Discovery of Molecular Mechanisms Underlying Lysosomal and Mitochondrial Defects in Parkinson’S Disease

Discovery of Molecular Mechanisms Underlying Lysosomal and Mitochondrial Defects In Parkinson’s Disease

Brigitte Phillips

Supervisor: Associate Professor Antony Cooper

A thesis in fulfillment of the requirements for the degree of Doctor of Philosophy

St Vincent’s Clinical School, Faculty of Medicine The University of New South Wales & The Garvan Institute of Medical Research April, 2018 THE UNIVERSITY OF NEW SOUTH WALES Thesis/Dissertation Sheet Surname or Family name: Phillips First name: Brigitte Other names/s: Radinovic Abbreviation for degree as given in the University calendar: PhD School: St Vincent’s Clinical School Title: The emerging contributions of the lysosome and mitochondria to Parkinson’s diseases

Abstract 350 words maximum Parkinson’s disease (PD) is a common, debilitating neurodegenerative disease yet the causes of cell dysfunction in PD remain unclear. By integrating available patient data with data from an unbiased assessment of proteomic changes in multiple cellular PD models, this study has identified new aspects of mitochondrial and lysosomal dysfunction that likely contribute to PD. Two areas were investigated in detail; the mitochondrial protein CHCHD2 and the V-ATPase complex, which acidifies endolysosomal compartments. CHCHD2 had not been well characterized or associated with PD when identified in this study, however PD-causative variants have since been described and CHCHD2 has recently been proposed to regulate mitochondrial cristae structure and interact with cytochrome c. Data from this thesis extends CHCHD2 dysfunction to sporadic PD patients, where its reduced expression was identified in the brain. In exploring the potential function of CHCHD2, mitochondrial impairment resulted in rapid translational up- regulation of CHCHD2 and its specific accumulation in depolarised mitochondria, suggesting that CHCHD2 plays a targeted role in aiding mitochondrial recovery or quarantining cytochrome c in response to mitochondrial damage. A recently identified PINK1/PUF protein-mediated translational control mechanism is proposed to regulate this response and is under investigation. This thesis also identified and explored the contributions of V-ATPase dysfunction to PD. Protein and mRNA levels of multiple V-ATPase subunits were down-regulated in patients’ brain regions and these reduced V-ATPase protein levels inversely correlated with increased levels of αSynuclein, the toxic protein that accumulates in PD. Depletion of V-ATPase subunits in vitro increased αSynuclein protein levels and produced many PD-associated lysosomal and mitochondrial dysfunctions, including increased mitochondrial Ca2 and reactive oxygen species. V-ATPase depletion also impaired lysosomal clustering in the cell perinuclear region, aligning with previous indications that the V-ATPase facilitates microtubule- mediated lysosomal transport and suggesting that lysosomal transport and maturation may be impaired in PD, a dysfunction evident in similar neurodegenerative diseases but not yet explored in PD. These results characterise new consequences of V-ATPase dysfunction and implicate V-ATPase depletion as a contributing factor to αSynuclein accumulation, lysosome and mitochondrial dysfunction in PD.

Declaration relating to disposition of project thesis/dissertation I hereby grant to the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or in part in the University libraries in all forms of media, now or here after known, subject to the provisions of the Copyright Act 1968. I retain all property rights, such as patent rights. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertation. I also authorise University Microfilms to use the 350-word abstract of my thesis in Dissertation Abstracts International (this is applicable to doctoral theses only).

…………………………… …………………………… ……………………………23/10/2017 Signature Witness Date

The University recognises that there may be exceptional circumstances requiring restrictions on copying or conditions on use. Requests for restriction for a period of up to 2 years must be made in writing. Requests for a longer period of restriction may be considered in exceptional circumstances and require the approval of the Dean of Graduate Research.

FOR OFFICE USE ONLY Date of completion of requirements for Award:

i ORIGINALITY STATEMENT

‘I hereby declare that this submission is my own work and to the best of my knowledge it contains no materials previously published or written by another person, or substantial proportions of material which have been accepted for the award of any other degree or diploma at UNSW or any other educational institution, except where due acknowledgement is made in the thesis. Any contribution made to the research by others, with whom I have worked at UNSW or elsewhere, is explicitly acknowledged in the thesis. I also declare that the intellectual content of this thesis is the product of my own work, except to the extent that assistance from others in the project's design and conception or in style, presentation and linguistic expression is acknowledged.’

Signed......

Date...... 23/10/17

ii COPYRIGHT STATEMENT ‘I hereby grant the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or part in the University libraries in all forms of media, now or here after known, subject to the provisions of the Copyright Act 1968. I retain all proprietary rights, such as patent rights. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertation. I also authorise University Microfilms to use the 350-word abstract of my thesis in Dissertation Abstract International (this is applicable to doctoral theses only). I have either used no substantial portions of copyright material in my thesis or I have obtained permission to use copyright material; where permission has not been granted I have applied/will apply for a partial restriction of the digital copy of my thesis or dissertation.' Signed ...... Date ...... 23/10/17

AUTHENTICITY STATEMENT ‘I certify that the Library deposit digital copy is a direct equivalent of the final officially approved version of my thesis. No emendation of content has occurred and if there are any minor variations in formatting, they are the result of the conversion to digital format.’

Signed ...... Date ...... 23/10/17

iii ABSTRACT

Parkinson’s disease (PD) is a common, debilitating neurodegenerative disease yet the causes of cell dysfunction in PD remain unclear. By integrating available patient data with data from an unbiased assessment of proteomic changes in multiple cellular PD models, this study has identified new aspects of mitochondrial and lysosomal dysfunction that likely contribute to PD. Two areas were investigated in detail; the mitochondrial protein CHCHD2 and the V-ATPase complex, which acidifies endolysosomal compartments.

CHCHD2 had not been well characterized or associated with PD when identified in this study, however PD-causative variants have since been described and CHCHD2 has recently been proposed to regulate mitochondrial cristae structure and interact with cytochrome c. Data from this thesis extends CHCHD2 dysfunction to sporadic PD patients, where its reduced expression was identified in the brain. In exploring the potential function of CHCHD2, mitochondrial impairment resulted in rapid translational up-regulation of CHCHD2 and its specific accumulation in depolarised mitochondria, suggesting that CHCHD2 plays a targeted role in aiding mitochondrial recovery or quarantining cytochrome c in response to mitochondrial damage. A recently identified PINK1/PUF protein-mediated translational control mechanism is proposed to regulate this response and is under investigation.

This thesis also identified and explored the contributions of V-ATPase dysfunction to PD. Protein and mRNA levels of multiple V-ATPase subunits were down-regulated in patients’ brain regions and these reduced V-ATPase protein levels inversely correlated with increased levels of αSynuclein, the toxic protein that accumulates in PD. Depletion of V-ATPase subunits in vitro increased αSynuclein protein levels and produced many PD-associated lysosomal and mitochondrial dysfunctions, including increased mitochondrial Ca2 and reactive oxygen species. V-ATPase depletion also impaired lysosomal clustering in the cell perinuclear region, aligning with previous indications that the V-ATPase facilitates microtubule-mediated lysosomal transport

and suggesting that lysosomal transport and maturation may be impaired in PD, a dysfunction evident in similar neurodegenerative diseases but not yet explored in PD. These results characterise new consequences of V-ATPase dysfunction and implicate V-ATPase depletion as a contributing factor to αSynuclein accumulation, lysosome and mitochondrial dysfunction in PD.

v PUBLICATIONS FROM THIS THESIS

Oral Presentations

Live Cell Imaging of mitochondrial function in Parkinson’s disease-like cells, EMBL Australia PhD Symposium, Sydney, Nov 2014

Poster Presentations

The Role of the Vacuolar ATPase in Mediating Mitochondrial-Lysosomal Communication in Parkinson’s Disease, Mitochondrial Communication Keystone Meeting, Taos, New Mexico, 14-18 Jan 2017

Reduced Lysosomal V-ATPase expression results in mitochondrial dysfunction in Parkinson’s disease, AussieMit, Sydney, 26-28 Nov 2016

Unravelling the Cascade of Cellular Dysfunction in Parkinson’s Disease: A SILAC labelling Mass Spectrometry Approach, EMBL Australia PhD Symposium, Melbourne, 25-27 Nov 2015

Unravelling the Cascade of Cellular Dysfunction in Parkinson’s Disease: A SILAC labelling Mass Spectrometry Approach, EMBL International PhD Symposium, Heidelberg, Germany, 21-23 Oct 2015

ACKNOWLEDGEMENTS

Foremost, I would like to thank my supervisor A/Prof Antony Cooper, without whose support this thesis would not have been possible. Thank you for your time, jokes, enthusiasm, encouragement … and the coffee machine.

To Kathryn, I am so grateful that you are in the lab. Thank you for being my lab rock, keeping me calm, always knowing how to fix whatever problem happens to arise. I wouldn’t have been able to do this without you.

To all Cooper Doopers past and present. Louise, Tom, Alastair, Lee, Steph, Fiona, Boris, Hayley and anyone who has ever contributed to the writing on the Lab glass wall. You have made it a pleasure to come to work every day. I am so lucky to work alongside such a talented and wonderful group of people.

To my family. Mum, dad, Vanessa, Leah and deda. Thank you for your constant encouragement and support. Thank you for believing in me, inspiring me and seeing the benefit on my work even when I couldn’t.

Finally, to Hector. No words will ever be able to express how grateful I am for everything that you do for me. The past 4 years have been a challenge and you have constantly been there for me, through all the highs and lows. I can’t even imagine where I would be without you. I am so lucky to have you in my life.

vii

LIST OF ABBREVIATIONS

αSynuclein αSyn βGal βGalactosidase ACC Anterior cingulate cortex AD Alzheimer’s disease BF Basal forebrain Ca2+ Calcium cDNA Complementary DNA CH Coiled-helix CHCHD2 Coiled-helix coiled-helix domain containing 2 protein CLEAR Coordinated Lysosomal Expression and Regulation CMA Chaperone mediated autophagy Cyt c Cytochrome c DOX Doxycycline ER Endoplasmic reticulum Fe2+ Iron GWAS Genome wide association study fPD Familial Parkinson’s disease LAMP Lysosome associated membrane protein LYNUS Lysosome nutrient sensing complex M6PR Mannse-6-phosphate receptor mRNA Messenger RNA MPTP 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine MTFR Methylenetetrahydrofolate reductase mTORC Mammalian target of rapamycin complex 1 mtUPR Mitochondrial unfolded protein response OR Odds Ratio PD Parkinson’s disease

viii

PKA Protein kinase A PUF Pumulo and FBF qRT-PCR Quantitative real time PCR RA Retinoic acid RAVE Regulator of ATPase of vacuoles and endosomes RILP Rab interacting lysosomal protein RNAseq RNA sequencing ROS Reactive oxygen species SFC Superior frontal cortex SILAC Stable isotope labelled amino acids in cell culture SNARE Soluble NSF Attachment Protein Receptor SOC Superior occipital cortex sPD Sporadic Parkinson’s disease TET Tetracycline TFEB Transcription Factor EB Tg Transgene TGN Trans Golgi network TOM Translocase of the outer membrane V-ATPase Vacuolar ATPase WebGestalt WEB-based Gene SeT AnaLysis Toolket WT Wild Type

TABLE OF CONTENTS ABSTRACT ...... iv PUBLICATIONS FROM THIS THESIS ...... vi ACKNOWLEDGEMENTS ...... vii LIST OF ABBREVIATIONS ...... viii

1 INTRODUCTION ...... 1 1.1 Parkinson’s disease ...... 1 1.2 Prevalence and symptoms of PD ...... 1 1.3 Parkinson’s disease progression ...... 2 1.4 Treatment of PD ...... 5 1.5 Genetic contributors to PD ...... 6 1.5.1 Confirmed PD associated gene mutations and risk factors ...... 7 1.5.2 Other genetic factors implicated in PD ...... 10 1.6 Environmental and lifestyle contributors to PD ...... 13 1.6.1 Pesticide exposure increases PD risk ...... 13 1.6.2 MPTP injection induces PD in humans ...... 14 1.6.3 Other risk factors for PD ...... 14 1.7 Ageing is the primary risk factor for PD ...... 16 1.8 αSynuclein accumulation is a pathological hallmark of PD ...... 16 1.9 The endolysosomal system is perturbed in PD ...... 18 1.9.1 Macroautophagy ...... 20 1.9.2 Chaperone Mediated Autophagy ...... 22 1.10 Elevated production of mitochondrial reactive oxygen species is a hallmark of PD 22 1.11 Mitochondrial damage is prevalent in PD ...... 24 1.11.1 Mitophagy ...... 25 1.11.2 Other mechanisms of mitochondrial quality control ...... 26 1.12 Project Aims ...... 27

2 MATERIALS AND METHODS ...... 29 2.1 General Materials and Methods ...... 29 2.1.1 Chemicals, consumables and equipment ...... 29 2.1.2 Plasmids used in this study ...... 31

2.1.3 Culturing of bacterial strains and plasmid preparation ...... 31 2.1.3.1 Bacterial strains and culture conditions ...... 31 2.1.3.2 Bacterial transformation ...... 32 2.1.3.3 Plasmid DNA purification ...... 32 2.1.4 Culturing of mammalian cells ...... 33 2.1.4.1 SH-SY5Y cells ...... 33 2.1.4.2 SH-SY5Y-αSyn and SH-SY5Y-βGal DOX-off cells ...... 33 2.1.4.3 Seeding density ...... 33 2.1.4.4 Experimental timetable ...... 34 2.1.4.5 HEK293FT cells ...... 34 2.1.5 Transfection into SH-SY5Y cells ...... 35 2.1.5.1 Transient plasmid transfection ...... 35 2.1.5.2 Transfection of siRNAs for knockdown ...... 35 2.1.6 RNA isolation, cDNA synthesis and qRT-PCR ...... 36 2.1.6.1 RNA isolation from tissue culture samples ...... 36 2.1.6.2 Conversion of RNA to cDNA ...... 37 2.1.6.3 Quantification of transcript levels by qRT-PCR ...... 37 2.1.7 Human brain samples used for quantification of transcript levels ...... 40 2.1.7.1 Isolation of RNA and preparation of cDNA from post-mortem samples ...... 40 2.1.8 Protein manipulation for immunoblotting ...... 41 2.1.8.1 Preparation of protein samples from tissue culture ...... 41 2.1.8.2 SDS-Polyacrylamide gel electrophoresis and immunoblotting ...... 41 2.1.8.3 Immunoblot densitometry analyses ...... 43 2.1.9 Assessment of cells by flow cytometry...... 43 2.1.10 Assessment of cell viability using AlamarBlueTM ...... 44 2.1.11 Assessing indications of mitochondrial damage ...... 45 2.1.11.1 Mitochondrial Reactive Oxygen Species (MitoSOX) ...... 45 2.1.11.2 Lysosomal uptake of damaged mitochondria (mito-keima) ...... 45 2.1.12 Immunofluorescence and microscopy ...... 47 2.1.12.1 Indirect immunofluorescnce ...... 47 2.1.12.2 Epifluorescent and confocal imaging ...... 48 2.2 Chapter 3 Specific Materials and Methods ...... 48 2.2.1 Setup of sample labelling...... 48 2.2.2 Preparation of cells for SILAC mass spectrometry ...... 51 2.3 Chapter 4 Specific Materials and Methods ...... 52

2.3.1 Generation of DOX inducible CHCHD2 overexpressing SH-SY5Y cells [DOX- CHCHD2] ...... 52 2.3.1.1 Construction of the CHCHD2 lentiviral plasmid ...... 53 2.3.1.2 Virus preparation and harvest ...... 55 2.3.1.3 Establishing CHCHD2-DOX and GFP-DOX stable SH-SY5Y cell lines ...... 57 2.3.2 Assays for mitochondrial functions and features ...... 57 2.3.2.1 Imaging of mitochondrial membrane potential ...... 57 2.3.2.2 Measurement of mitochondrial mass ...... 57 2.4 Chapter 5 Specific Materials and Methods ...... 58 2.4.1 Immunoblotting of human brain tissue ...... 58 2.4.2 Nuclear translocation of TFEB ...... 59 2.4.3 Measurement of Cytosolic and Mitochondrial Ca2+ ...... 59 2.4.4 Measurement of lysosomal pH ...... 60 2.4.4.1 ImageJ LysoSensor blue macro ...... 60 2.4.5 LC3-mitochondria co-localisation as a measure of mitophagy initiation ...... 61 2.4.5.1 ImageJ Tom20-LC3 co-localisation Macro ...... 62 2.4.6 Lamp1 Imaging analysis: Lysosomal size and perinuclear clustering ...... 63 2.4.6.1 Image J Measurement of LAMP1 compartments macro ...... 64 2.4.6.2 Image J macro for the measurement of lysosomal density in the perinuclear region 65

3 PROTEOMIC PROFILING OF A COMPLEX DISEASE BY INTEGRATION OF DATA DERIVED FROM IN VITRO MODELS AND PD PATIENTS ...... 67 3.1 Introduction ...... 67 3.1.1 The rotenone and elevated αSyn expression models of PD recapitulate disease characteristics in many systems ...... 68 3.1.1.1 The rotenone model of PD ...... 68 3.1.1.2 The αSyn model of PD ...... 70 3.1.1.3 The combined rotenone and αSyn model of PD may represent additional aspects of the disease ...... 74 3.1.2 Identification of early, potentially causative proteomic changes in PD ...... 75 3.1.3 Cross-referencing with patient data ...... 78 3.1.3.1 PD patient proteomic data...... 78 3.1.3.2 Available PD patient RNAseq data ...... 79 3.1.4 Specific Aims ...... 80 3.2 Results and Discussion ...... 81

xii

3.2.1 Overview of SILAC mass spectrometry ...... 81 3.2.2 Key dysregulated proteins ...... 82 3.2.2.1 KIDINS220 ...... 87 3.2.2.2 KPNA2 ...... 87 3.2.2.3 GGH ...... 88 3.2.3 Pathways analysis of dysregulated proteins ...... 92 3.2.4 OXPHOS pathway proteins are enriched amongst proteins dysregulated by elevated αSyn expression and/or rotenone treatment ...... 95 3.2.4.1 Low-dose rotenone treatment increases OXPHOS protein levels ...... 97 3.2.4.2 The effect of αSyn on mitochondrial protein levels may be a result of altered doxyxycline concentration ...... 98 3.2.5 Glycolysis, galactose metabolism and amino sugar metabolism are perturbed by the combination of rotenone treatment and elevated αSyn expression ...... 101 3.2.5.1 Central carbon-energy metabolism is perturbed in PD ...... 103 3.2.6 Amino acid metabolism is a significantly deregulated pathway ...... 111 3.3 Conclusions ...... 117

4 POST-TRANSCRIPTIONAL REGULATION OF PD-ASSOCIATED PROTEIN CHCHD2 FACILITATES A MITOCHONDRIAL STRESS RESPONSE ...... 120 4.1 Introduction ...... 120 4.1.1 CHCH Domain Containing Proteins ...... 120 4.1.2 Proposed Cellular Functions of CHCHD2 and association with PD ...... 124 4.1.3 Specific aims ...... 128 4.2 Results ...... 129 4.2.1 CHCHD2 expression is reduced in sporadic PD patient brain regions ...... 129 4.2.2 Mitochondrial stress increases in CHCHD2 protein abundance in vitro ...... 130 4.2.3 Reduced CHCHD2 expression results in diminished OXPHOS Complex IV subunit levels 133 4.2.3.1 CHCHD2 transcript and protein levels are reduced by CHCHD2 siRNA ...... 134 4.2.3.2 CHCHD2 depletion reduces the protein abundance of the OXPHOS complex IV subunit COX2, but does not reduce COX2 transcript levels ...... 135 4.2.4 CHCHD2 accumulates in depolarized mitochondria ...... 137 4.2.5 Elevated CHCHD2 expression does not perturb mitochondrial functions but highlights a potential post-transcriptional regulatory mechanism in modulating CHCHD2 protein levels ...... 139

xiii

4.2.5.1 The increase in CHCHD2 protein caused by CHCHD2 overexpression is disproportionate to the increase in CHCHD2 mRNA ...... 142 4.2.5.2 CHCHD2 overexpression does not perturb mitochondria or trigger mitophagy ...... 143 4.3 Discussion ...... 145 4.3.1 CHCHD2 is involved in the maintenance of mitochondrial cristae structure and the sequestration of potentially toxic cytochrome c within mitochondria...... 145 4.3.2 CHCHD2 is regulated at the transcriptional and translational level ...... 150 4.3.3 A proposed model for localised post-transcriptional regulation of CHCHD2 in response to decreased mitochondrial membrane potential ...... 154 4.4 Summary and future directions...... 160

5 THE VACUOLAR ATPase AND PD ...... 162 5.1 Introduction ...... 162 5.1.1 Structure of the Vacuolar ATPase ...... 163 5.1.2 V-ATPase dependent functions of the endolysosomal system ...... 165 5.1.2.1 Lysosomal degradation and recycling of cellular components ...... 166 5.1.2.2 Delivery of cargo from the trans-Golgi network the endolysosomal network ...... 167 5.1.2.3 Receptor mediated endocytosis and recycling of plasma membrane receptors ...... 168 5.1.2.4 Storage of ions and amino acids ...... 169 5.1.2.5 Synaptic vesicle loading ...... 170 5.1.2.6 Membrane fusion ...... 171 5.1.2.7 Amino acid sensing and regulation of endolysosomal signalling ...... 172 5.1.2.1 Cytoskeleton mediated transport of endolysosomal compartments ...... 173 5.1.3 Regulation of the V-ATPase ...... 176 5.1.3.1 Acidification of endolysosomal compartments is regulated by the density of V-ATPase subunits on the membrane ...... 176 5.1.3.2 V-ATPase activity is regulated by reversible dissociation of the V0 and V1 domains 177 5.1.3.3 Availability of specific amino acids impacts V-ATPase activity ...... 178 5.1.3.4 V-ATPase activity is regulated by altering proton-coupling efficiency ...... 179 5.1.3.5 Transcriptional regulation of the V-ATPase ...... 179 5.1.4 Evidence of V-ATPase dysfunction in neurodegenerative diseases including PD 181 5.1.5 Aims ...... 183 5.2 Results and Discussion ...... 184 5.2.1 V-ATPase mRNA and protein levels are reduced in brain regions of sporadic PD patients...... 184

xiv

5.2.1.1 V-ATPase transcript levels are reduced in sporadic PD patient brain regions ...... 184 5.2.1.2 V-ATPase protein abundance is reduced in sporadic PD patient brain tissue ...... 187 5.2.1.3 Exploring the relationship between the V-ATPase and αSyn in sPD patient brain tissue 188 5.2.1.4 Single Nucleotide Variants in the ATP6V0A1 gene may be associated with increased risk of PD 190 5.2.2 Modelling mild V-ATPase impairment in SH-SY5Y cells ...... 192 5.2.2.1 Establishing a chronic, low-dose BafilomycinA1 induced model of V-ATPase impairment ...... 193 5.2.2.2 Establishing an siRNA mediated knock-down model of V-ATPase impairment ...... 194 5.2.3 V-ATPase impairment causes lysosomal dysfunction in SH-SY5Y cells ...... 196 5.2.3.1 V-ATPase impairment results in the accumulation of chaperone mediated autophagy (CMA) substrates ...... 196 5.2.3.2 V-ATPase impairment increases translocation of TFEB to the nucleus ...... 199 5.2.4 V-ATPase impairment impacts mitochondria in SH-SY5Y cells ...... 201 5.2.4.1 V-ATPase impairment increases mitochondrial Ca2+ levels ...... 202 5.2.4.2 V-ATPase impairment increases the level of mitochondrial reactive oxygen species205 5.2.5 V-ATPase impairment impacts mitophagy ...... 207 5.2.5.1 V-ATPase impairment increases the initiation of mitophagy ...... 207 5.2.5.2 V-ATPase impairment does not inhibit autolysosome formation...... 210 5.2.5.3 V-ATPase depletion induces accumulation of specific mitochondrial proteins ...... 212 5.2.6 V-ATPase impairment may perturb lysosomal and mitochondrial functions through a pH independent mechanism ...... 216 5.2.6.1 Chronic, low-dose BafA1 and acute, high-dose BafA1 treatments have distinct effects on lysosomal acidification ...... 216 5.2.6.2 siRNA-mediated depletion of the V-ATPase V1 domain does not impair lysosomal acidification ...... 222 5.2.7 Depletion of V-ATPase V1 domain subunits may impair endolysosomal transport and maturation ...... 224 5.2.8 A relationship between the V-ATPase and αSynuclein may contribute to PD pathology ...... 231 5.2.8.1 αSynuclein overexpression increases expression of V-ATPase subunit mRNA and protein levels ...... 231 5.2.8.2 Elevated αSyn expression induces transcription of TFEB target genes ...... 233 5.2.8.3 V-ATPase depletion results in increased αSyn protein abundance, phenocopying the relationships observed in PD patients ...... 235

xv 5.2.8.4 A proposed mechanism through which V-ATPase depletion may result in increased αSyn protein...... 237 5.2.9 V-ATPase depletion causes a reduction in αSynuclein mRNA levels and reveals a novel relationship between αSyn and V-ATPase transcript levels in the brain ...... 238 5.3 Summary and future directions ...... 244 5.3.1 siRNA-mediated V-ATPase depletion and BafA1-mediated V-ATPase impairment have different effects on cellular functions ...... 244 5.3.2 siRNA-mediated V-ATPase depletion recapitulates many established PD phenotypes and identifies potential novel aspects of the disease ...... 247

6 Thesis Conclusions and Future directions ...... 250 6.1 CHCHD2 ...... 250 6.1.1 CHCHD2 dysfunction is implicated in both familial and sporadic PD ...... 250 6.1.2 CHCHD2 is translationally up-regulated in response to reduced mitochondrial membrane potential ...... 251 6.2 The V-ATPase ...... 252 6.2.1 The emerging role of the V-ATPase in PD ...... 252 6.2.2 The V-ATPase is involved in in mediating lysosomal transport: implications for PD 253 6.2.3 Discovery of a novel relationship between the V-ATPase and αSyn ...... 254 6.2.4 Future directions: regulation of αSyn expression through V-ATPase mediated lysosomal Fe2+ release and the GATA2 transcription factor ...... 254 REFERENCES ...... 256 APPENDIX 1 ...... 301 APPENDIX 2 ...... 327

xvi

TABLE OF FIGURES

Figure 1.1 Lewy body and Lewy neurite pathology in the dorsal motor nucleus of PD patients...... 3 Figure 1.2 Parkinson’s disease progression as defined by Braak staging...... 4 Figure 1.3 Overview of PD associated genes indicating their relative frequency and associated disease risk as a measure of allele penetrance...... 6 Figure 1.4 Impaired endolysosomal function is likely to play a central role in PD. .... 19 Figure 1.5 Macroautophagy is the process through which protein and lipid aggregates and defective organelles are degraded in the lysosome...... 21 Figure 1.6 Mitochondrial Reactive Oxygen Species (ROS). ROS are produced as a byproduct of mitochondrial OXPHOS at complexes I and/or III...... 23 Figure 1.7 PINK1/PARKIN-mediated mitophagy...... 26 Figure 2.1 Schematic of cell population gating for assessment of mKeima fluorescence by flow cytometry...... 47 Figure 2.2 Cloned CHCHD2 fragment sequence...... 54 Figure 3.2 Number of proteins identified by SILAC mass spectrometry ...... 82 Figure 3.3 Transcriptional profile of the 7 prioritized targets in the superior frontal cortex of PD patients relative to 10 controls...... 85 Figure 3.4 Schematic representation of the role of GGH in folate uptake and retention...... 89 Figure 3.5 GGH transcription is significantly reduced in the SFC of PD patients.). .... 90 Figure 3.6 The OXPHOS machinery is comprised of 5 protein complexes embedded in the inner mitochondrial membrane...... 95 Figure 3.7 Proteins associated with the OXPHOS metabolic pathway were significantly enriched among proteins dysregulated by rotenone, elevated αSyn expression or the combination of these treatments in SH-SY5Y ...... 96 Figure 3.8 NDUFB8 protein abundance is significantly increased in response to DOX withdrawal in DOXoff-αSyn SHSY-5Y cells...... 99 Figure 3.9 NDUFA8 protein abundance is increased in response to DOX withdrawal in DOXoff-βGal SHSY-5Y cells...... 100 Figure 3.10 Proteins involved in glycolysis, galactose metabolism and amino sugar metabolism were significantly enriched among proteins dysregulated by rotenone, elevated αSyn expression or the combination of these treatments .. 102 Figure 3.11 The combination of αSyn, but not βGal, and low-dose rotenone causes a significant increase in the level of the glycolytic protein GAPDH...... 103 Figure 3.12 Central carbon energy metabolic processes are centred on the glycolytic pathway...... 104

xvii

Figure 3.13 ...... 107 Figure 3.14 Transcripts encoding proteins involved in central carbon-energy metabolism are perturbed in PD patients...... 111 Figure 3.15 Proteins involved in amino acid metabolism are perturbed by rotenone and/or moderately elevated αSyn expression...... 112 Figure 3.16 Transcript levels of genes involved in dopamine detoxification are not changed in PD patient brain regions...... 116 Figure 4.1 CHCH protein domain structure and structures, proposed functions and disease associations of the 9 CHCH family proteins that have been identified in mammalian cells...... 121 Figure 4.2 The mitochondrial contact site and cristae organizing system (MICOS) complex regulates mitochondrial inner membrane architecture...... 123 Figure 4.3 CHCHD2 transcript levels are significantly reduced in the Basal Forebrain (BF), Superior Frontal Cortex (SFC) and Superior Occipital Cortex (SOC) of sporadic PD patients compared to age/sex matched controls...... 129 Figure 4.4 Rotenone, Antimycin A, CCCP and Thapsigargin concentrations were titred to determine concentration that mildly impact cell viability...... 131 Figure 4.5 Mitochondrial stress increases CHCHD2 protein levels but does not impact CHCHD2 transcript level...... 132 Figure 4.6 siRNA mediated knockdown of CHCHD2 results in significant reduction of CHCHD2 transcript and protein levels in SH-SY5Y cells...... 134 Figure 4.7 CHCHD2 knockdown reduces COX2 protein levels but does not affect COX2 transcript levels...... 136 Figure 4.8 Mitochondrial impairment mediated by rotenone, antimycinA and CCCP results in increased levels of CHCHD2 protein in the mitochondria...... 138 Figure 4.9 CHCHD2 accumulates in depolarised mitochondria...... 141 Figure 4.10 Characterization of DOX inducible CHCHD2 and GFP overexpressing SH- SY5Y cell lines...... 143 Figure 4.11 CHCHD2 overexpression does not effect levels of mitochondrial ROS, induce initiation of mitophagy, perturb mitochondrial mass or reduce cell viability...... 145 Figure 4.12 CHCHD2 regulates mitochondrial cristae structure and cytochrome c release.. 147 Figure 4.13 A model for the role of CHCHD2 as a mitochondrial stress responsive protein.. 150 Figure 4.14 Model of CHCHD2 translational regulation...... 159 Figure 5.1 The endolysosomal network is a dynamic and interrelated network of early endosomes, late endosomes and lysosomes...... 162 Figure 5.2 The V-ATPase complex is comprised of a V0 membrane and V1 cytosolic domain...... 163

xviii

Figure 5.3 The V- ATPase affects many aspects of cell biology ...... 165 Figure 5.4 The V-ATPase regulates lysosomal hydrolase activity by regulating the pH of the lysosomal lumen...... 166 Figure 5.5 The delivery of cargo from the trans-Golgi Network to endolysosomal compartments is dependent on the luminal acidity of the compartment...... 167 Figure 5.6 Uptake of extracellular cargo and recycling of plasma membrane receptors is dependent of the luminal acidity of recycling endosome compartments...... 169 Figure 5.7 Uptake of Ca2+ ions into endolysosomal compartments is mediated by Ca2+/H+ exchange channels or Ca2+ ATPase channels...... 170 Figure 5.8 The V-ATPase indirectly regulates synaptic vesicle loading...... 171 Figure 5.9 The V-ATPase V0 domain plays a role in facilitating membrane fusion.. . 172 Figure 5.10 The V-ATPase signals lysosomal nutrient status through sensing amino acid levels...... 174 Figure 5.11 The V-ATPase interacts directly and indirectly with microtubules...... 175 Figure 5.12 Together with the lysosome nutrient sensing (LYNUS) machinery, TFEB senses lysosomal stress/insufficiency and acts as a master regulator of lysosomal gene transcription...... 180 Figure 5.13 The expression of core V-ATPase subunits is significantly reduced in brain regions of sporadic PD patients...... 186 Figure 5.14 ATP6V1B2 protein abundance is reduced in the Anterior Cingulate Cortex of PD patients...... 188 Figure 5.15 The reduction in ATP6V1B2 abundance in PD patients inversely correlates with αSynuclein protein abundance...... 189 Figure 5.16 Multiple PD-associated single nucleotide variants (SNVs) were identified across the ATP6V0A1 locus. s...... 191 Figure 5.17 BafilomycinA1 treatment does not significantly reduce cell viability at concentrations at 2.5nM or lower...... 194 Figure 5.18 Knockdown of ATP6V1A and ATP6V1B2 results in reduced transcript levels of these subunits without reducing cell viability...... 195 Figure 5.19 siRNA-mediated knockdown of ATP6V1A or ATP6V1B2 reduces ATP6V1B2 protein abundance in SH-SY5Y cells...... 196 Figure 5.20 V-ATPase impairment causes an increase in chaperone mediated autophagy substrates...... 198 Figure 5.21 Neither (A) siRNA-mediated V-ATPase knockdown or (B) BafA1 treatment causes a significant change in GAPDH or MEF2D mRNA level...... 199 Figure 5.22 V-ATPase impairment triggers nuclear translocation of TFEB...... 200 Figure 5.23 Mitochondrial calcium levels are elevated by V-ATPase impairment with BafA1 or siRNA-mediated depletion of the ATP6V1A and ATP6V1B2 subunits. .. 204 Figure 5.24 Mitochondrial Reactive Oxygen Species (ROS) are increased in response to V-ATPase impairment...... 206

xix

Figure 5.25 V-ATPase impairment increases initiation of mitophagy...... 210 Figure 5.26 Mitophagy can be detected using the coral-derived fluorescent protein Keima. The Keima protein exhibits distinct excitation wavelengths dependent on the pH environment of the probe. When targeted to the mitochondria, the neutral mitochondrial pH results in an excitation maxima at 420 nM. When mitochondria are delivered to the lysosome through mitophagy, the acidic lysosomal pH results in a shift in the excitation maxima to 580 nM...... 211 Figure 5.27 V-ATPase impairment increases the proportion of mitochondria in autolysosomes...... 212 Figure 5.28 ATP5A protein abundance is significantly increased in response to V- ATPase depletion...... 214 Figure 5.29 ATP5A mRNA levels are not significantly changed by V-ATPase depletion...... 215 Figure 5.30 Inhibition of the V-ATPase using low-doses of BafA1 results in lysosomal hyperacidification...... 217 Figure 5.31 Inhibition of the V-ATPase using a high-dose of BafA1 or either of the protonophores chloroquine or nigericin results in reduced lysosomal acidity. ... 219 Figure 5.32 Low dose BafA1 treatment induces lysosomal fusion...... 221 Figure 5.33 Depletion of V-ATPase V1 domain subunits does not change endolysosomal acidity...... 223 Figure 5.34 Depletion of V-ATPase V1 domain subunits causes an apparent decrease in lysosomal size...... 227 Figure 5.35 Perinuclear clustering of endolysosomal compartments is reduced by depletion of V-ATPase V1 domain subunits but not BafA1 treatment...... 228 Figure 5.36 The correlation between increased αSyn and decreased V-ATPase protein abundance observed in PD patient brain tissue...... 231 Figure 5.37 αSyn overexpression results in increased ATP6V1B2 protein levels and increased expression of V-ATPase subunits...... 233 Figure 5.38 αSyn overexpression results in increased expression of TFEB target genes...... 235 Figure 5.39 siRNA-mediated V-ATPase depletion results in αSyn protein accumulation...... 236 Figure 5.40 V-ATPase depletion significantly reduces αSyn mRNA levels...... 239 Figure 5.41 αSyn mRNA is reduced in PD patient brain regions and correlates with the mRNA levels of V-ATPase transcripts...... 240 Figure 5.42 ATP6V1A mRNA levels are positively correlated with αSyn mRNA levels in brain regions of PD patients and controls...... 242 Figure 6.1 Proposed mechanism of V-ATPase-mediated modulation of αSyn transcription ...... 255

TABLE OF TABLES

Table 1.1 Frequency, mechanism of inheritance, and clinical characteristics associated with confirmed PD-causative gene mutations...... 8 Table 1.2 Genes with unconfirmed or controversial association with PD. I...... 11 Table 2.1 List of materials and reagents used in this study...... 29 Table 2.2 Plasmids used in this study...... 31 Table 2.3 E. Coli strains used in this study...... 31 Table 2.4 SH-SY5Y seeding volume according to plate size...... 34 Table 2.5 SH-SY5Y seeding and experimental timeline...... 34 Table 2.6 Sequences and sources of the siRNAs used in this study...... 36 Table 2.7 Thermal cycling conditions used for qRT-PCR...... 37 Table 2.8 Sequences of primers used in this study...... 38 Table 2.9 PD patient and control cohort used for transcript level assessment...... 40 Table 2.10 Polyacrylamyde gel electrophoresis conditions and antibody details...... 42 Table 2.11 Flow cytometry laser and filter setting used for flow cytometry assays in this study ...... 44 Table 2.12 Concentration and sources of antibodies used for indirect immunofluorescence...... 48 Table 2.13 SILAC mass spectrometry labelled amino acids obtained from Silantes, Germany (medium and heavy) or SigmaAldrich, Australia (light)...... 49 Table 2.14 Treatments of SH-SY5Y-αSyn cells for SILAC mass spectrometry...... 49 Table 2.15 Experimental setup indicating treatments combined to create each sample for mass spectrometry analysis and the labels used for each treatment at the 3 time-points assessed...... 50 Table 2.16 Plasmids used for production of CHCHD2-DOX ON lentivirus...... 52 Table 2.17 Primers used for construction of the CHCHD2 lentiviral plasmid...... 54 Table 2.18 Primers used for assessing lentiviral incorporation...... 56 Table 2.19 PD patient and control cohort used for assessment of ATP6V1B2 protein levels in the anterior cingulate cortex...... 58 Table 3.1 Elevated αSyn levels have been used to model PD in multiple in vivo and in vitro systems...... 72 Table 3.2 Proteomic datasets from PD patient brain regions...... 79 Table 3.3 7 proteins were prioritized based on significant changes identified in 3 or more of the 9 datasets...... 84 Table 3.4 Cross referencing of 7 proteins of interest with available proteomic data from PD affected patient brain regions...... 85 Table 3.5 Pathway analysis of proteins that were significantly dysregulated by greater than 0.2 fold (log2) conducted using the web based gene set analysis toolkit platform...... 94 Table 4.1 Genetic variants in the CHCHD2 gene that have been associated with PD. 127 Table 4.2 CHCHD2 mRNA and protein responses causes by various mitochondrial or cellular stresses...... 151 Table 5.1 Overview of the 13 V-ATPase genes,...... 164

xxi

Table 5.2 RNAseq analysis identified that V-ATPase gene expression is reduced in PD. 185 Table 5.3 Interactions identified between V-ATPase subunits and Ser-129 phosphorylated and non-phosphorylated αSynuclein...... 189 Table 5.4 Versatility Gene Based Association Study (VEGAS) analysis assessing PD- associated SNVs within V-ATPase genes...... 190 Table 5.5 αSyn-V-ATPase correlation analysis based on RNAseq data obtained from the basal forebrain (BF), superior frontal cortex (SFC) and superior occipital cortex (SOC) of 10 patients and 10 controls...... 241 Table 5.6 Multiple known cellular and molecular characteristics of PD were reproduced following V-ATPase depletion in vitro ...... 247

xxii 1. Introduction

1 INTRODUCTION

1.1 Parkinson’s disease

Parkinson’s disease (PD) is a common progressive neurodegenerative disorder caused by the loss of dopaminergic neurons in the substantia nigra region of the brain, as well as the progressive degeneration of neurons in other brain regions (Braak et al., 2004). This neuronal dysfunction underlies the characteristic symptoms of the disease, which encompass both movement-related difficulties and neuropsychological symptoms (Braak et al., 2003).

PD is the second most common neurodegenerative disorder worldwide, after Alzheimer’s disease (AD) (Pringsheim et al., 1996). However despite the high prevalence of the disease and the severity of the symptoms, there are no reliable long- term treatment options and there is no cure. The development of PD therapeutics has been limited by the lack of understanding of the molecular mechanisms that underlie the disease cause and progression. Improving the understanding of the early disease mechanisms will therefore help to develop more targeted and effective treatment regimes.

1.2 Prevalence and symptoms of PD

PD affects approximately 70,000 Australians (Deloitte, 2015) and over 5 million people globally (WHO, 2008). However, because the primary risk factor for the disease is ageing, as populations age this number is expected to increase, with some predictions estimating a doubling in the incidence of PD in developed nations between 2010 and 2040 (Dorsey et al., 2007; Kowal et al., 2013).

1 1. Introduction

PD is characterised by a number of typical movement related symptoms. These movement symptoms are caused by the degeneration of dopaminergic neurons in the substantia nigra region of the brain and the consequent decrease in dopamine levels (Halliday & McCann, 2010). This deficiency causes in a range of motor symptoms, including resting tremor, bradykinesia (movement slowing), loss of balance and muscle rigidity. The clinical standard for diagnosis requires the presence of bradykinesia as well as at least one of these other 3 movement related symptoms (Hughes et al., 1992).

Impaired neuronal function in other brain regions contributes to the many non-motor symptoms of PD that can substantially impair patient quality of life. These symptoms include depression and anxiety, hyposmia (olfactory dysfunction), sleep disturbances and autonomic dysfunction, which manifests in constipation, hypotension and sexual dysfunction (Davie, 2008). These non-motor symptoms typically occur at an earlier disease stage (Bonnet et al., 2012), prior to the loss of substantia nigra neurons and could potentially be used to aid earlier disease diagnosis, however this diagnostic approach is complicated due to the overlap of these symptoms with other neurodegenerative conditions, such as AD, Dementia with Lewy Bodies and Multiple System Atrophy (Koga et al., 2015).

1.3 Parkinson’s disease progression

Neuronal damage and dysfunction in PD follows a pattern of spreading pathology throughout the brain. This spread is characterized by the accumulation of large Lewy body or fibrillar Lewy neurite protein aggregates that are primarily composed of αSynuclein (αSyn) (Spillantini et al., 1997).

The extent of pathology and neuronal degeneration in PD worsens throughout the disease (Figure 1.1) and is associated with six progressive stages of the disease, which

2 1. Introduction are known as Braak stages (Figure 1.2). During the early, pre-diagnostic disease stages (Braak stage I/II) pathology develops from the vagus nerve and damage is limited to the brain stem and olfactory bulb, resulting in olfactory and autonomic disturbances that often go unrecognized by the patient (Doty, 2012; Ferrer et al., 2011). As the disease progresses this pathology spreads to the midbrain. At stage III Lewy pathology becomes evident in the dopaminergic neurons of the substantia nigra pars compacta (Braak et al., 2004) and by stage IV there is evident neuronal loss in this region, which results in the development of classical PD motor symptoms (Braak et al., 2004). Consequently, diagnosis typically occurs at this stage.

Figure 1.1 Lewy body and Lewy neurite pathology in the dorsal motor nucleus of PD patients. Adapted from Braak et al., (2004). Arrowheads in (A) indicate (i) Lewy body and (ii) Lewy neurite pathology in a stage 2 disease case. (B) Pathology worsens in stage 3.

At later disease stages (Braak stage V/VI) additional brain regions become dysfunctional and display pathology, including the putamen and caudate nucleus, amygdala, hippocampus and eventually the prefrontal and neocortex (Braak et al., 2004). Pathology in these regions manifests in severe cognitive and emotional symptoms, including dementia, confusion, depression, apathy and psychosis (Coelho & Ferreira, 2012).

PD diagnosis generally occurs between 55-70 years of age and progression to the severe disease state occurs over the following 10-15 years (Forsaa et al., 2010). Is it unknown how much earlier the initial Braak I and II stages of pathology become

3 1. Introduction evident, however a recent long-term study has shown evidence of initial disease pathology as early as 23 years prior to diagnosis and development of movement symptoms (Darweesh et al., 2017).

Figure 1.2 Parkinson’s disease progression as defined by Braak staging. The 6 Braak stages of PD are charaterised by the progression of neuronal damage in particular brain regions and the development of accompanying symptoms. Adapted from Braak et al., (2004) and Doty (2012).

Atypical variants of PD have been described which do not follow this characteristic progression or age of onset. In particular, many instances of PD that have a genetic basis, as will be discussed in section 1.5, have a much younger age of onset. While classical PD is diagnosed between 55 and 70, diagnosis between 21 and 40 years of age is considered early onset and diagnosis at <21 years of age is considered to be juvenile- onset parkinsonism (Davis et al., 2016; Di Fonzo et al., 2007). Age of onset appears to influence the progression of the disease, with patients diagnosed later experiencing more rapid symptomatic progression, more severe motor defects and displaying more pronounced postural instability (Diederich et al., 2003; Schrag & Schott, 2006).

4 1. Introduction

1.4 Treatment of PD

Current treatments for PD are targeted towards controlling the symptoms of the disease. No treatments are available to slow or stop disease progression and there are no treatments capable of regenerating lost neurons.

The benchmark therapeutic treatment for PD is supplementation with levodopa (L- dopa), the precursor to dopamine. As PD is characterized by the loss of dopaminergic neurons and subsequent reduction in dopamine levels, supplementation with L-dopa targets the symptoms that are caused by the loss of dopamine, primarily the movement associated symptoms (Nagatsua & Sawadab, 2009). However the effectiveness of L-dopa treatment is limited and deteriorates over time (Marsden & Parkes, 1977). Additionally, some reports suggest that treatment with L-dopa accelerates disease progression (Morgan & Sethi, 2005; Parkkinen et al., 2011), making it a controversial treatment option.

Treatment options for PD are limited by the fact that diagnosis typically doesn’t occur until the onset of motor symptoms, at which point substantial neuronal loss and damage has already occurred (Marsden, 1990; Ross et al., 2004). Extensive research efforts are being put towards developing diagnostic approaches that enable earlier diagnosis and improved diagnostic accuracy. This would substantially expand the possible treatment options as it would allow earlier, pre-symptomatic stages of the disease to be targeted, expanding therapeutic options to include preventative treatments that slow or stop progression prior to the degenerative, symptomatic phases of the disease.

Additionally, it is becoming apparent that there are likely to be numerous subtypes of PD that have unique responses to different treatments (Coelln & Shulman, 2016; Thenganatt & Jankovic, 2014). This probable existence of multiple disease subtypes limits the development of new therapeutic strategies as there is currently no reliable method for stratifying patients into likely responders and non-responders. Improved

5 1. Introduction patient stratification will enable the development of more personalized treatments and will significantly accelerate therapeutic developments to improve overall patient care.

1.5 Genetic contributors to PD

PD can be broadly categorized into two forms; the familial form (fPD), which accounts for approximately 10% of cases (Thomas & Beal, 2007) and can be traced back to a specific gene mutation, and the sporadic form (sPD), for which there is no known cause. Studying the gene mutations that are associated with fPD has provided significant insight into the underlying molecular factors that contribute to the disease. A number of PD-associated gene mutations have been described to date that range from low penetrance/high frequency to high penetrance/low frequency mutations (Figure 1.3). These PD-causative genes, known as PARK genes, as well as other PD- associated risk genes are discussed in detail below.

Figure 1.3 Overview of PD associated genes indicating their relative frequency and associated disease risk as a measure of allele penetrance. Adapted from Manolio et al.,

6 1. Introduction

(2009) with permission. Bubble size approximates population allele frequencies, blue bubbles represent dominant mutations, orange bubbles represent recessive mutations and green bubbles represent risk loci, while grey bubbles represent genetic associations that are yet to be fully characterised.

1.5.1 Confirmed PD associated gene mutations and risk factors

The identification of a missense mutation in the SNCA gene in 1997 (Polymeropoulos et al., 1997) was the first major indication of a genetic basis for PD and SNCA allele duplications and triplications have since been identified as a cause of autosomal dominant PD (Farrer et al., 1999). However, the full complexity of the genetic basis of fPD is only beginning to be uncovered due to the low penetrance or low frequency of the mutations involved. As whole genome and exome sequencing has become more accessible multiple PD causative gene mutations and variants have been identified. Table 1.1 describes the gene mutations that have been confirmed to cause autosomal recessive or autosomal dominant PD and their proposed functions.

Investigating the function of the proteins encoded by PD-associated genes has substantially expanded the understanding of the molecular mechanisms that contribute to the disease. For instance, the observation that many of the mutations occur in genes encoding proteins that have mitochondrial roles, such as PINK1, PARKIN and DJ1, implicates a central role of mitochondrial dysfunction in PD pathogenesis, as will be discussed in section 1.11. Other PD genes encode proteins that are associated with the endolysosomal network or regulate vesicular trafficking, such as LRRK2, ATP13A2, PLA2G6, VPS35, DNAJC16 and GBA (Table 1.1), suggest that lysosomal dysfunction may also be a key aspects of the disease, as will be discussed in section 1.9.

In addition to heritable PD-causative gene mutations, multiple genetic susceptibility risk factors for PD have also been described, such as mutations in the glucocerebrosidase (GBA) gene. GBA is a lysosomal enzyme that cleaves glucoceramide. Homozygous mutations in the GBA gene cause Gaucher’s disease, a

7 1. Introduction

Table 1.1 Frequency, mechanism of inheritance, and clinical characteristics associated with confirmed PD-causative gene mutations. Information obtained from Ferreira & Massano, 2016; Hernandez et al., 2016; Puschmann, 2013 and Lu et al., 2012. Classical PD refers to cases in which diagnosis occurs at an age > 50 years.

Disease Gene Protein Model Phenotype Population Function and clinical notes Designa- Frequency tion

PARK1/ SNCA αSynuclein Autosomal Autosomal Point mutations: Function unknown. May play a role in vesicular PARK4 dominant dominant: Very rare neurotransmission, synaptic vesicle regulation & risk locus Early-onset Duplications & or dopamine transmission, as detailed in section PD triplications: Rare 1.8 PARK2 PARK2 PARKIN Autosomal Early-onset Relatively E3 ubiquitin ligase involved in PINK1/PARKIN- recessive PD frequent among mediated mitophagy and ubiquination of patients with very proteins for proteasome-dependent young onset degradation PARK6 PINK1 Pten-induced Autosomal Early-onset Very rare Mitochondrial protein that detects putative kinase 1 recessive PD mitochondrial dysfunction and initiates (PINK1) PINK1/PARKIN-mediated mitophagy PARK7 PARK7 DJ-1 Autosomal Early-onset Very rare to Protein of several functions, including recessive PD exceedingly rare protecting cells from oxidative stress and serving as a chaperone molecule

8 1. Introduction

PARK8 LRRK2 Leucine-rich Autosomal Classical PD Founder effects in Unknown function. Belongs to the Roco protein repeat kinase 2 dominant specific family, with a Roc GTPase domain, a COR Lysosomal type populations. Most dimerization region, and a protein kinase 5 common form of domain. Recent results demonstrate a role of familial PD LRRK2 in cytoskeletal dynamics, vesicular transport, and autophagy PARK9 ATP13A ATPase Autosomal Early-onset Exceedingly rare P type ATPase. Possible metal translocation 2 recessive PD capabilities. Associated with Kufor-Rakeb disease, a rapidly progressive, juvenile-onset parkinsonism. PARK14 PLA2G6 Phospholipase Autosomal Early-onset Unknown Release of fatty acids from phospholipids A2 recessive PD PARK17 VPS35 Sorting 35 Autosomal Classical PD Very rare Component of the retromer cargo-recognition Eukaryotic dominant complex critical for the endosome to trans-Golgi translation trafficking and the recycling of membrane- associated proteins PARK18 EIF4G1 Initiation factor Autosomal Classical PD Very rare Operates as a scaffold protein that interacts 4 dominant with the 40S ribosome through translation initiation factors, including PABP, eIF3, and two eIF4F components PARK19 DNAJC1 Gamma 1 Autosomal Early-onset Very rare Neuronal protein that plays a role in the 6 DNAJ/HSP40 recessive PD pathway of clathrin-mediated endocytosis homolog subfamily C member 6

9 1. Introduction juvenile-onset lysosomal storage disease characterized by accumulation of glucoceramides within the cell (Rosenbloom & Weinreb, 2013). Gauchers disease is associated with liver cirrhosis, enlargement of the spleen and neurological symptoms that can include impaired cognition, hypertonia and dementia (Rosenbloom & Weinreb, 2013). The observation that Gaucher’s patients have increased risk of PD (Neudorfer et al., 1996) led to the finding that heterozygous GBA mutations are the most common genetic risk factor for PD (Rana et al, 2013). The occurrence of heterozygous GBA mutations in PD patients is estimated to confer a 30% increase in the risk of developing PD (Anheim et al., 2012) and have an Odds Ratio (OR) of 5.43 (Sidransky et al., 2009), where OR describes the association between presence of the mutation and development of PD. However, despite this significant association the molecular basis of this link between Gaucher’s disease, PD and GBA remains unknown.

1.5.2 Other genetic factors implicated in PD

Advances in sequencing technology and data analysis has resulted in the identification of a number of additional PD associated genes (Table 1.2). Although the association of these gene mutations with PD is yet to be fully verified, and some appear to be present only in specific populations (Rubino et al., 2017; Yang et al., 2016), further characterization of these genetic variants and the possible mechanisms through which they could contribute to PD will add to our overall understanding of disease etiology. Of particular relevance to this project, mutations within the CHCHD2 gene have recently been identified to cause late-onset autosomal dominant PD (Funayama et al., 2015). The proposed molecular mechanism through which impaired CHCHD2 activity could contribute to PD is discussed in detail in section 4 of this thesis. Additionally, mutations that perturb the splicing of ATP6AP2 have been associated with X-linked parkinsonism with spasticity (Gupta et al., 2015; Korvatska et al., 2013). ATP6AP2 is an accessory protein involved in regulating assembly of the Vacuolar ATPase (V-ATPase). The function of the V-ATPase and its possible contribution to PD pathogenesis will be discussed in depth in section 5 of this thesis.

10 1. Introduction

Table 1.2 Genes with unconfirmed or controversial association with PD. Information obtained from Ferreira & Massano, 2016; Hernandez et al., 2016; Puschmann, 2013; Burchell et al., 2013; Korvatska et al., 2013; Miyake et al., 2012; Oliveira et al., 2005; West et al., 2001 and Yuan Zhang et al., 2015b.

Former Gene Protein Model Phenotype Gene function and clinical notes Disease Designation

PARK3 PARK3 Unknown Autosomal Classical PD Unknown dominant PARK5 UCHL1 Ubiquitin c Autosomal Classical PD Thiol protease that hydrolyzes a peptide bond at the C-terminal terminal hydrolase dominant glycine of ubiquitin PARK10 PARK10 Unknown Risk locus Unconfirmed May play a role in determining age of onset of PD

PARK11 GIGYF2 GRB interacting Autosomal Classical PD May be involved in tyrosine kinase receptor signalling GYF protein 2 dominant PARK12 PARK12 Unknown X-linked Unconfirmed Unknown PARK13 HTRA2 HTRA serine Autosomal Unconfirmed An endoplasmic reticulum and mitochondrially localised serine peptidase 2 dominant protease thought to play a role in regulation of apoptosis. PARK15 FBXO7 F-box only protein Autosomal Early-onset May play a role in mitochondrial maintenance through direct 7 recessive PD interaction with PINK1 and PARKIN - CHCHD2 CHCHD2 Autosomal Classical PD Proposed transcription factor that may bind to and activates a dominant conserved oxygen response element in the COX4I2 gene, a nuclear gene that encodes a subunit of OXPHOS complex IV (cytochrome c oxidase). Alternatively, may regulate mitochondrial cristae structure and cytochrome c availability

11 1. Introduction

- ATP6AP2 ATPase H+ X-linked X-linked Accessory protein for the assembly of the vacuolar ATPase Transporting dominant adult-onset Patients develop early onset spasticity, which develops into Accessory Protein parkinsonism symptoms 2/ Prorenin AD-like Tau pathology has been observed post-mortem Receptor - RAB39B RAB interacting X-linked X-linked A small GTPases involved in the regulation of vesicular protein 29B dominant early-onset trafficking between membrane compartments PD

12 1. Introduction

1.6 Environmental and lifestyle contributors to PD

The low penetrance associated with many of the known PD gene mutations and the observation that there is no apparent genetic basis for the disease in the approximately 90% of PD cases that are sporadic (Thomas & Beal, 2007) suggests that environmental and lifestyle factors also contribute to the development of the disease. The initial indication that there is an environmental aspect to the disease was based on the observation that the Spanish influenza pandemic of 1918 was associated with a spike in PD diagnoses, indicating that exposure to infectious agents contributes to increased PD risk (Poskanzer & Schwab, 1963; Vlajinac et al., 2013). A number of environmental and lifestyle factors are now recognized to contribute to increased or decreased PD risk, as discussed in detail below.

1.6.1 Pesticide exposure increases PD risk

Prolonged exposure to the pesticides rotenone and paraquat was first recognized as a potential environmental risk factors for PD in the late 1980s (Ho et al., 1989) and has been confirmed in multiple large-scale epidemiological studies (Dhillon et al., 2008; Furlong et al., 2015; Tanner et al., 2011). Exposure may occur through inhalation during pesticide application, skin-to-skin contact or consumption of contaminated food or water and studies indicate that the level of PD risk increases with increasing levels of exposure (Anderson & Meade, 2014; Gorell, et al., 1998; Smargiassi et al., 1998).

Rotenone and paraquat have different modes of toxicity and therefore appear to increase PD risk through different mechanisms. While rotenone is a high affinity inhibitor of the mitochondrial OXPHOS complex I (Lambert & Brand, 2004), paraquat toxicity is caused by elevated levels of oxidative stress (Blanco-Ayala et al., 2014), suggesting both oxidative stress and impaired mitochondrial OXPHOS activity

13 1. Introduction contribute to PD. These features of PD are discussed in detail in sections 1.11 and 1.10, respectively.

1.6.2 MPTP injection induces PD in humans

1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) first recognized as a PD mimic in 1984 when a group of intravenous drug users in California spontaneously developed parkinsonism following injection of MPTP contaminated synthetic opioids (Langston & Ballard, 1984). These patients displayed typical Parkinson’s symptoms with L-Dopa responsive movement symptoms (Langston & Ballard, 1984).

The mechanism through which MPTP induces the death of dopaminergic neurons has been studied in detail and has provided significant insight into possible mechanisms underlying PD. MPTP is metabolized to neurotoxic MPP+ in glial cells and enters neuronal cells through interaction with the dopamine receptor, thereby specifically accumulating in dopaminergic neurons (Watanabe et al., 2005). MPP+ then mediates a toxic response in these cells by impairing OXPHOS complex I activity, resulting in decreased ATP production and increasing the levels of mitochondrial oxidative stress, which eventually triggers a cell death response (Hare et al., 2013). These data have added substantial support to the proposed role of mitochondrial impairment in PD.

1.6.3 Other risk factors for PD

A number of other lifestyle and environmental factors have also been proposed to increase PD risk. For instance, increased PD risk is associated with a number of other medical conditions, including diabetes (Bohnen et al., 2014; Lu et al., 2014; Santiago & Potashkin, 2014) and head trauma (Jafari et al., 2013; Noyce et al., 2012). Alternatively

14 1. Introduction the incidence of certain cancers, including lung, larynx and urinary bladder cancers (Olsen et al., 2006), has an inverse correlation with PD diagnosis (Driver, 2014).

Findings that PD patients accumulate higher levels of iron (Dexter et al., 1987), copper and zinc (Dexter et al., 1992) in the substantia nigra has led to the proposal that metal exposure may play a role in PD susceptibility (Smargiassi et al., 1998; Weisskopf et al., 2010). However, although there are rare reports of parkinsonism developing following magnesium (Jankovic, 2005) and lead (Kuhn et al., 1998) exposure, there is to date no convincing epidemiological evidence supporting an association between metal exposure and PD.

A number of factors that appear to abrogate PD risk have also been described. For instance, the lifestyle factors smoking and coffee consumption appear to reduce PD risk (Hernán et al., 2002). An inverse relationship between smoking and PD was first identified in 1960 (Dorn, 1960). Since then numerous studies have confirmed this inverse correlation and it is suggested that smoking confers up to a 50% reduced risk of PD, as reviewed by (Wirdefeldt et al., 2011). The mechanism through which smoking reduces PD risk has been suggested to be related to the neuroprotective activity of nicotine that is thought to be related to its ability to stimulate dopaminergic neurons (Quik, 2004) and inhibit αSyn fibrillation (Hong et al., 2009).

High levels of coffee consumption are also observed to reduce PD risk (Costa et al., 2010), however this effect is more prevalent in men than women (Ascherio et al., 2004). The protective mechanism through which caffeine may reduce PD risk is thought to involve interaction of caffeine with the Adenosine A2A receptor (Fernández-Dueñas et al., 2014), which plays a role in regulating dopamine and glutamate neurotransmission, however further work is required to better define this protective mechanism.

15 1. Introduction

1.7 Ageing is the primary risk factor for PD

Ageing is widely regarded as the primary risk factor for PD (Collier et al., 2011). This relationship between ageing and PD has been well characterized and is reflected in many in vivo models of the disease (Burré et al., 2010; Kamp et al., 2010; Kilpatrick et al., 2016; Varga et al., 2014). Furthermore, there is substantial overlap between the pathology and neuronal dysfunction observed in PD and that observed in normal ageing (Collier et al., 2017). For instance, the αSyn enriched Lewy body protein deposits that are the hallmark feature of PD are also evident in the substantial nigra of up to 20% of elderly individuals who are not affected by movement disorders (Markesbery et al., 2009). Similarly, the neuronal loss evident in the substantia nigra of PD patients is also evident in ageing, albeit not to the same extent. Substantia nigra neuronal loss associated with normal ageing is estimated to occur at a rate of 7% per decade (Stark & Pakkenberg, 2004) and may account for the impairment of motor function that is also often associated with ageing (Seidler et al., 2010). Therefore, although ageing alone cannot explain the extent of cell loss evident in PD, the observation of these overlapping features suggests that PD may be caused by additional environmental or genetic factors that accelerate the normal ageing process. Consequently understanding the molecular basis of ageing may provide significant insight into the cause of PD and other ageing related degenerative conditions.

1.8 αSynuclein accumulation is a pathological hallmark of PD

Studying the pathological consequences of PD has provided significant insight into the underlying molecular mechanisms that contribute to the disease. The accumulation of Lewy bodies and Lewy neurites in affected brain regions is the primary pathological feature of PD and is a requirement for PD diagnosis (Braak et al., 2004; Gelb et al., 1999). Both familial and sporadic PD are characterized by the accumulation of these large protein aggregates, which are primarily composed of the protein αSyn (Kalia &

16 1. Introduction

Kalia, 2015), suggesting that αSyn plays a central role in the disease. Additionally, familial forms of PD that are caused by duplication or triplication of the SNCA allele result in a 50% and 100% increase in protein abundance (Singleton et al., 2003), respectively, indicating that increased αSyn abundance is disease causative.

αSyn is a 140 amino acid protein that is expressed throughout the body, with the highest levels of expression observed in the brain (Jakes et al., 1994; Maroteaux et al., 1988; Taguchi et al., 2016). The protein has 3 distinct regions; an N-terminal apolipoprotein binding domain, a central hydrophobic region and a highly charged C- terminal region (Bisaglia et al., 2006; Bussell & Eliezer, 2003) and occurs in multiple forms, including monomers, dimers, tetramers, oligomers and fibrils (Lashuel et al., 2013). The protein is also observed to undergo multiple post-translational modifications, including oxidation, nitrosylation, non-enzymatic glycosylation and phosphorylation (Barrett & Timothy Greenamyre, 2015). Of particular interest, αSyn phosphorylated at the Serine129 residue is a major constituent of Lewy bodies (Fujiwara et al., 2002) and phosphorylation of this residue is observed to promote fibrilisation of the protein in vitro (Fujiwara et al., 2002), suggesting that this post- translational modification may alter the aggregation propensity of the protein.

The identity of the toxic species of αSyn in PD remains controversial. It has been suggested that αSyn toxicity is mediated by the oligomeric and fibrillar forms of the protein, which are the protein species primarily observed in Lewy body protein aggregates. Additionally, elevated expression of an αSyn variant that preferentially forms oligomers was observed to enhance neurodegeneration in vivo (Rockenstein et al., 2014) relative to expression of αSyn variants that do not oligomerize. Application of αSyn pre-formed fibrils in vitro or injection of pre-formed fibrils into the brain in vivo causes cell death and promotes cell-to-cell transmission of αSyn (Luk et al., 2012; Volpicelli-Daley et al., 2011), implicating the fibrillar species as a toxic form of the protein and suggesting that this form may act in a prion-like manner to spread αSyn pathology to connected brain regions (Allen et al., 2015; Xu et al., 2015a).

17 1. Introduction

Despite the substantial evidence that αSyn plays a role in both familial and sporadic PD, the endogenous function and the identity of the toxic species of the protein remains undetermined. Due to the predominant localisation of αSyn in the synaptic region of the cell (Maroteaux et al., 1988), αSyn was initially proposed to play a role in regulating synaptic function or neurotransmission. Supporting this proposal, αSyn is observed to interact with synaptic proteins to enhance formation of the SNARE protein complex that facilitate synaptic vesicle fusion with the synaptic membrane (Burré et al., 2010; Diao et al., 2013; Lai et al., 2014) and to regulate synaptic vesicle recycling (Cheng et al., 2011).

Non-synaptic related functions of αSyn have also been proposed. For instance the protein has been suggested to play a role in mediating innate and adaptive immune responses (Allen et al., 2015) and αSyn is also observed to impair many other aspects of cell biology. Excessive amounts of the protein are observed to impair mitochondrial functions (Cieri et al., 2017; Lindström et al., 2017; Noelker et al., 2015) possibly by preventing mitochondrial protein import (Di Maio et al., 2016) or perturbing mitochondria-Endoplasmic Reticulum (ER) contact sites (Guardia-Laguarta et al., 2015). These observations indicate that there may be a toxic interrelationship between αSyn and mitochondria in PD, which is explored in detail in section 3 of this thesis. Elevated αSyn levels also impede lysosomal transport (Volpicelli-Daley et al., 2014) and impair lysosomal degradation capacity through impairing trafficking of lysosomal enzymes from the trans-Golgi network (Mazzulli et al., 2016). These data implicate a role of lysosomal dysfunction in PD, as discussed below.

1.9 The endolysosomal system is perturbed in PD

The endolysosomal system is a complex network of compartments which includes early endosomes, late endosomes, multi-vesicular bodies, and mature lysosomes (Huotari & Helenius, 2011), as well as synaptic vesicles in neuronal cells (Watanabe et

18 1. Introduction al., 2014). This network of compartments regulates the processes of endocytosis, exocytosis, membrane trafficking, lysosomal degradation and lysoexocytosis, as well as synaptic vesicle maturation in neuronal cells (Goldenring, 2015; Scott et al., 2014). Consequently, endolysosomal dysfunction perturbs many cellular functions and metabolic pathways (Figure 1.4).

Figure 1.4 Impaired endolysosomal function is likely to play a central role in PD. The endolysosomal network is a system of compartments, including early and late endosomes and lysosomes, which regulate endocytosis and degradation of cellular components for turnover. The accumulation of αSyn in PD is suggestive impaired endolysosomal degradation capacity. Additionally, multiple PD-causative gene mutations (blue) are predicted to impair aspects of endolysosomal function.

Many confirmed PD-associated genes encode proteins that are associated with the endolysosomal network, implicating a role of endolysosomal dysfunction in PD (Figure 1.4) (Gan-Or et al., 2015). For instance the well-established PD genes ATP13A2 and GBA encode proteins that function in the late endosome and lysosome (Kong et al., 2014; Sidransky & Lopez, 2012), while depletion of VPS35 is observed to perturb endocytosis (Miura et al., 2014). Similarly, LRRK2 mutations, which are the most common cause of fPD, also perturb endolysosomal function (Gómez-Suaga & Hilfiker, 2012). Overexpression of wild type (WT) LRRK2 or expression of the common PD- causative G2019S variant decreases the number of lysosomes in the cell and impairs

19 1. Introduction the formation of the protein complex required to translocate cytosolic proteins into the lysosomal lumen for degradation (Henry et al., 2015). Additionally, PD-associated proteins that are more generally viewed as having mitochondrial functions, including PINK1 and PARKIN also affect lysosomes as these proteins play a direct role in regulating mitophagy (Truban et al., 2017), the process by which defective mitochondria are degraded in the lysosome.

One of the primary functions of the endolysosomal network is the degradation of macromolecules, such as proteins, lipids and nucleic acids through a process known as autophagy and impaired autophagy has been proposed to contribute to the accumulation of αSyn protein deposits in both sporadic and familial forms of PD (Alvarez-Erviti et al., 2013; Bae et al., 2015; Cuervo et al., 2004). In addition to being degraded by the proteasome, αSyn is also degraded through two separate lysosomal autophagy pathways; macroautophagy and chaperone mediated autophagy (CMA) (Vogiatzi et al., 2008). The molecular mechanisms of these degradation pathways and their proposed contributions to PD pathogenesis are outlined below.

1.9.1 Macroautophagy

Macroautophagy occurs over a number of tightly regulated stages, namely initiation, nucleation, elongation, fusion and degradation (Figure 1.5). During initiation a double membrane structure known as the phagophore develops (Bernard & Klionsky, 2014; Biazik et al., 2015). This membrane appears to be sequestered from the Golgi apparatus, mitochondria or endoplasmic reticulum (ER) (Chan & Tang, 2013; Lamb et al., 2013). During the subsequent nucleation stage, Atg proteins on the phagophore membrane recognise adaptor proteins that accumulate on the surface of the defective organelle, protein or lipid aggregate (Feng et al., 2014; Sica et al., 2015). The membrane then encloses around the cargo to form an autophagosome, which is subsequently fuses with a lysosome to form an autolysosome. This hybrid organelle

20 1. Introduction facilitates degradation of the autophagosome cargo by the lysosomal enzymes (Klionsky et al., 2014; Rong et al., 2015b).

Figure 1.5 Macroautophagy is the process through which protein and lipid aggregates and defective organelles are degraded in the lysosome. Macroautophagy occurs over 5 distinct stages; initiation, nucleation, elongation, fusion and degradation. Adapted from Delbridge et al., (2015) with permission.

In addition to the proposal that impaired macroautophagy contributes to αSyn accumulation in PD (Vogiatzi et al., 2008), αSyn accumulation is also proposed to impair macroautophagy at the initiation stage and lead to reduced autophagosome clearance. This effect is proposed to be caused by αSyn-mediated inhibition of Rab1a, which controls trafficking of enzymes, including autophagy receptor proteins, from the ER to the trans-Golgi network (Tanik et al., 2013; Winslow et al., 2010). Therefore the accumulation of αSyn in PD may initiate a feed-forward cycle of degeneration wherein αSyn accumulation perturbs macroautophagy, which in turn exacerbates the accumulation of αSyn protein deposits.

21 1. Introduction

1.9.2 Chaperone Mediated Autophagy

CMA is the process by which specific cytosolic proteins are degraded in the lysosome by direct transfer of these substrates into the lysosomal lumen (Arias & Cuervo, 2011). Proteins for degradation through this pathway are recognised by cytosolic chaperone proteins through the presence of a KFERQ protein domain (Dice, 1990). αSyn contains the KFERQ domain necessary for CMA (Vogiatzi et al., 2008) and has been confirmed to undergo CMA-mediated degradation (Cuervo et al., 2004). Consequently, increasing CMA activity has been suggested as a potential therapeutic approach for PD (Xilouri & Stefanis, 2015).

CMA substrates are translocated into the lysosomal lumen for degradation through interaction with the lysosomal membrane protein LAMP2A. This translocation of substrates into the lysosome is a rate-limiting step in CMA (Cuervo & Dice, 1996) and reduced LAMP2A levels observed in affected brain regions of PD patients (Murphy et al., 2015) are likely to be indicative of reduced CMA activity in PD. Additionally, the decrease in LAMP2A levels in patient brain tissue inversely correlates with increased abundance of αSyn protein (Murphy et al., 2015), suggesting that reduced CMA activity in PD contributes to the accumulation of αSyn protein that is evident in the disease.

1.10 Elevated production of mitochondrial reactive oxygen species is a hallmark of PD

Mitochondria are the major site of reactive oxygen species (ROS) production within the cell, which are produced as a by-product of mitochondrial electron transport (Figure 1.6). ROS are formed when electrons are transferred from OXPHOS complexes I

- and/or III to molecular oxygen, producing superoxide radicals (O2 ) that can subsequently react to form other ROS, including hydrogen peroxide (H2O2) and

22 1. Introduction hydroxyl radicals (HO-) (Muller et al., 2004). At low levels these ROS can act as signalling molecules that activate protective autophagy processes (Guan-yu Liu et al., 2015a; Haiyuan Zhang et al., 2014b) and increase expression of antioxidant defence ROS scavengers, such as glutathione and NADPH (D'Autréaux & Toledano, 2007). However, above a threshold level the detrimental effects of ROS outweigh the antioxidant defence system, resulting in ROS-induced toxicity. In particular, high levels of ROS result in oxidation of the side chains of lysine, arginine, proline and threonine amino acid residues on proteins, causing protein carbonylation that can result of loss of function (Suzuki et al.,2010). Increased levels of protein carbonylation are evident in PD patient brain tissue (Alam et al., 1997; Floor & Wetzel, 1998), strongly suggesting that levels of ROS are increased in the disease. Further supporting this suggestion, ROS can also react with polyunsaturated lipids to cause lipid peroxidation or react with DNA to cause DNA damage, both of which are also observed to be increased in PD patients (Dexter et al., 1989; 1994; J Zhang et al., 1999).

Figure 1.6 Mitochondrial Reactive Oxygen Species (ROS). ROS are produced as a byproduct of mitochondrial OXPHOS at complexes I and/or III.

23 1. Introduction

Further supporting the suggestion that ROS plays a pathogenic role in PD, levels of antioxidant molecules appear to be reduced in affected brain regions of patients. For instance, reduced levels of the antioxidant glutathione are one of the first molecular alterations evident in the substantia nigra of PD patients, occurring prior to the appearance of αSyn pathology (Jenner, 1993; Zeevalk et al., 2008). This suggests that perturbed ROS homeostasis plays an early role in the disease. However the failure of numerous antioxidant based clinical trials to improve PD symptoms (Parkinson Study Group QE3 Investigators et al., 2014; Shoulson et al., 2002; Snow et al., 2010; Storch et al., 2007) suggests that the mechanism through which ROS contributes to PD is complex and it is possible that while ROS is pathogenic in some brain regions, in other affected brain regions sub-toxic increases in ROS levels may have a protective effect by activating antioxidant defences and/or autophagy (D'Autréaux & Toledano, 2007). Improved understanding of the mechanisms that regulate ROS detoxification and signalling will significantly improve the current understanding of how ROS contributes to PD and mechanisms through which damaging ROS could be targeted therapeutically.

1.11 Mitochondrial damage is prevalent in PD

Mitochondrial dysfunction appears to play a central and pathogenic role in PD. In particular, defective turnover of damaged mitochondria through a form of autophagy referred to as mitophagy is implicated as a pathogenic mechanism in the disease due to the identification of PD-causative mutations in the genes encoding PINK1 and PARKIN, which are proteins that regulate an early stage of this process. Reduced PARKIN activation (Dawson & Dawson, 2014) and accumulation of damaged mitochondria has also been observed in patients with non-PINK1/PARKIN associated fPD, implicating perturbed mitochondrial quality control more broadly in sPD (Pickrell & Youle, 2015).

24 1. Introduction

1.11.1 Mitophagy

Characterization of the functions of PINK1 and PARKIN has significantly enhanced understanding of the molecular mechanisms that underlie PINK1/PARKIN-mediated mitophagy and has implicated the loss of mitochondrial membrane potential as a key event in the removal of defective mitochondria by the cell. Healthy mitochondria maintain a positive electrochemical membrane potential through the activity of the proton pumping OXPHOS complexes I, III and IV (Figure 1.6). Impaired activity of these complexes, accelerated complex V-mediated usage of the proton gradient, or increased membrane permeability can dissipate mitochondrial membrane potential. This impairs ATP production and inhibits the membrane-potential dependent import of proteins into the mitochondria, including the serine/threonine kinase PINK1. Consequently, depolarisation of mitochondria leads to PINK1 accumulation on the mitochondrial surface (Jin et al., 2010; Meissner et al., 2011). PINK1 subsequently phosphorylates ubiquitin and the E3 ubiqutin ligase PARKIN, leading to selective recruitment of activated PARKIN to the surface of damaged mitochondria (Koyano et al., 2014). Thereby PINK1/PARKIN accumulation on the surface of damaged mitochondria plays an early role in “tagging” defective mitochondria for degradation (Figure 1.7).

PINK1 and PARKIN on the surface of damaged mitochondria facilitate the recruitment of the autophagosome membrane during the nucleation phase of mitophagy (Figure 1.5). Specifically, PINK1 generated phospho-ubiquitin, as well as ubiquitinated PARKIN substrates on the outer mitochondrial membrane act as signals to recruit autophagy receptors to the damaged mitochondrion (Lazarou et al., 2015). These autophagy receptors interact with LC3 on the phagosome membrane to mediate nucleation (Jin & Youle, 2012). The autophagosome membrane subsequently elongates to isolate the mitochondria in a mature autophagosome, which fuses with an acidic lysosome to form an autolysosome (Feng et al., 2014), thereby enabling degradation of the mitochondrial contents through the activity of lysosomal hydrolases (Figure 1.7).

25 1. Introduction

Figure 1.7 PINK1/PARKIN-mediated mitophagy. Damaged mitochondria are recognised through their loss of membrane potential resulting in PINK1 and PARKIN accumulation on the mitochondrial surface. This leads to ubiquitination of mitochondrial outer membrane proteins, which recruits the autophagosome membrane and facilitates degradation of mitochondrial contents.

1.11.2 Other mechanisms of mitochondrial quality control

In addition to mitophagy, a number of mechanisms exist to enable mitochondrial quality to be monitored and maintained. In eukaryotes, a mitochondrial quality control mechanisms known as the mitochondrial unfolded protein response (mtUPR) has

26 1. Introduction recently been described. The mtUPR enables cells to sense mitochondrial damage and transcriptionally activate expression of a subset of genes (Haynes et al., 2010; Martinus et al., 1996), including mitochondrial proteostasis genes, to alleviate this stress (Aldridge et al, 2007; Fiorese et al., 2016). If mitochondrial function cannot be recovered a number of mechanisms, such as PINK1/PARKIN-mediated mitophagy, exist to remove mitochondria. Additionally, alternative degradation mechanisms have also been described more recently that facilitate selective turnover of mitochondrial contents. For instance, the formation of mitochondrial-derived vesicles (McLelland et al, 2014; Soubannier et al., 2012) and mitochondrial-derived compartments (Hughes et al, 2016) enable the degradation of a particular subset of mitochondrial proteins in response to mitochondrial damage or lysosomal stress, respectively (Hughes et al., 2016; Soubannier et al., 2012). Although the potential contribution of these pathways to PD is yet to be determined, enhanced formation of MDCs is associated with ageing (Hughes et al., 2016), suggesting that this specialised mitochondrial degradation pathway may also be associated with ageing related diseases, such as PD.

1.12 Project Aims

The overall aim of this study was to identify and explore novel and early molecular and cellular changes that contribute to the degeneration evident in brain regions of sporadic PD patients. To achieve this outcome this study explored the contributions of both mitochondrial dysfunction (induced by rotenone) and αSyn accumulation to PD. The proteins and cellular pathways that these PD-relevant insults perturb individually and in combination were assessed at multiple time points using a SILAC mass spectrometry approach. Assessment of multiple time points enabled early changes to be distinguished from later events, which are more likely to be downstream consequences. The proteomic data obtained from this assessment was subsequently intersected with proteomic and transcriptomic data obtained from patient brain regions in order to obtain a prioritized list of candidate proteins and cellular pathways

27 1. Introduction that are likely to be perturbed early in the disease course and are also evident in patient tissue (section 3). Therefore, the specific aims of this thesis were:

1: To identify early proteomic changes that occur in in vitro PD models, namely a rotenone model, an elevated αSyn expression model and a model wherein these two insults are combined.

2: To confirm that there is evidence that the proteomic changes identified in vitro also occur in affected patient brain regions in order to prioritize key areas of interest.

3: To explore in detail the contributions and potential mechanism(s) through which dysfunction of two of these prioritized areas of interest contribute to PD, namely CHCHD2 (section 4) and the vacuolar ATPase (section 5).

28 2. Materials and methods

2 MATERIALS AND METHODS

2.1 General Materials and Methods

2.1.1 Chemicals, consumables and equipment

The source of all materials and reagents used in this study are listed in Table 2.1, unless otherwise stated.

Table 2.1 List of materials and reagents used in this study.

Material/Reagent Source 1 KB DNA ladder ThermoFisher Scientific, AUS 2-mercaptoethanol Sigma-Aldrich, NSW 2-propanol Sigma-Aldrich, AUS 2xYT Fisher Scientific, USA Acrylamide USB Corporation, USA Agarose USB Corporation, USA Antimycin A Sigma-Aldrich, AUS Ammonium Persulphate Sigma-Aldrich, AUS BafilomycinA1 Sigma-Aldrich, AUS Bis-acrylamide USB Corporation, USA Blasticidine S Sigma-Aldrich, AUS Bovine Serum Albumin (BSA) Sigma-Aldrich, AUS Carbenicillin Gold Biotechnology Inc, USA Carbonylcyanide-m-chlorophenylhydrazone Sigma-Aldrich, AUS (CCCP) Chloroform Sigma-Aldrich, AUS Dimethyl sulfoxide (DMSO) Sigma-Aldrich, AUS DMEM basal media ThermoFisher Scientific, AUS DMEM/F12 basal media ThermoFisher Scientific, AUS Doxycycline (DOX) BD, AUS Dithiothreitol Sigma-Aldrich, AUS Ethanol (absolute) Sigma-Aldrich, AUS Ethidium bromide Sigma-Aldrich, AUS Ethylenediaminetetraacetic acid (EDTA) Sigma-Aldrich, AUS FBS ThermoFisher Scientific, AUS GlutaMAX ThermoFisher Scientific, AUS Glycerol Sigma-Aldrich, AUS GlycoBlue TM Life Technologies, AUS

29 2. Materials and methods

Guanosine Sigma-Aldrich, AUS HEPES Sigma-Aldrich, AUS Immobilon HRP substrate Millipore, MA, USA Iodoacetamide Sigma-Aldrich, AUS Kanamycin Sulphate Sigma-Aldrich, AUS Lipofectamine 2000 ThermoFisher Scientific, AUS Lipofectamine 3000 ThermoFisher Scientific, AUS Lys-C Roche, AUS Methanol Univar, AUS MilliQ water Millipore, MA, USA MitoSoxTM Red Life Technologies, AUS Non-Essential Amino Acids ThermoFisher Scientific, AUS Oligofectamine ThermoFisher Scientific, AUS OptiMEM ThermoFisher Scientific, AUS PageRuler Plus pre-stained protein ladder ThermoFisher Scientific, AUS Paraformaldehyde Sigma-Aldrich, AUS Phosphate Buffered Saline (PBS) Sigma-Aldrich, AUS Phusion polymerase New England Biolabs, AUS Protease inhibitors- Complete EDTA-free Roche, AUS Restriction Endonucleases New England Biolabs, AUS Retinoic Acid Sigma-Aldrich, AUS Rotenone Sigma-Aldrich, AUS RPMI basal media ThermoFisher Scientific, AUS Saponin Sigma-Aldrich, AUS SensiMix NoROX Bioline, AUS Sodium acetate ThermoFisher Scientific, AUS Sodium azide Sigma-Aldrich, AUS sodium dodecyl sulphate (SDS) Sigma-Aldrich, AUS Sodium hydroxide Sigma-Aldrich, AUS Superscript III First-Strand Synthesis Life Technologies, AUS TBS Sigma-Aldrich, AUS TEMED Sigma-Aldrich, AUS Thapsigargin Sigma-Aldrich, AUS Thiourea ThermoFisher Scientific, AUS Triethylammonium bicarbonate (TEAB) Sigma-Aldrich, AUS Tris Sigma-Aldrich, AUS Trizol TM ThermoFisher Scientific, AUS TrypLE ThermoFisher Scientific, AUS Trypsin Sigma-Aldrich, AUS Tween-20 (Polysorbate 20) Sigma-Aldrich, AUS Urea ThermoFisher Scientific, AUS

30 2. Materials and methods

2.1.2 Plasmids used in this study

Table 2.2 Plasmids used in this study.

Plasmid Source Experimental usage pEGFP-N1-TFEB A gift from Shawn Ferguson TFEB translocation to the (Addgene plasmid # 38119) nucleus (Chapter 3) CMV-mito-GEM- A gift from Robert Campbell Measurement of GECO1 (Addgene plasmid # 32461) mitochondrial calcium level (Chapter 3) CMV-mito-GEM- A gift from Robert Campbell Measurement of cytosolic GECO1 (Addgene plasmid # 32461) calcium level (Chapter 3) MKeima Obtained with permission Measurement of lysosomal from (Katayama et al., 2011) uptake of defective mitochondria (Chapters 2 and 3)

2.1.3 Culturing of bacterial strains and plasmid preparation

2.1.3.1 Bacterial strains and culture conditions

The Escherichia coli (E. coli) MC1061 strain was used for cloning and propagation of most plasmids. STBL3 (ThermoFisher) or Top10 (ThermoFisher) were used for propagation of lentiviral plasmids or when high efficiency transformation was required. Details of these strains are outlined in Table 2.3. Strains were cultured in 2xYT liquid (1.6% tryptone, 1% yeast extract, 0.5% NaCl) with constant shaking, or on solid LB agar (1% milk peptones, 0.5% NaCl, 0.18% NaOH, 1.5% agar, 0.5% yeast extract) at 37°C overnight. For plasmid selection culture media was supplemented with either 100 μg/ml carbenicillin (ampicillin) or 50 μg/ml kanamycin.

Table 2.3 E. Coli strains used in this study.

Strain Description Source MC1061 F-Δ(ara-leu)7697 [araD139]B/r ThermoFisher Scientific Δ(cod-B-lacI)3 galK16 galE15(GalS) Λ- Australia

31 2. Materials and methods

e14-mcrA0 relA1 rpsL150(strR) spoT1 mcrB1 hsdR2 STBL3 (F–mcrB mrr hsdS20 (rB–, mB–) recA13 ThermoFisher Scientific supE44 ara-14 galK2 lacY1 proA2 rpsL20 Australia (StrR ) xyl-5 λ– leu mtl-1) Top10 (F-mcrAΔ(mrr-hsdRMS-mcrBC) ThermoFisher Scientific φ80lacZΔM15 ΔlacX74 recA1 araD139 Australia Δ(araleu)7697 galU galK rpsL (StrR) endA1 nupG

2.1.3.2 Bacterial transformation

Transformation competent Top10 and STBL3 cells were purchased ready to use (ThermoFisher Scientific), while MC1061 cells were made competent using the calcium chloride method (Mandel & Higa, 1970).

For MC1061 transformations, cells were thawed on ice and 500 ng of plasmid DNA was added. The reaction was incubated for a further 10 minutes on ice and then heat shocked at 37°C for 10 minutes. For recovery 300 μL of 2xYT liquid was added and cells were incubated for a final 10 minutes at 37°C. Cells were then plated onto selective LB agar plates and incubated overnight at 37°C.

STBL3 and Top10 cells were transformed as per the manufacturers protocol (ThermoFisher Scientific). Briefly, cells were thawed on ice, 500 ng of plasmid DNA was added and the reaction was incubated on ice for 30 minutes. Cells were heat shocked at 42°C for 45 seconds, 300 μL 2xYT liquid was added and cells were incubated while rotating at 37°C for a further 30 minutes to allow recovery. Cells were plated onto selective LB agar plates and incubated overnight at 37°C.

2.1.3.3 Plasmid DNA purification

Plasmids were extracted from overnight E. coli liquid cultures using the Wizard SV miniprep kit (Promega) for small-scale extractions or the PureLink Midiprep kit

32 2. Materials and methods

(ThermoFisher Scientific) for large-scale plasmid extractions, as per the manufacturer’s instructions. A NanoDrop ND-1000 spectrophotometer (ThermoFisher Scientific) was used to determine plasmid concentration and purity.

2.1.4 Culturing of mammalian cells

All tissue culture reagents, including FBS, RPMI media, DMEM media, GlutaMAX and Non-Essential Amino Acids were obtained from ThermoFisher Scientific, Australia.

2.1.4.1 SH-SY5Y cells

SH-SY5Y human neuroblastoma cells were obtained from the American Type Culture

Collection (VA, USA). Cells were cultured at 37°C with 5% CO2 in DMEM/F12 media supplemented with 10% FBS and 5mM glutaMAX. To induce cell differentiation FBS was reduced to 2% and 10 μM Retinoic Acid (RA) was added. Cells were passaged to a maximum of passage 19.

2.1.4.2 SH-SY5Y-αSyn and SH-SY5Y-βGal DOX-off cells

SH-SY5Y cells transduced to stably express wild type (WT) αSynuclein (or βGalactosidase control) under the control of a TET-OFF promoter were a gift from Kostas Vekrellis (University of Athens, Greece). Cells were cultured at 37°C with 5%

CO2 and passaged in RPMI media supplemented with 10% FBS and 5 mM glutaMAX. As per SH-SY5Y cells, cell differentiation was induced by reducing FBS to 2% and adding 10 μM RA. Cells were passaged to a maximum of passage 16.

2.1.4.3 Seeding density

SH-SY5Y, SH-SY5Y-αSyn and SH-SY5Y-βGal cells were seeded at a density of 16 x 104 cells/mL, with seeding volume varying dependent on plate size (Table 2.4).

33 2. Materials and methods

Table 2.4 SH-SY5Y seeding volume according to plate size.

Plate Size Seeding Volume (μL) 6-well 2500 12-well 1000 24-well 5000 96-well 90

2.1.4.4 Experimental timetable

All experimental work using SH-SY5Y and SH-SY5Y-αSyn/βGal cells was conducted using differentiated cells. Table 2.5 indicates the timetable of differentiation, cell treatment and sample harvest/assessment. All experiments were conducted using samples harvested from three or more biological replicates obtained from independent experimental set-ups.

Table 2.5 SH-SY5Y seeding and experimental timeline.

Day Treatment -1 Cells seeded in media with 10% FBS (as per Table 2.4) 0 Media changed to 2% FBS with RA Chemical inhibition added or siRNA-mediated knock-down initiated 1 24 hour samples harvested/assessed 2 Media refreshed as per day 0 or 48 hour samples harvested/assessed 3 72 hour samples harvested/assessed 4 96 hour samples harvested/assessed

2.1.4.5 HEK293FT cells

Human kidney epithelial HEK293 LentiX cells were a gift from Marc Giry-Laterriere (Garvan Institute of Medical Research). HEK293 LentiX cells were cultured for use in virus production (section 4) due to their high transfection efficiency. Cells were

34 2. Materials and methods

cultured at 37°C with 5% CO2 in DMEM/F12 media with 10% FBS, 5 mM glutaMAX and 5 mM Non-Essential Amino Acids.

2.1.5 Transfection into SH-SY5Y cells

2.1.5.1 Transient plasmid transfection

Transfection of plasmid DNA into SH-SY5Y cells was achieved using Lipofectamine2000 reagent (ThermoFisher Scientific, Australia) as per the manufacturers protocol. Briefly seeding media containing 10% FBS was replaced with 2% FBS media with RA immediately prior to transfection. Plasmid DNA and Lipofectamine2000 reagent were diluted in optiMEM media, combined by inversion and incubated for 20 minutes. DNA complexes were added to cells dropwise, with DNA at a final concentration of 1.6 μg/mL and lipofectamine2000 reagent at a final concentration of 4 μL/mL. Cells were then incubated for 4 hours before media was changed to fresh 2% media with RA and the chemical treatment of interest or siRNA (as per the siRNA transfection protocol outlined below).

2.1.5.2 Transfection of siRNAs for knockdown

Transfection of siRNA into SH-SY5Y cells was achieved using Oligofectamine reagent (ThermoFisher Scientific, Australia) as per the manufacturers protocol. Briefly, 10% seeding media was replaced with 2% media with RA immediately prior to transfection. SiRNA and oligofectamine were diluted in optiMEM media, combined and incubated for 20 minutes. siRNA complexes were added to cells dropwise, with siRNA at a final concentration of 25 nM and oligofectamine at a final concentration of 1.5 μL/mL.

SiRNAs used for knockdown of CHCHD2 in section and ATP6V1A and ATP6V1B2 in section 5 are outlined in Table 2.6.

35 2. Materials and methods

Table 2.6 Sequences and sources of the siRNAs used in this study.

siRNA Target Sequence Source Negative Non- Proprietary ThermoFisher control No. 5 targeting Scientific SilencerTM J-019120-09 CHCHD2 CGGAUUGGCCUAAUGAAGA Dharmacon ON- TARGETplus J-19120-10 CHCHD2 GUGAGGGUAUAAAGUGUAA Dharmacon ON-TARGETplus M-017590-00 ATP6V1A AGAAAUCUCUGGUCGUUUA Dharmacon GAAACUAGCUCAACGUAAG Smart Pool CAAUGGAGGUUGAUGGUAA GACGAAAGCUAAGGAAAUU M-011589-01 ATP6V1B2 GAGAAACAUGUAUUGGUA Dharmacon CGAUGUAUCUAACCAGCUA Smart Pool UCGAUGGGAUGAACAGUAU AGAAACGGCUCGAUUACUC

2.1.6 RNA isolation, cDNA synthesis and qRT-PCR

2.1.6.1 RNA isolation from tissue culture samples

Extraction of RNA from cell culture samples was conducted using the TrizolTM (ThermoFisher Scientific, Australia) method of purification. Media was removed from cells, replaced with 1 mL Trizol per 6-well and cells incubated at room temperature for 5 minutes. Samples were then moved to nuclease free microcentrifuge tubes and 200 μL chloroform was added. Samples were shaken vigorously for 15 seconds, incubated at room temperature for 3 minutes and centrifuged at 12,000 g for 15 minutes at 4°C. The upper aqueous phase was then transferred to a new tube with the addition of 0.5 μL GlycoBlueTM (ThermoFisher Scientific) and 500 μL 2-propanol and inverted 5 times to mix. This was incubated for 10 minutes at RT and centrifuged at 12,000 g for 10 minutes at 4°C. The supernatant was discarded and samples were centrifuged for a further 30 seconds to remove any residual supernatant. Pellets

36 2. Materials and methods were then washed twice by re-suspension in 500 μL cold 80% ethanol and centrifugation at 7,500 g for 5 minutes at 4°C. The supernatant was removed and the pellet was air dried for 5 minutes and resuspended in nuclease free water. RNA concentration and quality were determined using a NanoDrop1000 (ThermoFisher Scientific) and samples were stored at -70°C.

2.1.6.2 Conversion of RNA to cDNA

Conversion of RNA to cDNA was conducted using the SuperScriptIII first strand cDNA synthesis kit (ThermoFisher Scientific). 2 μg of each RNA sample was combined with 5 μM Oligo (dT) and 1 μM dNTPs in a final volume of 5 μL and incubated at 65°C for 5 minutes. Samples were incubated on ice for a further 2 minutes prior to the addition of 10 mM MgCl2, 20 mM DTT, 20 Units RNase Out and 100 Units SuperScriptIII made up in 1X Reverse Transcription buffer to a total volume of 10 μL. Samples were then incubated at 50°C for 50 minutes, supplemented with 1 Unit RNase H and incubated at 37°C for a further 20 minutes. cDNA samples were stored at -20°C until further use.

2.1.6.3 Quantification of transcript levels by qRT-PCR

Quantification of transcript levels was achieved by qRT-PCR using the SensiFAST SYBR no-ROX Kit (Bioline, Australia), as per the manufactures protocol. Samples were assessed in triplicate using 384-well PCR plates (ThermoFisher Scientific). A 10 μL reaction volume was used which contained 10 ng cDNA, 0.5 μM of each primer and 1X SensiFAST mix. Thermal cycling was conducted using a LightCycler480 Thermocycler (Roche), with cycling conditions outlined in Table 2.7.

Table 2.7 Thermal cycling conditions used for qRT-PCR.

Stage Temperature Time (sec) Ramp Rate Cycles (°C) (°C/sec) Initial 95 120 4.8 1 Denaturing

37 2. Materials and methods

Denaturing 95 5 4.8 45 (incremental Annealing 65-60 10 4.8 decrease in annealing Extension 72 15 4.8 temperature by 0.7°C per cycle) Melt Curve 95 120 4.8 1 40 180 4.8 40-95 Incremental temperature increase of 0.04 °C/sec with 16 measurements per sec Cooling 40 30 2.5 1

Gene expression was normalized to housekeeping gene expression using the ΔΔCT method described in the Minimum Information for Publication of qRT-PCR Experiments (MIQE) guidelines (Bustin et al., 2009). The expression of the gene of interest relative to the expression of housekeeping gene was calculated, following which the expression of the gene of interest in treated relative to untreated samples was calculated. As per the guidelines a minimum of two housekeeping genes were used as a reference for each test.

Primers used for qRT-PCR in this study are listed in Table 2.8. As noted in the table, primers were either designed using the NCBI Primer3 online platform or were custom designed by Sigma-Aldrich. Self-designed primers were optimized for qRT-PCR by ensuring primers were separated by at least one intron and amplified a fragment of 70-150 base pairs. Primer efficiency and specificity was assessed by melt curve analyses.

Table 2.8 Sequences of primers used in this study.

Gene Sequence (5’-3’) Additional notes B2M F TTCTGGCCTGGAGGCTATC Housekeeping Gene R TCAGGAAATTTGACTTTCCATTC HPRT F TGACCTTGATTTATTTTGCATACC Housekeeping Gene R CATCTCGAGCAAGACGTTCA

38 2. Materials and methods

MALAT F GACGGAGGTTGAGATGAAGC Housekeeping Gene R ATTCGGGGCTCTGTAGTCCT NFX1 F CAAACCTGCGCTAGAGTCCA Housekeeping Gene R GGGGATGTTGCTCCGAAACT ADNP F GGATGTAGGACTGTGGGACC Housekeeping Gene R CGCTGCAGCAGAAAGGTTTT ZFR F TGCTGTATCTGAAGCGGCAA Housekeeping Gene R TGGGTCTTTCACCATACCCG NDUFB8 F TACCCTCTCTGTCTCACC Chapter 4 & 5 R GCGTCCTGACCTGTTCTC OXPHOS complex I subunit SDHB F CTCCCACTTGGTTGCTCG Chapter 4 & 5 R GGCTTTCCTGACTTTTCCCT OXPHOS complex II subunit UQCRC2 F GCTCGTTATCATGTGGCAGA Chapter 4 & 5 R TCATGTCCAGCATCCTCTTG OXPHOS complex III subunit COX2 F CCCCACATTAGGCTTAAAAACAGAT Chapter 4 & 5 R TATACCCCCGGTCGTGTAGCGGT OXPHOS complex IV subunit ATP5A F TGGGTTCATCTTTCATTGCTGC Chapter 4 & 5 R GACACGCCCAGTTTCTTCAA OXPHOS complex V subunit SNCA F AGAAGCAGCAGGAAAGACAAAAG Chapter 3, 4 & 5 R TCTTTGGTCTTCTCAGCCACTGT ATP6V1A F AAACTTGATCCACATCACAC Chapter 5 R AAAACATAATGCCACTCTGG Sigma KiCQ Start oligo ATP6V1B2 F GATGGGATGAACAGTATTGC Chapter 5 R CATGTTTACACCCATAGCAG Sigma KiCQ Start oligo ATP6V0A F AATGGTTCTTCTGAATCAAACGG Chapter 5 R CATCTATGTTGCACAGGTTC Sigma KiCQ Start oligo TPP F GATCCCAGCTCTCCTCAATACG Chapter 5 R GCCATTTTTGCACCGTGTG CLEAR Gene CTSD F AACTGCTGGACATCGCTTGCT Chapter 5 R CATTCTTCACGTAGGTGCTGGA CLEAR Gene GNS F CCCATTTTGAGAGGTGCCAGT Chapter 5 R TGACGTTACGGCCTTCTCCTT CLEAR Gene MEF2D F GACATTAAAGTGACGATTCCC Chapter 5 R ACAAATGTAACCGTCAACAG CMA Substrate Sigma KiCQ Start oligo GAPDH F ACAGTTGCCATGTAGACC Chapter 5 R TTTTTGGTTGAGCACAGG CMA Substrate Sigma KiCQ Start oligo CHCHD2 F CACATTGGGTCACGCCATTAC Chapter 4

39 2. Materials and methods

R ATCTCATAGAGGCAAGGCTGC GGH F TGC TTT CTC CTC TTC AGA TTC AG Chapter 3 R TTT CCC ATG CAC CTA ATG CTG T

2.1.7 Human brain samples used for quantification of transcript levels

Protein extracts from PD patient and control brain regions used to assess transcript levels in sections 3, 4 and 5 of this thesis were obtained from the Sydney Brain Bank (Neuroscience Research Australia) and the New South Wales Tissue Resource Centre (University of Sydney) (Table 2.9). This study was supported by the Human Research Ethics Committee of the University of NSW (PID0245).

Table 2.9 PD patient and control cohort used for transcript level assessment.

Sample Condition Braak Sex Age Disease Stage duration CTRL_1 CTRL - F 85 - CTRL_2 CTRL - M 79 - CTRL_3 CTRL - F 84 - CTRL_4 CTRL - M 73 - CTRL_5 CTRL - M 86 - CTRL_6 CTRL - M 73 - CTRL_7 CTRL - M 88 - CTRL_8 CTRL - M 81 - CTRL_9 CTRL - M 83 - CTRL_10 CTRL - F 73 - PD_1 PD 4 M 72 7 PD_2 PD 4 M 84 17 PD_3 PD 5 M 90 15 PD_4 PD 4 M 72 9 PD_5 PD 5 F 83 14 PD_6 PD 5 M 75 14 PD_7 PD 5 F 85 17 PD_8 PD 4 F 78 21 PD_9 PD 4 M 84 12 PD_10 PD 5 M 79 6

2.1.7.1 Isolation of RNA and preparation of cDNA from post-mortem samples

RNA from patient and control brain samples was isolated by Lee Marshall (Garvan

40 2. Materials and methods

Institute of Medical Research). Briefly, 100 mg of frozen brain tissue was added to 1 mL of Trizol and homogenised with a Polytron at half speed for 1 min. Homogenised tissue samples were thawed and incubated at RT for 5 min and RNA extracted as outlined in section 2.1.6.1. Preparation of cDNA from these samples was conducted using the protocol outlined in section 2.1.6.2.

2.1.8 Protein manipulation for immunoblotting

2.1.8.1 Preparation of protein samples from tissue culture

Tissue culture samples were harvested for Western blotting following removal of cells from culture dishes by incubation in trypLETM dissociation solution (ThermoFisher Scientific, Australia) for 3 minutes. Following dissociation cells were resuspended in 10% FBS media and pelleted by centrifugation at 200 g for 5 minutes. The cell pellet was then resuspended in 60 μL HES Solution (20 mM HEPES, 1 mM EDTA, 250 mM sucrose, 2% SDS, pH 7.4 with 1X Complete EDTA-free protease inhibitors and incubated for 1 hour at room temperature prior to homogenization by sonication. Sonication was conducted using a Sonicator Ultrasonic Processor XL (Heat Systems Inc.) at 10% amplitude 3 times for 7 seconds separated by 5-second pauses. Protein concentration was estimated using a BCA protein estimation kit (Pierce, IL, USA) as per the manufacturers protocol and samples were stored at -70°C until further use.

2.1.8.2 SDS-Polyacrylamide gel electrophoresis and immunoblotting

For SDS-polyacrylamide gel electrophoresis (PAGE) 30 μg of protein sample was combined with sample loading buffer (12 mM Tris HCl, 5% glycerol, 1% SDS, 1% β- mercaptoethannol pH 6.8) and denatured at 95°C for 5 minutes, except for samples used for mitoprofile blots, which were denatured at 70°C for 10 minutes. Samples were then loaded onto a PAGE gel of appropriate acrylamide percentage (375 mM Tris, pH 8.8, 12- 15% acrylamide, 0.15% bis-acrylamide, 0.1% SDS, 0.084% APS and 0.034% TEMED) with a

41 2. Materials and methods stacking gel (125 mM Tris, pH 6.8, 5.6% acrylamide, 0.14% bis-acrylamide, 0.1% SDS, 0.2% APS an 0.06% TEMED). PageRuler Plus (ThermoFisher Scientific) pre-stained molecular weight marker was also loaded onto each gel for size comparison. PAGE gels were run at 125 V for 75 - 100 minutes (Table 2.10) in running buffer (1.5% Tris, 7.2% glycine, 5% SDS in H20) using a BioRad power pack.

Table 2.10 Polyacrylamyde gel electrophoresis conditions and antibody details. *The βtubulin antibody was obtained from the Developmental Studies Hybridoma Bank, created by the NICHD of the NIH and maintained at The University of Iowa, Department of Biology,

Iowa City, IA 52242.

Gel % Running Running Primary Source Secondary Antibody voltage time antibody antibody (V) (mins) dilution αSynuclein 15 135 75 1:300 BD Mouse Bioscience GAPDH 15 135 75 1:1000 Life Mouse Research βTubulin 12-15 125-135 125-135 1:1000 DSHB, Mouse University of Iowa* 14_3_3 12-15 125-135 125-135 1:1000 Rabbit Mitoprofile 12 125 100 1:500 AbCam Mouse OXPHOS cocktail NDUFB8 12 125 100 1:300 AbCam Mouse ATP6V1B2 15 135 75 1:1000 Sigma Rabbit Aldrich MEF2D 15 135 75 1:200 AbCam Mouse Caspase 15 135 75 1:200 Nova Mouse CHCHD2 15 135 75 1:500 Sigma Rabbit Aldrich

Proteins were transferred from PAGE gels onto PVDF membrane (Merck) that was activated by incubation in methanol for 15 seconds, followed by H20 for 2 minutes and transfer buffer (10.5% glycine, 2.25% Tris, in H20 and 20% methanol) for 2 minutes. Transfer was conducted at 90 V for 90 minutes in transfer buffer. Membranes were then fixed by incubation in 0.4% paraformaldehyde (PFA) solution for 45 minutes and blocked

42 2. Materials and methods by incubation in 10% skim milk powder in TTBS (0.12% Tris, 0.87% NaCl, 0.1% Tween20) for 1 hour. Blocked membranes were incubated in blocking solution containing 10 mM sodium azide and primary antibody at the dilution listed in Table 2.10 overnight at room temperature. Membranes were then washed 3 times in TTBS for 7 minutes prior to incubation in blocking solution containing secondary antibody listed in Table 2.10 at a dilution of 1:15,000 for 90 minutes. Blots were washed a further 2 times with TTBS and once with PBS, for 7 minutes each, and developed using HRP chemiluminescent substrate (Millipore, USA). Images were obtained using a FusionFX (VILBER Fisher Biotech, AUS) chemiluminescence detection system.

Denaturing conditions, gel acrylamide percentages and running setting are detailed in Table 2.10 for each of the antibodies used.

2.1.8.3 Immunoblot densitometry analyses

Immunoblot densitometry analyses was performed using ImageJ analysis software (imagej.nih.gov) (Schneider, Rasband, & Eliceiri, 2012). Statistical analyses were performed using GraphPad Prism 7 (GraphPad Software Inc., USA) and figures were assembled using Adobe Illustrator CS5 Version 15.1.0 (Adobe Systems Incorporated).

2.1.9 Assessment of cells by flow cytometry

A number of the assays used in this study, including the mito-Keima mitophagy assay (sections 4 and 5), the mitochondrial mass assay (section 4), measurement of mitochondrial ROS (sections 4 and 5) and the mitochondrial and cytosolic Ca2+ measurement assays (section 5) required assessment of cell fluorescence by flow cytometry. To prepare samples for flow cytometry cells were washed once with PBS, dissociated from plates using trypLE and pelleted as per section 2.1.6 and pelleted at 200 g for 5 minutes. Cell pellets were resuspended in fluorescence activated cell sorting (FACS) buffer (2 mM EDTA, 25 mM HEPES, 1% FBS in PBS, pH 7.0) and

43 2. Materials and methods fluorescence was assessed using a SORP LSRII FACS machine (BD Biosciences) with the excitation and emission settings outlined in Table 2.11. Data were analysed using FlowJo (FlowJo, LLC, USA) software (version 10.2).

Table 2.11 Flow cytometry laser and filter setting used for flow cytometry assays in this study

Assay Measurement Chapter (s) Excitation Emission wavelength filter (nM) MitoSox Mitochondrial 2 and 3 561 670/20 (ThermoFisher ROS Scientific) Mito-Keima Mitophagy 2 and 3 561 610/20 405 610/20 MitoTracker Green Mitochondrial 2 488 535/30 (ThermoFisher Mass Scientific) GECO/mGECO Mitochondrial 3 405 450/50 and cytosolic 405 525/50 Ca2+

2.1.10 Assessment of cell viability using AlamarBlueTM

Cells were seeded in 96-well plates with 6 replicates per treatment. Viability was assessed via oxidation of the alamarBlueTM (ThermoFisher Scientific) reagent as per the manufacturers protocol. Briefly, culture media was replaced with fresh 2% FBS culture media containing a 1/10 dilution of alamarBlueTM detection reagent and cells were incubated at 37°C for 1 hour. Fluorescence was then measured using a

R ClarioStar plate reader (BGM LabTech) (570nmEx/585nmEm). The assay was conducted with three biological replicates that each contained 4-6 separate wells as technical replicates. Data from individual replicates was compiled using the error propagation method whereby standard deviation is calculated as the square root of the average of the individual standard deviations squared (the root mean square deviation method).

44 2. Materials and methods

2.1.11 Assessing indications of mitochondrial damage

2.1.11.1 Mitochondrial Reactive Oxygen Species (MitoSOX)

MitoSoxTM (ThermoFisher Scientific) reagent was resuspended in DMSO to a final concentration of 5 mM. Culture media was removed from cells and replaced with fresh 2% FBS media containing MitoSoxTM at a final concentration of 5 μM. Cells were incubated with MitoSoxTM reagent for 30 minutes, media was then removed and cells were harvested for assessment by flow cytometry according to section 2.1.9.

2.1.11.2 Lysosomal uptake of damaged mitochondria (mito-keima)

The mito-Keima (mKeima) probe was used to measure the lysosomal uptake of damaged mitochondria, according to the method used by Katayama et al., (2011), adjusted for use with flow cytometry detection, as previously published (Sargsyan et al., 2015). The mKeima plasmid encodes a fluorophore that is targeted to the mitochondria. When damaged mitochondria that contain mKeima undergo autophagy and fuse with lysosomes the fluorophore is exposed to the acidic lysosomal environment, which triggers a conformational change that alters the excitation wavelength of the fluorophore. The ratio of neural (mitochondrial) fluoresce to acidic (lysosomal) fluorescence is assessed to determine the percentage of mitochondria undergoing mitophagy, referred to as the mitophagy index.

Cells were transfected with the mKeima plasmid as per section 2.1.5.1 and incubated with the transfection reagent for 4 hours. To assess of the effect of CHCHD2 overexpression on mitophagy (section 4), following this 4 hours incubation media was replaced with DOX supplemented media to induce CHCHD2 expression. Cells were cultured for a further 2 days and mitophagy index was assessed as detailed below. Alternatively, to assess the effect of BafA1 treatment or siRNA-mediated depletion of ATP6V1A or ATP6V1B2 on mitophagy (section 5), following mKeima transfection media was replaced with BafA1 supplemented media, or siRNA-

45 2. Materials and methods transfection was performed using the protocol outlined in section 2.1.5.2. Cells were cultured in these conditions for 48 hours and mitophagy index was determined as outlined below.

To measure the ratio of mKeima neutral to acidic fluoresce (mitophagy index) treated cells were harvested and assessed by flow cytometry using a SORP LSRII FACS machine (BD Biosciences) as per section 2.1.9. A 561 nm excitation laser with a 610/20 nm bandwidth filter was used for detection of the lysosomal signal and 405 nm excitation laser with 610/20 bandwidth filter for detection of the mitochondrial signal. Mitophagy index was calculated as a ratio of these two signals. Data was assessed using FlowJo software with the cell population gating is depicted in Figure 2.1. As a positive control 30 μM CCCP was added to untreated, mKeima expressing cells immediately prior to FACS analysis to induce rapid acidification of the mitochondrial matrix (Abad et al., 2004; Llopis et al., 1998), thereby shifting the excitation spectrum of mitochondrially localised mKeima into the acidic range (610/20 nm excitation wavelength). The shift in signal induced by CCCP treatment was used to determine the location of the gate used to differentiate between acidic and neutral mKeima signal, such that 100% of the CCCP signal fell within the gate (Figure 2.1).

46 2. Materials and methods

Figure 2.1 Schematic of cell population gating for assessment of mKeima fluorescence by flow cytometry. The CCCP positive control was used to determine the position of the gate denoting lysosomal localized mKeima such that 100% of CCCP treated cells fell within this gate.

2.1.12 Immunofluorescence and microscopy

2.1.12.1 Indirect immunofluorescnce

Cells were cultured on glass coverslips in 24-well plates. At the desired time-point media was removed and cells were washed once with PBS and then fixed by incubation in 4% PFA for 30 minutes at room temperature. PFA was removed; coverslips were washed once with PBS and then incubated in blocking buffer (10% BSA, 0.1% saponin in PBS) for 1 hour. Coverslips were then incubated with primary antibodies diluted in antibody buffer (2% BSA, 0.1% saponin in PBS) for 2 hours at concentrations listed in Table 2.12. Following 20 washes by dipping in PBS, coverslips were incubated for 1 hour in an appropriate fluorescent secondary antibody (Jackson ImmunoResearch Laboratories, Inc., USA), as listed in Table 2.12 used at a 1:300 dilution in 2% BSA blocking buffer. Where appropriate coverslips were additionally DAPI stained by washing a further 20 times in PBS and incubating in 50 ng/mL DAPI

47 2. Materials and methods for 10 minutes. Coverslips were then mounted onto slides for microscopy using Fluoromount-G medium (Southern Biotech).

Table 2.12 Concentration and sources of antibodies used for indirect immunofluorescence.

Antibody Secondary Dilution Source antibody Comments LAMP1 1:300 DHSB Mouse Assessment of lysosomal morphology LC3 1:300 MBL Mouse Mitophagy initiation assessment TOM20 1:400 Sapphire Rabbit Mitophagy initiation Bioscience assessment CHCHD2 1:400 Sigma Aldrich Rabbit CHCHD2 localisation assessment

2.1.12.2 Epifluorescent and confocal imaging

Epifluorescent images for assessment of TFEB translocation (section 5) were obtained using a 40X objective on a DM5500B microscope (Leica Microsystems) while confocal imaging of CHCHCD2 localization (section 4), MitoTracker (section 4), TOM20-LC3 colocalisation (section 5) and LysoSensor (section 5) was conducted using a DMI6000 SP8 confocal microscope (Leica Microsystems). Image processing and analyses were conducted using FIJI (ImageJ) as detailed in the appropriate sections.

2.2 Chapter 3 Specific Materials and Methods

2.2.1 Setup of sample labelling

SH-SY5Y-αSyn cells were cultured in RPMI media with 10% dialysed FBS (ThermoFisher Scientific) supplemented with 5 mM glutaMAX (ThermoFisher Scientific) and light, medium or heavy arginine and lysine (Silantes, Germany) (Table

48 2. Materials and methods

2.13). Cells were passaged in media containing these labels for >5 passages, at which point samples from heavy and medium labelled cells were harvested and processed as per the standard protocol (described below in section 2.2.2) for assessment of labelling efficiency. The proportions of labelled peptides relative to the remaining non-labelled (light) peptides in the samples were determined by mass spectrometry and analysis using MaxQuant software (version 1.5.3.25) (Cox and Mann, 2008). Using this method the incorporation of medium and heavy labels were determined to be to be 92% and 91.5%, respectively. Following this assessment of label incorporation cells were passaged in labelled amino acids for a further 2 passages prior to experimental set up.

Table 2.13 SILAC mass spectrometry labelled amino acids obtained from Silantes, Germany (medium and heavy) or SigmaAldrich, Australia (light).

Weight Lysine (mass shift) Arginine (mass shift)

Light C6 H14 N2 O2 . H1Cl1 C6 H14 N4 O2 . H1 Cl1 2 13 Medium C6 H10 H4N2 O2 . H2Cl2 (4) C6 H14 N4 O2 . H1 Cl1 (6) 13 15 13 15 Heavy C6 H14 N2 O2 . H1Cl1 (8) C6 H14 N4 O2 . H1 Cl1 (10)

Mass spectrometry analysis was conducted of SH-SY5Y-αSyn cells (described in section 2.1.4.2) treated with low-dose rotenone, moderately elevated αSyn expression or a combination of these insults (Table 2.14). SILAC labelling enables samples from different treatments to be combined prior to processing and these treatment combinations are outlined in Table 2.15. A replicate containing a label switch was included for each treatment, for instance light and medium amino acid labelled SH-SY5Y cells were treated with low dose rotenone in samples 1 and 2, respectively. The use of different labels for each replicate removes the potential artefacts caused by growth of cells in media containing heavy or medium SILAC labelled amino acids (Piechura et al., 2012).

Table 2.14 Treatments of SH-SY5Y-αSyn cells for SILAC mass spectrometry.

Treatment DOX Rotenone RA (ng/mL) (nM) (μM)

49 2. Materials and methods

Control 2000 0 10 Low-dose rotenone 2000 10 10 Moderate αSyn expression 1000 0 10 Combined low-dose rotenone 1000 10 10 and moderate αSyn expression

Table 2.15 Experimental setup indicating treatments combined to create each sample for mass spectrometry analysis and the labels used for each treatment at the 3 time-points assessed. Heavy amino acid labelled cells are highlighted in red, medium amino acid labelled cells are highlighted in orange and light amino acid labelled cells are highlighted in blue.

Time harvested Sample 1 Sample 2 Sample 3 Sample 4 Control Control Control Control Combination Combination Combination Combination treated treated treated treated Synuclein Synuclein Rotenone Rotenone 24 hours only only only only Control Control Control Control Combination Combination Combination Combination treated treated treated treated Synuclein Synuclein Rotenone Rotenone 48 hours only only only only Control Control Control Control Combination Combination Combination Combination treated treated treated treated Synuclein Synuclein Rotenone Rotenone 72 hours only only only only

Samples were harvested at the required time points (24, 28 and 72 hours) for mass spectrometry and additional samples were harvested at 96 hours and assessed using the FACS based ethidium homodimer cell death assay (ThermoFisher Scientific) by Charmaine Lang (Garvan Institute of Medical Research) to confirm that cells treated with the combination of low-dose rotenone and elevated αSyn expression exhibited a significant increase in cell death, as described in Lang (2015). 2 replicates for each the low-dose rotenone and elevated αSyn expression treatment groups and 4 replicates for the combination treated samples were assessed by mass spectrometry (Table

2.15).

50 2. Materials and methods

2.2.2 Preparation of cells for SILAC mass spectrometry

At the appropriate time point (24, 48 or 72 hours) samples were harvested and processed as per the method described in (Minard et al., 2016), with some minor adjustments. Briefly, cells were washed once with PBS, dissociated using trypLE and resuspended in PBS. Cells were pelleted at 600 g for 8 minutes at 4°C and resuspended in chilled buffer containing 6 M Urea, 2 M thiourea, 20 mM triethylammonium bicarbonate (TEAB) buffer, 2% SDS and Complete EDTA-free protease inhibitors. Samples were homogenized by sonication 3 times for 15 seconds with 10-second pauses between each sonication using a Sonicator Ultrasonic Processor XL (Heat Systems Inc.) on power level 4. SDS was then removed by methanol- chloroform purification (Wessel & Flügge, 1984). Briefly, 1 part methanol and 4 parts chloroform were added to the sample and vortexed to mix, flowing which 3 parts water was added and the samples were centrifuged at 15,000 g for 1 minute. The upper aqueous layer was removed and 3 parts methanol were added back, the samples vortexed and re-centrifuged at 15,000 g for 1 minute. The supernatant was removed and the protein pellet resuspended in buffer containing 6 M Urea, 2 M thiourea, 20 mM TEAB buffer and protease inhibitors.

Protein content was assessed by Qubit Protein assay (ThermoFisher Scientific), as per the manufacturers protocol and protein samples containing equal amounts of protein (50 μg) were pooled according to Table 2.15. To prepare the pooled protein samples for mass spectrometry proteins were reduced using 10 mM DTT for 1 hour at room temperature followed by inactivation by incubation of samples at 95°C for 2 minutes. Cooled samples were then alkylated by 30 minute incubation in 25 mM iodoacetamide, which was quenched by addition of 10 mM dithiothreitol (DTT). Samples were then digested by incubation with Lys-C (2 μg/ 100 μg protein) for 4 hours at RT, diluted 1:5 in TEAB buffer and further digested by incubation with trypsin (1 μg/ 100 μg protein) for 16 hours at 30°C with shaking. Peptides were desalted using SepPak tC18 cartridges (Waters, Germany) and protein content was again assessed by Qubit Protein assay (ThermoFisher Scientific). Sample fractionation

51 2. Materials and methods by Strong Cation Exchange (SCX) and LC-MS/MS analysis was subsequently conducted by Benjamin Parker (Charles Perkins Institute, University of Sydney) using a Q- Exactive Orbitrap HCD-MS/MS (ThermoFisher Scientific). Raw mass spectrometry data were processed by Daniel Fazakerley (Charles Perkins Institute, University of Sydney) using MaxQuant software (version 1.5.3.25) (Cox and Mann, 2008). Protein ratios were expressed relative to untreated controls and all subsequent data processing and analysis were conducted using Perseus software package (Tyanova et al., 2016), as detailed in section 3.2.1. Sources of the chemicals used are detailed in Table 2.1.

2.3 Chapter 4 Specific Materials and Methods

2.3.1 Generation of DOX inducible CHCHD2 overexpressing SH-SY5Y cells [DOX- CHCHD2]

Details of plasmids used for construction of the CHCHD2 DOX-on lentiviral vector and production of CHCHD2 lentivirus are outlined in Table 2.16

Table 2.16 Plasmids used for production of CHCHD2-DOX ON lentivirus.

Plasmid Source/Reference Notes

pCLX-pTF-R1-DEST- A gift from Marc Establishing CHCHD2 DOX-on R2-EBR65 GiryLaterriere (Addgene stable cell line plasmid # 45952) {GiryLaterriere:2011tj} pCMV-pSORT6- GE Dharmacon MGC Human CHCHD2 CHCHD2 MHS6278-202760062 sequence-verified cDNA CloneID:6192342 pJET1.2 ThermoFisher Scientific Blunt end cloning vector psPAX2 {GiryLaterriere:2011tj} Lentiviral production pCAG-VSVG {GiryLaterriere:2011tj} Lentiviral production

52 2. Materials and methods

2.3.1.1 Construction of the CHCHD2 lentiviral plasmid

A 902 bp CHCHD2 DNA fragment was amplified by polymerase chain reaction (PCR) from the CMV-pSPORT6-CHCHD2 plasmid (GE Dharmacon) (Table 2.16) using primers 2702/2703 (Table 2.17) to add BsrG1 restriction endonuclease sites (Figure 2.2). PCR was conducted using the Phusion DNA polyermase kit (New England Biolabs) as per the manufacturers protocol. Briefly, a 20 μL reaction was prepared containing 2 ng DNA, 1 X Phusion HF buffer, 200 μM deoxynucleotide solution mix, 0.5 μM each primer and 0.4 U Phusion polymerase. Thermal cycling was conducted using a PTC- 200 Thermocycler (MJ research) with thermocycling conditions as follows; 98°C for 30 sec., then 30 cycles of 98°C for 10 sec, 56°C for 20 seconds, 72°C for 1 min. Final extension of 72°C for 1 min. The PCR product obtained was separated by agarose gel electrophoresis using a 1% agarose gel [1% agarose; 1X TAE buffer (0.484% Tris base, 0.114% acetic acid, 0.2% EDTA, 0.283 g/L guanosine), 1 mg/ml ethidium bromide] at 90 V for 40 minutes using a BioRad power pack. The correct product size was confirmed using GeneRuler DNA ladder (ThermoFisher Scientific) for size comparison and the gel fragment was isolated using the ZymoClean Gel DNA Recovery kit (ZymoResearch, CA, USA) according to the manufacturers instructions.

The isolated amplified CHCHD2 fragment was inserted into the pJet1.2 cloning vector using the CloneJet PCR cloning kit (ThermoFisher Scientific) as per the manufacturers instructions and ligated DNA was transformed into MC1061 competent cells (Table 2.3). Colonies obtained were tested by colony PCR. Briefly, a PCR reaction was prepared containing 0.5 μM of primers 2706 and pJetF (Table 2.17) and 1X GoTaq green master mix (Promega, USA). Using a sterile pipette tip single colonies were resuspended in 10 μL of PCR reaction solution and the reaction was cycled as follows: initial melt 95°C for 4 min. 30 cycles of 95°C for 20 sec, 55°C for 30 seconds, 72°C for 1 min. Final extension of 95°C for 5 minutes. PCR products visualized after agarose gel electrophoresis (as above) and positive colonies containing the desired pJet-CHCHD2(BsrG1) plasmid were cultured further.

53 2. Materials and methods

Figure 2.2 Cloned CHCHD2 fragment sequence. The CHCHD2 transgene expressed consisted of the WT human CHCHD2 protein coding cDNA sequence (blue) (CloneID:6192342) [protein sequence (black)], as well as 33 base pairs from the 5’ UTR (red) and 294 base pairs from the 5’ UTR (green). BsrG1 cutting sites used for subsequent cloning steps are indicated in yellow.

Table 2.17 Primers used for construction of the CHCHD2 lentiviral plasmid.

Plasmid Plasmid Sequence number 2702 pSPORT6_CHCHD2_BsrG1 CAA TTG TAC ACC GGA ATT CCC GGG _F ATA TCG TC 2703 pSPORT6_CHCHD2_BsrG1 GTA CTG TAC ACG TTG TAA AAC GAC _R GGC CAG 2706 CHCHD2_colonyPCR_R CCT GGG TGC AGC TCT CAT CT pJetF pJet1.2_F CGA CTC ACT ATA GGG AGA GCG GC 2650 DEST_colonyPCR_F ACT TTG CCT AAT CAC TTA GCA ACT

To clone the CHCHD2 fragment into the pCLX-pTF-R1-DEST-R2-EBR65 lentiviral vector (Table 2.16) the pJet-CHCHD2(BsrG1) and pCLX-pTF-R1-DEST-R2-EBR65 plasmids were first digested with the BsrG1 restriction enzyme (New England Biolabs). Briefly, 750 ng DNA was combined with 10 U BsrG1, 1 X NEB CutSmart buffer (New England Biolabs) and 100 μg/mL BSA in a total volume of 10 μL and incubated at 37°C for 1 hour. The desired 880 bp CHCHD2 and 9238 bp pCLX-pTF-R1-DEST-R2-EBR65 vector

54 2. Materials and methods fragments were isolated and extracted by gel electrophoresis and gel purification (as above). The pCLX-pTF-R1-DEST-R2-EBR65 fragment was then ligated with the CHCHD2 fragment using Quick ligase (New England Biolabs). 20 U Quick ligase was combined with 200 ng DNA with a vector to insert ratio of 1:3 and 1X supplied Quick ligase buffer in a total volume of 10 μL and incubated at room temperature for 5 minutes. DNA was then transformed into chemically competent STBL3 cells as per manufacturers protocol. Transformants containing the desired pCLX-YTF-DEST-EBR- CHCHD2 (pCHCHD2-DOX) lentiviral plasmid were assessed by colony PCR (as above) using primers 2706/2650 (Table 2.17). The selected clone was further assessed by Sanger sequencing, performed by the Garvan Molecular Genetics (GMG) facility at The Garvan Institute of Medical Research. Sequence files were analysed using Seqbuilder (DNASTAR, Inc., USA) and ApE (University of Utah, USA).

2.3.1.2 Virus preparation and harvest

Lentivirus was prepared according to the protocol of Giry-Laterrière et al., (2011). pCHCHD2-DOX and the pCLX-YTF-DEST-EBR-GFP (pGFP-DOX) lentiviral plasmid, which was a kind gift from Dr Marc Giry-Laterriere (Garvan Institute) were transfected into HEK293 LentiX cells along with the pSPAX2 and pCAG-VSG packaging plasmids (Table 2.16) using Lipofectamine 3000 (ThermoFisher Scientific). Briefly, plasmid DNA (8.5 μg pSPAX2, 4.25 μg pCAG-VSVG along with 8.5 μg pCHCHD2-DOX or pGFP-DOX lentiviral plasmid) and 34 μL P3000 reagent (ThermoFisher Scientific) were diluted in 2 mL optiMEM media and combined with 2 mL optiMEM media containing 29 μL Lipofectamine 3000 reagent. These solutions were inverted to mix, incubated at room temperature for 15 minutes and added dropwise onto HEK293 LentiX cells in DMEM media with 10% FBS grown to 70% confluence in a T75 vented flask. Media was replaced with fresh media after 6 hours.

Viral particles were harvested from the cell growth media following a further 24 hours of cell growth. Briefly, media was removed and cell debris was pelleted by centrifugation at 100 g for 10 minutes. The supernatant was isolated and viral

55 2. Materials and methods particles were concentrated using centrifugal filter units with Ultracel-100 membrane (Merck), as per the manufacturers instructions.

To determine virus titre, HEK293 LentiX cells seeded at 16 x 104 cells per well in 6- well plates and were transduced with increasing amounts of CHCHD2-DOX or GFP- DOX lentivirus. 2 sets of HEK293 LentiX cells were transduced with 0.5, 5 or 50 μL GFP-DOX lentivirus and one set of cells was transduced with 0.5, 5 or 50 μL of CHCHD2-DOX lentivirus. GFP transduced cells were supplemented with 2000 ng/mL DOX to induce GFP expression. Following 2 days of growth cells were harvested by cell scraping and centrifuged at 100 g for 10 minutes to pellet. One set of GFP-DOX infected cells was resuspended in 1% PFA in PBS and GFP fluorescence was assessed by flow cytometry using a CantoII FACS machine (BD biosciences) with 488 nm excitation laser and 525/30 emission filter to determine the Multiplicity of Infection (MOI) of the GFP-DOX virus. For the remaining GFP-DOX and CHCHD2-DOX lentiviral infected samples, genomic DNA was extracted using the QIAamp DNA mini kit (Qiagen) as per the manufacturers instructions for DNA purification from cultured cells. Incorporation of viral DNA into the genome was assessed by qRT-PCR using the primers listed in Table 2.18, enabling the MOI of the CHCHD2-lentivirus to be back- calculated based on the GFP virus MOI determined by flow cytometry and the genomic incorporation determined by qRT-PCR, as outlined in Giry-Laterrière et al. (2011).

Table 2.18 Primers used for assessing lentiviral incorporation.

Number Name Sequence 1996 Actin_F TCATGAAGTGTGACGTGGACATCCGT 1997 Actin_R CCTAGAAGCATTTGCGGTGCACGATG 2041 GAG_F GGAGCTAGAACGATTCGCAGTTA 2042 GAG_R GGTTGTAGCTGTCCCAGTATTTGTC

56 2. Materials and methods

2.3.1.3 Establishing CHCHD2-DOX and GFP-DOX stable SH-SY5Y cell lines

GFP-DOX and CHCHD2-DOX virus particles at 0.4 MOI were used to transduce SH- SY5Y cells to create SH-SY5Y [DOX-GFP] and SH-SY5Y [DOX-CHCHD2] stable cells respectively. After 3 days transduced cells were selected by the addition of 2 µg/mL blasticidine S (Sigma-Aldrich). Cells were routinely cultured in blasticidine S to ensure selectivity. Cells were cultured as per SH-SY5Y cells, as outlined in section 2.1.4.1.

2.3.2 Assays for mitochondrial functions and features

2.3.2.1 Imaging of mitochondrial membrane potential

Assessment of mitochondrial membrane potential was achieved using the CMX/ROS mitoTracker RedTM probe (ThermoFisher Scientific). Cells were cultured on glass coverslips for 48 hours in media containing the desired treatments (control, 20 nM rotenone, 5 nM antimycinA or 200 nM CCCP or 250 nM thapsigargin). Media was removed, cells were washed once with PBS and incubated for 20 minutes at 37°C in 2% media containing 100 nM mitoTracker RedTM . Media was then removed, cells washed with PBS and fixed in 4% PFA for 20 minutes at 37°C. Indirect immunofluorescence of CHCHD2 and DAPI staining was then conducted as per section 2.1.12.1.

2.3.2.2 Measurement of mitochondrial mass

Mitochondrial mass was determined using the mitoTracker Green FMTM probe (ThermoFisher Scientific) by flow cytometry. Briefly, media was removed, cells were washed once with PBS and media replaced with fresh 2% media containing 200 nM mitoTracker Green FMTM. Cells were incubated in this solution for 30 min at 37°C and harvested for flow cytometry assessment as per section 2.1.9.

57 2. Materials and methods

2.4 Chapter 5 Specific Materials and Methods

2.4.1 Immunoblotting of human brain tissue

PD patient and healthy control brain tissue used to assess ATP6V1B2 protein levels were obtained from the Sydney Brain Bank at Neuroscience Research Australia and the New South Wales Tissue Resource Centre (University of Sydney) (Table 2.19). Protein extracts from these samples were generously supplied by Dr Karen Murphy and Prof Glenda Halliday (Neuroscience Research Australia).

Table 2.19 PD patient and control cohort used for assessment of ATP6V1B2 protein levels in the anterior cingulate cortex.

Sample Condition Braak Sex Age Disease Stage duration CTRL_1 CTRL - F 78 - CTRL_2 CTRL - F 82 - CTRL_3 CTRL - M 78 - CTRL_4 CTRL - M 60 - CTRL_5 CTRL - M 69 - CTRL_6 CTRL - M 73 - PD_1 PD 4 M 75 18 PD_2 PD 4 M 71 16 PD_3 PD 4 M 80 16 PD_4 PD 4 F 75 10 PD_5 PD 4 M 74 16 PD_6 PD 6 F 74 17

Gel electrophoresis and immunoblotting of human brain protein samples were conducted by Anouk Spruit and Louise Cottle (Garvan Institute), according to the method outlined in section 2.1.8. The membrane was probed for ATP6V1B2, as per section 2.1.8.2 and densitometry analyses conducted as outlined in section 2.1.8.3.

58 2. Materials and methods

2.4.2 Nuclear translocation of TFEB

The eGFP-TFEB plasmid (Table 2.2) was transfected into cells as per the protocol outlined in section 2.1.5.1 and coverslips were fixed and imaged as per section 2.1.12. The ratio of nuclear to total TFEB signal was quantified using Image J, with nuclear area determined based on DAPI staining. A total of 30 cells from 3 biological replicates were assessed from each treatment.

2.4.3 Measurement of Cytosolic and Mitochondrial Ca2+

Mitochondrial and cytosolic calcium levels were assessed using the Genetically Encoded Ca2+ Indicators for Optical Imaging (GECO) created by Zhao et al. (2011). The cytosolic probe (GECO) accumulates in the cytosol, while the mitochondrial probe (mGECO) is targeted to mitochondria. The emission spectrum of the probe is sensitive to changes in Ca2+ concentration, displaying a 11,000 % ratio change upon binding with Ca2+ (Zhao et al., 2011), enabling ratiometric assessment of Ca2+ levels at these subcellular locations.

For assessment of cytosolic and mitochondrial Ca2+ levels plasmids encoding GECO or mGECO (Table 2.2), respectively, were transfected into cells (section 2.1.5.1) prior to siRNA-mediated V-ATPase depletion (section 2.1.5.2) or addition of BafilomycinA1. Cells were grown to the desired time-point (48 or 72 hours), harvested for FACS analyses (section 2.3.2.2) and fluorescence intensity was assessed according to Table 2.11, with cell population gating as indicated in Figure 2.1. As a positive control GECO or mGECO expressing cells were treated with 20 mM potassium chloride prior to flow cytometry assessment. Potassium chloride induces plasma membrane depolarisation that leads to increased cytosolic and mitochondrial calcium levels (Challet et al., 2001; Mueller et al., 2013).

59 2. Materials and methods

2.4.4 Measurement of lysosomal pH

Lysosomal pH was determined using LysoSensorTM Blue (ThermoFisher Scientific), a fluorescent probe that accumulates in the lumen of lysosomes and late endosomes and displays pH dependent fluorescence intensity (Dolman et al., 2013). Cells were cultured on glass coverslips for 48 hours in the desired treatments (control, BafilomycinA1 or siRNA depletion), media was removed and cells were washed once with PBS. Cells were then stained by incubation in 2% media containing 5 µM LysoSensor blue for 1 hour in the dark at 37°C. Media was removed and cells were washed again with PBS and fixed in 4% PFA at 37°C for 15 minutes. Cells were imaged immediately following fixation using a 63X objective on a DMI 6000 SP8 confocal microscope (Leica Microsystems). 7 images from each of 4 biological replicates for each treatment were processed and fluorescence intensity determined according to the ImageJ macro I developed, as outlined below (section 2.4.4.1). Total corrected lysosomal florescence was calculated as total fluorescence intensity – (average background fluorescence intensity x total lysosomal area), where the average background fluorescence intensity was calculated as the average fluorescence intensity of 4 regions of each image that did not exhibiting LysoSensor staining, according to the protocols of (Burgess et al., 2010) and (McCloy et al., 2014). Total corrected lysosomal fluorescence was then determined relative to total lysosomal area, providing an indication of lysosomal fluorescence per unit lysosomal area, so as to remove the possible confounding effect caused by changes in lysosomal area.

2.4.4.1 ImageJ LysoSensor blue macro

//LysoSensor automated fluorescence intensity analysis- for LysoSensor Blue imaging in SHSY-5Y cells //select source directory dir1 = getDirectory("select the source directory"); // select a destination directory dir2 = getDirectory("Select the destination directory"); //generate a list of files in the source directory list = getFileList(dir1);

60 2. Materials and methods

Array.sort(list); //loop the code through all images in selected directory for(i=0; i

2.4.5 LC3-mitochondria co-localisation as a measure of mitophagy initiation

Cells treated with low-dose Bafilomycin A1 or siRNA mediated depletion of ATP6V1A or ATP6V1B2 were stained for TOM20, LC3 and DAPI fluorescent detection according to section 2.1.12 and imaged using a 63X objective on a DMI 6000 SP8 confocal microscope (Leica Microsystems). Images were processed and co-localisation area was determined using the ImageJ colocalization plugin as per the ImageJ macro I developed, as outlined below (section 2.4.5.1). The total area of TOM20-LC3 colocalisation was calculated as a percentage of total TOM20 area based on imaging data obtained from 7 images per treatment from each of 3 biological replicates.

61 2. Materials and methods

2.4.5.1 ImageJ Tom20-LC3 co-localisation Macro

//This is designed for analyzing the colocalized points for LAMP1 TOM20 -- mitophagy by colocalisation macro "Background subtraction [F5]" { print("Automated Red Green Colocalisation - Brigitte Phillips - 22/07/16"); wait(100); //ask user to select source directory dir1 = getDirectory("select the source directory"); //ask the user to select a destination directory dir2 = getDirectory("Select the destination directory"); //generate a list of files in the source directory list = getFileList(dir1); Array.sort(list); //loop the code through all images in selected directory for(i=0; i

//store images in the namestore variable and process images DAPIImage = nameStore +" (blue)"; TOMImage = nameStore +" (green)"; LC3Image = nameStore +" (red)"; selectWindow(DAPIImage); close(); selectWindow(TOMImage); run("Duplicate...", " "); rename("original"); selectWindow(TOMImage); run("Window/Level..."); run("Enhance Contrast", "saturated=0.35"); setMinAndMax(139, 226); selectWindow(LC3Image); run("Window/Level..."); setMinAndMax(167, 140);

62 2. Materials and methods

//select green window and threshold selectWindow(TOMImage); run("Subtract Background...", "rolling=20"); run("Median...", "radius=1"); setAutoThreshold("Otsu dark"); run("Convert to Mask");

//select red window and threshold selectWindow(LC3Image); run("Subtract Background...", "rolling=50"); run("Median...", "radius=1"); setAutoThreshold("Moments dark"); run("Convert to Mask");

//colocalisation analysis using colicalization plugin run("Colocalization ", "channel_1=[TOMImage] channel_2=[LC3Image] ratio=30 threshold_channel_1=15 threshold_channel_2=15 display=255 colocalizated"); saveAs("Tiff", dir2+nameStore+" - ColocRGB.tif"); selectWindow("Colocalizated points (8-bit) "); rename(nameStore+"_Coloc"); run("Make Binary"); run("Set Measurements...", "area mean min integrated area_fraction limit display add redirect=[original] decimal=5"); run("Analyze Particles...", "size=0-infinity pixel circularity=0.00-1.00 show=Outlines display summarize"); //save composite image using namestore variable saveAs("Tiff", dir2+nameStore+"_colocOutlines.tif"); run("Close All"); } //close all open images run("Close All");

2.4.6 Lamp1 Imaging analysis: Lysosomal size and perinuclear clustering

Cells treated with low-dose Bafilomycin A1 or siRNA mediated depletion of ATP6V1A or ATP6V1B2 were stained for LAMP1 and DAPI flourescent detection, according to

63 2. Materials and methods section 2.1.12 and imaged using a 63X objective on a DMI 6000 SP8 confocal microscope (Leica Microsystems). Images were processed and LAMP1 area was determined using the ImageJ macro I developed, as detailed in section 2.4.6.1. Data were obtained from 7 images per treatment from each of 3 biological replicates.

Using this same dataset the proportion of lysosomes in the perinuclear region of the cell was quantified using the Image J macro I developed, as detailed in section 2.4.6.2. Briefly, DAPI staining was used to identify the nucleus and this region was expanded automatically by 5 μM on all sides. This expanded DAPI selection was then used as a mask to measure the lysosomal area within 5 μM of the nucleus. This perinuclear lysosomal area was then calculated as a proportion of the total lysosomal area per image.

2.4.6.1 Image J Measurement of LAMP1 compartments macro

//This macro will prompt the user to select a directory containing the images of interest. It will then open the images, remove background and duplicate the image. It will take the duplicate image and threshold it. It will use this as a mask to measure intensity in the original image //ask user to select source directory dir1 = getDirectory("select the source directory"); //ask the user to select a destination directory dir2 = getDirectory("Select the destination directory"); //generate a list of files in the source directory list = getFileList(dir1); Array.sort(list); //loop the code through all images in selected directory for(i=0; i

64 2. Materials and methods

setMinAndMax(50, 96); run("Subtract Background...", "rolling=50"); run("Median...", "radius=1"); run("Maximum...", "radius=1"); setAutoThreshold("Otsu dark"); run("Convert to Mask"); run("Watershed"); run("Set Measurements...", "area mean standard modal min integrated median skewness area_fraction limit display add redirect=None decimal=5"); run("Analyze Particles...", "size=0-infinity pixel circularity=0.00-1.00 show=[Bare Outlines] display summarize"); saveAs("Tiff", dir2+nameStore+" -Mask.tif");

//close all open images run("Close All"); }

2.4.6.2 Image J macro for the measurement of lysosomal density in the perinuclear region

//this macro is for assessing the intensity of lysosomal signal in close proximility to the nucleus. //ask user to select source directory dir1 = getDirectory("select the source directory"); //ask the user to select a destination directory dir2 = getDirectory("Select the destination directory"); //generate a list of files in the source directory list = getFileList(dir1); Array.sort(list); //loop the code through all images in selected directory for(i=0; i

//store images in the namestore variable

65 2. Materials and methods

DAPIImage = "C2-"+nameStore; LAMP1Image = "C1-"+nameStore;

//process images selectWindow(LAMP1Image); run("Window/Level..."); run("Enhance Contrast", "saturated=0.35"); setMinAndMax(50, 96); run("Subtract Background...", "rolling=50"); run("Median...", "radius=1"); run("Maximum...", "radius=1"); setAutoThreshold("Otsu dark"); run("Convert to Mask");

selectWindow(DAPIImage); setAutoThreshold("Default dark"); run("Convert to Mask"); run("Maximum...", "radius=25"); run("Convert to Mask"); selectWindow(DAPIImage); run("Create Selection"); selectWindow(LAMP1Image); run("Restore Selection");

run("Set Measurements...", "area mean min integrated median area_fraction limit display add redirect=" + LAMP1Image + " decimal=5"); run("Analyze Particles...", "size=0-infinity circularity=0.00-1.00 show=Outlines display summarize"); saveAs("Tiff", dir2+nameStore+" -Mask.tif"); //close all open images run("Close All"); }

66 3. Proteomic profiling of a complex disease

3 PROTEOMIC PROFILING OF A COMPLEX DISEASE BY INTEGRATION OF DATA DERIVED FROM IN VITRO MODELS AND PD PATIENTS

3.1 Introduction

Despite the high prevalence of PD and the severity of the symptoms, the molecular mechanisms underpinning the disease remain controversial. Therefore, to identify early, potentially causative events in the disease proteomic analysis of 2 in vitro PD models was conducted. The first model selected was an elevated αSyn expression model that is based on the observation that duplications and triplications of the αSyn gene cause PD (Polymeropoulos et al., 1997). The second PD model that was assessed recapitulates environmental aspects of the disease, whereby PD-like dysfunctions develop following exposure to the pesticide rotenone, an established PD risk factor (Dhillon et al., 2008; Furlong et al., 2015; Tanner et al., 2011). The effects of these two PD relevant stresses (αSyn and rotenone) were also assessed in combination, with the purpose of exploring how interplay between these environmental and genetic risk factors may contribute to this complex disease.

The proteomic changes induced by rotenone, αSyn or the combination of these stresses were assessed at multiple time-points of treatment (24, 48 and 72 hours) using Stable Isotope Labeling by Amino Acids in Cell Culture (SILAC) mass spectrometry. This assessment of multiple time-points allowed the changes that were identified to be distinguished into early changes, that are more likely to be causative, and late changes, that are more likely to be downstream consequences. Additionally, to determine whether there is evidence that the proteomic changes identified also occur in PD patients, the proteomic data obtained was cross-referenced against available published patient proteomic data, as well as transcriptomic data obtained from RNA sequencing of PD patient brain regions. This approach of combining in vitro

67 3. Proteomic profiling of a complex disease and patient data enabled the changes that were identified in vitro to be confirmed as patient relevant and provided manipulable model systems in which these changes could be further assessed. Through this approach a number of key areas of interest were established that will be discussed in this chapter and explored in depth in the flowing sections of this thesis.

3.1.1 The rotenone and elevated αSyn expression models of PD recapitulate disease characteristics in many systems

PD is a complex disease associated with many genetic and environmental risk factors, discussed in sections 1.5 and 1.6, and many disease models have been developed that reflect these different aspects of the disease. The rotenone exposure and the elevated αSyn expression models of PD are among the most well characterised and frequently used PD models and were therefore selected for inclusion in this study. The characteristics of each of these models are discussed in detail below.

The αSyn and rotenone models examined in this study were based in differentiated SH-SY5Y cells, a neuroblastoma cell line that exhibits many neuronal characteristics (Xicoy et al., 2017). This in vitro approach was selected due to the considerable expense and time associated with conducting SILAC based proteomic analyses in vivo. However, these models are readily translatable into other in vitro or in vivo systems for further testing, with equivalent models, including rodent models, already established and well characterized.

3.1.1.1 The rotenone model of PD

Rotenone is an OXPHOS complex 1 inhibitor (Lambert & Brand, 2004) that became a commonly used PD mimic following the discovery of a significant correlation between exposure to rotenone based pesticides and increased PD risk (Dhillon et al., 2008; Furlong et al., 2015; Tanner et al., 2011), as discussed in section 1.6.1. Rotenone was

68 3. Proteomic profiling of a complex disease first described to trigger PD-like phenotypes in a rat model (Betarbet et al., 2000) and has since been used to model aspects of PD in many in vitro and in vivo systems, including cell culture, C. elegans, zebrafish, drosophila and mice (Betarbet et al., 2000; Bretaud et al., 2004; Coulom & Birman, 2004; Murakami et al., 2015; Pan-Montojo et al., 2010; Zhou et al., 2013b).

Rodent models are among the best-characterized rotenone PD model systems. Rotenone treated rodents develop striatal nerve terminal damage, damage to substantia nigra nerve cell bodies and accumulation of nigral cytosolic αSyn deposits (Betarbet et al., 2000; Cannon et al., 2009); replicating the Lewy pathology and brain region specific neuronal damage observed in patients. This neuronal damage is generally considered to be the basis for many of the motor symptoms associated with PD. Accordingly, rotenone treated mice develop many PD related movement symptoms, including postural instability, reduced activity and endurance and delayed initiation of movement (Fleming et al., 2004; Murakami et al., 2015; Richter et al., 2007). Rotenone treatment in rodents has also been shown to cause many non-motor PD symptoms and non-motor associated pathologies, such as hyposmia (Yan Liu et al., 2015d), gastrointestinal and olfactory αSyn deposits (Drolet et al., 2009), cognitive and memory impairment (Kaur et al., 2011), sleep disturbances (García-García et al., 2005), anxiety-like behaviours (Bassani et al., 2014) and reduced striatal serotonin levels (Santiago et al., 2010), which is indicative of depression. Together these results suggest that in vivo rotenone models accurately recapitulate many of the movement and non- movement related aspects of PD.

Many in vitro rotenone PD models have also been described, including models that utilize primary mouse and rat neuronal cultures (Sanders & Greenamyre, 2013), induced pluripotent stem cells (Peng et al., 2013) and cultured cell lines (Wu et al., 2013a; Chiu et al., 2015). These in vitro systems have been used to examine in detail the molecular and cellular changes that contribute to rotenone toxicity. For instance, assessment of the effects of rotenone treatment of cultured midbrain or cortical primary mouse neurons suggested that the brain region specific toxicity of rotenone,

69 3. Proteomic profiling of a complex disease which is limited to the midbrain in rodent models, may be a result of increased sensitivity of midbrain neurons to the oxidative damage that is produced as a side- effect of rotenone treatment, rather than differential sensitivities of these brain regions to the effect of rotenone as an OXPHOS complex I inhibitor (Sanders et al., 2014). This observation suggests that a similar sensitivity to oxidative stress may underlie the brain region specific damage that is observed in patients.

An in vitro rotenone model of PD was used in this study as a well-characterized and accepted environmental-based model of the disease. The rotenone model used was established in the lab by a previous PhD student, Charmaine Lang, using differentiated SH-SY5Y cells (Lang, 2015). Cells were treated with a low and chronic dose of rotenone (10 nM for up to 4 days) to recapitulate the slow degenerative process that occurs in the disease and to mimic the long-term, chronic environmental rotenone exposure that is associated with increased PD risk (Betarbet et al., 2000). The development of multiple PD-like cellular phenotypes, including OXPHOS defects, production of reactive oxygen species (ROS) and endoplasmic reticulum (ER) stress, were assessed in this model over a time course of 4 days (Lang, 2015) to confirm that this model recapitulates many of the molecular dysfunctions that are associated with PD.

3.1.1.2 The αSyn model of PD

As discussed in section 1.8, αSyn is implicated in PD as it is the primary component of the Lewy body and Lewy neurite protein aggregates that are the hallmark feature of PD (Surmeier & Sulzer, 2013; Xu et al., 2015a). Additionally, PD-causative mutations in and multiplications of the αSyn gene were the first identified genetic basis of PD (Polymeropoulos et al., 1997) and variants in the αSyn gene are the most significant hit in Genome Wide Association Studies that assess PD-associated genetic mutations (Simón-Sánchez et al., 2009). Together these observations strongly implicate a role of αSyn dysregulation in both familial and sporadic forms of the disease.

70 3. Proteomic profiling of a complex disease

Due to this substantial evidence implicating a role of αSyn in PD, multiple αSyn-based in vivo and in vitro models of the disease have been developed. Characterization of these models has confirmed that elevated expression of wild type (WT) αSyn causes PD-associated symptoms, pathologies and molecular changes in vivo and in vitro, as summarized in Table 3.1. Examination of αSyn-based models has also helped to determine the basis of αSyn toxicity in PD, suggesting that elevated levels of αSyn cause cell dysfunction by interfering with lysosomal function and membrane trafficking (Dehay et al., 2013; Mazzulli et al., 2016), inhibiting endosomal transport (Volpicelli- Daley et al., 2014), dysregulating synaptic transmission (Diao et al., 2013) or perturbing mitochondrial function (Lindström et al., 2017), possibly by impairing mitochondrial protein import (Di Maio et al., 2016).

The in vitro αSyn model used in this study develops PD-relevant cellular characteristics following elevated expression of WT αSyn and is based in differentiated neuronal-like SH-SY5Y cells, as per the rotenone model used. SH-SY5Y cells were transduced to stably express WT αSyn (or βGal as a control) under the regulation of a TET-Off promoter, which enables αSyn expression to be regulated in a DOX dependent manner, such that reduced DOX concentrations induce higher transgene (Tg) expression. This model was established and originally characterized by Vekrellis et al. (2009) and has since been further characterized by others. Mimicking PD, induction of elevated αSyn expression in these cells results in time-dependent loss of viability and neurite retraction (Gouarné et al., 2015; Vekrellis et al., 2009), increased markers of oxidative stress (Perfeito et al., 2017), proteasomal dysfunction (Vekrellis et al., 2009) and increased levels of αSyn protein, albeit not in Lewy body-like aggregates (Vekrellis et al., 2009). This model has additionally been extensively characterized in our lab (Lang, 2015) and has been used to establish a model of moderately increased αSyn expression, whereby moderately elevated αSyn levels are achieved by supplementing growth media with a low level of DOX, as opposed to the absence of DOX (Lang, 2015). This model exhibits similar loss of cell viability as was observed by Vekrellis et al., (2009) and Gouarne et al., (2015), however the loss of viability is less substantial and more gradual. This slow-degenerating, moderately-elevated αSyn expression model

71 3. Proteomic profiling of a complex disease

Table 3.1 Elevated αSyn levels have been used to model PD in multiple in vivo and in vitro systems. Key characteristics of these models are outlined.

Model Method of αSyn Characteristics that mimic PD in humans Reference(s) elevation Mouse Transgenic Thy-1 • Age dependent decline in function {Chesselet:2012gi} promoter driven • Progressive formation of insoluble αSyn protein aggregates expression of resembling Lewy bodies human WT αSyn • Increased levels of 129S-phosphorylated αSyn • Progressive loss of striatal dopamine • Early olfactory deficits and progressive PD like motor deficits Rodents Striatal injection of • Age dependent degeneration of DA neurons {Paumier:2015cp} pre-formed αSyn • Accumulation of Lewy bodies and Lewy neurites {MasudaSuzukake:2013by} fibrils • Prion-like spreading of αSyn pathology

Drosophila Transgenic • Reduced memory and learning skills {Zhao:2015iz} overexpression of • Impaired neuronal calcium handling and synaptic transmission {Chen:2014kx} human WT αSyn • Olfactory deficits

72 3. Proteomic profiling of a complex disease

Patient derived iPS αSyn gene • Elevated αSyn levels {Devine:2011fr} cells differentiated triplication • Increased expression of oxidative stress genes {Byers:2012ds} to DA neurons • Increased levels of 129S-phosphorylated αSyn {Flierl:2014dw} • Increased production of reactive oxygen species {Lin:2016kh}

Primary Neuronal Application of αSyn • Time dependent accumulation of Lewy body and Lewy neurite {VolpicelliDaley:2011da} Cultures pre-formed fibrils like pathology and neuronal loss {Lindstrom:2017ef} or oligomers • Reduced neural network activity {Karpowicz:2017gw} • Mitochondrial damage and lysosomal dysfunction Cell culture (PC12 Transgenic • Reduced dopamine release {Larsen:2006cf} cells) overexpression of • Impaired exocytosis human WT αSyn

73 3. Proteomic profiling of a complex disease was established to mimic the gradual degeneration observed in PD patients, as per the low-dose rotenone model used, and has been extensively characterized previously by Charmaine Lang (Lang, 2015).

3.1.1.3 The combined rotenone and αSyn model of PD may represent additional aspects of the disease

The observation that most PD-causative gene mutations have low penetrance (Schulte & Gasser, 2011) suggests that their pathogenicity is dependent on additional environmental or genetic factors. Consequently, PD models have been developed that reflect a combination of genetic and environmental aspects of PD. For instance, the pathogenicity of PINK1 deletion in zebrafish is exacerbated by exposure to the pesticide rotenone (Yuxi Zhang et al., 2017). Zebrafish that contain a homozygous loss of function PINK1 mutation develop PD-like movement phenotypes only upon exposure to rotenone (Yuxi Zhang et al., 2017), suggesting that a gene-environment interaction between PINK1 and rotenone may play a role in the disease. Similarly, there is also accumulating evidence from our own lab and others that there may be interaction between mitochondrial dysfunction induced by rotenone and αSyn.

Our lab identified a negatively synergistic relationship between αSyn expression and OXPHOS dysfunction following a screening approach to identify genetic mutations that render yeast sensitive to αSyn expression (Lang, 2015). Yeast were highly sensitive to the combination of OXPHOS mutations and αSyn overexpression, while yeast exposed to αSynuclein or OXPHOS mutation alone were relatively unaffected (Lang, 2015). Additionally, yeast challenged with this combination of OXPHOS mutation and αSyn overexpression also showed a significant increase in the level of mitochondrial ROS (Lang, 2015), a cellular dysfunction implicated to play a central role in PD (Hwang, 2013). These results suggest that combined αSyn overexpression and OXPHOS inhibition may substantially increase αSyn related pathogenicity, suggesting that exposure to environmental toxins that impair mitochondrial function, such as

74 3. Proteomic profiling of a complex disease rotenone, may increase the penetrance of αSyn gene mutations or result in increased cell sensitivity to elevated αSyn levels.

Consistent with the likelihood that there is an interrelationship between αSyn and rotenone, rotenone exposure has been observed to increase αSyn levels (Sala et al., 2013; Wu et al., 2015), enhance αSyn aggregation (Yuan et al., 2015) and induce the formation of αSyn fibrils (Uversky et al., 2001), which are thought to be the highly toxic form of the protein (Mahul-Mellier et al., 2015). There is also some evidence that the toxicity of elevated αSyn levels or PD-associated αSyn mutations is increased by simultaneous treatment with rotenone. For instance, dopaminergic neurons derived from induced pluripotent stem cells harbouring a PD-causative A53T αSyn mutation exhibited oxidative and nitrosative stress only upon addition of a low level of rotenone, while control cells did not (Ryan et al., 2013). Similarly, the behavioural and pathological impairments caused by injection of αSyn adeno-associated virus into the substantia nigra of rats were significantly enhanced by the subsequent injection of rotenone (Mulcahy et al., 2012) (Mulcahy et al., 2013). Alternatively, rats injected with only rotenone displayed minimal pathology or behavioural deficits. These results further implicate a potential relationship between these 2 PD relevant stresses and suggest that the combined rotenone and αSyn model of PD may provide a more accurate representation of certain aspects of the disease. Therefore mass spectrometry proteomic analysis was conducted upon cells treated with the combination of these insults, in addition to assessing the effects of each individual insult alone, to further explore the molecular basis of the potential relationship between these stresses.

3.1.2 Identification of early, potentially causative proteomic changes in PD

The primary aim of this chapter was to identify early and potentially causative events in PD. The approach of using the in vitro rotenone and/or αSyn disease models to achieve this aim circumvents many of the limitations that are associated with the use

75 3. Proteomic profiling of a complex disease of post-mortem human samples. In particular, identifying early disease triggers using patient tissue is difficult because the required samples from affected brain regions are only available post-mortem, meaning that these samples can only provide a single snapshot of the disease at the time of death, which generally occurs following a significant time-period of disease progression, often decades (Braak et al., 2004). At this point substantial pathological damage is present that may mask the earlier changes that could have contributed to degeneration. For example, many patient- based studies assess changes that occur in the substantia nigra, a brain region that exhibits substantial neuronal loss in PD and accumulates extensive Lewy pathology (Lewy bodies and Lewy neurites) (Braak et al., 2004). This pathology and neuronal loss may mask the earlier molecular changes that contributed to the formation of this damage. Thus post-mortem sample analysis does not enable early, potentially disease causative events to be differentiated from the later consequences of pathological damage.

Unlike post-mortem tissue, animal and cell models of PD enable disease related molecular changes to be assessed temporally, such that the early and likely causative changes can be differentiated from the changes that occur later and are more likely to be downstream consequences. To identify the early proteomic changes that occur in the established low-dose rotenone and/or moderately elevated αSyn expression models of PD, mass spectrometry was conducted on samples harvested at multiple time-points during treatment; 24, 48 and 72 hours (Figure 3.1). These time-points were selected based on the observation that 96 hours of treatment with the combination of these insults results in a significant increase in cell death, as determined by Charmaine Lang, while no significant increase in cell death was observed at earlier time-points (Figure 3.1) (Lang, 2015).

Assessment of proteomic changes was conducted using SILAC mass spectrometry. This approach involves culturing cells in media that contains heavy or medium amino acids that become incorporated into >95% of proteins. The heavy label used in this study consisted of arginine and lysine that contained 13C/15N, while the medium consisted of

76 3. Proteomic profiling of a complex disease arginine and lysine containing 13C/2H. Incorporation of these labels into the proteome results in a mass shift in the peptides that can be differentiated upon analysis (Ong et al., 2002). Peptide labeling strategies, such as SILAC, have significantly improved the sensitivity, accuracy and reproducibility of quantitative proteomics by enabling samples to be combined prior to protein extraction, thereby minimizing variability caused by sample processing (Mann, 2006; Ong, 2012).

Figure 3.1 Proteomic analysis of genetic (elevated αSyn expression) and environmental (rotenone treatment) models of PD was conducted at 3 consecutive time-points to establish a timeline of proteomic changes that could contribute to disease progression. The time-points assessed were based on previous assessment of the cell death induced by these treatments, as determined by Charmaine Lang and outlined in detail in (Lang, 2015). 2000 ng/ml DOX represses αSyn transgene expression (αSyn OFF). 1000 ng/ml DOX partially induces αSyn transgene expression (moderately elevated αSyn).

77 3. Proteomic profiling of a complex disease

3.1.3 Cross-referencing with patient data

Despite the many advantages of model systems, including their compatibility with SILAC mass spectrometry and their ability to assess changes temporally, additional information is required to provide evidence that the proteomic changes that occur in vitro also occur in affected PD patient brain regions. Therefore, this study adopted a two-stage approach to identify proteomic changes that are likely to occur early in the disease. Foremost the established in vitro low-dose rotenone and/or moderately elevated αSyn expression models were used to discover early occurring proteomic changes by conducting SILAC mass spectrometry at multiple time-points. Subsequently, the results obtained from this discovery approach were assessed against available patient datasets to determine whether there is evidence that these changes occur in patients. The patent datasets that were used for this purpose are described below.

3.1.3.1 PD patient proteomic data

The published data available from quantification of PD-associated proteomic changes that were used for cross-referencing purposes in this study are described in Table 3.2. These analyses detail the proteomic changes identified in multiple affected patient brain regions, including the heavily pathologically affected substantia nigra (Licker et al., 2014), as well as the less damaged prefrontal cortex (Dumitriu et al., 2016) and striatum (Riley et al., 2014). However, the low sample numbers used in these analyses (as few as 3 patients) significantly limited their statistical power. Additionally, variability between the disease stages of individuals, as well as the heterogeneous nature of the disease adds to variability in the data and substantially limits its usefulness. For this reason available transcriptomic RNA sequencing (RNAseq) data was used as an additional resource for determining whether there is evidence that the molecular changes identified in the in vitro models assessed are also present in patient tissue.

78 3. Proteomic profiling of a complex disease

Table 3.2 Proteomic datasets from PD patient brain regions used for determining whether there is evidence that the proteomic changes identified in vitro also occur in patients, indicating the number of patient samples included.

Licker et al., Reference Dumitriu et al., 2016 Riley et al., 2014 2014 Brain region Substantia Nigra Prefrontal Cortex Cortex Striatum Samples 5 12 5 3 (Patient) Samples 5 12 7 3 (Control)

3.1.3.2 Available PD patient RNAseq data

Recent advances in RNAseq technologies mean that high-depth assessment of transcriptomic changes can be achieved using a relatively small amount of sample. Consequently, the transcriptomic data from patient brain regions that was available for cross-referencing purposes in this study provides a substantially more in-depth assessment of PD-associated changes than the currently available proteomic datasets. This analysis, which is currently unpublished, was conducted by Lee Marshal and Boris Gunnewig (Garvan Institute of Medical Research) and assesses transcriptomic changes in 3 brain regions from 10 PD patients and 10 age-sex matched controls. The patient cohort consisted of Braak stage 4 and 5 patients (Table 2.9) and the three brain regions assessed were selected to represent different stages of disease pathology; the superior occipital cortex (SOC) is a pathologically unaffected region, while in patients at these disease stages the superior frontal cortex (SFC), which becomes affected at a later disease stage, displays a low level of pathology and the basal forebrain (BF), which is affected at an earlier disease stage, exhibits a greater level of damage and accumulation of αSyn, but does not display neuronal loss.

The data obtained from the transcriptomic assessment of these brain regions was of substantially greater sequencing depth than any other available PD transcriptomic assessment conducted to date (Chatterjee & Roy, 2017; Riley et al., 2014) and hence provided an invaluable tool for assessing evidence of changes in PD patients. Proteomic analysis of corresponding samples from this cohort of patients and controls

79 3. Proteomic profiling of a complex disease is currently being conducted using high-throughput mass spectrometry. This data will additionally be used for cross-referencing purposes when it becomes available.

3.1.4 Specific Aims

This study used an approach of combining data obtained from multiple established in vitro PD models with data obtained from affected brain regions of PD patients to identify key areas of interest for further assessment. The primary aim of this chapter can be divided into 3 subsections:

Aim 1: To identify proteomic changes induced by low-dose rotenone and/or moderately elevated αSyn expression at progressive time-points in order to distinguish between early changes, which are more likely to be causative, and later changes, which are more likely to be downstream consequences.

Aim 2: To prioritize the proteomic changes identified in these in vitro models according to evidence of their dysregulation in affected PD patient brain regions by cross- referencing the proteomic data generated from these models with available proteomic and transcriptomic data obtained from patient brain regions.

Aim 3: To investigate the molecular basis of the possible synergistic relationship between αSyn and rotenone by examining the proteomic changes induced by the combination of these insults.

80 3. Proteomic profiling of a complex disease

3.2 Results and Discussion

3.2.1 Overview of SILAC mass spectrometry

SILAC labelled mass spectrometry was conducted to assess proteomic changes induced by treatment of SH-SY5Y cells with 10 nM rotenone and/or moderately elevated αSyn expression (according to the models developed and characterized by Lang, 2015). Proteomic changes were assessed at 3 progressive time points (24, 48 and 72 hours) to enable early, potentially causative events to be distinguished from subsequent changes, which are more likely to be a downstream consequence.

Mass Spectrometry was conducted in collaboration with Benjamin Parker and Daniel Fazakerley at the Charles Perkins Centre, University of Sydney. Sample preparation was conducted by myself with training from Daniel Fazakerley and sample fractionation by Strong Cation Exchange (SCX) and LC-MS/MS analysis was conducted by Benjamin Parker using a Q-Exactive Orbitrap HCD-MS/MS (ThermoFisher scientific). Daniel Fazakerley aligned the resulting peptide reads with the human peptide library database using the MaxQuant software system. All subsequent analysis, as well as confirmation of incorporation of the heavy and medium amino acid labels, was conducted by myself using the Perseus software package as detailed in section 2.2.1.

Using this method over 2000 individual proteins were detected for each treatment at each time point, as summarized in Figure 3.2. A greater number of proteins were consistently detected in cells treated with the combination of elevated αSyn expression and rotenone than in those treated with elevated αSyn expression or rotenone alone due to the increased number of replicates in this group (combined treatment n=4; elevated αSyn expression n=2; rotenone n=2).

81 3. Proteomic profiling of a complex disease

Figure 3.2 Number of proteins identified by SILAC mass spectrometry in the 3 treatment groups at the 3 time-points assessed, indicating the protein overlap identified between the treatments.

3.2.2 Key dysregulated proteins

Quantitative assessment of proteomic changes was achieved by filtering changes for a stringent p-value cut off of 0.01 (based on Students t-test). The proteomic changes induced by low-dose rotenone and/or moderately elevated αSyn expression typically fell within a 15 to 25% range of increased or decreased abundance, relative to control cells. Large changes in protein abundance were not expected due to the low-level and chronic nature of the stresses applied.

82 3. Proteomic profiling of a complex disease

Among the 9 separate datasets obtained, corresponding to the 3 treatment groups at each of the 3 time-points (24, 48 and 96-hours), the levels of 440 individual proteins were significantly altered by greater than 0.2 or less than -0.2 Log2 fold in one or more datasets (p<0.01) (Appendix 1). To prioritize this list into a consolidated set of candidate proteins with likely relevance to PD the 440 proteins were filtered for changes that occurred in 3 or more of the 9 data sets. This approach yielded 7 candidate proteins (KIDINS220, KPNA2, GGH, GLUL, NDUFA9, CHCHD2 and DBH; Table 3.3).

To determine whether these proteins are similarly dysregulated in PD patients the changes associated with these 7 proteins were assessed using previously published proteomic data (prefrontal cortex, Dumitriu et al., 2016; substantia nigra, Licker et al., 2014; cortex and striatum, Riley et al., 2014). However, due to the low coverage obtained by these published proteomic analyses, the majority of these studies did not detect the 7 prioritized candidate proteins (Table 3.4).

Because of the lack of consistent proteomic data supporting the hypothesis that these 7 candidate proteins are dysregulated in PD patients, evidence of their dysregulation in patients was further evaluated by assessing the levels of their corresponding mRNA transcripts in patient brain regions using available RNAseq data. Recognizing that a transcriptional changes do not necessarily correlate with protein changes, this transcriptomic data was viewed as tool to provide the initial necessary evidence that these proteins may be dysregulated in PD, however further assessment of their protein levels in patients may also be necessary.

83 3. Proteomic profiling of a complex disease

Table 3.3 7 proteins were prioritized based on significant changes identified in 3 or more of the 9 datasets with a log2 fold change (FC) greater then 0.2 or less than -0.2. Datasets were filtered for a p-value (P) cutoff of 0.01, as determined by Students t-test. ND= not detected.

24 hours 48 hours 10 nM Rotenone αSyn 10 nM Rotenone 10 nM Rotenone + αSyn 10 nM Rotenone Gene + αSyn αSyn Symbol FC P FC P FC P FC P FC P FC P NDUFA9 0.286 3.14E-07 0.068 0.766 0.124 0.654 0.283 0.006 0.087 0.008 0.278 1.66E-06 CHCHD2 0.499 5.59E-05 0.08 0.028 0.407 0.13 0.496 0.007 ND 0.456 0.037 DBH 0.547 0.001 0.292 0.784 0.197 0.818 0.311 0.002 0.255 0.702 0.168 0.843 GGH 0.088 0.686 0.219 0.005 0.39 0.031 0.142 0.486 0.213 0.025 0.272 0.001 1.11E- GLUL -0.573 1.94E-05 -0.107 0.007 -0.636 0.052 -0.982 -0.04 0.52 -0.786 0.11 05 KIDINS220 0.111 0.74 0.297 0.052 0.417 0.0003 0.108 0.808 0.374 0.005 0.409 0.089 KPNA2 0.012 0.817 0.028 0.218 -0.121 0.429 -0.3 0.008 0.115 0.583 -0.379 0.004

72 hours 10 nM Rotenone + αSyn 10 nM Rotenone Gene αSyn Symbol FC P FC P FC P NDUFA9 0.292 0.006 -0.051 0.036 0.237 0.206 CHCHD2 0.461 0.002 ND 0.486 0.038 DBH 0.48 0.003 0.576 0.502 0.092 0.917 GGH 0.102 0.382 0.242 0.001 0.336 0.014 GLUL -1.072 5.08E-06 -0.449 0.091 -0.992 0.065 KIDINS220 -0.205 0.603 0.555 0.001 ND

KPNA2 -0.272 0.066 0.286 0.005 -0.394 0.019

84 3. Proteomic profiling of a complex disease

Table 3.4 Cross referencing of 7 proteins of interest with available proteomic data from PD affected patient brain regions.

(Licker et al., (Dumitriu et al., 2016) (Riley et al., 2014) Reference 2014) Brain Substantia Nigra Prefrontal Cortex Cortex Striatum region Log2 Fold PD/Control ratio FDR* Fold Change Fold Change Change NDUFA9 0.95 -0.097 0.027 Not detected Not detected CHCHD2 Not detected Not detected Not detected Not detected DBH Not detected Not detected Not detected Not detected GGH 0.64 0.015 0.902 Not detected Not detected GLUL 0.91 -0.052 0.649 Not detected 1.618 KIDINS220 Not detected Not detected Not detected Not detected KPNA2 Not detected Not detected Not detected Not detected *adjusted p-value (Benjamini-Hochberg method)

Using RNAseq data obtained from the SFC, described in section 3.1.3.2, differential expression analysis was conducted to assess the levels of the transcripts encoding the 7 prioritized candidate proteins. This assessment revealed that near significant transcriptional changes are associated with 6 of these proteins in PD patients (KIDINS220, KPNA2, GGH, GLUL, NDUFA9 and CHCHD2) (Figure 3.3). DBH was the exception to this finding and was therefore excluded from the list of candidates.

200 p=0.07 p=0.06 p=0.07 p=0.11 p=0.06 p=0.07 p=0.70 Control

) PD

i r

K 150

(

r o

T 100

a l

p 50

R E

0 KIDINS220 KPNA2 GGH GLUL NDUFA9 CHCHD2 DBH Gene

Figure 3.3 Transcriptional profile of the 7 prioritized targets in the superior frontal cortex of PD patients relative to 10 controls. Data obtained from analysis of RNAseq data. Statistical significance (indicated p value) determined by Students t-test corrected for multiple

85 3. Proteomic profiling of a complex disease comparisons using the two-stage step-up method of Benjamini, Krieger and Yekutieli. Error bars represent standard deviation.

Interestingly, it was noted that the directional change of 5 (KIDINS220, KPNA2, GGH, NDUFA9 and CHCHD2) of the 6 transcripts in the SFC followed the opposite trend to the proteomic change identified in the in vitro models assessed; that is, while the transcript level was reduced in the SFC in PD patients (Figure 3.3), the protein level increased in response to rotenone treatment and/or elevated αSyn expression, or visa- versa in the case of GLUL (Table 3.3). KPNA2 was the exception to this observation and was significantly increased in response to elevated αSyn expression, significantly decreased in response to rotenone treatment and decreased at the transcript level in the brain.

The basis of this apparent paradox between the direction of change of the in vitro protein levels and patient transcript levels is currently not clear, however there is also evidence of a similar discordance between mRNA and protein levels within patients. For example, as will be outlined in section 5 of this thesis, the mRNA level of αSyn was observed to be significantly reduced in the SFC of patients, despite the increase in the αSyn protein level in the same brain region. This inverse correlation between protein and transcript abundance could be a result of positive feedback regulation, wherein protein over-abundance, possibly caused by impaired protein degradation, acts as a signal to reduce transcription (Mazurie et al., 2005; Yeger-Lotem et al., 2004). The correlation between transcript and protein abundance in these PD samples will be further assessed when the corresponding proteomic data for these samples becomes available

This analysis of patient transcriptomic data provides initial evidence that 6 of the 7 prioritized candidate genes/proteins (KIDINS220, KPNA2, GGH, GLUL, NDUFA9 and CHCHD2) are transcriptionally perturbed in patients. To further determine the likelihood that these proteins may be dysregulated in PD patient brain regions, published literature was assessed for evidence of their association with PD or with other PD model systems. Evidence for the possible involvement of KIDINS220, KPNA2

86 3. Proteomic profiling of a complex disease and GGH dysregulation in PD is discussed below, while NDUFA9 and GLUL will be discussed in the later the sections 3.2.4 and 3.2.6, respectively. CHCHD2, which was linked to autosomal dominant cases of familial PD soon after the completion of this mass spec analysis (Funayama et al., 2015), has consequently become a key interest in PD research and is the primary focus of section 4 of this thesis.

3.2.2.1 KIDINS220

Kinase D-Interacting substrate 220kD, KIDINS220, is a synaptic protein that plays a role in neurotrophin signalling and is required for neuronal survival (Gamir-Morralla et al., 2015; López-Menéndez et al., 2009; Neubrand et al., 2012), suggesting that the increase in KIDINS220 levels observed in response to both rotenone and elevated αSyn expression (Table 3.3) may be a protective response. Alternatively, the decrease in KINDINS220 transcript levels observed in patients (Figure 3.3) may sensitize the brain to the increase in αSyn protein levels or OXHPOS dysfunction that are known to occur in PD by impairing this response, predisposing this tissue to neuronal degeneration.

KIDINS220 protein levels have recently been shown to accumulate in affected brain regions in Alzheimer’s disease (AD) and this accumulation was found to strongly correlate with the level of Tau protein, the major protein component of the neurofibriallary tangles that are a pathological hallmark of AD (Gamir-Morralla et al., 2017; López-Menéndez et al., 2013). The finding that KIDINS220 protein levels increase in response to αSyn (Table 3.3) suggests that increased KIDINS220 levels may be a response to amyloidosis, a pathological process that underlies both AD and PD, and is consistent with the view that these two neurodegenerative conditions have overlapping molecular characteristics.

3.2.2.2 KPNA2

Karyopherin Subunit Alpha 2, KPNA2, also known as importin-α1, is a component of the large protein complex that regulates the import of proteins into the nucleus

87 3. Proteomic profiling of a complex disease

(Goldfarb et al., 2004). KPNA2 selects proteins to be imported into the nucleus through recognition of their nuclear localization signal (Lange et al., 2007). The finding that KPNA2 levels are decreased in response to rotenone treatment, but increased in response to moderately elevated αSyn expression (Table 3.3) suggests that these two PD-related stresses may have opposing influences on nuclear protein import. Together with the finding that the KPNA2 transcript levels is reduced in the SFC of PD patients, this data suggests that perturbed nuclear protein import as a potential novel aspect of PD pathobiology.

It has previously been suggested that accumulation of αSyn in the cell nucleus, which has been observed in multiple cell and in vivo systems (Goers et al., 2003; Xianpeng Liu et al., 2011; Surguchov, 2015; Zhou et al., 2013a) and is proposed to induce cell death (Ma et al., 2014), is facilitated by direct interaction with KPNA2 (Ma et al., 2014). Additionally, further supporting a possible relationship between nuclear protein import and αSyn, depletion of another structural component of the nuclear import protein complex, KAP1, has been shown to reduce αSyn protein accumulation (Rousseaux et al., 2016). However, the nuclear localization of αSyn protein has not been observed consistently and therefore remains controversial (Stefanis, 2012), indicating that this potential relationship requires further validation. KPNA2 levels have also previously been observed to decrease in response to rotenone treatment in HeLa cells (Gielisch & Meierhofer, 2015), closely aligning with the results observed in SH-SY5Y cells this study, however the mechanism and function of this response are yet to be explored. Therefore, further understanding this interaction and identification of the specific subset of proteins that are imported through this channel is required to postulate how altered KPNA2 levels could contribute to the degenerative process in PD.

3.2.2.3 GGH

Gamma Glutamyl Hydrolase (GGH) is a lysosomal enzyme that plays a role in the folate biosynthetic pathway. Folate is essential for a number of cellular processes, including

88 3. Proteomic profiling of a complex disease amino and nucleic acid biosynthesis, as it is a co-factor in one-carbon (1-C) metabolism, the process that provides 1-C groups for these biosynthetic reactions (Akhtar et al., 2010). Folate is utilized in the cytosol and the mitochondria but can be stored within the lysosome in a polyglutamated form (Galivan et al., 2000). GGH catalyzes the removal of glutamate from this stored lysosomal pool, enabling the release of lysosomal folate and thereby facilitating its utilization by the cell (Figure 3.4). Consequently, increased GGH levels, as is observed in response to rotenone and elevated αSyn expression from an early time-point (24 hours) (Table 3.3), may facilitate folate release from the lysosome, indicating increased folate requirement. Alternatively, increased GGH levels could instead impair the retention of folate in the lysosome, subsequently resulting in folate depletion.

Figure 3.4 Schematic representation of the role of GGH in folate uptake and retention. Folate is stored in the lysosome in the polyglutamated form. Folypolyglutamate synthase (FPGS) adds polyglutamate attachments to folate, while gamma glutamyl hydrolase (GGH) removes these polyglutamate attachments. Monoglutamated folate can be exported from the lysosome, making it available for biosynthetic reactions in the cytosol or mitochondria.

GGH dysregulation is also evident in affected brain regions of PD patients. RNAseq analysis identified that GGH expression is reduced in the SFC of PD patients (Figure 3.3) and this transcriptional reduction was validated by qRT-PCR (Figure 3.5). Corresponding with this observation, reduced GGH protein levels have also previously

89 3. Proteomic profiling of a complex disease been observed in the substantia nigra of PD patients (Licker et al., 2014), further suggesting a potential role of GGH dysregulation in PD. However further work is required to determine whether these observed reductions in GGH result in increased lysosomal retention of polyglutamated folate or impair the bio-availability of non- glutamated folate for 1-C dependent metabolic processes.

Figure 3.5 GGH transcription is significantly reduced in the SFC of PD patients. GGH transcript levels were assessed by qRT-PCR relative to the expression of the housekeeping gene B2M. Statistical significance determined by Students t-test with Welch’s correction (p<0.05). Error bars represent standard deviation. (n=10 patients/10 age-sex matched controls).

The observation that GGH levels are dysregulated in PD patient brain regions suggests that folate metabolism is likely to also be perturbed, a suggestion that is supported by accumulating evidence, as reviewed by Araújo et al., 2015; Hooshmand et al., 2012; McGarel et al., 2015 and Ramos et al., 2005. Most notably, polymorphisms in the methylenetetrahydrofolate reductase (MTHFR) gene, a central enzyme in folate metabolism which converts 5,10-methylenetetrahydrofolate to 5- methyltetrahydrofolate, have been associated with increased PD risk in a number of populations (Rozycka et al., 2013; Wu et al., 2013b). Additionally, reduced folate levels are associated with the cognitive decline that is prevalent during the later stages of PD (Xie et al., 2017), as well as with dementia and impaired cognitive function in normal ageing (Araújo et al., 2015; Hooshmand et al., 2012; Ramos et al., 2005), suggesting

90 3. Proteomic profiling of a complex disease that impaired folate metabolism may contribute to similar non-movement related symptoms of PD.

This potential contribution of perturbed folate metabolism to PD and other ageing related conditions has only recently begun to be addressed due to difficulties associated with studying the folate metabolic machinery in post-mitotic cells. While the essential role of folate in supporting biosynthetic processes during development is well established (Dhur et al., 1991; Osterhues et al., 2013; Reynolds, 2014), the embryonic lethal nature of mutations in key folate metabolism genes means that the role of folate metabolism in adult, non-proliferative tissues remains largely unexplored. Alternatively, examining the effects of folate supplementation and withdrawal has provided some insight into the essential role of folate metabolism in adult tissue. There is accumulating evidence that reduced folate intake and intestinal malabsorption in the elderly may contribute to the development of neuropsychiatric conditions, such as dementia and cognitive impairment (Araújo et al., 2015). However, the link between folic acid supplementation and PD remains unclear. While folic acid supplementation has been shown to ameliorate motor and sub-cellular defects in PINK1 and Parkin KO drosophila models of PD (Srivastav et al., 2015; Tufi et al., 2014), possibly by enhancing the mitochondrial nucleotide salvage pathway (Tufi et al., 2014), studies examining the effect of folate supplementation and dietary intake on PD risk have failed to establish a link between these factors (Chen et al., 2004; McGeer et al., 1972; Murakami et al., 2010; Schwarz et al., 1992). However, the negative results obtained by folic acid clinical trials for PD may be due to the short clinical trial times used, for instance in the case of Schwarz et al. the effect of folate supplementation on patient motor performance was evaluated following a 2 week trial period. Alternatively, these negative results may be a reflection of the clinical trial features assessed as a measure of the treatment effectiveness. These trials assessed motor performance as a measure of disease improvement (McGeer et al., 1972; Schwarz et al., 1992), however, the finding that increased folate intake is associated with improved cognition in the elderly (Walker et al., 2012) suggests that the effectiveness of folic acid supplementation as a PD therapeutic should be instead addressed in

91 3. Proteomic profiling of a complex disease relation to the neuropsychiatric symptoms that are prevalent in the later stages of the disease, including impaired cognition and dementia. Such trials would substantially improve the current understanding of the potential association between folate metabolism and PD and the regulation of folate metabolism in post-mitotic cells, such as neurons.

3.2.3 Pathways analysis of dysregulated proteins

To gain further insight into the biological pathways that are affected by rotenone and/or elevated αSyn expression, KEGG pathway enrichment analysis was conducted of the dysregulated proteins identified using the WEB based Gene SEt AnaLysis Toolkit (WebGestalt). Pathway enrichment was conducted for each the 3 treatments (elevated αSyn expression, rotenone treatment and elevated αSyn expression + rotenone), assessing proteins dysregulated by greater than 0.2 fold with a p value <0.01 at any of the 3 time-points assessed (Table 3.5). Due to the small numbers of proteins that met these stringent criteria, pathway analysis was also conducted in parallel using a data combination approach, assessing pathways enriched among proteins that met these criteria in any of the 9 separate datasets (Table 3.5; combined datasets, Appendix 2). These approaches identified a number of enriched pathways, some of which have previously been associated with PD and some that present potentially novel aspects of PD pathobiology.

Among the KEGG pathways identified the Oxidative phosphorylation (OXPHOS) pathway, which has previously been associated with PD, stands out as a possible aspect of cell biology that rotenone and αSyn could converge upon to cause dysfunction in PD. The implications of this result are discussed in detail below in section 3.2.4. Intriguingly, among the OXPHOS proteins identified as dysregulated using this pathway analysis approach were components of the V-ATPase. The V-ATPase is not an OXPHOS component, but is likely included in this KEGG pathway due to the

92 3. Proteomic profiling of a complex disease role the protein complex plays in proton translocation, using energy from the hydrolysis of ATP to pump protons into the lumen of endolysosomal compartments. The finding that subunits of the V-ATPase are affected by both rotenone and elevated αSyn expression has interesting implications for PD due to the central role that the V- ATPase plays in regulating lysosomal functions (Nelson, 1992), which are known to be perturbed in PD (section 1.9). The potential involvement of the V-ATPase in PD and the consequences of V-ATPase impairment are explored in depth in section 5 of this thesis.

In addition to the OXPHOS pathway, a number of other pathways were also significantly dysregulated by rotenone, elevated αSyn expression or the combination of these stresses (Table 3.5). Interactome mapping using the String database (EMBL version 10.5) was used to identify overlap between the remaining identified pathways to provide a systems biology approach for assessing the changes triggered by these stresses. This approach identified a further two areas of interest; central carbon- energy metabolism (particularly glycolysis, galactose and nucleotide sugar metabolism) and amino acid metabolism, which are discussed in sections 3.2.5 and 3.2.6, respectively.

93 3. Proteomic profiling of a complex disease

Table 3.5 Pathway analysis of proteins that were significantly dysregulated by greater than 0.2 fold (log2) conducted using the web based gene set analysis toolkit platform indicating the number (#) of proteins identified as dysregulated in each pathway and the significance of the pathway enrichment. Grey values represent datasets where this pathway was not identified. Colours are an indication of pathway significance, as indicated below. Individual proteins that contributed to each pathway in each treatment are detailed in Appendix 2.

10 nM rotenone & Combined data sets 10 nM Rotenone Elevated αSyn elevated αSyn # Adjusted P # Adjusted P # Adjusted P # Adjusted P proteins Value proteins Value proteins Value proteins Value Metabolic pathways 80 1.01E-41 20 2.61E-10 19 1.37E-08 49 1.98E-29 Oxidative phosphorylation 21 1.48E-17 3 1.12E-02 5 4.00E-04 16 3.18E-16 Parkinson's disease 18 4.67E-14 4 1.60E-03 4 1.60E-03 13 2.49E-12 Huntington's disease 20 9.72E-14 4 4.00E-03 5 1.10E-03 14 9.31E-12 Alzheimer's disease 19 1.82E-13 3 1.59E-02 4 3.60E-03 15 2.13E-13 Galactose metabolism 8 3.27E-09 6 3.92E-08 Protein processing in endoplasmic reticulum 13 2.21E-07 2 6.92E-02 2 8.59E-02 10 1.14E-07 Glycolysis / Gluconeogenesis 9 2.49E-07 9 4.95E-10 Amino sugar and nucleotide sugar metabolism 8 3.09E-07 3 1.40E-03 6 1.16E-06 Arginine and proline metabolism 8 7.27E-07 3 1.60E-03 4 8.00E-04 Valine, leucine and isoleucine degradation 7 2.69E-06 3 1.40E-03 2 1.26E-02 4 3.00E-04 Lysosome 10 3.93E-06 4 1.40E-03 5 1.30E-03

p<5.0E-11 p<5.0E-7 p<5.0E-5 p<5.0E-3 p<5.0E-1

94 3. Proteomic profiling of a complex disease

3.2.4 OXPHOS pathway proteins are enriched amongst proteins dysregulated by elevated αSyn expression and/or rotenone treatment

OXPHOS is the primary process through which cells produce ATP. The OXPHOS metabolic machinery is comprised of 5 multi-subunit protein complexes assembled within the inner mitochondrial membrane that are encoded by genes on both the nuclear and mitochondrial genomes. Electrons are supplied from NADH and FADH2 to complexes I and II and are transported between the first four complexes by the electron carriers quinone/CoQ10 or cytochrome c (Alcázar-Fabra et al., 2016). These electrons are ultimately donated to oxygen to produce H2O. This electron transport process is coupled to the translocation of protons from the mitochondrial matrix to the intermembrane space to produce an electrochemical proton gradient that is used by the terminal complex, ATP synthase (complex V), to phosphorylate ADP to ATP (Lenaz & Genova, 2009) (Figure 3.6). Treatment of SH-SY5Y cells with rotenone and/or elevated αSyn expression resulted in increased levels of OXPHOS proteins associated with complex I (NADH dehydrogenase), III (cytochrome c reductase) and IV (cytochrome c oxidase), while Complexes II (succinate dehydrogenase) and V (ATP synthase) subunits were unaffected (Figure 3.7), suggesting that these treatments may impair the proton-translocating components of the OXPHOS machinery.

Figure 3.6 The OXPHOS machinery is comprised of 5 protein complexes embedded in the inner mitochondrial membrane. Transport of electrons between complexes I to IV,

95 3. Proteomic profiling of a complex disease mediated by CoQ/Quinone (Q) and cytochrome c (CytC), is coupled to the translocation of protons from the mitochondrial matrix to the intermembrane space. This produces a proton gradient that it utilized by complex V to generate ATP from ADP.

Figure 3.7 Proteins associated with the OXPHOS metabolic pathway were significantly enriched among proteins dysregulated by rotenone, elevated αSyn expression or the

96 3. Proteomic profiling of a complex disease combination of these treatments in SH-SY5Y cells at each of the 3 time-points assessed (24, 48 and 72 hours). Protein interactions were assessed using the String database to visualize key dysregulated aspects of this pathway. Black nodes represent proteins that were not detected in that dataset.

Reduced levels and/or activity of OXPHOS complexes has been identified in brain tissue of sporadic PD patients, however the identity of the complexes that are affected remains controversial. For instance, while levels of core complex I, II and IV subunits have been observed to be reduced in substantia nigra neurons (Grünewald et al., 2016) and the frontal cortex (Arthur et al., 2009) of PD patients, others have previously identified a decrease in the levels of complex I subunits, but not complex II or IV (Mizuno et al., 1989; Parker et al., 1989; Janetzky et al., 1994; Keeney et al., 2006), or in complex I and IV, but not complex II or III (Schapira et al., 1990). Therefore improved characterization of these changes and understanding of the factors that contribute to this dysregulation is a key area of study and the finding that the PD relevant stresses of rotenone and elevated αSyn expression could contribute to perturbed OXPHOS protein levels may significantly contribute to this area of research.

3.2.4.1 Low-dose rotenone treatment increases OXPHOS protein levels

As a known OXPHOS inhibitor it is somewhat unsurprising that rotenone affects the levels of OXPHOS proteins, however this result may provide additional insight into the mechanism through which rotenone affects OXPHOS and through which environmental rotenone exposure may increase PD risk. In particular, the finding that low-dose rotenone treatment led to an increase in the levels of OXPHOS subunits suggests that this up-regulation may be a protective response that compensates for a loss of activity of these complexes. A similar increase in complex I subunit levels has previously been observed in response to chronic rotenone treatment in dopaminergic MES cells (Jin et al., 2007), however this is not a well recognized effect of rotenone because it is not seen in response to the high-dose, acute levels of rotenone that are more generally used (Annunen-Rasila et al., 2007; Gielisch & Meierhofer, 2015). This highlights a distinct difference between the effects of low and high dose rotenone

97 3. Proteomic profiling of a complex disease treatment and suggests that a compensatory up-regulation may dominate in response to low, non-toxic levels of rotenone, however when the level of rotenone surpasses a threshold this response is likely to be superseded by an apoptotic response (Li et al., 2003).

Another interesting observation regarding the effect of rotenone treatment on OXPHOS protein levels is that, in addition to increasing levels of subunits belonging to complex I, which is the complex known to be directly inhibited by rotenone (Lambert & Brand, 2004), rotenone treatment additionally increased levels of complex III and IV subunits (Figure 3.9). Complexes I, III and IV are known to form a close physical association that constitute a super-complex known as a respirasome (Melber & Winge, 2016). The assembly of this super-complex enhances the efficiency of electron transfer between complex I and III (Althoff et al., 2011; Genova & Lenaz, 2014) and reduces the production of reactive oxygen species (Maranzana et al., 2013; Melber & Winge, 2016; Schägger & Pfeiffer, 2000). It has recently been shown that in addition to inhibiting complex I, rotenone can also cause respirasome disintegration (Jang & Javadov, 2017), which may be a trigger for the observed increase in subunits belonging to all three of these OXPHOS complexes. Impaired respirasome assembly has also been suggested as a possible pathogenic mechanism in PD (Arthur et al., 2009), highlighting the need for increased understanding of the mechanisms that regulate the assembly/disassembly of this super-complex.

3.2.4.2 The effect of αSyn on mitochondrial protein levels may be a result of altered doxyxycline concentration

The observation that elevated αSyn expression results in an increase in complex I, III and IV subunit levels (Figure 3.7) aligns with published evidence indicating that elevated αSyn levels can cause mitochondrial dysfunction (Kamp et al., 2010; Winslow et al., 2010; Nakamura et al., 2011) and suggests that αSyn may influence respirasome assembly and/or stability. To confirm the observation that only specific OXPHOS complexes are affected by elevated αSyn expression, levels of a protein subunit from

98 3. Proteomic profiling of a complex disease each of the five OXPHOS complexes (NDUFB8, SDHB, UQCRC2, COXII and ATP5A, corresponding to complexes I-V, respectively) were assessed by Western blot in response to αSyn overexpression. To obtain a more robust phenotype Tg αSyn expression was induced by the complete withdrawal of DOX to obtain a level of αSyn overexpression that was greater than the moderate elevation of αSyn induced in the mass spectrometry experiments. This assessment confirmed that αSyn overexpression resulted in a significant and substantial increase in level of the complex I subunit NDUFB8 (Figure 3.8). To confirm that this increase is a result of elevated αSyn levels as opposed the reduced DOX concentration used to induce αSyn expression, NDUFB8 levels were additionally assessed in control cells expressing Tg βGal under the regulation of the same DOX inducible system. Surprisingly this assessment indicated that there is a similar increase in NDUFB8 levels in response to DOX withdrawal in βGal control cells (Figure 3.9).

Figure 3.8 NDUFB8 protein abundance is significantly increased in response to DOX withdrawal in DOXoff-αSyn SHSY-5Y cells. Western blot analysis of the levels of 5 OXPHOS subunits corresponding to components of each of the 5 OXPHOS complexes, relative to the level of βtubulin. Statistical significance determined by Students t-test corrected for multiple comparisons using the two-stage step-up method of Benjamini, Krieger and Yekutieli (** p<0.01). Data from 4 biological replicates. Error bars represent standard deviation.

99 3. Proteomic profiling of a complex disease

Figure 3.9 NDUFA8 protein abundance is increased in response to DOX withdrawal in DOXoff-βGal SHSY-5Y cells. Western blot analysis of NDUFA8 protein level relative to the level of βtubulin. Statistical significance assessed by Students t-test. Data from 3 biological replicates. Error bars represent standard deviation.

This observation that the removal of low doses of DOX results in increased levels of the complex I subunit NDUFB8 (Figure 3.9) suggests that DOX can impact mitochondrial function, even when used at a very low concentration. This potentially confounding effect of DOX is beginning to be more widely recognized, with recent reports highlighting the effect that similar low doses of DOX have on inducing mitochondrial fragmentation and the production of mitochondrial ROS (Xing et al., 2017), impairing mitochondrial oxygen consumption (Moullan et al., 2015), suppressing the immune responses mediated by mitochondria (Xing et al., 2017) and down-regulating transcription of genes involved in ATP synthesis and electron transport (Moullan et al., 2015). This ability of DOX to perturb mitochondrial functions is likely a result of the bacterial origin of mitochondria making these organelles sensitive to antibiotic compounds. Specifically, DOX, which exerts antibacterial action by impairing bacterial translation (Nguyen et al., 2014), is thought to also impair mitochondrial translation in eukaryotes (Moullan et al., 2015), a suggestion that aligns with the observation from this study that the increased levels of DOX in the control (Tg repressed) cells are

100 3. Proteomic profiling of a complex disease associated with reduced levels of the nuclear encoded OXPHOS subunit NDUFB8 (Figure 3.9). These combined results therefore urge a high level of caution when using DOX inducible systems to assess mitochondrial functions and highlight the need to confirm that the proteomic changes induced by αSyn in this study are the result of elevated Tg expression, as opposed to changed DOX concentration.

3.2.5 Glycolysis, galactose metabolism and amino sugar metabolism are perturbed by the combination of rotenone treatment and elevated αSyn expression

KEGG pathway analysis indicated that many of the proteomic changes induced by treatment of SH-SY5Y cells with the combination of moderately elevated αSyn expression and low-dose rotenone treatment were associated with glycolysis, galactose metabolism and amino sugar metabolism (Figure 3.10). The intriguing observation that these pathways were dysregulated by the combination of these treatments, but not by either of the individual insults alone (Table 3.5) suggests that these two PD-related stresses may converge to perturb this aspect of cell biology. Additionally, changes in key rate limiting enzymes, including phosphofructokinase L (PFKL) (Beatty et al., 1976) were evident from an early time point (24 hours), suggesting that this may be an early event in the disease.

To exclude the possibility that the observed synergistic effect of combined rotenone and αSyn treatment on these pathways could be the result of the marginally reduced DOX concentration used to induce αSyn expression, the abundance of the glycolytic protein GAPDH was assessed by Western blot in response to the combination of low- dose rotenone treatment and elevated αSyn or βGal expression. Supporting the proposal that this result is an αSyn dependent effect, GAPDH levels were elevated in response to combined treatment with moderately elevated αSyn expression and rotenone, but not perturbed by combined moderately elevated βGal expression and rotenone treatment (Figure 3.11). Additionally, further excluding the possibility that

101 3. Proteomic profiling of a complex disease

Figure 3.10 Proteins involved in glycolysis, galactose metabolism and amino sugar metabolism were significantly enriched among proteins dysregulated by rotenone, elevated αSyn expression or the combination of these treatments in SH-SY5Y cells at each of the 3 time-points assessed (24, 48 and 72 hours). Protein interactions were assessed using the String database to visualize key dysregulated aspects of this pathway. Black nodes represent proteins that were not detected in that dataset.

102 3. Proteomic profiling of a complex disease this is a DOX dependent effect, previous assessment of the transcriptomic changes induced by treatment of RT112 cells with 1000 ng/mL DOX (the same level as was used in this study) did not find evidence of changes in glycolytic pathway genes (Moullan et al., 2015). Taken together this evidence suggests that the increase in glycolytic proteins observed in response to treatment with the combination of low-dose rotenone and moderately elevated αSyn expression is not an effect of the marginal decrease in DOX concentration.

Figure 3.11 The combination of αSyn, but not βGal, and low-dose rotenone causes a significant increase in the level of the glycolytic protein GAPDH. GAPDH protein level was assessed in DOXoff-αSyn and DOXoff-βGal SH-SY5Y cells treated with low-dose (20 nM) rotenone and a moderate induction of transgene expression (2000 ng/mL DOX relative to 1500 ng/mL DOX) for 72 hours. Protein abundance was calculated relative to the level of βTubulin. Statistical significance determined by Students t-test corrected for multiple comparisons using the two-stage step-up method of Benjamini, Krieger and Yekutieli (** p<0.01). Error bars represent standard deviation.

3.2.5.1 Central carbon-energy metabolism is perturbed in PD

Central carbon-energy metabolism is the process through which simple sugars are broken down to produce energy or converted to complex metabolic building blocks, such as nucleic acids. Glycolysis is a central pathway that integrates multiple aspects of

103 3. Proteomic profiling of a complex disease carbon-energy metabolism. The glycolytic pathway enables cells to convert glucose into pyruvate through a series of enzymatic reactions that produce ATP for energy and NADH for OXPHOS reducing power (Figure 3.12A). The glucose degraded through glycolysis can be absorbed in the diet or provided from the degradation of galactose to glucose-6-phophate through the Lelior pathway (Holden et al., 2003) (Figure 3.12B). A single glucose molecule produces 2 ATP molecules through glycolysis. Although this number appears minimal relative to the 34 ATP molecules produced though the mitochondrial OXPHOS pathways, the glycolytic rate is 10-100 times faster than mitochondrial processing (Dienel & Cruz, 2016), meaning that glycolysis can be a significant energy source within cells.

Figure 3.12 Central carbon energy metabolic processes are centred on the glycolytic pathway. Central carbon-energy metabolism is the process through which simple sugars are

104 3. Proteomic profiling of a complex disease converted to complex metabolic building blocks, such as nucleic acids, or broken down to produce energy. Multiple pathways converge on this process, including glycolysis (A), The Leloir/galactose pathways (B), OXPHOS/TCA (C, D), The pentose phosphate pathways (E), Hexosamine metabolism (F), Amino sugar metabolism (G) and the interconversion of sugars pathway (H).

Glycolysis is also directly linked to many other metabolic pathways. For instance, the pyruvate produced through glycolysis feeds into the Krebs/TCA cycle, which is connected to the OXPHOS pathway directly through succinate dehydrogenase {Fernie:2004jn} (Figure 3.12C,D) and indirectly as a provider of NADH, which is an electron donor for OXPHOS complex I. Glycolysis is also linked to the process of amino sugar metabolism through the pentose phosphate pathway (PPP) (often referred to as the hexose monophosphate shunt) (Marin-Valencia et al., 2012) (Figure 3.12E). The PPP is provided with glocose-6-phosphate that is produced in the first step of glycolysis. Shunting of glocose-6-phosphate towards the PPP and away from the glycolytic pathway favors increased production of nucleotides and the antioxidant cofactor NADPH over ATP production, providing a mechanism to coordinate the production of ATP and antioxidants (Marin-Valencia et al., 2012). Similarly fructose-6- phosphate produced in the second step of glycolysis can be channeled towards the hexosamine pathway to produce uridine diphosphate N-acetylglucosamine (UDP- GlcNAc), which is used for N and O linked glycosylation of proteins and lipids (Denzel & Antebi, 2015) (Figure 3.13F). Thus glycolysis is central to the many branches of central carbon-energy metabolism.

Glycolysis, galactose and amino sugar metabolism were perturbed by combined rotenone and αSyn treatment. As each of these pathways contributes to central carbon-energy metabolism it was predicted that other overlapping pathways, such as hexamine metabolism and the PPP could additionally be perturbed. Therefore, proteomic expression analysis was conducted to assess the effects of combined rotenone and αSyn treatment on additional aspects of central carbon-energy metabolism. As predicted, this analysis revealed widespread changes in the abundance of proteins involved in these parallel metabolic pathways (Figure 3.13).

105 3. Proteomic profiling of a complex disease

106 3. Proteomic profiling of a complex disease

Figure 3.13 Proteins involved in central carbon-energy metabolism pathways are perturbed by treatment of SH-SY5Y cells with the combination of low-dose rotenone and moderately elevated αSyn following 72 hours of treatment.

Assessment of these results revealed a number of interesting proteomic changes associated with these overlapping pathways. For instance, the abundance of key rate- limiting glycolytic enzymes (Beatty et al., 1976; Berg et al., 2002), including hexokinase 2 (HK2) and phosphofructokinase (PFKM) were significantly increased, suggesting that there may be increased glycolytic flux. However, the abundance of 3 of the 4 subunits of the pyruvate dehydrogenase (PDH) complex were decreased, while the 4th was not detected by the analysis. This result mimics the up-regulation of glycolytic enzymes and repression of PDH activity that is observed in response to mitochondrial impairment caused by mild hypoxia (Kim et al., 2006). This hypoxic response acts as a metabolic switch between glycolysis and OXPHOS, shunting pyruvate away from the impaired mitochondria to reduce consumption of limited O2 levels and prevent hypoxic mitochondrial ROS production, while maintaining ATP production through glycolysis (Golias et al., 2016; Kim et al., 2006). The observation that a similar response is induced following treatment of cells with the combination of rotenone and elevated αSyn expression suggests that these combined insults may induce mitochondrial dysfunction that triggers a compensatory increase in glycolytic ATP production and reduction of OXPHOS activity. In the case of rotenone, enhanced glycolysis may compensate for the mitochondrial impairment caused by this established OXPHOS inhibitor. However, the observation that rotenone treatment alone does not cause an increase in glycolytic pathway proteins (Table 3.5) indicates that αSyn likely causes an additional increase in glycolytic requirement or impairment of mitochondrial OXPHOS function. Further exploring this relationship may therefore provide significant insight into the molecular function of αSyn and the impact of αSyn on mitochondrial function.

Emerging data suggests that central carbon-energy metabolism pathways are also dysregulated in affected brain regions of PD patients and in other PD model systems. Hypometabolism in the brain, which is characterized by reduced glucose uptake and is indicative of dysregulated glycolysis, is evident early in PD, often prior to the

107 3. Proteomic profiling of a complex disease development of movement symptoms (Berti et al., 2012; Borghammer et al., 2010; Zilberter & Zilberter, 2017), however the basis of this dysregulation remains unknown. Similar hypometabolism has also been observed in multiple PD model systems, including affected brain regions of 6-hydroxydopa treated rats (Casteels et al., 2008; Jang et al., 2012) and in ATP13A2-deficient patient derived olfactory neurospheres and fibroblasts (Park et al., 2016). Additionally, enhancing glycolysis pharmacologically is protective in many PD models. For instance, meclizine, a drug that enhances glycolysis through increasing PFK activity, has been shown to ameliorate cell death in the PD 6- hydroxydopa toxicity model in SH-SY5Y cells and rat cortical neurons (Hong et al., 2016). Similarly, enhancing glycolysis through the addition of glucose can rescue MPTP induced toxicity (Mazzio & Soliman, 2003) and overexpression of the glycolytic enzyme glucose-6-phosphate isomerase (GPI), which catalyses the conversion of glucose-6- phosphate to fructose-6-phosphate, is protective against αSyn induced toxicity in worms, flies and mouse primary neurons (Knight et al., 2014). The protective effect of enhanced glycolysis is thought to be due to the ability of these treatments to restore cellular ATP production through glycolysis, while reducing the mitochondrial ROS production that is associated with OXPHOS-mediated ATP production (Hong et al., 2016). Together these results suggest that impaired glucose uptake and glycolytic dysregulation plays a role in PD and that glycolysis may be an effective therapeutic target in the disease.

Other pathways that converge on glycolysis have also been implicated in PD pathogenesis in patients. For instance, shunting of glucose-6-phosphate towards the PPP to increase production of the NADPH (Figure 3.12) needed for antioxidant production has been suggested to contribute to degeneration in PD. Pathologically affected brain tissue in late stage patients shows increases in the levels of NADPH and increased glucose-6-phosphate dehydrogenase (G6PD) enzyme levels, suggesting that glucose-6-phospate is shunted away from glycolysis and towards the PPP (Dunn et al., 2014). This could result in energy depletion as carbon sugar intermediates are shunted away from energy production and towards antioxidant production. However, despite this accumulating evidence that central carbon-energy metabolism is dysregulated in

108 3. Proteomic profiling of a complex disease

PD, no comprehensive analysis of these pathways in PD has been conducted. Therefore, to better examine this area of interest, RNAseq data obtained from the SFC of PD patients was assessed for evidence of transcriptional changes associated with key central carbon-energy metabolism genes. Supporting the hypothesis that perturbations in these overlapping pathways are likely to contribute to PD, this analysis indicated that there is widespread transcriptional down-regulation of genes involved in glycolysis in PD patients (Figure 3.14). This suggests that glucose uptake and glycolytic energy production are likely to be impaired in affected patient brain tissue. Additionally, corresponding with the result of Dunn et al. (2014), an increase in the transcript levels of enzymes in the pentose phosphate pathway, including Phosphogluconolactonase (PGLS), which is involved in the production of the antioxidant NADPH (Miclet et al., 2001), was also observed. This suggests that carbon sources may be being shunted away from energy production and towards antioxidant production.

The observed decrease in glycolytic pathway genes in the SFC of PD patients (Figure 3.14) suggests that there could be a glycolytic energy deficit in this brain region. Of particular relevance to the neuronal cells that degenerate in PD, glycolytic enzymes are enriched at nerve terminals (Knull, 1980) and among pre-synaptic compartments (Ikemoto et al. , 2003) and glycolysis is suggested to play an essential role in maintaining ATP levels at the synapse (Ashrafi & Ryan, 2017). The down-regulation of glycolytic enzymes that is observed in patients (Figure 3.14) may therefore impair this localized energy production, impairing synaptic transmission and contributing to the degeneration of neuronal cells. Corresponding with this suggestion, impaired synapse localized glycolysis has previously been suggested to contribute to the nerve terminal damage and impaired synaptic vesicle loading (Ashrafi & Ryan, 2017) that have been observed in PD patient brain tissue (Mu et al., 2013; Picconi et al., 2012).

109 3. Proteomic profiling of a complex disease

110 3. Proteomic profiling of a complex disease

Figure 3.14 Transcripts encoding proteins involved in central carbon-energy metabolism are perturbed in PD patients. Data obtained from RNAseq of the superior frontal cortex of 10 Braak stage 4/5 patients and 10 controls (described in section 3.1.3.2).

Intriguingly, cytosolic αSyn is also localized to synaptic nerve terminals and interacts with synaptic vesicles (Bussell & Eliezer, 2003) and synaptic proteins (Burré, 2015; Burré et al., 2010). Additionally, αSyn has been observed to induce clustering of synaptic vesicles at the synaptic membrane in vitro (Diao et al., 2013; Lai et al., 2014), suggesting that the protein plays a role in synaptic vesicle regulation. Although the specific role of αSyn at the synapse remains undetermined, ATP availability appears to promote interaction of αSyn with synaptic membranes (Wislet-Gendebien et al., 2008), suggesting that there may be some relationship between the synaptic function of αSyn and localized glycolytic ATP production. Further examination of this relationship may therefore provide significant insight into the synaptic function of αSyn and the contribution of perturbed glycolysis to PD.

3.2.6 Amino acid metabolism is a significantly deregulated pathway

Proteins involved in amino acid metabolic pathways were dysregulated following treatment of SH-SY5Y cells with rotenone, αSyn and the combination of these two insults (Table 3.5). While rotenone appeared to have a greater effect on proteins involved in valine, leucine and isoleucine metabolism (Figure 3.15A), αSyn had a greater effect on the arginine and proline metabolic pathway (Figure 3.15B). Proteins in these pathways were generally up-regulated by these insults (Figure 3.15A, B), with the exception of Methylcrotonoyl-CoA Carboxylase (MCCC1 and MCCC2) which plays a role in leucine catabolism (Shepard et al., 2015) and was down-regulated. Alternatively, there was no clear up or down-regulation of proteins in the alanine, aspartate and glutamate degradation pathway by any of the treatments (Figure 3.15C).

111 3. Proteomic profiling of a complex disease

Figure 3.15 Proteins involved in amino acid metabolism are perturbed by rotenone and/or moderately elevated αSyn expression. KEGG pathway analysis of proteins that were significantly dysregulated (p<0.01, based on Students t-test) with a fold change of greater than

0.2 or less than -0.2 (log2) by any of the three treatments at any of the 3 time points assessed

112 3. Proteomic profiling of a complex disease

(24, 48 or 72 hours). Black nodes represent proteins that were not detected in that experiment.

Amino acids are the basic building blocks for protein synthesis and of the 20 amino acids used for protein synthesis, 9 are essential and must be obtained through the diet, while the others can be synthesized by the cell. However, dietary amino acid uptake substantially exceeds the amount required for protein synthesis. Therefore the majority of amino acids consumed are instead catabolized for energy production and it has been suggested that the energy obtained from amino acid degradation plays an essential role in compensating for perturbed glucose availability in neurodegenerative conditions, such as AD (Gueli & Taibi, 2013). Reduced glucose uptake has also been demonstrated to occur in PD (Berti et al., 2012; Borghammer et al., 2010; Zilberter & Zilberter, 2017), suggesting that a similar compensatory degradation of amino acids for energy production may also occur in PD. However, degradation of amino acids for energy supply can have detrimental effects that may be particularly pertinent to neuronal cells that have been observed to degenerate in PD. For instance, amino acid catabolism results in a build up of toxic ammonia, which neuronal cells are not able to detoxify (Rao et al., 2005). Consequently, astrocytes support neuronal cells by sequestering the ammonia that is produced and converting it to glutamine. Impaired astrocytic support has been suggested as a potential pathogenic mechanism in PD (Finsterwald et al., 2015; Niranjan et al., 2014) which, together with increased amino acid catabolism, could have a particularly detrimental effect.

Although perturbed amino acid metabolism is not a well studied area of PD, there is evidence that amino acid levels are decreased in PD, however the cause of this decrease remains unknown. Assessment of individual amino acid levels in PD patient plasma and cerebrospinal fluid (CSF) indicates that the levels of valine, leucine and isoleucine are reduced in PD (Molina et al., 1997). This result closely aligns with the observed increase in valine, leucine and isoleucine degradation proteins in response to rotenone treatment (Figure 3.15), however additional experiments are required to

113 3. Proteomic profiling of a complex disease confirm that the increased levels of these proteins corresponds to a decrease in the levels of these amino acids.

Levels of other amino acids are also perturbed in PD patients. For instance, reduced plasma aspartate and glutamate levels have also been observed in PD patients (Tong et al., 2015; Yuan et al., 2013) and lowered glutamate levels have been associated with depression in PD (Tong et al., 2015), while lowered aspartate is correlated with sleep disturbances among patients (Tong et al., 2015). These results suggest that amino acid metabolism is perturbed in PD and that this dysfunction may have symptomatic effects. However, because amino acids play many roles within the cell, in addition to protein synthesis, it can be difficult to interpret the impact that changes in amino acid metabolism can have. For instance, the amino acids glutamic acid, cysteine and methionine are also used to create glutathione, an essential cellular antioxidant molecule (Lu, 2000). Reduced glutathione levels are evident in PD and have previously been suggested as the basis of the increased oxidative damage that occurs at an early stage of the disease (Zeevalk et al., 2008). Additionally, in neuronal cells amino acids also play an essential role in neurotransmission. Glutamate, glycine and aspartate act as neurotransmitters themselves (Brosnan & Brosnan, 2013; Pradhan et al., 2014), while other amino acids are essential neurotransmitter precursors. For instance, serotonin is formed from the amino acid tryptophan (O'Mahony et al., 2015), while tyrosine is used to form dopamine, norepinephrine and adrenalin (O'Mahony et al., 2015; Segura-Aguilar et al., 2014). The case of dopamine is particularly relevant to PD as the loss of dopamine producing cells is a clinical characteristic of the disease.

Supporting the likelihood that perturbed amino acid metabolism could impair dopamine processing in PD, rotenone or elevated αSyn expression were observed to perturb a number of key dopamine processing proteins. In particular, monoamine oxidase A (MAO-A) was increased by these treatments at the 24, 48 and 72 hour time- points (Figure 3.14i). MAO-A is one of two outer mitochondrial membrane associated enzymes, the other being MOA-B, involved in the first step of dopamine detoxification (Shih et al., 1999). While MOA-A and MOA-B are both expressed in glial cells and

114 3. Proteomic profiling of a complex disease astrocytes, MOA-A is additionally expressed in neuronal cells (Youdim et al., 2006). MOA oxidizes dopamine to highly toxic 3,4-dihyroxyphenylacetylaldehyde (DOPAL), producing ammonia and hydrogen peroxide as byproducts. DOPAL is subsequently converted to a less toxic product by aldehyde dehydrogenase (ALDH) (Giovanni et al., 2009). These reactions plays an essential role in the detoxification of cytosolic dopamine, which may accumulate due to impaired sequestration of dopamine into synaptic vesicles as it is taken up from the pre-synaptic cleft or synthesized within the cell (Segura-Aguilar et al., 2014). Therefore perturbations in MAO-A and ALDH levels may be particularly detrimental to the dopaminergic neurons that are known to degenerate in PD.

To further explore the possibility that the PD-relevant stresses of rotenone and αSyn may perturb this dopamine detoxification pathway, levels of ALDH, required for the critical second stage of this detoxification pathway, were examined. Of the 19 functional ALDH genes, 3 are confirmed to oxidize DOPAL; ALDH2, ALDH9A1 and ALDH1A1, (Marchitti et al., 2008). ALDH1A1 was not detected by mass spectrometry. Alternatively, assessment of ALDH9A1 and ALDH2 levels indicate that ALDH2 is significantly increased by 16% following 72 hours of rotenone treatment (p<0.0001) and that this treatment also increased ALDH9A1 levels by 12% (p=0.09). This suggests that low-dose rotenone treatment may up-regulate dopamine detoxification pathways to compensate for increased production of toxic dopamine metabolites. In contrast, despite elevated αSyn expression resulting in a significant 43% increase in MAO-A levels at 72 hours (p=0.004), this treatment did not increase in ALDH levels. This suggests that elevated αSyn expression may accelerate dopamine conversion to DOPAL but not concurrently increase the cells ability to detoxify DOPAL. Supporting the likelihood that impaired detoxification of DOPAL could contribute to αSyn- associated degeneration in PD, accumulation of toxic DOPAL has previously been observed in the substantially degenerated putamen of end-stage PD patients (Goldstein et al., 2013) and down-regulation of ALDH1 has been reported in the heavily pathologically affected substantia nigra (Galter et al., 2003; Grünblatt et al., 2004). The expression of dopamine detoxification pathway genes was also assessed in the SFC, BF

115 3. Proteomic profiling of a complex disease or SOC of PD patients using RNAseq data. Although no significant changes in these transcript levels were identified in these brain regions (Figure 3.16), ALDH is known to be regulated post-trascriptionally (Xu et al., 2015b), suggesting that assessment of transcript levels may not identify changes in ALDH function. Alternatively, transcriptional changes in these genes may not be evident until a late stage of the disease at which substantial degeneration and αSyn accumulation is already evident, such as is observed in the substantia nigra and putamen samples previously observed to display perturbed dopamine detoxification (Galter et al., 2003; Goldstein et al., 2013; Grünblatt et al., 2004).

Figure 3.16 Transcript levels of genes involved in dopamine detoxification are not changed in PD patient brain regions. Data obtained from RNAseq of the superior frontal cortex of 10 Braak stage 4/5 patients and 10 controls (described in section 3.1.3.2).

116 3. Proteomic profiling of a complex disease

This study highlights the potential role of αSyn in contributing to dopamine toxicity and suggests that elevated αSyn expression enhances MOA-A mediated conversion of dopamine to toxic DOPAL, but does not trigger a simultaneous increase in the levels of ALDH proteins required for DOPAL detoxification. This finding therefore suggests that enhancing ALDH activity may be protective in PD and may complement therapeutic intervention strategies currently being trailed that target the dopamine metabolism pathway, such as the MOA-B inhibitor rasagiline (Stocchi et al., 2015).

In summary, this work presents an interesting link between altered amino acid metabolism and PD. Enhanced amino acid catabolism may provide an additional energy source for the cell, however obtaining energy in this way has knock-on effects that may be particularly detrimental to neuronal cells, which do not possess the necessary detoxification mechanism to sequester the ammonia produced as a byproduct of amino acid catabolism (Griffin & Bradshaw, 2017), or in the case of dopaminergic neurons, produced during dopamine detoxification (Giovanni et al., 2009). Additionally, neurons may be particularly sensitive to changes in amino acid metabolism as they rely on amino acid precursors for the synthesis of neurotransmitters including dopamine. The results of this study support the likelihood that amino acid metabolism is perturbed in PD and substantially contribute to this novel and rapidly developing area of PD research.

3.3 Conclusions

The results presented here provide substantial insight into the proteomic changes induced by moderately elevated αSyn expression, low-dose rotenone treatment and the combination of these insults. This approach identified a number of key proteins of interest, such as the folate metabolic enzyme gamma glutamyl hydrolase (GGH), which is increased by rotenone and αSyn treatment from an early time-point (24 hours) (Table 3.3). GGH levels were further confirmed to be transcriptionally perturbed in the

117 3. Proteomic profiling of a complex disease

SFC of PD patients (Figure 3.5). These findings suggest that perturbed folate metabolism may be an early pathogenic mechanism in PD.

This proteomic analysis also identified a number of perturbed pathways of interest (Table 3.5). In particular, these results suggest that rotenone and αSyn treatment converge to perturb central carbon-energy metabolism in the cell and highlight the glycolytic pathway, which was dysregulated by the combination of these stresses, but not by rotenone treatment or moderately elevated αSyn expression alone (Figure 3.11), as a key area for future research. Additionally, the increase in key glycolytic pathway proteins was evident from an early time-point (24 hours), suggesting that perturbed glycolysis may be an early event in the disease. This suggestion is supported by the observation that hypometabolism, which is indicative of dysregulated glycolysis, is evident in the brain tissue of early stage PD patients (Berti et al., 2012; Borghammer et al., 2010; Zilberter & Zilberter, 2017).

This observed dysregulation of glycolytic proteins in response to combined elevated αSyn expression and rotenone treatment also provides insight into the relationship between αSyn and rotenone and highlights glycolytic energy production as a pathway that these insults converge upon. Additionally, this result suggests a possible role of glycolytic energy production in regulating the synaptic function of αSyn. Glycolysis has been proposed to play a role in facilitating localized ATP production at nerve terminals (Ashrafi & Ryan, 2017), an undisputed localization of αSyn (Burré, 2015) and ATP availability is observed to enhance the membrane binding capacity of αSyn (Wislet- Gendebien et al., 2008). Further investigation of this relationship will provide significant insight into this emerging aspect of PD.

Pathway enrichment analysis also indicated that rotenone and αSyn treatment altered the levels of proteins involved in amino acid metabolism (Figure 3.15), including proteins associated with the metabolism of amino acid based neurotransmitters, such as dopamine. In particular, while low-dose rotenone treatment appeared to enhance dopamine detoxification pathways, moderate elevation of αSyn expression may

118 3. Proteomic profiling of a complex disease instead lead to an accumulation of the toxic intermediate product of dopamine degradation, DOPAL. This result closely aligns with the observed accumulation of DOPAL in PD patient brain regions (Goldstein et al., 2013), however additional experimentation is required to confirm that this effect of αSyn is not dependent on the moderate reduction in DOX concentration used to induce Tg expression.

This approach to identifying PD relevant proteomic changes using in vitro models of the disease has fulfilled the aims initially proposed, identifying multiple novel proteomic changes that are evident in affected patient brain regions and are responsive to the PD-relevant stresses of rotenone and/or elevated αSyn expression. Additionally, this approach enabled these results to be prioritized for proteins and pathways that are likely to be dysregulated early in the disease and warrant further investigation, such as GGH and glycolysis. Additional areas of interest that were identified by this approach, namely CHCHD2 and the vacuolar ATPase are examined in depth in sections 4 and 5 of this thesis.

119 4. Post-transcriptional regulation of PD-associated protein CHCHD2

4 POST-TRANSCRIPTIONAL REGULATION OF PD-ASSOCIATED PROTEIN CHCHD2 FACILITATES A MITOCHONDRIAL STRESS RESPONSE

Analysis of the proteomic changes induced by treatment of differentiated SH-SY5Y cells with PD-relevant stresses (section 3) revealed that the level of coiled-helix coiled- helix domain containing 2 protein (CHCHD2) was significantly up-regulated by rotenone from an early time-point onwards (24, 48 and 72 hours). Additionally, assessment of RNAseq data obtained from PD patients indicated that CHCHD2 transcript levels are reduced in PD (section 3). Shortly after these discoveries it was published that mutations in the CHCHD2 gene are associated with autosomal- dominant late-onset PD in Asian populations (Funayama et al., 2015), further supporting the likelihood that CHCHD2 dysfunction contributes to PD.

This chapter confirms that CHCHD2 expression is significantly reduced in brain regions of sporadic PD patients and explores the potential mechanism through which CHCHD2 dysfunction could contribute to PD. Specifically, as CHCHD2 had previously been associated with maintaining complex IV activity and/or stability (Aras et al., 2013), the functional effects of both CHCHD2 knock-down and overexpression on mitochondrial functions and cell viability were assessed.

4.1 Introduction

4.1.1 CHCH Domain Containing Proteins

CHCHD2 belongs to a class of proteins that contain twin CX(9)C “coiled-helix” (CH) domains, which are characterized by 2 cysteine residues separated by 9 amino (Longen

120

4. Post-transcriptional regulation of PD-associated protein CHCHD2 et al., 2009) (Figure 4.1). These cysteine residues form a disulfide bridge that stabilizes the CH fold.

Figure 4.1 CHCH protein domain structure and structures, proposed functions and disease associations of the 9 CHCH family proteins that have been identified in mammalian cells. Green region in CHCHD2 represents the transmembrane element, red in CHCHD4 represents a CPC motif and green in CHCHD5 represents an α-helical segment. CHCH; coiled-

121

4. Post-transcriptional regulation of PD-associated protein CHCHD2 helix coiled-helix, DUF; domain of unknown function, MTS; mitochondrial targeting sequence. Adapted from Zhou et al., (2016) with permission.

There are 9 human CHCH domain family proteins and of the 8 that have been characterised, all have mitochondrial associations, typically related to OXPHOS function (Figure 4.1)(Cavallaro, 2010). It has been suggested that the CHCH conformation plays a role in ROS scavenging by providing free thiols under the reducing conditions of the mitochondria (Liu & Zhang, 2015), however there is also evidence that CHCH proteins play additional roles in regulating other aspects of mitochondrial biology. In particular, many CHCH domain containing proteins, including CHCHD2, 3, 4, 6 and 10 are considered to play a role in the regulation or assembly of cytochrome c oxidase (COX), the 4th complex of the OXPHOS system (Modjtahedi et al., 2016). Additionally, many of the known COX subunit proteins contain CHCH domains, including COX6B1, COX6B2, COX17 and COX19 (Khalimonchuk & Winge, 2008).

Other mitochondrial functions of CHCH proteins have also been described. For instance CHCHD1 (also known as MRPS37) has been identified as a component of the small subunit of the mitochondrial ribosome (Koc et al., 2013). Alternatively, CHCHD3, CHCHD6 and CHCHD10 appear to play a role in regulating mitochondrial cristae morphology and may be subunits of the mitochondrial contact site and cristae organizing system (MICOS) complex (Darshi et al., 2011; C. Ding et al., 2015; Genin et al., 2016). This complex plays a critical role in regulating inner mitochondrial membrane architecture, which is required to sequester cytochrome c (Cyt c) and the proton gradient established by OXPHOS within the mitochondrial inter membrane space and also regulates the assembly of OXPHOS super-complexes (Cogliati et al., 2016). MICOS complex disassembly and the subsequent impaired cristae architecture can therefore lead to diminished OXPHOS capacity and mitochondrial release of Cyt c, which activates cytosolic caspases to trigger apoptotic cell death (Figure 4.2).

122

4. Post-transcriptional regulation of PD-associated protein CHCHD2

Figure 4.2 The mitochondrial contact site and cristae organizing system (MICOS) complex regulates mitochondrial inner membrane architecture. This membrane architecture regulates assembly of OXPHOS super-complexes and is required to sequester the H+ gradient established by the activity of the first 4 complexes of the OXPHOS system (I-IV), to sequester cytochrome c (c), which mediates electron transfer between complexes III and IV but can induce apoptosis if released into the cytosol.

A subset of CHCH proteins has also been described to have additional non- mitochondrial functions. For instance, CHCHD1 (Westerman et al., 2004), CHCHD2 (Aras et al., 2013) and CHCHD3 (Hongyu Liu et al., 2012), have also been localized to the nucleus and are proposed to play a role in regulating transcription of nuclear encoded mitochondrial genes.

Among the CHCH family proteins, CHCHD2 displays high homology to CHCHD10 (Figure 4.1), which has previously been associated with frontotemporal dementia (FTD) and Amyloid Lateral Sclerosis (ALS) (Bannwarth et al., 2014), conditions that both share similar pathogenic markers and degenerative phenotypes with PD. CHCHD10 has been described to play a role in both the regulation of COX assembly and the maintenance of mitochondrial cristae architecture as a component of the MICOS machinery (Genin et al., 2016). Supporting this likely role of CHCHD10 in maintaining cristae integrity, fibroblasts obtained from FTD patients expressing the p.Ser59Leu CHCHD10 clinical

123

4. Post-transcriptional regulation of PD-associated protein CHCHD2 mutation exhibited MICOS complex disassembly and cristae deformation. Additionally, p.Ser59Leu patient fibroblasts exhibited inhibition of apoptosis, which was proposed to be a caused by impaired mitochondrial Cyt c release (Genin et al., 2016). Due to the structural similarity between CHCHD2 and CHCHD10 it is likely that CHCHD2 has a similar function in maintaining cristae structure or regulating mitochondrial Cyt c release.

4.1.2 Proposed Cellular Functions of CHCHD2 and association with PD

CHCHD2 was first associated with mitochondrial OXPHOS capacity through a computational expression screening approach (Baughman et al., 2009). This co- expression based analysis utilized >1400 microarray datasets to identify genes that are co-expressed with OXPHOS pathway genes. CHCHD2 was additionally shown to have a functional impact on OXPHOS, with CHCHD2 knockdown (KD) in human fibroblasts resulting in a significant reduction in COX activity, COX2 protein levels and Oxygen Consumption Rate (OCR; a measure of mitochondrial OXPHOS capacity) (Baughman et al., 2009).

The role of CHCHD2 in regulating OXPHOS capacity was further characterized by Grossman and colleagues, who used a luciferase reporter system to show that CHCHD2 overexpression increases activation of the oxygen responsive element upstream of the OXPHOS complex IV subunit COX4I2 (Aras et al., 2013). Furthermore, they identified that CHCHD2 also contains a putative oxygen responsive element in its own promoter region, which is also activated by CHCHD2 overexpression (Aras et al., 2013). This suggested that CHCHD2 may be a transcription factor that regulates expression of itself and COX4I2 in response to hypoxic conditions. In accordance with this hypothesis they found that CHCHD2 mRNA and protein levels were both increased by hypoxia in HEK293 cells and that, while the majority of CHCHD2 appears to be mitochondrial, a small fraction of the protein is localized to the nucleus (Aras et al., 2015). However this

124

4. Post-transcriptional regulation of PD-associated protein CHCHD2 analysis of CHCHD2 localisation was conducted using a cell fractionation approach, as opposed to visualisation of nuclear CHCHD2 in intact cells by microscopy, and the relative amount of mitochondrial to nuclear CHCHD2 was not quantified. Additionally, the mechanism through which CHCHD2 could be transported from mitochondria back to the nucleus to initiate this transcriptional regulation remains unknown, adding further doubt to the proposal that CHCHD2 acts as a transcription factor.

In addition to proposing that CHCHD2 acts as a transcription factor, Grossman and colleagues assessed the mitochondrial function(s) of CHCHD2. In line with previous reports (Baughman et al., 2009), they found that CHCHD2 depletion reduces mitochondrial OCR as well as OXPHOS complex IV activity and protein levels. Furthermore, they observed that CHCHD2 depletion additionally increased mitochondrial ROS production, decreased mitochondrial membrane potential and decreased the abundance of ROS scavengers (Mn-SOD and Cu-Zn-SOD) and OXPHOS complex I subunit proteins (Aras et al., 2015). They subsequently proposed that CHCHD2 acts in a multi-functional, bi-organellar manner, both regulating OXPHOS capacity in the mitochondria by stabilising complex IV, and additionally functioning in the nucleus as a transcription factor to modulate the expression of a subset of genes in response to hypoxic stress. These CHCHD2 responsive genes were proposed to include CHCHD2 itself and COX4I2, although only activation of the COX4I2 promoter region was confirmed to decline upon CHCHD2 depletion (Aras et al., 2015). Any additional CHCHD2-regulated genes remain yet to be identified.

My interest in CHCHD2 stemmed from the earlier finding that CHCHD2 protein levels are increased by treatment of SH-SY5Y cells with a mild and chronic level of the PD- associated OXPHOS complex I inhibitor, rotenone (section 3). At the time of this discovery (2015) CHCHD2 had been identified as a potential regulator of COX activity (Baughman et al., 2009) or a possible mitochondrial retrograde effector (Aras et al., 2013). I therefore aimed to determine whether CHCHD2 dysregulation is evident in affected sporadic PD patient brain regions and if so, whether this dysregulation could

125

4. Post-transcriptional regulation of PD-associated protein CHCHD2 contribute to the mitochondrial dysfunction evident in PD (Dexter et al., 1994; Dölle et al., 2016; Hauser & Hastings, 2013; Zhang et al., 1999).

During the course of this study interest in the role of CHCHD2 has increased significantly due to the finding that polymorphisms in the CHCHD2 gene are associated with late-onset autosomal dominant PD (Funayama et al., 2015), substantiating the hypothesis that CHCHD2 may play a causal role in PD pathogenesis. To date a number CHCHD2 genetic variants have been associated with increased risk of PD (Table 4.1). These very rare genetic variants appear to be predominantly associated with familial PD in Asian populations (Funayama et al., 2015; Yang et al., 2016), and multiple association analyses have failed to establish a similar connection between CHCHD2 variants and PD in Italian (Rubino et al., 2017), French (Jansen et al., 2015), Swedish (Puschmann et al., 2015) and Canadian (Ming Zhang et al., 2016a) populations or among datasets derived from patients of US origin (Jansen et al., 2015). Additionally, multiple PD associated CHCHD2 variants have been identified in unaffected individuals (Ogaki et al., 2015), suggesting that the occurrence of CHCHD2 variants in the general population is very rare and that the penetrance of these variants is low and may be dependent on additional genetic or environmental factors. Nevertheless, characterization of these rare genetically inherited forms of PD has the potential to substantially improve the current understanding of the disease and could provide addition insight into the underlying disease mechanisms that contribute to the sporadic condition.

126

4. Post-transcriptional regulation of PD-associated protein CHCHD2

Table 4.1 Genetic variants in the CHCHD2 gene that have been associated with PD, indicating the nature of the genetic variant and the population(s) that it has been identified in. *Odds Ratio (OR) has not yet been assessed.

Amino Form of PD OR; p value Population Additional Comments Genetic Variant Acid 5C > T Pro2Leu sPD Japanese 4.69; 0.0025 Japanese (Funayama et al., Variant lies within the proteins (rs142444896) Chinese 4.95, 0.021 2015), Chinese (Foo, Liu, & mitochondrial targeting sequence Tan, 2015) Asian (Yang et al., 2016) 182C > T Thr61Ile Autosomal * Japanese (Funayama et al., Overexpression of this variant does dominant FPD 2015), not rescue CHCHD2 KO phenotypes in drosophila {Meng:2017eb} 434G > A Arg145Gln Autosomal * Japanese (Funayama et al., Overexpression of this variant does dominant fPD 2015), Chinese (X. Yang et al., not rescue CHCHD2 KO phenotypes and sPD 2016) in drosophila {Meng:2017eb}

–9T > G Intron Autosomal 2.51; 0.0004 Japanese (Funayama et al., (rs10043) dominant FPD 2015) 300 + 5G > A Intron Autosomal * Japanese (Funayama et al., Causes exon 2 skipping in SH-SY5Y dominant FPD 2015) cells (Funayama et al., 2015)

127

4. Post-transcriptional regulation of PD-associated protein CHCHD2

4.1.3 Specific aims

Aim 1: To confirm that CHCHD2 expression is significantly reduced in brain regions of sporadic patients

CHCHD2 variants have been associated with familial PD (Funayama et al., 2015; X. Yang et al., 2016) and reduced CHCHD2 transcript levels were evident in brain regions of sporadic PD patients, based on analysis of RNAseq data (section 3). I aimed to confirm that CHCHD2 expression is significantly reduced sporadic PD patients by assessing CHCHD2 transcript levels in multiple patient brain regions using qRT-PCR.

Aim 2: To investigate the mitochondrial function(s) of CHCHD2 and the factors that regulate CHCHD2 expression

CHCHD2 has been suggested to play a role in regulating COX stabilization (Aras et al., 2013; Baughman et al., 2009) and transcriptional regulation in response to hypoxia (Aras et al., 2015). Mass spectrometry analysis also indicated that CHCHD2 protein increases in response to treatment of SH-SY5Y with the PD mimic, rotenone (section 3), suggesting that CHCHD2 up-regulation may act as a mitochondrial stress response. Numerous questions remain surrounding the nature of the cellular signals that mediate these changes in CHCHD2 expression and the mechanisms through which CHCHD2 could respond to these stimuli. Validating these functions and understanding the mechanisms that regulate the different roles of this seemingly multi-functional protein is fundamental to understanding how CHCHD2 dysfunction could contribute to PD and thus may specify novel underlying aspects of the disease biology.

128

4. Post-transcriptional regulation of PD-associated protein CHCHD2

4.2 Results

4.2.1 CHCHD2 expression is reduced in sporadic PD patient brain regions

Variants within the CHCHD2 gene have been shown to cause PD in Asian populations (Funayama et al., 2015). Furthermore, as outlined in section 3, assessment of RNAseq data identified that the level of CHCHD2 transcripts is reduced in the SFC of sporadic PD patients. To confirm that CHCHD2 transcript levels are significantly reduced in PD patient brain tissue, the level of CHCHD2 mRNA was assessed by qRT-PCR in 3 brain regions (BF, SOC and SFC). This assessment revealed that CHCHD2 mRNA was significantly reduced in sPD patients to approximately 30-50% of the level of controls in all 3 brain regions assessed (Figure 4.3). The observation that CHCHD2 expression is significantly reduced in sPD suggests that CHCHD2 is not only a causative factor for rare genetic forms of PD (Funayama et al., 2015), but may also contribute to sporadic cases of PD. Additionally, the observation of reduced CHCHD2 mRNA levels in the SFC, a brain region that degenerates in the later stages of the disease, but did not exhibit overt pathology at the time of death of these patients, suggests that reduced CHCHD2 transcription may be an early event in the disease that occurs before the development of pathology.

Control 4

) * PD

g 2

o * * D

L 2

(

C e

L 0

l s

e -2

r T -4 BF SFC SOC

Figure 4.3 CHCHD2 transcript levels are significantly reduced in the Basal Forebrain (BF), Superior Frontal Cortex (SFC) and Superior Occipital Cortex (SOC) of sporadic PD

129

4. Post-transcriptional regulation of PD-associated protein CHCHD2 patients compared to age/sex matched controls. CHCHD2 transcript levels were assessed by qRT-PCR relative to expression of the housekeeping gene B2M using the delta delta CT method. BF; n=5; SFC n= 10; SOC, n=6. Statistical significance determined by Students t-test using the two-stage step-up method of Benjimini, Krieger and Yekutieli’s correction for multiple comparisons (*p<0.05). Error bars represent standard deviation.

4.2.2 Mitochondrial stress increases in CHCHD2 protein abundance in vitro

CHCHD2 was prioritised as a protein of interest flowing mass spectrometry analysis (section 3.2.2), which identified that CHCHD2 abundance increases rapidly in response to exposure to the PD-relevant mitochondrial inhibitor, rotenone, with significantly elevated levels of CHCHD2 protein observed from 24 hours of treatment onwards (Table 3.3). As discussed in section 3, rotenone is a commonly used PD mimic that induces OXPHOS impairment by inhibiting complex I. Together with the previous suggestion that CHCHD2 may regulate OXPHOS complex IV function (Aras et al., 2013; 2015), this observed increase in CHCHD2 protein in response to rotenone suggests that CHCHD2 accumulation may act as a mitochondrial stress response to compensate for OXPHOS impairment.

To validate that CHCHD2 protein abundance increases in response to rotenone and to determine if other mitochondrial stress inducing treatments result in a similar increase, the effects of various mitochondrial inhibitors on CHCHD2 protein abundance were assessed by Western bot. In addition to rotenone, these inhibitors included antimycin A, an inhibitor of OXPHOS complex III, and CCCP, a mitochondrial proton ionophore that collapses the mitochondrial membrane proton gradient, resulting in mitochondrial depolarisation. To assess whether other cellular stresses also increase the abundance of CHCHD2 the effect of thapsigargin, a chemical that induces ER stress by inhibiting ER calcium uptake, was also tested. Furthermore, to evaluate whether the anticipated increase in CHCHD2 abundance is a response to specific mitochondrial/ER defects, as opposed to a secondary stress caused by a cell death response, the

130

4. Post-transcriptional regulation of PD-associated protein CHCHD2 inhibitors used were titred to determine a concentration that caused a mild impairment (<15%) of cell viability, but did not cause an overt cell death response (Figure 4.4). Accordingly, CHCHD2 protein abundance was assessed in response to treatment with 40 nM rotenone, 5 nM antimycin A, 200 nM CCCP or 250 nM thapsigargin.

Figure 4.4 Rotenone, Antimycin A, CCCP and Thapsigargin concentrations were titred to determine concentration that mildly impact cell viability. SH-SY5Y cells were treated with increasing concentrations of rotenone (A), antimycinA (B), CCCP (C) or thapsigargin (D) for 48 hours. Cell viability was determined by alamar blue assay. Viability was significantly impaired at rotenone concentrations > 60 nM, antimycinA concentrations > 20 nM, CCCP concentrations > 125 nM and Thapsigargin concentrations > 125 nM. Statistical significance determined by one-way ANOVA with using the two-stage step-up method of Benjimini, Krieger and Yekutieli’s correction for multiple comparisons (*p<0.05, ** p<0.01, ***p<0.001, **** p<0.0001; data obtained from 3 biological replicates). Error bars represent standard deviation.

In accordance with the hypothesis that CHCHD2 accumulation is a mitochondrial stress response, protein levels were anticipated to increase in response to treatment with

131

4. Post-transcriptional regulation of PD-associated protein CHCHD2 the mitochondrial inhibitors rotenone, antimycin A and CCCP, but not in response to ER stress induced by addition of thapsigargin. Supporting this hypothesis, rotenone, antimycin A and CCCP treatment resulted in an increase in CHCHD2 protein abundance, although the increase induced by rotenone was deemed non-significant following correction for multiple comparisons (Figure 4.5 A,B). Alternatively, thapsigargin had no effect on CHCHD2 protein levels (Figure 4.5 A,B), suggesting that CHCHD2 protein abundance is specifically responsive to mitochondrial stress and is not effected by ER stress.

Figure 4.5 Mitochondrial stress increases CHCHD2 protein levels but does not impact CHCHD2 transcript level. (A) The effect of mitochondrial stress induced by rotenone (roten.), antimycinA (AntiA) or CCCP or ER stress induced by thapsigargin (thaps) on CHCHD2 protein abundance was determined by Western blot relative to level of the loading control βtubulin. (B) Representative Western blot image. (C) CHCHD2 transcript levels assessed in response to rotenone, antimycinA or CCCP treatment by qRT-PCR, calculated according to the delta delta CT method using the housekeeping gene B2M. Statistical significance determined by one-way ANOVA with using the two-stage step-up method of Benjimini, Krieger and Yekutieli’s

132

4. Post-transcriptional regulation of PD-associated protein CHCHD2 correction for multiple comparisons (*p<0.05, ** p<0.01; Western blot n=6; qRT-PCR n=3). Error bars represent standard deviation.

Intriguingly, CCCP treatment appeared to cause the most substantial increase in CHCHD2 abundance (increased to 350%), despite causing a reduction in cell viability similar to that caused by treatment with thapsigargin, which had no effect on CHCHD2 levels. This provides evidence that the increase in CHCHD2 in response to CCCP is unlikely to be caused by a mild loss of cell viability and suggests that mitochondrial depolarisation, the mechanism through which CCCP induces mitochondrial impairment, may be an overarching trigger for CHCHD2 accumulation. The mitochondrial inhibitors rotenone, antimycinA also induce mitochondrial depolarisation as a secondary consequence of their effects on impairing mitochondrial OXPHOS (Li et al., 2014). Therefore mitochondrial depolarisation may be a common mechanism that regulates CHCHD2 protein levels.

To determine whether the increase in CHCHD2 protein level in response to mitochondrial impairment is caused by increased transcription, the effects of rotenone, antimycinA and CCCP treatment on CHCHD2 transcript levels were assessed. These stresses did not cause an increase in the CHCHD2 transcript level (Figure 4.5C), indicating that the increase in CHCHD2 protein in response to these mitochondrial inhibitors is a result of altered post-transcriptional regulation.

4.2.3 Reduced CHCHD2 expression results in diminished OXPHOS Complex IV subunit levels

To explore the mechanism through which CHCHD2 could mediate a mitochondrial stress response the effects of reduced CHCHD2 expression were assessed in vitro using neuronal-like SH-SY5Y cells. In particular, because CHCHD2 has previously been suggested to maintain OXPHOS complex IV stability and/or assembly (Aras et al., 2013), the effect of CHCHD2 depletion on the levels of OXPHOS proteins was assessed.

133

4. Post-transcriptional regulation of PD-associated protein CHCHD2

4.2.3.1 CHCHD2 transcript and protein levels are reduced by CHCHD2 siRNA

CHCHD2 depletion was accomplished by siRNA-mediated knockdown. To minimise potential off-target effects two distinct siRNAs were selected that each targeted distinct regions in the transcript; CHCHD2-09 targets the open reading frame and CHCHD2-10 targets the 3’ untranslated region of the transcript.

Transfected cells were cultured in the presence of each individual siRNA or a control non-targeting siRNA in differentiating conditions (2% FBS supplemented with retinoic acid). Following 4 days of this treatment CHCHD2 transcript levels were reduced to approximately 25% of control levels by each of the siRNAs (Figure 4.6A). Similarly, CHCHD2 protein levels were reduced to approximately 15% of the level of non- targeting control treated cells (Figure 4.6B).

Figure 4.6 siRNA mediated knockdown of CHCHD2 results in significant reduction of CHCHD2 transcript and protein levels in SH-SY5Y cells. SH-SY5Y cells were separately transfected with two distinct siRNAs targeting the CHCHD2 transcript or a non-targeting (NT) control. (A) CHCHD2 transcript level was assessed by qRT-PCR relative to expression of the

134

4. Post-transcriptional regulation of PD-associated protein CHCHD2 housekeeping gene B2M. (B) CHCHD2 protein abundance was assessed by Western blot relative to the level of the loading control protein βtubulin. (C) Representative Western blot image. Significance determined by One-way ANOVA using the two-stage step-up method of Benjimini, Krieger and Yekutieli’s correction for multiple comparisons (***p<0.001, **** p<0.0001; n=5). Error bars represent standard deviation.

4.2.3.2 CHCHD2 depletion reduces the protein abundance of the OXPHOS complex IV subunit COX2, but does not reduce COX2 transcript levels

To test the hypothesis that CHCHD2 stabilizes or regulates OXPHOS complex IV (cytochrome c oxidase, COX), the abundance of the COX subunit COX2 was assessed in response to CHCHD2 depletion. Supporting this hypothesis, COX2 protein abundance was significantly reduced following depletion of CHCHD2 mediated by either of the 2 siRNAs (Figure 4.7A). Due to the close functional and physical association between complex IV and other OXPHOS complexes, in particular with complex III and complex I (as discussed in section 3.2.4.1), the abundance of other OXPHOS complex subunits was additionally assessed. These subunits included NDUFB8, SDHB, UQCRC2 and ATP5A, which represent subunits of Complex I, II, III and V, respectively. Although reductions in complex III and I subunit levels were evident (Figure 4.7B), these reductions were non-significant, suggesting that complex IV is most prominently effected by CHCHD2 depletion.

The mechanism through which CHCHD2 regulates COX levels and/or activity has not been well characterised. It has been suggested that CHCHD2 is a complex IV subunit of accessory protein and that its depletion may destabilise complex IV and lead to degradation of the subunits (Aras et al., 2013; Baughman et al., 2009). Alternatively, it has also been suggested that CHCHD2 is a transcription factor, which has been proposed to regulate expression of other complex IV subunits (Aras et al., 2015). This presents two possible mechanisms through which CHCHD2 depletion may reduce COX2 protein abundance; (1) transcriptional down-regulation or (2) complex IV

135

4. Post-transcriptional regulation of PD-associated protein CHCHD2 destabilisation. To differentiate between these two possibilities the transcript levels of the OXPHOS subunit genes tested were assessed in CHCHD2 depleted cells.

Figure 4.7 CHCHD2 knockdown reduces COX2 protein levels but does not affect COX2 transcript levels. (A) COX2 and (B) NDUFB8, SDHB, UQCRC2, ATP5A and CHCHD2 protein abundance in response to siRNA-mediated knockdown of CHCHD2 using two distinct siRNAs (CHCHD2-09 and CHCHD2-10) as determined by Western blot relative to the level of the loading control 14-3-3. (C) Representative Western blot image. (D) Levels of NDUFB8, SDHB, UQCRC2, COX2 and ATP5A transcripts in response to CHCHD2 knock-down as determined by qRT-PCR relative to expression if the housekeeping gene B2M. Statistical significance determined by (A) One-way ANOVA or (B, D) Two-way ANOVA using the two-stage step-up method of Benjimini, Krieger and Yekutieli’s correction for multiple comparisons (*p<0.05, ** p<0.01; A, B n=5, D n=3). Error bars represent standard deviation.

136

4. Post-transcriptional regulation of PD-associated protein CHCHD2

Supporting the proposal that CHCHD2 regulates complex IV stability as opposed to influencing complex IV transcription, neither the transcript level of COX2, nor any of the other OXPHOS subunit genes tested were altered by CHCHD2 depletion (Figure 4.7D). Although it remains to be fully elucidated whether this reduction in complex IV subunit levels has a functional impact of OXPHOS activity, this result suggests that CHCHD2 may regulate OXPHOS activity by modulating complex IV.

4.2.4 CHCHD2 accumulates in depolarized mitochondria

The above data indicates that CHCHD2 protein abundance increases in response to mitochondrial stresses that reduce mitochondrial membrane polarisation and that CHCHD2 depletion reduces COX2 protein levels, possibly due to impaired Complex IV stability. Taken together these results suggest that CHCHD2 protein accumulation may be a stress response that functions to stabilise or boost OXPHOS upon mitochondrial impairment. In accordance with this hypothesis, CHCHD2 would be predicted to accumulate in mitochondria following treatment with the mitochondrial inhibitors used above (rotenone, antimycin A or CCCP). To assess this hypothesis the effect of these mitochondrial inhibitors on CHCHD2 was visualised by confocal microscopy with co-staining for mitochondria. Supporting the above prediction CHCHD2 predominantly localised to mitochondria and the intensity of mitochondrial CHCHD2 signal increased in response to mitochondrial impairment mediated by rotenone, antimycin A and CCCP treatment (Figure 4.8). A very low level of CHCHD2 positive staining was also observed in the nucleus that did not appear to be affected by mitochondrial impairment.

Close observation of the subcellular localisation of CHCHD2 in response to rotenone, antimycin A and CCCP treatment also indicated that the increase in CHCHD2 protein abundance induced by these mitochondrial toxins predominantly localised to damaged mitochondria, as indicated based on their lower degree of membrane polarisation

137

4. Post-transcriptional regulation of PD-associated protein CHCHD2

Figure 4.8 Mitochondrial impairment mediated by rotenone, antimycinA and CCCP results in increased levels of CHCHD2 protein in the mitochondria. CHCHD2 protein localisation was determined by confocal microscopy using a DM5500 SP8 basic confocal microscope (Leica microsystems). Cells were stained with the mitochondrial marker MitoTracker, fixed and imaged for CHCHD2 and the nuclear marker DAPI using the immunofluorescence protocol detailed in section 2.1.12.1. Scale bar represents 15 μm.

138

4. Post-transcriptional regulation of PD-associated protein CHCHD2

(Figure 4.9). Mitochondria were stained using the MitoTracker (CMX/ROS) stain, which accumulates in mitochondria based on membrane polarisation. This membrane potential dependent accumulation enables less polarised mitochondria to be differentiated based on the intensity of MitoTracker staining, with lower MitoTracker staining intensity representing lower membrane potential. In cells treated with rotenone, antimycinA or CCCP, which are all treatments that induce mitochondrial depolarisation (Dispersyn et al., 1999; Li et al., 2014), mitochondria with higher membrane potential (intense MitoTracker staining) displayed little co-localisation with the CHCHD2 signal (Figure 4.9A), while mitochondria with lower membrane potential (faint MitoTracker staining) accumulated high levels of CHCHD2 (Figure 4.9B). The observation that CHCHD2 preferentially accumulates in mitochondria with a lower degree of membrane potential is consistent with the hypothesis that mitochondrial depolarisation is a common mechanism that triggers CHCHD2 accumulation in response to mitochondrial impairment. This is an exciting preliminary result that will need to be confirmed with the addition of a non-membrane potential dependent mitochondrial marker prior to publication.

4.2.5 Elevated CHCHD2 expression does not perturb mitochondrial functions but highlights a potential post-transcriptional regulatory mechanism in modulating CHCHD2 protein levels

CHCHD2 protein was found to increase in response to mitochondrial impairment and to accumulate in depolarised mitochondria. However, the function of this increased CHCHD2 abundance remains unknown. Because mitochondrial depolarisation plays an essential role in initiating PINK1/PARKIN mediated mitophagy, CHCHD2 accumulation was assessed as a possible mitophagy trigger. For this purpose a DOX inducible CHCHD2 SH-SY5Y stable cell line [DOX-CHCHD2] and a parallel GFP expressing control

139

4. Post-transcriptional regulation of PD-associated protein CHCHD2 line [DOX-GFP] were constructed to explore the effect of CHCHD2 up-regulation on mitophagy.

140

4. Post-transcriptional regulation of PD-associated protein CHCHD2

Figure 4.9 CHCHD2 accumulates in depolarised mitochondria. CHCHD2 was visualised by confocal microscopy using a DM5500 SP8 basic confocal microscope (Leica microsystems) in SH-SY5Y cells treated with 250 nM CCCP for 48 hours. Cells were stained with the mitochondrial marker MitoTracker, fixed and imaged for CHCHD2 and the nuclear marker DAPI using the immunofluorescence protocol detailed in section 2.1.12.1. A) Mitochondria with intense MitoTracker staining showed little co-localisation with CHCHD2. B) Mitochondria with

141

4. Post-transcriptional regulation of PD-associated protein CHCHD2 lower MitoTracker fluorescence intensity showed substantial co-localisation with CHCHD2 following artificial enhancement of the MitoTracker signal (mask image) using ImageJ software. Scale bar represents 10 um.

4.2.5.1 The increase in CHCHD2 protein caused by CHCHD2 overexpression is disproportionate to the increase in CHCHD2 mRNA

The DOX-CHCHD2 and DOX-GFP cell lines were characterised to ensure DOX dependent transgene overexpression and to confirm that DOX alone has minimal effect on CHCHD2 mRNA or protein levels. Interestingly, treatment of DOX-GFP control cells with 2000 ng/mL DOX caused an unexpected small but significant ~35% increase in CHCHD2 transcript levels (Figure 4.10A). However, no corresponding increase in CHCHD2 protein was evident (Figure 4.10B). Based on the proposed suggestion that CHCHD2 protein levels accumulate in response to mitochondrial depolarisation, this result suggests that this level of DOX treatment does not impair mitochondrial membrane potential.

Alternatively, confirming DOX dependent transgene expression in the DOX-CHCHD2 line, increasing concentrations of DOX resulted in a dose dependent increase in both CHCHD2 mRNA and protein in DOX-CHCHD2 cells. Addition of 125 ng/mL or 2000 ng/mL DOX increased CHCHD2 mRNA levels by 62% and 148%, respectively (Figure 4.10). A corresponding increase in CHCHD2 protein was also evident and intriguingly the increase in protein abundance was substantially greater than the observed mRNA increase, with 125 ng/mL or 2000 ng/mL DOX resulting in a 2580% and 3270% increase in CHCHD2 protein, respectively. This disproportionately greater increase in CHCHD2 protein than mRNA suggests that a post-transcriptional regulatory mechanism may be present which results in CHCHD2 overexpression inducing translational up-regulation of itself.

142

4. Post-transcriptional regulation of PD-associated protein CHCHD2

Figure 4.10 Characterization of DOX inducible CHCHD2 and GFP overexpressing SH-SY5Y cell lines. A) Effect of DOX supplementation on CHCHD2 mRNA level in DOX inducible GFP or CHCHD2 overexpressing cells as determined by qRT-PCR relative to expression of the housekeeping gene B2M. B) CHCHD2 protein level in DOX inducible GFP or CHCHD2 overexpressing cells as determined by Western blot using the loading control βtubulin. Statistical significance determined by two-way ANOVA using the two-stage step-up method of Benjimini, Krieger and Yekutieli’s for correction for multiple comparisons (**p<0.01, ***p<0.001, **** p<0.0001; data obtained from 3 biological replicates).

4.2.5.2 CHCHD2 overexpression does not perturb mitochondria or trigger mitophagy

The established DOX-CHCHD2 and DOX-GFP cell lines were used to assess the effects of CHCHD2 up-regulation on mitochondria. As a proposed initiation signal for mitophagy, CHCHD2 overexpression was predicted to reduce mitochondrial mass and increase translocation of mitochondrial contents to the lysosome. However, contrary to this prediction, the overexpression of CHCHD2 had no effect on either mitochondrial mass or mitophagy index (a measure of the accumulation of

143

4. Post-transcriptional regulation of PD-associated protein CHCHD2 mitochondrial contents in the lysosome) (Figure 4.11A,B). Similarly, although CHCHD2 depletion has previously been observed to increase mitochondrial ROS production (Aras et al., 2013), overexpression of CHCHD2 had no effect on mitochondrial ROS levels (Figure 4.11C). Cell viability was also assessed in response to CHCHD2 overexpression. Although a slight reduction in cell viability was evident when CHCHD2 overexpression was induced by addition of 2000 ng/mL DOX, there was no significant difference evident between these cells and GFP overexpressing cells treated with the identical level of 2000 ng/mL DOX (Figure 4.11D), suggesting that this reduction in viability is likely to be, at least in part, a result of DOX supplementation. The absence of mitochondrial phenotypes in response to CHCHD2 overexpression suggests that if CHCHD2 does play a role in mediating mitophagy, the endogenous level of the protein is sufficient to fulfil these functions and increased CHCHD2 abundance causes no additional mitochondrial perturbations.

144

4. Post-transcriptional regulation of PD-associated protein CHCHD2

Figure 4.11 CHCHD2 overexpression does not effect levels of mitochondrial ROS, induce initiation of mitophagy, perturb mitochondrial mass or reduce cell viability. A) Mitochondrial ROS was assessed by flow cytometry using the mitochondrial superoxide indicator mitoSOX. B) Mitophagy index was calculated based on the proportion of mitochondria in the acidic environment of the lysosomal lumen, relative to free cytosolic mitochondria using the mitophagy probe mKeima by flow cytometry. C) Mitochondrial mass was assessed by flow cytometry using MitoTracker green. D) Cell viability was assessed using the Alamar blue assay. Statistical significance was assessed by (A,B,C) One-way ANOVA or (D) two-way ANOVA with two-stage step-up method of Benjimini, Krieger and Yekutieli’s for correction for multiple comparisons (**p<0.01, ***p<0.001, **** p<0.0001; data obtained from 3 biological replicates).

4.3 Discussion

4.3.1 CHCHD2 is involved in the maintenance of mitochondrial cristae structure and the sequestration of potentially toxic cytochrome c within mitochondria

Interest in the function of CHCHD2 has rapidly developed following the discovery that CHCHD2 variants are significantly associated with rare familial forms of PD (Funayama et al., 2015). Furthermore, the results of this study additionally demonstrate that CHCHD2 dysfunction may also play a broader role in non-familial forms of PD, with significantly reduced CHCHD2 expression observed in affected brain regions of sporadic patients (Figure 4.3). Consequently, understanding the function of CHCHD2 has the potential to expose novel aspects of PD pathobiology that may be of relevance to both rare familial forms of the disease as well as more common sporadic cases.

During the writing of this thesis the pathogenicity of the initially described PD- associated CHCHD2 variants were confirmed using an in vivo drosophila model (Meng et al., 2017). Huattori and colleagues, who identified the first PD-associated CHCHD2 variants (Funayama et al., 2015), showed that whole body depletion of the CHCHD2 ortholog in drosophila led to multiple PD-like phenotypes, including increased oxidative stress, age-dependent loss of DA neurons, climbing defects and shorter

145

4. Post-transcriptional regulation of PD-associated protein CHCHD2 lifespan (Meng et al., 2017). Furthermore, while overexpression of human wild type CHCHD2 ameliorated these PD-associated phenotypes in drosophila, overexpression of the PD-causative clinical missense mutations 434G>A or 182C>T did not (Meng et al., 2017). Additionally, drosophila with a mutation in the CHCHD2 gene (CHCHD2-h43) that was proposed to recapitulate a PD-associated human clinical mutation also exhibited PD-associated characteristics, including reduced ATP production and climbing ability (Meng et al., 2017), thereby further confirming the causative role of CHCHD2 dysfunction in PD.

In addition to confirming the pathogenicity of PD-associated CHCHD2 variants, Huattori and colleagues also explored the possible function of CHCHD2 and the mechanism through which CHCHD2 dysfunction contributes to PD. In contrast to previous cell culture studies (Aras et al., 2015), a direct physical association between CHCHD2 and complex IV was not observed in drosophila (Meng et al., 2017). However, the reduction in mitochondrial oxygen consumption rate and ATP production observed upon CHCHD2 depletion (Meng et al., 2017) suggested that the protein is required for full OXPHOS function. Additionally, CHCHD2 was found to directly interact with both the electron carrier Cyt c as well as MICS1 (Meng et al., 2017), a component of the MICOS complex that is involved in maintaining mitochondrial cristae structure (discussed in section 4.1.2). Furthermore, CHCHD2 depletion resulted in impaired mitochondrial cristae structure in drosophila, suggesting that CHCHD2 is required to maintain cristae structure through the observed interaction with MICS1 (Meng et al., 2017). Based on these results Huattori and colleagues proposed a model of CHCHD2 function wherein the protein regulates OXPHOS capacity by facilitating Cyt c-mediated electron transfer between complexes III and IV and through playing a role in maintaining mitochondrial cristae structure, which is required to support OXPHOS super-complex assembly and maintain the proton gradient (Cogliati et al., 2016) (Figure 4.12). Supporting this proposal, overexpression of MICS1 was found to ameliorate the defects in ATP production and climbing ability caused by PD-associated CHCHD2 variants. Additionally, it was further suggested that CHCHD2 also plays a role

146

4. Post-transcriptional regulation of PD-associated protein CHCHD2 in regulating Cyt c-induced apoptosis by sequestering Cyt c within the mitochondria (Meng et al., 2017) (Figure 4.12). The suggestion that CHCHD2 regulates Cyt c-induced apoptosis is supported by the previous observation that CHCHD2 depletion sensitises cells to apoptotic stimuli (Liu et al., 2015c).

Figure 4.12 CHCHD2 regulates mitochondrial cristae structure and cytochrome c release. Huattori and colleagues proposed a model of CHCHD2 function wherein CHCHD2 interacts with MICS1 and cytochrome c (Cyt c) to maintain mitochondrial cristae structure and sequester cytochrome c within mitochondria. Release of cytochrome c into the cytosol leads to activation of the caspase cascade, which initiates apoptosis. Adapted from Meng et al. (2017) with permission.

This very recently (June 2017) proposed CHCHD2-Cytc-MICS1 model is an interesting novel aspect of mitochondrial biology that aligns with the results of this thesis. For instance, the CHCHD2-Cytc-MICS1 interaction model may explain the finding that CHCHD2 overexpression had no effect on the mitochondrial parameters assessed, including mitochondrial mass or mitophagy index (Figure 4.11). If CHCHD2 is a structural component of a CHCHD2-Cytc-MICS1 complex, elevated levels of CHCHD2 alone may have no impact on assembly or function of this complex. Similarly, the observation that complex IV subunit levels are reduced upon CHCHD2 depletion

147

4. Post-transcriptional regulation of PD-associated protein CHCHD2

(Figure 4.7), which has also been observed by others (Baughman et al., 2009), may also be explained by the CHCHD2-Cytc-MICS1 model. This model suggests that CHCHD2 depletion may impair mitochondrial retention of Cyt c and reduce its availability for facilitating electron transfer during OXPHOS. However, reduced Cyt c availability is reported to impair complex IV assembly, causing complex IV destabalisatiation and degradation of subunits (Fontanesi et al., 2008; Vempati et al., 2009). Therefore the observed reduction in the level of the complex IV subunit COX2 upon CHCHD2 depletion (Figure 4.7) may be due to impaired Cyt c availability, rather than the direct interaction between CHCHD2 and Complex IV that has previously been suggested to be the basis for this relationship (Aras et al., 2013; Baughman et al., 2009).

The proposed CHCHD2-Cyt c-MICS1 model suggests that CHCHD2 impairment contributes to cell dysfunction in PD my impairing mitochondrial OXPHOS capacity and sensitising cells to apoptotic stimuli by impairing mitochondrial retention of Cyt c. The results of this thesis build on this model and provide substantial additional insight into the function of CHCHD2. In particular, the results described in this chapter suggest that the interaction between CHCHD2, Cyt c and MICS1 may underlie a more specialised function of CHCHD2 than has previously been recognised. When exploring the possible function of CHCHD2, the protein was found to be up-regulated in response to mitochondrial impairment with rotenone, antimycin A and CCCP (Figure 4.5), treatments that are all known to induce mitochondrial depolarisation (Dispersyn et al., 1999; Li et al., 2014). Additionally, CHCHD2 was observed to accumulate in less polarised mitochondria (Figure 4.9), suggesting that the loss of mitochondrial membrane potential is a common trigger for increasing CHCHD2 protein abundance. This finding suggests that, in addition to the proposed model of CHCHD2 function wherein the protein plays a generalised role in all mitochondria, CHCHD2 may additionally mediate a mitochondrial stress response to restore mitochondrial membrane potential in a subset of damaged mitochondria. This restoration of mitochondrial membrane potential could be facilitated by CHCHD2-mediated enhanced availability of Cyt c for OXPHOS and/or restoration of inner membrane

148

4. Post-transcriptional regulation of PD-associated protein CHCHD2 structural integrity, which is required to retain the mitochondrial proton gradient and regulate the assembly of OXPHOS super-complexes (Cogliati et al., 2016). Furthermore, if mitochondrial membrane potential cannot be restored CHCHD2 may play an additional role in PINK1/PARKIN-mediated mitophagy, a process that is also initiated by loss of mitochondrial membrane potential and has been implicated as a pathogenic mechanism in PD (as discussed in section 1.11.1). In particular, CHCHD2 accumulation in damaged mitochondria may facilitate quarantining of Cyt c during the early stages of mitophagy as the mitochondrion is being isolated within an autophagosome compartment. This Cyt c quarantining capacity may be facilitated either by direct CHCHD2-Cyt c interaction, or by sequestering Cyt c within the folds of the inner mitochondrial membrane through MICOS-mediated closure of cristae junctions. Alternatively, failure to sufficiently sequester Cyt c during mitophagy could instead lead to leakage of Cyt c into the cytosol, which subsequently triggers apoptotic cell death by activation of the caspase cascade (Kulikov et al., 2012).

This proposed amended model of CHCHD2 function (Figure 4.13) suggests that CHCHD2 accumulation in depolarised mitochondrial may have a dual purpose; primarily mediating restoration of mitochondrial membrane potential by restoring cristae structure and OXPHOS capacity, however if this cannot be achieved CHCHD2 may additionally facilitate Cyt c sequestration during the early stages of mitophagy. Depletion or mutation of CHCHD2 may contribute to PD by perturbing either or both of these functions, thereby preventing recovery from depolarisation and/or sensitising cells to inadvertent mitochondrial Cyt c release during mitophagy. This model of CHCHD2 as a mitochondrial damage response factor adds to the previously published CHCHD2-Cyt c-MICS1 model and suggests that CHCHD2 dysfunction could contribute to PD by impairing mitochondrial recovery in response to moderate levels of mitochondrial impairment and/or impairing the isolation of Cyt c within the mitochondrion during mitophagy.

149

4. Post-transcriptional regulation of PD-associated protein CHCHD2

Figure 4.13 A model for the role of CHCHD2 as a mitochondrial stress responsive protein. CHCHD2 accumulates in depolarised mitochondria to facilitate restoration of membrane potential or mediate cytochrome c sequestration during mitophagy. Reduced levels of CHCHD2 impair sequestration of cytochrome c during mitophagy, leading to cytosolic accumulation of cytochrome c and apoptosis through activation of the caspase cascade.

4.3.2 CHCHD2 is regulated at the transcriptional and translational level

Rotenone, antimycin A and CCCP, but not thapsigargin treatment caused an increase in CHCHD2 protein levels, indicating that CHCHD2 is up-regulated by mitochondrial

150

4. Post-transcriptional regulation of PD-associated protein CHCHD2 stress, but not by ER stress (Figure 4.5A). Additionally, these mitochondrial stresses increased CHCHD2 protein, but not mRNA levels (Figure 4.5B), suggesting that CHCHD2 mRNA and protein levels may be regulated separately. Understanding the different mechanisms through which CHCHD2 mRNA and protein levels are regulated may provide substantial insight into the cause of the reduced CHCHD2 mRNA levels observed in sporadic PD patient brain regions and the mechanism through which PD- associated CHCHD2 variants may contribute to the disease. To explore the potential mechanisms through which CHCHD2 expression is regulated, available published data was assessed to consolidate information regarding different cellular stresses that regulate CHCHD2 mRNA and/or protein levels (Table 4.2).

Table 4.2 CHCHD2 mRNA and protein responses causes by various mitochondrial or cellular stresses.

CHCHD2 CHCHD2 Treatment Model mRNA protein Reference Stress change change Not Meng et PINK1 KO Genetic deletion Drosophila 6 X assessed al., 2017 Not Meng et PARKIN KO Genetic deletion Drosophila 9 X assessed al., 2017 Not Meng et DJ1 KO Genetic deletion Drosophila 2.5 X assessed al., 2017 Mutation of DNA Mitochondria polymerase g Meng et l DNA Drosophila 1.4 X 1.4 X subunit (b al., 2017 depletion subunit) Overexpression Mitochondria of DNA Meng et l DNA Drosophila 2.5-3 X 1.3 X polymerase g al., 2017 depletion subunit A Substantial 4% vs 20% O2 (48 HEK293 increase Aras et al., Hypoxia 3 X hours) cells (not 2013 quantified)

151

4. Post-transcriptional regulation of PD-associated protein CHCHD2

3% vs 21% O2 (5 Not Meng et Hypoxia Drosophila No Change days) assessed al., 2017 90% vs 21% O2 (4 Not Meng et Hyperoxia Drosophila 0.7 X days) assessed al., 2017 Overexpression Mitochondria of truncated Meng et l unfolded version of Drosophila 2.5 X 1.3 X al., 2017 protein stress ornithine transcarbamylase 1.3-1.5 X by OXPHOS RNAi of various Not complex V Meng et subunit Drosophila OXPHOS subunits assessed subunits al., 2017 depletion depletion Substantial Not decreased Lui et al., UV damage UV irradiation U2OS cells assessed (not 2015 quantified) OXPHOS complex SH-SY5Y No Rotenone 1.9 X Figure 4.5 I impairment cells Change OXPHOS complex SH-SY5Y No Antimycin A 2.8 X Figure 4.5 III impairment cells Change Mitochondrial SH-SY5Y No CCCP 3.5 X Figure 4.5 depolarisation cells Change

A number of stresses, including hypoxia and mtDNA depletion, caused a change in both CHCHD2 mRNA and protein levels. Hypoxia was observed to cause the greatest transcriptional increase in CHCHD2 (3X) in mammalian cells and this increase was proposed to be caused by the presence of a putative hypoxia responsive element (HRE) in the promoter region of the gene (Aras et al., 2013). Interestingly, this response was not reproduced in drosophila and the HRE was not evident in the drosophila gene promoter (Meng et al., 2017), suggesting that the transcriptional regulation of CHCHD2 may differ among organisms. Hypoxia has been implicated as a potential pathogenic mechanism in PD (Seccombe et al., 2013), suggesting that the

152

4. Post-transcriptional regulation of PD-associated protein CHCHD2 observed dysregulation of CHCHD2 expression in the disease may be associated with hypoxia. However the increase in CHCHD2 transcription that would be predicted to result from increased hypoxia in PD contradicts the observed decrease in CHCHD2 mRNA level in patient brain tissue (Figure 4.3). This suggests that hypoxia is not the cause of the observed reduction in CHCHD2 mRNA level in PD.

CHCHD2 mRNA and protein levels (Meng et al., 2017) were also up-regulated by mitochondrial DNA (mtDNA) depletion. This observation suggests that mtDNA destabilisation may trigger a CHCHD2 transcriptional response and that mtDNA levels may be inversely correlated with CHCHD2 mRNA levels. In line with this discovery, mtDNA accumulation is evident in the frontal cortex of sporadic PD patients (Dölle et al., 2016), a brain region that additionally demonstrated reduced CHCHD2 mRNA levels (Figure 4.3), suggesting that the down-regulation of CHCHD2 mRNA in this brain region may be related to changes in mitochondrial DNA content. Supporting this suggestion, assessment of RNAseq data from the superior frontal cortex of PD patients identified a significant inverse correlation between the level of CHCHD2 mRNA and the mRNA levels of 5 of the 13 mitochondrially encoded protein genes (MT-ND1, MT-ND2, MT- ND4, MT-ND4L and MT-CYB), however further assessment is required to confirm that there is also an inverse correlation between CHCHD2 transcript and mtDNA levels in the brain.

CHCHD2 was also observed to respond to a number of other PD-associated mitochondrial impairments. In particular, this analysis revealed that the most substantial increase in CHCHD2 protein levels occurred in response to deletion of the PD-associated genes PINK1 and PARKIN in drosophila, which resulted in a substantial 6X and 9X increase in CHCHD2 protein, respectively (Meng et al., 2017) (Table 4.2). These proteins play a central role in mitophagy by initiating autophagosome formation around the damaged mitochondrion following loss of mitochondrial membrane potential, as discussed in section 1.11.1 (Clark et al., 2006; Zhang et al., 2017). Therefore the observation that deletion of these proteins results in a substantial

153

4. Post-transcriptional regulation of PD-associated protein CHCHD2 increase in CHCHD2 protein aligns with the proposal that CHCHD2 accumulation may play an early role in mitophagy, functioning upstream of PINK1 and PARKIN to restore OXPHOS capacity and/or quarantine potentially toxic Cyt c within the mitochondrion (Figure 4.13).

CHCHD2 was also observed to increase in response to deletion of the PD-associated gene DJ1 (2.5X). DJ1 deletion causes increased oxidative stress (Puschmann, 2013) and reduces mitochondrial membrane potential (Giaime et al., 2012; Toyoda et al., 2014). This result therefore further suggests that up-regulation of CHCHD2 protein levels may be a response to reduced mitochondrial membrane potential. Unfortunately, the impact of PINK1, PARKIN or DJ1 deletion upon CHCHD2 mRNA was not measured and as a result it was not possible to determine whether this increase in CHCHD2 levels was a result of transcriptional up-regulation, or alternatively caused by a post- transcriptional response, similar to that observed in response to the mitochondrial impairments assessed in this thesis (Figure 4.5).

Table 4.2 suggests that CHCHD2 is regulated at both the transcriptional and translational levels and identifies changes in mtDNA copy number as a potential contributor to the reduced CHCHD2 mRNA levels observed in PD patients. Determining the mechanisms that underlie these transcriptional and translational responses will significantly increase the understanding of how CHCHD2 dysfunction contributes to PD and provide insight into factors that should be considered when assessing potential strategies for restoring CHCHD2 function in PD.

4.3.3 A proposed model for localised post-transcriptional regulation of CHCHD2 in response to decreased mitochondrial membrane potential

The observations that mitochondrial toxins cause an increase in CHCHD2 protein, but not mRNA levels (Figure 4.5) and that CHCHD2 preferentially accumulates in

154

4. Post-transcriptional regulation of PD-associated protein CHCHD2 mitochondria with a lower degree of polarisation (Figure 4.9) suggests that CHCHD2 can be post-transcriptionally up-regulated in a localisation specific manner. Post- transcriptional regulation of CHCHD2 is also evident in response to its own overexpression, where protein levels are substantially increased (33X increase in protein) compared to mRNA levels (2.1X increase in mRNA levels) (Figure 4.10). Similar evidence of self-amplifying up-regulation of CHCHD2 protein levels was observed by Aras et al. (2015), who showed that overexpression of Tg Flag tagged CHCHD2 caused a significant and substantial increase in the level of the endogenous protein, as differentiated based on protein size, but had minimal effect on activating CHCHD2 transcription (Aras et al., 2015). Also aligning with the results of this thesis (Figure 4.11), overexpression of the Flag tagged version of CHCHD2 did not cause mitochondrial dysfunction, indicating that this response is not caused by the transgenic version of the protein inducing mitochondrial stress (Aras et al., 2015).

There are a number of molecular mechanisms that may contribute to the increase in CHCHD2 protein abundance that is observed upon mitochondrial stress or CHCHD2 overexpression. The most likely mechanisms are reduced protein turnover and increased translation and there are multiple examples of both of these mechanisms regulating the abundance and localisation of proteins within the cell. For example, reduced protein turnover underlies the rapid and localised accumulation of PINK1 protein on the surface of depolarised mitochondria, which occurs in the initial “damage recognition” stage of mitophagy (Figure 1.7), as discussed in section1.11.1 (Jin & Youle, 2012). The protein is continually imported into healthy mitochondria and then degraded (Jin et al., 2010), resulting in low basal levels of the protein. However, when mitochondria are damaged, loss of membrane potential inhibits this import and subsequent degradation process, resulting in PINK1 accumulation on the surface of depolarised mitochondria where it recruits other components for mitophagy initiation (Jin et al., 2010). However, although this membrane potential-dependent accumulation of PINK1 is reminiscent of the accumulation of CHCHD2 in mitochondria with a lower degree of polarisation, it is unlikely that these responses are regulated by

155

4. Post-transcriptional regulation of PD-associated protein CHCHD2 the same mechanism as the mitochondrial import of these two proteins is fundamentally different. While PINK1 is imported through the membrane potential- dependent TOM20 channel (Jin et al., 2010) and thus would accumulate on the surface in response to depolarisation, CHCHD2 is instead imported through the MIA/IMS import pathway (Mordas & Tokatlidis, 2015; Wasilewski et al., 2017), which is not dependent on mitochondrial membrane potential (Stojanovski et al., 2008). This import of CHCHD2 through a channel that is not membrane potential dependent aligns with the proposal that CHCHD2 accumulates within depolarised mitochondria, rather than on the surface, in order to aid recovery of membrane potential and/or to sequester potentially damaging Cyt c (Figure 4.13). In summary, the non-membrane potential-dependent uptake of CHCHD2 into mitochondria indicates that the protein is not regulated through a PINK1-like import/degradation mechanism.

An alternative regulatory mechanism that may be responsible for the localised post- transcriptional regulation of CHCHD2 involves PUmilo and FBF (PUF) protein mediated translational regulation. PUF proteins are RNA binding proteins that can directly interact with the 3’ UTR of target mRNAs to reduce translation by inhibiting translation initiation or destabilising the mRNA (Blewett & Goldstrohm, 2012; Kong & Lasko, 2012; Quenault et al., 2011). PUF protein interaction is known to underlie many rapid response stress pathways. For instance, the mRNA encoding the nitric oxide stress response factor YHB1 is released from PUF protein translational repression upon elevation of nitric oxide levels, facilitating rapid induction of the nitric oxide stress response (Russo & Olivas, 2015). Additionally, PUF proteins can also regulate localised translation of mRNAs at specific subcellular regions, which has a number of advantages. For example, in neuronal cells localised translation at the axonal terminal eliminates the need to transport proteins across large axonal distances, reducing the time and energy cost associated with protein transport (Cooke et al., 2011; Donnelly et al., 2010). Localised translation is also commonly observed in the mitochondrial vicinity, where localisation of mRNA in close proximity to mitochondria enables translation to be coupled to mitochondrial import, minimising the likelihood of protein

156

4. Post-transcriptional regulation of PD-associated protein CHCHD2 misfolding (Fox, 2012). Therefore, PUF proteins mediated translational regulation facilitates both rapid and localised translational responses that may underlie the observed rapid increase in CHCHD2 protein levels in response to mitochondrial impairment and the accumulation of the protein in mitochondria with reduced membrane potential.

There is substantial evidence supporting the proposal that CHCHD2 mRNA associates with PUF proteins and that this interaction may underlie the rapid and localisation specific increase in CHCHD2 protein observed in response to mitochondrial depolarisation (Figure 4.5; Figure 4.9). Firstly, microarray data was obtained from PUF protein RIP-Chip analysis and assessment of this data identified that the mRNAs encoding the CHCHD2 homologues in C. elegans (glod-4) and S. cerevisiae (MIC17) interact with PUF proteins (Freeberg et al., 2013; Kershner & Kimble, 2010). Secondly, MIC17 mRNA was identified to interact with Puf3p (Freeberg et al., 2013), a protein known to localise its target mRNAs to the mitochondrial surface (Quenault et al., 2011). It is likely that a similar regulatory mechanism underlies the localised translation of CHCHD2 at the mitochondrial surface in mammalian cells as there is evidence that the targets of PUF proteins are highly conserved across species (Hogan et al., 2015; Kershner & Kimble, 2010). However, neither of the 2 PUF proteins that have been identified in mammalian cells, PUM1 and PUM2, localise to the mitochondria. This suggests that while an association between CHCHD2 mRNA and PUM1 or PUM2 could regulate CHCHD2 translation, an additional pathway may have evolved to mediate localised translation in the vicinity of less polarised mitochondria. One recently identified mechanism that could facilitate mitochondrially localised translation in mammalian cells involves the localisation of specific mRNAs to the mitochondrial surface through interaction with the mitophagy regulating proteins PINK1 an PARKIN (Gehrke et al., 2015). The accumulation of PINK1 and PARKIN on the surface of depolarised mitochondria (Jin et al., 2010) would provide an ideally positioned stress response mechanism to alleviate the proposed translational repression of CHCHD2 in the vicinity of mitochondria with lowered membrane polarisation.

157

4. Post-transcriptional regulation of PD-associated protein CHCHD2

PINK1 and PARKIN have been demonstrated to interact with nuclear encoded mRNAs that encode mitochondrial proteins and are normally bound in a translationally repressed state within the cytosol to RNA binding proteins, including PUM1 (Gehrke et al., 2015). Association of these mRNAs with PINK1 at the mitochondrial surface results in PINK1-mediated phosphorylation of the RNA binding proteins, removing this translational block (Gehrke et al., 2015). PINK1/PARKIN association also increases association of these mRNAs with the translation initiation complex, thereby facilitating activation of translation of these mRNAs (Gehrke et al., 2015). Accordingly, if CHCHD2 translation were regulated through a PINK1/PARKIN-dependent mechanism, under normal conditions CHCHD2 mRNA would be bound in a translationally repressed state in the cytosol, likely through interaction with an as yet unidentified PUF-like RNA binding protein (Figure 4.14A). This may enable a low basal rate of translation to account for the low level of CHCHD2 protein present in non-stressed conditions. Upon loss of mitochondrial membrane potential CHCHD2 mRNA would associate with the PINK1 that also accumulates on the mitochondrial surface in response to reduced membrane potential. This would result in accumulation of CHCHD2 mRNA and activation of CHCHD2 translation specifically in the vicinity of less polarised mitochondria (Figure 4.14B), as is observed in this study (Figure 4.9). Alternatively, in the case of CHCHD2 protein levels increasing in response to its own overexpression (Figure 4.10), this self-amplifying capacity may be associated with the mechanism of translational repression. If CHCHD2 mRNA were to increase even sightly beyond the capacity of the available PUF-like repressor protein, excess mRNA would be available for translation. Released from translational repression, these free CHCHD2 mRNAs could be rapidly translated, subsequently resulting in the observed increases in endogenous and transgenic protein levels (Figure 4.14C).

158

4. Post-transcriptional regulation of PD-associated protein CHCHD2

Figure 4.14 Model of CHCHD2 translational regulation. CHCHD2 mRNA is translationally repressed in the cytosol through association with an mRNA binding protein (RBP), such as PUM1. Upon mitochondrial depolarisation, PINK1 accumulates on the mitochondrial surface. Association of the repressed mRNA with PINK1 results in phosphorylation of the RBP, releasing the translational block and facilitating CHCHD2 translation. Alternatively, in response to CHCHD2 overexpression, mRNA levels in excess of the available RBP mRNA binding capacity result in rapid protein translation from non-bound mRNAs and the observed increase in endogenous and transgenic protein.

This proposed model of CHCHD2 translational regulation presents an intriguing aspect of mitochondrial biology that integrates multiple observations made through this study and by others. Furthermore, emerging evidence suggest that translational dysregulation may contribute to PD (reviewed in Martin et al., 2016), highlighting this aspect of CHCHD2 regulation as a key area for further investigation. Additional

159

4. Post-transcriptional regulation of PD-associated protein CHCHD2 experimentation is ongoing towards confirming this model by demonstrating that CHCHD2 mRNA localises to the mitochondria in a membrane potential dependent manner and that interaction with PINK1/PARKIN may underlie this response.

4.4 Summary and future directions.

The results of this chapter indicate that CHCHD2 mRNA expression is reduced in multiple brain regions of sporadic PD patients, suggesting that CHCHD2 dysfunction is likely to play a role in sporadic PD. Additional assessment is ongoing towards determining whether CHCHD2 protein levels are similarly perturbed in patient brain tissue.

The results of this chapter additionally build on the model of CHCHD2 function recently proposed by Huattori and colleagues by suggesting that CHCHD2 is post- transcriptionally up-regulated and accumulates in damaged mitochondria as a stress response that is likely to be triggered by decreased mitochondrial membrane potential. This accumulation of CHCHD2 in less polarised mitochondria is likely to facilitate the recovery of mitochondrial membrane potential and/or the retention of potentially toxic Cyt c within mitochondria during mitophagy (Figure 4.13).

Post-transcriptional up-regulation is a key feature of the rapid and localised CHCHD2 response to mitochondrial impairment. Assessment of post-transcriptional regulation of CHCHD2 homologues in other organisms indicated that CHCHD2 is translationally regulated in C. elegans and S. cerivisea through association of the mRNA transcript with PUF RNA binding proteins, which repress translation of the mRNA. Furthermore, in mammalian cells PINK1 and PARKIN have recently been identified to facilitate translational activation by removing PUF protein-mediated suppression. Together these discoveries support a model of CHCHD2 regulation in mammalian cells wherein

160

4. Post-transcriptional regulation of PD-associated protein CHCHD2 translationally active CHCHD2 mRNA is localised to the vicinity of less polarised mitochondria through association with PINK1/PARKIN (Figure 4.14). Experiments are ongoing that are aimed at visualising this proposed membrane potential dependent translocation of CHCHD2 mRNA using fluorescent in situ hybridisation (FISH) and assessing whether this response is impaired by the absence of PINK1. Additionally, the effects of truncations in the 3’ and 5’ UTR regions of CHCHD2 will also be assessed. These truncations would be predicted to impair association of the mRNA with regulatory RBPs, thereby inhibiting translational repression of the mRNA and resulting in substantial increase in CHCHD2 protein levels even in non-mitochondrially perturbed conditions.

This ongoing project presents a novel mechanism through which a rapid and localisation specific mitochondrial stress response involving CHCHD2 may be regulated. Furthermore, the additional insight into the function and regulation of CHCHD2 that this study has provided has significantly improve the understanding of how CHCHD2 dysfunction could contribute to PD, suggesting that CHCHD2 dysfunction could impair mitochondrial recovery in response to moderate levels of mitochondrial impairment and/or impair the isolation of Cyt c within the mitochondrion during mitophagy (Figure 4.13).

161

5. The Vacuolar ATPase and PD

5 THE VACUOLAR ATPase AND PD

5.1 Introduction

The Vacuolar ATPase (V-ATPase) is a large proton-translocating, multi-subunit super- complex that regulates the acidity of the compartments that make up the endolysosomal system (Maxson & Grinstein, 2014). These dynamic and interconnected cellular compartments include early endosomes, late endosomes and lysosomes. The endolysosomal system is additionally intimately connected to the trans-Golgi network (TGN), which provides hydrolases, membrane and receptors in a bi-directional manner to each of these compartments (Figure 5.1) (Bonifacino & Rojas, 2006; Huotari & Helenius, 2011). The many functions of the endolysosomal system and the roles that the V-ATPase plays in regulating these functions are outlined below.

Figure 5.1 The endolysosomal network is a dynamic and interrelated network of early endosomes, late endosomes and lysosomes and the V-ATPase regulates the acidity of these compartments. Incoming extracellular cargo enters the network through endocytosis, while cytoplasmic cargo is taken up by chaperone mediated autophagy (CMA), micro-autophagy or macro-autophagy. Enzymes and proteins are supplied to these compartments though exchanges with the trans-Golgi network.

162 5. The Vacuolar ATPase and PD

5.1.1 Structure of the Vacuolar ATPase

The 13 subunits of the V-ATPase assemble into either a membrane bound V0 domain or a cytosolic V1 domain, which resemble and are evolutionarily related to the F1 and F0 domains of mitochondrial ATP synthase (Futai et al., 2012). However, in the reverse fashion to the mitochondrial ATP synthase, the V-ATPase complex uses energy from the hydrolysis of ATP to drive protons against an electrochemical gradient into the lysosomal lumen (Figure 5.2A). The V1 domain mediates hydrolysis of ATP, which drives rotation of the V0 domain and thus mediates the transport of protons into the lumen of endosomes and the lysosome (Maxson & Grinstein, 2014).

Figure 5.2 The V-ATPase complex is comprised of a V0 membrane and V1 cytosolic domain. A) The V-ATPase complex is localized to the membranes of endolysosomal compartments and uses energy from ATP hydrolysis to translocate protons into the compartment’s lumen. B) 13 protein subunits assemble to form the 2 V-ATPase domains. Figure adapted from MazhabJafari et al. (2016).

By convention the subunits of the V-ATPase are named according to their domain. The 8 subunits of the V1 cytosolic domain are denoted in capital letters (A-H), while the 6 subunits of the V0 domain are denoted in lower case italicised letters (a, c, c’’, d, e) (Mazhab-Jafari et al., 2016; Zhao et al., 2015) (Figure 5.2B; Table 5.1). Additional assembly chaperones have also been identified, including ATP6AP1 and ATP6AP2,

163 5. The Vacuolar ATPase and PD which appear to play a role in assembly of the superstructure (Jansen & Martens, 2012).

Table 5.1 Overview of the 13 V-ATPase genes, the protein subunits that they encode and the number of gene paralogues they encode in humans.

Gene Paralogues Gene Protein subunit Domain in humans ATP6V0A a V0 4 ATP6V0B c’’ V0 0 ATP6V0C c V0 0 ATP6V0D d V0 2 ATP6V0E e V0 2 ATP6V1A A V1 0 ATP6V1B B V1 2 ATP6V1C C V1 2 ATP6V1D D V1 0 ATP6V1E E V1 2 ATP6V1F F V1 0 ATP6V1G G V1 3 ATP6V1H H V1 0

Several subunits of both the V0 (a, d, e) and V1 domains (B, C, E and G) exist in the human genome as multiple gene paralogues that are expressed in a cell type specific manner (Smith et al., 2003). For example, the ATP6V1B2 paralogue is primarily expressed in a brain specific manner. Additionally, for many V-ATPase subunits multiple splicing variants exist, such that to date multiple splicing isoforms have been identified for ATP6V0A, ATP6V0D, ATP6V0E, ATP6V1C, ATP6V1G and ATP6V1H, which encode the a, d, e, C, G and H subunits respectively (Miranda et al., 2010). The presence of these multiple gene paralogues and/or splicing isoforms encoding individual V-ATPase subunits is indicative of the complexity of V-ATPase function and regulation.

164 5. The Vacuolar ATPase and PD

5.1.2 V-ATPase dependent functions of the endolysosomal system

The endolysosomal system is a network of compartments that can be difficult to categorise into distinct organelles. Early endosomes mature to form late endosomes and compartments in this maturation pathway can fuse with vesicles from the TGN to form lysosomes (Huotari & Helenius, 2011). These inter-related compartments can be roughly differentiated based on the specific subset of membrane proteins and luminal enzymes that they contain (Hu et al., 2015), as well as their distinct luminal acidities (Johnson et al., 2016; Korolchuk et al., 2011). The V-ATPase is present on each of these compartments and acts to establish a pH gradient whereby compartments become more acidified as they mature from early to late endosomes and lysosomes (Huotari & Helenius, 2011). The role the V-ATPase plays in sensing, regulating and maintaining the different pH of each endolysosomal compartment is essential to maintaining the many functions of this system. These functions and the consequences of failing to maintain the pH environment within the endolysosomal network are summarised in Figure 5.3 and discussed in detail below.

Figure 5.3 The V- ATPase affects many aspects of cell biology through its role in acidifying endolysosomal compartments, sensing of nutrient status and mediating membrane fusion events.

165 5. The Vacuolar ATPase and PD

5.1.2.1 Lysosomal degradation and recycling of cellular components

The lysosome is the primary site of protein, lipid and nucleic acid degradation within the cell and enables unwanted cellular substrates, such as misfolded proteins, or entire organelles, such as damaged mitochondria, to be catabolised and their constituent components (eg: amino acids) to be recycled to the cytosol for future use. As discussed in section 1.8.2, these substrates can be delivered into the lysosome through macroautophagy, microautophagy or chaperone mediated autophagy.

Degradation of substrates within the lysosome is mediated by a set of lysosomal proteases and lipases that are optimally active at an acid pH, similar to that of the lysosome (pH 4.6-5) (Mellman et al., 1986). Therefore V-ATPase-mediated establishment and maintenance of lysosomal acidity is essential for creating the optimal pH environment required by these enzymes, thus coupling lysosomal acidification to lysosomal degradation capacity (Coffey & De Duve, 1968; Fowler & De Duve, 1969). If V-ATPase-mediated acidification of the lysosome is impaired the activity of lysosomal enzymes is reduced (Baltazar et al., 2012; Wolfe et al., 2013; Yoshimori et al., 1991), leading to reduced lysosomal degradation capacity and accumulation of non-degraded substrates within the cell (Figure 5.4).

Figure 5.4 The V-ATPase regulates lysosomal hydrolase activity by regulating the pH of the lysosomal lumen. Lysosomal hydrolases are optimally active in acidic environments. Reduced V-ATPase activity impairs the acidification of the lysosome, leading to reduced activity of lysosomal hydrolases and accumulation of non-degraded substrates.

166 5. The Vacuolar ATPase and PD

5.1.2.2 Delivery of cargo from the trans-Golgi network the endolysosomal network

The enzymes that mediate lysosomal degradation are synthesised in the endoplasmic reticulum and exported to the TGN for further sorting (Kornfeld, 1987). In the TGN enzymes destined for delivery to endolysosomal compartments are bound to the mannose-6-phosphate receptor (M6PR) (Griffiths et al., 1988) and sorted into vesicles that bud from the TGN and fuse with maturing endolysosomal compartments. Upon delivery of M6PR bound enzymes to the compartments the acidity of the endolysosomal compartment lumen, which is established by the V-ATPase, facilitates the dissociation of the enzyme from M6PR, allowing the receptor to be recycled back to the TGN (Olson et al., 2008)(Figure 5.5). Therefore, failure to acidify the lysosomal compartment impairs the delivery of lysosomal hydrolase(s), resulting in a subsequent loss of lysosomal degradation capacity as well as impaired recycling of the M6PR back to the TGN, which further exacerbates the impaired enzyme delivery.

Figure 5.5 The delivery of cargo from the trans-Golgi Network to endolysosomal compartments is dependent on the luminal acidity of the compartment. Lysosomal enzymes that mediate degradation are delivered to maturing endolysosomal compartments from the trans-Golgi network. These enzymes are bound to mannose-6-phosphate receptors and the acidic environment of the endolysosomal lumen is required to facilitate dissociation of the

167 5. The Vacuolar ATPase and PD enzyme from the receptor. If luminal acidification is impaired the cargo cannot dissociate and the receptor cannot be recycled back to the TGN.

5.1.2.3 Receptor mediated endocytosis and recycling of plasma membrane receptors

Receptor-mediated endocytosis, also referred to as clathrin-mediate endocytosis, is a process through which the cell absorbs metabolites, nutrients and hormones through their recognition by plasma membrane surface receptors. For example, extracellular cholesterol, in the form of LDL particles, is imported through interaction with the LDL receptor (Brown & Goldstein, 1986; Pietiäinen et al., 2013). Similarly, iron in a complex with transferrin is imported through interaction with the transferrin receptor (Dautry- Varsat et al., 1983). The dissociation of cargo from these receptors is reliant on a pH dependent mechanism similar to that used for the delivery of enzymes from the TGN to the lysosome (Maxfield & McGraw, 2004). Endocytosed vesicles containing the cargo-bound receptors fuse with specialised endosomal compartments known as recycling endosomes. The acidity of the recycling endosome lumen allows the receptor to dissociate from its cargo and be recycled back to the cell surface for binding further ligand (Goldenring, 2015)(Figure 5.6). Failure to establish endosomal acidity prevents this dissociation, resulting in depletion of cell surface receptors and associated impaired cargo delivery.

168 5. The Vacuolar ATPase and PD

Figure 5.6 Uptake of extracellular cargo and recycling of plasma membrane receptors is dependent of the luminal acidity of recycling endosome compartments. Receptor mediated endocytosis allows extracellular nutrients, hormones or metabolites to be absorbed by the cell. Vesicles containing receptor bound cargo are endocytosed and fuse with recycling endosomes. The acidity of the endosomal compartment is required for the receptor to dissociate from the cargo and be recycled back to the cell surface.

5.1.2.4 Storage of ions and amino acids

Endolysosomal compartments are recognised as important storage sites for metal ions, including Ca2+ (Lloyd-Evans et al., 2010; Michelangeli et al., 2005). The endolysosomal network contains concentrations of Ca2+ 3 to 4 orders or magnitude greater that that of the cytosol (Christensen et al., 2002) and, together with the ER and mitochondria, plays a central role in regulating Ca2+ levels within the cell (Lloyd-Evans & Platt, 2011). Ca2+ entry into endolysosomal compartments is regulated through two pathways; ATP dependent Ca2+-ATPases and Ca2+/H+ exchangers, which load Ca2+ into the compartment lumen in exchange for protons (Pittman, 2011) (Figure 5.7). Through this second pathway Ca2+ uptake is coupled to V-ATPase-mediated establishment of the endolysosomal proton gradient (Figure 5.7). Consequently, inhibition of the V- ATPase using the specific inhibitor Bafilomycin or disruption of the lysosome pH gradient using proton ionophores results in depletion of lysosomal Ca2+ stores (Christensen et al., 2002), exemplifying the indirect role of the V-ATPase in regulating endolysosomal Ca2+ uptake.

169 5. The Vacuolar ATPase and PD

Figure 5.7 Uptake of Ca2+ ions into endolysosomal compartments is mediated by Ca2+/H+ exchange channels or Ca2+ ATPase channels. The Ca2+ ATPase channel used energy from ATP hydrolysis to drive Ca2+ into the lumen. The Ca2+/H+ exchange channel loads Ca2+ ions into the lumen in exchange for H+ ions that are provided by the pumping capacity of the V- ATPase.

Impaired endolysosomal Ca2+ storage is implicated as a pathogenic mechanism in AD, a neurodegenerative condition that has symptomatic and pathological overlaps with PD. Abnormal release of lysosomal Ca2+ stores has been suggested to contribute to neuronal dysfunction in both sporadic and familial AD (McBrayer & Nixon, 2013), suggesting that therapeutic strategies which target endolysosomal Ca2+ regulation may be effective treatments for many forms of AD. Impaired Ca2+ homeostasis has also been implicated as a likely pathogenic mechanism in PD (Surmeier, 2007), however the potential contribution of perturbed endolysosomal Ca2+ storage to this dysfunction remains unknown. Therefore further understanding the mechanisms that regulate endolysosomal Ca2+ uptake and the effects that Ca2+ dyshomeostasis may have on other Ca2+ storage compartments in the cell would significantly enhance the current understanding of this aspect of both AD and PD.

5.1.2.5 Synaptic vesicle loading

The V-ATPase also has a number of cell specific functions, including indirectly regulating the process of synaptic vesicle loading in neuronal cells. Neurotransmitters, such as dopamine (DA), are synthesized in the cytosol from where they are loaded into synaptic vesicles via the membrane vesicular monoamine transporter (VMAT2) in

170 5. The Vacuolar ATPase and PD exchange for 2 protons from within the SV lumen (Figure 5.8)(Blakely & Edwards, 2012). These protons are provided by V-ATPase proton pumping activity (Gasnier, 2000). Consequently, if V-ATPase proton translocation is reduced the loading of synaptic vesicles is impaired, diminishing neurotransmission and leading to an accumulation of neurotransmitters within the cytosol. In the case of dopamine, cytosolic accumulation leads to oxidation of the neurotransmitter, resulting in the formation of damaging quinones that react with cysteine residues on proteins within the cytosol (Segura-Aguilar et al., 2014).

Figure 5.8 The V-ATPase indirectly regulates synaptic vesicle loading. Neurotransmitters, such as dopamine, are loaded into synaptic vesicles through the vesicular monoamine transport (VMAT) channels in exchange for protons that are provided by V-ATPase proton pump.

5.1.2.6 Membrane fusion

The V-ATPase has also been described to have a number of functions that are independent of its proton pumping capacity. For instance, the V-ATPase V0 domain may play a role in facilitating membrane fusion (Bayer et al., 2003), a process that enables the delivery of endocytic and TGN-derived vesicles to endolysosomal compartments and is also required to mediate the formation of autolysosomes during macroautophagy (Huotari & Helenius, 2011). Fusion of membranes is initiated by

171 5. The Vacuolar ATPase and PD trans-SNARE (soluble N-ethylmaleimide-sensitive factor attachment protein receptors) pairing (Bayer et al., 2003), which brings the two membranes into close proximity. The V-ATPase V0 domain has been suggested to facilitate membrane fusion subsequently to this pairing, possibly by forming a pore between the two compartments (Peters et al., 2001)(Figure 5.9). This suggests that V-ATPase depletion may inhibit membrane fusion, however the precise mechanism through which this process is regulated remains controversial, with conflicting results as to whether this mechanisms acts independently of V-ATPase proton translocating capacity or not (Bayer et al., 2003; Mauvezin et al., 2015).

Figure 5.9 The V-ATPase V0 domain plays a role in facilitating membrane fusion. Fusion of membranes is initiated by trans-SNARE pairing, which draws membranes into close proximity. This enables V0 domains on adjacent membranes to form a pore, which is thought to facilitate fusion.

Through a similar mechanism the V-ATPase V0 domain also appears to play a role in regulating the fusion of synaptic vesicles with the plasma membrane at the synaptic terminal in neuronal cells (Vavassori et al., 2014). The V-ATPase V0 domain present on synaptic vesicles directly interacts with synaptobrevin, a SNARE protein present on the presynaptic membrane, to enable fusion of these membranes and facilitate neurotransmitter release into the synapse (Giovanni et al., 2010).

5.1.2.7 Amino acid sensing and regulation of endolysosomal signalling

The V-ATPase has emerged as an essential regulator of cellular nutrient status through interaction with mTORC1 (Zhang et al., 2014a). mTORC1 is considered to be a master

172 5. The Vacuolar ATPase and PD regulator of cell growth and metabolism (Laplante & Sabatini, 2009) that interacts with the V-ATPase and a subset of other lysosomal proteins to form a complex known as the lysosomal nutrient sensing (LYNUS) machinery, which senses and responds to changes in cell nutrient status ( Zhang et al., 2014a). Under favourable, nutrient rich conditions lysosomal amino acid reserves are high and mTORC1 is tethered to the lysosomal surface through the Ragulator complex and Rag proteins. This causes activation of mTORC1, which activates biosynthetic growth processes while blocking autophagy (Zhang et al., 2014a). Alternatively, in unfavourable or starvation conditions lysosomal amino acid reserves become depleted. This causes a conformational change in the V-ATPase, which induces an interaction between the V-ATPase and Regulator that results in the release of mTORC1 from the lysosomal surface (Stransky & Forgac, 2015; Zoncu et al., 2011). Non-lysosomal associated mTORC1 is subsequently inactivated, removing the autophagy block and enabling autophagy to proceed in an attempt to replenish lysosomal amino acid stores (Jewell et al., 2013; Zoncu et al., 2011)(Figure 5.10).

5.1.2.1 Cytoskeleton mediated transport of endolysosomal compartments

The V-ATPase may also play an important role in regulating the transport of endolysosomal compartments within the cell by interacting with both actin microfilaments and microtubules, which facilitate short and long-range movement within the cell, respectively. For instance, the transport of endocytosed vesicles is mediated by actin filaments, which drive these vesicles off the plasma membrane and into the cytosol (Lamaze et al., 1997). Alternatively, long range movements, such as the movement of maturing endosomal compartments towards the perinuclear region (Figure 5.1) is mediated by microtubule tracks (Granger et al., 2014), however actin appears to continue to play a role in regulating the directionality of this movement (Cordonnier et al., 2001).

173 5. The Vacuolar ATPase and PD

Figure 5.10 The V-ATPase signals lysosomal nutrient status through sensing amino acid levels. Reduced lysosomal amino acid levels result in a change in V-ATPase conformation, which induces association between the V-ATPase and regulator. This releases mTORC1 from the lysosomal surface, resulting in mTOC1 inactivation, which switches the cell from a catabolic to anabolic state.

The interaction between endosomal compartments and actin filaments appears to be mediated by interaction between actin and the V-ATPase V1 subunit B (Holliday et al., 2000) or C (Vitavska et al., 2005). The V-ATPase V1 domain been shown to interact with actin in a number of organisms including humans (Holliday et al., 2000), and this interaction appears to promote f-actin stablisation in Arabidopsis (Ma et al., 2012), suggesting that the V-ATPase may play a role in controlling the dynamics of the actin cytoskeleton.

The V-ATPase has also been demonstrated to interact directly with microtubules through the V1C subunit (Tabke et al., 2014) and indirectly through Rab interacting

174 5. The Vacuolar ATPase and PD lysosomal protein (RILP). RILP binds to the V1G subunit (De Luca et al., 2014) and recruits dynein-dynactin motors to the membrane (Jordens et al., 2001). This facilitates the minus end directed movement of endosomal compartments along microtubule tracks (Jordens et al., 2001) (Figure 5.11). Failure to transport and correctly position endolysosomal compartments has a number of consequences. For instance, impaired delivery of lysosomes to the perinuclear region of the cell, which is a primary site of autophagosome accumulation, results in impaired macroautophagy (Pu et al., 2016) (Yang & Klionsky, 2010). Impaired lysosomal transport may also impairing the delivery of enzymes from the TGN, with peripheral lysosomes appearing to be supplied with lower levels of enzymes than their perinuclear counterparts (Gowrishankar et al., 2015; Johnson et al., 2016). Therefore impaired lysosomal transport can reduce lysosomal degradation capacity both though impairing lysosome-autophagosome fusion and by reducing the efficiency with which the necessary degradation enzymes are delivered from the TGN.

Figure 5.11 The V-ATPase interacts directly and indirectly with microtubules. The ATP6V1C may bind directly to microtubules. Alternatively the V-ATPase may also interact indirectly with microtubules. The ATP6V1G subunit of the V-ATPase binds to Rab Interacting Lysosomal protein (RILP), which recruits dynactin motors to facilitate endosome transport along microtubules. Mature endosomes become concentrated at the perinuclear region of the cell, bringing them into close proximity with the autophagosomes that are to be degraded.

175 5. The Vacuolar ATPase and PD

5.1.3 Regulation of the V-ATPase

V-ATPase proton pumping activity is controlled at a number of levels that include regulation of activity and/or assembly of the complex as well as transcriptional and translational regulation. The multitude of regulatory factors that influence V-ATPase function is indicative of the central and essential role of the V-ATPase in regulating many cellular processes.

5.1.3.1 Acidification of endolysosomal compartments is regulated by the density of V-ATPase subunits on the membrane

The density of V-ATPase complexes on the membrane contributes to the range of endolysosomal compartment acidities that are apparent within the cell. Early endosomes mature to late endosomes (Huotari & Helenius, 2011). Further maturation, coupled to the delivery of lysosomal enzymes from the TGN leads to the formation of mature lysosomes (Scott et al., 2014). These lysosomes fuse with autophagosomes or late endosomes to produce autolysosomes or endolysosomes with full hydrolytic potential capable of degrading substrates. This maturation process is associated with increasing acidification (early endosome  late endosome  lysosome), which is mediated by the V-ATPase. To achieve this increasing acidification, late endosomes accumulate higher density of V-ATPase’s than early endosomes (Lafourcade et al., 2008). There are a number of mechanisms that may underlie this differential accumulation of V-ATPase complex. For instance, targeting of the V-ATPase to specific endosome compartments may enable the complex to accumulate to higher levels on the surface of late endosomes and lysosomes relative to early endosomes. Differential splicing is suggested as a mechanism through which the V-ATPase could be targeted to specific endosomal compartments, with ATP6V0A splice variants having been shown to preferentially target the complex to the late endosome (Hernando et al., 1999; Poëa- Guyon et al., 2006). Alternatively, gradual accumulation of V-ATPase complexes on late endosomes may also occur through regulated fission of early endosome membrane regions into two separate compartments, with one compartment containing a higher

176 5. The Vacuolar ATPase and PD proportion of V-ATPase complexes than the other (Huotari & Helenius, 2011). The mechanisms through which V-ATPase enriched membrane regions are recognised and isolated remain unknown. Due to the many consequences of perturbed lysosomal pH, better understanding the mechanisms through which the V-ATPase accumulates on specific compartment surfaces may provide significant insight into how this process may become dysregulated in disease situations.

5.1.3.2 V-ATPase activity is regulated by reversible dissociation of the V0 and V1 domains

The proton pumping capacity of the V0 domain of the V-ATPase is coupled to ATP hydrolysis, which is mediated by the V1 domain. Reversible dissociation the V0 and V1 domains therefore enables rapid modulation of V-ATPase proton pumping activity (Kane, 1995; Sumner et al., 1995; Xu & Forgac, 2001). These domains reversibly dissociate in response to certain stresses, including glucose deprivation (Xu & Forgac, 2001; Sumner et al., 1995), serotonin withdrawal (Dames et al., 2006) and salt stress (Silva & Gerós, 2009), suggesting that this is a mechanism through which cells regulate V-ATPase activity in response to a variety of signals.

The reversible dissociation of the V-ATPase domains was first characterised as a response to glucose withdrawal (Sumner et al., 1995). Glucose deprivation limits glycolysis and OXPHOS and therefore impairs ATP production (Parra & Kane, 1998). As the V-ATPase is a major consumer of ATP, the dissociation of it’s two domains enables the complex to be rapidly inactivated when ATP levels are low (Kane, 1995; Xu & Forgac, 2001). The mechanism through which low glucose levels trigger V-ATPase dissociation is unknown, however evidence suggests that the V1C subunit plays a role. The V1C subunit appears to mediate bridging of the V0 and V1 domains by binding to the E and G subunits of the V1 domain and the a subunit of the V0 domain (Inoue & Forgac, 2005). Reduced ATP availability causes a structural change in the V1C subunit (Armbrüster et al., 2005), which impairs this bridging capacity and leads to dissociation of the V0 and V1 domains (Beyenbach & Wieczorek, 2006), although in vivo only the

177 5. The Vacuolar ATPase and PD

V1C subunit appears to fully dissociate from the complex while the remaining V1 subunits remain attached (Tabke et al., 2014).

In addition to the role of ATP availability in regulating V-ATPase domain association, particular glycolytic pathway enzymes also appear to regulate V-ATPase assembly and disassembly. For instance, the glycolytic enzyme aldolase is required for V-ATPase domain re-assembly following glucose restoration (Lu et al., 2007) by bridging the supercomplex domains between the B and E subunits on the V1 domain and the a subunit on the V0 domain (Lu et al., 2007; 2004). Similarly, interaction between the V0 domain and the glycolytic enzyme phosphofructokinase have also been reported (Su et al., 2003). This suggests that V-ATPase activity may be connected to glycolysis and cellular energy metabolism at multiple points.

A number other cytosolic proteins have also been suggested to regulate association/dissociation of the V-ATPase subunits, including the Regulator of the H+- ATPase of Vacuolar and Endosomal (RAVE) complex (Seol et al., 2001), RILP (De Luca et al., 2014), Protein Kinase A (PKA) (Voss et al., 2007) and ATP6AP2 (Jansen & Martens, 2012). Of particular interest, RILP may play a role in both activating the V-ATPase (De Luca et al., 2014) and regulating the microtubule-mediated trafficking of these compartments to the perinuclear region of the cell (Li et al., 2016) (Figure 5.11). Both acidification and transport to the perinuclear region are involved in the formation of mature lysosomes (Huotari & Helenius, 2011), suggesting that this RILP/V-ATPase interaction may coordinate these aspects of endosome maturation.

5.1.3.3 Availability of specific amino acids impacts V-ATPase activity

The central signalling role of the V-ATPase is exemplified by the identification of the V- ATPase as a lysosomal amino acid sensor (Zoncu et al., 2011), as discussed in section 5.1.2.7. There is evidence that the levels of different amino acids in the cell differentially effect the amino acid sensing capacity of the V-ATPase, with specific amino acids exerting different effects on V-ATPase activity and domain association

178 5. The Vacuolar ATPase and PD

(Stransky & Forgac, 2015). For instance, withdrawal of arginine, leucine and lysine increase V-ATPase activity and domain association, while withdrawal of glutamate and serine has the opposite consequence (Stransky & Forgac, 2015). Alternatively both aspartate and methionine withdrawal alter V-ATPase activity, but have no affect on domain association (Stransky & Forgac, 2015). Although the mechanisms through which these specific amino acids exert their opposing affects on V-ATPase activity and domain association are yet to be elucidated, this complex regulatory mechanism highlights the complexity of V-ATPase regulation.

5.1.3.4 V-ATPase activity is regulated by altering proton-coupling efficiency

Activity of the V-ATPase complex can also be regulated at the level of proton coupling efficiency (Kettner et al., 2003), that is, the number of protons transported as a result of a single ATP hydrolysis. In yeast S. cerevisiae coupling efficiency of the V-ATPase appears to be controlled by the V0 a subunit (Kawasaki-Nishi et al., 2001a) and/or the V1 D subunit (Xu & Forgac, 2000) and differential coupling efficiency is evident between different subcellular localizations in the cell (Kawasaki-Nishi et al., 2001a). This may be a mechanism that contributes to the establishment of a pH gradient along the endolysosomal compartment maturation pathway, as the V-ATPase on the less acidic Golgi complex has lower coupling efficiency than the V-ATPase on the more acidic vacuole (yeast lysosome) (Kawasaki-Nishi et al., 2001b).

5.1.3.5 Transcriptional regulation of the V-ATPase

Expression of V-ATPase transcripts is tightly regulated and coordinated. Several transcription factors have been identified which regulate V-ATPase genes in response to different stimuli, the best-characterized being Transcription Factor EB (TFEB). TFEB activates transcription of a set of genes, termed coordinated lysosomal expression and regulation (CLEAR) genes, which include the V-ATPase subunits (Sardiello et al., 2009) (Peña-Llopis et al., 2011) as well as genes involved in autophagy, lysosomal biogenesis

179 5. The Vacuolar ATPase and PD and activity and membrane repair (Sardiello et al., 2009). TFEB senses lysosomal dyshomeostasis and insufficient degradation capacity through interaction with mTORC1 (Peña-Llopis et al., 2011). In nutrient rich conditions TFEB is tethered to the lysosome through the same activated Rag proteins that recruit mTORC1 to the lysosome surface (see Figure 5.10 above). This close proximity of TFEB to mTORC1 allows TFEB to be inactivated by mTORC1-mediated phosphorylation. Once inactivated, TFEB either remains tethered to the lysosomal surface or is sequestered within the cytosol by 14-3-3 proteins (Roczniak-Ferguson et al., 2012). Alternatively, under conditions of starvation or lysosomal stress mTORC1 is inactivated and TFEB becomes dephosphorylated. This dephosphoylation allows TFEB to translocate to the nucleus, where it activates transcription of genes through recognition of the CLEAR element (Martini-Stoica et al., 2016) (Figure 5.12). This TFEB-dependent regulation enables V-ATPase subunit genes to be transcriptionally up-regulated in a coordinated fashion in response to lysosomal stress.

Figure 5.12 Together with the lysosome nutrient sensing (LYNUS) machinery, TFEB senses lysosomal stress/insufficiency and acts as a master regulator of lysosomal gene

180 5. The Vacuolar ATPase and PD transcription. TFEB associates with activated Rag proteins on the lysosomal surface. This brings TFEB into contact with mTORC1, which phosphorylates TFEB and prevents its translocation to the nucleus. Upon lysosomal stress mTORC1 is inactivated and no longer phosphorylates TFEB. This allows TFEB to translocate to the nucleus, where it up-regulates expression of a subset of genes termed coordinated lysosomal expression and regulation (CLEAR) genes, which include the V-ATPase subunits as well as genes involved in autophagy, lysosomal biogenesis and activity and membrane repair.

5.1.4 Evidence of V-ATPase dysfunction in neurodegenerative diseases including PD

Mutations in specific V-ATPase genes are associated with a number of disease conditions that are characterised by neurodegeneration and cognitive impairment. For instance, recessive tubular acidosis (Stover et al., 2002), Zimmerman loband syndrome (Kortüm et al., 2015), autosomal recessive osteopetrosis (Bhargava et al., 2012) and cutis laxa type II (Van Damme et al., 2017) can be attributed to mutations in the V0A4, V1B2, V0A3 and V0A2 genes, respectively. These conditions are associated with neurological symptoms, including mental retardation, neurodegeneration or, in less severe cases, susceptibility to ageing associated mental deterioration (Colacurcio & Nixon, 2016). The association of neurological features with these V-ATPase-linked diseases suggests that the brain may be particularly susceptible to perturbed V-ATPase activity. This may be particularly true for neurons, which are long lived, post-mitotic cells that cannot divide or regenerate to avoid damage. Therefore well-regulated and efficient lysosomal degradation and recycling of cellular components is essential to ensuring long-term neuronal survival.

Mutations in the gene encoding the V-ATPase associated protein ATP6AP2, have also been specifically associated with PD, providing evidence that altered V-ATPase activity may contribute to PD. Variants in the ATP6AP2 gene, which acts as a an accessory protein regulating V-ATPase assembly have been linked to X-linked PD (Gupta et al., 2015; Poorkaj et al., 2010). These PD associated ATP6AP2 variants were shown to cause altered splicing when replicated in vitro and exon 4 skipping was evident in affected patients (Korvatska et al., 2013). Furthermore, replication of these mutations

181 5. The Vacuolar ATPase and PD in vivo has been shown to increase cognitive impairment and neurodegeneration in flies and mice (Dubos et al., 2015). This form of Parkinsonism has variable age of onset and is characterized by typical PD symptoms, including resting tremor, bradykinesia and rigidity. However patients appear to also display pathological overlaps with AD. Post mortem analysis of one affected individual identified tau protein accumulation, a pathological hallmark of AD, while αSyn pathology was absent (Poorkaj et al., 2010).

Perturbed lysosomal acidification, a process that is primarily mediated by the V- ATPase, is also implicated as a pathogenic mechanism in other familial forms of PD. Mutations in a number of PD genes, including LRRK2 (Henry et al., 2015), ATP13A2 (PARK9) (Dehay et al., 2012a; Dehay et al., 2012b), GBA (Bourdenx et al., 2016) and overexpression of αSyn (Stefanis et al., 2001) impair lysosomal acidification. Additionally, rotenone and methamphetamine exposure, which are both PD risk factors (Curtin et al., 2015; Dhillon et al., 2008), also impair lysosomal acidification (Funakoshi-Hirose et al., 2013; Pal et al., 2016). These results suggest that impaired lysosomal acidification may play a role in both familial and sporadic forms of PD.

There is also evidence that V-ATPase dysfunction contributes to neuronal degeneration in AD, a neurodegenerative condition that has significant symptomatic overlaps with PD. The most common form of early onset familial AD is caused by mutations in the presenillin 1 gene (Chouraki & Seshadri, 2014). Presenillin 1 is a component of the gamma secratase complex, which cleaves integral membrane proteins in their transmembrane domain. Processing of the integral membrane protein amyloid precursor protein is impaired by pathogenic Presenillin 1 mutations. This results in accumulation of amyloid β, the predominant component of the amyloid deposits that characterize AD (Suh & Checler, 2002). However, more recently these Presenillin 1 mutations have also been shown to affect the processing of other integral membrane proteins, including the V-ATPase membrane bound subunit ATP6V0A1. In vitro and in vivo depletion of Presenillin 1 has been shown to impair N-linked glycosylation of ATP6V0A1 in the endoplasmic reticulum (ER), which in turn inhibits ER to lysosome transport of the subunit (Lee et al., 2015). As a result, Presenillin 1

182 5. The Vacuolar ATPase and PD mutations cause reduced V-ATPase activity, reduced lysosomal acidity and impaired lysosomal proteolysis and autophagy (Lee et al., 2010). Additionally, restoration of lysosomal pH in Presenillin 1 KO cells ameliorated these lysosomal defects and has been proposed as a novel therapeutic approach for AD (Lee et al., 2015).

The association of variants in the ATP6AP2 gene with X-linked parkinsonism and the finding that lysosomal acidification may be perturbed in other forms of familial PD and in response to environmental risk factors for PD is compelling evidence that V-ATPase impairment may contribute to both sporadic and familial forms of the disease. Additionally, the association between impaired V-ATPase trafficking and AD, a disease that has many symptomatic and molecular overlaps with PD, may implicate a broader l role of V-ATPase dysfunction in neurodegeneration.

5.1.5 Aims

Perturbed V-ATPase protein abundance was observed in response to treatment of SH- SY5Y cells with rotenone and/or elevated Syn expression (section 3). Additionally, RNAseq analysis of PD patient brain regions identified that the expression of V-ATPase genes is reduced in sporadic PD. I therefore aimed to confirm the finding that V- ATPase mRNA expression is reduced in sporadic PD patients and to determine whether V-ATPase protein levels are similarly reduced in patient brain regions.

Following validation that V-ATPase mRNA expression and protein abundance is reduced in PD patient brain regions, I aimed to assess whether reduced V-ATPase expression contributes to known PD cellular phenotypes in vitro using differentiated SH-SY5Y neuronal-like cells. Such PD-associated phenotypes included both lysosomal and mitochondrial phenotypes, as dysfunction of both of these organelles appears to play a central role in the degenerative process (Martini-Stoica et al., 2016; Schapira, 2008). I also aimed to determine if V-ATPase impairment could contribute to Syn accumulation, a pathological hallmark of PD (Spillantini et al., 1997).

183 5. The Vacuolar ATPase and PD

5.2 Results and Discussion

5.2.1 V-ATPase mRNA and protein levels are reduced in brain regions of sporadic PD patients

Mass spec analysis of SH-SY5Y cells chronically treated with low levels of rotenone or mild αSyn elevation revealed that the abundance of specific V-ATPase subunits increased in response to these distinct stresses and that this response is amplified when these stresses are combined (Section 3.2.5; Figure 3.5). I aimed to determine whether there is evidence of perturbed V-ATPase functionality in patient brain regions by assessing V-ATPase mRNA and protein levels using available patient samples and RNAseq data.

5.2.1.1 V-ATPase transcript levels are reduced in sporadic PD patient brain regions

V-ATPase gene expression was assessed using existing RNAseq analysis (Marshall & Cooper, unpublished). Gene expression was assessed in 3 different brain regions; the Basal Forebrain (BF), Superior Frontal Cortex (SFC) and Superior Occipital Cortex (SOC). As discussed in section 3.2.3, these brain regions are pathologically affected in progressive disease stages; the BF displays substantial pathology at a relatively early stage of the disease, prior to the time of death of these patients, while the SFC was yet to display overt pathology in these patients and the SOC is not pathologically affected in PD.

Assessment of this RNAseq data revealed that transcription of the primarily expressed gene paralogue for every V-ATPase subunit is reduced in all 3 brain regions examined (Table 5.2), relative to age/sex-matched controls. This suggests that the reduction in V- ATPase expression in sporadic PD patients may be a prevailing feature in the disease that is not limited to a specific disease stage and is not a direct result of accumulating pathology. Although the statistical significance of the transcriptional changes in these

184 5. The Vacuolar ATPase and PD

Table 5.2 RNAseq analysis identified that V-ATPase gene expression is reduced in the Basal Forebrain, Superior Frontal Cortex and Superior Occipital Cortex of sporadic PD patients. Data obtained from 10 patients and 10 controls. P value determined by unpaired Students t-tests.

Superior Frontal Superior Occipital Basal Forebrain Cortex Cortex

Fold Fold Fold Associated Change P Value Change P Value Change P Value Gene Name Gene ID Description (Log2) (Log2) (Log2) ATP6V0A1 ENSG00000033627 ATPase H+ transporting V0 subunit a1 -0.399 3.62E-02 -0.397 1.81E-02 -0.786 4.40E-05 ATP6V0B ENSG00000117410 ATPase H+ transporting V0 subunit b -0.212 2.32E-01 -0.299 6.39E-02 -0.487 9.07E-03 ATP6V0C ENSG00000185883 ATPase H+ transporting V0 subunit c -0.397 3.36E-02 -0.432 1.66E-02 -0.808 2.19E-05 ATP6V0D1 ENSG00000159720 ATPase H+ transporting V0 subunit d1 -0.486 2.23E-02 -0.401 2.09E-02 -0.678 4.62E-05 ATP6V0E2 ENSG00000171130 ATPase H+ transporting V0 subunit e2 -0.533 9.06E-02 -0.514 2.41E-02 -0.775 3.48E-04 ATP6V1A ENSG00000114573 ATPase H+ transporting V1 subunit A -0.795 1.93E-03 -0.649 6.93E-03 -0.921 2.00E-04 ATP6V1B2 ENSG00000147416 ATPase H+ transporting V1 subunit B2 -0.701 2.94E-03 -0.689 5.85E-03 -1.050 2.30E-05 ATP6V1C1 ENSG00000155097 ATPase H+ transporting V1 subunit C1 -0.617 6.74E-03 -0.585 4.05E-03 -0.538 5.52E-03 ATP6V1D ENSG00000100554 ATPase H+ transporting V1 subunit D -0.539 1.57E-03 -0.463 1.19E-02 -0.752 4.85E-04 ATP6V1E1 ENSG00000131100 ATPase H+ transporting V1 subunit E1 -0.450 3.02E-02 -0.530 9.62E-03 -0.929 7.19E-05 ATP6V1F ENSG00000128524 ATPase H+ transporting V1 subunit F -0.150 5.18E-01 -0.382 2.35E-02 -0.609 1.16E-03 ATP6V1G1 ENSG00000136888 ATPase H+ transporting V1 subunit G1 -0.353 4.20E-02 -0.214 9.15E-02 -0.328 7.70E-03 ATP6V1H ENSG00000047249 ATPase H+ transporting V1 subunit H -0.474 1.27E-01 -0.526 5.69E-03 -0.902 1.49E-05

185 5. The Vacuolar ATPase and PD

13 V-ATPase genes fell outside the false discovery rate (FDR) cut-off of 0.05, the statistically significant p values of the transcriptional changes and their coordinated reduced expression, evident across multiple brain regions warranted further investigation. Subsequently, further transcriptomic analysis by qRT-PCR confirmed that expression of 3 core V-ATPase subunits (ATP6V1A, ATP6V1B2 and ATP6V0A) is significantly reduced in the SOC and that ATP6V1A expression is additionally significantly reduced in the BF and SFC, while ATPV1B2 expression is additionally significantly reduced in the BF (Figure 5.13). Transcriptional expression of these subunits was reduced by approximately 50%, closely reflecting the extent of change identified by RNAseq.

Figure 5.13 The expression of core V-ATPase subunits is significantly reduced in brain regions of sporadic PD patients. qRT-PCR analysis of three core V-ATPase subunits (ATP6V0A, ATP6V1A and ATP6V1B2) in the basal forebrain, superior frontal cortex and superior occipital cortex of PD patients, as calculated using the delta delta CT method using the housekeeping gene NFX1. Statistical significance determined by Students t-tests corrected for multiple

186 5. The Vacuolar ATPase and PD comparisons using the Holm-Sidak test (*p<0.01, ** p<0.001). Error bars represent standard deviation, n=10/10 patients/controls.

5.2.1.2 V-ATPase protein abundance is reduced in sporadic PD patient brain tissue

To determine if V-ATPase protein abundance is also reduced in PD patient brain regions the abundance of the ATP6V1B2 subunit, an integral component of the V- ATPase V1 domain, was assessed. As protein samples matching the samples and brain regions previously assessed for transcriptomic changes (BF, SFC and SOC) were unavailable, V-ATPase protein abundance was assessed in Anterior Cingulate Cortex (ACC) protein fraction samples obtained from a separate cohort of 6 sporadic PD patients and 6 age-sex matched controls. The ACC is a brain region that is affected relatively late in PD disease progression (BF  SFC  ACC) and exhibited substantial αSyn protein accumulation at the time of death of these 6 patients (Braak stage IV), but did not display overt Lewy pathology or neuronal loss (Murphy et al., 2015). Therefore assessment of the ACC enables the early effects of αSyn accumulation to be examined without results being distorted by alterations in the cell types present due to neuronal loss.

Assessment of ATP6V1B2 protein levels in the ACC revealed that the abundance of this core V-ATPase subunit is reduced by approximately 50% in sPD patients (Figure 5.14A), correlating with the reduced mRNA expression level identified in other brain regions. This result suggests that V-ATPase activity is likely to be reduced in sPD patient brain regions due to decreased protein abundance.

187 5. The Vacuolar ATPase and PD

Figure 5.14 ATP6V1B2 protein abundance is reduced in the Anterior Cingulate Cortex of PD patients. (A) ATP6V1B2 protein density calculated relative to abundance of the loading control βtubulin. Statistical significance determined by unpaired Students t-test with Welch’s correction (*p<0.05). Error bars represent standard deviation (n= 6 patients and 6 controls). (B) Western blot image indicating patient disease stage.

5.2.1.3 Exploring the relationship between the V-ATPase and αSyn in sPD patient brain tissue

Supporting the hypothesis that reduced V-ATPase levels are likely to play a role in PD pathogenesis, a significant (p= 0.032) inverse correlation was observed between the abundance of the ATP6V1B2 and αSyn proteins in sPD patients, but not in controls (Figure 5.15). αSyn protein accumulation in PD is a central pathological feature of the disease, however the function of αSyn remains unknown, as does the cause of its accumulation in PD. The observation that increased αSyn abundance is significantly correlated with reduced ATP6V1B2 abundance in sPD patients positions reduced V- ATPase levels as a potential contributing factor to αSyn based pathogenesis in PD and highlights V-ATPase dysregulation in PD as a key area for further study.

188 5. The Vacuolar ATPase and PD

Figure 5.15 The reduction in ATP6V1B2 abundance in PD patients inversely correlates with αSynuclein protein abundance. αSynuclein protein abundance determined by Murphy et al., 2015. Results of statistical analysis indicated in the figure, as per Pearson’s correlation analysis.

The identification of a relationship between V-ATPase and αSyn protein levels is supported by published evidence that indicates that these proteins directly associate in vivo. Pull-down of αSyn identified an interaction between αSyn and particular V- ATPase subunits in mouse brain synaptosomes (Table 5.3) (McFarland et al., 2008), which is perhaps not surprising given the synaptic localisation of αSyn (Maroteaux et al., 1988). Additionally, the association between αSyn and V-ATPase subunits was dependent on phosphorylation of αSyn (Table 5.3) (McFarland et al., 2008), a modification that is associated with increased αSyn toxicity (Sugeno et al., 2008). This suggests not only that the V-ATPase and αSyn are likely to interact, but also that αSyn phosphorylation, which is increased in PD patient brain regions (Sato et al., 2013), could potentially functionally affect the V-ATPase.

Table 5.3 Interactions identified between V-ATPase subunits and Ser-129 phosphorylated and non-phosphorylated αSynuclein. Data obtained from McFarland et al (2008).

V-ATPase Subunit Interaction with α-Synuclein Appeared equally in pull-downs with Ser-129 and non-phosphorylated Atp6v1b2 α-synuclein

189 5. The Vacuolar ATPase and PD

Appeared equally in pull-downs with Ser-129 and non-phosphorylated Atp6v0a1 α-synuclein Atp6v0d1 Preferential affinity for the Ser-129 phosphorylated α-synuclein Atp6v1a Preferential affinity for the Ser-129 phosphorylated α-synuclein

5.2.1.4 Single Nucleotide Variants in the ATP6V0A1 gene may be associated with increased risk of PD

Further evidence supporting the hypothesis that impaired V-ATPase activity may be a contributing factor to PD was obtained by assessing whether single nucleotide variants (SNVs) within V-ATPase genes are associated with increased risk of PD. For this purpose Scott Youlton (Garvan Institute, Bone Biology Division) conducted a Versatility Gene Based Association Study (VEGAS) to summarise the risk associated with multiple SNVs across each of the V-ATPase genes (defined as the region between initiation codon and translational stop codon), thus giving a single P value overview for each gene of interest (Table 5.4). This analysis revealed there are a number of near- significant SNVs within the ATP6V0A1 gene locus that are significantly associated with increased risk of PD (p=0.000003) when assessed using the VEGAS approach (Liu et al., 2010).

Table 5.4 Versatility Gene Based Association Study (VEGAS) analysis assessing PD- associated SNVs within V-ATPase genes. Corrected p value calculated based on the top 10% most significant single nucleotide variants (SNVs) within a gene.

Corrected p TopSNV p Chromosome Gene Start Stop value Value 5 ATP6AP1L 81601165 81614147 0.872127872 0.501187 17 ATP6V0A1 40610861 40674597 3.00E-06 6.46E-07 12 ATP6V0A2 124196864 124246301 0.098901099 0.0501187 7 ATP6V0A4 138391038 138482941 0.967032967 0.0501187 1 ATP6V0B 44440601 44443972 0.555444555 0.501187 16 ATP6V0D1 67471916 67515089 0.118881119 0.0501187 8 ATP6V0D2 87111138 87166454 0.87012987 0.0501187 5 ATP6V0E1 172410762 172461900 0.173826174 0.0501187 7 ATP6V0E2 149570056 149577801 0.807192807 0.501187

190 5. The Vacuolar ATPase and PD

ATP6V0E2- 7 AS1 149564782 149570951 0.662337662 0.501187 3 ATP6V1A 113465865 113530905 0.655344655 0.501187 2 ATP6V1B1 71162997 71192561 0.751248751 0.0501187 8 ATP6V1B2 20054703 20079207 0.985014985 0.501187 8 ATP6V1C1 104033247 104085285 0.28971029 0.0501187 2 ATP6V1C2 10861774 10925236 0.754245754 0.0501187 14 ATP6V1D 67804580 67826720 0.08991009 0.00616595 22 ATP6V1E1 18074902 18111588 0.603396603 0.0501187 2 ATP6V1E2 46738985 46747096 0.134865135 0.0501187 7 ATP6V1F 128502856 128505903 0.636363636 0.501187 9 ATP6V1G1 117349993 117361152 0.185814186 0.0501187 6 ATP6V1G2 31512227 31514625 0.925074925 0.501187 6 ATP6V1G2 31497995 31514625 0.274725275 0.0501187 1 ATP6V1G3 198492351 198510075 0.851148851 0.501187 8 ATP6V1H 54628102 54755871 0.070929071 0.0501187

Closer inspection of the individual SNVs within the ATP6V0A1 locus revealed that there are over 10 PD-associated SNVs within or close to this locus (Figure 5.16) that fall slightly outside of the accepted p value cut-off threshold for genome wide association

-7 studies of -log10(p)<7.3 (p<5 x 10 ) (Qu et al., 2010). The ability of VEGAS analysis to summarise these multiple near significant SNVs into a single p value has enabled the identification of ATP6V0A1 as a novel PD risk locus, which otherwise would have been overlooked by a traditional GWAS analysis.

Figure 5.16 Multiple PD-associated single nucleotide variants (SNVs) were identified across the ATP6V0A1 locus. SNVs were identified within and in close proximity to the

191 5. The Vacuolar ATPase and PD

ATP6V0A1 gene that fell slightly below the significance threshold of –log10(p)<7.3 for GWAS studies.

The above data indicates that V-ATPase mRNA and protein levels are reduced in affected brain regions of PD patients (Figure 5.13, Figure 5.14), strongly implicating a role of V-ATPase dysfunction in PD. Additionally, the identified interaction between the V-ATPase and αSyn (Table 5.3), the observed inverse correlation between the levels of these proteins (Figure 5.14) and the identification of ATP6V0A1 as a novel PD- risk locus (Table 5.4) further suggests that V-ATPase dysfunction contributes to the disease. To further explore this I assessed whether V-ATPase impairment in vitro leads to the development of PD-associated cellular phenotypes, including lysosomal and mitochondrial dysfunctions that are known to occur in the disease. Such information would provide additional evidence that V-ATPase depletion is a contributing factor to the disease.

5.2.2 Modelling mild V-ATPase impairment in SH-SY5Y cells

Inhibition of the V-ATPase using high levels of the specific chemical inhibitors Bafilomycin and Concanamycin has been shown to impair lysosomal function and result in cell death (Mangieri et al., 2014; Woo et al., 1992; Yoshimori et al., 1991). However, the finding that V-ATPase levels are depleted in sPD patient brain regions that are yet to show significant cell loss (Figure 5.13, Figure 5.14) suggests that the level of V-ATPase depletion in patients is below a lethal threshold. The few studies have been conducted into the effects of moderate or low-level V-ATPase impairment suggest that the cellular responses to low doses of V-ATPase inhibitors are very distinct from the responses to high doses (Pivtoraiko et al., 2010; Shacka et al., 2006; Sobota et al., 2009). To explore the possible consequences of the moderately reduced V-ATPase expression that is observed in sPD patients, the effect of non-toxic V-ATPase impairment on PD-relevant cellular phenotypes was assessed in differentiated neuronal-like SH-SY5Y cells.

192 5. The Vacuolar ATPase and PD

Two complementary V-ATPase inhibition methods were used to model V-ATPase impairment:

1) Chemical inhibition of the V-ATPase using BafilomycinA1 (BafA1), a specific inhibitor of the c subunit of the V0 domain (Bowman et al., 1988; Forgac, 2007).

2) siRNA mediated knock-down (KD) of two structurally and functionally distinct subunits of the V1 domain (ATP6V1A and ATP6V1B2).

5.2.2.1 Establishing a chronic, low-dose BafilomycinA1 induced model of V- ATPase impairment

To mimic the moderate reduction in V-ATPase activity that is likely to be present in sPD patient brain regions, as suggested by the observed moderate reduction in transcript and protein levels (Figure 5.13, Figure 5.14), the BafA1 concentration used in this study was titred to determine a dosage that induced no significant loss of cell viability following 48 hours of treatment. This treatment time period is considered chronic for this experimental approach as other BafA1 based studies generally assess the effect of 1-12 hours of treatment with BafA1 at concentrations >100nM (Hong et al., 2006; Mauvezin et al., 2015; Zhdanov et al., 2011).

Cell viability was not significantly affected at BafA1 concentrations at or below 2.5 nM (Figure 5.17). Therefore a chronic, low-dose model of 1 nM treatment for 48 hours was selected for further experiments.

193 5. The Vacuolar ATPase and PD

150

*** y

t ***

l i

b 100

v i

t 50

e R

0 0 0.625 1.25 2.5 5 10 BafilomycinA1 Concentration (nM)

Figure 5.17 BafilomycinA1 treatment does not significantly reduce cell viability at concentrations at 2.5nM or lower. Differentiated SH-SY5Y cells were treated with increasing concentrations of BafA1 for 48 hours, following which cell viability was assessed by Alamar blue assay. Significance determined by one-way ANOVA with two-stage step-up Benjimini, Krieger and Yekutieli’s correction for multiple comparisons (*** p<0.001). Error bars represent standard deviation. Data are from 4 biological replicates.

5.2.2.2 Establishing an siRNA mediated knock-down model of V-ATPase impairment

The second method of V-ATPase impairment used in this study involved siRNA- mediated depletion of the V-ATPase to mimic the reduction in V-ATPase expression observed in patient brain regions (Figure 5.13). The effect of depletion of 2 V-ATPase subunits (ATP6V1A and ATP6V1B2) was assessed as they exhibited the most substantial reduction in expression in patient brain regions, as determined by RNAseq analysis (Table 5.1), as well as the added advantage that depletion of 2 distinct genes minimises the possibility that the results obtained are due to siRNA off-target effects.

ATP6V1A and ATP6V1B2 depletion each resulted in a 67% and 84% reduction in the respective transcript level at 48 hours (Figure 5.18A). ATP6V1A KD did not significantly affect transcript levels of ATP6V1B2, nor visa-versa. Additionally, reflecting the intent to assess the effects of non-toxic levels of V-ATPase impairment, this level of V-ATPase depletion had no significant effect on cell viability at this time-point (Figure 5.18B).

194 5. The Vacuolar ATPase and PD

Figure 5.18 Knockdown of ATP6V1A and ATP6V1B2 results in reduced transcript levels of these subunits without reducing cell viability. (A) qRT-PCR analysis of ATP6V1A and ATP6V1B2 transcript levels in response to siRNA-mediated depletion of ATP6V1A or ATP6V1B2, or a non- targeting control siRNA. Fold Change (log2) calculated using the delta delta CT method using the housekeeping gene NFX1. (B) ATP6V1A and ATP6V1B2 depletion do not reduce cell viability as assessed by Alamar blue assay. Significance determined by one-way ANOVA with two-stage step-up Benjimini, Krieger and Yekutieli’s correction for multiple comparisons (*** p<0.001, **** p<0.0001). Error bars represent standard deviation. Data are from 4 biological replicates.

The effects of siRNA-mediated V-ATPase depletion on subunit protein levels were also assessed using the available ATP6V1B2 antibody. Knockdown of ATP6V1B2 reduced ATP6V1B2 protein levels, as expected, by approximately 50% (Figure 5.19). More surprisingly, depletion of the ATP6V1A subunit caused a similar decrease in ATP6V1B2 protein levels (Figure 5.19). This suggests that loss of ATP6V1A may destabilize the domain assembly and causes generalized degradation of multiple V1 domain subunits, without affecting subunit transcription. This result provides substantial insight into the assembly mechanisms that may regulate the stability of the V1 domain. Additionally, this 50% reduction in protein abundance closely mimics the reduction in protein abundance observed in the ACC of sporadic PD patients (Figure 5.14), making this a cellular model that closely recapitulates the V-ATPase reduction present in PD patient brain regions.

195 5. The Vacuolar ATPase and PD

Figure 5.19 siRNA-mediated knockdown of ATP6V1A or ATP6V1B2 reduces ATP6V1B2 protein abundance in SH-SY5Y cells. (A) Quantification of ATP6V1B2 protein levels as assessed by Western blot relative to the level of the loading control 14-3-3. (B) Representative western blot image. Significance determined by one-way ANOVA with two-stage step-up Benjimini, Krieger and Yekutieli’s correction for multiple comparisons (* p<0.05, ** p<0.01). Error bars represent standard deviation. Data are from 4 biological replicates.

5.2.3 V-ATPase impairment causes lysosomal dysfunction in SH-SY5Y cells

As discussed in section 1.9, there is substantial evidence supporting a pathological role of lysosomal dysfunction in PD. To explore whether V-ATPase depletion could contribute to this lysosomal dysfunction the effects of non-toxic, low-level V-ATPase impairment on lysosomal functions were assessed in differentiated SH-SY5Y cells.

5.2.3.1 V-ATPase impairment results in the accumulation of chaperone mediated autophagy (CMA) substrates

One of the primary functions of the lysosome is to degrade macromolecules, including proteins, lipids and nucleic acids, via autophagy lysosome pathways that encompass macro-autophagy, chaperone mediated autophagy (CMA) and micro-autophagy {MartiniStoica:2016fb}. Declining autophagy activity has previously been reported to play a role in PD pathogenesis (Colacurcio & Nixon, 2016; Gan-Or et al., 2015; Zhang et

196 5. The Vacuolar ATPase and PD al., 2015a) as well as ageing (Carmona-Gutierrez et al., 2016), which is a key PD risk factor. Furthermore, declining CMA activity, the process through which unneeded or damaged cytosolic proteins are degraded by the lysosome, is a suggested pathogenic mechanism in PD (Alvarez-Erviti et al., 2013; Murphy et al., 2015; Pang et al., 2012), with accumulation of CMA substrate proteins having been observed in affected patient brain regions (Murphy et al., 2015; Vogiatzi et al., 2008).

To determine whether V-ATPase impairment results in CMA substrate accumulation, mimicking the accumulation of CMA substrates observed in PD patient brain regions, the effect of BafA1 treatment or siRNA-mediated depletion of V-ATPase subunits on levels of the known CMA substrate proteins was assessed. The CMA substrates assessed were MEF2D (Yang et al., 2009) and GAPDH (Napolitano et al., 2015). Additionally, levels of the non-CMA degraded control protein, 14-3-3, were also assessed.

Depletion of ATP6V1A or ATP6V1B2 levels resulted in significantly increased abundance of the CMA substrates MEF2D and GAPDH by 80% and 50%, respectively, but did not effect the abundance of the non-CMA substrate 14-3-3 (Figure 5.20A). Similarly, levels of MEF2D and GAPDH, but not 14-3-3, were also increased in response to BafA1 treatment, albeit not to a significant level (Figure 5.20B).

To confirm that the increased CMA substrate abundance observed in response to V- ATPase impairment is likely to be the result of impaired degradation of these proteins, as opposed to increased transcription, the transcript levels of GAPDH and MEF2D were assessed in response to these treatments. Neither BafA1 nor V-ATPase KD altered the mRNA level of GAPDH or MEF2D (Figure 5.21). These results indicate that non-toxic levels of V-ATPase depletion reduce lysosomal degradation capacity and reproduce indications of lysosomal dysfunction that are evident in PD, positioning reduced V- ATPase expression as a possible contributing factor to lysosomal dysfunction in PD.

197 5. The Vacuolar ATPase and PD

Figure 5.20 V-ATPase impairment causes an increase in chaperone mediated autophagy substrates. (A) The abundance of the CMA substrates MEF2D and GAPDH, as well as the non- CMA substrate 14-3-3 were assessed in response to (A) siRNA-mediated depletion of ATP6V1A or ATP6V1B2 or (B) low-dose (1 nM) BafA1 treatment. Error bars represent standard deviation. Data are from (A) 8 or (B) 6 biological replicates. Representative Western blot images showing 2 biological replicates are shown. Error bars represent standard deviation. Statistical significance determined by two-way ANOVA (knockdown) or Students t-test (BafA1) with two- stage correction for multiple comparisons using the method of Benjamini, Kreger and Yekutieli (** p<0. 1, *** p<.0.001).

198 5. The Vacuolar ATPase and PD

Figure 5.21 Neither (A) siRNA-mediated V-ATPase knockdown or (B) BafA1 treatment causes a significant change in GAPDH or MEF2D mRNA level. Fold Change (log2) calculated relative to the level of the housekeeping gene NFX1. Error bars represent standard deviation (knockdown; n=4, BafA1 n= 3). Statistical significance determined by two-way ANOVA (knockdown) or Students t-test with two-stage correction for multiple comparisons using the method of Benjamini, Kreger and Yekutieli (BafA1).

5.2.3.2 V-ATPase impairment increases translocation of TFEB to the nucleus

To further explore the extent of lysosomal perturbation triggered by chronic, moderate V-ATPase impairment the sub-cellular localisation of lysosomal transcription factor EB (TFEB) was assessed. As discussed in section 5.1.3.5 monitoring the relative abundance of nuclear localised TFEB, a master regulator of lysosomal function and autophagy, provides an indication of lysosomal dysfunction and stress (Napolitano & Ballabio, 2016). Under homeostasis the transcription factor is phosphorylated and sequestered within the cytosol. However upon lysosomal stress TFEB translocates to the nucleus where it activates the transcription of a subset of CLEAR genes to compensate for the lysosomal dysfunction (Roczniak-Ferguson et al., 2012) (Figure 5.12). Therefore the relative abundance of nuclear TFEB was assessed in response to V-ATPase depletion or low-dose BafA1 treatment to determine the effects of these treatments on inducing lysosomal dysfunction.

Treatment of SH-SY5Y cells with either low-dose BafA1 or ATP6V1A or ATP6V1B2 subunit depletion significantly increased TFEB translocation to the nucleus (Figure

199 5. The Vacuolar ATPase and PD

5.22). A BafA1 dosage effect was also observed, with a 4 nM BafA1 treatment resulting in a greater increase in TFEB nuclear translocation than 1 nM treatment. No increase in nuclear size was evident (Figure 5.22), confirming that this result is not a consequence of increased relative nuclear area.

Figure 5.22 V-ATPase impairment triggers nuclear translocation of TFEB. Nuclear TFEB fluorescence intensity and area relative to total TFEB fluorescence intensity and area were quantified in response to (A) siRNA-mediated knockdown of ATP6V1A or ATP6V1B2 or (B) low- dose BafA1 treatment. Statistical significance determined by two-way ANOVA with two-stage correction for multiple comparisons using the method of Benjamini, Kreger and Yekutieli (**p<0.01, *** p<0.001). Data obtained from measurement of 30 cells per treatment. Error bars represent SEM. Scale bar represents 10 um.

200 5. The Vacuolar ATPase and PD

The identification of accumulation of CMA substrates (Figure 5.20) and activation of TFEB (Figure 5.22) in response to low levels of V-ATPase impairment confirms that these treatments result in lysosomal stress, suggesting that the similar levels of V- ATPase depletion observed in patients (Figure 5.13; RNA, Figure 5.14; protein) may contribute to the lysosomal dysfunction that is a characteristic of PD.

5.2.4 V-ATPase impairment impacts mitochondria in SH-SY5Y cells

There is accumulating evidence suggesting that in addition to perturbing endolysosomal functions, V-ATPase impairment may also impact the mitochondria. In particular, emerging evidence based on observations of yeast suggests that a relationship exists wherein V-ATPase impairment perturbs the pH regulation of the vacuole (the yeast lysosome equivalent), which subsequently causes mitochondrial dysfunction (Hughes et al., 2016). This V-ATPase-mediated lysosome-mitochondria relationship suggests that the V-ATPase depletion identified in affected PD patient brain regions (Figure 5.13, Figure 5.14) may contribute to both the lysosomal and mitochondrial dysfunctions that are characteristics of the disease. Additionally, the mitochondrial and lysosomal dysfunctions induced by V-ATPase impairment in yeast, including the loss of lysosomal acidity, were associated with ageing (Hughes & Gottschling, 2012), implicating a role of V-ATPase dysfunction in ageing. This finding is also of particular relevance to PD as ageing is the primary risk factor for the disease.

To explore this lysosome-mitochondria relationship and, in particular, to assess the possibility that V-ATPase depletion could contribute to mitochondrial dysfunction in PD, the effects of V-ATPase impairment on mitochondrial parameters were assessed in vitro. Specifically, as mitochondrial dysfunction in PD is characterised by perturbed Ca2+ homeostasis (Hurley et al., 2013; Surmeier, 2007; Surmeier & Sulzer, 2013) and elevated mitochondrial ROS (discussed in section 1.10), the effects of V-ATPase impairment on mitochondrial Ca2+ and ROS were assessed.

201 5. The Vacuolar ATPase and PD

5.2.4.1 V-ATPase impairment increases mitochondrial Ca2+ levels

Perturbed Ca2+ handling is implicated as a pathogenic mechanism in PD based on the observations that the levels of calcium binding proteins, including calbindin (Hurley et al., 2013), calreticulin (Hurley et al., 2013; Wilhelmus et al., 2011) and calmodulin (Hurley et al., 2013; Reynolds et al., 2008), and the expression of plasma membrane calcium channels are altered in brain regions of PD patients (Hurley et al., 2013; Watson et al., 1988). Additionally, these changes are observed prior to the appearance of PD pathology or neuronal degeneration (Hurley et al., 2013), suggesting that perturbed Ca2+ homeostasis is an early feature of the disease.

Maintaining cellular Ca2+ homeostasis requires coordinated regulation of the Ca2+ levels in the multiple organelles that act as Ca2+ storage sites. Endolysosomal compartments provide substantial Ca2+ storage capacity (Christensen et al., 2002) (discussed in section 5.1.2.4) and together with the endoplasmic reticulum and mitochondria these compartments play a central role in regulating Ca2+ levels within the cell (Lloyd-Evans & Platt, 2011). It is currently unknown if the levels of Ca2+ in these subcellular storage sites are altered brain regions of PD patients due to difficulties associated with assessing subcellular Ca2+ levels in post-mortem tissue, however multiple studies conducted on PD model systems suggest that this is likely to be the case. Most convincingly, there is evidence of perturbed subcellular Ca2+ levels in PD patient-derived induced pluripotent stem cells (iPSC) differentiated to midbrain dopaminergic neurons. In these cells while there was no overall change in the total cellular Ca2+ level, the level of lysosomal Ca2+ was reduced, while the level of mitochondrial calcium was increased (Kilpatrick et al., 2016; Schöndorf et al., 2014).

V-ATPase impairment has previously been associated with impaired cellular Ca2+ homeostasis. Depletion of the ATP6V0A subunit was observed to trigger lysosomal Ca2+ release (Lee et al., 2015), resulting in reduced lysosomal Ca2+ levels that are reminiscent of the reduced lysosomal Ca2+ levels observed in PD patent iPSC-derived neurons (Kilpatrick et al., 2016; Schöndorf et al., 2014). However the possible effects

202 5. The Vacuolar ATPase and PD of V-ATPase dysfunction on perturbing other aspects of cellular Ca2+ homeostasis are unknown. In particular, it is unknown if the loss of lysosomal Ca2+ induced by V-ATPase depletion (Lee et al., 2015) may be accompanied by a compensatory increase in mitochondrial Ca2. To explore this possibility, mitochondrial Ca2+ levels were assessed in response to V-ATPase impairment mediated by low-dose BafA1 treatment or siRNA- mediated depletion of ATP6V1A or ATP6V1B2.

V-ATPase impairment caused a significant increase in mitochondrial Ca2+ following either 48 hours of treatment with BafA1 or ATP6V1A depletion (Figure 5.23A&B) or 72 hours of ATP6V1B2 depletion (Figure 5.23C). The time delay in the effect of ATP6V1B2 may be due to compensatory up-regulation of the ATP6V1B1 gene paralogue, making the depletion of this subunit less effective at perturbing V-ATPase activity. This finding that V-ATPase impairment increases mitochondrial Ca2+ levels indicates that V-ATPase impairment has consequences beyond the endolysosomal network.

To determine whether V-ATPase impairment additionally increases cytosolic Ca2+ levels, cytosolic Ca2+ was also assessed in response to V-ATPase impairment. Neither BafA1 treatment nor V-ATPase depletion had an effect on cytosolic Ca2+, while the positive control KCl resulted in the expected significant increase in cytosolic Ca2+ levels (Figure 5.23). This result demonstrates that V-ATPase impairment can elevate mitochondrial Ca2+ without impacting cytosolic Ca2+ levels. Additionally, increased cytosolic Ca2+ is known to induce apoptotic cell death through activation of the calpain- caspase pathway (La Rovere et al., 2016). This result therefore indicates the BafA1 concentration used and the extent of siRNA-mediated V-ATPase depletion do not activate apoptosis, correlating with the previous observations that these treatments do not cause a significant loss of cell viability (Figure 5.17; Figure 5.18).

203 5. The Vacuolar ATPase and PD

Figure 5.23 Mitochondrial calcium levels are elevated by V-ATPase impairment with BafA1 or siRNA-mediated depletion of the ATP6V1A and ATP6V1B2 subunits. (A) Mitochondrial and cytosolic calcium were measured following siRNA-mediated depletion of ATP6V1A or ATP6V1B2 following 48 or (B) 72 hours or (C) treatment with BafA1 or the positive control KCl following 48 hours. Statistical significance determined by two-way ANOVA with Dunnett’s correction for multiple comparisons (***p<0.001 ****p<0.0001). Data is from 4 biological replicates, error bars represent standard deviation. (D) Calcium levels were measured by FACS analysis of a ratiometric calcium sensitive probe targeted to the cytosol (GECO) or mitochondria (mGECO) as per the cell population gating indicated.

Together with the previously published result indicating that V-ATPase depletion reduces lysosomal Ca2+ levels (Lee et al., 2015), the results of this study suggest that reduced V-ATPase expression perturbs cellular Ca2+ homeostasis and both reduces lysosomal Ca2+ levels and increases mitochondrial Ca2+ levels, recapitulating the changes in subcellular Ca2+ levels that are observed in patient derived iPSC neurons

204 5. The Vacuolar ATPase and PD

(Kilpatrick et al., 2016; Schöndorf et al., 2014). This result indicates that V-ATPase impairment has consequences outside of the endolysosomal network and in particular on mitochondria and suggests that the V-ATPase depletion observed in sporadic PD patient brain regions (Figure 5.13, Figure 5.14) could contribute to the perturbed Ca2+ handling that has been implicated as a pathogenic mechanism in PD (Kilpatrick et al., 2016; Luth et al., 2014).

5.2.4.2 V-ATPase impairment increases the level of mitochondrial reactive oxygen species

Increased mitochondrial Ca2+ levels are typically accompanied by an increase in mitochondrial reactive oxygen species (ROS) (Brookes et al., 2004; Görlach et al., 2015; Hansson et al., 2008; Peng & Jou, 2010). This relationship is due to the effect that mitochondrial Ca2+ has on stimulating aerobic metabolism (Zhang et al., 2016) through the activation of a number of key Krebs cycle enzymes {Wan:1989vs}, including pyruvate dehydrogenase phosphatase, a-ketoglutarate dehydrogenase and isocitrate dehydrogenase {Denton:2009kd} (Raffaello et al., 2016). Activation of these enzymes increases the supply of the OXPHOS substrates NAHD and succinate, which has the subsequent effect of boosting OXPHOS activity. However, beyond a critical limit threshold, elevated mitochondrial Ca2+ levels and the associated increased OXPHOS activity can result in ROS production that exceeds the mitochondria’s antioxidant capacity, leading to a net increase in damaging ROS production (Adam-Vizi & Starkov, 2010) and oxidative stress. Due to this close association between increased mitochondrial Ca2+ and ROS, and the well-established connection between oxidative stress and PD (discussed in section 1.10), the effect of V-ATPase impairment on mitochondrial ROS production was assessed, with the prediction that mitochondrial Ca2+ overload induced by V-ATPase impairment may result in increased mitochondrial ROS production.

Both BafA1 treatment and depletion of the ATP6V1A subunit significantly increased mitochondrial ROS levels (Figure 5.24) as indicated by increased staining with the

205 5. The Vacuolar ATPase and PD mitochondrially targeted ROS dye mitoSox (ThermoFisher). Depletion of the ATP6V1B2 subunit also increased mitochondrial ROS, although not to a significant extent. A similar association between V-ATPase impairment and increased mitochondrial ROS has previously been observed in yeast (Milgrom et al., 2007), however no mechanism has yet been identified to explain this effect. The above results suggest that the increase in mitochondrial ROS caused by V-ATPase impairment is likely a downstream effect of mitochondrial Ca2+ overloading, however additional experimental manipulations are required to confirm this.

Figure 5.24 Mitochondrial Reactive Oxygen Species (ROS) are increased in response to V- ATPase impairment. Mitochondrial ROS was assessed in response to (A) siRNA-mediated V- ATPase depletion or (B) BafA1 treatment using the mitochondrially targeted ROS dye mitoSOX. Statistical significance determined by (B) one-way ANOVA with Dunnets correction for multiple comparisons or (A) Students t-test with Welch’s correction (***p<0.001, **p<0.01). Error bars represent standard deviation. MFI; Median Fluorescence Intensity. (C) Example flow cytometry cell population gating and mitoSOX fluorescence intensity.

206 5. The Vacuolar ATPase and PD

5.2.5 V-ATPase impairment impacts mitophagy

Perturbed mitophagy is implicated as a pathogenic mechanism in PD (as discussed in section 1.11.1) based on the discoveries that mutations in the genes encoding the PINK1 and PARKIN, two proteins that are central to mitophagy, cause familial forms of PD (Truban et al., 2017) and the observation that PARKIN activation is reduced in sporadic cases of the disease (Dawson & Dawson, 2014). Furthermore, mitochondrial damage is prevalent in patient brain tissue in both sporadic and familial forms of the disease and is proposed to accumulate due to impaired degradation of these defective organelles (Alam et al., 1997; Dexter et al., 1994; Zhang et al., 1999).

The observed increases in mitochondrial Ca2+ and ROS resulting from V-ATPase impairment (Figure 5.23, Figure 5.24) are known to occur during the early stages of mitophagy (Rimessi et al., 2013; Wang et al., 2012; Zhang et al., 2016b), suggesting that V-ATPase impairment is likely to increase the initiation of mitophagy. However, the final stages of mitophagy are dependent on the function the lysosome, which was perturbed by V-ATPase impairment (Figure 5.20, Figure 5.22). This suggests that V- ATPase impairment may have two detrimental influences on mitophagy, both increasing mitophagy initiation and compromising the final degradation stage of the process. Therefore, to determine the effect of V-ATPase impairment on mitophagy and to assess whether V-ATPase depletion could contribute to perturbed mitochondrial quality control in patients, multiple stages of mitophagy were assessed in response to V-ATPase depletion or BafA1 treatment. The stages of mitophagy assessed were (1) mitophagy initiation through recognition of damaged mitochondria, (2) fusion of autophagosomes with lysosomes and (3) degradation of the mitochondrial contents in the lysosome.

5.2.5.1 V-ATPase impairment increases the initiation of mitophagy

The accumulation of LC3 on the surface of damaged mitochondria is the first committed step of mitophagy. As detailed in section 1.11.1, damaged mitochondria

207 5. The Vacuolar ATPase and PD are recognised by PINK1, which recruits PARKIN to mediate ubiquitination of mitochondrial membrane proteins (Jin & Youle, 2012). This in turn recruits autophagy adaptor proteins that interact with LC3 on the autophagosome membrane to facilitate sequestration of the damaged mitochondrion within an autophagosome compartment (Chu et al., 2014; Zhu et al., 2011). Therefore the colocalisation of mitochondria and LC3 was measured as a method of assessing whether V-ATPase impairment increases the initiation of mitophagy.

Treatment of SH-SY5Y cells with either BafA1, siRNA-mediated depletion of the ATP6V1A subunit or rotenone, a known mitophagy initiating toxin (Chu et al., 2013), caused a significant increase in the colocalisation of LC3 and the mitochondrial marker TOM20 (Figure 5.25). This result demonstrates that V-ATPase impairment increases the initiation of mitophagy and confirms that impairment of the endolysosomal V- ATPase complex impacts the mitochondria.

208 5. The Vacuolar ATPase and PD

209 5. The Vacuolar ATPase and PD

Figure 5.25 V-ATPase impairment increases initiation of mitophagy. Automated imaging analysis was conducted to assess colocalisation between LC3 and TOM20, relative to total TOM20 area in differentiated SH-SH5Y (described in 2.4.5.1) treated with either (A) depletion of ATP6V1A or (B) BafA1 or rotenone (rot). Statistical significance determined by (A) one-way ANOVA using the method of Benjamini, Kreger and Yekutieli or (B) Students t-test with Welch’s correction. Data from 10 images per treatment taken from each of 3 biological replicates. Error bars represent SEM. (B, D) Representative images from the different treatment groups showing the overlay images obtained from automated image processing, where white represents the area of colocalisation. Scale bar represents 20 μm.

5.2.5.2 V-ATPase impairment does not inhibit autolysosome formation

Despite the evident increase in mitophagy initiation in response to V-ATPase impairment (Figure 5.25), it is possible that the fusion of autophagosomes and lysosomes (to form autolysosomes) may be reduced due to the underlying lysosomal impairment present. The V-ATPase V0 domain is believed to play a role in facilitating fusion of the lysosome and autophagosome membranes (Bayer et al., 2003). Additionally, reduced lysosomal degradation capacity, which is observed in response to V-ATPase depletion (Figure 5.20), can also impair autolysosome formation (Peters et al., 2001). Therefore the increased initiation of mitophagy that results from V-ATPase impairment may result in an accumulation of non-fused autophagosomes. To assess this possibility the mitophagy probe mito-Keima (mKeima) was used to quantify autolysosome formation during mitophagy. mKeima is a genetically encoded fluorophore that is targeted to the mitochondrial matrix and can be used to monitor fusion of autophagsosomes with lysosomes during mitophagy (Katayama et al., 2011). When damaged or defective mitochondria are targeted for mitophagy the mitochondrion, along with the mKeima probe, is enclosed in an autophagosome that subsequently fuses with a lysosome, resulting in exposure of mKeima to the acidic lysosomal pH. The pH change that is associated with the movement of the probe from the neutral pH of the mitochondrion to the acidic pH of the lysosome shifts the excitation wavelength of the probe (Figure 5.26) (Katayama et al., 2011). Therefore the ratio of acidic to neutral pH signal provides an indication of

210 5. The Vacuolar ATPase and PD the proportion of mitochondria that are delivered to the lysosome through the process of mitophagy.

Figure 5.26 Mitophagy can be detected using the coral-derived fluorescent protein Keima. The Keima protein exhibits distinct excitation wavelengths dependent on the pH environment of the probe. When targeted to the mitochondria, the neutral mitochondrial pH results in an excitation maxima at 420 nM. When mitochondria are delivered to the lysosome through mitophagy, the acidic lysosomal pH results in a shift in the excitation maxima to 580 nM.

Treatment of SH-SY5Y cells with rotenone, a known inducer of mitophagy (Chu et al., 2013), resulted in an expected increase in the proportion of mitochondria contained in autolysosomes (Figure 5.27A), confirming the ability of the mKeima probe to monitor mitophagy in SH-SY5Y cells. However, contrary to the above prediction that V-ATPase impairment may impair the fusion of autophagosomes and lysosomes during mitophagy, BafA1 treatment and ATP6V1A depletion both caused a significant increase in the formation of autolysosomes that contained mitochondria (Figure 5.27). This result suggests that V-ATPase impairment increases autophagosome-lysosome fusion, despite the lysosomal impairment that is known to be present.

211 5. The Vacuolar ATPase and PD

Figure 5.27 V-ATPase impairment increases the proportion of mitochondria in autolysosomes. MKeima was used to assess the proportion of mitochondria enclosed within lysosomes in differentiated SH-SY5Y cells treated with BafA1, ATP6V1A depletion or the known mitophagy initiating toxin rotenone. Mitophagy index was calculated as the ratio of the fluorescence intensity of mKeima in neutral environment relative to the acidic environment of the lysosome. Statistical significance determined by (A) one-way ANOVA using the method of Benjamini, Kreger and Yekutieli or (B) Students t-test with Welch’s correction. Error bars represent standard deviation. Data are from 3 biological replicates.

5.2.5.3 V-ATPase depletion induces accumulation of specific mitochondrial proteins

The observed increase in the proportion of mKeima contained within autolysosomes in response to V-ATPase impairment (Figure 5.27) may be an indication of reduced autolysosome degradation capacity during mitophagy. Autolysosomes accumulate when lysosomal proteolytic capacity that is insufficient to degrade the autophagosome cargo (Ganley, 2013; Luzio at al., 2007), which would be an expected consequence of V-ATPase impairment, and accumulation of defective autolysosomes is a prevalent feature in the substantia nigra of PD patients (Anglade et al., 1997).

Accumulation of defective autolysosomes that contain non-degraded mitochondrial components would be expected to lead to increased abundance of mitochondrial proteins within the cell. Therefore to determine if V-ATPase impairment reduces autolysosome degradation capacity during the final stage of mitophagy, the

212 5. The Vacuolar ATPase and PD abundance mitochondrial proteins was assessed in response to V-ATPase impairment. As certain proteins are more resistant to lysosomal degradation than others (Ciechanover, 2005; McGlinchey & Lee, 2015; Parsell & Sauer, 1989) the abundance of 5 different mitochondrial proteins was assessed, with each protein being a subunit of each of the 5 OXPHOS complexes (NDUFB8, SDHB, UQCRC2, COX2 and ATP5A, corresponding to subunits of complex I-V, respectively).

The abundance of each of these mitochondrial proteins was assessed in response to V- ATPase impairment mediated by siRNA depletion of the ATP6V1A or ATP6V1B2 subunit. This method of V-ATPase impairment was selected as V-ATPase depletion was found to cause a substantially greater increase in mitophagy initiation and autolysosome formation than low-dose BafA1 treatment (Figure 5.25, Figure 5.27), suggesting that any accumulation of mitochondrial contents may be more apparent in response to V-ATPase depletion than BafA1 treatment. Additionally, this method of V- ATPase impairment mimics the reduced V-ATPase transcription observed in PD patient brain regions (Figure 5.13).

Depletion of ATP6V1A and ATP6V1B2 had minimal effect on the abundance of subunits belonging to OXPHOS complexes I-IV (NDUFB8, SDHB, UQCRC2 and COX2) (Figure 5.28). Intriguingly however, the abundance of ATP5A, a subunit of the OXPHOS complex V (ATP synthase) F1 domain {Godbout:1997vo}, was significantly and substantially increased by > 4 fold following V-ATPase depletion (Figure 5.28).

The observed increase in the ATP5A protein abundance in response to V-ATPase depletion is intriguing for a number of reasons. Foremost, the increase in ATP5A levels mimics the accumulation of ATP5A that has previously been observed in the frontal cortex of sPD patients (Ferrer et al., 2007), however no explanation for this apparent dysregulation of a specific OXPHOS complex in PD has been proposed. The discovery that ATP5A accumulates in response to V-ATPase depletion in vitro (Figure 5.28) suggests that ATP5A accumulation in the frontal cortex of PD patients may be related to the V-ATPase depletion that is observed in the same brain region (Figure 5.13).

213 5. The Vacuolar ATPase and PD

Figure 5.28 ATP5A protein abundance is significantly increased in response to V-ATPase depletion. (A) The effects of siRNA-mediated depletion of ATP6V1A or ATP6V1B2 on the levels of 5 mitochondrial OXPHOS complex subunits were quantified by Western blot, relative to the level of the loading control βTubulin. Statistical significance determined by two-way ANOVA using the method of Benjamini, Kreger and Yekutieli to correct for multiple comparisons. Data obtained from 4 biological replicates. Error bars represent SEM (*p<0.05, **p<0.01). (B) Representative Western blot image.

There are a number of possible mechanisms through which V-ATPase depletion may result in ATP5A accumulation. Primarily, this result may be an indication of reduced ATP5A degradation during mitophagy, suggesting that ATP5A may be more difficult to degrade and therefore more stable in the lysosome than the other mitochondrial proteins assessed. Alternatively, increased ATP5A levels may be a result of transcriptional up-regulation and, due to the evolutionary and structural relationship between the V-ATPase and mitochondrial ATP synthase (Futai et al., 2012), it cannot be ruled out that a feedback mechanism may exist between the two complexes such that V-ATPase depletion triggers a compensatory up-regulation response that also effects the related ATP synthase subunits. Therefore, to determine whether the increased ATP5A protein abundance observed in response to V-ATPase depletion is a result of transcriptional up-regulation, the ATP5A mRNA level was assessed in response to V-ATPase depletion by qRT-PCR. This analysis identified no significant change in the ATP5A mRNA level following siRNA-mediated depletion of ATP6V1A or

214 5. The Vacuolar ATPase and PD

ATP6V1B2 (Figure 5.29), suggesting that the increase in ATP5A protein abundance is not a result of transcriptional up-regulation and instead supporting the suggestion that ATP5A protein accumulates as a result of impaired degradation.

) 2

o L

P 1

(

o e

i 0

a l

r -1

x R

E -2

-3 NT ATP6V1A ATP6V1B2 Control KD KD

Figure 5.29 ATP5A mRNA levels are not significantly changed by V-ATPase depletion. ATP5A mRNA levels was determined by qRT-PCR relative to expression of the housekeeping gene NFX1. Lack of statistical significance determined by one-way ANOVA. Error bars represent standard deviation.

The results above demonstrate that V-ATPase impairment increases the initiation of mitophagy (Figure 5.25) and in parallel results in an increase in the proportion of mitochondria contained within autolysosomes (Figure 5.27), as well as an evident accumulation of ATP5A protein within the cell (Figure 5.28). These results suggest that while the early stages of mitochondrial turnover are enhanced by V-ATPase impairment, reduced lysosomal degradation capacity may perturb the final stages of mitophagy, leading to an accumulation of defective autolysosomes. Although additional experiments are required to determine why ATP5A accumulates in response to V-ATPase depletion while other OXPHOS components do not, these results confirm that V-ATPase impairment impacts the mitochondria and identify perturbed mitophagy as an additional consequence of V-ATPase dysfunction. Additionally, these results also suggest that V-ATPase depletion could contribute to the observed increased levels of ATP5A protein (Ferrer et al., 2007) and accumulation of

215 5. The Vacuolar ATPase and PD autolysosomes that are observed in the brain regions of PD patients (Anglade et al., 1997).

5.2.6 V-ATPase impairment may perturb lysosomal and mitochondrial functions through a pH independent mechanism

The V-ATPase plays a central role in sensing, establishing and maintaining the acidity of endolysosomal compartments and it is likely that perturbed endolysosomal acidification is the mechanism through which V-ATPase impairment results in the observed lysosomal (Figure 5.20, Figure 5.22) and mitochondrial (Figure 5.23, Figure 5.24) perturbations. To confirm that perturbed lysosomal acidification is the mechanisms through which V-ATPase impairment results in the observed cellular phenotypes, lysosomal pH was assessed in response to both low-dose BafA1 treatment and siRNA mediated V-ATPase depletion.

5.2.6.1 Chronic, low-dose BafA1 and acute, high-dose BafA1 treatments have distinct effects on lysosomal acidification

BafA1 is an inhibitor of the proton translocating V0 domain of the V-ATPase complex (Bowman et al., 1988) that has been shown to impair lysosomal acidification (Christensen et al., 2002; Guha et al., 2014; Yoshimori et al., 1991). However, these studies assessed the effect of acute, high-doses of BafA1 (> 30 nM). Therefore, to confirm that low-dose BafA1 treatment has a similar effect on endolysosomal acidification, the pH of endolysosomal compartments was assessed in response to low- dose BafA1 treatment using the LysoSensorTM blue probe, which accumulates in both late endosomes and lysosomes and exhibits increased fluorescence intensity in more acidic compartments (Dolman et al., 2013).

Surprisingly, and contrary to the prediction that chronic, low-dose BafA1 would mimic the effect of acute, high-dose BafA1 treatment by reducing lysosomal acidity

216 5. The Vacuolar ATPase and PD

(increasing pH), treatment with a chronic (48h) low dose of 1 nM BafA1, the same level used throughout this study, resulted in a non-significant increase in lysosomal acidity (lower pH) (Figure 5.30). Furthermore, chronic 48h treatment with the slightly higher level of 4 nM BafA1 enhanced this lysosomal hyperacidification phenotype to a significant level (Figure 5.30).

Figure 5.30 Inhibition of the V-ATPase using low-doses of BafA1 results in lysosomal hyperacidification. Imaging of LysoSensor blue was used to evaluate lysosomal pH in SH-SY5Y cells treated with 1 nM or 4 nM BafA1 for 48 hours. (A) Representative images of control and

217 5. The Vacuolar ATPase and PD

BafA1 treated cells stained with LysoSensor Blue. (B) Automated imaging analysis of LysoSensor Blue fluorescence intensity corrected for area and background, as detailed in section 2.4.4.1. Data from 30 images per treatment obtained from 3 biological replicates. Scale bar represents 20 μM. Significance determined by one-way ANOVA with two-stage step-up Benjimini, Krieger and Yekutieli’s correction for multiple comparisons (**** p<0.0001). Error bars represent SEM.

To confirm that low (1-4 nM) and high-dose (30 nM) BafA1 have opposing effects on acidification of endolysosomal compartments, and to confirm the reliability of the LysoSensorTM approach used, endolysosomal pH was assessed in response to treatments that have been previously published to reduce endolysosomal acidification. These treatments included an acute, high-dose of BafA1 (30 nM for 2 hours), as well as acute doses of the lysosomal protonophores nigericin (Won et al., 2015) and chloroquine (Homewood et al., 1972). Aligning with previously published results, these treatments resulted in the predicted diminishment of lysosomal acidity (Figure 5.31), demonstrating the reliability of the LysoSensor approach used and confirming a difference between the effects of chronic low-dose and acute high-dose BafA1- mediated V-ATPase impairment.

The possible mechanism through which low-dose BafA1 causes hyperacidification may be related to the recently described pH sensing capacity of the V-ATPase. The V- ATPase plays a dual role in regulating the acidification of endolysosomal compartments through both pumping protons to establish the pH gradient and through sensing endolysosomal acidity, which enables pH to be maintained at an optimal level (Dechant et al., 2010; Hurtado-Lorenzo et al., 2006). The V0 domain, which is inhibited by BafA1, is thought to play a central role in this pH sensing process (Hurtado-Lorenzo et al., 2006) and the observation that endolysosomal compartments become hyperacidified in response to low-doses of BafA1 raises the possibility that low-doses of BafA1 impair the sensing and shut-off capacity of the V-ATPase without perturbing the pumping activity. This could result in protons continuing to be pumped

218 5. The Vacuolar ATPase and PD into the lumen beyond the optimal concentration, leading to the hyperacidification observed.

Figure 5.31 Inhibition of the V-ATPase using a high-dose of BafA1 or either of the protonophores chloroquine or nigericin results in reduced lysosomal acidity. (A) Imaging of LysoSensor blue was used to evaluate lysosomal pH in SH-SY5Y cells treated with 30 nM or 200 uM chloroquine (CQ) or 5 ug/mL nigericin for 2 hours. (A) Representative images of control and BafA1 treated cells stained with LysoSensor Blue. Brightness and exposure have been artificially enhanced equally across all images to improve visibility. (B) Automated imaging

219 5. The Vacuolar ATPase and PD analysis of LysoSensor Blue fluorescence intensity corrected for lysosomal area and background, as detailed in section 2.4.4.1. Data from 10 images per treatment obtained from each of 3 biological replicates. Scale bar represents 20 μm. Significance determined by one- way ANOVA with two-stage step-up Benjimini, Krieger and Yekutieli’s correction for multiple comparisons (p*< 0.05, **** p<0.0001). Error bars represent SEM.

Lysosomal acidity was initially assessed as a possible mechanism through which low- dose BafA1 treatment results in the lysosomal and mitochondrial perturbations observed. Although the seemingly paradoxical increase in lysosomal acidity was unexpected, this result identifies a novel mechanism through which this treatment could induce the cellular dysfunctions observed. Hyperacidification of endolysosomal compartments is damaging (Cooper & Sunderland, 2000) and can trigger lysosomal Ca2+ release through a similar TRPML1 dependent mechanism that is induced by loss of lysosomal acidity (Cao et al., 2017). Therefore hyperacidification is likely to cause perturbed Ca2+ homeostasis, which could in turn lead to the mitochondrial perturbations observed (Figure 5.24, Figure 5.25). Hyperacidification can also perturb lysosomal degradation capacity by promoting fusion of autophagosomes with immature lysosomes (Cao et al., 2017). This subsequently results in accumulation of large autolysosomes that have diminished degradation capacity and the depletion of mature lysosomal compartments within the cell, which could explain the observed decrease in lysosomal degradation of CMA substrates (Figure 5.20), as well as the lysosomal stress implicated by the observed activation of TFEB (Figure 5.22). To determine whether inappropriate lysosomal fusion caused by lysosomal hyperacidification is the likely cause of these observed lysosomal perturbations, the effect of low-dose BafA1 treatment on the lysosomal fusion/fission cycle was assessed. For this purpose, lysosomal size and quantity were evaluated in response to low-dose BafA1 treatment by immunofluorescence imaging of the lysosomal membrane protein LAMP1, with the prediction that inappropriate lysosomal fusion would result larger, and potentially fewer lysosomes. Confirming this prediction, low-dose BafA1 treatment significantly increased lysosomal size by approximately 15% and resulted in a non-significant decline in the number of lysosomal compartments (Figure 5.32), suggesting that lysosomal fusion is induced by low-dose BafA1 treatment. This result

220 5. The Vacuolar ATPase and PD supports the suggestion that the lysosomal hyperacidification induced by low-dose BafA1 treatment has the expected functional effect on inducing lysosomal fusion, which is likely to contribute to perturbed lysosomal degradation capacity that is also observed in response to these treatment. Therefore, although lysosomal hyperacidification was an unexpected effect of low-dose BafA1 treatment, it is likely to be the mechanism through which low-dose BafA1 treatment perturbs lysosomal function.

Figure 5.32 Low dose BafA1 treatment induces lysosomal fusion. Lysosomal size, area and count were quantified in response to 1 nM BafA1 treatment based on immunofluorescence labelling of LAMP1. (A) Representative images indicating the mask image achieved by automated image processing using the method outlined in section 2.4.6.1. (B, C) Data from 10 images per treatment obtained from each of 3 biological replicates was used to determine the average size, area per cell and lysosomal count per cell of LAMP1 positive compartments. Scale bar represents 10 μm. Significance determined by Students t-test with two-stage step-up Benjimini, Krieger and Yekutieli’s correction for multiple comparisons where necessary (q*< 0.05). Error bars represent SEM.

221 5. The Vacuolar ATPase and PD

5.2.6.2 siRNA-mediated depletion of the V-ATPase V1 domain does not impair lysosomal acidification

V-ATPase depletion mediated by siRNA knockdown more accurately reflects the reduced V-ATPase expression observed in patient brain regions and thus is likely to more closely recapitulate the consequences of the V-ATPase depletion observed in patients. Therefore to determine whether V-ATPase depletion perturbs the acidification of endolysosomal compartments, the LysoSensorTM approach used above was employed to assess endolysosomal pH in response to siRNA-mediated depletion of the ATP6V1A and ATP6V1B2 subunits. Although the V1 domain is not directly involved in proton pumping, V1 domain depletion would be expected impair lysosomal acidification as the ATP hydrolysis mediated by the V1 domain is coupled to V0 domain mediated proton translocation (Maxson & Grinstein, 2014). However, while it has been previously published that depletion of subunits of the proton-translocating V0 domain impairs endolysosomal acidification (Mangieri et al., 2014), the effect of depletion of V1 domain subunits has not been studied in mammalian cells.

Surprisingly, depletion of the ATP6V1A and ATP6V1B2 subunits did not alter the pH of endolysosomal compartments (Figure 5.33). Although unexpected, there are a number of possible explanations for this finding. Foremost, this method of assessing lysosomal pH may be insufficiently sensitive to detect the small, predicted pH increase. Similarly, changes in pH in specific subsets of lysosomes, for instance among the perinuclear lysosomes that are known to be more acidic than those close to the cell periphery (Gowrishankar & Ferguson, 2016), may be undetectable. Alternatively, due to the many mechanisms through which V-ATPase activity is regulated, the loss of V1 subunits may not impact lysosomal acidity. In particular, because the V1 domain is present in excess (Kane, 2012), reduced levels of V1 domain subunits may reduce the maximal V-ATPase capacity of the cell without reducing the capacity required in normal conditions, therefore resulting in no change in pH. However the latter explanation would not explain the perturbations in lysosomal and mitochondrial functions that are also observed upon depletion of these subunits.

222 5. The Vacuolar ATPase and PD

Figure 5.33 Depletion of V-ATPase V1 domain subunits does not change endolysosomal acidity. (A) Imaging of LysoSensor blue was used to evaluate lysosomal pH in SH-SY5Y cells treated with a non-targeting (NT) siRNA or siRNA targeted towards ATP6V1A or ATP6V1B2. (A) Representative images of treated cells stained with LysoSensor Blue. (B) Automated imaging analysis of LysoSensor Blue fluorescence intensity corrected for area and background. Data from 10 images per treatment obtained from each of 3 biological replicates. Scale bar represents 20 um. Significance determined by one-way ANOVA with two-stage step-up Benjimini, Krieger and Yekutieli’s correction for multiple comparisons. Error bars represent SEM.

223 5. The Vacuolar ATPase and PD

The acidity of endolysosomal compartments could be maintained in response to V- ATPase depletion by reducing the activity of the lysosomal membrane channels that consume the proton gradient supplied by the V-ATPase. For example, the NHE6 channel is a Na+/H+ antiporter that is highly expressed in the brain (Kondapalli et al., 2014) and has been localised to the membrane of endolysosomal compartments (Brett et al., 2002). Together with the V-ATPase, NHE6 channels regulate lumenal pH by providing a proton leak pathway (Brett et al., 2005; Prasad & Rao, 2015), such that reduced activity of the NHE6 channels would enable pH to be maintained at the expense of impaired lysosomal Na+ uptake. Recent advances in understanding lysosomal pH regulation by ion transporters such as NHE6, have highlighted the complexity of the regulatory mechanisms that converge to maintain lysosomal acidity (Lucien et al., 2017; Prasad & Rao, 2015). The finding that non-toxic levels of V-ATPase depletion do not impair lysosomal acidification suggests that reduced activity of lysosomal membrane transporters that consume the proton gradient may compensate for impaired V-ATPase activity, enabling endolysosomal acidity to be maintained in spite of the V-ATPase depletion present, but likely perturbing other endolysosomal functions, such as ion storage. Similarly, in the case of synaptic vesicles, reduced consumption of the V-ATPase-provided proton gradient could impair the loading of neurotransmitter precursors into these vesicles, a process that is also dependent on proton exchange (Gasnier, 2000) and has additionally been implicated as a toxic mechanism in PD (Goldstein et al., 2013). This suggests that compensation for perturbed V-ATPase pumping capacity has the potential to perturb multiple aspects of cell biology.

5.2.7 Depletion of V-ATPase V1 domain subunits may impair endolysosomal transport and maturation

Following the observation that the depletion of the V-ATPase V1 domain subunits ATP6V1A and ATP6V1B2 does not cause an apparent loss of endolysosomal acidity

224 5. The Vacuolar ATPase and PD

(Figure 5.33), alternative mechanisms through which depletion of these subunits could cause the lysosomal and mitochondrial phenotypes observed were assessed. A clue as to a possible pH independent mechanism through which V-ATPase depletion could result in these phenotypes was obtained when assessing the effect of ATP6V1A and ATP6V1B2 depletion on the size and subcellular distribution of lysosomal compartments.

Immunofluorescent labelling of LAMP1 was used to visualize lysosomal compartments in cells depleted of the V-ATPase subunits ATP6V1A and ATP6V1B2 (Figure 5.34A&B). Assessment of the abundance and size of LAMP1 positive compartments indicated that while depletion of ATP6V1A or ATP6V1B2 had no affect on the total LAMP1 content per cell (Figure 5.34C), these treatments resulted in an increase in the quantity of LAMP1 compartments (Figure 5.34D), while individual LAMP1 positive compartments appeared significantly smaller (Figure 5.34E). However, closer examination of these images suggested that this apparent reduction in lysosomal size in V-ATPase depleted cells was caused by increased dispersal of endolysosomal compartments. In V-ATPase depleted cells lysosomes were more predominantly localised to the periphery of the cell, whereas control cells contained substantially more clusters of lysosomes in the perinuclear region. These dense perinuclear clusters of lysosomes in control cells could not be resolved into individual compartments by confocal microscopy, resulting in apparently larger lysosomal size in control cells, relative to V-ATPase depleted cells (Figure 5.34E).

225 5. The Vacuolar ATPase and PD

226 5. The Vacuolar ATPase and PD

Figure 5.34 Depletion of V-ATPase V1 domain subunits causes an apparent decrease in lysosomal size. Representative images of Lamp1/DAPI stained cells at 1X (A) and 1.5X (B) indicating the mask image generated following image filtering as described in section 2.4.6.1. Scale bars represent 10 μm. Total lysosomal area (C), lysoaomal count (D) and lysosomal size (E) were calculated based on the unfiltered dataset. Statistical significant determined by One- way ANOVA. (p*<0.05, **<0.01). Error bars represent SEM. Data obtained from 5 images from each of 4 biological replicates.

To explore the possibility that V-ATPase depletion increases the spatial distribution of lysosomal throughout the cell, the proportion of LAMP1 positive compartments within a 2 μm distance of the nucleus, as determined by DAPI staining, was quantified (Figure 5.35A). This distance from DAPI was selected as it included approximately 50% of the cell lysosomes. Confirming the observation that depletion of V1 domain subunits impairs perinuclear clustering of lysosomes, the percentage of endolysosomal compartments in the perinuclear region was significantly reduced following depletion of ATP6V1A or ATP6V1B2 (Figure 5.35B). The effect of low-dose BafA1 on lysosomal clustering was also assessed and was found to cause no change in the proportion of lysosomes within the perinuclear region (Figure 5.35C), indicating that perturbed perinuclear clustering of lysosomes is caused by depletion of V1 domain subunits, but not impairment of the V0 domain of the complex.

227 5. The Vacuolar ATPase and PD

Figure 5.35 Perinuclear clustering of endolysosomal compartments is reduced by depletion of V-ATPase V1 domain subunits but not BafA1 treatment. (A) Representative images of LAMP1 (red) and DAPI (blue) stained cells indicating the perinuclear perimeter at a distance of 5 μM from the DAPI signal. Scale bars represent 5 μm. (B, C) The proportion of lysosomes within a 5 μM distance of the nucleus was quantified using FIJI (ImageJ) as per the method outlined in section 2.4.6.2. Statistical significance was determined by (B) One-way ANOVA (p*<0.05, ** p<0.01) or (C) Student’s t-test. Data obtained from 20 images from 4 biological replicates per treatment. Error bars represent SEM.

Lysosomal compartments are actively transported from the cell periphery into the perinuclear region of the cell along microtubule tracks, which brings them into close

228 5. The Vacuolar ATPase and PD proximity with autophagosomes that are also transported there (Huotari & Helenius, 2011). The discovery that V-ATPase depletion results in reduced perinuclear clustering of lysosomes suggests that V-ATPase depletion impairs this transport of lysosomes and is consistent with the previous finding that the V-ATPase V1 domain indirectly interacts with microtubule motors (De Luca et al., 2014). As discussed in section 5.1.2.1, the V- ATPase interacts with microtubules through RILP (De Luca et al., 2014), a cytoplasmic protein that is acquired by maturing endosomes and regulates endosomal transport (Huotari & Helenius, 2011; Johansson et al., 2007). RILP tethers lysosomal compartments to a set of dynein motors that regulate lysosomal transport along microtubule tracks to the microtubule organising centre (MTOC) in the perinuclear region of the cell (De Luca et al., 2014; Hyttinen et al., 2013; Jordens et al., 2001; Progida et al., 2007). This transport of lysosomes is coupled to their increasing acidity (Huotari & Helenius, 2011) and additionally increases the efficiency with which lysosomes are supplied with the necessary hydrolytic enzymes from the TGN (Johnson et al., 2016; Scheel et al., 1990; Scott et al., 2014) such that peripheral lysosomes are both less acidic and less active than their perinuclear counterparts due to reduced acidification and a reduced supply of lysosomal hydrolytic enzymes (Johnson et al., 2016). Therefore perturbed lysosomal transport has the potential to severely impact the function and maturation of lysosomal compartments.

The observation that LAMP1 compartments are more peripheral in V-ATPase depleted cells suggests that V1 domain depletion impairs microtubule-mediated endosomal transport, likely due to reduced V-ATPase/RILP interaction. Furthermore, the observation of reduced perinuclear clustering of lysosomes also suggests that lysosomal degradation capacity is likely to be reduced due to both impaired delivery of hydrolytic enzymes transported from the TGN, and the presence of fewer lysosomes at the MTOC, which is a major site of autophagy within the cell (Jahreiss et al., 2008). Therefore, V-ATPase depletion may result in the observed lysosomal dysfunctions (Figure 5.20, Figure 5.22) by reducing the V-ATPase-microtubule interaction that facilitates both endolysosomal transport towards the MTOC and the accompanying lysosomal maturation.

229 5. The Vacuolar ATPase and PD

From the results presented here it is not possible to determine whether the combination of reduced expression of V-ATPase subunits from both the V0 and V1 domain, as is observed in sPD patient brain tissue (Figure 5.13), might result in a lysosomal acidification defect, however these results highlight that, in addition to a possible lysosomal acidification defect, the reduced level of V-ATPase expression present in PD patients may also impair the microtubule dependent transportation of endolysosomal compartments. The role of microtubule dependent transport of organelles, such as lysosomes and mitochondria, is an underexplored area of PD. Although to date no studies have addressed transportation defects or assessed the distribution of endolysosomal compartments in affected brain regions of PD patients, there is significant evidence from model systems that impaired microtubule dependent transport of endosomes may play a role in the disease. For instance, application of pre-formed αSyn fibrils, a commonly used model of αSyn accumulation, result in perturbed axonal transport of endosomes in primary neurons (Volpicelli-Daley et al., 2014) and the organisation and stability of the microtubule cytoskeleton is perturbed in fibroblast cultures derived from sporadic and PARKIN or LRRK2 familial PD patients (Cartelli et al., 2012). Additionally, perturbed endolysosomal transport is evident and implicated as a pathogenic mechanism in Alzheimer’s disease (Cataldo et al., 1991; Gowrishankar et al., 2015), a neurodegenerative condition that is pathologically similar to PD, and abnormal distribution of key lysosomal cathepsin enzymes, which is indicative of perturbed lysosomal transport, is also associated with ageing (Sato et al., 2006), which is the primary risk factor for PD. Taken together with the observation that V-ATPase depletion impair perinuclear accumulation of lysosomal (Figure 5.35), these results strongly suggest that the decreased V-ATPase levels apparent in patient brain regions (Figure 5.13, Figure 5.14) may impair endolysosomal transport in PD, highlighting this as a key area for further research.

230 5. The Vacuolar ATPase and PD

5.2.8 A relationship between the V-ATPase and αSynuclein may contribute to PD pathology

Initial assessment of V-ATPase expression in patient brain regions (section 5.2.1) demonstrated that there is an inverse relationship between the levels of αSyn and a V- ATPase subunit protein in patients but not in controls (Figure 5.14), with increased αSyn levels significantly correlating with reduced levels of the V-ATPase subunit ATP6V1B2. To determine if there is a causative basis to this relationship two hypotheses were assessed in vitro as potential explanations for this observation. The first hypothesis suggests that accumulating αSyn levels cause a reduction in V-ATPase expression (Figure 5.36; hypothesis 1). Alternatively, the second hypothesis suggests that reduced V-ATPase expression is the causative factor in the relationship, with reduced V-ATPase levels contributing to accumulation of αSyn, suggesting that V- ATPase depletion may be a preceding factor in the development of αSyn pathology in PD (Figure 5.36; hypothesis 2).

Hypothesis 1: αSynuclein V-ATPase

Hypothesis 2: V-ATPase  αSynuclein 

Figure 5.36 The correlation between increased αSyn and decreased V-ATPase protein abundance observed in PD patient brain tissue could be a result of αSynuclein accumulation leading to reduced V-ATPase levels (hypothesis 1) or V-ATPase depletion leading to αSyn accumulation (hypothesis 2).

5.2.8.1 αSynuclein overexpression increases expression of V-ATPase subunit mRNA and protein levels

Accumulation of αSyn protein is a hallmark feature of PD that is considered by many to be a primary cause of neuronal degeneration in the disease. To determine if the inverse relationship observed between V-ATPase and αSyn protein levels in PD patient

231 5. The Vacuolar ATPase and PD brain tissue could be caused by αSyn accumulation (Figure 5.36; hypothesis 1), the protein level of the V-ATPase subunit, ATP6V1B2, and the mRNA levels of several key V-ATPase subunits were assessed in response to elevated αSyn expression. For this purpose the DOXoff-αSyn stable SH-SY5Y cell line described in section 3 was used to increase αSyn levels in a DOX dependent manner.

This assessment revealed that αSyn overexpression does not cause a reduction in either V-ATPase protein or RNA levels (Figure 5.37A,C), and thus does not support the proposed hypothesis 1 (Figure 5.36). Instead, elevated αSyn expression caused a significant increase in ATP6V1B2 protein abundance (Figure 5.37A) as well as significant increases in the mRNA levels of the assessed V-ATPase subunits (ATP6V0A, ATP6V1A and ATP6V1B2) (Figure 5.37C). The mRNA levels of these subunits were unaltered in control βGal overexpressing cells [DOXoff-βGal] (Figure 5.37D), confirming that this increase in V-ATPase levels is caused by elevated αSyn expression, and not due to an effect of the reduction in DOX concentration used to induce expression of the transgene.

232 5. The Vacuolar ATPase and PD

Figure 5.37 αSyn overexpression results in increased ATP6V1B2 protein levels and increased expression of V-ATPase subunits. (A) ATP6V1B2 and αSyn protein levels were assessed by western blot in DOXoff-αSyn SH-SY5Y cells relative to the abundance of the loading control βTubulin. Data is from 4 biological replicates (B) Representative western blot image. (C) Levels of ATP6V1A, ATP6V1B2, ATP6V0A1 and αSyn mRNA were assessed by qRT- PCR in response to elevated αSyn expression [DOXoff-αSyn] or (D) elevated βGal expression [DOXoff- βGal cells]. Data is from 3 biological replicates. Statistical significance determined by Students t-tests corrected for multiple comparisons using the method of Benjamini, Kreger and Yekutieli. Error bars represent standard deviation.

5.2.8.2 Elevated αSyn expression induces transcription of TFEB target genes

Transcription of the V-ATPase genes is regulated by the transcription factor TFEB, which is considered to be a master-regulator of lysosomal biogenesis (Napolitano &

233 5. The Vacuolar ATPase and PD

Ballabio, 2016), as discussed in section 5.1.3.5. TFEB up-regulates the expression of a subset of CLEAR genes, which include the V-ATPase subunits, that are required to maintain lysosomal function in response to lysosomal stress or increased lysosomal degradation requirement (Sardiello et al., 2009). Therefore, the observation that elevated αSyn expression results in up-regulation of V-ATPase genes suggests that αSyn causes lysosomal stress, which in turn induces TFEB activation. This suggestion corresponds with previous indications that elevated αSyn expression reduces lysosomal degradation capacity (Mazzulli et al., 2016) and induces lysosomal dysfunction (Koch et al., 2015; Winslow et al., 2010).

To confirm that the αSyn-induced up-regulation of V-ATPase subunits is a result of TFEB activation, the effect of αSyn overexpression on the levels of other TFEB target genes was assessed. These results suggest that in addition to increasing the expression of V-ATPase subunit genes, elevated αSyn expression may also increase the transcript levels of other TFEB target genes, including cathepsin D (CTSD) and tripeptidyl peptisase (TPP) (Figure 5.38A), however the increase in the levels of these CLEAR genes did reach the significance level that was observed in the case of the V-ATPase subunits assessed. The increase in TFEB targets was not mimicked in DOXoff-βGal cells (Figure 5.38B), confirming that increased TFEB target gene expression is αSyn- dependent.

234 5. The Vacuolar ATPase and PD

Figure 5.38 αSyn overexpression results in increased expression of TFEB target genes. (A) Levels of the TFEB target genes GNS (glucosamine sulphatase), CTSD (cathepsin D) and TPP (tripeptidyl peptisase) were assessed by qRT-PCR in response to elevated αSyn expression or (D) elevated βGal expression relative to the expression of the housekeeping gene B2M. Statistical significance determined by Students t-tests corrected for multiple comparisons using the method of Benjamini, Kreger and Yekutieli. Error bars represent standard deviation. Data is from 3 biological replicates.

This result suggests that elevated αSyn expression results in lysosomal stress that triggers TFEB activation, causing up-regulation of CLEAR genes, including the V-ATPase subunits. Although this data presents an interesting association between αSyn, lysosomal stress and TFEB transcriptional regulation, this in vitro result contradicts the observation that increased αSyn levels are correlated with reduced V-ATPase levels in patient tissue (Figure 5.14). This suggests that elevated αSyn levels are not responsible for the decreased V-ATPase levels observed in patients and indicates that the alternative hypothesis (Figure 5.36; hypothesis 2), that reduced V-ATPase levels contribute to elevated αSyn levels, may be the basis for the observed inverse relationship.

5.2.8.3 V-ATPase depletion results in increased αSyn protein abundance, phenocopying the relationships observed in PD patients

To assess the hypothesis that elevated αSyn levels result from V-ATPase depletion (Figure 5.36; hypothesis 2), αSyn protein abundance and mRNA levels were assessed in response to siRNA-mediated V-ATPase depletion, as well as BafA1-mediated V-ATPase impairment. According to this hypothesis, reduced V-ATPase expression would be predicted to increase αSyn protein levels, potentially by impairing the lysosomal degradation of αSyn. αSyn can be degraded in the lysosome through CMA (Cuervo et al., 2004; Vogiatzi et al., 2008). Therefore an accumulation of αSyn in response to V- ATPase depletion would mirror the observed accumulation of other CMA substrate proteins in response to this treatment (Figure 5.20).

235 5. The Vacuolar ATPase and PD

Supporting the proposed hypothesis 2, siRNA-mediated depletion of the V-ATPase subunits ATP6V1A and ATP6V1B2 in SH-SY5Y cells was found to significantly increase αSyn protein abundance (Figure 5.39A). The finding that reducing V-ATPase levels in vitro triggers a response that mimics the relationship observed in patient brain tissue supports the proposal that V-ATPase down-regulation is a driving factor in this relationship. However, somewhat surprisingly, BafA1 mediated impairment of the V- ATPase activity had no effect on αSyn protein levels (Figure 5.39B).

Figure 5.39 siRNA-mediated V-ATPase depletion results in αSyn protein accumulation. αSyn protein abundance was assessed in response to (A) siRNA mediated V-ATPase depletion or (B) low-dose BafA1 treatment by Western blot, relative to the abundance of the loading control 14-3-3. Representative western blot images are shown. Statistical significance determined by (A) one-way ANOVA or (B) Students t-test. Error bars represent standard deviation. Data are from 4 biological replicates.

236 5. The Vacuolar ATPase and PD

5.2.8.4 A proposed mechanism through which V-ATPase depletion may result in increased αSyn protein siRNA-mediated V-ATPase depletion resulted in accumulation of αSyn protein (Figure 5.39A), mirroring the relationship observed in patients and suggesting that this treatment may perturb lysosomal degradation of αSyn. However, low-dose BafA1 mediated V-ATPase impairment had no effect on αSyn protein levels (Figure 5.39B). Understanding the mechanism through which αSyn is degraded may reveal the basis of the different effects of these two methods of V-ATPase impairment.

αSyn is degraded in the lysosome both through CMA (Cuervo et al., 2004) and macroautophagy (Vogiatzi et al., 2008), and cathepsin D (cathD) is the main lysosomal protease involved in αSyn degradation (Cullen et al., 2009; Sevlever et al., 2008). Consequently, impaired cathD activity results in αSyn accumulation (Sevlever et al., 2008), generation of toxic αSyn fibrils (Tsujimura et al., 2015) and enhanced cell-to-cell transmission of αSyn (Bae et al., 2015). Additionally, reduced cathD levels have been observed in the heavily pathologically affected substantia nigra of PD patients (Chu et al., 2009), further implicating reduced cathD activity as a factor that may contribute to αSyn accumulation in PD.

CathD is synthesised in the rough ER as a zymogen (pro-enzyme) and transported to the TGN, where it is recognised by mannose-6-phosphate or sortillin receptors for subsequent trafficking to the lysosome (Canuel et al., 2008). Once trafficked to the lysosome, pro-CathD is proteolytically processed by other proteases (cathepsin B and L) (Laurent-Matha et al., 2006) to yield mature and active cathD (Richo & Conner, 1991). Due to this dependence of cathD on the presence of other lysosomal proteases for processing, perturbations in lysosomal maturation would be expected to have a more profound effect on cathD than other lysosomal enzymes. This suggests that the defect in lysosomal transport that is associated with V-ATPase depletion, but not BafA1 treatment (Figure 5.34), may additionally reduce the trafficking and/or activation of cathD.

237 5. The Vacuolar ATPase and PD

Supporting the suggestion that impaired lysosomal transport may reduce cathD activity, perturbed lysosomal transport to the perinuclear region is observed to impair the delivery of enzymes from the TGN to the lysosome (Johnson et al., 2016) and, in particular, cathepsin D levels are observed to be reduced in peripheral lysosomes (Gowrishankar et al., 2015). This proposed association between reduced V-ATPase levels, perturbed lysosomal transport and perturbed cathD maturation is currently under investigation, with the prediction that V-ATPase depletion will result in accumulation of pro-cathD and diminished levels of mature cathD, but likely not change the total cellular level of CathD, while BafA1 will have no effect on cathD processing.

Abnormal trafficking of enzymes to the endolysosomal network has been implicated as a pathogenic mechanism in PD and other neurodegenerative conditions (reviewed by (Wang et al., 2014b)). In particular and supporting the suggestion that impaired cathD trafficking and processing may be an underlying pathogenic mechanism in PD, disruption of cathD processing is implicated as a pathogenic mechanism in VPS35 (Miura et al., 2014) and ATP13A2 (Matsui et al., 2013) models of PD, and is also evident in rodent neurons in response to ageing (Y. Sato et al., 2006), which is the primary risk factor for the disease. Confirmation that V-ATPase depletion also contribute to perturbed cathD trafficking would significantly contribute to knowledge of the enzyme trafficking pathway and support the proposal that V-ATPase depletion perturbs cellular functions by impairing lysosomal transport and maturation.

5.2.9 V-ATPase depletion causes a reduction in αSynuclein mRNA levels and reveals a novel relationship between αSyn and V-ATPase transcript levels in the brain

The above results demonstrate that V-ATPase depletion causes an increase in αSyn protein levels. As described above, this increase is likely due to reduced lysosomal degradation, however increased αSyn transcription could also contribute to this

238 5. The Vacuolar ATPase and PD phenotype. To address this possibility the αSyn mRNA level was assessed in V-ATPase depleted cells.

Depletion of the V-ATPase ATP6V1A and ATP6V1B2 subunits did not result in an increase in αSyn mRNA levels (Figure 5.40A). Surprisingly however, V-ATPase depletion was observed to cause a significant > 50% decrease in αSyn transcription, suggesting that a feedback mechanism may exist that results in transcriptional down-regulation of αSyn in response to V-ATPase depletion.

Figure 5.40 V-ATPase depletion significantly reduces αSyn mRNA levels. αSyn mRNA levels were assessed by qPCR in response to depletion of ATP6V1A or ATP6V1B2. Log2 fold change was calculated relative to the expression of the housekeeping gene NFX1. Statistical analysis as per (A) two-way ANOVA corrected for multiple comparisons using the method of Benjamini, Kreger and Yekutieli. Error bars represent standard deviation.

This identification of this relationship between reduced V-ATPase and reduced αSyn transcript levels in vitro raised the question of whether a similar relationship is evident in patient brain tissue. Although increased αSyn protein levels are a pathological hallmark of PD (Spillantini et al., 1997), the level of αSyn mRNA in brain regions of sporadic PD patients remains controversial. For instance, in the heavily pathologically affected substantia nigra of sporadic patients, the αSyn mRNA level was initially reported to be decreased (Kingsbury et al., 2004), however subsequent studies have reported the level to be increased (Chiba-Falek et al., 2006) or unchanged (Fuchs et al., 2008). Therefore, to better characterise αSyn mRNA levels in patient brain regions that

239 5. The Vacuolar ATPase and PD display differing levels of pathological damage (BF > SFC; SOC not pathologically affected) and to explore whether the transcriptional relationship observed between V- ATPase and αSyn mRNA levels in vitro is mirrored in patients, αSyn mRNA levels were assessed using the RNAseq transcriptomic data described in section 5.2.1.

Surprisingly, a significant reduction in the αSyn mRNA level was observed in the brain regions of sporadic PD patients relative to controls (Figure 5.41A). This transcriptomic reduction was evident in all 3 brain regions assessed, including the BF, a brain region that exhibited increased αSyn protein levels at the time of death of these patients. This decrease in αSyn mRNA levels was further validated by qRT-PCR of αSyn mRNA levels in the SFC (Figure 5.41B).

Figure 5.41 αSyn mRNA is reduced in PD patient brain regions and correlates with the mRNA levels of V-ATPase transcripts. (A) Assessment of αSyn mRNA levels in 3 brain regions of PD patients and controls based on RNAseq analysis. Statistical significance determined by Students t-test corrected for multiple comparisons using the method of Benjamini, Kreger and Yekutieli (q*<0.05, **<0.01) (B) Assessment of the αSyn mRNA level in the superior frontal cortex by qRT-PCR relative to the expression of the housekeeping gene B2M. Statistical significance determined by Students t-test (p*<0.05). Error bars represent standard deviation. Data obtained from 10 patients and 10 controls.

The finding that αSyn mRNA levels are reduced in PD patient brain regions aligns with the discovery that V-ATPase subunit expression is also reduced in these brain regions (Figure 5.13) and thus suggests that there may be a positive correlation between

240 5. The Vacuolar ATPase and PD reduced αSyn and reduced V-ATPase mRNA levels in the human brain. Such a relationship would mimic the relationship identified between the levels of these transcripts in vitro following V-ATPase depletion (Figure 5.40). To explore this possibility, correlation analysis of αSyn mRNA levels relative to the mRNA levels of the 13 V-ATPase subunit genes was conducted using RNAseq data. This analysis identified multiple significant positive correlations between the expression of many V-ATPase subunit genes and αSyn in the SFC, SOC and BF (Table 5.5) and this positive correlation was evident both in patients and controls (Figure 5.42).

Table 5.5 αSyn-V-ATPase correlation analysis based on RNAseq data obtained from the basal forebrain (BF), superior frontal cortex (SFC) and superior occipital cortex (SOC) of 10 patients and 10 controls. P value and r value determined by Pearson’s correlation analysis and statistical significance determined using Bonferroni’s correction for multiple comparisons (p*<0.005 **<0.0005, ***<0.0001).

BF SFC SOC r value p value r value p value r value p value <0.000 ATP6V0A1 0.6702 0.0012 * 0.8954 <0.0001 *** 0.8876 1 *** ATP6V0B 0.4352 0.0552 ns 0.6752 0.0011 * 0.4204 0.065 ns <0.000 ATP6V0C 0.6259 0.0032 * 0.8787 <0.0001 *** 0.7794 1 *** ATP6V0D1 0.658 0.0016 * 0.8573 <0.0001 *** 0.7202 0.0003 ** <0.000 ATP6V0E2 0.7525 0.0001 ** 0.7775 <0.0001 *** 0.8219 1 *** <0.000 <0.000 ATP6V1A 0.8108 1 *** 0.9504 <0.0001 *** 0.9276 1 *** <0.000 ATP6V1B2 0.7038 0.0005 * 0.9157 <0.0001 *** 0.8184 1 *** <0.000 <0.000 ATP6V1C1 0.8847 1 *** 0.922 <0.0001 *** 0.8646 1 *** <0.000 ATP6V1D 0.7073 0.0005 * 0.9006 <0.0001 *** 0.8795 1 *** <0.000 ATP6V1E1 0.7324 0.0002 ** 0.8792 <0.0001 *** 0.8678 1 *** ATP6V1F 0.3628 0.1159 ns 0.7008 0.0006 * 0.3687 0.1097 ns ATP6V1G1 0.01988 0.9337 ns 0.6883 0.0008 * 0.2933 0.2095 ns <0.000 <0.000 ATP6V1H 0.8299 1 *** 0.8681 <0.0001 *** 0.8546 1 ***

241 5. The Vacuolar ATPase and PD

Figure 5.42 ATP6V1A mRNA levels are positively correlated with αSyn mRNA levels in brain regions of PD patients and controls. Linear regression and Pearson’s correlation analysis of αSyn and ATP6V1A mRNA levels, based on RNAseq data, in 3 brain regions of 10 PD patients and 10 controls. Pearson’s r value and p value are displayed in the figure.

The discovery of this novel relationship between V-ATPase and αSyn expression levels, together with the previous identification of interaction between subunits of the V- ATPase (ATP6V1A, ATP6V1B2, ATP6V0D1 and ATP6V0A) and αSyn (Table 5.3) suggests that these proteins may function in overlapping pathways. Both the V-ATPase and αSyn localise to synaptic terminals and appear to play a role in regulating synaptic vesicles. The V-ATPase indirectly regulates synaptic vesicle loading (section 5.1.2.5)

242 5. The Vacuolar ATPase and PD and additionally plays a role in facilitating fusion of synaptic vesicles with the synaptic membrane (Vavassori & Mayer, 2014; Wang et al., 2014a). αSyn is also proposed to play a role in regulating the synaptic vesicle pool (section 1.8) (Burré, 2015; Burré et al., 2010), however the exact role of αSyn at the synaptic terminal remains undetermined. Characterisation of the V-ATPase-αSyn relationship may therefore provide further insight into the functions of the V-ATPase and αSyn in neurobiology.

The discovery that V-ATPase depletion both increased αSyn protein levels (Figure 5.39) and decreased αSyn mRNA levels (Figure 5.40) raises a number of additional questions regarding the regulation of αSyn expression. While the increased protein level is likely caused by reduced degradation of αSyn, the molecular basis of the decrease in αSyn mRNA level is not as forthcoming. A similar reduction in αSyn mRNA has been observed in PD patient blood (Locascio et al., 2015) and this result was attributed to reduced activation of αSyn expression by the transcription factor GATA1, which co- regulates expression of αSyn as well as genes involved in biosynthesis of the Fe2+ cofactor, heme (Scherzer et al., 2008). While GATA1 is not expressed in the brain, closely related GATA2 is highly expressed in the dopaminergic neurons that degenerate in PD (Scherzer et al., 2008) and is observed to regulate αSyn expression in mammalian cells (Scherzer et al., 2008). GATA2 is also observed to negatively regulate the expression of the mitochondrial Fe2+ storage protein ferritin (FtMt) in tissues with high metabolic activity (Guaraldo et al., 2016), such as the brain. These data suggest that GATA2 may mimic the role of GATA1 in co-regulating αSyn in an Fe2+ dependent manner. Intriguingly, the lysosome is a major Fe2+ storage site (Terman & Kurz, 2013) and V-ATPase impairment is observed to remodel cellular Fe2+ stores, reducing lysosomal Fe2+ levels and increasing mitochondrial Fe2+ levels (Uchiyama et al., 2008). This raises the possibility that Fe2+ released from the lysosome upon V-ATPase depletion may increase the requirement for mitochondrial Fe2+ storage proteins; such as FtMt. Therefore, V-ATPase depletion may result in reduced GATA2 activation to facilitate up-regulation of FtMt, consequently additionally reduce αSyn transcription.

243 5. The Vacuolar ATPase and PD

Fe2+ dyshomeostasis is a well-described feature of PD (reviewed in Rhodes & Ritz, 2008), with significantly elevated Fe2+ levels evident in the substantia nigra of PD patients (Dexter et al., 1987; Michaeli et al., 2007; Mochizuki & Yasuda, 2012) and the extent of Fe2+ accumulation in this brain region associated with disease severity (Ryvlin et al., 1995). However the cause of the elevated Fe2+ levels in PD remains unknown. Investigation of this proposed V-ATPase-aSyn-Fe2+ relationship may therefore provide significant insight into this elusive aspect of PD.

5.3 Summary and future directions

The molecular mechanisms that underpin Parkinson’s disease (PD) remain controversial. This study identified reduced mRNA levels of subunits of the endolysosomal V-ATPase in multiple brain regions of sPD patients (Figure 5.13), as well as reduced protein abundance of the ATP6V1B2 subunit (Figure 5.14). These data suggest that V-ATPase dysfunction may contribute to cell dysfunction in sPD.

5.3.1 siRNA-mediated V-ATPase depletion and BafA1-mediated V-ATPase impairment have different effects on cellular functions

To explore the possibility that V-ATPase depletion is a contributing factor to PD, the effects of V-ATPase impairment on known PD-related cellular phenotypes were assessed in vitro. Two methods of V-ATPase impairment were assessed; siRNA- mediated depletion of two distinct V-ATPase subunits and chemical impairment of the V-ATPase using chronic, low-dose BafA1 treatment. Intriguingly, it was noted that these two methods of V-ATPase impairment did not phenocopy each other in all instances. As summarised in Table 5.6, although BafA1 and V-ATPase depletion both impaired lysosomal function (Figure 5.20, Figure 5.22), increased mitochondrial

244 5. The Vacuolar ATPase and PD damage (Figure 5.23, Figure 5.24) and perturbed mitophagy (Figure 5.25), only V- ATPase depletion reproduced the accumulation of αSyn protein (Figure 5.39) that is observed in PD patients. Additionally, the two methods of V-ATPase impairment had differing effects on endolysosomal pH, with BafA1 treatment resulting in hyperacidification (Figure 5.30), while V-ATPase depletion caused no detectable change in pH (Figure 5.33). In agreement with this observation, the increase in lysosomal size that is expected to accompany an increase in lysosomal acidity (Cao et al., 2017) was observed in BafA1 treated cells (Figure 5.32), but not in V-ATPase depleted cells (Figure 5.34). Alternatively, the reduced perinuclear clustering of lysosomes observed upon siRNA-mediated V-ATPase depletion was not observed in BafA1 treated cells (Figure 5.35).

Table 5.6 Summary of the effects of V-ATPase impairment mediated by either siRNA- mediated depletion or BafA1-mediated inhibition on aspects of cell biology.

Category Cellular Phenotype V-ATPase KD 1 nM BafA1 Non- CMA substrate significant accumulation increase

Increased TFEB activation Lysosomal Hyper- Lysosomal pH No change dysfunction acidification Increased lysosomal Size No change Reduced Lysosomal distribution perinuclear No change clustering

Increased mitochondrial ROS

Mitochondrial Increased mitochondrial impairment Ca2+

Accumulation of ATP5A Not assessed protein

Increased mitophagy Perturbed initiation (Tom20-LC3 mitophagy colocalisation)

Increased mitophagy Index

245 5. The Vacuolar ATPase and PD

(mKeima)

αSyn protein accumulation No change αSynuclein αSyn mRNA reduction Not assessed

The differing effects of the two methods of V-ATPase impairment are likely due to the different mechanisms through which they impair V-ATPase activity. While the canonical proton-pumping activity of the V-ATPase is expected to be inhibited by both BafA1 and siRNA-mediated V-ATPase depletion, V-ATPase depletion may additionally perturb other non-canonical functions of the V-ATPase, such as the structural functions that the protein complex plays (reviewed in section 5.1.2). Consequently, perturbations that were observed upon both BafA1 treatment and V-ATPase depletion are likely to be a result of impaired proton translocation. Perturbations that were evident upon V-ATPase depletion but not BafA1 impairment are therefore likely to be dependent on structural functions of the V-ATPase complex. For instance, the reduced perinuclear lysosomal clustering that is observed upon siRNA-mediated V-ATPase depletion is likely to be dependent on the structural, tethering role that the V-ATPase has been suggested to play in mediating transport of endolysosomes along microtubule tracks (section 5.1.2.1). Consistent with BafA1 not perturbing the structure of the V-ATPase, reduced perinuclear clustering of lysosomes was not observed in response to BafA1 treatment.

Impaired structural functions of the V-ATPase and the consequent reduction in perinuclear lysosomal clustering may also underlie the observation that while V- ATPase depletion caused an increase in αSyn protein levels, BafA1 treatment did not (Figure 5.34). As discussed in section 5.2.8.4, reduced transport of lysosomes into the perinuclear region has previously been associated with impaired delivery of lysosomal hydrolases, such as cathD, the hydrolase known to degrade αSyn (Cullen et al., 2009; Sevlever et al., 2008). This suggests that impaired lysosomal transport caused by V- ATPase depletion could be an underlying mechanism that contributes to the accumulation of αSyn. Perturbed spatial distribution of lysosomes is associated with other neurodegenerative diseases (Gowrishankar et al., 2015) and ageing (Sato et al.,

246 5. The Vacuolar ATPase and PD

2006), which is the primary risk factor for PD, however this cellular phenotype has not previously been assessed in affected neurons of PD patients. This is therefore a continuing area of study that will provide significant insight into this novel consequence of V-ATPase dysfunction and its implications in PD.

5.3.2 siRNA-mediated V-ATPase depletion recapitulates many established PD phenotypes and identifies potential novel aspects of the disease

As discussed above, siRNA-mediated V-ATPase depletion impairs both the proton- pumping and structural functions of the complex. This method of V-ATPase impairment recapitulates the reduction of V-ATPase mRNA levels observed in PD patient brain regions and provides a more complete representation of the V-ATPase depletion evident in patient brain tissue than BafA1 treatment. V-ATPase depletion recapitulated many established PD cellular phenotypes and characteristics including mitochondrial and lysosomal dysfunctions, as well as elevated αSyn protein levels (summarised in Table 5.7).

Table 5.7 Multiple known cellular and molecular characteristics of PD were reproduced following siRNA-mediated V-ATPase depletion in vitro

Characteristic Comments In Patients In vitro

Accumulation of CMA dysfunction is implicated in (Murphy et al., Figure 5.20 CMA substrates PD as the cause of the 2015) accumulation of CMA substrate (Sala et al., proteins including MEF2D 2016)

247 5. The Vacuolar ATPase and PD

Ca2+ Reduced lysosomal Ca2+ and (Hurley et al., Figure 5.23 dyshomeostasis increased mitochondrial Ca2+ 2013) evident in patient derived iPSC (Schöndorf et neurons al., 2014) 2+ Altered levels of Ca binding (Kilpatrick et al., proteins evidence in patient 2016) brain regions

Increased ROS Strongly implicated in PD due to (Dexter et al., Figure 5.24 the presence of oxidative 1994) damage to mitochondrial DNA, (J Zhang et al., lipids and proteins 1999) (Alam et al., 1997)

Increased αSyn The accumulation of αSyn {Kalia:2015bw} Figure 5.39 protein levels protein and the presence of αSyn enriched Lewy bodies and Lewy neurites is a hallmark of PD

Increased ATP5A Observed in the frontal cortex of {Ferrer:2007cq} Figure 5.28 levels sporadic PD patients

Decreased αSyn The potential change in αSyn Figure 5.41 Figure 5.40 mRNA levels mRNA levels in patient brain {Kingsbury:2004 regions has not been well ce} characterised, with multiple studies reporting different findings. This study identified a significant decrease in αSyn mRNA in patient brain tissue.

In addition to recapitulating many known PD cellular phenotypes, V-ATPase depletion also resulted in a number of cellular dysfunctions that have not previously been associated with PD, including the novel observation that V-ATPase depletion led to reduced perinuclear clustering of lysosomes (as discussed above). Assessment of the

248 5. The Vacuolar ATPase and PD effects of V-ATPase depletion also led to the discovery of a relationship between V- ATPase and αSyn mRNA levels. Depletion of ATP6V1A and ATP6V1B2 subunit levels in vitro resulted in a reduction in αSyn mRNA levels (Figure 5.40), mirroring the correlation between αSyn and V-ATPase mRNA levels that was subsequently identified in the brain (Figure 5.41). Despite the central role of αSyn in PD, the mechanisms through which αSyn gene expression are regulated remain poorly understood. The discovery of this interrelationship between V-ATPase and αSyn expression provides significant insight into this aspect of PD and is a primary focus of future experiments, as discussed in further detail in section 6.2.4.

249 6. Conclusions

6 Thesis Conclusions and Future directions

This thesis aimed to identify and explore early molecular and cellular changes that contribute to degeneration in affected brain regions of sporadic PD patients. Using an approach that integrated proteomic data from multiple in vitro PD model systems with transcriptomic data available from patient brain regions (section 3) two key areas were prioritized for detailed investigation of their potential contributions to PD; the mitochondrial protein CHCHD2 (section 4) and the V-ATPase protein complex (section 5).

6.1 CHCHD2

6.1.1 CHCHD2 dysfunction is implicated in both familial and sporadic PD

CHCHD2 variants have been identified to cause fPD (Funayama et al., 2015) and these variants have recently been proposed to cause mitochondrial dysfunction by destabilising inner membrane cristae structure and/or impairing mitochondrial retention of Cyt c (Meng et al., 2017). CHCHD2 dysfunction would therefore be predicted to lead to impaired OXPHOS super-complex formation, impaired sequestration of the mitochondrial proton gradient and leakage of Cyt c into the cytosol, where it induces apoptosis. This thesis additionally identifies reduced CHCHD2 transcript levels in sPD patient brain regions (Figure 4.3), suggesting that these CHCHD2-dependent mitochondrial perturbations may additionally contribute to sPD.

250 6. Conclusions

6.1.2 CHCHD2 is translationally up-regulated in response to reduced mitochondrial membrane potential

CHCHD2 was prioritized for detailed investigation in this thesis based on the observation that the protein is rapidly increased following treatment of SH-SY5Y cells with the mitochondrial toxin and commonly used PD-mimic rotenone (Table 3.3). Investigation of the effects of other mitochondrial toxins indicated that CHCHD2 protein is up-regulated in response to rotenone, antimycin A and CCCP, but not affected by ER stress inducer thapsigargin, suggesting that the protein plays a role in responding to mitochondrial dysfunction. These mitochondrial toxins significantly increased CHCHD2 protein levels without altering CHCHD2 mRNA levels (Figure 4.5) and additionally resulted in preferential accumulation of CHCHD2 in mitochondria with a lower degree of membrane polarisation (Figure 4.9), where its increased level would potentially help restore mitochondrial cristae structure and/or prevent cytochrome c release and associated apoptosis. These observations emphasize the precision of the CHCHD2 response and implicate the loss of mitochondrial membrane potential as a common mechanism through which the tested mitochondrial toxins induce translational up-regulation CHCHD2.

In investigating the regulatory mechanism that facilitates the rapid and localised increase in CHCHD2 protein levels in response to mitochondrial toxins, the CHCHD2 homologues in C. elegans and S. cerevisiae were identified to interact with translationally-repressive PUF RNA binding proteins (Freeberg et al., 2013; Kershner & Kimble, 2010), suggesting that mammalian CHCHD2 expression may be regulated through a similar mechanism. Furthermore, in mammalian cells PD-associated PINK1 and PARKIN have recently been identified to localise the translation of nuclear encoded mitochondrial proteins to the mitochondrial vicinity by facilitating the localised removal of PUF-protein translational suppression (Gehrke et al., 2015). Based on these data this thesis presents a model for the regulation of CHCHD2 translation wherein protein levels are controlled through PUF protein-mediated repression of CHCHD2 mRNA and removal of this translational block by PINK1/PARKIN. Through this

251 6. Conclusions mechanism the accumulation of PINK1/PARKIN on the mitochondrial surface that occurs in response to loss of membrane potential (Jin et al., 2010), is specifically coupled to increased CHCHD2 translation, thereby targeting translational up- regulation of CHCHD2 to less polarised mitochondria.

This model of CHCHD2 translational regulation proposes that there is a biological link between multiple PD-associated proteins (PINK1, PARKIN and CHCHD2), suggesting that impaired mitochondrial response to loss of membrane potential is a pathway that these multiple PD-associated proteins converge upon. Furthermore, the reduced CHCHD2 mRNA expression and decreased PARKIN activation (Dawson & Dawson, 2014) evident in sPD patient brain regions suggest that activation of this mitochondrial stress response may be impaired in sPD. Experiments are currently ongoing towards confirming this model and determining whether CHCHD2 protein levels are perturbed in sPD patient brain tissue. Together with the results of this thesis these experiments will substantially enhance understanding of this aspect of mitochondrial biology and the consequences of its dysfunction in PD.

6.2 The V-ATPase

6.2.1 The emerging role of the V-ATPase in PD

This thesis provides substantial evidence that the V-ATPase is dysregulated in sPD, with reduced levels of transcripts encoding multiple subunits and reduced protein levels of the ATP6V1B2 subunit identified in sPD patient brain regions. Additionally, assessment of PD-associated SNVs using the VEGAS approach identified ATP6V0A1 as a novel-PD associated risk locus and this discovery was very recently (October 2017) confirmed by a meta-analysis of PD GWAS studies that identified ATP6V0A1 amongst 17 newly identified PD risk loci (Chang et al., 2017). These data therefore implicate for the first time that V-ATPase dysfunction contributes to sPD.

252 6. Conclusions

V-ATPase impairment mediated by either chemical inhibition with BafA1 or siRNA- mediated depletion was assessed in vitro to identify the potential consequences of the reduced V-ATPase expression observed in PD patients. V-ATPase impairment was found to recapitulate multiple PD-associated cellular phenotypes, including reduced protein degradation (Figure 5.21), Ca2+ dyshomeostasis (Figure 5.24), increased ROS production (Figure 5.23) and perturbed mitochondrial turnover (Figure 5.25). These results position the V-ATPase depletion observed in patients as a potential contributor to the development of these phenotypes in patients.

6.2.2 The V-ATPase is involved in in mediating lysosomal transport: implications for PD

Assessment of V-ATPase impairment mediated by either low-doses of BafA1 or siRNA- mediated depletion of subunits revealed a distinct difference between the consequences of these two methods of V-ATPase impairment on the subcellular spatial distribution of lysosomes. While depletion of V-ATPase subunit levels reduced the percentage of lysosomes in the perinuclear region of the cell, BafA1 did not have this effect (Figure 5.35). These data suggest that the V-ATPase plays a structural role in facilitating the transport of lysosomes towards the microtubule organising centre where the majority of lysosomes are located (Huotari & Helenius, 2011), aligning with the previous discovery that through RILP subunits of the V-ATPase V1 domain associate with dynein motors that regulate this directional movement along microtubules (De Luca et al., 2014). Impaired lysosomal transport has been associated with AD (Cataldo et al., 1991; Gowrishankar et al., 2015), a neurodegenerative condition that has many overlaps with PD, but has not yet been explored in PD. This result implicates impaired lysosomal transport as a possible additional consequence of the reduced V-ATPase levels observed in patients and has led to investigation of this possibility using patient brain sections. Characterisation of this novel aspect of PD

253 6. Conclusions pathobiology will significantly enhance the understanding of the underlying causes of lysosomal dysfunction in PD.

6.2.3 Discovery of a novel relationship between the V-ATPase and αSyn

This thesis describes the discovery of a novel relationship between the V-ATPase and αSyn that is evident in human brain tissue. Depletion of V-ATPase subunits in vitro increased αSyn protein levels, mimicking the same inverse correlation observed between these proteins in the human brain. Furthermore, V-ATPase depletion resulted in a significant reduction in αSyn transcription, leading to the discovery of a positive correlation between the levels of αSyn and V-ATPase transcripts in the brain. These discoveries suggest that the V-ATPase and αSyn may be functionally related, possibly through their mutual involvement in regulating the synaptic vesicle pool (Wang et al., 2014a) (Burré, 2015; Burré et al., 2010; Vavassori & Mayer, 2014). Further investigation of this novel relationship (discussed below) will significantly improve understanding of this aspect of neurobiology.

6.2.4 Future directions: regulation of αSyn expression through V-ATPase mediated lysosomal Fe2+ release and the GATA2 transcription factor

The transcriptional regulation of αSyn in the brain is currently poorly understood. Investigation of the relationship between V-ATPase and αSyn transcript levels may provide significant insight into this elusive area of PD research. This thesis proposes a model to explain the observed V-ATPase-αSyn transcriptional relationship. V-ATPase depletion is proposed to cause release of lysosomal Fe2+ that inactivates the Fe2+ responsive transcription factor GATA2, which both negatively regulates expression of mitochondrial Fe2+ storage genes (Guaraldo et al., 2016) and positively regulate αSyn transcription (Scherzer et al., 2008). Release of lysosomal Fe2+ upon V-ATPase

254 6. Conclusions depletion may reduce GATA2 activity, which facilitates an increase in the expression of mitochondrial Fe2+ storage genes, consequently reducing the transcriptional expression of αSyn. To test this proposal the effect of V-ATPase depletion on the levels of other known GATA2 target genes and the abundance and nuclear localisation of GATA2 is being assessed. Characterisation of this relationship will significantly enhance understanding of the factors that regulate transcription of αSyn and provide invaluable insight into the mechanism of Fe2+ toxicity in PD.

Figure 6.1 Proposed mechanism of V-ATPase-mediated modulation of αSyn transcription

255 References

REFERENCES

Abad, M. F. C., Di Benedetto, G., Magalhães, P. J., Filippin, L., & Pozzan, T. (2004). Mitochondrial pH monitored by a new engineered green fluorescent protein mutant. The Journal of Biological Chemistry, 279(12), 11521–11529. doi:10.1074/jbc.M306766200 Adam-Vizi, V., & Starkov, A. A. (2010). Calcium and mitochondrial reactive oxygen species generation: how to read the facts. Journal of Alzheimer's Disease : JAD, 20 Suppl 2(s2), S413–26. doi:10.3233/JAD-2010-100465 Akhtar, T. A., Orsomando, G., Mehrshahi, P., Lara-Núñez, A., Bennett, M. J., Gregory, J. F., & Hanson, A. D. (2010). A central role for gamma-glutamyl hydrolases in plant folate homeostasis. The Plant Journal : for Cell and Molecular Biology, 64(2), 256– 266. doi:10.1111/j.1365-313X.2010.04330.x Alam, Z. I., Daniel, S. E., Lees, A. J., Marsden, D. C., Jenner, P., & Halliwell, B. (1997). A generalised increase in protein carbonyls in the brain in Parkinson's but not incidental Lewy body disease. Journal of Neurochemistry, 69(3), 1326–1329. Alcázar-Fabra, M., Navas, P., & Brea-Calvo, G. (2016). Coenzyme Q biosynthesis and its role in the respiratory chain structure. Biochimica Et Biophysica Acta, 1857(8), 1073–1078. doi:10.1016/j.bbabio.2016.03.010 Aldridge, J. E., Horibe, T., & Hoogenraad, N. J. (2007). Discovery of genes activated by the mitochondrial unfolded protein response (mtUPR) and cognate promoter elements. PloS One, 2(9), e874. doi:10.1371/journal.pone.0000874 Allen Reish, H. E., & Standaert, D. G. (2015). Role of α-synuclein in inducing innate and adaptive immunity in Parkinson disease. Journal of Parkinson's Disease, 5(1), 1–19. doi:10.3233/JPD-140491 Althoff, T., Mills, D. J., Popot, J.-L., & Kühlbrandt, W. (2011). Arrangement of electron transport chain components in bovine mitochondrial supercomplex I1III2IV1. The EMBO Journal, 30(22), 4652–4664. doi:10.1038/emboj.2011.324 Alvarez-Erviti, L., Seow, Y., Schapira, A. H. V., Rodriguez-Oroz, M. C., Obeso, J. A., & Cooper, J. M. (2013). Influence of microRNA deregulation on chaperone-mediated autophagy and α-synuclein pathology in Parkinson's disease. Cell Death & Disease, 4(3), e545. doi:10.1038/cddis.2013.73 Anderson, S. E., & Meade, B. J. (2014). Potential health effects associated with dermal exposure to occupational chemicals. Environmental Health Insights, 8(Suppl 1), 51–62. doi:10.4137/EHI.S15258 Anglade, P., Vyas, S., Javoy-Agid, F., Herrero, M. T., Michel, P. P., Marquez, J., et al. (1997). Apoptosis and autophagy in nigral neurons of patients with Parkinson's disease. Histology and Histopathology, 12(1), 25–31. Anheim, M., Elbaz, A., Lesage, S., Dürr, A., Condroyer, C., Viallet, F., et al. (2012). Penetrance of Parkinson disease in glucocerebrosidase gene mutation carriers. Neurology, 78(6), 417–420. doi:10.1212/WNL.0b013e318245f476 Annunen-Rasila, J., Ohlmeier, S., Tuokko, H., Veijola, J., & Majamaa, K. (2007). Proteome and cytoskeleton responses in osteosarcoma cells with reduced OXPHOS activity. Proteomics, 7(13), 2189–2200. doi:10.1002/pmic.200601031 Aras, S., Bai, M., Lee, I., Springett, R., Hüttemann, M., & Grossman, L. I. (2015). MNRR1

256 References

(formerly CHCHD2) is a bi-organellar regulator of mitochondrial metabolism. Mitochondrion, 20, 43–51. doi:10.1016/j.mito.2014.10.003 Aras, S., Pak, O., Sommer, N., Finley, R., Hüttemann, M., Weissmann, N., & Grossman, L. I. (2013). Oxygen-dependent expression of cytochrome c oxidase subunit 4-2 gene expression is mediated by transcription factors RBPJ, CXXC5 and CHCHD2. Nucleic Acids Research, 41(4), 2255–2266. doi:10.1093/nar/gks1454 Araújo, J. R., Martel, F., Borges, N., Araújo, J. M., & Keating, E. (2015). Folates and aging: Role in mild cognitive impairment, dementia and depression. Ageing Research Reviews, 22, 9–19. doi:10.1016/j.arr.2015.04.005 Arias, E., & Cuervo, A. M. (2011). Chaperone-mediated autophagy in protein quality control. Current Opinion in Cell Biology, 23(2), 184–189. doi:10.1016/j.ceb.2010.10.009 Armbrüster, A., Hohn, C., Hermesdorf, A., Schumacher, K., Börsch, M., & Grüber, G. (2005). Evidence for major structural changes in subunit C of the vacuolar ATPase due to nucleotide binding. FEBS Letters, 579(9), 1961–1967. doi:10.1016/j.febslet.2005.02.042 Arthur, C. R., Morton, S. L., Dunham, L. D., Keeney, P. M., & Bennett, J. P. (2009). Parkinson's disease brain mitochondria have impaired respirasome assembly, age- related increases in distribution of oxidative damage to mtDNA and no differences in heteroplasmic mtDNA mutation abundance. Molecular Neurodegeneration, 4(1), 37. doi:10.1186/1750-1326-4-37 Ascherio, A., Weisskopf, M. G., O'Reilly, E. J., McCullough, M. L., Calle, E. E., Rodriguez, C., & Thun, M. J. (2004). Coffee consumption, gender, and Parkinson's disease mortality in the cancer prevention study II cohort: the modifying effects of estrogen. American Journal of Epidemiology, 160(10), 977–984. doi:10.1093/aje/kwh312 Ashrafi, G., & Ryan, T. A. (2017). Glucose metabolism in nerve terminals. Current Opinion in Neurobiology, 45, 156–161. doi:10.1016/j.conb.2017.03.007 Bae, E.-J., Yang, N. Y., Lee, C., Kim, S., Lee, H.-J., & Lee, S.-J. (2015). Haploinsufficiency of cathepsin D leads to lysosomal dysfunction and promotes cell-to-cell transmission of α-synuclein aggregates. Cell Death & Disease, 6(10), e1901. doi:10.1038/cddis.2015.283 Baltazar, G. C., Guha, S., Lu, W., Lim, J., Boesze-Battaglia, K., Laties, A. M., et al. (2012). Acidic nanoparticles are trafficked to lysosomes and restore an acidic lysosomal pH and degradative function to compromised ARPE-19 cells. PloS One, 7(12), e49635. doi:10.1371/journal.pone.0049635 Bannwarth, S., Ait-El-Mkadem, S., Chaussenot, A., Genin, E. C., Lacas-Gervais, S., Fragaki, K., et al. (2014). A mitochondrial origin for frontotemporal dementia and amyotrophic lateral sclerosis through CHCHD10 involvement. Brain : a Journal of Neurology, 137(Pt 8), 2329–2345. doi:10.1093/brain/awu138 Barrett, P. J., & Timothy Greenamyre, J. (2015). Post-translational modification of α- synuclein in Parkinson's disease. Brain Research, 1628(Pt B), 247–253. doi:10.1016/j.brainres.2015.06.002 Bassani, T. B., Gradowski, R. W., Zaminelli, T., Barbiero, J. K., Santiago, R. M., Boschen, S. L., et al. (2014). Neuroprotective and antidepressant-like effects of melatonin in

257 References

a rotenone-induced Parkinson's disease model in rats. Brain Research, 1593, 95– 105. doi:10.1016/j.brainres.2014.09.068 Baughman, J. M., Nilsson, R., Gohil, V. M., Arlow, D. H., Gauhar, Z., & Mootha, V. K. (2009). A computational screen for regulators of oxidative phosphorylation implicates SLIRP in mitochondrial RNA homeostasis. PLoS Genetics, 5(8), e1000590. doi:10.1371/journal.pgen.1000590 Bayer, M. J., Reese, C., Buhler, S., Peters, C., & Mayer, A. (2003). Vacuole membrane fusion: V0 functions after trans-SNARE pairing and is coupled to the Ca2+-releasing channel. The Journal of Cell Biology, 162(2), 211–222. doi:10.1083/jcb.200212004 Beatty, C. H., Young, M. K., & Bocek, R. M. (1976). Control of glycolysis in skeletal muscle from fetal rhesus monkeys. Pediatric Research, 10(3), 149–153. doi:10.1203/00006450-197603000-00001 Berg, J. M., Tymoczko, J. L., & Stryer, L. (2002). Biochemistry, 5th Edition. In 16.2 The Glycolytic Pathway is Tightly Controlled (5 ed.). New York: W H Freeman. Bernard, A., & Klionsky, D. J. (2014). Defining the membrane precursor supporting the nucleation of the phagophore. Autophagy, 10(1), 1–2. doi:10.4161/auto.27242 Berti, V., Polito, C., Borghammer, P., Ramat, S., Mosconi, L., Vanzi, E., et al. (2012). Alternative normalization methods demonstrate widespread cortical hypometabolism in untreated de novo Parkinson's disease. The Quarterly Journal of Nuclear Medicine and Molecular Imaging : Official Publication of the Italian Association of Nuclear Medicine (AIMN) [and] the International Association of Radiopharmacology (IAR), [and] Section of the Society of, 56(3), 299–308. Betarbet, R., Sherer, T. B., MacKenzie, G., Garcia-Osuna, M., Panov, A. V., & Greenamyre, J. T. (2000). Chronic systemic pesticide exposure reproduces features of Parkinson's disease. Nature Neuroscience, 3(12), 1301–1306. doi:10.1038/81834 Beyenbach, K. W., & Wieczorek, H. (2006). The V-type H+ ATPase: molecular structure and function, physiological roles and regulation. The Journal of Experimental Biology, 209(Pt 4), 577–589. doi:10.1242/jeb.02014 Bhargava, A., Voronov, I., Wang, Y., Glogauer, M., Kartner, N., & Manolson, M. F. (2012). Osteopetrosis mutation R444L causes endoplasmic reticulum retention and misprocessing of vacuolar H+-ATPase a3 subunit. The Journal of Biological Chemistry, 287(32), 26829–26839. doi:10.1074/jbc.M112.345702 Biazik, J., Ylä-Anttila, P., Vihinen, H., Jokitalo, E., & Eskelinen, E.-L. (2015). Ultrastructural relationship of the phagophore with surrounding organelles. Autophagy, 11(3), 439–451. doi:10.1080/15548627.2015.1017178 Bisaglia, M., Trolio, A., Bellanda, M., Bergantino, E., Bubacco, L., & Mammi, S. (2006). Structure and topology of the non-amyloid-beta component fragment of human alpha-synuclein bound to micelles: implications for the aggregation process. Protein Science : a Publication of the Protein Society, 15(6), 1408–1416. doi:10.1110/ps.052048706 Blakely, R. D., & Edwards, R. H. (2012). Vesicular and plasma membrane transporters for neurotransmitters. Cold Spring Harbor Perspectives in Biology, 4(2), a005595– a005595. doi:10.1101/cshperspect.a005595 Blanco-Ayala, T., Andérica-Romero, A. C., & Pedraza-Chaverri, J. (2014). New insights into antioxidant strategies against paraquat toxicity. Free Radical Research, 48(6),

258 References

623–640. doi:10.3109/10715762.2014.899694 Blewett, N. H., & Goldstrohm, A. C. (2012). A eukaryotic translation initiation factor 4E- binding protein promotes mRNA decapping and is required for PUF repression. Molecular and Cellular Biology, 32(20), 4181–4194. doi:10.1128/MCB.00483-12 Bohnen, N. I., Kotagal, V., Müller, M. L. T. M., Koeppe, R. A., Scott, P. J. H., Albin, R. L., et al. (2014). Diabetes mellitus is independently associated with more severe cognitive impairment in Parkinson disease. Parkinsonism & Related Disorders, 20(12), 1394–1398. doi:10.1016/j.parkreldis.2014.10.008 Bonifacino, J. S., & Rojas, R. (2006). Retrograde transport from endosomes to the trans-Golgi network. Nature Reviews. Molecular Cell Biology, 7(8), 568–579. doi:10.1038/nrm1985 Bonnet, A. M., Jutras, M. F., Czernecki, V., Corvol, J. C., & Vidailhet, M. (2012). Nonmotor symptoms in Parkinson's disease in 2012: relevant clinical aspects. Parkinson's Disease, 2012(3), 198316–15. doi:10.1155/2012/198316 Borghammer, P., Chakravarty, M., Jonsdottir, K. Y., Sato, N., Matsuda, H., Ito, K., et al. (2010). Cortical hypometabolism and hypoperfusion in Parkinson's disease is extensive: probably even at early disease stages. Brain Structure & Function, 214(4), 303–317. doi:10.1007/s00429-010-0246-0 Bourdenx, M., Daniel, J., Genin, E., Soria, F. N., Blanchard-Desce, M., Bezard, E., & Dehay, B. (2016). Nanoparticles restore lysosomal acidification defects: Implications for Parkinson and other lysosomal-related diseases. Autophagy, 12(3), 472–483. doi:10.1080/15548627.2015.1136769 Bowman, E. J., Siebers, A., & Altendorf, K. (1988). Bafilomycins: a class of inhibitors of membrane ATPases from microorganisms, animal cells, and plant cells. Proceedings of the National Academy of Sciences of the United States of America, 85(21), 7972–7976. Braak, H., Del Tredici, K., Rüb, U., de Vos, R. A. I., Jansen Steur, E. N. H., & Braak, E. (2003). Staging of brain pathology related to sporadic Parkinson's disease. Neurobiology of Aging, 24(2), 197–211. Braak, H., Ghebremedhin, E., Rüb, U., Bratzke, H., & Del Tredici, K. (2004). Stages in the development of Parkinson's disease-related pathology. Cell and Tissue Research, 318(1), 121–134. doi:10.1007/s00441-004-0956-9 Bretaud, S., Lee, S., & Guo, S. (2004). Sensitivity of zebrafish to environmental toxins implicated in Parkinson's disease. Neurotoxicology and Teratology, 26(6), 857–864. doi:10.1016/j.ntt.2004.06.014 Brett, C. L., Tukaye, D. N., Mukherjee, S., & Rao, R. (2005). The yeast endosomal Na+K+/H+ exchanger Nhx1 regulates cellular pH to control vesicle trafficking. Molecular Biology of the Cell, 16(3), 1396–1405. doi:10.1091/mbc.E04-11-0999 Brett, C. L., Wei, Y., Donowitz, M., & Rao, R. (2002). Human Na(+)/H(+) exchanger isoform 6 is found in recycling endosomes of cells, not in mitochondria. American Journal of Physiology. Cell Physiology, 282(5), C1031–41. doi:10.1152/ajpcell.00420.2001 Brookes, P. S., Yoon, Y., Robotham, J. L., Anders, M. W., & Sheu, S.-S. (2004). Calcium, ATP, and ROS: a mitochondrial love-hate triangle. American Journal of Physiology. Cell Physiology, 287(4), C817–33. doi:10.1152/ajpcell.00139.2004

259 References

Brosnan, J. T., & Brosnan, M. E. (2013). Glutamate: a truly functional amino acid. Amino Acids, 45(3), 413–418. doi:10.1007/s00726-012-1280-4 Brown, M. S., & Goldstein, J. L. (1986). A receptor-mediated pathway for cholesterol homeostasis. Science (New York, N.Y.), 232(4746), 34–47. Burchell, V. S., Nelson, D. E., Sanchez-Martinez, A., Delgado-Camprubi, M., Ivatt, R. M., Pogson, J. H., et al. (2013). The Parkinson's disease-linked proteins Fbxo7 and Parkin interact to mediate mitophagy. Nature Neuroscience, 16(9), 1257–1265. doi:10.1038/nn.3489 Burgess, A., Vigneron, S., Brioudes, E., Labbé, J.-C., Lorca, T., & Castro, A. (2010). Loss of human Greatwall results in G2 arrest and multiple mitotic defects due to deregulation of the cyclin B-Cdc2/PP2A balance. Proceedings of the National Academy of Sciences of the United States of America, 107(28), 12564–12569. doi:10.1073/pnas.0914191107 Burré, J. (2015). The Synaptic Function of α-Synuclein. Journal of Parkinson's Disease, 5(4), 699–713. doi:10.3233/JPD-150642 Burré, J., Sharma, M., Tsetsenis, T., Buchman, V., Etherton, M. R., & Südhof, T. C. (2010). Alpha-synuclein promotes SNARE-complex assembly in vivo and in vitro. Science (New York, N.Y.), 329(5999), 1663–1667. doi:10.1126/science.1195227 Bussell, R., & Eliezer, D. (2003). A structural and functional role for 11-mer repeats in alpha-synuclein and other exchangeable lipid binding proteins. Journal of Molecular Biology, 329(4), 763–778. Bustin, S. A., Benes, V., Garson, J. A., Hellemans, J., Huggett, J., Kubista, M., et al. (2009, April). The MIQE guidelines: minimum information for publication of quantitative real-time PCR experiments. Clinical Chemistry. Clinical Chemistry. doi:10.1373/clinchem.2008.112797 Cannon, J. R., Tapias, V., Na, H. M., Honick, A. S., Drolet, R. E., & Greenamyre, J. T. (2009). A highly reproducible rotenone model of Parkinson's disease. Neurobiology of Disease, 34(2), 279–290. Canuel, M., Korkidakis, A., Konnyu, K., & Morales, C. R. (2008). Sortilin mediates the lysosomal targeting of cathepsins D and H. Biochemical and Biophysical Research Communications, 373(2), 292–297. doi:10.1016/j.bbrc.2008.06.021 Cao, Q., Yang, Y., Zhong, X. Z., & Dong, X.-P. (2017). The lysosomal Ca2+ release channel TRPML1 regulates lysosome size by activating calmodulin. The Journal of Biological Chemistry, jbc.M116.772160. doi:10.1074/jbc.M116.772160 Carmona-Gutierrez, D., Hughes, A. L., Madeo, F., & Ruckenstuhl, C. (2016). The crucial impact of lysosomes in aging and longevity. Ageing Research Reviews, 32, 2–12. doi:10.1016/j.arr.2016.04.009 Cartelli, D., Goldwurm, S., Casagrande, F., Pezzoli, G., & Cappelletti, G. (2012). Microtubule destabilization is shared by genetic and idiopathic Parkinson's disease patient fibroblasts. PloS One, 7(5), e37467. doi:10.1371/journal.pone.0037467 Casteels, C., Lauwers, E., Bormans, G., Baekelandt, V., & Van Laere, K. (2008). Metabolic-dopaminergic mapping of the 6-hydroxydopamine rat model for Parkinson's disease. European Journal of Nuclear Medicine and Molecular Imaging, 35(1), 124–134. doi:10.1007/s00259-007-0558-3 Cataldo, A. M., Paskevich, P. A., Kominami, E., & Nixon, R. A. (1991). Lysosomal

260 References

hydrolases of different classes are abnormally distributed in brains of patients with Alzheimer disease. Proceedings of the National Academy of Sciences of the United States of America, 88(24), 10998–11002. Cavallaro, G. (2010). Genome-wide analysis of eukaryotic twin CX9C proteins. Molecular bioSystems, 6(12), 2459–2470. doi:10.1039/c0mb00058b Challet, C., Maechler, P., Wollheim, C. B., & Ruegg, U. T. (2001). Mitochondrial calcium oscillations in C2C12 myotubes. The Journal of Biological Chemistry, 276(6), 3791– 3797. doi:10.1074/jbc.M006209200 Chan, S. N., & Tang, B. L. (2013). Location and membrane sources for autophagosome formation - from ER-mitochondria contact sites to Golgi-endosome-derived carriers. Molecular Membrane Biology, 30(8), 394–402. doi:10.3109/09687688.2013.850178 Chang, D., Nalls, M. A., Hallgrímsdóttir, I. B., Hunkapiller, J., van der Brug, M., Cai, F., et al. (2017). A meta-analysis of genome-wide association studies identifies 17 new Parkinson's disease risk loci. Nature Genetics, 49(10), 1511–1516. doi:10.1038/ng.3955 Chatterjee, P., & Roy, D. (2017). Comparative analysis of RNA-Seq data from brain and blood samples of Parkinson's disease. Biochemical and Biophysical Research Communications, 484(3), 557–564. doi:10.1016/j.bbrc.2017.01.121 Chen, H., Zhang, S. M., Schwarzschild, M. A., Hernán, M. A., Logroscino, G., Willett, W. C., & Ascherio, A. (2004). Folate intake and risk of Parkinson's disease. American Journal of Epidemiology, 160(4), 368–375. doi:10.1093/aje/kwh213 Cheng, F., Vivacqua, G., & Yu, S. (2011). The role of α-synuclein in neurotransmission and synaptic plasticity. Journal of Chemical Neuroanatomy, 42(4), 242–248. doi:10.1016/j.jchemneu.2010.12.001 Chiba-Falek, O., Lopez, G. J., & Nussbaum, R. L. (2006). Levels of alpha-synuclein mRNA in sporadic Parkinson disease patients. Movement Disorders : Official Journal of the Movement Disorder Society, 21(10), 1703–1708. doi:10.1002/mds.21007 Chiu, C.-C., Yeh, T.-H., Lai, S.-C., Wu-Chou, Y.-H., Chen, C.-H., Mochly-Rosen, D., et al. (2015). Neuroprotective effects of aldehyde dehydrogenase 2 activation in rotenone-induced cellular and animal models of parkinsonism. Experimental Neurology, 263, 244–253. doi:10.1016/j.expneurol.2014.09.016 Chouraki, V., & Seshadri, S. (2014). Genetics of Alzheimer's disease. Advances in Genetics, 87, 245–294. doi:10.1016/B978-0-12-800149-3.00005-6 Christensen, K. A., Myers, J. T., & Swanson, J. A. (2002). pH-dependent regulation of lysosomal calcium in macrophages. Journal of Cell Science, 115(Pt 3), 599–607. Chu, C. T., Bayır, H., & Kagan, V. E. (2014). LC3 binds externalized cardiolipin on injured mitochondria to signal mitophagy in neurons: implications for Parkinson disease. Autophagy, 10(2), 376–378. doi:10.4161/auto.27191 Chu, C. T., Ji, J., Dagda, R. K., Jiang, J. F., Tyurina, Y. Y., Kapralov, A. A., et al. (2013). Cardiolipin externalization to the outer mitochondrial membrane acts as an elimination signal for mitophagy in neuronal cells. Nature Cell Biology, 15(10), 1197–1205. doi:10.1038/ncb2837 Chu, Y., Dodiya, H., Aebischer, P., Olanow, C. W., & Kordower, J. H. (2009). Alterations in lysosomal and proteasomal markers in Parkinson's disease: relationship to

261 References

alpha-synuclein inclusions. Neurobiology of Disease, 35(3), 385–398. doi:10.1016/j.nbd.2009.05.023 Ciechanover, A. (2005). Proteolysis: from the lysosome to ubiquitin and the proteasome. Nature Reviews. Molecular Cell Biology, 6(1), 79–87. doi:10.1038/nrm1552 Cieri, D., Brini, M., & Calì, T. (2017). Emerging (and converging) pathways in Parkinson's disease: keeping mitochondrial wellness. Biochemical and Biophysical Research Communications, 483(4), 1020–1030. doi:10.1016/j.bbrc.2016.08.153 Clarimón, J., & Kulisevsky, J. (2013). Parkinson's disease: from genetics to clinical practice. Current Genomics, 14(8), 560–567. doi:10.2174/1389202914666131210212305 Clark, I. E., Dodson, M. W., Jiang, C., Cao, J. H., Huh, J. R., Seol, J. H., et al. (2006). Drosophila pink1 is required for mitochondrial function and interacts genetically with parkin. Nature, 441(7097), 1162–1166. doi:10.1038/nature04779 Coelho, M., & Ferreira, J. J. (2012). Late-stage Parkinson disease. Nature Reviews. Neurology, 8(8), 435–442. doi:10.1038/nrneurol.2012.126 Coelln, von, R., & Shulman, L. M. (2016). Clinical subtypes and genetic heterogeneity: of lumping and splitting in Parkinson disease. Current Opinion in Neurology, 29(6), 727–734. doi:10.1097/WCO.0000000000000384 Coffey, J. W., & De Duve, C. (1968). Digestive activity of lysosomes. I. The digestion of proteins by extracts of rat liver lysosomes. The Journal of Biological Chemistry, 243(12), 3255–3263. Cogliati, S., Enriquez, J. A., & Scorrano, L. (2016). Mitochondrial Cristae: Where Beauty Meets Functionality. Trends in Biochemical Sciences, 41(3), 261–273. doi:10.1016/j.tibs.2016.01.001 Colacurcio, D. J., & Nixon, R. A. (2016). Disorders of lysosomal acidification-The emerging role of v-ATPase in aging and neurodegenerative disease. Ageing Research Reviews. doi:10.1016/j.arr.2016.05.004 Collier, T. J., Kanaan, N. M., & Kordower, J. H. (2011). Ageing as a primary risk factor for Parkinson's disease: evidence from studies of non-human primates. Nature Reviews. Neuroscience, 12(6), 359–366. doi:10.1038/nrn3039 Collier, T. J., Kanaan, N. M., & Kordower, J. H. (2017). Aging and Parkinson's disease: Different sides of the same coin? Movement Disorders : Official Journal of the Movement Disorder Society, 14(suppl. 1), 317. doi:10.1002/mds.27037 Cooke, A., Prigge, A., Opperman, L., & Wickens, M. (2011). Targeted translational regulation using the PUF protein family scaffold. Proceedings of the National Academy of Sciences of the United States of America, 108(38), 15870–15875. doi:10.1073/pnas.1105151108 Cooper, G. M., & Sunderland, M. A. (2000). The Cell: A Molecular Approach. (2nd ed.). Sinauer Associates. Cordonnier, M. N., Dauzonne, D., Louvard, D., & Coudrier, E. (2001). Actin filaments and myosin I alpha cooperate with microtubules for the movement of lysosomes. Molecular Biology of the Cell, 12(12), 4013–4029. Costa, J., Lunet, N., Santos, C., Santos, J., & Vaz-Carneiro, A. (2010). Caffeine exposure and the risk of Parkinson's disease: a systematic review and meta-analysis of

262 References

observational studies. Journal of Alzheimer's Disease : JAD, 20 Suppl 1(s1), S221– 38. doi:10.3233/JAD-2010-091525 Coulom, H., & Birman, S. (2004). Chronic exposure to rotenone models sporadic Parkinson's disease in Drosophila melanogaster. The Journal of Neuroscience : the Official Journal of the Society for Neuroscience, 24(48), 10993–10998. doi:10.1523/JNEUROSCI.2993-04.2004 Cuervo, A. M., & Dice, J. F. (1996). A receptor for the selective uptake and degradation of proteins by lysosomes. Science (New York, N.Y.), 273(5274), 501–503. Cuervo, A. M., Stefanis, L., Fredenburg, R., Lansbury, P. T., & Sulzer, D. (2004). Impaired degradation of mutant alpha-synuclein by chaperone-mediated autophagy. Science (New York, N.Y.), 305(5688), 1292–1295. doi:10.1126/science.1101738 Cullen, V., Lindfors, M., Ng, J., Paetau, A., Swinton, E., Kolodziej, P., et al. (2009). Cathepsin D expression level affects alpha-synuclein processing, aggregation, and toxicity in vivo. Molecular Brain, 2(1), 5. doi:10.1186/1756-6606-2-5 Curtin, K., Fleckenstein, A. E., Robison, R. J., Crookston, M. J., Smith, K. R., & Hanson, G. R. (2015). Methamphetamine/amphetamine abuse and risk of Parkinson's disease in Utah: a population-based assessment. Drug and Alcohol Dependence, 146, 30– 38. doi:10.1016/j.drugalcdep.2014.10.027 D'Autréaux, B., & Toledano, M. B. (2007). ROS as signalling molecules: mechanisms that generate specificity in ROS homeostasis. Nature Reviews. Molecular Cell Biology, 8(10), 813–824. doi:10.1038/nrm2256 Dames, P., Zimmermann, B., Schmidt, R., Rein, J., Voss, M., Schewe, B., et al. (2006). cAMP regulates plasma membrane vacuolar-type H+-ATPase assembly and activity in blowfly salivary glands. Proceedings of the National Academy of Sciences of the United States of America, 103(10), 3926–3931. doi:10.1073/pnas.0600011103 Darshi, M., Mendiola, V. L., Mackey, M. R., Murphy, A. N., Koller, A., Perkins, G. A., et al. (2011). ChChd3, an inner mitochondrial membrane protein, is essential for maintaining crista integrity and mitochondrial function. The Journal of Biological Chemistry, 286(4), 2918–2932. doi:10.1074/jbc.M110.171975 Darweesh, S. K. L., Verlinden, V. J. A., Stricker, B. H., Hofman, A., Koudstaal, P. J., & Ikram, M. A. (2017). Trajectories of prediagnostic functioning in Parkinson's disease. Brain : a Journal of Neurology, 140(Pt 2), 429–441. doi:10.1093/brain/aww291 Dautry-Varsat, A., Ciechanover, A., & Lodish, H. F. (1983). pH and the recycling of transferrin during receptor-mediated endocytosis. Proceedings of the National Academy of Sciences of the United States of America, 80(8), 2258–2262. Davis, A. A., Andruska, K. M., Benitez, B. A., Racette, B. A., Perlmutter, J. S., & Cruchaga, C. (2016). Variants in GBA, SNCA, and MAPT influence Parkinson disease risk, age at onset, and progression. Neurobiology of Aging, 37, 209.e1–7. doi:10.1016/j.neurobiolaging.2015.09.014 Dawson, T. M., & Dawson, V. L. (2014). Parkin plays a role in sporadic Parkinson's disease. Neuro-Degenerative Diseases, 13(2-3), 69–71. doi:10.1159/000354307 De Luca, M., Cogli, L., Progida, C., Nisi, V., Pascolutti, R., Sigismund, S., et al. (2014). RILP regulates vacuolar ATPase through interaction with the V1G1 subunit. Journal of Cell Science, 127(Pt 12), 2697–2708. doi:10.1242/jcs.142604

263 References

Dechant, R., Binda, M., Lee, S. S., Pelet, S., Winderickx, J., & Peter, M. (2010). Cytosolic pH is a second messenger for glucose and regulates the PKA pathway through V- ATPase. The EMBO Journal, 29(15), 2515–2526. doi:10.1038/emboj.2010.138 Dehay, B., Martinez-Vicente, M., Caldwell, G. A., Caldwell, K. A., Yue, Z., Cookson, M. R., et al. (2013). Lysosomal impairment in Parkinson's disease. Movement Disorders : Official Journal of the Movement Disorder Society, 28(6), 725–732. doi:10.1002/mds.25462 Dehay, B., Martinez-Vicente, M., Ramirez, A., Perier, C., Klein, C., Vila, M., & Bezard, E. (2012a). Lysosomal dysfunction in Parkinson disease: ATP13A2 gets into the groove. Autophagy, 8(9), 1389–1391. doi:10.4161/auto.21011 Dehay, B., Ramirez, A., Martinez-Vicente, M., Perier, C., Canron, M.-H., Doudnikoff, E., et al. (2012b). Loss of P-type ATPase ATP13A2/PARK9 function induces general lysosomal deficiency and leads to Parkinson disease neurodegeneration. Proceedings of the National Academy of Sciences of the United States of America, 109(24), 9611–9616. doi:10.1073/pnas.1112368109 Delbridge, L. M. D., Mellor, K. M., Taylor, D. J. R., & Gottlieb, R. A. (2015). Myocardial autophagic energy stress responses--macroautophagy, mitophagy, and glycophagy. American Journal of Physiology. Heart and Circulatory Physiology, 308(10), H1194–204. doi:10.1152/ajpheart.00002.2015 Deloitte. (2015). Living with Parkinson's disease. Parkinsons Australia Inc. Denzel, M. S., & Antebi, A. (2015). Hexosamine pathway and (ER) protein quality control. Current Opinion in Cell Biology, 33, 14–18. doi:10.1016/j.ceb.2014.10.001 Dexter, D. T., Carter, C. J., Wells, F. R., Javoy-Agid, F., Agid, Y., Lees, A., et al. (1989). Basal lipid peroxidation in substantia nigra is increased in Parkinson's disease. Journal of Neurochemistry, 52(2), 381–389. Dexter, D. T., Holley, A. E., Flitter, W. D., Slater, T. F., Wells, F. R., Daniel, S. E., et al. (1994). Increased levels of lipid hydroperoxides in the parkinsonian substantia nigra: an HPLC and ESR study. Movement Disorders : Official Journal of the Movement Disorder Society, 9(1), 92–97. doi:10.1002/mds.870090115 Dexter, D. T., Jenner, P., Schapira, A. H., & Marsden, C. D. (1992). Alterations in levels of iron, ferritin, and other trace metals in neurodegenerative diseases affecting the basal ganglia. The Royal Kings and Queens Parkinson's Disease Research Group. Annals of Neurology, 32 Suppl, S94–100. Dexter, D. T., Wells, F. R., Agid, F., Agid, Y., Lees, A. J., Jenner, P., & Marsden, C. D. (1987). Increased nigral iron content in postmortem parkinsonian brain. Lancet, 2(8569), 1219–1220. Dhillon, A. S., Tarbutton, G. L., Levin, J. L., Plotkin, G. M., Lowry, L. K., Nalbone, J. T., & Shepherd, S. (2008). Pesticide/environmental exposures and Parkinson's disease in East Texas. Journal of Agromedicine, 13(1), 37–48. doi:10.1080/10599240801986215 Dhur, A., Galan, P., & Hercberg, S. (1991). Folate status and the immune system. Progress in Food & Nutrition Science, 15(1-2), 43–60. Di Fonzo, A., Chien, H. F., Socal, M., Giraudo, S., Tassorelli, C., Iliceto, G., et al. (2007). ATP13A2 missense mutations in juvenile parkinsonism and young onset Parkinson disease. Neurology, 68(19), 1557–1562. doi:10.1212/01.wnl.0000260963.08711.08

264 References

Di Giovanni, G., Di Matteo, V., & Esposito, E. (2009). Birth, life and death of dopaminergic neurons in the substantia nigra. Journal of Neural Transmission. Supplementum, (73), 1 p preceeding table of contents. Di Giovanni, J., Boudkkazi, S., Mochida, S., Bialowas, A., Samari, N., Lévêque, C., et al. (2010). V-ATPase membrane sector associates with synaptobrevin to modulate neurotransmitter release. Neuron, 67(2), 268–279. doi:10.1016/j.neuron.2010.06.024 Di Maio, R., Barrett, P. J., Hoffman, E. K., Barrett, C. W., Zharikov, A., Borah, A., et al. (2016). α-Synuclein binds to TOM20 and inhibits mitochondrial protein import in Parkinson's disease. Science Translational Medicine, 8(342), 342ra78–342ra78. doi:10.1126/scitranslmed.aaf3634 Diao, J., Burré, J., Vivona, S., Cipriano, D. J., Sharma, M., Kyoung, M., et al. (2013). Native α-synuclein induces clustering of synaptic-vesicle mimics via binding to phospholipids and synaptobrevin-2/VAMP2. eLife, 2, e00592. doi:10.7554/eLife.00592 Dice, J. F. (1990). Peptide sequences that target cytosolic proteins for lysosomal proteolysis. Trends in Biochemical Sciences, 15(8), 305–309. Diederich, N. J., Moore, C. G., Leurgans, S. E., Chmura, T. A., & Goetz, C. G. (2003). Parkinson disease with old-age onset: a comparative study with subjects with middle-age onset. Archives of Neurology, 60(4), 529–533. doi:10.1001/archneur.60.4.529 Dienel, G. A., & Cruz, N. F. (2016). Aerobic glycolysis during brain activation: adrenergic regulation and influence of norepinephrine on astrocytic metabolism. Journal of Neurochemistry, 138(1), 14–52. doi:10.1111/jnc.13630 Ding, C., Wu, Z., Huang, L., Wang, Y., Xue, J., Chen, S., et al. (2015). Mitofilin and CHCHD6 physically interact with Sam50 to sustain cristae structure. Scientific Reports, 5(1), 16064. doi:10.1038/srep16064 Dispersyn, G., Nuydens, R., Connors, R., Borgers, M., & Geerts, H. (1999). Bcl-2 protects against FCCP-induced apoptosis and mitochondrial membrane potential depolarization in PC12 cells. Biochimica Et Biophysica Acta, 1428(2-3), 357–371. Dolman, N. J., Kilgore, J. A., & Davidson, M. W. (2013). A review of reagents for fluorescence microscopy of cellular compartments and structures, part I: BacMam labeling and reagents for vesicular structures. Current Protocols in Cytometry, Chapter 12, Unit 12.30–12.30.27. doi:10.1002/0471142956.cy1230s65 Donnelly, C. J., Fainzilber, M., & Twiss, J. L. (2010). Subcellular communication through RNA transport and localized protein synthesis. Traffic (Copenhagen, Denmark), 11(12), 1498–1505. doi:10.1111/j.1600-0854.2010.01118.x DORN, H. F. (1960). Tobacco consumption and mortality from cancer and other diseases. Acta - Unio Internationalis Contra Cancrum, 16, 1653–1665. Dorsey, E. R., Constantinescu, R., Thompson, J. P., Biglan, K. M., Holloway, R. G., Kieburtz, K., et al. (2007). Projected number of people with Parkinson disease in the most populous nations, 2005 through 2030. Neurology, 68(5), 384–386. doi:10.1212/01.wnl.0000247740.47667.03 Doty, R. L. (2012). Olfactory dysfunction in Parkinson disease. Nature Reviews. Neurology, 8(6), 329–339. doi:10.1038/nrneurol.2012.80

265 References

Dölle, C., Flønes, I., Nido, G. S., Miletic, H., Osuagwu, N., Kristoffersen, S., et al. (2016). Defective mitochondrial DNA homeostasis in the substantia nigra in Parkinson disease. Nature Communications, 7, 13548. doi:10.1038/ncomms13548 Driver, J. A. (2014). Inverse association between cancer and neurodegenerative disease: review of the epidemiologic and biological evidence. Biogerontology, 15(6), 547–557. doi:10.1007/s10522-014-9523-2 Drolet, R. E., Cannon, J. R., Montero, L., & Greenamyre, J. T. (2009). Chronic rotenone exposure reproduces Parkinson's disease gastrointestinal neuropathology. Neurobiology of Disease, 36(1), 96–102. doi:10.1016/j.nbd.2009.06.017 Dubos, A., Castells-Nobau, A., Meziane, H., Oortveld, M. A. W., Houbaert, X., Iacono, G., et al. (2015). Conditional depletion of intellectual disability and Parkinsonism candidate gene ATP6AP2 in fly and mouse induces cognitive impairment and neurodegeneration. Human Molecular Genetics, 24(23), 6736–6755. doi:10.1093/hmg/ddv380 Dumitriu, A., Golji, J., Labadorf, A. T., Gao, B., Beach, T. G., Myers, R. H., et al. (2016). Integrative analyses of proteomics and RNA transcriptomics implicate mitochondrial processes, protein folding pathways and GWAS loci in Parkinson disease. BMC Medical Genomics, 9(1), 5. doi:10.1186/s12920-016-0164-y Dunn, L., Allen, G. F., Mamais, A., Ling, H., Li, A., Duberley, K. E., et al. (2014). Dysregulation of glucose metabolism is an early event in sporadic Parkinson's disease. Neurobiology of Aging, 35(5), 1111–1115. doi:10.1016/j.neurobiolaging.2013.11.001 Farrer, M., Gwinn-Hardy, K., Hutton, M., & Hardy, J. (1999). The genetics of disorders with synuclein pathology and parkinsonism. Human Molecular Genetics, 8(10), 1901–1905. Feng, Y., He, D., Yao, Z., & Klionsky, D. J. (2014). The machinery of macroautophagy. Cell Research, 24(1), 24–41. doi:10.1038/cr.2013.168 Fernández-Dueñas, V., Gómez-Soler, M., López-Cano, M., Taura, J. J., Ledent, C., Watanabe, M., et al. (2014). Uncovering caffeine's adenosine A2A receptor inverse agonism in experimental parkinsonism. ACS Chemical Biology, 9(11), 2496–2501. doi:10.1021/cb5005383 Ferreira, M., & Massano, J. (2016). An updated review of Parkinson's disease genetics and clinicopathological correlations. Acta Neurologica Scandinavica. doi:10.1111/ane.12616 Ferrer, I., Martinez, A., Blanco, R., Dalfó, E., & Carmona, M. (2011). Neuropathology of sporadic Parkinson disease before the appearance of parkinsonism: preclinical Parkinson disease. Journal of Neural Transmission (Vienna, Austria : 1996), 118(5), 821–839. doi:10.1007/s00702-010-0482-8 Ferrer, I., Perez, E., Dalfó, E., & Barrachina, M. (2007). Abnormal levels of prohibitin and ATP synthase in the substantia nigra and frontal cortex in Parkinson's disease. Neuroscience Letters, 415(3), 205–209. doi:10.1016/j.neulet.2007.01.026 Finsterwald, C., Magistretti, P. J., & Lengacher, S. (2015). Astrocytes: New Targets for the Treatment of Neurodegenerative Diseases. Current Pharmaceutical Design, 21(25), 3570–3581. Fiorese, C. J., Schulz, A. M., Lin, Y.-F., Rosin, N., Pellegrino, M. W., & Haynes, C. M.

266 References

(2016). The Transcription Factor ATF5 Mediates a Mammalian Mitochondrial UPR. Current Biology : CB, 26(15), 2037–2043. doi:10.1016/j.cub.2016.06.002 Fleming, S. M., Zhu, C., Fernagut, P.-O., Mehta, A., DiCarlo, C. D., Seaman, R. L., & Chesselet, M.-F. (2004). Behavioral and immunohistochemical effects of chronic intravenous and subcutaneous infusions of varying doses of rotenone. Experimental Neurology, 187(2), 418–429. doi:10.1016/j.expneurol.2004.01.023 Floor, E., & Wetzel, M. G. (1998). Increased protein oxidation in human substantia nigra pars compacta in comparison with basal ganglia and prefrontal cortex measured with an improved dinitrophenylhydrazine assay. Journal of Neurochemistry, 70(1), 268–275. Fontanesi, F., Soto, I. C., & Barrientos, A. (2008). Cytochrome c oxidase biogenesis: new levels of regulation. IUBMB Life, 60(9), 557–568. doi:10.1002/iub.86 Foo, J. N., Liu, J., & Tan, E.-K. (2015). CHCHD2 and Parkinson's disease. The Lancet. Neurology, 14(7), 681–682. doi:10.1016/S1474-4422(15)00098-8 Forgac, M. (2007). Vacuolar ATPases: rotary proton pumps in physiology and pathophysiology. Nature Reviews. Molecular Cell Biology, 8(11), 917–929. doi:10.1038/nrm2272 Forsaa, E. B., Larsen, J. P., Wentzel-Larsen, T., Goetz, C. G., Stebbins, G. T., Aarsland, D., & Alves, G. (2010). A 12-year population-based study of psychosis in Parkinson disease. Archives of Neurology, 67(8), 996–1001. doi:10.1001/archneurol.2010.166 Fowler, S., & De Duve, C. (1969). Digestive activity of lysosomes. 3. The digestion of lipids by extracts of rat liver lysosomes. The Journal of Biological Chemistry, 244(2), 471–481. Fox, T. D. (2012). Mitochondrial protein synthesis, import, and assembly. Genetics, 192(4), 1203–1234. doi:10.1534/genetics.112.141267 Freeberg, M. A., Han, T., Moresco, J. J., Kong, A., Yang, Y.-C., Lu, Z. J., et al. (2013). Pervasive and dynamic protein binding sites of the mRNA transcriptome in Saccharomyces cerevisiae. Genome Biology, 14(2), R13. doi:10.1186/gb-2013-14-2- r13 Fuchs, J., Tichopad, A., Golub, Y., Munz, M., Schweitzer, K. J., Wolf, B., et al. (2008). Genetic variability in the SNCA gene influences alpha-synuclein levels in the blood and brain. FASEB Journal : Official Publication of the Federation of American Societies for Experimental Biology, 22(5), 1327–1334. doi:10.1096/fj.07-9348com Fujiwara, H., Hasegawa, M., Dohmae, N., Kawashima, A., Masliah, E., Goldberg, M. S., et al. (2002). alpha-Synuclein is phosphorylated in synucleinopathy lesions. Nature Cell Biology, 4(2), 160–164. doi:10.1038/ncb748 Funakoshi-Hirose, I., Aki, T., Unuma, K., Funakoshi, T., Noritake, K., & Uemura, K. (2013). Distinct effects of methamphetamine on autophagy-lysosome and ubiquitin-proteasome systems in HL-1 cultured mouse atrial cardiomyocytes. Toxicology, 312, 74–82. doi:10.1016/j.tox.2013.07.016 Funayama, M., Ohe, K., Amo, T., Furuya, N., Yamaguchi, J., Saiki, S., et al. (2015). CHCHD2 mutations in autosomal dominant late-onset Parkinson's disease: a genome-wide linkage and sequencing study. The Lancet. Neurology, 14(3), 274– 282. doi:10.1016/S1474-4422(14)70266-2 Furlong, M., Tanner, C. M., Goldman, S. M., Bhudhikanok, G. S., Blair, A., Chade, A., et

267 References

al. (2015). Protective glove use and hygiene habits modify the associations of specific pesticides with Parkinson's disease. Environment International, 75, 144– 150. doi:10.1016/j.envint.2014.11.002 Futai, M., Nakanishi-Matsui, M., Okamoto, H., Sekiya, M., & Nakamoto, R. K. (2012). Rotational catalysis in proton pumping ATPases: from E. coli F-ATPase to mammalian V-ATPase. Biochimica Et Biophysica Acta, 1817(10), 1711–1721. doi:10.1016/j.bbabio.2012.03.015 Galivan, J., Ryan, T. J., Chave, K., Rhee, M., Yao, R., & Yin, D. (2000). Glutamyl hydrolase. pharmacological role and enzymatic characterization. Pharmacology & Therapeutics, 85(3), 207–215. Galter, D., Buervenich, S., Carmine, A., Anvret, M., & Olson, L. (2003). ALDH1 mRNA: presence in human dopamine neurons and decreases in substantia nigra in Parkinson's disease and in the ventral tegmental area in schizophrenia. Neurobiology of Disease, 14(3), 637–647. Gamir-Morralla, A., Belbin, O., Fortea, J., Alcolea, D., Ferrer, I., Lleó, A., & Iglesias, T. (2017). Kidins220 Correlates with Tau in Alzheimer's Disease Brain and Cerebrospinal Fluid. Journal of Alzheimer's Disease : JAD, 55(4), 1327–1333. doi:10.3233/JAD-160639 Gamir-Morralla, A., López-Menéndez, C., Ayuso-Dolado, S., Tejeda, G. S., Montaner, J., Rosell, A., et al. (2015). Development of a neuroprotective peptide that preserves survival pathways by preventing Kidins220/ARMS calpain processing induced by excitotoxicity. Cell Death & Disease, 6(10), e1939. doi:10.1038/cddis.2015.307 Gan-Or, Z., Dion, P. A., & Rouleau, G. A. (2015). Genetic perspective on the role of the autophagy-lysosome pathway in Parkinson disease. Autophagy, 11(9), 1443–1457. doi:10.1080/15548627.2015.1067364 Ganley, I. G. (2013). Autophagosome maturation and lysosomal fusion. Essays in Biochemistry, 55, 65–78. doi:10.1042/bse0550065 García-García, F., Ponce, S., Brown, R., Cussen, V., & Krueger, J. M. (2005). Sleep disturbances in the rotenone animal model of Parkinson disease. Brain Research, 1042(2), 160–168. doi:10.1016/j.brainres.2005.02.036 Gasnier, B. (2000). The loading of neurotransmitters into synaptic vesicles. Biochimie, 82(4), 327–337. Gasser, T. (2015). Usefulness of Genetic Testing in PD and PD Trials: A Balanced Review. Journal of Parkinson's Disease, 5(2), 209–215. doi:10.3233/JPD-140507 Gehrke, S., Wu, Z., Klinkenberg, M., Sun, Y., Auburger, G., Guo, S., & Lu, B. (2015). PINK1 and Parkin control localized translation of respiratory chain component mRNAs on mitochondria outer membrane. Cell Metabolism, 21(1), 95–108. doi:10.1016/j.cmet.2014.12.007 Gelb, D. J., Oliver, E., & Gilman, S. (1999). Diagnostic criteria for Parkinson disease. Archives of Neurology, 56(1), 33–39. Genin, E. C., Plutino, M., Bannwarth, S., Villa, E., Cisneros-Barroso, E., Roy, M., et al. (2016). CHCHD10 mutations promote loss of mitochondrial cristae junctions with impaired mitochondrial genome maintenance and inhibition of apoptosis. EMBO Molecular Medicine, 8(1), 58–72. doi:10.15252/emmm.201505496 Genova, M. L., & Lenaz, G. (2014). Functional role of mitochondrial respiratory

268 References

supercomplexes. Biochimica Et Biophysica Acta, 1837(4), 427–443. doi:10.1016/j.bbabio.2013.11.002 Giaime, E., Yamaguchi, H., Gautier, C. A., Kitada, T., & Shen, J. (2012). Loss of DJ-1 does not affect mitochondrial respiration but increases ROS production and mitochondrial permeability transition pore opening. PloS One, 7(7), e40501. doi:10.1371/journal.pone.0040501 Gielisch, I., & Meierhofer, D. (2015). Metabolome and proteome profiling of complex I deficiency induced by rotenone. Journal of Proteome Research, 14(1), 224–235. doi:10.1021/pr500894v Goers, J., Manning-Bog, A. B., McCormack, A. L., Millett, I. S., Doniach, S., Di Monte, D. A., et al. (2003). Nuclear localization of alpha-synuclein and its interaction with histones. Biochemistry, 42(28), 8465–8471. doi:10.1021/bi0341152 Goldenring, J. R. (2015). Recycling endosomes. Current Opinion in Cell Biology, 35, 117– 122. doi:10.1016/j.ceb.2015.04.018 Goldfarb, D. S., Corbett, A. H., Mason, D. A., Harreman, M. T., & Adam, S. A. (2004). Importin alpha: a multipurpose nuclear-transport receptor. Trends in Cell Biology, 14(9), 505–514. doi:10.1016/j.tcb.2004.07.016 Goldstein, D. S., Sullivan, P., Holmes, C., Miller, G. W., Alter, S., Strong, R., et al. (2013). Determinants of buildup of the toxic dopamine metabolite DOPAL in Parkinson's disease. Journal of Neurochemistry, 126(5), 591–603. doi:10.1111/jnc.12345 Golias, T., Papandreou, I., Sun, R., Kumar, B., Brown, N. V., Swanson, B. J., et al. (2016). Hypoxic repression of pyruvate dehydrogenase activity is necessary for metabolic reprogramming and growth of model tumours. Scientific Reports, 6(1), 31146. doi:10.1038/srep31146 Gorell, J. M., Johnson, C. C., Rybicki, B. A., Peterson, E. L., & Richardson, R. J. (1998). The risk of Parkinson's disease with exposure to pesticides, farming, well water, and rural living. Neurology, 50(5), 1346–1350. Gossen, M., & Bujard, H. (1992). Tight control of gene expression in mammalian cells by tetracycline-responsive promoters. Proceedings of the National Academy of Sciences of the United States of America, 89(12), 5547–5551. Gouarné, C., Tracz, J., Paoli, M. G., Deluca, V., Seimandi, M., Tardif, G., et al. (2015). Protective role of olesoxime against wild-type α-synuclein-induced toxicity in human neuronally differentiated SHSY-5Y cells. British Journal of Pharmacology, 172(1), 235–245. doi:10.1111/bph.12939 Gowrishankar, S., & Ferguson, S. M. (2016). Lysosomes relax in the cellular suburbs. The Journal of Cell Biology, 212(6), 617–619. doi:10.1083/jcb.201602082 Gowrishankar, S., Yuan, P., Wu, Y., Schrag, M., Paradise, S., Grutzendler, J., et al. (2015). Massive accumulation of luminal protease-deficient axonal lysosomes at Alzheimer's disease amyloid plaques. Proceedings of the National Academy of Sciences of the United States of America, 112(28), E3699–708. doi:10.1073/pnas.1510329112 Gómez-Suaga, P., & Hilfiker, S. (2012). LRRK2 as a modulator of lysosomal calcium homeostasis with downstream effects on autophagy. Autophagy, 8(4), 692–693. doi:10.4161/auto.19305 Görlach, A., Bertram, K., Hudecova, S., & Krizanova, O. (2015). Calcium and ROS: A

269 References

mutual interplay. Redox Biology, 6, 260–271. doi:10.1016/j.redox.2015.08.010 Granger, E., McNee, G., Allan, V., & Woodman, P. (2014). The role of the cytoskeleton and molecular motors in endosomal dynamics. Seminars in Cell & Developmental Biology, 31, 20–29. doi:10.1016/j.semcdb.2014.04.011 Griffin, J. W. D., & Bradshaw, P. C. (2017). Amino Acid Catabolism in Alzheimer's Disease Brain: Friend or Foe? Oxidative Medicine and Cellular Longevity, 2017(6), 5472792–15. doi:10.1155/2017/5472792 Griffiths, G., Hoflack, B., Simons, K., Mellman, I., & Kornfeld, S. (1988). The mannose 6- phosphate receptor and the biogenesis of lysosomes. Cell, 52(3), 329–341. Grünblatt, E., Mandel, S., Jacob-Hirsch, J., Zeligson, S., Amariglo, N., Rechavi, G., et al. (2004). Gene expression profiling of parkinsonian substantia nigra pars compacta; alterations in ubiquitin-proteasome, heat shock protein, iron and oxidative stress regulated proteins, cell adhesion/cellular matrix and vesicle trafficking genes. Journal of Neural Transmission (Vienna, Austria : 1996), 111(12), 1543–1573. doi:10.1007/s00702-004-0212-1 Grünewald, A., Rygiel, K. A., Hepplewhite, P. D., Morris, C. M., Picard, M., & Turnbull, D. M. (2016). Mitochondrial DNA Depletion in Respiratory Chain-Deficient Parkinson Disease Neurons. Annals of Neurology, 79(3), 366–378. doi:10.1002/ana.24571 Guaraldo, M., Santambrogio, P., Rovelli, E., Di Savino, A., Saglio, G., Cittaro, D., et al. (2016). Characterization of human mitochondrial ferritin promoter: identification of transcription factors and evidences of epigenetic control. Scientific Reports, 6(1), 33432. doi:10.1038/srep33432 Guardia-Laguarta, C., Area-Gomez, E., Schon, E. A., & Przedborski, S. (2015). A new role for α-synuclein in Parkinson's disease: Alteration of ER-mitochondrial communication. Movement Disorders : Official Journal of the Movement Disorder Society, 30(8), 1026–1033. doi:10.1002/mds.26239 Gueli, M. C., & Taibi, G. (2013). Alzheimer's disease: amino acid levels and brain metabolic status. Neurological Sciences : Official Journal of the Italian Neurological Society and of the Italian Society of Clinical Neurophysiology, 34(9), 1575–1579. doi:10.1007/s10072-013-1289-9 Guha, S., Coffey, E. E., Lu, W., Lim, J. C., Beckel, J. M., Laties, A. M., et al. (2014). Approaches for detecting lysosomal alkalinization and impaired degradation in fresh and cultured RPE cells: evidence for a role in retinal degenerations. Experimental Eye Research, 126, 68–76. doi:10.1016/j.exer.2014.05.013 Gupta, H. V., Vengoechea, J., Sahaya, K., & Virmani, T. (2015). A splice site mutation in ATP6AP2 causes X-linked intellectual disability, epilepsy, and parkinsonism. Parkinsonism & Related Disorders, 21(12), 1473–1475. doi:10.1016/j.parkreldis.2015.10.001 Halliday, G. M., & McCann, H. (2010). The progression of pathology in Parkinson's disease. Annals of the New York Academy of Sciences, 1184(1), 188–195. doi:10.1111/j.1749-6632.2009.05118.x Hansson, M. J., Månsson, R., Morota, S., Uchino, H., Kallur, T., Sumi, T., et al. (2008). Calcium-induced generation of reactive oxygen species in brain mitochondria is mediated by permeability transition. Free Radical Biology & Medicine, 45(3), 284–

270 References

294. doi:10.1016/j.freeradbiomed.2008.04.021 Hare, D. J., Adlard, P. A., Doble, P. A., & Finkelstein, D. I. (2013). Metallobiology of 1- methyl-4-phenyl-1,2,3,6-tetrahydropyridine neurotoxicity. Metallomics : Integrated Biometal Science, 5(2), 91–109. doi:10.1039/c2mt20164j Hauser, D. N., & Hastings, T. G. (2013). Mitochondrial dysfunction and oxidative stress in Parkinson's disease and monogenic parkinsonism. Neurobiology of Disease, 51, 35–42. doi:10.1016/j.nbd.2012.10.011 Haynes, C. M., Yang, Y., Blais, S. P., Neubert, T. A., & Ron, D. (2010). The matrix peptide exporter HAF-1 signals a mitochondrial UPR by activating the transcription factor ZC376.7 in C. elegans. Molecular Cell, 37(4), 529–540. doi:10.1016/j.molcel.2010.01.015 Henry, A. G., Aghamohammadzadeh, S., Samaroo, H., Chen, Y., Mou, K., Needle, E., & Hirst, W. D. (2015). Pathogenic LRRK2 mutations, through increased kinase activity, produce enlarged lysosomes with reduced degradative capacity and increase ATP13A2 expression. Human Molecular Genetics, 24(21), 6013–6028. doi:10.1093/hmg/ddv314 Hernandez, D. G., Reed, X., & Singleton, A. B. (2016). Genetics in Parkinson disease: Mendelian versus non-Mendelian inheritance. Journal of Neurochemistry. doi:10.1111/jnc.13593 Hernando, N., David, P., Tarsio, M., Bartkiewicz, M., Horne, W. C., Kane, P. M., & Baron, R. (1999). The presence of the alternatively spliced A2 cassette in the vacuolar H+-ATPase subunit A prevents assembly of the V1 catalytic domain. European Journal of Biochemistry, 266(1), 293–301. Hernán, M. A., Takkouche, B., Caamaño-Isorna, F., & Gestal-Otero, J. J. (2002). A meta- analysis of coffee drinking, cigarette smoking, and the risk of Parkinson's disease. Annals of Neurology, 52(3), 276–284. doi:10.1002/ana.10277 Ho, S. C., Woo, J., & Lee, C. M. (1989). Epidemiologic study of Parkinson's disease in Hong Kong. Neurology, 39(10), 1314–1318. Hogan, G. J., Brown, P. O., & Herschlag, D. (2015). Evolutionary Conservation and Diversification of Puf RNA Binding Proteins and Their mRNA Targets. PLoS Biology, 13(11), e1002307. doi:10.1371/journal.pbio.1002307 Holden, H. M., Rayment, I., & Thoden, J. B. (2003). Structure and function of enzymes of the Leloir pathway for galactose metabolism. The Journal of Biological Chemistry, 278(45), 43885–43888. doi:10.1074/jbc.R300025200 Holliday, L. S., Lu, M., Lee, B. S., Nelson, R. D., Solivan, S., Zhang, L., & Gluck, S. L. (2000). The amino-terminal domain of the B subunit of vacuolar H+-ATPase contains a filamentous actin binding site. The Journal of Biological Chemistry, 275(41), 32331–32337. doi:10.1074/jbc.M004795200 Homewood, C. A., Warhurst, D. C., Peters, W., & Baggaley, V. C. (1972). Lysosomes, pH and the anti-malarial action of chloroquine. Nature, 235(5332), 50–52. Hong, C. T., Chau, K.-Y., & Schapira, A. H. V. (2016). Meclizine-induced enhanced glycolysis is neuroprotective in Parkinson disease cell models. Scientific Reports, 6(1), 25344. doi:10.1038/srep25344 Hong, D.-P., Fink, A. L., & Uversky, V. N. (2009). Smoking and Parkinson's disease: does nicotine affect alpha-synuclein fibrillation? Biochimica Et Biophysica Acta, 1794(2),

271 References

282–290. doi:10.1016/j.bbapap.2008.09.026 Hong, J., Nakano, Y., Yokomakura, A., Ishihara, K., Kim, S., Kang, Y.-S., & Ohuchi, K. (2006). Nitric oxide production by the vacuolar-type (H+)-ATPase inhibitors bafilomycin A1 and concanamycin A and its possible role in apoptosis in RAW 264.7 cells. The Journal of Pharmacology and Experimental Therapeutics, 319(2), 672–681. doi:10.1124/jpet.106.109280 Hooshmand, B., Solomon, A., Kåreholt, I., Rusanen, M., Hänninen, T., Leiviskä, J., et al. (2012). Associations between serum homocysteine, holotranscobalamin, folate and cognition in the elderly: a longitudinal study. Journal of Internal Medicine, 271(2), 204–212. doi:10.1111/j.1365-2796.2011.02484.x Houtkooper, R. H., Mouchiroud, L., Ryu, D., Moullan, N., Katsyuba, E., Knott, G., et al. (2013). Mitonuclear protein imbalance as a conserved longevity mechanism. Nature, 497(7450), 451–457. doi:10.1038/nature12188 Hu, Y.-B., Dammer, E. B., Ren, R.-J., & Wang, G. (2015). The endosomal-lysosomal system: from acidification and cargo sorting to neurodegeneration. Translational Neurodegeneration, 4(1), 18. doi:10.1186/s40035-015-0041-1 Hughes, A. J., Daniel, S. E., Kilford, L., & Lees, A. J. (1992). Accuracy of clinical diagnosis of idiopathic Parkinson's disease: a clinico-pathological study of 100 cases. Journal of Neurology, Neurosurgery, and Psychiatry, 55(3), 181–184. Hughes, A. L., & Gottschling, D. E. (2012). An early age increase in vacuolar pH limits mitochondrial function and lifespan in yeast. Nature, 492(7428), 261–265. doi:10.1038/nature11654 Hughes, A. L., Hughes, C. E., Henderson, K. A., Yazvenko, N., & Gottschling, D. E. (2016). Selective sorting and destruction of mitochondrial membrane proteins in aged yeast. eLife, 5, 913. doi:10.7554/eLife.13943 Huotari, J., & Helenius, A. (2011). Endosome maturation. The EMBO Journal, 30(17), 3481–3500. doi:10.1038/emboj.2011.286 Hurley, M. J., Brandon, B., Gentleman, S. M., & Dexter, D. T. (2013). Parkinson's disease is associated with altered expression of CaV1 channels and calcium-binding proteins. Brain : a Journal of Neurology, 136(Pt 7), 2077–2097. doi:10.1093/brain/awt134 Hurtado-Lorenzo, A., Skinner, M., Annan, El, J., Futai, M., Sun-Wada, G.-H., Bourgoin, S., et al. (2006). V-ATPase interacts with ARNO and Arf6 in early endosomes and regulates the protein degradative pathway. Nature Cell Biology, 8(2), 124–136. doi:10.1038/ncb1348 Hwang, O. (2013). Role of oxidative stress in Parkinson's disease. Experimental Neurobiology, 22(1), 11–17. doi:10.5607/en.2013.22.1.11 Hyttinen, J. M. T., Niittykoski, M., Salminen, A., & Kaarniranta, K. (2013). Maturation of autophagosomes and endosomes: a key role for Rab7. Biochimica Et Biophysica Acta, 1833(3), 503–510. doi:10.1016/j.bbamcr.2012.11.018 Ikemoto, A., Bole, D. G., & Ueda, T. (2003). Glycolysis and glutamate accumulation into synaptic vesicles. Role of glyceraldehyde phosphate dehydrogenase and 3- phosphoglycerate kinase. The Journal of Biological Chemistry, 278(8), 5929–5940. doi:10.1074/jbc.M211617200 Inoue, T., & Forgac, M. (2005). Cysteine-mediated cross-linking indicates that subunit C

272 References

of the V-ATPase is in close proximity to subunits E and G of the V1 domain and subunit a of the V0 domain. The Journal of Biological Chemistry, 280(30), 27896– 27903. doi:10.1074/jbc.M504890200 Jafari, S., Etminan, M., Aminzadeh, F., & Samii, A. (2013). Head injury and risk of Parkinson disease: a systematic review and meta-analysis. Movement Disorders : Official Journal of the Movement Disorder Society, 28(9), 1222–1229. doi:10.1002/mds.25458 Jahreiss, L., Menzies, F. M., & Rubinsztein, D. C. (2008). The itinerary of autophagosomes: from peripheral formation to kiss-and-run fusion with lysosomes. Traffic (Copenhagen, Denmark), 9(4), 574–587. doi:10.1111/j.1600- 0854.2008.00701.x Jakes, R., Spillantini, M. G., & Goedert, M. (1994). Identification of two distinct synucleins from human brain. FEBS Letters, 345(1), 27–32. Janetzky, B., Hauck, S., Youdim, M. B., Riederer, P., Jellinger, K., Pantucek, F., et al. (1994). Unaltered aconitase activity, but decreased complex I activity in substantia nigra pars compacta of patients with Parkinson's disease. Neuroscience Letters, 169(1-2), 126–128. Jang, D. P., Min, H.-K., Lee, S.-Y., Kim, I. Y., Park, H. W., Im, Y. H., et al. (2012). Functional neuroimaging of the 6-OHDA lesion rat model of Parkinson's disease. Neuroscience Letters, 513(2), 187–192. doi:10.1016/j.neulet.2012.02.034 Jang, S., & Javadov, S. (2017). Association between ROS production, swelling and the respirasome integrity in cardiac mitochondria. Archives of Biochemistry and Biophysics, 630, 1–8. doi:10.1016/j.abb.2017.07.009 Jankovic, J. (2005). Searching for a relationship between manganese and welding and Parkinson's disease. Neurology, 64(12), 2021–2028. doi:10.1212/01.WNL.0000166916.40902.63 Jansen, E. J. R., & Martens, G. J. M. (2012). Novel insights into V-ATPase functioning: distinct roles for its accessory subunits ATP6AP1/Ac45 and ATP6AP2/(pro) renin receptor. Current Protein & Peptide Science, 13(2), 124–133. Jansen, I. E., Bras, J. M., Lesage, S., Schulte, C., Gibbs, J. R., Nalls, M. A., et al. (2015). CHCHD2 and Parkinson's disease. The Lancet. Neurology, 14(7), 678–679. doi:10.1016/S1474-4422(15)00094-0 Jenner, P. (1993). Altered mitochondrial function, iron metabolism and glutathione levels in Parkinson's disease. Acta Neurologica Scandinavica. Supplementum, 146, 6–13. Jewell, J. L., Russell, R. C., & Guan, K.-L. (2013). Amino acid signalling upstream of mTOR. Nature Reviews. Molecular Cell Biology, 14(3), 133–139. doi:10.1038/nrm3522 Jin, J., Davis, J., Zhu, D., Kashima, D. T., Leroueil, M., Pan, C., et al. (2007). Identification of novel proteins affected by rotenone in mitochondria of dopaminergic cells. BMC Neuroscience, 8(1), 67. doi:10.1186/1471-2202-8-67 Jin, S. M., & Youle, R. J. (2012). PINK1- and Parkin-mediated mitophagy at a glance. Journal of Cell Science, 125(Pt 4), 795–799. doi:10.1242/jcs.093849 Jin, S. M., Lazarou, M., Wang, C., Kane, L. A., Narendra, D. P., & Youle, R. J. (2010). Mitochondrial membrane potential regulates PINK1 import and proteolytic

273 References

destabilization by PARL. The Journal of Cell Biology, 191(5), 933–942. doi:10.1083/jcb.201008084 Johansson, M., Rocha, N., Zwart, W., Jordens, I., Janssen, L., Kuijl, C., et al. (2007). Activation of endosomal dynein motors by stepwise assembly of Rab7-RILP- p150Glued, ORP1L, and the receptor betalll spectrin. The Journal of Cell Biology, 176(4), 459–471. doi:10.1083/jcb.200606077 Johnson, D. E., Ostrowski, P., Jaumouillé, V., & Grinstein, S. (2016). The position of lysosomes within the cell determines their luminal pH. The Journal of Cell Biology, 212(6), 677–692. doi:10.1083/jcb.201507112 Jordens, I., Fernandez-Borja, M., Marsman, M., Dusseljee, S., Janssen, L., Calafat, J., et al. (2001). The Rab7 effector protein RILP controls lysosomal transport by inducing the recruitment of dynein-dynactin motors. Current Biology : CB, 11(21), 1680– 1685. Kalia, L. V., & Kalia, S. K. (2015). α-Synuclein and Lewy pathology in Parkinson's disease. Current Opinion in Neurology, 28(4), 375–381. doi:10.1097/WCO.0000000000000215 Kamp, F., Exner, N., Lutz, A. K., Wender, N., Hegermann, J., Brunner, B., et al. (2010). Inhibition of mitochondrial fusion by α-synuclein is rescued by PINK1, Parkin and DJ-1. The EMBO Journal, 29(20), 3571–3589. doi:10.1038/emboj.2010.223 Kane, P. M. (1995). Disassembly and reassembly of the yeast vacuolar H(+)-ATPase in vivo. The Journal of Biological Chemistry, 270(28), 17025–17032. Kane, P. M. (2012). Targeting reversible disassembly as a mechanism of controlling V- ATPase activity. Current Protein & Peptide Science, 13(2), 117–123. Katayama, H., Kogure, T., Mizushima, N., Yoshimori, T., & Miyawaki, A. (2011). A sensitive and quantitative technique for detecting autophagic events based on lysosomal delivery. Chemistry & Biology, 18(8), 1042–1052. doi:10.1016/j.chembiol.2011.05.013 Kaur, H., Chauhan, S., & Sandhir, R. (2011). Protective effect of lycopene on oxidative stress and cognitive decline in rotenone induced model of Parkinson's disease. Neurochemical Research, 36(8), 1435–1443. doi:10.1007/s11064-011-0469-3 Kawasaki-Nishi, S., Bowers, K., Nishi, T., Forgac, M., & Stevens, T. H. (2001a). The amino-terminal domain of the vacuolar proton-translocating ATPase a subunit controls targeting and in vivo dissociation, and the carboxyl-terminal domain affects coupling of proton transport and ATP hydrolysis. The Journal of Biological Chemistry, 276(50), 47411–47420. doi:10.1074/jbc.M108310200 Kawasaki-Nishi, S., Nishi, T., & Forgac, M. (2001b). Yeast V-ATPase complexes containing different isoforms of the 100-kDa a-subunit differ in coupling efficiency and in vivo dissociation. The Journal of Biological Chemistry, 276(21), 17941– 17948. doi:10.1074/jbc.M010790200 Keeney, P. M., Xie, J., Capaldi, R. A., & Bennett, J. P. (2006). Parkinson's disease brain mitochondrial complex I has oxidatively damaged subunits and is functionally impaired and misassembled. The Journal of Neuroscience : the Official Journal of the Society for Neuroscience, 26(19), 5256–5264. doi:10.1523/JNEUROSCI.0984- 06.2006 Kershner, A. M., & Kimble, J. (2010). Genome-wide analysis of mRNA targets for

274 References

Caenorhabditis elegans FBF, a conserved stem cell regulator. Proceedings of the National Academy of Sciences of the United States of America, 107(8), 3936–3941. doi:10.1073/pnas.1000495107 Kettner, C., Bertl, A., Obermeyer, G., Slayman, C., & Bihler, H. (2003). Electrophysiological analysis of the yeast V-type proton pump: variable coupling ratio and proton shunt. Biophysical Journal, 85(6), 3730–3738. doi:10.1016/S0006- 3495(03)74789-4 Khalimonchuk, O., & Winge, D. R. (2008). Function and redox state of mitochondrial localized cysteine-rich proteins important in the assembly of cytochrome c oxidase. Biochimica Et Biophysica Acta, 1783(4), 618–628. doi:10.1016/j.bbamcr.2007.10.016 Kilpatrick, B. S., Magalhaes, J., Beavan, M. S., McNeill, A., Gegg, M. E., Cleeter, M. W. J., et al. (2016). Endoplasmic reticulum and lysosomal Ca²⁺ stores are remodelled in GBA1-linked Parkinson disease patient fibroblasts. Cell Calcium, 59(1), 12–20. doi:10.1016/j.ceca.2015.11.002 Kim, J.-W., Tchernyshyov, I., Semenza, G. L., & Dang, C. V. (2006). HIF-1-mediated expression of pyruvate dehydrogenase kinase: a metabolic switch required for cellular adaptation to hypoxia. Cell Metabolism, 3(3), 177–185. doi:10.1016/j.cmet.2006.02.002 Kingsbury, A. E., Daniel, S. E., Sangha, H., Eisen, S., Lees, A. J., & Foster, O. J. F. (2004). Alteration in alpha-synuclein mRNA expression in Parkinson's disease. Movement Disorders : Official Journal of the Movement Disorder Society, 19(2), 162–170. doi:10.1002/mds.10683 Klionsky, D. J., Eskelinen, E.-L., & Deretic, V. (2014). Autophagosomes, phagosomes, autolysosomes, phagolysosomes, autophagolysosomes... wait, I'm confused. Autophagy, 10(4), 549–551. doi:10.4161/auto.28448 Knight, A. L., Yan, X., Hamamichi, S., Ajjuri, R. R., Mazzulli, J. R., Zhang, M. W., et al. (2014). The glycolytic enzyme, GPI, is a functionally conserved modifier of dopaminergic neurodegeneration in Parkinson's models. Cell Metabolism, 20(1), 145–157. doi:10.1016/j.cmet.2014.04.017 Knull, H. R. (1980). Compartmentation of glycolytic enzymes in nerve endings as determined by glutaraldehyde fixation. The Journal of Biological Chemistry, 255(13), 6439–6444. Koc, E. C., Cimen, H., Kumcuoglu, B., Abu, N., Akpinar, G., Haque, M. E., et al. (2013). Identification and characterization of CHCHD1, AURKAIP1, and CRIF1 as new members of the mammalian mitochondrial ribosome. Frontiers in Physiology, 4, 183. doi:10.3389/fphys.2013.00183 Koch, J. C., Bitow, F., Haack, J., d'Hedouville, Z., Zhang, J.-N., Tönges, L., et al. (2015). Alpha-Synuclein affects neurite morphology, autophagy, vesicle transport and axonal degeneration in CNS neurons. Cell Death & Disease, 6(7), e1811. doi:10.1038/cddis.2015.169 Koga, S., Aoki, N., Uitti, R. J., van Gerpen, J. A., Cheshire, W. P., Josephs, K. A., et al. (2015). When DLB, PD, and PSP masquerade as MSA: an autopsy study of 134 patients. Neurology, 85(5), 404–412. doi:10.1212/WNL.0000000000001807 Kondapalli, K. C., Prasad, H., & Rao, R. (2014). An inside job: how endosomal

275 References

Na(+)/H(+) exchangers link to autism and neurological disease. Frontiers in Cellular Neuroscience, 8, 172. doi:10.3389/fncel.2014.00172 Kong, J., & Lasko, P. (2012). Translational control in cellular and developmental processes. Nature Reviews. Genetics, 13(6), 383–394. doi:10.1038/nrg3184 Kong, S. M. Y., Chan, B. K. K., Park, J.-S., Hill, K. J., Aitken, J. B., Cottle, L., et al. (2014). Parkinson's disease-linked human PARK9/ATP13A2 maintains zinc homeostasis and promotes α-Synuclein externalization via exosomes. Human Molecular Genetics, 23(11), 2816–2833. doi:10.1093/hmg/ddu099 Kornfeld, S. (1987). Trafficking of lysosomal enzymes. FASEB Journal : Official Publication of the Federation of American Societies for Experimental Biology, 1(6), 462–468. Korolchuk, V. I., Saiki, S., Lichtenberg, M., Siddiqi, F. H., Roberts, E. A., Imarisio, S., et al. (2011). Lysosomal positioning coordinates cellular nutrient responses. Nature Cell Biology, 13(4), 453–460. doi:10.1038/ncb2204 Kortüm, F., Caputo, V., Bauer, C. K., Stella, L., Ciolfi, A., Alawi, M., et al. (2015). Mutations in KCNH1 and ATP6V1B2 cause Zimmermann-Laband syndrome. Nature Genetics, 47(6), 661–667. doi:10.1038/ng.3282 Korvatska, O., Strand, N. S., Berndt, J. D., Strovas, T., Chen, D.-H., Leverenz, J. B., et al. (2013). Altered splicing of ATP6AP2 causes X-linked parkinsonism with spasticity (XPDS). Human Molecular Genetics, 22(16), 3259–3268. doi:10.1093/hmg/ddt180 Kowal, S. L., Dall, T. M., Chakrabarti, R., Storm, M. V., & Jain, A. (2013). The current and projected economic burden of Parkinson's disease in the United States. Movement Disorders : Official Journal of the Movement Disorder Society, 28(3), 311–318. doi:10.1002/mds.25292 Koyano, F., Okatsu, K., Kosako, H., Tamura, Y., Go, E., Kimura, M., et al. (2014). Ubiquitin is phosphorylated by PINK1 to activate parkin. Nature, 510(7503), 162– 166. doi:10.1038/nature13392 Kuhn, W., Winkel, R., Woitalla, D., Meves, S., Przuntek, H., & Müller, T. (1998). High prevalence of parkinsonism after occupational exposure to lead-sulfate batteries. Neurology, 50(6), 1885–1886. Kulikov, A. V., Shilov, E. S., Mufazalov, I. A., Gogvadze, V., Nedospasov, S. A., & Zhivotovsky, B. (2012). Cytochrome c: the Achilles' heel in apoptosis. Cellular and Molecular Life Sciences : CMLS, 69(11), 1787–1797. doi:10.1007/s00018-011-0895- z La Rovere, R. M. L., Roest, G., Bultynck, G., & Parys, J. B. (2016). Intracellular Ca(2+) signaling and Ca(2+) microdomains in the control of cell survival, apoptosis and autophagy. Cell Calcium, 60(2), 74–87. doi:10.1016/j.ceca.2016.04.005 Lafourcade, C., Sobo, K., Kieffer-Jaquinod, S., Garin, J., & van der Goot, F. G. (2008). Regulation of the V-ATPase along the endocytic pathway occurs through reversible subunit association and membrane localization. PloS One, 3(7), e2758. doi:10.1371/journal.pone.0002758 Lai, Y., Kim, S., Varkey, J., Lou, X., Song, J.-K., Diao, J., et al. (2014). Nonaggregated α- synuclein influences SNARE-dependent vesicle docking via membrane binding. Biochemistry, 53(24), 3889–3896. doi:10.1021/bi5002536 Lamaze, C., Fujimoto, L. M., Yin, H. L., & Schmid, S. L. (1997). The actin cytoskeleton is

276 References

required for receptor-mediated endocytosis in mammalian cells. The Journal of Biological Chemistry, 272(33), 20332–20335. Lamb, C. A., Yoshimori, T., & Tooze, S. A. (2013). The autophagosome: origins unknown, biogenesis complex. Nature Reviews. Molecular Cell Biology, 14(12), 759–774. doi:10.1038/nrm3696 Lambert, A. J., & Brand, M. D. (2004). Inhibitors of the quinone-binding site allow rapid superoxide production from mitochondrial NADH:ubiquinone oxidoreductase (complex I). The Journal of Biological Chemistry, 279(38), 39414–39420. doi:10.1074/jbc.M406576200 Lang, C. (2015). The pathogenic relationship of Alpha Synuclein and Mitochondrial dysfunction in sporadic Parkinson’s Disease. (A. Cooper). Lange, A., Mills, R. E., Lange, C. J., Stewart, M., Devine, S. E., & Corbett, A. H. (2007). Classical nuclear localization signals: definition, function, and interaction with importin alpha. The Journal of Biological Chemistry, 282(8), 5101–5105. doi:10.1074/jbc.R600026200 Langston, J. W., & Ballard, P. (1984). Parkinsonism induced by 1-methyl-4-phenyl- 1,2,3,6-tetrahydropyridine (MPTP): implications for treatment and the pathogenesis of Parkinson's disease. The Canadian Journal of Neurological Sciences. Le Journal Canadien Des Sciences Neurologiques, 11(1 Suppl), 160–165. Laplante, M., & Sabatini, D. M. (2009). mTOR signaling at a glance. Journal of Cell Science, 122(Pt 20), 3589–3594. doi:10.1242/jcs.051011 Lashuel, H. A., Overk, C. R., Oueslati, A., & Masliah, E. (2013). The many faces of α- synuclein: from structure and toxicity to therapeutic target. Nature Reviews. Neuroscience, 14(1), 38–48. doi:10.1038/nrn3406 Laurent-Matha, V., Derocq, D., Prébois, C., Katunuma, N., & Liaudet-Coopman, E. (2006). Processing of human cathepsin D is independent of its catalytic function and auto-activation: involvement of cathepsins L and B. Journal of Biochemistry, 139(3), 363–371. doi:10.1093/jb/mvj037 Lazarou, M., Sliter, D. A., Kane, L. A., Sarraf, S. A., Wang, C., Burman, J. L., et al. (2015). The ubiquitin kinase PINK1 recruits autophagy receptors to induce mitophagy. Nature, 524(7565), 309–314. doi:10.1038/nature14893 Lee, J.-H., McBrayer, M. K., Wolfe, D. M., Haslett, L. J., Kumar, A., Sato, Y., et al. (2015). Presenilin 1 Maintains Lysosomal Ca(2+) Homeostasis via TRPML1 by Regulating vATPase-Mediated Lysosome Acidification. Cell Reports, 12(9), 1430–1444. doi:10.1016/j.celrep.2015.07.050 Lee, J.-H., Yu, W. H., Kumar, A., Lee, S., Mohan, P. S., Peterhoff, C. M., et al. (2010). Lysosomal proteolysis and autophagy require presenilin 1 and are disrupted by Alzheimer-related PS1 mutations. Cell, 141(7), 1146–1158. doi:10.1016/j.cell.2010.05.008 Lenaz, G., & Genova, M. L. (2009). Structural and functional organization of the mitochondrial respiratory chain: a dynamic super-assembly. The International Journal of Biochemistry & Cell Biology, 41(10), 1750–1772. Li, N., Oquendo, E., Capaldi, R. A., Robinson, J. P., He, Y. D., Hamadeh, H. K., et al. (2014). A systematic assessment of mitochondrial function identified novel signatures for drug-induced mitochondrial disruption in cells. Toxicological

277 References

Sciences : an Official Journal of the Society of Toxicology, 142(1), 261–273. doi:10.1093/toxsci/kfu176 Li, N., Ragheb, K., Lawler, G., Sturgis, J., Rajwa, B., Melendez, J. A., & Robinson, J. P. (2003). Mitochondrial complex I inhibitor rotenone induces apoptosis through enhancing mitochondrial reactive oxygen species production. The Journal of Biological Chemistry, 278(10), 8516–8525. doi:10.1074/jbc.M210432200 Li, X., Rydzewski, N., Hider, A., Zhang, X., Yang, J., Wang, W., et al. (2016). A molecular mechanism to regulate lysosome motility for lysosome positioning and tubulation. Nature Cell Biology, 18(4), 404–417. doi:10.1038/ncb3324 Licker, V., Turck, N., Kövari, E., Burkhardt, K., Côte, M., Surini-Demiri, M., et al. (2014). Proteomic analysis of human substantia nigra identifies novel candidates involved in Parkinson's disease pathogenesis. Proteomics, 14(6), 784–794. doi:10.1002/pmic.201300342 Lindström, V., Gustafsson, G., Sanders, L. H., Howlett, E. H., Sigvardson, J., Kasrayan, A., et al. (2017). Extensive uptake of α-synuclein oligomers in astrocytes results in sustained intracellular deposits and mitochondrial damage. Molecular and Cellular Neurosciences, 82, 143–156. doi:10.1016/j.mcn.2017.04.009 Liu, Guan-yu, Jiang, X.-X., Zhu, X., He, W.-Y., Kuang, Y.-L., Ren, K., et al. (2015a). ROS activates JNK-mediated autophagy to counteract apoptosis in mouse mesenchymal stem cells in vitro. Acta Pharmacologica Sinica, 36(12), 1473–1479. doi:10.1038/aps.2015.101 Liu, Hongyu, Li, Y., Li, Y., Liu, B., Wu, H., Wang, J., et al. (2012). Cloning and functional analysis of FLJ20420: a novel transcription factor for the BAG-1 promoter. PloS One, 7(5), e34832. doi:10.1371/journal.pone.0034832 Liu, Jimmy Z, McRae, A. F., Nyholt, D. R., Medland, S. E., Wray, N. R., Brown, K. M., et al. (2010). A versatile gene-based test for genome-wide association studies. American Journal of Human Genetics, 87(1), 139–145. doi:10.1016/j.ajhg.2010.06.009 Liu, Rong, Zhi, X., & Zhong, Q. (2015b). ATG14 controls SNARE-mediated autophagosome fusion with a lysosome. Autophagy, 11(5), 847–849. doi:10.1080/15548627.2015.1037549 Liu, Xianpeng, Lee, Y. J., Liou, L.-C., Ren, Q., Zhang, Z., Wang, S., & Witt, S. N. (2011). Alpha-synuclein functions in the nucleus to protect against hydroxyurea-induced replication stress in yeast. Human Molecular Genetics, 20(17), 3401–3414. doi:10.1093/hmg/ddr246 Liu, Y, Clegg, H. V., Leslie, P. L., Di, J., Tollini, L. A., He, Y., et al. (2015c). CHCHD2 inhibits apoptosis by interacting with Bcl-x L to regulate Bax activation. Cell Death and Differentiation, 22(6), 1035–1046. doi:10.1038/cdd.2014.194 Liu, Yan, Sun, J.-D., Song, L.-K., Li, J., Chu, S.-F., Yuan, Y.-H., & Chen, N.-H. (2015d). Environment-contact administration of rotenone: A new rodent model of Parkinson's disease. Behavioural Brain Research, 294, 149–161. doi:10.1016/j.bbr.2015.07.058 Liu, Yong, & Zhang, Y. (2015). CHCHD2 connects mitochondrial metabolism to apoptosis. Molecular & Cellular Oncology, 2(4), e1004964. doi:10.1080/23723556.2015.1004964

278 References

Llopis, J., McCaffery, J. M., Miyawaki, A., Farquhar, M. G., & Tsien, R. Y. (1998). Measurement of cytosolic, mitochondrial, and Golgi pH in single living cells with green fluorescent proteins. Proceedings of the National Academy of Sciences of the United States of America, 95(12), 6803–6808. Lloyd-Evans, E., & Platt, F. M. (2011). Lysosomal Ca(2+) homeostasis: role in pathogenesis of lysosomal storage diseases. Cell Calcium, 50(2), 200–205. doi:10.1016/j.ceca.2011.03.010 Lloyd-Evans, E., Waller-Evans, H., Peterneva, K., & Platt, F. M. (2010). Endolysosomal calcium regulation and disease. Biochemical Society Transactions, 38(6), 1458– 1464. doi:10.1042/BST0381458 Locascio, J. J., Eberly, S., Liao, Z., Liu, G., Hoesing, A. N., Duong, K., et al. (2015). Association between α-synuclein blood transcripts and early, neuroimaging- supported Parkinson's disease. Brain : a Journal of Neurology, 138(Pt 9), 2659– 2671. doi:10.1093/brain/awv202 Longen, S., Bien, M., Bihlmaier, K., Kloeppel, C., Kauff, F., Hammermeister, M., et al. (2009). Systematic analysis of the twin cx(9)c protein family. Journal of Molecular Biology, 393(2), 356–368. doi:10.1016/j.jmb.2009.08.041 López-Menéndez, C., Gamir-Morralla, A., Jurado-Arjona, J., Higuero, A. M., Campanero, M. R., Ferrer, I., et al. (2013). Kidins220 accumulates with tau in human Alzheimer's disease and related models: modulation of its calpain-processing by GSK3β/PP1 imbalance. Human Molecular Genetics, 22(3), 466–482. doi:10.1093/hmg/dds446 López-Menéndez, C., Gascón, S., Sobrado, M., Vidaurre, O. G., Higuero, A. M., Rodríguez-Peña, A., et al. (2009). Kidins220/ARMS downregulation by excitotoxic activation of NMDARs reveals its involvement in neuronal survival and death pathways. Journal of Cell Science, 122(Pt 19), 3554–3565. doi:10.1242/jcs.056473 Lu, C.-S., Lai, S.-C., Wu, R.-M., Weng, Y.-H., Huang, C.-L., Chen, R.-S., et al. (2012). PLA2G6 mutations in PARK14-linked young-onset parkinsonism and sporadic Parkinson's disease. American Journal of Medical Genetics. Part B, Neuropsychiatric Genetics : the Official Publication of the International Society of Psychiatric Genetics, 159B(2), 183–191. doi:10.1002/ajmg.b.32012 Lu, L., Fu, D.-L., Li, H.-Q., Liu, A.-J., Li, J.-H., & Zheng, G.-Q. (2014). Diabetes and risk of Parkinson's disease: an updated meta-analysis of case-control studies. PloS One, 9(1), e85781. doi:10.1371/journal.pone.0085781 Lu, M., Ammar, D., Ives, H., Albrecht, F., & Gluck, S. L. (2007). Physical interaction between aldolase and vacuolar H+-ATPase is essential for the assembly and activity of the proton pump. The Journal of Biological Chemistry, 282(34), 24495– 24503. doi:10.1074/jbc.M702598200 Lu, M., Sautin, Y. Y., Holliday, L. S., & Gluck, S. L. (2004). The glycolytic enzyme aldolase mediates assembly, expression, and activity of vacuolar H+-ATPase. The Journal of Biological Chemistry, 279(10), 8732–8739. doi:10.1074/jbc.M303871200 Lu, S. C. (2000). Regulation of glutathione synthesis. Current Topics in Cellular Regulation, 36, 95–116. Lucien, F., Pelletier, P.-P., Lavoie, R. R., Lacroix, J.-M., Roy, S., Parent, J.-L., et al. (2017). Hypoxia-induced mobilization of NHE6 to the plasma membrane triggers

279 References

endosome hyperacidification and chemoresistance. Nature Communications, 8, 15884. doi:10.1038/ncomms15884 Luk, K. C., Kehm, V., Carroll, J., Zhang, B., O'Brien, P., Trojanowski, J. Q., & Lee, V. M.-Y. (2012). Pathological α-synuclein transmission initiates Parkinson-like neurodegeneration in nontransgenic mice. Science (New York, N.Y.), 338(6109), 949–953. doi:10.1126/science.1227157 Luth, E. S., Stavrovskaya, I. G., Bartels, T., Kristal, B. S., & Selkoe, D. J. (2014). Soluble, prefibrillar α-synuclein oligomers promote complex I-dependent, Ca2+-induced mitochondrial dysfunction. The Journal of Biological Chemistry, 289(31), 21490– 21507. doi:10.1074/jbc.M113.545749 Luzio, J. P., Pryor, P. R., & Bright, N. A. (2007). Lysosomes: fusion and function. Nature Reviews. Molecular Cell Biology, 8(8), 622–632. doi:10.1038/nrm2217 Ma, B., Qian, D., Nan, Q., Tan, C., An, L., & Xiang, Y. (2012). Arabidopsis vacuolar H+- ATPase (V-ATPase) B subunits are involved in actin cytoskeleton remodeling via binding to, bundling, and stabilizing F-actin. The Journal of Biological Chemistry, 287(23), 19008–19017. doi:10.1074/jbc.M111.281873 Ma, K.-L., Song, L.-K., Yuan, Y.-H., Zhang, Y., Han, N., Gao, K., & Chen, N.-H. (2014). The nuclear accumulation of alpha-synuclein is mediated by importin alpha and promotes neurotoxicity by accelerating the cell cycle. Neuropharmacology, 82, 132–142. doi:10.1016/j.neuropharm.2013.07.035 Mahul-Mellier, A.-L., Vercruysse, F., Maco, B., Ait-Bouziad, N., De Roo, M., Muller, D., & Lashuel, H. A. (2015). Fibril growth and seeding capacity play key roles in α- synuclein-mediated apoptotic cell death. Cell Death and Differentiation, 22(12), 2107–2122. doi:10.1038/cdd.2015.79 Mandel, M., & Higa, A. (1970). Calcium-dependent bacteriophage DNA infection. Journal of Molecular Biology, 53(1), 159–162. Mangieri, L. R., Mader, B. J., Thomas, C. E., Taylor, C. A., Luker, A. M., Tse, T. E., et al. (2014). ATP6V0C knockdown in neuroblastoma cells alters autophagy-lysosome pathway function and metabolism of proteins that accumulate in neurodegenerative disease. PloS One, 9(4), e93257. doi:10.1371/journal.pone.0093257 Mann, M. (2006). Functional and quantitative proteomics using SILAC. Nature Reviews. Molecular Cell Biology, 7(12), 952–958. doi:10.1038/nrm2067 Manolio, T. A., Collins, F. S., Cox, N. J., Goldstein, D. B., Hindorff, L. A., Hunter, D. J., et al. (2009). Finding the missing heritability of complex diseases. Nature, 461(7265), 747–753. doi:10.1038/nature08494 Maranzana, E., Barbero, G., Falasca, A. I., Lenaz, G., & Genova, M. L. (2013). Mitochondrial respiratory supercomplex association limits production of reactive oxygen species from complex I. Antioxidants & Redox Signaling, 19(13), 1469– 1480. doi:10.1089/ars.2012.4845 Marchitti, S. A., Brocker, C., Stagos, D., & Vasiliou, V. (2008). Non-P450 aldehyde oxidizing enzymes: the aldehyde dehydrogenase superfamily. Expert Opinion on Drug Metabolism & Toxicology, 4(6), 697–720. doi:10.1517/17425255.4.6.697 Marin-Valencia, I., Cho, S. K., Rakheja, D., Hatanpaa, K. J., Kapur, P., Mashimo, T., et al. (2012). Glucose metabolism via the pentose phosphate pathway, glycolysis and

280 References

Krebs cycle in an orthotopic mouse model of human brain tumors. NMR in Biomedicine, 25(10), 1177–1186. doi:10.1002/nbm.2787 Markesbery, W. R., Jicha, G. A., Liu, H., & Schmitt, F. A. (2009). Lewy body pathology in normal elderly subjects. Journal of Neuropathology and Experimental Neurology, 68(7), 816–822. doi:10.1097/NEN.0b013e3181ac10a7 Maroteaux, L., Campanelli, J. T., & Scheller, R. H. (1988). Synuclein: a neuron-specific protein localized to the nucleus and presynaptic nerve terminal. The Journal of Neuroscience : the Official Journal of the Society for Neuroscience, 8(8), 2804– 2815. Marsden, C. D. (1990). Parkinson's disease. Lancet, 335(8695), 948–952. Marsden, C. D., & Parkes, J. D. (1977). Success and problems of long-term levodopa therapy in Parkinson's disease. Lancet, 1(8007), 345–349. Martini-Stoica, H., Xu, Y., Ballabio, A., & Zheng, H. (2016). The Autophagy-Lysosomal Pathway in Neurodegeneration: A TFEB Perspective. Trends in Neurosciences, 39(4), 221–234. doi:10.1016/j.tins.2016.02.002 Martinus, R. D., Garth, G. P., Webster, T. L., Cartwright, P., Naylor, D. J., Høj, P. B., & Hoogenraad, N. J. (1996). Selective induction of mitochondrial chaperones in response to loss of the mitochondrial genome. European Journal of Biochemistry, 240(1), 98–103. Matsui, H., Sato, F., Sato, S., Koike, M., Taruno, Y., Saiki, S., et al. (2013). ATP13A2 deficiency induces a decrease in cathepsin D activity, fingerprint-like inclusion body formation, and selective degeneration of dopaminergic neurons. FEBS Letters, 587(9), 1316–1325. doi:10.1016/j.febslet.2013.02.046 Mauvezin, C., Nagy, P., Juhász, G., & Neufeld, T. P. (2015). Autophagosome-lysosome fusion is independent of V-ATPase-mediated acidification. Nature Communications, 6, 7007. doi:10.1038/ncomms8007 Maxfield, F. R., & McGraw, T. E. (2004). Endocytic recycling. Nature Reviews. Molecular Cell Biology, 5(2), 121–132. doi:10.1038/nrm1315 Maxson, M. E., & Grinstein, S. (2014). The vacuolar-type H⁺-ATPase at a glance - more than a proton pump. Journal of Cell Science, 127(Pt 23), 4987–4993. doi:10.1242/jcs.158550 Mazhab-Jafari, M. T., Rohou, A., Schmidt, C., Bueler, S. A., Benlekbir, S., Robinson, C. V., & Rubinstein, J. L. (2016). Atomic model for the membrane-embedded VO motor of a eukaryotic V-ATPase. Nature, 539(7627), 118–122. doi:10.1038/nature19828 Mazurie, A., Bottani, S., & Vergassola, M. (2005). An evolutionary and functional assessment of regulatory network motifs. Genome Biology, 6(4), R35. doi:10.1186/gb-2005-6-4-r35 Mazzio, E., & Soliman, K. F. A. (2003). D-(+)-glucose rescue against 1-methyl-4- phenylpyridinium toxicity through anaerobic glycolysis in neuroblastoma cells. Brain Research, 962(1-2), 48–60. Mazzulli, J. R., Zunke, F., Isacson, O., Studer, L., & Krainc, D. (2016). α-Synuclein- induced lysosomal dysfunction occurs through disruptions in protein trafficking in human midbrain synucleinopathy models. Proceedings of the National Academy of Sciences of the United States of America, 113(7), 1931–1936.

281 References

doi:10.1073/pnas.1520335113 McBrayer, M., & Nixon, R. A. (2013). Lysosome and calcium dysregulation in Alzheimer's disease: partners in crime. Biochemical Society Transactions, 41(6), 1495–1502. doi:10.1042/BST20130201 McCloy, R. A., Rogers, S., Caldon, C. E., Lorca, T., Castro, A., & Burgess, A. (2014). Partial inhibition of Cdk1 in G 2 phase overrides the SAC and decouples mitotic events. Cell Cycle (Georgetown, Tex.), 13(9), 1400–1412. doi:10.4161/cc.28401 McFarland, M. A., Ellis, C. E., Markey, S. P., & Nussbaum, R. L. (2008). Proteomics analysis identifies phosphorylation-dependent alpha-synuclein protein interactions. Molecular & Cellular Proteomics : MCP, 7(11), 2123–2137. doi:10.1074/mcp.M800116-MCP200 McGarel, C., Pentieva, K., Strain, J. J., & McNulty, H. (2015). Emerging roles for folate and related B-vitamins in brain health across the lifecycle. The Proceedings of the Nutrition Society, 74(1), 46–55. doi:10.1017/S0029665114001554 McGeer, P. L., Zeldowicz, L., & McGeer, E. G. (1972). A clinical trial of folic acid in Parkinson's disease. Canadian Medical Association Journal, 106(2), 145–6 passim. McGlinchey, R. P., & Lee, J. C. (2015). Cysteine cathepsins are essential in lysosomal degradation of α-synuclein. Proceedings of the National Academy of Sciences of the United States of America, 112(30), 9322–9327. doi:10.1073/pnas.1500937112 McLelland, G.-L., Soubannier, V., Chen, C. X., McBride, H. M., & Fon, E. A. (2014). Parkin and PINK1 function in a vesicular trafficking pathway regulating mitochondrial quality control. The EMBO Journal, 33(4), 282–295. doi:10.1002/embj.201385902 Meissner, C., Lorenz, H., Weihofen, A., Selkoe, D. J., & Lemberg, M. K. (2011). The mitochondrial intramembrane protease PARL cleaves human Pink1 to regulate Pink1 trafficking. Journal of Neurochemistry, 117(5), 856–867. doi:10.1111/j.1471- 4159.2011.07253.x Melber, A., & Winge, D. R. (2016). Inner Secrets of the Respirasome. Cell, 167(6), 1450–1452. doi:10.1016/j.cell.2016.11.025 Mellman, I., Fuchs, R., & Helenius, A. (1986). Acidification of the endocytic and exocytic pathways. Annual Review of Biochemistry, 55(1), 663–700. doi:10.1146/annurev.bi.55.070186.003311 Meng, H., Yamashita, C., Shiba-Fukushima, K., Inoshita, T., Funayama, M., Sato, S., et al. (2017). Loss of Parkinson's disease-associated protein CHCHD2 affects mitochondrial crista structure and destabilizes cytochrome c. Nature Communications, 8, 15500. doi:10.1038/ncomms15500 Michaeli, S., Oz, G., Sorce, D. J., Garwood, M., Ugurbil, K., Majestic, S., & Tuite, P. (2007). Assessment of brain iron and neuronal integrity in patients with Parkinson's disease using novel MRI contrasts. Movement Disorders : Official Journal of the Movement Disorder Society, 22(3), 334–340. doi:10.1002/mds.21227 Michelangeli, F., Ogunbayo, O. A., & Wootton, L. L. (2005). A plethora of interacting organellar Ca2+ stores. Current Opinion in Cell Biology, 17(2), 135–140. doi:10.1016/j.ceb.2005.01.005 Miclet, E., Stoven, V., Michels, P. A., Opperdoes, F. R., Lallemand, J. Y., & Duffieux, F. (2001). NMR spectroscopic analysis of the first two steps of the pentose-

282 References

phosphate pathway elucidates the role of 6-phosphogluconolactonase. The Journal of Biological Chemistry, 276(37), 34840–34846. doi:10.1074/jbc.M105174200 Milgrom, E., Diab, H., Middleton, F., & Kane, P. M. (2007). Loss of vacuolar proton- translocating ATPase activity in yeast results in chronic oxidative stress. The Journal of Biological Chemistry, 282(10), 7125–7136. doi:10.1074/jbc.M608293200 Minard, A. Y., Tan, S.-X., Yang, P., Fazakerley, D. J., Domanova, W., Parker, B. L., et al. (2016). mTORC1 Is a Major Regulatory Node in the FGF21 Signaling Network in Adipocytes. Cell Reports, 17(1), 29–36. doi:10.1016/j.celrep.2016.08.086 Miranda, K. C., Karet, F. E., & Brown, D. (2010). An extended nomenclature for mammalian V-ATPase subunit genes and splice variants. PloS One, 5(3), e9531. doi:10.1371/journal.pone.0009531 Miura, E., Hasegawa, T., Konno, M., Suzuki, M., Sugeno, N., Fujikake, N., et al. (2014). VPS35 dysfunction impairs lysosomal degradation of α-synuclein and exacerbates neurotoxicity in a Drosophila model of Parkinson's disease. Neurobiology of Disease, 71, 1–13. doi:10.1016/j.nbd.2014.07.014 Miyake, Y., Tanaka, K., Fukushima, W., Kiyohara, C., Sasaki, S., Tsuboi, Y., et al. (2012). UCHL1 S18Y variant is a risk factor for Parkinson's disease in Japan. BMC Neurology, 12(1), 62. doi:10.1186/1471-2377-12-62 Mizuno, Y., Ohta, S., Tanaka, M., Takamiya, S., Suzuki, K., Sato, T., et al. (1989). Deficiencies in complex I subunits of the respiratory chain in Parkinson's disease. Biochemical and Biophysical Research Communications, 163(3), 1450–1455. Mochizuki, H., & Yasuda, T. (2012). Iron accumulation in Parkinson's disease. Journal of Neural Transmission (Vienna, Austria : 1996), 119(12), 1511–1514. doi:10.1007/s00702-012-0905-9 Modjtahedi, N., Tokatlidis, K., Dessen, P., & Kroemer, G. (2016). Mitochondrial Proteins Containing Coiled-Coil-Helix-Coiled-Coil-Helix (CHCH) Domains in Health and Disease. Trends in Biochemical Sciences, 41(3), 245–260. doi:10.1016/j.tibs.2015.12.004 Molina, J. A., Jiménez-Jiménez, F. J., Gomez, P., Vargas, C., Navarro, J. A., Ortí-Pareja, M., et al. (1997). Decreased cerebrospinal fluid levels of neutral and basic amino acids in patients with Parkinson's disease. Journal of the Neurological Sciences, 150(2), 123–127. Mordas, A., & Tokatlidis, K. (2015). The MIA pathway: a key regulator of mitochondrial oxidative protein folding and biogenesis. Accounts of Chemical Research, 48(8), 2191–2199. doi:10.1021/acs.accounts.5b00150 Morgan, J., & Sethi, K. D. (2005). Levodopa and the progression of Parkinson's disease. Current Neurology and Neuroscience Reports, 5(4), 261–262. Moullan, N., Mouchiroud, L., Wang, X., Ryu, D., Williams, E. G., Mottis, A., et al. (2015). Tetracyclines Disturb Mitochondrial Function across Eukaryotic Models: A Call for Caution in Biomedical Research. Cell Reports, 10(10), 1681–1691. doi:10.1016/j.celrep.2015.02.034 Mu, L., Sobotka, S., Chen, J., Su, H., Sanders, I., Adler, C. H., et al. (2013). Alpha- synuclein pathology and axonal degeneration of the peripheral motor nerves innervating pharyngeal muscles in Parkinson disease. Journal of Neuropathology and Experimental Neurology, 72(2), 119–129.

283 References

doi:10.1097/NEN.0b013e3182801cde Mueller, B. H., Park, Y., Daudt, D. R., Ma, H.-Y., Akopova, I., Stankowska, D. L., et al. (2013). Sigma-1 receptor stimulation attenuates calcium influx through activated L-type Voltage Gated Calcium Channels in purified retinal ganglion cells. Experimental Eye Research, 107, 21–31. doi:10.1016/j.exer.2012.11.002 Mulcahy, P., O'Doherty, A., Paucard, A., O'Brien, T., Kirik, D., & Dowd, E. (2012). Development and characterisation of a novel rat model of Parkinson's disease induced by sequential intranigral administration of AAV-α-synuclein and the pesticide, rotenone. Neuroscience, 203, 170–179. doi:10.1016/j.neuroscience.2011.12.011 Mulcahy, P., O'Doherty, A., Paucard, A., O'Brien, T., Kirik, D., & Dowd, E. (2013). The behavioural and neuropathological impact of intranigral AAV-α-synuclein is exacerbated by systemic infusion of the Parkinson's disease-associated pesticide, rotenone, in rats. Behavioural Brain Research, 243, 6–15. doi:10.1016/j.bbr.2012.12.051 Muller, F. L., Liu, Y., & Van Remmen, H. (2004). Complex III releases superoxide to both sides of the inner mitochondrial membrane. The Journal of Biological Chemistry, 279(47), 49064–49073. doi:10.1074/jbc.M407715200 Murakami, K., Miyake, Y., Sasaki, S., Tanaka, K., Fukushima, W., Kiyohara, C., et al. (2010). Dietary intake of folate, vitamin B6, vitamin B12 and riboflavin and risk of Parkinson's disease: a case-control study in Japan. The British Journal of Nutrition, 104(5), 757–764. doi:10.1017/S0007114510001005 Murakami, S., Miyazaki, I., Miyoshi, K., & Asanuma, M. (2015). Long-Term Systemic Exposure to Rotenone Induces Central and Peripheral Pathology of Parkinson's Disease in Mice. Neurochemical Research, 40(6), 1165–1178. doi:10.1007/s11064- 015-1577-2 Murphy, K. E., Gysbers, A. M., Abbott, S. K., Spiro, A. S., Furuta, A., Cooper, A., et al. (2015). Lysosomal-associated membrane protein 2 isoforms are differentially affected in early Parkinson's disease. Movement Disorders : Official Journal of the Movement Disorder Society, 30(12), 1639–1647. doi:10.1002/mds.26141 Nagatsua, T., & Sawadab, M. (2009). L-dopa therapy for Parkinson's disease: past, present, and future. Parkinsonism & Related Disorders, 15 Suppl 1, S3–8. doi:10.1016/S1353-8020(09)70004-5 Nakamura, K., Nemani, V. M., Azarbal, F., Skibinski, G., Levy, J. M., Egami, K., et al. (2011). Direct membrane association drives mitochondrial fission by the Parkinson disease-associated protein alpha-synuclein. The Journal of Biological Chemistry, 286(23), 20710–20726. doi:10.1074/jbc.M110.213538 Napolitano, G., & Ballabio, A. (2016). TFEB at a glance. Journal of Cell Science, 129(13), 2475–2481. doi:10.1242/jcs.146365 Napolitano, G., Johnson, J. L., He, J., Rocca, C. J., Monfregola, J., Pestonjamasp, K., et al. (2015). Impairment of chaperone-mediated autophagy leads to selective lysosomal degradation defects in the lysosomal storage disease cystinosis. EMBO Molecular Medicine, 7(2), 158–174. doi:10.15252/emmm.201404223 Nelson, N. (1992). Structure and function of V-ATPases in endocytic and secretory organelles. The Journal of Experimental Biology, 172, 149–153.

284 References

Neubrand, V. E., Cesca, F., Benfenati, F., & Schiavo, G. (2012). Kidins220/ARMS as a functional mediator of multiple receptor signalling pathways. Journal of Cell Science, 125(Pt 8), 1845–1854. doi:10.1242/jcs.102764 Neudorfer, O., Giladi, N., Elstein, D., Abrahamov, A., Turezkite, T., Aghai, E., et al. (1996). Occurrence of Parkinson's syndrome in type I Gaucher disease. QJM : Monthly Journal of the Association of Physicians, 89(9), 691–694. Nguyen, F., Starosta, A. L., Arenz, S., Sohmen, D., Dönhöfer, A., & Wilson, D. N. (2014). Tetracycline antibiotics and resistance mechanisms. Biological Chemistry, 395(5), 559–575. doi:10.1515/hsz-2013-0292 Niranjan, R. (2014). The role of inflammatory and oxidative stress mechanisms in the pathogenesis of Parkinson's disease: focus on astrocytes. Molecular Neurobiology, 49(1), 28–38. doi:10.1007/s12035-013-8483-x Noelker, C., Lu, L., Höllerhage, M., Vulinovic, F., Sturn, A., Roscher, R., et al. (2015). Glucocerebrosidase deficiency and mitochondrial impairment in experimental Parkinson disease. Journal of the Neurological Sciences. doi:10.1016/j.jns.2015.06.030 Noyce, A. J., Bestwick, J. P., Silveira-Moriyama, L., Hawkes, C. H., Giovannoni, G., Lees, A. J., & Schrag, A. (2012). Meta-analysis of early nonmotor features and risk factors for Parkinson disease. Annals of Neurology, 72(6), 893–901. doi:10.1002/ana.23687 O'Mahony, S. M., Clarke, G., Borre, Y. E., Dinan, T. G., & Cryan, J. F. (2015). Serotonin, tryptophan metabolism and the brain-gut-microbiome axis. Behavioural Brain Research, 277, 32–48. doi:10.1016/j.bbr.2014.07.027 Ogaki, K., Koga, S., Heckman, M. G., Fiesel, F. C., Ando, M., Labbé, C., et al. (2015). Mitochondrial targeting sequence variants of the CHCHD2 gene are a risk for Lewy body disorders. Neurology, 85(23), 2016–2025. doi:10.1212/WNL.0000000000002170 Oliveira, S. A., Li, Y.-J., Noureddine, M. A., Zuchner, S., Qin, X., Pericak-Vance, M. A., & Vance, J. M. (2005). Identification of risk and age-at-onset genes on chromosome 1p in Parkinson disease. American Journal of Human Genetics, 77(2), 252–264. doi:10.1086/432588 Olsen, J. H., Friis, S., & Frederiksen, K. (2006). Malignant melanoma and other types of cancer preceding Parkinson disease. Epidemiology (Cambridge, Mass.), 17(5), 582– 587. doi:10.1097/01.ede.0000229445.90471.5e Olson, L. J., Hindsgaul, O., Dahms, N. M., & Kim, J.-J. P. (2008). Structural insights into the mechanism of pH-dependent ligand binding and release by the cation- dependent mannose 6-phosphate receptor. The Journal of Biological Chemistry, 283(15), 10124–10134. doi:10.1074/jbc.M708994200 Ong, S.-E. (2012). The expanding field of SILAC. Analytical and Bioanalytical Chemistry, 404(4), 967–976. doi:10.1007/s00216-012-5998-3 Ong, S.-E., Blagoev, B., Kratchmarova, I., Kristensen, D. B., Steen, H., Pandey, A., & Mann, M. (2002). Stable isotope labeling by amino acids in cell culture, SILAC, as a simple and accurate approach to expression proteomics. Molecular & Cellular Proteomics : MCP, 1(5), 376–386. Osterhues, A., Ali, N. S., & Michels, K. B. (2013). The role of folic acid fortification in

285 References

neural tube defects: a review. Critical Reviews in Food Science and Nutrition, 53(11), 1180–1190. doi:10.1080/10408398.2011.575966 Pal, R., Bajaj, L., Sharma, J., Palmieri, M., di Ronza, A., Lotfi, P., et al. (2016). NADPH oxidase promotes Parkinsonian phenotypes by impairing autophagic flux in an mTORC1-independent fashion in a cellular model of Parkinson's disease. Scientific Reports, 6(1), 22866. doi:10.1038/srep22866 Pan-Montojo, F., Anichtchik, O., Dening, Y., Knels, L., Pursche, S., Jung, R., et al. (2010). Progression of Parkinson's disease pathology is reproduced by intragastric administration of rotenone in mice. PloS One, 5(1), e8762. doi:10.1371/journal.pone.0008762 Pang, S., Chen, D., Zhang, A., Qin, X., & Yan, B. (2012). Genetic analysis of the LAMP-2 gene promoter in patients with sporadic Parkinson's disease. Neuroscience Letters, 526(1), 63–67. doi:10.1016/j.neulet.2012.07.044 Park, J.-S., Koentjoro, B., Davis, R. L., & Sue, C. M. (2016). Loss of ATP13A2 impairs glycolytic function in Kufor-Rakeb syndrome patient-derived cell models. Parkinsonism & Related Disorders, 27, 67–73. doi:10.1016/j.parkreldis.2016.03.018 Parker, W. D., Boyson, S. J., & Parks, J. K. (1989). Abnormalities of the electron transport chain in idiopathic Parkinson's disease. Annals of Neurology, 26(6), 719– 723. doi:10.1002/ana.410260606 Parkinson Study Group QE3 Investigators, Beal, M. F., Oakes, D., Shoulson, I., Henchcliffe, C., Galpern, W. R., et al. (2014). A randomized clinical trial of high- dosage coenzyme Q10 in early Parkinson disease: no evidence of benefit. JAMA Neurology, 71(5), 543–552. doi:10.1001/jamaneurol.2014.131 Parkkinen, L., O'Sullivan, S. S., Kuoppamäki, M., Collins, C., Kallis, C., Holton, J. L., et al. (2011). Does levodopa accelerate the pathologic process in Parkinson disease brain? Neurology, 77(15), 1420–1426. doi:10.1212/WNL.0b013e318232ab4c Parra, K. J., & Kane, P. M. (1998). Reversible association between the V1 and V0 domains of yeast vacuolar H+-ATPase is an unconventional glucose-induced effect. Molecular and Cellular Biology, 18(12), 7064–7074. Parsell, D. A., & Sauer, R. T. (1989). The structural stability of a protein is an important determinant of its proteolytic susceptibility in Escherichia coli. The Journal of Biological Chemistry, 264(13), 7590–7595. Peng, J., Liu, Q., Rao, M. S., & Zeng, X. (2013). Using human pluripotent stem cell- derived dopaminergic neurons to evaluate candidate Parkinson's disease therapeutic agents in MPP+ and rotenone models. Journal of Biomolecular Screening, 18(5), 522–533. doi:10.1177/1087057112474468 Peng, T.-I., & Jou, M.-J. (2010). Oxidative stress caused by mitochondrial calcium overload. Annals of the New York Academy of Sciences, 1201(1), 183–188. doi:10.1111/j.1749-6632.2010.05634.x Peña-Llopis, S., Vega-Rubin-de-Celis, S., Schwartz, J. C., Wolff, N. C., Tran, T. A. T., Zou, L., et al. (2011). Regulation of TFEB and V-ATPases by mTORC1. The EMBO Journal, 30(16), 3242–3258. doi:10.1038/emboj.2011.257 Perfeito, R., Ribeiro, M., & Rego, A. C. (2017). Alpha-synuclein-induced oxidative stress correlates with altered superoxide dismutase and glutathione synthesis in human neuroblastoma SH-SY5Y cells. Archives of Toxicology, 91(3), 1245–1259.

286 References

doi:10.1007/s00204-016-1788-6 Peters, C., Bayer, M. J., Bühler, S., Andersen, J. S., Mann, M., & Mayer, A. (2001). Trans- complex formation by proteolipid channels in the terminal phase of membrane fusion. Nature, 409(6820), 581–588. doi:10.1038/35054500 Picconi, B., Piccoli, G., & Calabresi, P. (2012). Synaptic dysfunction in Parkinson's disease. Advances in Experimental Medicine and Biology, 970(Chapter 24), 553– 572. doi:10.1007/978-3-7091-0932-8_24 Pickrell, A. M., & Youle, R. J. (2015). The roles of PINK1, parkin, and mitochondrial fidelity in Parkinson's disease. Neuron, 85(2), 257–273. doi:10.1016/j.neuron.2014.12.007 Piechura, H., Warscheid, B., & Oeljeklaus, S. (2012). Quantitative Methods in Proteomics. In K. Marcus, SILAC for the Study of Mammalian Cell Lines and Yeast Protein Complexes (Vol. 893, pp. 201–221). Germany: Humana Press. Pietiäinen, V., Vassilev, B., Blom, T., Wang, W., Nelson, J., Bittman, R., et al. (2013). NDRG1 functions in LDL receptor trafficking by regulating endosomal recycling and degradation. Journal of Cell Science, 126(Pt 17), 3961–3971. doi:10.1242/jcs.128132 Pittman, J. K. (2011). Vacuolar Ca(2+) uptake. Cell Calcium, 50(2), 139–146. doi:10.1016/j.ceca.2011.01.004 Pivtoraiko, V. N., Harrington, A. J., Mader, B. J., Luker, A. M., Caldwell, G. A., Caldwell, K. A., et al. (2010). Low-dose bafilomycin attenuates neuronal cell death associated with autophagy-lysosome pathway dysfunction. Journal of Neurochemistry, 114(4), 1193–1204. doi:10.1111/j.1471-4159.2010.06838.x Poëa-Guyon, S., Amar, M., Fossier, P., & Morel, N. (2006). Alternative splicing controls neuronal expression of v-ATPase subunit a1 and sorting to nerve terminals. The Journal of Biological Chemistry, 281(25), 17164–17172. doi:10.1074/jbc.M600927200 Polymeropoulos, M. H., Lavedan, C., Leroy, E., Ide, S. E., Dehejia, A., Dutra, A., et al. (1997). Mutation in the alpha-synuclein gene identified in families with Parkinson's disease. Science (New York, N.Y.), 276(5321), 2045–2047. Poorkaj, P., Raskind, W. H., Leverenz, J. B., Matsushita, M., Zabetian, C. P., Samii, A., et al. (2010). A novel X-linked four-repeat tauopathy with Parkinsonism and spasticity. Movement Disorders : Official Journal of the Movement Disorder Society, 25(10), 1409–1417. doi:10.1002/mds.23085 POSKANZER, D. C., & SCHWAB, R. S. (1963). Cohort analysis of Parkinson's syndrome: Evidence for a single etiology related to subclinical infection about 1920. Journal of Chronic Diseases, 16, 961–973. Pradhan, T., Jung, H. S., Jang, J. H., Kim, T. W., Kang, C., & Kim, J. S. (2014). Chemical sensing of neurotransmitters. Chemical Society Reviews, 43(13), 4684–4713. doi:10.1039/c3cs60477b Prasad, H., & Rao, R. (2015). The Na+/H+ exchanger NHE6 modulates endosomal pH to control processing of amyloid precursor protein in a cell culture model of Alzheimer disease. The Journal of Biological Chemistry, 290(9), 5311–5327. doi:10.1074/jbc.M114.602219 Preface. Lysosomes and lysosomal diseases. (2015). Preface. Lysosomes and lysosomal

287 References

diseases. Methods in Cell Biology, 126, xvii–xviii. doi:10.1016/S0091- 679X(15)00046-1 Pringsheim, T., Jette, N., Frolkis, A., & Steeves, T. D. L. (2014). The prevalence of Parkinson's disease: a systematic review and meta-analysis. Movement Disorders : Official Journal of the Movement Disorder Society, 29(13), 1583–1590. doi:10.1002/mds.25945 Progida, C., Malerød, L., Stuffers, S., Brech, A., Bucci, C., & Stenmark, H. (2007). RILP is required for the proper morphology and function of late endosomes. Journal of Cell Science, 120(Pt 21), 3729–3737. doi:10.1242/jcs.017301 Pu, J., Guardia, C. M., Keren-Kaplan, T., & Bonifacino, J. S. (2016). Mechanisms and functions of lysosome positioning. Journal of Cell Science, 129(23), 4329–4339. doi:10.1242/jcs.196287 Puschmann, A. (2013). Monogenic Parkinson's disease and parkinsonism: clinical phenotypes and frequencies of known mutations. Parkinsonism & Related Disorders, 19(4), 407–415. doi:10.1016/j.parkreldis.2013.01.020 Puschmann, A., Dickson, D. W., Englund, E., Wszolek, Z. K., & Ross, O. A. (2015). CHCHD2 and Parkinson's disease. The Lancet. Neurology, 14(7), 679. doi:10.1016/S1474-4422(15)00095-2 Qu, H.-Q., Tien, M., & Polychronakos, C. (2010). Statistical significance in genetic association studies. Clinical and Investigative Medicine. Medecine Clinique Et Experimentale, 33(5), E266–70. Quenault, T., Lithgow, T., & Traven, A. (2011). PUF proteins: repression, activation and mRNA localization. Trends in Cell Biology, 21(2), 104–112. doi:10.1016/j.tcb.2010.09.013 Quik, M. (2004). Smoking, nicotine and Parkinson's disease. Trends in Neurosciences, 27(9), 561–568. doi:10.1016/j.tins.2004.06.008 Raffaello, A., Mammucari, C., Gherardi, G., & Rizzuto, R. (2016). Calcium at the Center of Cell Signaling: Interplay between Endoplasmic Reticulum, Mitochondria, and Lysosomes. Trends in Biochemical Sciences, 41(12), 1035–1049. doi:10.1016/j.tibs.2016.09.001 Ramos, M. I., Allen, L. H., Mungas, D. M., Jagust, W. J., Haan, M. N., Green, R., & Miller, J. W. (2005). Low folate status is associated with impaired cognitive function and dementia in the Sacramento Area Latino Study on Aging. The American Journal of Clinical Nutrition, 82(6), 1346–1352. Rana, H. Q., Balwani, M., Bier, L., & Alcalay, R. N. (2013). Age-specific Parkinson disease risk in GBA mutation carriers: information for genetic counseling. Genetics in Medicine : Official Journal of the American College of Medical Genetics, 15(2), 146– 149. doi:10.1038/gim.2012.107 Rao, K. V. R., Panickar, K. S., Jayakumar, A. R., & Norenberg, M. D. (2005). Astrocytes protect neurons from ammonia toxicity. Neurochemical Research, 30(10), 1311– 1318. doi:10.1007/s11064-005-8803-2 Reynolds, A. D., Glanzer, J. G., Kadiu, I., Ricardo-Dukelow, M., Chaudhuri, A., Ciborowski, P., et al. (2008). Nitrated alpha-synuclein-activated microglial profiling for Parkinson's disease. Journal of Neurochemistry, 104(6), 1504–1525. doi:10.1111/j.1471-4159.2007.05087.x

288 References

Reynolds, E. H. (2014). The neurology of folic acid deficiency. Handbook of Clinical Neurology, 120, 927–943. doi:10.1016/B978-0-7020-4087-0.00061-9 Rhodes, S. L., & Ritz, B. (2008). Genetics of iron regulation and the possible role of iron in Parkinson's disease. Neurobiology of Disease, 32(2), 183–195. doi:10.1016/j.nbd.2008.07.001 Richo, G., & Conner, G. E. (1991). Proteolytic activation of human procathepsin D. Advances in Experimental Medicine and Biology, 306, 289–296. Richter, F., Hamann, M., & Richter, A. (2007). Chronic rotenone treatment induces behavioral effects but no pathological signs of parkinsonism in mice. Journal of Neuroscience Research, 85(3), 681–691. doi:10.1002/jnr.21159 Riley, B. E., Gardai, S. J., Emig-Agius, D., Bessarabova, M., Ivliev, A. E., Schüle, B., et al. (2014). Systems-based analyses of brain regions functionally impacted in Parkinson's disease reveals underlying causal mechanisms. PloS One, 9(8), e102909. doi:10.1371/journal.pone.0102909 Rimessi, A., Bonora, M., Marchi, S., Patergnani, S., Marobbio, C. M. T., Lasorsa, F. M., & Pinton, P. (2013). Perturbed mitochondrial Ca2+ signals as causes or consequences of mitophagy induction. Autophagy, 9(11), 1677–1686. doi:10.4161/auto.24795 Rockenstein, E., Nuber, S., Overk, C. R., Ubhi, K., Mante, M., Patrick, C., et al. (2014). Accumulation of oligomer-prone α-synuclein exacerbates synaptic and neuronal degeneration in vivo. Brain : a Journal of Neurology, 137(Pt 5), 1496–1513. doi:10.1093/brain/awu057 Roczniak-Ferguson, A., Petit, C. S., Froehlich, F., Qian, S., Ky, J., Angarola, B., et al. (2012). The transcription factor TFEB links mTORC1 signaling to transcriptional control of lysosome homeostasis. Science Signaling, 5(228), ra42–ra42. doi:10.1126/scisignal.2002790 Rosenbloom, B. E., & Weinreb, N. J. (2013). Gaucher disease: a comprehensive review. Critical Reviews in Oncogenesis, 18(3), 163–175. Ross, G. W., Petrovitch, H., Abbott, R. D., Nelson, J., Markesbery, W., Davis, D., et al. (2004). Parkinsonian signs and substantia nigra neuron density in decendents elders without PD. Annals of Neurology, 56(4), 532–539. doi:10.1002/ana.20226 Rousseaux, M. W., de Haro, M., Lasagna-Reeves, C. A., De Maio, A., Park, J., Jafar- Nejad, P., et al. (2016). TRIM28 regulates the nuclear accumulation and toxicity of both alpha-synuclein and tau. eLife, 5, 18748. doi:10.7554/eLife.19809 Rozycka, A., Jagodzinski, P. P., Kozubski, W., Lianeri, M., & Dorszewska, J. (2013). Homocysteine Level and Mechanisms of Injury in Parkinson's Disease as Related to MTHFR, MTR, and MTHFD1 Genes Polymorphisms and L-Dopa Treatment. Current Genomics, 14(8), 534–542. doi:10.2174/1389202914666131210210559 Rubino, E., Brusa, L., Zhang, M., Boschi, S., Govone, F., Vacca, A., et al. (2017). Genetic analysis of CHCHD2 and CHCHD10 in Italian patients with Parkinson's disease. Neurobiology of Aging, 53, 193.e7–193.e8. doi:10.1016/j.neurobiolaging.2016.12.027 Russo, J., & Olivas, W. M. (2015). Conditional regulation of Puf1p, Puf4p, and Puf5p activity alters YHB1 mRNA stability for a rapid response to toxic nitric oxide stress in yeast. Molecular Biology of the Cell, 26(6), 1015–1029. doi:10.1091/mbc.E14-10- 1452

289 References

Ryan, S. D., Dolatabadi, N., Chan, S. F., Zhang, X., Akhtar, M. W., Parker, J., et al. (2013). Isogenic human iPSC Parkinson's model shows nitrosative stress-induced dysfunction in MEF2-PGC1α transcription. Cell, 155(6), 1351–1364. doi:10.1016/j.cell.2013.11.009 Ryvlin, P., Broussolle, E., Piollet, H., Viallet, F., Khalfallah, Y., & Chazot, G. (1995). Magnetic resonance imaging evidence of decreased putamenal iron content in idiopathic Parkinson's disease. Archives of Neurology, 52(6), 583–588. Sala, G., Arosio, A., Stefanoni, G., Melchionda, L., Riva, C., Marinig, D., et al. (2013). Rotenone upregulates alpha-synuclein and myocyte enhancer factor 2D independently from lysosomal degradation inhibition. BioMed Research International, 2013(7), 846725–10. doi:10.1155/2013/846725 Sanders, L. H., & Greenamyre, J. T. (2013). Oxidative damage to macromolecules in human Parkinson disease and the rotenone model. Free Radical Biology & Medicine, 62, 111–120. doi:10.1016/j.freeradbiomed.2013.01.003 Sanders, L. H., McCoy, J., Hu, X., Mastroberardino, P. G., Dickinson, B. C., Chang, C. J., et al. (2014). Mitochondrial DNA damage: molecular marker of vulnerable nigral neurons in Parkinson's disease. Neurobiology of Disease, 70, 214–223. doi:10.1016/j.nbd.2014.06.014 Santiago, J. A., & Potashkin, J. A. (2014). System-based approaches to decode the molecular links in Parkinson's disease and diabetes. Neurobiology of Disease, 72 Pt A, 84–91. doi:10.1016/j.nbd.2014.03.019 Santiago, R. M., Barbieiro, J., Lima, M. M. S., Dombrowski, P. A., Andreatini, R., & Vital, M. A. B. F. (2010). Depressive-like behaviors alterations induced by intranigral MPTP, 6-OHDA, LPS and rotenone models of Parkinson's disease are predominantly associated with serotonin and dopamine. Progress in Neuro- Psychopharmacology & Biological Psychiatry, 34(6), 1104–1114. doi:10.1016/j.pnpbp.2010.06.004 Sardiello, M., Palmieri, M., di Ronza, A., Medina, D. L., Valenza, M., Gennarino, V. A., et al. (2009). A gene network regulating lysosomal biogenesis and function. Science (New York, N.Y.), 325(5939), 473–477. doi:10.1126/science.1174447 Sargsyan, A., Cai, J., Fandino, L. B., Labasky, M. E., Forostyan, T., Colosimo, L. K., et al. (2015). Rapid parallel measurements of macroautophagy and mitophagy in mammalian cells using a single fluorescent biosensor. Scientific Reports, 5(1), 12397. doi:10.1038/srep12397 Sato, H., Kato, T., & Arawaka, S. (2013). The role of Ser129 phosphorylation of α- synuclein in neurodegeneration of Parkinson's disease: a review of in vivo models. Reviews in the Neurosciences, 24(2), 115–123. doi:10.1515/revneuro-2012-0071 Sato, Y., Suzuki, Y., Ito, E., Shimazaki, S., Ishida, M., Yamamoto, T., et al. (2006). Identification and characterization of an increased glycoprotein in aging: age- associated translocation of cathepsin D. Mechanisms of Ageing and Development, 127(10), 771–778. doi:10.1016/j.mad.2006.07.001 Schapira, A. H. V. (2008). Mitochondria in the aetiology and pathogenesis of Parkinson's disease. The Lancet. Neurology, 7(1), 97–109. doi:10.1016/S1474- 4422(07)70327-7 Schapira, A. H., Cooper, J. M., Dexter, D., Clark, J. B., Jenner, P., & Marsden, C. D.

290 References

(1990). Mitochondrial complex I deficiency in Parkinson's disease. Journal of Neurochemistry, 54(3), 823–827. Schägger, H., & Pfeiffer, K. (2000). Supercomplexes in the respiratory chains of yeast and mammalian mitochondria. The EMBO Journal, 19(8), 1777–1783. doi:10.1093/emboj/19.8.1777 Scheel, J., Matteoni, R., Ludwig, T., Hoflack, B., & Kreis, T. E. (1990). Microtubule depolymerization inhibits transport of cathepsin D from the Golgi apparatus to lysosomes. Journal of Cell Science, 96 ( Pt 4), 711–720. Scherzer, C. R., Grass, J. A., Liao, Z., Pepivani, I., Zheng, B., Eklund, A. C., et al. (2008). GATA transcription factors directly regulate the Parkinson's disease-linked gene alpha-synuclein. Proceedings of the National Academy of Sciences of the United States of America, 105(31), 10907–10912. doi:10.1073/pnas.0802437105 Schneider, C. A., Rasband, W. S., & Eliceiri, K. W. (2012). NIH Image to ImageJ: 25 years of image analysis. Nature Methods, 9(7), 671–675. Schöndorf, D. C., Aureli, M., McAllister, F. E., Hindley, C. J., Mayer, F., Schmid, B., et al. (2014). iPSC-derived neurons from GBA1-associated Parkinson's disease patients show autophagic defects and impaired calcium homeostasis. Nature Communications, 5, 4028. doi:10.1038/ncomms5028 Schrag, A., & Schott, J. M. (2006). Epidemiological, clinical, and genetic characteristics of early-onset parkinsonism. The Lancet. Neurology, 5(4), 355–363. doi:10.1016/S1474-4422(06)70411-2 Schulte, C., & Gasser, T. (2011). Genetic basis of Parkinson's disease: inheritance, penetrance, and expression. The Application of Clinical Genetics, 4, 67–80. doi:10.2147/TACG.S11639 Schwarz, J., Trenkwalder, C., Gasser, T., Arnold, G., & Oertel, W. H. (1992). Folinic acid therapy fails to improve early Parkinson's disease: a two week placebo controlled clinical trial. Journal of Neural Transmission. Parkinson's Disease and Dementia Section, 4(1), 35–41. Scott, C. C., Vacca, F., & Gruenberg, J. (2014). Endosome maturation, transport and functions. Seminars in Cell & Developmental Biology, 31, 2–10. doi:10.1016/j.semcdb.2014.03.034 Segura-Aguilar, J., Paris, I., Muñoz, P., Ferrari, E., Zecca, L., & Zucca, F. A. (2014). Protective and toxic roles of dopamine in Parkinson's disease. Journal of Neurochemistry, 129(6), 898–915. doi:10.1111/jnc.12686 Seidler, R. D., Bernard, J. A., Burutolu, T. B., Fling, B. W., Gordon, M. T., Gwin, J. T., et al. (2010). Motor control and aging: links to age-related brain structural, functional, and biochemical effects. Neuroscience and Biobehavioral Reviews, 34(5), 721–733. doi:10.1016/j.neubiorev.2009.10.005 Seol, J. H., Shevchenko, A., & Deshaies, R. J. (2001). Skp1 forms multiple protein complexes, including RAVE, a regulator of V-ATPase assembly. Nature Cell Biology, 3(4), 384–391. doi:10.1038/35070067 Sevlever, D., Jiang, P., & Yen, S.-H. C. (2008). Cathepsin D is the main lysosomal enzyme involved in the degradation of alpha-synuclein and generation of its carboxy- terminally truncated species. Biochemistry, 47(36), 9678–9687. doi:10.1021/bi800699v

291 References

Shacka, J. J., Klocke, B. J., Shibata, M., Uchiyama, Y., Datta, G., Schmidt, R. E., & Roth, K. A. (2006). Bafilomycin A1 inhibits chloroquine-induced death of cerebellar granule neurons. Molecular Pharmacology, 69(4), 1125–1136. doi:10.1124/mol.105.018408 Shepard, P. J., Barshop, B. A., Baumgartner, M. R., Hansen, J.-B., Jepsen, K., Smith, E. N., & Frazer, K. A. (2015). Consanguinity and rare mutations outside of MCCC genes underlie nonspecific phenotypes of MCCD. Genetics in Medicine : Official Journal of the American College of Medical Genetics, 17(8), 660–667. doi:10.1038/gim.2014.157 Shih, J. C., Chen, K., & Ridd, M. J. (1999). Role of MAO A and B in neurotransmitter metabolism and behavior. Polish Journal of Pharmacology, 51(1), 25–29. Shoulson, I., Oakes, D., Fahn, S., Lang, A., Langston, J. W., LeWitt, P., et al. (2002). Impact of sustained deprenyl (selegiline) in levodopa-treated Parkinson's disease: a randomized placebo-controlled extension of the deprenyl and tocopherol antioxidative therapy of parkinsonism trial. Annals of Neurology, 51(5), 604–612. doi:10.1002/ana.10191 Sica, V., Galluzzi, L., Bravo-San Pedro, J. M., Izzo, V., Maiuri, M. C., & Kroemer, G. (2015). Organelle-Specific Initiation of Autophagy. Molecular Cell, 59(4), 522–539. doi:10.1016/j.molcel.2015.07.021 Sidransky, E., & Lopez, G. (2012). The link between the GBA gene and parkinsonism. The Lancet. Neurology, 11(11), 986–998. doi:10.1016/S1474-4422(12)70190-4 Sidransky, E., Nalls, M. A., Aasly, J. O., Aharon-Peretz, J., Annesi, G., Barbosa, E. R., et al. (2009). Multicenter analysis of glucocerebrosidase mutations in Parkinson's disease. The New England Journal of Medicine, 361(17), 1651–1661. doi:10.1056/NEJMoa0901281 Silva, P., & Gerós, H. (2009). Regulation by salt of vacuolar H+-ATPase and H+- pyrophosphatase activities and Na+/H+ exchange. Plant Signaling & Behavior, 4(8), 718–726. Simón-Sánchez, J., Schulte, C., Bras, J. M., Sharma, M., Gibbs, J. R., Berg, D., et al. (2009). Genome-wide association study reveals genetic risk underlying Parkinson's disease. Nature Genetics, 41(12), 1308–1312. doi:10.1038/ng.487 Singleton, A. B., Farrer, M., Johnson, J., Singleton, A., Hague, S., Kachergus, J., et al. (2003). alpha-Synuclein locus triplication causes Parkinson's disease. Science (New York, N.Y.), 302(5646), 841–841. doi:10.1126/science.1090278 Smargiassi, A., Mutti, A., De Rosa, A., De Palma, G., Negrotti, A., & Calzetti, S. (1998). A case-control study of occupational and environmental risk factors for Parkinson's disease in the Emilia-Romagna region of Italy. Neurotoxicology, 19(4-5), 709–712. Smith, A. N., Lovering, R. C., Futai, M., Takeda, J., Brown, D., & Karet, F. E. (2003). Revised nomenclature for mammalian vacuolar-type H+ -ATPase subunit genes. Molecular Cell, 12(4), 801–803. Snow, B. J., Rolfe, F. L., Lockhart, M. M., Frampton, C. M., O'Sullivan, J. D., Fung, V., et al. (2010). A double-blind, placebo-controlled study to assess the mitochondria- targeted antioxidant MitoQ as a disease-modifying therapy in Parkinson's disease. Movement Disorders : Official Journal of the Movement Disorder Society, 25(11), 1670–1674. doi:10.1002/mds.23148

292 References

Sobota, J. A., Bäck, N., Eipper, B. A., & Mains, R. E. (2009). Inhibitors of the V0 subunit of the vacuolar H+-ATPase prevent segregation of lysosomal- and secretory- pathway proteins. Journal of Cell Science, 122(Pt 19), 3542–3553. doi:10.1242/jcs.034298 Soubannier, V., McLelland, G.-L., Zunino, R., Braschi, E., Rippstein, P., Fon, E. A., & McBride, H. M. (2012). A vesicular transport pathway shuttles cargo from mitochondria to lysosomes. Current Biology : CB, 22(2), 135–141. doi:10.1016/j.cub.2011.11.057 Spillantini, M. G., Schmidt, M. L., Lee, V. M., Trojanowski, J. Q., Jakes, R., & Goedert, M. (1997). Alpha-synuclein in Lewy bodies. Nature, 388(6645), 839–840. doi:10.1038/42166 Srivastav, S., Singh, S. K., Yadav, A. K., & Srikrishna, S. (2015). Folic acid supplementation rescues anomalies associated with knockdown of parkin in dopaminergic and serotonergic neurons in Drosophila model of Parkinson's disease. Biochemical and Biophysical Research Communications, 460(3), 780–785. doi:10.1016/j.bbrc.2015.03.106 Stark, A. K., & Pakkenberg, B. (2004). Histological changes of the dopaminergic nigrostriatal system in aging. Cell and Tissue Research, 318(1), 81–92. doi:10.1007/s00441-004-0972-9 Stefanis, L. (2012). α-Synuclein in Parkinson's disease. Cold Spring Harbor Perspectives in Medicine, 2(2), a009399–a009399. doi:10.1101/cshperspect.a009399 Stefanis, L., Larsen, K. E., Rideout, H. J., Sulzer, D., & Greene, L. A. (2001). Expression of A53T mutant but not wild-type alpha-synuclein in PC12 cells induces alterations of the ubiquitin-dependent degradation system, loss of dopamine release, and autophagic cell death. The Journal of Neuroscience : the Official Journal of the Society for Neuroscience, 21(24), 9549–9560. Stocchi, F., Fossati, C., & Torti, M. (2015). Rasagiline for the treatment of Parkinson's disease: an update. Expert Opinion on Pharmacotherapy, 16(14), 2231–2241. doi:10.1517/14656566.2015.1086748 Stojanovski, D., Müller, J. M., Milenkovic, D., Guiard, B., Pfanner, N., & Chacinska, A. (2008). The MIA system for protein import into the mitochondrial intermembrane space. Biochimica Et Biophysica Acta, 1783(4), 610–617. doi:10.1016/j.bbamcr.2007.10.004 Storch, A., Jost, W. H., Vieregge, P., Spiegel, J., Greulich, W., Durner, J., et al. (2007). Randomized, double-blind, placebo-controlled trial on symptomatic effects of coenzyme Q(10) in Parkinson disease. Archives of Neurology, 64(7), 938–944. doi:10.1001/archneur.64.7.nct60005 Stover, E. H., Borthwick, K. J., Bavalia, C., Eady, N., Fritz, D. M., Rungroj, N., et al. (2002). Novel ATP6V1B1 and ATP6V0A4 mutations in autosomal recessive distal renal tubular acidosis with new evidence for hearing loss. Journal of Medical Genetics, 39(11), 796–803. Stransky, L. A., & Forgac, M. (2015). Amino Acid Availability Modulates Vacuolar H+- ATPase Assembly. The Journal of Biological Chemistry, 290(45), 27360–27369. doi:10.1074/jbc.M115.659128 Su, Y., Zhou, A., Al-Lamki, R. S., & Karet, F. E. (2003). The a-subunit of the V-type H+-

293 References

ATPase interacts with phosphofructokinase-1 in humans. The Journal of Biological Chemistry, 278(22), 20013–20018. doi:10.1074/jbc.M210077200 Sugeno, N., Takeda, A., Hasegawa, T., Kobayashi, M., Kikuchi, A., Mori, F., et al. (2008). Serine 129 phosphorylation of alpha-synuclein induces unfolded protein response- mediated cell death. The Journal of Biological Chemistry, 283(34), 23179–23188. doi:10.1074/jbc.M802223200 Suh, Y.-H., & Checler, F. (2002). Amyloid precursor protein, presenilins, and alpha- synuclein: molecular pathogenesis and pharmacological applications in Alzheimer's disease. Pharmacological Reviews, 54(3), 469–525. Sumner, J. P., Dow, J. A., Earley, F. G., Klein, U., Jäger, D., & Wieczorek, H. (1995). Regulation of plasma membrane V-ATPase activity by dissociation of peripheral subunits. The Journal of Biological Chemistry, 270(10), 5649–5653. Surguchov, A. (2015). Intracellular Dynamics of Synucleins: "Here, There and Everywhere". International Review of Cell and Molecular Biology, 320, 103–169. doi:10.1016/bs.ircmb.2015.07.007 Surmeier, D. J. (2007). Calcium, ageing, and neuronal vulnerability in Parkinson's disease. The Lancet. Neurology, 6(10), 933–938. doi:10.1016/S1474- 4422(07)70246-6 Surmeier, D. J., & Sulzer, D. (2013). The pathology roadmap in Parkinson disease. Prion, 7(1), 85–91. doi:10.4161/pri.23582 Suzuki, Y. J., Carini, M., & Butterfield, D. A. (2010). Protein carbonylation. Antioxidants & Redox Signaling, 12(3), 323–325. doi:10.1089/ars.2009.2887 Tabke, K., Albertmelcher, A., Vitavska, O., Huss, M., Schmitz, H.-P., & Wieczorek, H. (2014). Reversible disassembly of the yeast V-ATPase revisited under in vivo conditions. The Biochemical Journal, 462(1), 185–197. doi:10.1042/BJ20131293 Taguchi, K., Watanabe, Y., Tsujimura, A., & Tanaka, M. (2016). Brain region-dependent differential expression of alpha-synuclein. The Journal of Comparative Neurology, 524(6), 1236–1258. doi:10.1002/cne.23901 Tanik, S. A., Schultheiss, C. E., Volpicelli-Daley, L. A., Brunden, K. R., & Lee, V. M.-Y. (2013). Lewy body-like α-synuclein aggregates resist degradation and impair macroautophagy. The Journal of Biological Chemistry, 288(21), 15194–15210. doi:10.1074/jbc.M113.457408 Tanner, C. M., & Goldman, S. M. (1996). Epidemiology of Parkinson's disease. Neurologic Clinics, 14(2), 317–335. Tanner, C. M., Kamel, F., Ross, G. W., Hoppin, J. A., Goldman, S. M., Korell, M., et al. (2011). Rotenone, paraquat, and Parkinson's disease. Environmental Health Perspectives, 119(6), 866–872. doi:10.1289/ehp.1002839 Terman, A., & Kurz, T. (2013). Lysosomal iron, iron chelation, and cell death. Antioxidants & Redox Signaling, 18(8), 888–898. doi:10.1089/ars.2012.4885 Thenganatt, M. A., & Jankovic, J. (2014). Parkinson disease subtypes. JAMA Neurology, 71(4), 499–504. doi:10.1001/jamaneurol.2013.6233 Thomas, B., & Beal, M. F. (2007). Parkinson's disease. Human Molecular Genetics, 16 Spec No. 2(R2), R183–94. doi:10.1093/hmg/ddm159 Tong, Q., Xu, Q., Xia, Q., Yuan, Y., Zhang, L., Sun, H., et al. (2015). Correlations between plasma levels of amino acids and nonmotor symptoms in Parkinson's disease.

294 References

Journal of Neural Transmission (Vienna, Austria : 1996), 122(3), 411–417. doi:10.1007/s00702-014-1280-5 Toyoda, Y., Erkut, C., Pan-Montojo, F., Boland, S., Stewart, M. P., Müller, D. J., et al. (2014). Products of the Parkinson's disease-related glyoxalase DJ-1, D-lactate and glycolate, support mitochondrial membrane potential and neuronal survival. Biology Open, 3(8), 777–784. doi:10.1242/bio.20149399 Truban, D., Hou, X., Caulfield, T. R., Fiesel, F. C., & Springer, W. (2017). PINK1, Parkin, and Mitochondrial Quality Control: What can we Learn about Parkinson's Disease Pathobiology? Journal of Parkinson's Disease, 7(1), 13–29. doi:10.3233/JPD- 160989 Tsujimura, A., Taguchi, K., Watanabe, Y., Tatebe, H., Tokuda, T., Mizuno, T., & Tanaka, M. (2015). Lysosomal enzyme cathepsin B enhances the aggregate forming activity of exogenous α-synuclein fibrils. Neurobiology of Disease, 73, 244–253. doi:10.1016/j.nbd.2014.10.011 Tufi, R., Gandhi, S., de Castro, I. P., Lehmann, S., Angelova, P. R., Dinsdale, D., et al. (2014). Enhancing nucleotide metabolism protects against mitochondrial dysfunction and neurodegeneration in a PINK1 model of Parkinson's disease. Nature Cell Biology, 16(2), 157–166. doi:10.1038/ncb2901 Tyanova, S., Temu, T., Sinitcyn, P., Carlson, A., Hein, M. Y., Geiger, T., et al. (2016). The Perseus computational platform for comprehensive analysis of (prote)omics data. Nature Methods, 13(9), 731–740. doi:10.1038/nmeth.3901 Uchiyama, A., Kim, J.-S., Kon, K., Jaeschke, H., Ikejima, K., Watanabe, S., & Lemasters, J. J. (2008). Translocation of iron from lysosomes into mitochondria is a key event during oxidative stress-induced hepatocellular injury. Hepatology (Baltimore, Md.), 48(5), 1644–1654. doi:10.1002/hep.22498 Uversky, V. N., Li, J., & Fink, A. L. (2001). Pesticides directly accelerate the rate of alpha-synuclein fibril formation: a possible factor in Parkinson's disease. FEBS Letters, 500(3), 105–108. Van Damme, T., Gardeitchik, T., Mohamed, M., Guerrero-Castillo, S., Freisinger, P., Guillemyn, B., et al. (2017). Mutations in ATP6V1E1 or ATP6V1A Cause Autosomal- Recessive Cutis Laxa. American Journal of Human Genetics, 100(2), 216–227. doi:10.1016/j.ajhg.2016.12.010 Varga, S. J., Qi, C., Podolsky, E., & Lee, D. (2014). A new Drosophila model to study the interaction between genetic and environmental factors in Parkinson's disease. Brain Research, 1583, 277–286. doi:10.1016/j.brainres.2014.08.021 Vavassori, S., & Mayer, A. (2014). A new life for an old pump: V-ATPase and neurotransmitter release. The Journal of Cell Biology, 205(1), 7–9. doi:10.1083/jcb.201403040 Vekrellis, K., Xilouri, M., Emmanouilidou, E., & Stefanis, L. (2009). Inducible overexpression of wild type alpha-synuclein in human neuronal cells leads to caspase- dependent non-apoptotic death. Journal of Neurochemistry, 109(5), 1348–1362. doi:10.1111/j.1471-4159.2009.06054.x Vempati, U. D., Han, X., & Moraes, C. T. (2009). Lack of cytochrome c in mouse fibroblasts disrupts assembly/stability of respiratory complexes I and IV. The Journal of Biological Chemistry, 284(7), 4383–4391. doi:10.1074/jbc.M805972200

295 References

Vitavska, O., Merzendorfer, H., & Wieczorek, H. (2005). The V-ATPase subunit C binds to polymeric F-actin as well as to monomeric G-actin and induces cross-linking of actin filaments. The Journal of Biological Chemistry, 280(2), 1070–1076. doi:10.1074/jbc.M406797200 Vlajinac, H., Dzoljic, E., Maksimovic, J., Marinkovic, J., Sipetic, S., & Kostic, V. (2013). Infections as a risk factor for Parkinson's disease: a case-control study. The International Journal of Neuroscience, 123(5), 329–332. doi:10.3109/00207454.2012.760560 Vogiatzi, T., Xilouri, M., Vekrellis, K., & Stefanis, L. (2008). Wild type alpha-synuclein is degraded by chaperone-mediated autophagy and macroautophagy in neuronal cells. The Journal of Biological Chemistry, 283(35), 23542–23556. doi:10.1074/jbc.M801992200 Volpicelli-Daley, L. A., Gamble, K. L., Schultheiss, C. E., Riddle, D. M., West, A. B., & Lee, V. M.-Y. (2014). Formation of α-synuclein Lewy neurite-like aggregates in axons impedes the transport of distinct endosomes. Molecular Biology of the Cell, 25(25), 4010–4023. doi:10.1091/mbc.E14-02-0741 Volpicelli-Daley, L. A., Luk, K. C., Patel, T. P., Tanik, S. A., Riddle, D. M., Stieber, A., et al. (2011). Exogenous α-synuclein fibrils induce Lewy body pathology leading to synaptic dysfunction and neuron death. Neuron, 72(1), 57–71. doi:10.1016/j.neuron.2011.08.033 Voss, M., Vitavska, O., Walz, B., Wieczorek, H., & Baumann, O. (2007). Stimulus- induced phosphorylation of vacuolar H(+)-ATPase by protein kinase A. The Journal of Biological Chemistry, 282(46), 33735–33742. doi:10.1074/jbc.M703368200 Walker, J. G., Batterham, P. J., Mackinnon, A. J., Jorm, A. F., Hickie, I., Fenech, M., et al. (2012). Oral folic acid and vitamin B-12 supplementation to prevent cognitive decline in community-dwelling older adults with depressive symptoms--the Beyond Ageing Project: a randomized controlled trial. The American Journal of Clinical Nutrition, 95(1), 194–203. doi:10.3945/ajcn.110.007799 Wang, D., Epstein, D., Khalaf, O., Srinivasan, S., Williamson, W. R., Fayyazuddin, A., et al. (2014a). Ca2+-Calmodulin regulates SNARE assembly and spontaneous neurotransmitter release via v-ATPase subunit V0a1. The Journal of Cell Biology, 205(1), 21–31. doi:10.1083/jcb.201312109 Wang, X., Huang, T., Bu, G., & Xu, H. (2014b). Dysregulation of protein trafficking in neurodegeneration. Molecular Neurodegeneration, 9(1), 31. doi:10.1186/1750- 1326-9-31 Wang, Y., Nartiss, Y., Steipe, B., McQuibban, G. A., & Kim, P. K. (2012). ROS-induced mitochondrial depolarization initiates PARK2/PARKIN-dependent mitochondrial degradation by autophagy. Autophagy, 8(10), 1462–1476. doi:10.4161/auto.21211 Wasilewski, M., Chojnacka, K., & Chacinska, A. (2017). Protein trafficking at the crossroads to mitochondria. Biochimica Et Biophysica Acta, 1864(1), 125–137. doi:10.1016/j.bbamcr.2016.10.019 Watanabe, Shigeki, Trimbuch, T., Camacho-Pérez, M., Rost, B. R., Brokowski, B., Söhl- Kielczynski, B., et al. (2014). Clathrin regenerates synaptic vesicles from endosomes. Nature, 515(7526), 228–233. doi:10.1038/nature13846 Watanabe, Yu, Himeda, T., & Araki, T. (2005). Mechanisms of MPTP toxicity and their

296 References

implications for therapy of Parkinson's disease. Medical Science Monitor : International Medical Journal of Experimental and Clinical Research, 11(1), RA17– 23. Watson, D. L., Carpenter, C. L., Marks, S. S., & Greenberg, D. A. (1988). Striatal calcium channel antagonist receptors in Huntington“s disease and Parkinson”s disease. Annals of Neurology, 23(3), 303–305. doi:10.1002/ana.410230316 Weisskopf, M. G., Weuve, J., Nie, H., Saint-Hilaire, M.-H., Sudarsky, L., Simon, D. K., et al. (2010). Association of cumulative lead exposure with Parkinson's disease. Environmental Health Perspectives, 118(11), 1609–1613. doi:10.1289/ehp.1002339 Wessel, D., & Flügge, U. I. (1984). A method for the quantitative recovery of protein in dilute solution in the presence of detergents and lipids. Analytical Biochemistry, 138(1), 141–143. West, A. B., Zimprich, A., Lockhart, P. J., Farrer, M., Singleton, A., Holtom, B., et al. (2001). Refinement of the PARK3 locus on chromosome 2p13 and the analysis of 14 candidate genes. European Journal of Human Genetics : EJHG, 9(9), 659–666. doi:10.1038/sj.ejhg.5200698 Westerman, B. A., Poutsma, A., Steegers, E. A. P., & Oudejans, C. B. M. (2004). C2360, a nuclear protein expressed in human proliferative cytotrophoblasts, is a representative member of a novel protein family with a conserved coiled coil- helix-coiled coil-helix domain. Genomics, 83(6), 1094–1104. doi:10.1016/j.ygeno.2003.12.006 WHO. (2008). The global burden of disease: 2004 update. World Health Organization. Wilhelmus, M. M. M., Verhaar, R., Andringa, G., Bol, J. G. J. M., Cras, P., Shan, L., et al. (2011). Presence of tissue transglutaminase in granular endoplasmic reticulum is characteristic of melanized neurons in Parkinson's disease brain. Brain Pathology (Zurich, Switzerland), 21(2), 130–139. doi:10.1111/j.1750-3639.2010.00429.x Winslow, A. R., Chen, C.-W., Corrochano, S., Acevedo-Arozena, A., Gordon, D. E., Peden, A. A., et al. (2010). α-Synuclein impairs macroautophagy: implications for Parkinson's disease. The Journal of Cell Biology, 190(6), 1023–1037. doi:10.1083/jcb.201003122 Wirdefeldt, K., Adami, H.-O., Cole, P., Trichopoulos, D., & Mandel, J. (2011). Epidemiology and etiology of Parkinson's disease: a review of the evidence. European Journal of Epidemiology, 26 Suppl 1(S1), S1–58. doi:10.1007/s10654- 011-9581-6 Wislet-Gendebien, S., Visanji, N. P., Whitehead, S. N., Marsilio, D., Hou, W., Figeys, D., et al. (2008). Differential regulation of wild-type and mutant alpha-synuclein binding to synaptic membranes by cytosolic factors. BMC Neuroscience, 9(1), 92. doi:10.1186/1471-2202-9-92 Wolfe, D. M., Lee, J.-H., Kumar, A., Lee, S., Orenstein, S. J., & Nixon, R. A. (2013). Autophagy failure in Alzheimer's disease and the role of defective lysosomal acidification. The European Journal of Neuroscience, 37(12), 1949–1961. doi:10.1111/ejn.12169 Woo, J. T., Shinohara, C., Sakai, K., Hasumi, K., & Endo, A. (1992). Isolation, characterization and biological activities of concanamycins as inhibitors of

297 References

lysosomal acidification. The Journal of Antibiotics, 45(7), 1108–1116. Wu, F., Wang, Z., Gu, J.-H., Ge, J.-B., Liang, Z.-Q., & Qin, Z.-H. (2013a). p38(MAPK)/p53- Mediated Bax induction contributes to neurons degeneration in rotenone-induced cellular and rat models of Parkinson's disease. Neurochemistry International, 63(3), 133–140. doi:10.1016/j.neuint.2013.05.006 Wu, F., Xu, H.-D., Guan, J.-J., Hou, Y.-S., Gu, J.-H., Zhen, X.-C., & Qin, Z.-H. (2015). Rotenone impairs autophagic flux and lysosomal functions in Parkinson's disease. Neuroscience, 284, 900–911. doi:10.1016/j.neuroscience.2014.11.004 Wu, Y.-L., Ding, X.-X., Sun, Y.-H., Yang, H.-Y., & Sun, L. (2013b). Methylenetetrahydrofolate reductase (MTHFR) C677T/A1298C polymorphisms and susceptibility to Parkinson's disease: a meta-analysis. Journal of the Neurological Sciences, 335(1-2), 14–21. doi:10.1016/j.jns.2013.09.006 Xicoy, H., Wieringa, B., & Martens, G. J. M. (2017). The SH-SY5Y cell line in Parkinson's disease research: a systematic review. Molecular Neurodegeneration, 12(1), 10. doi:10.1186/s13024-017-0149-0 Xie, Y., Feng, H., Peng, S., Xiao, J., & Zhang, J. (2017). Association of plasma homocysteine, vitamin B12 and folate levels with cognitive function in Parkinson's disease: A meta-analysis. Neuroscience Letters, 636, 190–195. doi:10.1016/j.neulet.2016.11.007 Xilouri, M., & Stefanis, L. (2015). Chaperone mediated autophagy to the rescue: A new- fangled target for the treatment of neurodegenerative diseases. Molecular and Cellular Neurosciences, 66(Pt A), 29–36. doi:10.1016/j.mcn.2015.01.003 Xing, Y., Liqi, Z., Jian, L., Qinghua, Y., & Qian, Y. (2017). Doxycycline Induces Mitophagy and Suppresses Production of Interferon-β in IPEC-J2 Cells. Frontiers in Cellular and Infection Microbiology, 7, 21. doi:10.3389/fcimb.2017.00021 Xu, T., & Forgac, M. (2000). Subunit D (Vma8p) of the yeast vacuolar H+-ATPase plays a role in coupling of proton transport and ATP hydrolysis. The Journal of Biological Chemistry, 275(29), 22075–22081. doi:10.1074/jbc.M002983200 Xu, T., & Forgac, M. (2001). Microtubules are involved in glucose-dependent dissociation of the yeast vacuolar [H+]-ATPase in vivo. The Journal of Biological Chemistry, 276(27), 24855–24861. doi:10.1074/jbc.M100637200 Xu, W., Tan, L., & Yu, J.-T. (2015a). Link between the SNCA gene and parkinsonism. Neurobiology of Aging, 36(3), 1505–1518. doi:10.1016/j.neurobiolaging.2014.10.042 Xu, X., Chai, S., Wang, P., Zhang, C., Yang, Y., Yang, Y., & Wang, K. (2015b). Aldehyde dehydrogenases and cancer stem cells. Cancer Letters, 369(1), 50–57. doi:10.1016/j.canlet.2015.08.018 Xu, Z., Zhang, D., He, X., Huang, Y., & Shao, H. (2016). Transport of Calcium Ions into Mitochondria. Current Genomics, 17(3), 215–219. doi:10.2174/1389202917666160202215748 Yang, Q., She, H., Gearing, M., Colla, E., Lee, M., Shacka, J. J., & Mao, Z. (2009). Regulation of neuronal survival factor MEF2D by chaperone-mediated autophagy. Science (New York, N.Y.), 323(5910), 124–127. doi:10.1126/science.1166088 Yang, X., Zhao, Q., An, R., Zheng, J., Tian, S., Chen, Y., & Xu, Y. (2016). Mutational scanning of the CHCHD2 gene in Han Chinese patients with Parkinson's disease

298 References

and meta-analysis of the literature. Parkinsonism & Related Disorders, 29, 42–46. doi:10.1016/j.parkreldis.2016.05.032 Yang, Z., & Klionsky, D. J. (2010). Mammalian autophagy: core molecular machinery and signaling regulation. Current Opinion in Cell Biology, 22(2), 124–131. doi:10.1016/j.ceb.2009.11.014 Yeger-Lotem, E., Sattath, S., Kashtan, N., Itzkovitz, S., Milo, R., Pinter, R. Y., et al. (2004). Network motifs in integrated cellular networks of transcription-regulation and protein-protein interaction. Proceedings of the National Academy of Sciences of the United States of America, 101(16), 5934–5939. doi:10.1073/pnas.0306752101 Yoshimori, T., Yamamoto, A., Moriyama, Y., Futai, M., & Tashiro, Y. (1991). Bafilomycin A1, a specific inhibitor of vacuolar-type H(+)-ATPase, inhibits acidification and protein degradation in lysosomes of cultured cells. The Journal of Biological Chemistry, 266(26), 17707–17712. Youdim, M. B. H., Edmondson, D., & Tipton, K. F. (2006). The therapeutic potential of monoamine oxidase inhibitors. Nature Reviews. Neuroscience, 7(4), 295–309. doi:10.1038/nrn1883 Yuan, Y.-H., Yan, W.-F., Sun, J.-D., Huang, J.-Y., Mu, Z., & Chen, N.-H. (2015). The molecular mechanism of rotenone-induced α-synuclein aggregation: emphasizing the role of the calcium/GSK3β pathway. Toxicology Letters, 233(2), 163–171. doi:10.1016/j.toxlet.2014.11.029 Yuan, Y.-S., Zhou, X.-J., Tong, Q., Zhang, L., Zhang, L., Qi, Z.-Q., et al. (2013). Change in plasma levels of amino acid neurotransmitters and its correlation with clinical heterogeneity in early Parkinson's disease patients. CNS Neuroscience & Therapeutics, 19(11), 889–896. doi:10.1111/cns.12165 Zeevalk, G. D., Razmpour, R., & Bernard, L. P. (2008). Glutathione and Parkinson's disease: is this the elephant in the room? Biomedicine & Pharmacotherapy = Biomedecine & Pharmacotherapie, 62(4), 236–249. doi:10.1016/j.biopha.2008.01.017 Zhang, Chen-Song, Jiang, B., Li, M., Zhu, M., Peng, Y., Zhang, Y.-L., et al. (2014a). The lysosomal v-ATPase-Ragulator complex is a common activator for AMPK and mTORC1, acting as a switch between catabolism and anabolism. Cell Metabolism, 20(3), 526–540. doi:10.1016/j.cmet.2014.06.014 Zhang, H, Duan, C., & Yang, H. (2015a). Defective autophagy in Parkinson's disease: lessons from genetics. Molecular Neurobiology, 51(1), 89–104. doi:10.1007/s12035-014-8787-5 Zhang, Haiyuan, Lei, Y., Yuan, P., Li, L., Luo, C., Gao, R., et al. (2014b). ROS-mediated autophagy induced by dysregulation of lipid metabolism plays a protective role in colorectal cancer cells treated with gambogic acid. PloS One, 9(5), e96418. doi:10.1371/journal.pone.0096418 Zhang, J, Perry, G., Smith, M. A., Robertson, D., Olson, S. J., Graham, D. G., & Montine, T. J. (1999). Parkinson's disease is associated with oxidative damage to cytoplasmic DNA and RNA in substantia nigra neurons. The American Journal of Pathology, 154(5), 1423–1429. doi:10.1016/S0002-9440(10)65396-5 Zhang, Ming, Xi, Z., Fang, S., Ghani, M., Sato, C., Moreno, D., et al. (2016a). Mutation

299 References

analysis of CHCHD2 in Canadian patients with familial Parkinson's disease. Neurobiology of Aging, 38, 217.e7–217.e8. doi:10.1016/j.neurobiolaging.2015.10.038 Zhang, Xiaoli, Yu, L., & Xu, H. (2016b). Lysosome calcium in ROS regulation of autophagy. Autophagy, 12(10), 1954–1955. doi:10.1080/15548627.2016.1212787 Zhang, Yuan, Sun, Q.-Y., Yu, R.-H., Guo, J.-F., Tang, B.-S., & Yan, X.-X. (2015b). The contribution of GIGYF2 to Parkinson's disease: a meta-analysis. Neurological Sciences : Official Journal of the Italian Neurological Society and of the Italian Society of Clinical Neurophysiology, 36(11), 2073–2079. doi:10.1007/s10072-015- 2316-9 Zhang, Yuxi, Nguyen, D. T., Olzomer, E. M., Poon, G. P., Cole, N. J., Puvanendran, A., et al. (2017). Rescue of Pink1 Deficiency by Stress-Dependent Activation of Autophagy. Cell Chemical Biology, 24(4), 471–480.e4. doi:10.1016/j.chembiol.2017.03.005 Zhao, J., Benlekbir, S., & Rubinstein, J. L. (2015). Electron cryomicroscopy observation of rotational states in a eukaryotic V-ATPase. Nature, 521(7551), 241–245. doi:10.1038/nature14365 Zhao, Y., Araki, S., Wu, J., Teramoto, T., Chang, Y.-F., Nakano, M., et al. (2011). An expanded palette of genetically encoded Ca²⁺ indicators. Science (New York, N.Y.), 333(6051), 1888–1891. doi:10.1126/science.1208592 Zhdanov, A. V., Dmitriev, R. I., & Papkovsky, D. B. (2011). Bafilomycin A1 activates respiration of neuronal cells via uncoupling associated with flickering depolarization of mitochondria. Cellular and Molecular Life Sciences : CMLS, 68(5), 903–917. doi:10.1007/s00018-010-0502-8 Zhou, M., Xu, S., Mi, J., Uéda, K., & Chan, P. (2013a). Nuclear translocation of alpha- synuclein increases susceptibility of MES23.5 cells to oxidative stress. Brain Research, 1500, 19–27. doi:10.1016/j.brainres.2013.01.024 Zhou, S., Wang, Z., & Klaunig, J. E. (2013b). Caenorhabditis elegans neuron degeneration and mitochondrial suppression caused by selected environmental chemicals. International Journal of Biochemistry and Molecular Biology, 4(4), 191– 200. Zhou, Z.-D., Saw, W.-T., & Tan, E.-K. (2016). Mitochondrial CHCHD-Containing Proteins: Physiologic Functions and Link with Neurodegenerative Diseases. Molecular Neurobiology, 25(7), 319–13. doi:10.1007/s12035-016-0099-5 Zhu, J., Dagda, R. K., & Chu, C. T. (2011). Monitoring mitophagy in neuronal cell cultures. Methods in Molecular Biology (Clifton, N.J.), 793(Chapter 21), 325–339. doi:10.1007/978-1-61779-328-8_21 Zilberter, Y., & Zilberter, M. (2017). The vicious circle of hypometabolism in neurodegenerative diseases: Ways and mechanisms of metabolic correction. Journal of Neuroscience Research, 134(Pt 6), 1658. doi:10.1002/jnr.24064 Zoncu, R., Bar-Peled, L., Efeyan, A., Wang, S., Sancak, Y., & Sabatini, D. M. (2011). mTORC1 senses lysosomal amino acids through an inside-out mechanism that requires the vacuolar H(+)-ATPase. Science (New York, N.Y.), 334(6056), 678–683. doi:10.1126/science.1207056

300 Appendix 1

APPENDIX 1

Table A 1 Among the 9 separate datasets obtained, corresponding to the 3 treatment groups at each of the 3 time-points (24, 48 and 96-hours), the levels of 440 individual proteins were significantly altered by greater than 0.2 or less than -0.2 Log2 fold in one or more datasets (p<0.01)

24 hours 48 hours 72 hours Moderately 10 nM Synuclein + Moderately 10 nM Synuclein + Moderately 10 nM Synuclein + elevated Rotenone Rotenone elevated Rotenone Rotenone elevated Rotenone Rotenone Gene αSynuclein αSynuclein αSynuclein names FC p FC p FC p FC p FC p FC p FC p FC p FC p ABCF2 0.074 0.595 0.036 0.639 0.018 0.845 0.014 0.784 0.040 0.056 - 0.053 0.119 0.274 - 0.035 - 0.006 0.050 0.211 0.279 ABI1 0.355 0.004 0.081 0.231 0.133 0.119 0.020 0.784 - 0.971 0.113 0.218 0.060 0.381 - 0.840 0.004 0.067 ACO2 0.019 0.816 0.072 0.021 0.122 0.190 0.083 0.290 0.228 0.039 0.207 0.001 0.098 0.054 0.073 0.396 0.085 0.001 ADCK3 0.012 0.913 - 0.527 0.073 0.602 - 0.003 - 0.021 0.227 0.236 0.295 ADPGK 0.228 0.465 - 0.695 - 0.744 0.131 0.719 0.232 0.005 0.118 0.050 ADSL - 0.333 0.055 0.455 0.008 0.937 - 0.002 - 0.009 - 0.411 - 0.009 - 0.518 - 0.009 0.186 0.058 0.138 0.087 0.303 0.034 0.105 AGAP3 0.132 0.459 0.189 0.014 0.173 0.028 - 0.817 0.127 0.096 0.005 0.936 0.099 0.013 0.239 0.010 0.133 0.420 0.025 AGO2 0.318 0.006 - 0.713 - 0.030 - 0.265 0.111 0.066 - 0.238 0.101 0.049 0.233 0.420 AIDA 0.006 0.944 - 0.002 0.099 0.072 - 0.833 0.060 0.268 0.094 0.781 - 0.335 0.609 0.061 0.081 AK3 - 0.186 - 0.008 - 0.851 0.021 0.779 - 0.392 - 0.308 0.113 0.005 0.052 0.280 0.294 0.291 0.012 0.305 0.062 AKR1C1 0.288 0.371 0.089 0.793 0.263 0.005 0.087 0.876 - 0.872 0.228 0.081

301 Appendix 1

0.052 ALDH1B1 - 0.142 0.155 0.405 0.214 0.000 - 0.523 0.025 0.327 0.162 0.582 0.170 0.112 AP2M1 0.207 0.010 0.066 0.777 0.150 0.278 0.056 0.812 - 0.528 0.070 0.414 - 0.643 0.047 0.321 0.199 0.066 ARFGAP2 0.207 0.000 - 0.809 0.082 0.401 0.012 0.862 - 0.192 0.080 0.307 0.013 0.868 0.028 0.722 - 0.332 0.063 0.073 0.093 ARHGAP4 0.666 0.487 0.000 0.998 0.404 0.000 2 ARHGEF12 - 0.740 - 0.008 0.198 0.781 ARIH1 - 0.041 0.146 0.190 0.083 0.248 0.050 0.739 0.076 0.536 - 0.597 - 0.001 - 0.701 0.046 0.829 0.087 0.022 0.267 0.113 ARMC10 - 0.525 0.366 0.327 0.161 0.520 - 0.722 - 0.321 0.468 0.006 - 0.970 0.130 0.077 0.134 0.014 ARPIN 0.004 0.991 - 0.007 0.548 ASAH1 0.382 0.093 0.041 0.853 0.069 0.084 0.144 0.473 0.068 0.290 0.005 0.943 0.241 0.006 0.188 0.246 ASMTL 0.007 0.964 - 0.659 0.149 0.682 0.017 0.872 0.298 0.001 0.101 ASNS - 0.690 - 0.610 - 0.578 0.266 0.194 - 0.939 - 0.594 0.270 0.004 - 0.460 - 0.136 0.055 0.096 0.104 0.028 0.139 0.103 0.146 ATG3 0.025 0.924 - 0.003 0.111 0.441 0.160 0.243 - 0.701 0.406 0.080 ATP6V0A1 - 0.758 0.194 0.007 0.053 0.678 0.021 0.901 0.152 0.536 0.178 0.008 0.070 0.712 0.221 0.405 0.232 0.007 0.059 ATP6V0D1 0.097 0.629 0.179 0.386 0.208 0.000 0.005 0.961 0.149 0.450 0.161 0.072 0.105 0.230 0.321 0.123 0.250 0.090 ATP6V1C1 0.224 0.004 0.081 0.020 0.106 0.412 0.144 0.153 - 0.346 1.061 ATP6V1D 0.200 0.000 0.032 0.720 0.151 0.058 - 0.749 0.210 0.225 0.088 0.074 0.111 0.446 0.108 0.657 0.211 0.005 0.043 ATP6V1G1 0.014 0.861 - 0.404 0.094 0.022 0.262 0.001 0.166 0.074

302 Appendix 1

0.259 ATPAF1 - 0.010 - 0.079 0.322 0.650 BAG2 - 0.421 - 0.251 - 0.285 0.016 0.957 0.057 0.463 0.195 0.008 0.299 0.002 0.096 0.788 0.143 0.147 0.211 0.369 0.130 BAK1 0.345 0.000 0.013 0.933 - 0.042 - 0.888 0.116 0.528 - 0.898 - 0.370 0.181 0.033 0.013 0.250 BCLAF1 - 0.514 - 0.578 - 0.067 0.054 0.646 - 0.305 - 0.447 0.041 0.028 - 0.001 - 0.906 0.049 0.086 0.058 0.296 0.071 0.491 0.011 BCR - 0.249 0.118 0.692 - 0.751 - 0.924 - 0.228 - 0.006 - 0.075 0.053 0.059 0.033 1.258 0.288 0.545 BID - 0.533 - 0.229 - 0.101 - 0.583 - 0.399 - 0.005 - 0.676 - 0.153 - 0.354 0.045 0.092 0.060 0.111 0.130 0.210 0.021 0.278 0.075 BPTF 0.243 0.032 - 0.174 - 0.001 0.080 0.606 0.125 0.292 0.057 0.485 0.045 0.594 - 0.135 0.030 0.218 0.974 BTAF1 - 0.533 - 0.345 0.028 0.826 - 0.252 0.049 0.520 - 0.001 - 0.407 0.195 0.162 0.140 0.291 0.232 BTF3L4 0.005 0.975 0.133 0.054 0.040 0.546 - 0.809 - 0.201 - 0.633 - 0.002 0.026 0.524 0.146 0.649 BYSL - 0.009 - 0.001 - 0.694 - 0.934 - 0.335 - 0.000 0.005 0.935 - 0.064 - 0.000 0.184 0.170 0.048 0.013 0.150 0.178 0.321 0.228 BZW2 - 0.975 0.329 0.280 0.246 0.004 0.058 0.021 0.183 0.393 0.144 0.247 0.004 C11orf54 - 0.929 0.036 0.803 0.124 0.302 0.030 0.919 0.019 0.954 0.148 0.320 0.033 0.804 0.230 0.336 0.214 0.004 0.034 C11orf95 0.039 0.554 - 0.007 0.487 C14orf166 0.049 0.410 - 0.608 0.058 0.227 - 0.535 - 0.015 - 0.323 - 0.002 - 0.585 - 0.248 0.042 0.054 0.049 0.026 0.202 0.031 0.048 C9orf114 - 0.021 0.028 0.819 - 0.097 - 0.149 - 0.046 0.309 0.004 - 0.138 0.116 0.093 0.173 0.439 0.894 C9orf142 - 0.453 - 0.003 - 0.032 0.115 0.269 0.187

303 Appendix 1

CACNA2D - 0.002 - 0.885 0.058 0.750 0.129 0.048 0.276 0.012 0.181 0.171 0.201 0.009 0.149 0.266 0.092 0.500 1 0.041 0.042 CAMSAP2 0.090 0.003 0.178 0.196 0.041 0.783 0.287 0.003 - 0.494 0.180 CAPRIN1 0.033 0.182 - 0.191 - 0.168 0.061 0.042 - 0.000 - 0.551 0.084 0.439 - 0.287 - 0.103 0.137 0.041 0.201 0.066 0.218 0.151 CASP3 0.066 0.339 - 0.376 - 0.242 - 0.008 - 0.004 - 0.003 - 0.037 - 0.807 - 0.294 0.180 0.117 0.089 0.260 0.160 0.075 0.022 0.066 CBX8 - 0.005 0.026 0.821 0.045 0.225 0.035 0.061 - 0.068 0.022 0.853 0.000 0.998 - 0.157 - 0.470 0.259 0.265 0.089 0.056 CCDC86 - 0.405 - 0.005 0.360 0.593 CCDC88A - 0.834 - 0.969 0.433 0.002 0.241 0.076 0.057 0.006 CCDC94 - 0.279 - 0.005 - 0.513 0.093 0.298 0.064 CCNB1 0.206 0.469 - 0.005 0.535 CCNY 0.336 0.478 - 0.009 0.040 0.296 - 0.937 - 0.002 0.468 0.006 0.124 CD63 - 0.915 0.182 0.694 0.337 0.022 - 0.958 0.192 0.573 0.261 0.000 0.067 0.272 0.217 0.686 0.267 0.127 0.056 0.019 CD81 - 0.952 - 0.447 0.117 0.089 0.388 0.001 - 0.872 0.007 0.290 0.014 CDC42EP3 - 0.366 - 0.007 0.277 0.849 CDK16 0.725 0.003 0.458 0.504 CHCHD2 0.080 0.028 0.407 0.130 0.499 0.000 0.456 0.037 0.496 0.007 0.486 0.038 0.461 0.002 CHD1 0.020 0.373 0.097 0.691 0.052 0.259 0.048 0.425 0.248 0.240 - 0.940 0.295 0.022 - 0.005 - 0.318 0.004 0.242 0.131 CHD3 0.432 0.015 - 0.276 - 0.550 - 0.250 - 0.267 - 0.643 - 0.024 - 0.009 0.213 0.060 0.409 0.223 0.129 0.271 0.260

304 Appendix 1

CHGB 1.353 0.235 0.980 0.378 0.487 0.648 1.203 0.023 0.396 0.720 1.513 0.005 0.971 0.014 0.012 0.994 CHMP6 0.214 0.003 - 0.463 0.015 0.795 - 0.353 - 0.437 - 0.923 0.039 0.864 0.201 0.106 0.169 0.017 CHP1 0.182 0.020 0.055 0.238 0.041 0.333 0.112 0.615 0.195 0.470 0.229 0.002 CHPF2 0.031 0.924 0.055 0.822 0.212 0.000 CIRBP 0.057 0.446 - 0.617 - 0.626 0.344 0.002 0.061 0.758 0.010 0.941 0.109 0.049 - 0.013 - 0.657 0.079 0.042 0.488 0.082 CKMT1A 0.375 0.167 0.467 0.070 - 0.059 1.277 0.072 1.488 0.038 0.758 0.523 1.112 0.000 1.393 0.038 0.366 0.741 0.805 CKS1B 0.038 0.740 - 0.662 0.270 0.129 0.142 0.125 0.089 0.506 0.207 0.142 - 0.006 - 0.566 0.139 0.222 0.089 COG8 0.064 0.165 0.158 0.410 0.129 0.050 0.388 0.008 0.024 0.913 COPS5 0.270 0.000 0.078 0.022 0.048 0.535 - 0.146 0.080 0.226 - 0.532 0.163 0.066 0.095 0.715 0.056 0.727 0.107 0.049 COQ5 0.039 0.916 - 0.408 0.108 0.574 0.032 0.851 - 0.614 - 0.344 - 0.014 - 0.002 - 0.096 0.195 0.113 0.058 0.129 0.235 0.169 CORO7 0.123 0.135 0.137 0.170 0.234 0.178 0.116 0.273 0.191 0.003 0.043 0.742 0.184 0.052 0.229 0.000 0.178 0.261 COX4I1 0.033 0.236 0.102 0.236 0.072 0.116 0.142 0.101 0.243 0.085 0.232 0.000 0.182 0.025 0.197 0.350 0.184 0.007 COX6A1P2 0.264 0.006 - 0.395 0.966 COX6B1 0.427 0.009 0.182 0.682 0.508 0.000 0.108 0.533 0.199 0.145 0.180 0.003 CPD 0.207 0.003 0.219 0.081 0.190 0.296 0.116 0.578 0.179 0.467 0.185 0.002 0.116 0.045 0.029 0.808 0.116 0.224 CPNE3 0.213 0.011 0.106 0.427 0.097 0.506 0.237 0.000 0.338 0.079 0.068 0.647 0.116 0.173 0.068 0.202 0.003 0.985 CPSF3L - 0.242 - 0.006 0.089 0.292 CSPG4 0.260 0.001 0.302 0.158 0.071 0.881 0.432 0.026 0.589 0.028 0.326 0.430 0.327 0.029 0.325 0.153 0.188 0.581 CTDSPL2 0.388 0.605 - 0.111 0.059 0.091 0.324 0.192 0.281 0.000 0.344 CYP26A1 - 0.567 0.090 0.509 0.007 0.952 0.245 0.086 0.321 0.001 0.229 0.260 0.344 0.201 0.401 0.136 0.260 0.272 0.008 CYP26B1 0.566 0.006 - 0.398 - 0.000

305 Appendix 1

0.586 0.828 DBH 0.292 0.784 0.197 0.818 0.547 0.001 0.255 0.702 0.168 0.843 0.311 0.002 0.576 0.502 0.092 0.917 0.480 0.003 DCP2 - 0.003 - 0.635 0.763 0.135 DCX 1.183 0.000 - 0.913 0.086 DDAH2 0.180 0.058 0.180 0.256 0.120 0.257 0.075 0.557 0.067 0.527 0.069 0.330 0.065 0.818 - 0.841 0.210 0.000 0.044 DDX28 - 0.009 0.467 DDX51 0.341 0.000 - 0.681 0.008 0.912 - 0.237 - 0.984 - 0.114 0.013 0.167 0.004 0.213 DDX59 - 0.010 0.576 DGCR14 0.128 0.020 - 0.007 0.537 DGKI - 0.380 - 0.005 0.190 0.702 DHCR7 - 0.719 0.045 0.497 0.024 0.179 - 0.532 - 0.569 0.245 0.018 0.103 0.651 0.225 0.002 0.006 0.259 0.196 DHRS7 0.367 0.151 - 0.197 0.094 0.260 0.255 0.376 0.223 0.002 0.077 0.425 0.213 0.344 0.186 0.006 0.380 DHX57 - 0.454 - 0.193 0.095 0.000 0.001 0.995 0.079 0.115 - 0.808 0.254 0.008 0.092 0.012 0.239 0.263 0.023 DIAPH1 0.391 0.155 0.123 0.199 0.140 0.475 0.181 0.499 0.050 0.595 - 0.515 0.355 0.007 - 0.240 0.128 0.510 DLD 0.046 0.352 - 0.002 - 0.024 0.057 0.265 - 0.020 - 0.013 0.076 0.493 - 0.075 - 0.002 0.146 0.203 0.215 0.283 0.368 0.271 DNAH8 - 0.008 - 0.924 - 0.881 - 0.035 1.201 0.124 0.198 0.250 DNAJC3 0.082 0.620 0.191 0.488 0.209 0.004 0.083 0.774 0.549 0.082 0.440 0.042 DUT 0.028 0.652 - 0.328 - 0.130 0.163 0.164 - 0.339 - 0.181 0.138 0.116 - 0.003 - 0.079 0.116 0.027 0.082 0.109 0.260 0.153

306 Appendix 1

DYNC1LI2 0.054 0.120 0.220 0.002 0.049 0.782 - 0.801 0.188 0.482 0.087 0.119 0.077 0.617 0.237 0.013 0.202 0.008 0.021 EARS2 - 0.789 - 0.003 0.086 0.594 EDF1 0.022 0.641 - 0.727 - 0.044 - 0.320 - 0.002 0.081 0.127 0.335 0.219 EEF2K 0.079 0.770 - 0.002 0.519 EHD4 0.141 0.635 - 0.738 0.204 0.008 - 0.257 0.095 1.796 EIF4A2 0.078 0.008 - 0.201 0.056 0.265 - 0.006 - 0.888 - 0.847 - 0.764 0.064 0.515 0.084 0.367 0.015 0.426 0.029 0.036 0.070 ELAC2 0.023 0.888 - 0.791 0.059 0.545 0.041 0.541 - 0.531 - 0.019 - 0.463 - 0.000 - 0.060 0.033 0.070 0.110 0.085 0.267 0.106 ELAVL3 0.542 0.068 0.616 0.033 0.228 0.713 0.543 0.056 - 0.502 0.660 0.001 0.410 0.004 0.278 0.542 0.258 ELF1 - 0.000 - 0.919 - 0.017 0.465 0.006 0.624 ELMO2 - 0.002 0.072 0.858 - 0.759 0.417 0.136 - 0.243 0.362 0.181 0.809 ENO2 - 0.948 - 0.330 0.042 0.384 - 0.299 0.060 0.619 0.105 0.134 0.001 0.996 0.212 0.112 0.246 0.000 0.003 0.047 0.083 EPB41L5 - 0.771 - 0.014 - 0.000 - 0.228 0.061 0.601 - 0.291 0.108 0.454 0.360 0.295 0.202 EPHB4 - 0.037 - 0.012 0.057 0.258 - 0.350 - 0.838 - 0.958 - 0.148 0.091 0.419 0.058 0.026 0.003 0.529 EPS15 0.386 0.246 - 0.888 0.652 0.001 0.012 ERCC5 0.306 0.005 - 0.293 0.333 ERO1L 0.242 0.087 0.260 0.005 0.214 0.117 0.064 0.792 0.318 0.003 0.267 0.018 0.228 0.014 0.278 0.106 0.350 0.001 EXOSC8 - 0.117 0.108 0.002 0.053 0.666 - 0.000 - 0.331 0.142 0.617 0.396

307 Appendix 1

FADS2 0.098 0.738 - 0.665 0.039 0.810 0.164 0.221 - 0.035 - 0.086 0.186 0.222 - 0.177 - 0.001 0.208 0.164 0.235 0.242 0.242 FAM134C - 0.927 - 0.503 - 0.670 0.410 0.001 1.012 0.297 0.197 0.230 - 0.968 0.096 0.696 0.101 0.344 0.024 0.688 0.207 0.008 FAM20B 0.016 0.937 0.673 0.002 0.347 0.044 FASTKD2 - 0.807 - 0.012 - 0.680 - 0.552 - 0.002 - 0.690 - 0.006 0.082 0.491 0.114 0.220 0.251 0.122 0.570 FBLIM1 - 0.849 - 0.002 0.233 0.682 FBXO30 0.301 0.465 - 0.465 0.201 0.008 - 0.449 0.275 0.129 FKBP2 0.014 0.238 0.138 0.137 0.214 0.004 FMR1 0.242 0.004 - 0.864 0.009 0.968 0.034 0.799 0.041 0.710 0.088 0.151 - 0.993 0.605 0.345 - 0.324 0.076 0.001 0.138 FN1 - 0.005 0.543 FN3K 0.129 0.350 - 0.769 0.166 0.101 0.014 0.908 0.164 0.466 0.193 0.036 - 0.415 0.356 0.109 0.231 0.001 0.083 0.098 FTH1 0.216 0.530 0.187 0.639 0.268 0.002 FTL - 0.931 0.254 0.627 0.357 0.025 0.163 0.559 0.271 0.562 0.280 0.004 - 0.636 0.041 0.479 FTSJD2 - 0.163 - 0.003 0.259 0.435 GAA 0.013 0.863 0.106 0.529 - 0.995 0.092 0.364 0.113 0.311 0.063 0.507 0.122 0.209 0.235 0.004 0.020 0.829 0.001 GALE - 0.518 - 0.427 - 0.000 0.071 0.386 0.088 0.339 0.117 0.144 0.239 GALK1 0.079 0.358 0.109 0.458 0.179 0.001 - 0.916 0.024 0.937 0.304 0.008 - 0.687 0.219 0.080 0.337 0.001 0.026 0.139 GALNT2 0.085 0.802 0.330 0.055 0.330 0.015 0.149 0.292 0.156 0.589 0.231 0.000 - 0.756 0.304 0.069 - 0.489 0.083 0.088 GAN 0.250 0.007 - 0.355 0.329

308 Appendix 1

GATA3 - 0.055 - 0.001 - 0.019 - 0.002 0.145 0.273 0.283 0.708 GDAP1 0.324 0.001 0.239 0.171 0.155 0.548 0.258 0.102 0.184 0.200 0.125 0.417 0.261 0.010 0.166 0.049 0.111 0.657 GFPT1 0.245 0.057 0.151 0.049 0.112 0.463 0.306 0.001 0.168 0.006 0.146 0.477 0.106 0.349 0.130 0.392 0.031 0.826 GFRA2 - 0.071 - 0.009 - 0.579 - 0.000 - 0.225 - 0.676 - 0.153 - 0.022 0.061 0.896 0.517 0.545 0.228 0.563 0.335 0.159 0.190 0.567 GGH 0.219 0.005 0.390 0.031 0.088 0.686 0.213 0.025 0.272 0.001 0.142 0.486 0.242 0.001 0.336 0.014 0.102 0.382 GID8 - 0.002 0.398 GLB1 0.141 0.249 0.142 0.132 0.153 0.039 0.141 0.118 0.221 0.000 0.177 0.102 0.128 0.181 0.231 0.154 0.204 0.002 GLIPR2 - 0.412 0.231 0.300 0.197 0.497 0.213 0.006 0.273 GLUL - 0.007 - 0.052 - 0.000 - 0.520 - 0.110 - 0.000 - 0.091 - 0.065 - 0.000 0.107 0.636 0.573 0.040 0.786 0.982 0.449 0.992 1.072 GMPPA 0.321 0.301 - 0.973 - 0.006 - 0.154 - 0.098 0.278 0.002 - 0.171 0.010 0.473 0.194 0.509 0.693 GNPAT - 0.291 - 0.646 - 0.698 - 0.860 - 0.812 0.132 0.263 0.148 0.325 0.303 0.009 0.097 0.168 0.080 0.053 0.058 GNS 0.117 0.005 0.125 0.632 0.050 0.780 0.178 0.563 0.365 0.203 0.187 0.407 0.053 0.091 0.431 0.001 0.201 0.065 GOLGA1 0.463 0.009 0.188 0.045 0.389 0.016 GORAB 0.326 0.005 0.298 0.169 - 0.761 - 0.756 0.056 0.798 - 0.823 0.054 0.055 0.017 GPN1 0.524 0.409 0.611 0.017 - 0.538 - 0.009 0.177 0.543 0.085 0.705 0.108 0.270 GRHPR - 0.934 - 0.732 - 0.951 0.019 0.867 0.021 0.858 0.038 0.331 - 0.397 0.316 0.004 0.093 0.413 0.005 0.129 0.006 0.110 GRIPAP1 - 0.645 - 0.621 0.032 0.612 0.090 0.498 - 0.012 0.074 0.132 - 0.690 0.348 0.004 0.187 0.146 0.069 0.030 0.079 0.111 GRSF1 - 0.263 - 0.351 - 0.000 - 0.916 - 0.676 - 0.192 - 0.353 - 0.045 - 0.002 0.200 0.143 0.285 0.034 0.149 0.172 0.284 0.132 0.366 GSR 0.221 0.007 0.053 0.706 0.109 0.122 - 0.190 0.121 0.146 - 0.838 - 0.110 0.100 0.127 0.021 0.553 0.034 0.008 0.122

309 Appendix 1

GYS1 - 0.001 - 0.982 - 0.773 0.000 0.997 0.017 0.819 - 0.306 0.402 0.004 0.075 0.396 H1FX 0.070 0.739 0.148 0.054 0.141 0.029 0.058 0.664 0.242 0.208 0.203 0.001 0.047 0.806 0.080 0.810 0.184 0.052 HADHA 0.173 0.009 0.228 0.009 0.141 0.477 0.213 0.000 0.256 0.074 0.131 0.539 0.220 0.022 0.334 0.001 0.101 0.584 HADHB 0.174 0.005 0.243 0.008 0.170 0.450 0.201 0.003 0.208 0.000 0.100 0.694 0.233 0.073 0.228 0.002 0.097 0.654 HBS1L 0.122 0.419 0.043 0.883 - 0.901 - 0.299 - 0.001 0.004 0.154 0.217 HDHD2 - 0.333 - 0.620 - 0.001 - 0.484 0.118 0.124 0.229 0.196 HEATR5B 0.204 0.377 0.118 0.008 0.272 0.003 HIST3H2B 0.201 0.106 0.426 0.003 0.217 0.140 0.143 0.068 - 0.373 B 0.156 HK2 - 0.647 0.197 0.270 0.257 0.113 0.100 0.510 0.398 0.021 0.465 0.000 0.332 0.312 0.243 0.209 0.560 0.005 0.075 HMGCS1 0.031 0.572 - 0.010 0.727 HNRNPA0 0.020 0.787 - 0.000 - 0.865 0.065 0.393 - 0.173 0.046 0.856 0.175 0.162 - 0.009 - 0.402 0.073 0.005 0.349 0.449 0.093 HNRNPK 0.007 0.801 0.007 0.943 - 0.250 0.153 0.302 - 0.672 - 0.695 0.132 0.315 - 0.000 - 0.379 0.023 0.045 0.095 0.203 0.246 HOGA1 - 0.782 - 0.005 0.148 0.546 HOMER3 0.097 0.075 - 0.303 - 0.793 - 0.001 0.197 0.040 0.575 HRAS 0.202 0.001 0.116 0.253 - 0.328 0.082 HSD17B12 0.153 0.094 0.039 0.501 0.236 0.095 0.062 0.784 0.639 0.014 0.311 0.002 0.132 0.173 - 0.604 0.177 0.002 0.089 HSD17B8 - 0.589 - 0.008 0.161 0.059 - 0.257 0.123 0.327 0.231 HSP90B1 0.108 0.637 0.181 0.307 0.172 0.059 0.047 0.794 0.210 0.314 0.207 0.000 0.134 0.604 0.109 0.698 0.180 0.010

310 Appendix 1

HSPA5 0.112 0.718 0.220 0.344 0.242 0.028 0.130 0.611 0.218 0.490 0.215 0.000 0.180 0.571 0.148 0.686 0.211 0.006 HSPBP1 0.033 0.567 - 0.317 0.046 0.284 - 0.159 - 0.157 - 0.200 - 0.214 - 0.224 - 0.000 0.088 0.109 0.123 0.113 0.067 0.031 0.229 HSPH1 - 0.132 - 0.500 - 0.805 - 0.166 - 0.000 - 0.355 - 0.181 - 0.005 - 0.824 0.065 0.064 0.021 0.047 0.148 0.069 0.058 0.233 0.020 HTT 0.046 0.027 - 0.498 0.139 0.064 0.162 0.232 0.111 0.300 - 0.650 0.280 0.091 0.277 0.005 0.241 0.161 IDI1 0.142 0.534 - 0.060 0.048 0.689 - 0.803 - 0.001 0.009 0.900 0.190 0.360 - 0.000 - 0.373 0.047 0.056 0.231 0.279 0.082 IGSF1 0.270 0.815 0.271 0.835 0.821 0.006 - 0.990 - 0.997 0.588 0.161 0.087 0.949 - 0.993 0.832 0.067 0.014 0.005 0.011 ILVBL - 0.132 0.059 0.573 0.251 0.142 0.302 0.003 0.049 0.808 0.149 0.426 0.388 0.006 0.344 IPO13 - 0.241 - 0.009 - 0.268 - 0.074 - 0.809 - 0.960 0.276 0.208 0.245 0.288 0.043 0.017 ITGA1 0.199 0.388 0.164 0.541 0.140 0.265 0.076 0.786 0.232 0.552 0.191 0.101 0.285 0.518 0.279 0.446 0.215 0.008 IVNS1ABP 0.083 0.881 - 0.006 - 0.546 - 0.348 0.497 0.074 2.042 KCTD1 0.397 0.004 - 0.050 0.451 KCTD12 - 0.589 - 0.579 - 0.009 - 0.582 - 0.515 - 0.190 - 0.300 - 0.773 - 0.178 0.325 0.317 0.388 0.281 0.277 0.284 0.563 0.143 0.261 KCTD15 - 0.685 - 0.062 - 0.004 0.121 0.486 0.538 KEAP1 - 0.004 0.090 0.694 - 0.528 - 0.011 - 0.436 0.370 0.194 0.242 0.396 KIAA1279 0.327 0.234 0.073 0.651 0.303 0.001 0.099 0.545 0.008 0.928 0.100 0.408 0.182 0.144 - 0.766 0.021 0.851 0.078 KIAA1524 - 0.138 - 0.007 0.020 0.716 0.148 0.474 - 0.611 0.021 0.840 0.058 0.263 0.089 KIDINS220 0.297 0.052 0.417 0.000 0.111 0.740 0.374 0.005 0.409 0.089 0.108 0.808 0.555 0.001 - 0.603 0.205 KIF22 0.204 0.159 0.007 0.854 0.009 0.755 - 0.271 - 0.001

311 Appendix 1

0.446 0.748 KIF4A 0.175 0.406 - 0.485 0.049 0.206 0.211 0.002 - 0.344 - 0.677 0.217 0.286 - 0.211 - 0.119 0.101 0.116 0.026 0.291 0.107 KIF5A 0.142 0.436 0.266 0.215 0.441 0.015 0.523 0.004 - 0.889 0.038 KLHL13 - 0.693 - 0.007 0.028 0.865 0.003 0.984 - 0.174 - 0.741 - 0.768 - 0.286 0.070 0.329 0.119 0.092 0.026 0.279 KPNA2 0.028 0.218 - 0.429 0.012 0.817 0.115 0.583 - 0.004 - 0.008 0.286 0.005 - 0.019 - 0.066 0.121 0.379 0.300 0.394 0.272 KPNA3 0.015 0.403 0.070 0.000 0.064 0.531 0.026 0.677 0.042 0.020 0.031 0.665 0.075 0.143 0.206 0.003 - 0.955 0.011 KRI1 0.304 0.000 - 0.097 0.450 KRT87P 1.548 0.656 - 0.005 0.794 KTI12 - 0.015 - 0.000 - 0.108 0.134 0.273 0.222 LAMA5 0.126 0.756 0.131 0.496 0.090 0.110 - 0.985 0.166 0.699 0.211 0.002 0.210 0.544 - 0.970 0.355 0.001 0.007 0.014 LAMB1 0.173 0.606 0.109 0.661 0.203 0.001 0.049 0.840 0.071 0.597 0.154 0.143 0.031 0.950 0.056 0.809 0.217 0.035 LAMC1 0.303 0.513 0.184 0.229 0.231 0.353 0.126 0.469 0.188 0.465 0.323 0.003 0.067 0.850 0.149 0.671 0.286 0.053 LCLAT1 0.087 0.766 0.238 0.034 0.218 0.007 LDB1 0.151 0.846 0.303 0.418 0.363 0.000 - 0.139 0.482 LDHA - 0.942 0.044 0.763 0.097 0.006 - 0.878 0.032 0.624 0.126 0.073 - 0.858 0.109 0.306 0.224 0.002 0.007 0.012 0.017 LENG1 - 0.343 - 0.002 0.158 0.504 LEPREL2 0.169 0.016 0.022 0.901 0.122 0.316 0.126 0.004 0.193 0.462 0.256 0.000 0.077 0.060 0.015 0.958 LIG1 - 0.848 - 0.023 - 0.985 - 0.076 0.232 0.410 - 0.235 - 0.005 0.054 0.242 0.003 0.663 0.256 0.227 LMCD1 0.461 0.002 - 0.138 - 0.574

312 Appendix 1

0.381 0.207 LPP - 0.008 0.039 0.279 - 0.581 - 0.560 - 0.266 - 0.570 0.495 0.140 0.038 0.079 0.059 LRPAP1 - 0.233 - 0.865 0.060 0.739 0.102 0.308 0.229 0.008 - 0.057 - 0.347 0.286 0.037 0.167 0.973 LRRC58 - 0.003 - 0.476 - 0.110 0.016 0.878 0.193 0.037 0.020 0.904 0.299 0.076 0.140 LTV1 - 0.286 - 0.009 - 0.242 - 0.033 - 0.025 - 0.721 - 0.076 0.180 0.363 0.221 0.176 0.320 0.021 0.441 LZTFL1 0.045 0.488 - 0.001 - 0.555 - 0.231 - 0.092 - 0.492 - 0.482 0.062 0.564 0.226 0.105 0.104 0.355 0.147 0.090 MAN2C1 - 0.000 0.840 MAOA 0.251 0.002 0.239 0.001 0.100 0.695 0.292 0.076 0.387 0.054 0.221 0.397 0.429 0.005 0.427 0.058 0.277 0.365 MAP1LC3B 0.131 0.653 - 0.743 - 0.971 0.061 0.894 0.275 0.000 0.088 0.588 0.259 0.140 0.109 0.196 0.067 0.010 MAP2K6 - 0.000 - 0.101 - 0.133 0.041 0.209 - 0.359 0.083 0.038 - 0.268 0.217 0.294 0.108 0.300 0.509 MAP7 0.341 0.000 - 0.709 0.095 MAP7D1 0.101 0.179 - 0.996 0.012 0.864 0.169 0.544 0.046 0.846 0.202 0.006 0.115 0.859 - 0.773 0.132 0.336 0.001 0.057 MARCKSL1 - 0.046 - 0.963 - 0.742 0.326 0.412 - 0.556 - 0.452 - 0.000 0.244 0.431 0.094 0.008 0.030 0.399 0.155 0.392 MCCC1 - 0.672 - 0.210 - 0.117 - 0.715 - 0.264 - 0.587 - 0.481 - 0.004 0.056 0.456 0.141 0.118 0.067 0.184 0.140 0.259 MCCC2 0.084 0.700 - 0.002 - 0.658 - 0.718 - 0.883 - 0.003 - 0.440 - 0.002 - 0.011 0.104 0.013 0.014 0.015 0.173 0.060 0.405 0.320 MCM3 - 0.246 - 0.074 - 0.008 0.067 0.430 - 0.424 - 0.010 - 0.304 - 0.015 - 0.000 0.090 0.157 0.076 0.123 0.094 0.044 0.249 0.210 MCM5 - 0.767 - 0.014 - 0.044 0.051 0.104 - 0.116 - 0.006 - 0.546 - 0.059 - 0.000 0.017 0.156 0.073 0.155 0.104 0.028 0.272 0.221 MCM7 - 0.562 - 0.000 - 0.250 0.080 0.133 - 0.064 - 0.011 - 0.111 - 0.053 - 0.000

313 Appendix 1

0.016 0.199 0.061 0.199 0.110 0.012 0.244 0.211 MCTS1 - 0.218 - 0.000 0.000 0.997 - 0.455 0.128 0.663 - 0.444 0.254 0.002 - 0.663 0.257 0.116 0.065 0.034 0.127 MDC1 0.145 0.003 - 0.298 0.047 0.719 0.241 0.001 0.004 0.964 0.155 0.154 0.062 MDN1 0.023 0.540 0.097 0.733 - 0.822 0.059 0.002 - 0.005 - 0.400 0.317 0.507 - 0.204 0.027 0.249 0.072 0.331 MEMO1 0.106 0.549 0.051 0.558 - 0.748 - 0.007 - 0.588 - 0.158 - 0.474 - 0.965 - 0.005 0.029 0.028 0.073 0.072 0.181 0.005 0.262 METAP2 - 0.090 - 0.085 - 0.039 - 0.155 - 0.138 - 0.000 - 0.381 - 0.445 - 0.005 0.089 0.199 0.123 0.042 0.209 0.282 0.148 0.169 0.340 MFF - 0.572 - 0.007 0.015 0.883 - 0.114 0.065 0.296 0.503 MOCS3 - 0.450 - 0.005 - 0.385 - 0.032 - 0.001 0.203 0.401 0.147 0.299 0.567 MPC2 - 0.656 0.226 0.185 0.121 0.464 0.035 0.572 0.241 0.007 0.071 0.554 0.063 MRPL16 0.055 0.152 0.013 0.852 - 0.873 - 0.004 - 0.285 - 0.091 - 0.004 0.011 0.139 0.207 0.352 0.562 MT-CO2 0.181 0.554 0.222 0.501 0.398 0.000 0.146 0.319 0.126 0.644 0.266 0.002 MVP 0.252 0.776 0.377 0.733 0.457 0.003 0.164 0.828 0.520 0.477 0.772 0.000 MYO6 0.052 0.676 0.142 0.335 0.229 0.041 0.111 0.384 0.166 0.000 0.192 0.009 0.082 0.664 0.211 0.012 0.242 0.008 N4BP1 - 0.006 - 0.069 0.214 0.537 NAA25 - 0.002 - 0.133 0.100 0.323 - 0.173 - 0.046 - 0.112 0.349 0.360 0.376 0.454 0.305 NAA35 0.316 0.009 - 0.309 - 0.191 0.372 1.530 NAB2 - 0.006 - 0.000 0.299 0.524 NACC1 0.114 0.604 - 0.001 - 0.014 0.072 0.768 - 0.198 - 0.553 0.215 0.087 0.314 0.057

314 Appendix 1

NADK2 0.088 0.033 - 0.745 - 0.010 0.097 0.187 - 0.415 - 0.903 0.172 0.574 - 0.619 0.044 0.231 0.076 0.002 0.376 NCOA7 0.009 0.977 0.008 0.977 0.253 0.056 0.187 0.465 0.383 0.009 0.376 0.172 0.318 0.265 NCOR2 0.071 0.569 - 0.699 - 0.661 0.105 0.012 - 0.008 - 0.879 0.035 0.044 0.329 0.015 NDUFA10 0.214 0.084 0.136 0.366 0.235 0.000 0.106 0.444 - 0.782 0.171 0.047 0.024 NDUFA11 0.327 0.349 0.027 0.854 0.176 0.270 0.264 0.000 0.059 0.838 NDUFA6 0.221 0.040 0.076 0.017 0.195 0.085 0.183 0.116 - 0.795 0.158 0.189 - 0.419 0.334 0.006 0.054 0.119 NDUFA9 0.068 0.766 0.124 0.654 0.286 0.000 0.087 0.008 0.278 0.000 0.283 0.006 - 0.036 0.237 0.206 0.292 0.006 0.051 NDUFB4 0.447 0.000 - 0.638 0.307 0.217 - 0.389 - 0.986 0.140 0.461 0.004 NDUFB7 0.232 0.196 0.245 0.476 0.391 0.000 0.157 0.045 0.398 0.339 0.410 0.000 NDUFB8 0.283 0.424 0.242 0.327 0.310 0.015 0.280 0.053 0.177 0.620 0.472 0.003 0.235 0.317 0.309 0.017 0.467 0.001 NDUFB9 - 0.022 - 0.807 0.235 0.009 0.221 0.631 0.365 0.000 0.246 0.058 NDUFS1 0.281 0.519 - 0.785 0.207 0.003 0.165 0.265 - 0.805 0.184 0.016 0.186 0.049 0.020 0.823 0.146 0.002 0.075 0.010 NDUFS2 - 0.754 0.093 0.655 0.253 0.033 0.115 0.488 0.184 0.479 0.257 0.004 0.273 0.044 - 0.902 0.186 0.006 0.105 0.026 NDUFS8 0.002 0.992 0.135 0.224 0.187 0.023 - 0.908 0.117 0.410 0.229 0.003 0.289 0.299 - 0.547 0.137 0.273 0.014 0.108 NDUFV1 0.322 0.003 0.194 0.072 0.148 0.030 0.117 0.517 - 0.841 0.102 0.045 0.311 0.122 - 0.812 0.239 0.024 0.063 0.032 NME4 - 0.192 - 0.011 - 0.180 - 0.034 - 0.414 - 0.001 - 0.542 - 0.004 - 0.004 0.152 0.136 0.102 0.068 0.019 0.158 0.063 0.187 0.255 NMRAL1 0.000 0.991 - 0.006 - 0.007 0.353 0.253 NOC2L - 0.143 0.005 0.977 - 0.298 0.117 0.289 - 0.788 - 0.514 0.024 0.794 - 0.304 - 0.001 0.141 0.042 0.059 0.039 0.244 0.222

315 Appendix 1

NOP14 0.082 0.593 0.107 0.611 - 0.444 0.035 0.920 - 0.043 - 0.002 0.219 0.542 - 0.374 - 0.428 0.048 0.208 0.288 0.070 0.154 NOSIP 0.031 0.635 - 0.000 - 0.061 0.028 0.707 - 0.113 - 0.028 0.000 0.990 - 0.197 0.233 0.138 0.059 0.159 0.845 NOVA2 - 0.056 - 0.227 0.061 0.751 0.101 0.270 - 0.004 - 0.435 0.021 0.606 - 0.087 - 0.478 0.210 0.153 0.452 0.070 0.328 0.070 NPC2 0.313 0.001 - 0.417 0.237 0.618 - 0.855 0.126 0.055 NR2C2 - 0.804 - 0.000 0.014 0.938 0.073 0.408 NSUN5 - 0.973 - 0.497 0.127 0.023 - 0.225 - 0.685 0.113 0.000 - 0.000 0.004 0.187 0.197 0.030 0.494 NT5C2 - 0.965 - 0.838 - 0.871 - 0.206 0.119 0.609 0.220 0.016 - 0.621 0.191 0.012 0.281 0.004 0.019 0.085 0.027 0.108 0.075 NT5DC3 0.319 0.002 - 0.387 0.247 NUDT1 - 0.137 - 0.021 0.104 0.021 0.217 0.002 - 0.921 0.107 0.263 - 0.401 - 0.192 0.083 0.068 0.005 0.077 0.986 NVL - 0.630 0.184 0.065 - 0.800 - 0.007 - 0.013 0.019 0.906 - 0.490 - 0.074 0.078 0.040 0.259 0.473 0.201 0.260 OAT 0.084 0.055 0.246 0.009 0.149 0.357 0.189 0.001 0.503 0.005 0.373 0.104 0.201 0.302 0.605 0.005 0.337 0.181 OCIAD2 0.050 0.746 - 0.211 - 0.199 0.024 0.753 0.034 0.184 - 0.327 - 0.621 - 0.010 - 0.000 0.053 0.156 0.051 0.062 0.160 0.227 OGFOD1 - 0.002 - 0.211 - 0.003 0.200 0.191 0.234 ORMDL3 - 0.260 0.232 0.262 0.104 0.282 0.063 0.800 - 0.822 0.127 0.508 0.130 0.187 0.293 0.000 0.043 0.063 OSGEP - 0.001 0.021 0.611 0.015 0.877 0.070 0.778 - 0.967 - 0.000 - 0.006 - 0.326 0.013 0.818 0.101 0.010 0.227 0.331 0.089 P4HA1 - 0.459 0.182 0.475 0.071 0.325 0.247 0.326 0.091 0.588 0.265 0.003 0.306 0.186 - 0.832 0.192 0.134 0.180 0.068 P4HA2 0.610 0.475 0.247 0.329 0.357 0.008 0.320 0.422 0.316 0.418 0.366 0.000 0.238 0.552 0.096 0.759 P4HB 0.018 0.917 0.086 0.519 0.092 0.077 0.046 0.830 0.198 0.400 0.218 0.000 0.124 0.652 0.104 0.718 0.177 0.017

316 Appendix 1

PAF1 - 0.011 - 0.319 0.112 0.003 - 0.451 - 0.088 0.016 0.890 0.057 0.263 - 0.875 0.317 0.354 0.271 0.182 0.023 PAFAH1B1 0.063 0.551 0.089 0.165 0.117 0.001 0.029 0.826 0.025 0.893 0.087 0.007 0.040 0.617 0.006 0.970 0.211 0.001 PARVB 0.308 0.009 0.262 0.050 0.176 0.001 PCBP3 0.577 0.009 - 0.394 0.611 0.045 0.264 0.437 0.203 0.589 0.287 PCGF3 - 0.628 - 0.005 0.136 0.570 PCNA - 0.212 - 0.035 - 0.378 0.041 0.465 - 0.015 - 0.007 - 0.574 - 0.018 - 0.000 0.025 0.106 0.038 0.240 0.159 0.016 0.291 0.239 PDIA3 0.096 0.559 0.148 0.267 0.203 0.000 0.124 0.407 0.118 0.553 0.181 0.091 0.135 0.508 0.112 0.689 0.145 0.016 PDIA4 0.160 0.436 0.247 0.179 0.243 0.116 0.118 0.503 0.241 0.328 0.222 0.043 0.237 0.336 0.157 0.614 0.213 0.000 PDIA6 0.081 0.571 0.264 0.042 0.130 0.305 0.116 0.344 0.252 0.159 0.248 0.020 0.194 0.280 0.185 0.444 0.208 0.000 PDXK - 0.434 - 0.426 0.121 0.451 0.011 0.872 - 0.003 - 0.325 - 0.004 - 0.281 - 0.855 0.100 0.102 0.156 0.073 0.387 0.108 0.031 PDZRN3 - 0.003 - 0.134 0.449 0.545 PEG10 0.005 0.918 - 0.007 0.370 PES1 - 0.215 - 0.277 - 0.105 0.001 0.992 - 0.001 - 0.039 0.007 0.926 - 0.356 - 0.000 0.072 0.097 0.105 0.201 0.112 0.185 0.122 PEX14 - 0.000 - 0.158 0.004 0.938 - 0.314 - 0.763 0.382 0.327 0.128 0.034 PFDN2 - 0.535 - 0.466 - 0.036 0.031 0.859 - 0.780 - 0.518 - 0.008 - 0.042 - 0.775 0.044 0.081 0.111 0.053 0.065 0.265 0.063 0.028 PFDN4 - 0.211 - 0.000 0.015 0.882 0.025 0.260 PFKL 0.020 0.753 0.105 0.382 0.074 0.015 - 0.889 0.149 0.292 0.227 0.002 0.010 0.962 0.066 0.483 0.303 0.000 0.017 PFKP 0.034 0.730 0.038 0.759 0.023 0.802 0.087 0.116 0.113 0.356 0.204 0.008 0.032 0.606 0.265 0.082 0.272 0.011 PGM1 0.078 0.293 0.040 0.812 0.088 0.029 0.043 0.721 0.058 0.663 0.087 0.100 0.111 0.491 0.123 0.457 0.220 0.000

317 Appendix 1

PHAX - 0.247 - 0.000 0.098 0.415 - 0.193 0.100 0.631 0.452 PHKB 0.206 0.009 - 0.193 0.227 0.079 0.168 PHRF1 0.135 0.651 - 0.278 - 0.008 0.340 0.263 PISD - 0.003 - 0.222 0.260 0.345 PLOD1 0.155 0.481 0.118 0.610 0.194 0.038 0.217 0.311 0.304 0.266 0.409 0.009 0.271 0.312 0.248 0.483 0.259 0.000 PNMA1 0.220 0.000 - 0.161 0.433 PNPLA6 0.830 0.202 0.032 0.943 0.042 0.851 0.102 0.780 0.215 0.002 0.001 0.994 0.227 0.434 0.315 0.116 POFUT1 - 0.001 - 0.017 - 0.784 0.034 0.119 0.127 0.330 - 0.098 0.253 0.379 0.129 0.197 0.693 0.511 0.039 0.111 POLR2C 0.003 0.949 - 0.024 - 0.822 - 0.827 - 0.347 0.088 0.388 - 0.003 - 0.092 0.202 0.034 0.011 0.125 0.361 0.363 POMGNT1 - 0.779 - 0.000 - 0.437 - 0.175 0.125 0.362 0.170 0.408 PPAT 0.085 0.489 0.103 0.368 0.072 0.278 0.204 0.001 0.161 0.006 - 0.934 - 0.584 0.098 0.383 0.010 0.906 0.009 0.044 PPHLN1 - 0.000 - 0.462 - 0.917 0.375 0.001 - 0.250 - 0.630 - 0.092 0.174 0.222 0.014 0.239 0.034 0.393 PPIP5K2 - 0.813 - 0.082 0.221 0.000 - 0.486 0.127 0.048 0.027 0.765 0.249 0.029 - 0.986 0.055 0.332 0.227 0.002 PPP2R4 0.352 0.001 - 0.093 0.012 0.936 0.128 0.471 - 0.626 - 0.247 - 0.317 - 0.584 - 0.405 0.193 0.039 0.123 0.133 0.172 0.084 PRDX4 0.103 0.036 0.119 0.196 0.075 0.365 - 0.967 0.163 0.015 0.244 0.000 0.114 0.191 0.135 0.297 0.201 0.000 0.002 PRKAB1 - 0.685 - 0.002 0.036 0.875 0.111 0.698 0.015 0.699 0.006 0.991 - 0.321 - 0.155 0.048 0.306 0.149 0.179 PRKD2 0.053 0.736 - 0.907 0.069 0.283 0.148 0.490 - 0.320 - 0.370 0.087 0.560 0.273 0.002 0.053 0.383 0.013 0.288 0.090 PRPF3 - 0.692 0.032 0.587 - 0.004 - 0.055 - 0.083 - 0.289 0.052 0.613 - 0.706 - 0.899

318 Appendix 1

0.047 0.201 0.047 0.083 0.061 0.121 0.022 PRPF4B - 0.000 0.110 0.602 0.130 0.481 0.062 0.420 0.204 0.382 0.189 0.179 - 0.541 0.069 0.043 - 0.363 0.304 0.163 0.110 PSME2 0.035 0.538 - 0.008 0.033 0.290 0.022 0.672 - 0.480 0.014 0.878 - 0.641 - 0.058 0.076 0.350 0.248 0.041 0.038 0.072 PSMF1 0.024 0.812 - 0.002 - 0.912 - 0.118 - 0.128 - 0.171 - 0.177 - 0.048 - 0.147 0.200 0.013 0.105 0.336 0.132 0.217 0.216 0.109 PTGES2 - 0.383 0.007 0.656 0.178 0.251 0.168 0.003 0.049 0.670 0.237 0.007 0.221 0.000 0.285 QPRT - 0.507 - 0.056 - 0.938 - 0.032 - 0.119 0.061 0.688 - 0.009 0.109 0.266 0.062 0.636 0.055 0.106 0.014 0.172 0.125 0.304 RAB18 0.019 0.918 0.234 0.243 0.074 0.447 0.116 0.600 0.375 0.137 0.312 0.009 0.220 0.270 0.164 0.275 0.084 0.109 RAB1A 0.076 0.761 0.164 0.493 0.237 0.017 0.164 0.617 0.217 0.368 0.205 0.000 0.133 0.454 0.226 0.312 0.205 0.003 RAB2B - 0.967 - 0.996 0.045 0.690 0.046 0.671 - 0.515 0.205 0.001 0.486 0.295 0.006 0.981 0.010 0.001 0.124 RAB32 0.106 0.776 0.564 0.351 0.222 0.134 - 0.429 0.284 0.083 0.253 0.005 0.117 0.715 0.086 0.674 0.371 0.119 0.061 RAB33B - 0.366 0.366 0.000 0.378 0.301 0.085 0.620 - 0.377 0.038 0.495 RAB3D - 0.670 - 0.497 0.108 0.485 0.515 0.049 0.328 0.004 0.073 0.313 RABAC1 0.395 0.005 - 0.660 0.182 RABL6 0.272 0.001 0.075 0.358 0.016 0.902 0.093 0.691 - 0.054 - 0.143 0.120 0.074 RACGAP1 0.515 0.000 - 0.878 0.024 RALA - 0.116 0.069 0.103 0.036 0.373 - 0.024 - 0.598 0.113 0.220 0.219 0.002 - 0.143 0.178 0.098 0.057 0.064 0.107 0.162 RASA1 - 0.398 - 0.935 0.044 0.309 - 0.994 0.319 0.000 0.167 0.004 0.011 0.012 0.002 RASA4 - 0.016 - 0.000 - 0.680 - 0.176 0.038 0.674 0.071 0.043 0.188 0.272 0.110 0.044

319 Appendix 1

RB1 - 0.144 - 0.654 - 0.001 - 0.138 0.397 0.093 0.319 0.180 RBM14 0.017 0.368 0.020 0.796 - 0.435 0.016 0.294 - 0.170 - 0.389 0.012 0.293 - 0.051 - 0.003 0.033 0.086 0.068 0.097 0.203 RBM15B - 0.621 - 0.034 - 0.506 0.260 0.000 0.057 0.553 0.116 0.216 0.002 0.994 - 0.004 - 0.141 0.001 0.101 0.038 0.152 0.094 RBM19 - 0.155 0.109 0.599 - 0.008 0.427 0.481 RBM3 0.036 0.239 - 0.117 - 0.409 0.233 0.078 0.015 0.953 0.012 0.922 0.072 0.321 - 0.004 - 0.394 0.198 0.035 0.290 0.084 RBM5 0.194 0.080 - 0.000 - 0.010 0.024 0.938 - 0.138 - 0.003 0.057 0.121 0.467 0.309 RBP1 0.032 0.794 - 0.114 - 0.691 - 0.314 - 0.000 - 0.001 - 0.354 - 0.681 - 0.429 0.055 0.019 0.016 0.078 0.080 0.030 0.041 0.044 RBPJ 0.260 0.000 0.207 0.255 0.080 0.389 - 0.543 - 0.243 0.029 0.459 RCC1 0.039 0.873 0.005 0.942 - 0.796 - 0.476 0.069 0.539 0.014 0.753 - 0.930 - 0.004 - 0.016 0.021 0.030 0.021 0.229 0.138 REC8 - 0.021 - 0.047 - 0.003 0.166 0.570 0.377 REEP5 0.023 0.224 0.213 0.147 0.131 0.058 0.049 0.687 - 0.921 0.285 0.005 0.019 0.456 0.254 0.308 0.181 0.000 0.035 RENBP - 0.533 0.594 0.256 0.219 0.001 0.281 RIF1 0.825 0.042 - 0.005 - 0.453 - 0.018 - 0.384 0.621 0.062 0.121 0.343 RPL22L1 - 0.995 - 0.233 - 0.000 0.116 0.757 - 0.018 0.001 0.189 0.262 0.603 RPRD1B - 0.044 - 0.009 - 0.565 - 0.000 - 0.005 - 0.218 - 0.007 - 0.214 - 0.011 0.112 0.199 0.061 0.156 0.198 0.121 0.246 0.362 0.226 RPRD2 - 0.411 0.049 0.426 - 0.523 0.013 0.945 - 0.329 - 0.086 0.409 0.000 - 0.974 0.047 0.023 0.024 0.243 0.028 RPS19 - 0.632 - 0.385 - 0.184 - 0.840 - 0.506 0.002 0.661 0.076 0.459 - 0.003 0.131 0.452

320 Appendix 1

0.022 0.066 0.022 0.010 0.032 0.411 RPS19BP1 0.287 0.007 - 0.102 0.406 RPUSD4 - 0.020 - 0.002 0.100 0.593 RQCD1 0.201 0.034 - 0.162 0.004 0.964 - 0.002 0.067 0.777 0.013 0.874 0.040 0.419 0.006 0.970 0.296 0.010 0.151 0.126 RRM1 - 0.464 - 0.007 - 0.104 0.074 0.500 - 0.275 - 0.037 0.074 0.278 - 0.398 - 0.000 0.135 0.261 0.070 0.235 0.162 0.277 0.372 RTFDC1 - 0.005 - 0.306 - 0.409 0.305 0.209 0.167 S100A11 0.310 0.115 0.237 0.343 0.134 0.646 0.261 0.008 0.286 0.086 0.097 0.658 0.122 0.416 0.223 0.031 0.171 0.398 SACM1L - 0.389 0.279 0.000 0.198 0.000 - 0.766 0.148 0.566 0.149 0.023 - 0.809 0.216 0.278 0.209 0.000 0.075 0.043 0.027 SAR1A 0.299 0.304 0.084 0.610 0.284 0.000 - 0.816 0.133 0.374 0.135 0.070 - 0.133 0.120 0.519 0.199 0.213 0.033 0.033 SARM1 0.203 0.373 0.101 0.311 0.089 0.034 0.083 0.516 0.158 0.268 0.052 0.698 0.148 0.085 0.261 0.003 0.082 0.526 SART1 - 0.239 - 0.362 - 0.037 0.057 0.537 - 0.260 - 0.153 0.041 0.523 - 0.004 - 0.577 0.067 0.079 0.060 0.133 0.089 0.367 0.076 SDE2 0.022 0.078 - 0.003 - 0.040 - 0.062 - 0.229 0.278 0.080 0.275 0.317 SEPT10 0.310 0.001 SERBP1 - 0.753 0.001 0.985 - 0.907 0.076 0.589 - 0.777 0.069 0.371 0.043 0.781 - 0.001 0.127 0.742 0.021 0.005 0.045 0.958 SETD3 - 0.046 - 0.007 0.194 0.419 SF1 - 0.620 - 0.497 - 0.229 - 0.802 - 0.129 - 0.027 - 0.873 - 0.402 - 0.031 0.054 0.047 0.051 0.086 0.055 0.538 0.010 0.153 0.071 SF3B2 0.030 0.759 - 0.575 - 0.000 0.051 0.461 - 0.320 - 0.109 0.039 0.329 - 0.001 - 0.884 0.062 0.065 0.032 0.046 0.202 0.010 SGSH - 0.006 - 0.107 0.165 0.290 0.063 0.009 0.067 0.667 - 0.126 0.106 0.617 0.096 0.021 0.261 0.221 0.094

321 Appendix 1

SH3PXD2B 0.213 0.002 0.128 0.181 0.089 0.566 0.263 0.092 0.174 0.125 0.051 0.724 0.174 0.175 0.059 0.100 0.091 0.677 SHD - 0.009 - 0.032 - 0.585 - 0.280 0.266 0.751 - 0.227 - 0.060 - 0.408 0.220 0.611 0.202 0.245 0.125 0.138 0.215 0.341 0.255 SLC27A3 0.094 0.802 0.228 0.590 0.114 0.172 - 0.979 0.213 0.525 0.280 0.015 0.097 0.758 0.059 0.894 0.333 0.001 0.009 SLC27A4 0.049 0.675 0.038 0.093 0.024 0.710 0.082 0.715 0.242 0.485 0.243 0.009 0.005 0.984 0.099 0.608 - 0.934 0.007 SMAP 0.069 0.361 - 0.165 - 0.007 0.120 0.207 SMAP2 - 0.712 - 0.004 0.232 0.627 SMARCA1 0.116 0.269 0.066 0.013 0.044 0.504 0.052 0.559 0.108 0.578 0.004 0.926 0.175 0.323 0.122 0.140 0.221 0.003 SMARCD2 0.147 0.595 0.240 0.311 0.105 0.236 0.080 0.604 - 0.706 0.008 0.927 - 0.005 - 0.889 0.146 0.594 0.027 SMG6 - 0.585 - 0.005 0.057 0.570 SMNDC1 0.424 0.195 - 0.947 - 0.001 - 0.451 0.013 0.295 0.433 SNIP1 - 0.003 - 0.406 - 0.429 - 0.398 0.291 0.204 0.485 0.114 SNRPG - 0.757 - 0.732 0.040 0.641 - 0.000 0.009 0.052 0.780 SNX21 - 0.007 - 0.350 0.304 0.778 SNX27 0.038 0.402 0.146 0.137 - 0.693 0.040 0.731 0.029 0.504 0.134 0.061 - 0.560 0.128 0.030 0.242 0.005 0.036 0.078 SNX4 0.119 0.294 0.069 0.329 0.037 0.749 0.215 0.001 0.121 0.135 0.109 0.153 0.074 0.122 0.001 0.995 - 0.780 0.012 SOGA3 0.549 0.152 0.852 0.074 0.379 0.299 0.453 0.003 0.423 0.040 0.191 0.618 0.096 0.421 0.116 0.784 0.165 0.401 SPAG7 0.290 0.003 - 0.423 0.298 SPECC1L 0.556 0.000 0.164 0.395 - 0.726 0.220 0.069 - 0.478 - 0.319 0.030 0.057 0.090

322 Appendix 1

SREBF1 0.531 0.000 0.233 0.742 SRM 0.326 0.000 0.190 0.166 - 0.476 0.046 0.645 0.020 0.776 - 0.067 - 0.914 - 0.843 - 0.316 0.029 0.078 0.024 0.051 0.109 SRSF11 0.002 0.991 - 0.001 - 0.363 0.049 0.313 0.032 0.510 - 0.676 0.014 0.961 - 0.489 - 0.062 0.264 0.092 0.039 0.239 0.072 STAG1 - 0.701 0.033 0.926 - 0.975 0.287 0.286 - 0.002 - 0.497 0.146 0.003 0.593 0.115 STAT1 0.190 0.220 - 0.003 - 0.982 - 0.497 - 0.015 - 0.590 - 0.617 - 0.407 - 0.546 0.232 0.001 0.276 0.201 0.060 0.079 0.051 0.072 STK24 0.394 0.002 - 0.791 0.017 0.609 - 0.275 - 0.742 0.153 0.619 0.309 0.014 0.075 0.035 0.078 0.053 0.033 STMN1 0.037 0.150 0.026 0.698 0.088 0.017 0.067 0.290 0.080 0.497 0.208 0.001 - 0.971 - 0.957 0.054 0.758 0.009 0.005 STX12 0.285 0.003 0.152 0.567 0.120 0.133 0.030 0.528 0.059 0.361 0.088 0.163 - 0.489 0.237 0.077 0.102 0.074 0.063 SUCLG2 0.025 0.814 0.071 0.073 0.101 0.007 0.073 0.332 0.153 0.241 0.065 0.642 0.101 0.450 0.280 0.002 0.099 0.468 SUGT1 - 0.592 - 0.286 - 0.755 - 0.020 - 0.000 - 0.022 - 0.797 - 0.001 - 0.563 0.042 0.127 0.033 0.208 0.199 0.513 0.062 0.414 0.109 SUN1 0.001 0.975 0.127 0.222 0.150 0.076 0.029 0.473 0.232 0.175 0.127 0.026 0.206 0.041 0.383 0.128 0.203 0.007 SYNE3 0.390 0.157 - 0.809 0.233 0.007 0.359 0.091 0.182 0.475 0.244 0.095 0.044 0.870 0.134 SYP 0.714 0.000 0.009 0.985 - 0.464 0.473 TAF2 0.103 0.410 - 0.004 0.521 TES 0.522 0.387 0.294 0.770 0.683 0.001 0.126 0.767 0.052 0.874 0.030 0.866 0.138 0.530 0.296 0.000 TH 1.388 0.080 1.358 0.121 0.511 0.729 1.470 0.052 1.543 0.068 0.617 0.657 1.385 0.029 1.346 0.005 0.742 0.538 TM9SF2 - 0.681 0.146 0.386 0.217 0.005 0.085 0.700 - 0.959 0.149 0.017 0.028 0.811 - 0.244 - 0.305 0.133 0.007 0.042 0.075 TMED7 0.171 0.240 0.034 0.658 0.071 0.506 0.042 0.878 0.104 0.440 0.153 0.000 0.228 0.008 0.207 0.113 0.189 0.003 TMEM9 0.002 0.968 - 0.558 - 0.000 0.361 0.075 - 0.741 0.133 0.883 0.128

323 Appendix 1

TMOD3 0.155 0.509 - 0.904 0.144 0.187 0.175 0.567 0.230 0.018 0.203 0.403 0.037 0.540 0.339 0.000 0.039 TONSL - 0.973 - 0.001 0.009 0.970 - 0.059 - 0.851 0.016 0.336 0.087 0.022 TPP1 0.010 0.893 0.141 0.000 0.103 0.027 - 0.422 0.087 0.258 0.085 0.001 0.014 0.747 0.205 0.087 0.231 0.001 0.038 TPPP3 0.076 0.907 0.341 0.645 0.334 0.007 0.350 0.620 0.357 0.442 0.494 0.001 TRAFD1 - 0.340 - 0.382 0.180 0.068 - 0.006 0.027 0.720 0.198 0.140 0.518 TRIM5 0.039 0.977 - 0.009 0.486 TRMT2A - 0.038 - 0.022 - 0.001 - 0.132 0.221 0.544 0.301 0.822 TRPT1 - 0.017 - 0.053 0.375 0.455 TS - 0.008 0.633 TSNAX - 0.578 - 0.255 - 0.920 0.079 0.635 - 0.001 - 0.181 - 0.223 - 0.757 0.038 0.187 0.013 0.223 0.119 0.150 0.011 TTC28 0.002 0.915 - 0.460 0.028 0.895 0.027 0.823 - 0.831 0.248 0.005 - 0.702 - 0.567 0.089 0.009 0.412 0.152 TXN 0.011 0.771 - 0.067 0.053 0.335 - 0.149 - 0.276 - 0.265 - 0.001 0.048 0.026 - 0.444 0.070 0.103 0.071 0.077 0.263 0.049 TXNDC12 0.133 0.447 0.208 0.002 0.175 0.338 0.041 0.690 0.211 0.000 0.082 0.297 - 0.000 - 0.087 0.242 0.353 TXNDC15 - 0.900 - 0.211 0.057 0.765 0.097 0.172 0.073 0.001 0.070 0.906 - 0.702 - 0.009 0.075 0.284 0.174 0.295 UBE2O 0.088 0.098 0.144 0.180 0.079 0.076 0.105 0.331 0.154 0.004 0.102 0.000 0.035 0.821 - 0.135 0.214 0.003 0.013 UBE2S 0.056 0.708 - 0.489 0.084 0.021 0.086 0.867 - 0.088 0.265 0.002 - 0.438 0.180 0.556 0.198 UGDH 0.025 0.736 0.041 0.519 - 0.450 0.005 0.859 - 0.032 - 0.022 - 0.279 - 0.064 - 0.000 0.019 0.116 0.107 0.062 0.058 0.230

324 Appendix 1

UHRF1 0.104 0.155 - 0.172 0.177 0.176 0.040 0.931 - 0.001 0.185 0.434 - 0.093 0.456 0.312 0.464 UNG - 0.433 - 0.000 0.158 0.520 UNK 0.264 0.172 - 0.682 0.204 0.003 - 0.339 0.033 0.455 UQCRB 0.192 0.449 0.025 0.812 0.331 0.000 0.261 0.376 0.083 0.245 0.135 0.181 0.109 0.657 0.178 0.449 0.159 0.021 UQCRH - 0.893 - 0.194 - 0.816 - 0.967 0.223 0.008 0.003 0.983 - 0.162 0.064 0.615 0.028 0.137 0.019 0.006 0.077 URGCP - 0.673 - 0.002 0.141 0.198 - 0.099 0.110 0.225 0.584 USP11 - 0.019 - 0.516 0.010 0.899 0.342 0.062 - 0.006 - 0.220 0.118 0.146 0.305 0.201 USP34 0.143 0.300 0.286 0.511 0.256 0.000 UTP6 - 0.000 - 0.851 0.002 0.994 0.085 0.331 0.060 0.056 0.109 0.422 - 0.240 - 0.028 0.252 0.067 0.051 0.503 VAMP2 0.201 0.698 - 0.474 0.095 0.549 0.161 0.421 0.212 0.000 0.187 0.204 - 0.827 0.224 0.054 VAT1 0.134 0.097 0.105 0.000 0.118 0.126 0.093 0.563 0.244 0.126 0.197 0.001 0.070 0.626 0.262 0.062 0.251 0.001 VGF 1.210 0.025 1.361 0.003 0.696 0.585 1.002 0.022 1.013 0.001 0.713 0.466 0.741 0.019 - 0.706 0.374 WDFY3 - 0.380 - 0.002 0.333 0.565 ZC3HAV1L 0.234 0.370 0.206 0.277 0.192 0.045 0.157 0.372 0.008 0.953 0.144 0.032 0.224 0.312 - 0.526 0.273 0.000 0.015 ZDHHC5 - 0.013 - 0.031 0.482 0.585 ZMYM2 - 0.574 - 0.168 0.042 0.607 - 0.002 - 0.211 - 0.695 - 0.387 0.099 0.156 0.465 0.147 0.018 1.133 ZNF511 - 0.032 - 0.155 0.256 0.527 ZNF512B 0.016 0.984 - 0.006 0.604

325 Appendix 1

ZNF581 0.324 0.074 - 0.000 0.646 ZNF787 - 0.324 - 0.295 - 0.233 - 0.288 - 0.004 - 0.478 0.016 0.501 - 0.794 - 0.004 0.077 0.149 0.082 0.051 0.122 0.078 0.303 0.216 ZSWIM8 0.019 0.983 - 0.395 - 0.501 0.230 0.005 - 0.313 0.226 0.086 3.692

326 Appendix 2

18 HADHB hydroxyacyl-CoA APPENDIX 2 dehydrogenase/3-ketoacyl- CoA thiolase/enoyl-CoA hydratase (trifunctional Table A2 KEGG pathway analysis of protein), beta subunit proteins dysregulated by treatment of SH- 19 NDUFB8 NADH dehydrogenase SY5Y expressing moderately elevated levels (ubiquinone) 1 beta of αSynuclein, treated with 10 nM subcomplex, 8, 19kDa 20 SUCLG2 succinate-CoA ligase, GDP- rotenone or a combination of these insults. forming, beta subunit Proteins dysregulated by greater than 0.2 21 NDUFB9 NADH dehydrogenase fold with a p value <0.01 at any of the 3 (ubiquinone) 1 beta time-points assessed (24, 48 or 72 hours) in subcomplex, 9, 22kDa response to any of the 3 treatments were 22 NDUFA6 NADH dehydrogenase assessed. (ubiquinone) 1 alpha subcomplex, 6, 14kDa 23 DLD dihydrolipoamide Database:KEGG pathway Name:Metabolic dehydrogenase pathways ID:01100 24 NDUFA10 NADH dehydrogenase (ubiquinone) 1 alpha C=1130; O=80; E=11.27; R=7.10; rawP=8.89e-44; subcomplex, 10, 42kDa adjP=1.01e-41 25 NDUFS2 NADH dehydrogenase 1 SRM spermidine synthase (ubiquinone) Fe-S protein 2, 2 GAA glucosidase, alpha; acid 49kDa (NADH-coenzyme Q 3 PAFAH1B platelet-activating factor reductase) 1 acetylhydrolase 1b, regulatory 26 ATP6V1G ATPase, H+ transporting, subunit 1 (45kDa) 1 lysosomal 13kDa, V1 subunit 4 UQCRB ubiquinol-cytochrome c G1 reductase binding protein 27 NDUFA9 NADH dehydrogenase 5 UGDH UDP-glucose 6- (ubiquinone) 1 alpha dehydrogenase subcomplex, 9, 39kDa 6 MCCC1 methylcrotonoyl-CoA 28 EARS2 glutamyl-tRNA synthetase 2, carboxylase 1 (alpha) mitochondrial (putative) 7 ADPGK ADP-dependent glucokinase 29 COQ5 coenzyme Q5 homolog, 8 OAT ornithine aminotransferase methyltransferase (S. 9 LCLAT1 lysocardiolipin acyltransferase cerevisiae) 1 30 HADHA hydroxyacyl-CoA 10 GLUL glutamate-ammonia ligase dehydrogenase/3-ketoacyl- CoA thiolase/enoyl-CoA 11 PISD phosphatidylserine hydratase (trifunctional decarboxylase protein), alpha subunit 12 QPRT quinolinate 31 ATP6V0A1 ATPase, H+ transporting, phosphoribosyltransferase lysosomal V0 subunit a1 13 DBH dopamine beta-hydroxylase 32 GMPPA GDP-mannose (dopamine beta- pyrophosphorylase A monooxygenase) 33 PTGES2 prostaglandin E synthase 2 14 IDI1 isopentenyl-diphosphate delta isomerase 1 34 POLR2C polymerase (RNA) II (DNA 15 ASNS asparagine synthetase directed) polypeptide C, (glutamine-hydrolyzing) 33kDa 16 DUT deoxyuridine triphosphatase 35 ACO2 aconitase 2, mitochondrial

17 CKMT1A creatine kinase, mitochondrial 36 GRHPR glyoxylate 1A reductase/hydroxypyruvate

327 Appendix 2

reductase 60 NME4 NME/NM23 nucleoside 37 P4HA2 prolyl 4-hydroxylase, alpha diphosphate kinase 4 polypeptide II 61 NDUFS8 NADH dehydrogenase 38 DGKI diacylglycerol kinase, iota (ubiquinone) Fe-S protein 8, 23kDa (NADH-coenzyme Q 39 NDUFA11 NADH dehydrogenase reductase) (ubiquinone) 1 alpha 62 ATP6V0D1 ATPase, H+ transporting, subcomplex, 11, 14.7kDa lysosomal 38kDa, V0 subunit 40 P4HA1 prolyl 4-hydroxylase, alpha d1 polypeptide I 63 HSD17B12 hydroxysteroid (17-beta) 41 MAOA monoamine oxidase A dehydrogenase 12 64 MCCC2 methylcrotonoyl-CoA 42 UQCRH ubiquinol-cytochrome c carboxylase 2 (beta) reductase hinge protein 65 GFPT1 glutamine--fructose-6- 43 TH tyrosine hydroxylase phosphate transaminase 1 66 PGM1 phosphoglucomutase 1 44 PDXK pyridoxal (pyridoxine, vitamin B6) kinase 67 HSD17B8 hydroxysteroid (17-beta) 45 NDUFB7 NADH dehydrogenase dehydrogenase 8 (ubiquinone) 1 beta 68 PPAT phosphoribosyl subcomplex, 7, 18kDa pyrophosphate 46 CHPF2 chondroitin polymerizing amidotransferase factor 2 69 LDHA lactate dehydrogenase A 47 PFKP phosphofructokinase, platelet 70 ADSL adenylosuccinate lyase 48 COX6B1 cytochrome c oxidase subunit VIb polypeptide 1 (ubiquitous) 71 ALDH1B1 aldehyde dehydrogenase 1 family, member B1 49 GALK1 galactokinase 1 72 NT5C2 5'-nucleotidase, cytosolic II

50 NDUFS1 NADH dehydrogenase 73 SGSH N-sulfoglucosamine (ubiquinone) Fe-S protein 1, sulfohydrolase 75kDa (NADH-coenzyme Q 74 ATP6V1D ATPase, H+ transporting, reductase) lysosomal 34kDa, V1 subunit 51 ENO2 enolase 2 (gamma, neuronal) D 75 GALNT2 UDP-N-acetyl-alpha-D- 52 RRM1 ribonucleotide reductase M1 galactosamine:polypeptide N- acetylgalactosaminyltransfera 53 ASAH1 N-acylsphingosine se 2 (GalNAc-T2) amidohydrolase (acid 76 HK2 hexokinase 2 ceramidase) 1 54 COX4I1 cytochrome c oxidase subunit 77 GLB1 galactosidase, beta 1 IV isoform 1 55 GNS glucosamine (N-acetyl)-6- 78 ATP6V1C1 ATPase, H+ transporting, sulfatase lysosomal 42kDa, V1 subunit 56 HMGCS1 3-hydroxy-3-methylglutaryl- C1 CoA synthase 1 (soluble) 79 DHCR7 7-dehydrocholesterol 57 NDUFV1 NADH dehydrogenase reductase (ubiquinone) flavoprotein 1, 80 GALE UDP-galactose-4-epimerase 51kDa 58 NDUFB4 NADH dehydrogenase (ubiquinone) 1 beta

subcomplex, 4, 15kDa 59 PFKL phosphofructokinase, liver Database:KEGG pathway Name:Oxidative phosphorylation ID:00190

328 Appendix 2

C=132; O=21; E=1.32; R=15.96; rawP=2.60e-19; 49kDa (NADH-coenzyme Q adjP=1.48e-17 reductase) UserID Gene Name 19 ATP6V0D1 ATPase, H+ transporting, 1 NDUFA9 NADH dehydrogenase lysosomal 38kDa, V0 subunit (ubiquinone) 1 alpha d1 subcomplex, 9, 39kDa 20 COX6B1 cytochrome c oxidase subunit 2 NDUFS1 NADH dehydrogenase VIb polypeptide 1 (ubiquitous) (ubiquinone) Fe-S protein 1, 75kDa (NADH-coenzyme Q 21 ATP6V1G ATPase, H+ transporting, reductase) 1 lysosomal 13kDa, V1 subunit 3 UQCRB ubiquinol-cytochrome c G1 reductase binding protein 4 ATP6V0A1 ATPase, H+ transporting, lysosomal V0 subunit a1 Database:KEGG pathway Name:Parkinson's 5 COX4I1 cytochrome c oxidase subunit disease ID:05012 IV isoform 1 C=130; O=18; E=1.30; R=13.89; rawP=1.23e-15; 6 ATP6V1D ATPase, H+ transporting, adjP=4.67e-14 lysosomal 34kDa, V1 subunit UserID Gene Name D 1 NDUFA9 NADH dehydrogenase 7 NDUFV1 NADH dehydrogenase (ubiquinone) 1 alpha (ubiquinone) flavoprotein 1, subcomplex, 9, 39kDa 51kDa 2 CASP3 caspase 3, apoptosis-related 8 NDUFB4 NADH dehydrogenase cysteine peptidase (ubiquinone) 1 beta 3 NDUFS1 NADH dehydrogenase subcomplex, 4, 15kDa (ubiquinone) Fe-S protein 1, 9 NDUFA11 NADH dehydrogenase 75kDa (NADH-coenzyme Q (ubiquinone) 1 alpha reductase) subcomplex, 11, 14.7kDa 4 UQCRB ubiquinol-cytochrome c 10 NDUFB8 NADH dehydrogenase reductase binding protein (ubiquinone) 1 beta 5 COX4I1 cytochrome c oxidase subunit subcomplex, 8, 19kDa IV isoform 1 11 NDUFS8 NADH dehydrogenase 6 NDUFV1 NADH dehydrogenase (ubiquinone) Fe-S protein 8, (ubiquinone) flavoprotein 1, 23kDa (NADH-coenzyme Q 51kDa reductase) 7 NDUFB4 NADH dehydrogenase 12 UQCRH ubiquinol-cytochrome c (ubiquinone) 1 beta reductase hinge protein subcomplex, 4, 15kDa 13 NDUFB9 NADH dehydrogenase 8 NDUFA11 NADH dehydrogenase (ubiquinone) 1 beta (ubiquinone) 1 alpha subcomplex, 9, 22kDa subcomplex, 11, 14.7kDa 14 ATP6V1C1 ATPase, H+ transporting, 9 NDUFB8 NADH dehydrogenase lysosomal 42kDa, V1 subunit (ubiquinone) 1 beta C1 subcomplex, 8, 19kDa 15 NDUFA6 NADH dehydrogenase 10 NDUFS8 NADH dehydrogenase (ubiquinone) 1 alpha (ubiquinone) Fe-S protein 8, subcomplex, 6, 14kDa 23kDa (NADH-coenzyme Q 16 NDUFA10 NADH dehydrogenase reductase) (ubiquinone) 1 alpha 11 UQCRH ubiquinol-cytochrome c subcomplex, 10, 42kDa reductase hinge protein 17 NDUFB7 NADH dehydrogenase 12 NDUFB9 NADH dehydrogenase (ubiquinone) 1 beta (ubiquinone) 1 beta subcomplex, 7, 18kDa subcomplex, 9, 22kDa 18 NDUFS2 NADH dehydrogenase 13 TH tyrosine hydroxylase (ubiquinone) Fe-S protein 2,

329 Appendix 2

14 NDUFA6 NADH dehydrogenase 12 NDUFB8 NADH dehydrogenase (ubiquinone) 1 alpha (ubiquinone) 1 beta subcomplex, 6, 14kDa subcomplex, 8, 19kDa 15 NDUFA10 NADH dehydrogenase 13 NDUFS8 NADH dehydrogenase (ubiquinone) 1 alpha (ubiquinone) Fe-S protein 8, subcomplex, 10, 42kDa 23kDa (NADH-coenzyme Q 16 NDUFB7 NADH dehydrogenase reductase) (ubiquinone) 1 beta 14 UQCRH ubiquinol-cytochrome c subcomplex, 7, 18kDa reductase hinge protein 17 NDUFS2 NADH dehydrogenase 15 NDUFB9 NADH dehydrogenase (ubiquinone) Fe-S protein 2, (ubiquinone) 1 beta 49kDa (NADH-coenzyme Q subcomplex, 9, 22kDa reductase) 16 NDUFA6 NADH dehydrogenase 18 COX6B1 cytochrome c oxidase subunit (ubiquinone) 1 alpha VIb polypeptide 1 (ubiquitous) subcomplex, 6, 14kDa 17 NDUFA10 NADH dehydrogenase

(ubiquinone) 1 alpha subcomplex, 10, 42kDa Database:KEGG 18 NDUFB7 NADH dehydrogenase pathway Name:Huntington's (ubiquinone) 1 beta disease ID:05016 subcomplex, 7, 18kDa C=183; O=20; E=1.82; R=10.96; rawP=3.41e-15; 19 NDUFS2 NADH dehydrogenase adjP=9.72e-14 (ubiquinone) Fe-S protein 2, UserID Gene Name 49kDa (NADH-coenzyme Q 1 NDUFA9 NADH dehydrogenase reductase) (ubiquinone) 1 alpha 20 COX6B1 cytochrome c oxidase subunit subcomplex, 9, 39kDa VIb polypeptide 1 (ubiquitous) 2 CASP3 caspase 3, apoptosis-related cysteine peptidase Database:KEGG pathway Name:Alzheimer's 3 NDUFS1 NADH dehydrogenase disease ID:05010 (ubiquinone) Fe-S protein 1, C=167; O=19; E=1.67; R=11.41; rawP=7.98e-15; 75kDa (NADH-coenzyme Q adjP=1.82e-13 reductase) UserID Gene Name 4 AP2M1 adaptor-related protein complex 2, mu 1 subunit 1 NDUFA9 NADH dehydrogenase 5 UQCRB ubiquinol-cytochrome c (ubiquinone) 1 alpha reductase binding protein subcomplex, 9, 39kDa 6 COX4I1 cytochrome c oxidase subunit 2 CASP3 caspase 3, apoptosis-related IV isoform 1 cysteine peptidase 7 POLR2C polymerase (RNA) II (DNA 3 NDUFS1 NADH dehydrogenase directed) polypeptide C, (ubiquinone) Fe-S protein 1, 33kDa 75kDa (NADH-coenzyme Q 8 NDUFV1 NADH dehydrogenase reductase) (ubiquinone) flavoprotein 1, 4 UQCRB ubiquinol-cytochrome c 51kDa reductase binding protein 9 NDUFB4 NADH dehydrogenase 5 COX4I1 cytochrome c oxidase subunit (ubiquinone) 1 beta IV isoform 1 subcomplex, 4, 15kDa 6 BID BH3 interacting domain death 10 HTT huntingtin agonist 7 NDUFV1 NADH dehydrogenase 11 NDUFA11 NADH dehydrogenase (ubiquinone) flavoprotein 1, (ubiquinone) 1 alpha 51kDa subcomplex, 11, 14.7kDa 8 NDUFB4 NADH dehydrogenase (ubiquinone) 1 beta subcomplex, 4, 15kDa

330 Appendix 2

9 NDUFA11 NADH dehydrogenase C=165; O=13; E=1.65; R=7.90; rawP=1.36e-08; (ubiquinone) 1 alpha adjP=2.21e-07 subcomplex, 11, 14.7kDa UserID Gene Name 10 NDUFB8 NADH dehydrogenase 1 BAK1 BCL2-antagonist/killer 1 (ubiquinone) 1 beta 2 HSPH1 heat shock 105kDa/110kDa subcomplex, 8, 19kDa protein 1 11 NDUFS8 NADH dehydrogenase 3 HSP90B1 heat shock protein 90kDa (ubiquinone) Fe-S protein 8, beta (Grp94), member 1 23kDa (NADH-coenzyme Q 4 HSPBP1 HSPA (heat shock 70kDa) reductase) binding protein, cytoplasmic 12 UQCRH ubiquinol-cytochrome c cochaperone 1 reductase hinge protein 5 SAR1A SAR1 homolog A (S. 13 NDUFB9 NADH dehydrogenase cerevisiae) (ubiquinone) 1 beta 6 P4HB prolyl 4-hydroxylase, beta subcomplex, 9, 22kDa polypeptide 14 NDUFA6 NADH dehydrogenase 7 DNAJC3 DnaJ (Hsp40) homolog, (ubiquinone) 1 alpha subfamily C, member 3 subcomplex, 6, 14kDa 8 ERO1L ERO1-like (S. cerevisiae) 15 NDUFA10 NADH dehydrogenase (ubiquinone) 1 alpha 9 HSPA5 heat shock 70kDa protein 5 subcomplex, 10, 42kDa (glucose-regulated protein, 16 NDUFB7 NADH dehydrogenase 78kDa) (ubiquinone) 1 beta 10 PDIA3 protein disulfide isomerase subcomplex, 7, 18kDa family A, member 3 17 NDUFS2 NADH dehydrogenase 11 BAG2 BCL2-associated athanogene (ubiquinone) Fe-S protein 2, 2 49kDa (NADH-coenzyme Q 12 PDIA6 protein disulfide isomerase reductase) family A, member 6 18 COX6B1 cytochrome c oxidase subunit 13 PDIA4 protein disulfide isomerase VIb polypeptide 1 (ubiquitous) family A, member 4

19 CHP1 calcineurin-like EF hand Database:KEGG pathway Name:Glycolysis / protein 1 Gluconeogenesis ID:00010

C=65; O=9; E=0.65; R=13.89; rawP=1.75e-08; Database:KEGG pathway Name:Galactose adjP=2.49e-07 metabolism ID:00052 UserID Gene Name C=27; O=8; E=0.27; R=29.72; rawP=1.72e-10; 1 PGM1 phosphoglucomutase 1 adjP=3.27e-09 UserID Gene Name 2 ENO2 enolase 2 (gamma, neuronal) 1 PGM1 phosphoglucomutase 1 3 PFKL phosphofructokinase, liver 2 GALK1 galactokinase 1 4 LDHA lactate dehydrogenase A 3 GAA glucosidase, alpha; acid 5 HK2 hexokinase 2 4 PFKL phosphofructokinase, liver 6 ADPGK ADP-dependent glucokinase 5 HK2 hexokinase 2 7 ALDH1B1 aldehyde dehydrogenase 1 family, member B1 6 GLB1 galactosidase, beta 1 8 DLD dihydrolipoamide 7 PFKP phosphofructokinase, platelet dehydrogenase 8 GALE UDP-galactose-4-epimerase 9 PFKP phosphofructokinase, platelet

Database:KEGG pathway Name:Amino sugar Database:KEGG pathway Name:Protein and nucleotide sugar metabolism ID:00520 processing in endoplasmic C=48; O=8; E=0.48; R=16.72; rawP=2.44e-08; reticulum ID:04141 adjP=3.09e-07

331 Appendix 2

UserID Gene Name 6 HADHB hydroxyacyl-CoA 1 GFPT1 glutamine--fructose-6- dehydrogenase/3-ketoacyl- phosphate transaminase 1 CoA thiolase/enoyl-CoA 2 PGM1 phosphoglucomutase 1 hydratase (trifunctional protein), beta subunit 3 GALK1 galactokinase 1 7 MCCC1 methylcrotonoyl-CoA 4 UGDH UDP-glucose 6- carboxylase 1 (alpha) dehydrogenase 5 HK2 hexokinase 2 Database:KEGG 6 RENBP renin binding protein pathway Name:Lysosome ID:04142 7 GMPPA GDP-mannose C=121; O=10; E=1.21; R=8.29; rawP=4.14e-07; pyrophosphorylase A adjP=3.93e-06 8 GALE UDP-galactose-4-epimerase UserID Gene Name 1 TPP1 tripeptidyl peptidase I Database:KEGG pathway Name:Arginine and 2 GNS glucosamine (N-acetyl)-6- proline metabolism ID:00330 sulfatase C=54; O=8; E=0.54; R=14.86; rawP=6.38e-08; 3 GAA glucosidase, alpha; acid adjP=7.27e-07 4 SGSH N-sulfoglucosamine UserID Gene Name sulfohydrolase 1 SRM spermidine synthase 5 CD63 CD63 molecule 2 P4HA2 prolyl 4-hydroxylase, alpha 6 ASAH1 N-acylsphingosine polypeptide II amidohydrolase (acid 3 CKMT1A creatine kinase, mitochondrial ceramidase) 1 1A 7 ATP6V0A1 ATPase, H+ transporting, 4 MAOA monoamine oxidase A lysosomal V0 subunit a1 5 P4HA1 prolyl 4-hydroxylase, alpha 8 GLB1 galactosidase, beta 1 polypeptide I 9 NPC2 Niemann-Pick disease, type 6 OAT ornithine aminotransferase C2 7 ALDH1B1 aldehyde dehydrogenase 1 10 ATP6V0D1 ATPase, H+ transporting, family, member B1 lysosomal 38kDa, V0 subunit 8 GLUL glutamate-ammonia ligase d1

Database:KEGG pathway Name:Valine, leucine and isoleucine degradation ID:00280 Table A 3 KEGG pathway analysis of C=44; O=7; E=0.44; R=15.96; rawP=2.60e-07; proteins dysregulated by treatment of SH- adjP=2.69e-06 SY5Y expressing moderately elevated levels UserID Gene Name of αSynuclein and treatment with 10 nM 1 HMGCS1 3-hydroxy-3-methylglutaryl- rotenone. Protein levels were changed by CoA synthase 1 (soluble) greater than 0.2 fold with a p value <0.01 2 ALDH1B1 aldehyde dehydrogenase 1 family, member B1 at any of the 3 time-points assessed (24, 48 3 DLD dihydrolipoamide or 72 hours). dehydrogenase 4 HADHA hydroxyacyl-CoA dehydrogenase/3-ketoacyl- Database:KEGG pathway Name:Metabolic CoA thiolase/enoyl-CoA pathways ID:01100 hydratase (trifunctional C=1130; O=49; E=5.71; R=8.58; rawP=3.04e-31; protein), alpha subunit adjP=1.98e-29 5 MCCC2 methylcrotonoyl-CoA UserID Gene Name carboxylase 2 (beta) 1 GALK1 galactokinase 1

332 Appendix 2

2 NDUFS1 NADH dehydrogenase 24 NDUFS2 NADH dehydrogenase (ubiquinone) Fe-S protein 1, (ubiquinone) Fe-S protein 2, 75kDa (NADH-coenzyme Q 49kDa (NADH-coenzyme Q reductase) reductase) 3 ENO2 enolase 2 (gamma, 25 HSD17B12 hydroxysteroid (17-beta) neuronal) dehydrogenase 12 4 UQCRB ubiquinol-cytochrome c 26 NDUFA9 NADH dehydrogenase reductase binding protein (ubiquinone) 1 alpha 5 PAFAH1B1 platelet-activating factor subcomplex, 9, 39kDa acetylhydrolase 1b, 27 PGM1 phosphoglucomutase 1 regulatory subunit 1 (45kDa) 28 HSD17B8 hydroxysteroid (17-beta) 6 UGDH UDP-glucose 6- dehydrogenase 8 dehydrogenase 29 EARS2 glutamyl-tRNA synthetase 2, 7 RRM1 ribonucleotide reductase mitochondrial (putative) M1 8 MCCC1 methylcrotonoyl-CoA 30 LDHA lactate dehydrogenase A carboxylase 1 (alpha) 31 ATP6V0A1 ATPase, H+ transporting, 9 COX4I1 cytochrome c oxidase lysosomal V0 subunit a1 subunit IV isoform 1 32 ALDH1B1 aldehyde dehydrogenase 1 10 ADPGK ADP-dependent glucokinase family, member B1 11 LCLAT1 lysocardiolipin 33 GMPPA GDP-mannose acyltransferase 1 pyrophosphorylase A 12 GLUL glutamate-ammonia ligase 34 PTGES2 prostaglandin E synthase 2 13 DBH dopamine beta-hydroxylase 35 NT5C2 5'-nucleotidase, cytosolic II (dopamine beta- 36 ACO2 aconitase 2, mitochondrial monooxygenase) 37 GALNT2 UDP-N-acetyl-alpha-D- 14 HMGCS1 3-hydroxy-3-methylglutaryl- galactosamine:polypeptide CoA synthase 1 (soluble) N- 15 PFKL phosphofructokinase, liver acetylgalactosaminyltransfe rase 2 (GalNAc-T2) 16 NDUFB8 NADH dehydrogenase 38 ATP6V1D ATPase, H+ transporting, (ubiquinone) 1 beta lysosomal 34kDa, V1 subunit subcomplex, 8, 19kDa D 17 NME4 NME/NM23 nucleoside 39 P4HA2 prolyl 4-hydroxylase, alpha diphosphate kinase 4 polypeptide II 18 NDUFS8 NADH dehydrogenase 40 DGKI diacylglycerol kinase, iota (ubiquinone) Fe-S protein 8, 23kDa (NADH-coenzyme Q 41 HK2 hexokinase 2 reductase) 42 P4HA1 prolyl 4-hydroxylase, alpha 19 NDUFB9 NADH dehydrogenase polypeptide I (ubiquinone) 1 beta 43 GLB1 galactosidase, beta 1 subcomplex, 9, 22kDa 44 UQCRH ubiquinol-cytochrome c 20 NDUFA6 NADH dehydrogenase reductase hinge protein (ubiquinone) 1 alpha 45 DHCR7 7-dehydrocholesterol subcomplex, 6, 14kDa reductase 21 DLD dihydrolipoamide 46 NDUFB7 NADH dehydrogenase dehydrogenase (ubiquinone) 1 beta 22 NDUFA10 NADH dehydrogenase subcomplex, 7, 18kDa (ubiquinone) 1 alpha 47 CHPF2 chondroitin polymerizing subcomplex, 10, 42kDa factor 2 23 ATP6V0D1 ATPase, H+ transporting, 48 PFKP phosphofructokinase, lysosomal 38kDa, V0 subunit platelet d1 49 COX6B1 cytochrome c oxidase subunit VIb polypeptide 1

333 Appendix 2

(ubiquitous) 16 COX6B1 cytochrome c oxidase subunit VIb polypeptide 1 (ubiquitous)

Database:KEGG pathway Name:Oxidative Database:KEGG pathway Name:Alzheimer's phosphorylation ID:00190 disease ID:05010 C=132; O=16; E=0.67; R=23.98; rawP=9.78e-18; C=167; O=15; E=0.84; R=17.77; rawP=9.85e-15; adjP=3.18e-16 adjP=2.13e-13 UserID Gene Name UserID Gene Name 1 NDUFA9 NADH dehydrogenase 1 NDUFA9 NADH dehydrogenase (ubiquinone) 1 alpha (ubiquinone) 1 alpha subcomplex, 9, 39kDa subcomplex, 9, 39kDa 2 NDUFS1 NADH dehydrogenase 2 NDUFS1 NADH dehydrogenase (ubiquinone) Fe-S protein 1, (ubiquinone) Fe-S protein 1, 75kDa (NADH-coenzyme Q 75kDa (NADH-coenzyme Q reductase) reductase) 3 UQCRB ubiquinol-cytochrome c 3 UQCRB ubiquinol-cytochrome c reductase binding protein reductase binding protein 4 ATP6V0A1 ATPase, H+ transporting, 4 NDUFB8 NADH dehydrogenase lysosomal V0 subunit a1 (ubiquinone) 1 beta 5 COX4I1 cytochrome c oxidase subcomplex, 8, 19kDa subunit IV isoform 1 5 NDUFS8 NADH dehydrogenase 6 ATP6V1D ATPase, H+ transporting, (ubiquinone) Fe-S protein 8, lysosomal 34kDa, V1 subunit 23kDa (NADH-coenzyme Q D reductase) 7 NDUFB8 NADH dehydrogenase 6 COX4I1 cytochrome c oxidase (ubiquinone) 1 beta subunit IV isoform 1 subcomplex, 8, 19kDa 7 UQCRH ubiquinol-cytochrome c 8 NDUFB9 NADH dehydrogenase reductase hinge protein (ubiquinone) 1 beta 8 NDUFB9 NADH dehydrogenase subcomplex, 9, 22kDa (ubiquinone) 1 beta 9 UQCRH ubiquinol-cytochrome c subcomplex, 9, 22kDa reductase hinge protein 9 NDUFA6 NADH dehydrogenase 10 NDUFS8 NADH dehydrogenase (ubiquinone) 1 alpha (ubiquinone) Fe-S protein 8, subcomplex, 6, 14kDa 23kDa (NADH-coenzyme Q 10 NDUFB7 NADH dehydrogenase reductase) (ubiquinone) 1 beta 11 NDUFA6 NADH dehydrogenase subcomplex, 7, 18kDa (ubiquinone) 1 alpha 11 BID BH3 interacting domain subcomplex, 6, 14kDa death agonist 12 NDUFA10 NADH dehydrogenase 12 NDUFA10 NADH dehydrogenase (ubiquinone) 1 alpha (ubiquinone) 1 alpha subcomplex, 10, 42kDa subcomplex, 10, 42kDa 13 NDUFB7 NADH dehydrogenase 13 NDUFS2 NADH dehydrogenase (ubiquinone) 1 beta (ubiquinone) Fe-S protein 2, subcomplex, 7, 18kDa 49kDa (NADH-coenzyme Q 14 NDUFS2 NADH dehydrogenase reductase) (ubiquinone) Fe-S protein 2, 14 COX6B1 cytochrome c oxidase 49kDa (NADH-coenzyme Q subunit VIb polypeptide 1 reductase) (ubiquitous) 15 ATP6V0D1 ATPase, H+ transporting, 15 CHP1 calcineurin-like EF hand lysosomal 38kDa, V0 subunit protein 1 d1

Database:KEGG pathway Name:Parkinson's

334 Appendix 2

disease ID:05012 2 NDUFS1 NADH dehydrogenase C=130; O=13; E=0.66; R=19.78; rawP=1.53e-13; (ubiquinone) Fe-S protein 1, adjP=2.49e-12 75kDa (NADH-coenzyme Q UserID Gene Name reductase) 3 HTT huntingtin 1 NDUFA9 NADH dehydrogenase (ubiquinone) 1 alpha 4 UQCRB ubiquinol-cytochrome c subcomplex, 9, 39kDa reductase binding protein 2 NDUFS1 NADH dehydrogenase 5 NDUFB8 NADH dehydrogenase (ubiquinone) Fe-S protein 1, (ubiquinone) 1 beta 75kDa (NADH-coenzyme Q subcomplex, 8, 19kDa reductase) 6 NDUFS8 NADH dehydrogenase 3 UQCRB ubiquinol-cytochrome c (ubiquinone) Fe-S protein 8, reductase binding protein 23kDa (NADH-coenzyme Q 4 NDUFB8 NADH dehydrogenase reductase) (ubiquinone) 1 beta 7 COX4I1 cytochrome c oxidase subcomplex, 8, 19kDa subunit IV isoform 1 5 NDUFS8 NADH dehydrogenase 8 NDUFB9 NADH dehydrogenase (ubiquinone) Fe-S protein 8, (ubiquinone) 1 beta 23kDa (NADH-coenzyme Q subcomplex, 9, 22kDa reductase) 9 UQCRH ubiquinol-cytochrome c 6 COX4I1 cytochrome c oxidase reductase hinge protein subunit IV isoform 1 10 NDUFA6 NADH dehydrogenase 7 NDUFB9 NADH dehydrogenase (ubiquinone) 1 alpha (ubiquinone) 1 beta subcomplex, 6, 14kDa subcomplex, 9, 22kDa 11 NDUFB7 NADH dehydrogenase 8 UQCRH ubiquinol-cytochrome c (ubiquinone) 1 beta reductase hinge protein subcomplex, 7, 18kDa 9 NDUFA6 NADH dehydrogenase 12 NDUFA10 NADH dehydrogenase (ubiquinone) 1 alpha (ubiquinone) 1 alpha subcomplex, 6, 14kDa subcomplex, 10, 42kDa 10 NDUFB7 NADH dehydrogenase 13 NDUFS2 NADH dehydrogenase (ubiquinone) 1 beta (ubiquinone) Fe-S protein 2, subcomplex, 7, 18kDa 49kDa (NADH-coenzyme Q 11 NDUFA10 NADH dehydrogenase reductase) (ubiquinone) 1 alpha 14 COX6B1 cytochrome c oxidase subcomplex, 10, 42kDa subunit VIb polypeptide 1 12 NDUFS2 NADH dehydrogenase (ubiquitous) (ubiquinone) Fe-S protein 2, 49kDa (NADH-coenzyme Q Database:KEGG pathway Name:Glycolysis / reductase) Gluconeogenesis ID:00010 13 COX6B1 cytochrome c oxidase C=65; O=9; E=0.33; R=27.39; rawP=4.57e-11; subunit VIb polypeptide 1 adjP=4.95e-10 (ubiquitous) UserID Gene Name 1 PGM1 phosphoglucomutase 1 Database:KEGG 2 ENO2 enolase 2 (gamma, pathway Name:Huntington's neuronal) disease ID:05016 3 PFKL phosphofructokinase, liver C=183; O=14; E=0.93; R=15.13; rawP=7.16e-13; 4 HK2 hexokinase 2 adjP=9.31e-12 UserID Gene Name 5 LDHA lactate dehydrogenase A 1 NDUFA9 NADH dehydrogenase 6 ADPGK ADP-dependent glucokinase (ubiquinone) 1 alpha 7 ALDH1B1 aldehyde dehydrogenase 1 subcomplex, 9, 39kDa family, member B1

335 Appendix 2

8 DLD dihydrolipoamide 1 PGM1 phosphoglucomutase 1 dehydrogenase 2 GALK1 galactokinase 1 9 PFKP phosphofructokinase, 3 GMPPA GDP-mannose platelet pyrophosphorylase A

4 UGDH UDP-glucose 6- Database:KEGG pathway Name:Galactose dehydrogenase metabolism ID:00052 5 HK2 hexokinase 2 C=27; O=6; E=0.14; R=43.96; rawP=4.22e-09; 6 RENBP renin binding protein adjP=3.92e-08 UserID Gene Name 1 GLB1 galactosidase, beta 1 Database:KEGG pathway Name:Valine, leucine and isoleucine degradation ID:00280 2 PGM1 phosphoglucomutase 1 3 GALK1 galactokinase 1 C=44; O=4; E=0.22; R=17.98; rawP=7.36e-05; adjP=0.0003 4 PFKL phosphofructokinase, liver UserID Gene Name 5 PFKP phosphofructokinase, 1 HMGCS1 3-hydroxy-3-methylglutaryl- platelet CoA synthase 1 (soluble) 6 HK2 hexokinase 2 2 ALDH1B1 aldehyde dehydrogenase 1 family, member B1 Database:KEGG pathway Name:Protein 3 DLD dihydrolipoamide processing in endoplasmic dehydrogenase reticulum ID:04141 4 MCCC1 methylcrotonoyl-CoA C=165; O=10; E=0.83; R=11.99; rawP=1.40e-08; carboxylase 1 (alpha) adjP=1.14e-07

UserID Gene Name Database:KEGG pathway Name:Arginine and 1 HSPBP1 HSPA (heat shock 70kDa) proline metabolism ID:00330 binding protein, cytoplasmic C=54; O=4; E=0.27; R=14.65; rawP=0.0002; cochaperone 1 adjP=0.0008 2 HSP90B1 heat shock protein 90kDa UserID Gene Name beta (Grp94), member 1 3 SAR1A SAR1 homolog A (S. 1 P4HA1 prolyl 4-hydroxylase, alpha cerevisiae) polypeptide I 4 P4HB prolyl 4-hydroxylase, beta 2 ALDH1B1 aldehyde dehydrogenase 1 polypeptide family, member B1 5 DNAJC3 DnaJ (Hsp40) homolog, 3 P4HA2 prolyl 4-hydroxylase, alpha subfamily C, member 3 polypeptide II 6 ERO1L ERO1-like (S. cerevisiae) 4 GLUL glutamate-ammonia ligase 7 HSPA5 heat shock 70kDa protein 5 (glucose-regulated protein, Database:KEGG 78kDa) pathway Name:Lysosome ID:04142 8 PDIA3 protein disulfide isomerase C=121; O=5; E=0.61; R=8.17; rawP=0.0004; family A, member 3 adjP=0.0013 9 PDIA6 protein disulfide isomerase UserID Gene Name family A, member 6 1 ATP6V0A1 ATPase, H+ transporting, 10 PDIA4 protein disulfide isomerase lysosomal V0 subunit a1 family A, member 4 2 TPP1 tripeptidyl peptidase I 3 GLB1 galactosidase, beta 1 Database:KEGG pathway Name:Amino sugar 4 CD63 CD63 molecule and nucleotide sugar metabolism ID:00520 5 ATP6V0D1 ATPase, H+ transporting, C=48; O=6; E=0.24; R=24.73; rawP=1.60e-07; lysosomal 38kDa, V0 subunit adjP=1.16e-06 d1 UserID Gene Name

336 Appendix 2

Table A 4 KEGG pathway analysis of 17 SUCLG2 succinate-CoA ligase, GDP- proteins dysregulated by treatment od SH- forming, beta subunit SY5Y cells with 10 nM rotenone by greater 18 TH tyrosine hydroxylase than 0.2 fold with a p value <0.01 at any of 19 MCCC2 methylcrotonoyl-CoA the 3 time-points assessed (24, 48 or 72 carboxylase 2 (beta) hours). 20 ATP6V1G ATPase, H+ transporting, 1 lysosomal 13kDa, V1 subunit G1 Database:KEGG pathway Name:Metabolic pathways ID:01100 Database:KEGG pathway Name:Valine, C=1130; O=20; E=2.88; R=6.94; rawP=9.31e-12; leucine and isoleucine degradation ID:00280 adjP=2.61e-10 C=44; O=3; E=0.11; R=26.73; rawP=0.0002; UserID Gene Name adjP=0.0014 1 NDUFA9 NADH dehydrogenase UserID Gene Name (ubiquinone) 1 alpha 1 HADHA hydroxyacyl-CoA subcomplex, 9, 39kDa dehydrogenase/3-ketoacyl- 2 GAA glucosidase, alpha; acid CoA thiolase/enoyl-CoA 3 COQ5 coenzyme Q5 homolog, hydratase (trifunctional methyltransferase (S. protein), alpha subunit cerevisiae) 2 MCCC2 methylcrotonoyl-CoA 4 HADHA hydroxyacyl-CoA carboxylase 2 (beta) dehydrogenase/3-ketoacyl- 3 HADHB hydroxyacyl-CoA CoA thiolase/enoyl-CoA dehydrogenase/3-ketoacyl- hydratase (trifunctional CoA thiolase/enoyl-CoA protein), alpha subunit hydratase (trifunctional 5 RRM1 ribonucleotide reductase M1 protein), beta subunit 6 ASAH1 N-acylsphingosine amidohydrolase (acid Database:KEGG ceramidase) 1 pathway Name:Lysosome ID:04142 7 OAT ornithine aminotransferase C=121; O=4; E=0.31; R=12.96; rawP=0.0003; 8 PTGES2 prostaglandin E synthase 2 adjP=0.0014 9 GNS glucosamine (N-acetyl)-6- UserID Gene Name sulfatase 1 GNS glucosamine (N-acetyl)-6- 10 POLR2C polymerase (RNA) II (DNA sulfatase directed) polypeptide C, 2 ASAH1 N-acylsphingosine 33kDa amidohydrolase (acid 11 GRHPR glyoxylate ceramidase) 1 reductase/hydroxypyruvate 3 GAA glucosidase, alpha; acid reductase 4 SGSH N-sulfoglucosamine 12 SGSH N-sulfoglucosamine sulfohydrolase sulfohydrolase 13 IDI1 isopentenyl-diphosphate delta isomerase 1 Database:KEGG pathway Name:Parkinson's 14 DUT deoxyuridine triphosphatase disease ID:05012 15 NDUFA1 NADH dehydrogenase C=130; O=4; E=0.33; R=12.06; rawP=0.0004; 1 (ubiquinone) 1 alpha adjP=0.0016 subcomplex, 11, 14.7kDa UserID Gene Name 16 HADHB hydroxyacyl-CoA 1 NDUFA9 NADH dehydrogenase dehydrogenase/3-ketoacyl- (ubiquinone) 1 alpha CoA thiolase/enoyl-CoA subcomplex, 9, 39kDa hydratase (trifunctional protein), beta subunit

337 Appendix 2

2 CASP3 caspase 3, apoptosis-related 3 NDUFA1 NADH dehydrogenase cysteine peptidase 1 (ubiquinone) 1 alpha subcomplex, 11, 14.7kDa 3 TH tyrosine hydroxylase

4 NDUFA1 NADH dehydrogenase Database:KEGG pathway Name:Protein 1 (ubiquinone) 1 alpha processing in endoplasmic subcomplex, 11, 14.7kDa reticulum ID:04141

C=165; O=2; E=0.42; R=4.75; rawP=0.0667; Database:KEGG adjP=0.0692 pathway Name:Huntington's UserID Gene Name disease ID:05016 1 HSPH1 heat shock 105kDa/110kDa C=183; O=4; E=0.47; R=8.57; rawP=0.0013; protein 1 adjP=0.0040 2 ERO1L ERO1-like (S. cerevisiae) UserID Gene Name 1 POLR2C polymerase (RNA) II (DNA directed) polypeptide C, 33kDa 2 NDUFA9 NADH dehydrogenase Table A 5 KEGG pathway analysis of (ubiquinone) 1 alpha proteins dysregulated by treatment od SH- subcomplex, 9, 39kDa SY5Y expressing moderately elevated levels 3 CASP3 caspase 3, apoptosis-related of αSynuclein by greater than 0.2 fold with cysteine peptidase a p value <0.01 at any of the 3 time-points 4 NDUFA1 NADH dehydrogenase assessed (24, 48 or 72 hours). 1 (ubiquinone) 1 alpha subcomplex, 11, 14.7kDa

Database:KEGG pathway Name:Metabolic Database:KEGG pathway Name:Oxidative pathways ID:01100 phosphorylation ID:00190 C=1130; O=19; E=3.20; R=5.94; rawP=4.91e-10; C=132; O=3; E=0.34; R=8.91; rawP=0.0048; adjP=1.37e-08 adjP=0.0112 UserID Gene Name UserID Gene Name 1 GFPT1 glutamine--fructose-6- 1 NDUFA9 NADH dehydrogenase phosphate transaminase 1 (ubiquinone) 1 alpha subcomplex, 9, 39kDa 2 SRM spermidine synthase 2 NDUFA1 NADH dehydrogenase 3 PPAT phosphoribosyl pyrophosphate 1 (ubiquinone) 1 alpha amidotransferase subcomplex, 11, 14.7kDa 3 ATP6V1G ATPase, H+ transporting, 4 HADHA hydroxyacyl-CoA 1 lysosomal 13kDa, V1 subunit dehydrogenase/3-ketoacyl-CoA G1 thiolase/enoyl-CoA hydratase (trifunctional protein), alpha subunit Database:KEGG pathway Name:Alzheimer's 5 ADSL adenylosuccinate lyase disease ID:05010 C=167; O=3; E=0.43; R=7.04; rawP=0.0091; 6 GMPPA GDP-mannose adjP=0.0159 pyrophosphorylase A UserID Gene Name 7 PISD phosphatidylserine decarboxylase 1 NDUFA9 NADH dehydrogenase 8 QPRT quinolinate (ubiquinone) 1 alpha phosphoribosyltransferase subcomplex, 9, 39kDa 9 ASNS asparagine synthetase 2 CASP3 caspase 3, apoptosis-related (glutamine-hydrolyzing) cysteine peptidase 10 NDUFV1 NADH dehydrogenase (ubiquinone) flavoprotein 1,

338 Appendix 2

51kDa 2 NDUFV1 NADH dehydrogenase (ubiquinone) flavoprotein 1, 51kDa 11 NDUFB4 NADH dehydrogenase 3 AP2M1 adaptor-related protein complex (ubiquinone) 1 beta 2, mu 1 subunit subcomplex, 4, 15kDa 12 CKMT1A creatine kinase, mitochondrial 4 NDUFB4 NADH dehydrogenase 1A (ubiquinone) 1 beta 13 HADHB hydroxyacyl-CoA subcomplex, 4, 15kDa dehydrogenase/3-ketoacyl-CoA 5 COX6B1 cytochrome c oxidase subunit thiolase/enoyl-CoA hydratase VIb polypeptide 1 (ubiquitous) (trifunctional protein), beta

subunit 14 MAOA monoamine oxidase A Database:KEGG pathway Name:Amino sugar and nucleotide sugar metabolism ID:00520 15 NDUFB9 NADH dehydrogenase (ubiquinone) 1 beta C=48; O=3; E=0.14; R=22.09; rawP=0.0003; subcomplex, 9, 22kDa adjP=0.0014 16 PDXK pyridoxal (pyridoxine, vitamin UserID Gene Name B6) kinase 1 GFPT1 glutamine--fructose-6- 17 ATP6V1 ATPase, H+ transporting, phosphate transaminase 1 C1 lysosomal 42kDa, V1 subunit C1 2 GMPPA GDP-mannose 18 COX6B1 cytochrome c oxidase subunit pyrophosphorylase A VIb polypeptide 1 (ubiquitous) 3 GALE UDP-galactose-4-epimerase 19 GALE UDP-galactose-4-epimerase Database:KEGG pathway Name:Arginine and Database:KEGG pathway Name:Oxidative proline metabolism ID:00330 phosphorylation ID:00190 C=54; O=3; E=0.15; R=19.64; rawP=0.0005; C=132; O=5; E=0.37; R=13.39; rawP=3.87e-05; adjP=0.0016 adjP=0.0004 UserID Gene Name UserID Gene Name 1 MAOA monoamine oxidase A 1 NDUFB9 NADH dehydrogenase 2 SRM spermidine synthase (ubiquinone) 1 beta 3 CKMT1A creatine kinase, mitochondrial subcomplex, 9, 22kDa 1A 2 NDUFV1 NADH dehydrogenase (ubiquinone) flavoprotein 1, Database:KEGG pathway Name:Parkinson's 51kDa disease ID:05012 3 ATP6V1 ATPase, H+ transporting, C=130; O=4; E=0.37; R=10.88; rawP=0.0005; C1 lysosomal 42kDa, V1 subunit C1 adjP=0.0016 4 NDUFB4 NADH dehydrogenase UserID Gene Name (ubiquinone) 1 beta 1 NDUFB9 NADH dehydrogenase subcomplex, 4, 15kDa (ubiquinone) 1 beta 5 COX6B1 cytochrome c oxidase subunit subcomplex, 9, 22kDa VIb polypeptide 1 (ubiquitous) 2 NDUFV1 NADH dehydrogenase (ubiquinone) flavoprotein 1, 51kDa Database:KEGG pathway Name:Huntington's 3 NDUFB4 NADH dehydrogenase disease ID:05016 (ubiquinone) 1 beta C=183; O=5; E=0.52; R=9.66; rawP=0.0002; subcomplex, 4, 15kDa adjP=0.0011 4 COX6B1 cytochrome c oxidase subunit UserID Gene Name VIb polypeptide 1 (ubiquitous) 1 NDUFB9 NADH dehydrogenase (ubiquinone) 1 beta subcomplex, 9, 22kDa Database:KEGG pathway Name:Alzheimer's

339 Appendix 2

disease ID:05010 2 BAG2 BCL2-associated athanogene 2 C=167; O=4; E=0.47; R=8.47; rawP=0.0013; adjP=0.0036 UserID Gene Name 1 NDUFB9 NADH dehydrogenase (ubiquinone) 1 beta subcomplex, 9, 22kDa 2 NDUFV1 NADH dehydrogenase (ubiquinone) flavoprotein 1, 51kDa 3 NDUFB4 NADH dehydrogenase (ubiquinone) 1 beta subcomplex, 4, 15kDa 4 COX6B1 cytochrome c oxidase subunit VIb polypeptide 1 (ubiquitous)

Database:KEGG pathway Name:Valine, leucine and isoleucine degradation ID:00280 C=44; O=2; E=0.12; R=16.07; rawP=0.0069; adjP=0.0126 UserID Gene Name 1 HADHA hydroxyacyl-CoA dehydrogenase/3-ketoacyl-CoA thiolase/enoyl-CoA hydratase (trifunctional protein), alpha subunit 2 HADHB hydroxyacyl-CoA dehydrogenase/3-ketoacyl-CoA thiolase/enoyl-CoA hydratase (trifunctional protein), beta subunit

Database:KEGG pathway Name:Insulin signaling pathway ID:04910 C=138; O=3; E=0.39; R=7.68; rawP=0.0072; adjP=0.0126 UserID Gene Name 1 HRAS v-Ha-ras Harvey rat sarcoma viral oncogene homolog 2 PHKB phosphorylase kinase, beta 3 SREBF1 sterol regulatory element binding transcription factor 1

Database:KEGG pathway Name:Protein processing in endoplasmic reticulum ID:04141 C=165; O=2; E=0.47; R=4.28; rawP=0.0798; adjP=0.0859 UserID Gene Name 1 BAK1 BCL2-antagonist/killer 1

340