observed in FBXL4 patient fibroblast lines. DCA treatment (10mM, 24hrs) was well-tolerated in control and patient fibroblasts. Notably, while DCA treatment led to a significant reduction of ECAR to control levels in both patient fibroblast lines (Fig 3.2 F), DCA failed to rescue mtDNA depletion or restore bioenergetic profiles (Fig 3.2 A-E). In addition, though DCA treatment promoted mitochondrial fusion in control cells, there was no change in the morphology for FBXL4 patient cells treated with DCA (Fig 3.5 B). Thus, despite the clinical benefits that were observed, while DCA improves ECAR, it does not seem to fix the underlying defects caused by FBXL4 impairment.

3.2.6 Impaired mitochondrial fusion in FBXL4 patient fibroblasts

The structural features observed in FBXL4 patient fibroblasts, including mitochondrial network fragmentation, mtDNA nucleoid clustering and overall mtDNA loss were reminiscent of the phenotypes present in cells lacking fusion [22]. In addition to the fact that DCA treatment promoted mitochondrial fusion in control cells but not in FBXL4 patient cells, our findings led us to test the hypothesis that FBXL4 may promote mitochondrial fusion. To this end, we examined mitochondrial fusion rates in control and FBXL4 patient fibroblasts using a photo-activatable

GFP (PA-GFP) fusion assay [169, 180, 181]. Following photoactivation of a small region of the mitochondrial network in live cells, this assay allows tracking of the diffusion of the GFP signal throughout the network as fusion of mitochondrial membranes and matrix content mixing occur.

Compared to control fibroblasts, in which mitochondria fusion was ongoing, fibroblasts with the

C584R FBXL4 mutation displayed significantly reduced fusion (Fig 3.6). This was quantified by two methods. First, we quantified the percentage of the photoactivated region relative to the total mitochondrial network, and measured the increase in distribution of the GFP signal over time (16 mins) (Fig 3.6 C, Fig 3.7). The initial photo-activation region of the mitochondrial

76 network was comparable in control and patient fibroblasts (approx. 3%). After 16 minutes, the

GFP signal spread to about 25% of the mitochondrial network in control cells. In patient cells, however, the GFP signal spread only to 5% of the network. Secondly, we measured the relative fluorescence intensity of the GFP signal in the photo-activated region over time [169, 180], where loss of signal reflects spread of the GFP signal due to fusion events (Fig 3.6 B, D).

During the period of imaging, we observed that GFP intensity decreased by approximately 80% in control cells, compared to only 20% in FBXL4 patient cells (Fig 3.6 B, D). These data highlight a previously unrecognized mitochondrial fusion impairment in cells harbouring mutant

FBXL4 protein.

Figure 3.6. Mitochondrial fusion is reduced in FBXL4 patient fibroblasts. A) Representative confocal images of live control and patient cells expressing a matrix-targeted PA-GFP and stained with TMRE. White boxes indicated region of photoactivation. Images were acquired pre-photoactivation, 5s after photoactivation and then every 120s, up to 960s (16mins). Scalebars: 10µm. B) Photoactivated region of cells shown in A, zoomed in, and showing initial photoactivate frame (t=5s) and final frame at 960s. C) Quantification of area of photoactivated mitochondrial network at t=5s and the percentage of the network to which the GFP signal had spread at t=960s. Analysis was performed on binarized images, shown in Fig 3.7. Data represent analyses in 5 cells per fibroblast line. P-values were determined by an unpaired student t-test, compared to the control. D) Relative fluorescence intensity over time in the photoactivated region indicated in panel A, at t=0 (pre-photoactivation), t=5s (post- photoactivation) and 120s intervals. Data points represents relative fluorescence intensity in the representative images shown in panel A.

Figure 3.7. Example of binarized confocal images of control and FBXL4 patient cells expressing a matrix-targeted PA-GFP and stained with TMRE as in Figure 3.6. Images shown were at t=5s post-photoactivation (A) and t=960s at the final acquisition frame (B). Images are representative of how PA-GFP signal diffusion was analyzed. The percentage of the network that was photoactivated at t=5s relative to the total network (TMRE signal) as well as the % network containing GFP signal at t=960s relative to the total network in TMRE channel were measured using ImageJ FIJI. Percentages displayed on respective images indicate measurements for these frames.

3.2.7 Overexpression of wildtype FBXL4, but not C584R mutant, promotes mitochondrial fusion

We then asked whether wild-type FBXL4 could promote mitochondrial fusion. To this end, we overexpressed a C-terminal HA-tagged FBXL4 protein in HEK cells and assessed mitochondrial network morphology 24hrs post-transfection. Consistent with a role in promoting fusion, nearly 50% of HEK cells overexpressing wildtype FBXL4 contained fused mitochondrial networks, compared to ~20 % in empty vector-transfected cells (Fig 3.8 A, B). However, overexpression of the C584R mutant FBXL4 did not lead to more cells with hyperfused

79 mitochondrial networks (Fig 3.8 A, B). These findings support a novel role for FBXL4 in promoting fusion, which is abrogated by the pathogenic C584R mutation.

Figure 3.8. FBXL4 overexpression shifts mitochondrial networks into a fused state. A) Representative confocal images of HEK cells transfected with an empty vector, wildtype FBXL4- HA, or C584R-HA, immunolabelled with antibodies against TOMM20 (red) and the epitope tag, HA (green). Scalebars: 10µm. B) Quantification of mitochondrial morphology from cells as in panel A. P-values were determined by an unpaired student t-test, compared to the percentage of cells with a fused morphology in the indicated lines.

3.2.8 Restoring mitochondrial morphology in FBXL4 fibroblasts rescues mtDNA depletion

In order to determine whether impaired mitochondrial fusion is the cause of mtDNA depletion in FBXL4 patient fibroblasts, we asked whether restoring the mitochondrial network was able to replenish levels of mtDNA. Control and FBXL4 fibroblasts were treated with the mitochondrial fission inhibitor mdivi-1 [182] (25 µM) for 1 day and for 7 days, which improved

80 mitochondrial morphology in patient fibroblast lines (Fig 3.9, Fig 3.10). As picogreen staining can be influenced by supercoiling [183], we imaged nucleoids in fixed cells using and anti-DNA antibody in order to get a more accurate analysis of mtDNA nucleoids. In addition, we used a microscope with a higher resolution and performed a z-stack in order to look at all of the nucleoids within each cell, rather than just through a single cross section of the cell. As such, the size and number of nucleoids is different than with the previous picogreen analysis (Fig 3.5).

However, the trends remained the same, with both patient fibroblast lines exhibiting fewer but larger mtDNA nucleoids than control (Fig 3.9). However, upon mdivi-1 treatment, we observed a rescue in the number of mtDNA nucleoids via microscopy in FBXL4 fibroblasts (Fig 3.9 D), consistent with the increased mtDNA copy number as assessed by qPCR. Unexpectedly, we did not observe any statistical changes in the average size of mtDNA nucleoids following mdivi-

1 treatment (Fig 3.9 E).

Figure 3.9. Restoration of fused mitochondrial networks in FBXL4 patient fibroblasts rescues mtDNA copy number. A) Representative confocal images of control and patient fibroblasts treated with mdivi-1 or DMSO for 7 days, fixed and immunolabeled with antibodies against TOMM20 (red) and DNA (green). Scalebars: 10µm. B) Mitochondrial morphology analysis of control and patient cells treated with mdivi-1 or a vehicle control for 7 days. Data represents two independent biological replicates. C) Relative mtDNA copy number in control and patient cells treated with mdivi-1 or a vehicle control for 7 days. D) Average mtDNA nucleoid counts from cells treated as indicated, fixed and immunolabeled with an anti-DNA antibody. E) Average mtDNA nucleoid sizes from cells as in D. P-values were determined by an unpaired student t-test, comparing each group to the vehicle control-treated control fibroblasts.

Figure 3.10. Mdivi-1 treatment (25uM, 24hrs) in FBXL4 patient fibroblasts does not rescue mtDNA copy number. A) Relative mtDNA copy number in control and patient cells treated with mdivi-1 or a vehicle control for 24hrs. P-values were determined by an unpaired student t-test, comparing each group to the vehicle control-treated control fibroblasts.

3.3 Discussion

Mutations in the nuclear-encoded mitochondrial protein, FBXL4, are recognized to cause mtDNA depletion syndrome (over 50 mutations reported to date) [56, 57, 153-158, 172-174].

However, prior to this study, the molecular function of FBXL4, and the mechanism leading to mtDNA depletion were unknown. Here, we report a previously uncharacterized variant in FBXL4

(c.1750T>C; p.C584R) in two siblings presenting with early onset multisystemic defects including encephalomyopathy, lactic acidosis and cardiac hypertrophy. Characterization of patient-derived fibroblasts showed severe metabolic deficiencies, fragmented mitochondrial networks, fewer mtDNA nucleoids that were enlarged, as well as mtDNA depletion. Collectively, these phenotypes are consistent with previously reported pathogenic FBXL4 mutations [56, 57,

156] and indicate that the C584R variant is indeed pathogenic.

Given that there are many pathogenic mutations throughout the FBXL4 protein that have been described to cause the same patient phenotype [173, 174], it is likely that they act as loss of function mutations. The C584R mutation investigated here lies within the leucine rich repeat

(LRR) domain of FBXL4. LRR domains are known to provide a structural basis for protein- protein interactions and have been implicated in a diverse array of functions [184]. Given the amino acid change from cysteine, with a polar side chain containing a reactive sulfhydryl group that can form disulfide bonds, to arginine, with a basic positively charged side chain, it is possible that the structure of the LLR is altered such that is prevents normal protein-protein interactions required for FBXL4 to promote mitochondrial fusion.

Congenital lactic acidosis has been reported in all patients carrying pathogenic variants in FBXL4 to date [56, 57, 153-158, 172-174]. However, there have not been any reports of effective interventions for lactic acidosis in these patients. Notably, DCA was administered to P2 at 6 weeks of age. In response to DCA treatment venous lactate levels were reduced, and a remarkable reversal of the cardiac hypertrophy was noted. Importantly, the prolonged survival for P2 (2 years and 10 months) is also notable, compared to P1 carrying the same mutation, who passed at 7 months of age. Thus, we set out to leverage these positive clinical outcomes and ask whether DCA could reverse mitochondrial dysfunction in FBXL4 patient fibroblasts.

Consistent with the known mechanism of action for DCA, extracellular acidification was reversed in patient cells. However, none of the other parameters of mitochondrial dysfunction we investigated were improved by DCA treatment. This finding suggests that the clinical benefits reported for DCA encompass managing only the lactic acidosis, while the underlying mechanism causing mitochondrial dysfunction was not resolved by DCA. The fact that DCA also reversed the cardiovascular dysfunction in P2, suggests that the increase in lactate specifically was most likely responsible for the cardiovascular dysfunction in this patient.

Though rare, normalization of cardiac dysfunction in mitochondrial disease patients has been reported previously [185]. Notably, in a patient with lactic acidosis and hypertrophic cardiomyopathy, who harbored pathogenic mutations in the mitochondrial translation optimization 1 protein (MTO1) [186, 187], DCA treatment was also beneficial, and correlated

84 with reduced lactate levels and reversal of cardiac dysfunction. A continued regimen of DCA and co-factors was suggested to have prolonged survival in this MTO1 patient into adulthood

[187]. In these rare cases, restoration of cardiac function occurs in patients surviving past one year of age. Together with our FBXL4 patient, these observations suggest that there is a critical developmental period early in life, where high levels of lactate are detrimental to cardiac function.

While DCA has been used to treat mitochondrial disease in the past, patients’ responses have been extremely variable, and DCA can even be deleterious. Positive clinical outcomes following DCA treatment have been reported in several studies involving patients with lactic acidosis [186-190]. Chronic DCA administration, in patients with congenital causes of lactic acidosis (deficiencies of PDHC, Respiratory chain complex I, IV or I+IV), into adulthood has been reported to be tolerable and beneficial at maintaining blood lactate levels [186, 187, 190].

Meanwhile, a clinical trial examining the efficacy of DCA in mitochondrial myopathy, encephalopathy, lactic acidosis and stroke-like episodes (MELAS) patients was prematurely terminated due peripheral nerve toxicity in 17/19 participants [191]. As such, clinical use of DCA has declined and it has orphan drug status [192, 193]. The variability in outcomes of mitochondrial disease patients treated with DCA likely reflects the genetic and phenotypic heterogeneity of mitochondrial diseases in general. To this end, we argue that the use of DCA to manage elevated blood lactate should be practiced with caution, considering the underlying pathogenesis mechanism leading to lactatemia. Importantly, DCA treatment may be a viable option to manage symptoms in mitochondrial disease patients presenting with pathogenic variants in FBXL4.

In order to understand how mutations in FBXL4 cause mtDNA depletion, we sought to investigate the molecular function of FBXL4 in mitochondria. Considering the mechanisms known to lead to mtDNA depletion, we reasoned that FBXL4, proposed as an IMS protein [57],

85 would be unlikely to directly regulate replication and maintenance of mitochondrial genomes localized in the matrix [6]. Moreover, the mitochondrial fragmentation and enlarged mtDNA nucleoids that we and others observed in FBXL4 patient fibroblasts were particularly intriguing

[56, 57, 156]. Together with the mtDNA depletion phenotype, these features have also been reported in cells lacking mitochondrial fusion [22, 40, 116, 117]. Additionally, the fact that DCA promoted fused mitochondrial networks in control fibroblasts, but not in FBXL4 patient cells, also suggested an impairment in the dynamic regulation of mitochondrial structure. Hence, we speculated that FBXL4 had a previously unrecognized role in promoting mitochondrial fusion.

Our study is the first to demonstrate that FBXL4 promotes mitochondrial fusion. We found that mitochondrial fusion rates were greatly reduced in FBXL4 patient fibroblast cells, which explains the fragmented mitochondrial networks and enlarged mtDNA nucleoids in these cells. In fact, our observations in FBXL4 patient fibroblasts were comparable to impaired fusion in MFN2 knockout mouse embryonic fibroblasts described previously [181]. Moreover, overexpression of FBXL4 led to a marked increase in the number of cells with reticular mitochondrial networks, demonstrating that FBXL4 promotes mitochondrial fusion. Finally, the

C584R mutation abrogates the pro-fusion activity of FBXL4. Taken together, these results demonstrate a novel role for FBXL4 in promoting mitochondrial fusion. The exact mechanism through which FBXL4 protein promotes fusion is still an open area for investigation. Notably, it is intriguing that the total amounts of fusion proteins (MFN1/2 and OPA1) are largely unaltered in

FBXL4 patient cells compared to control, which suggests that the lower fusion rates are not due to reduced stability of fusion GTPases, but by some other mechanism. Future work studying how FBXL4 promotes fusion could lead to new approaches to mediate mitochondrial dynamics as a therapeutic approach for a growing list of pathologies where mitochondrial fragmentation is implicated [194-196]

Our data also suggest that the reduction in mitochondrial fusion is the most likely explanation for why FBXL4 impairments causes mtDNA depletion and disease. When we restored the mitochondrial network to a more fused state in patient fibroblasts, we observed a rescue of both the mtDNA copy number and the number of mtDNA nucleoids. If the mtDNA depletion had been due to reduced levels of mitochondrial dNTP nucleotide pools, then mdivi-1 treatment should not have rescued the mtDNA copy number. Our finding that restoring the network morphology rescued mtDNA copy number in FBXL4 fibroblasts is consistent with recent work showing that a fused network is important for proper distribution of the mtDNA replisome components in order to maintain mtDNA copy number [20]. It is also notable that restoring the mitochondrial network does not cause an immediate rescue to the mtDNA copy number, suggesting it takes more than 1 day for copy number to be re-established following restoration of the mitochondrial network.

Unexpectedly, we did not see a rescue of the mtDNA nucleoid size suggesting that there may not be a direct correlation between nucleoid clumping and mtDNA copy number. However, this lack of rescue could be due to the fact that large nucleoids persist longer than 7 days.

Alternatively, mdivi-1 treatment has been shown to induce nucleoid clumping [24], which is potentially a confounding factor. Thus, at this juncture we cannot make any definitive conclusions regarding the correlation between mtDNA nucleoid clumping and mtDNA depletion.

Nonetheless, it is intriguing that loss of the mitochondrial fission protein Drp1 has also been shown to lead to larger mtDNA nucleoids and mtDNA depletion [24, 197]. Meanwhile, a pathogenic mutation in MYH14, encoding a myosin protein recently implicated in mitochondrial fission, also leads to hyperfused mitochondrial networks and larger mtDNA nucleoids [198]. In this context, recent work suggests that mitochondrial fission is important to segregate newly synthesized mtDNA nucleoids [23, 199]. However, exactly how loss of fusion relates to clumping of mtDNA nucleoids remains unresolved (Fig 3.11).

Figure 3.11. Proposed model for mtDNA depletion and altered nucleoids distribution in cells with impaired mitochondrial dynamics regulation. Under normal conditions, balanced fusion and fission events ensure even distribution of mtDNA nucleoids throughout reticular mitochondrial networks or fragmented puncta. Notably, mitochondrial fission at sites of mtDNA replication facilitates segregation of newly synthesized mtDNA, thus maintaining mtDNA copies throughout the network. An imbalance in mitochondrial dynamics arising from impaired fusion (e.g. in mutant FBXL4, OPA1 or MFN2) leads to excessive fragmentation of the mitochondrial network, clustering of mtDNA nucleoids and mtDNA depletion. Meanwhile, cells with defective fission (e.g. mutant DNM1L), mtDNA nucleoids form large aggregates likely due to lack of efficient segregation following replication. Impaired fission also leads to mtDNA depletion, however the exact mechanisms are poorly understood.

Collectively, our data highlight impaired mitochondrial fusion as the most likely mechanism underlying pathogenic mtDNA depletion caused by mutations in FBXL4. Identifying

FBXL4 as a novel protein promoting mitochondrial fusion adds mechanistic insight into the field of mitochondrial dynamics, particularly given the importance of mitochondrial fusion in health and disease. Thus, future studies are warranted in order to develop a comprehensive mechanistic understanding of FBXL4 as a pro-fusion factor. Finally, though our research suggests that DCA may be an effective treatment option to improve elevated lactate levels in patients with mutations in FBXL4, DCA does not restore FBXL4 function. Nonetheless, novel therapeutic approaches to restore mitochondrial morphology may be beneficial for patients with

FBXL4 dysfunction.

MSTO1 mutations cause mtDNA depletion, manifesting as muscular dystrophy with cerebellar involvement.

This chapter is based on a published manuscript with permission from Donkervoort et al., 2019 (see Appendix D).

4.1 Introduction

Mitochondria maintain and express their own genome (mtDNA), typically present in 100-

1000 copies per cell and organized into nucleoprotein structures known as nucleoids [124, 200].

The mtDNA encodes some subunits of the oxidative phosphorylation (OXPHOS) machinery that are essential for mitochondrial respiration and ATP production [2]. The relative amount of mtDNA per cell varies in a tissue-specific manner, as an adequate number of mtDNA copies must be maintained to support aerobic respiration and meet cellular energetic demands [138].

As such, reduction in the total amount of mtDNA clinically manifests as severe multi-systemic abnormalities to which energy demanding organs, such as the brain and muscles, are particularly susceptible.

MtDNA depletion syndromes are a clinically and genetically diverse class of mitochondrial diseases characterized by a reduction of mitochondrial genomes [114, 201]. Of the 15 formally defined mtDNA depletion syndromes listed in The Online Mendelian Inheritance in Man database (OMIM) [171], most are caused by pathogenic variants in proteins that are required for mtDNA replication (POLG, C10orf2, MGME1, and TFAM) or those that are necessary to maintain mitochondrial deoxyribonucleoside triphosphates (dNTP) pools (TK2,

DGUOK, RRM2B, TYMP, SUCLA2, SUCLG1, AGK, MPV17 and SLC25A) [113, 114]. However, pathogenic variants in proteins that regulate the processes of mitochondrial fusion and fission have also been linked to mtDNA depletion. These include the fusion protein OPA1 (OMIM

616896), as well as MFN2 and DNM1L, which are essential for fusion and fission, respectively

[115-117, 119, 175].

While defects in mtDNA replication or impaired maintenance of mitochondrial dNTP pools are expected to lead to mtDNA depletion, it is less clear how mitochondrial fusion and fission are involved in the regulation of mtDNA. Nonetheless, abnormalities in mtDNA integrity and nucleoid distribution have been demonstrated in several models of defective fusion.

Mitochondrial membrane fusion is orchestrated by the activity of large GTPases, including

Mitofusin 1 and 2 (MFN1, MFN2) localized to the outer membrane, and Optic Atrophy 1 (OPA1) in the inner mitochondrial membrane. In yeast, cells lacking the Mitofusin homolog Fzo1p suffer from complete loss of mtDNA [202]. Meanwhile, knockout fibroblasts for OPA1, MFN1 and/or

MFN2, display fragmented mitochondrial networks with altered nucleoid distribution whereby some fragments are devoid of mitochondrial genomes [22]. Finally, in humans, pathogenic variants in MFN2 and OPA1 cause severe mtDNA depletion and multiple mtDNA deletions [115-

117]. Patients with pathogenic variants in MFN2 or OPA1 have been reported to present with a phenotype of early-onset ataxia, hypotonia, axonal sensorimotor neuropathy, optic atrophy and hearing loss [105, 106, 115-117, 203].

Recently, MSTO1 was described as a cytosolic mitochondrial fusion protein, and pathogenic variants in MSTO1 have been reported to cause ataxia, muscle weakness, cerebellar atrophy and pigmentary retinopathy [61, 62, 112, 204]. Notably, these phenotypic features have all been previously reported in mtDNA depletion syndromes [105, 106, 114]. To date, five independent studies have described MSTO1 variants in 12 patients from seven families [61, 62, 112, 204, 205]. While pathogenic MSTO1 variants have been linked to impairments in mitochondrial fusion, the consequence of these variants on mtDNA maintenance and its associated clinical spectrum has not been studied extensively.

In this study, we present an extensive phenotypic characterization of 15 new patients from 12 families harbouring a broad array of bi-allelic pathogenic variants in MSTO1 confirming a remarkably consistent, and ultimately recognizable clinical phenotype. Additionally, in seven cultured fibroblasts from patients with MSTO1-related disease, we demonstrate loss of MSTO1 protein, significantly fragmented mitochondrial networks, enlarged lysosomal vacuoles, depletion of mtDNA, and alterations to mtDNA nucleoids. Therefore, we demonstrate that bi- allelic loss-of-function variants in the mitochondrial fusion protein MSTO1 impair mtDNA

91 maintenance and fusion and result in mtDNA depletion in fibroblasts, which we establish is associated with a remarkably consistent clinical spectrum of MSTO1-deficiency.

4.2 Results

4.2.1 Human Subjects and Samples

Patients were ascertained through their local neurology and genetics clinics. F2 was identified through GeneMatcher [206]. Written informed consent and age-appropriate assent for study procedures and photographs were obtained by a qualified investigator (protocol 12-N-

0095 approved by the National Institute of Neurological Disorders and Stroke, National Institutes of Health, Research Ethics Board of the Hospital for Sick Children, REB # 1000009004: SCHN

Human Ethics Committee 10/CHW/45, University of Calgary Conjoint Health Research Ethics

Board). Medical history was obtained and clinical evaluations, including brain MRI and muscle biopsy, were performed as part of the standard diagnostic evaluation. Muscle biopsy slides and available electron microscopy images (EM) were reviewed by investigators. DNA, muscle and skin biopsy samples were obtained according to standard procedures.

4.2.2 Clinical Characteristics

The clinical presentation of the 15 patients, which includes nine females and six males, is summarized in Table 4.1 (Appendix A), with ages ranging from 6-52 years at the time of most recent examination. There were two families with more than one affected relative: Family 1 (F1) consists of three affected sisters (P1, P2, P3; p.(Asp236His); p.(Arg279His)), and Family 2 (F2) consists of two affected siblings (P4, P5; p.(Gly420ValfsX2); p.(Arg256Glu)). Family history was non-contributory in the remaining patients of the cohort.

Patients typically presented with hypotonia and delayed motor milestones with first symptoms recognized between birth and three years of age. All but five patients presented with delayed motor milestones and then subsequently achieved independent ambulation with gaits

92 characterized as waddling-like and wide-based across the entire cohort. All 15 patients reported either relatively slow progression or no progression of their muscle weakness. At the time of the last examination, all were found to have predominantly proximal weakness. Cerebellar symptoms manifested as dysmetria in 11, gait ataxia in nine and abnormal speech in 12, which included a history of speech delay in five and dysarthria in eight patients, respectively.

Corticospinal tract manifestations including increased tone, the presence of a spastic catch, clonus and/or increased deep tendon reflexes were observed in eight patients. A history of learning difficulties was reported in nine patients while none of the patients were found to have major cognitive involvement. None of the patients had a history of seizures, cataracts, hearing or cardiac involvement.

Electrophysiological studies were available for four patients. Three (P4: p.(Gly420ValfsX2); p.(Arg256Glu), P5: p.(Gly420ValfsX2); p.(Arg256Glu), P10; p.(Asp236Gly); p.(Arg279His)) of the four had evidence of myotonia on EMG, and complex repetitive discharges were present in P8 (p.(Phe217Leu); p.(Asp236His)). Lower extremity muscle MRI imaging was available for four patients and revealed a spectrum of involvement ranging from severe involvement of the muscles of the upper leg with apparent fatty replacement of all muscles except for the semimembranosus and the biceps femoris (P6; p.(Asp236His); p.(Phe217Leu)) to mild fatty infiltration of upper and lower leg muscles (P10; p.(Asp236Gly); p.(Arg279His)) (Table 4.1 (Appendix A) & Fig 4.1 a). Pulmonary function testing was performed in nine patients and revealed reduced forced vital capacity (FVC) measurements, ranging from

60% to 79% predicted. Serum creatine kinase (CK) levels were significantly increased in all patients, except for P15, ranging from 300 to 5000 U/L. Echocardiogram was normal in the eleven patients who underwent this screening evaluation.

Figure 4.1. Muscle and brain imaging in MSTO1 patients. (a) Lower extremity muscle MRI of patients P6 (p.(Asp236His); p.(Phe217Leu)), P7 (p.(Leu450Phe); deletion), P8 (p.(Phe217Leu); p.(Asp236His)) and P10 (p.(Asp236Gly); p.(Arg279His)) at ages 37 years, 9 years, 6 years and 16 years, respectively. Abnormal signaling of muscles such as the posterior gastrocnemius muscle in patient P6 (white arrow), reflects muscle breakdown with replacement with adipose tissue (b) Brain MRI completed in twelve patients consistently demonstrates moderate-to-severe cerebellar volume loss involving the vermis and both hemispheres. Repeat MRI images available in patients P4 (p.(Gly420ValfsX2); p.(Arg256Glu)), P8 (p.(Phe217Leu); p.(Asp236His)), P10 (p.(Asp236Gly); p.(Arg279His)) and P12 (p.(Arg345His); p.(Thr324Ile)) demonstrate mild (P12: p.(Arg345His); p.(Thr324Ile)) to no progression (P4: p.(Gly420ValfsX2); p.(Arg256Glu), P8: p.(Phe217Leu); p.(Asp236His) and P10: p.(Asp236Gly); p.(Arg279His)) of cerebellar volume loss over time.

Overall, we recorded a remarkably consistent clinical phenotype of primary motor developmental delay, fairly stable mostly proximal muscle weakness caused by a dystrophic myopathy, mild cerebellar findings of dysmetria, ataxia and dysarthria based on a stable congenital cerebellar atrophy, mild pyramidal signs and evidence for some degree of speech delay and learning disability in some. Meanwhile, major cognitive involvement, seizures, retinopathy, optic atrophy or hearing loss were not seen.

4.2.3 Neuroimaging Characteristics

Brain MR imaging was available for 11 patients and consistently showed moderate to severe cerebellar atrophy/hypoplasia involving the vermis and both hemispheres in all ten (Fig

4.1 b). Four of the brain MRIs had been performed in patients before age two years, which revealed significant decrease in cerebellar volume. A lack of progression of cerebellar volume loss was confirmed through repeat imaging available in four patients. Patient 12 (p.(Arg345His); p.(Thr324Ile)) showed mild progression of cerebellar volume loss between ages one and six years.

4.2.4 Muscle Histopathology and Electron Microscopy

Muscle biopsies were performed in ten patients and were consistent with a dystrophic process with evidence of variation in fiber type size, a mild degree of necrosis and regeneration, internalized nuclei and whorled fibers (Fig 4.2 a). Electron microscopy analysis was performed in three patients. Aggregates of subsarcolemmal mitochondria were noted in patients P9

(p.(Tyr478Cys); missing) and P10 (p.(Asp236Gly); p.(Arg279His)). There was also evidence of non-specific mitochondrial morphologic abnormalities (variations in mitochondrial shape and size) seen in patients P9 and P10 (Fig 4.2 b).

Figure 4.2. Muscle biopsy, MSTO1 pathogenic variants and pedigrees. (a) Histology findings from the vastus lateralis muscle biopsy of P7 (p.(Leu450Phe); deletion) at age 20 months include internalized nuclei on hematoxylin and eosin (H&E) staining (white arrow) (i) and variation in fiber size on nicotinamide dinucleotide (NADH) staining (ii) and whorled fibers evident on Gömöri trichrome (inset) (iii) and COX staining (white arrow) (iv). (b) Muscle biopsy electron microscopy (EM) findings are notable for aggregates of subsarcolemmal mitochondria in both P9 (p.(Tyr478Cys); missing)) (i and ii) and P10 (p.(Asp236Gly); p.(Arg279His)) (iii and iv) and non-specific mitochondrial morphologic abnormalities (variations in mitochondrial shape and size) (black boxes) in P10. (c) Schematic of new and reported human MSTO1 pathogenic variants. Shown in numbered light blue squares are cDNA exons (RefSeq isoform NM_018116.3 of MSTO1). Corresponding known protein domains are shown in orange (tubulin 3 domain) and beige (Misato segment II tublin-like domain). Variants written in black text are recessive; the single mutation in red has been previously reported to cause dominantly inherited MSTO1-related disease. The top half of the figure depicts novel variants reported in this publication; the bottom half of the figure depicts variants which have been previously reported. Bolded variants depict previously reported mutations that were also present in our cohort. The dotted line depicts a large deletion (exons 9-14). (d) Pedigree of two families consistent with recessive inheritance of MSTO1 pathogenic variants.

4.2.5 Molecular Results

Using whole exome sequencing (WES), we identified compound heterozygous missense variants in MSTO1 (NM_018116.3) in 15 patients from 12 independent families consistent with the bi-allelic recessive mode of inheritance (Fig 4.2 c). Parental DNA for segregation testing was not available for P13, who was found to be apparently homozygous for the common p.(Phe217Leu) variant. Six of these missense variants have not yet been reported. The p.(Gly420ValfsX2) frameshift variant was recently reported [204]. There were three apparent recurring mutation hotspots (p.Asp236, p.Arg279, p.Phe217), which were identified as five, four and four independent alleles, respectively. The p.(Arg279His) variant had been previously reported in heterozygosity with a second pathogenic allele in three families [112, 204].

Trio WES identified an apparently homozygous p.(Leu450Phe) MSTO1 pathogenic variant in P7. Targeted sequencing confirmed that this variant was paternally inherited, while the mother was found to be negative for the variant. Subsequent testing using exon-level oligo CGH array identified a presumed maternally inherited deletion in P7 encompassing at least exons 9-

14 of the MSTO1 gene and extending to include both the YY1AP1 as well as the DAP3 genes

(Genomic Coordinates: arr[GRCh37] 1q22(155582110_155708204)x1). Recessive pathogenic variants in YY1AP1 have been reported in association with Grange syndrome (OMIM 607860).

DAP3 has not yet been associated with human disease.

Trio WES for P9 identified a heterozygous p.(Tyr478Cys) MSTO1 variant inherited from a clinically unaffected father. Whole genome sequencing (WGS) in P9 confirmed this heterozygous missense variant but did not identify a second MSTO1 allele in compound heterozygosity. Exon-level oligo CGH testing for P9 was normal. RNA sequencing (RNA-seq) analysis to identify any possible transcriptional aberrations in MSTO1 for P9 was inconclusive, and no splice aberrations were identified in P9. Attempts to evaluate allele balance at the hg19:

97 chr1:155583319 variant via RNA-seq were unsuccessful due to insufficient coverage in the region, likely due to the presence of a highly homologous pseudogene.

Parental segregation testing for all MSTO1 variants was consistent with bi-allelic recessive inheritance, except for P9 (p.(Tyr478Cys); missing) in whom the presumed maternally inherited allele has not yet been identified, and for P13 (p.(Phe217Leu); p.(Phe217Leu)) in whom parental DNA was not available. Variants identified were predicted to be damaging and either absent or extremely rare (allele frequency below 0.00005) in Genome Aggregation

Database (GnomAD) and Exome Aggregation Consortium (ExAC) except for the p.(Arg279His) variant. This particular variant was listed with an allele frequency of 0.00019 in ExAC and

0.00026 in GnomAD with one reported homozygous individual [207]. MSTO1 variants are scattered throughout the gene and do not seem to cluster in a specific MSTO1 domain (Fig.

2c).

4.2.6 Characterization of Patient Fibroblasts

In order to further investigate the effect of MSTO1 mutations, we examined the expression of MSTO1 protein by immunoblotting. As the MSTO1 antibody we used detects multiple bands in fibroblast cells, we confirmed the size of the correct band, which migrates at the predicted size of 62 kDa, corresponding to MSTO1 protein in HeLa cells overexpressing a

V5-epitope-tagged-MSTO1 or an empty vector (Fig 4.3 a). Our data show that the MSTO1 protein is undetectable in all patient fibroblasts, (n=7) suggesting that pathogenic variants affect protein expression and/or stability (Fig 4.3 a). Although this observation was expected in cells with large deletions encompassing the MSTO1 gene (e.g. P7; p.(Leu450Phe); deletion)), it is intriguing that all patient fibroblasts containing combinations of missense mutations exhibit a similar cellular phenotype. It is also worth noting that MSTO1 variants do not have gross effects on the expression of mitochondrial fusion proteins (MFN1/2 and OPA1) (Fig 4.3 b).

Figure 4.3. Pathogenic variants lead to MSTO1 protein instability. (a) Western blot analysis of total cell lysates from control and patient fibroblast. As a control, total cell lysates from HeLa cells overexpressing MSTO1-V5 or empty vector were also included. Blots were probed with antibodies against endogenous MSTO1, VDAC1, HSP60 and V5. Black arrow corresponds with endogenous MSTO1 protein further verified in HeLa cell lysates, meanwhile bands underneath are nonspecific. (b) Western blot analysis of cell lysates as in (a). Blots were probed against fusion proteins (Mfn1, Mfn2 and Opa1) and loading controls.

Given the established role of MSTO1 as a mitochondrial fusion regulator [61, 62], we examined mitochondrial morphology in seven MSTO1 patient fibroblasts. Consistent with previous reports, visualized mitochondrial networks in patient fibroblasts were fragmented compared to control fibroblasts (Fig 4.4). In addition, we examined lysosomal structures in

MSTO1 patient fibroblast lines by immunofluorescence, as mitochondrial dysfunction can also

99 cause alterations to lysosomes [208]. Compared to unaffected control cells, we observed markedly enlarged lysosomal vacuoles across all MSTO1 patient lines (Fig 4.5), an observation that has not been previously reported in the context of MSTO1 dysfunction [61, 62, 112].

Figure 4.4. Characteristics of MSTO1 patient fibroblasts. (a) Representative confocal microscopy images of control and patient cells. Mitochondrial networks in MSTO1 patient cells are more fragmented and contain fewer but larger mtDNA nucleoids compared to the control cells. Live cells were stained with MitoTracker Red (Red, mitochondria) and PicoGreen (Green, nuclear and mitochondrial DNA). (b) Quantification of mitochondrial morphology from control and patient cells performed from three independent replicates. Statistical analysis was performed on the number of cells with partly fragmented mitochondrial morphology in control versus patient cells; Student T-test, *p < 0.05, **p < 0.001.

100

Figure 4.5. Enlarged lysosomal vacuoles in MSTO1 patient fibroblasts. (a) Representative confocal images of control and patient cells fixed and stained with antibodies against TOMM20 (Red, mitochondria) and LAMP1 (Green, lysosomes). Compared to an unaffected control, patient cells contain distinct lysosomal clusters. (b) Quantification of cells containing enlarged lysosomes in control and patient fibroblasts performed from two independent replicates. Statistical analysis was performed; Student T-test, *p < 0.05.

101

As pathogenic variants in mitochondrial fusion proteins MFN2 and OPA1 have been shown to cause mtDNA depletion [105, 116], we analyzed mitochondrial genomes in patient cells. We observed a significant reduction in mtDNA copy number across all fibroblast lines, ranging from 30-70% depletion (Fig 4.6 a). MtDNA nucleoids are nucleoprotein assemblies involved in the organization and segregation of mtDNA. While examining the size and distribution of mtDNA nucleoids within the mitochondrial network in MSTO1 patient fibroblasts, we found that the patient fibroblasts contained fewer nucleoids, which were larger in size compared to control lines (Fig 4.6 b & c). Notably, several mitochondrial fragments were devoid of mitochondrial genomes in patient cells (Fig 4.4 a), a phenotype previously reported in cells lacking fusion regulation [22]. Together, these observations demonstrate significant alterations of the mitochondrial genome in all patient fibroblast lines evaluated. Unfortunately, muscle tissue was not available for further mtDNA content studies.

102

Figure 4.6. Pathogenic variants in MSTO1 are linked to mtDNA depletion. (a) Relative mtDNA copy number normalized to the nuclear-encoded 18S gene. Data represents at least three independent biological replicates. (b) Analysis of mtDNA nucleoid counts per cell from 35 cells for each group). (c) Quantification of nucleoid sizes in control and patient cells. Data represents average nucleoid sizes from the same cells as in (b). Average mtDNA nucleoid size is presented in a violin plot. K-S test was performed to determine statistical significance. (d) Frequency of nucleoids larger than 0.2µm2 in all 35 cells quantified per fibroblast line. Student T-test was performed as indicated for (a), (c) and (d). *p < 0.05, **p < 0.01, *** p < 0.0001.

The similarity and consistency of the cellular phenotypes described across all seven

MSTO1 patient fibroblast lines strongly supports the notion that loss of MSTO1 function is the underlying cause responsible for these observations. In order to further confirm that the cellular phenotypes were in fact due to the loss of MSTO1, we transiently expressed wild-type MSTO1 in two of the patient cell lines (P4 and P7) (Fig 4.7 a, b). Similar to previous reports [61], we found that expression of wild-type MSTO1 restored mitochondrial morphology after 48 hours

(Fig 4.7 c). Notably, we also observed more fused mitochondrial networks in control cells overexpressing MSTO1, further validating the role of MSTO1 in promoting fusion. In addition, we also see that lysosome abnormalities are restored (Fig 4.7 d). While we observed a

103 significant rescue in MSTO1 fibroblasts with regards to mtDNA nucleoid size (Fig 4.7 e), the number of mtDNA nucleoids did not change in MSTO1 fibroblasts (Fig 4.7 f). A potential confounding factor is that the transfection protocol itself causes a decrease in mtDNA nucleoid counts, which could be masking a rescue. However, mtDNA copy number was also not rescued after only 48 hours (Fig 4.7 g). This incomplete rescue of mtDNA nucleoid abundance and copy number likely reflects that fact that it may take longer than 48 hours for the mtDNA copy number to be re-established.

Collectively, our data suggests that various pathogenic variants in MSTO1 behave in a similar fashion and lead to mitochondrial abnormalities in patient cells, in particular with regards to mtDNA, providing novel mechanistic insight into the disease pathogenesis associated with

MSTO1 mutations.

104

Figure 4.7. Expression of wild-type MSTO1 rescues cellular phenotypes in MSTO1 patient fibroblasts. Control, P4 and P7 fibroblast cells were transfected with MSTO1-P2A-mCherry or the mCherry empty vector control. Representative images of fibroblasts transfected with MSTO1-P2A-mCherry, for (a) live cells stained with picogreen and MitoTracker Deep Red, or (b) fixed cells stained with antibodies against TOMM20 (Green, mitochondria) and LAMP1 (Blue, lysosomes). Scalebars: 10µm. Transfected fibroblasts, expressing cytosolic mCherry or MSTO1-P2A-mCherry, were characterized as described above for the following cellular phenotypes: (c) mitochondrial morphology, (d) lysosome morphology, (e) average mtDNA nucleoid size (f) mtDNA nucleoid counts, and (g) relative mtDNA copy number. Student T-test was performed as indicated for (c), (d), (f), and (g). K-S test was performed to determine statistical significance for (e). *p < 0.05, **p < 0.01, *** p < 0.0001.

105

4.3 Discussion

Bi-allelic pathogenic variants in the nuclear-encoded cytosolic protein MSTO1 have been reported in seven patients from five families to date [62, 112, 204]. This study characterizes 12 families with an additional 15 affected patients, and thus presents the largest single cohort of patients with variants in the cytosolic mitochondrial fusion regulator, MSTO1. The series includes several novel pathogenic variants and allows for further delineation of the recessive

MSTO1-related disease associated phenotype. We found this phenotype in our cohort to be remarkably consistent with childhood-onset, fairly non-progressive muscle weakness, and clinical evidence of corticospinal tract and cerebellar involvement. As a corollary to this clinical presentation, an elevated CK level associated with a histologically dystrophic myopathy and early-onset/congenital yet stable cerebellar atrophy/hypoplasia is seen on testing. Clinical findings are similar to a recently reported small case series of seven patients with recessive

MSTO1-related disease [62, 112, 204]. Pigmentary retinopathy was previously observed as part of the recessive phenotype [62, 112]. This finding was not reported in any of the patients in our cohort; however, formal ophthalmologic examination was not pursued in all patients.

Arthrogryposis was also previously reported in one patient [112] while congenital onset hypotonia was noted in three of our patients no other abnormalities were reported at birth.

The delayed motor development seen in our patients could be due to the cerebellum volume loss, the muscular dystrophy or more likely a combination of both. Review of brain MRI imaging of in our cohort demonstrated that the decrease in cerebellar volume is evident in both hemispheres; the volume loss in the vermis is present at a very young age (as observed on the first MRIs obtained) and is non-progressive on follow-up imaging in the majority of patients.

Hence, it is possible that the cerebellar volume loss reflects more of a hypoplasia rather than an early-onset progressive atrophy, or a combination thereof. In early cerebellar hypoplasias that interfere with normal cerebellar development, granule cell proliferation deficiency as well as

106 some granular cell loss may underlie an overall smaller cerebellar cortical volume [209]. This may be in keeping with the slight progression of cerebellum volume loss seen in consecutive scans obtained for P12 (p.(Arg345His); p.(Thr324Ile)) at ages 1 and 2 years, which may be disproportionate to the overall brain volume and reflective of a developmental process given that cerebellum growth continues after birth [210]. Without further morphological autopsy data, the relative contributions of hypoplasia versus atrophy will have to remain undetermined. Clinically, it is reassuring that all patients achieved ambulation, which has been maintained at age 52 years in the oldest patient reported to date with MSTO1-related disease (P14; p.(Phe217L); p.(Arg279Cys)).

One of the most important diagnostic considerations in patients presenting with childhood-onset muscle weakness, elevated CK and structural brain abnormalities with prominent cerebellar involvement, includes the -dystroglycanopathies (DGs). The DGs are a clinical and genetic heterogenous sub-group within the congenital muscular dystrophies

(CMDs) that manifest as an early-onset dystrophic muscle disease with central nervous system involvement, including abnormal neuronal migration resulting in cortical malformations as well as impaired synaptic function [211, 212]. Specifically, DGs caused by mutations in ISPD and

GMPPB may manifest with phenotypes reminiscent of recessive MSTO1-related disease [212-

214]. However, hypoglycosylation of a -dystroglycan is a distinctive marker for the DGs that can be detected using specific antibodies against the matriglycan glycoepitope of - dystroglycan on muscle immunohistochemistry and western blot, which would be normal in

MSTO1-related disease. In fact, immunofluorescence analysis of the muscle biopsy from P8

(p.(Phe217Leu); p,(Asp236His) and P9 (p.(Tyr478Cys); missing) showed normal -dystroglycan glycoepitope staining (data not shown).The other highly relevant differential diagnosis with a reminiscent clinical spectrum is Marinesco-Sjogren syndrome (MSS) caused by bi-allelic mutations in SIL1. MSS is characterized by intellectual disability, early onset cataracts, ataxia

107 with cerebellar atrophy and myopathy. The absence of cataracts and severe intellectual disability appear to distinguish MSTO1-related disorders from MSS. Approximately 60% of patients with the classic clinical features of MSS harbour SIL1 pathogenic variants, whereas only 3% (1/37) of those with atypical features have readily identifiable SIL1 pathogenic variants.

Notably, in the study of SIL1 negative, atypical MSS patients, one patient was ultimately diagnosed with an AGK-related mtDNA depletion syndrome [215]. Therefore, it is possible that other atypical MSS patients may also harbour pathogenic MSTO1 variants. In order to help facilitate an accurate genetic diagnosis, MSTO1 should be included in targeted next-generation- based neuromuscular, mitochondrial and ataxia-related panels, where MSTO1 is currently not included.

In this cohort we report six novel MSTO1 missense variants, a single base pair deletion and a large genomic deletion. Variants are scattered throughout the MSTO1 gene and do not preferentially impact specific domains. We identified apparent recurring variants in three specific residues (p.(Phe217Leu), p.(Arg279His) and p.(Asp236His/Gly)). The p.(Arg279His) variant was previously reported as pathogenic in three families in heterozygosity with a truncating allele

[112, 204]. There is one individual listed in ExAC who is homozygous for the p.(Arg279His) variant, which would be unusual for a childhood-onset disease [207]. It is therefore likely that in its own right this may be a much milder pathogenic allele that needs to occur in compound heterozygosity with a more severe allele in order to manifest as early-onset MSTO1-deficiency

[112]. As noted, all patients thus presented with a remarkably homogeneous phenotype; therefore, no clear genotype-phenotype correlations emerged for recessive MSTO1 variants beyond the p.(Arg279His) observation, and the absence of bi-allelic null mutations.

The MSTO1 gene is part of a large tandem segmental duplication of approximately

240kb located on chromosome 1q22. This arrangement is the result of an evolutionary duplication event estimated to have occurred 37 million years ago in the human evolutionary

108 lineage [216], which also resulted in the derivation of the pseudogene MSTO2P (NR_024117).

This locus generates a long non-coding RNA with a nucleotide identity degree of 99.5% and

98.1% of exonic and intronic regions, respectively [216]. This high sequence similarity results in ambiguity of alignment of the typical short-read sequences generated in next-generation based genetic testing approaches, as the alignment of short reads to their proper genomic location maps equally well with both MSTO1 and MSTO2P. Consequently, the coverage of MSTO1 is significantly reduced in WES, WGS and RNA-seq data from our patients, resulting in diagnostic challenges. In fact, even targeted Sanger sequencing can be challenging to unambiguously sequence MSTO1 without contribution of the pseudogene sequence. In this context we also report the first presumed multi-exon deletion of MSTO1, which, as expected, was not identified through WES. In this patient (P7; (p.(Leu450Phe); deletion)), WES identified an apparently homozygous MSTO1 missense variant in the absence of consanguinity; however, subsequent

WGS and RNA sequencing failed to identify the deletion due to the inherent difficulties of mapping highly homologous regions. Targeted array CGH analysis with exon-level resolution was able to identify the exon 9-14 deletion. Given the duplicated and thus highly similar regions in this analysis, there were only three probes discriminating MSTO1 from MSTO2P; therefore, the mapping of the deletions to MSTO1 is still ambiguous. Given the diagnostic confidence in the disease phenotype, the reduced MSTO1 protein in this patient’s fibroblasts and the previously identified rare, predicted to be damaging missense MSTO1 variant, we suspect that the deletion is likely encompassing MSTO1. In contrast, in patient P9 (p.(Tyr478Cys); missing), who also presented with the disease phenotype and reduced MSTO1 protein levels, extensive next-generation based-sequencing including array CGH analysis only yielded a single heterozygous rare, predicted to be damaging missense MSTO1 variant. A pathogenic variant on the other allele was not readily detectable with available technology, suggesting a more complex genomic re-arrangement which may be copy number neutral. Validation work is in progress; however, this work is significantly complicated by the high sequence similarity

109 between the duplicated genomic regions of the locus. Thus proper diagnosis of MSTO1- deficiency may require specialized sequencing strategies, triggered by proper phenotypic recognition through detailed clinical examination, brain MRI and if needed MSTO1 protein analysis in fibroblasts [217].

In all seven novel MSTO1 fibroblast lines characterised, MSTO1 protein was reduced

(Fig 4.3 a), and mitochondrial network fragmentation was observed (Fig 4.4). Consistent with our findings, it has been shown previously that the levels of the mitochondrial fusion MFN1/2 and OPA1 are not affected in MSTO1 patient cells (Fig 4.3 b), [61], suggesting that the observed fragmented mitochondrial network phenotype in patient cells is related to MSTO1- deficiency. While defects in mitochondrial fusion have been linked to abnormalities in mtDNA, the specific consequences of pathogenic MSTO1 variants regarding mtDNA integrity have not been thoroughly investigated thus far [218]. Of the previous MSTO1 studies, three did not investigate mtDNA [61, 112, 204]. Meanwhile, Nasca et al., reported mtDNA depletion in muscle from patient A1 and fibroblasts from patients A1 and A2 (sisters) while mtDNA nucleoid size was not quantified [62]. Analysis of mtDNA copy number across our patient cohort now provides further insights in the role of impaired mitochondrial fusion and mtDNA depletion in MSTO1- deficient patient cells. It would be of interest to study this phenomenon in muscle tissue from

MSTO1-deficient patients, which unfortunately were not available in this cohort.

While the mechanistic link between mitochondrial fusion/fission dynamics and loss of mtDNA content remains unresolved, the present study adds another mitochondrial fusion protein, MSTO1, to the list of mitochondrial dynamics proteins that are implicated in the maintenance of mtDNA content. Our observation of enlarged nucleoids using PicoGreen staining is of interest in this context (Fig 4.4). The enlarged nucleoids could reflect changes in the topology of mtDNA that affect the ability of the dye to intercalate [183]. Alternatively, impaired segregation of mtDNA nucleoids, which typically contain only a single copy of the

110 mtDNA genome [219], could lead to multiple mtDNA molecules within close spatial proximity that appear as larger nucleoids [24]. We favor the latter hypothesis given that loss of fusion is known to impair nucleoid distribution with some mitochondrial fragments containing a relatively higher concentration of mtDNA molecules while others completely lack mtDNA [22], which we also observe in MSTO1 fibroblasts (Fig 4.4 a). Meanwhile recent work suggesting that reduced fusion causes mtDNA depletion due to insufficient distribution of the mtDNA replication machinery [20], is also consistent with our findings in MSTO1 fibroblasts. In particular, the fact that we see a full recovery of mitochondrial and lysosome morphology at 48 hours, but only a partial recovery of mtDNA (in the form of restoration of nucleoid size), further suggests that impairments in mitochondrial dynamics are upstream of mtDNA depletion in MSTO1 patient fibroblasts. Nonetheless, the exact link between mtDNA copy number and enlarged nucleoids remains unresolved.

It is also notable that mitochondrial fission was recently shown to be important for mtDNA nucleoid segregation following mtDNA replication [23], and that impairments in fission are known to lead to enlarged nucleoid structures [24]. Taken together, these observations provide further evidence that mitochondrial fusion and fission are important for mtDNA segregation and distribution. We posit that altered distribution and segregation of mtDNA (i.e. nucleoid clumping), in conjunction with reduced mitochondrial fusion, impairs mtDNA maintenance leading to mtDNA depletion.

Importantly, the degree of mtDNA depletion we see in MSTO1 patient fibroblasts (30-

70% of the normal content) is consistent with the levels of mtDNA depletion reported in fibroblasts from OPA1 or MFN2 patients [115-117] as well as in mtDNA depletion syndromes

(Fig 4.6) [220]. The clinical significance of this observation requires further studies as the remarkably consistent and homogenous phenotype of recessive MSTO1-related disease reported within our cohort is in contrast to the frequently highly variable clinical spectrum observed in mitochondrial diseases with childhood onset due to either nuclear DNA or mtDNA

111 pathogenic variants. Thus, the consequences of abnormal mitochondrial dynamics caused by recessive MSTO1 pathogenic variants may be less susceptible to acutely variable changes in metabolic demands, while likely incorporating a developmental component given the early clinical manifestations and neuroimaging abnormalities. This consistent phenotype of MSTO1- deficiency reported to date also seems in contrast to the single family with a dominantly-acting

MSTO1 variant in whom psychiatric manifestations including schizophrenia and autism were leading clinical features, while muscle weakness was mild and cerebellar involvement was not reported, thus suggesting a different underlying pathogenic mechanism. It is notable that the pathogenic recessive variants in MSTO1 investigated to date seem to impair protein stability in fibroblasts, which is consistent with a loss of function mechanism, while not providing further insight into functional subdomains of the protein. Since haploinsufficiency, as is present in some of the heterozygous parents in our cohort, does not cause a clinical phenotype, the mechanism for the dominantly acting heterozygous mutation may be dominant-negative in nature with perhaps different physiological consequences [61].

Despite evidence for the role of MSTO1 in mitochondrial fusion, the molecular structure of the MSTO1 protein and the exact mechanism by which MSTO1 performs this function directly or via mediators are still unknown. Our findings that bi-allelic loss-of-function variants in MSTO1 result in fragmented mitochondrial networks with mtDNA depletion and nucleoid abnormalities highlight a previously unappreciated role for MSTO1 in the maintenance of the mitochondrial genome. Taken together, our findings newly characterize MSTO1-deficiency as a syndrome of both mitochondrial fusion as well as of mtDNA depletion, while clinically manifesting with a remarkably consistent phenotype.

112

The R941L mutation in MYH14 disrupts mitochondrial fission and alters the distribution of mtDNA nucleoids.

This chapter contains sections from a published manuscript, with permission from Almutawa et al., 2019 (see Appendix E).

113

5.1 Introduction

In order to promote mitochondrial fragmentation, an arsenal of proteins and molecular motors are recruited to mitochondria, where they are active at precise sites and in successive steps. Paradoxically, the first identified molecular component of the fission machinery is the dynamin related GTPase, DRP1, which is required for the final scission step [89]. However, a huge body of work has identified many additional regulators, each playing a specific role at various stages of mitochondrial fission. For instance, we now appreciate that a pre-constriction step, mediated by the ER, is required to reduce mitochondrial circumference allowing DRP1 oligomeric rings to encircle mitochondria [221]. Importantly, the mechanical force required for pre-constricting mitochondria has been shown to be mediated by non-muscle myosin family II proteins A and B (NMIIA/B), which are ATP-dependent molecular motors that interact with actin filaments [222-224].

In contrast to muscle myosins, which form the contractile filaments in smooth, skeletal, and cardiac muscle, NMII proteins are present in all cells, where they regulate various intracellular processes. The human genome contains three NMII isoforms: NMIIA, NMIIB and

NMIIC, encoded by the genes MYH9, MYH10, and MYH14, respectively. These NMII isoforms are thought to be partially redundant. However, while NMIIA/B are known to regulate cytokinesis, cell motility, cell polarity, mtDNA nucleoids, and mitochondrial fission, the cellular functions of NMIIC have not been thoroughly investigated [79, 143, 225].

Notably, the NMIIC isoform was recently implicated in human disease, where a reoccurring pathogenic mutation, c.2822G>T; p.ArgR941Leu (referred to as R941L) has been associated with autosomal dominant hearing loss and peripheral neuropathy [123, 226, 227].

However, as the exact molecular function of NMIIC was unknown, the underlying mechanism of pathology remained elusive. Notably, given that the closely related isoforms NMIIA and NMIIB have a role in fission regulation, we hypothesized that NMIIC may also play a role in

114 mitochondrial fission. In this study, we characterized patient fibroblasts harbouring the R941L pathogenic variant in MYH14. Our work identified a novel connection between NMIIC protein and mitochondrial fission regulation. Importantly, the R941L variant not only disrupts fission, but also affects mtDNA regulation. Further characterization of the R941L fibroblasts reveals a prominent fission defect at the cell periphery which may begin to explain the pathology mechanism of this variant.

5.2 Results

In order to test the potential role of NMIIC as a mediator of mitochondrial fission, we examined mitochondrial morphology in dermal fibroblasts from patients carrying the R941L variant in MYH14. Our prediction was that mitochondrial morphology would be altered in cells harbouring the R941L mutation, in line with perturbations to mitochondrial fission regulation.

5.2.1 Hyperconnected mitochondrial networks at the cell periphery in R941L patient fibroblasts

Consistent with a predicted defect in mitochondrial fission regulation, patient fibroblasts harbouring mutant NMIIC displayed hyperfused mitochondrial networks [80]. Intriguingly, a close examination of mitochondrial morphology in patient fibroblasts also revealed an enrichment of hyper-connected mitochondrial networks in the periphery of greater than 60% of cells, compared to approximately 20% of control fibroblasts (Fig 5.1). Based on this observation, we hypothesized that fission may be more impaired at the cell periphery.

115

Figure 5.1. Hyperconnected mitochondrial networks at the cell periphery in R941L patient fibroblasts. A) Representative confocal images of control and patient fibroblast cells taken with an Olympus SD-OSR microscope. Live cells were stained with MitoTracker Red (Red, mitochondria) and PicoGreen (Green, nuclear and mitochondrial DNA). Scale bars indicate 10 μm. B) Quantification of control and patient cells containing hyperconnected mitochondrial networks at the cell periphery. At least 70 cells were quantified from two independent replicates. Error bars indicate standard deviations, and p-values (Student's t-test) were determined by comparison to the number of control cells with hyperfused mitochondria.

116

5.2.2 Hyperfused mitochondria at the cell periphery are resistant to phototoxicity- induced fission

In order to examine whether mitochondrial fission was more impaired at the cell periphery, we devised a fission assay that allowed us to induce mitochondrial fission using phototoxic stress and monitor consequent changes in network morphology. Using the phototoxicity-induced fission assay we specifically examined if central vs. peripheral mitochondria displayed different sensitivities to high laser intensity, a signal that typically causes mitochondrial networks to fragment. Interestingly, in the R941L patient cells, hyper-connected mitochondria at the cell periphery remained intact although central mitochondria fragmented readily (Fig 5.2). On the other hand, mitochondrial networks in control fibroblasts underwent fission throughout the entire cell, including peripheral regions containing hyper-connected mitochondria (Fig 5.2). Taken together, this finding suggests that the hyper-connected region of the network at the cell periphery in NMIIC patient cells possesses unique properties and that this phenotype may be specific to the R941L mutation.

117

Figure 5.2. Hyperconnected mitochondrial networks at the cell periphery in R941L patient fibroblasts are resistant to phototoxicity-induced fission. A) Representative confocal images of mitochondrial networks taken with an Olympus SD-OSR microscope. Control and patient cells stained with MitoTracker Red were imaged continuously over 5 min with high laser power to induce fission. Zoomed boxes represent regions with fragmented mitochondria (green hashed boxes) or resistant to fission (magenta hashed boxes) when imaging commenced, and at the end of 5 min. Signal intensity was enhanced for later frames to adjust for photobleaching. Scale bars indicate 10 μm.

5.2.3 Altered mtDNA nucleoids in R941L patient fibroblasts

Given the established link between mitochondrial fission and mtDNA maintenance [23,

24, 228, 229], including a role for NMIIA/B in mtDNA regulation [143], we set out to examine whether the R941L mutation affects mitochondrial genomes. Notably, impaired fission has been previously linked to enlarged mtDNA nucleoids [24]. Using live-cell imaging to visualize nucleoid size and distribution, we observed fewer, yet larger nucleoids in patient cells (Fig 5.3).

Furthermore, mtDNA nucleoid distribution throughout the mitochondrial network was altered in

MYH14 fibroblasts. In control cells, nucleoids were evenly distributed throughout the entire network, including hyper-connected mitochondria at the periphery. However, there was an

118 apparent lack of nucleoids in the hyper-connected peripheral mitochondria in R941L patient fibroblasts (Fig 5.1). Quantifying the average size and number of mtDNA nucleoids per cell,

MYH14 fibroblasts indeed contained fewer, but larger mtDNA nucleoids.

Notably, a reduction in the total number of mtDNA nucleoids could be due to either loss of mtDNA or clumping of several nucleoids (many nucleoids in close proximity). Examining relative mtDNA copy number suggested that there were no significant changes in R941L patient fibroblasts compared to control (Fig 5.4). Finally, we performed long-range PCR to exclude the possibility of mtDNA deletions in NMIIC patient fibroblasts (Fig 5.4). Taken together, these observations are consistent with impaired mitochondrial fission leading to mtDNA nucleoid alterations.

119

Figure 5.3. Altered mtDNA nucleoids in R941L patient fibroblasts. A) Representative confocal images of control and patient fibroblast cells taken with an Olympus SD-OSR microscope. Live cells were stained with MitoTracker Red (Red, mitochondria) and PicoGreen (Green, nuclear and mitochondrial DNA). Scale bars indicate 10 μm. B) Quantification of nucleoid size and C) number from 10 cells for each line. Error bars indicate standard deviations, and p-values (Student's t-test) were determined by comparison to control fibroblasts.

120

Figure 5.4. Analysis of mtDNA copy number and mtDNA deletions in R941L patient fibroblasts. A) Copy number of mtDNA as determined by quantitative PCR. Error bars indicate standard deviations, and p-values (Student's t-test) were determined by comparison to control fibroblasts. B) Long range PCR of mtDNA in control and patient fibroblasts showing 16.3 kb amplicons and no mtDNA deletions.

5.3 Discussion

While the R941L mutation in NMIIC has been previously associated with peripheral neuropathy [123, 226, 227], we did not fully understand the exact function of NMIIC protein nor the underlying pathogenesis mechanism. The work outlined in this chapter highlights a previously unrecognized role for NMIIC in regulating mitochondrial fission, a unique fission defect in cells harbouring the R941L mutation and a novel link between NMIIC and mtDNA regulation.

As NMIIA and NMIIB have been recently implicated in the regulation of mitochondrial fission [79], we hypothesized that the closely related isoform, NMIIC, is also a fission regulator.

Indeed, overexpression of wildtype NMIIC was shown to localize to sites of mitochondrial fission and induce fragmentation [80]. Moreover, cells with mutant NMIIC contained primarily hyperfused mitochondrial networks, in line with an impairment in fission [80].

121

Intriguingly, the fission defect was even more pronounced at the cell periphery in R941L patient fibroblasts, as evident from hyper-connected peripheral mitochondrial networks. When challenged with a fission-inducing stress, hyper-connected mitochondria at the cell periphery remained intact in R941L patient cells, while central mitochondria fragmented readily. Taken together, our data strongly suggest that NMIIC may play an important role in mediating mitochondrial fission at the cell periphery. The fission defect at the cell periphery is particularly interesting given that the R941L variant in NMIIC is associated with peripheral neuropathy in patients. Although it remains to be experimentally validated, we posit that the relative distribution of NMII proteins in the cell may be varied such that, NMIIC primarily localizes at peripheral ends (i.e. in close proximity to the plasma membrane), while NMIIA/B may be localized more centrally.

An important outstanding question pertains to the functional overlap between the three

NMII isoforms (NMIIA, NMIIB, NMIIC) in regulating mitochondrial fission. We still do not fully understand if the three isoforms have unique roles at fission sites or perform redundant functions. Elucidating the exact role of each isoform has important implications for understanding disease pathogenesis. To this end, it is notable that the three isoforms are differentially expressed in various tissues. For instance, neurons predominantly express the

NMIIB and NMIIC isoforms [230, 231] while NMIIA is abundant in platelets, amongst other tissues [232]. Thus, given their expression profiles, it is plausible that NMIIB and NMIIC work in concert to modulate mitochondrial fission in neuronal tissues, while the pro-fission activity of

NMIIA is critical elsewhere. As such, it would be expected that neural cells may be more susceptible to mutations in NMIIC than other cell types. Interestingly, this varied expression pattern is similar to that of the functionally redundant fusion proteins MFN1 and MFN2. Although both MFNs have a role in regulating mitochondrial fusion, MFN2 mutations manifest

122 predominantly as peripheral neuropathies because MFN1 is not highly expressed in neurons

[233].

In addition to mitochondrial network changes, mtDNA nucleoids were altered in cells with mutant NMIIC. The R941L patient fibroblasts contained fewer yet larger mtDNA nucleoids, consistent with a mitochondrial fission defect [24]. The presence of fewer but larger mtDNA nucleoids, with no concomitant change in mtDNA copy number, argues that these enlarged structures contain multiple clumped mtDNA genomes. This observation is in fact consistent with previous studies where blocking fission by acute knockdown of DRP1 leads to fewer, but larger nucleoids composed of multiple genomes, and no change in relative mtDNA levels [24]. Thus, our data suggests that NMIIC plays a role in regulating the distribution of mtDNA throughout the network by mediating mitochondrial fission. The link to mtDNA nucleoid distribution is also evident from the reduced number of nucleoids at the hyper-connected peripheral mitochondria observed in the R941L fibroblasts. To this end, it would be informative to examine if the few mtDNA nucleoids present in hyper-connected peripheral mitochondria appear larger than elsewhere in the network. As R941L NMIIC exhibits reduced fission activity at the cell periphery, the prediction would be that peripheral nucleoids are more aggregated.

Overall, our data highlight a previously unrecognized role for NMIIC as a regulator of mitochondrial fission. Through characterization of the R941L pathogenic variant, we also identified a novel link between NMIIC and maintenance of mtDNA nucleoids. This finding is exciting as it expands the current list of mitochondrial fission proteins involved in mtDNA regulation. Intriguingly, fibroblasts harbouring the R941L NMIIC variant display a prominent fission defect at the cell periphery which may be relevant to the peripheral neuropathy phenotype reported in patients. Further characterization of the localization, activity and interaction of NMIIC with other NMII isoforms is key to understand how these proteins regulate fission.

123

DISCUSSION

124

Mitochondrial fusion, fission and the regulation of mtDNA

The overarching theme of this project was to further understand how impaired fusion or fission regulation affects mtDNA maintenance. As outlined in Chapter 1, mitochondrial dynamics regulators are crucial for the maintenance of mtDNA. This is evidenced by the strong link between pathogenic mutations and knockout models of fusion proteins (MFN1, MFN2, OPA1) and mtDNA depletion, deletions and abnormal nucleoid distribution [20, 21, 105, 106, 118].

Similarly, aberrant nucleoid clumping and mtDNA depletion are evident in the absence of the fission regulator DRP1 [24].

The work outlined in this thesis expands the list of mitochondrial fusion and fission proteins implicated in the regulation of mitochondrial genomes (Table 6.1). Firstly, chapter 3 highlighted the mitochondrial disease gene FBXL4. While mutations in FBXL4 are established causes of mtDNA depletion syndrome [56, 57, 153, 155-158, 174], the underlying mechanisms leading to mtDNA loss as well as FBXL4 protein function were unknown. Here, through investigating a previously uncharacterized pathogenic variant (C584R), we showed that the

FBXL4 protein functions as a mitochondrial fusion regulator. Further, we demonstrated that restoring fused mitochondrial networks (by inhibiting fission) rescued mitochondrial network morphology and mtDNA depletion in FBXL4 patient cells.

Next, chapter 4 discussed a recently identified cytosolic regulator of mitochondrial fusion, MSTO1 [61, 62]. Notably, pathogenic mutations in MSTO1 have been shown to disrupt the pro-fusion function of MSTO1 protein [61, 62], however the consequences of impaired fusion on mtDNA regulation were not investigated. In this project, we characterized various novel pathogenic MSTO1 mutations and showed that defective fusion leads to mtDNA depletion and enlarged mtDNA nucleoids in MSTO1 patient cells. Additionally, we showed that transiently restoring MSTO1 protein function rescued mitochondrial network morphology as well as nucleoid distribution throughout the network. However, mtDNA copy number remained depleted

125 following 48hrs of re-expressing wildtype MSTO1 protein. Taken together, the findings from characterizing these mitochondrial disease genes (FBXL4 and MSTO1) suggest that there is indeed more to learn about mitochondrial fusion machinery and how these regulators participate in the maintenance of mitochondrial genomes. How these findings integrated with what we know about mitochondrial fusion and the regulation of mitochondrial genomes will be discussed in the following section.

Finally, chapter 5 showcased a member of the nonmuscle myosin II family, NMIIC, encoded by MYH14. Our group recently showed that the R941 pathogenic MYH14 variant impairs mitochondrial fission regulation in patient cells [80]. My contribution to this study was investigating the connection between defective fission regulation and mtDNA maintenance in patient cells harbouring the R941L variant. Certainly, the fission defect was associated with enlarged nucleoids consistent with an impairment in mtDNA distribution throughout the network.

Interestingly, the fission defect in cells with mutant NMIIC was more prominent at the cell periphery which was also accompanied by localized abnormal mtDNA nucleoid distribution.

Thus, these results implicate other components of the fission machinery, namely NMIIC, in the regulation of mitochondrial genomes.

126

Table 6.1. Summary of mitochondrial network morphology and mtDNA phenotypes in patient primary dermal fibroblasts harbouring pathogenic mutations in FBXL4, MSTO1 or NMIIC.

mtDNA Mitochondrial Mitochondrial Network Morphology & Defective Mechanism Copy Disease Gene Nucleoids Number

Fusion Defect Mito: Fragmented networks Depletion FBXL4 (homozygous C584R Nucleoids: Fewer, larger mtDNA nucleoids (40-60%) variant) and some fragments devoid of mtDNA

Fusion Defect Mito: Fragmented networks Depletion MSTO1 (Compound Nucleoids: Fewer, larger mtDNA nucleoids (30-70%) heterozygous variants) and some fragments devoid of mtDNA Mito: Hyperfused networks Fission Defect Nucleoids: Fewer, larger nucleoids with MYH14 (heterozygous R941L Normal altered distribution in hyperfused variant) mitochondria at the cell periphery

How does restoring fused mitochondrial networks affect mtDNA? mtDNA copy number

Characterization of mitochondrial dysfunction in patient fibroblasts harbouring mutant

FBXL4, MSTO1 or NMIIC has provided excellent models to better understand the interplay between fusion, fission and mtDNA dynamics. To this end, rescue experiments begin to shed some light on how these morphological changes affect mitochondrial genomes (Table 6.2). The primary focus here was to particularly understand how fusion influences mtDNA.

Considering the order of events following network morphology rescue, we asked whether mitochondrial network morphology was restored before mtDNA copy or vice versa. In

MSTO1 patient cells rescued by overexpression of wildtype MSTO1 protein, mitochondrial network morphology was restored readily following 48hrs of re-expressing MSTO1 protein.

However, mtDNA levels remained depleted. As such, we can deduce that the fusion defect is upstream of mtDNA depletion.

127

Table 6.2. Summary of mitochondrial network morphology and mtDNA phenotypes following various approaches to rescue fused mitochondrial networks in patient primary dermal fibroblasts harbouring pathogenic mutations in FBXL4 or MSTO1.

Mitochondrial mtDNA mtDNA Mitochondrial Rescue Approach mtDNA Copy Network nucleoid nucleoid Disease Gene & Duration Number Morphology size count Blocking Fission Not Fused FBXL4 (mdivi-1) Rescued Rescued Rescued Networks 7 days (Clumped) Promoting Fusion (Transient MSTO1 Fused Not Rescued Not Rescued MSTO1 Rescued overexpression) Networks (low) ~40% Depletion 48 hours

Intriguingly, this result suggests that mtDNA copy number rescue needs more than

48hrs to be re-established. Supporting this logic is the fact that prolonged mdivi-1 treatment

(7days) rescued mtDNA copies in the FBXL4 impaired fusion model. Taken together, this data suggests that in patient cells with pathogenic FBXL4 and MSTO1 mutations, mtDNA copy is re- established anywhere between 2-7days when fused mitochondrial network are restored.

Nonetheless, we still do not know the exact kinetics of mtDNA re-population nor the mechanisms responsible for sensing/communicating changes in mtDNA copy number.

In the light of recent work on MFN1/2 and OPA1 defective fusion models [20], our findings corroborate the importance of intact fusion machinery for maintaining mitochondrial genomes. Notably, mtDNA depletion in the absence of MFN1/2 (double knockouts) or OPA1 was attributed to a defect in mtDNA replication. Further, the authors report that blocking fusion alters the composition and distribution of mtDNA replisomes [20]. The underlying mechanism leading to mtDNA loss in the FBXL4 and MSTO1 mutant models remains unknown. However, it is particularly interesting that restoring fused networks (e.g. with mdivi-1) was enough to restore mtDNA copy number. Certainly, this might argue that having a fused network mediates proper distribution of replisome components required for mtDNA replication.

128

Enlarged mtDNA nucleoids

Another prominent phenotype observed in models of impaired mitochondrial fusion and fission are enlarged mtDNA nucleoids (Table 6.1). Notably, based on reports from super resolution imaging, enlarged nucleoids or nucleoid aggregates/clumps are multiple nucleoids present within close proximity relative to one another [19, 20]. The appearance of mtDNA nucleoid clumps in cells with defective fission regulation is attributed to impaired distribution of nucleoids, as seen following acute knockdown of DRP1 and in our NMIIC model of impaired fission regulation. However, the cause of mtDNA nucleoid clumping in cells with impaired fusion regulation is less clear.

Notably, as described following re-expression of wildtype MSTO1 protein, promoting fused mitochondrial networks in the presence of active fission, restores nucleoid size and alleviates nucleoid aggregates. Meanwhile, restoring fused mitochondrial networks via prolonged inhibition of fission (mdivi-1, 7days), as described in the FBXL4 model, rescues mitochondrial network morphology but nucleoids remained aggregated. To this end, it is worth noting that mdivi-1 has been shown to give rise to enlarged mtDNA nucleoids [24]. This likely reflects the important role fission plays in this dynamic equation, as fission is known to be key for redistributing mtDNA nucleoids throughout the network thus preventing aggregation. One way to specifically tell if restoring fusion without inhibiting fission rescues mtDNA nucleoids in the FBXL4 model would be to promote fusion by MSTO1 protein overexpression, as described above. The prediction here would be that enlarged mtDNA nucleoids would be resolved due to restoration of fused networks and having simultaneous fission activity.

Nonetheless, why nucleoids appear clumped in cells with impaired fusion is still an open question for investigation. Given that mitochondrial fusion and fission events are coupled (i.e. decreased fusion is coupled with decreased fission) [234], there may be an overall decrease in the number of fission events in models of impaired mitochondrial fusion regulation. As such, this

129 secondary reduction in fragmentation could be responsible for the nucleoid aggregates (Fig

6.1).

Figure 6.1. Proposed model for the contribution of fusion and fission to mtDNA nucleoid distribution and the formation of mtDNA clumps. A) Under steady state conditions, there are continuous, balanced fusion and fission events, and even distribution of mtDNA nucleoids. B) In cells with a fission defect, mitochondrial fission events are reduced, which is coupled with a reduction in the number of fusion events. As such, reduced fission impairs segregation of nucleoids giving rise to clumps while the reduction in fusion can impair mtDNA replication affecting total copy number. C) Cells with impaired fusion also exhibit a reduced number of fission events. Defective fusion may contribute to reduced synthesis (mtDNA depletion) while reduced fission impaired segregation of nucleoids. RESCUE EXPERIMENTS D) Restoring fused mitochondrial networks by inhibiting fragmentation results in a net increase in the number of fusion events relative to fission events, restoring network morphology and mtDNA copy number. However, the lack of fission events impaired nucleoid segregation. E) Restoring fused mitochondrial networks by transiently promoting fusion, rescues network morphology and because fission is active, mtDNA nucleoid clumps are resolved. KO: knockout, LOF: loss of function.

130 mtDNA deletions in fusion-defective cells

Another intriguing connection between mitochondrial fusion and mtDNA is the observation of mtDNA deletions in patients with mutations in MFN2 and OPA1 (Table 1.1).

Though we did not detect mtDNA deletions in our fibroblasts with mutations in FBXL4, MSTO1 or NMIIC, the fact that impaired fusion can lead to largescale deletions is interesting. MtDNA deletions are proposed to arise due to double-strand breaks (DSBs), slippage during replication as well as replication stalling or pausing [235]. Notably, perturbations to mitochondrial fusion have been shown to alter the composition of replisomes [20]. Thus, it is likely that when fusion is impaired, mtDNA replication is more erroneous with an increased frequency of stalled or paused replication for instance due to the reduced levels of POLγ, TWINKLE or mtSSBP1 [20].

Moreover, several mtDNA replication proteins also have a role in the turnover of mitochondrial genomes as evident from the role of POLγ, MGME1 and TWINKLE in degrading linear mtDNA following DSBs [236-238]. Notably, accumulation of linear mtDNA fragments (e.g. in cells lacking functional POLγ, MGME1 or TWINKLE) has been suggested increase mtDNA rearrangements [236, 237]. As such, given that impaired fusion regulation is implicated in altering the levels of these enzymes [20], this may contribute to increased levels of linear mtDNA fragments and the formation of deletions.

Furthermore, DSBs can occur following exposure to various DNA damaging agents (e.g. ionizing radiation, chemotherapeutic agents and ROS). The significance of elevated ROS is particularly relevant here as impaired fusion is associated with elevated ROS production [239].

Therefore, defective fusion regulation could contribute to an increased susceptibility to mtDNA

DSBs and subsequently mtDNA deletions.

131

MtDNA phenotypes in models of impaired fusion are discrepant – why?

It remains unclear why some pathogenic mutations in mitochondrial fusion proteins (e.g.

MFN2 variants, Table 1.1) lead to mtDNA depletion and defective maintenance of mitochondrial genomes while others do not. Indeed, there are many possible reasons for these disparities including experimental techniques and detection sensitivity, available patient specimen, types of mutations and specific protein function(s) being disrupted.

Given that mitochondrial dysfunction typically manifests in a tissue-specific manner, it is possible that changes in mtDNA would be more prominent in certain tissues (e.g. muscle biopsies vs. dermal fibroblasts). The tissue-specific mtDNA loss may be explained by the fact that defects in fusion regulation manifest differently in various tissues. For instance, as highlighted in the Silva Ramos report, conditional cardiac double knockout of MFN1/2 leads to compensatory upregulated expression of the mitochondrial helicase, TWINKLE [20]. Meanwhile, in MFN1/2 KO fibroblasts, TWINKLE levels are significantly reduced [20]. As such, these differences likely echo why mtDNA depletion/deletions are evident in some tissues but not others. However, it is not clear why or how these changes occur.

Another intriguing angle that might begin to explain these differences pertains to the fact that fusion GTPases have multiple roles in addition to regulating fusion. For instance, in addition to mediating OMM fusion, MFN2 functions as a mitochondria-ER tethering protein [34] and mediates mito-ER lipid transfer [35]. Similarly, OPA1 plays a role in IMM fusion [36], cristae junction maintenance [39] and nucleoid dynamics [40]. Notably, pathogenic mutations could disrupt any one of these functions, however, we do not fully understand which of these roles is key for the maintenance of mtDNA. As such, defects in mtDNA maintenance will depend greatly on how a fusion protein function is altered. Further studies investigating these mutations side by side would be key to decipher exactly how mutations in fusion proteins influence mtDNA

132 maintenance. Similarly, as more fusion factors are being discovered, it is also crucial to consider how alterations to other components of the fusion machinery affect mitochondrial genomes.

Limitations

There are several notable limitations to the approaches used to examine mtDNA nucleoids. Firstly, mtDNA nucleoid size quantification. While line scanning or spinning disc confocal systems were used to visualize mtDNA nucleoids, it is notable that the resolution limit of confocal microscopy is approximately 250nm while individual mtDNA nucleoids are observed to be around 100nm in diameter using stimulated emission depletion (STED) super resolution microscopy (resolution limit: ~50nm) [19, 126]. Thus, nucleoid sizes were overestimated in the acquired images from control and patient fibroblasts. The differences in average nucleoid size

(reported as area in Fig 3.5, Fig 3.9, Fig 4.6, Fig 4.7, Fig 5.3) were significantly different, however, given the resolution limits of the systems used, nucleoids in control cells are likely much smaller than observed and the difference in size is even more drastic than currently reported. One possible solution could involve image processing techniques such as deconvolution. In this case, image processing could be performed on the acquired images where deconvolution algorithms can de-blur the signal (minimize out of focus light) and enhance the resolution of nucleoid structures [240]. Alternatively, and perhaps ideally, one could image control and patient fibroblasts using a super resolution system such as STED. However, this technology was not available to us at the time of completing this project.

Secondly, staining live cells with PicoGreen, a DNA intercalating dye, could have influenced the enlarged appearance of mtDNA nucleoids as the topology of DNA (e.g. supercoiling) and packaging state can be affected by intercalation [168, 183, 241]. However, visualizing mtDNA nucleoids in fixed cells with an anti-DNA antibody confirmed the appearance of enlarged mtDNA nucleoids and argues against the phenotype being an artifact of the

PicoGreen dye. Another advantage to immunolabeling nucleoids with an anti-DNA antibody is

133 the photostability of Alexafluor conjugated secondary antibodies compared to the rapid photobleaching observed with PicoGreen dye. As such, imaging fixed, immunolabelled samples offers greater flexibility in terms of imaging the samples for longer periods of time and acquiring z-stacks.

Furthermore, there are a few notable caveats to the approaches used to restore fused mitochondrial networks in the FBXL4 and MSTO1 models of impaired fusion. In cells with mutant FBXL4, fibroblasts were treated with mdivi-1 for 7 days in order to restore fused networks. While blocking fission with mdivi-1 for 7 days did not have notable effects on cells viability, this approach is problematic because mdivi-1 is a pharmacological compound that has off-target effects (e.g. reversible inhibitor of complex I activity) [242].

As for re-expression of wildtype MSTO1 protein in MSTO1 patient cells, transfecting primary dermal fibroblasts is a significantly challenging task. While electroporation is a feasible approach, there remains a caveat of low plasmid uptake efficiency and a transient nature of the transfection approach. Thus, moving forward, a more optimal approach would be to use a retroviral transduction system to introduce the MSTO1 expression vector into primary fibroblasts. Using a viral system would allow us to create stable rescue lines and examine the cells for a prolonged period. By including a fluorescent epitope tag, the transduced cells could also be selected by flow cytometry to obtain an enriched rescue population and circumvent low transfection efficiency. Alternatively, we could use a number of other recently reported pharmacological compounds that promote fusion (e.g. Leflunomide [243], S3 [54], BGP15 [244],

M1 [245]) or inhibit fission (e.g. P110 [246], Dynasore [247]).

134

Future directions

How does impaired fusion lead to mtDNA loss?

While this work highlights a novel connection between mitochondrial fusion regulators

(FBXL4 and MSTO1) and mtDNA copy number maintenance, the underlying mechanism responsible for mtDNA depletion remains an outstanding question. Notably, the possible mechanisms leading to mtDNA loss are increased turnover (degradation), decreased synthesis

(replication), or both. Thus, to determine whether mtDNA synthesis and/or turnover are affected in cells with impaired mitochondrial fusion, one approach could be to label mtDNA with EdU

(thymine analog) and track newly replicated, EdU-labeled mtDNA. By visualizing EdU-labelled nucleoids, one can compare control cells to cells with impaired fusion and determine if mtDNA synthesis rates are different. Additionally, an EdU pulse-chase experiment could be performed to compare the turnover rates of labeled mtDNA molecules in control and fusion defective cells.

Importantly, it would be informative to test multiple models of defective fusion (MFN KO, FBXL4,

MSTO1) simultaneously as the underlying mtDNA depletion mechanisms may be different.

Considering the mtDNA turnover hypothesis, we tested whether upregulated mitophagy in cells with defective fusion could be leading to mtDNA loss. Given that impaired fusion gives rise to small mitochondrial fragments, we predicted that there maybe an increased uptake of fragmented puncta by the autophagosome, which in turn results in loss of mtDNA. However, treating FBXL4 patient fibroblasts as well as MFN double knockout cells with chloroquine (an autophagy inhibitor) did not rescue mtDNA levels (Fig 6.2). This suggests that mitochondrial elimination is not the likely explanation for the observed mtDNA loss when fusion is defective, at least in the FBXL4 and MFN1/2 impaired fusion models, in the examined timeframe. However, it is notable that the exonuclease domain of POLγ as well as the mitochondrial endonuclease G

(EndoG) have been implicated in mtDNA turnover [236, 237, 248]. Therefore, independent of

135 mitophagic clearance of mitochondria, there are other mechanisms that may be contributing to mtDNA turnover in fusion defective cells.

Figure 6.2. The autophagy inhibitor chloroquine does not rescue mtDNA copy number in MFN1/2 double knockout cells or in FBXL4 patient fibroblasts. A) QPCR analysis of relative mtDNA copy number in control and MFN1/2 double knockout mouse embryonic fibroblasts. Cells were treated with 25uM chloroquine (CQ) for 24hrs. B) QPCR analysis of mtDNA copy number from control and patient fibroblasts under the same conditions as panel A.

Fused Mitochondrial Networks – are they all equal?

Another intriguing unanswered question pertains to whether fused mitochondrial networks that arise from upregulated fusion or downregulated fission are functionally different.

With respect to nucleoid organization and mtDNA copy number, it would be valuable to learn whether restoring fused mitochondrial networks by directly promoting fusion (e.g. by overexpression of MSTO1, FBXL4) or inhibiting fission (e.g. mdivi-1 or P110 treatment) has different effects on mtDNA. As described earlier in the FBXL4 model of impaired fusion, restoring fused mitochondrial networks by inhibiting fission only partially rescued mtDNA nucleoid phenotypes. Whether this applies to various models of impaired fusion (e.g. MSTO1,

FBXL4, MFN KO, OPA1 KO) is an open area for investigation. It is worth mentioning that blocking fission has inhibitory effects on mitophagic clearance of mitochondria [16] which may interfere with the efficiency of mitochondrial quality control. Clarifying this distinction would be fascinating in and of itself from a biological point of view. Moreover, such work would have

136 significant implications for drug development and efforts to restore mitochondrial network morphology therapeutically.

How does FBXL4 promote mitochondrial fusion?

Starting the FBXL4 project, there were many unanswered questions about the exact protein function and why pathogenic FBXL4 variants cause mtDNA depletion syndromes.

Through the work presented earlier, we now appreciate that FBXL4 is a regulator of mitochondrial fusion and that defective fusion is the likely underlying mechanism for mtDNA loss in FBXL4 patient cells. However, we still do not understand how FBXL4 promotes fusion. One way to address this question is to start by examining the likely functions of an F-box protein. F- box proteins are known to function within E3 ubiquitin ligase complexes to mediate protein ubiquitination [249]. Given that FBXL4 has been proposed to localize to the IMS, we asked whether protein ubiquitination in the IMS is important for mitochondrial fusion. To this end, targeting a constitutively active deubiquitilase to the IMS (IMS-GFP-USP8) gave rise to fragmented mitochondrial networks (Fig 6.3). However, targeting a catalytically inactive deubiquitilase did not affect mitochondrial morphology. While this data does not answer questions about how FBXL4 promotes fusion, it is clear that (de)ubiquitination in the IMS plays an important role in remodeling mitochondrial networks.

137

Figure 6.3. Targeting a constitutively active deubiquitilase to the mitochondrial intermembrane space disrupts mitochondrial networks and may phenocopy FBXL4 mutations. A) Representative confocal images of HEK cells transfected with an empty vector or IMS-targeted, GFP-tagged catalytic domain of USP8. Scalebars represent 5um. B) Quantification of mitochondrial morphology from cells as in panel A. Statistical analysis: no. of cells with fused mitochondrial morphology in control vs. IMS-USP8-overexpressing cells or in control vs. IMS-C784A overexpressing cells; Student T.test, *p<0.05. C) Western analyses of total cell lysates from cells overexpressing the USP8-GFP constructs or an empty vector. Blots were probed with antibodies against GFP or β-actin.

138

Where is FBXL4 protein in the mitochondria?

Another outstanding question with respect to FBXL4 protein pertains to its localization within the mitochondria. While initially proposed to localize to the mitochondrial IMS [57], recent studies suggest that FBXL4 may localize on the OMM. This is evidenced by BioID proximity biotinylation assays where FBXL4 interactome was shown to include primarily OMM, peroxisomal and ER proteins [250]. Intriguingly, our data also indicate that FBXL4 patient fibroblasts exhibit an ER stress (ER swelling) phenotype observed in electron micrographs and immunolabelled fibroblasts; a phenotype that is similar to cells lacking MFN2 [239] (Fig 6.4).

This further suggests that FBXL4 may be anchored to the OMM at the mito-ER interface.

Identifying the exact mitochondrial location of FBXL4 protein is critical for understanding its pro- fusion function. It is possible that if FBXL4 is in the OMM, it interacts directly or through other protein complexes with MFN1/2 proteins and modulates their fusion activity and/or mito-ER contact sites via interactions with MFN2.

139

Figure 6.4. Distinctive structural abnormalities of mitochondria and the endoplasmic reticulum in FBXL4 patient fibroblasts. A) Electron micrographs of control and FBXL4 patient cells showing mitochondria (top) and endoplasmic reticulum structures (bottom). Structural abnormalities in mitochondria are notable including swelling and abnormal cristae. As well, ER appears swollen in FBXL4 patient cells. B) Confocal images of control and FBXL4 patient fibroblasts stained with antibodies against TOMM20 (magenta) and Calnexin (green) to label mitochondrial and the endoplasmic reticulum, respectively. Abnormal ER networks in patient cells compared to control are notable. ER Tubules are indicated with an asterisk, and ER sheets with an arrowhead.

In terms of experimental approaches, FBXL4 localization can be verified via immunolabelling and super resolution microscopy. Using superresolution imaging platforms such as structured illumination microscopy (SIM) or STED, we can tease apart OMM and matrix compartments and a similar approach can be employed with cells expressing FBXL4. Notably,

FBXL4 protein has not been shown to contain a bona fide transmembrane domain [56, 57]. As such, if it is localized to the OMM, it may be anchored or associated with other OMM proteins.

Given that the C-terminal LRR domain in FBXL4 is proposed to mediate protein-protein interactions; a possible way to show if it is involved in localizing the protein to the OMM is to create a LRR-deletion construct and compare its localization to full length FBXL4 protein.

140

Additionally, we could take advantage of the C584R mutation identified in our FBXL4 patients, which resides within the LRR domain and ask if it disrupts protein localization. It is indeed possible that FBXL4 gets imported into the IMS and is shuttled between the OMM and IMS. To this end, employing protease protection assays and in vitro protein import assays on isolated mitochondria could help further clarify the behaviour and localization of FBXL4 protein.

Finally, it would be important to specifically decipher if FBXL4 mediates protein ubiquitination. Using Western analyses, we could ask if overexpression of FBXL4 alters the ubiquitination status of mitochondrial proteins. Notably, MFN1/2 ubiquitination is a pro-fusion modification [54, 251, 252], however the E3 ligase complex involved is unknown. One possibility is that FBXL4 promotes fusion by MFN1/2 ubiquitination. To test this, we could immune- precipitate MFNs and examine Ub levels following FBXL4 overexpression or knockdown.

How does MSTO1 promote mitochondrial fusion?

The cytosolic protein MSTO1 was recently identified as a mitochondrial fusion regulator

[61, 62]; however, we still do not fully understand how MSTO1 regulates fusion. Consistent with published reports [61, 62], our data indicates that overexpressing MSTO1 in HeLa cells or in control fibroblasts promotes fused mitochondrial networks (Fig 6.5 and Fig 4.7). While this may be due to direct stimulation of fusion or inhibition of fission, we used an in vitro, cell-free fusion assay in order to tease these mechanisms apart [170]. Notably, this assay depends on the fusion-mediated assembly of a split reporter, which is targeted to the mitochondrial matrix.

Mitochondria are isolated from two independent HeLa cell lines, each stably expressing one half of the split YFP-split luciferase reporter. Importantly, when mixed in vitro, fusion of the isolated mitochondria drives the assembly of a full reporter protein and YFP signal or luciferase activity could be examined as a readout of fusion activity [170]. This assay is ideal for testing the effects of cytosolic fusion factors which could be directly added to the fusion reaction. As such, we tested the effects of cytosolic extracts from cells overexpressing MSTO1 protein and found that

141

MSTO1 indeed directly promotes fusion (Fig 6.5). Thus, our preliminary data from the cell-free fusion assay are consistent with previous reports where they show increased fusion activity following MSTO1 overexpression, using a PA-GFP fusion assay [61].

Figure 6.5. MSTO1 protein promotes mitochondrial fusion. A) Overexpression of MSTO1 in human embryonic kidney (HEK) cells gives rise to hyperfused mitochondrial networks. HEK cells were transfected with an empty vector control or MSTO1-V5 for 24hrs. Mitochondrial morphology was scored from at least 50 cells, from three independent replicates. B) Western Analyses from HEK cells overexpressing MSTO1-V5 or an empty vector control probed with antibodies against MSTO1, V5 and β-actin. Arrow indicates MSTO1 band. C) Cytosolic MSTO1 protein directly promotes mitochondrial fusion in a cell-free fusion assay. Isolated mitochondria were incubated with cytosol from control cells or cells overexpressing MSTO1 protein and fusion activity measured based on complementation of a split-luciferase protein and luciferase signal. Negative controls N-Luci, C-Luci refer to mitochondria containing only the N- or C-terminal end of the luciferase protein and Neg refers to fusion reactions with no mitochondria. P-values are based on Student’s t.tests.

Does MSTO1 interact with MFN1 or MFN2?

Being a cytosolic protein, a possible pro-fusion mechanism could be through interactions with the OMM fusion machinery. To this end, we asked if overexpressing MSTO1 protein in

MFN1 or MFN2 knockout fibroblasts influences mitochondrial fusion. Here, our data shows that

MSTO1 promotes mitochondrial fusion in the absence of MFN1 or MFN2, suggesting possible interactions with both MFNs (Fig 6.6).

142

Figure 6.6. MSTO1 protein promotes fusion via interactions with MFN1 and MFN2. A) Model depicting mitofusin interactions to promote fusion and a putative role for MSTO1 in promoting homo- or heterotypic MFN interactions. B) Western analyses of total cell lysates from control, MFN1 or MFN2 knockout mouse embryonic fibroblasts overexpressing an empty vector or MSTO1-V5. Blots were probed with antibodies against endogenous MSTO1, V5, MFN1, MFN2 and β-actin. C) Mitochondrial morphology analysis from cells transfected as in B. At least 50 cells were scored for mitochondrial network morphology from three independent replicates. P-values are based on Student’s t.tests.

Moving forward, it would be important to examine how MSTO1 interacts with MFNs and dissect which step of membrane fusion MSTO1 mediates. For example, MSTO1 could be facilitating membrane docking, tethering or stimulating MFNs oligomerization or GTPase activity. Certainly, we could tease apart these steps using the cell-free fusion assay in the presence and absence of MSTO1 cytosols. Additionally, while little is known about MSTO1 protein, it would be worthwhile to employ BioID proximity ligation approach [253] and mass

143 spectrometry to learn more about MSTO1’s interactome. This information would be relevant to further understand the role of MSTO1 protein as a fusion promoter.

Does MSTO1 have a role in oxidative stress-induced mitochondrial hyperfusion?

Our data suggests that a subset of MSTO1 protein in mitochondria-enriched fractions forms high molecular weight oligomers under non-reducing conditions (monomeric MSTO1:

~62kDa) (Fig 6.7). This observation is indeed reminiscent of MFN1/2 oligomers under acute oxidative stress conditions [25] and led us to hypothesize that MSTO1 protein may be redox regulated. Notably, mitochondrial fusion is known to be regulated by acute redox stress, whereby an oxidizing environment promotes mitochondrial hyperfusion. Mechanistically, acute oxidative stress mediates MFNs oligomerization in cis and primes mitochondria for subsequent fusion events [25]. As such, we asked if MSTO1 protein may also be redox regulated, and if this regulation may be relevant for its pro-fusion activity. When examining MSTO1 protein sequence, we identified a highly conserved cysteine residue (C222) which is also predicted to be involved in disulfide bonds based on in silico predictions (Metal Detector v2.0) [254]. Notably, while overexpressing wildtype MSTO1 promotes fusion, introducing the C222A mutation appears to dampen this pro-fusion activity (Fig 6.7). Moving forward, it would be crucial to discern if this mutation also affects MSTO1 oligomers or its ability to interact with MFNs. We posit that this residue may be important for MSTO1’s role in driving membrane fusion.

144

Figure 6.7. MSTO1 may be a redox regulated protein and Cys222 is important for its pro- fusion activity. A) Western analyses of mitochondria enriched fractions from HEK cells overexpressing an empty vector or MSTO1-V5. Samples were prepared under non-reducing conditions (without β-mercaptoethanol) and blots were probed with antibodies against MSTO1 and VDAC1. Note MSTO1 bands appear at a significantly higher molecular weight than monomeric protein (monomeric MSTO1: ~62kDa). B) Sequence alignment of MSTO1 homologs in Homo sapiens NP_060586.2, Mus musculus NP_659147.2, Drosophila melanogaster NP_523435.1, Saccharomyces cerevisiae NP_013938.1, and Arabidopsis thaliana NP_195436.1. Protein sequences were aligned using COBALT [255]. A highly conserved cysteine residue is highlighted in red. C) Western analyses of total cell lysates from HEK cells expressing wildtype MSTO1-V5, C222A-V5 MSTO1 mutant or empty vector control. Blots were probes against endogenous MSTO1, V5 and β-actin. D) Mitochondrial morphology analysis from cells transfected as in C. At least 50 cells were scored for mitochondrial network morphology from three independent replicates. P-values are based on Student’s t.tests.

145

CONCLUSIONS

Our understanding of mitochondrial functions has certainly transcended their role in bioenergetics and ATP supply. Indeed, there are numerous aspects of mitochondrial biology that we are only beginning to unravel. Notably, mitochondria house over 1500 proteins, many of which with poor functional characterization or completely unknown functions. While there are many routes to deciphering novel protein roles, I believe that there is a huge advantage to studying disease relevant genes. Firstly, the connection to human disease speaks to the importance of perturbed mechanisms in normal mitochondrial function. Secondly, characterizing mitochondrial dysfunction in disease models can directly elucidate normal protein function.

Finally, the disease relevance opens up possibilities to translate basic findings back to the clinic and inform patient care.

In the work presented here, my main goal was to identify novel regulators of mitochondrial dynamics and further understand the connection between fusion, fission and mitochondrial genomes. Through characterizing mitochondrial dysfunction in patient cells harbouring mutant FBXL4 protein, my work uncovered a novel role for FBXL4 as a fusion regulator. Importantly, this provided important insights into the underlying mechanisms leading to mtDNA depletion in disease cases. Furthermore, through studying several biallelic variants in the cytosolic fusion regulator, MSTO1, my work highlighted a novel link between mutant MSTO1 and mtDNA depletion. These projects contribute important insights into the interplay between mitochondrial fusion and mtDNA dynamics, findings that will likely influence the diagnostic and therapeutic options for patients presenting with mtDNA depletion syndromes. Further, characterizing the role of NMIIC in the maintenance of mtDNA nucleoids expands the list of fission proteins implicated in regulating mtDNA nucleoid segregation throughout the network.

The implications of unraveling novel mitochondrial fusion or fission regulators are vast especially given the wide array of cellular functions affected by remodeling mitochondrial

146 networks. Additionally, perturbations in mitochondrial network morphology are prevalent in many conditions, even in diseases that are not primarily mitochondrial. Thus, improving our understanding on the complex machinery governing mitochondrial dynamics has broad consequences considering the cytoprotection afforded by restoring balanced fusion and fission.

Finally, we often think of mitochondrial fusion or fission factors as proteins that serve a sole function. However, it is becoming increasingly clear than many of these regulators mediate other cellular functions, whether it be contact with other organelles, phospholipid homeostasis, communicating mitochondrial stress, or regulating genome dynamics. As such, one must keep a critical eye open for unexpected results!

147

BIBLIOGRAPHY

1. Sabouny, R. and T.E. Shutt, Reciprocal Regulation of Mitochondrial Fission and Fusion. Trends in Biochemical Sciences, 2020. 2. Nunnari, J. and A. Suomalainen, Mitochondria: in sickness and in health. Cell, 2012. 148(6): p. 1145-59. 3. Chan, D.C., Mitochondrial Dynamics and Its Involvement in Disease. Annu Rev Pathol, 2019. 4. Archer, S.L., Mitochondrial dynamics--mitochondrial fission and fusion in human diseases. N Engl J Med, 2013. 369(23): p. 2236-51. 5. Wai, T. and T. Langer, Mitochondrial Dynamics and Metabolic Regulation. Trends Endocrinol Metab, 2016. 6. Nicholls, T.J. and C.M. Gustafsson, Separating and Segregating the Human Mitochondrial Genome. Trends in Biochemical Sciences, 2018. 43(11): p. 869-881. 7. Mitra, K., et al., A hyperfused mitochondrial state achieved at G1-S regulates cyclin E buildup and entry into S phase. Proceedings of the National Academy of Sciences of the United States of America, 2009. 106(29): p. 11960-11965. 8. Kashatus, D.F., et al., RALA and RALBP1 regulate mitochondrial fission at mitosis. Nat Cell Biol, 2011. 13(9): p. 1108-15. 9. Taguchi, N., et al., Mitotic phosphorylation of dynamin-related GTPase Drp1 participates in mitochondrial fission. J Biol Chem, 2007. 282(15): p. 11521-9. 10. Li, Z., et al., The Importance of Dendritic Mitochondria in the Morphogenesis and Plasticity of Spines and Synapses. Cell, 2004. 119(6): p. 873-887. 11. Verstreken, P., et al., Synaptic mitochondria are critical for mobilization of reserve pool vesicles at Drosophila neuromuscular junctions. Neuron, 2005. 47(3): p. 365-378. 12. Gilkerson, R.W., et al., Mitochondrial nucleoids maintain genetic autonomy but allow for functional complementation. The Journal of cell biology, 2008. 181(7): p. 1117-1128. 13. Schon, E.A. and R.W. Gilkerson, Functional complementation of mitochondrial DNAs: mobilizing mitochondrial genetics against dysfunction. Biochimica et biophysica acta, 2010. 1800(3): p. 245-249. 14. Gomes, L.C., G. Di Benedetto, and L. Scorrano, During autophagy mitochondria elongate, are spared from degradation and sustain cell viability. Nat Cell Biol, 2011. 13(5): p. 589-98. 15. Rambold, A.S., et al., Tubular network formation protects mitochondria from autophagosomal degradation during nutrient starvation. Proc Natl Acad Sci U S A, 2011. 108(25): p. 10190-5. 16. Twig, G., et al., Fission and selective fusion govern mitochondrial segregation and elimination by autophagy. Embo j, 2008. 27(2): p. 433-46. 17. Pickles, S., P. Vigié, and R.J. Youle, Mitophagy and Quality Control Mechanisms in Mitochondrial Maintenance. Current Biology, 2018. 28(4): p. R170-R185. 18. Springer, M.Z. and K.F. Macleod, Mitophagy: Mechanisms and Role in Human Disease. J Pathol, 2016. 19. Kukat, C., et al., Super-resolution microscopy reveals that mammalian mitochondrial nucleoids have a uniform size and frequently contain a single copy of mtDNA. Proceedings of the National Academy of Sciences, 2011. 108(33): p. 13534. 20. Silva Ramos, E., et al., Mitochondrial fusion is required for regulation of mitochondrial DNA replication. PLoS Genet, 2019. 15(6): p. e1008085. 21. Chen, H., et al., Mitochondrial fusion is required for mtDNA stability in skeletal muscle and tolerance of mtDNA mutations. Cell, 2010. 141(2): p. 280-9.

148

22. Chen, H., J.M. McCaffery, and D.C. Chan, Mitochondrial fusion protects against neurodegeneration in the cerebellum. Cell, 2007. 130(3): p. 548-62. 23. Lewis, S.C., L.F. Uchiyama, and J. Nunnari, ER-mitochondria contacts couple mtDNA synthesis with mitochondrial division in human cells. Science, 2016. 353(6296): p. aaf5549. 24. Ban-Ishihara, R., et al., Dynamics of nucleoid structure regulated by mitochondrial fission contributes to cristae reformation and release of cytochrome c. Proc Natl Acad Sci U S A, 2013. 110(29): p. 11863-8. 25. Shutt, T., et al., The intracellular redox state is a core determinant of mitochondrial fusion. EMBO Rep, 2012. 13(10): p. 909-15. 26. Tondera, D., et al., SLP-2 is required for stress-induced mitochondrial hyperfusion. Embo j, 2009. 28(11): p. 1589-600. 27. Gomes, L.C., G. Di Benedetto, and L. Scorrano, Essential amino acids and glutamine regulate induction of mitochondrial elongation during autophagy. Cell Cycle, 2011. 10(16): p. 2635-9. 28. Liesa, M. and O.S. Shirihai, Mitochondrial dynamics in the regulation of nutrient utilization and energy expenditure. Cell Metab, 2013. 17(4): p. 491-506. 29. Khacho, M., et al., Acidosis overrides oxygen deprivation to maintain mitochondrial function and cell survival. Nat Commun, 2014. 5: p. 3550. 30. Mattie, S., M. Krols, and H.M. McBride, The enigma of an interconnected mitochondrial reticulum: new insights into mitochondrial fusion. Curr Opin Cell Biol, 2019. 59: p. 159- 166. 31. Chen, H., et al., Mitofusins Mfn1 and Mfn2 coordinately regulate mitochondrial fusion and are essential for embryonic development. J Cell Biol, 2003. 160(2): p. 189-200. 32. Eura, Y., et al., Two mitofusin proteins, mammalian homologues of FZO, with distinct functions are both required for mitochondrial fusion. J Biochem, 2003. 134(3): p. 333-44. 33. Hoppins, S., et al., The soluble form of Bax regulates mitochondrial fusion via MFN2 homotypic complexes. Mol Cell, 2011. 41(2): p. 150-60. 34. de Brito, O.M. and L. Scorrano, Mitofusin 2 tethers endoplasmic reticulum to mitochondria. Nature, 2008. 456(7222): p. 605-10. 35. Hernandez-Alvarez, M.I., et al., Deficient Endoplasmic Reticulum-Mitochondrial Phosphatidylserine Transfer Causes Liver Disease. Cell, 2019. 177(4): p. 881-895.e17. 36. Cipolat, S., et al., OPA1 requires mitofusin 1 to promote mitochondrial fusion. Proc Natl Acad Sci U S A, 2004. 101(45): p. 15927-32. 37. Ishihara, N., et al., Regulation of mitochondrial morphology through proteolytic cleavage of OPA1. Embo j, 2006. 25(13): p. 2966-77. 38. Anand, R., et al., The i-AAA protease YME1L and OMA1 cleave OPA1 to balance mitochondrial fusion and fission. J Cell Biol, 2014. 204(6): p. 919-29. 39. Frezza, C., et al., OPA1 controls apoptotic cristae remodeling independently from mitochondrial fusion. Cell, 2006. 126(1): p. 177-189. 40. Elachouri, G., et al., OPA1 links human mitochondrial genome maintenance to mtDNA replication and distribution. Genome Res, 2011. 21(1): p. 12-20. 41. Oanh, N.T.K., Y.Y. Park, and H. Cho, Mitochondria elongation is mediated through SIRT1-mediated MFN1 stabilization. Cell Signal, 2017. 38: p. 67-75. 42. Lee, J.Y., et al., MFN1 deacetylation activates adaptive mitochondrial fusion and protects metabolically challenged mitochondria. J Cell Sci, 2014. 127(Pt 22): p. 4954-63. 43. Samant, S.A., et al., SIRT3 deacetylates and activates OPA1 to regulate mitochondrial dynamics during stress. Mol Cell Biol, 2014. 34(5): p. 807-19. 44. Sooyeon, L., K.L. Go, and J.S. Kim, Deacetylation of mitofusin-2 by sirtuin-1: A critical event in cell survival after ischemia. Mol Cell Oncol, 2016. 3(2): p. e1087452.

149

45. Pyakurel, A., et al., Extracellular regulated kinase phosphorylates mitofusin 1 to control mitochondrial morphology and apoptosis. Mol Cell, 2015. 58(2): p. 244-54. 46. Makino, A., et al., Regulation of mitochondrial morphology and function by O- GlcNAcylation in neonatal cardiac myocytes. Am J Physiol Regul Integr Comp Physiol, 2011. 300(6): p. R1296-302. 47. Sugiura, A., et al., MITOL regulates endoplasmic reticulum-mitochondria contacts via Mitofusin2. Mol Cell, 2013. 51(1): p. 20-34. 48. Narendra, D., et al., Parkin is recruited selectively to impaired mitochondria and promotes their autophagy. J Cell Biol, 2008. 183(5): p. 795-803. 49. Tanaka, A., et al., Proteasome and p97 mediate mitophagy and degradation of mitofusins induced by Parkin. J Cell Biol, 2010. 191(7): p. 1367-80. 50. Yun, J., et al., MUL1 acts in parallel to the PINK1/parkin pathway in regulating mitofusin and compensates for loss of PINK1/parkin. Elife, 2014. 3: p. e01958. 51. Leboucher, G.P., et al., Stress-induced phosphorylation and proteasomal degradation of mitofusin 2 facilitates mitochondrial fragmentation and apoptosis. Mol Cell, 2012. 47(4): p. 547-57. 52. Fu, M., et al., Regulation of mitophagy by the Gp78 E3 ubiquitin ligase. Mol Biol Cell, 2013. 24(8): p. 1153-62. 53. Shankar, J., et al., Raft endocytosis of AMF regulates mitochondrial dynamics through Rac1 signaling and the Gp78 ubiquitin ligase. J Cell Sci, 2013. 126(Pt 15): p. 3295-304. 54. Yue, W., et al., A small natural molecule promotes mitochondrial fusion through inhibition of the deubiquitinase USP30. Cell Res, 2014. 24(4): p. 482-96. 55. Mukherjee, R. and O. Chakrabarti, Regulation of Mitofusin1 by Mahogunin Ring Finger-1 and the proteasome modulates mitochondrial fusion. Biochim Biophys Acta, 2016. 1863(12): p. 3065-3083. 56. Bonnen, P.E., et al., Mutations in FBXL4 cause mitochondrial encephalopathy and a disorder of mitochondrial DNA maintenance. Am J Hum Genet, 2013. 93(3): p. 471-81. 57. Gai, X., et al., Mutations in FBXL4, encoding a mitochondrial protein, cause early-onset mitochondrial encephalomyopathy. Am J Hum Genet, 2013. 93(3): p. 482-95. 58. Eura, Y., et al., Identification of a novel protein that regulates mitochondrial fusion by modulating mitofusin (Mfn) protein function. J Cell Sci, 2006. 119(Pt 23): p. 4913-25. 59. Karbowski, M., et al., Role of Bax and Bak in mitochondrial morphogenesis. Nature, 2006. 443(7112): p. 658-62. 60. Donkervoort, S., et al., MSTO1 mutations cause mtDNA depletion, manifesting as muscular dystrophy with cerebellar involvement. Acta Neuropathol, 2019. 61. Gal, A., et al., MSTO1 is a cytoplasmic pro-mitochondrial fusion protein, whose mutation induces myopathy and ataxia in humans. EMBO Mol Med, 2017. 9(7): p. 967-984. 62. Nasca, A., et al., Recessive mutations in MSTO1 cause mitochondrial dynamics impairment, leading to myopathy and ataxia. Human Mutation, 2017. 38(8): p. 970-977. 63. Sabouny, R., et al., Characterization of the C584R variant in the mtDNA depletion syndrome gene FBXL4, reveals a novel role for FBXL4 as a regulator of mitochondrial fusion. Biochim Biophys Acta Mol Basis Dis, 2019. 1865(11): p. 165536. 64. Bahat, A., et al., MTCH2-mediated mitochondrial fusion drives exit from naive pluripotency in embryonic stem cells. Nat Commun, 2018. 9(1): p. 5132. 65. Newman, L.E., et al., The ARL2 GTPase regulates mitochondrial fusion from the intermembrane space. Cell Logist, 2017. 7(3): p. e1340104. 66. Newman, L.E., et al., The ARL2 GTPase is required for mitochondrial morphology, motility, and maintenance of ATP levels. PLoS One, 2014. 9(6): p. e99270. 67. Frohman, M.A., Role of mitochondrial lipids in guiding fission and fusion. J Mol Med (Berl), 2015. 93(3): p. 263-9.

150

68. Kameoka, S., et al., Phosphatidic Acid and Cardiolipin Coordinate Mitochondrial Dynamics. Trends Cell Biol, 2018. 28(1): p. 67-76. 69. Zhang, Y., et al., Mitoguardin Regulates Mitochondrial Fusion through MitoPLD and Is Required for Neuronal Homeostasis. Mol Cell, 2016. 61(1): p. 111-24. 70. Ban, T., et al., Relationship between OPA1 and cardiolipin in mitochondrial inner- membrane fusion. Biochim Biophys Acta Bioenerg, 2018. 1859(9): p. 951-957. 71. Ban, T., et al., Molecular basis of selective mitochondrial fusion by heterotypic action between OPA1 and cardiolipin. Nat Cell Biol, 2017. 19(7): p. 856-863. 72. Zhao, T., et al., PISD is a mitochondrial disease gene causing skeletal dysplasia, cataracts, and white matter changes. Life Sci Alliance, 2019. 2(2). 73. Janer, A., et al., SLC25A46 is required for mitochondrial lipid homeostasis and cristae maintenance and is responsible for Leigh syndrome. EMBO molecular medicine, 2016. 8(9): p. 1019-1038. 74. Hoppins, S., et al., Mitochondrial outer and inner membrane fusion requires a modified carrier protein. The Journal of cell biology, 2009. 184(4): p. 569-581. 75. Kraus, F. and M.T. Ryan, The constriction and scission machineries involved in mitochondrial fission. J Cell Sci, 2017. 130(18): p. 2953-2960. 76. Manor, U., et al., A mitochondria-anchored isoform of the actin-nucleating spire protein regulates mitochondrial division. Elife, 2015. 4. 77. Korobova, F., V. Ramabhadran, and H.N. Higgs, An actin-dependent step in mitochondrial fission mediated by the ER-associated formin INF2. Science, 2013. 339(6118): p. 464-7. 78. Rehklau, K., et al., Cofilin1-dependent actin dynamics control DRP1-mediated mitochondrial fission. Cell Death Dis, 2017. 8(10): p. e3063. 79. Korobova, F., T.J. Gauvin, and H.N. Higgs, A role for myosin II in mammalian mitochondrial fission. Curr Biol, 2014. 24(4): p. 409-14. 80. Almutawa, W., et al., The R941L mutation in MYH14 disrupts mitochondrial fission and associates with peripheral neuropathy. EBioMedicine, 2019. 45: p. 379-392. 81. Yang, C. and T.M. Svitkina, Ultrastructure and dynamics of the actin-myosin II cytoskeleton during mitochondrial fission. Nat Cell Biol, 2019. 21(5): p. 603-613. 82. Araujo, J.S., et al., A novel role for Myosin-Va in mitochondrial fission. bioRxiv, 2019: p. 655803. 83. Loson, O.C., et al., Fis1, Mff, MiD49, and MiD51 mediate Drp1 recruitment in mitochondrial fission. Mol Biol Cell, 2013. 24(5): p. 659-67. 84. James, D.I., et al., hFis1, a novel component of the mammalian mitochondrial fission machinery. J Biol Chem, 2003. 278(38): p. 36373-9. 85. Otera, H., et al., Mff is an essential factor for mitochondrial recruitment of Drp1 during mitochondrial fission in mammalian cells. J Cell Biol, 2010. 191(6): p. 1141-58. 86. Osellame, L.D., et al., Cooperative and independent roles of the Drp1 adaptors Mff, MiD49 and MiD51 in mitochondrial fission. J Cell Sci, 2016. 129(11): p. 2170-81. 87. Otera, H., et al., Drp1-dependent mitochondrial fission via MiD49/51 is essential for apoptotic cristae remodeling. J Cell Biol, 2016. 212(5): p. 531-44. 88. Frank, S., et al., The role of dynamin-related protein 1, a mediator of mitochondrial fission, in apoptosis. Dev Cell, 2001. 1(4): p. 515-25. 89. Smirnova, E., et al., Dynamin-related protein Drp1 is required for mitochondrial division in mammalian cells. Mol Biol Cell, 2001. 12(8): p. 2245-56. 90. Lee, J.E., et al., Multiple dynamin family members collaborate to drive mitochondrial division. Nature, 2016. 540(7631): p. 139-143. 91. Fonseca, T.B., et al., Mitochondrial fission requires DRP1 but not dynamins. Nature, 2019. 570(7761): p. E34-e42.

151

92. Kamerkar, S.C., et al., Dynamin-related protein 1 has membrane constricting and severing abilities sufficient for mitochondrial and peroxisomal fission. Nat Commun, 2018. 9(1): p. 5239. 93. Sabouny, R., et al., The Keap1-Nrf2 Stress Response Pathway Promotes Mitochondrial Hyperfusion Through Degradation of the Mitochondrial Fission Protein Drp1. Antioxid Redox Signal, 2017. 27(18): p. 1447-1459. 94. Tilokani, L., et al., Mitochondrial dynamics: overview of molecular mechanisms. Essays Biochem, 2018. 62(3): p. 341-360. 95. Hu, Q., et al., Increased Drp1 Acetylation by Lipid Overload Induces Cardiomyocyte Death and Heart Dysfunction. Circ Res, 2020. 96. Gawlowski, T., et al., Modulation of dynamin-related protein 1 (DRP1) function by increased O-linked-beta-N-acetylglucosamine modification (O-GlcNAc) in cardiac myocytes. J Biol Chem, 2012. 287(35): p. 30024-34. 97. Tondera, D., et al., The mitochondrial protein MTP18 contributes to mitochondrial fission in mammalian cells. J Cell Sci, 2005. 118(Pt 14): p. 3049-59. 98. Spier, A., et al., Bacterial FtsZ induces mitochondrial fission in human cells. bioRxiv, 2020: p. 2020.01.24.917146. 99. Francy, C.A., et al., Cryo-EM Studies of Drp1 Reveal Cardiolipin Interactions that Activate the Helical Oligomer. Sci Rep, 2017. 7(1): p. 10744. 100. Adachi, Y., et al., Coincident Phosphatidic Acid Interaction Restrains Drp1 in Mitochondrial Division. Mol Cell, 2016. 63(6): p. 1034-43. 101. Davies, V.J., et al., Opa1 deficiency in a mouse model of autosomal dominant optic atrophy impairs mitochondrial morphology, optic nerve structure and visual function. Hum Mol Genet, 2007. 16(11): p. 1307-18. 102. Wakabayashi, J., et al., The dynamin-related GTPase Drp1 is required for embryonic and brain development in mice. J Cell Biol, 2009. 186(6): p. 805-16. 103. Chen, H., et al., Titration of mitochondrial fusion rescues Mff-deficient cardiomyopathy. J Cell Biol, 2015. 211(4): p. 795-805. 104. Zuchner, S., et al., Mutations in the mitochondrial GTPase mitofusin 2 cause Charcot- Marie-Tooth neuropathy type 2A. Nat Genet, 2004. 36(5): p. 449-51. 105. Amati-Bonneau, P., et al., OPA1 mutations induce mitochondrial DNA instability and optic atrophy 'plus' phenotypes. Brain, 2008. 131(Pt 2): p. 338-51. 106. Hudson, G., et al., Mutation of OPA1 causes dominant optic atrophy with external ophthalmoplegia, ataxia, deafness and multiple mitochondrial DNA deletions: a novel disorder of mtDNA maintenance. Brain, 2008. 131(Pt 2): p. 329-37. 107. Alexander, C., et al., OPA1, encoding a dynamin-related GTPase, is mutated in autosomal dominant optic atrophy linked to chromosome 3q28. Nature genetics, 2000. 26(2): p. 211-215. 108. Waterham, H.R., et al., A lethal defect of mitochondrial and peroxisomal fission. N Engl J Med, 2007. 356(17): p. 1736-41. 109. Fahrner, J.A., et al., A novel de novo dominant negative mutation in DNM1L impairs mitochondrial fission and presents as childhood epileptic encephalopathy. Am J Med Genet A, 2016. 170(8): p. 2002-11. 110. Vanstone, J.R., et al., DNM1L-related mitochondrial fission defect presenting as refractory epilepsy. Eur J Hum Genet, 2016. 24(7): p. 1084-8. 111. von Spiczak, S., et al., DNM1 encephalopathy: A new disease of vesicle fission. Neurology, 2017. 89(4): p. 385-394. 112. Iwama, K., et al., Novel recessive mutations in MSTO1 cause cerebellar atrophy with pigmentary retinopathy. Journal of Human Genetics, 2018. 63(3): p. 263-270. 113. Suomalainen, A. and P. Isohanni, Mitochondrial DNA depletion syndromes--many genes, common mechanisms. Neuromuscul Disord, 2010. 20(7): p. 429-37.

152

114. El-Hattab, A.W. and F. Scaglia, Mitochondrial DNA depletion syndromes: review and updates of genetic basis, manifestations, and therapeutic options. Neurotherapeutics, 2013. 10(2): p. 186-98. 115. Renaldo, F., et al., MFN2, a new gene responsible for mitochondrial DNA depletion. Brain, 2012. 135(Pt 8): p. e223, 1-4; author reply e224, 1-3. 116. Vielhaber, S., et al., Mitofusin 2 mutations affect mitochondrial function by mitochondrial DNA depletion. Acta Neuropathol, 2013. 125(2): p. 245-56. 117. Spiegel, R., et al., Fatal infantile mitochondrial encephalomyopathy, hypertrophic cardiomyopathy and optic atrophy associated with a homozygous OPA1 mutation. J Med Genet, 2016. 53(2): p. 127-31. 118. Rouzier, C., et al., The MFN2 gene is responsible for mitochondrial DNA instability and optic atrophy 'plus' phenotype. Brain, 2012. 135(Pt 1): p. 23-34. 119. Nasca, A., et al., Biallelic Mutations in DNM1L are Associated with a Slowly Progressive Infantile Encephalopathy. Hum Mutat, 2016. 37(9): p. 898-903. 120. Koch, J., et al., Disturbed mitochondrial and peroxisomal dynamics due to loss of MFF causes Leigh-like encephalopathy, optic atrophy and peripheral neuropathy. J Med Genet, 2016. 53(4): p. 270-8. 121. Nasca, A., et al., Clinical and Biochemical Features in a Patient With Mitochondrial Fission Factor Gene Alteration. Frontiers in genetics, 2018. 9: p. 625-625. 122. Bartsakoulia, M., et al., A novel mechanism causing imbalance of mitochondrial fusion and fission in human myopathies. Human molecular genetics, 2018. 27(7): p. 1186- 1195. 123. Choi, B.O., et al., A complex phenotype of peripheral neuropathy, myopathy, hoarseness, and hearing loss is linked to an autosomal dominant mutation in MYH14. Hum Mutat, 2011. 32(6): p. 669-77. 124. Gustafsson, C.M., M. Falkenberg, and N.G. Larsson, Maintenance and Expression of Mammalian Mitochondrial DNA. Annu Rev Biochem, 2016. 85: p. 133-60. 125. Alam, T.I., et al., Human mitochondrial DNA is packaged with TFAM. Nucleic Acids Research, 2003. 31(6): p. 1640-1645. 126. Jakobs, S., et al., Light Microscopy of Mitochondria at the Nanoscale. Annual review of biophysics, 2020: p. 10.1146/annurev-biophys-121219-081550. 127. Alán, L., T. Špaček, and P. Ježek, Delaunay algorithm and principal component analysis for 3D visualization of mitochondrial DNA nucleoids by Biplane FPALM/dSTORM. European biophysics journal : EBJ, 2016. 45(5): p. 443-461. 128. Vidoni, S., et al., Why Mitochondria Must Fuse to Maintain Their Genome Integrity. Antioxidants & Redox Signaling, 2013. 19(4): p. 379-388. 129. Margineantu, D.H., et al., Cell cycle dependent morphology changes and associated mitochondrial DNA redistribution in mitochondria of human cell lines. Mitochondrion, 2002. 1(5): p. 425-435. 130. Korhonen, J.A., et al., Reconstitution of a minimal mtDNA replisome in vitro. The EMBO journal, 2004. 23(12): p. 2423-2429. 131. Zhang, H. and Y. Pommier, Mitochondrial topoisomerase I sites in the regulatory D-loop region of mitochondrial DNA. Biochemistry, 2008. 47(43): p. 11196-203. 132. Miralles Fusté, J., et al., In vivo occupancy of mitochondrial single-stranded DNA binding protein supports the strand displacement mode of DNA replication. PLoS genetics, 2014. 10(12): p. e1004832-e1004832. 133. Farge, G., et al., The accessory subunit B of DNA polymerase γ is required for mitochondrial replisome function. Nucleic Acids Research, 2007. 35(3): p. 902-911. 134. Roos, S., et al., Subnormal levels of POLγA cause inefficient initiation of light-strand DNA synthesis and lead to mitochondrial DNA deletions and autosomal dominant

153

progressive external ophthalmoplegia. Human Molecular Genetics, 2013. 22(12): p. 2411-2422. 135. Gorman, G.S., et al., Mitochondrial diseases. Nature Reviews Disease Primers, 2016. 2(1): p. 16080. 136. Fasullo, M. and L. Endres, Nucleotide salvage deficiencies, DNA damage and neurodegeneration. International journal of molecular sciences, 2015. 16(5): p. 9431- 9449. 137. Hahn, A. and S. Zuryn, The Cellular Mitochondrial Genome Landscape in Disease. Trends Cell Biol, 2019. 29(3): p. 227-240. 138. Kelly, R.D., et al., Mitochondrial DNA copy number is regulated in a tissue specific manner by DNA methylation of the nuclear-encoded DNA polymerase gamma A. Nucleic Acids Res, 2012. 40(20): p. 10124-38. 139. Jones, B.A. and W.L. Fangman, Mitochondrial DNA maintenance in yeast requires a protein containing a region related to the GTP-binding domain of dynamin. Genes Dev, 1992. 6(3): p. 380-9. 140. Pelloquin, L., et al., Identification of a Fission Yeast Dynamin-Related Protein Involved in Mitochondrial DNA Maintenance. Biochemical and Biophysical Research Communications, 1998. 251(3): p. 720-726. 141. Rapaport, D., et al., Fzo1p is a mitochondrial outer membrane protein essential for the biogenesis of functional mitochondria in Saccharomyces cerevisiae. J Biol Chem, 1998. 273(32): p. 20150-5. 142. Parone, P.A., et al., Preventing mitochondrial fission impairs mitochondrial function and leads to loss of mitochondrial DNA. PLoS One, 2008. 3(9): p. e3257. 143. Reyes, A., et al., Actin and myosin contribute to mammalian mitochondrial DNA maintenance. Nucleic Acids Research, 2011. 39(12): p. 5098-5108. 144. DiMauro, S. and E.A. Schon, Mitochondrial respiratory-chain diseases. N Engl J Med, 2003. 348(26): p. 2656-68. 145. Valenci, I., et al., Parkin modulates heteroplasmy of truncated mtDNA in Caenorhabditis elegans. Mitochondrion, 2015. 20: p. 64-70. 146. Lin, Y.F., et al., Maintenance and propagation of a deleterious mitochondrial genome by the mitochondrial unfolded protein response. Nature, 2016. 533(7603): p. 416-9. 147. Gitschlag, B.L., et al., Homeostatic Responses Regulate Selfish Mitochondrial Genome Dynamics in C. elegans. Cell metabolism, 2016. 24(1): p. 91-103. 148. Suen, D.-F., et al., Parkin overexpression selects against a deleterious mtDNA mutation in heteroplasmic cybrid cells. Proceedings of the National Academy of Sciences, 2010. 107(26): p. 11835. 149. Kandul, N.P., et al., Selective removal of deletion-bearing mitochondrial DNA in heteroplasmic Drosophila. Nature Communications, 2016. 7(1): p. 13100. 150. Pickrell, A.M., et al., Endogenous Parkin Preserves Dopaminergic Substantia Nigral Neurons following Mitochondrial DNA Mutagenic Stress. Neuron, 2015. 87(2): p. 371-81. 151. Malena, A., et al., Inhibition of mitochondrial fission favours mutant over wild-type mitochondrial DNA. Hum Mol Genet, 2009. 18(18): p. 3407-16. 152. Twig, G., B. Hyde, and O.S. Shirihai, Mitochondrial fusion, fission and autophagy as a quality control axis: The bioenergetic view. Biochimica et Biophysica Acta (BBA) - Bioenergetics, 2008. 1777(9): p. 1092-1097. 153. Ballout, R.A., et al., FBXL4-Related Mitochondrial DNA Depletion Syndrome 13 (MTDPS13): A Case Report With a Comprehensive Mutation Review. Frontiers in Genetics, 2019. 10(39). 154. Morton, S.U., et al., Hyperammonemia as a Presenting Feature in Two Siblings with FBXL4 Variants. JIMD Rep, 2017. 35: p. 7-15.

154

155. Baroy, T., et al., A novel mutation in FBXL4 in a Norwegian child with encephalomyopathic mitochondrial DNA depletion syndrome 13. Eur J Med Genet, 2016. 59(6-7): p. 342-6. 156. Antoun, G., et al., Detailed Biochemical and Bioenergetic Characterization of FBXL4- Related Encephalomyopathic Mitochondrial DNA Depletion. JIMD Rep, 2015. 157. Huemer, M., et al., Clinical, morphological, biochemical, imaging and outcome parameters in 21 individuals with mitochondrial maintenance defect related to FBXL4 mutations. J Inherit Metab Dis, 2015. 38(5): p. 905-14. 158. Ebrahimi-Fakhari, D., et al., Recurrent Stroke-Like Episodes in FBXL4-Associated Early- Onset Mitochondrial Encephalomyopathy. Pediatr Neurol, 2015. 53(6): p. 549-50. 159. Karbowski, M., et al., Quantitation of mitochondrial dynamics by photolabeling of individual organelles shows that mitochondrial fusion is blocked during the Bax activation phase of apoptosis. J Cell Biol, 2004. 164(4): p. 493-9. 160. Potorac, I., et al., A vital region for human glycoprotein hormone trafficking revealed by an LHB mutation. J Endocrinol, 2016. 231(3): p. 197-207. 161. Faesen, A.C., et al., The differential modulation of USP activity by internal regulatory domains, interactors and eight ubiquitin chain types. Chem Biol, 2011. 18(12): p. 1550- 61. 162. Reincke, M., et al., Mutations in the deubiquitinase gene USP8 cause Cushing's disease. Nat Genet, 2015. 47(1): p. 31-8. 163. Alexopoulou, Z., et al., Deubiquitinase Usp8 regulates α-synuclein clearance and modifies its toxicity in Lewy body disease. Proc Natl Acad Sci U S A, 2016. 113(32): p. E4688-97. 164. Eaton, J.S., et al., Ataxia-telangiectasia mutated kinase regulates ribonucleotide reductase and mitochondrial homeostasis. J Clin Invest, 2007. 117(9): p. 2723-34. 165. Livak, K.J. and T.D. Schmittgen, Analysis of relative gene expression data using real- time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods, 2001. 25(4): p. 402-8. 166. Hirano, M., R. Martí, and Y. Nishigaki, ND5 is a hot-spot for multiple atypical mitochondrial DNA deletions in mitochondrial neurogastrointestinal encephalomyopathy. Human Molecular Genetics, 2004. 13(1): p. 91-101. 167. Schindelin, J., et al., Fiji: an open-source platform for biological-image analysis. Nat Methods, 2012. 9(7): p. 676-82. 168. Ashley, N., D. Harris, and J. Poulton, Detection of mitochondrial DNA depletion in living human cells using PicoGreen staining. Exp Cell Res, 2005. 303(2): p. 432-46. 169. Molina, A.J. and O.S. Shirihai, Monitoring mitochondrial dynamics with photoactivatable [corrected] green fluorescent protein. Methods Enzymol, 2009. 457: p. 289-304. 170. Schauss, A.C., et al., A novel cell-free mitochondrial fusion assay amenable for high- throughput screenings of fusion modulators. BMC Biol, 2010. 8: p. 100. 171. Amberger, J.S., et al., OMIM.org: Online Mendelian Inheritance in Man (OMIM(R)), an online catalog of human genes and genetic disorders. Nucleic Acids Res, 2015. 43(Database issue): p. D789-98. 172. Kuptanon, C., et al., Whole exome sequencing revealed mutations in FBXL4, UNC80, and ADKin Thai patients with severe intellectual disabilities. Gene, 2019. 173. El-Hattab, A.W., et al., Molecular and clinical spectra of FBXL4 deficiency. Hum Mutat, 2017. 38(12): p. 1649-1659. 174. Dai, H., et al., FBXL4 defects are common in patients with congenital lactic acidemia and encephalomyopathic mitochondrial DNA depletion syndrome. Clin Genet, 2017. 91(4): p. 634-639. 175. Kim, J.Y., et al., Mitochondrial DNA content is decreased in autosomal dominant optic atrophy. Neurology, 2005. 64(6): p. 966-72.

155

176. Shutt, T.E. and H.M. McBride, Staying cool in difficult times: mitochondrial dynamics, quality control and the stress response. Biochim Biophys Acta, 2013. 1833(2): p. 417-24. 177. Whitehouse, S., R.H. Cooper, and P.J. Randle, Mechanism of activation of pyruvate dehydrogenase by dichloroacetate and other halogenated carboxylic acids. Biochem J, 1974. 141(3): p. 761-74. 178. Stacpoole, P.W., The pharmacology of dichloroacetate. Metabolism, 1989. 38(11): p. 1124-44. 179. Sarnat, H.B., et al., Endothelial ultrastructural alterations of intramuscular capillaries in infantile mitochondrial cytopathies: "mitochondrial angiopathy". Neuropathology, 2012. 32(6): p. 617-27. 180. Lovy, A., et al., A faster, high resolution, mtPA-GFP-based mitochondrial fusion assay acquiring kinetic data of multiple cells in parallel using confocal microscopy. J Vis Exp, 2012(65): p. e3991. 181. Karbowski, M., M.M. Cleland, and B.A. Roelofs, Photoactivatable green fluorescent protein-based visualization and quantification of mitochondrial fusion and mitochondrial network complexity in living cells. Methods Enzymol, 2014. 547: p. 57-73. 182. Cassidy-Stone, A., et al., Chemical inhibition of the mitochondrial division dynamin reveals its role in Bax/Bak-dependent mitochondrial outer membrane permeabilization. Dev Cell, 2008. 14(2): p. 193-204. 183. He, J., et al., The AAA+ protein ATAD3 has displacement loop binding properties and is involved in mitochondrial nucleoid organization. J Cell Biol, 2007. 176(2): p. 141-6. 184. Kobe, B. and A.V. Kajava, The leucine-rich repeat as a protein recognition motif. Curr Opin Struct Biol, 2001. 11(6): p. 725-32. 185. Boczonadi, V., B. Bansagi, and R. Horvath, Reversible infantile mitochondrial diseases. J Inherit Metab Dis, 2015. 38(3): p. 427-35. 186. Ghezzi, D., et al., Mutations of the mitochondrial-tRNA modifier MTO1 cause hypertrophic cardiomyopathy and lactic acidosis. Am J Hum Genet, 2012. 90(6): p. 1079-87. 187. Baruffini, E., et al., MTO1 mutations are associated with hypertrophic cardiomyopathy and lactic acidosis and cause respiratory chain deficiency in humans and yeast. Hum Mutat, 2013. 34(11): p. 1501-9. 188. Stacpoole, P.W., et al., Controlled clinical trial of dichloroacetate for treatment of congenital lactic acidosis in children. Pediatrics, 2006. 117(5): p. 1519-31. 189. Barshop, B.A., et al., Chronic treatment of mitochondrial disease patients with dichloroacetate. Mol Genet Metab, 2004. 83(1-2): p. 138-49. 190. Abdelmalak, M., et al., Long-term safety of dichloroacetate in congenital lactic acidosis. Molecular genetics and metabolism, 2013. 109(2): p. 139-143. 191. Kaufmann, P., et al., Dichloroacetate causes toxic neuropathy in MELAS: a randomized, controlled clinical trial. Neurology, 2006. 66(3): p. 324-30. 192. Parikh, S., et al., A modern approach to the treatment of mitochondrial disease. Current treatment options in neurology, 2009. 11(6): p. 414-430. 193. Avula, S., et al., Treatment of mitochondrial disorders. Curr Treat Options Neurol, 2014. 16(6): p. 292. 194. Tang, W.X., et al., Early protective effect of mitofusion 2 overexpression in STZ-induced diabetic rat kidney. Endocrine, 2012. 41(2): p. 236-47. 195. Jahani-Asl, A., et al., The mitochondrial inner membrane GTPase, optic atrophy 1 (Opa1), restores mitochondrial morphology and promotes neuronal survival following excitotoxicity. J Biol Chem, 2011. 286(6): p. 4772-82. 196. Chen, H. and D.C. Chan, Mitochondrial dynamics--fusion, fission, movement, and mitophagy--in neurodegenerative diseases. Hum Mol Genet, 2009. 18(R2): p. R169-76.

156

197. Ishihara, T., et al., Dynamics of mitochondrial DNA nucleoids regulated by mitochondrial fission is essential for maintenance of homogeneously active mitochondria during neonatal heart development. Mol Cell Biol, 2015. 35(1): p. 211-23. 198. Almutawa, W., et al., The R941L mutation in MYH14 disrupts mitochondrial fission and associates with peripheral neuropathy. EBioMedicine, 2019. 199. Nicholls, T.J., et al., Topoisomerase 3alpha Is Required for Decatenation and Segregation of Human mtDNA. Mol Cell, 2018. 69(1): p. 9-23 e6. 200. Anderson, S., et al., Sequence and organization of the human mitochondrial genome. Nature, 1981. 290(5806): p. 457-65. 201. Moraes, C.T., et al., mtDNA depletion with variable tissue expression: a novel genetic abnormality in mitochondrial diseases. Am J Hum Genet, 1991. 48(3): p. 492-501. 202. Hermann, G.J., et al., Mitochondrial fusion in yeast requires the transmembrane GTPase Fzo1p. J Cell Biol, 1998. 143(2): p. 359-73. 203. Amati-Bonneau, P., et al., OPA1 R445H mutation in optic atrophy associated with sensorineural deafness. Ann Neurol, 2005. 58(6): p. 958-63. 204. Li, K., R. Jin, and X. Wu, Whole-exome sequencing identifies rare compound heterozygous mutations in the MSTO1 gene associated with cerebellar ataxia and myopathy. Eur J Med Genet, 2019. 205. Ardicli, D., et al., A novel case of MSTO1 gene related congenital muscular dystrophy with progressive neurological involvement. Neuromuscul Disord, 2019. 206. Sobreira, N., et al., GeneMatcher: a matching tool for connecting investigators with an interest in the same gene. Hum Mutat, 2015. 36(10): p. 928-30. 207. Lek, M., et al., Analysis of protein-coding genetic variation in 60,706 humans. Nature, 2016. 536(7616): p. 285-91. 208. Demers-Lamarche, J., et al., Loss of Mitochondrial Function Impairs Lysosomes. J Biol Chem, 2016. 291(19): p. 10263-76. 209. Harding, B.N., Copp, A. J., Malformations, in Greenfield’s neuropathology, L.D. Love S, Ellison DW, Editor. 2008, CRC Press: London: Hodder Arnold. p. 365. 210. Adle-Biassette, H., J.A. Golden, and B. Harding, Developmental and perinatal brain diseases. Handb Clin Neurol, 2017. 145: p. 51-78. 211. Nickolls, A.R. and C.G. Bonnemann, The roles of dystroglycan in the nervous system: insights from animal models of muscular dystrophy. Dis Model Mech, 2018. 11(12). 212. Godfrey, C., et al., Dystroglycanopathies: coming into focus. Curr Opin Genet Dev, 2011. 21(3): p. 278-85. 213. Cirak, S., et al., ISPD gene mutations are a common cause of congenital and limb-girdle muscular dystrophies. Brain, 2013. 136(Pt 1): p. 269-81. 214. Carss, K.J., et al., Mutations in GDP-mannose pyrophosphorylase B cause congenital and limb-girdle muscular dystrophies associated with hypoglycosylation of alpha- dystroglycan. Am J Hum Genet, 2013. 93(1): p. 29-41. 215. Krieger, M., et al., SIL1 mutations and clinical spectrum in patients with Marinesco- Sjogren syndrome. Brain, 2013. 136(Pt 12): p. 3634-44. 216. Kuryshev, V.Y., et al., An anthropoid-specific segmental duplication on human chromosome 1q22. Genomics, 2006. 88(2): p. 143-51. 217. Foley, A.R.D., S.; Bonnemann, C.G. , Next-generation sequencing still needs our generation's clinicians. Next-generation sequencing still needs our generation's clinicians, 2015. 1(2): p. 2376-7839. 218. Suarez-Rivero, J.M., et al., Mitochondrial Dynamics in Mitochondrial Diseases. Diseases, 2016. 5(1). 219. Kukat, C., et al., Cross-strand binding of TFAM to a single mtDNA molecule forms the mitochondrial nucleoid. Proc Natl Acad Sci U S A, 2015. 112(36): p. 11288-93.

157

220. Nogueira, C., et al., Syndromes associated with mitochondrial DNA depletion. Italian journal of pediatrics, 2014. 40: p. 34-34. 221. Friedman, J.R., et al., ER tubules mark sites of mitochondrial division. Science, 2011. 334(6054): p. 358-62. 222. Hatch, A.L., et al., Actin filaments as dynamic reservoirs for Drp1 recruitment. Molecular Biology of the Cell, 2016. 27(20): p. 3109-3121. 223. Ji, W.-k., et al., Actin filaments target the oligomeric maturation of the dynamin GTPase Drp1 to mitochondrial fission sites. eLife, 2015. 4: p. e11553. 224. Prudent, J. and H.M. McBride, Mitochondrial Dynamics: ER Actin Tightens the Drp1 Noose. Curr Biol, 2016. 26(5): p. R207-9. 225. Golomb, E., et al., Identification and characterization of nonmuscle myosin II-C, a new member of the myosin II family. J Biol Chem, 2004. 279(4): p. 2800-8. 226. Iyadurai, S., et al., Variable phenotypic expression and onset in MYH14 distal hereditary motor neuropathy phenotype in a large, multigenerational North American family. Muscle Nerve, 2017. 56(2): p. 341-345. 227. Gonzaga-Jauregui, C., et al., Exome Sequence Analysis Suggests that Genetic Burden Contributes to Phenotypic Variability and Complex Neuropathy. Cell reports, 2015. 12(7): p. 1169-1183. 228. Garrido, N., et al., Composition and dynamics of human mitochondrial nucleoids. Mol Biol Cell, 2003. 14(4): p. 1583-96. 229. Iborra, F.J., H. Kimura, and P.R. Cook, The functional organization of mitochondrial genomes in human cells. BMC Biology, 2004. 2(1): p. 9. 230. Kneussel, M. and W. Wagner, Myosin motors at neuronal synapses: drivers of membrane transport and actin dynamics. Nat Rev Neurosci, 2013. 14(4): p. 233-47. 231. Newell-Litwa, K.A., R. Horwitz, and M.L. Lamers, Non-muscle myosin II in disease: mechanisms and therapeutic opportunities. Disease models & mechanisms, 2015. 8(12): p. 1495-1515. 232. Hu, A., F. Wang, and J.R. Sellers, Mutations in human nonmuscle myosin IIA found in patients with May-Hegglin anomaly and Fechtner syndrome result in impaired enzymatic function. J Biol Chem, 2002. 277(48): p. 46512-7. 233. Detmer, S.A. and D.C. Chan, Complementation between mouse Mfn1 and Mfn2 protects mitochondrial fusion defects caused by CMT2A disease mutations. J Cell Biol, 2007. 176(4): p. 405-14. 234. Twig, G., et al., Tagging and tracking individual networks within a complex mitochondrial web with photoactivatable GFP. Am J Physiol Cell Physiol, 2006. 291(1): p. C176-84. 235. Nissanka, N., M. Minczuk, and C.T. Moraes, Mechanisms of Mitochondrial DNA Deletion Formation. Trends in genetics : TIG, 2019. 35(3): p. 235-244. 236. Peeva, V., et al., Linear mitochondrial DNA is rapidly degraded by components of the replication machinery. Nat Commun, 2018. 9(1): p. 1727. 237. Nissanka, N., et al., The mitochondrial DNA polymerase gamma degrades linear DNA fragments precluding the formation of deletions. Nat Commun, 2018. 9(1): p. 2491. 238. Yang, C., et al., Structural insights into DNA degradation by human mitochondrial nuclease MGME1. Nucleic Acids Research, 2018. 46(20): p. 11075-11088. 239. Muñoz, J.P., et al., Mfn2 modulates the UPR and mitochondrial function via repression of PERK. The EMBO journal, 2013. 32(17): p. 2348-2361. 240. Sibarita, J.-B., Deconvolution microscopy. Advances in biochemical engineering/biotechnology, 2005. 95: p. 201-243. 241. Fukuoh, A., et al., Screen for mitochondrial DNA copy number maintenance genes reveals essential role for ATP synthase. Molecular systems biology, 2014. 10(6): p. 734- 734.

158

242. Bordt, E.A., et al., The Putative Drp1 Inhibitor mdivi-1 Is a Reversible Mitochondrial Complex I Inhibitor that Modulates Reactive Oxygen Species. Developmental cell, 2017. 40(6): p. 583-594.e6. 243. Miret-Casals, L., et al., Identification of New Activators of Mitochondrial Fusion Reveals a Link between Mitochondrial Morphology and Pyrimidine Metabolism. Cell chemical biology, 2018. 25(3): p. 268-278.e4. 244. Szabo, A., et al., Activation of mitochondrial fusion provides a new treatment for mitochondria-related diseases. Biochem Pharmacol, 2018. 150: p. 86-96. 245. Wang, D., et al., A small molecule promotes mitochondrial fusion in mammalian cells. Angewandte Chemie (International ed. in English), 2012. 51(37): p. 9302-9305. 246. Qi, X., et al., A novel Drp1 inhibitor diminishes aberrant mitochondrial fission and neurotoxicity. J Cell Sci, 2013. 126(Pt 3): p. 789-802. 247. Macia, E., et al., Dynasore, a cell-permeable inhibitor of dynamin. Dev Cell, 2006. 10(6): p. 839-50. 248. Wiehe, R.S., et al., Endonuclease G promotes mitochondrial genome cleavage and replication. Oncotarget, 2018. 9(26): p. 18309-18326. 249. Kipreos, E.T. and M. Pagano, The F-box protein family. Genome Biol, 2000. 1(5): p. Reviews3002. 250. Go, C.D., et al., A proximity biotinylation map of a human cell. bioRxiv, 2019: p. 796391. 251. Nakamura, N. and S. Hirose, Regulation of mitochondrial morphology by USP30, a deubiquitinating enzyme present in the mitochondrial outer membrane. Mol Biol Cell, 2008. 19(5): p. 1903-11. 252. Anton, F., et al., Two deubiquitylases act on mitofusin and regulate mitochondrial fusion along independent pathways. Mol Cell, 2013. 49(3): p. 487-98. 253. Kim, D.I., et al., An improved smaller biotin ligase for BioID proximity labeling. Mol Biol Cell, 2016. 27(8): p. 1188-96. 254. Passerini, A., M. Lippi, and P. Frasconi, MetalDetector v2.0: predicting the geometry of metal binding sites from protein sequence. Nucleic Acids Research, 2011. 39(suppl_2): p. W288-W292. 255. Papadopoulos, J.S. and R. Agarwala, COBALT: constraint-based alignment tool for multiple protein sequences. Bioinformatics (Oxford, England), 2007. 23(9): p. 1073- 1079.

159

Appendix A Table 4.1: Clinical Characteristics of patients carrying pathogenic mutations in MSTO1. (With permission from Donkervoort et al., 2019) (See Appendix D)

160

Table 4.1: Clinical Characteristics of patients carrying pathogenic mutations in MSTO1 (Continued) (With permission from Donkervoort et al., 2019) (See Appendix D)

161

Table 4.1: Clinical Characteristics of patients carrying pathogenic mutations in MSTO1 (Continued) (With permission from Donkervoort et al., 2019) (See Appendix D)

162

Appendix B – Copyright Permission (Chapter 1)

163

Appendix C – Copyright Permission (Chapter 3)

164

Appendix D – Copyright Permission (Chapter 4)

165

Appendix E – Copyright Permission (Chapter 5)

166

“Mischief Managed.”

167