ABSTRACT

KHAREL, PRAKASH, Ph.D., December 2018 Chemistry & Biochemistry

RNA OXIDATION IN MULTIPLE SCLEROSIS NEURODEGENERATION (204 PP.)

Dissertation advisor: Soumitra Basu

An increase in the production of reactive oxygen species (ROS) and the inability of cellular machinery to adequately neutralize the ROS thus produced, lead the cells towards oxidative stress.

ROS can damage all major classes of biomolecules including proteins, DNA and RNA. Recent findings show the evidence of high level of RNA oxidation in the neuronal cells of many neurological disorders suggesting a link between RNA oxidation and neurodegeneration. We have recently discovered the presence of extensive oxidative damage in the RNA molecules of neuronal cells in postmortem brains of multiple sclerosis (MS) patients.

Working on the MS model system, we have established the identity of selectively oxidized mRNA molecules under MS microenvironment in human neuronal cells. Our study reveals that many of the mRNAs linked to various neurological pathways are selectively oxidized under oxidative stress. We have demonstrated the functional consequences of mRNA oxidation in two neuropathology related mRNAs (namely Nat8l and Nlrp3 mRNA). N-acetyl aspartate transferase

8 like protein (NAT8L) is a key trans-mitochondrial membrane protein that catalyzes the transfer of acetyl group from acetyl-CoA to aspartate to form N-acetyl aspartate (NAA) and transports

NAA to neuronal cytoplasm. We have discovered that the oxidation in Nat8l mRNA molecule

i results in the reduced expression of NAT8L enzyme. We also observed a reduced level of NAA present in the neuronal cells under oxidative stress. Using the neuronal tissue culture and MS animal model (cuprizone mouse model), we have established that a higher mRNA oxidation results in a reduced expression of NAT8L protein, which presumably contributes to the MS disease progression by weakening the myelin production machinery. In a different context, we have also observed a unique situation where an oxidation in the Nlrp3 mRNA could lead towards the activation of alternative inflammasome pathway in human neurons under MS microenvironment.

To the best of our knowledge, our study is the first to establish the direct connection between mRNA oxidation and MS neurodegeneration.

In another study, we defined the impact of base oxidation in the structure and function of RNA molecules. We worked on an engineered version of the Tetrahymena group 1 intron (a ribozyme) to evaluate its enzymatic activity under oxidative stress. Our in vitro findings suggest the progressive loss of ribozyme function with increasing oxidation. Additionally, our investigation revealed that RNA oxidation is detrimental in RNA folding. We have also synthesized a novel 8-

OHG analog (phosphorothioate) molecule with a potential to map the interference in the RNA structure due to the presence of oxidatively modified nucleosides.

ii

RNA OXIDATION IN MULTIPLE SCLEROSIS NEURODEGENERATION

A dissertation submitted

to Kent State University in partial

fulfillment of the requirements for the

degree of Doctor of Philosophy

by

Prakash Kharel

December 2018

© Copyright

All rights reserved

Except for the previously published materials

iii

Dissertation written by

Prakash Kharel

B.S., Tribhuvan University, 2006

M.S., Tribhuvan University, 2010

Ph.D., Kent State University, 2018

Approved by

______, Chair, Doctoral Dissertation Committee Soumitra Basu, Ph.D., MBA.

______, Members, Doctoral Dissertation Committee Sanjaya Abeysirigunawardena, Ph.D.

______Jacob Shelley, PhD.

______Jennifer McDonough, Ph.D.

______Ernest Freeman, Ph.D.

Accepted by

______, Chair, Department of Chemistry and Biochemistry Soumitra Basu, Ph.D., MBA.

______, Dean, College of Arts and Sciences James L. Blank, Ph.D.

i

TABLE OF CONTENTS

Page

TABLE OF CONTENTS ………………………………………………………………...v

LIST OF FIGURES ……………………………………………………..….…………....x

LIST OF TABLES ..……………………………………………….………...... xv

LIST OF ABBREVIATIONS…..……………………………………………………....xvii

DEDICATION…………………………………………………………………………..xix

ACKNOWLEDGEMENTS………..………………………………….……..……....…..xx

CHAPTERS

CHAPTER 1. Introduction ……………………………………………….……………….1

1.1. Multiple sclerosis……………………………………………………………..1

1.2. Multiple sclerosis genetics…………………………………………………....4

1.3. Autoimmune attacks in multiple sclerosis…………………………………....5

1.4. Mitochondrial dysfunction in multiple sclerosis……………………………..6

1.5. Oxidative stress in multiple sclerosis and other neurological disorders

………………………………………………………………………………....8

1.6. Metal dysregulation in multiple sclerosis brain………………………….….12

1.7. ROS in CNS inflammation……………………………………………….….13

1.8. Damage in the biomolecules due to ROS…………………………………….15

1.8.1. Lipid oxidation…………………………………………………...15

v

1.8.2. Protein oxidation…….…………………………………...... 16

1.8.3. DNA oxidation…………………………………………………...17

1.8.4. RNA oxidation in MS and other neurological disorders………....17

1.8.5. Coping with RNA oxidative damage………………………….….24

1.9. Hypotheses of the study……………………………………………………...27

1.10. References………………………………………………………………..28

CHAPTER 2. Evidence of extensive RNA oxidation in the neurons of normal appearing cortex of Multiple Sclerosis brain…..…………………………………………………………...41

2.1. Introduction………………………………………………………………….41

2.2. Materials and methods………………….……...…...... 43

2.3. Results………………………………………………...……………………..49

2.4. Discussion……………………………………………….…………………..59

2.5. Conclusion…………………………………………………………………...62

2.6. References…………………………………………………………………...62

CHAPTER 3. Investigation of mRNA oxidation in human neuronal cells reveals selective oxidation of mRNAs could be linked to MS neuropathology………...... 65

3.1. Introduction…………………………………………………..……………...65

3.2. Materials and methods……………………………………....……………….68

3.3. Results………………………………………………….…….……………...73

3.4. Discussion………………………………………………………..………….88

3.5. Conclusion…………………………………………………………………...92

vi

3.6. References…………………………………………………………...... 92

CHAPTER 4. Oxidative damage to Nat8l mRNA in human neurons is linked to multiple sclerosis pathogenesis ……………………………………………………………...….95

4.1. Introduction………………………………………………………………..95

4.2. Materials and methods………………….……...…...... 100

4.3. Results ……………………………………………...…………………....105

4.4. Discussion …………………………………………….…………...... 120

4.5. Conclusion ……………………………………………………………….125

4.6. References………………………………………………………………..126

CHAPTER 5. Evidence of existence of NLRP3 inflammasome pathway in human neuronal cells which could be altered under SNP mediated oxidative stress ...………………..130

5.1. Introduction ..…………………………………………………………….130

5.2. Materials and methods………………….……...…...... 134

5.3. Results ……………………………………………...…………………....136

5.4. Discussion …………………………………………….………………….143

5.5. Conclusion ……………………………………………………………….145

5.6. References………………………………………………………………..145

CHAPTER 6. RNA oxidation impairs RNA function: Synthesis of a novel 8- hydroxyguanosine phosphorothioate analog with a potential to probe the interference of

RNA oxidation in structure and function ……………….…………………………...148

vii

6.1. Introduction ..……………………………………………………………148

6.2. Materials and methods………………….……...…...... 153

6.3. Results ……………………………………………...…………………...160

6.4. Discussion …………………………………………….………………...169

6.5. Conclusion ………………………………………………………………171

6.6. References……………………………………………………………….172

CHAPTER 7. Concluding Remarks………………..………………..…….………….176

CHAPTER 8. Future Perspectives…………………………………………………….178

APPENDIX..………………………………………………………………………….181

viii

LIST OF FIGURES

Page

Chapter 1. Introduction

Figure 1.1. Overview of MS symptoms, neurons & non-neuronal brain cells and neurodegeneration…………………………………………….…………………...... 3

Figure 1.2. Players in MS pathogenesis………………………………………………15

Figure 1.3. Common oxidative base modifications in DNA and RNA……………….20

Figure 1.4. Mutagenic nature of 8-OHG …………………………….………………..21

Figure 1.5. Proposed role of RNA oxidation in the pathogenesis of neurological disorders and cancer……………………………………………………………………………....23

Figure 1.6. A hypothesis to decipher the role of selective RNA oxidation in MS neurodegeneration…………………………………………………...... 28

Chapter 2. Evidence of extensive RNA oxidation in the neurons of normal appearing cortex of multiple sclerosis brain

Figure 2.1. PLP staining to distinguish between lesion and non-lesion areas of MS brain…………………………………...…………………………..…...... 50

Figure 2.2. Immunohistochemical analyses of RNA oxidation in MS brain...... 51

Figure 2.3. Comparative analysis of RNA and DNA oxidation in MS and non-MS brains………..…………...... 53

ix

Figure 2.4. RNA oxidation is predominant in neuronal cells……………………..…..54

Figure 2.5. RNA oxidation is less prominent in oligodendrocyte cells……………….55

Figure 2.6. HPLC analyses of 8-OHG in MS vs non-MS RNA and DNA…………...57

Figure 2.7. Comparison between RNA oxidation in lesion vs non-lesion areas of MS brain

……………………………………………………………………………………...….58

Figure 2.8. mRNA molecules are highly oxidized in MS brain…………...………….59

Chapter 3. Investigation of mRNA oxidation in human neuronal cells reveals selective oxidation of mRNAs could be linked to MS neuropathology

Figure 3.1. A schematic of a proposed path to study the link between mRNA oxidation to

MS neurodegeneration………………………………………………………………... 66

Figure 3.2. Sodium nitroprusside (SNP) and ROS generation scheme…………….....74

Figure 3.3. SNP treatment to SH-SY5Y cells creates oxidative stress in the cells

…………………………………………………………………………………… 75

Figure 3.4. SNP treatment results in RNA oxidation in SH-SY5Y cells ………….....76

Figure 3.5. A flow chart demonstrating stress induction in human neuronal cells with SNP, RNA isolation and immunoprecipitation, quality test of mRNA, cDNA library preparation, RNA-seq analysis, and bioinformatics analysis steps…………………………………………….78

Figure 3.6. mRNA oxidation in SNP stressed cells and relative change in OS related mRNAs in the sequenced samples ……………………………………………………79

Figure 3.7. RNA-seq data analysis and RT-qPCR validation ...………... …..80-81

Figure 3.8. Transcript analysis based on protein class ……………………...... 82-83

x

Figure 3.9. Transcript analysis based on biological processes ………….…………....83

Figure 3.10. Transcript analysis based on molecular function …………………..…...85

Figure 3.11. Transcript analysis based on cellular components ……..………….……86

Figure 3.12. The selective oxidation of mRNA as a function of G enrichment in the transcripts and coding regions…………………………………………………………91

Chapter 4. Oxidative damage to Nat8l mRNA in human neurons is linked to multiple sclerosis pathogenesis

Figure 4.1. NAT8L is central to CNS NAA metabolism.. ……………..………..……98

Figure 4.2. NAT8L protein sequence is well conserved in mammals…………………99

Figure 4.3. Selective mRNA oxidation in SNP treated SH-SY5Y cells is detrimental in

NAT8L protein expression……..…………………………..…………………………107

Figure 4.4. SNP treatment results in reduced NAA level in neuronal cells………….109

Figure 4.5. Cuprizone mouse as an MS model ………………………………….…...110

Figure 4.6. Selective Nat8l mRNA oxidation in cuprizone mice brain ………...... 111

Figure 4.7. NAT8L protein expression is compromised in cuprizone fed mice brain

………………………………………………………………………………………...112

Figure 4.8. Immunohistochemical analysis revealed the lack of NAT8L protein in the neurons of cuprizone-fed mice brains………………….……………………….113-114

Figure 4.9. N-acetyl transferase activity of NAT8L is compromised under MS microenvironment …………………………………………………………….…...115

Figure 4.10. NAT8L protein level is reduced in normal appearing gray areas of MS brain

xi

………………………………………………………………………………………..116

Figure 4.11. In vitro translation showed the loss of translational ability of Nat8l mRNA upon oxidation……………………………………………………………………………..118

Figure 4.12. G-quadruplexes in Nat8l mRNA could play regulatory roles…………120

Figure 4.13. A proposed model of a link between Nat8l mRNA oxidation and MS neurodegeneration …………………………………………………………………..124

Chapter 5. Evidence of existence of NLRP3 inflammasome pathway in human neuronal cells which could be altered under SNP mediated oxidative stress

Figure 5.1. NLRP3 inflammasome pathway ……………..………………...…….…132

Figure 5.2. RT-qPCR analysis to analyze the relative change in the expression and oxidation of NLRP3 inflammasome path related mRNAs …………………………..137

Figure 5.3. LPS treatment activates NLRP3 inflammasome in neurons ………..…..138

Figure 5.4. SNP treatment induced Nlrp3 mRNA oxidation impacts NLRP3 protein production ……………………………………………………………………...... 139

Figure 5.5. Caspase-1 glow assay showed the activation of inflammasome assay under

SNP treatment …………………………………….…………….…………………....141

Figure 5.6. A schematic of a proposed model for mRNA oxidation induced alternative inflammasome path in neurons under oxidative stress ……………………………....142

xii

Chapter 6. RNA oxidation impairs RNA function: Synthesis of a novel 8- hydroxyguanosine phosphorothioate analog with a potential to probe the interference of

RNA oxidation in structure and function

Figure 6.1. Tetrahymena Group I intron and schematic of catalytic reaction ……....151

Figure 6.2. P4-P6 RNA and tertiary interactions crucial for L-21G414 and P4-P6 RNA folding ……………………………………...... 152

Figure 6.3. Synthetic scheme of 8-OHGTPαS ……………………………………...157

Figure 6.4. Mutagenesis to get P4-P6 RNA and mutated T7 RNA polymerase…….159

Figure 6.5. RNA oxidation impairs L-21 G414 ribozyme function.………………...161

Figure 6.6. RNA oxidation results in the unfolding of P4-P6 RNA………………...162

Figure 6.7. Synthesis reaction and TLC chromatogram to track the reaction ………164

Figure 6.8. Mass spectrum of purified 8-OHGTPαS …………………………...... 165

Figure 6.9. 31P-NMR spectrum of purified 8-OHGTPαS ……………...... 166

Figure 6.10. 8-OHGTPαS can be incorporated in transcript with the help of mutated T7

RNA polymerase with difficulty ………………………………………………….…..169

Figure 6.11. A schematic representation of NAIM technique …………………….….171

xiii

LIST OF TABLES

Page

Chapter 1. Introduction

Table 1.1. Half-life of ROS in our cells…………………………………………….10

Table 1.2. Reactions centered around free radical metabolism……………………..12

Chapter 2. Evidence of extensive RNA oxidation in the neurons of normal appearing cortex of multiple sclerosis brain

Table 2.1. Table 2.1 Demographics of the donors used in the study………...……..44

Chapter 3. Investigation of mRNA oxidation in human neuronal cells reveals selective oxidation of mRNAs could be linked to MS neuropathology

Table 3.1. Primers used for RT-qPCR analyses……………………………………...73

Table 3.2. Catalytic proteins whose mRNAs are selectively oxidized ……………....84

Table 3.3. Mitochondrial proteins whose mRNAs are selectively oxidized...... 87-88

Chapter 4. Oxidative damage to Nat8l mRNA in human neurons is linked to multiple sclerosis pathogenesis

Table 4.1. Primers used for RT-qPCR analyses……………………………………..102

xiv

Table 4.2. Primers used for mutagenesis…………………………………………...104

Chapter 5. Evidence of existence of NLRP3 inflammasome pathway in human neuronal cells which could be altered under SNP mediated oxidative stress

Table 5.1. Primers for RT-qPCR analyses ………………………………………...135

Chapter 6. RNA oxidation impairs RNA function: Synthesis of a novel 8- hydroxyguanosine phosphorothioate analog with a potential to probe the interference of

RNA oxidation in structure and function

Table 6.1. Primers used for mutagenesis. ………………………………………...158

xv

LIST OF ABBREVIATIONS

MS, Multiple sclerosis

ALS, amyotrophic lateral sclerosis

AD, Alzheimer’s disease

PD, Parkinson’s disease

CNS, Central nervous system

RRMS, Relapsing remitting MS

SPMS, Secondary progressive MS

PPMS, Primary progressive MS

PRMS, Progressive relapsing MS

HLA, Human leukocyte antigen

CD, Cluster of antigens

GWAS, Genome wide association studies

Th, T helper

TNF, Tumor necrosis factor

COX, Cytochrome oxidase

OS, Oxidative stress

ROS, Reactive oxygen species

TLR, Toll-like receptors

GQ, G-Quadruplex

PGQS, Potential G-quadruplex forming sequence

PAGE, Polyacrylamide gel electrophoresis

xvi

RNS, Reactive nitrogen species

8-OHG, 8-hydroxyguanosine

G, Guanosine

C, Cytosine

A, Adenosine

TH, Tyrosine hydroxylase

GSH, Glutathione

NMR, Nuclear magnetic resonance

CSF, Cerebrospinal fluid

MRI, Magnetic resonance imaging

MRS, Magnetic resonance spectroscopy

CyPD, Cyclophilin D

EAE, Experimental autoimmune encephalomyelitis

SNP, Sodium nitroprusside

HPLC, High performance liquid chromatography

TLC, Thin layer chromatography

NAA, N-acetyl aspartate

NAAG, N-acetyalaspartylglutamate

SOD, Superoxide dismutase

PLP, Proteolipoprotein

MOG, Myelin oligodendrocyte glycoprotein

APC, Antigen presenting cells

IL, Interleukin

xvii

CSF, Cerebrospinal fluid

EAE, Experimental autoimmune encephalomyelitis

ATP, Adenosine triphosphate

BBB, Blood brain barrier

RNA-seq, High throughput RNA sequencing

NLRP3, NLR family, pyrin domain containing 3; NLR, nucleotide-binding domain, leucine-rich repeat

NLRP1, NLR family, pyrin domain containing 1; NLR, nucleotide-binding domain, leucine-rich repeat

AIM2, absent in melanoma 2

xviii

DEDICATION

This dissertation is dedicated to my mother Sumkala Kharel and my late father Keshab Raj

Kharel for their struggle and support to ensure my education.

xix

ACKNOWLEDGEMENTS

First and foremost, I am indebted to my advisor, Dr. Soumitra Basu, for being a mentor, a friend, and a motivating person who always allowed me to fly on my own yet giving proper directions whenever my compass jammed. Dr. Basu was open to any kind of discussions and taught values that go far beyond the laboratory work. I came to join the Basu lab with no biochemistry and molecular biology experience, but Dr. Basu never was scared to take on a challenge. I doubt that one can get a biochemistry and molecular biology training along with organic synthesis, neurobiology, translational research and many more diverse trainings in a single lab if it were not in Basu lab. I believe over the past 5 and half years we have truly been able to explore the role of

RNA in neurobiology, working on projects that span from fundamental biochemistry to neurobiology to translational research and a huge part of that is the vision and confidence Dr. Basu had in myself and the members of our group.

I would like to acknowledge those who collaborated or helped on the projects described in each chapter within this dissertation. Without the help of these collaborators this work would have not been able to be completed or answered the questions we are attempting to solve. Chapter 2, Drs.

Jennifer McDonough, Naveen Singhal, Rob Clements and Mike Model of Kent State University

(KSU) Department of Biological Sciences; Chapter 3, Dr. Jennifer McDonough of KSU

Department of Biological Sciences, Dr. Anshuman Chattopadhyay of University of Pittsburgh, Dr.

Venkat Sundar Gadepalli of Virginia Commonwealth University and Girihlet Non-Canonical

Genomics; Chapter 4, Drs. Jennifer McDonough and Naveen Singhal of the KSU Department of

Biological Sciences, Joram Rana, Lindsey Smith and Brintha Croos of KSU Department of

Chemistry and Biochemistry; Chapter 5, Lindsey Smith and Brintha Croos of the KSU Department of Chemistry and Biochemistry; Chapter 6, Brintha Croos, Rick Dunn, Dr. Jacob Shelley, Sunil

xx

Badal, Courtney Walton, and Mohamad Rahman of KSU Department of Chemistry and

Biochemistry. Past lab members Drs. Mohamed Farhath, Gayan Mirihana Arachchilage and

Debmalya Bhattacharyya were influential in setting a vibrant research discussion environment in lab and helped me a lot when I started learning the ABCs in Basu lab. Each of these great fellow scientists are also my collaborators in the projects which are not discussed in this dissertation.

I would like to thank my committee member and collaborator Dr. McDonough, for helping to guide my research, my understanding of neurobiology, and teaching me immunohistochemistry skills; to the committee member and collaborator Dr. Jacob Shelley and committee members Dr.

Sanjaya Abeysirigunawardena for helping to guide my research. I would also like to thank my previous committee member Dr. Roger Gregory for the valuable suggestions and critical inputs in my research. I am grateful to Drs. Robert Twieg and Prabin Rai of KSU Department of Chemistry and Biochemistry for always being there to help quench my Ochem thirst.

Nathan, Sumirtha and I have been together since the second semester of our first year. Anthony has also been around for the entirety of my graduate school experience. Each of them is as brilliant as the next but more importantly contributed in creating an incredibly high standard of research work in Basu lab. No result seemed good enough in their eyes till it became perfect, but that level of scrutiny is what ended up making my work down the road successful and I can’t thank them enough. Nate is also a collaborator of mine in a couple of projects and always a partner of mine in some crazy research ideas, Thanks Nate. I would also like to thank Brintha, Brad and Dipen, the future of Basu Lab, especially Brintha who will continue down the RNA in neurobiology path in the Basu lab. Rick, Lindsey and Joram are incredible undergraduate students each of whom shared more than a year working with me, thank you all. Briana and Jessica, they followed me to join

Basu lab all the way from OChem teaching labs, you girls are incredible.

xxi

I would like to thank Erin, the heart of the Chemistry Graduate Program, who has never turned down a question or a chance to help a graduate student in need. Janie, Crystal, Kristen, Catherine and the other student staff that worked at KSU during my tenure should also be thanked for their help on a daily basis. I am also thankful to Dr. Mahinda Gangoda and Larry Maurer of KSU

Department of Chemistry and Biochemistry for their help with instrumental problem solving.

Finally, I would like to thank my family and most importantly my wife Srijana and our 15 months old son Prabhas. In addition to their role in my family life, the stress within this Ph.D. process was greatly reduced by their support and unconditional love. Friends, family, co-workers, colleagues, all in some way or another have played a part in my graduation and I cannot thank each of them enough for being a part of this journey.

xxii

CHAPTER 1

1. Introduction

1.1 Multiple sclerosis

Multiple sclerosis (MS) is a demyelinating disease of the central nervous system (CNS), beginning most often in early adulthood1,2. MS involves an autoimmune-mediated inflammatory demyelination that is coupled with chronic neurodegeneration, resulting in progressive cognitive and physical disability3. The pathological hallmark of MS consists of focal demyelination, inflammation, scar formation, and axonal degeneration. These features in the brain and the spinal cord eventually result in the disturbance of normal conduction of electrical impulses in the CNS4. It is one of the most common non-traumatic causes of chronic neurodegenerative disorders in young adults, with an average onset age of 30 years5. The reported data show that MS affects more than

2.3 million people worldwide and 400,000 individuals in the USA alone6,7.

There are four different types in MS, each based on the progression of the disease 2. Relapsing- remitting MS (RRMS), which affects more than 85% of the total MS population, is characterized by a course of recurrent and reversible neurological deficits called ‘relapses’ which are supposed to result from an autoimmune-mediated inflammatory attack on myelin, oligodendrocytes, and neuronal axons2. Inflammatory signals are believed to initiate from the inflammation mediator molecules consisting of blood-derived lymphocytes and monocytes, which enter the brain through a compromised blood-brain barrier and appear to target specific myelin proteins8.

1

Relapses usually last for a few months when patients frequently regain normal neurological function representing a period of clinical recovery or remission. The mechanisms of remission include resolution of inflammation and edema, limited remyelination and restoration of normal electrical conduction9. Further, in cases of prolonged demyelination, upregulation and redistribution of sodium channels along axons has been documented, suggesting an additional mechanism for re-establishing signal conduction10. Despite temporary periods of remission and recovery of function, it has been reported that an irreversible and progressive loss of brain tissue begins early in the disease course2. Usually within a decade, a majority of RRMS patients transition to a secondary progressive stage of MS (SPMS) in which the number of inflammatory episodes diminishes, but continuous irreversible neuron degeneration and progressive disability increases11.

As a result, it appears that the overall axonal loss and neurodegeneration are independent of immune attack, and contribute to permanent neurological disability and further brain atrophy at this stage in MS patients11.

About 10% of MS patients are diagnosed with primary progressive MS (PPMS) at the initial onset of the disease2. Compared with RRMS, PPMS involves fewer inflammatory attacks, with far less brain lesions12. However, greater numbers of spinal cord lesions have been identified in patients with PPMS, which likely underlies the more progressive accumulation of physical disabilities noted in this disease. The least common form of MS is progressive relapsing MS (PRMS), which is diagnosed in approximately 5% of MS patients13,14. People with PRMS experience steady neurological dysfunction progression from onset in addition to occasional inflammatory relapses and demyelination14.

2

Figure 1.1 Overview of MS symptoms, neurons and non-neuronal brain cells and neurodegeneration, a. Main symptoms of MS patients (By Mikael Häggström, used with permission), b. A representative network between neuron and different non-neuronal cells in the brain (Allen and Barres, Nature, 2009, used with permission), and c. A schematic representation of neurodegeneration, 1-a healthy neuron, 2-A neuron under initial attack by ROS or inflammatory molecules, and 3-degenerated neuronal axon (Trapp and Nave, Annual Reviews of Neuroscience, 2008, used with permission).

3

1.2 Multiple sclerosis genetics

Whole genome screening of different MS populations has recently identified discrete chromosomal regions potentially harboring MS susceptibility genes15-19. However, apart from the major histocompatibility complex on 6p21, no single locus generated overwhelming evidence of genetic linkage. These results suggest complex genetic etiology for MS, including multiple genes having a small to moderate effect and probable genetic heterogeneity15-18,20. On the other hand, the human leukocyte antigen (HLA) was found to control immune response genes in MS, with HLA associations indicating the involvement of autoimmunity19. Further, MS was one of the first diseases proven to be HLA-associated, primarily linked to HLA class II factors21. Furthermore,

CD45 or protein tyrosine phosphatase receptor-type C (PTPRC) has been reported as a candidate in some families with MS22.

Polymorphism in two cytokines receptors have found to be associated with MS, namely interleukin

7 receptor alpha chain gene (IL7RA) on chromosome 5p13 and the interleukin 2 receptor alpha chain gene (IL2RA or CD25) on chromosome 10p15. It is estimated that the C allele of a single nucleotide polymorphism within the alternative spliced exon 6 of IL7RA is involved in about 30% of MS cases23. These results indicate that MS has a strong genetic component. Some of these findings (such as CD25) were confirmed by pathway and network-based genome-wide association studies (GWAS)20,24. As reviewed by Baranzini and Oksenberg, GWAS also revealed more extensive immune antigens, cell adhesion, and signaling molecules associated with MS, such as

CD4, CD11b, CD58, CD82, ITGB2, and STAT3, as well as glutamate receptors, multimeric scaffold molecular DLG1, and DLG220. Using a pooling-based, genome-wide approach, and high- density, single-nucleotide polymorphisms arrays, GWAS also identified a novel risk locus for MS on chromosome 13, in addition of the HLA class II genes (such as HLA-DRB1)20,24. Although

4 more studies need to be done in the MS context to clearly establish the cause effect relationship of genetic factors in MS development, the involvement of rare variants and endophenotypes in many genes in MS looks very possible.

1.3 Autoimmune attacks in multiple sclerosis

Proinflammatory cytokines, such as interferon and tumor necrosis factor beta (TNFβ) released by activated T helper 1 cells (Th1) upregulate the expression of cell-surface molecules on neighboring lymphocytes and antigen-presenting cells25. The binding of putative MS antigens may potentially trigger an enhanced immune response against the bound antigens26,27. The infiltration of Th cells results in microglia activation in MS brain. The activated state of microglial cells was also reflected by increased expression of human leukocyte antigen-DR (HLA-DR) and CD6828.

Components of myelin, such as myelin basic protein, myelin-associated basic glycoprotein, myelin oligodendrocyte glycoprotein (MOG), proteolipid protein (PLP), the T cell receptor, and major histocompatibility complex class II molecules on antigen presenting cells (APCs) are putative MS antigens29.

In addition to the autoimmune response, oligodendrocyte death, axon damage, and neuronal loss has also been associated with inflammatory attacks on the CNS in MS patients30. The exact reason for these attacks is largely unknown, although genetic factors have been speculated to influence immune-mediated inflammation as well as neuronal and glial survival by modulating the MS phenotype31. Therefore, autoreactive T cells are believed to be generated in response to the interplay of immune triggers and genetic susceptibility factors31. Differentiation of such CD4+ T cells results in pro-inflammatory Th1, Th17 cells and/or regulatory Th2 cells, all of which produce

5 cytokines such as interferon-gamma, IL-17, IL-4 and IL-1032. After activation, myelin-specific T cells can cross the blood-brain-barrier via interaction of adhesion molecules, such as, vascular cell adhesion molecule 133. Reactivation of these T cells involves local APCs in the CNS which initiate a cascade that typically involves the attraction of microglia, macrophages, CD8+ T cells, and plasma cells which produce myelin-specific antibodies34. It has been speculated that these tiered mechanisms in combination may lead to mitochondrial dysfunction, demyelination, and irreversible tissue damage in MS35. Recent studies showed that myelin-specific T cells also recognize the neuronal autoantigen in mouse models of MS, further indicating that multiple autoantigens may be involved in spontaneously developing human MS disease36,37.

1.4 Mitochondrial dysfunction in multiple sclerosis

Increasing evidence suggests the involvement of mitochondrial dysfunction in MS and other neurological disorders35,38-40. Analysis of demyelinated spinal cord lesions from MS patients showed dramatically reduced numbers of mitochondria and microtubules40. The study also found

Ca2+ mediated destruction of chronically demyelinated axons and axonal swelling in the CNS. As reduced energy production is a major contributor to Ca2+ mediated axonal degeneration, the alteration in oxidative phosphorylation is proposed to be involved in this destruction40. It has been observed that twenty-six nuclear-encoded mitochondrial genes and the functional activities of mitochondrial respiratory chain complexes I and III were decreased in the MS motor cortex, especially in neurons38. Immunohistochemical analysis demonstrated that functionally important defects of mitochondrial respiratory chain complex IV (cytochrome c oxidase, COX) including its catalytic component (COX-I) are present in some of the active MS lesions41. The same study also demonstrated the lack of immunohistochemically detectable COX-I in oligodendrocytes,

6 astrocytes and neuronal axons. In the inactive areas of chronic MS lesions, complex IV activity is increased within approximately half of the large chronically demyelinated axons compared with large myelinated axons in the brain and spinal cord42,43. The axon-specific mitochondrial docking protein (syntaphilin) and phosphorylated neurofilament-H were increased in chronic lesions42,43.

These results clearly indicate an adaptive change of mitochondrial function and morphology in chronic MS lesions.

Regenold et al. have investigated the relationship between disturbed CNS mitochondrial energy metabolism and MS disease progression by measuring the concentrations of sorbitol, fructose, and lactate in cerebrospinal fluid (CSF)44. They found that concentrations of all three metabolites, but not concentrations of glucose or myo-inositol, were significantly increased in the CSF of MS patients compared to healthy controls. Furthermore, CSF concentrations of sorbitol and fructose

(polyol pathway metabolites), but not lactate (anaerobic glycolysis metabolite), correlated positively and significantly with Expanded Disability Status Scale (EDSS) score, an index of neurologic disability in MS patients. These findings suggest that abnormal mitochondrial glucose metabolism is increased in MS patients and is associated with disease progression. This suggests that under MS circumstances there is a problem within the Kreb’s cycle.

Interestingly, analysis of mitochondrial enzymes on MS muscle showed that there were fewer type

I fibers31. The same study also demonstrated that the fibers of all types were smaller and had lower succinate dehydrogenase (SDH, component of the respiratory chain complex II) and SDH/alpha- glycerol-phosphate dehydrogenase (GPDH) but not GPDH activity, suggesting that muscle in this disease is smaller and relies more on anaerobic than aerobic-oxidative energy supply compared to muscle of healthy individuals. Like brain, muscles are also highly dependent on mitochondrial oxidative energy metabolism, so it is reasonable that there are weaker muscles in MS patients,

7 indicating muscle is also one of the targets of MS31. In some rare cases, MS could have a mitochondrial myopathy combination, in which MRI showed widespread white matter lesions where muscle biopsy displayed ragged fibers and COX deficiency. A large deletion of mitochondrial DNA (mtDNA) was observed in the MS patients indicating mitochondrial genomic deletion could be the key cause or initiation factor for this special case31,45.

The mitochondrial permeability transition pore (PTP) consists of the adenine nucleotide translocator, a voltage-dependent anion channel, and cyclophilin D (CyPD) which is a prolyl isomerase located within the mitochondrial matrix. CyPD is a key regulator of the PTP and required for mediating Ca2+ signaling and oxidative damage induced cell death46. In the experimental animal MS disease model (EAE mouse model), the neurons missing CyPD are resistant to oxidative agents and are thought to be the mediators of axonal degeneration observed in both EAE and MS47. Such neurons have mitochondria that can more effectively handle elevated

Ca2+. Consistent with this neuronal resistance, animals missing CyPD can recover clinically, following the induction of EAE47. These results connect pathological activation of the mitochondrial PTP in the axonal damage occurring during MS progression. A reduced level of neuronal health marker metabolite N-acetyl aspartate (NAA) as detected by magnetic resonance spectroscopy (MRS) in MS patients and by HPLC analysis in MS postmortem brain is another indicator of mitochondrial dysfunction in MS brain48,49.

1.5 Oxidative stress in multiple sclerosis and other neurological disorders

A free radical is a highly reactive chemical species that contains one or more unpaired electrons50.

The most common cellular free radicals are hydroxyl radical (·OH), superoxide radical (O2 –·), and

8

· 51 nitric monoxide radical (NO ) . Other molecules or ions, such as, hydrogen peroxide (H2O2) and peroxynitrite (ONOO–) are not free radicals but can lead to the generation of free radicals through various chemical reactions51. Free radicals and free radical producing reactive molecules are often classified together as reactive oxygen species (ROS) or reactive nitrogen species (RNS) (both will collectively be called ROS in this dissertation) to signify their ability to cause oxidative changes within the cell52. The ROS do have a wide range of half-lives ranging from less than a nanosecond for ·OH to a stable species like H2O2 (Table 1.1). ROS are mostly the by-products of the mitochondrial respiratory mechanism52. Besides being a by-product of respiration, ROS can also be synthesized in activated macrophages and microglia by enzymes like myeloperoxidase, xanthine and NADPH oxidases (NOX)53. Activated microglia are considered to lead the way to ongoing neurodegeneration via the production of ROS in MS and other classical neurological disorders54.

At low levels, ROS can play an essential role in cell division and survival, cell signaling, inflammation and immune functions, autophagy, and stress response. However, a redox imbalance caused by exposure to oxidants or metal dysregulation alters the cellular ability to detoxify ROS or to repair any damage caused by them. ROS are not totally harmful since they are the integral part of cellular defense mechanism and can act as signaling agents55. It is the overproduction of

ROS that is harmful to the cells. Cells normally have several mechanisms to defend against damage induced by free radicals. Problems occur when production of ROS exceeds their elimination by the antioxidant defense system, or when the latter is damaged55. This imbalance between cellular production of ROS and the inability of cells to defend against them is called oxidative stress

(OS)55,56.

9

Table 1.1 Half-life of ROS in our cells.

Recent studies revealed that OS plays a major role in the pathogenesis of many neurological disorders including Alzheimer’s disease (AD), Parkinson’s disease (PD), Amyotrophic lateral sclerosis (ALS), and MS57-60. ROS have been implicated as mediators of demyelination and axonal damage in MS61. ROS cause damage to each type of biomolecules, such as, lipids, proteins and nucleic acids resulting in cell death by necrosis or apoptosis56. The weakened cellular antioxidant defense systems in the central nervous system (CNS) in MS, and its vulnerability to ROS effects may increase the effective damage to the neuronal components61. The destruction of the myelin sheath due to microglial ROS attack (among many other events) in MS neurons will damage the trophic relationship between oligodendrocytes, astrocytes and neurons, eventually leading to neuronal atrophy and degeneration that is associated with secondary disease progression in

MS62,63.

Mitochondria are the major site for the generation of free radicals64. The electron transfer from the donors produced in the tricarboxylic acid cycle to molecular oxygen causes proton translocation

10 and energy release for ATP synthesis. Electrons can leak out from the energy-transducing

-· sequences and then bind with molecular oxygens to form superoxide radicals (O2 ) in the normal process of mitochondrial respiration64. Superoxide radicals have a dual effect. They are involved in cell defense pathway against infectious microorganisms, but they can also damage cells or form the most reactive species, hydroxyl radicals or peroxynitrite65. Superoxide radicals cross the cell membrane through anion channels. They can be converted into hydrogen peroxide by superoxide dismutase (SOD) as shown in Table 1.265. There are two types of SODs; one is manganese- dependent SOD (Mn SOD) located in mitochondria, where it converses superoxide radicals derived from the electron transfer chain into H2O2. The other is copper- and zinc dependent SOD

(Cu/Zn SOD) located in the cytosol with a more generic catalytic function66. Hydrogen peroxides can also be produced from the oxidative deamination of monoamines catalyzed by monoamine oxidase, which is associated with the outer mitochondrial membrane67. Despite being chemically less reactive, H2O2 can act as a more serious oxidant due to high stability (Table 1.1). Catalase and glutamate peroxidase are two major enzymes in the cytoplasm to decompose intracellular hydrogen peroxides (Reactions 1.2 & 1.3, Table 1.2). Glutathione peroxidase converts reduced glutathione (GSH) and hydrogen peroxide into oxidized glutathione (GSSG) and water65.

Although hydrogen peroxide itself is not as reactive, it can diffuse freely within a long distance and form hydroxyl radicals in the presence of redox metal ions by Fenton reaction (Reaction 1.4,

Table 1.4). Another ROS fighting platform is composed of nuclear factor erythroid related factor

2 (NRF2), a transcription factor that induces antioxidant enzymes as heme oxygenase 1 (HMOX1), which scavenges free radicals and remove damaged proteins68. The ability of HMOX1 to catabolize free heme and produce carbon monoxide (CO) also gives its anti-inflammatory

11 properties by up-regulation of interleukin 10 (IL-10) and interleukin 1 receptor antagonist (IL-

1RA) expression69.

Table 1.2 Reactions centered around free radical metabolism.

1.6 Metal dysregulation in multiple sclerosis brain

Redox active metals like iron (Fe) and copper (Cu) have been linked to neurodegeneration in MS,

AD, ALS and other neurological disorders70. These metals are essential cofactors for many enzymes and structural elements for stabilizing proteins. They also participate in principal biological metabolisms of the brain, including neurotransmitter synthesis, nerve transition, and oxygen transport70,71. Iron is a cofactor in the catalytic center of various enzymes for normal brain metabolism, including oxidative phosphorylation, myelination and neurotransmitter formation.

Moreover, Fe2+ participates in normal physiological processes of oligodendrocytes for the construction of myelin71. Tight homeostasis of cellular iron is required since excessive concentrations can become deleterious for cell function71,72. Iron mismanagement can cause microglia activation, induction of mitochondrial dysfunction, and generation of free radicals in the brain73. An injury of high Fe2+ containing oligodendrocyte cells during neuropathological

12 situations releases Fe2+ into the extracellular space where it is taken up by activated microglia74.

They degenerate eventually contributing to Fe2+ accumulation in extracellular space and axons74,75.

Indeed, the redox capacity of free iron to carry out one-electron reactions, catalyzing the formation of ROS, is proposed as a key factor in MS. Iron accumulation correlates with early axonal injury and is prominent in active lesions of patients with acute MS and short disease duration implying the involvement of ROS in the early phase of disease progression76. Iron accumulation, which presumably comes from myelin, oligodendrocytes, and the breakdown of the blood–brain barrier

(BBB) was found during active and recovery phases of EAE mouse model as well70.

Copper (Cu2+) dysregulation in CNS can also have a similar effect compared to Fe2+ dysregulation.

Cuprizone, a copper chelator, fed mouse model (a toxicant induced MS model) shows a severe demyelination in mouse brain which supports the importance of copper dysregulation in MS pathophysiology77,78. Cuprizone is proposed to carry copper into the CNS to induce prominent demyelination lesions through oxidative stress and oligodendrocyte toxicity77.

1.7 ROS in central nervous system inflammation

Inflammation, in general, is a protective response to various cell and tissue injuries79. The purpose of this process is to destroy and remove the detrimental agents and injured tissues, thereby benefiting tissue repair79,80. When this helpful response is uncontrolled, the effect initiates excessive cell and tissue damages that result in destruction of normal tissue and chronic inflammation79,80. Moreover, MS and other neurological disorders are characterized by redox imbalance and chronic inflammation, a major cause of cell damage and death50. Reactive oxygen species are widely recognized as key mediators of cell survival, proliferation, differentiation, and

13 apoptosis50. Excessive production of ROS is responsible for tissue injury associated with a range of brain injury, inflammation, and neurodegenerative diseases50. Moreover, many of the well- known inflammatory target proteins, including Nucleotide-binding oligomerization domain,

Leucine Rich Repeat and Pyrin domain containing 3 (NLRP3), Matrix Metalloproteinase-9

(MMP-9), Cycloxygenase-2 (COX-2), and inducible nitric oxide synthase (iNOS) are associated with excessive oxidative stress induced by proinflammatory factors such as cytokines, peptides, bacterial infections, and prooxidants81. Several studies have shown that activated microglia cells induce and release diverse inflammatory mediators in response to oxidative stress in CNS82.

Nevertheless, the inflammation response mechanism is active in other brain cell types including neurons and astrocytes83. In addition, ROS act as critical signaling species to trigger inflammatory responses in the central nervous system (CNS) through the activation of the redox-sensitive transcription factors, including Nuclear Factor-κB (NF-κB) and activator protein-1 (AP-1)84. The relationship between OS and the inflammatory response in cells, therefore, is often incredibly complicated since one produces the other and under the chronic situation the cycle just keeps going.

In contrast to the classical view of MS being a completely autoimmune based disease where immune responses lead toward neuroinflammation and neurodegeneration, many recent studies suggest the involvement of mitochondrial dysfunction and ROS as the causative factors for MS neurodegeneration along with autoimmune attacks. In either case, all three components are present in MS CNS: autoimmune attacks, inflammation/ROS and neurodegeneration (Figure 1.2).

14

Figure 1.2 Players in MS pathogenesis.

1.8 Damage in the biomolecules due to ROS

1.8.1 Lipid damage

Reaction of ·OH radicals with lipids can produce lipid radicals following hydrogen abstraction.

The products which are the lipid radicals can react to molecular oxygen to form highly damaging lipid peroxide radicals and respective adducts85. Lipid peroxidation is one of the most studied components of oxidative damage within neuropathological circumstances. Free radical-mediated lipid peroxidation proceeds by a chain mechanism, that is, one initiating free radical can oxidize both lipid molecules in bio-membranes and low-density lipoproteins86,87. Lipid peroxidation has been shown to induce disturbance in membrane organization and functional loss or modification of proteins and nucleic acids86. However, lipid peroxidation products also act as redox signaling mediators86,87. Using an animal model of AD it has been shown that oxidative stress and lipid peroxidation can induce amyloid b (Ab) accumulation during the progression of AD88. It has been reported that the lipid peroxidation adducts, 4-hydroxy-2-alkenal, 4-hydroxy-2-nonenal (HNE) and N6-(carboxymethyl) lysine, are localized in Lewy bodies in post-mortem PD brain tissue89,90.

15

Lipid oxidation product HNE reacts with a small but highly conserved protein α-synuclein which can result in conformational changes and oligomerization of the later. The HNE-modified protein oligomers are potentially toxic and could contribute to the demise of neurons when subjected to oxidative damage in PD88,90. Oxidized phospholipids were also located in myelin, in oligodendrocytes and in glial fibrillary acidic protein (GFAP) positive astrocytes in addition to neuronal cell body in postmortem MS brain91. A granular reactivity for oxidized phospholipids was also observed in the cytoplasm of macrophages, consistent with their content of lipofuscin.

Quantitative analysis showed the highest density of axonal spheroids containing oxidized phospholipids in active lesions, followed by slowly expanding lesions, inactive lesions and normal-appearing white matter in MS33,34,91. Dense cytoplasmic staining for oxidized phospholipids was also present in the perinuclear cell bodies and dendrites of some neurons in active lesions in the cortex and the basal ganglia91. Many of the neurons with cytoplasmic staining for oxidized phospholipids revealed irregularities or fragmentation of their dendritic processes91.

1.8.2 Protein damage

Many reports have shown a prominent protein oxidation and nitration under oxidative stress76,91,92.

Excessive protein oxidation, carbonylation and nitration is a hallmark of an AD brain93. Protein nitration involves the components of mitochondrial respiratory chain complexes I and IV, glycolysis, and chaperones critical to the stabilization and import of proteins into the mitochondria92,94. They were identified as the NDUFA6 subunit of complex I, cytochrome oxidase subunit IV, GAPDH, and mitochondrial Hsp70 chaperone95. The NDUFA subunits are critical components involved in assembly of the holoenzyme NADH dehydrogenase complex. This subunit of complex I was reported to be susceptible to inactivation by peroxynitrite anion mediated oxidation94. Cytochrome oxidase subunit IV is a key subunit of complex IV of the respiratory

16 chain. Although the nitration of complex IV is believed to have a lesser impact on respiration than nitration of complex I, involvement of both complexes contributes to loss of mitochondrial ATP synthesis95. Many protein residues from the cerebrospinal fluid (CSF) of patients with neurological disorders were found to be nitrated, oxidized or glycated94. Tyrosine hydroxylase nitration is proposed to be correlated with a low dopamine processing in PD96. A critical ROS neutralizing enzyme, SOD1, is shown to aggregate and lose its function under oxidative stress in vitro ALS microenvironment97. Oxidation of proteins has been compared as heat induced protein denaturation indicating a critical functional impact of oxidative damage to proeins98.

1.8.3 DNA damage

A plethora of research has shown a higher level of DNA oxidation under OS99. It has been reported that there is a higher level of DNA oxidation both in nuclear and mitochondrial genome in the lesion areas of MS brains76,91. There is evidence of extensive DNA oxidation in AD, ALS, PD and bipolar disorder99. DNA oxidation could activate p53 pathway leading to apoptosis or cell cycle arrest which could lead towards neurodegeneration in CNS. DNA oxidation can also lead to strand breakage and mutation99. In a different context, DNA oxidation has recently been shown to have an epigenetic function100. Fleming et al. have recently demonstrated that 8-oxo guanine in DNA has an epigenetic role where it can signal transcription on switch in VEGF gene100.

1.8.4 RNA oxidation in neurological disorders

Oxidative RNA damage has been shown to be involved in the pathogenesis of several neurological disorders including AD, PD, dementia with Lewy bodies, ALS, and prion diseases50,101. An increased level of oxidative DNA and RNA damage and alteration of ribosome function was reported in postmortem brains of patients with AD and PD 102. Interestingly, RNA oxidative

17 damage was greatest early in the disease and decreased with disease progression, thus suggesting that increased oxidative damage is an early event in Alzheimer’s and other neurodegenerative diseases103. The most common RNA and DNA base modification, 8-hydroxy guanosine (8-OHG), was also observed in cerebrospinal fluid from patients with AD and PD and in serum of PD patients104,105. Different studies detected a regional distribution of RNA oxidative damage that mirrors the selective neuronal vulnerability that characterizes each neurological disease105. Indeed, an increase in the levels of 8-OHG was observed in the hippocampus and cerebral neocortex in

AD, in the substantia nigra in PD, and in the motor cortex and spinal cord in ALS. No alteration in 8-OHG levels was detected in the cerebellum in Alzheimer’s disease, Parkinson’s disease, and amyotrophic lateral sclerosis compared with the controls106,107. Further studies reported that oxidative RNA damage was predominantly localized in neurons compared with glial cells in AD and PD107. The analysis of postmortem tissues from patients with Alzheimer’s disease and amyotrophic lateral sclerosis provides with evidence that mRNA oxidation does not represent a random phenomenon but is highly selective108,109. The hypothesis that the selective mRNA oxidation was due to the abundance of mRNA species was also excluded. Although 훽-actin and

MAP-2 mRNAs are abundant since they are housekeeping mRNA species, only very small amounts of their oxidized products were detected109. The involvement of RNA base modifications induced by ROS has been suggested as a possible initial event in the etiology of sporadic prion disease103,110. RNA damage has also been described in advanced human atherosclerotic plaques111.

As reported in the same study, oxidatively damaged mRNAs are translated in proteins forming aggregates or generate truncated proteins, which must be degraded by the proteasome108,109.

Ribosome stalling is another consequence of mRNA oxidation, as suggested by the decreased rate

18 of protein synthesis reported for polyribosomes isolated from neurons vulnerable to oxidative damage of patients with AD compared to polyribosomes from healthy neurons109.

Cellular RNA is more prone to oxidative damage than DNA for following obvious reasons29,101,103,109,112,113. The primary reason is the abundance where mammalian cells have RNA levels more than four times higher than that of DNA by weight. Secondly, RNA molecules are relatively less associated with proteins. The third reason is that single stranded RNA, the predominant form, are not protected by hydrogen bonding and are thus more easily accessible to

ROS. Another factor that may account for the high susceptibility of RNA to oxidative damage is the iron-binding properties of certain classes of RNAs103. Lastly, RNA has an extensive cytoplasmic distribution, where the majority of ROS is generated. Consistent with the above reasoning several studies have established that the levels of oxidative damage in RNA can be 10–

100 times higher than that of DNA from the same origin in different models including rat liver, human leukocytes, human neurons, and lung epithelial cells113,114. Moreover, a study quantified nucleic acid damage on livers of rats treated with doxorubicin, a known ROS generator, resulted in a significant increase in liver RNA oxidation but no significant increased DNA oxidation101.

Purified ribosomes from Alzheimer’s patients were found to have elevated levels of associated redox-active iron. Moreover, in vitro studies demonstrated that rRNA had higher iron binding than tRNA or mRNA. Accordingly, iron-rich rRNA had a 13-fold greater formation of 8-OHG in oxidation experiments as compared to the iron-poor tRNA115. The higher levels of oxidization observed for RNA could also be due to different rates of removal of RNA and DNA damage.

Indeed, in contrast to the many repair mechanisms for DNA, little is known about how and to what extent oxidatively damaged RNA is removed or repaired113.

19

Alterations of the ribose, base excision, and strand breakage can represent other impacts induced by ROS on RNA113. Apart from direct effects on translation through RNA base oxidation, oxidative stress can indirectly affect RNA metabolism through affecting RNA binding proteins

(RBP) aggregation116. Specifically, oxidative stress can affect proteins like TDP-43 and SOD1 in multiple ways97,116,117. TDP-43 crosslinking can be induced through cysteine oxidation and disulfide bond formation leading to reduced solubility117. It further induces TDP-43 acetylation, impairing its binding to RNA and enhancing its aggregation propensity which contributes to ALS pathophysiology117. SOD1 is another protein which is known to misfold under oxidative stress in addition to the damage it receives through selectively oxidized mRNA in ALS97.

Figure 1.3 Common oxidative base modifications in DNA and RNA.

Many different types of oxidatively altered bases have been recorded in both DNA and RNA

(Figure 1.3)118. Due to its favorable standard reduction potential, guanine is the most oxidizable

20 base both in DNA and in RNA where 8-hydroxyguanine (8-OHG) is found to be most abundant oxidative base modification 118. The other most common modifications are summarized in Figure

1.3118.

Figure 1.4 Mutagenic nature of 8-OHG. Top: Formation of 8-OHG and bottom: 8-OHG can non- canonically base pair with adenine along with its canonical partner cytosine resulting in the mutations. This non-canonical base pairing is responsible genetic transversion from G.C to T.A.

21

The most common RNA base modification 8-OHG is the most deleterious base modification since it can potentially alter genetic information by incorrectly pairing with adenine at similar or higher efficiency than with cytosine in both DNA and RNA to produce mutations at both the genomic and transcriptional levels 118,119. A crystallographic study has revealed that 8-hydroxy guanine in

DNA induces an inversion of the mismatch recognition mechanisms that normally proofread DNA, such that the 8-hydroxyguanine-adenine mismatch (8-OHG:A) mimics a cognate base pair whereas the 8-hydroxyguanine-cytosine (8-OHG:C) base pair behaves as a mismatch. These studies reveal a fundamental mechanism of error-prone replication and show how 8-OHG in DNA can form mismatches that evade polymerase error-detection mechanisms, potentially leading to the stable incorporation of lethal mutations119. Quite interestingly, 8-OHG is more readily oxidized than guanine because of its lower oxidation potential, and thus various oxidative lesions are produced by the oxidation of 8-OHG (Figure 1.3). The formation of 2,5-Diamino-4H-imidazol-4- one (Iz) and 2,2,4-triamino-5(2H)-oxazolone (Oz) in DNA could account for the G.C to C.G transversion observed in gene mutations112. The presence of Oz correlates with the presence of methyl cytosine in DNA indicating more important biological roles of oxidized guanine and its oxidation derivatives112. Therefore, it is necessary to study both 8-oxoG and other oxidized guanine in lesions of different pathological and physiological contexts in order to understand the various phenomena caused by guanine oxidation in both DNA and RNA.

Oxidative damage can potentially alter RNA structure and function and also can interfere with the interaction between RNA and other cellular molecules. Oxidation of mRNA leads to reduced translation efficiency and abnormal protein production and also leads to ribosome dysfunction109.

In addition to reduced protein production, RNA oxidation is speculated to contribute to the process of neurodegeneration via production of misfolded and/or truncated proteins leading towards

22 protein aggregation109. In addition to contribute in the mechanism of neurodegeneration, the presence of 8-OHG is speculated to contribute in cancer progression. There are evidences wherein involvement of 8-OHdG can lead to cell cycle arrest in cancer suggesting the possible connection of RNA oxidation in cancer development120.

Figure 1.5 The proposed role of RNA oxidation in the pathogenesis of neurological disorders and cancer.

A recent study has shown that the fate of oxidized mRNA is determined by ribosomal machinery where researchers have shown that the presence of 8-OHG in the start codon of mRNA stalls the ribosome, slowing down the translational process121. In a similar context, it has been demonstrated that under H2O2 stress, the ribosomal catalytic center is damaged thereby affecting the translational machinery122. The reduction of protein production can also be attested to mRNA oxidation as oxidized mRNAs are highly associated with polysomes in cellulo causing the lack of availability for mRNA correction/degradation machinery123. A recent finding showed that RNA oxidation results in the poor substrate recognition by a riboswitch hence debilitating the riboswitch

23

124 function . The study also demonstrates that 8-OHG can reduce the melting temperature (Tm) of certain RNA duplex or RNA-DNA duplex by 10 C per lesion124. There are evidences suggesting miRNA oxidation affects miRNA mediated gene regulation, for example, oxidation of miRNA-

184 makes it recognize the non-native target mRNAs, Bcl-XL and Bcl-w 125.

1.8.5 Coping with RNA oxidative damage

Ribonuclease (RNase) enzymes play a crucial role in RNA integrity and quality control126. It has been known for decades that the damaged RNA molecules can be removed through degradation by these enzymes, but the selective degradation activity for oxidized RNA has not yet been established126. Cells form cytoplasmic mRNA processing bodies (P-bodies), the site of active degradation of mRNA, under oxidative stress127. Recent findings have highlighted the formation of riboprotein complexes called stress granules (S-granules) under cellular stress conditions128,129.

In most of the cases, however, it appears more stable RNAs consisting of rRNA and tRNA are more commonly protected in such granules along with translating mRNA128,129. In contrast to mRNAs with rapid turnover, stable RNAs, consisting primarily of rRNAs and tRNAs encompass

90% of total cellular RNA, may be protected against RNase action by tertiary structure, assembly into ribonucleoprotein complex or even blocking the RNA’s 3´ terminus in stress granules129.

Damaged RNA molecules were speculated to undergo degradation rather than being repaired. A recent finding suggested the existence of at least one specific mechanism to repair RNA damage, indicating that cells may have a greater investment in the protection of RNA than previously thought130. Alkylation in RNA is repaired by the same mechanism as a DNA repair, catalyzed in the E. coli by the enzyme AlkB and in humans by hABH3130. Alk B and its homologue hABH3 cause hydroxylation of the methyl group on damaged DNA and RNA bases which can directly

24 reverse alkylation damage130. In addition to damage reversal, DNA damage can be repaired by a base excision repair mechanism, but it remains unknown whether RNA damage can be repaired by base excision34,130,131.

Cells have mechanisms of dealing with nucleotide damage other than direct repair and excision repair, which seems to be useful for defense against oxidative damage to both DNA and

RNA132,133. Because oxidation of nucleotides can occur in the cellular nucleotide pool and the oxidized nucleotide can be incorporated into DNA and RNA, the mechanism of avoiding such incorporation of the oxidized nucleotide is dependent on coping with nucleic acid damage132. MutT protein in E. coli and its mammalian homologues MutT homologue 1 (MTH1) and Nudix type 5

(NUDT5) proteins participate in this error-avoiding mechanism by hydrolyzing the oxidized nucleoside diphosphates and/or triphosphates to the monophosphates134. It has been reported that the erroneous protein production by oxidative damage is at least 20-fold higher than in the wild- type cells E. coli mutT-deficient cells, which is reduced to 1.2- or 1.4-fold by the expression of

MTH1 or NUDT5, respectively134. Correspondingly, MTH1 deficiency augments the RNA oxidation induced by kainic acid treatment in MTH1-null mouse134. In addition to the degradative activity of MTH1 and NUDT5, several enzymes involved in nucleotide metabolism show a discriminatory activity against the oxidized nucleotides. Guanylate kinase (GK), an enzyme that converts GMP to GDP, is inactive on 8-OH-GMP, while nucleotide diphosphate kinase (NDK), an enzyme that converts GDP to GTP, fails to show such discriminating function106. Similarly, ribonucleotide reductase (RNR), an enzyme that catalyzes reduction of four naturally occurring ribonucleoside diphosphates, is inactive on conversion of 8-OH-GDP to 8-OH-dGDP, which avoids incorporation of the oxidized nucleotide into DNA synthesis. RNA polymerases could act

25 as final checkpoint that incorporates 8-OH-GTP into RNA at a much lower rate compared to the normal GTP incorporation135.

It is not very clear whether cells have machinery to deal with oxidatively damaged nucleotides that are contained in RNA, because RNA can be directly oxidized even if the incorporation of oxidized nucleotides into RNA is blocked. Recently, proteins that bind specifically to 8-OHG-containing

RNA have been reported, namely, E. coli polynucleotide phosphorylase (PNPase) protein and human PNPase, as well as human Y box-binding protein 1 (YB-1)106. The binding of the specific protein likely makes the 8-OHG-containing RNA resistant to nuclease degradation. However, it has been proposed that these proteins may recognize and discriminate the oxidized RNA molecule from normal ones, thus contributing to the fidelity of translation in cells by sequestrating the damaged RNA from the translational machinery. The human PNPase binds to 8-OHG containing

RNA preferentially and cellular amounts of human PNPase decrease rapidly by exposure to agents inducing oxidative stress, while amounts of other proteins in the cells do not change after the treatments136. Within the same study, it was demonstrated that overexpression of human PNPase reduces RNA oxidation and increases cell viability against the oxidative insult, while human

PNPase knockdown increases RNA oxidation and decreases cell viability136. Furthermore, human

YB-1 has been elucidated to be a component of both P-bodies and S-granules137. YB-1 is translocated from P-bodies to stress granules during oxidative stress, which suggests a dynamic link between P-bodies and stress granules under oxidative stress137. It is also possible that the RNA quality control mechanisms are defective or inefficient under neuropathological circumstances.

Messenger RNA (mRNA) containing 8-OHG in the start codons have shown to stall the ribosome and there is a suggestion that a no go decay (NGD) pathway has been established to clear these and other aberrant mRNAs. Simms et al have shown that the presence of 8-OHG in the start codon

26 inhibits tRNA selection with rates of peptide-bond formation that are three to four orders of magnitude slower than those measured on intact complexes121,126. The same study also showed that the effects of 8-OHG on the accuracy of tRNA selection appear to be minimal; the rates with near-cognate ternary complexes in the presence of 8-OHG are slightly faster (10-100 fold) relative to intact complexes. The presence of 8-OHG even in the wobble position has been shown to have a detrimental effect on the decoding process highlighting the severity of mRNA oxidation on the protein synthesis machinery114,119.

1.9 Hypotheses of the study

A. Firstly, we hypothesize that RNA oxidation in MS neurons could contribute in MS

neurodegeneration. The assumption is that the recent discoveries supporting a role within

neurodegenerative aspect of MS is true.

B. Second, we hypothesize that the identification of readily oxidized mRNA molecules in MS

neurons and their impact on respective protein synthesis, will allow us to decipher the role of

mRNA oxidation in MS pathogenesis.

C. Finally, we hypothesize that, in addition to the transcriptional challenge due to the presence of

oxidized nucleotides in the cells, the presence of oxidized bases in RNA could weaken the

function of the RNA molecule by affecting its cognate folded state. The synthesis of a novel 8-

OHG phosphorothioate, its incorporation in the RNA and functional clustering of these RNA

molecules followed by chemical-electrophoretic assays will enable us to map the impact of

individual 8-OHG modifications in RNA structure and function.

27

Figure 1.6 A hypothesis to decipher the role of selective RNA oxidation in MS neurodegeneration.

1.10 References

1 Peterson, L. K. & Fujinami, R. S. Inflammation, demyelination, neurodegeneration and

neuroprotection in the pathogenesis of multiple sclerosis. J Neuroimmunol 184, 37-44, (2007).

2 Baecher-Allan, C., Kaskow, B. J. & Weiner, H. L. Multiple Sclerosis: Mechanisms and

Immunotherapy. Neuron 97, 742-768, (2018).

3 Koudriavtseva, T. & Mainero, C. Neuroinflammation, neurodegeneration and regeneration in

multiple sclerosis: intercorrelated manifestations of the immune response. Neural Regen Res 11,

1727-1730, (2016).

4 Pittock, S. J. & Lucchinetti, C. F. The pathology of MS: new insights and potential clinical

applications. Neurologist 13, 45-56, (2007).

5 Pérez-Cerdá, F., Sánchez-Gómez, M. V., Matute, C. J., D. The link of inflammation and

neurodegeneration in progressive multiple sclerosis. M. S. & Disorders 1, 9, (2016).

6 Dilokthornsakul, P., Valuck, R. J., Nair, K. V., Corboy, J. R., Allen, R. R. & Campbell J. D..

Multiple sclerosis prevalence in the United States commercially insured population. Neurology

86, 1014-1021, (2016).

7 Wallin, M. T., The prevalence of multiple sclerosis in the United States: a population-based

healthcare database approach. Mult Scler J 23, 125-125 (2017).

28

8 Wu, G. F. & Alvarez, E. The immuno-pathophysiology of multiple sclerosis. Neurologic clinics

29, 257-278, (2011).

9 Ross, A. P., Halper, J. & Harris, C. J. Assessing Relapses and Response to Relapse Treatment in

Patients with Multiple Sclerosis: A Nursing Perspective. Intl J MS Care 14, 148-159, (2012).

10 Moll, C., Mourre, C., Lazdunski, M. & Ulrich, J. Increase of sodium channels in demyelinated

lesions of multiple sclerosis. Brain Res 556, 311-316, (1991).

11 Trapp, B. D. & Nave, K.-A. Multiple Sclerosis: An Immune or Neurodegenerative Disorder?

Annu Rev Neurosci 31, 247-269, (2008).

12 Fitzner, D. & Simons, M. Chronic Progressive Multiple Sclerosis – Pathogenesis of

Neurodegeneration and Therapeutic Strategies. Curr Neuropharma 8, 305-315, (2010).

13 Andersson, P. B., Waubant, E. E., Gee, L. L. & Goodkin, D. E. Multiple sclerosis that is

progressive from the time of onset: Clinical characteristics and progression of disability. Arch

Neurol 56, 1138-1142, (1999).

14 Tullman, M. J., Oshinsky, R. J., Lublin, F. D. & Cutter, G. R. Clinical characteristics of

progressive relapsing multiple sclerosis. Mult Scler J 10, 451-454, (2004).

15 Baranzini, S. E. Revealing the genetic basis of multiple sclerosis: Are we there yet? Curr Opin

Genet Dev 21, 317-324, (2011).

16 Lin, X., Deng, F.-Y., Lu, X. & Lei, S.-F. Susceptibility Genes for Multiple Sclerosis Identified in

a Gene-Based Genome-Wide Association Study. J Clin Neurol 11, 311-318, (2015).

17 Bhatia, R., Bali, P. & Chaudhari, R. M. Epidemiology and genetic aspects of multiple sclerosis in

India. Ann Indian Acad Neurol 18, S6-S10, (2015).

18 Oksenberg, J. R., Barcellos, L. F. & Hauser, S. L. Genetic aspects of multiple sclerosis. Semin

Neurol 19, 281-288, (1999).

19 Kular, L. et al. DNA methylation as a mediator of HLA-DRB1*15:01 and a protective variant in

multiple sclerosis. Nat Commun 9, 2397, (2018).

29

20 Baranzini, S. E. & Oksenberg, J. R. The Genetics of Multiple Sclerosis: From 0 to 200 in 50

Years. Trends Genet 33, 960-970, (2017).

21 Gough, S. C. L. & Simmonds, M. J. The HLA Region and Autoimmune Disease: Associations

and Mechanisms of Action. Curr Genom 8, 453-465, (2007).

22 Vyshkina, T., Leist, T. P., Shugart, Y. Y. & Kalman, B. CD45 (PTPRC) as a candidate gene in

multiple sclerosis. Mult Scler J 10, 614-617, (2004).

23 Majdinasab, N., Hosseini Behbahani, M., Galehdari, H. & Mohaghegh, M. Association of

interleukin 7 receptor gene polymorphism rs6897932 with multiple sclerosis patients in

Khuzestan. Iran J Neurol 13, 168-171 (2014).

24 Baranzini, S. E. et al. Pathway and network-based analysis of genome-wide association studies in

multiple sclerosis. Hum Mol Genet 18, 2078-2090, (2009).

25 Zhang, J.-M. & An, J. Cytokines, Inflammation and Pain. Int Anesthesiol Clin 45, 27-37, (2007).

26 Riedhammer, C. & Weissert, R. Antigen Presentation, Autoantigens, and Immune Regulation in

Multiple Sclerosis and Other Autoimmune Diseases. Front Immunol 6, 322, (2015).

27 Hernández-Pedro, N. Y., Espinosa-Ramirez, G., de la Cruz, V. P., Pineda, B. & Sotelo, J. Initial

Immunopathogenesis of Multiple Sclerosis: Innate Immune Response. Clinical and Devel

Immunol 2013, 413465, (2013).

28 Walker, D. G. & Lue, L.-F. Immune phenotypes of microglia in human neurodegenerative

disease: challenges to detecting microglial polarization in human brains. Alzheimers Res Ther 7,

56, (2015).

29 Kaushansky, N., Eisenstein, M., Zilkha-Falb, R. & Ben-Nun, A. The myelin-associated

oligodendrocytic basic protein (MOBP) as a relevant primary target autoantigen in multiple

sclerosis. Autoimmun Reviews 9, 233-236, (2010).

30 Dutta, R. & Trapp, B. D. Mechanisms of Neuronal Dysfunction and Degeneration in Multiple

Sclerosis. Prog Neurobiol 93, 1-12, (2011).

30

31 Mao, P. & Reddy, P. H. Is multiple sclerosis a mitochondrial disease? Biochim biophys acta

1802, 66-79, (2010).

32 Martinez, G. J., Nurieva, R. I., Yang, X. O. & Dong, C. Regulation and Function of

Proinflammatory TH17 Cells. Ann N Y Acad Sci 1143, 188-211, (2008).

33 Goverman, J. Autoimmune T cell responses in the central nervous system. Nat Rev Immunol 9,

393, (2009).

34 Høglund, R. A. & Maghazachi, A. A. Multiple sclerosis and the role of immune cells. World J

Exp Med 4, 27-37, (2014).

35 Andrews, H. E., Nichols, P. P., Bates, D. & Turnbull, D. M. Mitochondrial dysfunction plays a

key role in progressive axonal loss in Multiple Sclerosis. Med Hypotheses 64, 669-677, (2005).

36 Krishnamoorthy, G. et al. Myelin-specific T cells also recognize neuronal autoantigen in a

transgenic mouse model of multiple sclerosis. Nat Med 15, 626, (2009).

37 Pöllinger, B. et al. Spontaneous relapsing-remitting EAE in the SJL/J mouse: MOG-reactive

transgenic T cells recruit endogenous MOG-specific B cells. J Exp Med 206, 1303-1316, (2009).

38 Dutta, R., McDonough, J., Yin, X., Peterson, J., Chang, A., Torres, T., Gudz, T., Macklin, W. B.,

Lewis, D. A., Fox, R. J., Rudick, R., Mirnics, K. & Trapp, B. D.. Mitochondrial dysfunction as a

cause of axonal degeneration in multiple sclerosis patients. Ann Neurol 59, 478-489 (2006).

39 van Horssen, J., Witte, M. E. & Ciccarelli, O. The role of mitochondria in axonal degeneration

and tissue repair in MS. Mult Scler J 18, 1058-1067, (2012).

40 Campbell, G. R. & Mahad, D. J. Mitochondria as Crucial Players in Demyelinated Axons:

Lessons from Neuropathology and Experimental Demyelination. Autoimmune Dis 2011, 262847,

(2011).

41 Mao, P. & Reddy, P. H. Is multiple sclerosis a mitochondrial disease? Biochim Biophys Acta

1802, 66-79, (2010).

31

42 Campbell, G. R., Worrall, J. T. & Mahad, D. J. The central role of mitochondria in axonal

degeneration in multiple sclerosis. Mult Scler J 20, 1806-1813, (2014).

43 Mahad, D. J., Ziabreva, I., Campbell, G., Lax, N., White, K., Hanson, P. S., Lassmann, H. &

Turnbull, D. M.. Mitochondrial changes within axons in multiple sclerosis. Brain 132, 1161-

1174, (2009).

44 Regenold, W. T., Phatak, P., Makley, M. J., Stone, R. D. & Kling, M. A. Cerebrospinal fluid

evidence of increased extra-mitochondrial glucose metabolism implicates mitochondrial

dysfunction in multiple sclerosis disease progression. J Neurol Sci 275, 106-112, (2008).

45 Pinto, M. & Moraes, C. T. Mitochondrial genome changes and neurodegenerative diseases.

Biochim Biophys Acta 1842, 1198-1207, (2014).

46 Vianello, A., Casolo, V., Petrussa, E., Peresson, C., Patui, S., Bertolini, A., Passamonti, S.,

Braidot, E. & Zancani M.. The mitochondrial permeability transition pore (PTP)-An example of

multiple molecular exaptation? Biochim Biophys Acta 1817, 2072-2086, (2012).

47 Forte, M., Gold, B. G., Marracci, G., Chaudhary, P., Basso, E., Johnsen, D., Yu, X., Fowlkes, J.,

Rahder, M., Stem, K., Bernardi, P. & Bourdette, D. Cyclophilin D inactivation protects axons in

experimental autoimmune encephalomyelitis, an animal model of multiple sclerosis. PNAS USA

104, 7558-7563, (2007).

48 Aboul-Enein, F., Krssak, M., Hoftberger, R., Prayer, D. & Kristoferitsch, W. Reduced NAA-

levels in the NAWM of patients with MS is a feature of progression. A study with quantitative

magnetic resonance spectroscopy at 3 Tesla. PloS one 5, e11625 (2010).

49 Moffett, J. R., Ross, B., Arun, P., Madhavarao, C. N. & Namboodiri, A. M. A. N-Acetylaspartate

in the CNS: from neurodiagnostics to neurobiology. Prog Neurobiol 81, 89-131 (2007).

50 Uttara, B., Singh, A. V., Zamboni, P. & Mahajan, R. T. Oxidative Stress and Neurodegenerative

Diseases: A Review of Upstream and Downstream Antioxidant Therapeutic Options. Curr

Neuropharmacol 7, 65-74, (2009).

32

51 Phaniendra, A., Jestadi, D. B. & Periyasamy, L. Free Radicals: Properties, Sources, Targets, and

Their Implication in Various Diseases. Indian J Clin Biochem 30, 11-26, (2015).

52 Kim, G. H., Kim, J. E., Rhie, S. J. & Yoon, S. The Role of Oxidative Stress in Neurodegenerative

Diseases. Exp Neurobiol 24, 325-340, (2015).

53 Maraldi, T. Natural Compounds as Modulators of NADPH Oxidases. Oxid Med Cell Longev

2013, 271602, (2013).

54 da Fonseca, A. C. C. et al. The impact of microglial activation on blood-brain barrier in brain

diseases. Front Cell Neurosci 8, 362, (2014).

55 Labunskyy, V. M. & Gladyshev, V. N. Role of reactive oxygen species-mediated signaling in

aging. Antiox Red Signal 19, 1362-1372, (2013).

56 Nita, M. & Grzybowski, A. The role of the reactive oxygen species and oxidative stress in the

pathomechanism of the age-related ocular diseases and other pathologies of the anterior and

posterior eye segments in adults. Oxid Med Cell Long 2016, 3164734, (2016).

57 Glass, C. K., Saijo, K., Winner, B., Marchetto, M. C. & Gage, F. H. Mechanisms underlying

inflammation in neurodegeneration. Cell 140, 918-934, (2010).

58 Chen, W.-W., Zhang, X. I. A. & Huang, W.-J. Role of neuroinflammation in neurodegenerative

diseases. Mol Med Reports 13, 3391-3396, (2016).

59 Kovacic, P. & Somanathan, R. Redox processes in neurodegenerative disease involving reactive

oxygen species. Curr Neuropharmacol 10, 289-302, (2012).

60 Liu, Z., Zhou, T., Ziegler, A. C., Dimitrion, P. & Zuo, L. Oxidative stress in neurodegenerative

diseases: from molecular mechanisms to clinical applications. Oxid Med Cell Long 2017, (2017).

61 Gilgun-Sherki, Y., Melamed, E. & Offen, D. The role of oxidative stress in the pathogenesis of

multiple sclerosis: The need for effective antioxidant therapy. J Neurol 251, 261-268, (2004).

62 Haines, J. D., Inglese, M. & Casaccia, P. Axonal damage in multiple sclerosis. Mt Sinai J Med

78, 231-243, (2011).

33

63 Bradl, M. & Lassmann, H. Oligodendrocytes: biology and pathology. Acta Neuropathol 119, 37-

53, (2010).

64 Cadenas, E. & Davies, K. J. A. Mitochondrial free radical generation, oxidative stress, and aging.

Free Rad Biol Med 29, 222-230, (2000).

65 Imlay, J. A. Cellular defenses against superoxide and hydrogen peroxide. Ann rev biochem 77,

755-776, (2008).

66 Fukai, T. & Ushio-Fukai, M. Superoxide dismutases: Role in redox signaling, vascular function,

and diseases. Antioxid Red Signal 15, 1583-1606, (2011).

67 Bortolato, M., Chen, K. & Shih, J. C. Monoamine oxidase inactivation: From pathophysiology to

therapeutics. Adv drug deliv reviews 60, 1527-1533, (2008).

68 Merry, T. L. & Ristow, M. Nuclear factor erythroid‐ derived 2‐ like 2 mediates exercise‐

induced mitochondrial biogenesis and the anti‐ oxidant response in mice. J Physiol 594, 5195-

5207, (2016).

69 Piantadosi, C. A. et al. Heme oxygenase-1 couples activation of mitochondrial biogenesis to anti-

inflammatory cytokine expression. J Biol Chem 286, 16374-16385, (2011).

70 Sheykhansari, S. et al. Redox metals homeostasis in multiple sclerosis and amyotrophic lateral

sclerosis: a review. Cell Death Dis 9, 348-348 (2018).

71 Ward, R. J., Zucca, F. A., Duyn, J. H., Crichton, R. R. & Zecca, L. The role of iron in brain

ageing and neurodegenerative disorders. The Lancet Neurol 13, 1045-1060 (2014).

72 Pantopoulos, K., Porwal, S. K., Tartakoff, A. & Devireddy, L. Mechanisms of mammalian iron

homeostasis. Biochem 51, 5705-5724 (2012).

73 Sian-Hülsmann, J., Mandel, S., Youdim, M. B. H. & Riederer, P. The relevance of iron in the

pathogenesis of Parkinson’s disease. J Neurochem 118, 939-957 (2011).

74 Gomes-Leal, W. Microglial physiopathology: how to explain the dual role of microglia after

acute neural disorders? Brain Behav 2, 345-356 (2012).

34

75 Schonberg, D. L. & McTigue, D. M. Iron is essential for oligodendrocyte genesis following

intraspinal macrophage activation. Exp neurol 218, 64-74, (2009).

76 Haider, L. et al. The topograpy of demyelination and neurodegeneration in the multiple sclerosis

brain. Brain 139, 807-815, (2016).

77 Norkute, A., Hieble, A., Braun, A., Johann, S., Clarner, T., Baumgartner, W., Beyer, C. & Kipp

M. Cuprizone treatment induces demyelination and astrocytosis in the mouse hippocampus. J

Neurosci Res 87, 1343-1355, (2009).

78 Torkildsen, Ø., Brunborg, L. A., Myhr, K. M. & Bø, L. The cuprizone model for demyelination.

Acta Neurol Scand 117, 72-76, (2008).

79 Chen, L., Deng, H., Cui, H., Fang, J., Zuo, Z., Deng, J., Li, Y., Wang, X. & Zhao, L.

Inflammatory responses and inflammation-associated diseases in organs. Oncotarget 9, 7204-

7218, (2018).

80 Ramesh, G., MacLean, A. G. & Philipp, M. T. Cytokines and chemokines at the crossroads of

neuroinflammation, neurodegeneration, and neuropathic pain. Mediators Inflamm 2013, 480739

(2013).

81 Alfonso-Loeches, S., Ureña-Peralta, J. R., Morillo-Bargues, M. J., Oliver-De La Cruz, J. &

Guerri, C. Role of mitochondria ROS generation in ethanol-induced NLRP3 inflammasome

activation and cell death in astroglial cells. Front Cell Neurosci 8, 1-18 (2014).

82 Valdearcos, M. et al. Microglial inflammatory signaling orchestrates the hypothalamic immune

response to dietary excess and mediates obesity susceptibility. Cell Metab 26, 185-197 (2017).

83 Srinivasan, K. et al. Untangling the brain’s neuroinflammatory and neurodegenerative

transcriptional responses. Nat Commun 7, 11295 (2016).

84 Hsieh, H.-L. & Yang, C.-M. Role of redox signaling in neuroinflammation and neurodegenerative

diseases. Biomed Res Intl 2013, 484613 (2013).

35

85 Sultana, R., Perluigi, M. & Butterfield, D. A. Lipid peroxidation triggers neurodegeneration: A

redox proteomics view into the alzheimer disease brain. Free rad biol med 62, 157-169 (2013).

86 Lobo, V., Patil, A., Phatak, A. & Chandra, N. Free radicals, antioxidants and functional foods:

Impact on human health. Pharmaco Rev 4, 118-126 (2010).

87 Yin, H., Xu, L. & Porter, N. A. Free radical lipid peroxidation: mechanisms and analysis. Chem

Rev 111, 5944-5972 (2011).

88 Arimon, M. et al. Oxidative stress and lipid peroxidation are upstream of amyloid pathology.

Neurobiol dis 84, 109-119 (2015).

89 Shichiri, M. The role of lipid peroxidation in neurological disorders. J Clin Biochem Nutr 54,

151-160 (2014).

90 Shamoto-Nagai, M., Hisaka, S., Naoi, M. & Maruyama, W. Modification of α-synuclein by lipid

peroxidation products derived from polyunsaturated fatty acids promotes toxic oligomerization:

its relevance to Parkinson disease. J Clin Biochem Nutr 62, 207-212, 18-25 (2018).

91 Haider, L., Fischer, M. T., Frischer, J. M., Bauer, J., Höftberger, R., Botond, G., Esterbauer, H.,

Binder, C. J., Witztum, J. L. & Lassmann, H. Oxidative damage in multiple sclerosis lesions.

Brain 134, 1914-1924 (2011).

92 Bulteau, A.-L., Szweda, L. I. & Friguet, B. Mitochondrial protein oxidation and degradation in

response to oxidative stress and aging. Exp Gerontol 41, 653-657 (2006).

93 Smith, C. D., Carney, J. M., Starke-Reed, P. E., Oliver, C. N., Stadtman, E. R., Floyd, R. A. &

Markesbery, W. R., Excess brain protein oxidation and enzyme dysfunction in normal aging and

in Alzheimer disease. PNAS USA 88, 10540-10543 (1991).

94 Ahmed, N., Ahmed, U., Thornalley, P. J., Hager, K., Fleischer, G. & Münch, G. Protein

glycation, oxidation and nitration adduct residues and free adducts of cerebrospinal fluid in

Alzheimer's disease and link to cognitive impairment. 92, 255-263 (2005).

36

95 Qi, X., Lewin, A. S., Sun, L., Hauswirth, W. W. & Guy, J. Mitochondrial protein nitration primes

neurodegeneration in experimental autoimmune encephalomyelitis. J Biol Chem 281, 31950-

31962 (2006).

96 Ara, J., Przedborski, S., Naini, A. B., Jackson-Lewis, V., Trifiletti, R. R., Horwitz, J. &

Ischiropoulos, H. Inactivation of tyrosine hydroxylase by nitration following exposure to

peroxynitrite and 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine. PNAS USA 95, 7659-7663

(1998).

97 Martins, D. & English, A. M. SOD1 oxidation and formation of soluble aggregates in

yeast: Relevance to sporadic ALS development. Redox Biology 2, 632-639 (2014).

98 Grune, T., Shringarpure, R., Sitte, N. & Davies, K. Age-related changes in protein oxidation and

proteolysis in mammalian cells. J Gerontol 56, B459-B467 (2001).

99 Chen, X., Guo, C. & Kong, J. Oxidative stress in neurodegenerative diseases. Neural Regen Res

7, 376-385 (2012).

100 Fleming, A. M., Ding, Y. & Burrows, C. J. Oxidative DNA damage is epigenetic by regulating

gene transcription via base excision repair. PNAS USA 114, 2604-2609 (2017).

101 Kong, Q. & Lin, C. G. Oxidative damage to RNA: mechanisms, consequences, and diseases. Cell

mol life sci 67, 1817-1829 (2010).

102 Ding, Q., Markesbery, W. R., Chen, Q., Li, F. & Keller, J. N. Ribosome dysfunction is an early

event in Alzheimer's disease. J Neurosci 25, 9171-9175 (2005).

103 Fimognari, C. Role of oxidative RNA damage in chronic-degenerative diseases. Oxid med cell

long 2015, 358713 (2015).

104 Loeffler, D. A., Klaver, A. C., Coffey, M. P., Aasly, J. O. & LeWitt, P. A. Increased oxidative

stress markers in cerebrospinal fluid from healthy subjects with Parkinson’s disease-associated

LRRK2 gene mutations. Front Aging Neurosci 9, 89 (2017).

105 Kikuchi, A. et al. Systemic increase of oxidative nucleic acid damage in Parkinson's disease and

multiple system atrophy. Neurobiol Dis 9, 244-248 (2002).

37

106 Castellani, R. J., Nunomura, A., Rolston, R. K., Moreira. P. I., Takeda, A., Perry, G. & Smith, M.

A. Sublethal RNA oxidation as a mechanism for neurodegenerative disease. Intl J Mol Sci 9, 789-

806 (2008).

107 Niedzielska, E., Smaga, I., Gawlik, M., Moniczewski, A., Stankowicz, P., Pera, J. & Filip, M..

Oxidative stress in neurodegenerative diseases. Mol Neurobiol 53, 4094-4125 (2016).

108 Chang, Y. et al. Messenger RNA oxidation occurs early in disease pathogenesis and promotes

motor neuron degeneration in ALS. PLOS ONE 3, e2849 (2008).

109 Shan, X., Chang, Y. & Lin, C. G. Messenger RNA oxidation is an early event preceding cell

death and causes reduced protein expression. FASEB J 21, 2753-2764 (2007).

110 Baskakov, I. V. & Breydo, L. Converting the prion protein: What makes the protein infectious?

Biochim Biophys Acta 1772, 692-703 (2007).

111 Martinet, W., Meyer, G. R. Y. D., Herman, A. G. & Kockx, M. M. RNA damage in human

atherosclerosis: pathophysiological significance and implications for gene expression studies.

RNA Biol 2, 4-7 (2005).

112 Wang, W.-X. et al. Increased oxidative damage of RNA in early-stage nephropathy in db/db

Mice. Oxid Med Cell Long 2017, 2353729 (2017).

113 Li, Z., Wu, J. & Deleo, C. J. RNA damage and surveillance under oxidative stress. IUBMB Life

58, 581-588 (2006).

114 Wurtmann, E. J. & Wolin, S. L. RNA under attack: Cellular handling of RNA damage. Crit Rev

Biochem Mol Biol 44, 34-49 (2009).

115 Honda, K. et al. Ribosomal RNA in Alzheimer disease is oxidized by bound redox-active iron. J

Biol Chem 280, 20978-20986 (2005).

116 Chen, L. & Liu, B. Relationships between stress granules, oxidative stress, and neurodegenerative

diseases. Oxid med cell long 2017, 1809592 (2017).

38

117 Cohen, T. J., Hwang, A. W., Unger, T., Trojanowski, J. Q. & Lee, V. M. Y. Redox signalling

directly regulates TDP-43 via cysteine oxidation and disulphide cross-linking. EMBO J 31,

1241-1252 (2012).

118 Küpfer, P. A. & Leumann, C. J. Chemical Biology of Nucleic Acids: Fundamentals and Clinical

Applications (eds Volker A. Erdmann, Wojciech T. Markiewicz, & Jan Barciszewski, Springer

Berlin Heidelberg), 75-94 (2014).

119 Hsu, G. W., Ober, M., Carell, T. & Beese, L. S. Error-prone replication of oxidatively damaged

DNA by a high-fidelity DNA polymerase. Nature 431, 217 (2004).

120 Bellacosa, A. & Moss, E. G. RNA repair: damage control. Curr Biol 13, R482-R484 (2003).

121 Simms, C. L., Hudson, B. H., Mosior, J. W., Rangwala, A. S. & Zaher, H. S. An active role for

the ribosome in determining the fate of oxidized mRNA. Cell rep 9, 1256-1264, (2014).

122 Willi, J., Küpfer, P., Evéquoz, D., Fernandez, G., Katz, A., Leumann, C. & Polacek, N. Oxidative

stress damages rRNA inside the ribosome and differentially affects the catalytic center. Nucleic

Acids Res 46, 1945-1957 (2018).

123 Tanaka, M., Chock, P. B. & Stadtman, E. R. Oxidized messenger RNA induces translation errors.

PNAS USA 104, 66-71 (2007).

124 Choi, Y. J., Gibala, K. S., Ayele, T., Deventer, K. V. & Resendiz, Marino J. E. Biophysical

properties, thermal stability and functional impact of 8-oxo-7,8-dihydroguanine on

oligonucleotides of RNA- a study of duplex, hairpins and the aptamer for preQ1 as models.

Nucleic Acids Res 45, 2099-2111 (2017).

125 Wang, J.-X. et al. Oxidative modification of miR-184 enables it to target Bcl-xL and Bcl-w. Mol

Cell 59, 50-61 (2015).

126 Simms, C. L. & Zaher, H. S. Quality control of chemically damaged RNA. Cell Mol Life Sci 73,

3639-3653 (2016).

39

127 Borbolis, F. & Syntichaki, P. Cytoplasmic mRNA turnover and ageing. Mech Ageing Dev 152,

32-42 (2015).

128 Protter, D. S. W. & Parker, R. Principles and properties of stress granules. Trends Cell Biol 26,

668-679 (2016).

129 Anderson, P. & Kedersha, N. Stress granules. Curr Biol 19, R397-R398 (2009).

130 Drabløs, F. et al. Alkylation damage in DNA and RNA-repair mechanisms and medical

significance. DNA Repair 3, 1389-1407 (2004).

131 Liu, P. et al. Removal of oxidative DNA damage via FEN1-dependent long-patch base excision

repair in human cell mitochondria. Mol Cell Biol 28, 4975-4987 (2008).

132 Foti, J. J., Devadoss, B., Winkler, J. A., Collins, J. J. & Walker, G. C. Oxidation of the guanine

nucleotide pool underlies cell death by bactericidal antibiotics. Science 336, 315-319 (2012).

133 Li, Z., Malla, S., Shin, B. & Li, J. M. Battle against RNA oxidation: Molecular mechanisms for

reducing oxidized RNA to protect cells. WIRE RNA 5, 335-346 (2014).

134 Sekiguchi, T., Ito, R., Hayakawa, H. & Sekiguchi, M. Elimination and utilization of oxidized

guanine nucleotides in the synthesis of RNA and its precursors. J Biol Chem 288, 8128-8135,

(2013).

135 Tornaletti, S., Maeda, L. S., Kolodner, R. D. & Hanawalt, P. C. Effect of 8-oxoguanine on

transcription elongation by T7 RNA polymerase and mammalian RNA polymerase II. DNA

Repair 3, 483-494 (2004).

136 Hayakawa, H. & Sekiguchi, M. Human polynucleotide phosphorylase protein in response to

oxidative stress. Biochem 45, 6749-6755 (2006).

137 Reineke, L. C. & Lloyd, R. E. Diversion of stress granules and P-bodies during viral infection.

Virology 436, 255-267 (2013).

40

CHAPTER 2

"Adapted with permission from: P. Kharel, J. McDonough, and S. Basu, Neurochemistry

International, 92 (2016) 43-48. © Copyright (2016) Elsevier."

2. Evidence of extensive RNA oxidation in the neurons of normal appearing

cortex of multiple sclerosis brain

2.1 Introduction

Multiple sclerosis (MS) is the most common chronic demyelinating disease of the central nervous system (CNS)1. It is widely recognized as an immune mediated inflammatory disease of the CNS, which leads to demyelination resulting in axonal degeneration and hence neuronal miscommunication in the CNS1,2. Although the mechanism of axonal injury in MS is believed to be inflammation-led neurodegeneration, accumulating evidence supports the possibility of a reverse mechanism2,3. Multiple factors play their roles in the process of neurodegeneration, including oxidative stress (OS) and inflammation3. The cause of high oxidative stress in cells is either the inability of cellular defense mechanism to detoxify the reactive oxygen species (ROS),

·− · such as, superoxide (O2 ), hydrogen peroxide (H2O2) or hydroxyl radicals ( OH) or its failure to repair the resulting oxidative damage caused by the generation of excess ROS4,5. A disturbance in the physiological balance of the redox state of a cell thus produces toxic effects via the accumulation of the ROS that triggers damage in different biomolecules including proteins, lipids,

41 and nucleic acids6. Brain, being the most vibrant site of oxidative phosphorylation, produces more

ROS as a byproduct of energy metabolism5. Since MS is primarily an immune mediated inflammatory disorder, its neurodegenerative aspects have received less priority in the past7,8. Over the last decade, significant effort has been spent to investigate the contribution of oxidative stress and mitochondrial dysfunction in MS progression7,8. Multiple studies have shown a significant increase in cellular oxidative markers namely 8-hydroxy-2'-deoxyguanosine (8-OHdG) and 8- hydroxyguanosine (8-OHG) levels in the brains of patients with classical neurological disorders, such as, Alzheimer’s disease (AD), amyotrophic lateral sclerosis (ALS) and Parkinson’s disease

(PD)5,9,10. Some of the studies showed that most of the oxidized nucleosides are associated with cytoplasmic RNA and are restricted to vulnerable neurons10. Interestingly, there are reports showing the selective susceptibility of specific mRNA species toward oxidative damage; some of which have shown to be involved in the disease pathogenesis in AD, PD, and ALS11. Oxidized mRNAs not only produce truncated proteins but also, they cannot be translated properly leading to reduced protein expression and the loss of protein function11. Although there are some studies that showed DNA oxidation in MS lesions12, to the best of our knowledge, there is no known study describing RNA oxidative damage in normal appearing areas of MS brain. Since selective RNA oxidation can be a preceding event that contributes to neurodegeneration, we hypothesize that OS induced RNA oxidation is linked to MS pathology.

Using an in situ approach, we have detected hitherto unidentified RNA oxidative damage in the neuronal cells of normal appearing cortex of postmortem MS brains. We analyzed the presence of oxidative damage marker nucleoside 8-hydroxyguanosine (8-OHG) to determine the presence of oxidized RNA in MS brain. Immunohistochemical analyses with anti 8-OHG antibody showed significant oxidation in the cytoplasm and to a conspicuously lesser extent in the nucleus of

42 neuronal cells within the normal appearing cortex of MS brain, whereas similar areas were weakly immunopositive in control brain tissues. Pretreatment with RNase 1 greatly reduced the immune reaction with anti 8-OHG antibody while it was only slightly diminished by DNase 1 pre- treatment, indicating extensive oxidative damage in the RNA pool of MS brain. The abundance of

8-OHG, hence the high extent of RNA oxidative damage was further confirmed by immunoprecipitation and HPLC analyses of total RNA isolated from MS brain. To our knowledge, this is the first evidence of increased RNA oxidation in normal appearing cortex of MS brain. The current study begins to define the link of RNA oxidation to MS pathophysiology.

2.2 Materials and methods

2.2.1 Demographic details of the tissue donors

The details of the tissue donors’ demographic information are summarized in Table 1. Frozen tissue blocks from postmortem MS and non-MS brains were obtained from the Rocky Mountain

MS Center and the Human Brain and Spinal Fluid Resource Center at UCLA. The average age and average postmortem intervals (PMI) of the MS donors were 62.3 years (standard deviation SD

±10.5) and 8.68 h (SD ±6.6) respectively while those of non-MS donors were 66.8 (SD ±6.1) years and 10.5 h (SD ±5.8). The average age and PMI and sex were matched as close as possible to compare the results.

43

Table 2.1 Demographics of the donors used in the study

Tissue storage Donor ID Age Sex PMI Brain region Lesion status period (years)

MS1 67 F NA NA FC NAGM MS2 67 M 5.5 3 FC NAGM MS3 69 M 12.5 6 FC GML MS4 52 F 20.5 6 PC NAGM MS5 78 M 15.5 10 FC GML MS6 74 M 4 9 MC NAGM MS7 50 M 3.5 9 MC NAGM MS8 51 F 5 9 FC NAGM MS9 53 F 3 10 PC NAGM C1 63 F 21 4 FC C2 65 M 6.3 9 FC C3 68 M 10.3 6 FC C4 NA F NA 9 PC C5 73 F 12 5 MC C6 74 F 4.43 6 PC C7 58 M 9 10 FC

NA: Not available, F: female, M: male, FC: frontal cortex, PC: parietal cortex, MC: motor cortex,

NAGM: normal appearing gray matter, GML: gray matter lesion, PMI: Postmortem interval

2.2.2 Immunohistochemistry

2.2.2.1 DAB Staining: Frozen tissue blocks from postmortem MS and non-MS brains were fixed in 4% paraformaldehyde (PFA) in phosphate buffered saline (PBS, 137 mM NaCl, 2.7 mM KCl,

10 mM Na2HPO4 and 1.8 mM KH2PO4) for 48 h and washed in PBS (2×24 h). Then the tissue blocks were sliced 20 μm thick sections perpendicular to the outer edge of the cortex on a PBS media cooled Vibratome®. Tissue sections were stored in PBS in 24-well plates at 4 °C in a humidified chamber overnight. Following which the sections were washed with PBS (3×5 min).

44

Next, the TS were treated with 1% H2O2 to quench the endogenous peroxidase activity and washed with PBS (3×5 min). The sections were then incubated in 3% normal donkey serum in PBS for 30 min and washed with PBS (3×5 min), which was followed by incubation in a primary antibody to myelin proteolipid protein (PLP) (Chemicon, MAB388) in a 1:200 dilution with 3% donkey serum in PBS with 0.5% Triton X-100 as the diluent. Each section was covered with diluted antibody solution and incubated at 4˚C for 48 h. The sections were rinsed with PBS (3*10 min). A biotinylated anti-mouse secondary antibody was applied at a 1:500 dilution in PBS containing 3% donkey serum and 0.5% Triton X-100 for 1 h at room temperature. The sections were then rinsed with PBS again (3*10 min) and then incubated in freshly prepared Vector Elite ABC solution

(Vector labs) for 30 min. Then the TS were rinsed in PBS (3*10 min) and then peroxidase substrate solution, diaminobenzidine (DAB) was applied until the desired staining level was achieved. The sections were rinsed three times in PBS for 10 min each, blotted and allowed to air dry. Each slide was protected from UV radiation using Permount® under the coverslip and allowed to dry overnight in a hood. Images were acquired using an Olympus BX53 microscope using bright field imaging mode.

2.2.2.2 Immunofluorescence staining: Ethanol fixed TS were incubated in diluted primary anti

DNA and RNA oxidative damage marker antibody (8-OHG antibody, QED Biosciences) solution overnight at 4˚C, washed three times in PBS (10 min each) and were incubated in donkey anti- mouse Alexa 488 secondary fluor antibody (1:500, Invitrogen) in 3% normal donkey serum in

PBS with 0.5% Triton-X 100, for 2 h. For coimmunostaining, the TS were incubated in diluted primary antibody solutions (anti 8-OHG antibody and anti NeuN antibody (abcam) or anti ASPA antibody (EMD Millipore) overnight at 4˚C, washed three times in PBS (10 min each) and were incubated in donkey anti-rabbit Alexa fluor 488 secondary fluor antibody (1:500, abcam) and

45 donkey anti-mouse Alexa Fluor 555 secondary antibody (1:500, Invitrogen) in 3% normal donkey serum in PBS with 0.5% Triton-X 100, for 2 h. Following three 10 min washes in PBS, the secondary antibody treated sections were incubated in lipofuscin auto-fluorescence quenching solution composed up of 50 mM ammonium acetate and 10 mM cupric sulfate for 1.5 h. Sections were next washed three times in PBS (10 min each), placed below coverslips under

Vectashield mounting media, sealed with clear nail polish, and kept refrigerated until microscopically imaged. Images were acquired using an Olympus FV1000 confocal microscope equipped with two lasers (Ar 488 nm and HeNe 555 nm) and analyzed with ImageJ.

2.2.3 Nucleic acid isolation and immunoprecipitation

Nucleic acids were isolated from postmortem brain tissues using Tripure isolation reagent (Roche

Lifesciences) according to the manufacturer’s protocol and quantified using a NanoDrop®. RNA and DNA were immunoprecipitated separately with anti 8-OHG antibody by using some modifications in the previous protocol. (Shan et al., 2007) Briefly, 25 µg of total RNA (each from both MS and control brain tissue) were incubated with 30 µg of 8-OHG antibody at room temperature for 2 h. Then 35 µL of immobilized protein L agarose gel beads (Pierce) were added to the RNA-antibody mixture and incubated overnight at 4 ºC. The beads were washed three times

(3×5 min) with 200 µL of 0.04% (v/v) Nonidet P-40 (Roche Applied Science) solution in sterile

PBS. The oxidized RNA: antibody: protein L agarose beads complexes were separated from non- oxidized RNAs (which remained in the supernatant) by centrifugation at 1500 rpm for 5 min at 4

ºC. Non-oxidized RNAs were recovered after ethanol precipitation. The oxidized RNAs were mixed with following reagents: 3 mL of PBS with 0.04% Nonidet P-40, 300 µL of 10% (w/v) sodium dodecyl sulfate (SDS), and 3 mL of PCI (phenol: chloroform: isoamyl alcohol; 25:24:1), and the mixture was incubated at 37 °C for 30 min (with occasional vortexing) and separated to

46 an aqueous and an organic phase by spinning at 13,200 rpm for 15 min at 4 ºC. The aqueous layer containing oxidized RNA was separated and mixed with 40 µL of 3 M sodium acetate buffer (pH

5.3), 2 µL of 10 µg/µL glycogen and 1 mL of absolute ethanol. The sample was then frozen at -80

ºC for 1 h and centrifuged for 20 min at 13200 rpm at 4 ºC. The pellet was washed with 70% ethanol, vacuum dried and quantified with the NanoDrop (Thermo Scientific). DNA was also treated similarly; only difference being the selection of organic and interphase after PCI extraction before processing further.

2.2.4 Hydrolysis of RNA/ DNA and HPLC detection of 8-OHG/ 8-OHdG

Total RNA (50 μg) was treated with 10 μL of Nuclease P1 (US Biologicals, stock of 0.68 U/μL in

300 mM sodium acetate, 0.2 mM ZnCl2, pH 5.3) and 1 μL of calf intestinal alkaline phosphatase

(New England Biolabs, 10 U/μL), and the reaction mixture was incubated at 60 ºC for 1 h. The hydrolysate was filtered and placed into an autosampler vial for HPLC-ECD analysis on a Thermo

Scientific Ultimate 3000 system consisting of an ESA Model 582 pump set at 1 mL/min solvent flow. C-18 reversed-phase column (Agilent technologies) was used for the analysis. The DNA was also treated similarly before the HPLC analysis with relatively longer incubation time during enzymatic treatment (1.5 h). 8-Hydroxyguanosine (8-OHG) and 8-hydroxy-2ʹ-deoxyguanosine (8-

OHdG) were detected with an electrochemical detector (Coulochem III, ESA) with a PEEK filter- protected 5011A analytical cell (ESA, 5 nA; guard electrode, 205 mV; analytical electrode, 275 mV). Chromatograms were recorded using Chromeleon® software, which also controlled the pump, autosampler, and detector. The HPLC solvent consisted of 7% v/v acetonitrile in 50 mM sodium acetate buffer set to pH 5.3 with acetic acid. The HPLC solvents were vacuum degassed before use.

47

2.2.5 Poly-A RNA separation and gel analysis

Oligo-dT was attached to the agarose beads by adopting the previously published protocol with

13 some modifications . 1.5 nmol of (dT)24U oligo (Integrated DNA Technologies) was mixed in a

400 μL reaction mixture containing 0.1 M sodium acetate, pH 5.3, and freshly prepared 25 mM sodium m-periodate solution and the reaction mixture was incubated for 1 h in the dark at room temperature. The oxidized oligo was then ethanol precipitated and resuspended in 500 μL of 0.1

M sodium acetate, pH 5.3. 400 μL of adipic acid dihydrazide-agarose bead 50% slurry (Sigma

Aldrich) was washed four times in 10 mL of 0.1 M sodium acetate, pH 5.3, and pelleted after each wash at 300 rpm for 3 min at 4 °C. After the final wash, 0.5 mL of 0.1 M sodium acetate, pH 5.3, were added to the beads, and the slurry was then mixed with the periodate-treated oligo (dT)24U and agitated gently for 12 h at 4 °C. The beads with the bound oligo (dT)24U were then pelleted and washed three times in 1 mL of 2 M NaCl and three times in 1mL of buffer D (20 mM HEPES-

KOH, pH 7.6, 5% v/v glycerol, 0.1 M KCl, 0.2 mM EDTA, 0.5 mM dithiothreitol).

The binding efficiency of RNA to the beads was 39% as determined using 5′ 32P-end-labeled

(dT)24U. To label the oligo-dT(U), the sequence was incubated along with T4 polynucleotide kinase (PNK) in 1X PNK buffer in presence of γ-labelled ATP. The reaction was quenched with urea loading buffer in 45 min. The reaction mixture was purified in 12% denaturing gel and the labelled oligomer was eluted overnight in elution buffer (300 mM NaCl in TE buffer). The eluted oligo was sec-butanol concentrated and ethanol precipitated, and the label quantified before loading to the column.

The beads containing immobilized (dT)24U were incubated in a reaction mixture containing 40 µg of total RNA (isolated either from MS or non-MS postmortem brain tissues) and 400 μL of buffer

D for 1 h at 37 °C. Beads were then pelleted by centrifugation at 1000 rpm for 3 min at 4 °C and

48 washed four times with 1 mL of buffer D. The supernatant was chloroform extracted and ethanol precipitated to remove all other RNAs except poly-A RNAs. After the final centrifugation, the poly-A RNAs bound to the immobilized (dT)24U were eluted by denaturing with urea solution.

The beads were removed after the centrifugation at 5000 rpm for 5 min at 4 ºC and the eluent was ethanol precipitated to get the mRNA pellet. Poly-A RNA pool was then immunoprecipitated as mentioned above. The oxidized mRNA pool from both MS and control total RNA were 5ʹ end labelled with γ-32P labelled ATP after a 30 min phosphatase reaction followed by PNK reaction

(as described before). The reaction mixture was run into a 4 % denaturing polyacrylamide gel, the gel was exposed into a phosphoimaging screen and the gel image was taken with Typhoon FLA

9500 Phosphorimager (GE Healthcare Lifesciences).

2.3 Results

2.3.1 A higher level of 8-OHG immune reactivity in the normal appearing areas of MS brain

Before analyzing the presence of RNA oxidative damage in MS postmortem brain, we analyzed the MS gray matter for the presence and absence of MS lesions. Since the lesion biomolecules are already modified a lot, we intended to study RNA damage in normal appearing areas of MS brain.

To detect RNA oxidation, we used normal appearing gray matter from seven different MS and seven different non-MS (as controls) donor brain tissues for immuno-histochemical analyses. The tissues were age, sex and postmortem interval (PMI) matched as closely as possible. The non- lesion areas were identified by PLP staining (Figure 2.1, a representative image) and the adjacent tissue slices were used to study the nucleic acid oxidation in normal appearing MS brain. We also analyzed three MS tissue blocks, which showed the presence of some gray matter lesions (MS3,

49

MS4, and MS5). Tissue sections were treated with anti 8-OHG primary antibody and the immune reaction was detected by an Alexa® fluor-labeled secondary antibody by confocal microscopy.

From immunohistochemical analyses, we discovered that nucleic acids are extensively oxidized in normal appearing MS brain tissues compared to the age, sex and PMI matched non-MS brain tissues (Figure 2.2. a-c). We also detected a high level of nucleic acids oxidation in the analyzed lesion areas (data not shown here).

Figure 2.1. PLP staining to distinguish between the lesion and non-lesion areas in MS brain a. PLP staining of MS cortex showing a representative non-lesion area taken for the study unless otherwise mentioned, b. PLP staining of MS cortex showing a representative lesion area (**).

2.3.2 RNA oxidation is more abundant than DNA oxidation in MS brain

There are evidences of both DNA and RNA being oxidized in the brain of patients with neurological disorders like AD, PD and ALS. There are studies showing selective oxidation of

DNA or RNA in some instances. To distinguish between RNA and DNA oxidation in MS brain

50 we treated the TS with DNase 1 before the treatment with anti 8-OHG primary antibody. We observed only a slight decrease in fluorescence intensity (Figure 2.2.e); however, the fluorescence intensity almost disappeared when the TS were pre-treated with RNase 1 (Figure 2.2.f). These results clearly indicate that among the nucleic acids pool the cytoplasmic RNA undergoes significantly higher oxidation in the normal appearing cortex of MS brain. When analyzed in the higher resolution, the higher cytoplasmic stain was obvious (Figure 2.2.d) in MS brain cells. In combination, these results clearly establish that there is significantly more oxidation in the RNA pool of brain cells within the normal appearing MS cortex.

51

Figure 2.2 Immunohistochemical analyses of normal appearing cortex of postmortem brain tissue show high level of RNA oxidation in MS brain. Top panel: High level of RNA oxidation in postmortem MS brain (a) 20 μm thick tissue section from a non-MS (control) postmortem brain (motor cortex) was immunostained with anti 8-OHG monoclonal primary antibody; (b) and (c) Identical staining was performed on 20 μm thick TS from postmortem brain of two different MS regions (normal appearing motor and parietal cortices); (d) a part of (c) in a higher resolution to decipher the signals at the cellular level: it clearly shows (e) DNase 1 treated TS from MS postmortem brain prior to antibody incubation (f) RNase 1 treated TS from the MS postmortem brain prior to antibody incubation. In all cases primary antibody treated sections were incubated with fluorescence secondary antibody and the images were taken by Olympus FV 1000 confocal microscope at the same resolution (20 X) in each case except in d (60 X); Scale bar: 50 μm except in d (10 μm).

To quantitatively determine the RNA oxidation in MS brain, we isolated the total RNA from different MS and non-MS brains (n=5, in each case) and the same amount of the total RNA (25

μg) were treated with DNase 1 and immunoprecipitated in each case with anti 8-OHG antibody.

Quantification of the immunoprecipitated total RNA by UV spectroscopy showed at least 2-fold more oxidation in the RNA isolated from MS brain compared to the RNA isolated from the non-

MS pool (Figure 2.3.a), with a strong statistical significance (data represent the mean values ± standard error of mean, n = 5, p<0.0005 Student’s t-test). Similar analysis on DNA oxidation revealed lesser oxidation in the DNA pool (Fig 2.3.b, data represent the mean values ± standard error of the mean, n = 3, p<0.005 Student’s t-test). Although the extent of damage in the DNA pool was very less, the difference between two population was very significant. Isolated total DNA was treated with RNase 1 before proceeding for immunoprecipitation.

52

Figure 2.3 Comparative analysis of RNA and DNA oxidation in MS and non-MS brains. a. 25 μg of total RNA were isolated from age, sex and PMI matched normal appearing cortices of MS and non-MS donors’ postmortem brains (n=5, in each case). The oxidized RNAs were immunoprecipitated with anti 8-OHG antibody and quantified and b. 10 µg of DNA was isolated from age, sex and PMI matched normal appearing cortices of MS and non-MS donors’ postmortem brains (n=3, in each case). The oxidized DNAs were immunoprecipitated with anti 8-OHG antibody and quantified. Data represent mean values ± standard error of mean (s.e.m), * p<0.0005 Student’s t-test.

2.3.3 RNA oxidation is more abundant in neuronal cells is MS brain

In the classical neurological disorders like AD, PD, and ALS, it has been demonstrated that there is a higher level of RNA oxidative damage in neuronal cells5,11. Neuronal RNA damage is

53 considered as a causative agent to the progressive loss of neurons in such diseases. We also wanted to test the specificity of RNA oxidation in the neurons of postmortem MS brain. We analyzed the differential level of RNA oxidation in neuronal cells in MS brain by coimmunostaining (anti NeuN antibody as a neuronal marker along with anti 8-OHG antibody as an RNA oxidation marker) and observed that almost 90% of the neuronal cells have a prominent RNA oxidation.

Figure 2.4 RNA oxidation is predominant in neuronal cells. Coimmunostaining with anti NeuN antibody (ab177488, abcam®) and anti 8-OHG antibody clearly showed higher level of RNA oxidation in neuronal cells. Our data showed that almost 90% of neuronal cells are positive for RNA oxidation: white circles: neuronal cells that are positive for 8-OHG; yellow circles: neuronal cells, which have no (or very less) 8-OHG immune reactivity; and dark orange circles: non- neuronal cells with high 8-OHG reactivity.

Next, we asked ourselves whether the extensive RNA oxidation we observed in MS neurons is specific to the neurons or is prevalent in all cell types. As a representative of glial cells, this time we performed co-immunostaining of anti 8-OHG antibody with anti-aspartoacylase (ASPA) antibody, the latter is a marker of oligodendrocytes. Interestingly we observed about 30% of oligodendrocytes were positive for RNA oxidation. This observation demonstrates the RNA damage in MS brain is more prominent in the neurons.

54

Figure 2.5 RNA oxidation is less prominent in oligodendrocyte cells. Coimmunostaining with anti ASPA antibody (ABN1698, EMD Millipore®) and anti 8-OHG antibody showed that nearly 30% of oligodendrocyte cells are positive for RNA oxidation, white circles: oligodendrocytes that are not positive for 8-OHG; and yellow circles: oligodendrocytes which are positive for 8-OHG.

2.3.4 8-OHG is abundant in MS brain

In a separate assay, to look at the extent of oxidative damage on the most mutagenic oxidative modification 8-OHG, we took the same amount of total RNA (50 μg each) isolated from MS and non-MS postmortem brains and treated them with nuclease P1 and calf intestinal phosphatase to hydrolyze the RNA into nucleoside level. After cleavage we analyzed the presence of oxidized guanosine molecules by HPLC-ECD (Figure 2.6.a-b) and quantified them by comparing with a standard 8-OHG chromatogram. We detected 2-3 fold more 8-OHG in MS RNA in comparison to non-MS RNA isolated from age and PMI matched donor brains. Nevertheless, the quantification of 8-OHdG shows relatively more oxidation in the DNA pool in MS brain than in non-MS brain, it is less pronounced than the level of 8-OHG in RNA oxidation in the similar region (Fig 2.6c-d).

These results corroborate well with our observations made using the UV quantitation of the immunoprecipitated oxidized total RNA and the DNA (Figure 2.3).

55

2.3.5 The extent of RNA oxidation in lesion area is similar to that in normal appearing area in

MS brain

Once we confirmed the selective RNA oxidation in normal appearing areas in gray matter of MS brain, we aimed to analyze whether or not there is a difference in the level of RNA damage in the lesion areas. The reports in the AD and PD neurodegeneration have demonstrated that the RNA oxidation in these diseases is mainly involved in the early pathogenesis process as shown from the higher level of RNA oxidation in the early plaque forming areas5,11. When quantified after immunoprecipitation, the amount of oxidized RNA in the tissue block with GML was not significantly higher (Figure 2.7) than that in NAGM. Nonetheless, there are reports suggesting a higher level of molecular damage in lesion areas, our observation in RNA oxidative damage suggests that the extent of RNA damage is not higher in the lesion areas implying RNA oxidation could be acting only in earlier stage of disease progression as in case of AD, ALS, and PD.

56

Figure 2.6 HPLC analyses at nucleoside level to detect the relative amount of 8-OHG in RNA and DNA. a & b. Top panel show a presence of twofold more 8-OHG in the MS RNA. Total RNA was isolated from normal appearing cortex area of brain tissue and analyzed in HPLC-ECD after nuclease P1 and CIP hydrolyses. Left panel: RNA from MS brain. Right panel: RNA from non- MS brain. C & d. Bottom panel show there is slightly higher 8-OHG in MS DNA than in control DNA but overall 8-OHG distribution in DNA in comparison to that in RNA is lower.

57

Figure 2.7 Comparison between RNA oxidation in lesion vs non-lesion areas of MS brain. Although statistically non-significant, we observed a 9% increase in the level of RNA oxidation in MS lesion areas.

2.3.7 mRNA pool is highly oxidized in MS brain

Oxidation of specific mRNA molecules has been linked to the progression of classical neurological disorders5,11. Next, we wanted to investigate whether the coding transcriptome is selectively targeted by oxidative stress in MS brain. We analyzed the presence of oxidized mRNA in the MS isolated total RNA pool. We separated the oxidized poly-A RNA from the total RNA using oligo- dT agarose beads and immunoprecipitated with 8-OHG antibody to analyze the differential level of oxidized poly-A RNA in MS and non-MS brains by denaturing acrylamide gel electrophoresis of 5ʹ-end radiolabeled immunoprecipitated poly-A RNA. The data in Figure 2.8. show

58 considerably more labeled mRNA in the MS poly-A sample providing additional proof for more

RNA oxidation in MS brain sample.

Figure 2.8 mRNA molecules are highly oxidized in MS brain. As demonstrated in the schematic in the left panel (a), total RNA isolated from MS and non-MS brain was immunoprecipitated and enriched for mRNA using oligo-dT beads. The immunoprecipitated mRNA thus obtained was labelled with γ-32P-ATP and resolved in 4% denaturing polyacrylamide gel (b).

2.4 Discussion

This study shows at a molecular level that the neuronal cells in the normal appearing cortex of MS brains have more oxidized RNA than the sex, age, PMI matched non-MS brain tissues based upon

59 immunohistochemical detection and quantification of the major oxidized RNA product 8-OHG.

Unless otherwise mentioned, throughout the study we avoided the lesion containing tissue areas as far as possible. As a control, we did the immunoprecipitation of RNA from three tissue blocks that contained gray matter lesions adjacent to the tissue analyzed and found that the yield of RNA in the oxidized pool went up by nearly 9 % in comparison to that from tissue blocks without lesions

(Figure 2.7). The study of DNA oxidation in the NAGM of MS brain showed that though there is some oxidation in the DNA pool of MS NAGM, RNA oxidation is much more abundant (Figure

2.2 and 2.3). There are many studies on the oxidative damage of lipids, proteins, and on both nuclear and mitochondrial DNA in MS brain, not only in lesions but also in normal appearing white and gray matter14-18. However, to the best of our knowledge, there is no study on RNA oxidative damage and their possible impacts on MS pathophysiology. Our data indicate that oxidative stress induces damage to RNA in the neuronal cells of normal appearing areas of MS brain. Such damage can contribute to impaired energy metabolism and neuronal degeneration in

MS.

In this study we detected the presence of a higher level of oxidized guanosine product 8-OHG in post-mortem MS brain. By using co-immunohistochemistry, we have shown that there is more prominent RNA oxidation in neuronal cells and a relatively lesser extent of RNA damage in oligodendrocyte cells (Figure 2.4 and 2.5). The higher level of RNA oxidation in neurons could be related to a higher respiration rate in the neurons. Our analyses suggested a 2-3 fold increase in the mutagenic 8-OHG modification in the RNA isolated from MS non-lesion areas. Oxidized nucleotides can lead to non-Watson-Crick base pairing and some of the mismatched pairs are shown to be as strong as the canonical base pairs in some instances19. In a recent study, it has been shown that the incorporation of a single 8-OHG in the codon reduces the rate of peptide-bond

60 formation by more than threefold independent of its position suggesting that the oxidative base modification can stall the translational machinery resulting in low protein expression. Previous studies on the chemical consequences of 8-OHdG:A mismatch pairing and 8-OHdG:C pairing showed the possibility of perturbation in the local structure of double stranded DNA. Based on the presence of considerable single stranded structure in RNA, one can speculate that the impact of non-canonical base pairing with oxidized bases on overall RNA structure may be more pronounced and this structural change can impact the RNA function and might contribute to the disease progression.

The assessment of differential RNA oxidation in postmortem brain samples from MS and control brain may lead to a better understanding of OS induced damage in RNA in MS brain. Our results are compatible with the report of high levels of RNA oxidation found in the vulnerable neurons of classical neurodegenerative diseases like AD, PD and ALS It has been well established from in vitro studies that oxidized mRNAs result in the production of aberrant proteins resulting in protein aggregation which can contribute to neurodegeneration. In some cases, oxidation of non-random and specific mRNAs has been observed. The oxidation in specific sets of mRNA pool in MS brain is also possible, can have downstream effects in protein expression, and can contribute to disease progression in MS. Since neurodegeneration has also been proved an important contributing factor in MS pathogenesis, we reason that RNA oxidation can play a role in MS pathophysiology. In addition, there may be impacts of oxidative damage on non-coding RNA that might affect the target gene regulation and contribute towards the disease progression. For example, rRNA oxidation has been shown to be associated with neurodegeneration in AD. Owing to the presence of more than a two-fold increase in RNA oxidation in normal appearing cortex of MS brain and the established connection between RNA oxidation and disease progression in other

61 neurodegenerative diseases, we postulate that OS induced RNA oxidation contributes to MS pathogenesis by producing aberrant proteins or by dysregulating the target gene expression. The identification of the damaged RNA molecules and an investigation to understand their functional consequences would help in understanding the mechanism of MS neurodegeneration from a different window.

2.5 Conclusion

To the best of our knowledge, this is the first study deciphering the presence of extensive RNA oxidation in the normal appearing area of the MS brain. We found at least two-fold more oxidative damage in the total RNA pool and same level of production of the most mutagenic modification,

8-OHG in MS brain. In addition, we have shown a different level of RNA oxidation in different cell types of NAGM of MS brain. The present study begins to identify the possible contribution of

RNA oxidation in MS pathogenesis.

2.6 References

1 Fitzner, D. & Simons, M. Chronic progressive multiple cclerosis – Pathogenesis of

neurodegeneration and therapeutic strategies. Curr Neuropharmacol 8, 305-315 (2010).

2 Nakahara, J., Maeda, M., Aiso, S. & Suzuki, N. Current concepts in multiple sclerosis:

Autoimmunity versus oligodendrogliopathy. Clinl Rev Allergy Immunol 42, 26-34, (2012).

3 Glass, C. K., Saijo, K., Winner, B., Marchetto, M. C. & Gage, F. H. Mechanisms underlying

inflammation in neurodegeneration. Cell 140, 918-934 (2010).

62

4 Nita, M. & Grzybowski, A. The role of the reactive oxygen species and oxidative stress in the

pathomechanism of the age-related ocular diseases and other pathologies of the anterior and

posterior eye segments in adults. Oxid Med Cell Long 2016, 3164734 (2016).

5 Uttara, B., Singh, A. V., Zamboni, P. & Mahajan, R. T. Oxidative stress and neurodegenerative

diseases: A review of upstream and downstream antioxidant therapeutic options. Curr

Neuropharmacol 7, 65-74 (2009).

6 Lobo, V., Patil, A., Phatak, A. & Chandra, N. Free radicals, antioxidants and functional foods:

Impact on human health. Pharmacog Rev 4, 118-126 (2010).

7 Dutta, R., McDonough, J., Yin, X., Peterson, J., Chang, A., Torres, T., Gudz, T., Macklin, W. B.,

Lewis, D. A., Fox, R. J., Rudick, R., Mirnics, K. & Trapp, B. D.. Mitochondrial dysfunction as a

cause of axonal degeneration in multiple sclerosis patients. Ann Neurol 59, 478-489 (2006).

8 Bolaños, J. P., Almeida, A., Stewart, V., Peuchen, S., Land, J. M., Clark, J. B. & Heales, S. J.

Nitric oxide-mediated mitochondrial damage in the brain: mechanisms and implications for

neurodegenerative diseases. J Neurochem 68, 2227-2240 (1997).

9 Nunomura, A., Hofer, T., Moreira, P. I., Castellani, R. J., Smith, M. A. & Perry, G.. RNA

oxidation in Alzheimer disease and related neurodegenerative disorders. Acta Neuropathol 118,

151-166 (2009).

10 Poulsen, H. E. et al. RNA modifications by oxidation: a novel disease mechanism? Free Radic

Biol Med 52, 1353-1361 (2012).

11 Shan, X., Chang, Y. & Lin, C. G. Messenger RNA oxidation is an early event preceding cell

death and causes reduced protein expression. FASEB J 21, 2753-2764 (2007).

12 Haider, L. et al. Oxidative damage in multiple sclerosis lesions. Brain 134, 1914-1924 (2011).

13 Burrell, H. R. & Horowitz, J. Binding of ribosomal proteins to RNA covalently coupled to

agarose. Eur J Biochem 75, 533-544 (1977).

14 Oliveira, S. R. et al. Oxidative stress in multiple sclerosis patients in clinical remission:

Association with the expanded disability status scale. J Neurol Sci 321, 49-53, (2012).

63

15 van Horssen, J., Witte, M. E., Schreibelt, G. & de Vries, H. E. Radical changes in multiple

sclerosis pathogenesis. Biochim Biophys Acta 1812, 141-150 (2011).

16 Kutzelnigg, A. & Lassmann, H. Cortical demyelination in multiple sclerosis: A substrate for

cognitive deficits? J Neurol Sci 245, 123-126 (2006).

17 Bizzozero, O. A., DeJesus, G., Callahan, K. & Pastuszyn, A. Elevated protein carbonylation in

the brain white matter and gray matter of patients with multiple sclerosis. J Neurosci Res 81, 687-

695 (2005).

18 Gilgun-Sherki, Y., Melamed, E. & Offen, D. J. The role of oxidative stress in the pathogenesis of

multiple sclerosis: The need for effective antioxidant therapy. J Neurol 251, 261-268 (2004).

19 Delaney, S., Jarem, D. A., Volle, C. B. & Yennie, C. J. Chemical and biological consequences of

oxidatively damaged guanine in DNA. Free Rad Res 46, 420-441 (2012).

64

CHAPTER 3

3. Investigation of RNA oxidation in neuronal cells reveals selective oxidation of mRNAs could be linked to MS neuropathology

3.1 Introduction

Multiple sclerosis (MS) is the most common chronic demyelinating disease of the central nervous system (CNS). Multiple factors are supposed to contribute in the process of MS pathogenesis, including oxidative stress (OS), mitochondrial dysfunction, and inflammation1-3. Brain being the most active site of oxidative phosphorylation, produces more ROS as a byproduct of energy metabolism which eventually contribute towards the oxidative damage of biomolecules4,5. Recent studies have been focused on the investigation of DNA, RNA, and protein oxidative modifications in the MS brain1,2,6. Our recent investigation has shown a possibility of involvement of RNA oxidative damage in the progression of MS neurodegeneration where we have demonstrated a high level oxidative damage of RNA in the neurons of normal appearing areas of MS brain7.

Although accumulating evidences demonstrate the presence of a higher level of RNA damage in the neuronal cells in neurological disorders, the high-throughput identification of such damaged mRNA molecules and analysis of their downstream effect has not gotten sufficient attention in the field8-10. The new discoveries about the roles of p-bodies and stress granules are compelling us to think about the importance of RNA modifications more seriously and more thoroughly better than

65 ever before11,12. In general p-bodies are to degrade the bad mRNAs while the S-granules are to save the translating machinery both of which involve a complex interaction of different RNA and protein molecules12. Based on the evidences that oxidized RNA not only affects the translation10 but also is detrimental for substrate binding in vitro13, we speculate that the RNA oxidative modification present in mRNAs under neuropathological environment could have detrimental effect in neurobiology from multiple angles, for example, translational inhibition, production of bad proteins or disturbance in the formation of ribonucleoprotein (RNP) complexes which are essential to maintain smooth RNA turnover. This relation is going to be more complex and more important than was previously thought. With these important findings in cell biology, ‘oh it is just a damaged RNA’ is not a thing anymore.

Figure 3.1 A schematic of proposed path to delineate the link between selective RNA oxidation and MS neurodegeneration.

Oxidized mRNA has shown to stall ribosomal machinery and produce truncated proteins in different contexts10,14. The oxidative base modifications in mRNA could also impact the formation of ribonucleoprotein complexes (RNPs) like p-bodies and stress granules. Indeed, there are evidences of the instability of stress granules under H2O2 influence while sodium arsenate mediated oxidative stress induces the stress granules formation. The in cellulo experiments show that the oxidized mRNAs are as stable inside the cells as normal mRNAs15 indicating the

66 detrimental impact of oxidative base modifications in mRNA function. In sunflower plants selective oxidation of mRNAs has been shown to regulate seed dormancy16. In addition to mRNA, there is an evidence of selective oxidation of miRNAs which have shown to impact gene regulation by identifying the ‘off-targets’ as ‘targets’17. In the context of expanding RNA oxidation biology, it would be of tremendous importance to identify the oxidatively modified RNA molecules under neurodegenerative circumstances. This will allow us to understand MS and other neurological disorders from a novel mechanistic aspect of disease development. In addition, the finding of selectively oxidized mRNAs and interpretation of their involvement in disease pathways could hint about the possibility of conditional regulation of gene expression via RNA oxidative modification.

We hypothesize that the selective RNA damage in certain transcripts could lead towards an erroneous translation or gene (mys)regulation in MS. We aimed to identify the selectivity in the mRNA oxidative damage in human neurons under MS microenvironment and their possible link in disease progression. By identifying the oxidatively modified mRNA molecules under oxidative stress in the human neuronal cells (SH-SY5Y), we aimed to delineate the relationships between specific mRNA oxidation and neurodegeneration (Figure 3.1). We developed an MS microenvironment by exerting oxidative stress/inflammatory stress to human neuroblastoma (SH-

SY5Y) cells using sodium nitroprusside (SNP) as a stressing agent. Here we describe a transcriptome-wide assay to identify oxidized mRNA transcripts that relies on the affinity purification of oxidized mRNA on anti-8-OHG antibody-coupled agarose beads (IP+ve RNA) followed by high-throughput Illumina® sequencing (RNA-seq) of a cDNA library prepared from these mRNA. In parallel, we assessed total RNA abundance by construction and sequencing of

DNA libraries from RNA that had not undergone the affinity purification for 8-OHG (IP-ve RNA).

67

The data thus obtained was analyzed by using Qiagen Ingenuity® platform to get the selectively oxidized pool of mRNA that is altered during the SNP-induced oxidative stress mimicking the situation of the stress level in MS brain during inflammatory and oxidative stress situation. The selectively oxidized mRNA clusters were further analyzed by using PANTHER® pathway analysis platform.

We have discovered that many of the neuropathologically relevant mRNAs are among the pool of selectively oxidized mRNAs which mostly consists of mRNAs for nucleic acid binding proteins, transcription factors and many other important classes of mRNAs. We have also discovered that the level of RNA-oxidation does not perfectly correlate with the proximity towards the ROS sites nor does completely correlate with the mRNA abundance indicating a possibility of controlled regulatory role of RNA oxidation.

3.2 Materials and methods

3.2.1 Immunofluorescence staining

The 4% paraformaldehyde fixed SH-SY5Y cells were incubated in diluted primary anti-DNA and

RNA oxidative damage marker antibody (8-OHG antibody, QED Biosciences) solution overnight at 4 ˚C, washed three times in PBS (10 min each) and were incubated in donkey anti-mouse Alexa

488 secondary fluor antibody (1:250, Invitrogen) in 3% normal donkey serum in PBS with 0.5%

Triton-X 100, for 12 h. Following three 10 min washes in PBS, the sections were incubated in lipofuscin auto-fluorescence quenching solution composed of 50 mM ammonium acetate and

10 mM cupric sulfate for 30 min. Sections were next washed three times in PBS (10 min each), placed below coverslips under Vectashield® mounting media, sealed with clear nail polish, and

68 refrigerated for three days. Images were acquired using an Olympus FV1000 confocal microscope equipped with two lasers (Ar 488 nm and HeNe 555 nm) and analyzed with ImageJ.

3.2.2 RNA isolation and immunoprecipitation

RNA was isolated from SNP treated SH-SY5Y cells using Tripure isolation reagent (Roche

Lifesciences) according to the manufacturer’s protocol with slight modification and quantified using a NanoDrop®. Total RNA was immunoprecipitated with anti 8-OHG antibody. Briefly, 25

µg of total RNA (each from both MS and control brain tissue) were incubated with 30 µg of 8-

OHG antibody at room temperature for 2 h. Then 35 µL of immobilized protein L agarose gel beads (Pierce) were added to the RNA-antibody mixture and incubated overnight at 4 ºC. The beads were washed three times (3×5 min) with 200 µL of 0.04% (v/v) Nonidet P-40 (Roche

Applied Science) solution in sterile PBS. The oxidized RNA: antibody: protein L agarose beads complexes were separated from non-oxidized RNAs (which remained in the supernatant) by centrifugation at 1500 rpm for 5 min at 4 ºC. Non-oxidized RNAs were recovered after ethanol precipitation. The oxidized RNAs were mixed with following reagents: 3 mL of PBS with 0.04%

Nonidet P-40, 300 µL of 10% (w/v) sodium dodecyl sulfate (SDS), and 3 mL of PCI (phenol: chloroform: isoamyl alcohol; 25:24:1), and the mixture was incubated at 37 °C for 30 min (with occasional vortexing) and separated to an aqueous and an organic phase by spinning at 13,200 rpm for 15 min at 4 ºC. The aqueous layer containing oxidized RNA was separated and mixed with 40

µL of 3 M sodium acetate buffer (pH 5.3), 2 µL of 10 µg/µL glycogen and 1 mL of absolute ethanol. The sample was then frozen at -80 ºC for 1 h and centrifuged for 20 min at 13200 rpm at

4 ºC. The pellet was washed with 70% ethanol, vacuum dried and quantified with the NanoDrop

(Thermo Scientific).

69

3.2.3 Hydrolysis of RNA and HPLC detection of 8-OHG

As detailed in chapter 2.2.4.

3.2.4 Cell culture and SNP treatment

Human neuroblastoma SH-SY5Y cells, obtained from ATCC were cultured in Dulbecco’s modified Eagle’s medium supplemented with 10% (v/v) fetal bovine serum, and 0.01% penicillin/streptomycin at 37 °C under 5% CO2 in air. To determine the effect of SNP on retinoic acid differentiated SH-SY5Y cells, the cells were treated with various doses of SNP for 8 h. In a single experiment each treatment was performed in triplicate.

3.2.5 Cell viability assay

Cell viability was determined by MTS assay. SH-SY5Y cells were seeded in 96-well plates at density of 1 × 105 cells/well and incubated for 48 h prior to experimental treatments. The cells were incubated with or without SNP for 8 h. The cultured medium was removed, and the cell viability was measured by Promega assay following the manufacturer’s protocol. Absorbance of formazan was measured at 490 nm using a microplate reader (Spectramax).

3.2.6 Cell treatment

Differentiated SH-SY5Y cells were treated with SNP. The cells were seeded in a 6-well plate

(500K cells/well) and they were allowed to settle for 24 h before treatment.

3.2.6 Oxidative stress measurement

Oxidative stress was assayed by using Promega H2O2 glow assay. The cells were treated with SNP and incubated in 37 °C. The media was taken out in certain intervals and was assayed for the presence of H2O2.

70

3.2.7 Library construction and high-throughput RNA sequencing (RNA-seq) analysis

Library construction and deep sequencing analysis were performed with the help of gilhrit.inc.

RNA sequencing analysis was carried out on an Illumina® HiSeq 1000 sequencing system

(Illumina Inc.). Poly-A enriched RNA was selected from total RNA (1 μg) using oligo dT magnetic beads (NEB). Bound RNA was eluted by incubation at 95 °C for 2 min in 20 μL of elution buffer.

First and second-strand synthesis, adapter ligation, and amplification of the library were performed using the Illumina TruSeq RNA sample preparation kit per the manufacturer's instructions

(www.girihlet.com). Library quality was evaluated utilizing an Agilent DNA-1000 chip on an

Agilent 2100 Bioanalyzer. Library DNA templates were quantitated by qPCR and a known-size reference standard. Cluster formation of the library DNA templates was performed using the

TruSeq PE Cluster Kit v3 (Illumina) and the Illumina cBot work station under the conditions recommended by the manufacturer. Single-end, 75-base sequencing was performed with a TruSeq

SBS kit v3 on an Illumina HiSeq 1000 by using protocols defined by the manufacturer. Base call conversion to sequence reads was performed using CASAVA-1.8.2. Sequence data were analyzed with the Bowtie2, Tophat, and Cufflinks programs using NCBI's human (Homo sapiens) genome build reference mm10. Reads per kilobase transcript per million (RPKM) were normalized for the experimental group to its corresponding control. The reads that resulted were compressed by removing duplicates but keeping track of how many times each sequence occurred in each sample in a database. The unique reads were then mapped to the human genome, using exact matches.

Each mapped read was then assigned annotations from the underlying genome. This was then used to identify the reads belonging to each transcript and coverage over each position on the transcript was established. In order to compare the expression in different samples, quantile normalization

71 was used. The ratios of expression levels were then calculated to estimate the log (to base 2) of the fold-change.

3.2.8 Gene expression analysis

Heat maps and hierarchical clusters from whole transcriptomes were constructed with Ingenuity pathway analysis from Quiagen (http://www.broadinstitute.org/ cancer/ software/ GENE-E/).

Gene ontologies (GO) and signaling pathways were analyzed by the PANTHER (Protein ANalysis

THrough Evolutionary Relationships) classification system (http://www.pantherdb.org/) version

9.0. The lists of differentially expressed genes (fold change) were processed via GO annotations in this database (22160 genes, for Homo sapiens). PANTHER's overrepresentation test uses the binomial method with Bonferroni correction for multiple comparisons to annotate classification categories for a list of genes. Significance was considered at p values > 0.05. The overrepresentation test was used to identify functional classes from the submitted gene lists according to PANTHER's reference lists.

3.2.9 RT-qPCR assay

Total RNA (1 μg) or mRNA was reverse-transcribed using the quanta cDNA mastermix (Quanta

Biosciences) per the manufacturer's instructions. To evaluate transcript levels for selected genes, we used the ΔΔCt method, as previously described. The commercial (IDT) validated primers were qRT-PCR was performed on an Eppendorf thermal cycler. The amplification thermal profile was

2 min 50 °C, 10 min 95 °C, and 15 s 95°C followed by 1 min 60 °C (40 cycles). To confirm the presence of a single product after amplification, a dissociation stage was carried out: 15 s, 95 °C;

20 s, 60 °C; and 15 s, 95 °C. The primers used for the study are listed in Table 3.1.

72

Table 3.1 Primers used for RT-qPCR analyses.

Transcripts Primers Nlrp3 Forward 5'-CTTCTCTGATGAGGCCCAAG Reverse 5'-GCAGCAAACTGGAAAGGAAG Nat8l Forward 5'-TGTGCATCCGCGAGTTCCGT Reverse 5'-CGGAAGGCCGTGTTAGGGAT Sox4 Forward 5'-GAGTCCAGCATCTCCAACC Reverse 5'-ACTCTTCGTCTGTCCTTTTCG Caspase 1 Forward 5′-GCCTGTTCCTGTGATGTGGAG Reverse 5′-TGCCCACAGACATTCATACAGTTTC Cox7a1 Forward 5′-CCGACAATGACCTCCCAGTA Reverse 5′-TGTTTGTCCAAGTCCTCCAA β-actin Forward 5'-TGAGAGGGAAATCGTGCGTG Reverse 5'-TGCTTGCTGATCCACATCTGC Sod1 Forward 5’-ACTGGTGGTCCATGAAAAAGC Reverse 5’-AACGACTTCCAGCGTTTCCT Cr Forward 5′-CTGATCCCACACCCTTTCAT Reverse 5′-TTAAGGGCTCTGACGCTCAT

3.2.11 Statistical analysis

Differences were considered statistically significant at p < 0.05, student’s t-test. Each experiment was repeated at least three times unless otherwise mentioned.

3.3 Results

3.3.1 Sodium nitroprusside (SNP) treatment introduces oxidative stress in the human neuronal (SH-SY5Y) cells

Sodium nitroprusside (SNP) is known to generate a mild oxidative stress in the human cells by producing hydroxyl (•OH) and nitroxyl radicals (NO•). It also produces Peroxynitrite anion as an oxidative stress causing intermediate which is more potent oxidizer due a prolonged half-life. It

73 has been proposed that the metabolites of SNP may undergo redox cycling with oxygen to form

•- • • superoxide anion (O2 ) that in turn produces OH and NO. The iron moiety of SNP appears to mediate the pro-oxidative properties of SNP (Figure 3.2). Human neuroblastoma cells (SH-SY5Y) are known to show the characteristics of mature neurons after retinoic acid treatment. The retinoic acid differentiated human neuronal cells were treated with different concentration of SNP for 8h and an optimal SNP concentration was established which would produce a significant amount of oxidative stress in the cells with a minimum cell death (IC50: 500 µM, working concentration 250

μM, Figure 3. 3).

Figure 3.2 a. Structure of sodium nitroprusside (SNP) molecule and b. A proposed mechanism of free radical generation by SNP.

Once we established the working concentration of SNP, we quantified the level of oxidative stress produced in the SNP treated neuronal cells. As expected based on the literature data and ROS generating property of SNP, we observed a nearly two-fold increase in ROS level in the treated cells (Figure 2. b). Next, we wanted to text whether SNP induced oxidative stress is a constant stress or a burst stress as in case of peroxide based stressing agents. We used ROS-Glow assay to quantitatively estimate the level of ROS present in the cells, which measures the relative amount of H2O2 present in the media. We observed an increase in cellular ROS production immediately

74 after the SNP treatment, which stayed nearly constant for another 24 h implying that SNP treatment to neuronal cells mimics the mild but constant oxidative stress the neurons are facing in CNS under neuropathological circumstances (Figure 3.3d).

Figure 3.3. a. SNP is cytotoxic to SHSY5Y cells at higher concentration. IC50 value for the 8h treatment is 700 µM (data not shown here) which reduces to 450 µM over overnight treatment and b. SNP treatment alleviates the ROS production in the cells, c. A schematic of ROS detection by ROS-Glow assay (Promega), and d. SNP induces a mild but steady ROS production in SH-SY5Y cells.

75

Next, we aimed to investigate whether the level of stress generated in the cells is translated in the transcriptome in the similar level. We isolated the total RNA from the SNP treated neuronal cells after 8 h of treatment and quantified the level of RNA oxidation by using HPLC-ECD detection of 8-OHG molecule following the RNA cleavage with nuclease P1 and CIP. We have discovered that the extent of oxidation in the isolated RNA increases with the increasing concentration of SNP treatment to the neuronal cells (Figure 3.4). Additionally, we also used an immunoprecipitation

(IP) procedure to isolate the oxidized mRNA; spectroscopic quantification of which revealed more than 3-fold increase in the level of oxidized mRNA in the SNP treated neuronal cells (Figure 3.6).

Figure 3.4 SNP treatment results in RNA oxidation in SH-SY5Y cells, a. A representative 8-OHG HPLC-ECD chromatogram and b. The extent of 8-OHG production increases with increase in SNP concentration.

3.3.2 RNA-seq analysis of oxidized mRNA reveals that certain transcripts are selectively oxidized under SNP stress

Once we confirmed that SNP induces oxidative stress in the neuronal cells resulting in the RNA oxidation, we aimed to decipher the identity of oxidatively modified transcripts under such

76 environment. For that we isolated the oxidized mRNA from the SNP treated differentiated

SHSY5Y cells as schematically illustrated in Figure 3.5. Following the selection, we made sure that there was indeed a greater extent of oxidative damage in the immunoprecipitated mRNA pool using HPLC-ECD and found more than threefold oxidation in SNP treated mRNA pool (Figure

3.6). To further confirm the oxidative stress induced in the cells, we analyzed the expression level of oxidative stress marker mRNAs, namely superoxide dismutase 1 (sod1) and carbonyl reductase

(cr) in the total mRNA isolated from SNP treated cells. We observed that these mRNAs are upregulated in the treated cells confirming high level of oxidative stress in the cells right before

RNA isolation (Figure 3.6). RNA integrity was tested in every step either by using agarose gel electrophoresis (data not shown here) to compare the 40S and 80S ribosomal RNA bands for the total RNA or by Bioanalyzer test (Girihlet non-canonical genomics) to check the mRNA integrity

(Figure 3.5 c). The iLumina® RNA-Seq platform was used to investigate the identity of selectively oxidized mRNAs in the pool of 8-OHG immunoprecipitated mRNA.

77

Figure 3. 5 A flow-chart demonstrating stress induction in human neuronal cells with SNP, RNA isolation and immunoprecipitation, quality test of mRNA, cDNA library preparation, RNA-seq analysis and bioinformatics analysis steps. a. SH-SY5Y cells undergoing SNP treatment, b. Preparation of oxidized mRNA for RNA-seq analysis, c. Quality control tests to make sure i. oxidative stress is induced in the cells, ii. mRNA thus separated is of high quality, and iii. Good cDNA library was prepared., and d. RNA-seq and data analysis.

78

Figure 3.6 Top. A representation of level of increase in mRNA oxidation when cells were stressed with 250 μM SNP (quantified for the mRNA used for RNA-seq analysis), and Bottom. Relative level of oxidative stress related mRNA expression change in the SNP treated neuronal cells measured by RT-qPCR assay.

79

3.3.3 SNP mediated oxidative stress induces a different level of oxidation in the mRNAs involved in various biological processes in human neuronal cells

RNA-seq data analysis was done using CLC Genomics Workbench version 10.0 (Qiagen Inc.)

Ingenuity® Pathway Analysis (Quiagen Bioinformatics) of RNA-seq mRNA pool identified and ranked 917 upregulated (> 1.5-fold, p < 0.05) transcripts in the IP (+) pool as normalized to control mRNA pool. The Heatmap was generated using hierarchical clustering algorithm (Figure 3.7.a), which shows the homogeneity in the cDNA library between the experimental duplicates.

80

Figure 3.7 RNA-seq data analysis and validation. a. A hit map representing the experimental replicates for each condition (control mRNA vs immunoprecipitated mRNA), b. PANTHER pathway analysis of highly oxidized transcripts showed many of the neuropathologically relevant transcripts are among the highly oxidized ones. c. RT-qPCR analysis of some of the transcripts validated the RNA-seq data.

Overrepresentation test of the transcripts abundant in the oxidized mRNA pool after SNP treatment and immunoprecipitation was performed by using Protein Analysis Through Evolutionary

Relationships (PANTHER, http://www.pantherdb.org). The transcripts were analyzed for 5 major categories: (A) Signaling pathways, (B) Protein class, (C) Cellular component, and (D) Molecular function, and (E) Biological process. Among signaling pathways, one of the most represented one was neurological pathways related transcripts as represented by Alzheimer’s, Parkinson’s,

Huntington diseases, inflammation signaling and apoptosis related pathways (Figure 3.7.b). These pathways included various significantly oxidized transcripts (1.5-100 fold). Since MS shares some of the characteristics of inflammatory diseases and classical neurodegenerative diseases any common transcripts found in common signifies the involvement of common route of disease progression in MS. Among these altered transcripts, Nlrp3 mRNA is a key inflammation signaling mRNA which is one of the highly oxidized molecules. Another key molecule of this category is

Mapk11, which is a key regulator of oxidative stress in Hela cells. As discussed in more details in the following sections and listed in Appendix of this dissertation, there are many more oxidized mRNAs which belong either to inflammation signaling pathways or one of the neuropathological pathways.

81

3.3.4 mRNAs coding for RNA binding proteins are among the highly oxidized RNA

Protein class analysis (Figure 3.8) of selectively oxidized mRNA revealed that many of the nucleic acid binding proteins (both DNA and RNA) are among the highly oxidized mRNAs. Transcription factors coding mRNAs are another major class impacted most by the stress. The presence of both categories of mRNA in the oxidized mRNA pool implies the detrimental effect of selective mRNA oxidation in transcriptional regulation and RNA decay mechanism (RNPs are essential components of p-bodies and S-granules).

82

Figure 3.8 Transcript analysis based on protein class reveals that many of the mRNAs for nucleic acid binding proteins and transcription factors are selectively oxidized in human neurons under SNP stress.

3.3.5 mRNAs related to metabolic pathways and cell processes are oxidized more

As summarized in Figure 3.9 and Table 3.2, we also observed that many transcripts involved in metabolic pathways and cell process pathways are oxidized more. Metabolism is directly linked to cellular energetics; the latter is obviously centered around mitochondrial functioning. N-acetyl aspartate (NAA) is a key neuronal metabolite, which is also indicative of neuronal health and energetics. Interestingly we found that Nat8l mRNA is among the highly oxidized mRNA molecules. Since Nat8l looks to be in a prime location of mitochondrial energetics and neuronal health we focused on investigating the correlation between mRNA oxidation, the mRNA expression level and protein level for this mRNA. We observed that there is a four-fold more oxidation in the Nat8l mRNA. After the immunoblot analysis of total cell extract, we observed a reduced level of expression of NAT8L protein in SNP treated SH-SY5Y cells. This clearly demonstrated the detrimental impact of mRNA oxidation in their translational efficiency.

83

Figure 3. 9 Transcript analysis based on biological processes reveals that many of the mRNAs for metabolic pathways and cell process pathways are selectively oxidized in human neurons under SNP stress.

Table 3.2 A list of catalytic proteins in human neuronal cells, the mRNAs of which are highly oxidized.

Gene symbols Gene names

B4GALNT3 Beta-1,4-N-acetylgalactosaminyltransferase 3

TRMT11 tRNA (guanine(10)-N2)-methyltransferase

CDC42BPG Serine/threonine-protein kinase MRCK gamma

NSUN2 tRNA (cytosine(34)-C(5))-methyltransferase

NAA50 N-alpha-acetyltransferase 50

CSGALNACT2 Chondroitin sulfate N-acetylgalactosaminyltransferase 2

CDC42BPB Serine/threonine-protein kinase MRCK beta;

RFK Riboflavin kinase;RFK;ortholog

MAPK11 Mitogen-activated protein kinase 11

FAS Fatty acid synthase

TAF1 Transcription initiation factor TFIID subunit 1

DUSP4 Dual specificity protein phosphatase 4

NAT8L N-acetylaspartate synthetase like 8

ZDHHC11B Probable palmitoyltransferase

SLK STE20-like serine/threonine-protein kinase

GGT1 Gamma-glutamyltranspeptidase 1

TYW3 tRNA wybutosine-synthesizing protein 3

84

RRP1 60S acidic ribosomal protein P1

FUT6 Alpha-(1,3)-fucosyltransferase 6

NTPCR Cancer-related nucleoside-triphosphatase

UBASH3B Ubiquitin-associated and SH3 domain-containing protein B

3.3.6 mRNAs related to catalytic functions and binding proteins are selectively oxidized

Nat8l mRNA is just an example but there are many other catalytic proteins, the mRNAs of which are selectively oxidized in human neurons under SNP treatment. Interestingly many of these are transferase proteins (Table 3.3 and Figure 3.10). Some other transcripts of significant interest are tRNA transferases and those which are involved in fatty acid synthesis. Our analysis also revealed that many of the binding proteins are among the highly oxidized mRNAs (Figure 3.10).

Figure 3. 10 Transcript analysis based on molecular function reveals that many of the mRNAs for binding proteins and catalytic proteins are selectively oxidized in human neurons under SNP stress.

85

3.3.7 Cellular component-based analysis

Cellular component-based analysis revealed that the transcripts which code for proteins all over the cells are amongst the highly oxidized transcripts. This implies that there might not be just the proximity factor which makes certain mRNAs to go to a higher level of oxidation. In the cellular component analysis, 138 out of 238 transcript hits are for mRNAs that code the nuclear proteins.

Similarly, we observed 26 endoplasmic reticulum proteins, 22 Golgi proteins, and 16 mitochondrial proteins on the oxidized transcript hit (Figure 3.11 and Table 3.3).

Figure 3.11 Transcript analysis based on cellular components reveals that mRNAs encoding proteins for different cellular compartments are selectively oxidized in human neurons under SNP stress.

86

Table 3.3 A list of mitochondrial proteins in human neuronal cells, the mRNAs of which are highly oxidized.

Gene ID Gene names

SPATA18 Mitochondria-eating protein

RAB21 Ras-related protein Rab-21

MRPL27 39S ribosomal protein L27

MRPL32 39S ribosomal protein L14

HARS Histidine--tRNA ligase

RGPD2 RANBP2-like and GRIP domain-containing protein 2

TXNRD1 Thioredoxin reductase 1

HSPD1 60 kDa heat shock protein1

RAB18 Ras-related protein Rab-18

CCT6A T-complex protein 1 subunit zeta

RGPD8 RANBP2-like and GRIP domain-containing protein 8

RAB1A Ras-related protein Rab-1A

LONP1 Lon protease homolog

RAB12 Ras-related protein Rab-12

RAB17 Ras-related protein Rab-17

MRPL27 39S ribosomal protein L41

BCL2L2 Bcl-2-like protein 2

FOXRED1 FAD-dependent oxidoreductase domain-containing protein 1

87

3.3.8 Validation of RNA-seq data

We performed RT-qPCR analyses of some of the representative mRNAs from RNA-seq data to validate the later. As represented in Figure 3.7.c the transcripts which were upregulated in the

RNA-seq analysis of immunoprecipitated mRNA were also upregulated in RT-qPCR analysis. To validate the impact of mRNA oxidation in their translation, we also performed the western blot analysis of a couple of proteins whose mRNAs are oxidized more (as described in detail in

Chapters 4 and 5). We observed a reduced level of protein for the highly oxidized transcripts clearly indicating the detrimental role of mRNA oxidation in cellular physiology.

3.4 Discussion

A plethora of reports have demonstrated the presence of higher RNA oxidation in the neurons of patients with various neurological disorders5,7-10,15. Some of the studies have demonstrated that

RNA oxidation is an earlier phenomenon in neurons just before neurodegeneration10,18. A couple of studies have demonstrated that some of the mRNAs are more susceptible to oxidative damage and the oxidative damage to selective mRNAs could be detrimental to particular neurological disorder19. Although there is a lot of speculation in the literature about the presence of selective

RNA oxidative damage and their possible connection to neurodegeneration5,8,10,15,19, the high- throughput evaluation of oxidized mRNA and their analysis based on different aspects with the possible connection to a different mechanistic understanding of disease development was lacking in MS and related diseases. This study has identified the selectively oxidized mRNA molecules in the human neuronal cells under oxidative stress/ inflammatory stress situation using immunoprecipitation and RNA-seq techniques. By doing pathway analysis, we have also identified

88 the possible involvement of those altered mRNAs in the MS and other neurodegenerative diseases.

The RNA-seq data were validated using RT-qPCR technique and the impact of mRNA oxidation in respective protein synthesis has been verified by western blotting analysis (Chapter 4 and 5).

This study well corroborates with our previous finding which showed at a molecular level that normal appearing cortex of MS brains have more oxidized RNA than the sex, age, PMI matched non-MS brain tissues based upon immunohistochemical detection and quantification of the major oxidized RNA product 8-OHG.

Our findings suggest that there is no proximity context in mRNA oxidation as being obvious by the highly oxidized mRNAs related to proteins in every part of cellular compartment (Figure 3.11).

One would speculate that the mRNA molecules which are located nearby ROS generation sites should probably get a higher oxidation. Though it needs to be concluded only after analyzing the mRNA oxidation in a bigger sample size, our data indicate no such correlation. This suggests possibly a targeted mRNA oxidation under oxidative stress which eventually implies a regulatory role of mRNA oxidation. Another speculation about the selectivity of mRNA oxidation was that the highly G-rich mRNAs are the one which get highly oxidized due to the favorable standard oxidation potential of guanines. In this study we have observed the situations towards the both ends of this prediction. There are some highly G-rich mRNAs like Nat8l which are oxidized more but there are other mRNAs like Nlrp3 which are not as G-rich but still highly oxidized. As summarized in Figure 3.12, the correlation between G-content and mRNA oxidation does not hold very strong. Nevertheless, more than 60% of highly oxidized mRNAs have >25% Gs in the whole transcript sequence. Quite interestingly nearly 75% of these selectively oxidized mRNAs do have

>25% Gs in their coding region. However, we also discovered that certain transcripts with only

17% Gs are also among the one, which go a higher oxidation (Complete list in Appendix 1).

89

We have discovered that a significantly high number of mRNAs which code for nucleic acid binding proteins are highly oxidized. EIF3A, EIF3C, EIF5B, and EIF4H some of the important proteins involved in translational initiation with a higher mRNA oxidation under SNP stress. Some of these proteins are also responsible in stress granule formation11. Based on our current understanding of detrimental effect of mRNA oxidation in translation and target recognition

(binding), this could be of tremendous importance in mRNA turnover since RNPs are the proteins which play a crucial role both in mRNA quality control and protein synthesis. We have also found that many of the mRNAs which code for enzymes, especially transferases, are highly oxidized.

The other class of coding RNA getting a higher oxidation is neuropathologically relevant mRNA molecules like SP-140, which is known to present in less amount in MS brain.

NLRP3 is another interesting molecule involved in inflammation signaling in CNS, whose mRNA is selectively oxidized in neurons under our study conditions. Similarly, mRNAs for proteins like

PRDX3, MYL1, NES, and ATP4A which are known to be downregulated in the MS brain are highly oxidized indicating a connection between a higher mRNA oxidation to damaged proteome in MS20,21.

90

Figure 3.12 The selective oxidation of mRNA as a function of G enrichment in the transcripts and coding regions. Top. A scatter plot showing the trend of G% in oxidized transcripts (arranged based on decreased number of Gs in the total transcript). B. A scatter plot showing the trend of G% in oxidized transcripts (arranged based on decreasing number of Gs in the coding region).

91

3.5 Conclusion

Using an in cellulo model, we investigated the selectively targeted transcript under MS like environment. We have demonstrated that the mRNAs related to neuropathological situation especially the mRNAs for nucleic acids binding proteins are among the highly oxidized mRNAs.

Our finding also correlates well with the MS proteome data in the literature suggesting a higher mRNA oxidation in the respective mRNAs could contribute in the damaged proteins in MS brain.

3.6 References

1 Fitzner, D. & Simons, M. Chronic progressive multiple sclerosis – Pathogenesis of

neurodegeneration and therapeutic strategies. Curr Neuropharmacol 8, 305-315 (2010).

2 Haider, L. et al. Oxidative damage in multiple sclerosis lesions. Brain 134, 1914-1924 (2011).

3 Nakahara, J., Maeda, M., Aiso, S. & Suzuki, N. Current concepts in multiple sclerosis:

Autoimmunity versus oligodendrogliopathy. Clinl Rev Allergy Immunol 42, 26-34, (2012).

4 Nita, M. & Grzybowski, A. The role of the reactive oxygen species and oxidative stress in the

pathomechanism of the age-related ocular diseases and other pathologies of the anterior and

posterior eye segments in adults. Oxid Med Cell Long 2016, 3164734 (2016).

5 Uttara, B., Singh, A. V., Zamboni, P. & Mahajan, R. T. Oxidative stress and neurodegenerative

diseases: A review of upstream and downstream antioxidant therapeutic options. Curr

Neuropharmacol 7, 65-74 (2009).

6 Dutta, R., McDonough, J., Yin, X., Peterson, J., Chang, A., Torres, T., Gudz, T., Macklin, W. B.,

Lewis, D. A., Fox, R. J., Rudick, R., Mirnics, K. & Trapp, B. D.. Mitochondrial dysfunction as a

cause of axonal degeneration in multiple sclerosis patients. Ann Neurol 59, 478-489 (2006).

7 Kharel, P., McDonough, J. & Basu, S. Evidence of extensive RNA oxidation in normal appearing

cortex of multiple sclerosis brain. Neurochem Intl 92, 43-48 (2016).

92

8 Nunomura, A. et al. RNA oxidation in Alzheimer disease and related neurodegenerative disorders.

Acta Neuropathol 118, 151-166 (2009).

9 Poulsen, H. E. et al. RNA modifications by oxidation: a novel disease mechanism? Free Radic Biol

Med 52, 1353-1361 (2012).

10 Shan, X., Chang, Y. & Lin, C. G. Messenger RNA oxidation is an early event preceding cell death

and causes reduced protein expression. FASEB J 21, 2753-2764 (2007).

11 Anderson, P., Kedersha, N. & Ivanov, P. Stress granules, P-bodies and cancer.Biochim Biophys

Acta 1849, 861-870 (2015).

12 Decker, C. J. & Parker, R. P-Bodies and stress granules: Possible roles in the control of translation

and mRNA degradation. CSH Perspect Biol 4, a012286, (2012).

13 Choi, Y. J., Gibala, K. S., Ayele, T., Deventer, K. V. & Resendiz, Marino J. E. Biophysical

properties, thermal stability and functional impact of 8-oxo-7,8-dihydroguanine on

oligonucleotides of RNA-a study of duplex, hairpins and the aptamer for preQ1 as models. Nucleic

Acids Res 45, 2099-2111 (2017).

14 Simms, C. L., Hudson, B. H., Mosior, J. W., Rangwala, A. S. & Zaher, H. S. An active role for the

ribosome in determining the fate of oxidized mRNA. Cell reports 9, 1256-1264 (2014).

15 Tanaka, M., Chock, P. B. & Stadtman, E. R. Oxidized messenger RNA induces translation errors.

PNAS USA 104, 66-71 (2007).

16 El-Maarouf-Bouteau, H., Meimoun, P., Job, C., Job, D. & Bailly, C. Role of protein and mRNA

oxidation in seed dormancy and germination. Front Plant Sci 4, 77 (2013).

17 Wang, J.-X. et al. Oxidative Modification of miR-184 Enables It to Target Bcl-xL and Bcl-w. Mol

Cell 59, 50-61 (2015).

18 Nunomura, A. et al. Oxidative damage is the earliest event in Alzheimer disease. J Neuropathol

Exp Neurol 60, 759-767 (2001).

19 Kong, Q. & Lin, C. G. Oxidative damage to RNA: mechanisms, consequences, and diseases. Cell

mol life sci 67, 1817-1829 (2010).

93

20 Satoh, J. I., Tabunoki, H. & Yamamura, T. Molecular network of the comprehensive multiple

sclerosis brain-lesion proteome. Mult Scler 15,531-541 (2009).

21 Fischer, M. T. et al. NADPH oxidase expression in active multiple sclerosis lesions in relation to

oxidative tissue damage and mitochondrial injury, Brain 135, 886–899 (2012).

94

CHAPTER 4

4. Oxidative damage to Nat8l mRNA in human neurons is linked to multiple sclerosis pathogenesis

4.1 Introduction

N-acetyl aspartate transferase 8 like protein (NAT8L) is a key trans-mitochondrial membrane protein that catalyzes the transfer of acetyl group from acetyl-CoA to aspartate to form N-acetyl aspartate (NAA) and transfers the product to cytoplasm (Figure 4.1)1,2. The sequence of NAT8L is very well conserved among mammals with a very high sequence similarity between human and mouse (Figure 4.2). NAA is one of the most abundant analytes in the human central nervous system

(CNS)3. NAA synthesizing enzyme NAT8L is expressed in neurons while the catabolic enzyme aspartoacylase (ASPA) is present primarily in oligodendrocytes in the CNS3. NAA in the brain not only serves as a source of metabolic acetate in oligodendrocyte for myelin production but also signals protein/histone acetylation and acts as a precursor for the N-acetylaspartylglutamate

(NAAG) synthesis4. NAA, NAAG as well as ASPA are significantly reduced in glioma tumors, suggesting a possible role for decreased acetate metabolism in tumorigenesis4. There are compelling evidences that NAA is essential for lipid synthesis and myelination in the CNS5. In oligodendrocytes, NAA is converted back into aspartate and acetyl CoA, the later involves in myelin synthesis which will eventually deposited onto the neuronal axons as an insulating material5. It has been shown that NAA-derived acetate has a 3-fold higher potential to be

95 incorporated into lipids than free acetate6. In addition to its link to lipogenesis and myelin synthesis in oligodendrocytes, NAA has also been proposed to act as an organic osmolyte that removes excess water from neurons by acting as a molecular water pump7. In this regard, it has been suggested that a reduced level of NAA leads to a bad neuronal health disturbing the neuronal osmolarity in addition to poor myelin synthesis7. On the other hand it has been proposed that excess

NAA leads to osmotic dysregulation or has other cytotoxic effects that are responsible for the pathology observed in Canavan disease (CD) patients8. Axonal degeneration in MS brains is supported by in vivo magnetic resonance spectroscopy (MRS) data describing decreased levels of

NAA9. Reduced NAA levels could reflect reversible neuronal or axonal damage caused by inflammatory demyelination, altered neuronal or axonal metabolism, or axonal loss9. Recently, it was reported that NAT8L knockout mice show decreased NAA content in the brain, reduced social interaction and shortened immobility time in the forced swimming test10. These results suggest the central regulatory role of NAT8L in mammalian brain.

The most abundant brain metabolite NAAG is widely distributed in the mammalian brains and acts as a highly selective neurotransmitter for group II metabotropic glutamate receptor 3

(mGluR3)11. Once being released to synaptic cleft, NAAG binds mGluR3 and is metabolized to

NAA and glutamate by glutamate carboxypeptidase II (GCPII)11. Studies on postmortem brains showed that the levels of NAA and NAAG are significantly lower in the brains of subjects with psychiatric disorders like major depressive disorder, schizophrenia and bipolar disorder11. In patients with Alzheimer’s disease, NAA levels are significantly reduced in the cingulate gyrus12,13.

NAA level is reduced in the CNS or PD and MS patients2,7,9,14. A clinical study using magnetic resonance spectroscopy (MRS) showed that NAA was significantly increased in adult patients with autism spectrum disorder when compared with a control group. In a different context like

96

Canavan’s disease, the accumulation of NAA due to the absence of downstream catabolic enzyme

(ASPA mutation) could be toxic to the CNS. NAA, recently has been demonstrated to have an epigenetic signaling role. A higher concentration of NAA signals histone methylation which eventually contributes in mitochondrial energetics and myelin synthesis15. Presence of a substantially high amount of NAA in the brain (20 mM- 50 mM) suggests its yet to be discovered diverse roles in CNS physiology10. It has been shown that elevated NAA levels detected by proton magnetic resonance spectroscopy (1H-MRS) brain is related to an increased neuronal activity7.

Uno’s lab has demonstrated that Shati/ Nat8l knock out mice exhibits hyper locomotion, anxiety behaviors, and social dysfunction16. The same lab has also demonstrated that the same mice have a high sensitivity to methamphetamine, as evidenced by their results in assessments of locomotor activity and conditioned place preference, as well as their elevated dopamine levels. They have shown that overexpression of accumbal NAT8L attenuates methamphetamine-induced behaviors16. Together, these observations suggest that the NAA synthesizing enzyme NAT8L may play important roles in psychiatric, neurodegenerative, and neurodevelopmental disorders. This led us to hypothesize that problems with NAT8L expression could hamper the NAA metabolism in CNS which could eventually impair the supply of NAA-derived acetate, which in turn results in decreased synthesis of myelin-related fatty acids and lipids.

Recent findings indicate the increased oxidative damage to mRNAs appears to be associated with many neurological disorders17. One previous finding reported a direct correlation between the extent of mRNA oxidation and the frequency of translation errors, as indicated by the accumulation of 8-oxoguanosine derivatives and short peptides18. There are reports indication the production of truncated proteins18 and ribosome stalling due to the presence of oxidized guanine in the coding region of mRNA19. The studies also suggested that oxidized mRNAs do have a similar lifetime as

97 the non-oxidized mRNAs in cellulo18. These findings clearly suggest a detrimental role of mRNA oxidation in cellular physiology.

Figure 4.1 NAT8L is central to the CNS NAA metabolism.

As discussed in chapter 2 of this dissertation, we have recently demonstrated the presence of three- fold more RNA damage in the neurons of normal appearing areas of MS brain. As detailed in chapter 3, using RNA-seq and RT-qPCR analyses of the oxidized mRNA pool from human neurons, we have identified the mRNAs which undergo selective oxidation under MS microenvironment. We have discovered that Nat8l mRNA is one of the selectively oxidized mRNAs which could contribute in MS pathogenesis by disrupting NAA metabolism.

In this study, we have demonstrated that the oxidative damage of Nat8l mRNA leads towards the reduced production of NAT8L protein in vitro. The in cellulo and in vivo study under the oxidative

98 stress situation revealed that indeed there is a higher Nat8l mRNA oxidation and a reduced production of NAT8L production under MS microenvironment. As mentioned earlier the presence of a reduced level of NAT8L reaction product NAA has already been reported in MS brain and

MS animal models9,20. The observation of a reduced NAT8L protein in the postmortem MS brain strongly suggests a correlation between Nat8l mRNA oxidation and a reduced NAA production.

Our study establishes the link between selective mRNA oxidation and MS neurodegeneration. To directly study the impact of mRNA oxidation in NAT8L protein expression, we engineered a plasmid-based platform and demonstrated that Nat8l mRNA oxidation results in reduced protein production in vitro.

Figure 4.2 NAT8L protein sequence is well conserved in mammals. Human (Homo sapiens) and mouse (Mus musculus) sequences are compared here (Source NCBI).

99

4.2 Materials and methods

4.2.1 Cell culture

Human neuroblastoma SH-SY5Y cells, obtained from ATCC were cultured in Dulbecco’s modified Eagle’s medium supplemented with 10% (v/v) fetal bovine serum, and 0.01% penicillin/streptomycin at 37°C under 5% CO2 in air.

4.2.2 Cell treatment

Differentiated SH-SY5Y cells were treated with different concentrations of SNP. The cells were seeded in a 6-well plate (500K cells/well) and they were allowed to settle for 24h before treatment.

4.2.3 Cuprizone mouse model

The mode of action of cuprizone-intoxication is believed to happen via the copper-chelation action of cuprizone which inhibits mitochondrial enzymes of the respiratory chain that require copper as co-factor. This leads to oxidative stress and in consequence to primary oligodendrocyte apoptosis which is closely followed by microglia and astrocyte activation. These events lead to demyelination and neurodegeneration. The C57BL/6 strain mice were fed either a regular diet or cuprizone containing diet for 6 weeks before being sacrificed for brain RNA/protein analysis. The male C57BL/6 mice at 6 to 9 weeks of age were fed a diet of chow mixed with 0.2% cuprizone over the course of 6 weeks.

4.2.4 Immunohistochemistry

Mice brain sections (30 µm, gray matter) were fixed with 4% paraformaldehyde in PBS (pH 7.4) containing for 20 min, washed with PBS, and then incubated with 0.25% Triton X-100 in PBS for

15 min. The sections were treated with 10 mM citrate solution (pH 6.0) for antigen retrieval at

100

95 °C for 15 min, washed with PBS. Afterwards, the sections were blocked in 5% normal donkey serum (Sigma-Aldrich, St. Louis, MO) in PBS for 1 h. Sections were incubated with primary antibody (8-OHG, NAT8L, NeuN) with 5% normal donkey serum in PBS at 4 °C overnight, washed with PBS-T, and then incubated with fluor secondary antibodies at room temperature for

2 h. After being washed with PBS, the sections were mounted using Vectashield mounting media and imaged later.

4.2.5 Organelle fractionation

Mitochondria were isolated from the cells with a commercially available kit (Thermo Scientific) using the Dounce homogenizer and 3000 × g to pellet the mitochondria. To isolate the mitochondria from the mice brain, the brains were homogenized using a Dounce homogenizer with about 50 strokes. The centrifugation steps to pellet the mitochondria were carried out at 3000 × g for 10 min to reduce peroxisomal contamination. Nuclear fraction, mitochondrial fraction, and post-mitochondrial supernatant containing cytosol and ER remnants were lysed in SDS-lysis buffer (50 mm Tris-HCl, pH 6.8, 10% glycerol, 2.5% SDS, 1× protease inhibitor mixture, 1 mm

PMSF) and used for further analysis. Cytosolic/ER proteins have been precipitated using the trichloroacetic acid (TCA) method. Briefly, cytolytic protein lysate was mixed with 50% ice-cold

TCA to obtain a concentration of 10% TCA and incubated for 1.5 h on ice. Then it was centrifuged for 10 min at 16000*g and 4 °C. The pellet was washed twice with ice-cold acetone, air-dried, and dissolved in SDS-lysis buffer. The protein concentrations were determined with the Bradford protein assay kit (Pierce).

4.2.6 RNA isolation and immunoprecipitation

101

RNA was isolated and immunoprecipitated from SH-SY5Y cells as described in previously in section 3.2.3. RNA from mice brain was isolated from gray areas following similar protocol after the tissues were homogenized in Dounce homogenizer.

4.2.7 RT-qPCR

Relative levels of mRNA for Nat8l, and gapdh, were quantitated by qRT-PCR in mRNA isolated from SH-SY5Y cells and gray matter of mice brains. qRT-PCR was performed with SYBR Green

(Quanta Biosciences) and gene-specific primers. Relative levels of mRNA expression in control

& treated cells and control & cuprizone mice were determined by the 2−ΔΔCt method after normalization to gapdh mRNA.

Table 4.1 Primers used for RT-qPCR analysis.

4.2.8 Western blotting

20 μg of total protein extracts were subjected to a 15% SDS-PAGE analysis and were blotted to nitrocellulose membranes. The following antibodies were used: anti-NAT8L (1:1000) (Novus

Biologicals, catalog no. NBP1-06599); anti-GAPDH (1:1000) (Calbiochem); anti-Aralar (1:1000).

102

For chemiluminescent detection, a horseradish peroxidase-conjugated secondary antibody was used (anti-rabbit; 1:5000, SantaCruz), and ECL (SantaCruz) served as substrate.

4.2.9 Enzymatic assay

Asp-NAT activity of NAT8L was assayed radiochemically in a mixture (200 μl final volume) comprising, 10 mM potassium phosphate, 20 mM Hepes, pH 7.4, 1 mM MgCl2, 50μM L-aspartate,

10000 cpm. L-[U-14C] aspartate (Moravek Inc.), containing approx. 20 mg of protein/ mL. After

30 min incubation at 30◦C, the reaction was stopped by a 5-min incubation at 80◦C and 1 mL of 5 mM Hepes, pH 7.4, was added. The sample was centrifuged for 5 min at 13200 rpm and the supernatant was applied on to a 1 mL Dowex AG1X8 column (Cl− form, 100–200 mesh; Acros

Organics) prepared in a Bradford column. The latter was washed with 5 mL of 5 mM Hepes, pH

7.4, followed by 5 mL of 100 mM NaCl in Hepes buffer to elute unreacted aspartate. Finally, 5 mL of 300 mM NaCl in Hepes was used to elute the NAA. The yield was radioactively quantified using a liquid-scintillation counter.

4.2.10 NAA measurement

Levels of NAA were measured in neuronal cells by HPLC-UV. A Whatman Partisil 10 SAX anion- exchange column (4.6 mm × 250 mm) was used in an Agilent 1100 Series HPLC Value System.

The mobile phase consisted of 0.1 M KH2PO4 and 0.025 M KCl at pH 4.5. Retention data were collected at a flow rate of 1.5 mL/ min, and the flow was monitored with an Agilent 1100 series

UV detector at 214 nm. The retention time was determined with an NAA standard (Sigma). Peak areas were acquired with Agilent Chemstation® software. NAA concentrations for SNP treated and control cells were determined in triplicate, and statistical significance of differences in NAA

103 concentration between control and treated samples was determined with a two-tailed Student's t test, with p < 0.05 considered statistically significant.

4.2.11 Mutagenesis

Mutations in the Nat8l expression were introduced to disrupt the GQ forming sequence within the

ORF using NEB Q5 site directed mutagenesis kit. Table 4.2 summarizes the list of primers used and intended purpose.

Table 4.2 Primers used for mutagenesis.

4.4.12 Nat8l mRNA expression vector design and mRNA oxidation

Human Nat8l expressing plasmid was designed and gotten from applied biomaterials. The plasmid was transcribed to get mRNA. The purified mRNA was oxidized using Fenton chemistry (100 µM

2+ H2O2, 10 µM Fe and 10 µM ascorbate). After an hour reaction, the mRNA was purified using the desalting columns and then 5% denaturing polyacrylamide gel electrophoresis. The amount of purified RNA was spectrophotometrically quantified and normalized before using the mRNA for in vitro translation.

104

4.2.13 In vitro translation

To translate mRNAs encoding for NAT8L protein, rabbit reticulocytes system, nuclease treated

(Promega) was employed, according to the manufacturer’s protocol. All reactions were done in volume of 50 μL at 30°C for 60 min. Relative quantification of protein thus produced was done by comparing the band intensities using immunoblot technique followed by luminol detection as mentioned above in western blot protocol.

4.2.14 Statistical Analysis

If not otherwise stated, results are mean values ± S.D. of at least three independent experiments, or results show one representative experiment of three. Statistical analysis was done on all available data. Statistical significance was determined using the two-tailed Student's t test. p

Values ≤ 0.05 were considered statistically significant.

4.3 Results

4.3.1 Many mRNAs linked to neuropathology are selectively oxidized under MS microenvironment

As detailed in the chapter 3, RNA-seq analysis of oxidized mRNA showed that many of the MS pathology relevant mRNAs are among the highly oxidized mRNAs. We analyzed the potential functional impact of oxidation in such molecules and decided to elaborate our analysis on some representative molecules with known functional relevance in MS pathophysiology. Nat8l and

Nlrp3 (discussed in chapter 5) are two key mRNA molecules we chose for the further analysis.

105

4.3.2 Nat8l is one of the selectively oxidized mRNAs under MS microenvironment

NAT8L is a single-pass membrane protein, which contains a conserved sequence of the NAT superfamily of N-acetyltransferases and is a member of the N-acyltransferase (NAT) superfamily21. It is a neuron-specific N-acetylaspartate (NAA) biosynthetic enzyme, catalyzing the NAA synthesis from L-aspartate and acetyl-CoA (Figure 4.1). NAA is a major storage and transport form of acetyl coenzyme A specific to the nervous system21. NAT8L protein is getting an increasing attention in different neuropathological situations and seems to have an involvement in the neurological disorders like MS, Canavan’s disease and Parkinson’s disease2,7,8,13,15. A reduced level of NAA has been reported in these diseases. NAA being a downstream metabolite of NAT8L enzyme, we were very interested to analyze the level and consequences of Nat8l mRNA damage. We have observed a two-fold increase of Nat8l mRNA in oxidized pool in RNA-seq data indicating two-fold more oxidative damage to this mRNA (Figure 4.3). Interestingly, this mRNA turned out to be a very G-rich molecule having nearly 35% G’s in coding region and 3’ UTR and

36% G’s in coding region.

4.3.3 NAT8L protein is downregulated in SNP treated neurons

To test the impact of RNA oxidative modification in the protein level we did immunoblotting analysis of NAT8L protein. The protein expression was normalized with the expression of

GAPDH protein as an internal control. Quite interestingly, we observed a 9-fold reduction in the protein level (Figure 4.3). This observation demonstrates a connection of Nat8l mRNA oxidative damage to a reduced NAT8L protein production.

106

Figure 4.3 Selective mRNA oxidation is detrimental in Nat8l expression. a. a comparison of RNA- seq data with RT-qCR demonstrates that though the overall mRNA expression does not change significantly, the level of oxidized Nat8l mRNA increases by almost 4-fold in SNP treated SH- SY5Y cells, b. A representative western blot showing a reduced level of NAT8L protein in SNP treated cells, and c. Quantification of 3 blots (as represented in b) show 90% reduction in NAT8L protein level.

107

4.3.4 NAA concentration is reduced in SNP treated neuronal cells

Next, we studied the generation of NAA in control versus SNP treated SH-SY5Y cells. NAA was quantitated in SH-SY5Y cells before and after treatment with the SNP as shown in Figure 4.4.

Representative HPLC chromatograms for the NAA standard and for NAA measurements in SH-

SY5Y cells are shown in Figure 4.4 a, b and c. Quantitation for NAA measured with our HPLC method in control and SNP treated SH-SY5Y cells is shown in Figure 4.4.d. Our data show that treatment with SNP reduced NAA levels in SH-SY5Y cells by 75 % after 8 h of treatment. Cell number was normalized before NAA isolation and after SNP treatment. The reduced amount of crucial neuronal metabolite NAA in the neuronal cells corroborates well with a reduction in

NAT8L observed via immunoblot analysis (Figure 4.4.c).

108

Figure 4.4 SNP treatment results in a reduced NAA level in neuronal cells. a & b. Representative chromatograms to show the presence of NAA in SH-SY5Y cells and a reduced level of NAA in

SNP treated SH-SY5Ycells- arrows indicate the NAA position in the chromatograms, c. A chromatogram representing standard NAA, and d. Quantification of NAA level in four independent experiments (control vs SNP treated) showed a 75% reduction in NAA level in SNP treated cells (p< 0.005).

109

4.3.5 Nat8l mRNA is highly oxidized in the brain of MS animal model

Cuprizone fed mouse model is one of the very common MS animal models. Having a copper chelator compound cuprizone in the mice diet is known to induce demyelination and MS like phenotype in mice brain (Figure 4.5). We aimed to analyze the extent of NAT8L mRNA damage in MS animal model. The mice were fed on cuprizone containing diet for 4 weeks (regular diet for control mice) and sacrificed.

Figure 4.5 a. Cuprizone chelates Cu2+ and b. The inclusion of cuprizone in mice diet induces demyelination and other MS symptoms.

Total RNA was isolated from the gray areas of mice brain and immunoprecipitated with 8-OHG antibody. We observed a nearly 2-fold increase in total RNA oxidation in cuprizone mice brain

(Figure 4.6.a). Next, we analyzed the relative change in Nat8l mRNA level in cuprizone mice brain

110 and see no significant change. Interestingly we noticed a 2-fold more oxidation in the Nat8l mRNA in cuprizone mice brain which is the exact same trend we observed in human neuronal cells under

MS microenvironment (Figure 4.6. b & c). This is a pretty exciting correlation between Nat8l mRNA damage in cellulo and in animal model.

Fig. 4.6 Selective mRNA oxidation is detrimental in Nat8l expression in cuprizone mice brain. a. There is a higher level of RNA oxidation in cuprizone mice brain, b. RT-qPCR analysis of total RNA extracted from gray matter of cuprizone mice brain shoed that there is no significant change in Nat8l mRNA expression in cuprizone mice brain, c. Although the overall mRNA expression does not change significantly, the level of oxidized Nat8l mRNA increases by almost 2-fold in cuprizone mice brain.

4.3.6 NAT8L protein is downregulated in cuprizone-fed mice brain

Next, we wanted to investigate whether NAT8L protein expression in influenced by the oxidation of mRNA molecules in vivo. Total protein was isolated from the gray areas of mice brain and immunoblotted against anti-NAT8L antibody. As in case of SNP treated neurons, we observed an

111

85% reduction in NAT8L protein level in cuprizone fed mice brain. Since there are at least a couple of studies claiming exclusively endoplasmic reticulum based localization of NAT8L, we also aimed to analyze the presence of NAT8L protein in mice brain. For that, we fractioned the total tissue extract into mitochondrial and cytosolic fractions and performed immunoblot analysis. As demonstrated in Figure 4.7, we observed a reduction in NAT8L protein level in both fractions.

Figure 4.7 NAT8L protein expression is compromised in cuprizone fed mice brain. Top. Immunoblot analysis of NAT8L level in fractions isolated from mice brain (GAPDH and ARALAR1 are loading controls which are abundant in cytosol and mitochondria respectively) and Bottom. Quantification of three such immunoblots.

112

There are at least a couple of conflicting suggestions in literature indicating the exclusive presence of NAT8L protein in endoplasmic reticulum (ER)21. Our data clearly indicate the impact of reduction of NAT8L protein in NAA production, which eventually could be detrimental in lipogenesis and myelin synthesis process in oligodendrocytes. The catalytic activity of NAT8L to produce NAA itself is a characteristic of mitochondrial location of NAT8L protein. To confirm the existence of NAT8L in mitochondria, we fractioned the total cell extract from both SH-SY5Y cells and mice brain into mitochondrial and cytosolic fractions. The immunoblot analysis of fractionated cell extracts reveal that NAT8L protein is indeed present in mitochondria along with its presence in cytosol presumably in ER (Figure 4.7). We have also observed the similar change in protein level in both mitochondrial and cytosolic fractions in SNP treated SH-SY5Y cells.

113

Figure 4.9 Immunohistochemical analysis revealed the lack of NAT8L protein in cuprizone-fed mice brains. Top panel, NeuN staining (left), NAT8L staining (middle) and merge (right), bottom panel, NeuN staining (left), NAT8L staining (middle) and merge (right). In the merged figures, there is almost no overlay between NeuN and NAT8L showing the absence of NAT8L protein in the neurons of cuprizone-fed mice brain.

In a separate experiment, we performed immunohistochemical analyses on the gray areas of mice brain tissue sections to analyze the change in NAT8L level in cuprizone-fed mice brain. As demonstrated in Figure 4.8, our findings showed a very less to no overlay of NAT8L staining to

NeuN staining clearly indicating the reduction of NAT8L protein in the neurons of cuprizone-fed mice brain.

4.3.7 NAT8L enzymatic activity is reduced in SNP treated neurons and cuprizone mice brain

Next, we aimed to analyze the impact of NAT8L reduction in the Asp-Nat catalytic activity of the enzyme in the neuronal cells and in the cuprizone mice brain. NAT8L enzyme catalyzes the transfer of acetyl group from acetyl CoA to aspartate resulting in NAA (Figure 4.9). It has been shown that, being a membrane bound protein, NAT8L and other NAT proteins preserve their catalytic activity in the total cell extract21. We used a previously established assay to analyze the catalysis of acetyl group transfer both in the cell extract and mice brain extract21. Interestingly we observed a reduction of enzymatic activity both in SNP treated neurons and cuprizone fed mice brain. The loss of catalytic activity was more pronounced in neuronal cells. This study implies that the lack of NAT8L is detrimental in brain NAA synthesis and there might not be an alternative mechanism that can synthesize brain NAA when NAT8L protein synthesis mechanism is wrecked.

114

Figure 4.9 N-Acetyl transferase activity (Asp-Nat) is compromised under MS environment, a. N- acetyl transferase reaction catalyzed by NAT8L results in the formation of NAA in human brain, b & c. N-acetyl transferase activity in the brain extract (b) or cell extract (c) is compromised in cuprizone mice brain and SNP treated SH-SY5Y cells.

4.3.8 There is a significant reduction of NAT8L protein in postmortem MS brain

After the observation of a reduced protein and a reduced Asp-Nat activity in MS cell model and

MS animal model, we investigated the presence of level of NAT8L protein in normal appearing

115 gray matter of MS postmortem brain. The immunoblot analysis in the protein extracted from three

MS and three control brain tissues revealed that there is indeed a nearly 50% reduction in the relative amount of NAT8L protein present in MS brain (Figure 4.10).

Figure 4.10 NAT8L protein level is reduced in normal appearing gray areas of MS postmortem brain, a. A representative immunoblot analysis showing a relative change in NAT8L level in MS vs non-MS postmortem brains and b. Quantification of average of three such blots.

4.3.9 In vitro translation of oxidized Nat8l mRNA show a reduced protein production

The presence of 8-OHG in mRNA has been shown to stall ribosome. Oxidized mRNAs produces truncated proteins. On top of that, oxidized mRNAs half-life is almost equal to its non-oxidized

116 counterpart in cellulo. These observations from different labs indicate a detrimental effect of Nat8l oxidation in protein production. To support the in cellulo and in vivo data connecting Nat8l mRNA oxidation and the reduced level of NAT8L protein detected, I performed a direct assay to show the effect of mRNA oxidation in translation. The effect of mRNA oxidation in NAT8L protein expression was investigated by using an in vitro translation assay. Nat8l mRNA was transcribed from the pPB-C-His-NAT8L vector. Transcribed mRNA was oxidized using Fenton reaction conditions as described in the Methods and Figure 4.11. The oxidized mRNA thus obtained was purified, quantified, and analyzed to determine the extent of oxidation (based on the presence of

8-OHG as measured by the quantification of 8-OHG immunoprecipitated mRNA). The amount of

RNA was normalized to non-oxidized RNA before translating them in vitro. We observed a 70% reduction in the amount of full length translated product compared to the protein from the non- oxidized mRNA (Figure 4.11).

117

Figure 4.11 In vitro translation showed the loss of translational ability of Nat8l mRNA upon oxidation. a. a schematic of an engineered vector that expresses human NAT8L protein (applied biomaterials), b. a schematic of generating oxidized and non-oxidized mRNAs for in vitro translation, c. western blot detection of in vitro translation product showed a reduction in NAT8L protein after mRNA oxidation, and d. quantification of three blots as represented in c.

118

4.3.10 Disruption of a potential G-quadruplex forming sequence (PGQS) in the coding region of Nat8l mRNA deteriorates protein synthesis

G-quadruplexes (GQs) are fairly stable DNA or RNA secondary structures formed by sequences rich in guanine22. These structures are implicated in many essential cellular processes, and the horizon of biological functions attributed to them continues to grow22. RNA GQs are known to play a context dependent role to regulate translation23. A general consensus in the field is that GQs or any other stable secondary structure forming regions are avoided in the open reading frames in mRNAs since the presence of stable structures in the coding region slows down the translation machinery24. Nevertheless, there exist substantial local secondary structures in both prokaryotes and eukaryotes25. Stable GQs are avoided evolutionarily in ORFs (unpublished data from our lab).

Interestingly human Nat8l mRNA is one of the exceptions. We observed that Nat8l mRNA is very

G-rich with a plethora of potential G-quadruplex forming sequences (PGQS) including one very strong sequence in the coding region. Interestingly a deletion of a 19-nucleotide long sequence from the very area is known to be responsible in an only known case of hypoacetylaspartia21. There are studies showing the relative instability of GQs under oxidative stress especially once 8-OHG is introduced in the sequence26,27. In an ongoing work we hypothesized that the disturbance in the

PGQS due to the presence of 8-OHG under oxidative stress could destabilize the GQ in Nat8l ORF and could disturb the regular translation machinery. This seemingly paradoxical idea looks to be working as postulated as we observed a reduction in protein synthesis (Figure 4.12) when we changed the mRNA sequence so that it loses PGQS. Nevertheless, this needs to be confirmed only after further experimentation to investigate the role of GQ in an apparent chaperon role.

119

Figure 4.12 G-quadruplexes in Nat8l mRNA could play regulatory role under oxidative stress. a. Nat8l mRNA (ORF and 3′-UTR) does have many PGQS. Notably the mutation in the ORF GQ area has been linked to hypoacetylaspartia and b. The mutation in the PGQS in ORF to disrupt GQ forming sequence is detrimental in Nat8l translation efficiency.

4.4 Discussion

NAT8L is a mitochondrial transmembrane protein with an enzymatic activity which is responsible in the synthesis of a key neuronal metabolite NAA21. There is a wide range of regulatory and

120 energetic mechanisms known or proposed to be connected to NAA metabolism in mammalian brain1-7,10,11,21. Looking at its central place in CNS and the concentration it is present in CNS, one could predict a higher level of monitoring mechanism for NAA synthesis. Very little is known about the regulation of NAT8L and the potential roles of NAA in brain. Looking at the high density of guanines present within the Nat8l mRNA and a high probability of GQ formation which could act as local regulatory elements, it is not unfair to predict a regulatory role of guanine oxidation in

Nat8l. Interestingly it turns out to be one of the highly oxidized mRNAs under MS environment.

Recent study has shown that oxidation of nucleotides could have a regulatory role. A study from

Burrows lab has demonstrated an epigenetic role of DNA guanine oxidation28. Wang et al. have shown that guanine oxidation in miRNAs result in mistargeting of mRNAs29. Tanaka et al. have demonstrated the critical impact of mRNA oxidation in translated products’ quality and quantity.

The critical role of NAT8L in CNS, the presence of a reduced level of NAA in the brain of patients with MS and other neurological disorders, and a selective oxidation of Nat8l mRNA led us to investigate the relationship between Nat8l mRNA oxidation and NAA production in CNS.

No change in mRNA expression both in cellulo and in vivo under MS environment but a significantly higher level of mRNA oxidation observed led us to critically think about the involvement of Nat8l mRNA oxidation in MS pathogenesis and overall mechanism of neurodegeneration. We expected to get a reduced level of NAT8L protein due to the damage in mRNA, but to our surprise, we observed almost 90% reduction in protein level via immunoblot analysis of cell extract. This finding corroborates well with the presence of a reduced level of NAA under neuropathological circumstances. We have also observed a 75% reduction in NAA level in

SNP treated neuronal cells. A reduced level of NAT8L protein or a defective NAT8L protein, both situations could be responsible in a reduced catalytic power of NAT8L, which eventually is

121 responsible for a reduced NAA in CNS. Next, we aimed to test the extent of damage Nat8l mRNA faces in MS animal model. Cuprizone fed mice model is a well-established toxicant induced animal model of MS. Cuprizone is known to mainly impact glial mitochondrial machinery (including oligodendrocytes, astrocytes, and microglia) which induces ROS production in brain. We have discovered that, there is a higher level of RNA oxidation in cuprizone mice brain. To our pleasant surprise we discovered that NAT8L expression is similarly affected in cuprizone mice brain as it behaved in SNP treated cells. We observed in mice brain that Nat8l mRNA faces a severe oxidative modification, but overall Nat8l mRNA level does not change. When we analyzed the protein level in the total extract isolated from gray area of mice brain it appeared that the protein level was suppressed by 80%. In a separate study, we also studied the Asp-Nat activity in the SNP treated cells and cuprizone fed mice brain, in both situations, we observed a reduced NAA synthesis clearly demonstrating a key role of NAT8L in NAA synthesis in neurons. Our result clearly established the link between higher ROS production, selective oxidation of Nat8l mRNA and its detrimental impact in protein production both in cellulo and in MS animal model.

Although there are relatively few studies to clearly understand the global role of Nat8L product

NAA in human brain, it has been shown to be a critical contributor of myelin production/myelin composition10,15 in oligodendrocytes especially in juvenile population along with its critical role in osmolarity maintenance3,4,12 and epigenetic signaling role where it regulates histone H3 methylation15. A reduced NAA concentration as detected in postmortem brain tissue as well as through MRI scanning in living patients is a hallmark of neurodegeneration in different neurological disorders including MS2-4,7,9,12-14,21. A reduced NAA is also a symbol of mitochondrial dysfunction and a poor neuronal health2,7,21.

122

Although we did not observe a global correlation between G content to a higher rate of mRNA oxidation in the high-throughput analysis, the presence of 36% G in the coding region and 3´-UTR of Nat8l is really a unique situation. In addition to a rocketing percentage of G, the mRNA sequence also consists of many potential GQ forming stretches (PGQS) including a very strong looking PGQS within the ORF. The deletion within the PGQS sequence in Nat8l ORF has been shown to have a deleterious effect in the only case reported in a patient with hypoacetylaspartia21.

Based on the reports about the weakening of GQs under oxidative stress, it could be possible that a disrupted GQ in the coding region could be detrimental for the stability of the resultant protein.

Though it needs to be tested by a tough experimentation, our unpublished data suggest that mutations in the ORF to abolish the potential GQ formation indeed results in the loss of protein production (Figure 4.12). Though it sounds ironic, a GQ structure in the coding region of Nat8l mRNA seems to have a crucial role in the protein production. We speculate that due to the presence of a stalling unit in the middle, the paused translating machinery could allow the nascent polypeptide a platform to fold properly, which might allow in the production of functional protein.

In addition to the PGQS in the ORF, as illustrated in Figure 4.12.a, Nat8l mRNA sequence have at least a dozen PGQS in the 3´ UTR. The GQs in the 3´-UTR of mRNAs have been suggested to play roles in mRNA stability. The disruption of one or more such GQ structures under oxidative stress could be responsible for mRNA instability or that could also affect the mRNA surveillance mechanism by disrupting RNP complexes designed to store translating machinery under stress.

Indeed, there are evidences of weakened stress granules under H2O2 mediated oxidative stress in

U2OS cells.

123

Figure 4.13 A proposed model of a link between Nat8l mRNA oxidation and MS neurodegeneration.

In accordance with the in cellulo and in vivo data, the in vitro mRNA oxidation-translation experiment showed that the oxidation in Nat8l mRNA results in the reduction in the NAT8L protein production. Next, we analyzed the presence of NAA in SNP treated cells and observed a

2-fold reduction in the NAA level in the extract isolated from the treated cells. Our finding clearly demonstrates the link between a higher Nat8l mRNA oxidation and a lower production of NAA in

MS circumstances which eventually could contribute in neurodegeneration. The NAT8L enzymatic activity in the cells extract demonstrated a reduction in the acetyl-transferase activity in the extracts isolated from both SNP treated neuronal cells and gray areas of cuprizone mice brain.

Previous studies have demonstrated that a reduced NAA concentration both in normal appearing white and gray matters of MS brain are a clear indicative of progressive MS9,14,15. Our discovery on the link between selective Nat8l mRNA oxidation to the low abundance of crucial neuronal metabolite NAA could provide a new direction in MS neurological disorder research. NAA concentration in brain is not only related to a healthy myelin production and neuroprotection in human CNS, it is also responsible for NAAG synthesis. NAAG is a key brain signaling molecule

124 especially playing role in the astrocytes which is known to be involved in pain signaling via glutamate receptors1,4,11. A suppression of Nat8l enzyme in neuron could also lead towards the reduced production of NAAG in neurons.

In a different context, a single nucleotide polymorphism (SNP) in the human NAT8L gene has shown to be related to reward dependence, a personality trait, and grey matter volume in the caudate nucleus in healthy subjects, suggesting that NAT8L might also affect human personality30.

The problem with NAT8L expression is detrimental in CNS but it is not lethal. Nat8l-/- mice look healthy but are weaker10. Their neurons do have myelin protection albeit a weaker one which could lead them toward developmental neuronal dysfunction25. On the other hand, the toxicant induced

MS model (cuprizone mice) show an extensively damaged myelination that corroborates well with a reduced NAA in cuprizone mice brain20. Cuprizone fed mice show more MS like phenotype than Nat8l -/- mice suggesting that the mice which born with no Nat8l sought other sources of acetate more easily for lipogenesis and myelin production. One would speculate that they must have to pay price for the overuse of other acetate source, but this needs further study before we can reach to a conclusion about an apparently healthy myelination in Nat8l-/- mice.

4.5 Conclusion

Based on our findings on the selective oxidation of Nat8l mRNA in MS neurons, the reduced amount of NA8L protein expressed under such environment and our in vitro data that show a direct inverse correlation between mRNA oxidation and protein expression of NAT8L, we conclude that

ROS induced oxidation of Nat8l mRNA results in the reduced production of NAT8L protein in

MS neurons hence debilitating the NAA metabolism. To the best of our knowledge our finding is

125 the first to address the direct pathophysiological connection between selective mRNA oxidative damage and MS neurodegeneration. Our finding opens up a new window to analyze the progression of MS and other neurological disorders through.

4.6 References

1 Miyamoto, Y. et al. Overexpression of Shati/Nat8l, an N-acetyltransferase, in the nucleus

accumbens attenuates the response to methamphetamine via activation of group II mGluRs in

mice. Intl J Neuropsychopharmacol 17, 1283-1294 (2014).

2 Moffett, J. R., Arun, P., Ariyannur, P. S. & Namboodiri, A. M. A. N-Acetylaspartate reductions

in brain injury: impact on post-injury neuroenergetics, lipid synthesis, and protein acetylation.

Front Neuroenerg 5, 11 (2013).

3 Baslow, M. H. J. N. R. N-Acetylaspartate in the Vertebrate Brain: Metabolism and Function.

Neurochem Res 28, 941-953 (2003).

4 Long, P. M., Moffett, J. R., Namboodiri, A. M., Viapiano, M. S., Lawler, S. E. & Jaworski, D. M.

N-acetylaspartate (NAA) and N-acetylaspartylglutamate (NAAG) promote growth and inhibit

differentiation of glioma stem-like cells. J Biol Chem 288, 26188-26200 (2013).

5 Madhavarao, C. N. et al. Defective N-acetylaspartate catabolism reduces brain acetate levels and

myelin lipid synthesis in Canavan's disease. PNAS USA 102, 5221-5226 (2005).

6 D'Adamo, A. F. & Yatsu, F. M. Acetate metabolism in the nervous system. N-acetyl aspartic acid

and the biosynthesis of brain lipids. J Neurochem 13, 961-965 (1966).

7 Moffett, J. R., Ross, B., Arun, P., Madhavarao, C. N. & Namboodiri, M. A. A. N-Acetylaspartate

in the CNS: From neurodiagnostics to neurobiology. Prog neurobiol 81, 89-131, (2007).

126

8 Sohn, J., Bannerman, P., Guo, F., Burns, T., Miers, L., Croteau, C., Singhal N. K., McDonough J.

A., Pleasure, D. Suppressing N-Acetyl-L-aspartate synthesis prevents loss of neurons in a murine

model of Canavan leukodystrophy. J Neurosci 37, 413-421 (2017).

9 Aboul-Enein, F., Krššák, M., Höftberger, R., Prayer, D. & Kristoferitsch, W. Reduced NAA-

levels in the NAWM of patients with MS is a feature of progression. A study with quantitative

Magnetic Resonance Spectroscopy at 3 Tesla. PLoS One 5, e11625 (2010).

10 Sumi, K., Uno, K., Noike, H., Tomohiro, T., Hatanaka, Y., Furukawa-Hibi, Y., Nabeshima, T.,

Miyamoto, Y. & Nitta, A. Behavioral impairment in SHATI/NAT8L knockout mice via

dysfunction of myelination development. Sci Rep 7, 16872 (2017).

11 Neale, J. H. N-Acetylaspartylglutamate (NAAG) is an agonist at mGluR3 in vivo and in vitro. J

Neurochem 119, 891-895 (2011).

12 Guo, Z. et al. Neurometabolic characteristics in the anterior cingulate gyrus of Alzheimer’s

disease patients with depression: a 1H magnetic resonance spectroscopy study. BMC Psych 15,

306 (2015).

13 Waragai, M., Moriya, M. & Nojo, T. Decreased N-acetyl aspartate/ myo-inositol ratio in the

posterior cingulate cortex shown by Magnetic Resonance Spectroscopy may be one of the risk

markers of preclinical Alzheimer’s disease: A 7-year follow-up study. J Alz Dis 60, 1411-1427

(2017).

14 Li, S., Clements, R., Sulak, M., Gregory, R., Freeman, E. & McDonough, J. Decreased NAA in

gray matter is correlated with decreased availability of acetate in white matter in postmortem

multiple sclerosis cortex. Neurochem Res 38, 1151-1158 (2013).

15 Singhal, N. K. et al. The neuronal metabolite NAA regulates histone H3 methylation in

oligodendrocytes and myelin lipid composition. Exp Brain Res 235, 279-292 (2017).

16 Nitta, A. et al. Nicotinic Acetylcholine Receptor Signaling in Neuroprotection (eds Akinori

Akaike, Shun Shimohama, & Yoshimi Misu) 89-111, Springer Singapore, 2018.

127

17 Fimognari, C. Role of oxidative RNA damage in chronic-degenerative diseases. Oxid Med Cell

Long 2015, 358713 (2015).

18 Tanaka, M., Chock, P. B. & Stadtman, E. R. Oxidized messenger RNA induces translation errors.

PNAS USA 104, 66-71, (2007).

19 Simms, C. L., Hudson, B. H., Mosior, J. W., Rangwala, A. S. & Zaher, H. S. An active role for

the ribosome in determining the fate of oxidized mRNA. Cell rep 9, 1256-1264 (2014).

20 Praet, J. et al. Cuprizone-induced demyelination and demyelination-associated inflammation

result in different proton magnetic resonance metabolite spectra. NMR Biomed 28, 505-513

(2015).

21 Wiame, E. et al. Molecular identification of aspartate N-acetyltransferase and its mutation in

hypoacetylaspartia. Biochem J 425, 127-136 (2009).

22 Bhattacharyya, D., Mirihana Arachchilage, G. & Basu, S. Metal cations in G-quadruplex folding

and stability. Front Chem 4, 38 (2016).

23 Bhattacharyya, D., Morris, M. J., Kharel, P., Mirihana, A. G., Fedeli, K. M., Basu, S. Engineered

domain swapping indicates context dependent functional role of RNA G-quadruplexes. Biochim

137, 147-150 (2017).

24 Klionsky, D. J., Skalnik, D. G. & Simoni, R. D. Differential translation of the genes encoding the

proton-translocating ATPase of Escherichia coli. J Biol Chem 261, 8096-8099 (1986).

25 Katz, L. & Burge, C. B. Widespread Selection for Local RNA Secondary Structure in Coding

Regions of Bacterial Genes. Genome Res 13, 2042-2051 (2003).

26 Lech, Christopher J. et al. Effects of site-specific guanine C8-modifications on an intramolecular

DNAG-quadruplex. Biophys J 101, 1987-1998 (2011).

27 Cheong, V. V., Heddi, B., Lech, C. J. & Phan, A. T. Xanthine and 8-oxoguanine in G-

quadruplexes: formation of a G·G·X·O tetrad. Nucleic Acids Res 43, 10506-10514 (2015).

28 Fleming, A. M., Ding, Y. & Burrows, C. J. Oxidative DNA damage is epigenetic by regulating

gene transcription via base excision repair. PNAS USA 114, 2604-2609 (2017).

128

29 Wang, J.-X. et al. Oxidative Modification of miR-184 Enables It to Target Bcl-xL and Bcl-w.

Mol Cell 59, 50-61 (2015).

30 Toriumi, K. et al. Deletion of SHATI/NAT8L increases dopamine D1 receptor on the cell surface

in the nucleus accumbens, accelerating methamphetamine dependence. Intl J

Neuropsychopharmacol 17, 443-453 (2014).

129

CHAPTER 5

5. Evidence of existence of NLRP3 inflammasome pathway in human neuronal cells which could be altered under SNP mediated oxidative stress

5.1 Introduction

Neuroinflammation is an important causative factor of multiple sclerosis (MS) and many other neurological disorders1,2. Neuroinflammatory autoimmune attacks are one of the key factors behind MS neurodegeneration1-3. The defense system in our body defends against many internal and external threats using innate immunity and adaptive immunity4. The innate immune response relies on pattern-recognition receptors (PRRs) to target pathogenic microbes and other endogenous or exogenous pathogens5. PRRs are expressed mainly in immune and inflammatory cells, such as, monocytes, neutrophils, and in microglia in the central nervous system (CNS)3,5. They present antigens to the adaptive immune system to generate long-lasting protection5. Pathogen-associated molecular patterns (PAMPs), which are antigens common to a given group of pathogens are normally recognized by at least three PRRs: Toll-like receptors (TLRs), C-type lectins (CTLs), and Galectins. The innate immune system is evolutionarily conserved across vertebrates and invertebrates6.

Neuroinflammation is a fundamental innate immune response in the CNS by which the brain and spinal cord react to diverse pathogens and host-derived signals of cellular damage7. The

130 inflammasomes are the new classes of PRR, first described only in 20026,8 and defined as caspase activating complexes. Several inflammasome complexes and inflammasome pathways have been identified since then, including NLRP1, NLRP2, NLRP3, double-stranded DNA (dsDNA) sensors absent in melanoma 2 (AIM2) and NLRC47-11. The best characterized is the NLRP3 inflammasome, so named because the NLRP3 protein in the complex belongs to the family of nucleotide-binding and oligomerization domain-like receptors (NLRs) and is also known as ‘pyrin domain-containing protein 3’ (P3)9. In addition to the NLRP3 protein, the NLRP3 inflammasome complex also contains adapter protein apoptosis-associated speck-like protein (ASC) and procaspase-1. Interactions among these three proteins tightly regulate inflammasome function in order to ensure immune activity only when appropriate12. In the absence of immune activators, an internal interaction occurs between the NACHT domain and leucine-rich repeats (LRRs), suppressing the interaction between NLRP3 and ASC, thus preventing assembly of the inflammasome6. In the presence of immune activators such as PAMPs, danger-associated molecular patterns (DAMPs), other exogenous invaders or environmental stress. NLRP3 opens up and allows interaction between the pyrin domains (PYDs) in NLRP3 and ASC6. Subsequently the caspase recruitment domain (CARD) of ASC binds to the CARD domain on procaspase-1, giving rise to the NLRP3 inflammasome (Figure 5.1)6,11-14.

Formation of this complex (or other similar complexes) triggers procaspase-1 (or other caspases) self-cleavage, generating the active caspase-1 p10/p20 tetramer and inducing the conversion of pro-inflammatory cytokines interleukin (IL)-1β and IL-18 from their immature ‘pro’ forms to the secretable active forms6,13. Formation of the inflammasome complex also triggers a process of inflammation-related cell death termed ‘pyroptosis’ which is characterized by rapid plasma- membrane rupture, DNA fragmentation, and the release of pro-inflammatory cytosolic contents

131 into the extracellular space15. Pyroptosis is both morphologically and mechanistically different from apoptosis and other forms of cell death15. Recent studies have identified that activated caspase-1 cleaves gasdermin D (GSDMD) to generate the gasdermin-N domain of GSDMD

(GSDMD-NT), which can directly bind phosphoinositides and cardiolipin14. GSDMD-NT then associates with the plasma membrane and oligomerizes to form non-selective pores, which triggers cell swelling and lysis14.

Figure 5.1 NLRP3 inflammasome pathway (Guo, Callaway and Ting, Nature Medicine, 2015, used with permission)11.

132

Both neurons and glial cells may trigger this process of cell death in response to a wide range of pathological stimuli7. These cytokines drive inflammatory responses through diverse downstream signaling pathways leading to neuronal damage. The important point of consideration is that, even if the inflammasome activation happened in the glial cells, it can still contribute in neurodegeneration via pyroptotic ejection of inflammatory material from the dying cells. Thus, the

NLRP3 inflammasome is considered a key contributor to the development of neuroinflammation.

Inflammasome related mRNAs are highly expressed in SNP stressed neuronal cells. Although, the presence of NLRP3 inflammasome in human neuronal cells seem debatable in literature, we observed an increase in Nlrp3 (and other inflammasome related mRNAs) transcript level in SNP treated neuronal cells. However, we did not observe a corresponding increase in the protein expression level suggesting mRNA oxidation of Nlrp3 is neutralizing its overexpression. RNA base oxidation could modulate gene expression mainly via two paths: i) a poor translation as observed by a reduced translational efficiency of oxidized mRNAs and ii) mis-target recognition resulting in a non-canonical RNAi interaction as evidenced from the recognition of non-canonical mRNAs due to the presence of 8-OHG in miRNAs. Based on these unique situations an oxidized

RNA could bring, we hypothesize that the selective mRNA oxidation can allow cells to choose a different path to signal the same information. We also seek for an alternative pathway that can signal the inflammatory responses when Nlrp3 transcript goes bad. Our data demonstrate the existence and activation of NLRP3 inflammasome in SHSY5Y cells under inflammatory circumstances. We also show the possibility of existence of an alternative inflammasome pathway under SNP mediated oxidative stress which need a more thorough study to exactly identify the components playing role in such pathway.

133

5.2. Materials and methods

5.2.1. Cell culture

Human neuroblastoma SHSY5Y cells, obtained from ATCC were cultured in 1:1 mixture of

Eagle’s Minimum Essential Medium (EMEM) and F-12 Ham’s medium supplemented with 10%

(v/v) fetal bovine serum, and penicillin/streptomycin antibiotics (0.01%) at 37°C under 5% CO2 in air. To determine the effect of lipopolysaccharide (LPS) on SH-SY5Y cells, SH-SY5Y cells were treated with various doses of LPS for 8h.

5.2.2 RNA isolation and immunoprecipitation

As described previously in 2.2.3.

5.2.3 RT-qPCR

Relative levels of mRNA for Nlrp3, Caspase 1, Caspase 4, Caspase 5, Nlrp1, Asc, Il18 and gapdh, were quantitated by qRT-PCR in mRNA isolated from SHSY5Y cells. qRT-PCR was performed with SYBR Green (Quanta Biosciences) and gene-specific primers. Relative levels of mRNA expression in control & treated cells and control & cuprizone mice were determined by the 2−ΔΔCt method after normalization to gapdh mRNA. The results are presented as ratios between the target gene and the GAPDH housekeeping mRNA.

134

Table 5.1. Primers for RT-qPCR analyses.

5.2.4 Western blotting

Cells were washed with ice-cold PBS and then lysed in RIPA buffer. After centrifugation at 5,000

× g for 15 min at 4◦C, the protein concentration was measured with a Bradford protein assay kit

(Santa Cruz). Lysates were separated using SDS-PAGE and transferred to polyvinylidene difluoride membranes (PVDF membranes). The membranes were blocked with 5% non-fat dry milk in Tris-buffered saline, pH 7.4, containing 0.05% Tween 20 (Sigma), and were incubated with primary rabbit- polyclonal antibodies (anti-NLRP3, anti-Caspase-1, anti-NLRP1, anti-

Caspase-5, anti-GAPDH) and horseradish peroxidase-conjugated secondary anti-rabbit antibodies

135 according to the manufacturer’s instructions. The protein of interest was visualized using Luminol reagent (Santa Cruz).

5.2.5 Caspase-1 glow assay

The Caspase-Glo-1 assay reagent (Promega) was used for caspase-1 detection in treated cells in vitro. The reagent provides a proluminescent caspase-1 substrate, which contains the tetrapeptide sequence YEVD, in combination with luciferase and a cell-lysing agent. The addition of the

Caspase-Glo-1 reagent directly to the assay well results in cell lysis, followed by caspase cleavage of the YEVD substrate, and the generation of luminescence. The amount of luminescence as displayed on the readout is proportional to the amount of caspase activity in the sample.

5.3 Results

5.3.1 NLRP3 inflammasome is activated in SNP treated neuronal cells

As detailed in Chapter 3, RNA-seq analysis of oxidatively damaged transcriptome from SNP stressed human neuronal cells revealed that Nlrp3 mRNA is one of the highly oxidized mRNAs.

By doing RT-qPCR analysis, we have also observed that Nlrp3 mRNA expression upregulates under SNP stress. This, apparently contradictory status of Nlrp3 led us to investigate more about the NLRP3 expression mechanism in human neurons under the stress situation. NLRP3 is a core activating molecule of NLRP3 inflammasome pathway, triggering of which signals caspase-1 activation downstream. To test the effect of SNP treatment on the Caspase-1 level, we analyzed the relative change in mRNA expression of Caspase-1 as well. As shown in Figure 5.2, we observed no apparent change. This was a surprising result based on nearly 3-fold upregulation of

Nlrp3 mRNA level. Figure5.3 summarizes the relative change in expression of other RNAs critical

136 in inflammasome activation along with their relative population distribution in oxidized vs non- oxidized mRNA pool. This clearly demonstrates that Nlrp3 is the one which gets a maximum oxidative impact under the SNP induced oxidative stress.

As mentioned in the literature the existence of NLRP3 inflammasome in human neurons is debatable. There are reports both in support and against to the existence of NLRP3 inflammasome activation in the neurons of different neurological disorders. In general, NLRP3 inflammasome is activated in response to bacterial toxins. Lipopolysaccharides (LPS) are the most commonly used chemicals to induce NLRP3 inflammasome in different cell types to study the mechanism of inflammation signaling. Before proceeding to investigate the impact of SNP treatment in NLRP3 expression in SHSY5Y cells, we wanted to prove the existence of this pathway by using LPS as an inflammasome activating agent. By RT-qPCR analysis of RNA extracted after LPS treatment, we observed that the Nlrp3 expression goes up in response to a higher concentration of LPS clearly supporting the expression of NLRP3 we observed in SNP treated cells (Figure 5.3).

137

Figure 5.2 A comprehensive RT-qPCR analysis to look at a. the relative change in respective mRNA expression in SHSY5Y cells after SNP treatment, numbers below x-axis, b. relative change in respective mRNA oxidation in SHSY5Y cells after SNP treatment, black bars and c. non- oxidized mRNA level under the same condition, red bars.

Figure 5.3 LPS treatment induces NLRP3 inflammasome activation in human neuronal cells.

5.3.2 mRNA oxidation leads towards the reduced protein expression of NLRP3

Next, we wanted to analyze the change in protein level of NLRP3 and Caspase-1 proteins using western blot analysis. Based on our data of the loss of RNA function after the oxidation both in vitro and in vivo (Chapters 4 and 6), we speculate that NLRP3 RNA oxidation can result in the inhibition of its downstream function. As shown in Figure 5.4 we observed no significant change

138 in the protein level of NLRP3 and its downstream pyroptosis signaling molecule Caspase-1.

Apparently, this observation is in contrast to the upregulation of mRNA expression for Nlrp3, but it does make sense if we consider the fact that Nlrp3 is a highly oxidized mRNA. This implies that oxidation in the Nlrp3 mRNA could be responsible for the compromised NLRP3 inflammasome activation in the human neuronal cells under oxidative stress.

Figure 5.4. The protein expression of NLRP3 and Caspase-1 does not change significantly after SNP treatment a. A representative western blot image to show the protein expression, b. Quantification of relative change in protein expression (3 independent experiments from a), and c. A schematic explaining the Nlrp3 expression situation in neuronal cells under stress.

5.3.3 Alternative inflammasome pathway in SNP treated neuronal cells

After the observation of quite contradicting inflammasome component expression in the SNP treated cells, we were eager the analyze the possibility of existence of alternative inflammasome pathway in the neurons under such conditions. Alternative inflammasome to NLRP3 can be expressed in mainly two ways. Firstly, NLRP3 inflammasome could still be active but it can use

139 an alternative caspase molecule as downstream inflammation signaling molecules, in such case,

Caspase 4 and 5 are the mostly studied candidates. The second situation could be a complete alternative, where the pathway other than NLRP3 could be activated (NLRP1, AIM3, etc.) using

Caspase 1 or other caspases as signaling molecules. To test these ideas, we analyzed the presence of activated inflammatory caspase molecules in the SNP treated SH-SY5Y cells. As detailed in

Figure 5.5, by using Inflammasome luciferase assay for Caspase 1, we have observed that the relative level of Caspase activities in SNP treated neurons go up by at least 15 folds. The remarkable change in the caspase luciferase activity was not suppressed much after the use of

Caspase 1 inhibitor molecule indicating the activation of other caspase molecules to account for the observed luciferase activity.

140

Figure 5.4 A high level of caspase activity in the SNP treated cells. a. A schematics of working principle of Caspase-1 detection system used in this study, b. A luciferase reaction catalyzed by active caspase-1, and c. SNP treated SHSY-5Y cells do have active caspase activity other than Caspase-1 as evident by the persistent activity even after the treatment of Caspase-1 inhibitor molecule.

141

Next, we wanted to study the expression level of NLRP1 under the same stress situation. NLRP1 has not only been known to be activated in human neurons but it has also been shown to work as an alternative to NLRP3 inflammasome pathway. Caspase-4 or 5 are shown to be involved as alternatives for Caspase-1 to signal inflammation and pyroptosis under certain circumstances. Both

NLRP3 and NLRP1 can use Caspase-5 as a non-canonical pyroptosis/inflammation signaling molecule. Our preliminary data show that although relative mRNA and protein expression of

NLRP1 does not change significantly, the production of matured Caspase-5 significantly goes up under SNP treatment demonstrating NLRP1- Caspase 5 could be an alternative signaling pathway of inflammasome signaling system when NLRP3 faces roadblocks like mRNA oxidation.

Figure 5.5 Inflammation signaling via NLRP3 pathway got disturbed under SNP mediated oxidative stress in human neuronal cells.

142

5.4 Discussion The NLRP3 inflammasome is the most extensively investigated inflammasome, and it is present central nervous system (CNS)6,7,10,16. Mutations in NLRP3 gene lead to several autoinflammatory disorders referred to as the cryopyrin-associated periodic syndromes17. NLRP3 inflammasome pathway has been shown to be activated in MS brain and MS animal models like cuprizone fed mouse model and EAE mouse model18. The cuprizone model is associated with demyelination of neuronal axons is NLRP3 pathway dependent process 16. Although, the activation of NLRP3 in microglia and monocytes is clearly documented, it remains debatable whether neurons express

NLRP36,7,16. In vitro studies suggest that the basal level of NLRP3 in resting cells is not enough to activate the inflammasome. It is widely accepted that successful NLRP3 inflammasome activation requires a two-checkpoint signal process10,19. A priming signal is provided by the NF-κB- activating stimuli to transcriptionally enhance the expression of NLRP3 and pro-IL-1β10. Many

TLR and NLR ligands, as well as endogenous cytokines such as IL-1α, have been demonstrated to prime cells10. The subsequent activating signal is provided by various NLRP3-activating agents to promote the formation of the inflammasome complex. A wide range of exogenous and endogenous stimuli including PAMPs, aggregated and misfolded proteins, ATP and crystalline substances induce NLRP3 activation. Among many other factors NLRP3 has been suggested to be activated by mitochondrial ROS production7.

Inflammasome activation is a complex mechanistic cellular effort that requires a proper interaction of multiple protein components17. We observed a very interesting change in NLRP3 expression in

SH-SY5Y cells after SNP treatment. The upregulation of Nlrp3 in the transcript level without translating the same change in the protein level indicates translational arrest of NLRP3 expression.

Interestingly relatively a higher level of mRNA oxidation as we observed, both in RNA-seq

143 analysis of oxidized mRNA (Chapter 3) and RT-qPCR analyses of immunoprecipitated mRNA from SNP treated SH-SY5Y cells (Figure 5.2) indicate a link between mRNA oxidation with the defective translation of Nlrp3 mRNA. The selective oxidation of Nlrp3 mRNA could be detrimental in inflammasome activation by reducing the amount of protein necessary to overcome the basal level or by producing the defective protein that will not be able to initiate or complete the protein oligomerization necessary to build the active complex. The inactivation of NLRP3 inflammasome pathway thus happened could be harmful as suggested by the mutations in NLRP3 in syndrome play role in autoimmune attacks in cryopyrin-associated periodic syndrome17. This could also direct the neuronal cell machinery to seek for an alternative inflammasome route. Under the SNP mediated oxidative stress, the cells’ attempt to activate NLRP3 pathway but their inability to do so completely, seem to be very costly energetically. Nevertheless, our observation of the activation of other inflammatory caspase (Caspase 5) under SNP stress indicates the activation of an alternative inflammasome pathway. Although it needs to be verified through further experimentation to analyze the components of the inflammasome complex and downstream processes, the activation of NLRP1 inflammasome and/or the participation of Caspase-5 as downstream singling molecule is plausible in the neurons under neurodegenerative circumstances.

Indeed, there are evidences of active NLRP1 inflammasome pathway in the neurons of many neurological disorders including MS.

In addition to the inability of oxidatively modified mRNA molecules to get translated properly, oxidation could impact the gene expression via the influence in RNA interference mechanism20.

There are evidences suggesting the presence of oxidized guanosine in miRNA make them misidentify the targets20. The future work to clearly understand the impact of Nlrp3 mRNA

144 oxidation should also focus to study whether oxidation of the nucleotides in the 3´-UTR of the mRNA mitigates its binding to regulatory miRNAs, for example miRNA-223.

In human monocytes, the alternative inflammasome consisting of a non-pyroptotic signaling does occur which goes by IL-18 maturation pathway21. There could be similar pathways active in neurons with or without the involvement of NLRP3. Although, it seems highly possible, we need more work to conclude the existence of the alternative or non-canonical inflammasome pathway in human neurons under oxidative stress.

5.5 Conclusion Here we have demonstrated the activation of NLRP3 inflammasome pathway in human neurons under SNP induced oxidative stress. We have shown that SNP induced oxidative stress could affect inflammasome signaling pathway. Especially, the selective oxidation of Nlrp3 mRNA could be detrimental for NLRP3 inflammasome activation, which eventually signals for an alternative inflammasome pathway most probably one through the involvement of NLRP1 and Caspase 5.

5.6 References 1 Chen, W.-W., Zhang, X. I. A. & Huang, W.-J. Role of neuroinflammation in neurodegenerative

diseases. Mol Med Rep 13, 3391-3396 (2016).

2 Lassmann, H. What drives disease in multiple sclerosis: Inflammation or neurodegeneration? Clin

Exp Neuroimmunol 1, 2-11 (2010).

3 Kempuraj, D. et al. Neuroinflammation induces neurodegeneration. J neurol neurosurg spine 1,

1003 (2016).

4 Chaplin, D. D. Overview of the immune response. J allerg clin immunol 125, S3-23 (2010).

145

5 Mogensen, T. H. Pathogen recognition and inflammatory signaling in innate immune defenses.

Clin Microbiol Rev 22, 240-273 (2009).

6 Shao, B.-Z., Xu, Z.-Q., Han, B.-Z., Su, D.-F. & Liu, C. NLRP3 inflammasome and its inhibitors:

a review. Front Pharmacol 6, 1-15 (2015).

7 Song, L., Pei, L., Yao, S., Wu, Y. & Shang, Y. NLRP3 Inflammasome in neurological diseases,

from functions to therapies. Front Cell Neurosci 11, 63 (2017).

8 Martinon, F., Burns, K. & Tschopp, J. The Inflammasome: A molecular platform triggering

activation of inflammatory caspases and processing of proIL-β. Mol Cell 10, 417-426 (2002).

9 Khare, S., Luc, N., Dorfleutner, A. & Stehlik, C. Inflammasomes and their activation. Critic rev

immunol 30, 463-487 (2010).

10 He, Y., Hara, H. & Núñez, G. Mechanism and regulation of NLRP3 inflammasome activation.

Trends biochem sci 41, 1012-1021 (2016).

11 Guo, H., Callaway, J. B. & Ting, J. P. Y. Inflammasomes: Mechanism of action, role in disease,

and therapeutics. Nature med 21, 677-687 (2015).

12 Guarda, G. & So, A. Regulation of inflammasome activity. Immunol 130, 329-336 (2010).

13 Bauernfeind, F. & Hornung, V. Of inflammasomes and pathogens – sensing of microbes by the

inflammasome. EMBO Mol Med 5, 814-826 (2013).

14 Liu, X. et al. Inflammasome-activated gasdermin D causes pyroptosis by forming membrane

pores. Nature 535, 153 (2016).

15 Bergsbaken, T., Fink, S. L. & Cookson, B. T. Pyroptosis: host cell death and inflammation.

Nature rev. Microbiol 7, 99-109 (2009).

16 Jha, S. et al. The Inflammasome sensor, NLRP3, regulates CNS inflammation and demyelination

via caspase-1 and interleukin-18. J Neurosci 30, 15811-15820 (2010).

17 Broderick, L., Nardo, D. D., Franklin, B. S., Hoffman, H. M. & Latz, E. The Inflammasomes and

autoinflammatory syndromes. Ann Rev Pathol 10, 395-424 (2015).

146

18 Inoue, M. & Shinohara, M. L. NLRP3 Inflammasome and MS/EAE. Autoimm Dis 2013, 859145

(2013).

19 Latz, E., Xiao, T. S. & Stutz, A. Activation and regulation of the inflammasomes. Nature rev

Immunol 13 (2013).

20 Wang, J.-X. et al. Oxidative modification of miR-184 enables it to target Bcl-xL and Bcl-w. Mol

cell 59, 50-61 (2015).

21 Gaidt, Moritz M. et al. Human monocytes engage an alternative inflammasome pathway.

Immunity 44, 833-846 (2016).

147

CHAPTER 6

6. RNA oxidation impairs their function: Synthesis of 8-hydroxyguanosine phosphorothioate analog with a potential to probe the interference of RNA oxidation in structure and function

6.1 Introduction

The sequence and structure of RNA are known to have self-regulatory effects, influencing the splicing, translation, cellular localization and longevity of the RNA life1. Both the sequence and structure of a molecule of RNA can be changed in its lifetime1. Self-regulation of RNA has been well studied in riboswitches2, which are fragments of RNA with a secondary structure that can be modulated by small molecules to alter splicing, translation and RNA stability and in ribozymes3, which are the catalytic RNA molecules which regulate their own splicing4. Structural alterations to the nucleobases have the ability to change functional RNA molecules as has long been known through modifications in transfer RNA and ribosomal RNA molecules1. Unlike highly structured

RNAs such as self-splicing introns, riboswitches, spliceosomes, and ribosomes, neither the coding region nor the UTRs of mRNAs are evolved to adopt single, well-defined structures. Instead they adopt an ensemble of conformations best described by a partition function, which is defined as the probabilities of all possible base-pairs5,6. Coding regions are evolutionarily evolved to have a minimum to no secondary structural elements6. Most mutations/modifications in an RNA only have local effects on the structural ensemble. A small subset of mutations, however, can have a

148 large and global effect, sometimes a deleterious one5,7. If a disease associated mutation/ modification belongs to the larger global effect, it can suggest a role for RNA structure in the molecular mechanism of the disease5. Nevertheless, there are evidences of involvement of single nucleotide polymorphism in the UTR of certain mRNAs in disease development including cancer and neurological disorders5,8. Under oxidative stress, the production of reactive oxygen species

(ROS) inside the cells, makes various adducts in biomolecules including oxidation of RNA bases.

As described in chapter 1 in detail, some of the oxidized bases including 8-OHG can have mutagenic impacts. In this context, the mutation due to RNA oxidation, though is not genetically coded, could still be responsible in the development/ progression of a wide range of neurological disorders. Recently, a lot of effort is being put to understand the effect of intentional and accidental

RNA base mutations in their base pairing paradigm9.In mRNA, the nucleoside modifications that are associated with are 5′- methyl cap modification, 6-methyl adenosine, and pseudo uridine modification10,11. These modifications are driven by enzymes and have been shown to impact RNA biology and cell biology overall through genetic, epigenetic and transcriptional regulations10.

Similarly, RNA base modification has been shown to impact the structure and folding dynamics of RNA in vitro8,11. Although, not regulated by enzymes, RNA oxidative modification, especially,

8-hydroxyguanine (8-OHG) formation is the most common RNA oxidative base modification12 with a great possibility to induce mutation and RNA structure defect.

To study the impact of oxidative stress in overall RNA oxidation and to evaluate its downstream effect (RNA function) we aimed to work on a fundamentally characterized RNA system namely

Tetrahymena Group I intron, an engineered version of which is known as L-21G414 ribozyme13 and its self-folding RNA domain called P4-P6 RNA7. Tetrahymena Group I intron is a ribozyme which catalyzes a self-cleavage reaction (splicing) to form a matured intervening sequence (IVS)

149

(Figure 6.1.b)14. This intron has been reengineered for handling purpose to be converted in L-

21G414 ribozyme molecule, which contains all the necessary components from the intron plus integrated guanine cofactor as G41415. The engineered ribozyme L-21 G414 catalyzes the reverse of second step of Tetrahymena Group I intron reaction15. The reengineered ribozyme allows mapping the catalytic activity of RNA by analyzing the differences between the label transfer reactions amongst different conditions (Figure 6.1c). Roughly 90% of the Tetrahymena pre-RNA misfolds under in vitro physiological conditions16. The refolding of misfolded pre-rRNAs is a very slow process in vitro resulting in a reduced splicing activity16. We hypothesized that oxidative stress induced formation of 8-OHG in the ribozyme molecule impacts RNA folding or refolding process thereby debilitating the ribozyme function of the RNA.

150

Figure 6.1.a. A secondary structure of Tetrahymena Group I Intron ribozyme (adapted from Guo, Gooding and Cech, RNA, 2006), b. reaction catalyzed by the ribozyme17 and c. mechanism of action of engineered L-21G414.

Assuming the impact of the presence of 8-OHG in the sequence could impact the H-bonding and

RNA-tertiary structure, we also hypothesized that severe RNA oxidation results in RNA misfolding/unfolding. To test our hypothesis, we worked in a self-folding middle domain of

Tetrahymena Group I ribozyme, known as P4-P6 RNA. We were motivated to pick these systems to analyze the impact of RNA oxidation in their structure function relationship due to their established folding/function relationship and relatively easier functional assays they offer.

To study the impacts of 8-OHG modifications on RNA folding in atomic level, we planned to incorporate the 8-OHG moieties in the L-21 G414 ribozyme and P4-P6 RNA molecules and analyze at the effects of incorporation of this modification by using Nucleotide analog interference mapping (NAIM) technique18,19. Nucleotide phosphorothioates are analogs of nucleotides in which a non-bridging oxygen atom of a phosphate group is replaced by sulfur atom19. We aim to synthesize a novel 8-OHG analog molecule 8-OHGTPαS (a triphosphate salt of phosphorothioate modification of 8-hydroxyguanosine). The synthesis of this compound could allow us to study the effect of the presence of 8-OHG in RNA structure mainly because after the incorporation of 8-

OHGTPαS in the transcript, the RNA molecules can be cleaved right from this site with iodine treatment and hence can be mapped easily by using NAIM technique18,19. After the successful synthesis of the proposed compound, we found that the incorporation of this modification in RNA is tough due to the high selectivity of the enzyme used to transcribe the RNA (T7 RNA polymerase). It has been known that the mutations of just one or two amino acids in the catalytic

151 center of T7 RNA polymerase enzyme greatly enhance the incorporation of modified nucleotides in RNA during transcription20-22. Based on the literature reports20,22, we hypothesize that the same mutations can help in 8-OHG incorporation. We have mutated the T7 polymerase enzyme in these two crucial positions (namely Y639F and H784A) so that the enzyme could potentially be less selective towards the modified nucleotide during transcription and we could get a good incorporation in the RNA and can look at the effect of this modification in RNA folding/ RNA activity. We have got some success in the incorporation of 8-OHGTPαS in P4-P6 RNA. The optimization of the incorporation and NAIM analysis of selected (functional vs non-functional)

RNA species could give a detailed idea about the involvement of 8-OHG in RNA structure and function.

Figure 6.2.a. A crystal structure of P4-P6 RNA (PDB ID 1GID), b. a part of Tetrahymena ribozyme crystal showing tertiary interactions between P4-P6 and P3-P9 domains to show the non-Watson- Crick faces’ interaction responsible in proper folding of the RNA (adapted from Guo, Gooding and Cech, RNA, 2006) and c. N-7 position of guanine (the electron localization of which is disturbed in 8-OHG) plays a critical role in tertiary interactions and hence could be crucial for

152 proper RNA folding and/or the catalytic activity of the folded ribozyme, top: coordination of the monovalent cation within the binding pocket of the tetraloop receptor- the metal ion is shown in orange and the metal-ligand interactions are shown in black (adapted from Basu et al)23 and bottom: the recognition of guanine as a cofactor for Tetrahymena ribozyme activity also involves

N7 (G264) to NH2 (cofactor G) interaction.

6.2 Materials and methods

6.2.1 Transcription reactions

L-21 G414 RNA (Figure 6.1) was transcribed from EarI digested pUC19L-21G414R by T7 RNA polymerase using 1 mM of each NTP. P4-P6 RNA (Figure 6.2) was transcribed from BSPHI digested pUC19-P4P6 plasmid by WT T7 RNA polymerase and its mutated versions. Following transcription, the RNAs were purified by PAGE and eluted into 10 mM Tris-HCl, pH 7.4/1 mM

EDTA (TE buffer). The RNAs were ethanol precipitated and resuspended in TE buffer.

6.2.2 Fenton reaction to get oxidized transcript

RNAs were oxidized using Fenton reaction. 50 picomoles of each RNA was 5´-end labelled using

32 P-γ-labelled ATP. Labelled RNA was incubated with a variable H2O2 concentration along with

10 µM sodium ascorbate and different concentrations of ferrous sulfate (as summarized in Figure

6.3 and 6.4) for an hour at 37 C. The reaction was quenched by using 20 µM DFOM and passed through a G-25 column and or purified by denaturing gel electrophoresis. The purified RNA was gel extracted using 300mM NaCl in TE buffer and ethanol precipitated before using for the functional assays.

6.2.3 HPLC estimation of level of oxidized RNA

153

The extent of RNA oxidation after Fenton chemistry was assayed by HPLC-UV analysis of 8-

OHG in RNA. The RNA was cleaved to nucleoside level using Nuclease P1 and CIP enzymes before HPLC analysis.

6.2.5 Ribozyme reaction

The L-21G414 RNA performs RNA ligation reactions that is analogous to the reverse of second step of Tetrahymena Group I intron RNA splicing. A typical ligation reaction with the L-21 G414 intron is performed as follows: The RNAs are prefolded by incubating in reaction buffer at 50°C for 10 min. After a quick spin, the RNAs are further heated for 2 min at 50°C, at this time the substrate dT(-1)S [CCCUC(dT)AAAAA], radiolabeled at 3'-end by yeast poly(A) polymerase

(NEB) and [α-32P] cordycepin (PerkinElmer), is dissolved in the same buffer and heated at 50°C for 2 min. After 2 min, dT(-1)S substrate, equal in volume to the ribozyme reaction, is added with proper mixing. The reaction is stopped after 10 min by adding 2 vol of loading buffer. The reaction was quenched by the addition of two volumes of stop solution (8 M urea/50 mM EDTA/0.01% bromophenol blue/0.01% xylene cyanol). The products were separated on a 5% polyacrylamide gel and imaged using PhosphorImager. The intensities of individual bands were quantitated by

ImageJ.

6.2.6 RNA folding

The P4-P6 RNAs were purified by PAGE (6% denaturing) and 5'-end labeled. Folding of P4-P6 were performed by incubating the RNAs in buffer containing 33 mM Tris-base, 66 mM HEPES, pH 7.4, 2 mM Mg(OAc)2, and 10% glycerol at 70° C for 10 min, followed by slow cooling to room temperature. Gel shift assay was used to analyze the folded and unfolded population of by 8% non-

154 denaturing PAGE in 33 mM Tris, 66 mM HEPES, pH 7.4, 0.1 mM EDTA and 2 mM Mg(OAc)2 at 4–6 °C.

6.2.7 Radiolabeling of RNA

The substrate of L-21 G414 ribozyme dT(-1)S CCCUC(dT)AAAAA was 3′ end labeled with [α-

32P] cordycepin using poly(A) polymerase. L-21 G414 and P4-P6 RNAs were reacted with CIP and 5′ end labeled with [γ-32P] ATP using T4 polynucleotide kinase. All oligonucleotides were purified on a denaturing polyacrylamide gel prior to use and eluted into TE buffer.

6.2.8 Synthesis of the novel phosphorothioate nucleotide

The 5'-O-(1-thio)-8-hydroxy-guanosine triphosphate nucleotide (8-OHGTPαS) was prepared from

8-hydroxyguanosine (8-OHG). The synthesis involves a one-pot, two-step reaction, adopted from previous methods for other phosphorothioates. 50 mg (0.167 mmoles) of 8-OHG was transferred to a 25 mL round-bottomed flask, and co-evaporated with 3 × 10 mL of anhydrous pyridine to convert it to a pyridinium salt. The nucleoside is further co-evaporated with 5 mL of toluene to remove any residual pyridine. After drying, the flask was capped with a septum, and purged with argon. The nucleoside was then dissolved in a 0.5 mL of triethylphosphate (TEP). The mixture was heated with air gun for 30 minutes to facilitate the dissolution. To the nucleoside, 1.5 equivalent of trioctylamine (Sigma Aldrich, MW 353.7, density 0.816 g/mL,176 μL) and 1.1 equivalent of PSCl3 (Sigma Aldrich, 45 μL, MW 169.4, density 1.67 g/mL) were added. The reaction was stirred for an hour under argon at room temperature. The progress of the reaction was monitored by analyzing a small portion of the reaction via TLC. Once the reaction was more towards the product side, the reaction was quenched with water, and the products were resolved by cellulose TLC in a solvent system of 0.5 M LiCl in water. The products on the TLC plate were

155 visualized by UV light. The monophosphate intermediate migrates slowly than the starting material. To the intermediate 5'-O-(1-thio-1, 1-dicloro) phosphoryl 8-hydroxy guanosine formed at the first step, tetra-butyl ammonium pyrophosphate, TBAP (Sigma Aldrich, MW 451.5, 4 equivalents of TBAP per mole of PSCl3) solution was added and stirred for 30 min at room temperature. The TBAP solution was prepared in anhydrous triethyl phosphate (TEP) at a concentration of 0.1 g/ mL. 50.0 μL of the reaction was removed and quenched by adding a few drops of triethylamine (TEA). The TEA precipitates the phosphates, which was collected by centrifuging the mixture @ 3000 rpm at 4° C and dissolved in 50 mM triethyl ammonium bicarbonate (TEAB), and the products were analyzed by silica TLC using a solvent system containing n-propanol: ammonium hydroxide: water (6:3:1, v/v). TLC analysis showed at least two products, and both of these migrate slower than the unreacted nucleoside precursor and the nucleoside monophosphate. One of the products is the desired triphosphate and the other is the cyclic triphosphate, which hydrolyzes to the linear triphosphate after 12 hours at room temperature.

After the reaction is over, the triphosphates are precipitated by adding about 50 equivalents of TEA

(MW 101.19, density 0.726 g/mL, calculated per mol of the nucleoside precursor). The resulting precipitate is collected by centrifugation and dissolved in 5 mL of 50 mM TEAB and kept overnight at room temperature. The nucleoside triphosphate is purified by DEAE Sephadex chromatography using a linear gradient from 50–1000 mM TEAB in a total volume of 1 L The fractions (approx 15 mL each) were checked for the presence of the nucleoside by UV measurements (abs @ 260 nm and 295 nm). Positive fractions were pooled as a batch of 2–3, and lyophilized. The reaction products were characterized by 31P NMR and mass spectrometry. The nucleotide analog amount is quantitated by measuring the UV absorbance @ 295 nm. The yield was 20%.

156

OHGTPαS.

- Figure 6.3 Synthesis and purification of 8 Synthesis 6.3 of and purification Figure

157

6.2.9 Mutagenesis pUC19L-21G414R plasmid (part of it is shown in the color text below) was mutated to get pUC19-

P4P6 plasmid, BSHP1 enzyme digestion of which gives p4-p6 RNA domain after transcription.

Wild type (WT) T7 polymerase was mutated to get three mutated variants: Y639F, H784A and a double mutant using following Agilent mutagenesis protocol (Lightening Site directed mutagenesis kit, Agilent).

Table 6.1 Primers used for P4-P6 and T7 polymerase mutagenesis.

158

Figure 6.4 Mutagenesis to get P4-P6 RNA and mutated T7 polymerase. a. Mutation in pUC19L-

21G414R plasmid to get pUC19L-21G414R plasmid (green color represents the P4-P6 RNA). b.

T7 polymerase (WT) sequence (mutated amino acids are highlighted).

159

6.3 Results

6.3.1 RNA oxidation results in loss of L-21G414 ribozyme function

We wanted to investigate the impact of RNA oxidation in the catalytic activity of a ribozyme. L-

21G414 RNA provides an easier platform to analyze the catalytic activity in presence of different modified bases in RNA. In this ribozyme reaction, the terminal guanosine (G414) attacks the oligonucleotide substrate via SN2 mechanism mimicking the reverse of the second step of ribozyme reaction and transfers the 3'-exon covalently to the 3'-end of the ribozyme. By using a

3'-end-radiolabeled substrate, the active variants among the ribozyme pool selectively label themselves (Figure 6 1.c). To test the impact of RNA base oxidation, we used the very system thereby we can track the fall in ribozyme activity by tracing the label transfer in the enzyme from the labelled substrate.

The oxidation reaction was optimized such that there would be enough oxidation but not much cleavage of the product. The extent of oxidation was established by HPLC-UV and with immunoprecipitation with anti 8-OHG antibody. As one would expect, we observed an increase in

8-OHG level with the increase in the concentration of H2O2 in Fenton reaction. The oxidized products were purified first by passing through desalting columns and using denaturing gel electrophoresis to make sure that the oxidized RNA enzyme molecules are as pure as the wild type

(WT) ones. The ribozyme assay was performed on the WT and oxidized L-21 G414 ribozymes as represented in Figure 6.3. Interestingly we observed a reduced enzymatic activity of the ribozyme with the increase in the severity of oxidative stress (Figure 6.4). These results clearly indicate that

RNA oxidation results in the loss of RNA function.

160

Figure 6.5 RNA oxidation impairs L-21 G414 ribozyme function. a. Top: PAGE analysis of the reaction between WT or oxidized ribozymes with the labelled substrate showed a reduced product formation with the increase in RNA oxidation indicating the availability of a smaller number of active ribozymes under oxidative stress, bottom: Fenton reaction conditions used to oxidize the RNA and b. Quantification of the labelled products from three independent experiments (as shown in a).

6.3.2. RNA oxidation disrupts RNA folding

As obvious from the structural and electronic change in 8-OHG due to oxidation, we speculated that the loss of ribozyme activity could be due to the alternation in the tertiary interactions in the

RNA molecules which are crucial for the stability and proper folding of structured RNA molecules.

If this is the case, we predicted that the change in the tertiary interactions due to the oxidation of

161 critical guanine residues (and oxidation of other bases) can result in RNA unfolding and/or misfolding. To investigate the impact of oxidative stress in RNA folding, we used P4-P6 RNA as a model molecule. The RNA was oxidized using a variable Fenton reaction conditions (Figure 6.5) and folded in the folding buffer. Absolute care was given to remove the residual cations and purify the oxidized molecules before performing the folding assay. The resulting folding mixture was separated using non-denaturing gel electrophoresis. We observed an increase in the relative amount of unfolded population under the oxidative stress. Our data also suggest that the intent of unfolding increases with the increase in the level of oxidation in the RNA molecules (Figure 6.5).

Figure 6.6 RNA oxidation results in the unfolding of P4-P6 RNA. a. Fenton reaction conditions used to study the impact of RNA oxidation in P4-P6 RNA folding, b. Non-denaturing PAGE analysis showed the amount of unfolded species increases with higher oxidation, and c. Quantification of three gel images from independent experiments show a significant reduction in the folded species with increase in RNA oxidation.

162

6.3.3 Synthesis of a novel phosphorothioate compound (8-OHGTPS)

Although extensive RNA oxidation has been observed in many neurological disorders, its effect on the RNA folding has never been directly studied at least to the best of our knowledge. In this study we have observed the detrimental effect of RNA oxidation in RNA folding and function. To study the impacts of 8-OHG modifications on RNA folding in the base/ atomic level, we plan to incorporate the phosphorothioate modification of 8-OHG molecules (8-OHGTPαS) in the functional RNA molecules to look at the effects of this modification by using NAIM technique.

The synthesis of 8-OHGTPαS is very crucial to study the effect of 8-OHG in RNA structure mainly because after the incorporation of 8-OHGTPαS, the RNA molecules can be cleaved right at this site with iodine treatment and hence can be mapped easily by NAIM. This could allow us to simultaneously yet individually analyze the effect of incorporation of this analog on the functional

RNA molecules. With this aim, using an approach as detailed in the methods section, we synthesized 8-OHGTPαS using a one-pot two step reaction scheme starting from the pure nucleoside (8-OHG). The first step results in a dichloro-intermediate which fives the final phosphorothioate product after reacting with TBAP. The reaction progress was monitored via TLC

(Figure 6.6.b) and the final product was purified using ion exchange chromatography (A-25 DEAE sephadex beads) and C-18 reverse phase column chromatography.

The identity of the product was confirmed by UV absorbance (data not shown here), TLC (Figure

6.6), mass spectrometry (Figure 6.7), and 31P NMR spectroscopy (Figure 6.8). The final product moves in between GTP and GTPαS in TLC, which makes perfect sense based on the polarity

- differences between these molecules. The [M-H] peak at m/z 553.954 (P3SO14C10N5O15, molecular weight 554.959) establishes the identity of the molecule. The mass spectrum data pattern does match nicely with the simulated mass spectrum for P3SO14C10N5O15 (Figure 6.7).

163

The identity of the compound was further confirmed by 31P-NMR spectroscopy (Figure 6.8) A huge downfield shift of α-P due to the replacement of O with S is a characteristic of α- phosphorothioates. We clearly observed the same pattern in the NMR spectrum, ∂ -23.1, 1P (β-P),

∂ -9.45, 1P (γ-P), and ∂ 44.08, 1P (α-P).

Figure 6.7.a. Synthesis of 8-OHGTPαS and b. a representative TLC chromatogram showing the reaction progress and product formation.

164

Figure 6.8 Mass spectrometric analysis of 8-OHGTPαS.

165

Figure 6.9 31P-NMR of 8-OHGTPαS.

6.3.4 Incorporation of 8-OHGTPαS in RNAis challenging

One of the basic requirements of NAIM is to incorporate the nucleotide analogs randomly within the RNA, which is usually accomplished enzymatically via T7 RNA polymerase. Certain analogs especially the minor groove modified analogs are incorporated better by the Y639F mutant version

166 of T7 RNA polymerase. The incorporation level is chosen to generate enough signals within the detectable range, and also to limit cooperative interference due to too many substitutions. Usually, the incorporation level is kept at about 5-10% to fulfill the above criteria. After the successful synthesis and purification of 8-OHGTPαS we proceed to incorporate the newly synthesized GTP analog in the RNA molecules via T7 RNA polymerase-based transcription reaction. Unlike the reports in the literature, unfortunately, we did not observe enough incorporation of 8-OHG in the transcription to give the full-length product in the conditions described in the method section. To get signals in NAIM study we need 1-10% incorporation of α-phosphorothioate analog nucleotides. Not only the newly synthesized phosphorothioate but also 8-OHGTP were unable to be sufficiently incorporated (to give any full-length products) in the transcripts of either L-21 G414 or P4-P6 RNAs under our experimental conditions.

6.3.5 Enzyme mutation allows a slightly better incorporation of 8-OHGTPαS

Based on the literature reports about the promiscuity of certain mutations of T7-polymerase, we were encouraged to mutate the enzyme in the hope that the mutant enzymes will be less-selective and allow 8-OHGTPαS incorporation in the better rate in the transcripts. We successfully prepared two three mutants in T7 RNA polymerase, namely, Y639F, H784A and a double mutant. Y639F mutation is known to eliminate discrimination of the hydrogen-bonding potential of the 2′- substituent of the substrate NTP and has been used to incorporate 2′-dNMPs, 2′-F-NMPs and 2′- amino-NMPs into transcripts to make them RNase resistant or for studies of RNA structure– function relationships. However, this mutant is not as useful for incorporating NTPs with bulkier substituents (2′-OMe groups) into transcripts. There are suggestions that the H784A mutation might relax the barrier to extension of transcripts containing non-canonical NMPs at the 3′-end of the RNA. Though, these criteria are not enough for our analog, we were hopeful that these

167 mutations will help in the better incorporation of the analog. To our surprise the double mutant and H784A mutant did not help in the incorporation at all, rather we observed a reduced activity than the wild type. The Y639F mutant on the other hand, helps little bit in 8-OHG incorporation.

Under a tough push of 90% 8-OHGTPαS in the transcription reaction and a relatively higher spermidine concentration, Y639F mutant allowed nearly 1% incorporation (based on the quantification of the parent bands) of 8-OHGTPαS in the P4-P6 transcripts. This indeed a great step towards the successful NAIM analysis of functional RNA molecules, is not quite enough to get enough signals yet. As shown in Fig. 6.8.b, we got cleavage signals for I2/EtOH cleavage of 8-

OHGTPαS incorporated P4-P6 transcripts adjacent to many guanine residues. The signals are less prevalent than the GTPαS ladder and are present in a non-G location in some cases indicating a poorer incorporation of 8-OHGTPαS and possibly its incorporation in other non-canonical positions, most probably in place of uracil. The next step in the study is to get a mutant that is better in 8-OHGTPαS incorporation and to find an optimized condition for the latter’s transcriptional incorporation.

168

Figure 6.10 Mutagenesis of T7 RNAP to get a better incorporation of 8-OHGTPαS in the transcripts, a. Mutations in T7RNAP resulting Y639F mutant (B), H784A mutant (not shown here), and a double mutant (C) could facilitate the nucleotide analog incorporation and b. Y639F mutant allows to incorporate 8-OHGTPαS in P4-P6 RNA as evident by the I2 cleavage of 8- OHGTPαS doped P4-P6 RNA.

6.4 Discussion

RNA oxidation has been shown as a disease feature of many neurological disorders, diabetes, cancer and psychiatric disorders12,25-33. There are evidences of translational impairment due to mRNA oxidation which are supposed to contribute in disease pathogenesis25,32,34. The presence of oxidized guanosine in pre-Q1 riboswitch has recently been shown to weaken the binding with analyte molecule35. With all these interesting finding one can speculate that RNA base oxidation play a detrimental role in RNA folding and

169

RNA-protein interactions. Yet the effect of RNA base oxidation and their effect on RNA folding and catalysis has never been directly studied based on our knowledge.

In our study, for the first time based on our knowledge, we have demonstrated that the RNA base modification due to oxidative stress result in the loss of catalytic function of L-21 G414 ribozyme. This could have happened either via misfolding/ unfolding of the ribozyme molecules once they are oxidized or due to the presence of modified bases in the critical position which could be detrimental in the catalytic reaction. We have also found that RNA oxidation results in the unfolding of P4-P6 domain of Tetrahymena group I intron. The crystal structures along with biochemical and biophysical assays of both L-21 G414 molecule and P4-P6 molecule have stablished the involvement of many critical tertiary base interactions which are responsible to hold the functional tertiary structure of these RNA molecules intact13,14,16-18. The presence of 8-OHG or any other base modification in the critical position could impair these native interactions and make these molecules misfold or unfold. To study the impacts of 8-OHG modifications on

RNA folding, we planned to incorporate the 8-OHG molecules in the L-21 G414 Tetrahymena ribozyme and P4-P6 domain to investigate the effects of this modification by using Nucleotide Analog Interference

Mapping (NAIM) technique. The synthesis of this compound is very crucial to study the effect of 8-OHG in RNA structure mainly because after the incorporation of 8-OHGTPαS, the RNA molecules can be cleaved right at this site with iodine treatment and hence can be mapped easily by NAIM.

Although we successfully synthesized and characterized the novel 8-OHG analog molecule (8-OHGTPαS), we had a hard time incorporating this modification in the transcripts. Nevertheless, we are in a right direction to incorporate this analog with the help of mutated T7 RNA polymerase enzyme. The successful incorporation of this analog in >1% level in the transcripts and their selection followed by the functional assays (ribozyme assay for L-21 G414 RNA and folding assay for P4-P6 RNA) could allow for the identification of the critical sites, oxidation of which are detrimental in the interference of RNA function.

170

Figure 6.11 A schematic representation of NAIM study.

6.5 Conclusion

Using biochemical assays, we have demonstrated that RNA oxidation is detrimental in RNA structure and function. RNA oxidation not only weakens the catalytic activity of a L-21 G414 ribozyme, it also disrupts P4-P6 RNA folding. We have also synthesized a novel 8-OHG analog phosphorothioate molecule, successful substantial incorporation of which could lead us towards the atomic mapping of the interference of 8-OHG in RNA structure and function using NAIM technique.

171

6.6 References

1 Harcourt, E. M., Kietrys, A. M. & Kool, E. T. Chemical and structural effects of base modifications

in messenger RNA. Nature 541, 339-346 (2017).

2 Topp, S. & Gallivan, J. P. Emerging applications of riboswitches in chemical biology. ACS Chem

Biol 5, 139-148 (2010).

3 Serganov, A. & Patel, D. J. Ribozymes, riboswitches and beyond: regulation of gene expression

without proteins. Nature rev Genetics 8, 776-790 (2007).

4 Jin, Y., Yang, Y. & Zhang, P. New insights into RNA secondary structure in the alternative splicing

of pre-mRNAs. RNA Biol 8, 450-457 (2011).

5 Halvorsen, M., Martin, J. S., Broadaway, S. & Laederach, A. Disease-associated mutations that

alter the RNA structural ensemble. PLOS Genetics 6, e1001074 (2010).

6 Meyer, I. M. & Miklós, I. Statistical evidence for conserved, local secondary structure in the coding

regions of eukaryotic mRNAs and pre-mRNAs. Nucleic Acids Res 33, 6338-6348 (2005).

7 Juneau, K., Podell, E., Harrington, D. J. & Cech, T. R. Structural basis of the enhanced stability of

a mutant ribozyme domain and a detailed view of RNA–solvent interactions. Structure 9, 221-231

(2001).

8 Wang, W. et al. A single-nucleotide polymorphism in the 3′-UTR region of the adipocyte fatty acid

binding protein 4 gene is associated with prognosis of triple-negative breast cancer. Oncotarget 7,

18984-18998 (2016).

9 Cheng, C. Y., Kladwang, W., Yesselman, J. D. & Das, R. RNA structure inference through

chemical mapping after accidental or intentional mutations. PNAS USA 114, 9876-9881 (2017).

10 Roundtree, I. A., Evans, M. E., Pan, T. & He, C. Dynamic RNA modifications in gene expression

regulation. Cell 169, 1187-1200 (2017).

11 Leamy, K. A., Assmann, S. M., Mathews, D. H. & Bevilacqua, P. C. Bridging the gap between in

vitro and in vivo RNA folding. Quart Rev Biophys 49, e10-e10 (2016).

172

12 Kong, Q. & Lin, C.-l. G. Oxidative damage to RNA: mechanisms, consequences, and diseases. Cell

Mol Life Sci 67, 1817-1829 (2010).

13 Guo, F., Gooding, A. R. & Cech, T. R. Comparison of crystal structure interactions and

thermodynamics for stabilizing mutations in the Tetrahymena ribozyme. RNA 12, 387-395 (2006).

14 Brehm, S. L. & Cech, T. R. Fate of an intervening sequence ribonucleic acid: excision and

cyclization of the Tetrahymena ribosomal ribonucleic acid intervening sequence in vivo. Biochem

22, 2390-2397 (1983).

15 Strobel, S. A. & Shetty, K. Defining the chemical groups essential for Tetrahymena group I intron

function by nucleotide analog interference mapping. PNAS USA 94, 2903-2908 (1997).

16 Russell, R. & Herschlag, D. Probing the folding landscape of the Tetrahymena ribozyme:

commitment to form the native conformation is late in the folding pathway. J Mol Biol 308, 839-

851 (2001).

17 Cech, T. The chemistry of self-splicing RNA and RNA enzymes. Science 236, 1532-1539 (1987).

18 Basu, S., Pazsint, C. & Chowdhury, G. in Ribozymes and siRNA Protocols (ed Mouldy Sioud)

57-75, Humana Press, 2004.

19 Meyer, A. J. et al. Transcription yield of fully 2′-modified RNA can be increased by the addition

of thermostabilizing mutations to T7 RNA polymerase mutants. Nucleic Acids Res 43, 7480-7488

(2015).

20 Guillerez, J., Lopez, P. J., Proux, F., Launay, H. & Dreyfus, M. A mutation in T7 RNA polymerase

that facilitates promoter clearance. PNAS USA 102, 5958-5963 (2005).

21 Padilla, R. & Sousa, R. A Y639F/H784A T7 RNA polymerase double mutant displays superior

properties for synthesizing RNAs with non-canonical NTPs. Nucleic Acids Res 30, e138-e138

(2002).

22 Guo, F., Gooding, A. R. & Cech, T. R. Comparison of crystal structure interactions and

thermodynamics for stabilizing mutations in the Tetrahymena ribozyme. RNA 12, 387-395 (2006).

173

23 Basu, S. et al. A specific monovalent metal ion integral to the AA platform of the RNA tetraloop

receptor. Nature Struct Biol 5, 986 (1998).

24 Castellani, R. J. et al. Sublethal RNA Oxidation as a mechanism for neurodegenerative disease. I J

Mol Sci 9, 789-806 (2008).

25 Chang, Y. et al. Messenger RNA oxidation occurs early in disease pathogenesis and promotes

motor neuron degeneration in ALS. PLOS ONE 3, e2849 (2008).

26 Ding, Q., Markesbery, W. R., Chen, Q., Li, F. & Keller, J. N. Ribosome dysfunction is an early

event in Alzheimer's disease. J Neurosci 25, 9171-9175 (2005).

27 Foti, J. J., Devadoss, B., Winkler, J. A., Collins, J. J. & Walker, G. C. Oxidation of the guanine

nucleotide pool underlies cell death by bactericidal antibiotics. Science 336, 315-319 (2012).

28 Honda, K. et al. Ribosomal RNA in Alzheimer disease is oxidized by bound redox-active iron. J

Biol Chem 280, 20978-20986 (2005).

29 Küpfer, P. A. & Leumann, C. J. in Chemical Biology of Nucleic Acids: Fundamentals and Clinical

Applications (eds Volker A. Erdmann, Wojciech T. Markiewicz, & Jan Barciszewski) 75-94,

Springer Berlin Heidelberg, 2014.

30 Shan, X., Chang, Y. & Lin, C.-l. G. Messenger RNA oxidation is an early event preceding cell

death and causes reduced protein expression. FASEB J 21, 2753-2764 (2007).

31 Tanaka, M., Chock, P. B. & Stadtman, E. R. Oxidized messenger RNA induces translation errors.

PNAS USA 104, 66-71 (2007).

32 Kharel, P., McDonough, J. & Basu, S. Evidence of extensive RNA oxidation in normal appearing

cortex of multiple sclerosis brain. Neurochem Int 92, 43-48 (2016).

33 Simms, C. L., Hudson, B. H., Mosior, J. W., Rangwala, A. S. & Zaher, H. S. An active role for the

ribosome in determining the fate of oxidized mRNA. Cell rep 9, 1256-1264 (2014).

34 Choi, Y. J., Gibala, K. S., Ayele, T., Deventer, K. V. & Resendiz, Marino J. E. Biophysical

properties, thermal stability and functional impact of 8-oxo-7,8-dihydroguanine on

174 oligonucleotides of RNA—a study of duplex, hairpins and the aptamer for preQ1 as models.

Nucleic Acids Res 45, 2099-2111 (2017).

175

CHAPTER 7 7. Concluding remarks

Oxidative stress induced biomolecule damage is a detrimental event in the development of many diseases including neurological disorders. There is a plethora of studies which demonstrate the critical role of lipid oxidation, protein oxidation/nitration, and DNA oxidation in the literature in the progression of many neurological disorders including multiple sclerosis (MS). Due a wrongly coined ‘just an intermediate’ identity of RNA, the oxidative modifications in the RNA and their impact in disease progression was largely overlooked. Multiple sclerosis, being primarily considered as an auto immune inflammatory disease of central nervous system, the role of oxidative stress and mitochondrial dysfunction was largely ignored until the early 21st century.

In this study we have discovered that mitochondrial dysfunction and oxidative stress induced RNA oxidation is a feature of MS brains. We have also demonstrated that RNA oxidation is mainly prevalent in MS neurons indicating a connection between RNA oxidation and MS neurodegeneration.

Using an in cellulo model of MS, we have established that mRNA oxidation in MS like environment is not random but there are certain classes of mRNAs including those involved in neuropathology, that are targeted more under oxidative stress. By picking a couple of highly oxidized mRNA molecules, namely Nat8l and Nlrp3, we have demonstrated that mRNA oxidation could be detrimental to protein synthesis. These two genes are members of different cellular

176 pathways. Any problem with Nat8l, as we have demonstrated, impacts N-acetyl aspartate (NAA) production which is responsible of a healthy myelin production, the maintenance of neuronal osmolarity and is a representative of neuronal health. Nlrp3 is a key gene of inflammasome signaling pathway, the problem with which could impact inflammation signaling. Interestingly, we have demonstrated the existence of NLRP3 inflammasome pathway at least in cellulo in human neurons which appears to be modulated by oxidative stress.

Next, we also aimed to delineate the impact of RNA oxidation in the structure function relationship of functional RNA molecules. By analyzing the ribozyme function of oxidized engineered

Tetrahymena Group I intron RNA, we have demonstrated that RNA loses its catalytic function under oxidative stress. We have also discovered that oxidation induces unfolding of P4-P6 domain of the same ribozyme molecule. To analyze the impact of individual base oxidation in RNA structure and function, we successfully synthesized a novel guanosine analog (8-OHGTPαS). The efficient incorporation of this molecule in the RNA and their structure probing followed by a functional assay could allow us to map the impact of guanine base oxidation in RNA function.

The work in this dissertation, for the first time based on our knowledge, demonstrated that selective mRNA oxidation in MS neurons could contribute in MS pathogenesis. Our work provides a new window to look at the mechanistic aspect of progression of MS and other neurological disorders.

The work here also discovered that RNA oxidation impact RNA folding and ribozyme activity clearly indicating the detrimental effect RNA oxidation could bring in non-coding RNA functioning.

177

CHAPTER 8

8. Summary and future perspective

RNA modification field and RNA biology overall has grown a lot in past 2-3 decades. With the advancement in the field, unique RNA mediated gene regulatory pathways are emerging. Based mainly on 6-methyl adenosine modification, epitranscriptomics field has been booming since the last decade. Overall, RNA field is growing tremendously outside ‘the messenger’ identity.

The defect in the cellular antioxidant machinery, metal dysregulation or overproduction of reactive oxygen species create oxidative stress in the cells which can damage every biomolecule. There is a glut of report highlighting the involvement of biomolecular oxidation in many pathological situations. RNA being single stranded nucleic acid and due to its presence around ROS generation cites, they are considered more vulnerable to oxidative damage under stress. The study presented in this dissertation and work from some other labs recently started a molecular understanding of

RNA oxidation and their impact in cell biology, which eventually shows a potential of RNA oxidation being involved in the progression of different kind of diseases. In this study we have demonstrated that RNA oxidation under MS environment is not completely random but is selective and the mRNAs responsible in many neuropathologically relevant pathways are amongst the highly oxidized mRNAs. We have also shown how the oxidation of mRNA could impact protein production debilitating the downstream metabolite pathways. Our current study, although provides a very solid pathway to dig deep in the role of RNA oxidation in neuropathology, needs to

178 be supplemented with high-throughput analysis of oxidatively modified transcripts not only in a broader range of in cellulo experiments but also with a broader study in the animal models and a wider analysis of postmortem brain tissues.

Based on our study on the impact of mRNA oxidation in translation, it is wise to speculate that oxidized RNA could mess up the canonical base pairing within the RNA or might alter RNA- protein interactions. These mis-interactions could change the course of non-coding RNA mediated gene regulation. One study already shows that oxidized miRNAs regulate a non-canonical mRNA target. Provided the current knowledge of small non-coding RNA based gene regulation mechanisms, a slight change in a base or bases could change the stability and hence functioning of RNA-protein complex and hence gene regulation. A high-throughput analysis of oxidatively modified non-coding RNA and mRNA along with a dream computational platform, which could mine up the big data and give the possible alternation in gene-expression interactome in a particular disease state, would be a next level in RNA oxidation pathobiology. This analysis compared with the relative level of gene expression in animal models and patients’ tissue could provide a basic understanding of progression of MS and other oxidative stress relevant diseases. Developing a probe which can selectively modify the oxidized RNA thereby giving a pause to polymerase would make a phenomenal progress to understand the positional selectivity of guanine oxidation.

Although guanine is the most highly oxidized base, oxidative modification of other nucleosides could be also as important in gene mys(regulation). With all this, the ideal incorporation of 8-OHG in the transcript and possibility of nucleotide interference mapping analysis in the functional RNA like ribozymes or riboswitches would allow us to chemically understand the nature of interactions responsible to kill the RNA function.

179

We should investigate the presence of RNA oxidation correcting enzymes in the cells and if they exist globally, whether they are deactivated under certain pathological situations. Next big question in the RNA oxidation field is whether 8-OHG in RNA does have an epitranscriptomics role? However, sounds improbable, this definitely is not impossible.

180

Appendix

Appendix 1. G% of highly oxidized mRNA analyzed in Chapter 3.

Name G% (total) G% (coding)

TMX3 19.547928 39.79643766

MRFAP1 29.20096852 39.0625

PALM3 36.85840708 38.72403561

CCDC88B 38.32552194 38.52190974

B3GALT6 33.15508021 38.48484848

LMNB2 32.62273902 38.43263553

ZNF316 32.87746926 38.3747927

CREBZF 21.9322771 37.84037559

IER5 32.9787234 37.5

ANKRD33B 27.30097503 37.30639731

SPTBN4 36.43882826 37.14100065

COL11A2 34.28468208 36.76164003

SART1 33.79616773 36.6625052

METRNL 33.74157991 36.53846154

PLEC 35.77418695 36.53675819

MRFAP1L1 27.60581175 36.45833333

EMILIN1 36.54963375 36.44706654

PPP1R9B 34.61666114 36.44514768

MIB2 36.26311313 36.33187773

KIF7 34.9726776 36.25992063

PACSIN3 35.41992697 36.23529412

OGFR 34.92842536 36.13569322

DDRGK1 33.78277154 36.08465608

BCL2L2 28.58326429 36.08247423

ZGPAT 34.7012987 36.06770833

181

CHGA 31.94154489 36.02620087

MAD1L1 35.16699411 36.00562588

NAT8L 35.03562945 35.97359736

PHLDA3 30.35844916 35.9375

COL7A1 35.94598006 35.90038314

PABPN1 28.3321454 35.80246914

CIRBP 30.89579525 35.79638752

MAFG 28.99862286 35.78732106

EPPK1 34.72940103 35.52831851

NUDT16 27.14117449 35.41666667

SAFB2 32.42236025 35.4

CCDC9 34.97372193 35.33834586

TPM2 32.82571912 35.32163743

EHBP1L1 36.17313638 35.20452567

SURF6 33.67741935 35.18518519

CACNG8 32.57913381 35.13302034

SAFB 32.45403677 34.91039427

BBC3 33.74193548 34.84848485

NES 33.50026829 34.79243732

AKAP17A 32.25906415 34.72222222

CCDC137 32.85917496 34.71264368

ZNF783 30.14590348 34.67397928

OBSCN 34.57034164 34.56762249

KCTD12 22.15827338 34.5603272

TSR2 27.71435511 34.54861111

HNRNPM 32.9218107 34.53757225

SCRIB 33.56709629 34.50080515

AGRN 33.63083164 34.47720316

FUS 27.62258254 34.47185326

IBA57 28.73210634 34.45378151

182

GAR1 29.765625 34.40366972

VEGFA 26.81533859 34.30185634

SYNM 29.63078361 34.29027113

RAP2B 22.23939142 34.23913043

CDC37 31.83697578 34.21284081

CD2BP2 30.73341094 34.21052632

ALPK3 32.45101409 34.17721519

FDXR 33.79679144 34.15977961

C6orf47 31.83838384 34.12429379

LMNA 33.89585342 34.08695652

GOLGA8B 25.67249496 34.05639913

JAG2 33.39606501 34.00591875

PDZD4 33.34210526 33.97766323

UHRF1 29.50737684 33.96305626

HHIPL1 30.17695529 33.92933163

ZSCAN25 26.45102088 33.70030581

TNKS1BP1 32.37915667 33.69942197

CTD 32.54585881 33.60215054

ENGASE 32.97730307 33.55734767

OPLAH 33.28395062 33.54021205

CEP250 32.99582331 33.53336191

GTF2F1 30.61544492 33.52638353

NINL 32.18138234 33.47794649

AEBP2 20.9540636 33.46560847

ZNF697 26.25984252 33.45543346

MAGEF1 29.88235294 33.44155844

CDC42BPG 33.18503539 33.44072165

HSPA1A 31.04294479 33.38525441

COL6A2 32.54419192 33.36960464

PPP1R26 32.32506554 33.33333333

183

ZSCAN2 31.88118812 33.33333333

HSPA1B 30.69384555 33.33333333

KHNYN 28.88430408 33.33333333

HNRNPA3 22.23752151 33.24538259

HNRNPAB 30.81110506 33.23323323

ARMCX4 32.32758621 33.21698662

TOP1MT 32.5140809 33.16722038

SYT5 28.8720993 33.16106804

GRIPAP1 33.47145488 33.15926893

ZNF707 32.10251955 33.15412186

RNF40 32.19860511 33.13373253

COL4A1 29.54927426 33.11377246

JAKMIP1 33.61009631 33.09294872

MVP 31.97417601 33.07233408

SEMA3B 33.82887347 33.06666667

LRPAP1 27.85687168 33.05400372

NUDC 31.32928847 33.03212851

FAM129C 31.11375347 32.99904489

COL20A1 32.70207852 32.95795796

TMPRSS9 32.85198556 32.95597484

AMZ1 29.72337733 32.93253173

HNRNPUL2 30.21722265 32.93226381

HNRNPD 28.43601896 32.8990228

TMEM64 20.33616933 32.89588801

IER2 32.61494253 32.88690476

ZBTB46 29.04643608 32.88135593

AAMP 32.30352304 32.87461774

FBXO45 22.77555285 32.86875726

GOLGA8A 26.02265576 32.86290323

WDR4 31.42722117 32.83395755

184

PHRF1 32.90978399 32.82736288

LONP1 32.76566757 32.8125

NEFH 31.0112964 32.81096964

MYH9 31.1490601 32.8063913

PNMA2 24.53569955 32.78538813

UBAP1L 30.66476054 32.72251309

RAB17 30.97125867 32.70735524

GOLGA3 28.55072464 32.69622834

TNXB 32.68173488 32.68158618

GNL1 27.89905363 32.6754386

MELTF 33.48473379 32.65674335

CDK5RAP3 30.87983567 32.62411348

RRBP1 32.2521097 32.61758691

FCHSD1 30.62543273 32.60974433

PDAP1 28.00147765 32.6007326

HMCN2 32.66271241 32.56800599

FBXL19 32.90508652 32.56594724

ZNF212 30.63255153 32.52688172

ARFRP1 35.35589265 32.50825083

FSCN2 34.17054264 32.49516441

MAPK8IP3 32.12653779 32.46073298

CDK11A 31.11041208 32.42542153

C9orf172 32.6789107 32.41214603

GOLGA2 30.69191211 32.30309073

MAMDC4 31.77977651 32.24956063

HNRNPA1 28.05538832 32.17158177

CTTN 29.30174564 32.16601816

RHBDL1 32.16216216 32.1460374

UPF3A 28.35137386 32.1453529

ZNF513 32.84132841 32.10332103

185

GDF15 30.06351447 32.03883495

HNRNPH3 25.33659731 32.02811245

PRNP 24.21652422 32.02099738

FAM212B 26.89269256 32.01911589

CYTH2 31.52432432 32.00332502

NPTX1 27.5703513 31.94765204

SMIM17 19.86624705 31.93277311

PNPLA7 31.16374871 31.89377017

TRIM73 33.21060383 31.87250996

TXLNA 27.44166839 31.87081048

NECTIN4 30.41697692 31.8275154

SLC22A7 30.63063063 31.80987203

NACC2 30.04783302 31.80272109

NUMA1 31.43015521 31.78694158

TERF2IP 27.95992714 31.75

C15orf59 27.60907504 31.74603175

GRIP2 30.17743816 31.67305236

ACTN3 31.4701897 31.59645233

SRRT 30.63984169 31.58494869

PRPF6 31.44230769 31.56404812

FLNC 31.33701657 31.56331229

CALCOCO1 28.00393959 31.55105973

STX4 31.26649077 31.53153153

GCC1 29.21483622 31.52920962

CHAF1A 31.33971292 31.52211773

COL3A1 28.25136612 31.49284254

CDKN1A 26.40860215 31.49078727

TPM3 24.15940224 31.45917001

ARHGAP27 29.96715928 31.45112325

AEN 28.82234379 31.39059305

186

AHRR 28.36504581 31.38888889

PACSIN2 28.9433848 31.34839151

TCEAL4 26.76878191 31.32716049

MAPK11 32.52131547 31.32420091

DGCR14 30.49028677 31.30677848

ING5 31.96069256 31.24183007

CHMP3 22.85535379 31.2406577

CHD5 31.22218773 31.23614663

CLEC18B 30.24564995 31.21345029

CCDC136 30.96041731 31.1914324

ERICH1 32.19991807 31.16987179

GGT1 31.28576879 31.16959064

PSD2 28.91422737 31.14478114

CLK3 30.18798378 31.14241002

HGFAC 30.31865042 31.12116642

ZNF469 30.71271681 31.11733741

TRIM3 30.02934464 31.09619687

GTPBP2 28.95169241 31.01160862

RYR1 30.79192343 30.97139055

ZNF622 30.23391813 30.9623431

PCNT 30.68203467 30.93418259

TNFRSF10A 29.64852608 30.91684435

HYOU1 28.90625 30.9

SDE2 22.01538843 30.89970501

EZR 28.43631778 30.8915389

DCTN1 29.99114653 30.88350274

CLEC18C 28.68852459 30.87248322

DHX38 30.96196868 30.86319218

SMIM12 25.29480153 30.82437276

SH2B1 30.41709054 30.8234258

187

GRIN2C 30.52921383 30.81617086

PLXNB2 30.65602004 30.81384811

PRKRIP1 31.54516277 30.81081081

AHNAK2 30.45305677 30.74534161

CHGB 29.21980495 30.74274471

PES1 30.70934256 30.73005093

HARS 30.36175711 30.71895425

NOP9 28.94910773 30.71690215

METTL13 28.26658872 30.71428571

IGDCC4 27.32459522 30.69544365

CBFB 21.65079365 30.67375887

NOM1 26.08052588 30.66511988

FOXRED1 29.99520843 30.66392882

REXO4 32.05918619 30.65405831

NOTCH1 30.13168896 30.58163798

HNRNPU 21.55268339 30.50847458

ITM2C 30.02364066 30.47263682

SQSTM1 28.55680655 30.4388422

EPHA2 29.85074627 30.43478261

RNF103 23.06552031 30.42328042

LPAR5 25.98949212 30.38427167

FTSJ3 29.55939766 30.34591195

MYO1F 29.3793434 30.33060358

SERPINH1 30.26146592 30.31026253

GADD45A 28.54938272 30.3030303

STBD1 28.6437247 30.26926648

CNPY2 28.92290869 30.23679417

ASB1 28.69287991 30.21276596

PDGFA 28.97828409 30.18867925

LMOD1 28.25812957 30.17193566

188

B4GALNT3 30.30831879 30.16349683

BCL2L2-PABPN1 27.46478873 30.13972056

TNFRSF10B 25.44873371 30.09708738

CD6 29.82774252 30.09466866

ERP29 28.05706522 30.02544529

PDCL 21.71799028 30.02207506

GPR35 30.31400966 30

DRAXIN 29.32882121 30

FKBP4 25.68538728 30

PTGFRN 26.25349334 29.94732882

MANF 28.90231621 29.92831541

MOAP1 26.98151951 29.92424242

SART3 27.48457848 29.87551867

NCL 27.67203514 29.81715893

TCEAL1 22.62295082 29.79166667

BTG2 24.4665195 29.76939203

DNAJC14 27.95795796 29.72972973

HNRNPA1L2 26.5961945 29.69885774

EFNB1 30.41267943 29.68299712

LONRF2 25.22270115 29.66887417

CDHR5 29.2898403 29.64015152

ATXN7L3B 25.26344231 29.59183673

C2orf48 28.70662461 29.58333333

HNRNPH2 25.45986622 29.55555556

ACBD3 23.01033231 29.552615

FUBP3 24.5984252 29.55206515

ANKRD11 29.50590763 29.52952953

NCBP2 23.63218391 29.51167728

XPC 28.61372813 29.49946752

DLL4 29.71395213 29.494655

189

VAV2 30.09449466 29.49308756

UBA7 28.7987988 29.48338269

MORN4 25.64522737 29.47845805

MEOX1 26.35193133 29.41176471

WNT2B 24.78632479 29.40125112

TMED7 18.68077775 29.39130435

UBTF 28.93211921 29.32461874

UTP3 27.29885057 29.30555556

MFAP1 26.369365 29.24242424

MYO15A 28.77231391 29.10412537

TMEM215 20.28153362 29.0960452

ZNF263 29.71255673 29.09356725

RASGRP2 29.35064935 29.07103825

AKAP12 25.53846154 29.07085437

FUT5 30.83501006 29.06666667

PRKX 24.72057857 29.06220984

AATF 29.46593002 29.05525847

EIF4H 27.49410841 29.04953146

BRD9 29.86291219 28.98550725

PDIA6 24.76489028 28.92376682

RAB21 20.19890261 28.90855457

NOP14 27.94117647 28.9044289

MYH3 28.62563933 28.90262751

ADAMTS1 26.95931478 28.89118457

GRK7 28.80870561 28.82069795

HNRNPH1 25.78125 28.81481481

CES2 28.48733043 28.78289474

UPB1 27.08814029 28.74458874

VWCE 29.55316742 28.69595537

FOSL1 25.46901649 28.67647059

190

MYO1A 28.04811372 28.67177522

ZNHIT6 21.54268194 28.66242038

EIF3C 27.8101072 28.55579869

DKC1 27.18558282 28.54368932

DNAJB6 25.82197273 28.54230377

LSM14B 29.37356761 28.49740933

SRSF5 25.7918552 28.44932845

NTPCR 25.5033557 28.44677138

UBE2M 34.80519481 28.44202899

CDC42BPB 28.30920758 28.40732087

DHX16 27.47997535 28.39838493

CLDN15 26.85314685 28.38427948

PRPF38B 22.97297297 28.37465565

C9orf64 25.41347317 28.3625731

HSF2BP 27.34864301 28.35820896

DUOX2 26.32233976 28.34086507

RTCB 27.77510812 28.32674572

NOP2 27.99556213 28.3065513

DUSP4 26.02484472 28.28947368

CD74 25.08833922 28.28282828

IFITM10 33.45874765 28.23871907

TMEM179 35.01199041 28.20512821

MRPL27 29.25824176 28.18791946

PNN 24.29284526 28.18012999

EIF3A 25.230449 28.17546397

HTATSF1 27.85643727 28.13051146

ZBTB10 21.88954076 28.10534591

ZSCAN29 25.09363296 28.09691286

MUC4 27.58452931 28.09402436

EIF3J 20.74592075 28.05662806

191

NBPF20 27.3495935 28.03379416

GOLIM4 25.15279737 28.00199302

FOLR2 26.12533098 27.99479167

DNAH17 28.01136778 27.98565987

MAN1A1 22.44250595 27.98165138

SYNJ2 25.09439051 27.9670452

FUT6 27.76538094 27.96296296

CALM2 23.38416848 27.94612795

NFE2L1 29.05336332 27.94307891

RRP8 28.03632236 27.93581327

ST13 22.04614957 27.92792793

IK 27.05882353 27.89725209

LEMD3 23.82242601 27.85087719

GSDMB 25.7287706 27.80528053

SCN4B 27.34221446 27.80203785

CPE 24.66012085 27.74283718

DSCAML1 26.81798468 27.73573005

CTSV 22.35538142 27.66169154

EDA2R 19.45416375 27.63819095

CDH23 28.0851446 27.61053154

SYF2 21.89082724 27.59562842

WDR60 27.05199261 27.58512965

AHNAK 27.40541991 27.52786737

IWS1 27.06666667 27.4796748

DNHD1 27.46942615 27.4786145

TG 27.30391577 27.47080775

ZNF768 30.79082826 27.45098039

FUT3 27.52834467 27.44014733

NUS1 21.20456906 27.43764172

3-Mar 24.35653003 27.42782152

192

WBP1 30.3152789 27.40740741

LUC7L3 23.26241135 27.40569669

DDX3X 23.73618409 27.40070387

NOLC1 25.24515967 27.37089202

GID8 26.19369369 27.36535662

DQX1 26.56603774 27.34447539

CCDC127 24.98260264 27.33077905

CRIP3 25.41841004 27.31707317

CHSY1 26.03459601 27.31423827

DNAJB11 27.3262662 27.29805014

SEMA4A 26.66885893 27.24987431

OPTN 25.5952381 27.22029988

TBX4 25.45878241 27.1483306

DDX42 27.34158041 27.12105076

APTX 25.37807183 27.11370262

RNASET2 31.6 27.1076524

KLHL2 24.96015301 27.1043771

NSUN2 26.79658952 27.10322874

SLC6A13 27.58310872 27.08678828

ZNF337 25.9235256 27.08333333

RBM6 27.30175625 27.07591934

UTP18 26.30744849 27.04967086

C1S 26.11050921 27.04402516

SULF2 26.11464968 26.9969278

MYH15 26.28111274 26.9375

MED4 20.41976388 26.93726937

CEBPG 23.44080338 26.93156733

MOBP 23.30569059 26.8921095

MUC19 26.7566714 26.85350825

DNAH2 27.34871281 26.83679615

193

UTP14A 26.60991858 26.80467091

SLTM 25.5994006 26.79474216

DNAH10 26.62218973 26.78352653

C5AR1 23.35610589 26.78062678

STAG3 27.6372315 26.77419355

SRGAP3 24.46569843 26.76870748

ZNF189 23.48116646 26.74418605

GPRASP2 26.76794133 26.73818037

PPM1B 26.36667906 26.73611111

BOD1L1 26.89463956 26.72098241

SLC30A1 25.5653884 26.70603675

PPM1D 22.92275574 26.67766777

LLPH 23.39228296 26.66666667

BMS1 24.36630219 26.65627436

SIRT1 22.48175182 26.64884135

INPP5D 27.0683793 26.58263305

OXER1 27.43265531 26.57232704

DGKE 21.01747174 26.5258216

UPF3B 25.86422324 26.51515152

GTPBP4 25.22664564 26.50918635

TSR1 26.52351738 26.5010352

SEC14L1 28.12670672 26.48975791

UTP14C 22.57540184 26.46675359

DDX46 22.33043478 26.46014843

NPC1L1 26.29889669 26.45661415

BAZ1B 25.9579325 26.41509434

SEC11A 27.39403454 26.41025641

MSANTD4 23.4771962 26.39691715

BNIP2 20.70810386 26.37614679

HSPA9 24.41528808 26.37254902

194

LIN7A 17.46810599 26.35327635

ZMAT3 19.77765425 26.32183908

SMARCE1 22.55670103 26.29449838

NARS 23.0474198 26.29022465

MAPRE1 23.90151515 26.27013631

ZDHHC11 26.59137577 26.25570776

MAK16 20.25069638 26.24584718

GSPT2 23.76791332 26.23211447

DNAJB13 25.06666667 26.1829653

THUMPD1 22.20500332 26.17702448

CANX 21.65052462 26.16772824

DNAJA1 23.81546135 26.13065327

C22orf46 30.19434983 26.09289617

PEG3 23.55961209 26.07509964

NR1I3 25.81329562 26.0707635

ABCA1 23.19012797 26.03655818

DDB2 26.57754011 26.01246106

ATXN7L3 28.28227571 26.00938967

ZNF331 24.86926206 26.00574713

ZC3H13 22.53919301 25.96806387

CCDC174 25.61344538 25.92592593

LIN7C 18.02301685 25.92592593

HSPD1 24.88242839 25.84204413

RPL5 25.19455253 25.83892617

CCT6A 24.79492916 25.81453634

TXNRD1 23.58123866 25.8

ORMDL3 31.50115473 25.7751938

RAB18 19.94803118 25.76489533

NAB1 21.44014818 25.7357974

TVP23B 21.95831314 25.72815534

195

DNAJC7 26.43129771 25.72390572

FMR1 22.60258445 25.69773565

TAF1 24.86423584 25.69517775

CRNKL1 22.34307727 25.67726737

RBM12B 22.24828935 25.64870259

H1F0 28.03938356 25.64102564

RASGRF1 25.10930668 25.64102564

ANKRD42 25.52543405 25.63131313

YTHDC1 24.99264057 25.61151079

ACTR8 24.41443924 25.6

ZNF24 20.42075293 25.56458898

NAPG 20.40761598 25.55910543

ZC3H15 23.10379242 25.52552553

SNW1 24.55822383 25.52447552

AP1AR 22.50292789 25.52255226

GPATCH2 19.29407772 25.4568368

PGA5 23.93465909 25.44987147

C12orf45 25.88424437 25.44802867

ST6GALNAC2 26.0807601 25.42222222

TMEM92 28.32180561 25.41666667

RABGGTB 22.14239059 25.40160643

MORF4L2 22.42305813 25.37485582

PIWIL2 26.14774624 25.35934292

TTC33 18.75339735 25.34854246

SEC62 19.33955053 25.33333333

SYNCRIP 20.27365014 25.31565657

NPM1 24.19716206 25.31073446

FAM180A 21.23212321 25.28735632

PKD1L1 24.92300924 25.26315789

FXR1 19.75020427 25.24115756

196

CYP2E1 24.53509298 25.23616734

DDX5 23.57269728 25.20325203

LAMB4 24.39219877 25.19661222

WTAP 23.3995585 25.18891688

RPL7 23.55658199 25.16733601

EIF4A2 23.62204724 25.16339869

SMU1 23.142011 25.09727626

GNL3 24.85351563 25.09090909

FAN1 24.933687 25.08185986

CDV3 24.36720816 25.0764526

EIF4ENIF1 25.95108696 24.98309669

VPS36 21.01497876 24.97846684

TAF7 23.93939394 24.95238095

PNISR 20.79783114 24.93796526

OSBPL3 20.64241852 24.88262911

FBXW7 21.06486864 24.85875706

RRM2B 19.45857928 24.8427673

ZNF514 24.19635788 24.82419128

MORF4L1 25.2375924 24.79423868

FYTTD1 21.73444654 24.76489028

EIF5B 23.75478927 24.76112476

CBWD1 22.98850575 24.74747475

SLU7 24.39252336 24.70187394

RSPH10B 24.40743113 24.68427095

C10ORF2 28.10457516 24.67532468

CBX3 20.34334764 24.63768116

DENND2D 23.63470874 24.57627119

WDR43 21.68266516 24.53294002

PHF10 25.23640662 24.51569806

TGFBR1 21.70246619 24.51209992

197

TCEA1 20.89871612 24.50331126

MUC5B 24.66510382 24.48955984

TRIM23 22.76657061 24.44444444

PRPF18 23.75963629 24.43693694

PSMC6 22.76422764 24.42244224

RABEP1 22.73182455 24.4057971

PTP4A2 22.74689007 24.4047619

CMBL 24.61082283 24.3902439

LMOD2 21.87367529 24.33090024

MSANTD3 20.95316352 24.32234432

TRIM37 23.91351824 24.31778929

LTV1 24.23296932 24.29971989

ZNF852 25.38071066 24.29378531

ITGB1 23.17607631 24.28035044

BRMS1L 20.89093702 24.27983539

MSL1 23.0829421 24.24242424

TYW3 21.34570766 24.23076923

UBASH3B 21.9823214 24.22764228

PDP1 21.93002515 24.21551214

CSGALNACT2 21.54944474 24.18661756

RNF168 24.73438956 24.18414918

E2F6 24.82983131 24.15458937

CETN3 19.57123098 24.13194444

ZNF345 20.11494253 24.13087935

RAB1A 22.24320242 24.11003236

ZDHHC11B 28.97410685 24.10394265

USP7 24.29464117 24.10303588

IMPA1 18.49234393 24.10071942

DYNLL2 30.74901445 24.07407407

HSPA4 24.01772526 24.05866033

198

DNM3 24.07407407 24.0201785

FBXO22 20.81005587 24

TFRC 22.06092824 23.97058824

MIA3 21.90499631 23.95178197

ZNF112 19.37561631 23.9169352

UGDH 22.19092332 23.90572391

CWF19L2 22.77016743 23.87337058

TIGAR 21.56124803 23.86223862

FNBP1L 21.27421759 23.852657

HSP90B1 24.27926363 23.79767828

ATP6AP1L 24.29842574 23.7037037

HRH2 24.00914634 23.70184255

BTBD10 22.63487099 23.69146006

ZRANB2 19.69549725 23.66565962

TRMT13 18.34019639 23.65145228

PPIG 23.62900258 23.62030905

MPHOSPH10 24.23633441 23.60703812

CFAP70 23.36124747 23.60583016

CATSPERD 23.81329114 23.57113058

SRP54 23.32463011 23.56435644

USP47 23.76720901 23.54965585

GOLGB1 22.77432712 23.54435974

KIAA0232 22.8248444 23.5434575

NAA50 21.62643771 23.52941176

ZNF268 20.76631122 23.52941176

PPP4R2 20.12817945 23.52472089

PUM3 24.14874552 23.52336929

DHX36 22.70344828 23.52163859

FMOD 24.30449404 23.51900973

SLC25A36 19.58520419 23.5042735

199

UBP1 24.59299457 23.49834983

STK31 23.96271798 23.49673203

MGEA5 23.12098992 23.49165597

ZFP69B 23.9588437 23.48909657

TRPM6 22.84426521 23.46350305

TGS1 22.61777873 23.45823575

SMC3 22.99685306 23.45374932

ZNF594 22.92392864 23.43234323

FBXO30 20.70135747 23.36907954

LYNX1 32.19861232 23.36182336

BIVM-ERCC5 23.07991445 23.35974

NEB 23.24631708 23.34037063

WAPL 22.48453273 23.33333333

EPRS 23.81995618 23.3311302

SESN1 22.01630838 23.32657201

PPIL4 21.23233909 23.32657201

CCZ1B 24.11003236 23.32643202

ESF1 22.26368159 23.31768388

BLOC1S2 21.02619304 23.26388889

CATSPER2 26.28140704 23.22661645

TPR 21.88752956 23.13874788

MSH2 23.40359578 23.1372549

GXYLT1 21.49420695 23.1292517

HSPH1 22.15125192 23.12757202

FGFR1OP2 20.59405941 23.09711286

BDP1 21.88879877 23.08571429

TFAM 20.06016168 23.07692308

CYLD 22.25658648 23.05816789

PSMD12 20.21544809 23.05343511

UPF2 23.05550399 23.01649647

200

SYTL2 22.40222171 23.01500183

SMARCA5 20.6787787 23.0136119

FAM71F2 23.22077922 23.01075269

PCMTD1 20.01427891 22.9981378

NRG4 18.08656036 22.98850575

CHORDC1 19.23076923 22.98325723

ARID4A 21.51243094 22.97297297

CCNL1 22.70105192 22.9601518

PPP1R12A 21.09675166 22.95505981

ROCK2 21.85318893 22.94216463

POLR2B 23.13986193 22.93939394

NAPB 24.05857741 22.92682927

NOL8 22.76291513 22.88812785

DYNC1LI2 23.82916768 22.87893232

KTN1 22.50526316 22.85223368

RFC1 22.73284314 22.85133566

MYSM1 18.84654995 22.79855247

LARP7 22.12041885 22.75586049

MUC5AC 22.8282019 22.74683171

SSB 22.57126236 22.73838631

SLK 22.13551827 22.71092669

KLHL9 20.44122716 22.70765912

STAG2 20.45096184 22.66876806

CEBPZ 22.11912601 22.65402844

ALG11 23.08285164 22.65043949

SERINC1 22.18769422 22.61380323

TRNT1 21.47533828 22.60536398

AGBL3 20.55464927 22.60273973

DRAM1 23.50126654 22.59414226

SCN2A 20.95359039 22.58225324

201

RBM27 20.56445461 22.55733585

APAF1 22.40282686 22.55389718

PRPF4B 20.83904579 22.55291005

UBA5 19.92296919 22.53968254

TMF1 19.50688905 22.51675807

CCDC82 20.23602916 22.50764526

RGPD8 21.83860803 22.4801812

TOP2B 22.91704649 22.43392747

IKBIP 24.00249377 22.41215575

CLK1 22.03828246 22.40549828

RPS27L 19.55719557 22.35294118

ROCK1 22.7518797 22.33702337

RANBP6 21.38857783 22.32537577

DNTTIP2 21.83486239 22.32496697

GOLGA4 21.80511182 22.29129663

FAM168B 26.16856163 22.28796844

BAZ1A 22.40092854 22.2864483

DIS3 21.00408486 22.28015294

MDM2 20.13363029 22.25672878

ARL14EP 19.91649269 22.22222222

PTP4A1 19.57588847 22.22222222

MUC6 22.95389229 22.21311475

FRS2 18.50937091 22.20039293

XPOT 21.55212716 22.04915196

AKAP9 21.94495707 22.0402593

ZBTB1 19.21315009 22.03548086

ZNF441 20.11185682 21.99807877

APPBP2 20.51482059 21.94174757

CHD1 22.0519316 21.93811377

CCDC30 22.7043673 21.8962585

202

ANKRD18B 23.91366906 21.83713722

HSPA4L 20.17960909 21.82637183

NAA35 23.07113286 21.80899908

OTUD6B 19.18796992 21.76165803

FBXL3 20.82866741 21.75602176

LAMTOR3 20.68560695 21.75141243

TAX1BP1 21.24928694 21.72995781

UFM1 19.91654021 21.70542636

SMC5 22.80166436 21.67134394

CMKLR1 26.05531295 21.65775401

U2SURP 19.5471597 21.65048544

PPP4R3A 24.11456056 21.64027609

CNOT6 20.89175989 21.5651135

ANKRD18A 22.85006196 21.55085599

KBTBD8 19.66268147 21.5393134

FAS 19.71652746 21.52777778

EEA1 18.75545579 21.48253069

ZFC3H1 21.76585638 21.45728643

TAF1D 23.20483749 21.38590203

ZNF322 21.89288635 21.33995037

USP16 20.7635633 21.18266505

TRIP11 20.14592623 21.17468574

KITLG 18.71794872 21.16788321

ZNF146 21.24142233 21.16040956

ZNF286A 21.98237093 21.1366539

ASCC3 21.5004458 21.12430721

CCP110 20.89819288 21.09246463

USP1 20.31827224 21.07718405

CPM 20.12569205 21.02102102

ANKRD26 21.49873702 21.01377953

203

TBK1 20.91672046 21.00456621

ZNF705E 18.91891892 20.93023256

KIAA1644 23.80396733 20.89314195

C1orf27 20.32603158 20.80586081

CCDC18 20.83680266 20.80183276

ZNF667 23.19714222 20.78559738

CEP290 21.03302707 20.77698223

KPNA4 19.74167687 20.75351213

DNAH12 20.13651877 20.66957787

BBS10 18.52578475 20.48802947

CLHC1 19.35722202 20.44293015

ZNF37A 22.57013318 20.40332147

PDCD10 21.25171939 20.34428795

GCC2 19.71146769 20.34329307

CASP8AP2 20.45154669 20.25550513

RBBP6 22.25302787 20.20821714

MAP9 17.83673469 19.9381762

MTERF1 20.91761223 19.91666667

ERVH48-1 22.18831735 19.8757764

TMTC3 17.52047758 19.3442623

KIF20B 18.84192599 19.33003844

LRIF1 18.72200657 18.91774892

SYCP2 18.46291332 18.81123449

ABHD13 18.34085779 17.94871795

KIAA1586 19.7705803 17.9357022

TRIM59 18.81111683 17.16171617

204