Biochemical Approaches for the Diagnosis and Treatment of Lafora Disease

University of Kentucky UKnowledge

Theses and Dissertations--Molecular and Cellular Biochemistry Molecular and Cellular Biochemistry

2019

BIOCHEMICAL APPROACHES FOR THE DIAGNOSIS AND TREATMENT OF LAFORA DISEASE

Mary Kathryn Brewer University of Kentucky, [email protected] Author ORCID Identifier: https://orcid.org/0000-0001-5089-3570 Digital Object Identifier: https://doi.org/10.13023/etd.2019.120

Right click to open a feedback form in a new tab to let us know how this document benefits ou.y

Recommended Citation Brewer, Mary Kathryn, "BIOCHEMICAL APPROACHES FOR THE DIAGNOSIS AND TREATMENT OF LAFORA DISEASE" (2019). Theses and Dissertations--Molecular and Cellular Biochemistry. 41. https://uknowledge.uky.edu/biochem_etds/41

This Doctoral Dissertation is brought to you for free and open access by the Molecular and Cellular Biochemistry at UKnowledge. It has been accepted for inclusion in Theses and Dissertations--Molecular and Cellular Biochemistry by an authorized administrator of UKnowledge. For more information, please contact [email protected]. STUDENT AGREEMENT:

I represent that my thesis or dissertation and abstract are my original work. Proper attribution has been given to all outside sources. I understand that I am solely responsible for obtaining any needed copyright permissions. I have obtained needed written permission statement(s) from the owner(s) of each third-party copyrighted matter to be included in my work, allowing electronic distribution (if such use is not permitted by the fair use doctrine) which will be submitted to UKnowledge as Additional File.

I hereby grant to The University of Kentucky and its agents the irrevocable, non-exclusive, and royalty-free license to archive and make accessible my work in whole or in part in all forms of media, now or hereafter known. I agree that the document mentioned above may be made available immediately for worldwide access unless an embargo applies.

I retain all other ownership rights to the copyright of my work. I also retain the right to use in future works (such as articles or books) all or part of my work. I understand that I am free to register the copyright to my work.

REVIEW, APPROVAL AND ACCEPTANCE

The document mentioned above has been reviewed and accepted by the student’s advisor, on behalf of the advisory committee, and by the Director of Graduate Studies (DGS), on behalf of the program; we verify that this is the final, approved version of the student’s thesis including all changes required by the advisory committee. The undersigned agree to abide by the statements above.

Mary Kathryn Brewer, Student

Dr. Matthew Shawn Gentry, Major Professor

Dr. Trevor Creamer, Director of Graduate Studies

BIOCHEMICAL APPROACHES FOR THE DIAGNOSIS AND TREATMENT OF LAFORA DISEASE

______

DISSERTATION ______A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the College of Medicine at the University of Kentucky

By Mary Kathryn Brewer Lexington, Kentucky Director: Dr. Matthew S. Gentry, Professor of Molecular and Cellular Biochemistry Lexington, Kentucky 2019

ABSTRACT OF DISSERTATION

BIOCHEMICAL APPROACHES FOR THE DIAGNOSIS AND TREATMENT OF LAFORA DISEASE

Glycogen is the sole carbohydrate storage molecule found in mammalian cells and plays an important role in cellular metabolism in nearly all tissues, including the brain. Defects in glycogen metabolism underlie the glycogen storage diseases (GSDs), genetic disorders with variable clinical phenotypes depending on the mutation type and affected gene(s). Lafora disease (LD) is a fatal form of progressive myoclonus epilepsy and a non-classical GSD. LD typically manifests in adolescence with tonic-clonic seizures, myoclonus, and a rapid, insidious progression. Patients experience increasingly severe and frequent epileptic episodes, loss of speech and muscular control, disinhibited dementia, and severe cognitive decline; death usually ensues in the second decade of life. LD, like one- third of all epilepsy disorders, is intractable and resistant to antiseizure drugs. A hallmark of LD is the accumulation of intracellular, insoluble carbohydrate aggregates known as Lafora bodies (LBs) in brain, muscle, and other tissues. LBs are a type of polyglucosan body, an insoluble aggregate of aberrant glycogen found in some GSDs and neurodegenerative disorders. Like most GSDs, LD is an autosomal recessive genetic disorder. Approximately 50% of LD patients carry mutations in the epilepsy, progressive myoclonus 2A (EPM2A) gene encoding laforin, a glycogen phosphatase. Remaining patients carry mutations in EPM2B, the gene that encodes malin, an E3 ubiquitin ligase. Laforin and malin play important roles in glycogen metabolism. In the absence of either enzyme, glycogen transforms into an insoluble, hyperphosphorylated and aberrantly branched polysaccharide reminiscent of plant starch. This abnormal polysaccharide precipitates to form LBs and has pathological consequences in the brain. Since a definitive LD diagnosis requires genetic testing, whole exome sequencing has been increasingly used to diagnose LD. As a result, numerous cases of more slowly progressing or late-onset LD have been discovered that are

associated with missense mutations in EPM2A or EPM2B. Over 50 EPM2A missense mutations have been described. These mutations map to many regions of the laforin X-ray crystal structure, suggesting they produce a spectrum of effects on laforin function. In the present work, a biochemical pipeline was developed to characterize laforin patient mutations. The mutations fall into distinct classes with mild, moderate or severe effects on laforin function, providing a biochemical explanation for less severe forms of LD. LBs drive LD pathology. As a result, LBs and glycogen metabolism have become therapeutic targets. Since LBs are starch-like, and starch is degraded by amylases, these enzymes are potential therapeutics for reducing LB loads in vivo. However, amylases are normally secreted enzymes. Degradation of intracellular LBs requires a cell-penetrating delivery platform. Herein, an antibody-enzyme fusion (AEF) technology was developed to degrade LBs in vitro, in situ in cell culture, and in vivo in LD mouse models. AEFs are a now putative precision therapy for LD, potentially the first therapeutic to provide a significant clinical benefit. Prior to this work, LD was considered a homogenous disorder and treatments were only palliative. The data herein support a spectrum of clinical progression, a potential therapy for LD, and mechanistic insights into LD pathophysiology. This work illustrates how personalized medicine, both in diagnosis and treatment, can be achieved through basic biochemical approaches to human disease.

KEYWORDS: Lafora disease, glycogen, epilepsy, neurodegenerative disease, phosphatase, personalized medicine.

Mary Kathryn Brewer

April 22, 2019 Date

BIOCHEMICAL APPROACHES FOR THE DIAGNOSIS AND TREATMENT OF LAFORA DISEASE

By Mary Kathryn Brewer

Dr. Matthew Shawn Gentry Director of Dissertation

Dr. Trevor Creamer Director of Graduate Studies

April 22, 2019 Date

DEDICATION

For my little brother Ben

ACKNOWLEDGMENTS

I must first and foremost acknowledge my extraordinary mentor, Dr. Matthew Gentry. I will never be able to thank him sufficiently for his kindness, grace and guidance throughout my undergraduate and graduate career. Not only has he done everything in his power to enable me and my colleagues to succeed as scientists, the goodness of his character has profoundly impacted my life. I am forever grateful for the lessons that I have learned from him. I know he will go on to train many more outstanding scientists like himself. I am also thankful for his wife Sandy, who cheers on the whole lab and helps to hold things together. I would like to thank my dissertation committee members and outside examiner for their valuable time, questions, guidance and feedback. Dr. Craig Vander Kooi has always challenged me and encouraged me to think more deeply. He is an indispensable collaborator and an unending source of knowledge and discussion, whom I could approach with technical or scientific questions at any time. I am always thankful for the big-picture perspective of Dr. Jim Geddes, and for a very successful rotation in his lab at the beginning of graduate school. I am also very thankful for Dr. Tianyan Gao, for her intelligence and gentle demeanor, and for mentoring me during my first departmental seminar. Finally, I am very grateful to my outside examiner Dr. Jim Pauly, whom I am honored to have as a part of my defense. It was very fortuitous for us to cross paths and I deeply value his skills, knowledge, and collaboration. I am deeply grateful for the friendship and mentorship of Satrio Husodo and Dr. Madushi Raththagala during my early years in the Gentry lab. Both are talented scientists and I am glad to have known and worked with them. I am also very thankful for Dr. Ramon Sun, whose vibrancy, energy, skills and brilliance are unmatched. Ramon has been an enormous asset to the Gentry lab. The lab continues to grow and flourish over the years and I am thankful for each and every individual who passes through. I am grateful for the foundational work of Amanda Sherwood and Dave Meekins, which paved the way for mine. I deeply appreciate the hard work, help and patience of Annette Uittenbogaard and Grant Austin. I want to thank Corey Brizzee, Zoe Simmons, Shane Emanuelle, Shirlly Zhou, Jeremiah Wayne, Veronica Torres, Sarah Sternbach, and Kit Donohue, with whom I have worked on many projects, and Marina Falaleeva, Bobby Murphy, Lyndsay Young, and Andrea Kuchtová whose friendship and conversation strengthened me and helped me grow as a scientist and as a person. I also want to thank John McCarthy of the Physiology Department; Martin Chow of the University of Kentucky (UK) Protein Core; Michael Mendenhall of the Genetic Technologies Core; Thomas Wilkop and the UK Light Microscopy Core; Azin Akbari, Nico Briot, and Dali Qian of the UK Electron Microscopy Core; and the Biospecimen Procurement & Translational Pathology Shared Resource Facility of the UK Markey Cancer Center. I am grateful for the help and support of our

iii department chair Becky Dutch, Director of Graduate Studies Trevor Creamer, Carole Moncman, Skip Waechter, Jeff Rush, Cheylene Plummer, Rachel Putty, and the UK Biochemistry department. This work would not have been possible without our outside collaborators. First and foremost, I want to thank the team at Valerion Therapeutics: Tracy McKnight, Dustin Armstrong, Deb Ramsdell, Beth Goad, Rob Schaffer, Hal Landy, and Nadine Aziz. It has been an amazing opportunity for me to work so closely with this company. They are deeply passionate about basic science and therapies and also a personable and enjoyable group to work with. I am thankful for the collaboration and hospitality of Pascual Sanz, Ada García-Gimenez, and members of the Sanz lab at Instituto de Biomedicina de Valencia, Spain. Peter Roach and Anna DePaoli-Roach of Indiana University Purdue University Indianapolis have been invaluable friends and collaborators. I have enjoyed and appreciated the efforts and insights of Jean-Luc Putaux of Université Grenoble Alpes, France, and the help of Nico Szydlowski and Stanislas Helle of Université Lille, France. I am also very grateful for valuable feedback and fruitful discussions with Jordi Duran and Joan Guinovart of IRB Barcelona, Helle Waagepetersen and Kia Markussen of Copenhagen University, Mitch Sullivan, and Felix Nitschke. I am also grateful for the technical support of Sheng Li at University of California San Diego, Lance Hellman at University of Notre Dame, and Srinivas Chakravarthy at Advanced Photon Source, Argonne National Laboratory. Finally, I am eternally grateful for the love and support of my longtime friends: Kristen Gabriel, Laura Holbrook, Rachael John, Genny Coburn, Catherine John, Holly and Daniel Epperly, Elizabeth Williams, and the Lexington Christian Fellowship church. They carried me through many ups and downs during my years in Lexington. I am also deeply thankful for my parents Charlie and Sue, stepmom Suzy, sister Paige, and my loving aunts and uncles, who have always encouraged and supported me in my endeavors.

TABLE OF CONTENTS

ACKNOWLEDGMENTS...... iii

LIST OF TABLES ...... ix

LIST OF FIGURES ...... x

LIST OF ADDITIONAL FILES ...... xiii

Brain glycogen structure and its associated proteins ...... 1 1.1 Introduction ...... 1 1.2 Discovery of glycogen and its associated proteins ...... 2 1.2.1 Claude Bernard and early reports of brain glycogen ...... 2 1.2.2 The Cori years: advances in glycogen enzymology and glycogen structure 4 1.2.3 Emerging technologies with staying power ...... 6 1.2.4 Sugar nucleotides, reversible phosphorylation, and a glycogen primer ...... 9 1.2.5 Reviving interest in brain glycogen metabolism ...... 12 1.3 Glycogen Structure ...... 13 1.3.1 Basic structure of glycogen and other glucose polymers ...... 13 1.3.2 Glycogen architecture ...... 14 1.3.3 Glycogen structure and phosphate: insights from Lafora disease ...... 16 1.3.4 Glucosamine ...... 19 1.4 Cytosolic glycogen synthesis and degradation ...... 19 1.4.1 Glycogenin ...... 19 1.4.2 Glycogen synthase and phosphorylase ...... 20 1.4.3 Glycogen branching and debranching enzymes ...... 21 1.5 Glycophagy ...... 23 1.5.1 Lessons from Pompe disease ...... 23 1.5.2 Lessons from Lafora disease ...... 25 1.5.3 The benefits of glycophagy ...... 28 1.6 Glycogen ultrastructure and subcellular distribution ...... 28 1.6.1 Glycogen α, β, and γ particles ...... 28 1.6.2 Proglycogen and macroglycogen ...... 31 1.6.3 Nuclear glycogen...... 33 1.7 The glycogen granule and its associated proteins ...... 34 1.8 The future of brain glycogen research ...... 37

Materials and Methods ...... 49

2.1 Structure-guided in vitro studies of laforin and laforin mutants ...... 49 2.1.1 Cloning, protein expression, and protein purification ...... 49 2.1.2 Glycogen phosphate determination and dephosphorylation assays ...... 49 2.1.3 Site-specific dephosphorylation assays ...... 50 2.1.4 Differential scanning fluorimetry (DSF) ...... 50 2.1.5 Yeast two-hybrid assays...... 51 2.1.6 Protein phosphate content determination ...... 51 2.1.7 Dialysis with phosphate binders ...... 51 2.1.8 Hydrogen deuterium exchange mass spectrometry (DXMS) ...... 51 2.1.9 Analytical ultracentrifugation (AUC) ...... 52 2.1.10 Structural analysis ...... 52 2.2 LB purification, characterization, and VAL-0417 studies...... 52 2.2.1 Mouse lines ...... 52 2.2.2 Generation of the antibody-enzyme fusion VAL-0417 ...... 53 2.2.3 Purification of native LBs from Epm2a-/- and Epm2b-/- mice ...... 54 2.2.4 Pflüger method for polysaccharide isolation ...... 54 2.2.5 Separation of glycogen and LBs in native LB purification ...... 55 2.2.6 Iodine-based measurements ...... 55 2.2.7 Determination of LB phosphate content ...... 56 2.2.8 In vitro degradation assays ...... 56 2.2.9 Scanning electron microscopy ...... 57 2.2.10 Profiling of VAL-0417 degradation product by HPAEC-PAD ...... 57 2.2.11 Generation of HEK293-PTG/PP1Cα cells and polysaccharide quantitation 58 2.2.12 Cell culture studies of VAL-0417 uptake ...... 59 2.2.13 In vivo mouse studies with VAL-0417 ...... 59 2.2.14 Statistical analyses ...... 60 2.2.15 Enzyme-linked immunosorbent assay (ELISA) ...... 60 2.2.16 Confocal microscopy ...... 61 2.2.17 Fluorescence-associated capillary electrophoresis (FACE) ...... 61 2.2.18 Wide angle and small angle X-ray scattering (WAXS and SAXS) ...... 61 2.2.19 LB treatments and light microscopy ...... 62 2.2.20 Transmission electron microscopy (TEM) and lintnerization ...... 62

Personalized biochemistry of Lafora disease patient mutations ...... 63 3.1 Introduction ...... 63 3.2 Results ...... 64 3.2.1 Pathogenicity predictions of EPM2A missense mutations ...... 64 3.2.2 Stability and activity of laforin missense mutations ...... 66 3.2.3 Mutation-induced decoupling of the CBM and DSP domains ...... 69 3.2.4 Effects of LD mutations on protein-protein interactions ...... 72 3.2.5 Laforin binding sites for malin and PTG are disparate ...... 74

3.3 Discussion ...... 76 3.3.1 Functional classes of EPM2A missense mutations ...... 76 3.3.2 Laforin stability and interaction, not activity, are the primary defects of EPM2A missense mutations ...... 79 3.3.3 A slower progression in LD and genetic effects ...... 80 3.3.4 A biochemical pipeline for personalized medicine...... 81

The catalytic activity and dynamics of the human glycogen phosphatase laforin provide insights into Lafora disease ...... 101 4.1 Introduction ...... 101 4.2 Results ...... 103 4.2.1 The laforin active site ...... 103 4.2.2 The DSP-DSP dimer interface ...... 106 4.2.3 Distinct binding properties of WT and C266S laforin ...... 108 4.3 Discussion ...... 113

Targeting pathogenic Lafora bodies in Lafora disease using an antibody-enzyme fusion ...... 136 5.1 Introduction ...... 136 5.2 Results ...... 138 5.2.1 Construction of an antibody-enzyme fusion (AEF) that degrades starch ... 138 5.2.2 Isolation and characterization of LBs from LD mice ...... 139 5.2.3 Degradation of LBs by VAL-0417 ...... 141 5.2.4 VAL-0417 uptake and polysaccharide degradation in situ ...... 142 5.2.5 In vivo uptake of VAL-0417 and LB reduction in Epm2a-/- mice ...... 143 5.3 Discussion ...... 146

The structure of Lafora disease polyglucosan bodies ...... 166 6.1 Introduction ...... 166 6.2 Results ...... 168 6.2.1 Confocal micrographs of iodine-stained LBs ...... 168 6.2.2 Differences in chain length distribution (CLD) between LB types ...... 169 6.2.3 LBs are B-type crystallites lacking long-range order ...... 170 6.2.4 Effects of mechanical, thermal and chemical treatments on LBs ...... 171 6.2.5 Transmission electron microscopy and lintnerization of LBs ...... 174 6.3 Discussion ...... 175 6.3.1 LB architecture depends on cellular origin ...... 175 6.3.2 Proposed model for LB architecture and formation ...... 177 6.3.3 Nomenclature: Polysaccharides, glycogen, polyglucosan, PGBs and LBs178

vii

6.3.4 Similarities of LBs with other PGBs and amyloid and relevance to neurological disease ...... 179

Concluding Remarks ...... 191 7.1 Summary ...... 191 7.2 Limitations and Future Directions ...... 193 7.3 Conclusion ...... 195

REFERENCES ...... 196

VITA ...... 244

viii

LIST OF TABLES

Table 3.1 Clinical data for patients carrying EPM2A missense mutations analyzed in this study...... 82 Table 3.2 Predicted pathogenicity of selected EPM2A missense mutations...... 83 Table 3.3 Classification of LD-causing EPM2A missense mutations based on biochemical and clinical data...... 84 Table 4.1 Comparison of DSP molecules and interfaces in 4R30 and 4RKK. . 119 Table 4.2 Apparent binding coefficients from DSF titration experiments with various substrates...... 120

LIST OF FIGURES

Figure 1.1 Timeline of landmark discoveries in the history of glycogen-related research...... 38 Figure 1.2 Glycogen in the CNS of invertebrates and vertebrates...... 39 Figure 1.3 The Cori cycle and the tree-like structure of glycogen...... 40 Figure 1.4 Early studies of brain glycogen...... 41 Figure 1.5 Glycogen ultrastructure as demonstrated by electron microscopy. ... 42 Figure 1.6 Structure and synthesis of polyglucans...... 43 Figure 1.7 Chain length distribution (CLD) of glycogen and amylopectin demonstrated by HPAEC...... 44 Figure 1.8 Insights from Lafora disease, a non-classical glycogen storage disease...... 45 Figure 1.9 Size and structure of glycogen α and β particles, proglycogen and macroglycogen...... 46 Figure 1.10 Nuclear-cytoplasmic shuttling and colocalization of glycogen associated enzymes...... 47 Figure 1.11 Crystal structures and relative sizes of glycogen-associated proteins...... 48 Figure 3.1 Distribution of LD cases and predicted pathogenicity of EPM2A missense mutations...... 85 Figure 3.2 Distribution of LD missense mutations in the laforin structure...... 86 Figure 3.3 Recombinant laforin mutants purified from E. coli...... 87 Figure 3.4 Stability and carbohydrate binding of LD mutants...... 88 Figure 3.5 Glycogen phosphate levels, kinetics, and phosphatase activity of LD mutants...... 89 Figure 3.6 Kinetics of C3- and C6-labeled starch dephosphorylation...... 90 Figure 3.7 Effects of LD mutants on C3- and C6 dephosphorylation...... 91 Figure 3.8 Melting profiles of laforin mutants determined by differential scanning fluorimetry...... 92 Figure 3.9 Effects of carbohydrate binding on domain decoupling in Y294N and P301L...... 93 Figure 3.10 LD mutation "hotspot" along the allosteric path between the CBM and DSP domain...... 94 Figure 3.11 Purification and characterization of laforin CBM and DSP domain. . 95 Figure 3.12 Expression levels of LD mutants fused to LexA...... 96 Figure 3.13 Effects of LD mutants on the interaction of laforin with malin and PTG...... 97 Figure 3.14 Laforin as a scaffolding protein...... 98 Figure 3.15 A threshold of laforin function in LD...... 99 Figure 3.16 A personalized biochemistry pipeline for LD patients...... 100

Figure 4.1 Comparison of glucan-free DSP domain (PDB: 4R30) and glucan- bound full-length laforin (PDB: 4RKK) structures...... 121 Figure 4.2 Characterization of putative and proposed catalytic residues at the laforin active site...... 122 Figure 4.3 Sequence and structural comparison of DSPs in the glucan phosphatase family...... 123 Figure 4.4 Contributions of active site residues to laforin stability, binding and activity...... 124 Figure 4.5 Oligomeric state of the laforin CBM...... 125 Figure 4.6 Residues of the DSP-DSP dimer interface contribute to cooperativity...... 126 Figure 4.7 Comparison of WT and C266S binding to glucans...... 127 Figure 4.8 C266S laforin scavenges phosphate from E. coli...... 128 Figure 4.9 Effects of sugars and phosphate binders on C266S phosphate...... 129 Figure 4.10 WT and C266S binding to phosphate and γ-cyclodextrin (γCD). ... 130 Figure 4.11 The effect of phosphate on WT and C266S binding to various carbohydrates...... 131 Figure 4.12 Model of phosphate-induced cooperativity switch...... 132 Figure 4.13 Glucan binding and phosphate scavenging of catalytic mutants. .. 133 Figure 4.14 C266S is conformationally distinct from WT laforin and interacts more strongly with malin and PTG...... 134 Figure 4.15 A model for C266S-mediated rescue of laforin deficiency in LD. ... 135 Figure 5.1 The antibody-enzyme fusion VAL-0417 degrades starch, a proxy for LBs...... 150 Figure 5.2 Starch degradation with different VAL-0417:starch ratios...... 151 Figure 5.3 A novel protocol for isolating native LBs from LD mice...... 152 Figure 5.4 Iodine staining of LB purification fractions...... 153 Figure 5.5 Size distribution of LBs purified from different tissues and appearance of dust-like particles...... 154 Figure 5.6 Scanning electron micrographs of LBs from brain, heart and skeletal muscle, washed in ethanol, dried, and applied to carbon tape...... 155 Figure 5.7 Scanning electron micrographs of isolated LBs...... 156 Figure 5.8 Soluble fractions of untreated and treated BrLBs after degradation reactions...... 157 Figure 5.9 Purification of LBs from skeletal muscle of 12 month old Epm2b-/- mice...... 158 Figure 5.10 VAL-0417 degrades LBs in vitro...... 159 Figure 5.11 HPAEC-PAD chromatograms of control samples and additional time points from LB degradation with VAL-0417...... 160 Figure 5.12 Polysaccharide levels relative to protein concentration in three cell lines...... 161 Figure 5.13 VAL-0417 uptake and polysaccharide reduction in cell culture. .... 162

Figure 5.14 VAL-0417 reduces LB load in vivo after intramuscular (IM) or intravenous (IV) injection...... 163 Figure 5.15 Comparison of intrathecal (IT) versus intracerebroventricular (ICV) administration in WT mice...... 164 Figure 5.16 VAL-0417 reduces LB load in vivo after continuous ICV infusion. . 165 Figure 6.1 Representative z-stack images of iodine-stained starch and LBs. .. 181 Figure 6.2 Additional z-stack images...... 182 Figure 6.3 Chain length distribution of LB types determined by FACE...... 183 Figure 6.4 Wide-angle X-ray scattering (WAXS) of LBs and phytoglycogen. ... 184 Figure 6.5 Small-angle X-ray scattering (SAXS) of LBs and potato amylopectin starch...... 185 Figure 6.6 Effects of various treatments on the structure of LBs...... 187 Figure 6.7 Structural transition of SmLBs during heating...... 188 Figure 6.8 Transmission electron micrographs of LBs exposed to various treatments...... 189 Figure 6.9 A model for LB structure and formation...... 190

xii

LIST OF ADDITIONAL FILES

Supplemental File 1 SASA…………………………….……..………[PDF 393 KB] Supplemental File 2 all_mutant_data………………….……..…….[PDF 65 KB]

xiii

BRAIN GLYCOGEN STRUCTURE AND ITS ASSOCIATED PROTEINS

1.1 Introduction

In the 1840s, the French physiologist Claude Bernard made an astonishing observation: that sugar could be synthesized de novo by the liver of a dog that was exclusively fed meat (1). In 1857, he described the isolation of this sugar- forming substance, which he aptly called la matière glycogène (2). Soon afterward, glycogen in muscle was reported, shown to diminish with exercise, and discovered to "ferment" into lactic acid. Over the century that followed, the structure and regulation of glycogen were extensively characterized in the context of liver and muscle, leading to multiple Nobel prizes throughout the 20th century. Although the mechanisms governing glycogen metabolism have been deeply studied, the work is not yet complete. Meticulous investigations of Nobel laureates Otto Meyerhof, Carl and Gerty Cori, Luis Leloir, Edwin Krebs and Edmond Fischer, and countless others have paved the way for a profound understanding of glycogen metabolism, even in organs where glycogen constitutes a smaller fraction of total weight, such as the brain. Despite glycogen metabolism being a mainstay in textbooks, new aspects of glycogen structure, metabolism and tissue-specific roles are still being defined, aided by modern technologies and the mechanistic investigations of glycogen storage diseases. This chapter discusses the historical progression of glycogen research and its current condition with particular emphasis on how it has benefitted the overarching field of biochemistry, why the study of brain glycogen has lagged, and what gaps remain in our understanding of glycogen metabolism. There are aspects of glycogen structure and metabolism that are applicable to all types of glycogen, but some aspects are specific to organisms, tissues, and/or nutritional state, and insights can be gleaned through study of the other types of carbohydrates, particularly plant starch. Given the similarities and connections, it is useful to review the literature on glycogen (and starch) to understand what can be extrapolated to brain glycogen. This chapter provides a detailed overview of our current understanding of glycogen structure, architecture, associated proteins, tissue-specific properties and subcellular distribution. Glycogen storage diseases inform our studies of glycogen in ways that would not otherwise be possible; insights gained from Lafora and Pompe disease in particular are discussed. An in- depth understanding of glycogen metabolism and the basic mechanisms of these diseases pave the way for more effective diagnoses and therapies, as discussed in later chapters.

1.2 Discovery of glycogen and its associated proteins

1.2.1 Claude Bernard and early reports of brain glycogen

Glycogen constitutes up to 8% of liver volume and up to 2% of muscle volume, and its content is higher in these tissues than in any other tissue of the adult mammal (3). It is not surprising that the fundamental study of its structure and metabolism began in liver and muscle. Claude Bernard, through a series of fundamental experiments, discovered the primary role of the liver in supplying glucose to other organs through the bloodstream (reviewed by (4-6). In 1848, Bernard noticed that after death, a large amount of glucose was released from the hepatic veins of a dog, even when the dog had not been fed any carbohydrates (6). He concluded that the liver could produce glucose, and by 1857, reported that he could extract glycogen (a term that literally means "sugar-former") by plunging the tissue in boiling water shortly after death and precipitating the material with alcohol (Figure 1.1) (2). Through fermentation with yeast and basic chemical tests, he confirmed that its main constituent was glucose (5). He also noticed that the glycogen yielded a red wine color in the presence of iodine, which facilitated its histological observation in other tissues (7, 8). Very shortly afterward, Sanson reported the isolation of "a glycogen-like substance analogous to dextrin" from the spleen, muscle and kidneys of a horse (translated, (9) (Figure 1.1). By 1875, Bernard and others reported glycogen in mammalian muscle, embryos, and trace amounts in other tissues, but it was difficult to know for certain that those trace amounts were indeed glycogen (10). Soon after Bernard's discovery of glycogen, it was noted that glycogen in muscle diminished upon contraction, and Bernard reported that "muscle glycogen always undergoes a lactic acid fermentation, and this is the only change that muscle glycogen ever undergoes, either in the living animal or after death" (5, 10). In the early twentieth century, a landmark paper from Fletcher and Hopkins demonstrated that lactic acid was formed in muscle under anaerobic conditions (11). Then, a German physician named Otto Meyerhof published a series of experiments demonstrating that the lactic acid produced by frog muscle was derived from glycogen, and that in the presence of oxygen, it could be either oxidized, or reconverted to glycogen (Figure 1.1) (12-14). "Meyerhof's brilliant analysis of the glycogen-lactic acid cycle and its relation to respiration explained the course of the heat production and, for the first time, established the cyclic character of energy transformations in the living cell" (15). For this discovery and his seminal work on glycolysis, Meyerhof and his colleague A.V. Hill were awarded the 1922 Nobel Prize in Physiology or Medicine. The German physiologist Eduard Pflüger is credited as the first to establish a method for obtaining pure fractions of glycogen (Figure 1.1), and modifications of his method are still used today (4, 16, 17). While Pflüger, Bernard and most others were unable to detect glycogen in the nervous system, traces of brain glycogen in

2 the normal and diabetic human brain were reported by Cramer in 1880 and Pavy in 1881 (18). Similar findings were published in an 1890 edition of The Lancet: "Dr. Fütterer has examined various organs of a diabetic person, finding glycogen in the medulla oblongata, spinal cord, and kidney in large quantities, and a little in the liver. A careful examination of the cerebral cortex showed that the vessels were full of glycogen. He concluded that extensive disturbances of nutrition were bound to result from this" (19). Although nearly all studies of glycogen in the early 1900s were in the liver and muscle of various organisms, as well as in yeast, a few groups reported glycogen in the central nervous system (CNS) of animals by staining tissues with an iodine solution. Glycogen was observed in the CNS, often in the retina, of the lancelet, lamprey, frog, pigeon, rabbit, and dog (18, 20). Gage asserted its presence in the CNS of lower organisms, in the dorsal root ganglia of the pig, and in the choroid plexus of human embryos (Figure 1.2). He validated the identity of the "mahogany-red substance" by incubating the sections with saliva, which contained amylases that digested the glycogen so that it no longer stained. "With the higher vertebrates, glycogen in demonstrable amount is not found in the nervous system after the embryonic period, the liver and muscles then assuming the main glycogenic function." Gage was hopeful that with the proper techniques, glycogen would also be found in the adult human CNS (18). With the relationship between glycogen and lactic acid demonstrated in muscle by Meyerhof, a few groups began investigating this relationship in brain. However, brain glycogen and lactic acid remained difficult to isolate and detect, and results were variable. Some groups reported a reduction in glycogen with insulin- stimulated convulsions or anesthetics, while others reported a rise in glycogen after convulsions produced by methylguanidine; Holmes and Holmes in reviewing these studies argued that the low values reported for brain glycogen were within the normal range of variation, and that brain glycogen is unlikely to produce lactic acid under normal conditions (20). They astutely commented that "it is possible that the low values which we, in common with other workers, find for the glycogen content of the brain, may be due, not to a slow or inadequate breakdown mechanism, but to an extremely rapid one." This was indeed the case, and methods for capturing brain glycogen improved over the years; however, it was not until the end of the 20th century that the most sophisticated methods for measuring brain glycogen emerged. "It seemed to us, therefore, that it is quite possible that glycogen plays a comparatively unimportant (or at least quite obscure) part in the carbohydrate metabolism of mammalian brain" (20). This sentiment prevailed for many years, and it did not become clear until the following century that the role of glycogen in brain was more obscure than it was unimportant. Meanwhile, studies on glycogen enzymology had just begun, which led to a series of fundamental discoveries with major impacts on the wider field of biochemistry.

1.2.2 The Cori years: advances in glycogen enzymology and glycogen structure

Soon after Meyerhof's Nobel prize was awarded, Carl and Gerty Cori, Czech- born medical doctors who had emigrated to the United States, started working on the effects of insulin and epinephrine on liver glycogen. They reported that the hyperglycemia observed in the blood after epinephrine injections could not be accounted for solely by liver glycogen, and that lactic acid, which would be derived from muscle, could give rise to newly formed liver glycogen (21). They theorized that there was a "cycle of carbohydrates" linking liver glycogen and muscle glycogen, with the intermediates between the two being glucose and lactic acid (Figure 1.3a). Their subsequent work demonstrated this theory to be true, and it became known as the Cori cycle. The Coris made a series of major discoveries in the 1930s (Figure 1.1). In 1936 they isolated a novel ester, glucose-1-phosphate, from a mixture of glycogen, inorganic phosphate, and aqueous muscle extract (22). So significant was this discovery that the compound became known as the Cori ester (7). In subsequent years, the Coris showed that a phosphorylating enzyme found in muscle, heart, liver, brain, and yeast extracts catalyzed the release of glucosyl moieties from glycogen by esterifying them with inorganic phosphate. The resulting glucose-1- phosphate was converted to glucose-6-phosphate by an enzyme they named phosphoglucomutase (23, 24). In a 1939 article in Science, they reported that the enzyme producing the Cori ester (which they aptly named phosphorylase) could also catalyze the reverse reaction, in order to synthesize polysaccharide (25). Almost simultaneously, phosphorolytic enzymes catalyzing polysaccharide synthesis were being described in the extracts of peas and potatoes and in yeast (26, 27). In 1943, Cori, Cori and Green crystallized glycogen phosphorylase and published a series of seminal papers on its properties (28-31). In reward for their pioneering work, the Coris shared the 1946 Nobel Prize in Physiology or Medicine with Bernardo Houssay, who also made important contributions to the field of carbohydrate metabolism through his study of the role of the pituitary gland in hormonal regulation. The discovery of phosphorylase in mammals, plants and yeast was a major breakthrough in the overarching field of biochemistry (7). Phosphorylase was the first enzyme demonstrated to synthesize polysaccharide, and glycogen was the first macromolecule to be synthesized in vitro (32, 33). However, while mammals and yeast utilize glycogen as an energy source, plants synthesize starch, a glucose polymer with very different properties. Investigations of plant starch had been underway years before Bernard's discovery of glycogen and were progressing in parallel with those of glycogen. By the 1940s it had become clear that starch was composed of two distinct polymeric fractions: amylose, which was composed of long linear chains of glucose and gave an intense blue-black color with iodine, and amylopectin, which was a branched macromolecule and gave a

4 purplish to reddish color (34). While the Coris were working out glycogen enzymology, others were studying the polymeric nature of both glycogen and starch. Haworth and Percival established a method for identifying the chemical linkages in glucose polymers through methylation of the free hydroxyls of the glucose polymer prior to hydrolysis with acid (35). They found that since most of the product had unmethylated hydroxyls at the 1- and 4- positions, the majority of the glucose units in glycogen must be linked by α-1,4 glycosidic bonds. A small proportion of glucose was methylated at the 4-hydroxyl position, which would represent what is called the non-reducing end of a glucose chain (i.e. the end lacking an aldehyde group). Based on the proportion of the methylated products, the average linear chain in glycogen was proposed to contain 12 glucose units. Similar work was done on several varieties of plant starch, and the average chain length in starch was estimated to be 24-30 units (36). In 1947 and 1949, the α- 1,6-linked disaccharide isomaltose was isolated from hydrolyzed starch and glycogen, providing definitive evidence for the chemical identity of the branch points (37, 38). For many years it was believed that phosphorylase catalyzed both the synthesis and degradation of glycogen and starch, although synthesis by phosphorylase always required a primer, i.e. the addition of a small amount of carbohydrate upon which phosphorylase could act (7, 39). Curiously, phosphorylase from brain, heart and liver extracts produced branched polysaccharides that yielded a similar color to glycogen when stained with iodine, while the enzyme from potato and muscle extracts synthesized a linear, unbranched polysaccharide resembling amylose (31). It was hypothesized that the former preparations contained a contaminating enzyme that was capable of introducing the α-1,6-linked branch points. A branching enzyme had been identified in potato, and in 1953, Larner characterized the branching enzyme from rat liver and muscle extracts. Using isotopic labeling, he showed that potato, liver and muscle branching enzymes transferred a 1,4-linked chain of 6-11 glucose moieties to create the 1,6-linked branch. Thus, branching enzyme is considered a transglucosidase (40). His approach was novel in that he utilized radioactive isotopes rather than measuring branching by a shift in iodine color, susceptibility to phosphorylase, or change in end-group. With the combination of phosphorylase and branching enzyme, glycogen could be synthesized in vitro, and for the most part, it resembled the glycogen purified from tissues (41). Similarly, it was observed that while muscle phosphorylase could completely digest glycogen, recrystallized preparations of muscle phosphorylase and potato phosphorylase degraded glycogen only partially, stopping at the branch points, and leaving what was called a limit dextrin (42). A second enzyme must be present in the crude extracts that could remove the branch points. In 1951, Cori and Larner discovered the contaminating enzyme, showing that it cleaved α-1,6 linkages and released free glucose (42). Walker and Whelan later found that the phosphorylase limit dextrin contained not one, but four glucose units attached to branch points

(43). It was then discovered that the enzyme they had identified, which they called amylo-1,6-glucosidase, also had transglucosidase activity, favoring the transfer of three glucose units from one chain to another (44). It is now well established that when phosphorolysis terminates four units away from a branch point, the glycogen debranching enzyme, which possesses both transferase and glucosidase activities, moves three glucose units to another chain and then releases the single branched glucosyl moiety (45). It appeared at that time that all of the enzymes required for glycogen synthesis and degradation had been identified: glycogen phosphorylase, branching enzyme, and debranching enzyme. Enzymes that were either analogous or distinct to these had also been identified in plants. It became possible to utilize these enzymes to more precisely understand the structure of the glycogen and amylopectin molecules. With such useful tools and insights in common, the fields of starch and glycogen metabolism continued to progress together and overlapped considerably. In 1940 a tree-like structure for glycogen and amylopectin was proposed by Meyer and Bernfeld based on serial degradation with various enzymes (Figure 1.1) (46). This model varied from the prior structures proposed by Staudinger and Haworth, which were based solely on data obtained by chemical methods. The Cori group used stepwise enzymatic degradation of various glycogens and amylopectins with phosphorylase and debranching enzyme to test the various models. It was evident that glycogen was composed of "tiers," and each successive tier could only be accessed by phosphorylase after removal of the branch points by the debranching enzyme; degradation of multiple tiers required alternating treatments of the two enzymes. "The Staudinger and the Haworth models would yield a constant percentage of the total branch points in each tier during successive enzymatic degradation, while the Meyer model would yield a diminishing percentage as one progresses from the outer to the inner tiers, such as is actually indicated by the data in Table II. Only one kind of model could be made to fit this arrangement of branch points; namely, one which represents the polysaccharides as multibranched, tree-like structures" (39) (Figure 1.3b). It was also noted in this paper that the average chain length of liver and muscle glycogen (15 glucose units) were shorter than those of wheat and corn amylopectin (18 and 24 glucose units, respectively). We now know that chain length is one of the major features that distinguish glycogen from amylopectin. Meyer's model was the accepted model for glycogen and amylopectin until experiments from Whelan's group in 1970 showed the absence of very long chains and refined the model (Figure 1.1) (47, 48). The currently accepted model for glycogen is based on Whelan's work, but recent data indicates this model requires further refinement (see Section 1.3.1).

1.2.3 Emerging technologies with staying power

When the Coris began their work, the most common procedure for purifying

6 glycogen was based on the 1909 Pflüger method. The protocol is similar to what Bernard used and involves boiling tissue in hot concentrated alkali and then precipitating the glycogen with alcohol (16). Modifications of this procedure, especially those of Somogyi (49), were used by the Coris and are still the most commonly utilized protocols for glycogen extraction. Another frequently used method, introduced in 1934 (50), is extraction with cold trichloroacetic acid (TCA); however, it has been observed that not all glycogen can be extracted by this method (see Section 1.6.2), and a slightly different molecular weight is observed for the purified glycogen (51, 52). In 1936, the first reliable procedure for extracting brain glycogen was established utilizing Somogyi's method (53). Kerr found that he obtained the highest yields when the brains of dogs were rapidly excised under amytal anesthesia and crushed directly into 60% potassium hydroxide. He also obtained high yields when the brains were frozen in situ in liquid air. Importantly, he showed that the purified brain glycogen "is indistinguishable from liver glycogen prepared by Somogyi’s method. It dissolves readily in water to form a transparent solution, which is opalescent in reflected light. This when treated with iodine solution gives a Burgundy red color identical with that obtained with a solution of liver glycogen." The similar chemical properties of brain and liver glycogen were important to establish, and made it reasonable to extend the principles of glycogen structure and regulation that were being so thoroughly defined in liver and muscle to brain glycogen. Quantifications of brain glycogen were now reproducible. Kerr and other groups showed that brain glycogen levels dropped during hypoglycemia in dogs, cats and rabbits (54). Once a robust method for quantifying low levels of brain lactate was also established (55), informative experiments on the effects of various stimuli on brain glycogen and lactate could begin. Using a circular saw to rapidly excise the brains of mice, Chance and Yaxley showed that lactate was elevated with both subconvulsive and convulsive levels of various seizure-inducing stimulants, while glycogen was only elevated when a convulsion was reached (56). They also carefully defined the rapid decrease of glycogen and lactate levels after death in both normal and convulsed mice, demonstrating that most of the glycogen was lost after five minutes (Figure 1.4a). Based on their own measurements and those of their contemporaries, they estimated that "a large number of vertebrates normally possess between 10 and 30 mg of glycogen per 100 g of brain tissue" and that glycogen distribution within the brain was heterogeneous. These values are in agreement with current estimates (3). A few groups also began studying glycogen metabolism in isolated cerebral tissues and the effects of sleep on brain glycogen (57-59). As glycogen quantification methods improved, so did microscopic techniques. Glycogen had long been visualized with iodine using the light microscope, but the stain was not sensitive enough to detect the small amounts found in the normal adult brain. Additionally, alcoholic fixation was also critical, as glycogen would dissolve in aqueous solutions (18). Some attempts were made to visualize

7 glycogen with Best's carnitine or the Bauer reaction, but these methods were also criticized for their lack of sensitivity (53, 60, 61). In 1946, McManus introduced the periodic acid-Schiff (PAS) technique as a delicate and convenient way to stain carbohydrates, including glycogen, in tissue sections: the periodic acid reacts with the 1,2 glycol linkage of carbohydrates, producing an aldehyde that can be colored with Schiff reagent (62, 63) (Figure 1.1). Although PAS stained most polysaccharides, including glycoproteins and mucins, its specificity for glycogen could be tested by incubation with diastase (a general term for amylase) or saliva (which contains amylase) to digest the glycogen and distinguish it from other types of polysaccharides, which were left intact after digestion (60). Using PAS and a modification of this stain (lead-tetra-acetate-Schiff), Shimizu and Kumamoto published a series of papers in the 1950s on glycogen deposition in the normal brain and with pathological insults (64-67). Glycogen was detected in the area postrema (an area populated primarily by glial cells) in mammalian brains, particularly around blood vessels, and in ependymal cells of the supraoptic crest in rodents; it could also be detected in nerve cells of the hypothalamic nucleus and other regions when the brain was perfused with appropriate fixatives (Figure 1.4b) (65). The PAS technique, usually in combination with diastase (PASD), is now the most widely used histochemical stain for glycogen detection in tissues (68). Another extremely useful technology in visualizing glycogen structure and distribution was born with the invention of the electron microscope in 1931 by Ernst Ruska (69). The electron microscope achieved superior resolution to that of the light microscope. In 1934, Marton developed a histological technique that allowed biological specimens to be visualized by the electron microscope without destruction (70). With this technique, the fine structure of subcellular organelles could be visualized. The first report of glycogen using electron microscopy (EM) was from Husemann and Helmut Ruska, Ernst Ruska's brother, in 1940 (71) (Figure 1.1). The spherical particles, which were derivatized to produce better scattering of the X-ray beam, were about 15-30 nm in diameter. But visualization of glycogen was difficult for a number of reasons. Firstly, polysaccharides contain primarily light atoms (carbon, oxygen and hydrogen) and are intrinsically not very electron dense. Secondly, after the identification of ribonucleoprotein particles by EM, particulate components with a diameter of ~150 Å free in the cytoplasm or associated with the endoplasmic reticulum were typically interpreted as ribonucleoprotein (72, 73). Unknown tissue components were sometimes identified by comparing the thin EM sections with thicker, stained sections viewed under the light microscope, but this too had its limits. The ultracentrifuge, invented by Theodor Svedburg in 1925, became a very powerful tool in the field of biochemistry and allowed glycogen to be isolated its native form. Lazarow showed in 1942 a species which he called "particulate glycogen" could be isolated from liver via ultracentrifugation without the use of harsh chemicals (Figure 1.1) (51). Its sedimentation constant suggested a molecular weight much larger than previous reports for Pflüger-purified glycogen,

8 and the particle, which he showed was made of pure glycogen, could be dispersed by heating, trichloroacetic acid (TCA), or potassium hydroxide treatment. Lazarow concluded that "clearly, then, particulate glycogen is an aggregate of smaller glycogen units." In 1962, Drochmans, combining these two state-of-the art technologies and utilizing a negative staining technique with phosphotungstate, elegantly described the ultrastructure of glycogen particles purified by differential ultracentrifugation (74) (Figure 1.1). He defined three levels of glycogen structure in the rat liver visible by EM: the largest structure was the α particle, which was 60-200 nm in diameter and had a rosette like appearance; the β particle, 20-40 nm in diameter, which associated to form α particles; and the γ particle, which referred to the 3 nm fine filaments comprising the β particle (Figure 1.5a,b). The α particles could be dispersed into β particles in low pH. It became clear that Drochmans's α particle was equivalent to Lazarow's particulate glycogen, and that these particles represent the primary native form of glycogen in the liver. This helped to explain the discrepancies in molecular weights that others had observed for glycogen purified by different methods. In 1968, Wanson and Drochmans also visualized differentially centrifuged glycogen from rabbit skeletal muscle, showing the presence of solely β particles (Figure 1.5c) (75). Dozens of EM studies on glycogen in various tissues and organisms followed. The use of EM on sections of the mammalian brain in the 1960s and later years made it clear that the vast majority of brain glycogen was present in the form of β particles in astrocytic processes (Figure 1.5d) (76, 77). Multiple groups showed striking accumulations of glycogen in reactive astrocytes in response to various traumas (Figure 1.5e) (66, 67, 76). Most groups reported that in the adult mammalian brain, neurons and microglia contained virtually no glycogen except under certain pathological conditions; however, in a few studies of normal tissue, glycogen was found in nerve cells (reviewed by (78)). During this time, numerous groups began using PAS and EM to visualize the microscopic characteristics of polyglucosan bodies, abnormal carbohydrate structures found in a variety of tissues. Polyglucosan bodies were identified in both astrocytes and neurons, in the context of glycogen storage diseases, epilepsy, neurodegenerative disorders, and aging (79, 80). Although glycogen still received little attention in the brain compared to other tissues, its significance was becoming increasingly evident. In 1961, A.W. Merrick stated, "A substantial amount of Russian work within the past decade has been directed toward brain glycogen changes during various function and biochemical states of the animal. The opinion is offered by many of these investigators (several of whom are cited in this paper) that glycogen takes an active part in brain metabolism" (81).

1.2.4 Sugar nucleotides, reversible phosphorylation, and a glycogen primer

In 1950, the Argentinian chemist Luis Leloir discovered the first sugar nucleotide: uridine-diphosphate (UDP) glucose (82). It soon became apparent that

9 the sugar nucleotides were found throughout nature and were essential for the interconversion of carbohydrates, occupying "a central position in carbohydrate metabolism" (83). In 1957, Leloir made the shocking discovery that the true catalyst of glycogen synthesis was not glycogen phosphorylase, but a glucosyl transferase utilizing UDP-glucose as a glucosyl donor, which became known as glycogen synthase (84) (Figure 1.1). In the words of Larner, "Nature seems to have found a new more powerful glucose donor by making a pyrophosphate derivative of the Cori ester" (7). Leloir's group went on to show that glycogen synthesized in vitro using glycogen synthase in combination with branching enzyme was identical to natively purified glycogen β particles from muscle; in contrast, glycogen synthesized by phosphorylase and branching enzyme was not exactly equivalent to native glycogen: it differed in its stability and response to various types of degradative treatments (85, 86). These results were confirmed and expanded upon by numerous groups, and Leloir was awarded the 1970 Nobel Prize in Chemistry for his discovery of the sugar nucleotides. It became apparent that both glycogen phosphorylase and synthase are peculiar enzymes. The Coris observed that glycogen phosphorylase existed in two forms, an active a form, and inactive b form, which were differentially activated by adenylic acid (87). The Coris had identified another enzyme capable of converting the a form to the b form, calling this the "prosthetic-group-removing" enzyme (29). Kinase activity, the covalent addition of phosphate to one protein by another, was first described in 1954 (88), and in 1955, Fisher and Krebs showed that phosphorylase a could be converted to phosphorylase b in the presence of adenosine triphosphate (ATP) by an enzyme from muscle that became known as phosphorylase kinase (PhK) (89). Concurrently, Wosilait and Sutherland also demonstrated the presence of an equivalent converting enzyme in liver (90). These were the first ever demonstrations of reversible phosphorylation, a momentous discovery that has reverberated throughout the life sciences. Fischer and Krebs published a series of papers on the details of this novel mechanism that earned them the 1970 Nobel Prize in Physiology or Medicine, and Sutherland received the 1971 Nobel Prize for his discovery of cyclic adenosine monophosphate (cAMP) and its role in hormonal regulation (91). A few years later, Larner and others showed that glycogen synthase is also subject to allosteric control, that it exists in two forms, and that the interconversion of these forms requires phosphorylation by another kinase (92). In fact, unlike phosphorylase, which has one phospho-site and one kinase, synthase is hierarchically phosphorylated at multiple sites by multiple kinases, inextricably linking glycogen metabolism to hormonal regulation, energy status, and a milieu of intracellular signals (93). One of the kinases discovered to phosphorylate synthase, glycogen synthase kinase 3 (GSK3), was later found to regulate a much greater array of cellular processes and play an important role in many pathologies including cancer, Alzheimer's disease, Parkinson's disease and diabetes (94, 95). Many of the roles of GSK3 are quite unrelated to glycogen metabolism, leading to an

10 unfortunately misleading keyword in literature searches. While more pieces in the puzzle of glycogen regulation had fallen into place, other pieces were not even known to be missing. One enzyme must be mentioned at this point that has received insufficient attention despite its discovery in 1963. It was discovered as the deficient enzyme in Pompe disease, one of the glycogen storage diseases (GSDs). GSDs, also known as glycogenoses, are a group of diverse pathologies characterized by glycogen accumulation in various tissues. GSDs were first described in the early 1900s, and over the decades, the enzyme deficiencies causing many of these disorders were identified through biochemical analyses of patient tissues. Typically, patients were lacking one of the basic enzymes involved in glycogen metabolism (reviewed in (45)). However, in one of the most severe GSDs, Pompe disease, all of the known glycogen-related enzymes had normal activity. Lysosomes had been recently discovered by de Duve with the use of the ultracentrifuge (96), and in 1963, Hers demonstrated that Pompe patients were lacking an enzyme called acid α-glucosidase or maltase, which was found in lysosomes and converts glycogen or maltose to glucose (97). He identified Pompe disease as the first of the lysosomal storage diseases, now known to have a combined incidence of 1 in 5,000-10,000 (98, 99). Brown, Brown and Jeffrey demonstrated that the lysosomal enzyme discovered by Hers could cleave both 1,4 and 1,6 glycosidic linkages (100-102). The role of this enzyme in the degradation of glycogen was unclear, but presumably it was important since its absence resulted in a fatal condition. It is now well established, but not broadly known, that glycogen can be degraded within lysosomes, but the details of this pathway are still being defined (see Section 1.5). The discoveries of Leloir, Krebs and Fischer spurred a major wave of glycogen- related research in the decade that followed (Figure 1.1). By 1970, these fervent investigations were waning, as it seemed to most that the work on glycogen was virtually complete. In 1971, Ryman and Whelan synthesized an exhaustive review of glycogen metabolism, which spanned 158 pages and cited nearly 900 publications. They stated in their introduction:

"The field of glycogen metabolism is one that to the outside observer has long seemed in a settled condition, with new discoveries being only likely to add gloss to existing facets. This is because it has been possible since the early 1940's to draw metabolic maps that seem to explain the process fully and satisfactorily, these maps being based on highly satisfactory in vitro experiments carried out with purified or semi-purified enzymes. This has been the polymer par excellence as far as in vitro work is concerned. This apparently settled condition is, in fact, illusory, and this has probably worked to the detriment of progress, since potentially interested investigators have almost certainly turned to other pursuits, feeling that no more major advances would be forthcoming. The true situation is the exact

opposite, and the ferment of activity now going on testifies to the growing realization that studies of glycogen metabolism have thrown up key discoveries that have the widest implications throughout biochemistry" (103).

One aspect that remained unclear was how glycogen synthesis was initiated in vivo. Both the Coris and Leloir observed that in vitro glucose polymerization by glycogen phosphorylase or glycogen synthase required the addition of a pre- formed carbohydrate primer (84, 104). Some investigators believed that glycogen synthesis could be initiated in vivo by glycogen synthase (105); others reported that a protein acted as the priming factor (106). In 1977, Whelan's group discovered a protein covalently bound to glycogen in liver, and later showed that the protein, which became known as glycogenin, was linked to glycogen via a novel tyrosine-glucose linkage (107, 108) (Figure 1.1). At first, Whelan's discovery was doubted, but his subsequent work and work from the laboratories of Cohen and Roach provided very strong support for this protein acting as the true primer of glycogen biogenesis. These groups established that glycogenin was a glucosyltransferase using UDP-glucose, and that after glycosylating itself, it could also extend the glucose chain to ~8 glucose residues, which would then be acted upon by glycogen synthase and branching enzyme (109-112). The initiation of glycogen synthesis by glycogenin is now widely accepted and has been extensively characterized, and similar protein primers for starch synthesis have also been described (reviewed by (113-115).

1.2.5 Reviving interest in brain glycogen metabolism

In recent years there has been a growing interest in studying brain metabolism, in part due to some important technological advancements. Firstly, in the 1970s, neurochemists began using focused microwave irradiation to rapidly and irreversibly inactivate enzymes in rodent brains in order to more accurately measure labile metabolites (reviewed by (116, 117). This technique was elegantly applied to determine the regional distribution of glycogen in the rat brain in the 1980s (118, 119) and more recently in the mouse brain (120). Secondly, after it was suggested that elevated glycogen in a GSD patient could be observed by 1H NMR, Choi et al. introduced the use of NMR spectroscopy to noninvansively measure the turnover of 13C-labelled glycogen in vivo in rodents (121, 122). Öz and colleagues applied this technique to humans, publishing a series of studies on the role of glycogen in normal human brain metabolism and during hypoglycemia (123-125). Another reason for the growing interest in brain glycogen is the emerging theme of metabolic coupling of astrocytes and neurons, particularly involving the transfer of lactate. These interactions are fascinating, complex, and highly debated. They have been recently discussed by multiple excellent reviews (126-130). The

12 coupling between astrocytes and neurons is reminiscent of the Cori cycle and the shuttling of lactate between liver and muscle; however, these interactions are more complicated and difficult to study. Glycogen metabolism is intimately linked to granule size, architecture, and its various associated proteins, which are still being investigated. Such topics are not frequently discussed in the context of brain glycogen metabolism. These nuanced aspects of glycogen regulation, in addition to its surprisingly dynamic nature, subcellular distribution, and alternative degradation pathways, are introduced in the next sections.

1.3 Glycogen Structure

1.3.1 Basic structure of glycogen and other glucose polymers

Glycogen is one of many types of polysaccharides that exist in biology. A polysaccharide refers to any large polymer comprised of many covalently linked monosaccharides; the bonds between them are known as glycosidic linkages. Glycogen and the two primary constituents of plant starch, amylopectin and amylose, are all homopolymers of glucose designed for energy storage in various organisms. We and others refer to these glucose homopolymers as polyglucans. Glycogen and amylopectin contain branches, while amylose is almost exclusively composed of linear chains. The linear chains of glucose are connected by α-1,4 glycosidic linkages, and the branch points are comprised of α-1,6 glycosidic linkages (Figure 1.6a). Because of the tree-like arrangement of branching, glycogen and amylopectin contain fewer reducing ends (i.e. ends with a free aldehyde group) than nonreducing ends, which are oriented outward. The α configuration of the glycosidic linkage causes a linear α-1,4 linked chain to twist, and when the chains are long enough, and uninhibited by branch points, they can form single or double helices (Figure 1.6b) (131). An alternative β-linked configuration of glucose polymers favors very straight chains that can associate into tight fibrils; this is characteristic of structural polyglucans such as cellulose. Mammals do not synthesize or hydrolyze β-glycosidic linkages, so they cannot digest cellulose, but they are well equipped to degrade glycogen and starch (132). Although the models of glycogen and amylopectin structure are both based on the tree-like arrangement originally proposed by Meyer and Bernfeld, the models were individually refined as it became increasingly evident that the arrangement and length of the branches of these glucose polymers were quite different. Glycogen contains a high degree of branching and short linear chains. By enzymatic analyses, there are, on average, 13 glucose units per α-1,4-linked linear chain, and 8% of the total glycosidic linkages are α-1,6 branch points (114, 133-136) (Figure 1.6c). Branching in glycogen is believed to be continuous, meaning the branch points are evenly distributed within the molecule. The presently accepted model for glycogen is based on the revisions of Whelan's group (52), but high resolution studies of chain length distribution (CLD) suggests

13 this model may still not be completely accurate (discussed below). Typically, a single glycogen is depicted as originating from a single chain covalently attached to a glycogenin molecule, which gives rise to all subsequent chains (Figure 1.6c) (114, 137). Enzymatic experiments indicate 3 to 4 glucose units between each branch point, so each linear chain would yield two branches (52, 133). Each new set of branches is referred to as a tier (Figure 1.6c), and since the number of linear chains doubles with each successive tier, the outermost tier would theoretically contain about one-third of the total glucose in the molecule, which is consistent with empirical observations (39, 136). Mathematical modeling has demonstrated that due to physical constraints, glycogen molecules can theoretically contain up to 12 tiers, corresponding to ~55,000 glucose molecules; beyond this size, the outer chains would become so crowded they would be inaccessible to enzymes (134). Indeed, the empirically observed upper limit of glycogen β-particles (44 nm) is consistent with the theoretical maximal diameter of a 12-tiered glycogen molecule (42 nm) (137, 138). The continuous branching within glycogen prevents, or dramatically limits, the formation of double helices, which would lead to insolubility of glycogen and inaccessibility of the glucan chains to enzymatic degradation (139). Mathematical modeling studies have also demonstrated that the empirically observed average chain length and branching degree are optimal for maintaining solubility and structural homogeneity (140). Amylopectin is also a branched polyglucan, but it contains longer chains of 20- 25 glucose units per chain with infrequent and clustered branching (Figure 1.6d) (141, 142). The clustering of branch points means there are regions of long, unbranched chains that associate to form crystalline, water-excluding double helices (Figure 1.6b). How these crystalline regions are arranged is still debated, but the most recent data suggests they are arranged along a polyglucan "backbone" (143). The alternating crystalline and amorphous layers within starch are advantageous as it allows for very dense glucose packing within starch granules, which can be as large as 100 μm in diameter (139). Amylose is believed to be interspersed among amylopectin molecules (Figure 1.6d). Glycogen, amylopectin, and amylose have diverse structural properties due to differences in chain length and degree of branching. Glycogen has the shortest average chain length and the highest degree of branching, usually 8%. The degree of branching in amylopectin is nearly half that of glycogen, only 4-6% (142, 144). Amylose is generally considered to be entirely composed of very long, linear chains, although infrequent branching has been reported (141).

1.3.2 Glycogen architecture

It is well established that amylopectin and glycogen differ in chain length and degree of branching. But a diverse array of polyglucans exist in nature with structures that vary beyond these two simple parameters. It is becoming increasingly apparent that the distribution of chain lengths and the arrangement of

14 branch points within a polyglucan are variable, producing distinct biological and physiochemical properties. Furthermore, not only do amylopectin and glycogen differ significantly from each other, but each polyglucan comes in a variety of shapes and properties based on species, tissue of origin, and even the nutritional state of the tissue. Polyglucan diversity in the context of starch has been exhaustively studied and thoroughly parameterized (145, 146). We will use the term "architecture" to refer to the unique chain length distribution, branching frequency and arrangement, and quantity and distribution of non-glucose moieties (notably phosphate) characteristic of a certain type of glycogen or amylopectin. Since the days of Bernard, investigators have utilized iodine to study polyglucan architecture. Bernard observed in 1877 that newly synthesized glycogen in rabbit muscle following exhaustive exercise gave a bluish coloration with iodine, in contrast to the reddish color of glycogen from rested muscle or well- fed liver (4). This colorimetric test can be made quantitative by an absorbance spectral scan: the wavelength of maximal absorption (λmax) increases with longer chains and less branching, reflected by the staining color (147-149). Glycogen from different sources yields a range of colors from yellow to reddish brown with iodine and produces a λmax of 420-490 nm; amylopectin gives a more intense color ranging from red to lavender with λmax of 490-570 nm; and amylose yields a blue to green color with a λmax of 580-640 nm (150-152). The iodine method is not perfectly quantitative, due to the various effects of the iodine and salt concentrations, temperature, and polysaccharide structure and source (151, 153). Additionally, although chain length and degree of branching describe different aspects of glycogen architecture, both parameters are related to λmax (150, 151, 154). It is difficult to deconvolve the various architectural parameters of polyglucans using just iodine, but it remains a convenient technique that yields useful information. High-performance anion exchange chromatography (HPAEC) was introduced in the 1990s to determine the chain length distribution (CLD) of enzymatically debranched polysaccharides with very high resolution (155). The CLD for various glycogens is asymmetrically unimodal: in bovine and rabbit liver glycogens, chains ranged from 3 to 35 degrees of polymerization (DP), peaking at DP 6-10 or 7-15, respectively (Figure 1.7a) (155, 156). This range of chain lengths suggests the architecture of glycogen may be intermediary between the Meyer and Whelan models. Amylopectin CLD profiles display a strikingly different bimodal CLD, with distinct groupings of short and long chains, further evidence for amylopectin chains being organized very differently than glycogen (Figure 1.7b) (157). CLD profiles for skeletal muscle and brain glycogen from mice have also been described recently. The Minassian group showed that the CLD profiles of mouse skeletal muscle and brain glycogen are nearly identical, ranging from DP 2-35 or more, with most chains being DP 4-15 (158, 159). Additionally, gradual isoamylolysis of muscle glycogen suggests that the internal chains are longer than the external chains (158). The Roach group showed a CLD for mouse skeletal

15 muscle glycogen ranging from DP 3-45, with most chains being 7-15 units long (160). After exercise and a 1 hour recovery, the CLD shifted toward long chains (Figure 1.7c). Similarly, muscle glycogen had an increased λmax after exercise and recovery, which was gradually restored over the course of 6 days. These data suggest that newly synthesized glycogen is less branched with longer chains that are remodeled over time (Figure 1.7d). It appears that glycogen molecule is not as homogenously structured as the Whelan model suggests, and chain length and branching appear to fluctuate based on nutritional state. Gerty Cori stated in a 1952 lecture that "the relationship of structure to nutritional state has not yet been fully explored, but it seems that glycogen freshly deposited after a fasting period is least branched and has long outer chains whereas the opposite is true of 'old' glycogens" (135). Additionally, mammalian and non-mammalian glycogens with similar average chain lengths produce differences in iodine spectra, suggesting that mammalian glycogens contain long chains in the interior of the granule that are not present in non- mammalian glycogens (161). Whole glycogen particles also display a continuum of sizes, and their size and ultrastructure vary with tissue type (see Section 2.6.1).

1.3.3 Glycogen structure and phosphate: insights from Lafora disease

Fontana first demonstrated that [32P]-labeled glycogen could be purified from the livers of rats after the administration of radioactive phosphoric acid (162). Whelan's group subsequently reported that mammalian muscle glycogen contains about 0.064% phosphorus by weight (163, 164). However, much like Whelan's reports of a proteinaceous component of glycogen that were later substantiated with the discovery of glycogenin, phosphate was initially treated as a contaminant. The physiological relevance of glycogen phosphate became evident through basic scientific investigations of Lafora disease (LD), a fatal, inherited childhood epilepsy characterized by accumulations of abnormal polysaccharides known as Lafora bodies (LBs). Historically, LD has not always listed among the glycogen storage diseases (GSDs) since a neurological phenotype predominates rather than a phenotype of muscle or liver dysfunction. However, like all other GSDs, LD is caused by genetic defect(s) leading to aberrant glycogen metabolism; it has therefore been included in recent reviews of GSDs and may be considered a non- classical GSD (165, 166). LBs have been shown to be the underlying cause of neurodegeneration and epilepsy in mouse models (recently reviewed by (167) (Figure 1.8a). In the 1960s and 1970s, it was demonstrated that LBs contain high levels of phosphorus and histologically resemble plant amylopectin (168, 169). In 1998 and 2003 the genes associated with the disease were identified, and it is now well established that virtually all LD patients carry recessive mutations in either the EPM2A or EPM2B gene (170-172). EPM2A encodes laforin, a glycogen phosphatase (173, 174), and EPM2B encodes malin, an E3 ubiquitin ligase (175, 176). Epm2a-/- and Epm2b-/- mouse models recapitulate the disease with respect

16 to LB formation and neurodegeneration (Figure 1.8b) (177-182). Purified polysaccharides from LD mice have a shifted λmax reminiscent of amylopectin (180), and their CLD profile shows a higher proportion of long chains than wild- type mice, most prominently in brain tissue (Figure 1.8c) (159). Unlike glycogen, purified LBs are quite large (>1 μm in diameter) and insoluble; they are presumed to be aggregates of a polysaccharide with abnormal architecture resulting from aberrant glycogen metabolism (167). Laforin and malin regulate two aspects of glycogen architecture that may be intertwined: chain length and phosphate level. Through studies of its role in LD, laforin was discovered to be the founding member of a class of enzymes known as glucan phosphatases that directly dephosphorylate carbohydrate substrates (173, 183-187). Laforin is the only known glucan phosphatase in mammals, and decreased laforin activity results in glycogen hyperphosphorylation. Malin ubiquitinates enzymes involved in glycogen metabolism, but the effects of ubiquitination are not clear (see Section 1.5.2). Surprisingly, the absence of malin also leads to hyperphosphorylation, and the absence of either enzyme leads to the accumulation of glycogen with abnormally long chains, which is puzzling (188). Recent results from mouse models overexpressing a catalytically inactive laforin demonstrate that the inactive laforin rescues the LD phenotype in the Epm2a- deficient mouse model (159, 189). As a result, the relevance of laforin's catalytic activity has been questioned, despite the very high conservation of its catalytic residues and the presence of glucan phosphatases across multiple kingdoms (183, 190). It has also been suggested that laforin and malin form a ubiquitination- targeting complex involved in the disposal of glycogen molecules with aberrantly long chains that could crystallize and cause the molecule to precipitate (188). A growing body of evidence suggest that laforin and malin are an integral part of an alternative route for glycogen degradation known as glycophagy, involving members of the autophagic pathway and lysosomal α-glucosidase (see Section 1.5.2). However, glycophagy is probably not reserved only for aberrantly structured glycogen molecules, since the accumulated lysosomal glycogen in acid α-glucosidase deficiency (i.e. Pompe disease) is of normal structure based on iodine staining (191, 192). Additional evidence for a physiological role of glycogen phosphate comes from the Roach lab and a substantial body of work from the field of starch metabolism. The Roach lab analyzed glycogen CLD and phosphate levels of laforin- deficient mice pre- and post-exercise. When muscle glycogen phosphate was depleted, phosphate levels remained suppressed even after total glycogen and CLD returned to normal (160). Laforin-deficient mice displayed exercise-induced glycogen depletion identical to wild- type mice, but phosphate levels remained elevated and iodine spectra suggested a delay in glycogen remodeling post- recovery. This study demonstrated that laforin dephosphorylates glycogen during exercise-induced cytosolic glycogeolysis, providing strong evidence for its physiological role as a glycogen phosphatase. Whelan suggested in 1994 that

17 phosphate was a marker for the age of a glycogen molecule; the suppression of glycogen phosphate after exhaustive exercise in wild-type mice does indeed suggest that phosphate accumulates with age, possibly from multiple cycles of glycogen degradation and re-synthesis (160, 163). The role of phosphate in glycogen metabolism is not clear, but studies from the starch field demonstrate that phosphate plays an important physiochemical and biological role in starch architecture and metabolism (193-195). In plants, phosphorylation is necessary for the proper synthesis of starch, and reversible phosphorylation is an integral part of starch breakdown. Only amylopectin contains significant amounts of phosphate, which is enriched in the amorphous regions (196, 197). Root and tuber starches, which exhibit less densely packed crystalline helices than other types of starches, have high phosphate levels (198, 199). Starch phosphate is covalently linked to the C3 and C6 hydroxyls of glucose moieties; about 70-80% of the phosphate is esterified to C6 (200). Phosphate, particularly at the C3 position, disrupts the crystalline helices in amylopectin by introducing steric hindrance, promoting their solubilization (201). Starch degradation is believed to be a cyclic process involving reversible phosphorylation and the concerted action of multiple enzymes: glucan dikinases, amylases, and glucan phosphatases (139). Two dikinases that respectively phosphorylate the C6 and C3 position in that order are glucan, water dikinase and phosphoglucan, water dikinase. Phosphorylation promotes solubilization of the glucan chains and facilitates their access by plant amylases. However, the amylases cannot proceed past a phosphate, so phosphate removal is achieved by the glucan phosphatases Starch Excess 4 (SEX4) and Like Sex Four 2 (LSF2), named after the plant phenotype that results from their deficiency and part of the same family as the glycogen phosphatase laforin (202-206). Crystal structures and structure-function studies of SEX4 and LSF2 preceded those of laforin (186, 187, 190, 207) and the cooperative study of both systems has led to a wealth of insight about polyglucan architecture and phosphorylation (139, 205, 208). Through studies of LD, the Roach and Minassian groups determined that normal glycogen contains about 1 phosphate moiety per 600-2500 glucose units, depending on species and tissue type (174, 181, 209-211). Phosphates are present as monoesters linked to the C2, C3 and C6 hydroxyls of glucose moieties, in approximately equal quantities (158, 210, 212). Data from multiple groups using various approaches strongly suggest the phosphate is concentrated at the interior of glycogen molecules (158, 160, 174). Liver glycogen contains less phosphate than muscle glycogen, while C6 phosphate has been detected in brain glycogen at similar levels to C6 phosphate in muscle glycogen (159, 163, 174, 211). The source of glycogen phosphate remains a mystery: although it has been reported that glycogen synthase can incorporate the β-phosphate of UDP-glucose into glycogen in a rare side reaction, this mechanism could only account for C2 and C3 phosphate, and the rate of incorporation is very low (1 in 10,000 catalytic

18 cycles) compared to the actual levels observed in glycogen (212, 213). Although no glucan dikinase has been identified in mammals, the physiological role of glycogen phosphate will eventually be defined, likely through the continued investigations of starch metabolism and the molecular mechanisms of LD.

1.3.4 Glucosamine

Glycogen also contains small amounts of covalent glucosamine. In the 1960s and 70s, Maley et al. reported the incorporation of glycosidically linked [14C]glucosamine into glycogen from [14C]galactosamine, and showed that glycogen synthase could catalyze glucosamine incorporation in vitro by using the substrate UDP-glucosamine instead of UDP-glucose (214, 215). Whelan also reported that intraperitoneal injection of [14C]galactosamine led to the incorporation of [14C]glucosamine into glycogen, replacing as much as 10% of the glucose residues (216). Covalently linked glucosamine in glycogen apparently would not block phosphorolysis, since it could be released as glucosamine-1- phosphate by phosphorylase. Whelan reported that normal pig and rabbit liver contained 86-266 nmol glucosamine per g glycogen, which was randomly distributed throughout the molecule (217). It is worth noting that an earlier study also demonstrating the incorporation of [14C] into glycogen from injected [14C]glucosamine showed that glucosyl residues were labeled, indicating [14C]glucosamine was converted to [14C]glucose through a deamination reaction prior to glycogen incorporation (218). Additional links between glycogen and glucosamine have been reported. Glycogen synthase can be modified by N-acetyl-glucosamine (GlcNAc), and it has been suggested that glycogenin could also be modified by GlcNAc (219, 220). Furthermore, studies have suggested that glycogen could be a carbohydrate source for protein glycosylation, for which glucosamine is a major precursor, and hypoglycosylation has been reported in GSDs (221-224). A physiological role for glycogen in protein glycosylation and the relevance of covalently linked glucosamine await further investigation.

1.4 Cytosolic glycogen synthesis and degradation

1.4.1 Glycogenin

Glycogen synthesis begins with glycogenin [UDP-α-glucose:glycogenin α- glucosyltransferase, EC 2.4.1.186], a member of glucosyltransferase family 8 (225) (www.cazy.org). The enzyme possesses two distinct enzymatic activities: self-glucosylation and chain elongation, both requiring UDP-glucose as a glucosyl donor and both occurring at the same active site, though the chemistries of the reactions are different (226, 227). For this reason the two activities were initially believed to belong to separate enzymes, until Whelan and Cohen discovered that both were accomplished by glycogenin (109, 110). Crystal structures and

19 biochemical studies show that glycogenin functions as an obligate dimer and requires Mn2+ for activity (228, 229). Autoglucosylation on tyrosine creates a glucose-1-O-tyrosyl linkage, a chemical linkage rarely found in nature, then glycogenin synthesizes a short oligosaccharide primer of at least 7-8 glucose units (226, 230). Even if this tyrosine is mutated, glycogenin is still capable of glucoslyation, just not of itself (231, 232). Whether glucosyl additions occur via intra- or intersubunit reactions has been debated, but recent studies suggest that the first four glucosyl units may be added via an intrasubunit reaction, and longer chains require intersubunit interaction (229). It has been suggested that although glycogenin functions as a dimer, at low enough concentrations the two subunits would dissociate after chain initiation, giving rise to two separate glycogen molecules (233). The presence of a single chain giving rise to an entire β-particle is supported by structural studies of glycogen (52). In most mammals, there is only one isoform of glycogenin that is widely expressed, but humans have two isoforms: glycogenin-1 (GYG1), expressed in all tissues, and glycogenin-2 (GYG2) which is primarily expressed only in liver, with minor expression in heart, pancreas, and adipose tissue (113, 234) (www.proteinatlas.org). There is some evidence that glycogenin is phosphorylated, but it has not been corroborated (227). Glycogenin can interact directly with glycogen synthase, and a crystal structure of the interaction provides evidence for their cooperation in the initiation of the glycogen granule (235, 236). Additionally, insufficiently glucosylated glycogenin does not serve as an efficient primer for glycogen synthase, and phosphorylase can reduce the glucosylation state of glycogenin, making it a less effective substrate for synthase (237, 238). Phosphorylase induces the dissociation of glycogen synthase from a proteoglycan fraction in hepatocytes, presumably containing glycogenin (219). Thus, priming through glycogenin may be regulated indirectly by the signals that govern glycogen synthase and phosphorylase activity, rather than by direct regulation of glycogenin. Glycogenin interacting proteins (GNIPs) have been suggested to stimulate the activity of glycogenin (239). Paradoxically, glycogenin mutations in humans and mice lead to glycogen accumulation in some tissues and muscle weakness, indicating that without glycogenin, glycogen synthesis can still occur, but in a dysregulated and pathogenic manner (240, 241).

1.4.2 Glycogen synthase and phosphorylase

Glycogen synthase [UDPglucose:glycogen α-4-glucosyltransferase, EC 2.4.1.11] is a member of the glycosyltransferase family 3 that catalyzes the addition of alpha-1,4-linked glucosyl units to a glycogen chain, using UDP-glucose as the donor and releasing UDP as product (www.cazy.com). Two isoforms of synthase exist in mammals: GYS1, encoding muscle glycogen synthase, highly expressed in all tissues including brain, and GYS2, restricted to liver (114) (www.proteinatlas.org). Muscle glycogen synthase has nine phosphorylation sites

20 and was one of the first examples of a hierarchically phosphorylated protein (93). Although liver glycogen synthase is also multiply phosphorylated, its activity appears to be regulated by only one phospho-site (242). Phosphorylation inhibits synthase activity, although activity can be fully restored in the presence of the potent allosteric activator glucose-6-phosphate. Numerous kinases are responsible for phosphorylating synthase, including protein kinase A, protein kinase C, AMP-activated protein kinase (AMPK), caseine kinase 2, and glycogen synthase kinase 3 (GSK3). Thus, control of glycogen metabolism is inextricably linked to intracellular energy status, glucose homeostasis, insulin signaling, and other external signals through diverse signaling cascades (93, 114, 243).. Glycogen phosphorylase [1,4-α-glucan:orthophosphate α-glycosyltransferase, EC 2.4.1.1] is a member of the glycosyltransferase family 35, and catalyzes the transfer of glucose moieties from the glycogen molecule to inorganic phosphate, releasing the product glucose-1-phospate (www.cazy.com). Glucose-1-phosphate is converted to glucose-6-phosphate by phosphoglucomutase, facilitating its entry into glycolysis. In the presence of excess glucose-1-phosphate, phosphorylase can catalyze the reverse reaction (31). There are three isoforms of phosphorylase corresponding to the tissues in which they are enriched, but not restricted: muscle (PYGM), liver (PYGL) and brain (PYGB) (www.proteinatlas.org). Only the brain isoform is expressed in fetal tissues, and all are expressed to some extent in adult brain (244). A single site on phosphorylase is phosphorylated by only one kinase: phosphorylase kinase (PhK), a multisubunit enzyme with muscle, liver and brain isoforms. Both phosphorylase and PhK are activated by phosphorylation and allosteric modulators, and an excellent review on their regulation in brain has been published recently (245). Crystal structures of all phosphorylase isoforms have been determined (246-248).

1.4.3 Glycogen branching and debranching enzymes

Like glycogenin, and unlike synthase and phosphorylase, glycogen branching enzyme (GBE) and debranching enzyme (GDE) do not appear to be regulated by phosphorylation or allosteric modulation. There is only one isoform of each, and both are highly expressed ubiquitously (www.proteinatlas.org). Due to their abundance, these enzymes are not considered to be rate-limiting under normal circumstances (114, 136, 249). However, they are clearly essential for maintaining proper glycogen structure, as deficiencies in either enzyme result in GSDs characterized by distinctly abnormal glycogen deposits with pathological consequences (165). Crystal structures of human GBE and yeast GDE have been determined and the effects of GSD-causing mutations defined, providing molecular level insights into how they affect glycogen structure (250, 251). GBE (α-1,4-glucan:α-1,4-glucan 6-glycosyltransferase, EC 2.4.1.18) belongs to the subfamily 8 of the GH13 family of glucosyl hydrolases. Like other members of this subfamily, it contains a carbohydrate binding module, CBM48 (252)

(www.cazy.com). Two reaction steps occur successively in its central catalytic core: hydrolysis and transglucosylation (250). In the first step, the enzyme cleaves an α-1,4 linkage on a glucan chain, forming a covalent enzyme-glycosyl intermediate; in the second step, an α-1,6 linkage is formed from the same chain or one nearby. The minimum length transferred is 6 or 7 glucosyl units (253, 254), and the average distance between chains is 3 or 4 glucose units (52, 255). This chain length requirement is supported by the crystal structure of GBE in complex with a heptasaccharide (250). Mammalian GBE can convert amylose or amylopectin to a glycogen-like polysaccharide. Krisman showed that incubation of amylopectin or amylose with GBE purified from rat liver led to a displacement of λmax to a value that coincided with that of glycogen (147). However, GBE did not change the λmax of rabbit liver glycogen, suggesting the molecule already had been maximally branched. Indeed, comparative studies on mammalian and plant branching enzymes indicate that the branching degree in polysaccharides is determined by intrinsic properties of the branching enzyme used to synthesize them (256, 257). In other words, the optimal structure of the mammalian glycogen molecule may be primarily attributed to the inherent "equilibrium" of the mammalian branching enzyme. Not surprisingly, synthetic glycogens produced in vitro by branching enzymes from different organisms have similar, but not identical, structural properties to natural glycogens (258-260). Perhaps these observations explain the delay in glycogen remodeling reported by the Roach lab after exercise (160): glycogen super-compensation, characterized by high glycogen synthase activity, results in longer chains because branching enzyme has not yet introduced the appropriate branch points. Only over time does branching enzyme act so that glycogen structure returns to an equilibrium, with CLD and λmax at basal levels. Indeed, it has been suggested that a chronic imbalance of glycogen synthase and GBE activities leads to abnormal polysaccharide formation in pathological conditions (209, 261, 262). Also, the ubiquitous expression of a single GBE isoform may help to explain why brain glycogen has a similar CLD to muscle glycogen (159). Presumably all healthy tissues in a single species would reach the same CLD at "equilibrium." GDE is required for cytosolic glycogen degradation since phosphorylase terminates four glucosyl residues away from a branch point. Mammalian GDE has two catalytic activities: 4-α-glucanotransferase activity (EC 2.4.1.25) involves transfer of a chain of three glucosyl units from the branch to a nearby nonreducing end, leaving a single α-1,6 glucosyl unit. Its amylo-α-1,6-glucosidase activity (EC 3.2.1.33) follows, and the α-1,6 linkage is hydrolyzed to release free glucose (103). The release of free glucose instead of glucose-1-phosphate makes it very convenient to compare the levels of α-1,6 and α-1,4 linkages in polysaccharides. GDE is considered an indirect debranching enzyme; direct debranching enzymes such as isoamylase exclusively cleave α-1,6-linkages and release intact oligosaccharide chains. Direct debranching enzymes occur in plants, bacteria and yeast but not in mammals (103). These enzymes are very useful in polysaccharide

22 analysis, especially in determining CLD. GDE catalytic mutants display a loss of either glucosidase or transferase activity, but not both, strongly suggesting disparate catalytic sites (263). This was confirmed when the crystal structure of yeast GDE was determined, revealing an elongated structure with catalytic domains at either end of the molecule (251). Structures of both the free wild-type enzyme and a transferase-deficient mutant in complex with maltopentaose were reported. Maltopentaose molecules were bound at both catalytic domains and other sites, illustrating the specificity of GDE transferase activity for a glucan of five units or less and the necessity of non- catalytic binding sites for glycogen association. The data also suggested that the product of the transferase reaction, a chain containing a single α-1,6-linked glucosyl unit, completely dissociates from GDE prior to recruitment to the glucosidase catalytic site. Multiple binding sites are a recurring theme among glycogen- and starch-associated proteins (187, 190, 250, 251, 264). In observing that oligosaccharides were bound both to catalytic and non-catalytic sites in the crystal structure of GBE, Froese et al. stated: "[Non-catalytic binding sites] may provide GBEs the capability to anchor a complex glycogen granule and, as proposed previously, determine the chain length specificity for the branching reaction as a ‘molecular ruler’. This agrees with the emerging concept of glycogen serving not only as the substrate and product of its metabolism but also as a scaffold for all acting enzymes" (2015) (see Section 1.7).

1.5 Glycophagy

1.5.1 Lessons from Pompe disease

Phosphorolysis is not the sole means of glycogen degradation in cells. This became evident through studies of Pompe disease, a lysosomal storage disorder characterized by the buildup of vacuole-bound glycogen in virtually all tissues (97, 98). The deficient enzyme in Pompe disease is lysosomal acid α-glucosidase (called GAA or acid maltase) [EC 3.2.1.3] which hydrolyzes both α-1,4 and α-1,6 linkages, optimally at low pH (265). It is a widely expressed enzyme, with a promoter characteristic of a housekeeping gene (266). GAA is synthesized as a precursor protein in the endoplasmic reticulum and undergoes extensive processing before it achieves optimal catalytic activity and reaches its lysosomal destination (265, 267). It can fully hydrolyze glycogen as well as short oligosaccharides, maltose and other polyglucans to glucose (97, 268, 269). GAA is exo-acting, cleaving single glucosyl units successively from the non-reducing ends of glycogen. However, rather than releasing glucose-6-phosphate like phosphorylase, GAA liberates free glucose, which can translocate into the cytosol via glucose transporters in the lysosomal membrane (270). How glycogen gets into the lysosome is not well defined, but recent work demonstrates that it involves the autophagic machinery. Autophagy, or 'self-

23 eating,' refers to multiple cellular pathways converging on the lysosomal breakdown of cellular components such as long-lived proteins and organelles (reviewed by (271-273). Autophagy is critical both for general cellular maintenance and for nutrient release in times of energy stress. It is especially critical for turning over worn or damaged cellular components in neurons and other post-mitotic cells (271, 272, 274). The best characterized autophagic pathway is macroautophagy, activated in response to starvation; 'autophagy' and 'macroautophagy' are frequently used interchangeably. Macroautophagy is a highly conserved pathway in eukaryotes involving a family of autophagy-related (Atg) proteins and the formation of a structure with a characteristic double membrane known as the autophagosome, which engulfs substrates and fuses with the lysosome. Microautophagy and chaperone-mediated autophagy are alternative pathways in which cellular components are targeted directly to lysosomes in the absence of autophagosome formation. In recent years, numerous terms for the selective degradation of specific substrates (via any of the autophagic pathways) have been introduced; e.g. mitophagy (of mitochondria), pexophagy (of perioxisomes), nucleophagy (of nuclear components), reticulophagy (of ER), and xenophagy (of pathogens) (271, 275). In 2011, the Roach group introduced the term 'glycophagy' to refer to the selective autophagic degradation of glycogen (276). Ironically, the term 'glycogenosome' had already been used by some investigators to refer to the glycogen-laden lysosomes found in newborn rat liver, aged neural tissue, and glycogen storage diseases (277-279), but the term did not gain traction. In addition to glycogenosomes, which have a single membrane characteristic of a lysosome, Pompe (GAA-/-) tissues also display accumulated autophagosomes filled with ubiquitinated proteins and cellular debris. In order to study the effect of autophagy on lysosomal glycogen, Raben and colleagues generated a muscle-specific knockout of a critical autophagosome gene, Atg5, on the GAA-/- background. These mice had just as many glycogenosomes as the GAA-/- mice, but they had a reduced number of autophagosomes and no improvement in pathology. Ubiquitinated proteins were increased, but they were not surrounded by the double membrane as in the GAA-/- mice (280, 281). The study suggested that the autophagy defect was caused by inefficient fusion of the autophagosomes with the glycogen-laden lysosomes, which formed by an independent pathway. In a later study, the group selectively inactivated a closely related gene, Atg7, in fast-twitch GAA-/- muscle (282). No autophagosomes were formed, glycogenosomes were still present, and there was no change in pathology; however, the glycogen level was reduced. These two studies demonstrate that macroautophagy contributes to some, but not all, of the delivery of glycogen to the lysosome (282). The group also illustrated that the muscle pathology of Pompe disease resulted largely from defective autophagic flux. The observation that Atg7, but not Atg5, abrogated glycogen accumulation may be explained by the fact that there are two primary conjugation systems in autophagy, and while Atg5 is only involved in one, Atg7 is required for both (283).

In 2002, Janeček identified a putative CBM20 in a protein of unknown function called starch binding domain-containing protein 1 (Stbd1, also called genethonin 1) (284). The only other proteins in mammals with a CBM20 are laforin and a glycosyltransferase (glycerophosphocholine phosphodiesterase 1, GPCPD1). It was known that the laforin CBM facilitates laforin binding to glycogen (285). The Roach group later showed that Stbd1 is highly expressed in glycogen-rich tissues such as muscle, liver, and heart, with trace amounts in brain, kidney and pancreas, and that Stbd1 levels mirrored glycogen levels in fed, fasted and transgenic mice (286). Stbd1 binds glycogen in vitro and co-localizes with the lysosomal- associated membrane protein LAMP1 and the Atg8 homolog GABARAPL1 in cell culture, but not with the microtubule-associated light chain 3 (LC3), a marker for auotophagosomes (276, 286). Demetriadou et al. later showed that when overexpressed in HeLa cells, Stbd1 was N-myristoylated, and this modification facilitates its recruitment with glycogen to subcellular domains associated with autophagy (287). The Sun group showed that Stbd1 is elevated in muscle tissue of GAA-/- mice, but not in liver or heart, but shRNA-mediated knockdown of Stbd1 did not alter glycogen levels in any tissues (288). A full knockout of Stbd1 in the Pompe mice (Stbd1-/- GAA-/-) did not change lysosomal accumulation of glycogen in muscle or heart tissue, but reduced liver glycogen by 73% (289). Glycogen accumulation could be restored by exogenous expression of human Stbd1, but not a mutant lacking the CBM20. Their study demonstrated that Stbd1 may significantly participate in glycophagy in the liver but not in the heart or skeletal muscle, and that the CBM20 is essential for its function.

1.5.2 Lessons from Lafora disease

Another link between glycogen and autophagy emerged through studies of LD. Laforin- and malin-deficient mouse models were generated and analyzed by six different groups. Most observed autophagy defects in the brain of these mice (179, 182, 189, 290). Some groups observed reduced conversion of LC3 from its cyotosolic form (LC3-I) to lipidated form (LC-II), which is associated with its recruitment to autophagosomal membranes (179, 189). Both mouse models also displayed an increase in p62 (179, 182, 290), a protein that normally binds to both ubiquitinated cargo and LC3, is incorporated into the autophagosome, and degraded (271). Reduced LC3-II and increased p62 levels strongly suggested defects in autophagic flux. Similar defects were also observed in liver of laforin- deficient mice, but not in muscle (160, 291). In contrast, the Minassian group did not observe autophagy defects in muscle or brain in either mouse model (159, 292). The Roach, Minassian and Guinovart groups generated LD mouse models with genetically reduced glycogen synthesis and found that LB formation was reduced and neuropathology was abrogated (182, 293-295). The Guinovart group elegantly demonstrated that the autophagy impairment was secondary to

25 polyglucosan accumulation by crossing malin-deficient mice with mice specifically lacking muscle glycogen synthase in the brain, which eliminated cerebral glycogen synthesis, LB formation, autophagy defects, gliosis, and susceptibility to epilepsy (182). These studies indicate that polyglucosan accumulation drives the autophagy impairment and neurological phenotype in LD. Although some groups have argued that laforin and malin regulate autophagy independent of glycogen, this is unlikely to be the case in vivo (167). However, as was demonstrated with the Pompe mice, insufficient clearance of glycogen via glycophagy may indeed provoke pathological damage by obstructing overall autophagic flux. This is a very prominent theme among lysosome storage disorders, which frequently have severe neurological phenotypes; this disease class is reviewed in depth elsewhere (296, 297). Corroborating this, recent studies from the Sanz group showed that in LD, mitophagy is impaired due to a global autophagic defect, and endocytic recycling of the astrocytic glutamate transporter GLT1 is altered (298, 299), likely also resulting from impaired autophagy. The possibility that laforin and malin are involved in glycophagy builds on work from many labs showing that malin, typically in concert with laforin, ubiquitinates glycogen-associated substrates. Malin is an E3 ubiquitin ligase that was initially demonstrated to bind and ubiquitinate laforin in cell culture (175, 176). Malin also promotes the ubiquitination of multiple proteins that bind to glycogen molecules: glycogen synthase, protein targeting to glycogen (PTG), R6, AMPK and GDE (see Section 1.7). In most of these studies, ubiquitination was shown to require laforin, which is believed to act as a scaffold for malin interaction with its substrate (300- 304). Although ubiquitination led to the degradation of these substrates in cell culture, PTG, R6 and AMPK levels were unchanged in LD mouse models; in contrast, glycogen synthase accumulated in the insoluble glycogen fraction, as did laforin in malin-deficient mice (178, 180, 181, 305) (Gentry, Sun, Dukhande, unpublished results). Importantly, whenever the ubiquitin linkages incorporated by malin are defined, they are K63-linked (302, 303, 306, 307). K63-linked ubiquitination is a substrate for p62 binding and specifically linked to selective autophagy (308, 309). The presence of a laforin-malin ubiquitination complex has been highly debated (167, 188, 310). Interestingly, the Roach group reported that there was no change in the levels of GABARAPL1 (a member of the Atg8 family and a homolog of LC3; see (311)) and LAMP1 in laforin-deficient fibroblasts, but these proteins were decreased in malin-deficient fibroblasts and in fibroblasts with both laforin and malin knocked out (312). These results suggest laforin is upstream of malin in the path from glycogen to the lysosome. The Sanz group showed that laforin, malin and p62 form a complex, and that malin mediates the interaction between laforin and p62, likely through ubiquitination (307). Additionally, malin and p62 do not co-localize in cells without the expression of laforin. It is well known that laforin binds to and co-localizes with glycogen in vitro, in cell culture, and in vivo. A model to explain these studies that is compatible with available data posits:

26 laforin binds to glycogen and recruits malin; the laforin-malin complex promotes K63-linked ubiquitination of various glycogen-bound proteins; p62 binds to K63- linked ubiquitin moieties, recruiting LC3 (or its homolog GABARAPL1) and the autophagosome machinery, which engulfs the glycogen particle and fuses with the lysosome. Since glycogen synthase and laforin expression parallel glycogen levels and these proteins accumulate with glycogen in LD mouse models, they may also be turned over inside the lysosome. The fact that a catalytically inactive form of laforin rescues laforin-deficient mouse models supports this scaffolding role for laforin (159, 189), but does not exclude the possibility that glycogen dephosphorylation by laforin is still relevant, as during cytosolic glycogenolysis (160). Also, phosphorylation of laforin by AMPK has been shown to enhance the interaction of laforin with malin (313); this regulation event could be one way to stimulate glycophagy in times of energy stress, when AMPK becomes activated. Finally, it is well established that laforin preferentially binds to polysaccharides with long chains and localizes to LBs in malin-deficient mice (179, 190, 314). Thus, laforin would selectively recruit the glycophagy machinery to precipitation-prone glycogen molecules with precariously long chains, as has been proposed (188). It is worth noting that delivery of glycogen to lysosomes may involve multiple pathways, possibly with cell-specific relevance. Perhaps Stbd1 and laforin, each containing a CBM20, function as distinct receptors for glycophagy. It is interesting that although LBs occupy most tissues in LD, they are enriched in certain cell types within these tissues. For example, in LD mice, 97% of LBs in skeletal muscle are found in fast-twitch type IIb muscle fibers and 2% in slow-twitch type I fibers (315). In human skin biopsy, LBs are only found in duct cells, but not secretory cells (316). In human liver, LBs are enriched in the periportal regions, but not the perivenous regions (317). All of these cell types appear to make glycogen (318- 320), but the differences in whether or not they make LBs suggest the glycogen is handled differently. Cell-specific glycogen metabolism and polyglucosan body formation are of particular interest in the brain. LBs were first reported in neurons (321), and historically were believed to occupy this cell type almost exclusively (322). However, this observation has been perplexing since the vast majority of brain glycogen is stored in astrocytes. Recent work has demonstrated that neurons possess an active glycogen metabolism (300, 323), and LD mouse models accumulate both LBs in neurons and a distinct polyglucosan body known as corpora amylacea in astrocytes (324, 325). Early literature on human patients corroborate these findings. Corpora amylacea in addition to LBs have been reported in multiple studies (326). Small inclusions, which may also correspond to corpora amylacea, have been identified in astrocytes and oligodendrocytes (327). Recently, it has been asserted that the striking appearance of neuronal LBs left glial LBs largely overlooked over the years; upon careful review of the literature, small LBs that were assumed to be in neuronal processes are likely to actually be astrocytic (328).

1.5.3 The benefits of glycophagy

The lysosomal glycogen pool plays an important role in circumstances when a burst of glucose is needed, such as during the neonatal starvation period (329- 331). Glycogen-rich autophagosomes have been observed in the neonatal liver, muscle and heart, but they are depleted within hours after birth. A unique role for glycophagy in the heart is also emerging (332, 333). While liver and muscle normally deplete their glycogen stores with fasting, cardiac tissues accumulate glycogen and upregulate glycogen autophagy under nutrient stress (333). An advantage of lysosomal glycogen is that it can be mobilized independently or in parallel with cytosolic glycogen by a distinct set of cellular stimuli. For example, glycophagy could be specifically regulated by lysosomal uptake of Ca2+, which enhances GAA activity (330, 332). Very little is known about the role of glycophagy in brain tissue, besides the report of glycogenosomes in the aging brain (277). It has been suggested that glycogen of large molecular weight (i.e. α particles) is specifically degraded in lysosomes, and proceeds by a random, rather than an ordered degradation mechanism (334, 335). Based on glycogen degradation patterns it has been speculated that brain glycogen is degraded primarily by phosphorolysis and not by lysosomal hydrolysis (336).

1.6 Glycogen ultrastructure and subcellular distribution

In over a century of glycogen-related research many researchers have debated the existence of multiple forms of glycogen that differ in size, structure, associated proteins, and metabolic activity. The distinction between glycogen α and β particles came into focus with studies using the electron microscope and ultracentrifugation, and some progress has been made on defining the nature of these particles. Acid-soluble and acid-insoluble fractions of glycogen have also been observed, but the physiological relevance of these distinct pools has been highly debated. These topics, as well as a possible role for glycogen metabolic enzymes in the nucleus, are discussed in the next sections.

1.6.1 Glycogen α, β, and γ particles

Drochmans and colleagues were the first to describe the various levels of glycogen structure in native glycogen isolated from liver and muscle, and their observations were corroborated by others (72, 74, 75, 337). The largest glycogen particle, named the α particle, is a rosette-shaped association of several smaller units, known as β particles. An α particle is usually more than 108 Daltons in molecular weight and up to 300 nm in diameter (137, 338). Glycogen β particles have a molecular weight of 106-107 Daltons and typically range in size from 10-44 nm (339, 340). It is generally accepted that a β particle corresponds to a single molecular unit of glycogen that can reach up to 12 tiers as modeled by Whelan and others (Figure 1.6c, see Section 1.3.1). Drochmans designated the 3 nm

28 filaments making up the β particles as γ particles (74). However, 'γ particle' has also been used to refer to protein-rich subunits on the glycogen β particles (see Section 1.7) (137). Generally speaking, glycogen is primarily stored as α particles (or a mixture of α and β particles) in the liver and as β particles in muscle, brain, and other tissues (reviewed below). Recent work has demonstrated a mixture of α and β particles in cardiac tissue (341). It is apparent that these norms become disrupted in pathological conditions. It is estimated that liver α particles contain, on average, 20 to 40 β particle subunits (342). Liver glycogen is typically cytoplasmic and often reported to closely associate with smooth endoplasmic reticulum (343, 344). In muscle, β particles are associated with sarcoplasmic reticulum, concentrated either in subsarcolemmal regions, between myofibrils, or within the myofibrils; these pools respond differently to metabolic cues (239). Aggregates up to 60 nm in diameter, apparently composed of 4-6 β particles, were also observed via EM by Wanson and Drochmans in muscle, and others have reported α particles in rat muscle and insect flight muscle (52, 75). In skin and adipose tissue, β particles appear as cytoplasmic granules interspersed between ribonucleoprotein particles (345, 346). In adipose tissue, they are only prominent after refeeding following a fast (347). In leukocytes and thymus, β particles are scattered throughout the cytoplasm and are fairly homogenous in size (348, 349). In the retina, cytoplasmic β particles and occasional small α particles (60 nm) have been observed in the cone, but not rod, photoreceptor cells, and in Müller cells (350). In the brain, glycogen exists primary as cytoplasmic β particles in the cell bodies and processes of astrocytes (76, 77, 351, 352). Normally, they are sparse to nonexistent in neurons, microglia, or oligodendrocytes, but β particles have appeared in neuronal axons and dendrites following fasting or trauma (77, 351). Astroglial β particles become apparent in early embryos and increase in both abundance and diameter with development (353). Small α particles (80-100 nm in diameter, composed of 4-8 β particle subunits by EM) have also been reported in rat astrocytes (352). Biochemical studies of glycogen isolated via a mild cold-water extraction method show significant heterogeneity in molecular weight: a fraction of particles had molecular weights suggestive of small α-particles composed of 3- 6 β particle subunits (336). In several cases of human glycogenoses, small (60- 120 nm in diameter) and large (150-350 nm in diameter) α particles have been reported in astrocytes or neurons (354-356). Aggregates of α or β particles have also been reported in Alzheimer's disease and the aging brain (357, 358), which are distinct from polyglucosan bodies (79). Reactive astrocytes are known to accumulate glycogen β particles in response to various pathological stimuli, including: brain infarction (359), methionine sulfoxide (351, 360), X-ray irradiation (76), and barbiturates (351). Occasionally, intramitochondrial glycogen has also been observed in rat retinal cells (361), in the human and canine myocardium (362, 363), and in astrocytes (352, 364). β particle size exists around a normal distribution, with most granules being 20-

30 nm (Figure 1.9a) (52, 365). Since each additional tier is estimated to contribute 3.8 nm to the diameter (134, 365), the observed size distribution implies that most β particles have only about 7-9 tiers (Figure 1.9a,b,c). It is still not clear why most β particles do not reach the maximum diameter, since such large granules can store a significantly greater amount of glucose (239). A growing body of evidence suggests that glycogenin regulates β particle size. In cell culture and in tissue, glycogenin does not appear to exist in a free form; it is always attached to glycogen (226, 241, 366). Overexpression of glycogenin in cell culture did not change total glycogen levels, but it increased the total number of glycogen molecules, which were of a smaller size (366). The Guinovart group showed that mice lacking glycogenin display reduced exercise endurance and an accumulation of glycogen particles in muscle that were larger than those in wild-type mice (241). Also, crystal structures and data from the Sicheri lab suggest glycogen particle size is regulated by the length of the linker between the glycogenin catalytic domain and the region that binds to glycogen synthase; this region is alternatively spliced and varies in length across species (236, 367). In healthy organisms, α particles seem to be a unique feature of the liver. It is well established that total liver glycogen depletes with fasting and exceeds basal levels upon refeeding, a phenomenon known as glycogen super-compensation that is also observed in muscle and brain (368-370). Additionally, hepatic glycogen levels and the activities of glycogen metabolizing enzymes undergo a diurnal rhythm (371, 372). The Gilbert lab analyzed glycogen particles by size exclusion chromatography and EM and proposed a model for diurnal glycogen cycling in liver based on their data and earlier work (373). Starved livers maintain low glycogen in the form of small β particles. Upon refeeding, glycogen is quickly synthesized on these preexisting β particles, which provide optimal surface-to- volume ratio for rapid resynthesis. Once glycogen levels peak, the β particles are assembled into α particles. The α particles persist as glycogen levels begin to gradually fall, but they are gradually disassembled into β particles, which are preferentially degraded. Thus it appears that the primarily role of α particles is to facilitate a slower release of glucose than β particles (374). This indeed seems optimal for the organ whose primary role is to maintain blood glucose homeostasis. Additionally, α particles in diabetic livers are more fragile than those of healthy mice, suggesting they could more easily be broken into β particles and contribute to hyperglycemia (375, 376). The nature of α particle assembly has remained elusive for decades. Drochmans observed that purified α particles could be dissociated into β particles under acidic conditions (74). Lazarow, Orrell and Bueding previously showed using ultracentrifugation that acid, alkali and heat drastically changed the sedimentation of liver glycogen, suggesting these treatments disrupt α particle structure (51, 377). The Gilbert lab has devoted a significant body of work to elucidating the nature of the α particle bond due to its relevance to diabetes (340, 376, 378-380). The lack of significant dissociation of normal α particles in dimethyl

30 sulfoxide, 2-mercaptoethanol, or sodium dodecyl sulfate suggested the association is not through hydrogen or disulfide bonds or weak protein-protein interactions (340, 378). It was reported that unexpectedly high levels of glycogenin were found in a proteomic study of liver glycogen, suggesting that glycogenin was present on the granule surface in addition to the granule core (381). The Gilbert lab then used proteomics to identify glycogenin as the molecular "glue" holding α particles together (379) (Figure 1.9d).

1.6.2 Proglycogen and macroglycogen

There is a lot of controversy surrounding proglycogen and macroglycogen, which have gone by different terms over the years and have been discussed in great detail by others (226, 339, 382, 383). Although the precise nature of these pools is debated, there does appear to be sufficient evidence for their distinct physiological roles in glycogen turnover (239, 339). The relationship between proglycogen/macroglycogen and α/β particles is not completely clear as these distinctions are discussed separately. However, since they are relevant to tissues generally lacking α particles, proglycogen and macroglycogen likely represent two states of β particle synthesis (226). When the TCA extraction method was first introduced for the purification of glycogen, it was observed that a portion of the glycogen was resistant to extraction by cold TCA (50). It could not be released unless the tissue was treated with heat, alkali, or protease, so it was concluded that this resistant fraction was attached to protein, and the two fractions were called lyo- ("free") and desmo- ("fixed") glycogen (50, 384). Such observations also gave rise to the terms acid-soluble, i.e. extractable, glycogen, and acid- insoluble, i.e. residual, glycogen (385, 386). In a 1960 literature review, Stetten and Stetten noted that while the quantity of free glycogen was widely variable with nutritional state, the quantity of fixed glycogen remained constant. However, experiments suggested the latter was more metabolically active, since injected [14C]glucose was more readily incorporated into fixed glycogen (386). For decades, researchers continued to debate whether such observations were artefactual, merely reflecting the association of known metabolic enzymes with glycogen (387). Whelan revisited the subject in the late 20th century while investigating the origin of glycogen synthesis. Using cultured primary astrocytes, his group showed that fixed glycogen, which was re-named proglycogen, contained 10% protein by weight and was rapidly resynthesized upon refeeding after glucose starvation, reaching a size of 400,000 Da. The proglycogen was later converted to macroglycogen, as demonstrated by pulse-chase labeling (383). The group proposed that proglycogen was an intermediate between glycogenin and the mature glycogen granule (226, 383). Resynthesis of proglycogen prior to macroglycogen following exhaustive exercise was also confirmed in vivo in humans {Adamo, 1998 #1640}. It has been suggested that the upper limit of proglycogen size corresponds to

31 a diameter of ~30 nm, a molecule of approximately 8 tiers; anything larger would be considered macroglycogen (Figure 1.9a,d) (136, 339, 365). However, Whelan's group estimated that the upper limit of proglycogen was 400 kDa (226, 383), which corresponds to a glycogen molecule with ~2500 glucose residues, i.e. approximately 5 tiers. Whelan's group also suggested that proglycogen and macroglycogen are synthesized by different enzymes, however it may be that the different activities associated with the pro- and macroglycogen fractions both correspond to glycogen synthase. Two independent groups have demonstrated that glycogen synthase in the proglycogen fraction has a higher affinity for UDP- glucose than synthase in the macroglycogen fraction {Curtino, 2000 #1461}{Tavridou, 2003 #1330}. Curtino and Lacoste proposed that the insolubility of proglycogen in TCA was due to its heavy association with glycogen synthase, not with glycogenin, as Whelan had proposed (388). Tavridou and Agnius showed that the hormone insulin increased the association of glycogen synthase with the proglycogen fraction, and glucagon had the opposite effect (219). Insulin is known to stimulate dephosphorylation (and activation) of glycogen synthase, and the effects of both hormones are well known to regulate glycogen synthesis through signaling cascades, so the difference in synthase activity may be due to its phosphorylation state. Indeed, phosphorylation-dependent stimulation of glycogen synthase and phosphorylase has been shown to induce translocation of these proteins to actin-rich structures proximal to sarcoplasmic reticulum in skeletal muscle (389, 390). In liver, newly synthesized glycogen has also been reported to be associated with a protein backbone (391) and smooth endoplasmic reticulum (343). Additionally, glycogenin has been shown to bind to actin (392). In vitro studies suggest that glycogen synthase is more active when it is bound to glycogenin (109, 393). When glycogen content decreases, glycogen synthase has been shown to translocate from a glycogen-enriched membrane fraction to the cytoskeleton, concurrent with a decrease in activity (394). In isolated hepatocytes, glucose stimulated the cytoskeleton-dependent translocation of synthase from the cytosol to the plasma membrane, and newly synthesized glycogen molecules always appeared first at the cell membrane, but moved inward as they were replaced by newer molecules (395-397). Proglycogen may represent a membrane-associated and/or actin-bound form of glycogen that is primed and ready for active anabolism/catabolism. When it grows to a certain size, it becomes macroglycogen, synthesis levels off, and it is released into the cytosol. A similar model has been proposed recently (137) and prompts further discussion of glycogen and its associated proteins as a dynamic subcellular entity rather than a static molecule (see Section 1.7) (382). It is still not clear whether synthesis and degradation can occur simultaneously on the same granule, but proximity of these activities is suggested (389). The spatiotemporal regulation of glycogen metabolism is an interesting area of research requiring further elucidation.

1.6.3 Nuclear glycogen

Dozens of microscopy studies have reported intranuclear glycogen deposits, most often in pathological conditions of the liver (398-400). In a 1975 case report, Ferrans et al. beautifully reviewed 24 prior studies on intranuclear glycogen either in the context of diabetes, glycogen storage diseases, or as an incidental feature in healthy humans and animals (401). In these reports, nuclear glycogen was usually in the form of β particles, sometimes dispersed, other times clustered; occasionally a few α particles were reported. The findings of Ferrans et al. were unusual in that nuclear glycogen was observed in the diseased myocardium. Intranuclear glycogen was only found in 6 out of 90 patients with cardiac diseases, and only in a few nuclei per patient. Most of the glycogen was present as β particles, often aggregated and filling the entire nucleoplasm. The presence of nuclear glycogen was not associated with any cellular damage. The authors proposed three mechanisms for the formation of nuclear glycogen, but shrewdly concluded that only one was possible: (1) there was no ultrastructural evidence supporting phagocytosis or endocytosis of glycogen particles by the nuclear membrane; (2) glycogen particles were too large to pass through the nuclear pore complex; (3) the glycogen must be synthesized in situ, which had been previously demonstrated in normal Müller cells of the rat retina (402) and in Novikoff hepatoma cells (403). Ferrans et al. proposed that conditions favoring increased permeability of the nuclear membrane facilitated the entry of the metabolic enzymes into the nucleus, but that they became trapped when permeability returned to normal. This implies that atypical trapping of glycogenic enzymes into the nucleus may result in excessive levels of glycogen. Nuclear α and β particles have also been reported in the context of various cancers: human gastric adenocarcinoma (404), sublines of the Ehrlich-Lettré mouse ascites tumor (405), chicken sarcoma (406) and human and mouse hepatoma (407, 408). Nuclear glycogen synthase activity was also detected in the Ehrlich-Lettré subline (409, 410). In the brain, intranuclear glycogen deposits have been observed in the abnormal astrocytes of Alzheimer's disease (411, 412) and in the pituicytes of aged rats (413). Nuclear polyglucosan was also reported in a neuronal cell culture model of GBE deficiency (262). In more recent years, Guinovart, Gentry and others have shown that nearly all the central glycogen metabolic enzymes can be found in the nucleus: glycogenin (392, 414), glycogen synthase (300, 415, 416) glycogen phosphorylase (Sun, Dukhande, and Gentry, unpublished results), and debranching enzyme (304). Guinovart's group published a series of studies demonstrating the translocation of the muscle isoform of glycogen synthase from the cytosol to the nucleus in conditions of glucose deprivation, which required glycogen binding (415-417). They also showed that neurons display primarily nuclear localization of glycogen synthase, characteristic of cells lacking glycogen, in contrast to glycogen-rich astrocytes with cytosolic synthase (300). Additionally, GDE and glucokinase, a key

33 glucose sensing enzyme, also shuttle to the nucleus upon glucose deprivation (304, 418). In contrast, glucose starvation has an opposite effect on the gluconeogenic enzyme fructose bisphosphatase, inducing its translocation to the cytosol (419). It has been suggested that the nuclear sequestration of these enzymes is yet another way to regulate their activity, preventing them from acting inappropriately in times of energy stress (420). Laforin and AMPK, which both contain CBMs and bind to glycogen in vitro and in cell culture, also translocate to the nucleus with glucose deprivation and glycogen depletion (304, 421) (Figure 1.10a,b). Nuclear-cytoplasmic shuttling has also been observed for yeast glycogen synthase. In a review discussing these data, it was postulated that "the freeing of glycogen synthase from its cytoplasmic tether to the glycogen particle as these stores reduced would therefore be a signal that carbon and energy reserves were low. The uptake of glycogen synthase into the nucleus might therefore represent a form of molecular 'fuel gauge.' It is possible that glycogen synthase could regulate transcription in response to energy availability by some as yet undetermined means" (422). 'Moonlighting' of metabolic enzymes in the nucleus is not a new theme. Moonlighting proteins perform multiple unrelated functions, expanding the functional repertoire of the cell without expanding the number of genes, and the first to be described was the lens structural protein ε-crystallin, also known as lactate dehydrogenase (423). All essential glycolytic enzymes have been observed in the nucleus, as well as some mitochondrial proteins and enzymes involved in methylation and acetylation (424, 425). These energy- sensing proteins have been shown to regulate transcription by diverse means in a "metabolism-epigenetic axis" that is critical for cellular homeostasis (425). Although the epigenetic role of glycolytic proteins is well described, it is still not clear whether glycolysis occurs in the nucleus, although its products ATP and NADH could certainly be useful. Glycogen-associated enzymes may also have both an epigenetic and energy producing role, and perhaps excessive glycogen deposits result from misregulated nuclear sequestration. AMP-activated protein kinase (AMPK), a master regulator of cellular energy known to bind to and sense glycogen levels (426) (see Section 1.7), also translocates to the nucleus and regulates transcription and the circadian clock through phosphorylation (427, 428).

1.7 The glycogen granule and its associated proteins

In 1968, Scott and Still recognized the dynamic nature of glycogen and introduced the term "glycosome" to refer to the molecule and its associated proteins, stating that the nature of glycogen as a cellular organelle had not been appreciated (348). Rybicka reintroduced the term in a detailed historical review of the evidence for such a classification (382). Unfortunately, the term "glycosome" has been used far more frequently to refer to the glycolytic enzyme-containing peroxisomes of trypanosomatids, a family of unicellular parasites (429), which

34 creates confusion. Also, in the strict definition, organelles are membrane-bound entities (430, 431), and glycogen is primarily cytosolic. Perhaps it is safest to simply recognize that a host of proteins are associated with the glycogen granule, its structure is complex and highly regulated, and it is a dynamic participant in cellular metabolism. The γ particle named by Drochmans was a description of the ramifying fine fibers (3 nm in diameter) comprising β particle structure, likely corresponding to glucan chains (74). Later authors came to identify protein-rich subdomains on the glycogen molecule as γ particles, which may be a misinterpretation (137). Protein is more electron dense than carbohydrate, so with lead and uranyl acetate staining, protein-rich regions appear darkly stained clusters about 2-3 nm in diameter; alternatively, staining with the histochemical method of Thiéry using silver proteinate results in electron-dense clusters of the same size corresponding to vicinal glycols (137, 382, 432). Rybicka refers to the electrodense vicinal glycols as γ particles, which is likely consistent with the meaning originally intended by Drochmans. Due to their very small size, both γ particles and protein-rich subdomains are only visible by EM. It is not yet clear why they exist as discrete, regular entities and which proteins correspond to the protein subdomains. However, it is well established that all of the primary glycogen metabolic enzymes (glycogenin, synthase, phosphorylase, GDE and GBE) bind to and colocalize with glycogen granules in cell culture and in vivo (251, 300, 304, 433, 434) (Figure 1.10c). In fact, due to its very tight association, glycogen synthase is frequently used as a marker for glycogen and polyglucosan bodies (179, 324). There are a few antibodies that directly bind to the polysaccharide chains of glycogen, but they are not yet commercially available and are still being characterized (120, 435). Other proteins known to closely associate with the glycogen molecule all contain a glycogen-binding CBM: laforin and Stbd1 (Section 1.5), the β-subunit of AMPK, and the glycogen-targeting subunits of protein phosphatase 1 (discussed below) (Figure 1.10b,c). Phosphorylase kinase also binds to glycogen (245, 436). Protein phosphatase 1 (PP1) catalyzes a vast array of dephosphorylation events in cells, regulation of which is achieved by diverse targeting subunits that bring the catalytic subunit (PP1c) into the vicinity of its target (437, 438). Seven glycogen-targeting subunits of PP1 have been identified (PPP1R3A-G), but four have been most extensively characterized as primary regulators of glycogen synthesis (114, 439). GM (also known as RGL, encoded by PPP1R3A) expression is restricted to heart and skeletal muscle; GL (encoded by PPP1R3B) is primarily expressed in liver; and Protein Targeting to Glycogen (PTG, also known as R5, encoded by PPP1R3C) and R6 (encoded by PPP1R3D) are more ubiquitously expressed, with enrichment in glycogen-rich tissues (www.proteinatlase.org) (114). All these subunits possess both a PP1-binding domain and a putative CBM21 and recruit PP1 to glycogen granules (www.cazy.com) (439, 440). Dephosphorylation by PP1 has opposing effects on synthase (activating) and phosphorylase (inhibitory), leading to net glycogen synthesis (300, 439, 441).

Overexpression of PTG, GM, GL, and R6 leads to glycogen accumulation in cell culture and mouse models (182, 301, 441-444). AMPK, a cellular energy sensor with multiple intracellular targets, is also considered an important glycogen sensor (445). AMPK is activated by elevated levels of AMP relative to ATP reflecting energy deficits. It is a heterotrimeric complex composed of a catalytic α subunit and regulatory β and γ subunits; the β subunit contains a CBM48 and targets AMPK to glycogen granules (252, 426, 446). AMPK phosphorylates both isoforms of glycogen synthase, inhibiting its activity (447). It has been speculated that recruitment of activated AMPK to glycogen granules would lead to decreased glycogen synthesis. There is also significant evidence for regulation of AMPK activity by glycogen via the β subunit (448). AMPK is particularly inhibited by the presence of a single α-1,6 linked glucose unit or by a glycogen molecule that has been partially degraded by phosphorylase (426, 449). A single glucose unit at a branch point could be exposed during debranching, since GDE is likely to dissociate from glycogen between its transferase and glucosidase activities (251). Thus, glycogen, especially when being actively degraded, is a potent inhibitor of AMPK, and there is likely to be reciprocal regulation between glycogen and AMPK. Proteomic studies have been performed on glycogen isolated from mouse and rat liver and the mouse adipocyte cell line 3T3-L1 (381, 434). Synthase, phosphorylase, GBE, GDE, and glycogenin were all abundant, as well as PP1 regulatory subunits. Stbd1, laforin, and proteins associated with endoplasmic reticulum were only found in the hepatic proteome; lysosomal α amylase was identified only in the 3T3-L1 proteome, indicative of cell-specific differences in glycogen metabolism. The lack of AMPK or PhK binding was consistent with earlier studies and attributed to the stringency of the isolation procedure and the nutritional state of the liver (450). A number of proteins associated with mitochondria, ribosomes and nuclei were also identified in lower abundance, suggestive of both a heterogeneous subcellular distribution and the link between glycogen and a variety of cellular processes, as has been discussed in previous sections. It should be noted that many structures of glycogen-associated enzymes have been determined, and they are quite large relative to the size of the glycogen granule (Figure 1.11). The dimeric structure of rabbit muscle glycogenin is ~80 kDa in solution and the crystal structure of the monomer is 63 Å at its widest point (228). This corresponds to one fourth of the diameter of the average glycogen β particle (25 nm, i.e. 250 Å). Yeast glycogen synthase is tetrameric, measuring 115 Å at its widest point, approximately half the diameter of a glycogen granule; each subunit is about 90 Å long (264). Monomeric human brain phosphorylase is 78 Å long, but it also is known to function as a dimer (246). Yeast GDE and human GBE monomers are 137 Å and 94 Å long, respectively. (250, 251). The elongated human laforin dimer measures 119 Å in length both in crystals and in solution (190). PhK and AMPK are both very large multisubunit complexes (245, 445).

Although a complete structure of AMPK is not yet available, structures of the PhK holoenzyme and PhK in complex with phosphorylase determined by cryo-EM reveal that these complexes are 270 Å by 225 Å and 310 Å by 250 Å, respectively, roughly the same dimensions as glycogen β particles (451). With this scale in mind, one can begin to envision that the surface of glycogen molecules is likely to be very crowded, and enzymes may compete for access to glucan chains. Molecular crowding has been shown to influence the interaction of PhK with phosphorylase and with the glycogen molecule (436). Glycogen and its associated proteins may function as molecular scaffolds, binding to one another and recruiting a variety of enzymes that respond to glycogen levels and regulate energy homeostasis.

1.8 The future of brain glycogen research

The groundbreaking discoveries of the 20th century greatly expanded our understanding of glycogen metabolism and monumentally impacted the broader field of biochemistry. As a result, there is a widespread belief that glycogen metabolism is an antiquated area of research that requires no further investigation. However, there is a milieu of interesting questions that remain unresolved regarding glycogen, particularly on its complex yet elusive role in the mammalian brain. Defects in glycogen metabolism are a hallmark of many diseases, including neurodegenerative diseases. The advent of new technologies makes it possible to better study brain glycogen and will most certainly lead to a better understanding of its architecture and metabolism in a variety of contexts.

Periodic acid-Schiff introduced Discovery of reversible phosphorylation McManus, 1946 of glycogen phosphorylase Krebs and Fischer, 1955 First isolation of native liver glycogen by ultracentrifugation Lazarow, 1942 Tree-like glycogen structure proposed Meyer and Bernfeld, 1940 First observation of glycogen via electron microscopy Husemann and Ruska, 1940 Glycogen first isolated Discovery of glycogen from dog liver synthase and UDP glucose Bernard, 1857 Leloir et al., 1957 Glycogen isolated from First description of α, β, and γ particles horse muscle and kidneys Drochmans, 1962 Sanson, 1857 Model of glycogen structure refined Gunja-Smith et al., 1970 Glycogenin discovered Whelan, Roach & Cohen, 1200 1977-1992 1000

Discovery of phosphorylase #

and the Cori cycle P

Cori, Cori, et al., 1936-1943 800 u

Glycogen l

i Glycogen-lactic acid cycle c Liver glycogen 600 a discovered in frog muscle t

i Meyerhof, 1920 Muscle glycogen o 400 n

s First robust method of glycogen purification established 200 Pflüger, 1909 Brain glycogen 0 1860 1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010 Figure 1.1 Timeline of landmark discoveries in the history of glycogen- related research. Notable discoveries and papers are shown. Claude Bernard, the physiologist who discovered glycogen and its function in liver, and Nobel laureates are pictured. PubMed (https://www.ncbi.nlm.nih.gov/pubmed/) searches were performed using the keywords shown (glycogen, liver glycogen, muscle glycogen, or brain glycogen), and the number of publications per year for each set of keywords are shown. Key terms related to glycogen synthase kinase (GSK, GSK3, GSK3 beta) were excluded from search results since much of the GSK-related literature is not directly relevant to glycogen metabolism.

Figure 1.2 Glycogen in the CNS of invertebrates and vertebrates. Drawings illustrating glycogen in the CNS of multiple organisms detected via iodine staining (18). Copyright © 2005 John Wiley and Sons, Inc. Used with permission.

39 a b first degradation second degradation

third degradation debranching enzyme

Figure 1.3 The Cori cycle and the tree-like structure of glycogen. (a) The cycle of carbohydrates between liver and muscle proposed by Cori and Cori (1929) when they observed that ingestion or subcutaneous injection of sodium d-lactate led to glycogen deposition in the liver. Their subsequent work validated and elaborated this model, which became known as the Cori cycle. (b) A segment of the tree-like structure of glycogen based on Meyer's model and the results of the Cori group. Glucose residues are represented by circles; dotted, bisected, and half-filled circles correspond to glucose residues released by the first, second, and third rounds of degradation with phosphorylase, as indicated. Filled circles represent glucose residues released by α-1,6 glucosidase (i.e. debranching enzyme). Tiers are numbered. Modified from Larner et al. 1952. Copyright © American Society for Biochemistry and Molecular Biology. Used with permission.

40 a b

10 μm

100 μm 10 μm

Figure 1.4 Early studies of brain glycogen. (a) Rapid glycogen loss following extraction from brain of normal and convulsed mice; glycogen was measured using a modification of the method by Kerr (1938) (from (56)). Copyright © Company of Biologists, Ltd. Used with permission. (b) Staining of glycogen (dark stain) in the perfused rabbit brain using the lead-tetra- acetate-Schiff method, a modification of PAS (65). The identity of stained regions as glycogen was confirmed by salivary digestion of comparable sections. Left: glycogen is abundant in the granular and molecular layers of the cerebellum, but Purkinje neurons lack glycogen. Upper right: ependymal cells of the hypothalamus lining the third ventricle show intense staining for glycogen, and granules are abundant in the neuropil, but nerve cells lack glycogen in this region. Lower right: some small nerve cells in the lateral hypothalamic nucleus appear to contain glycogen. Scale bars have been approximated based on magnification. Copyright © 2004 John Wiley and Sons, Inc. Used with permission.

41 a b c γ

α β β β α

200 nm 50 nm 50 nm

d e

Figure 1.5 Glycogen ultrastructure as demonstrated by electron microscopy. (a, b) Negative staining of natively purified rat liver glycogen from reveals the presence of α, β and γ particles (74). Copyright © 1962 Elsevier. Used with permission. (c) Typically, purified muscle glycogen exists only as β particles (75). Scale bars have been approximated based on magnification. Copyright © Rockerfeller University Press. Used with permission. (d,e) Glycogen in the rat cerebral cortex visualized by EM (76). (d) In sections of normal cortex, glycogen β particles (g) can be found in astrocyte processes (AP) and astrocytic end feet applied to the basement membrane (b). An astrocytic bundle of fibrils (f) is labeled. (e) One day after irradiation with ionizing particles, end feet (ef) are enlarged and contain numerous glycogen granules and mitochondria (m). An endothelial cell lining the capillary is also labeled (E). Copyright © Rockerfeller University Press. Used with permission.

g Amylopectin a n CH 2 OH 6 CH 2 OH c Glycogen d ci u O 5 O d s α-1,6 re d HO 4 HO 1 - n g n e O in o HO c n 3 2 O u d d n CH 2 OH HO CH 2 OH HO 6 CH 2 re e O O 5 O

HO HO 4 HO 1 O O HO 3 2 HO HO HO 1 2 3 4 5 α-1,4 GYG b

Amylose Figure 1.6 Structure and synthesis of polyglucans. (a) Starch and glycogen contain linear chains of glucose joined by α-1,4 glycosidic linkages, and α-1,6 glycosidic linkages constitute the branch points. The reducing ends (containing an aldehyde group) are oriented near the interior of the polysaccharide molecule, while metabolic enzymes work on the nonreducing ends, which are oriented outward. (b) The α conformation of α-1,4 linked polyglucans gives them a propensity to twist, and long unbranched chains can form single or double helices; a model of a double helix is shown. (c) A model of the first five tiers of the glycogen particle according to the Whelan model. A tyrosine residue of glycogenin (GYG) is shown, which is covalently linked to the glucan chain making up the first tier. (d) A recent model for starch structure according to Bertoft {Bertoft, 2017 #1132}. In amylopectin, the branched component of starch, branching is clustered, and the long linear chains form double helices (grey cylinders) that make up the crystalline regions of starch. Amylose, the unbranched component of starch, is believed to occupy areas around the amylopectin molecules.

a b

c d

Figure 1.7 Chain length distribution (CLD) of glycogen and amylopectin demonstrated by HPAEC. (a) CLD of bovine liver and rabbit liver glycogen determined by HPAEC (156). Copyright © 1993 Japan Society for Bioscience and Agrochemistry, reprinted by permission of Taylor & Francis, Ltd. (b) HPAEC profile and CLD of potato amylopectin (157). Copyright © 1996 Elsevier. Used with permission. (c) CLD of skeletal muscle glycogen from rested mice (not run) and 1 hour following exhaustive exercise (160). (d) Iodine spectra of skeletal muscle glycogen from rested mice and 1 hour, 3 hour, 1 day, and 6 days post-exercise (160). Copyright © 2015 American Society for Biochemistry and Molecular Biology. Used with permission.

a b

] %

c [

l e

1 μm R Degree of polymerisation Figure 1.8 Insights from Lafora disease, a non-classical glycogen storage disease. A typical LB visualized by electron microscopy in the human retina (452). Scale bar has been approximated based on magnification. Copyright © 1980 Springer Nature. Used with permission. (b) Neuronal (arrows) and astrocytic (arrowheads) polyglucosan accumulations visualized by immunostaining in the dentate gyrus (DG) and CA1 of the hippocampus in malin knockout (KO) mice (180). Scale bars = 10 μm. (c) Normal and abnormal CLD of purified polysaccharides from WT and LD (Epm2a-/-) mice determined by HPAEC (159).

45 a b d Macroglycogen

T Tier Diameter (nm) Proglycogen

4000 60,000 o Proglycogen Macro- t 1 0.2 synthase glycogen a

2 4 e

50,000 l

g c

3000 l 3 7.8 i

u y

t phosphorylase

c c

40,000 4 11.6 r

n a

s 5 15.4

p y u 2000 30,000 6 19.2

β q ~25 nm

r e

e 7 23 r

20,000 s

F 8 26.8 < 42 nm

1000 d 10,000 u 9 30.6 glycogenin e 10 34.4

s 0 0 11 38.2

1 2 3 4 5 6 7 8 9 10 11 12 12 42

e l

Tier c

t r

c a

1 2 3 4 5 6 7 8

~127 nm

Figure 1.9 Size and structure of glycogen α and β particles, proglycogen and macroglycogen. (a) Distribution of β glycogen particle size in muscle (365) compared to the estimated number of glucosyl residues per glycogen molecule (339). The theoretical threshold between proglycogen and macroglycogen is after tier 8. (b) Estimated diameter of the tiers of glycogen particles. (c) Schematic diagrams of an 8-tiered glycogen molecule based on the studies of Whelan and mathematical modeling of (134). (d) A possible relationship between proglycogen and macroglycogen and α and β particles and approximated sizes. Grey circles represent glycogenin molecules. Diagram of α particle was adapted from (379).

a Laforin Glycogen DAPI Merged c Glycogen Glycogen Merged

Synthase

l G

b Laforin AMPK DAPI Merged

) +

( Glycogen Glycogen Merged

c Synthase Phosphorylase

)

(

l G

Figure 1.10 Nuclear-cytoplasmic shuttling and colocalization of glycogen associated enzymes. (a) In media containing glucose and metformin, laforin colocalizes with glycogen in Neuro2A cells (421). (b) When expressed in COS-7 cells, laforin and AMPK colocalize (with glycogen) in the cytosol in the presence of glucose, and translocate to the nucleus upon glucose starvation (421). For (a,b), scale bars = 10 μm. Copyright © 2012 American Society for Microbiology. Used with permission. (c) Immunolabeling of glycogen synthase, phosphorylase, and glycogen itself in tilapia gill sections (433). Scale bars = 20 μm. Copyright © Company of Biologists. Used with permission.

GLYCOGEN

Glycogen Synthase 115 x 110 x 112 Å

Glycogen Debranching Enzyme Glycogenin 137 x 113 x 68 Å 39 x 50 x 63 Å

Glycogen Glycogen Phosphorylase Branching Enzyme 93 x 78 x 68 Å 50 x 64 x 94 Å

Laforin 50 Å 119 x 57 x 47 Å Figure 1.11 Crystal structures and relative sizes of glycogen-associated proteins. Crystal structures of yeast glycogen synthase (PDB: 3NAZ), yeast glycogen debranching enzyme (PDB:5D06), human branching enzyme (PDB: 4BZY), human brain glycogen phosphorylase (PDB: 5IKO), rabbit muscle glycogenin (PDB: 1LL2), and human laforin (PDB: 4RKK). Structures were superimposed manually in PyMol and are shown at equivalent zoom. The size of an average glycogen β particle (diameter = 250 Å) is shown for scale to illustrate the relative sizes of the proteins and glycogen particle. Crystal structure dimensions were calculated based on an inertia axis aligned bounding box in PyMol. In multimeric structures, individual subunits are colored separately.

MATERIALS AND METHODS

2.1 Structure-guided in vitro studies of laforin and laforin mutants

2.1.1 Cloning, protein expression, and protein purification

Laforin tends to aggregate when expressed in E. coli, but removal of 3 amino acids at the C-terminus reduces aggregation (453). We refer to this construct as laforin C329X. The crystal structure of laforin C329X containing a mutation of the catalytic cysteine (C266S) has been described (190). For the studies herein, wild- type and mutant laforin on the C329X background were expressed as His6-tagged fusions in pET28b (Novagen) as previously described (190, 453). Laforin is a bimodular protein with a carbohydrate-binding module (CBM) spanning amino acid residues 1-137 and a dual specificity phosphatase (DSP) domain spanning residues 138-331. A crystal structure of the laforin DSP domain (residues 150- 331) has been previously crystallized (454). An analogous DSP construct was generated by subcloning the DNA sequence for residues 150-331 from pET28b laforin (full length) into ppSUMO, a gift from the Dixon laboratory encoding a small ubiquitin-like modifier (SUMO), to generate the N-terminally His6-tagged SUMO- DSP fusion. The laforin CBM construct (amino acids 1-137) was generated with an N-terminal His6-tag in pET28b using the Gibson Assembly® Cloning Kit (New England Biolabs). All proteins were expressed in BL21-Codon Plus E. coli cells and purified using immobilized metal affinity chromatography and a Profinia Purification System (BioRad) and size exclusion chromatography via an ÄKTA fast protein liquid chromatography system (GE Healthcare). Purity of proteins was determined by SDS-polyacrylamide gel electrophoresis (PAGE) with Coomiassie staining. For yeast two-hybrid assays, pEG202 laforin encoding a LexA-laforin fusion protein and pACT2, pACT2-malin and pACT2-PTG encoding Gal4 activation domain (GAD) and GAD fusions have been previously described (313, 455-457). pWS- malin and pWS-PTG encoding HA-tagged proteins were generated by the Sanz Laboratory. All pET28b and pEG202 laforin mutants were generated by site- directed mutagenesis (QuickChange Lightning, Agilent; Q5 Site-Directed Mutagenesis, New England BioLabs; GENEWIZ Site-Directed Mutagenesis).

2.1.2 Glycogen phosphate determination and dephosphorylation assays

Glycogen purification from rabbit muscle and phosphate content determination were performed as previously described (174, 210). Glycogen was also dephosphorylated with purified laforin, amyloglucosidase (Sigma), -amylase (Sigma), and/or Antarctic phosphatase (New England Biolabs). Phosphate release was quantified using the Pi ColorLock Gold Phosphate Detection system (Innova Biosciences), a commercial reagent based on the malachite green assay

for detecting inorganic phosphate (174, 458). Kinetics of dephosphorylation were defined as previously (174). Assays were performed in 100 L reactions containing 2.5 g enzyme, 2 mM DTT, and phosphatase buffer (100 mM sodium acetate, 50 mM bis-Tris, 50 mM Tris-HCl, pH 6.5) at 25C, and time and glycogen concentration were varied as indicated for kinetic experiments. Mutants were assayed in the linear range with respect to time, enzyme amount, and substrate concentration. For specific activity determinations, reactions were performed for 30 minutes with 10mg/ml glycogen. Reactions were terminated with the addition of 25 L of the highly acidic Pi ColorLock Gold mix, and 10 L stabilizer was added after 5 minutes. After an additional 5 minutes, absorbance was measured at 635 nm. Phosphate release was quantified based on the Pi ColorLock standard curve.

2.1.3 Site-specific dephosphorylation assays

Radiolabeled starch was prepared as previously described (184, 206, 207). Briefly, phosphate-free Arabidopsis (sex1-3) starch was phosphorylated with purified glucan water dikinase (GWD) and phospho-glucan water dikinase (PWD). The C6-labeled starch was prepared by including [β 33P]-ATP during the GWD incubation and unlabeled ATP during the PWD incubation. The C3-labeld starch was prepared by including unlabeled ATP during the GWD incubation and [β 33P]- ATP during the PWD incubation. Labeled starch was washed thoroughly after each phosphorylation step to remove unbound phosphate. [β 33P]-ATP was obtained from Hartman Analytic. Dephosphorylation reactions were performed in a volume of 150 L with 50 ng of enzyme in dephosphorylation buffer (100 mM sodium acetate, 50 mM bis-Tris, 50 mM Tris-HCl, pH 6.5, 0.05% [v/v] Triton X-100, 1 g/L [w/v] BSA, and 2 mM DTT) and 3 mg/ml of either C6- or C3-labeled starch. After 2.5 minutes on a rotating wheel at 25C, reactions were quenched by adding 50 L of 10% SDS, and then centrifuged for 5 minutes at 13,000 rpm to pellet the starch. 150 L of the supernatant was added to 3 mL scintillation liquid, and 33P release was quantified using a 1900 TR liquid scintillation counter (Packard). For kinetic experiments, time and substrate concentration were varied as indicated. Kinetic parameters were determined using the Prism Software (Graphpad).

2.1.4 Differential scanning fluorimetry (DSF)

Experiments were performed using a CFX96 Real-Time PCR system (BioRad). Individual reactions contained 2 μM protein and 5X SYPRO Orange Protein Gel Stain (Invitrogen). DP7/maltoheptaose (Elicityl), DP24 maltodextrins (Elicityl), rabbit liver glycogen (Sigma), or -cyclodextrin (Sigma) were used as substrates in DSF reactions. Melting was monitored from 20 to 90°C at a ramp rate of 1°C/50sec. Melting temperature (Tm) was calculated from a Gaussian fit of the first derivative of the melting curve. Data analyses and binding fits were determined

50 using the Prism software (Graphpad). For DSF experiments in Chapter 4, buffers, water, and substrates were preincubated with PiBind resin to remove contaminating phosphate. Phosphate was added to DSF experiments from a 0.05 M K2HPO4/KH2PO4 stock for the indicated concentration.

2.1.5 Yeast two-hybrid assays

Saccharomyces cerevisiae were transformed with the indicated plasmids, and transformants were grown in selective SC medium. Extracts were prepared as described previously (313); samples were separated by SDS-PAGE and analyzed by Western blotting using the corresponding anti-LexA (Santa Cruz Biotechnology), anti-HA (haemagglutinin; Sigma) antibodies. Yeast two-hybrid assays were performed as previously described (456). Briefly, transformants were screened for β-galactosidase activity using a filter lift assay. The strength of the interaction was determined by measuring β-galactosidase activity in permeabilized yeast cells and expressed in Miller units.

2.1.6 Protein phosphate content determination

Pi ColorLock Gold (Innova Biosciences) was added to at least 100 g of protein, and then centrifuged at 13,000 rpm for 1 minute to pellet insoluble protein. The supernatant was removed and added to a 96-well plate for absorbance measurement at 635nm. Samples were boiled prior to phosphate measurement unless otherwise indicated. PiBind resin (Novus Biologicals) was used to remove contaminating phosphate from buffers, substrates and incubations, as indicated.

2.1.7 Dialysis with phosphate binders

Aliquots of 50 g of laforin C266S was added to Pur-A-Lyzer mini dialysis tubes (Sigma) and dialyzed into 100 mL of purification buffer (100 mM NaCl 20 mM Tris pH 7.5, 10% glycerol) with either 50-100 L PiBind or 1 mM phosphate binder. Samples were incubated at room temperature for 24 hours on a mixing plate with a slowly rotating stir bar. Exchanges of fresh buffer with PiBind or phosphate binder were performed approximately every four hours (4 exchanges total). Aluminum hydroxide [Al(OH)3], calcium carbonate [CaCO3], calcium acetate [CaAc2], lanthanum carbonate [La2(CO3)3], magnesium carbonate [MgCO3] were obtained from Sigma. Sevelamer HCl and CO3 were purchased from ApexBio.

2.1.8 Hydrogen deuterium exchange mass spectrometry (DXMS)

Deuteration experiments on laforin WT and C266S were performed as previously described (459, 460). Deuterated laforin samples were prepared by incubation of 10 g protein with D2O buffer (300 mM NaCl, 50 mM Tris-HCl, 3 mM

TCEP, pD (read) 7.1) at 0°C for 10, 30, 100, 300, 1000, 3000, and 10000 s, and then quenched (5.3% v/v/ formic acid, 15% v/v glycerol, 2.5M GuHCl, 39 mM TCEP, pH 2.4). Quenched samples were digested with pepsin and peptides were separated by chromatography. Mass spectrometric analysis of peptides was performed using an LCQ Classic (Thermo Fisher) electrospray ion trap-type mass spectrometer and an electrospray Q-TOF mass spectrometer (Waters Corp, Milford, MA). Data from all sample sets were acquired from a single automated run of 8 hours. Deuteration experiments were also performed with 5 mg/ml rabbit liver glycogen (Sigma). Maximum changes in deuteration for peptides were used to determine the difference between WT and C266S laforin and differences with each protein binding to glycogen.

2.1.9 Analytical ultracentrifugation (AUC)

Sedimentation velocity experiments were conducted using a Beckman XL-I analytical ultracentrifuge (Beckman-Coulter, Fullerton, CA) as previously described (190). Samples were analyzed at 40,000 and 4 °C and absorption measurements taken at 280 nm. Measurements of sedimentation coefficient distributions between 0.1 and 8 s were analyzed using numerical solutions of the Lamm equation implemented in Sedfit (461).

2.1.10 Structural analysis

PyMol 2.0 was used for structural analysis, measurements and generating molecular graphics (462). Solvent accessible surface area was calculated for 4RKK using the GetArea program (Supplemental File 1) (http://curie.utmb.edu/getarea.html) (463). DSP-DSP and CBM-DSP interface surface area and residue contributions were determined using PDBePISA (http://www.ebi.ac.uk/msd-srv/prot_int/cgi-bin/piserver) (464). Homology models for malin and PTG were generated using Phyre2 (http://www.sbg.bio.ic.ac.uk/phyre2/html/page.cgi?id=index) (465). 2D ligand- protein interaction diagrams were generated using the Ligplot+ software (https://www.ebi.ac.uk/thornton-srv/software/LigPlus/) (466). Structural comparisons were performed using the DALI server (http://ekhidna2.biocenter.helsinki.fi/dali/) (467). Electrostatic potential was mapped onto the laforin structure and malin homology model using the Adaptive Poisson-Boltzmann Solver (APBS) plugin in Pymol 2.0 (468).

2.2 LB purification, characterization, and VAL-0417 studies

2.2.1 Mouse lines

Epm2a-/- (177, 469) and Epm2b-/- (178) mice have been previously described. C56Bl/6 WT, Epm2a-/- and Epm2b-/- animals were maintained in a

12:12 hr light-dark cycle and were given ad libitum access to food and water. All procedures were approved by the UK Institutional Animal Care and Use Committee (IACUC) as specified by the 1985 revision of the Animal Welfare Act.

2.2.2 Generation of the antibody-enzyme fusion VAL-0417

VAL-0417 was designed and produced by Valerion Therapeutics (Concord, MA). The cDNA encoding the human IgG1 Fab-linker-AMY2A heavy chain and light chain were synthetically produced with codon optimization for mammalian cell expression and cloned into pTT5. HEK293 cells expressing a truncated variant of the Epstein Barr Virus nuclear antigen 1 (HEK293-6E) increase the volumetric yield of monoclonal antibodies and fragments and were used for VAL-0417 expression (470). 2 1L cultures of HEK293-6E cells in 2L shake flasks were transfected with 1 mg of total plasmid DNA/L (1:1 ratio heavy chain:light chain) culture using PolyPlus linear Q-PEI at a 1:1.5 (w/v) DNA:PEI ratio. Culture parameters were monitored using a ViCell XR (Beckman Coulter) for density and viability. Culture was harvested 5 days post transfection via centrifugation for 5 minutes at 1000×g. The conditioned culture supernatant was clarified by centrifugation for 30 minutes at 9300×g. Pre-packed CaptureSelect IgG-CH1 affinity columns (Fisher) were equilibrated in PBS (pH 7.2). VAL-0417 from 2 L of exhausted supernatant was top-loaded onto affinity columns (2 x 1 mL columns in tandem) at 4°C overnight. The column was washed with approximately 15 column volumes (CV) of PBS, 15 CV of buffer B (1×PBS with 500 mM NaCl, pH 7.2) and 15 CV of PBS. The resin- bound fusion protein was eluted with 10 CV of Buffer C (30 mM NaOAc, pH 3.5- 3.6), collecting the protein in 1 mL fractions diluted in 1/10th volume Buffer D (3 M NaAcetate pH~9.0) to neutralize. To minimize the elution volume, elution was paused for several minutes between each fraction collected. Fractions were analyzed by A280 prior to pooling fractions and pools were analyzed by SDS- PAGE. VAL-0417 remained in the non-bound pool from the first affinity chromatography pass. The above procedure was repeated to capture remaining fusion protein. The affinity pools were combined prior to dialysis. The combined CaptureSelect IgG-CH1 affinity pool (18 mL) was dialyzed against 3 x 1 L of dialysis buffer (20 mM Histidine, 150 mM NaCl, pH 6.5) at 4°C. The dialyzed pool was concentrated to 1 mg/mL using a VivaSpin 20 (10K MWCO, PES membrane) centrifugal device prior to final analysis and storage at -80°C. Purified Fab-AMY was analyzed by size exclusion chromatography (Agilent HP1100) showing a single peak before and after a freeze-thaw cycle, indicative of a single, stable species. Activity of the purified protein was quantified by an amylase activity colorimetric kit (BioVision) utilizing the amylase-specific E-G7-pNP substrate (51).

2.2.3 Purification of native LBs from Epm2a-/- and Epm2b-/- mice

19-24 month old Epm2a-/- mice were euthanized by CO2 and decapitation, and brain, heart and hindlimb skeletal muscle were immediately harvested, flash frozen and stored at -80°C. 12 month old Epm2b-/- mice were euthanized by cervical dislocation, and muscle tissues were similarly collected and stored. Tissues were pulverized over liquid N2 using a Freezer/Mill Cryogenic Grinder (SPEX SamplePrep). Powdered tissue was weighed and homogenized on ice in 4 vol. lysis buffer (100 mM Tris-HCl pH 8.0, 200 mM NaCl, 1mM CaCl2, 0.5% sodium azide) using a Dounce tissue grinder. For larger volumes, the grinding pestle was attached to a motorized drill. The homogenate was centrifuged for 10 min at 10,000×g at 4°C in a Ti-70 rotor and Optima XPN-90 ultracentrifuge (Beckman Coulter) and the supernatant was removed. The pellet was resuspended in an equivalent volume of lysis buffer, and 20% SDS was added for a final concentration of 0.2%, and 20 mg/ml Proteinase K (Invitrogen) was added for a final concentration of 0.4 mg/ml. Proteolytic digestion was performed in a 37°C water bath overnight. The digested samples were then syringe-filtered through 140 μm and 60 μm nylon net filters installed in Swinnex filter holders (Millipore), centrifuged at 16,000×g for 5 min, and the supernatant was removed. The LBs were resuspended in 10% SDS and then washed 5 times in LB buffer (10 mM HEPES-KOH pH 8.0, 0.1% sodium azide) each time with centrifugation at 16,000×g for 5 min. Final LB pellets were gently and thoroughly dispersed in LB buffer with a pipet. Polysaccharide yield at various steps of the purification were determined using the Pflüger method (see below). 2 μL of LB preparations were stained with 5 μL 20× Lugol's iodine, mounted on glass slides with a glass coverslip, and visualized at 100x using a Zeiss Axioimager Z1 equipped with an AxioCam 1Cc5 color camera. Individual LBs in 3-5 micrographs were measured manually using ImageJ and a frequency distribution was calculated for LBs from each tissue type using Prism 6.0 software (GraphPad) and a bin width of 1 μm. Only clearly delineated LBs were measured; clumps and dust-like particles were excluded. Purified LBs were stored at -20°C.

2.2.4 Pflüger method for polysaccharide isolation

Glycogen and polyglucosan from Epm2a-/- and Epm2b-/- mice have been purified and characterized by multiple groups using the Pflüger method, but a chaotropic salt must be added to enhance the efficiency of polyglucosan precipitation (209). We wanted to clarify that under native conditions, and prior to the Pflüger treatment, LBs in Epm2a-/- and Epm2b-/- tissues are intact, micron- sized structures. Pflüger treatment converts the LBs to much smaller polysaccharide molecules that are on the same nanometer scale as glycogen molecules, but they contain elevated phosphate and an altered chain length distribution (158, 209). Thus, the term "LB" refers to the native, micron-sized

54 polysaccharide-containing structures found in LD tissues, and "polyglucosan" refers to the abnormal polysaccharide comprising LBs, which is released with Pflüger treatment. The Pflüger method is direct and very sensitive and was used to track polysaccharide yield throughout the native LB purifications or to detect polysaccharide content of mouse tissues for in vivo studies. Aliquots from native LB purifications or tissue homogenates were added to 10 vol. 30% KOH, boiled for 2 hours, and allowed to cool. 2 vol. cold ethanol and 10 μL LiCl (1 M or 20 mM) were added, and samples were precipitated overnight at -20°C. Precipitated samples were centrifuged for 10 min at 16,000×g at 4°C, the supernatant was removed, and the polysaccharide pellet was resuspended in water. Two additional precipitations with cold ethanol and LiCl were carried out, each for 1-2 hours at - 20°C. The final pellet was washed in cold ethanol and resuspended with vortexing in 200 μL water. Polysaccharide was quantified by overnight hydrolysis with amyloglucosidase from Aspergillus niger (Sigma) and glucose determination was carried out using the R Biopharm Inc. D-glucose kit (Fisher). Fluorescence (ex340/em445) rather than absorbance of the NADPH product (stoichiometric to glucose) was measured for higher sensitivity.

2.2.5 Separation of glycogen and LBs in native LB purification

After the initial centrifugation step in the LB purification, a fraction of the heart and muscle glycogen was consistently present in the supernatant. In our preliminary purifications, we found that the homogenate and pellet fractions stained brown with Lugol's iodine after boiling, as did the final LBs and amylopectin, but the supernatants stained yellow like glycogen. We also stained the Pflüger-purified polysaccharide fractions from our large-scale preparations with Lugol's iodine, measured absorbance at 550 nm, and compared the values to the polysaccharide concentrations determined by hydrolysis and glucose measurement. While the polysaccharide from the supernatant did not absorb at 550 nm, polysaccharide from the pellet, filtrate and final fractions produced a high absorbance relative to concentration. Polysaccharide in the homogenate fraction was less absorbent relative to concentration. These data indicate that the polysaccharide in the heart and muscle tissue homogenates included both glycogen and polyglucosan (in the form of LBs), that were separated into supernatant and pellet fractions after low-speed centrifugation. No polysaccharide was detected in the supernatant fractions from brain. This result is consistent with the observation that in normal mice, glycogen levels are lower in brain than in other tissues in part due to its rapid catabolism after euthanasia (471).

2.2.6 Iodine-based measurements

Lugol's iodine was prepared as a 20× stock (1.5 M KI and 100 mM I2). To detect

LBs in preliminary purifications, 50 μL samples from purification fractions were boiled for 15 minutes on a 95°C heat block, clarified by centrifugation (16,000×g for 1 min), and 35 μL of the supernatant was added to 50 μL 1×Lugol's iodine and 15 μL water for 100 μL total. To analyze iodine absorbance of polysaccharide from LB purifications, Pflüger-isolated polyglucosan fractions were added to 50 μL 1×Lugol's iodine for 100 μL total and absorbance was measured at 550 nm. Absorbance was graphed alongside glucose-based concentrations to illustrate the fractionation of polyglucosan and glycogen into supernatant and pellet fractions, respectively. For spectral scans, 50 μg LBs, rabbit liver glycogen (Sigma), and potato amylopectin (Sigma) were solubilized by boiling for 30 min, and added to 50 μL 1×Lugol's iodine for 100 μL total. Absorbance scans were performed at 400- 800 nm in 10 nm steps.

2.2.7 Determination of LB phosphate content

Muscle glycogen and LBs contain covalent phosphate linked to the C2, C3, and C6 hydroxyls of glucose moieties, and the relative ratios of these modifications are equivalent (158, 210, 212). Boiling in mild HCl hydrolyzes glycosidic bonds and releases the acid-labile C2- and C3-linked phosphate leaving C6 phosphoesters intact in the form of glucose-6-phosphate (158). Thus, inorganic phosphate after mild acid hydrolysis represents only C2- and C3-linked phosphate, i.e. two-thirds of total phosphate. To determine phosphate content, skeletal muscle LBs were boiled for 2 hours at 95°C in 1 M HCl. Reactions were neutralized with NaOH and inorganic phosphate in the sample was determined using the Pi ColorLock Gold Phosphate Detection System (Innova Biosciences). Glycogen purified from rabbit skeletal muscle as previously described was used as a positive control, and the levels of phosphate detected using our method are consistent with two-thirds of published total phosphate levels (210).

2.2.8 In vitro degradation assays

Corn starch used for degradation assays and microscopy was purchased from Sigma. Starch and LB degradation experiments were performed in degradation buffer (30 mM HEPES-KOH pH 7.5, 5 mM MgCl2, 5 mM CaCl2) and a total volume of 100 μL. Reactions were performed in triplicate using PCR strip tubes and moderate agitation on a vortex to keep substrates in suspension, and we found that this level of agitation did not affect enzyme activity. Degradation reactions were allowed to proceed overnight (13-18 hours) unless otherwise indicated, after which tubes were centrifuged to pellet undigested substrate. 50 μL of the supernatant (containing the soluble degradation product) was transferred to a new tube with 50 μL 2 M HCl and boiled for 2 hours at 95°C in a C1000 thermocycler (Bio-Rad) to hydrolyze all glucans to glucose. Samples were neutralized with 50 μL 2 M NaOH, and glucose was determined using the Boehringer-Manheim D-

56 glucose kit (R-BioPharm). For visualization of LB degradation product, degradation reactions were allowed to proceed as above, and the reactions were centrifuged to pellet undigested substrate. After 75 μL of the supernatant was removed, the remaining 25 μL was resuspended, and 5 μL was stained with 5 μL 20× Lugol's iodine, mounted on glass slides with a glass coverslip, and visualized using a Nikon Eclipse E600 using DIC/Nomarski contrast and an AxioCam MRm camera at 100x. In addition to the R Biopharm D-glucose kit, the absorbance- based PGO assay (Sigma) was also used to quantify glucose in starch degradation assays. In vitro degradation of LBs in muscle homogenates was performed as follows: 300 mg of skeletal muscle from WT or laforin KO mice was pulverized over liquid N2 and homogenized in 4 vol. degradation buffer. Homogenate was split into 2 × 500 μL aliquots, and 25 μg VAL-0417 was spiked into one aliquot for a final concentration of 0.05 mg/ml. Samples were incubated on a rotator overnight at room temperature and the next morning centrifuged for 5 min at 16,000×g. The supernatant was removed, and the pelleted material containing LBs was resuspended in degradation buffer and boiled for 30 min on a 95°C heat block to solubilize LBs. The boiled samples were clarified by centrifugation (16,000×g for 5 min) and 25 μL of supernatant was added to a microplate. 50 μL of 1×Lugol's iodine and 25 μL of water were added to samples, and absorbance was measured at 550 nm.

2.2.9 Scanning electron microscopy

Starch granules were incubated with enzymes in degradation buffer overnight while agitating, washed in 1mL 100% ethanol, dried in a vacuum centrifuge, and applied to carbon tape on a pin mount. Samples were coated with gold and platinum to improve sample conductivity, and visualized under high vacuum at 2 kV using an FEI Quanta 250 field emission scanning electron microscope. Washing in ethanol, drying, and manual application to carbon tape did not yield satisfactory results for LB visualization, so alternatively, diluted LBs were applied to a mounted silicon wafer, lyophilized, coated in gold and platinum, and visualized at 2 kV with the FEI Quanta 250.

2.2.10 Profiling of VAL-0417 degradation product by HPAEC-PAD

80 μg of LBs were treated with 2.67 μg of VAL-0417 in degradation buffer to a final concentration of LBs at 1 μg/μL. The enzymatic reactions were done at 37°C in triplicates. 20 μL samples were removed at 24, 48, 72 hour intervals followed by addition of 2.67 μg of VAL-0417. The aliquots were stored at -20°C until they were run. The final reactions were continued at 37°C for total of 168 hours. All samples were profiled using CarboPac PA-100 column (Thermo-Dionex, 4 x 250 mm) and detected with PAD detector. 5 μL of samples from each time point

(corresponding to 5 μg LBs total) were injected on the column. Blank LBs and degradation buffer were injected as controls. Glucose and maltose in the degradation reactions were quantified by comparing with standards of known amount. 5 μg Maltrin100 was also included as a standard for oligosaccharide profiling.

2.2.11 Generation of HEK293-PTG/PP1Cα cells and polysaccharide quantitation

Protein Phosphatase 1 (PP1) stimulates glycogen synthesis by both dephosphorylating and activating glycogen synthase and inhibiting glycogen phosphorylase, the primary enzymes catalyzing glycogen synthesis and degradation, respectively (439). Protein targeting to glycogen (PTG) is a regulatory, glycogen-targeting subunit of PP1 that stimulates glycogen synthesis when overexpressed in cell lines (301, 442). Constitutive overexpression of PTG in WT mice has been shown to lead to cerebral polyglucosan accumulation that is similar to LBs in Epm2b-/- mice (182). We sought to design a cell line that accumulates polyglucosan like that which makes up LBs. HEK293 cells (472) were co-transfected with plasmids pCDH-FLAG-PTG and pCDH-(HA)3-PP1Cα-GFP harboring mouse PTG and the human PP1Cα (catalytic alpha-subunit of PP1) cDNAs, respectively. Mixed clones were selected for ~10 days in the presence of 1 μg/ml hygromycin and 0.2 mg/ml puromycin, expanded and stored in liquid N2. Analyses of protein expression and glycogen synthase activity ratio in the absence and presence of glucose-6-phosphate indicated that both proteins were expressed and the glycogen synthase activity ratio was increased 3-fold, from 0.02 in control cells to 0.06 in transfected cells. Quantitation of the expression of PTG and PP1Cα is difficult because the basal levels are very low, undetectable under our conditions. For quantitation of polysaccharide levels HEK293, HEK293-PTG/PP1Cα and Rat1fibroblasts were plated in 96-well plates at a density of 40,000 cells/well. After 48 hr, media was changed and the next day, the cells were washed 3 times with PBS followed by overnight incubation at 40°C in 50 μL of 0.2 M sodium acetate pH 4.8 containing 0.2% Triton and 0.3 mg/ml amyloglucosidase (Aspergillus niger; Sigma). Identical cell cultures were lysed in 50 μL of in 0.2 M sodium acetate pH 4.8 containing 0.2% Triton for protein determination by the Bradford procedure. After overnight digestion, polysaccharide was measured by transferring 40 μL samples to a new 96-well plate. A 160 μL reaction mixture consisting of 0.375 M ethanolamine pH7.6, 5 mM MgCl2, 1.12 mM NADP, 2.5 mM ATP and 0.2 Units of G6PDH (Roche Biochemicals 10127655001) was added and OD340 nm was recorded. Subsequently, 0.75 units of hexokinase (Roche 11426362001) were added, the reaction incubated at room temperature for 30 min and OD340 recorded again. Background absorbance was subtracted from sample absorbance and glucose equivalents were determined based a digested glycogen standard curve. The engineered HEK293 line accumulated >30-fold more polysaccharide

58 than the native HEK293 cells, and twice the level of normal glycogen found in Rat1 cells.

2.2.12 Cell culture studies of VAL-0417 uptake

PTG and the catalytic alpha-subunit of protein phosphatase 1 (PP1Cα) were stably expressed in HEK293 cells to generate a stable cell line that accumulates polyglucosan. To determine the effect of VAL-0417 on polysaccharide degradation, cells were plated in 96-well plates at a density of 40,000 cells/well, and after 4 days various concentrations of VAL-0417 were added for 20 hr to fresh growth media. Cells were then washed 3 times with PBS, fixed for 7 min with 3.7% formaldehyde in 90% ethanol, washed again 3 times with PBS and polysaccharide content was determined by hydrolysis and glucose determination. Statistical significance was assessed by using an unpaired Student t-test. For Western analyses, 20 μg of total cell lysates were separated on SDS-PAGE and protein transferred to nitrocellulose membranes that were probed with antibodies against human pancreatic α-amylase (Abcam #ab21156) and ENT2 (Alomone Labs #ANT-052) followed by appropriate HRP-conjugated secondary antibodies and chemiluminescence.

2.2.13 In vivo mouse studies with VAL-0417

For each IM injection, 20 μL VAL-0417 (30 mg/ml; 0.6 mg per injection) or 20 μL PBS was injected into one gastrocnemius. For the IV injections, mice were pretreated intraperitoneally with diphenhydramine (15mg/kg) before each IV injection to prevent anaphylaxis. After the diphenhydramine, the mice where kept under a heat lamp for twenty minutes. A Mouse Tail Illuminator (BrainTree Scientific, Inc.) was used to restrain the mouse and an insulin syringe/needle was used to inject PBS or VAL-0417 into the dorsal tail vein. 100 μL of PBS was injected into IV control mice on days 1, 5, 8, and 13. For the IV treatment group, 100 μL of 30 mg/ml VAL-0417 (3 mg total) was administered on days 1 and 5, 200 μL of 20 mg/ml VAL-0417 (2 mg total) on day 8, and 130 uL of 23 mg/ml VAL- 0417 (3 mg total) on day 13. At the indicated time points post-injection(s), mice were euthanized by cervical dislocation and decapitation. Tissues were harvested, flash frozen, powdered over liquid N2 using Freezer/Mill Cryogenic Grinder (SPEX SamplePrep), and stored at -80°C. Powdered tissues were weighed, homogenized in assay buffer from the BioVision amylase assay kit and protein concentration was determined using the Pierce BCA Protein Assay Kit (ThermoScientific). For the ELISA assay, tissue homogenates were diluted in PBS for a final concentration of 0.1 μg protein/μL in 100 μL. Diluted homogenates were added to 96-well plates for the ELISA. Polysaccharide content was determined using the Pflüger method. IT and ICV studies were performed by Northern Biomedical Research, Inc.

(Spring Lake, MI). Animals were anesthetized with a mixture of oxygen 0.5 to 1 L/min and isoflurane 1 to 5% prior to and during the catheter implantation procedure. The IT catheter attached to an osmotic pump was inserted at the cisterna magna and advanced caudally 2.5 cm to the lumbar region. The ICV cannula attached to an osmotic pump was inserted into the cerebral lateral ventricle and anchored. VAL-0417 (30 mg/ml) or PBS was continuously administered via the IT catheter or ICV cannula for 3 or 28 days (0.11 μL/hr). On the following day, mice were euthanized by isofluorane/oxygen sedation and perfusion via the left cardiac ventricle with 0.001% sodium nitrite in heparinized saline. Brains were harvested, weighed, sectioned using a Rodent Brain Matrix (RBM-2000C, ASI Instruments) with 2.0 mm coronal section slice intervals, flash frozen and stored at -80°C. Slices were directly homogenized in BioVision assay buffer using polypropylene pellet pestles in microcentrifuge tubes. VAL-0417 levels and amylase activity were determined by ELISA and the BioVision assay and normalized to protein content. Protein and polysaccharide content were determined by BCA assay and the Pflüger method, respectively.

2.2.14 Statistical analyses

Statistical analysis of correlation plots was performed using the Prism 6.0 software (GraphPad). Statistical significance of polysaccharide reduction was determined by one-way or two-way analysis of variation (ANOVA) using Prism.

2.2.15 Enzyme-linked immunosorbent assay (ELISA)

The capture antibody raised against the 3E10 Fab fragment was generated by Valerion Therapeutics (Concord, MA). Wells of a 96-well plate were incubated with 100 μL of capture antibody overnight (~16 hours) at a final concentration of 2 μg/ml in PBS. All incubations were done in a humidified chamber. Wells were then rinsed 3 times with 200 μL PBS followed by incubation for 1 hour with 200 μL of blocking solution (5% non-fat milk in PBS). Wells were then rinsed 3 times with 200 μL PBS. 100 μL of diluted tissue homogenates were added and incubated for 1 hour. Wells were then rinsed 3 times with 200 μL of Tris-buffered saline (TBS) and 100 μL of primary antibody (anti-AMY2A, Abcam #ab21156) in 5% non-fat milk and TBS was added and incubated for 1 hour. The wells were then washed (solution added and incubated for 5 minutes before being removed) 3 times with TBS. 100 μL of secondary antibody (anti-rabbit IgG, HRP-linked, Cell Signaling Technology 7074) in 5% non-fat milk in TBS was added and incubated for 1 hour. Wells were then washed 5 times with TBS. 100 μL of TMB substrate (ThermoFisher N301) was added and a timer started. After the highest concentration of standard curve saturated or a pre-set time was met, the reaction was stopped by adding 100 μL of stopping solution (0.18% H2SO4). The plate was then read for absorbance at 450 nm.

2.2.16 Confocal microscopy

For confocal microscopy, LBs or corn starch (Sigma) were stained with 20× Lugol's iodine, embedded in Mowiol 4-88 (Sigma cat #81381) on glass slides, and visualized using the TRITC channel (561nm excitation laser) and differential interference contrast (DIC) using a Nikon AR+ Scope. Z-stack images of LBs were captured in 0.1 μm steps, starch granules in 0.25 μm steps, and 3D deconvolution was performed using the Nikon Elements Advanced Research Software. The Richardson-Lucy deconvolution algorithm was utilized with 10 iterations per stack.

2.2.17 Fluorescence-associated capillary electrophoresis (FACE)

Chain length distribtuion was determined using FACE as previously described (473, 474). Samples were debranched with 20U isoamylase (Sigma) and 1U of pullulanase (Sigma) for 5h at 42°C. Enzymes were then precipitated at 100°C during 10 min, and samples centrifugated 5 min at full speed. Samples were desalted using Carbograph column with 5 mL of water. After desalting, debranched samples were eluted with 2 mL of Acetonitrile (ACN) 25% and dried using a speedvac overnight. Then, the debranched samples were labeled by adding 2 μL of 1M sodium cyanoborohydride (Sigma-Aldrich, France)/THF (tetrahydrofurane) and 2 μL of 200 mM APTS (Sigma-Aldrich, France)/15% acetic acid, with incubation overnight at 42°C. A Beckman Coulter PA800 plus instrument (Sciex Separations, Les Ulis, France) equipped with a laser induced fluorescence (LIF) detector and a 488 nm wavelength laser module was used to perform electrophoresis on reverse polarity. Separation was performed in a 50 μm I.D., 375 μm O.D. bare fused silica capillary of 60.2 cm in length (Sciex Separations, Les Ulis, France). Samples were injected for 10s at 0.5psi and separation was achieved at 30 kV in carbohydrate separation gel buffer (Sciex Separations, Les Ulis, France), diluted 1/3 in ultrapure water.

2.2.18 Wide angle and small angle X-ray scattering (WAXS and SAXS)

For WAXS studies, purified SmLBs, BrLBs and potato amylopectin starch were centrifuged and part of the wet pellet was poured into a 1 mm (outer diameter) glass capillary. The capillary was flamed-sealed and X-rayed in vacuum by a Ni- filtered CuKα radiation (λ = 0.1542 nm) using a Philips PW3830 generator operating at 30 kV and 20 mA. Two-dimensional diffraction diagrams were recorded on Fujifilm imaging plates read offline with a Fujifilm BAS 1800-II bioanalyzer. SAXS studies were performed at the European Synchrotron Radiation Facility (Grenoble, France) as previously described (475). 1D scattering profiles were calculated by rotationally averaging the 2D diffraction patterns.

2.2.19 LB treatments and light microscopy

Lugol's iodine was prepared by adding elemental iodine (for a final concentration of 100 mM) to a 1.5M potassium iodide solution to produce the triiodide anion (I3-). SmLBs and BrLBs were aliquoted (100 μg per tube) and washed in water by centrifugation at 16,000xg for 1 min. LB pellets were resuspending in 100 μL water or 100 μL 30% KOH for a final concentration of 1mg/ml. For sonication treatment, five pulses of 5 seconds each at 25% amplitude, with one minute on ice between each pulse to prevent sample heating. Samples were heated for 30 minutes or 2 hours as indicated. 5 μL of samples were stained with 2 μL 20× Lugol's iodine, mounted on glass slides with a glass coverslip, and visualized using a Nikon Eclipse E600 using DIC/Nomarski contrast and an AxioCam MRm or Zeiss 512 camera at 100x. LBs had a tendency to clump, after staining, samples were carefully titruated with a pipet, added to glass slides, and the coverslip was laid on top. The coverslip was very gently prodded and pressed to disperse the LB clumps and reduce the number of focal planes. Image thresholding was performed using ImageJ software (National Institutes of Health, USA).

2.2.20 Transmission electron microscopy (TEM) and lintnerization

Droplets of the dilute polysaccharide suspension were deposited onto PELCO Carbon support film with removable formvar baking (Ted Pella, Inc.). The excess liquid was blotted and samples were negatively stained with 2% uranyl acetate and allowed to dry. The grids were imaged with a JEOL 2010F field emission transmission electron microscope operating at 100kV using a CCD camera via Gatan DigitalMicrograph software. Lintnerization was performed by solubilizing LBs with 2.2 N HCl and incubation at 36°C. After 5 days, 15 μL aliquots were removed, neutralized with 0.1M aqueous NaOAc (2×) and washed with water. Samples were lyophilized, resuspended in water, and deposited onto glow-discharged carbon-coated TEM grids. After blotting of the excess liquid, the preparations have been negatively stained with 2 wt% uranyl acetate and allowed to dry. They were observed with a Philips (FEI) CM200 microscope operating at 200 kV. Images were recorded using a TVIPS TemCam F216 camera.

PERSONALIZED BIOCHEMISTRY OF LAFORA DISEASE PATIENT MUTATIONS

3.1 Introduction

Nearly all LD patients carry mutations in either the Epilepsy progressive myoclonus 2A (EPM2A) gene encoding laforin or EPM2B gene encoding malin. Laforin, mutated in approximately 50% of LD cases, is a bimodular protein comprised of a carbohydrate-binding module (CBM) from the CBM20 family and a dual-specificity phosphatase (DSP) domain. Laforin is the founding member of a class of enzymes known as the glucan phosphatases and has been shown to dephosphorylate glycogen, a soluble carbohydrate storage molecule synthesized by eukaryotic cells (173, 174, 183). In cell culture, laforin interacts with and is ubiquitinated by malin, an E3 ubiquitin ligase with a RING domain followed by six NHL repeats that comprise a substrate-interacting domain (175, 176). Malin has also been shown to ubiquitinate multiple enzymes involved in glycogen metabolism, and many of these interactions require laforin (300-304). It has been suggested that laforin may serve as an adaptor protein (310) and that the primary cause of LD is the lack of a functional laforin-malin complex (188). Over 171 distinct mutations affecting the EPM2A and EPM2B genes have been described in the LD population, including missense, nonsense, insertion, and deletion mutations (http://projects.tcag.ca/lafora/) (476, 477). Many patients are complex heterozygotes, meaning they have different mutations in the two loci of the same gene. LD is an orphan disease with an estimated prevalence fewer than one per million, though it is likely to be underdiagnosed (478). The number of reported mutations grows yearly, and many mutations have been found in only a single patient (479, 480). The genotypic variability and small number of patients make genotype-phenotype correlations very difficult, and descriptive clinical data for LD patients is not readily available. Although most LD patients degenerate rapidly, mild or late-onset cases of LD have been recently described (481, 482). A particularly mild clinical course has frequently been associated with the D146N mutation in EPM2B affecting the malin protein (483-487). This mutation decreases the laforin-malin interaction in vitro while not disrupting the ability of malin to ubiquitinate proteins (300). Other mild or late-onset cases have been linked to EPM2A mutations affecting laforin (456, 488). These patients were either homozygous for a missense mutation, carried two different missense mutations, or carried a missense and a nonsense mutation in EPM2A. A particularly striking case is a patient who lived to the age of 56, with a history of seizures that were relatively controlled with sodium valproate (489). This patient was homozygous for a novel laforin mutation, F321C. A vast amount of disease-related genetic information has emerged with the advancement of DNA technologies. Whole exome sequencing is frequently used to diagnose suspected genetic disorders or identify single nucleotide

polymorphisms (SNPs) predisposing patients to certain diseases or cancers (490). As a result, much effort has been directed at understanding the effects of the identified missense mutations or SNPs on protein stability, function, and dynamics and their contribution to disease (reviewed in (491). A popular approach is to use computational algorithms to predict pathogenicity of a variant based on sequence homology, protein structure, or both (492, 493). However, these algorithms range in performance from poor to moderately good, and results from different programs do not correlate well (494-496). Personalized medicine, the use of genetic information to diagnose and treat disease, is becoming more common in the clinic especially for the treatment of certain cancers (497). Therefore, the effects of nucleotide variants need to be determined with the greatest possible accuracy since they may be used to guide diagnostic and therapeutic decisions for patients. Cystic fibrosis (CF) is a genetic disease with variable clinical manifestations and a large number and variety of mutations that initially hindered genotype- phenotype studies (498-500). Effects of disease variants were difficult to predict using in silico tools, and were instead carefully defined through many basic biological studies (501, 502). Now, CF mutations are classified based on their functional effect(s), and mutation-specific therapies are used to treat patients (500). Although seemingly tedious and time-intensive, experimental studies defining the biological, biochemical, and biophysical defects of missense mutations are highly informative and have also been described for other diseases (491). Our lab previously determined the crystal structure of laforin, defined the structural basis for glycogen binding and dephosphorylation, and characterized its quaternary structure and dynamics (190). In this study, we analyze 27 missense mutations, some of which are associated with mild LD. These mutations reveal allosteric interactions between the CBM and DSP domain and clarify the role of structural dynamics in protein-protein interactions. The missense mutations are grouped into 5 classes based on the severity of their biochemical effects, and the data provide an explanation for mild or late-onset forms of LD. Moving forward, novel mutations can be rapidly profiled using the described biochemical pipeline. These data can be used to influence clinical decisions regarding personalized diagnoses and emerging LD therapies.

3.2 Results

3.2.1 Pathogenicity predictions of EPM2A missense mutations

As of 2017, 356 cases of LD associated with EPM2A or EPM2B mutations had been reported in the Lafora Progressive Myoclonus Epilepsy Mutation and Polymorphism Database (http://projects.tcag.ca/lafora/) (Figure 3.1a). These consist of missense mutations, premature termination codons (PTCs), and insertions and deletions (indels) in either EPM2A or EPM2B. Of the 94 different

64 mutations reported in EPM2A, 51 unique missense mutations were identified (Figure 3.1b). These 51 missense mutations accounted for 43% of LD cases caused by EPM2A mutations and 22% of all LD cases (Figure 3.1a). In 2009, 21 of these mutations were reviewed in a meta-analysis of LD mutations (477). The effects of some of these mutants on laforin function had been previously tested and were discussed, but many had not been tested. Some had no detectable defect in generic phosphatase activity. The present study was initiated with the prediction of mutation pathogenicity using three different computational tools. We selected 27 EPM2A mutations, some of which had been previously studied by our group or others (190, 477). Clinical data on many of these mutations are limited, but for most mutations their allelic context and the age of seizure onset for the patient(s) were known (Table 3.1). We also selected additional mutants identified more recently in patients with mild or late-onset LD (Table 3.1). PolyPhen-2 is a popular pathogenicity prediction programs that estimates the deleterious effects of mutations based on sequence homology, site conservation and structural features (503). The only query is amino acid sequence. PolyPhen-2 predicted that the majority of the mutations queried were "probably damaging," i.e. determined to be damaging with high confidence (Table 3.2). Many of the mutations were "possibly damaging," that is, predicted to be damaging with low confidence. The mutations F5S, V7A, F84L, N148Y, and E210K were all predicted with high confidence to be benign variants. However, most of these mutations have been identified in LD patients in the homozygous state, indicating that they are indeed pathogenic (Table 3.1). SDM and CUPSAT are commonly used to estimate changes in free energy (ΔΔG) induced by a missense mutation using the crystal structure as input (504, 505). Loss-of-function induced by destabilizing mutations is considered the most common cause of genetic disease, so stability predictions are often used to evaluate pathogenicity (491). Both programs estimated that most mutations would induce a negative ΔΔG, indicating they were moderately or severely destabilizing (Table 3.2). However, the algorithms employed by SDM and CUPSAT for calculating ΔΔG are distinct, and there was no correlation between the predicted values nor with the PolyPhen-2 pathogenicity scores (Figure 3.1c). Furthermore, we previously showed that W32G, Y294N, P301L, and F321S were significantly destabilized by a thermal melt assay, and Y294N was significantly more dynamic than wild-type (WT) laforin via hydrogen deuterium exchange (190). However, these mutations were predicted by either SDM or CUPSAT to produce a positive ΔΔG, leading to increased stability (Table 3.1). The results of prediction programs are inconsistent with patient data and our previous biochemical studies. Therefore, we initiated an experimental analysis of a panel of laforin missense mutations using the crystal structure as a roadmap for understanding their effects on laforin structure and function.

3.2.2 Stability and activity of laforin missense mutations

The 27 missense mutations selected for analysis are scattered throughout the primary structure of laforin (Figure 3.2a). When mapped onto on subunit of the dimeric laforin crystal structure (Figure 3.2b), they affect many regions with many clustered within the carbohydrate binding module (CBM) (Figure 3.2c). All mutations were generated by site-directed mutagenesis in a laforin construct lacking 3 amino acids at the C-terminus shown to reduce in vitro oligomerization (190, 453). The proteins were purified from E. coli via a two-step purification scheme using both affinity and size exclusion chromatography to very high purity (Figure 3.3). Structure-guided mutants were also generated and purified (Figure 3.3). Multiple attempts were made to purify F5S, Y86D, and R108C, affecting core residues of the CBM, and T187A affecting the DSP domain (Figure 3.3). However, these mutations consistently underwent significant proteolysis during the expression and/or purification process and were not further analyzed (Figure 3.3). We previously defined four regions of the laforin quaternary structure that possess different functional roles: the CBM, the CBM-DSP interface, the DSP domain, and the dimer interface (190). The LD mutants for this study were first divided into these four groups based on their location in the laforin structure for an initial assessment. We first studied the mutants by a thermal melt assay, i.e. differential scanning fluorimetry (DSF). In this assay melting temperature (Tm) is used to quantify protein stability. Additionally, since binding to ligands typically stabilizes a protein, the change in melting temperature (ΔTm) in the presence of a ligand can be used to quantify binding (506, 507). Wild-type (WT) laforin has a Tm of 49.6°C and incubation of laforin with short or long oligosaccharides (7 or 24 degrees of polymerization, i.e. DP7 or DP24) produced a ΔTm of 5°C or more (Figure 3.4a,b) (190). Incubation of WT laforin with saturating levels of DP7 and DP24 (10 mM) produced a ΔTm of 4.5 and 7.8°C, respectively (Figure 3.4b,c). The patient mutations W32G and K87T were somewhat destabilized with a Tm of 44-45°C (Figure 3.4a,b). These mutants produced minimal shifts of 1°C or less in the presence of glucans (Figure 3.4a,b). CBM20 domains utilize aromatic and hydrophilic residues to coordinate glucan binding (440). In the maltohexaose- bound crystal structure, laforin interacts with glucans via W32, K87, W99, and D107 (190). We and others have shown that W32G and K87A abolish or reduce the co-sedimentation of laforin with glycogen (190, 285, 508, 509). Additionally, it is well established that W32 and K87 are conserved residues of the CBM20 family and are directly involved in carbohydrate binding (190, 440). These results validate the use of DSF to assess binding to glucans. Mutations in core of the CBM (V7A, E28K, F84L, R91P) and at the CBM-DSP interface (E56K, I126T, Y294N, P301L) were highly destabilized, with a Tm of 40°C or less (Figure 3.4a,b). It must be noted that the melting profile of these mutants is quite different from WT laforin in that they possess two transitions, while other mutants had a WT-like transition. This is discussed in the next section (Section

3.2.3). The Tm values shown in Figure 3.4 correspond to the first transition. However, these mutations also produced a much greater ΔTm with DP7 and DP24, suggesting they actually bind more tightly to oligosaccharides than WT laforin (Figure 3.4b,c). Indeed, previous studies using affinity gel electrophoresis demonstrated that P301L and Y294N migrate more slowly than WT laforin through glycogen-rich polyacrylamide gels, indicative of a higher affinity for glycogen (460). The only CBM mutant with a similar stability and binding profile to WT was S25P (Figure 3.4a,b,c). This mutation is unique among the mutations in the CBM that we studied in that it does not have major effects on stability or glucan binding. Rather, it impacts a surface-exposed residue in the CBM that has been shown to be a phosphorylation site for AMPK (313). These data suggest a distinct disease mechanism for this mutation. Mutations in the DSP domain produced varying results. G279C, G279S, L310W, F321C and F321S, all affecting buried residues within the DSP, were all very destabilizing with a Tm of 36-41°C (Figure 3.4a,b). In contrast, other DSP mutations (A188G, K140N, N148Y, G240S, E210K, P211L, E224I, P246A) produced little to no destabilization with a Tm of 45-49°C. All of these mutations produced a binding shift with DP7 and DP24 that was similar to WT laforin with the exception of F321S and F321C (Figure 3.4b,c). Therefore, unlike the CBM, the DSP is less susceptible to destabilizing mutations, and with the exception of mutations at the dimer interface, mutations in the DSP domain do not affect carbohydrate binding. F321 is buried at the hydrophobic core of the DSP-DSP dimer interface (Figure 3.2c). We previously studied this mutant by analytical ultracentrifugation (AUC), and the data indicated that F321S is a monomer (190). Similarly, SEC coupled with multi-angle light scattering analysis of F321C also suggested that it is monomeric. (510). The lack of laforin dimerization results in impaired cooperative binding to long oligosaccharides such as DP24. We also determined the dimensions of laforin-F321S by small angle X-ray scattering and found that the size and shape parameters also correspond to a monomer (data not shown). Cysteine and serine mutations are likely to produce similar effects, and the DSF profiles of F321S and F321C are indeed nearly identical (Figure 4a,b,c). Both proteins are similarly destabilized and bind equally well to DP7 and DP24, reflecting their inability to cooperatively bind to long-chain glucans. To determine whether the melting temperature of the mutants correlated with the solvent exposure of the affected residues, we calculated solvent accessible surface area (SASA) using the GetArea program (Supplemental File 1) (http://curie.utmb.edu/getarea.html) (463). The ratio of side chain surface area to a value "random coil" determines whether the residue is solvent exposed (ratio > 50%) or buried (ratio < 20%). A significant positive correlation between Tm and SASA ratio was found (r = 0.4139), indicating mutations affecting buried residues produced a lower Tm (Figure 3.4d). However, no correlation was found between the experimentally determined Tm and the PolyPhen-2 score or ΔΔG values

67 predicted by CUPSAT and SDM (Figure 3.4e,f,g). We next tested the activity of the mutants using the biological substrate of laforin, glycogen, in a malachite green assay to measure phosphate release. Glycogen was purified from rabbit muscle as previously described (174). Covalently bound phosphate levels in the glycogen were determined by digesting the glycogen with perchloric and sulfuric acid, diluting in water, and then measuring phosphate using the malachite green assay. The levels of glycogen- bound phosphate were similar but slightly lower than reported in a precious preparation (Figure 3.5a) (174). However, these values were still in the linear range of detection by the malachite green assay. All of the phosphate could be released by treating glycogen with amyloglucosidase and Antarctic phosphatase (Figure 3.5a,b). Additionally, up to 96% of the phosphate could be released by treating glycogen with a combination of laforin, amylase and amyloglucosidase (Figure 3.5a,c). These results are in agreement with previous work (174). Glycogen dephosphorylation assays were performed in the linear range with respect to time, enzyme amount, and substrate concentration (Figure 3.5d). Given the low Tm of some laforin mutants, we tested the stability of the W32G, Y294N, P301L, and F321S mutants at both 25°C and 37°C. Laforin mutants with a low Tm, i.e. Y294N, P301L, and F321S, displayed precipitation after 30 minutes at 37°C, while these mutants were stable at 25°C (Figure 3.5e). Glycogen dephosphorylation assays were performed at 25˚ C to assess the phosphatase activity of these mutants. Consistent with previous studies of W32G and K87A, W32G and K87T had significantly impaired glycogen phosphatase activity due to a reduced affinity for glycogen (Figure 3.5f). Mutations in the CBM-DSP interface had variable effects on activity as seen for E56K, Y294N and P301L (Figure 3.5f). Surprisingly, the remaining 20 mutants produced a reduction in phosphatase activity of 50% or less. E28K, F84L, and K140N displayed no decrease in glycogen phosphatase activity. Additionally, no correlation between stability and phosphatase activity was observed (Figure 3.5g). Glycogen possesses phosphate bound to both C3- and C6-hydroxyls of glucose moieties, and increased phosphorylation at these sites is linked with aberrant polyglucan formation (210). Site-specific dephosphorylation assays were performed to define the ability of each mutant to dephosphorylate the C3- and C6- position. Arabidopsis starch free of phosphate was labeled with 33P at the C3- and C6-positions with starch dikinases, and the phosphorylated product was utilized as substrate in dephosphorylation assays. Laforin has previously been shown to exhibit a slight preference for C3 dephosphorylation when similar in vitro assays were performed (184). First, the kinetic parameters were defined for laforin WT and mutant proteins to determine the linear range of specific activity. Time course and substrate titration experiments were performed with both WT protein and the catalytically inactive mutant C/S to determine the appropriate assay parameters (Figure 3.6a,b). Kinetic studies were performed on both WT laforin and a subset of LD mutants.

W32G and F321S possessed a reduced Michaelis constant (Km) in assays with C3- and C6-labeled starch compared to WT laforin (Figure 3.6c,d,e). These data highlight how decreased affinity for the substrate results in reduced activity for W32G (Figure 3.5g). Conversely, the decreased activity for F321S is likely due to the lack of cooperative binding by F321S to substrates with long glucan chains (190). In contrast, P301L had a similar Km to WT but a reduced maximum reaction velocity (Vmax) (Figure 3.6c,d,e). This result indicates that the reduced activity of P301L is due to impaired catalytic turnover rather than reduced affinity for substrate. Other mutants were also analyzed for C3- and C6-specific dephosphorylation at a single substrate concentration (Figure 3.7a). P301L and the mutants with impaired affinity (W32G, K87T, and F321S) were the only mutants that produced a change in the C3 vs. C6 dephosphorylation ratio (Figure 3.7a). The change in activity for each mutant compared to WT laforin mirrored the change in glycogen dephosphorylation, with the exception of F321S due to its lack of cooperativity (Figure 3.7b). Therefore, most mutants do not change the preference of laforin for C3- vs. C6-phosphate. Glycogen contains C3 and C6 phosphate in approximately equal proportion (158, 210, 212). It is therefore unlikely that LD would result from a change in the ratio between C3 and C6 phosphate.

3.2.3 Mutation-induced decoupling of the CBM and DSP domains

In addition to defining the Tm, the melting profile from DSF experiments provides information regarding the structure and dynamics of multidomain proteins (511). Tightly integrated domains lead to cooperative thermal denaturation; in contrast, intermediate states during melting suggest discreetly folded domains. In other words, for a bimodular protein such as laforin, a single melt transition reflects tight integration between the domains, while the appearance of two peaks indicates the domains are decoupled and melt independently (512). WT laforin melts with a single sharp peak, reflecting tight coupling between the CBM and DSP (Figure 3.8). The shape of this peak shifts to the right but does not change with binding to DP7 and DP24 (Figure 3.8 and 5.9a,c). This result is consistent with the extensive hydrophobic CBM-DSP interface observed in the crystal structure (~1200 Å2) and the minimal solvent accessibility of residues in this interface observed in hydrogen deuterium exchange experiments (190, 205). Most laforin mutants also produced a melting profile with a single sharp transition, and the shape of the peak did not change with the shift induced by carbohydrate binding (Figure 3.8). Interestingly, mutations affecting the CBM core or CBM-DSP interface produced melting profiles that were either broader than WT (e.g. V7A, E56K, F84L) or had two distinct peaks (e.g. E28K, R91P, Y294N, P301L) (Figure 3.8). These data indicate that mutations in the core of the CBM and CBM-DSP interface induce domain decoupling. We previously observed increased deuteration of peptides at the CBM-DSP interface in Y294N compared to WT

69 laforin (190). The DSF results are perfectly consistent with those observations. However, no differences in the size and shape parameters of Y294N compared to WT laforin via small angle X-ray scattering were observed, indicating this is not a global effect (data not shown). These data indicate that the CBM and CBM-DSP is sensitive to stability perturbations and many LD mutations induce local domain decoupling. E210K also had an unusual melting profile, but this pattern was unique among the mutations that we studied. The decoupling mutations led to a broadening or separation of peaks during the denaturation phase (i.e. when fluorescence was increasing). In contrast, for E210K, a delay in the quenching phase (when fluorescence was decreasing) was observed. Quenching occurs at high temperatures when the denatured protein aggregates, which reduces the surface area interacting with dye and quenching its fluorescence. These data indicate that E210K affects the aggregation of laforin post-denaturation, which may or may not have biological relevance. Strikingly, these mutants melted with a single transition like WT laforin in the presence of 10 mM DP7 or DP24 (Figure 3.8 and 3.9a). The only exception was P301L, which even in the presence of DP7 still showed a profile suggesting two peaks (Figure 3.8). The data also suggest that decoupling is reversed by binding to a carbohydrate. When a lower concentration of DP24 was used, Y294N displayed an intermediate profile with a broad peak, and two peaks were still observed for P301L (Figure 3.9a). Interestingly, the peaks of the P301L profile were so distinct that each Tm could be determined: 31.8 and 52.5°C, likely corresponding to the decoupled domains (Figure 3.9b). The Tm of P301L in the presence of DP24 is between these two values, reflecting a distinct conformation that is substrate-bound. These data strongly suggest that binding to a sugar tightly integrates the two domains, rather than stabilizing one of the domains (such as the CBM) so much that it is masked by the peak of the other. The Y294N mutation produces an unambiguous two-state melting profile (Figure 3.8). We found that a Y294N yields from E. coli were generally similar to those of WT laforin. In contrast, more highly decoupled mutations (e.g. E28K, R91P and P301L) underwent significant proteolysis during expression and purification and yielded much less soluble protein. Since a higher quantity of this protein could be produced, which is necessary for DSF titration experiments, Y294N was selected for further study. To define how the DSF profile shifted with increasing concentrations of glycogen, a glycogen titration was performed with WT laforin and Y294N. While WT laforin exhibited a rightward shift with no change in peak shape, Y294N exhibited a gradual transition from two-state melting with low concentrations of glycogen to one-state melting with higher concentrations (Figure 3.9c). The ΔTm values from this experiment were used to graph a binding curve for WT and Y294N. Interestingly, a two-site binding equation provided a better fit for both WT and Y294N binding curves better than a one-site equation (Figure 3.9d). These data suggest that both WT and Y294N have two binding

70 sites with different affinities. We previously observed that the laforin dimer binds cooperatively to DP24, which implies two binding sites with interdependent affinities (190). Since cooperativity requires dimerization, the two binding sites likely correspond to the two CBM domains. Two-site binding to glycogen, a substrate with shorter linear chains than DP24, reflects two binding sites with different affinities that are independent rather than cooperative, It is also interesting to note that for Y294N, the second binding event occurs after the melting curves have shifted full to a single peak (Figure 3.8d, enlarged circle). This further suggests a binding at one CBM initiates allosteric conformational changes that propagate through the CBM-DSP and dimer interfaces to the other subunit. Binding to the second CBM is not achieved until the CBM and DSP are fully integrated and produce a single melt peak. In the crystal structure of laforin, the ends of the maltohexaose (DP6) molecules bound to the CBM and DSP domains are ~24 Å apart. This distance could be bridged by 4-5 additional glucose moieties (each unit is about ~5.5 Å long). There is continuous directionality the two sites so a glucan of least 16-17 units could potentially tether the two domains. A DP7 molecule would only be long enough to bind to one site, and it could not physically tether the two domains. Given these distances, an allosteric conformational change must be induced by DP7 binding that unites the domains in the decoupled mutants. Indeed, an alanine mutant of W32 surprisingly produced two-state melting, but unlike all other mutants with this type of profile, it did not shift to one state in the presence of sugars (Figure 3.10a). These results suggest that W32 is facilitating that conformational change via its interaction of substrate at the CBM glucan binding site. Y294 and P301 contribute 159 and 100 Å2, respectively, to the CBM-DSP interface and together comprise nearly 20% of the total interface surface area. It is not surprising then that mutation of these well-conserved and integral residues to very different amino acids would be detrimental to domain coupling (Figure 3.10b). Alanine mutants of P301 and Y294 also produced broad melting profiles, but they were not as exacerbated as the LD patient mutations (Figure 3.10c). In contrast, E56 contributes only 43 Å2 to the interface. E56K exhibits a slightly broadened peak indicative of mild decoupling, while E56A did not display a broadened peak. Although alanine mutants of these CBM-DSP interfacing residues still mildly destabilized laforin (Supplemental File 2), it is likely that these mutations would be less pathogenic since they are less structurally disruptive than replacing an acidic residue with a basic one (E56K), exchanging an aromatic with a hydrophilic residue (Y294N), or a substituting proline with a bulkier hydrophobic (P301L). Y294 makes multiple hydrogen bonds to surrounding residues, particularly with backbone hydroxyls of R4 and L124. These bonds are immediately adjacent to the hydrophobic CBM core and LD mutation "hotspot" containing LD-associated residues F5, V7, Y86, and F84 (Figure 3.10b). Hydrophilic residues affected by LD mutations, E28, R91 and R108, are also in this region. This hotspot appears

71 to bridge Y294 and the DSP domain with W32 of the CBM. Y294 mutated to a phenylalanine produced very mild decoupling and destabilization (Figure 3.10c). These data indicate that both the hydrophobic nature and hydrogen bonding capacity of Y294 are important for tethering between the DSP domain and CBM. Y294F abolishes this bond but still contains the aromatic benzene ring, therefore producing the mildest effect. Y294A lacks both aromatic and hydrogen bonding capacity and is partially decoupled. The LD mutation Y294N introduces a hydrophilic residue, inducing further structural perturbation and producing the most severe domain decoupling effect. In the laforin crystal structure, oligosaccharides are bound to both the CBM and DSP domain. However, data thus far suggests that only CBM residues associating with glucan contribute significantly to binding: W32, K87 and W99 (Figure 3.10c) (190). The individual laforin CBM and DSP domains were purified by affinity chromatography and size exclusion chromatography (Figure 3.11a,b). The DSP domain was expressed as a SUMO fusion protein to enhance stability. We were unable to express the DSP domain in the absence of the SUMO tag, and attempts to cleave the tag caused the DSP to become sticky and prone to aggregation. Neither domain had phosphatase activity, consistent with previous reports (data not shown) (454). The CBM Tm was 38.2˚ C, a lower Tm than full-length laforin. CBMs are flexible domains so they can coordinate ligand binding and presentation to an active site, and we have shown that CBM of laforin is more dynamic than the DSP domain (190, 440). These characteristics may explain the reduced stability of the laforin CBM. The SUMO-DSP fusion protein yielded a Tm of 51.7°C. While the CBM bound to both DP7 and DP24, SUMO-DSP exhibited no shift in Tm (Figure 3.11c). When these proteins were incubated with increasing concentrations of DP7, the CBM exhibited a similar binding curve to full-length laforin, while the DSP domain did not shift even at high concentrations of DP7 (Figure 3.11d). Interestingly, full-length laforin with the W32G mutation produced a right-shifted binding profile (Figure 3.11e). These data indicate that laforin binding to carbohydrates does rely on the CBM. Although W32 seems to be the primary residue involved in binding, even at high concentrations of substrate, other residues of the CBM can compensate.

3.2.4 Effects of LD mutations on protein-protein interactions

Thus far, the assays that we have utilized have revealed defects in some, but not all, of the LD-causing mutations. Mutations reducing or abolishing carbohydrate binding (W32G and K87T) have severe effects on glycogen phosphatase activity. Highly destabilizing mutations, i.e. having a Tm < 37°C, are likely too unstable in vivo to function, prone to unfolding and aggregation, and are likely degraded. Indeed, expression of Y294N in HeLa cells, which are cultured at 37°C, led to the formation of ubiquitin-positive aggregates, suggesting it was unfolded and targeting for degradation (513). Most of the mutations affecting the

CBM core or the CBM-DSP interface would fall into this category. However, mutations with a Tm > 37°C could still be functional in vivo. For example, the F321S and F321C mutations both have a Tm at or just above 37°C, and LD patient data strongly suggest that F321C is at least partially functional since a homozygous patient with F321C displayed a very mild LD phenotype (489). In contrast, W32G and K87T are only mildly destabilized, but these mutations severely impair carbohydrate binding (Figure 3.4b,c) and are found in patients with a typical LD progression, suggesting they are highly pathogenic (Table 3.1). Therefore, pathogenicity probably does not necessarily correlate with stability, unless the Tm of the mutant is below 37°C. Furthermore, the majority of the mutants analyzed yielded only mild defects in protein stability or phosphatase activity. Since laforin is known to interact with multiple proteins involved in glycogen metabolism, we next assessed the impact of the mutants on the association of laforin with its binding partners. Malin and Protein Targeting to Glycogen (PTG) are perhaps the best characterized laforin binding partners (175, 176, 301, 307, 313, 514). PTG is a regulatory subunit of protein phosphatase 1 (PP1) and contains a CBM21 that targets PP1 to glycogen. PTG indirectly promotes glycogen synthesis through PP1-mediated dephosphorylation of glycogen synthase (which is activating) and phosphorylase (which is inhibitory) (300, 439, 441). Malin and PTG are both difficult proteins to work with in vitro because they are largely insoluble when expressed as recombinant proteins (175, 301). To circumvent this issue, we employed a directed yeast two-hybrid analysis that has been used extensively to define the interaction of laforin with malin, PTG, and other proteins (175, 176, 313, 455, 515). Laforin mutations were generated in the bait yeast expression vector pEG202 and expressed as laforin fusion proteins with the DNA-binding protein LexA. Wildtype malin and PTG were expressed as prey fused with the Gal4 activation domain (GAD). The interaction of bait and prey proteins results in transcription of β-galactosidase (i.e. β-gal), so β-gal activity is used to quantify interactions. We also tested laforin mutations that we were unable to express in E. coli: F5S, Y86D, T187A and R108C. These proteins were all expressed at similar levels in yeast (Figure 3.12). The majority of laforin mutations reduced or abolished both PTG and malin interaction with laforin (Figure 3.13a). When co-expressed with GAD alone, only K87T, which ablates glycogen binding, yielded a significant interaction (Figure 3.13a). The other fusion proteins did not interact with this negative control protein (Figure 3.13a). The K87T result could be due to the finding that mutations affecting glycogen binding can localize to the nucleus in cell culture (421). In yeast, K87T may translocate to the nucleus and interact with GAD, promoting some auto- activation. However, this was not observed for any other mutants. Interestingly, mutations at the CBM glucan binding site, W32G, and K87T, reduced PTG interaction, while not affecting the interaction with malin (Figure 3.13a). In contrast, mutations at the dimer interface, F321S and F321C, specifically

73 abolished the malin interaction, and the PTG interaction was either unchanged or slightly increased (Figure 3.13a). Mutations that destabilize laforin (i.e. decrease its Tm) reduced the interaction of laforin with both binding partners despite normal expression levels of the mutants (Figure 3.12). Surprisingly, E210K, P211L, E224I, K140N, G240S, and P246A showed no defect in PTG or malin binding. Conversely, S25P did not interact with either binding partner. The effects of these mutations on multiple aspects of laforin have now been defined: stability, activity, and interaction with established binding partners. Based on these results and the location of the mutants in the crystal structure, the mutants were further subdivided into 5 groups. W32G and K87T affect residues at the glucan binding site and are the only mutations that lead to a complete loss of carbohydrate binding. F5S, V7A, E28K, F84L, Y86D, R91P and R018C clustered in the core of the CBM, and lead to domain decoupling and a reduction in the laforin-malin and laforin-PTG interactions. E56K, Y294N, and P301L affect the CBM-DSP interface and have similar effects as mutations at the CBM core. N148Y, T187A, A188G, G279S/C, and L310W affecting different regions of the DSP domain. These mutations do not affect phosphatase activity but are somewhat destabilizing and reduce the laforin- malin and laforin-PTG interaction. E210K, P211L and E224I clustering in a surface-exposed region of the DSP. They are very mildly destabilized and have little effect on the laforin-malin and laforin PTG interaction. F321S/C affects the dimer interface, destabilizing the protein and abolishing dimerization, and leading to a loss of malin binding only. We were not able to detect a defect in K140N, P246A and G240S using our assays. S25P displayed no interaction with malin or PTG, but the reason for this is not clear since this mutation is only very mildly destabilized. We performed pairwise correlation analyses between Tm, PTG interaction and malin interaction to determine whether these aspects of laforin function were interdependent. Importantly, the mutations clustered by group in these correlation graphs. A weak correlation between PTG interaction and Tm was observed, but it was not statistically significant (Figure 3.13c). A positive correlation was found between the malin interaction and Tm (Figure 3.13b). A very strong correlation was found between PTG and malin interaction (Figure 3.13d). These data indicate the primary defect of most LD mutations is a decreased interaction with malin and/or PTG. Binding to malin and PTG is tightly dependent on laforin stability, which is altered by many LD mutations. Furthermore, the effects of each group of mutations has a certain pattern of functional effects, leading to defined clusters in these graphs. The strong correlations between these functional parameters among LD mutations suggest they are the most relevant to LD pathogenesis.

3.2.5 Laforin binding sites for malin and PTG are disparate

It was striking that mutations at the CBM specifically affected the PTG interaction while mutations at the dimer interface specifically affected the

74 interaction of laforin with malin. More centrally located mutations affected both interactions. These data suggest that the binding sites for PTG and malin are spatially distinct. This arrangement seems plausible since PTG has been shown to be a substrate for malin in a laforin-dependent fashion, and therefore both PTG and malin are likely form a complex with laforin (188, 301). We tested this hypothesis by performing yeast triple hybrid experiments in which laforin was expressed in the presence of both binding partners. One partner was expressed as prey, and fused to GAD, while the other was expressed with only an HA tag. If the two binding partners had overlapping binding sites, the HA-tagged partner would displace the first and result in reduced β-gal activity. No change in β-gal activity was observed when HA-tagged PTG was expressed in addition to LexA- laforin and GAD-malin (Figure 3.14a). To account for the possibility that the laforin-malin interaction was too strong for malin to be displaced by PTG, we also performed the reverse experiment in which HA-tagged malin was expressed in addition to LexA-laforin and GAD-PTG. Again, no change was observed in β-gal activity. These data strongly suggest that PTG and malin have independent binding sites on laforin, and these binding sites are differentially affected by LD missense mutations. Mapping of electrostatic potential on the surface of the laforin crystal structure illustrates that the DSP active site is a highly positively charged pocket, ideal for binding to a negatively charged phospho-glucan (Figure 3.14b). The remaining surface of the molecule is fairly neutral. When rotated 180˚ along the longest axis, there are three discrete patches possessing charge: a positively charged region at each CBM-DSP interface and a patch of negatively charged residues comprised of the dimer interface. Each positively charged pocket corresponds to a CBM-DSP interface, particularly the region containing E56, Y294, and P301 (Figure 3.14c). F321 residues are located beneath either edge of the negatively charged patch (Figure 3.14c). The negatively charged patch is circular, approximately 30 Å in diameter. Malin possesses six NHL repeats that form a putative β-propeller fold (306). This type of fold is approximately 50 Å at its widest point (516). These β-propeller domains form a substrate-interacting domain, and we have shown that the NHL repeats are essential for the interaction of malin with laforin (175). A homology model of malin using Phyre2 shows the β-propeller fold of the NHL repeats and the location of D146, a known residue involved in the laforin-malin interaction, on the surface of this substrate-interacting domain (Figure 3.14d). Electrostatic mapping onto this homology model shows that the interior of the NHL β-propeller is highly negatively charged (Figure 3.14e). This negative charge may be suited to interact with the target lysine intended for ubiquitination. At the edges of the domain are regions of positive charge. It is possible that these may be involved in laforin binding. E3 ligases have been shown to dimerize (517, 518). Also, yeast two-hybrid assays and co-immunoprecipitation experiments in mammalian cells have shown that malin interacts with itself (519). Malin may also dimerize, and like

75 laforin, an interaction surface may form from the interaction between two malin subunits. PTG is a 36 kDa protein containing a PP1-binding motif at its N-terminus and a CBM21 spanning amino acids 149-257. An earlier study showed that a region encompassing the CBM21 (amino acids 141-263) is necessary for the interaction of PTG with laforin (514). Therefore, it is likely that laforin and PTG interact via their CBMs. Binding of both CBMs to a carbohydrate substrate would likely enhance their interaction by proximity effects. A homology model of PTG was also generated using Phyre2 (Figure 3.14f). Based on this information and our data, it is likely laforin acts as a glycogen-associated scaffold for PTG, malin, and other binding partners involved in glycogen metabolism (Figure 3.14g). These data suggest that a crucial detrimental effect of LD mutations is negatively impacting the interaction of laforin with malin and PTG and perhaps other proteins involved in glycogen metabolism.

3.3 Discussion

In the present study, we defined the molecular defects of LD-associated EPM2A missense mutations using biochemical tools. Genotype-phenotype correlations in the LD population are extremely difficult due to a small number of patients, large number of mutations, under diagnosis, and limited clinical data. Furthermore, computer algorithms do not yield satisfactory or consistent predictions regarding the pathogenicity of missense mutations. Therefore, we employed a reductionist biochemical approach to circumvent these limitations and identify molecular-level defects to provide a framework for evaluating the pathogenicity of LD mutations and predict their clinical course. These data provide a biochemical explanation for mild LD cases and provides insights into the physiological and etiological roles of laforin.

3.3.1 Functional classes of EPM2A missense mutations

Over the past three decades, the cystic fibrosis (CF) community has uncovered the molecular basis of CF and defined multiple classes of CFTR mutations through detailed basic biological and mechanistic studies (500, 502). These extensive investigations have led to mutation-specific therapeutics. The LD patient and research community are significantly smaller than the CF community and the reduced sample size poses additional challenges. Furthermore, the genetic basis of LD was only defined in 1998 with the identification of the laforin gene and 2003 with the identification of the malin gene (170-172, 328). This biochemical analysis of LD missense mutations was largely facilitated by the laforin crystal structure that was only determined in 2015 (190). Using the biochemical and biophysical pipeline that we developed, we identified five distinct classes of laforin mutations and propose a structure-guided, molecular explanation for why mutations in each class cause disease. 76

Class I mutations affect carbohydrate binding of the CBM (Table 3.3). We have shown that the CBM is essential for carbohydrate binding and that W32 and K87 each contribute significantly to binding (Figure 3.4 and 3.11) (190). These mutations are only slightly destabilizing to the laforin structure, but their phosphatase activity is greatly reduced (Figure 3.4, 3.5f, 3.7). Furthermore, these mutations abolished the laforin interaction with PTG (Figure 3.13). This defect may be partly due to their impaired association with the glycogen molecule, with which PTG is also associated. Although both mutant proteins are still able to bind to malin, they are likely not able to properly target malin to its glycogen-associated substrates. A patient homozygous for the W32G mutation was 13 years old at disease onset, typical of classic LD (Table 3.1); no other information about this patient was provided in the case report (520). A compound heterozygous patient carrying K87T and E56K mutations was 11 years old at seizure onset and 15 years old when cognitive decline began. These clinical features are characteristic of what is considered a classic LD course (479). Based on our biochemical results and these clinical data, it is reasonable to conclude that these mutations are highly detrimental to laforin function and lead to a very severe phenotype with rapid progression (Table 3.3). Class IIa mutations affect the CBM core and class IIb mutations affect the CBM- DSP interface (Table 3.3, Figure 3.2c). Although structurally distinct, the mutations produce similar biochemical defects: they dramatically destabilize laforin and impair or abolish interactions with both PTG and malin (Figure 3.4, 3.8, 3.9, 3.13). Class II mutations also appear to associate more tightly with carbohydrates than WT laforin (Figure 3.4) (460). Most of the mutations also have little or no effect on phosphatase activity, with the exception of P301L (Figure 3.5f, 3.7). Patients either homozygous or heterozygous for class II mutations also appear to have a classic clinical course based on available clinical data. Seizure onset is typically between the ages of 11 and 14 years old (Table 3.1). One exception is a family of 3 children all homozygous for R108C that experienced seizure onset at ages 5, 8 and 10 years old (513). These patients also had early onset learning disability that manifested at the age of 4 years old. Another patient homozygous for Y86D displayed early onset learning problems, although his brother did not (488). Both Y86D and R108C completely abolish interactions with PTG and malin. We were unable to purify these recombinant mutants, likely due to their severe instability. The major destabilization and impaired interactions of class II mutations likely lead to a very rapid disease progression, possibly even more rapid than class I mutations. Class III mutations affect the DSP domain (Table 3.3) and have variable effects on laforin stability and PTG/malin interaction (Figure 3.4, 4.13). All class IIII mutations have mild or moderate effects on phosphatase activity and they do not affect carbohydrate binding (Figure 3.4, 3.5f, 3.7). Patients carrying T187A, N148Y, and L310W mutations experienced seizure onset between the ages of 10 and 13 years old. These mutations also significantly impair the interaction with

PTG and malin (Figure 3.13). The clinical features of patients carrying the A188G mutation were not reported (477). Interestingly, G279C and G279S mutations have both been associated with a slower clinical course (488) (Jose Serratosa, personal communication). A compound heterozygous patient, carrying the G279C mutation and a common PTC (R241X) did not present with seizures until the age of 21 years old, and cognitive impairment did not develop until the patient was 24 years old. Since these mutations retain some interaction with malin and PTG, they may delay LD onset. Therefore, class III mutations are associated with variable phenotypes. We recently published a study on a compound heterozygous patient carrying the missense mutation N163D and a PTC (Y112X) (456). This patient had her first seizure at 16 but was still able to walk, still very cognitively engaged at the age of 28. We showed that N163D had no effect on laforin stability, carbohydrate binding, or activity, but impaired the interaction of laforin with its binding partners. N163 is in the DSP domain and the biochemical profile of the N163D mutation is consistent with other mutations and cases in class III (Table 3.3). Therefore, mutations in class III can produce both moderate and severe phenotypes, depending on how severely the malin and PTG interactions are affected. Class IV mutations affect a surface-exposed region of the DSP domain adjacent to the dimer interface (Table 3.3, Figure 3.2c). These mutations display no major defects in laforin stability, carbohydrate binding, phosphatase activity, or interaction with PTG or malin (Figure 3.4, 3.5f, 3.7, 3.13). This surface may be important for binding to another laforin substrate, such as AMP-activated protein kinase (AMPK), glycogen synthase, or other regulatory subunits of PP1 such as R6 (208, 300, 302, 303). A compound heterozygous patient carrying the E210K displayed seizure onset at the age of 17, which is later than classic LD. Another patient heterozygous for P211L experienced seizure onset at the age of 11 and lived to the age of 28, longer than the typical disease course (unpublished case). Therefore, it is likely that class IV mutations yield a more moderate effect on laforin function with a more moderate disease progression. Class V mutations affect laforin dimerization. Thus far we have only identified one residue affected by class V mutations, F321 (Figure 5.2c). F321 mutations reduce laforin stability, render laforin monomeric, and abolish the preference of laforin for binding long oligosaccharides (Figure 5.4) (190). They do not affect the interaction of laforin with PTG but significantly reduce the malin interaction (Figure 3.13). The origin of the F321S mutation is unknown; however, the F321C mutation is associated with a homozygous case of extremely mild LD (489). This case is comparable in clinical course only with LD patients homozygous for the malin D146N mutation. Disease onset for these patients is variable, reported between the ages of 13 and 29 years old, and the patients can live 10-20 years or more after onset, often into their 30s and 40s (483, 484, 486, 487, 521). One patient homozygous for D146N had his first seizures at 30 and lived beyond the age of 48 (522). Both laforin F321C and D146N have been shown to affect the laforin-

78 malin interaction while not affecting the other functions (Table 3.3) (300). F321C and F321S may also be mildly pathogenic due to their impaired ability to specifically target LB-like structures. We were unable to classify the mutants S25P, K140N, P246A and G240S because they did not affect the biochemical or biophysical properties of laforin that we measured (Table 3.3). All of these mutants had little to no effect on laforin stability, phosphatase activity, and carbohydrate binding (Figure 3.4, 3.5f, 3.7). Of these mutants, only S25P had impaired interaction with malin and PTG, but the reason for this is not clear. Laforin has been shown to be phosphorylated at S25 by AMPK, and phosphorylation has been shown to enhance the laforin-malin interaction via yeast two-hybrid (313, 455). It is possible that laforin must be phosphorylated in yeast in order to interact with malin in vivo (and perhaps also PTG). All of the mutants except S25P may be phosphorylated by Snf1 (the yeast homolog of AMPK) when expressed in yeast, facilitating an interaction with PTG and malin. Therefore, S25P may be pathogenic by preventing the phosphorylation of laforin by AMPK. K140N may also affect a post-translational modification site. Laforin has been shown by many groups to be ubiquitinated by malin (175, 176, 307). Although the specific lysine(s) that are ubiquitinated have not been identified, only 3 of the 11 lysine residues in laforin are surface exposed, according to SASA analysis of the crystal structure: K140, K219, and K323 (Supplemental File 1). It is possible that K140 is a primary ubiquitination site and that mutation of this residue leads to LD by impairing laforin ubiquitination. Our assays did not detect a defect in P246A and G240S. Both are buried in the core of the DSP between the active site and dimer interface (Figure 3.2c). Further study will be required to determine why these mutations lead to disease.

3.3.2 Laforin stability and interaction, not activity, are the primary defects of EPM2A missense mutations

It has been hypothesized that most genetic diseases are caused by loss-of- function resulting from destabilizing mutations (491). Herein we show that many LD-causing missense mutations are indeed destabilizing. The core of the laforin CBM and the CBM-DSP interface are the regions most susceptible to destabilizing mutations. At physiological temperatures, these mutations would likely lead to unfolding, aggregation, and entire ablation of protein function. Indeed, Y294N has been shown to form ubiquitin-positive aggregates when overexpressed in cell culture, although other mutants did not (R108C, E28L) (513). Like Y294N, L310W also forms aggregates in cell culture (523). However, pathogenicity of a mutation cannot be determined simply by its stability. Many LD-causing mutations have normal or nearly normal melting temperatures, such as S25P, K140N, N148Y, A188G, E210K, P211L, E224I, G240S, and P246A. Furthermore, some destabilizing mutations are associated with a milder phenotype, such as G279C and F321C.

It is very striking that most missense mutations have intact phosphatase activity at the permissible temperature (Figure 3.5f, 3.7). The only mutations we found with reduced glycogen phosphatase activity were W32G and K87T, which clearly have impaired binding to carbohydrate substrates. Furthermore, few LD mutations are found near the laforin active site, and none affect catalytic residues or are found in the active site loop (The Lafora Progressive Myoclonus Epilepsy Mutation and Polymorphism Database, http://projects.tcag.ca/lafora/). Recently, the relevance of laforin phosphatase activity has been questioned due to the observation that overexpression of the catalytically inactive laforin mutant C266S rescues the LD phenotype in mouse models (159, 189). Strikingly, C266S has possesses enhanced interaction with laforin binding partners compared to WT laforin in yeast two-hybrid studies (455, 514, 524). This result is consistent with our observation that most LD mutations do not affect laforin phosphatase activity but do negatively affect protein-protein interactions.

3.3.3 A slower progression in LD and genetic effects

LD was previously believed to produce a severe disease course irrespective of the mutation of the patient (477, 520). Before the genetic loci were identified, if patients presented with progressive myoclonic epilepsy and lived beyond the age of 30, a diagnosis of LD was ruled out (326). However, differences in clinical presentation have always been observed (322, 326). Even at present, adult patients with mild progressive myoclonus epilepsy typically do not undergo LD screening (484). With the increasing use of genetic testing to confirm LD diagnosis, only recently has the neurology community begun to accept that LD is a heterogeneous disease (328, 484). Modifier effects are certain to play a role in clinical manifestation. These effects could be either genetic or environmental. Patients carrying the same mutations do show phenotypic variability (525). Siblings carrying identical mutations sometimes display differences in disease progression, suggesting a role for genetic factors (481, 488, 513, 526). Furthermore, a PTG variant has been reported to contribute to a milder disease course (457). Differences in medical care can also influence clinical progression (520). However, a clinically homogeneous patient progression is often reported in families and among genetic isolates (487, 527). Another complicating factor for predicting disease progression is that many patients are compound heterozygotes, i.e. carrying two different chromosomal aberrations that affect the same gene. LD is a recessive disorder, meaning patients carrying only one mutated allele do not display any symptoms. It is likely that a disease threshold exists regarding decreased laforin function (Figure 3.15). Our data suggests that laforin function is defined as its ability to function as a protein scaffold on the glycogen molecule. When laforin function is above a specific threshold then individuals are completely healthy. Below this threshold, there is a gradation of how rapidly patients progress down a clinical path. Both the

80 type of mutation and whether the mutation is in the homozygous or heterozygous state may influence disease progression. For example, the mildest form of LD associated with an EPM2A mutation is a patient homozygous for F321C, a Class V mutation (489). Based on our analyses, this patient would have laforin function just below the disease threshold (Figure 3.15). In contrast, compound heterozygous patients carrying a class III or IV mutation in addition to a deleterious mutation such as a PTC or indel would experience a clinical course that is milder than classic LD and yet still more rapid than patients with class V mutations (456, 488). The most severe and rapidly progressing LD cases are homozygous or compound heterozygous patients with class I or II mutations leading to nonfunctional protein (Figure 3.15). Patients with these mutations are likely indistinguishable from patients with PTCs and/or indels.

3.3.4 A biochemical pipeline for personalized medicine

Genetic testing is being used with increasing frequency to detect or confirm a LD diagnosis. It is likely that in the near future, LD will be detected earlier, and that even more cases of mild or late-onset cases will be identified. A biochemical analysis of novel mutations using our biochemical and biophysical methods could be performed in a matter of weeks and would provide the most accurate prediction of mutation pathogenicity. New mutations would be classified based on the similarity of their biochemical profile with known mutations (Figure 3.16). Additionally, preclinical studies of LD therapeutics are currently underway (328). With the large number of novel mutations arising in new patients, such a biochemical pipeline could become extremely useful for providing patients with a personalized diagnosis and treatment strategy once therapies become available. Furthermore, genetic screening during pregnancy or at birth is already widely employed to identify genetic diseases and chromosomal abnormalities (528). Life- threatening disorders, even those as rare as LD, may soon be added to these early genetic screens. Pre-symptomatic detection of LD would be extremely valuable since early treatment could better ameliorate or even prevent LD onset. However, benign SNPs would also need to be correctly differentiated from mild, moderate or severely pathogenic mutations to prevent a false diagnosis and/or unnecessary treatment, which could be needlessly devastating, expensive, or even harmful for individuals.

Table 3.1 Clinical data for patients carrying EPM2A missense mutations analyzed in this study. Each line refers to a single patient unless noted below. Some patients have been cited multiple times. Asterisk(s) indicates this mutation has been explicitly associated with a milder phenotype, either in the compound heterozygous (*) or homozygous (**) state. a 2 cases in same family. b 3 cases in different families. Age of seizure Mutation Allelic context Reference onset F5S Unknown Unknown Unpublished case. V7A Hom. 17 (522) S25P Hom. 13 (513) E28K Comp. het. (R171H / E28K) 14 (520) W32G Hom. 13 (520) E56K Comp. het. (E56K/K87T) 11 (479) F84L Hom. Unknown (529) Y86D Hom.a 10,14 (488) K87T Comp. het. (E56K/K87T) 11 (479) R91P Unknown Unknown (530) Hom. 11 (520) R108C Hom. 5, 8,10 (513) K140N Comp. het. (L310W / K140N)a 13 (523) N148Y Hom. 11 (523) T187A Comp. het. (T187A / G75fs) 10 (531) A188G Unknown Unknown (477) E210K Comp. het. (E210K / unknown) 17 (523) P211L* Comp. het (P211L / unknown) 11 Unpublished case. E224I Unknown Unknown (513) G240S Comp. het. (G240S/H253fs) Unknown (529) P246A Unknown Unknown (477) G279C* Comp. het (G279C / R241X) 21-28 (488) comp. het. (G279S/R241X)b Unknown (172, 520, 529) G279S comp. het. (G279S/Ex2-del2) 14 (529) Comp. het. Y294N Unknown (172, 529) (Y294N / Ex1-32bpdel 2) P301L Comp. het. (P301L / unknown) Unknown (529) L310W Comp. het. (L310W / K140N)a 13 (523) F321C** Hom. Early 20s (489) F321S Unknown Unknown Unknown

Table 3.2 Predicted pathogenicity of selected EPM2A missense mutations. Asterisk(s) indicates this mutation has been explicitly associated with a milder phenotype, either in the compound heterozygous (*) or homozygous (**) state. CUPSAT and SDM cutoff scores: stabilizing/neutral >=0 (green); destabilizing, between -2 and 0 (orange); destabilizing <-2 (red). For CUPSAT results, ΔΔG was predicted for residues in both subunits (chain A and C) and the values in the table represent an average.

PolyPhen-2 CUPSAT SDM Prob. ΔΔG Predicted ΔΔG prediction Outcome Score (kcal/mol) Stability (kcal/mol) F5S 0.198 benign -4.045 Destabilizing -2.96 Reduced stability V7A 0.092 Benign -6.71 Destabilizing -2.71 Reduced stability S25P 1 probably damaging 0.56 Stabilizing -0.78 Reduced stability E28K 1 probably damaging -0.76 Destabilizing -1.09 Reduced stability W32G 1 probably damaging -0.595 Destabilizing 2.76 Increased stability E56K 1 probably damaging 2.605 Stabilizing -0.87 Reduced stability F84L 0.003 Benign -2.705 Destabilizing -1.89 Reduced stability Y86D 0.863 possibly damaging -1.68 Destabilizing -2.77 Reduced stability K87T 1 probably damaging -0.26 Destabilizing -0.63 Reduced stability R91P 0.988 probably damaging -0.145 Destabilizing -1.26 Reduced stability R108C 1 probably damaging 1.31 Stabilizing -0.28 Reduced stability K140N 0.915 possibly damaging -1.975 Destabilizing -0.86 Reduced stability N148Y 0.085 Benign 0 No effect 1.21 Increased stability T187A 0.999 probably damaging -1.06 Destabilizing 0.06 Increased stability A188G 1 probably damaging -5.965 Destabilizing -0.92 Reduced stability E210K 0.381 Benign 0.28 Stabilizing -0.04 Reduced stability P211L* 0.867 possibly damaging -1.74 Destabilizing -0.58 Reduced stability E224I 0.812 possibly damaging 0.585 Stabilizing -0.57 Reduced stability G240S 0.815 possibly damaging 1.11 Stabilizing -0.01 Reduced stability P246A 0.985 probably damaging -2.71 Destabilizing 2.19 Increased stability G279C* 1 probably damaging -1.76 Destabilizing -0.01 Reduced stability G279S 1 probably damaging -1.73 Destabilizing -2.32 Reduced stability Y294N 0.999 probably damaging 1.44 Stabilizing -2.3 Reduced stability P301L 1 probably damaging -4.335 Destabilizing 1.26 Increased stability L310W 1 probably damaging -2.275 Destabilizing -1.24 Reduced stability F321C** 1 probably damaging -2.5 Destabilizing -0.19 Reduced stability F321S 1 probably damaging -3.885 Destabilizing 0.38 Increased stability

Table 3.3 Classification of LD-causing EPM2A missense mutations based on biochemical and clinical data. Asterisk(s) indicates this mutation has been associated with a milder phenotype, either in the compound heterozygous (*) or homozygous (**) state. Mutants in parentheses are likely classifications but were not fully characterized due to extensive proteolysis.

Region Predicted Mutations General Effects affected Pathogenicity Severely impaired CBM W32G carbohydrate binding and Class I carbohydrate- Severe phosphatase activity; binding site K87T decreased PTG interaction (F5S) V7A E28K F84L Destabilization and Class IIa CBM core (Y86D) decoupled CBM and DSP; mild or moderate effects on F88L Severe phosphatase activity; R91P severely impaired PTG and (R108C) malin interaction. E56K CBM-DSP Class IIb Y294N interface P301L N148Y (T187A) Destabilized; variable effects A188G on phosphatase activity; Class III DSP Variable L310W impaired malin and PTG G279S interaction. G279C* E210K Slightly destabilized; slight Surface- decrease in activity; slight Class IV P211L* Moderate exposed regions decrease in malin/PTG E224I interaction. Destabilized; mild effects on F321S phosphatase activity; impaired malin interaction; Class V Dimer interface F321C** Mild loss of dimerization and preferential binding to long oligosaccharides S25P Mild or no effect on stability Possible K140N or activity. Unknown PTM/regulatory Unknown site G240S S25P displays no interaction P246A with malin or PTG.

Figure 3.1 Distribution of LD cases and predicted pathogenicity of EPM2A missense mutations. (a) Distribution of 356 LD cases by mutation type. (b) Distribution of 94 reported EPM2A mutations by mutation type. (c) Correlation analysis between PolyPhen-2 pathogenicity score and ΔΔG (kcal/mol) predictions by CUPSAT and SDM for 27 mutations selected for analysis (see Table 3.2).

Figure 3.2 Distribution of LD missense mutations in the laforin structure. (a) Laforin is a bimodular protein with a CBM and DSP domain. LD missense mutations selected for study are shown throughout the primary sequence of laforin. (b) The X-ray crystal structure of laforin bound to maltohexaose (PDB:4RKK) (190). Laforin forms an antiparallel homodimer with a DSP-DSP interface. The CBM and DSP domain of one subunit are shown in light and dark shades of blue, respectively. Bound maltohexaose molecules are shown in green. (c) Residues affected by missense mutations are shown in red. Phosphate at the active site is shown in orange. The CBM and DSP domain are shaded as in (a) and (b).

Figure 3.3 Recombinant laforin mutants purified from E. coli. LD-associated and structure-guided mutants were expressed in E. coli, purified by ion exchange and size exclusion chromatography, and then assessed for purity by SDS-polyacrylamide gel electrophoresis (PAGE). Approximately 5-20 g of purified protein was loaded per lane, and gels were stained with Coomassie blue to visualize protein content. Lanes containing molecular weight markers (MWM) are indicated.

Figure 3.4 Stability and carbohydrate binding of LD mutants. (a) Stability measured by Tm. (b) Absolute Tm for mutants without substrate, and in the presence of 10 mM DP7 or 10 mM DP24. (c) ΔTm for mutants with incubation with DP7 or DP24. In (a-c), assays were performed in triplicate, and graphs represent the average ± SD. (d-g) Correlation of Tm with SASA ratio (a), PolyPhen- 2 score (b), CUPSAT (c) and SDM (d) predictions for ΔΔG (kcal/mol). Correlation coefficients (r) are shown. Statistical significance is indicated: ns p > 0.05, * p ≤ 0.05.

Figure 3.5 Glycogen phosphate levels, kinetics, and phosphatase activity of LD mutants. (a) Comparison of glycogen phosphate determined by different methods in the present study and the previous one (174). (b) Phosphate release after treatment of glycogen with antarctic phosphatase (AP) and/or amyloglucosidase (AG) overnight at 37°C. nt = no treatment. Averages of triplicate reactions are shown ±SD. (c) 0.5 mg glycogen was treated with 2.5 μg laforin for 1-2 hours to saturation, then α-amylase (AA), AG, or AA + AG was added for a final concentration of each enzyme of 0.3 mg/ml. Reactions were performed in duplicate; averages ± SD are shown. (d) Time course of phosphate release from glycogen at two different substrate concentrations. 10 mg/ml glycogen was treated for 30 minutes in subsequent assays. (e) Stability of laforin mutants over time at 37°C or room temperature. WT laforin and the catalytic mutant C/S were used as positive controls. (f) Specific activity of LD mutants; average and SD of triplicate reactions are shown. (g) Correlation between Tm and specific activity.

Figure 3.6 Kinetics of C3- and C6-labeled starch dephosphorylation. (a) Time course of C3- and C6-phosphate release from labeled starch by WT and C266S (C/S) laforin. (b) Phosphate release by WT laforin with increasing concentration of labeled starch. (c) C6- and (d) C3-labeled starch dephosphorylation kinetic studies on laforin mutants. The catalytically inactive C/S laforin mutant (C266S) was used as a negative control. (e) Kinetic constants for starch dephosphorylation. Units for Vmax and Km are pmol min-1 μg-1 and mg/mL, respectively. For (a-d), average of triplicates ± SD are shown.

Figure 3.7 Effects of LD mutants on C3- and C6 dephosphorylation. (a) Specific activity of laforin mutants. C3 versus C6 percent phosphate release are shown in the top panels as a percent of total dephosphorylation. (b) Specific activity comparison of laforin mutants with C3-labeled starch (yellow), C6-labeled starch (blue), and glycogen (grey; data from Figure 3.4c). Activity of each mutant was normalized to WT activity for each substrate. All assays were performed in triplicate; averages ± SD are shown.

Figure 3.8 Melting profiles of laforin mutants determined by differential scanning fluorimetry. Melting profiles of proteins in the absence of substrate are shown in black. Melting profiles in the presence of 10 mM DP7 are shown in blue. Data are representative of 3 replicates. Only DP7 profiles are shown for clarity.

Figure 3.9 Effects of carbohydrate binding on domain decoupling in Y294N and P301L. (a) Raw melt curves for WT, Y294N, and P301L binding to different concentrations of DP24. (b) First derivative of melt curves in (a). Tm values (calculated from the peak maxima) are shown. Only curves for 0 and 10 mM DP24 are shown for clarity. (c) Melt curves of WT and Y294N binding to increasing concentrations of glycogen. Data in (a-c) are representative of 3 replicates. (d) Tm values from the experiment in (c) were graphed to generate a binding curve for WT and Y294N; averages of triplicates ± SD are shown. One-site and two-site fits were determined using the Prism Software (Graphpad).

Figure 3.10 LD mutation "hotspot" along the allosteric path between the CBM and DSP domain. (a) DSF melting profile of W32A in the absence of glucan (black) and in the presence of 10 mM DP7 (blue). Graphs are representative of 3 replicates. (b) The CBM core and CBM-DSP interface is a “hotspot” for LD mutations. Affected residues are shown in orange. Glucan is shown in green. Y294 makes hydrogen bonds (in yellow) with main chain hydroxyls of R4 and L124 (in blue). (c) Melting profiles of structure-guided mutants compared to LD mutants.

Figure 3.11 Purification and characterization of laforin CBM and DSP domain. (a) The laforin CBM (amino acids 1-137) was expressed as a His6-tagged recombinant protein in E. coli and purified by affinity and size exclusion chromatography. (b) The laforin DSP domain (amino acids 150-331) was expressed as a fusion protein with a His6-tagged SUMO protein for stability. The fusion was purified by affinity and size exclusion chromatography. (c) Stability of CBM and DSP constructs compared to full-length laforin (FL). (d) Binding of FL, CBM and DSP constructs to increasing concentration of DP7 determined by DSF. (e) Binding of W32G laforin to DP7 compared to WT (FL) laforin data from (d). DSF assays in (c-e) were performed in triplicate; averages ± SD are shown.

Figure 3.12 Expression levels of LD mutants fused to LexA. LD mutations were expressed as LexA fusion proteins and coexpressed with either HA-tagged GAD (Ø), GAD-malin (M), or GAD-PTG fusion proteins. Antibodies for LexA and HA were used to detect the levels of each fusion protein. The expected molecular weights for GAD-malin and GAD-PTG are 55 kDa and 49 kDa, respectively.

Figure 3.13 Effects of LD mutants on the interaction of laforin with malin and PTG. (a) Interaction of WT laforin and mutants with GAD, GAD-PTG and GAD-malin fusion proteins. Mutants are color coded by class. Average of triplicate reactions ± SD are shown. (b, c, d) Correlation between Tm and GAD-malin interaction (b), Tm and GAD-PTG interaction (c), and GAD-malin and GAD-PTG interactions (d). Correlation coefficients (r) and p-values are shown. Significance is indicated: ns = not significant, * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001. In (d), box corresponds to inset.

Figure 3.14 Laforin as a scaffolding protein. (a) Triple hybrid experiment demonstrating noncompetitive binding of malin and PTG to laforin. LexA-laforin was expressed with either GAD-malin and pWS-PTG or GAD-PTG and pWS-malin. A reduction in β-gal activity with expression of the second binding partner would indicate malin and PTG have overlapping binding sites. Average of triplicate reactions ± SD are shown. (b) Electrostatics mapped onto the laforin crystal structure. (c) LD-affected residues relative to charged patches on the laforin dimer. (d) Phyre2-generated homology model of malin. Zoom of D197 in the NHL domains is shown. (e) Electrostatics mapped onto the malin model. (f) Phyre2-generated homology model of PTG. CBM21 is shown in teal. (g) Proposed model of the assembly of laforin, malin and PTG on the glycogen molecule.

Figure 3.15 A threshold of laforin function in LD. LD progression is proposed to be dependent on the type of mutation(s) carried by LD patients and the overall threshold of laforin activity for patients carrying EPM2A mutations. In individuals who are homozygous or heterozygous for the wild-type allele, i.e. carriers and noncarriers, laforin function is above the disease threshold and are therefore healthy. LD patients carrying Class I or Class II mutations have a classic LD progression. This is likely to be the case whether the patients are homozygous for these missense mutations or complex heterozygotes, i.e. carrying a missense mutation and a nonsense mutation (or other mutation leading to complete loss-of-function). LD patients carrying Class III or Class I mutations have a slower progression; particularly if they are homozygous for one of these mutations. LD patients carrying Class V mutations will have the slowest progression, e.g. the patient who was homozygous for F321C (489).

Figure 3.16 A personalized biochemistry pipeline for LD patients. A suspected LD diagnosis will be confirmed by genetic testing. Skin biopsies will also be performed to facilitate comparisons of LB load between patients with different genotypes. Nonsense, frameshift, or indel mutations in either EPM2A or EPM2B leading to nonfunctional protein (i.e. total loss-of-function mutations) are highly pathogenic. If known loss-of-function or EPM2A missense mutation(s) are identified, patient progression could be predicted immediately. If novel EPM2A missense mutations are identified, mutant expression constructs could be generated, characterized, and classified within a matter of weeks. This classification would give rise to an estimated clinical progression for that patient.

100

THE CATALYTIC ACTIVITY AND DYNAMICS OF THE HUMAN GLYCOGEN PHOSPHATASE LAFORIN PROVIDE INSIGHTS INTO LAFORA DISEASE

4.1 Introduction

In the late 1990s, mutations in the gene encoding the novel phosphatase laforin were identified as the cause of Lafora disease (LD) (170, 172). Lafora bodies (LBs), the hallmark inclusions of LD, were assumed to form as a result of aberrant glycogen metabolism, but the molecular basis for LB formation and LD pathophysiology were unknown. Laforin contains a dual specificity phosphatase (DSP) domain and carbohydrate binding module (CBM) from the CBM20 family (285). It was assumed that laforin localized to glycogen via its CBM where it would dephosphorylate a protein substrate. However, a robust protein substrate for laforin was never identified. Subsequently, the Dixon and Roach groups independently discovered that laforin's substrate was not a protein but glycogen itself, the only carbohydrate storage molecule found in mammals (173, 174). Laforin was the first member of the glucan phosphatase family, a subgroup of atypical DSPs that dephosphorylate complex carbohydrates. The other known members of this family are the plant enzymes Starch Excess 4 (SEX4) and Like Sex Four 2 (LSF2). SEX4 also contains a CBM and a DSP but in the reverse orientation, and the SEX4 CBM is from the CBM48 family (183). LSF2 only contains a DSP domain and utilizes starch surface binding sites to engage its substrate (187, 205). All of glucan phosphatases directly dephosphorylate carbohydrates, and their X-ray crystal structures have been determined (186, 187, 190, 207, 454). A common active site motif has been defined among them, but each member possesses a unique architecture and binding platform (139, 184, 205). DSPs are part of the protein tyrosine phosphatase (PTP) superfamily that utilizes a catalytic triad of cysteine, arginine and aspartate residues (532, 533). The PTP-loop contains the Cx5R-motif in which the arginine side chain and main chain nitrogen moieties stabilize the phosphate moiety of the substrate (534). A flexible loop containing the catalytic aspartic acid, often called the WPD-loop or simply the D-loop, closes over the active site once the phospho-substrate is bound. The first step of catalysis is nucleophilic attack of the substrate by the thiolate of the active site cysteine (535). The aspartate acts as a general acid, donating its proton to the substrate leaving group. A cysteinyl-phosphate intermediate is formed, which in the second step is hydrolyzed by an attacking water molecule. The same aspartate acts as a general base during this step, and inorganic phosphate is released. Studies of the prototypical PTPs YopH and PTP1B show that during catalysis, closing of the D-loop involves a ~10 Å swinging motion that initiates catalysis (536). DSPs are aptly named for their ability to dephosphorylate both phospho-tyrosine and phospho-serine/threonine residues

using the same catalytic triad. However, in contrast to PTPs, which have a deep and narrow active site cleft, DSPs have a wide and shallow active site that can accommodate phospho-tyrosine (537). Glucan and lipid phosphatases are all considered atypical DSPs and have slightly different active site topologies suited to their specific substrates (139, 208, 537). One of the most common ways to produce a catalytically incompetent PTP or DSP is to mutate the active site cysteine to serine (538). Alternatively, the catalytic aspartate can be mutated, although the D/A mutation can harbor some residual activity (538, 539). Surprisingly, in 2014, it was reported that transgenic overexpression of a catalytic mutant of mouse laforin, C265S, could reverse LB formation and autophagy defects in a laforin-deficient mouse model (189). This result was confirmed by another group in 2017: overexpression of human laforin C266S rescued LD in laforin-deficient mouse models (159). Glycogen in the laforin-deficient mice expressing the laforin C266S transgene did not over accumulate nor was it abnormally branched, both hallmarks of LBs, but the glycogen was still hyperphosphorylated. These results suggested that either the C/S mutants possess residual phosphatase activity, or that laforin prevents LB formation independent of its catalytic activity. The latter was more likely since the glycogen was still hyperphosphorylated. X-ray crystal structures of the laforin DSP domain and full-length protein were both recently published (190, 454). The crystal structure of the DSP domain with the C266S mutation revealed a prototypical PTP-loop bound to a sulfate ion (454). The authors proposed that a second "triad" involving C169, R171, and D197 exists in laforin that is in close proximity to the original triad. It was suggested that laforin C266S might possess residual activity via this “second” triad, which could explain how C266S could rescue the LD phenotype. Furthermore, this group proposed that laforin forms a CBM-CBM dimer interface reminiscent of VH1, a prototypical DSP that dimerizes via its N-terminal helices (Koksal 2009). The laforin model was based on the crystal structure of the DSP domain, a homology model of the CBM, and manual fitting of the domains to small angle X-ray scattering data. We determined the crystal structure of laforin C266S lacking three C-terminal amino acids in complex with maltohexaose and phosphate (190). The two molecules in the asymmetric unit formed an antiparallel homodimer with an extensive 880 Å2 interface mediated exclusively by the two DSP domains. Multiple glucans were bound both to the CBMs and the DSP domains. We discovered that laforin C266S binds cooperatively to long oligosaccharides with an average length of 24 glucose units. The LD-associated mutant laforin F321S affecting the core of the hydrophobic DSP-DSP interface rendered laforin monomeric and abolished high-affinity cooperative binding. These data suggested that laforin dimerization mediates its specific binding to substrates with long glucan chains, such as those found in LBs. In the present study, we employ both crystal structures and an array of mutations coupled with biochemical analysis to characterize the laforin active site

102 and dimer interface. We show that a second catalytic triad is mechanistically impractical and incongruent with biochemical data. We define the architecture of the laforin active site in the context of the glucan phosphatase family and characterize the unique laforin dimer interface. Finally, we discover that the laforin C266S mutant scavenges phosphate and exhibits unique binding properties compared to WT laforin. The laforin C266S mutation confers rigidity to the protein, leading to increased interactions with laforin binding partners. These data support an important role for laforin as a protein scaffold and may explain why laforin C266S overexpression rescues the LD phenotype in mouse models.

4.2 Results

4.2.1 The laforin active site

The laforin DSP domain structure bound to sulfate was determined at 2.3 Å (PDB: 4R30) and the laforin CBM and DSP structure bound to phosphate and maltohexaose was determined at 2.4 Å (PDB: 4RKK) (190, 454). The isolated DSP domain is comprised of residues 153-331, and there are four molecules in the asymmetric unit (Figure 4.1a). There is very little conformational difference between each DSP domain with root mean square deviation (RMSD) values between the chains of 0.1-0.2 Å (Table 4.1). The CBM-DSP protein includes residues 1-329 and there are two molecules in the asymmetric unit, also with an RMSD of 0.1 Å within the DSP (Figure 4.1b, Table 4.1). The interface between chains B and D in 4R30 is the same interface as the one observed in 4RKK (Figure 4.1c); however, due to conformational differences, which are discussed later, the interface in 4R30 has slightly more surface area than in 4RKK (Table 4.1). The RMSD between the DSP domains in 4R30 and 4RKK is 1.7-1.8 Å (Table 4.1). DSPs share four common motifs that frame the PTP-loop at the catalytic core. The D-loop contains the catalytic aspartic acid; the recognition domain contributes to active site depth; the variable insert or V-loop, which is shorter in DSPs than PTPs; and the R-motif named for a highly conserved arginine that is surface- exposed in the DSPs (540). The DSP domains in 4R30 and 4RKK only deviate moderately in the PTP-loop, D-loop, R-motif, and helix α10 that immediately precedes the R-motif (Figure 4.1d). In 4R30, the majority of the recognition domain is missing (residues 137-153) and this absence results in an open D-loop; while the D-loop is closed over the glucan substrate in 4RKK (Figure 4.1d,e,f). D235 is the catalytic aspartate for laforin, and there is a 7 Å difference in its position between the open and closed conformations (Figure 4.1f). In both structures, the oxyanion interacts with the PTP-loop, but in the glucan-bound structure (4RKK) the phosphate is 2.0 Å deeper into the catalytic cleft than the sulfate in the glucan-free structure (4R30) (Figure 4.1f). Also, the alanine, glycine, and valine residues at position 2, 3, and 4 within the Cx5R-motif are inverted

103 inward in 4RKK in contrast to 4R30. Alanine and a long-chain aliphatic residue at positions 2 and 4 are hallmarks of the active site motif in glucan phosphatases, and they interact with glucans in all glucan phosphatase structures (184, 187, 190, 205, 207). The conformational difference in the PTP-loop of 4R30 is likely due to the absence of bound glucan but may also be due to the deleted recognition domain in this construct, which also makes contact with the PTP-loop (Figure 4.1d). In the glucan-bound structure (4RKK), the side chains of C266S, R272, and D235 are all within 7 Å of each other (Figure 4.1f). In the glucan-free structure (4R30), D235 is 12 Å away from C266S and 8 Å away from R272, but C266S and R272 are only ~5 Å apart. This positioning is ideal for coordinating a ~ 4 Å phosphate moiety. Mapping of surface electrostatic potential also reveals the strong positive charge of this pocket, suitable for accommodating phosphate (Figure 3.14b). Cingolani and colleagues predicted that C169, R171 and D197 form a second catalytic triad. In contrast, these residues are all at least 7 Å apart in both structures (Figure 4.1f) and are present in a region with no pocket or charge (Figure 3.14b). Furthermore, rather than being part of a flexible loop that is required for dephosphorylation, D197 is within an α-helical region of a highly conserved DSP structural motif and has very little solvent exposure (Supplemental File 1) (190). Therefore, it is structurally unlikely to rotate between an open and closed conformation upon binding to a substrate. Although C169, R171, and D197 are all highly conserved, upon close inspection of the laforin structures, the location and orientation of these residues indicates that they are unlikely to form a second catalytic triad and instead likely to be key structural elements of the protein. To test the contribution of each triad residue to stability and catalysis, individual alanine mutants in the context of the full-length protein were generated. These mutants were expressed and purified via nickel affinity and size exclusion chromatography to very high purity, similar to C266S (Figure 4.2a). Multiple attempts to purify D197A were necessary, since this protein had a tendency to undergo proteolysis during expression and purification. The mutants were first analyzed by differential scanning fluorimetry (DSF) to quantify stability and carbohydrate binding. D235A, C266S, and R272A all possess similar melting temperatures (Tm) compared to WT laforin, indicating these mutations did not affect laforin stability (Figure 4.2b). C266A was significantly destabilizing, reducing the Tm of laforin from 50.0 to 41.8°C. C169A, R171A, and D197A all reduced stability to a similar extent with a Tm of 44.3-45.6°C. WT laforin exhibits an increased Tm when it binds to maltoheptaose (i.e. DP7). None of the mutations negatively affected laforin binding to maltoheptaose with all exhibiting a ΔTm of 4- 5°C except for laforin C266A. Laforin C266A yielded an increase of 11.3°C in the presence of DP7. This result is consistent with our previous observation that destabilizing mutations exhibit a greater ΔTm in the presence of carbohydrates

104

(Figure 3.4). With the exception of C266A, the glucan phosphatase activity of these mutants has already been determined via malachite green assay utilizing amylopectin as the substrate (454). Amylopectin is the major component of plant starch and contains covalently bound phosphate (196, 197). C266S, D235A, R272A and D197A were all catalytically inactive, while C169A and R171A were nearly as active as wild-type. It was suggested that the malachite green assay is not very sensitive and that C266S might harbor some residual activity below the limit of detection, which could explain why a laforin C266S mutant could rescue laforin- deficient mice (454). Glycogen and starch can both be phosphorylated on C3- and C6-hydroxyls (139, 158, 200, 210). We have previously employed a highly sensitive radiolabeling assays to quantify specific phosphate release from C3- and C6- positions (184, 187, 207). We found that C266S, C266A, D235A, and R272A were entirely inactive and D197A activity was dramatically decreased, while C169A and R171A possessed slightly reduced activity compared to WT laforin (Figure 4.2c). No change in C3- versus C6-phosphate release was observed for those with high activity and D197A possessed only minimal activity toward C3-phosphate. These data indicate that residues of the first triad contribute primarily to activity rather than structural integrity. Conversely, residues in the proposed second triad contribute primarily to structural integrity rather than catalysis. The glucan phosphatase family members share at least 19% identity and 36% similarity in the DSP domain (Figure 4.3a) (183). Within this family, each DSP motif has at least one highly conserved residue, typically an aromatic, that contributes to glucan binding and substrate presentation to the active site. All three glucan phosphatases share a conserved tyrosine in the recognition domain that is a major contributor to overall glucan phosphatase activity in both SEX4 and LSF2 (Figure 4.3a,b) (187, 207). In the V-loop, SEX4 and LSF2 each have two adjacent aromatic residues that contribute to activity (Figure 4.3a,b) (187, 207). Laforin lacks an aromatic at this position, likely due to the fact that the glucan in laforin is positioned more deeply in the active site where it interacts instead with D197 (Figure 4.4a). Additionally, a nearby tryptophan, W196, interacts with the second glucan at the laforin active site, and this aromatic residue is unique to laforin (Figure 4.3a, 4,4a). In the D-loop, SEX4 and LSF2 possess a phenylalanine immediately adjacent to the catalytic aspartate, which facilitates closing of the D- loop over the bound phospho-substrate (Figure 4.3a). In LSF2, mutation of this residue (F162) reduces catalytic activity by over 50% (187). In contrast, the D-loop residue of laforin is M236, and actually forms a bridge-like interaction with M139 over the maltoheptaose molecule in the glucan-bound structure (Figure 4.4a) (190). In SEX4 and LSF2, residues of the R-motif contribute to C3- vs. C6-specific activity and the specificity was engineered by mutating these residues (187, 207). In SEX4, F235 promotes glucan binding so that phosphate at the C6 position is oriented toward the catalytic cysteine. In contrast, LSF2 possesses a glycine at this position that promotes the accommodating of glucan with C3-phosphate oriented toward the catalytic cysteine (139, 190). Laforin possesses a tryptophan

105 at this position in the R-motif (Y304), and a nearby aspartate (D306) also interacts with glucan at the active site (Figure 4.4a). However, the C2 and C3 hydroxyls of maltohexaose are oriented towards catalytic cysteine in the glucan-bound laforin structure (Figure 4.3b) (139, 190). We generated alanine mutations of these residues and other residues interacting with the glucan at the laforin active site and purified the recombinant proteins (Figure 4.4a,b). We then characterized each mutant using DSF, malachite green assays using glycogen as the substrate, and the radiolabeling assay. With the exception of D304A and M236A, all of these mutations had a Tm between 48 and 50°C, indicating normal stability (Figure 4.4c). Additionally, all mutations exhibited a normal binding profile to DP7 with the expected 3-5˚C increase. These data are consistent with the observation of multiple groups that only CBM residues dramatically contribute to glucan binding (Chapter 3) (190) (285). Interestingly, M139A, a residue of the recognition domain directly interacting with glucan, did not affect glycogen phosphatase activity (Figure 4.4d). However, other mutations affecting the recognition domain, T142A and Y146A, abolished activity (Figure 4.4d). These data support the importance of the recognition domain for catalysis, and may help explain why the purified DSP domain, which lacks most of the residues of the recognition domain, has no phosphatase activity even when small molecule substrates are used (Chapter 3) (454). Intriguingly, W196A slightly enhanced glycogen phosphatase activity (Figure 4.4d). The location of this residue relative to the two glucans at the laforin active site suggests it may promote accommodation of a branch point (Figure 4.4a). Mutations of the D-loop residue, M236A/F/N, partially reduced phosphatase activity (Figure 4.4d). M236F was the most detrimental mutation, suggesting a bulky aromatic like that of SEX4 and LSF2 prevents glucan access to the active site. In contrast, M236N was the least detrimental, indicating that a long and linear residue, even if hydrophilic, can accommodate a glucan better than a small residue like alanine. T238 and R241, proximal to the D-loop, and Y304 and D306 of the R-motif, were also important residues for activity (Figure 4.4d). Similar changes in activity were observed when radiolabeled starch was used as a substrate, and we did not see any major changes in C3 versus C6 dephosphorylation by these mutants (Figure 4.4e). These data further confirm that the C266 active site of the original PTP-loop is the laforin active site, and the second proposed triad lacks catalytic function in vitro or in vivo. Mutation of residues within the second triad instead affect the stability of laforin. These residues are likely to be highly conserved due to their importance in structural maintenance. D197 appears to interact with glucan at the active site, but it would not be able to perform a catalytic role due to its location within an -helix.

4.2.2 The DSP-DSP dimer interface

Although Sankhala et al. proposed a CBM-CBM dimerization interface, three

106 unique interfaces were found in the asymmetrical unit of the crystallized DSP domain (Figure 4.1a). First, the A:C and B:D interfaces were the most extensive interfaces in the structure each with encompassing over 1000 Å2 of surface area (Table 4.1). The B:D interface matches that of the DSP-DSP interface in the full- length crystal structure (Figure 4.1c). We previously demonstrated that the theoretical profile calculated for the dimeric crystal structure was consistent with small angle X-ray scattering data (190). Furthermore, the LD-associated mutation F321S falls within the DSP-DSP interface and this mutation renders laforin monomeric and abolishes laforin cooperativity towards long glucan substrates. To test if the CBM had a propensity for dimerization, we purified the laforin CBM (residues 1-137) and determined its oligomeric state by analytical ultracentrifugation (AUC). A single species of 21 kDa predominated, with a very small proportion of a 41 kDa species (Figure 4.5a,b). The molecular weight of the predominant species was in agreement with the predicted molecular weight of the monomeric CBM (15 kDa). These data indicate that the laforin CBM is predominantly monomeric and along with previously published data by Raththagala et al. that the DSP-DSP interface corresponding to the B:D interface in 4R30 is the physiologically relevant dimer interface. When the D-loop is in the open conformation (as in 4R30), it interacts with the D-loop of the other DSP subunit and extends the surface area of the dimer (Figure 4.6a, Table 4.1). M236 rotates outward and becomes buried at the interface (Figure 4.6a). When the D-loop closes (as in 4RKK), M236 rotates away from the dimer interface to associate with glucan, and the interface surface area decreases by over 100 Å (Figure 4.6a, Table 4.1). F321, I229, C250, and L251 are all residues with a high percentage of buried surface area at the interface in both conformations (Figure 4.6a, Supplemental File 1). To define the contribution of these residues to dimerization and cooperativity, we generated, purified and characterized their corresponding alanine mutants (Figure 4.6b). While I229A and C250A were mildly destabilizing, having a Tm of 45.8 and 47.8°C, L251A and F321A were severely destabilized with a Tm of 41.5 and 36.6°C (Figure 4.6c). All mutants exhibited a similar ΔTm with DP7 as WT, demonstrating that they can bind short glucans (Figure 4.6c). We also performed glycogen dephosphorylation and radiolabeling assays to determine C3 versus C6 dephosphorylation and substrate differences. All three activities of I229A were reduced by over 50%, while C250A did not reduce any of the activities (Figure 4.6d). In contrast, L251A and F321A showed substrate-specific effects: each exhibited a moderate decrease in glycogen dephosphorylation while starch dephosphorylation was significantly reduced (Figure 4.6d). The data for L251A and F321A are similar to what we have observed for LD mutations F321S and F321C (Chapter 3), and strongly suggest that L251A and F321A also impair cooperative binding to long glucan substrates. Cooperativity refers to a binding event that allosterically induces change in the affinity for subsequent binding events. Cooperativity is often facilitated by dimerization and can be quantified by the Hill equation (541, 542). A Hill coefficient

107 greater than one represents positive cooperativity; that is, the first binding event increases the affinity of the second binding event. A classic example of positive cooperativity is binding of oxygen to hemoglobin: the affinity of hemoglobin for oxygen increases as more oxygen binds (543). In contrast, a Hill coefficient less than one indicates negative cooperativity, where the first binding event reduces the affinity for the second event. A Hill coefficient equal to one represents no cooperativity, that is, no change in substrate affinity. Laforin C266S binds to long oligosaccharides with a Hill coefficient greater than one, reflecting positive cooperativity (190). Conversely, monomeric F321S has a Hill coefficient of approximately one, meaning it lacks cooperativity (190). Therefore, cooperative binding is impacted by laforin dimerization. In order to test the contribution of residues I229 and L251 to cooperative binding, we performed binding assays with a titration of long oligosaccharides (with an average length of 24 glucose units, i.e. DP24). We included the LD- associated mutation F321C, which has been shown to be monomeric like F321S and has a similar activity and binding profile (Chapter 3) (510), as well as F321A. Surprisingly, we found that none of the proteins exhibited the high affinity binding observed for C266S, including WT laforin (Figure 4.6e,f). The Hill coefficients for WT, L251A, F321A, and F321C were less than 1, indicative of negative cooperativity (Table 4.2). As previously reported, the Hill coefficient of C266S was greater than 1, indicative of positive cooperativity. I229A exhibited a Hill coefficient of approximately 1, reflecting a lack of cooperativity. Also, I229A possessed a lower maximum ΔTm (i.e. apparent Bmax). I229 is within a β-sheet adjacent to the D-loop (Figure 4.6a). Since this mutation abolishes cooperative binding, this β- sheet may assist in mediating interchain allosteric interaction. To further define the impact of I229A and L251A mutations on laforin structure, we also defined their oligomeric state using native polyacrylamide gel electrophoresis (PAGE) and AUC. Although the theoretical weight of the WT laforin dimer is 72 kDa, it migrated with the 66 kDa marker when analyzed by native PAGE, likely due to its elongated shape. I229A and L251A both yielded similar migration patterns as WT laforin (Figure 4.6g). Furthermore, both mutations had a sedimentation coefficient similar to WT laforin by AUC (Figure 4.6h). Collectively, the substrate-specific effects of L251A and F321A and the change in binding dynamics induced by I229A suggests all of these residues are involved in the cooperative binding event.

4.2.3 Distinct binding properties of WT and C266S laforin

The difference between WT and C266S binding to DP24 were unexpected and striking. Interestingly, no difference in binding was observed when DP7 was used as substrate (Figure 4.7a). A smaller but less pronounced difference was observed when glycogen was used with laforin C266S again demonstrating a higher binding affinity compared to WT laforin (Figure 4.7b). Strikingly, phosphate was found at the active site in the crystal structure of laforin C266S even though

108 there was no phosphate in any of the purification or crystallization buffers, suggesting it was scavenged during the purification procedure (190). Scavenged phosphate may alter laforin binding to long carbohydrates. To test this hypothesis, we added malachite green reagent to 100 μg of native and boiled WT, C266S and D236A proteins. Malachite green is a highly acidic reagent, so it denatures proteins and causes them to precipitate. Boiling the proteins prior to adding malachite green reagent would release any bound phosphate prior to precipitation. After malachite green was added, the samples were centrifuged, and the soluble fraction was added to a 96-well plate for spectrophotometric quantification. While WT and D235A contained no phosphate, phosphate was readily detected in both native and boiled C266S samples (Figure 4.8a,b). Malachite green reagent was also added to the remaining precipitated protein pellets. The pellets from the C266S samples that were not boiled stained more intensely, while the boiled C266S samples stained weakly (Figure 4.8a). These data suggest the phosphate is tightly bound in a pocket of the protein, and the rapid denaturation caused by the acidic reagent causes the bound phosphate to precipitate with the collapsed protein. In contrast, boiling the protein released the bound phosphate into the supernatant before acid was added. This experiment was performed on three different days, and consistently, WT laforin had no phosphate, while C266S had 0.5 nmol phosphate per nmol protein (Figure 4.8b). Since the protein concentration of laforin is determined based on the amino acid sequence, i.e. the theoretical molecular weight of the monomer, 0.5 nmol phosphate per nmol protein is equivalent to 1 phosphate per dimer. We then sought to identify the source of the scavenged phosphate. We performed a phosphate-strip experiment based on the method of Neznansky et al. used to remove phosphate from the well-known bacterial phosphate-scavenger PstS (544). PstS proteins were stripped of phosphate by washing with 70 column volumes of buffer during the metal-chelate chromatography step. We performed a similar procedure to strip phosphate from laforin C266S. To remove possible sources of phosphate contamination, all buffers were pre-incubated with PiBind resin, a commercial resin designed to remove contaminating phosphate from buffers and prevent high background in phosphate-sensitive enzymatic reactions. After expression of laforin C266S was induced, E. coli cells were harvested by centrifugation, resuspended in phosphate-free buffer, and lysed. Insoluble protein in the lysate was pelleted by centrifugation. The clarified lysate was then divided into two aliquots, which were separately loaded onto a nickel column for affinity chromatography, and the flow-through was collected for each (Figure 4.8b). After loading, the control sample was washed with 12 column volumes and then eluted with 300 mM imidazole according to our standard protocol. The stripped sample was washed with 70 column volumes, and then eluted (Figure 4.8c). Samples taken from each step of the purification show that the lysate and column flow- through fractions are phosphate-rich, and that the eluted proteins from both strip and control samples contained equal amounts of phosphate. When normalized to

109 protein content, the ratio of phosphate to protein was consistently 0.5. This experiment clearly demonstrates that all of the phosphate scavenged by laforin C266S comes from its bacterial host. The observation that C266S scavenges phosphate from the initial expression system suggested that this interaction is very tight, because the phosphate is carried all the way throughout the purification. We tested whether the addition of carbohydrates would cause C266S to release phosphate by incubating 100 μg of C266S with various carbohydrates in the presence of PiBind resin. Blank samples were included for the detection of phosphate contamination. After rotating for 1 hour at room temperature, the PiBind was allowed to settle, the supernatants were boiled and centrifuged, and then the soluble fraction was assayed for phosphate. Phosphate levels were unchanged in all of the samples with one exception: the sample in which C266S was incubated with 10 mM DP24 had much higher phosphate levels (Figure 4.9a). Although the DP24 was pre-incubated with PiBind resin, there was some residual phosphate in the samples prior to the addition of C266S, which accounts for the higher signal observed in this sample. We attempted to remove the phosphate by dialyzing C266S with PiBind and various phosphate binders used to treat hyperphosphatemia (545). 50 μg of C266S in 33 μL was dialyzed into 100 mL of buffer containing mixing PiBind resin or 1 mM of a phosphate binder. The buffers were exchanged four times over the course of 24 hours, and then phosphate was measured after boiling the dialyzed samples. While PiBind did remove a small fraction of the bound phosphate, the phosphate binders had no effect (Figure 4.9b). These data indicate the off-rate for phosphate binding of C266S is very low. We next tested binding of WT and C266S to phosphate or sugars in the presence of phosphate. While C266S exhibited an increasing ΔTm with increasing phosphate concentration, WT did not (Figure 4.10a). The Hill coefficient (h) of this binding curve was less than 1 (Table 4.2). In the presence of 50 mM γ-cyclodextrin (γCD), a circular glucan with seven glucose units, the Tm of WT and C266S increased by 7-8°C. The addition of phosphate to the reaction did not change the ΔTm of WT laforin with γCD, but the ΔTm of C266S with γCD increased with phosphate (Figure 4.10b). The phosphate-dependent increase in ΔTm was similar to the increase in ΔTm when phosphate was added in the absence of sugar (Figure 4.10b, black dots). To determine how phosphate changed the binding dynamics of WT and C266S laforin, titrations of DP7, glycogen and DP24 were performed with different phosphate concentrations. Phosphate had no effect on the binding curves of WT laforin with the various sugars (Figure 4.11a,b, Table 4.2). Although phosphate shifted the binding curves of C266S upward (Figure 4.11c), a minimal effect was observed on the shape of the C266S curves or binding coefficients with DP7 (Figure 4.11d, Table 4.2). For binding to DP7, both WT and C266S had Hill coefficients less than 1 regardless of phosphate concentration, indicative of negative cooperativity (Table 4.2). Higher concentrations of phosphate reduced

110 the steepness of the C266S binding curves with glycogen and DP24 (Figure 4.11d). Binding to DP24 went from positively cooperative in the absence of phosphate (h > 1) to negatively cooperative in the presence of 1 mM phosphate (h < 1) (Table 4.2). In contrast, binding to glycogen went from negatively cooperative in the absence of phosphate (h < 1) to positively cooperative in the presence of 1 mM phosphate, but the apparent Bmax was lowered (Table 4.2). It is important to note that although the steepness of the glycogen and DP24 binding curves for C266S changed with phosphate, the absolute maximum Tm reached at the highest glucan concentrations was the same for each substrate regardless of phosphate concentration (Figure 4.11c). The molar ratio of phosphate to C266S protein suggests that one C266S dimer scavenges only one phosphate molecule, i.e. one molecule in the dimer binds a phosphate (Figure 4.8b). The laforin C266S dimer could clearly bind to an additional phosphate, but the bound phosphate at one subunit appears to reduce the affinity of the other active site for phosphate (Figure 4.10a, Table 4.2). The data in Figure 4.11 indicate that both WT and C266S exhibit negatively cooperative binding to DP7 regardless of phosphate. This finding supports the preferential binding of laforin to long glucan substrates that many groups have observed in vitro and in vivo (181, 190, 314, 508). It is well established that glucan binding is mediated by the CBM. Therefore, it is likely that laforin is frequently "sampling" glucans for phosphorylation via both CBMs of the dimer. Binding of one site to short glucans such as DP7 reduces the affinity of the other site, so laforin is less likely to bind to other short glucans (Figure 4.12a). WT laforin also exhibits negative cooperativity in the presence of glycogen or DP24. The glycogen used in these assays is from rabbit liver. Liver glycogen contains little to no phosphate compared to muscle glycogen (163, 174, 211). DP24 also does not contain covalent phosphate. Therefore, binding of the laforin CBM to any non- phosphorylated carbohydrate, short or long, reduces the affinity of the other subunit for that glucan (Figure 4.12b). In contrast, laforin C266S, which binds at least one phosphate at the active site, shows positive cooperativity toward long glucans. This result means that when the CBM of one phosphate-associated dimer associates with DP24, the affinity of the adjacent CBM for DP24 is enhanced (Figure 4.12c). However, if both of the active sites are occupied with phosphate, then binding of one CBM to DP24 prevents the association of the second CBM with glucan (Figure 4.12d). The binding of a phosphate-bound C266S DSP to long glucans may mimic the binding of WT laforin to long phosphoglucan. Therefore, these data suggest that laforin preferentially binds to long phosphorylated glucan substrates, but a single dimer prefers to dephosphorylate only one glucan at a time even if it is associated with two separate glucan chains. D/A and C/A mutants are commonly used for substrate trapping. In our case, however, laforin C266S demonstrates product-trapping properties. It was unclear to us whether phosphate-scavenging was unique to the C266S mutant or common

111 in other catalytically incompetent mutants. We therefore performed titrations of DP24 with the mutants D235A and C266A. While D235A produced a similar binding curve to WT laforin, C266A exhibited a large shift with high concentrations of DP24 and a high apparent Bmax (Figure 4.13a, Table 4.2). However, binding was still negatively cooperative (Table 4.2). None of the other catalytic mutants exhibited the large shift of C266S and C266A. We also generated double mutants and found that mutation of R272 in addition to C266 abolished this large ΔTm and prevented phosphate scavenging (Figure 4.12b,c). Mutation of both D235 and C266 were stabilizing, raising the baseline Tm to 53°C, but the Tm in the presence of DP24 was still 66°C, similar to C266S, and the protein still scavenged phosphate (Figure 4.13b,c). The double mutant C169A/C266S was destabilized like C169A, but still exhibited a large ΔTm with DP24, and the protein scavenged phosphate. These data indicate it is the association with phosphate that causes the very large shift in the presence of DP24. This further corroborates the hypothesis that laforin preferentially binds long glucans that are phosphorylated. To determine how the association of phosphate impacts protein dynamics, we performed hydrogen deuterium exchange (DXMS) experiments on laforin C266S versus WT laforin. Interestingly, most peptides in C266S exhibited significantly less deuteration, indicating they were less solvent accessible (Figure 4.14a). A few peptides exhibited slightly increased deuteration in C266S compared to WT laforin (Figure 4.14a). The peptide containing C266S (257-275) was not identified due to the presence of the mutation. Peptides that were significantly less deuterated were part of the D-loop, R-motif and recognition domain (Figure 4.14b). These data suggest the regions framing the active site have closed in over the bound phosphate in C266S. We also performed DXMS experiments on WT and C266S in the presence of glycogen. The maximum percent change upon binding with glycogen was determined for each protein, and they were graphed together for comparison (Figure 4.14c). Interestingly, both WT and C266S exhibited similar changes in deuteration upon binding in the CBM. However, peptides in the DSP domain of WT and C266S laforin exhibited distinctly different deuteration changes. Multiple peptides in WT exhibited a major change in deuteration, particularly in the recognition domain and a peptide between the PTP- loop and R-motif (Figure 4.14c). However, C266S did not exhibit such large changes in deuteration in these regions, again suggesting that it is already in a more compact conformation (Figure 4.14c). In C266S, the DSP domain appears to undergo significantly less conformational change in the DSP domain than in WT laforin upon glycogen binding (Figure 4.14d). Furthermore, AUC experiments on WT and C266S show that WT produces a broader peak than C266S and a very minor monomeric fraction (Figure 4.14e). These data also indicate that WT and C266S laforin are conformationally distinct. These results also suggest that the dissociation constant for dimerization is lower in C266S compared to WT laforin, i.e. the C266S mutation produces a stronger dimerization interaction.

112

Cumulatively, these data suggest that the C266S mutation induces a significant conformational change in laforin that changes the way that it binds to glucans and phosphate. This conformational change is likely to also alter the ability of laforin to interact with other proteins. Malin is an E3 ubiquitin ligase that binds laforin, and some results suggest that a malin-laforin complex then ubiquitinates other proteins (175, 176, 188, 310). The malin-laforin complex has been shown to ubiquitinate enzymes involved in glycogen metabolism including glycogen synthase, glycogen debranching enzyme, and Protein Targeting to Glycogen (PTG), an indirect activator of glycogen synthesis. Malin deficiency also causes LD so it has been hypothesized that the common functions of the malin-laforin complex are more relevant to LD instead of the laforin glycogen phosphatase activity (159, 188). This hypothesis is supported by the finding that overexpression of laforin C266S in laforin-deficient mice rescues the LD phenotype (159, 189). However, it is possible that laforin C266S rescues the LD phenotype due to the preservation (or increase) of laforin C266S binding to malin, overcoming the lack of catalytic activity. To assess this hypothesis, we tested how the multiple laforin catalytic mutants affected the interaction of laforin with its binding partners, malin and PTG using a directed yeast two-hybrid assay. Laforin C266S enhanced both interactions by nearly two-fold (Figure 4.14f). In contrast, C266A and R272A reduced the PTG interaction by 35-62% and nearly abolished the interaction with malin. D235A minimally affected the malin interaction but reduced the PTG interaction by 63%. Since C266A and C266S do not have the same effects, it is not the bound phosphate that enhances laforin interaction with malin and PTG. Both C266A and C266S associate with phosphate and display high-affinity binding to long oligosaccharides (Figure 4.13). However, only C266S exhibits positive cooperativity, not C266A (Table 4.2). Furthermore, only C266S enhances the interaction of laforin with malin and PTG (Figure 4.14f). In contrast, laforin C266A, has a reduced Tm (Figure 4.2b), and fully or nearly abolishes the interaction of laforin with its protein substrates (Figure 4.14f). Therefore, it is likely that the conformation and positive cooperativity exhibited by C266S are linked to the enhanced interaction with laforin binding partners. These data demonstrate that laforin C266S possesses unique biophysical properties that enhance its binding to malin and PTG and suggest an alternative explanation for how laforin C266S rescues the LD phenotype in mice that are lacking laforin.

4.3 Discussion

The protein phosphatases are classified into completely separate superfamilies that independently evolved from distinct ancestral folds (546). The Cx5R active site motif is unique to the PTP superfamily (532). Alternatively, serine/threonine phosphatases utilize an entirely different catalytic domain that coordinates metal ions (547). The motifs utilized by the serine/threonine phosphatases involve multiple histidine residues and aspartic acid residues which vary by family. In

113 contrast, the haloacid dehalogenase phosphatase superfamily is characterized by a Rossmann-like fold, magnesium dependence and the active site sequence DxDx(V/T) (548). Herein, we show that laforin has only a single catalytic triad that adheres to the classic Cx5R motif of the PTP family. We define the unique architecture of the laforin active site in the context of the glucan phosphatase family. We show that laforin forms a DSP-DSP dimer interface and that residues at this interface contribute to the allostery between subunits and cooperative binding. Finally, we show that the laforin C266S mutant scavenges phosphate, exhibits unique binding properties and dynamics, and enhances the interaction of laforin with its protein substrates. The PTP consensus motif is comprised of a D/C/R triad where the aspartic acid is 30 residues upstream from the Cx5R active site (549). The proposed second triad is comprised of a cysteine and arginine that are only 1 residue apart and an aspartic acid 26 residues downstream from the arginine. It is evolutionarily unlikely that laforin would contain a second active site that does not adhere to the extended PTP consensus motif. Furthermore, catalysis is structurally implausible since the aspartate of the proposed non-classic triad is part of an α-helix instead of a flexible loop, and these residues do not reside in a positively charged pocket (Chapter 3). Finally, our experimental evidence demonstrates that mutation of any of the primary triad residues, D235, C266 or R272, completely ablates catalytic activity, while mutation of residues of the second triad does not (Figure 4.2). These data demonstrate that the laforin C266S protein does not maintain residual activity. Data from both crystal structures supports a DSP-DSP interface (Figure 4.1), the purified CBM does not dimerize (Figure 4.5), and mutations of residues at the DSP-DSP interface residues alter dimer-dependent cooperativity (Figure 4.6) (190). The laforin active site shares similarity with other members of the glucan phosphatase family, but it also differs in multiple ways. Like SEX4 and LSF2, a conserved tryptophan in the recognition domain (laforin Y146) is critical for the glucan phosphatase activity of laforin (Figure 4.3, 4.4). Indeed, the laforin recognition domain is essential for activity, since laforin T142A and the purified DSP domain lacking the recognition domain both lack phosphatase activity (Figure 4.4, Chapter 3) (454). Additionally, laforin contains an aromatic residue within the R-motif that interacts with glucan, similar to SEX4 (Figure 4.3). However, unlike SEX4, laforin does not exhibit a preference for the C3 or C6 position and mutation of this residue does not change specificity (Figure 4.4e) (184, 207). Interestingly, the glucans at the laforin active site interact with the V- loop residues W196 and D197, but W196 is not essential for activity like the V- loop aromatics of SEX4 and LSF2. Furthermore, the D-loop residue is a methionine (M239) in laforin rather than a phenylalanine. Both the shape and hydrophobicity of M239 are required for maximal activity (Figure 4.4d,e). The glucan is set further back in the laforin active site compared to SEX4 and LSF2, so the methionine may position the glucan more deeply in the catalytic pocket.

114

Strikingly, laforin C266S scavenges phosphate from its E. coli host while WT laforin and other catalytic mutants do not, and this requires coordination by R272 (Figure 4.8, 4.13c). Among the PTPs, C/S and D/A mutants are often used for substrate trapping (538, 539, 550). Although to our knowledge phosphate scavenging by other PTPs has not received significant attention, the combination of the scavenged phosphate with laforin's preferred substrate (long glucans) at the active site may mimic substrate-trapping by laforin C266S. It is not yet clear whether phosphate scavenging is unique to laforin or common among the glucan phosphatases. Despite the general assumption that substrate-trapping mutants retain the same structural and binding properties as their wild-type counterparts, conformational and dynamic differences in such mutants have been reported for other PTPs. The conformation of the PTP-loop in the apo structure of the PTP1B C/S mutant is dramatically different than in the apo structure of the WT enzyme (550). However, in the ligand-bound PTP1B WT and C/S structures, the PTP- loops adopt the same conformation. In WT PTP1B, the active site cysteine exists in a thiolate form, so the lack of charge in the C/S mutant could account for this conformational difference in the apo structure. The importance of a negative charge is supported by studies of a C/D mutant that possesses a PTP-loop conformation similar to the WT enzyme, but with a much lower catalytic turnover number (551). Although we observed conformational differences between the PTP-loops of the isolated DSP domain and full-length protein, this was likely not due to charge differences since both structures had the C266S mutation and a bound oxyanion (Figure 4.1). Instead, the difference may be due to the presence of the recognition domain and/or binding to a glucan, which likely pushes the oxyanion further into the active site. The movement of the D-loop is also quite different in C/S mutants of other PTPs. Studies have shown that the D-loop of WT PTP1B, YopH and SHP2 fluctuates between open and closed configurations, although the molecular details of this motion vary between enzymes (552-556). DXMS experiments on PTP1B show that the D-loop is less solvent accessible in the C/S mutant than in WT (550). The reduced solvent accessibility is likely due to locking of the D-loop in the open configuration, since closure would be sterically blocked by the altered PTP-loop. The C/S mutant of YopH also displayed reduced solvent accessibility of the D- loop (557). However, molecular dynamics simulations showed that the YopH C/S mutant spends more time in the closed conformation than the WT enzyme, which favors an open configuration, consistent with spectroscopic data (552). The laforin DXMS studies demonstrate that laforin C266S also has reduced solvent exposure in the D-loop (Figure 4.14a,b). Upon binding to glycogen, the D-loop of WT and C266S exhibit an identical increase in deuteration (Figure 4.14c). This result suggests that both enzymes favor the open configuration and that the change in deuteration upon glycogen binding is induced by D-loop closure. The DSP-DSP dimer interface increases by over 100 Å when the D-loop is open (Table 4.1, Figure 4.6). The AUC results suggest higher affinity at the dimer

115 interface of C266S than in WT laforin (Figure 4.14e). Therefore, like PTP1B, the laforin C266S D-loop may be locked in the open configuration, which increases interfacial surface area. Only upon glucan binding would it close, as in the crystal structure. Interestingly, laforin WT and C266S exhibit nearly identical deuteration changes in the CBM region upon binding to glycogen (Figure 4.14c,d). In contrast, there are major differences in the binding-induced deuteration changes in the DSP domains. Some of the regions in the DSP domain of glucan-free laforin C266S are much less solvent-exposed than in the WT enzyme; upon glucan binding, these regions also exhibit a smaller change in deuteration (Figure 4.14a-d). These data suggest that laforin C266S favors a conformation that, for the WT protein, is only induced by carbohydrate binding. Furthermore, the effect observed in the DSP must be allosteric since there is abundant evidence that glucan binding is mediated exclusively by the CBM (Chapter 3) (190, 285). These results are particularly profound because laforin C266S is also the only catalytic mutant found to enhance the interaction of laforin with malin and PTG (Figure 4.14f). Laforin C266S also interacts more strongly with other binding partners such as pyruvate kinase by yeast two-hybrid assay (524). Therefore, it is likely that the conformational rigidity induced by glucan binding also enhances laforin's function as a scaffold. Thus, laforin is more likely to bind to malin and PTG when it is associated with a glycogen molecule. The effects of phosphate on cooperative binding by C266S are difficult to fully interpret. This phenomenon calls for further investigations using additional techniques. However, allosteric effects on phosphatase activity induced by glucan binding have been previously reported. While glucans inhibit the generic phosphatase activity of laforin (where para-nitrophenylphosphate is used as substrate), laforin W32G, a mutation with abolished glucan binding, does not display this inhibition (508). The positive cooperativity of C266S binding to DP24 when it is associated with only one phosphate molecule suggests the laforin dimer may have a dual function: while one subunit dephosphorylates, the other subunit recruits malin and/or other proteins. This hypothesis leads to a very reasonable model: when laforin binds a long, phosphorylated glucan, in the process of dephosphorylating this glucan it also recruits malin and other proteins such as glycogen synthase and PTG. Ubiquitination of glycogen synthase and PTG by the laforin-malin complex may lead to slowing of glycogen synthesis and/or allow for remodeling by glycogen branching enzyme. Indeed, glycogen remodeling following exhaustive exercise is delayed in laforin-deficient mice, suggesting laforin plays an integral role in the remodeling process (160). Some have suggested that the glycogen phosphatase activity of laforin is irrelevant to LD. This hypothesis is flawed based on previously published work and data presented here. First, glucan phosphorylation is ubiquitous among the plant and animal kingdoms (139). Plants, algae and higher eukaryotes all possess glucan phosphatases, and although the source of glycogen phosphate has not

116 been identified in mammals, glucan dikinases have been identified in plants and algae. Secondly, glucan phosphorylation is essential for proper synthesis and degradation of plant and algal starch (139). Phosphate disrupts the formation of glucan α-helices by inducing steric hindrance (193). Cycles of phosphorylation and dephosphorylation are necessary for solubilizing the crystalline starch surface so that it can be broken down by plant amylases. It is possible – even likely – that in mammals, phosphate is also introduced (by an unidentified kinase) to prevent α- helix formation. This event may be especially important during glycogen super- compensation, when glycogen levels exceed basal levels upon refeeding after hypoglycemia or exercise (368-370). During this period of rapid re-synthesis, glycogen contains longer chains that at resting state (135, 160). The longer chains result because glycogen synthase extends glucan chains faster than glycogen branching enzyme can remodel them. Phosphate may be introduced during this super-compensation stage to prevent α-helices from forming and prevent glycogen from precipitating. Phosphates could then be removed upon glycogen remodeling; this may be the dual role facilitated by laforin. Furthermore, the catalytic residues of laforin are highly conserved among the higher vertebrates. If laforin's only role was to act as a scaffold, it would likely be a pseudophosphatase with the vestige of a PTP motif rather than an active phosphatase. Indeed, multiple members of the PTP family are pseudophosphatases, such as Styx, the Leukocyte common antigen-related (LAR) receptor PTP, and 6 out of the 14 known members of the myotubularin family (532). Finally, even in malin-deficient mouse models when laforin is fully functional, glycogen is both hyperphosphorylated and abnormally branched. These results suggest that hyperphosphorylation and abnormal branching are correlated and are likely functionally related. Glycogen degradation can proceed via two pathways: the major pathway is via cytosolic glycogenolysis facilitated by glycogen phosphorylase, and the second pathway is via lysosomal hydrolysis facilitated by acid α-glucosidase (Chapter 1). Laforin is a cytosolic protein that dephosphorylates glucans in vitro and dephosphorylates muscle glycogen during exercise-induced glycogenolysis (160, 173, 174, 184). Laforin also interacts with other enzymes of the cytosolic glycogen synthesis and degradation pathways: malin, PTG, glycogen synthase, and glycogen debranching enzyme. The laforin-malin complex may promote glycogen remodeling by ubiquitinating these enzymes (188). However, defective autophagic flux is a recurring theme in LD and recent studies point to a role of laforin and malin in glycophagy (179, 182, 291, 312). Malin incorporates K63-linked ubiquitin into its substrates, which is linked to autophagy rather than proteasomal degradation (303, 306). Furthermore, in cell culture, laforin and malin form a complex with the autophagy-related protein p62 (307). p62 is binds to polyubiquitin moieties via its ubiquitin-associated (UBA) domain and interacts directly with LC3 (558, 559). Additionally, malin was shown to ubiquitinate p62 in a laforin- dependent manner, and p62 enhanced laforin ubiquitination by malin (307). Furthermore, p62 expression induced the colocalization of malin and laforin with

117 the autophagosome marker LC3. These data strongly suggest that laforin and malin are involved in the selective autophagic degradation of glycogen via the lysosome. Hyperphosphorylated, aberrantly branched glycogen molecules may be shunted toward the glycophagy pathway when they become too precipitation- prone for the cytosol. Laforin is therefore likely to have multiple roles related to both cytosolic glycogenolysis and lysosomal glycophagy (Figure 4.15). In the cytosol, laforin is likely to both dephosphorylate glycogen and form a ubiquitin complex with malin. Glycogen synthase (GS), glycogen phosphorylase, PTG, debranching enzyme and branching enzyme are all cytosolic enzymes. Laforin and malin may form a complex and ubiquitinate some or all of these enzymes, leading indirectly to glycogen remodeling by changing the ratio of the activities of these enzymes. Although laforin C266S does not have phosphatase activity, it can still interact with malin to form a ubiquitination complex (Figure 4.15). In fact, this interaction is actually enhanced by C266S (Figure 4.15). When glycogen molecules escape this cytosolic remodeling and become hyperphosphorylated and precipitation- prone, they may be targeted for glycophagy by laforin and malin. Laforin preferentially binds to substrates with longer chains, and this is particularly evident for laforin C266S (Figure 4.11, 4.12). The laforin-malin complex may promote the ubiquitination of p62, which also interacts with ubiquitin moieties on laforin introduced by malin (Figure 4.15). This would promote the assembly of LC3 and the autophagy machinery around the aberrant glycogen molecule. This molecule would then be deposited in the lysosome where it is degraded at low pH by acid -glucosidase (GAA) (Figure 4.15). While Laforin C266S is catalytically inactive, our data demonstrate that it possesses an enhanced ability to interact with laforin binding partners. This enhanced interaction may rescue the LD phenotype in mice overexpressing laforin C266S due to its ability to bind carbohydrates and form a ubiquitination complex with malin. This ubiquitination complex may still promote glycogen remodeling in the cytosol and/or facilitate the recruitment of autophagic machinery for lysosomal glycogen degradation.

118

Table 4.1 Comparison of DSP molecules and interfaces in 4R30 and 4RKK. Structural comparisons between DSP domains (amino acids 153-329) in each PDB file were determined using the DALI server (467). Residues not common to both structures were excluded from the comparison. Interfacial atoms, residues and surface area were determined using PDBePISA (464).

Analysis of major intermolecular Structural comparison of DSP domains interfaces 4R30A: 4R30B: 4RKKA: PDB 4R30 4RKK Interface 4R30C 4R30D 4RKKC Chain A B C D A C Chain A B A A 37 0.2 0.2 0.1 1.7 1.7 Atoms 116 117 90

RMSD (Å) RMSD

B 36 37 0.1 0.2 1.7 1.8 Residues 32 31 27

4R30 C 35 35 37 0.2 1.8 1.8 Chain B D C D 35 35 35 37 1.7 1.8 Atoms 119 115 93

A 27 27 27 27 37 0.1 Residues 32 32 28

K 2 4RK C 27 27 27 27 36 37 Area (Å ) 1100.6 1084.5 879.95 DALI Z-score

119

Table 4.2 Apparent binding coefficients from DSF titration experiments with various substrates. Binding coefficients were determined using the Prism software (Graphpad). Units for Bmax are °C. Units for Kd are mM for phosphate, DP7 and DP24; for glycogen, Kd units are mg/ml.

Figure Protein Substrate Phosphate Bmax Hill Kd coefficient 4.6e WT DP24 0 mM 9.863 0.4966 0.4828 I229A DP24 0 mM 6.237 0.9339 0.01829 L251A DP24 0 mM 8.926 0.5945 0.09913 F321A DP24 0 mM 9.45 0.608 0.1357 4.6f C266S DP24 0 mM 15.83 1.744 0.09227 F321C DP24 0 mM 7.082 0.6101 0.0713 4.10a C266S Phosphate varied 9.459 0.4905 0.05471 4.11a,b WT DP7 0 mM 11.79 0.6406 14.71 DP7 0.1 mM 13.23 0.5884 21.41 DP7 1 mM 15.46 0.5322 35.31 Glycogen 0 mM 5.407 0.9633 0.6001 Glycogen 0.1 mM 6.055 0.7914 0.7724 Glycogen 1 mM 7.01 0.6111 1.05 DP24 0 mM 9.182 0.4946 0.4092 DP24 0.1 mM 9.73 0.4754 0.5048 DP24 1 mM 9.872 0.4787 0.4927 4.11c,d C266S DP7 0 mM 11.23 0.6362 9.942 DP7 0.1 mM 9.343 0.6267 15.58 DP7 1 mM 12.72 0.5107 63.78 Glycogen 0 mM 16.51 0.7495 2.951 Glycogen 0.1 mM 10.42 0.776 6.004 Glycogen 1 mM 2.861 1.53 1.247 DP24 0 mM 17.8 1.335 0.04648 DP24 0.1 mM 13.37 0.993 0.3474 DP24 1 mM 18.84 0.5556 5.486 4.13a WT DP24 0 mM 9.669 0.4993 0.5124 C266S DP24 0 mM 15.32 1.293 0.0653 C266A DP24 0 mM 17.56 0.6763 0.07973 D235A DP24 0 mM 15.44 0.4026 1.308

120

A 4r30 B 4rkk C

4rkkA 4r30A Chain A (DSP)

Chain C Chain A

Chain A (CBM) Chain C (CBM) 4r30D 4rkkC Chain C (DSP) Chain B Chain D D E F D235 4rkk 7 Å R motif 4r30 α10 R272

D loop

S266

D197 PTP loop C169

V loop R171

Recognition domain

Figure 4.1 Comparison of glucan-free DSP domain (PDB: 4R30) and glucan- bound full-length laforin (PDB: 4RKK) structures. (a) The four molecules in the asymmetric unit of the 4R30 crystal structure (amino acids 153-331) colored by chain. Sulfate is shown as yellow spheres. (b) The two molecules in the asymmetric unit of the 4RKK crystal structure (amino acids 1-329) colored by chain and domain. Phosphate is shown as orange spheres. Chain B consists of bound glucans. (c) Superimposed DSP domains forming the common interface between 4R30 and 4RKK. (d) Superimposed single DSP domains of 4R30 and 4RKK. DSP motifs are labeled. (e) Glucan at the active site in 4RKK. (f) Residues of the classic catalytic triad and the second triad proposed by Sankhala et al. in 4R30 and 4RKK. The difference in position of D235 between the two structures is shown. In (d, e, and f), phosphate and sulfate are shown in orange and yellow, respectively.

121

Figure 4.2 Characterization of putative and proposed catalytic residues at the laforin active site. (a) 5-10 μg of purified proteins were analyzed via SDS-polyacrylamide gel electrophoresis (PAGE). Protein was visualized by staining with Coomassie blue. (b) Mutant stability and binding to short glucans (DP7) determined via differential scanning fluorimetry (DSF). (c) C3- and C6-phosphate released from radiolabeled starch. Average of triplicate reactions ± SD are shown in (b,c).

122

Figure 4.3 Sequence and structural comparison of DSPs in the glucan phosphatase family. (a) Alignment of the DSPs of the glucan phosphatase family. DSP motifs and key residues are labeled. (b) Ligplot+ generated representation of hydrophobic and hydrophilic protein-glucan interactions at the CBM and active site in the glucan phosphatases (466). Maltohexaose is shown in green and hydrogen bonds are shown in black. Glucose moieties are numbered from the non-reducing end to the reducing end, which are indicated. Invariant CBM residues, DSP motifs and their corresponding key residues are highlighted.

123

Figure 4.4 Contributions of active site residues to laforin stability, binding and activity. (a) Stick representation of residues interacting with glucan at the laforin active site (PDB: 4RKK). (a) 10-20 μg of purified proteins were analyzed via SDS-PAGE with Coomassie staining. (c) Mutant stability and binding to short glucan (DP7). (d) Glycogen phosphatase activity of active site mutants. (e) C3- and C6- dephosphorylation of radiolabeled starch by active site mutants. In (c, d and e), the average of triplicate reactions ± SD are shown.

124

Figure 4.5 Oligomeric state of the laforin CBM. (A) C(S) and (B) c(M) plots from AUC experiments on the purified laforin CBM. Mapp indicates apparent molecular weight.

125

Figure 4.6 Residues of the DSP-DSP dimer interface contribute to cooperativity. (a) Comparison of the dimer interface in 4R30, where the D loop is open, and 4RKK, where the D loop is closed. (b) Purity of dimer mutants determined by SDS- PAGE and Coomassie staining. (c) Stability and binding to DP7 determined by DSF. (d) C6-P, C3-P and glycogen dephosphorylation relative to WT laforin. (e,f) Effects of dimer mutants on binding to DP24 compared to WT and C266S. (g) Oligomerization of laforin dimer mutants determined by native PAGE. (h) Sedimentation coefficient of dimer mutants compared to WT laforin. In (e, f, c, and d), the average of triplicate reactions ± SD are shown.

126

Figure 4.7 Comparison of WT and C266S binding to glucans. DSF titration experiments were performed with (A) DP7 and (B) glycogen. Average of triplicate reactions ± SD are shown.

127

Figure 4.8 C266S laforin scavenges phosphate from E. coli. (a) Visualization of phosphate in the soluble and insoluble protein fractions after the addition of malachite green reagent. (b) Quantification of phosphate in WT laforin and mutants. (c) Protein purity during C266S purification determined by SDS PAGE with Coomassie staining. (d) Visualization and quantification of phosphate during control and stripped purifications of C266S laforin from E. coli. Flow through (FT) and elution (E) are from the immobilized nickel affinity chromatography step. Average of triplicate reactions ± SD are shown.

128

Figure 4.9 Effects of sugars and phosphate binders on C266S phosphate. (a) Phosphate content of 100 μg C266S after 1 hour incubation with various sugars and PiBind resin. Blank controls were included to check for phosphate contamination. (b) Phosphate content of 100 μg C266S after dialysis with PiBind and phosphate binders.

129

Figure 4.10 WT and C266S binding to phosphate and γ-cyclodextrin (γCD). (a) WT and C266S phosphate binding curves determined by DSF. (b) WT and C266S binding to 50 mM γCD with various concentrations of phosphate. Data from (b) are shown as black dots in (b) for reference. Averages of triplicate reactions ± SD are shown.

130

Figure 4.11 The effect of phosphate on WT and C266S binding to various carbohydrates. (a) Absolute Tm and (b) ΔTm of WT binding to DP7, glycogen and DP24 with increasing phosphate. (c) Absolute Tm and (d) ΔTm of C266S binding with increasing phosphate. Averages of triplicate reactions ± SD are shown.

131

Figure 4.12 Model of phosphate-induced cooperativity switch. (a) Binding of DP7 to a CBM of WT (or C266S) laforin reduces the affinity of the adjacent CBM for glucan. (b) Binding of DP24 to WT reduces the affinity of the adjacent CBM for DP24. (c) Binding of DP24 to the CBM of C266S associated with a single phosphate enhances the binding of the adjacent CBM to DP24. (d) Binding of DP24 to the CBM of C266S associated with two phosphates prevents the binding of the adjacent CBM to glucan. Regions involved in glucan binding are based on the location of maltohexaose molecules in the crystal structure and previous DXMS experiments with glycogen (190).

132

Figure 4.13 Glucan binding and phosphate scavenging of catalytic mutants. (a) DP24 titration with catalytic mutants. (b) Stability and binding of catalytic mutants to DP7 and DP24. (c) Phosphate scavenging of catalytic mutants determined by malachite green assay. In (a, b), averages of triplicate reactions ± SD are shown.

133

Figure 4.14 C266S is conformationally distinct from WT laforin and interacts more strongly with malin and PTG. (a) Maximal percent change in deuteration of amino acids in WT laforin versus C266S. Negative change indicates decreased deuteration in C266S compared to WT. (b) Peptides with significantly decreased deuteration in C266S were mapped onto one subunit of the laforin crystal structure (4RKK). (c) Maximal percent change in deuteration upon WT and C266S binding to glycogen. Negative change indicates decreased deuteration in the presence of glycogen. Peptides with a significantly different deuteration change in WT binding to glycogen compared to C266S are indicated with asterisks: * indicates a difference of >10%; ** indicates a difference of >30%. (d) Peptides with a significant difference in deuteration change upon glycogen binding mapped onto the crystal structure (4RKK). In (b) and (d), the missing peptide is labeled in red. (e) C(S) plots from AUC experiments; both WT and C266S were run at ~0.3 mg/ml (OD ~0.5). (f) Binding of catalytic mutants to malin and PTG by yeast two-hybrid assay. Averages of triplicate reactions ± SD are shown.

134

Figure 4.15 A model for C266S-mediated rescue of laforin deficiency in LD. Modified from the schematic proposed by Irimia et al. (160) based on recent data from our group and others. C266S laforin ablates phosphatase activity but enhances the laforin-malin interaction and interaction with other proteins.

135

TARGETING PATHOGENIC LAFORA BODIES IN LAFORA DISEASE USING AN ANTIBODY-ENZYME FUSION

5.1 Introduction

The progressive myoclonic epilepsies (PMEs) are a group of inherited disorders characterized by recurrent seizures, myoclonus, and progressive neurological decline. There are currently no treatments for PMEs, and anti- epilepsy drugs are palliative at best (560). Lafora disease (LD; epilepsy, progressive myoclonus type 2, EPM2) is a severe form of PME that typically manifests with tonic-clonic seizures and myoclonic jerks in the early teen years followed by rapid neurological deterioration, increasingly severe and frequent epileptic episodes, dementia and death within ten years of onset (OMIM: 254780). LD is caused by mutations in the EPM2A or EPM2B genes that encode laforin, a glycogen phosphatase, and malin, an E3 ubiquitin ligase that ubiquitinates enzymes involved in glycogen metabolism (reviewed in (167)). LD is distinguishable from other PMEs by the presence of cytosolic polysaccharide inclusions known as Lafora bodies (LBs) most notably in the brain, where they are found in neuronal cell bodies and dendrites and astrocytic processes, and in other tissues such as muscle, heart, and liver (322, 324, 325). Among the PMEs LD is uniquely considered a non-classical glycogen storage disease (GSD) (165, 166). Independent studies from multiple groups demonstrate that Epm2a-/- and Epm2b- /- mice recapitulate disease symptoms with respect to LB formation, neurodegeneration, neurological impairments, and susceptibility to epilepsy (177- 181, 561, 562). LBs, which form in the absence of functional laforin or malin, are comprised of abnormal polysaccharide referred to as polyglucosan. The term polyglucosan refers to any glucose polymer that is abnormal, and other types of polyglucosan bodies have been described in some GSDs (79). Unlike glycogen, LBs are water- insoluble, hyperphosphorylated, contain long glucose chains, and range in size from a few microns to up to 35 or 40 μm (158, 168, 169, 209). Dr. Gonzalo Rodriguez Lafora first described these inclusions in patient autopsies, noting their histochemical similarity to plant starch (321). Decades later Sakai and colleagues isolated these inclusions from patient tissue and showed that LBs are indeed structurally more similar to starch than to glycogen (168, 169). Plant starch and LBs both contain significantly more covalently bound phosphate and longer glucan chains than mammalian glycogen (168, 208, 210). Loss of laforin or malin function results in the transformation of glycogen molecules into hyperphosphorylated polyglucosan molecules with an altered chain length distribution, and these aberrant molecules, prone to precipitation, accumulate to form LBs (188). How laforin and malin prevent this process is still being defined. Results from multiple labs have established that accumulation of polyglucosan drives neurodegeneration in both LD and non-LD models. Glycogen synthase is

the only mammalian enzyme able to catalyze glucose polymerization in vivo. Epm2b-/- mice lacking glycogen synthase (Gys1) in the brain are devoid of cerebral glycogen and LBs, exhibit no neurodegeneration, have normal hippocampal electrophysiology, and do not display an increased susceptibility to kainate-induced epilepsy (182). A similar phenotypic rescue was observed when Gys1-/- mice were crossed with Epm2a-/- mice (293). Furthermore, Epm2b-/- mice lacking just one Gys1 allele in the brain have reduced glycogen and also show near complete rescue of these phenotypes (182). Epm2a-/- and Epm2b-/- mice lacking Protein Targeting to Glycogen (PTG), a protein that promotes glycogen synthesis, also exhibit reduced LB accumulation, and neurodegeneration and myoclonic epilepsy are resolved in these animals (293-295). These results demonstrate that decreased or complete absence of the glycogen synthesis machinery ablates LB formation, neurodegeneration, and epilepsy in LD mouse models. The reverse has also been observed: overexpression of a constitutively active form of glycogen synthase in otherwise wild type animals drives neurodegeneration in both flies and mice (563). The accumulating polysaccharide in transgenic animals overexpressing glycogen synthase is a polyglucosan (i.e. abnormal) rather than normal glycogen (261). Collectively, the aforementioned studies illustrate that cerebral LB accumulation is pathogenic. These studies have both elucidated the molecular etiology of LD and have made LBs an obvious therapeutic target (564). These studies have motivated ongoing efforts to develop a targeted therapy (i.e. precision medicine) for LD (565). One form of precision medicine that has been successful with GSDs is the introduction of exogenous replacement enzymes. Enzyme replacement therapy has proven effective for Pompe disease (OMIM: 232300), an inherited GSD (566). Pompe patients are deficient in the lysosomal enzyme that degrades glycogen, acid -glucosidase (GAA), and are currently treated with a recombinant human form of this protein known as rhGAA or alglucosidase alfa (Myozyme®, Lumizyme®, Genzyme) (566, 567). The uptake of rhGAA is a receptor-mediated endocytic process that targets rhGAA to the lysosome. However, there is a significant portion of cytosolic glycogen in Pompe patients, and since rhGAA only targets lysosomal glycogen, those with large pools of cytoplasmic glycogen do not respond well to the current therapy (568). In LD, LBs are entirely cytosolic. Histological studies from patient tissues report that LBs are not membrane bound, and this observation has been confirmed in mouse models (179, 452, 569, 570). Thus, a therapeutic enzyme degrading LBs must be delivered to the cytosol. Although cytosolic targets remain challenging for protein therapeutics, antibody-based delivery platforms provide a means for penetrating the cell membrane (reviewed in (571)). The monoclonal anti-DNA autoantibody 3E10 and its antigen-binding (Fab) and variable domain (Fv) fragments can be fused to an enzyme to facilitate cytosolic delivery in multiple cell types (572-574). This antibody-enzyme fusion (AEF) strategy has proven successful for the replacement

137 of a cytosolic enzyme and rescue of X-linked myotubular myopathy in mice (575). A Fab-GAA fusion has similar efficacy to reduce glycogen in Pompe mice with the additional advantage of targeting both lysosomal and cytosolic glycogen and is now in Phase 1/2 clinical trials (576, 577). Rather than utilizing the AEF platform to deliver a replacement enzyme, we generated an AEF designed to penetrate cells and degrade pathogenic LBs. Herein, we show that an AEF comprised of the humanized 3E10 Fab fragment and pancreatic α-amylase degrades LBs in vitro, degrades glycogen and polyglucosan in situ, and reduces LB load in vivo.

5.2 Results

5.2.1 Construction of an antibody-enzyme fusion (AEF) that degrades starch

An important consideration in designing an LB-degrading enzyme is that LBs are less sensitive to enzymatic hydrolysis than normal glycogen. Resistance to digestion with diastase (a generic term for enzymes that degrade polysaccharides) is a defining feature of LBs and one factor that differentiates LBs from glycogen (570, 578). However, in routine histological preparations, embedded sections are only incubated with diastase for 15-30 minutes prior to staining. Early studies reported that LBs in tissue sections can be digested by a 5-10 hour incubation with pancreatic α-amylase and γ-amylase (i.e. amyloglucosidase), but they are somewhat resistant to digestion by β-amylase or glycogen phosphorylase (168, 169, 579). Glycogen, starch, and LBs are all polysaccharides with different properties. LBs are considered more similar to starch than glycogen due to their insolubility, elevated phosphate levels, and the increased chain length of their constituent glucose chains. We hypothesized that starch could serve as a proxy for LBs in a screen to identify a candidate therapeutic enzyme. We screened a panel of amylases to determine which possessed robust activity against starch, and therefore would likely degrade LBs. We found that only α-amylases showed significant activity, and pancreatic α-amylase had the highest capacity to degrade starch (Figure 5.1a). Pancreatic α-amylase is naturally secreted by the human pancreas for digestion of carbohydrates in the gut, cleaving α-1,4 linkages to yield maltose, maltotriose and low molecular weight oligosaccharides. It has previously been shown to digest LBs in tissue sections (579). Thus, pancreatic α-amylase is a putative candidate for degrading LBs. An AEF designated VAL-0417 was generated by fusing the humanized 3E10 IgG1 Fab heavy chain fragment with human pancreatic α-amylase and then coexpressed in HEK293-6E cells with the corresponding light chain as a secreted heterodimer (Figure 5.1b). The fusion was purified from HEK293-6E cell conditioned media by affinity chromatography, and the purity was determined by reducing and nonreducing SDS-PAGE (Figure 5.1c). The addition of β- mercaptoethanol (BME) dissociated the heavy and light chains, producing two

138 distinct bands of 75 and 25 kDa; in the absence of BME a single 100 kDa band was observed, corresponding to the intact Fab-amylase fusion. The specific activity of an α-amylase can be tested using a substrate that consists of seven glucose moieties and a chromophoric para-nitrophenyl moiety that is capped with an ethylidine group to block the action of other hydrolytic enzymes (Figure 5.1d) (580). Using this substrate, the specific activity of VAL-0417 was determined to be 10.5 units/mg. VAL-0417 activity was then tested against plant starch. After a 2-hour incubation, VAL-0417 degraded starch in a dose-dependent manner: 5 μg of enzyme released 61 μg of soluble glucan product (Figure 5.1e). We employed scanning electron microscopy (SEM) to visualize the effects of degradation on the starch granules. In the absence of enzymes, starch granules had a smooth, polyhedral or subspherical appearance (Figure 5.1f). Overnight treatment with VAL-0417 produced cavities in the granule surface, representing areas of active hydrolysis, and numerous partially degraded granules were visible (Figure 5.1f). When the ratio of VAL-4017 to starch was reduced by 10-fold, amylolysis was less extensive: fewer partially degraded particles were visible and most granules were only decorated with small pits (Figure 5.2). Other groups have similarly observed "pitting" of starch granules with α-amylase and other amylolytic enzymes (581, 582).

5.2.2 Isolation and characterization of LBs from LD mice

Polysaccharides (both glycogen and LBs) are typically purified from Epm2a-/- and Epm2b-/- mice based on a method described by Pflüger in 1909 where the tissue is heated in strong KOH and the glycogen is precipitated with alcohol (16, 583). The isolated particles are 15-65 nm in size when visualized by transmission electron microscopy (209), orders of magnitude smaller than the intact LBs observed in fixed tissue sections, and the method does not separate LBs from glycogen. It is well-established that heating changes the physiochemical properties of starch, leads to granule swelling, and enhances its sensitivity to enzymatic degradation (584). Therefore, it is likely that this protocol disrupts the native structure of LBs, intermixing the individual polyglucosan constituents of LBs with normal glycogen molecules. We developed a procedure for isolating pure and native LBs from tissue so that we could test the sensitivity of LBs to VAL-0417 degradation. A study from the 1960s showed that native LBs could be isolated from patient brain tissue by centrifugation steps and successive treatments with proteolytic enzymes (169). We designed a similar method for isolating native LBs from the brain, heart and skeletal muscle of Epm2a-/- mice (Figure 5.3a, see Section 2.1.3). Our procedure separates soluble glycogen (present in the supernatant) from the insoluble LBs (Figure 5.4, see Section 2.1.5). Final LB yields from each tissue type corresponded to ~30% of the total polysaccharide detected in the original tissue homogenate (Figure 5.3b,c). Phosphate content of

139 the final LBs from skeletal muscle was 5-fold higher than phosphate in normal muscle glycogen, which is in agreement with published results (Figure 5.3d) (209). Iodine spectral scans of the purified LBs were compared to those of commercial glycogen and amylopectin, the major component of plant starch to which LBs are most similar (Figure 5.3e). While glycogen has a peak maximum at ~440 nm when stained with Lugol's iodine, amylopectin has a right-shifted spectrum and a peak maximum at 560 nm due to its reduced branching (147). The spectra of the purified LBs were also shifted toward longer wavelengths, with peak maxima at 510-520 nm (Figure 5.3e). These data indicate that the purified LBs are less branched than glycogen, consistent with previous reports on Epm2a-/- and Epm2b-/- mice (180, 209). Final LB preparations were also stained with Lugol's iodine and examined by light microscopy. Minimal tissue debris was observed and the LBs tended to clump (Figure 5.3f,g,h). LBs from different tissues also had noticeably distinct morphologies. Brain LBs were generally spherical, sub-spherical or irregularly shaped (Figure 5.3f), and a range of sizes were observed: the maximum diameter of a single LB ranged from 1-10 μm, most being 2-4 μm (Figure 5.4a). These preparations also had a significant amount of granular material that stained less intensely with iodine; small amounts of this material were visible throughout the slide, but appeared to form a large aggregate in some fields of view (Figure 5.4d). Similar PAS-positive granular material has been reported in fixed tissue sections from patient brain biopsies or LD mice, which are referred to as dust-like particles, or simply "dust" (326, 570, 585, 586). Heart LBs were more often ellipsoid or ovoid in shape, frequently with pointed ends, although some were sub-spherical like the brain LBs (Figure 5.3g). Heart LBs also ranged in diameter/length from 1-10 μm, most of them being 2-4 μm (Figure 5.4b). Dust-like particles were also observed among the heart LBs, although no masses of granular material were observed as with the brain LBs (Figure 5.4e). LBs have been reported in LD patient cardiac tissue: in fixed and stained tissue sections, they are irregularly shaped or elongated with pointed ends (585, 587-589). Although these reports do not discuss dust-like particles in the heart, they appear to be present in the micrographs, particularly in the studies that involve immunostaining with an LB-specific antibody (587, 588). PAS-stained cardiac sections from malin KO mice show LBs of similar shape and size and diastase-resistant granular material (178, 179). LBs from skeletal muscle were the smallest and most homogenous in size and shape: the vast majority were ellipsoid and 1-3 μm long (Figures 3.3h, 3.4e). There appeared to be virtually no dust in these preparations. In LD patients, LBs are reported to similarly range in diameter from 0.2-2 μm in skeletal muscle (326). We also utilized SEM to visualize the purified LBs. We found that washing LBs in ethanol and allowing them to dry in a vacuum concentrator, the method used for visualization of starch granules, caused the LBs to aggregate and form very large clumps after being applied to carbon tape (Figure 5.5). It was difficult to

140 distinguish individual LBs within the clumps against the similarly textured carbon tape. As an alternative drying method, LBs were diluted in water, applied to a mounted silicon wafer and lyophilized prior to visualization. This method proved satisfactory. Micrographs showed that LBs from all three tissues had a similarly textured appearance (Figure 5.6), in contrast to the smooth surface of starch (Figure 5.1f). As observed via light microscopy, individual brain LBs up to 10 μm in diameter were observed, which were spherical or irregularly shaped (Figure 5.6a). Ovoid and ellipsoid LBs from heart and small ellipsoid LBs from skeletal muscle were also consistent with our light micrographic observations (Figure 5.6b,c). Visualization of skeletal muscle LBs at very high magnification (30,000×) revealed the fine texture of the bodies (Figure 5.6d).

5.2.3 Degradation of LBs by VAL-0417

Native LBs purified from different tissues possess unique morphology. To determine if VAL-0417 could degrade the different types of LBs, we incubated 50 μg of Epm2a-/- LBs from brain, heart and skeletal muscle with VAL-0417 and observed a consistent dose-response effect: 1 μg VAL-0417 released 6-9 μg of glucan into the soluble fraction, indicating 12-18% of the LBs had been degraded (Figure 5.7a). Increasing the dose to 10 μg of VAL-0417 released over twice as much, 15-19 μg of glucan, corresponding to 30-38% of the total substrate. A small amount of soluble glucan (5 μg or less) was observed in the untreated samples, so we visualized the soluble fractions of untreated and treated LBs by light microscopy. We observed some LB fragments in the soluble fraction of the untreated and treated brain LBs that did not sediment with centrifugation (Figure 5.8). We have observed that harsh mechanical agitation in vitro causes LBs to fragment, so the occasional fragments found in the soluble fraction result from the mild agitation that is necessary to keep the LBs in suspension during incubation with VAL-0417. We also purified LBs from skeletal muscle of 12-month-old Epm2b-/- mice and found they were indistinguishable from Epm2a-/- skeletal muscle LBs in appearance, size and iodine spectra (Figure 5.9). We observed a comparable dose-response effect of VAL-0417 on Epm2b-/- LBs: 1 μg and 10 μg VAL-0417 released 16% and 32% of the total, respectively (Figure 5.10a). In a similar experiment, we incubated LBs with 10 μg VAL-0417 overnight and then visualized the product by staining the insoluble fraction with Lugol's iodine. Untreated LBs had no change in appearance and seemed fully intact after the overnight agitation, but VAL-0417 treatment led to fewer iodine-positive LBs and the appearance of a filamentous degradation product (Figure 5.10b). Previous work has shown that various proteins are associated with LBs in vivo (179, 586). Our purification protocol removes these proteins, so it is possible that an intact protein coat could inhibit LB degradation by VAL-0417. To test this possibility, we measured LB load in crude muscle homogenates after incubation

141 with VAL-0417 using iodine absorbance. While virtually no absorbance was detected at 550nm in the WT samples with or without VAL-0417 treatment, treatment of the Epm2a-/- samples with VAL-0417 reduced iodine absorbance by >50%, indicating substantial LB degradation (Figure 5.10c). Since VAL-0417 is a potential therapeutic, it is important to define the soluble product(s) it releases from LBs. Pancreatic α-amylase is reported to release glucose, maltose, maltotriose and other short oligosaccharides (590). Amylolytic products can be separated and quantified via high-performance anion-exchange chromatography coupled with pulsed amperometric detection (HPAEC-PAD). We digested LBs from Epm2a-/- skeletal muscle with VAL-0417 for a total of 168 hours with additional enzyme added every 24 hours. At 24-, 48- and 72-hour time points, the soluble fractions of LB digestions were removed for HPAEC-PAD analysis. At all 4 time points, maltose and glucose were the major degradation products, followed by small amounts of oligosaccharides with 3, 4, 5 and 6 degrees of polymerization (DP-3, 4, etc., where each DP corresponds to one glucose unit) (Figure 5.10d, 5.11). Quantitation of the glucose and maltose peaks showed that maltose levels remained constant throughout degradation and glucose levels increased steadily over the 168-hour experiment (Figure 5.10e). Cumulatively, these data demonstrate that VAL-0417 robustly digests brain, heart and muscle LBs in vitro. It is not inhibited by endogenous proteins in muscle homogenates, and its degradation products are primarily glucose and maltose, with a few short oligosaccharides.

5.2.4 VAL-0417 uptake and polysaccharide degradation in situ

To degrade LBs in vivo, VAL-0417 must penetrate and remain active inside cells. Cell penetration of the parent 3E10 mAb and its fragments require functional expression of the membrane equilibrative nucleoside transporter 2 (ENT2), which is involved in a nucleoside salvage pathway and is ubiquitously expressed in rodents and humans (591-593). VAL-0417 uptake and activity were assessed in two cell lines: fibroblast-derived Rat1 cells accumulating normal glycogen and HEK293 cells engineered to accumulate polyglucosan (HEK293-PTG/PP1Cα) (Figure 3.12; see Section 2.1.11). After a 20-hour treatment with increasing doses of VAL-0417, cells were washed, lysed, and probed for ENT2 and VAL- 0417. ENT2 protein levels were stable in both lines, though Rat1 cells have slightly higher levels, and were unchanged with increasing concentrations of VAL-0417 (Figure 5.13a). Western analysis of VAL-0417 using an antibody for AMY2A revealed a dose-dependent increase in protein levels (Figure 5.13a). Polysaccharide per well (glycogen in Rat1 cells and polyglucosan in HEK293- PTG/PP1Cα cells) in the treated cells was also measured. Both Rat1 and HEK293-PTG/PP1Cα lines displayed a dose-dependent decrease in total polysaccharide, reductions of 30% and 26% of the untreated control, respectively (Figure 5.13b). Thus, VAL-0417 is both taken up by cells and degrades glycogen

142 and polyglucosan in situ.

5.2.5 In vivo uptake of VAL-0417 and LB reduction in Epm2a-/- mice

We have shown that VAL-0417 degrades LBs in vitro, penetrates cultured cells, and degrades glycogen and polyglucosan in situ. We next tested VAL-0417 uptake in an in vivo setting and biodistribution after peripheral injections. We performed intramuscular (IM) injections of VAL-0417 to determine if it can be detected in the injected muscle and other tissues and if it is quickly degraded and/or cleared. We injected 0.6 mg VAL-0417 into the left gastrocnemius muscles of WT C57BL/6 mice, then 2 and 24 hours post-injection, euthanized the mice and collected right and left gastrocnemii, right and left quadriceps, heart, liver, and brain. The same tissues were also collected 2 hours post-injection from control mice treated with phosphate-buffered saline (PBS). We designed a sandwich enzyme-linked immunosorbent assay (ELISA) to specifically detect VAL-0417 in tissue homogenates with very low background. At the 2-hour VAL-0417 post-injection time point, 3200 ng of VAL-0417 per mg protein (15% of the injected amount) was detected in the injected (left) gastrocnemius (Figure 5.14a). After 24 hours, VAL- 0417 levels dropped by 20-fold (to 151 ng per mg protein), 0.7% of the initial injected amount. At 2 hours, low levels of VAL-0417 (2-9 ng per mg protein, 0.02- 0.1% of the injected amount) were detected in the other muscles, heart, liver, and brain, indicating some VAL-0417 had entered circulation. After 24 hours, VAL- 0417 levels nearly returned to baseline in muscles, liver, and brain, but remained constant in the heart. Since blood circulates the entire body of a mouse within minutes, the very high levels of VAL-0417 in the injected muscle 2 and 24 hours post-injection strongly suggest that VAL-0417 is taken up by cells rather than being immediately cleared. The low levels of VAL-0417 in non-injected tissues indicates VAL-0417 entered the circulation. In mice, ENT2 expression levels are high in brain, muscle and heart, and low in liver (592). It is possible that these tissues took up a small amount of VAL-0417 from the circulation that was mostly cleared by 24 hours. ENT2 levels were probably not limiting since VAL-0417 levels were similar regardless of tissue differences in ENT2 expression. It is notable that VAL-0417 levels remained constant in heart even after 24 hours. We next tested whether IM injections of VAL-0417 reduced polysaccharide levels in vivo in Epm2a-/- mice. To determine polysaccharide levels in individual muscles, we utilized the Pflüger method rather than our native LB protocol because the Pflüger method recovers ~100% of the polysaccharides, enables more accurate and sensitive measurements from small amounts of tissue, and yields both normal glycogen and LBs. 10-month-old WT and Epm2a-/- mice were injected in the right gastrocnemius with either 0.6 mg VAL-0417 or PBS. Three injections were performed over the course of one week, on days 1, 4, and 7, and on day 8 the mice were euthanized. The injected muscles were collected for total polysaccharide quantification. Polysaccharide levels in the WT mice were similar

143 regardless of treatment, and consistent with reported skeletal muscle glycogen levels for C57BL/6 mice (Figure 5.14b) (209). In Epm2a-/- animals, VAL-0417 treatment reduced polysaccharide levels by 53% relative to the PBS-treated controls (Figure 5.14b). These data indicate that in vivo, VAL-0417 does not decrease normal glycogen levels, but degrades LBs in Epm2a-/- mice. Intravenous (IV) injections are sometimes preferable to IM injections to achieve rapid and extensive bioavailability of the drug to various tissues, but the preferred route depends on the drug (594). We tested whether a high dose of VAL-0417 administered intravenously could reduce LBs in the heart. Like skeletal muscle and brain, heart tissue has a very high LB load, and cardiac abnormalities have been reported in both LD mice and patients (595-597). Furthermore, our biodistribution experiment indicated there was less turnover of VAL-0417 in the heart compared to other tissues (Figure 5.14a). Four injections of either PBS or XX mg VAL-0417 were performed over the course of two weeks, on days 1, 5, 8 and 13. The mice were euthanized on day 14, and hearts were removed and flushed with PBS. A small section of tissue was excised for fixation and periodic acid-Schiff (PAS) staining, and the remaining tissue was flash frozen and analyzed for polysaccharide content via the Pflüger method. The glycogen content of the heart in WT animals is very low in comparison to skeletal muscle (598). We found that polysaccharide levels in PBS-treated WT animals were undetectable. Very low levels were detected in the WT animals treated with VAL-0417, which was not significantly different than the PBS-treated levels (Figure 5.14c). Since there was a delay prior to freezing the heart tissue in order to remove tissue for fixation, rapid post-mortem glycogen degradation may account for the absence of heart glycogen in the WT animals. In contrast, polysaccharide levels in the PBS-treated Epm2a-/- mice were very high in comparison to the WT animals and were reduced by 67% after the two-week treatment regimen with VAL-0417 (Figure 5.14c). Results from the fixed, PAS-stained heart sections were consistent with quantification via the Pflüger method. PAS-positive deposits, i.e. LBs, were abundant in PBS-treated Epm2a-/- tissue (Figure 5.14d). Epm2a-/- tissues treated with VAL-0417 showed a marked reduction in LBs (Figure 5.14d). No LBs were observed in WT tissues regardless of treatment (Figure 5.14e). HALO image analysis software was used to quantify the staining results. Similar to the biochemical analysis, we observed a 63% reduction in the average staining intensity of PAS-positive deposits in Epm2a-/- tissue with VAL-0417 treatment (Figure 5.14f). WT animals displayed undetectable levels, consistent with the biochemical results (Figure 5.14f). Since LBs in the Epm2a-/- mice are cytosolic, these data indicate that after IM or IV injection, VAL-0417 penetrates cells, remains active, and degrades LBs in vivo. Cerebral LB accumulation drives epilepsy and neurodegeneration in LD (182). Most drugs do not cross the blood-brain barrier, so any therapy targeting cerebral LBs must be deliverable directly into the CNS. There are multiple routes of drug

144 administration to the CNS that bypass the blood-brain barrier. The intrathecal (IT) route involves injection of the drug directly into the spinal cord or the subarachnoid space. Drugs can also be injected directly into the ventricles of the brain via intracerebroventricular (ICV) injection. Both IT and ICV are considered safe and well-tolerated drug delivery routes in both pediatric and adult populations, but the procedures have different risks and benefits (599). We first tested IT versus ICV administration of either PBS or VAL-0417 in WT mice over a period of 3 days. The IT catheter was surgically implanted intrathecally at the lumbar region. The ICV cannula was implanted in the cerebral lateral ventricle. The IT catheter and ICV cannula were each attached to an osmotic pump, through which a continuous infusion of VAL-0417 (0.08 mg/day) or PBS was administered for 3 days. Mice were sacrificed on the fourth day and brains were removed and sectioned as shown (Figure 5.15a). Homogenates from each section were used in ELISA assays to detect VAL-0417 levels (Figure 5.16a) and amylase assays were used to quantify VAL-0417 activity (Figure 5.16b). Little to no ELISA signal was detected in tissue treated with PBS either by IT or ICV, and amylase activity in the tissue was very low. In tissues treated with VAL-0417, VAL- 0417 levels detected by ELISA were markedly higher with ICV administration compared to IT. In most slices, amylase activity was also higher with ICV treatment of VAL-0417, with the exception of slice 6. After ICV administration, VAL-0417 levels and activity were nearly equivalent in all slices, indicating this route facilitates homogenous distribution of the drug throughout the brain. In contrast, after IT administration VAL-0417 levels and activity were high only in the most caudal sections, which are nearest the brain stem. Since LBs are abundant throughout the cortex and cerebellum, a thorough distribution is necessary and thus ICV administration is preferable to IT. Furthermore, these data are indicative of cellular uptake of VAL-0417 in the brain, since the drug was not cleared in the period after infusion (day 3) and before euthanasia (day 4). These data also demonstrate that the drug remains active after uptake. To target cerebral LBs, we administered VAL-0417 directly into the cerebral lateral ventricle of 5-month-old Epm2a-/- mice. A continuous infusion of VAL-0417 (0.08 mg/day) or PBS was administered for 28 days via the ICV cannula and pump. Mice were then euthanized, and brains were removed and sectioned as shown (Figure 5.15a). Polysaccharides were isolated from each section, quantified, and normalized to protein levels. Polysaccharide levels relative to protein content were lowest in the rostral section and highest in the caudal section (Figure 5.15b). This is the first-time polysaccharide distribution across different brain regions has been quantitated in Epm2a-/- mice, and is consistent with the high levels of LBs observed in histological sections from brainstem and cerebellum (179, 180, 325). Treatment with VAL-0417 reduced the polysaccharide load in all sections, with statistical significance in the two most caudal sections that contain the highest glucan load (Figure 5.15b). We also quantified whole brain polysaccharide relative to tissue weight (Figure 5.15c). Polysaccharide levels in

145

PBS-treated animals are consistent with what has been previously reported for Epm2a-/- animals of this age, i.e. 3-4 times higher than WT levels (2 μmol glucose/g tissue) (159, 209). Treatment with VAL-0417 reduced total brain polysaccharide in the Epm2a-/- mice by 43% (Figure 5.15c), only 2-fold higher than normal brain glycogen levels.

5.3 Discussion

LD is caused by mutations in one of two genes both encoding enzymes that impact glycogen structure. Efforts from multiple labs have shown that LBs are the pathogenic epicenter of LD and removal of LBs has become a therapeutic target for treating LD (564). In this study, we combine insights from the fields of antibody- based therapeutics, GSDs, carbohydrate metabolism, and LD to design an AEF that degrades LBs in vitro, penetrates cells and reduces polysaccharide levels in situ, and decreases polysaccharide load in vivo via both IM and ICV injection. This is the first time a significant LB reduction has been achieved in adult LD mice by the use of a drug. Antibody-based drugs have made a major impact on the treatment of cancer, autoimmune and inflammatory diseases with worldwide revenues of ~$89 billion in 2016 (600). Antibody-based drugs are typically whole IgGs designed to bind and block receptors, deplete target cells, downregulate a receptor or initiate signaling. Bispecific antibodies, antibody fragments and antibody-drug conjugates (ADCs) are also being developed across multiple diseases, further expanding the diversity of therapeutics utilizing antibody targeting (600-603). Herein, we report a novel class of antibody-based drugs with a similar construction to ADCs: antibody-enzyme fusions (AEFs). Rather than delivering a cytotoxin, AEFs deliver an enzyme to a specific target. The antibody or antibody fragment is essential both for targeting and cellular penetration. AEFs have recently been reported for the treatment of two congenital disorders: a muscle wasting disorder known as X- linked myotubular myopathy and the GSD Pompe disease. In the first study, the deficient lipid phosphatase myotubularin fused to an Fv fragment was delivered by intramuscular injection and improved muscle function in a mouse model of myotubular myopathy (575). In the second, a Fab fragment fused to GAA delivered intravenously reduced glycogen load and alleviated pathology in GAA-deficient mice (576). These two studies illustrate that enzyme replacement therapy, which can be problematic for cytosolic enzymes, can be enhanced by fusion of the replacement enzyme with a cell-penetrating antibody fragment. We now demonstrate that VAL-0417, an AEF comprised of a Fab fragment fused to pancreatic α-amylase, degrades LBs and reduces polysaccharide loads in vivo. Unlike the previous two studies, VAL-0417 does not replace a missing enzyme, but rather delivers a natural human enzyme to a new target. Since LBs are structurally similar to plant starch, and -amylase naturally degrades dietary starch in the human digestive tract, -amylase is a logical choice for the VAL-0417

146 payload. It is important to distinguish normal glycogen from polyglucosan. Polyglucosan bodies (PGBs) have historically been defined as "small, non-membrane-bound cytoplasmic structures composed largely of unusual glucose polymers" (261). They are typically a pathological feature of GSDs, although PGBs known as corpora amylacea are a hallmark of normal aging and neurodegenerative disorders (79). PGBs are not found in all GSDs, and not all PGBs are identical with respect to physiochemical structure. The abnormal polyglucosan found in Cori's disease is characterized by short outer chains, while the polyglucosan of LD, Anderson disease, Adult Polyglucosan Body Disease, and Tarui disease contain very long chains (261, 604). The Pflüger method, originally designed to purify normal glycogen, also precipitates polyglucosan. Herein we refer to Pflüger- isolated material as polysaccharide, since it contains both glycogen and polyglucosan. Furthermore, the Pflüger method includes a boiling step that solubilizes polyglucosan and destroys LB superstructure, which is not included in our LB purification procedure. LBs (and PGBs) may be considered a higher-order structure of polyglucosan. The LB purification protocol we designed does not destroy the native LB structure and separates LBs from normal, soluble glycogen. LBs from the different tissues displayed varied morphologies. Brain LBs were the most irregular in size and shape, included dust-like particles, and reached the largest size. LBs from brain exhibited the greatest variation in size, consistent with clinical observations: LBs from patient brain micrographs are 3-40 μm and LB dust-like granules are even smaller (322, 326). A recent microscopic study of Epm2b-/- mice showed that astrocytic LBs are on average smaller than neuronal LBs, more irregularly shaped, and clustered, while neuronal LBs are larger, more spherical, and isolated (324). LBs in the heart were also surprisingly large and included some dust-like particles. LBs from skeletal muscle appeared well organized: they were very small and homogenous in shape, and no dust-like particles were present. It is not clear why such variation in LB size and shape exists in the brain and heart, in contrast to the homogeneity found in skeletal muscle. One possibility is that LB size is limited by the dimensions of the cell or subcellular region it occupies. For example, in patients, canines, and mice, LBs can be found in the cell bodies of Purkinje neurons, which can be up to 40 μm in diameter (326, 586, 605). In astrocytes, LBs are typically found in processes that can be very long but are only a few microns wide at most (324, 325). However, cell size constraints do not explain the difference between heart and skeletal muscle LBs, since cardiac and skeletal myocytes are similar in size. It is well known that glycogen metabolism is remarkably different in astrocytes, neurons, cardiomyocytes and skeletal myocytes. Skeletal myocytes may sequester LBs more efficiently than cardiomyocytes, in smaller, discrete units. In brain and heart, the dust-like particles may represent a pre-aggregated form of polyglucosan, while the larger particles may represent polyglucosan aggregates that have been sequestered and compacted by the cell. Active cellular sequestration of

147 polyglucosan has been proposed to explain the large, dense, and spherical LBs observed in some neurons (586). Although LBs from all three tissues had iodine spectral peak maxima at similar wavelengths, indicative of similar degrees of branching, the higher amplitude of the spectra of brain and heart LBs suggests that they may contain longer chains compared to LBs from skeletal muscle (147, 148). Indeed, Nitschke et al. showed using HPAEC-PAD that in Epm2a-/- and Epm2b-/- mice, brain glycogen had a more exacerbated shift in chain length distribution than muscle glycogen (159). Although the brain is most severely affected in LD and a neurological phenotype predominates, heart failure and arrhythmia have been reported in LD patients, and Epm2a-/- and Epm2b-/- mice have metabolic cardiomyopathy (595-597). Glycogen metabolism is variable with tissue type: glycogen synthase and glycogen phosphorylase have tissue-specific isoforms that are disparately regulated by cellular signals (113); the ultrastructure of glycogen is different under different metabolic conditions and in different organs (337, 373); subcellular glycogen pools are variable (239); and both phosphate levels and average chain length differ with species and tissue type (133, 164). Due to these differences, the LBs may be more pathological in brain and heart than in skeletal muscle. Conversely, different cell types may be more sensitive to the presence of the LBs, and/or the metabolic perturbations induced by the LBs may have distinct consequences in different organs. Tissue-specific differences are important to consider throughout the clinical development of VAL-0417. ICV administration of VAL-0417 may attenuate neurological symptoms, but cardiac, liver and/or muscular pathology may require treatment by intravenous or intramuscular injection of VAL-0417. We found that VAL-0417 primarily releases maltose and glucose from LBs, and glucose levels gradually increased over time, as expected for an -amylase. Cytosolic free glucose would likely be phosphorylated to produce glucose-6- phosphate, entering central carbon metabolism. Although maltose is not considered a typical participant in intracellular energy utilization, studies have shown that maltose can be transported into and out of cells and metabolized in cell culture and humans after intravenous infusion (606, 607). The small oligosaccharides (DP-3, DP-4, etc.) would likely remain cytosolic and inert until they are further metabolized to glucose and maltose. Importantly, IM injections of VAL-0417 in Epm2a-/- mice reduced the polysaccharide load (i.e. LBs and glycogen) to WT levels but did not decrease glycogen levels in WT skeletal muscle. In contrast, we observed that VAL-0417 reduced glycogen levels in cell culture. It should be noted that the cell culture studies were performed in high glucose media, known to promote glycogen accumulation in cultured cells. Thus, it is possible that a reduction in polysaccharides with VAL-0417 treatment is only detected when the polysaccharide accumulates above a baseline level, as glycogen accumulates in cultured cells or LB in LD tissues. When glycogen levels drop below a certain threshold due to degradation by VAL-0417, the cellular

148 machinery may respond to restore those levels. Glycogen is likely to be maintained at a certain baseline level in vivo in muscle, and this may explain why no change was observed in WT muscle treated with VAL-0417. The observation that VAL-0417 reduced polysaccharide levels in Epm2a-/- muscle and brain indicates LBs are degraded in vivo. We have initiated further studies to define the metabolic changes invoked by VAL-0417 in WT and LD mice. Given these exciting results, we are now performing pre-clinical experiments with VAL-0417 to define the minimal in vivo efficacious dose, the therapeutic window for treatment, and the necessary dosing frequency. We have established a robust VAL-0417 ELISA-based quantitation method that will be utilized to define the pharmacokinetic and pharmacodynamic parameters. Further, we are actively developing biomarkers that could be used to assess VAL-0417 treatment efficacy and monitor ongoing therapy. In the current study, we observed dramatic effects with VAL-0417 in brain of 6-month old mice and muscle of 10-month old mice. Moving forward, we will define the range of doses that most effectively clear LBs in animals at various ages as well the dose that effectively prevents the appearance of LBs in younger animals. We have demonstrated that VAL-0417 is a putative precision therapy for LD, an intractable epilepsy. Currently, >30% of epilepsies, including LD, are resistant to available antiseizure drugs (ASDs) and are known as refractory epilepsies (608). More than 35 different iterations of ASDs have been discovered over the past 150 years with most of them acting primarily on ion channels or neurotransmitters. Due to this focus on one mode of action, the high proportion of refractory epilepsies has remained constant since the 1850s (608, 609). The underlying causes of these refractory epilepsies must be identified in order to develop more effective therapies. As the molecular basis of LD has been defined, we have now developed an AEF that targets its underlying cause. As molecular etiologies in other epilepsies become clear, precision therapies, possibly also utilizing the AEF platform, become feasible. LD is also a non-classical GSD. There are over 14 different types of GSDs characterized by glucan accumulations, and 1 in 20,000 people have some form of GSD though the individual types are considered rare (165, 610). While the glycogen in some GSDs is of normal structure like in Pompe disease, other GSDs such as Cori disease and Andersen disease are characterized by polyglucosans, like LD (165). Two AEF drugs (VAL- 0417 and Fab-GAA) have now been shown to target, respectively, cytosolic polyglucosan (the present study) and normal glycogen in both the cytosol and lysosome (576). These drugs could be repurposed and/or modified to target glycogen in other GSDs. Not only is VAL-0417 the first drug with the potential for providing a significant clinical benefit to LD patients, it is an example of a precision therapy and expands the repertoire of antibody-based drugs that can be used to treat human disease.

149

Figure 5.1 The antibody-enzyme fusion VAL-0417 degrades starch, a proxy for LBs. (A) Degradation of starch by a panel of amylases. 1 mg of starch was incubated with 5 μg of amylase overnight with constant agitation, then soluble and insoluble fractions were separated by centrifugation. Degradation product in the soluble fractions was quantified by measuring glucose equivalents (i.e. glucan). Reactions were performed in triplicate and data shown are means ± SD. (B) Schematic representation of VAL-0417. The gene encoding the human IgG1 Fab heavy chain fragment (H) was fused to AMY2A, encoding human pancreatic α-amylase, and coexpressed with the gene encoding the human light chain (L) in HEK293-6E cells. Heavy and light chain signal peptides were included to facilitate proper folding and assembly of the Fab fragment. The predicted molecular weight of each polypeptide is shown. (C) Purity of VAL-0417 assessed by reducing (+BME) and nonreducing (-BME) SDS-PAGE. (D) Chemical structure of the α-amylase-specific substrate E-G7-pNP. Reducing (inner) and non-reducing (outer) ends of the substrate are shown. (E) Degradation of 1 mg starch after a 2-hour incubation with increasing amounts of VAL-0417, expressed as glucan released in the soluble fraction. Mean ±SD of triplicate reactions are shown. (F) Scanning electron micrographs of starch granules in the absence of treatment and after overnight treatment with VAL-0417. Samples were visualized at 2kV by an FE Quanta 250 scanning electron microscope.

150

Figure 5.2 Starch degradation with different VAL-0417:starch ratios. (A) Degradation with low ratio (10 μg VAL-0417, 1 mg starch). (B) Degradation with high ratio (10 μg VAL-0417, 100 μg starch).

151

Figure 5.3 A novel protocol for isolating native LBs from LD mice. (A) LB purification scheme. (B) Polysaccharide was purified at different steps in the protocol via the Pflüger method and quantitated via glucose measurement following hydrolysis. Initial tissue weights and total polysaccharide at each step are shown. (C) Polysaccharide yields normalized to tissue weight. Triplicate samples were removed from each fraction and each measured in triplicate. Mean ±SD are shown. (D) Phosphate content of LBs from skeletal muscle and normal rabbit muscle glycogen. Mean ±SD of triplicate measurements are shown. (E) Normalized iodine spectra of purified LBs compared to commercial liver glycogen and amylopectin. Spectra shown are an average of 3 replicates. (F) Brain, (G) heart, and (H) skeletal muscle LBs stained with Lugol's solution and visualized using a Zeiss Axioimager Z1.

152

Figure 5.4 Iodine staining of LB purification fractions. (A) 50 μL samples were removed from homogenate, supernatant, and pellet fractions during preliminary LB purifications from skeletal muscle, boiled for 30 minutes, and clarified by centrifugation. 30 μL of the clarified sample was diluted with 15 μL water and stained with 50 μL 1x Lugol's solution in a microplate. Commercial liver glycogen and amylopectin were also stained as controls. (B) Concentration vs. iodine absorbance at 550 nm of Pflüger-isolated polysaccharide from skeletal muscle at each step of the purification scheme. Mean ±SD of triplicate reactions are shown.

153

Figure 5.5 Size distribution of LBs purified from different tissues and appearance of dust-like particles. Frequency distribution of LBs from each tissue type: (A) brain (total count: 473), (B) heart (total count: 247), and (C) skeletal muscle (total count: 252). Two fields of view from (D) brain and heart (E) LB preparations containing numerous dust- like particles are shown. Black boxes indicate areas that have been magnified and are labeled correspondingly (e1, e2, etc.). Occasionally very large bodies were observed (10-20 μm). They were difficult to distinguish from tightly packed clumps, so they were excluded from the size distribution histograms.

154

Figure 5.6 Scanning electron micrographs of LBs from brain, heart and skeletal muscle, washed in ethanol, dried, and applied to carbon tape. Samples were visualized at 2kV under high vacuum by an FE Quanta 250.

155

Figure 5.7 Scanning electron micrographs of isolated LBs. Scanning electron micrographs of LBs purified from brain (A), heart (B), and skeletal muscle (C) shown at identical scale. LBs from skeletal muscle are shown at a higher magnification in (D). Samples were visualized at 2 kV using an FE Quanta 250.

156

Figure 5.8 Soluble fractions of untreated and treated BrLBs after degradation reactions. Samples from the soluble fractions and total reactions were stained with Lugol's solution and visualized using a Zeiss Axioimager Z1.

157

Figure 5.9 Purification of LBs from skeletal muscle of 12 month old Epm2b- /- mice. (A) LB purification yields. Mean ± SD of triplicate measurements are shown. (B) Epm2b-/- LBs stained with Lugol's solution and visualized by light microscopy. (C) Comparison of size distribution of skeletal muscle LBs from Epm2a-/- and Epm2b-/- mice. (D) Normalized iodine spectra of Epm2b-/- and Epm2a-/- skeletal muscle LBs. The spectral data are an average of 3 replicates.

158

Figure 5.10 VAL-0417 degrades LBs in vitro. (A) Degradation of 50 μg LBs from different tissues after overnight incubation with 0, 1, or 10 μg VAL-0417. (B) LBs from different tissues after overnight incubation +/- 10 μg VAL-0417. Insoluble fractions were resuspended post- degradation, stained with Lugol's solution, and visualized using a Nikon Eclipse E600. (C) Iodine absorbance in pellet fractions after incubation of WT and Epm2a-/- skeletal muscle homogenates +/- 25 μg VAL-0417. In (A), (C), and (E) mean ± SD of triplicates are shown. (D) HPAEC-PAD chromatogram of degradation product after incubation with VAL-0417 for 168 hours. Peak identities were determined based on the elution profile of degradation buffer and glucan standards. The chromatogram is representative of triplicate experiments, and glucose/maltose quantification from the replicates is shown in (E). (E) Quantification of glucose and maltose at various time points throughout the degradation reaction of LBs with VAL-0417. Product released are expressed as a percentage of total LBs.

159

Figure 5.11 HPAEC-PAD chromatograms of control samples and additional time points from LB degradation with VAL-0417. (A) Overlay of control chromatograms: degradation buffer only (black line), LBs only (red line), and a mixture of 1 μg glucose and 1 μg maltose standards for quantitation (blue line). A small amount of maltose and low molecular weight oligosaccharides were detected in the LB control. Additional time points from the 168 hour LB degradation experiment with VAL-0417: (B) 24 hours, (C) 48 hours, and (D) 72 hours. Each chromatogram is representative of triplicate chromatograms. Small shifts in retention time (≤1 min) are typical between HPAEC-PAD runs. (E) Maltrin100, a mixture of low molecular weight oligosaccharides, was used as a standard for verifying degrees of polymerization (DP), i.e. number of glucose units.

160

Figure 5.12 Polysaccharide levels relative to protein concentration in three cell lines. Polysaccharide content of Rat1, HEK293, and HEK293 stably expressing PTG and PP1Cα. Data are expressed as the mean of 12 replicates ± SE.

161

Figure 5.13 VAL-0417 uptake and polysaccharide reduction in cell culture. (A) Rat1 cells and HEK293 cells stably expressing PTG and PP1Cα after 20-hour treatment with increasing concentrations of VAL-0417. VAL-0417 levels were assessed by Western blotting using an anti-AMY2A antibody. ENT2 levels were also quantified by Western blotting. (B) Polysaccharide level per well in Rat1 and HEK293-PTG/PP1Cα cells after 20-hour treatment with VAL-0417. Data shown are a mean of 4 measurements ± SE. Statistical significance is indicated: * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001.

162

Figure 5.14 VAL-0417 reduces LB load in vivo after intramuscular (IM) or intravenous (IV) injection. (A) Biodistribution of VAL-0417 levels determined by sandwich ELISA after IM injection of WT mice with PBS or 0.6 mg VAL-0417. (B) Quantification of polysaccharide in the injected gastrocnemii of WT and Epm2a-/- mice after 3 IM injections of PBS or VAL-0417 administered over the course of one week. (C) Quantification of polysaccharide in the heart of WT and Epm2a-/- mice after 4 IV injections of PBS or VAL-0417 administered over a two-week period. Polysaccharides in (B) and (C) were isolated via the Pflüger method and quantified by glucose measurement following hydrolysis. (D, E) PAS-stained heart tissue of WT (D) and Epm2a-/- (E) mice after IV treatment regimen. Intensely staining PAS- positive deposits are LBs (purple). Tissues were counterstained with hemotoxylin (blue). (F) Quantification of PAS staining using HALO image analysis software.

163

Figure 5.15 Comparison of intrathecal (IT) versus intracerebroventricular (ICV) administration in WT mice. (A) Distribution of VAL-0417 levels using the sandwich ELISA in PBS- and VAL- 0417-treated WT brain slices after intrathecal (IT) or intracerebroventricular (ICV) infusion. (B) Distribution of VAL-0417 activity in the same slices measured using the BioVision amylase activity assay. Data shown are a mean of 3 measurements ± SE.

164

Figure 5.16 VAL-0417 reduces LB load in vivo after continuous ICV infusion. (A) Representative photograph of six 2.0 mm coronal sections. Blue dye was injected through the cannula into the lateral ventricle to verify its location within slice 2. (B) Polysaccharide content normalized to protein content in Epm2a-/- brain slices after 28 days of continuous ICV infusion of VAL-0417 or PBS. (C) Total brain polysaccharide normalized to tissue weight of PBS and VAL-0417 treated Epm2a- /- mice. In (B) and (C), at least three technical replicates were performed for each isolated slice to determine an average per slice per animal, and data shown are the mean from each treatment group ± SE. The numbers of animals in each treatment group (n) are shown. Statistical significance is indicated: ns p > 0.05, * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001.

165

THE STRUCTURE OF LAFORA DISEASE POLYGLUCOSAN BODIES

6.1 Introduction

Polyglucosan bodies (PGBs) are a common feature of glycogen storage disorders (GSDs), neurodegenerative diseases, and physiological aging (79, 80, 611). PGBs range in size from 2 to 50 μm in diameter, and unlike the proteinacious inclusion bodies of neurodegenerative diseases they are primarily made of an abnormal glucose polymer called polyglucosan (261). Subtle chemical and ultrastructural differences distinguish the PGBs of various pathologies (79). What unites them is their chemical resemblance to mammalian glycogen and plant starch, the two major forms of glucose storage in living organisms (139). Starch-like structures in the brains of elderly patients were first described in 1836 by J. E. Purkinje, who named them corpora amylacea (literally "starch bodies"). In 1854, Rudolf Virchow observed a substance in the diseased nervous system that stained with iodine in a manner similar to plant starch, coining the term "amyloid" (612). It was later discovered that many amyloid deposits were proteinacious, so the term now typically refers to the inclusions of Alzheimer's and other amyloidoses (613, 614). But corpora amylacea and the later-described PGBs are, in fact, true amyloid: in addition to their similarity in size and shape and staining characteristics to starch, they are comprised primarily of insoluble glucose polymers containing significant amounts of phosphate (79, 615, 616). Since glycogen synthase is the only enzyme able to catalyze glucose polymerization in mammals, PGBs are considered a pathological result of aberrant glycogen metabolism (80, 114, 261). Like other PGBs, LBs from tissues of Lafora disease (LD) patients were described as starch-like in the early- and mid-1900s based on their chemical and structural characteristics (169, 321). A peculiar link between carbohydrate metabolism in humans and plants was made when it was discovered that laforin is a functional analog of a plant protein called starch excess 4 (SEX4) (183). SEX4 deficiency in plants leads to a starch excess phenotype whereby the plants synthesize a glucan that they cannot efficiently break down. Thus, the plant phenotype is very reminiscent of the LBs found in LD patients and mouse models (208, 617). It was later shown that polysaccharides purified from LD muscle and brain tissue had an abnormal chain length distribution (CLD) compared to wild- type glycogen, increased levels of covalent phosphate monoesters, and that laforin can remove this phosphate (158, 159, 209, 210). These important studies provided quantitative evidence that the polysaccharide in LD tissues resembles plant amylopectin, the branched component of starch, on a molecular level, and that phosphate is involved in its metabolism. To our knowledge, comparable structural studies on purified polyglucosan from other diseases have not yet been conducted.

The polysaccharides from Epm2a-/- and Epm2b-/- mice that have been characterized in vitro were purified based on a method first described in 1909 by Pflüger for purifying glycogen, in which the tissue is boiled in KOH and the polysaccharides are precipitated with ethanol (16, 583). The method does not separate glycogen from the LBs, and the Pflüger-purified polysaccharides differ in morphology from LBs observed via microscopy in LD tissue sections. Analysis of the Pflüger-purified particles by transmission electron microscopy (TEM) showed that they are larger and more aggregated than wild-type glycogen (209), but they are still much smaller (15-65 nm) than the micron-sized LBs observed in tissue sections. Another group showed that the average molecular weights of the Pflüger-purified particles from Epm2a-/- and Epm2b-/- mice were only 2- to 3-fold larger than glycogen, again much smaller than LBs observed in tissue sections (158). Thus, purification of LBs by the Pflüger method does not separate LBs from glycogen and likely changes the morphology of the LBs. It is conceivable that LBs are aggregates of polyglucosan (i.e. abnormal polysaccharide) molecules, and that the Pflüger method breaks apart the LBs and intermixes the polyglucosan molecules with normal glycogen particles. In the case of starch, many studies have shown that the application of heat and moisture results in solubilization and irreversible disruption of the starch granular structure (618). However, the deconstruction of LBs to polyglucosan molecules has not been investigated. An enormous body of work has been performed to define the chemical, physical and structural properties of different types of starch, which vary depending on plant species and tissue type (reviewed by (145)). One of the major features that distinguishes starch from glycogen is its semi-crystalline, lamellar organization, which can be detected with X-ray scattering techniques such as wide-angle and small-angle X-ray scattering (WAXS and SAXS, respectively). Treatment of starch granules with mild hydrochloric acid, a process known as lintnerization, preferentially hydrolyzes the amorphous regions, leaving crystalline regions intact and providing valuable information about granule architecture (619-621). Despite the similarities between PGBs and starch, comparable methods have not yet been applied to investigate the structure of PGBs. We have described a procedure for isolating native LBs from Epm2a-/- and Epm2b-/- mice using a novel protocol that does not rely on the Pflüger method and separates native LBs from glycogen (Chapter 3). Analysis of these LBs by light microscopy and scanning electron microscopy showed that native LBs from brain, heart, and skeletal muscle have distinct sizes and morphologies, ranging from 2 to 10 μm or more, depending on tissue type. In the present study, we use methods previously utilized to characterize starch to investigate the physiochemical and structural characteristics of LBs isolated from different tissues. Combining these methods with thermal, mechanical and chemical treatments to disassemble LBs allows for a model of how polyglucosan molecules aggregate to form LBs in LD. Although not all PGBs are identical, the corpora amylacea of aging and neurodegeneration and the PGBs of various GSDs bear

167 striking similarities to LBs. PGBs may represent a common pathological phenomenon resulting from misregulated glycogen metabolism, which has particularly detrimental consequences in the brain.

6.2 Results

6.2.1 Confocal micrographs of iodine-stained LBs

Lugol's iodine is a chemical stain that has been used for over a century to detect and differentiate polysaccharides such as glycogen, amylose and amylopectin. Lugol's iodine intercalates within the glucan chains to form a chromophoric complex that provides information about polysaccharide architecture. Amylose, which is α-1,4 linear chains, produces a deep blue-black color with Lugol's iodine, due to the very stable association of the triiodide ions with the long helical chains (622). Since amylopectin contains branches in addition to helix-forming glucan chains, it produces a less stable complex with triiodide, giving a reddish-purple color. Stained amylose and amylopectin maximally absorb around wavelengths of 650 nm and 550 nm, respectively (148). Since glycogen contains shorter chains and more frequent branching than amylopectin, it stains poorly with Lugol's iodine, and yields an absorbance maximum at 440 nm, although this varies slightly depending on glycogen source, i.e. glycogen from different tissues and/or organisms (151). LBs, corpora amylacea, and other PGBs stain brown or purplish-brown with Lugol's iodine, suggesting they have a branching degree that is more similar to amylopectin than glycogen (616, 623, 624). We showed that native LBs isolated from skeletal muscle, heart tissue, and brain of Epm2a-/- mice stained intensely reddish-brown in Lugol's iodine, with an absorbance maximum at 510-520 nm (Figure 5.3). The iodine-stained LBs not only absorb at this wavelength, but they also produce red fluorescence when excited by a 561 nm confocal laser. Confocal microscopy has previously been used as a sensitive method for visualizing starch granules in vitro and in planta (625, 626). Z-stack images of iodine-stained starch granules visualized using the TRITC channel and differential interference contrast (DIC) at high magnification show that only the smooth, outer edge of the granule is stained (Figure 6.1a, 6.2a,b). This staining pattern suggests the iodine does not penetrate beyond the outer edge, which is not surprising given the densely packed, semi-crystalline interior of starch granules. Isolated LBs from Epm2a-/- skeletal muscle (SmLBs) have a characteristic ovoid morphology and are typically 1-3 μm in length (Figure 5.3h, 5.5c). Like starch, they also fluoresce brightly with 561 nm excitation after iodine staining, but they stain more thoroughly, suggesting their smaller size and less dense interior facilitates greater penetration of the triiodide ions (Figure 4.1b, 6.2c,d). LBs isolated from Epm2a-/- hearts (HtLBs) are larger in size (typically 2-4 μm, up to 10 μm in length) and often more elongated than SmLBs (Figure 5.3g, 5.5b). They

168 also stained homogenously with iodine, with visible interior texture and a brighter rim (Figure 6.1c, 6.2e,f). The textured surface of the LBs was evident using DIC and consistent with what we observed via scanning electron microscopy (Figure 5.7). Brain LBs (BrLBs) were also larger in size (typically 2-4 μm, some >10 μm in diameter) and varied in morphology more than skeletal muscle and heart LBs. Some LBs were very round and homogenously stained, with a bright rim and textured appearance, similar to heart LBs except that they are not typically elongated (Figure 6.1d). Some LBs had a very jagged edge with a dark core (Figure 6.1e). Since both LB types were of comparable size, the lack of central staining in Figure 6.1e suggests that some bodies had a more densely packed center, like that of starch. Iodine-positive dust-like granules were also evident around LBs from heart and brain (Figure 6.1, 6.2), consistent with our previous observations (Figure 5.5). Distinct types of LBs have been observed in fixed brain tissue from LD patients and mouse models. Van Hoof's type I LBs have been described as homogenous and uniformly granular, while type II LBs have a dense core that stains more intensely with the carbohydrate-specific stain periodic acid-Schiff (PAS) (452, 570, 586). By TEM, type I LBs contain ramifying and randomly arranged, finely curved fibrils; in type II LBs, the fibrils are densely packed in the core, and were surrounded by radiating, less densely packed fibrillar layer (452). Small, dust-like granules appeared to be made of the same material as type I LBs. It is possible that the bodies in Figure 6.1d and e correspond to types I and II, respectively. Smaller, amorphous and granular BrLBs would likely also be type I.

6.2.2 Differences in chain length distribution (CLD) between LB types

In addition to the distinct morphologies of LBs from different tissues, data from our lab and others suggest that the CLD of LBs also varies depending on tissue origin (Chapter 5) (159). Polysaccharides purified via the Pflüger method from Epm2a-/- and Epm2b-/- skeletal muscle contain significantly more glucan chains with 14-36 degrees of polymerization (i.e. DP14-DP36), than polysaccharides from wild-type (WT) mice, and significantly fewer short chains (DP5-11). This shift in CLD was even more exacerbated in polysaccharides purified from Epm2a-/- and Epm2b-/- brain. As mentioned before, the Pflüger method does not separate glycogen from LBs, and thus the CLD represented a mixture of glycogen and LBs, not pure LBs. We showed that in our native LB purification, a low-speed centrifugation step after tissue homogenization separates the LBs and glycogen into pellet and supernatant fractions. Importantly, only the pellet fraction stains reddish-brown with iodine, indicative PGBs and not glycogen (Figure 5.4). All purified LB types maximally absorbed Lugol's iodine at 510-520 nm, suggesting they have a similar degree of branching, but equivalent amounts of LBs produced peaks of varying amplitude: BrLBs produced the highest amplitude, followed by HtLBs, then SmLBs. The higher amplitude of BrLBs suggests that they contain

169 even longer chains than HtLBs and SmLBs (147, 148), consistent with previous findings (159). In order to define these parameters for native LBs, we determined the CLD of purified SmLBs, HtLBs, and BrLBs using fluorescience-assisted capillary electrophoresis (FACE) (Figure 6.3). Samples were enzymatically debranched then linear oligosaccharides fluorescently labeled and separated using Beckman Coulter PA800 plus instrument equipped with a laser induced fluorescence detector. BrLBs had the highest proportion of long chains: 13.8% of the detected oligosaccharides had a length of 10 glucose units (i.e. 10 degrees of polymerization, DP10) or more (Figure 6.3). HtLBs had a lower proportion of long chains: 7.8% of the detected oligosaccharides were DP10 or longer. SmLBs had the lowest proportion of long chains, with only 2.0% of the detected oligosaccharides being DP10 or longer. These data demonstrate that native LBs in the different tissues have varying CLD, and that BrLBs are more amylopectin- like than HtLBs and SmLBs.

6.2.3 LBs are B-type crystallites lacking long-range order

Unlike glycogen, starch granules possess both short- and long-range order, which can be measured by X-ray scattering techniques. Short-range order refers to repeating units on the sub-naonometer scale, i.e. the crystalline double helices formed by the long glucan chains in the unbranched regions of amylopectin. Wide- angle X-ray scattering (WAXS) is used to probe sub-nanometer-sized structures and has been widely utilized to study crystallinity in starch and other polymers (627, 628). Long-range order refers to the alternation of crystalline and amorphous lamellae, which have a repeat distance of 9-10 nm (145, 628). Small-angle X-ray scattering is better suited for measuring these nanometer-sized distances (627). We utilized WAXS and SAXS to define the order of native, purified LBs. Crystallinity content and type of helical packing can be determined from WAXS patterns. In starch, amylopectin double helices crystallize in two different ways depending on the starch species (i.e. potato, corn, etc.), called A- or B-type polymorphs (143). Helices of A-type crystals are densely packed into a monoclinic unit cell that contains very little water. In the B-type crystal, the double helices form a hexagonal channel that fills up with water, rendering a polymorph that is more open and hydrated. Both SmLBs and BrLBs produced a WAXS pattern with an unambiguous B- type ring distribution (Figure 6.4a,b). This result is in sharp contrast to the typical scattering halos produced by amorphous samples such as glycogen or phytoglycogen, a soluble, glycogen-like polymer synthesized by isoaymlase-deficient plants (Figure 6.4c) (629). A scattering profile generated from the rotational average of the 2D patterns show characteristic B-type reflections at diffraction angles of 5.5, 16.6, 21.7 and 23.6° for SmLBs and BrLBs (Figure 6.4d) (142, 630). Peaks at approximately 15, 17, 18 and 23° are typical of A-type crystals (631) and WAXS studies of normal glycogen produce no distinct

170 peaks (632). These data clearly illustrate B-type crystallinity in both types of LBs. While WAXS data is presented as a function of diffraction angle (2θ), SAXS data is presented as a function of the scattering vector, q (628). The scattering vector is related to the wavelength of the incident radiation (λ) and the scattering angle (2θ). It can be converted to distance of the repeating unit (d) in real space using Bragg's law:

4휋 2휋 푞 = sin 휃 = 휆 푑

SAXS studies were performed on SmLBs and BrLBs. Potato amylopectin starch, a genetically modified starch that lacks amylose, was used as a positive control (633). The starch produced a SAXS diffraction pattern with two distinct rings: a large, outer ring, and a small, inner ring (Figure 6.5a). SmLBs produce a diffraction pattern with an outer ring that lacks the small, inner ring (Figure 6.5b). The outer ring represents a reflection at q = 4 nm-1 in the 1D scattering profile (Figure 6.5c). This ring is equivalent to the innermost ring on the WAXS diffraction pattern. From this value of q, a repeating distance, d, of 1.6 nm is calculated. This value is the distance between the helices in the B-type crystal. The inner ring of the starch diffraction pattern represents a reflection at q = 0.76 nm-1. This converts to a repeating distance of 8.3 nm, representing the lamellar repeats. SmLBs lack this reflection for lamellar repeats (Figure 6.5c). When starch is heated, the short-range and long-range order from the packing of amylopectin helices and lamellar units are lost (634). WAXS and SAXS experiments on cooked starch produce an amorphous scattering pattern. As the starch cools, the amylopectin retrogrades, gaining B-type crystallinity over time. Helices form from intra- and intermolecular interactions, and aggregate randomly into networks (635). The lamellar periodicity of the native starch granule never returns (636). In this way, LBs are quite similar to retrograded amylopectin. They are composed of crystalline glucan chains that are arrayed in a random, disordered fashion.

6.2.4 Effects of mechanical, thermal and chemical treatments on LBs

These data illustrate that LBs are comprised of long glucan chains that are crystalline. As discussed in the introduction, the Pflüger method (which involves both thermal and chemical treatment) converts micron-sized LBs to nanometer- sized polyglucosan particles that are on the same scale as glycogen particles. We previously showed that degradation of LBs with an amylolytic enzyme releases an iodine-positive filamentous material (Figure 5.10b). Since modification of starch and other polysaccharides by mechanical, thermal, chemical, or enzymatic means provides structural insights, we sought to explore the effects of various treatments on LB structure by light microscopy and iodine staining. Untreated SmLBs are uniformly small and ovoid and stain intensely with Lugol's

171 iodine (Figure 6.6a). As previously described, all LB types have a tendency to clump (Chapter 5). We found that mild sonication (5 x 5 seconds at 25% amplitude) disrupted SmLB structure almost entirely (Figure 6.6b). After this treatment, most of the LBs were fragmented into clumped, lace-like networks of polyglucosan material that only lightly stained with Lugol’s. Only a few intact SmLBs were left associated with the polyglucosan networks. This result is quite different from what has been observed with starch granules: treatment of starch granules with ultrasonication for minutes or longer produces only mild disruption of granular structure (637). Previously, we found that native LBs are solubilized after a short period of boiling (15-30 minutes), which facilitated the iodine absorbance spectral scans (Chapter 5). However, we did not visualize the effects of boiling on a microscopic level. For the present study, we heated SmLBs in water for 30 minutes at 95°C and found that the bodies were completely converted to the polyglucosan lacy networks (Figure 6.6c). The resulting material after heating appeared very similar to the material produced by sonication, except that no intact SmLBs were observed after the heat treatment. The clumps of material were even more dispersed with longer heating (2 hours at 95°C), and very small, freely floating particles with a diameter of approximately 0.2-0.6 μm were more abundant (Figure 6.6d). In the Pflüger method, tissue is heated at 95°C in 30% KOH for 2 hours, and then the polysaccharide is precipitated three times with 2 volumes of cold ethanol and then re-dissolved in water. LiCl is also added with the ethanol to ensure complete precipitation (209). The successive precipitation steps remove impurities and residual KOH. The effects of the Pflüger method on LB structure could be attributed to one or more components of the protocol: (1) heat in an aqueous environment, (2) heating in 30% KOH, and/or (3) precipitation with ethanol and LiCl. The results in Figure 6.6b and c demonstrate that heat alone is not sufficient to completely disassemble LBs. The SmLBs that were precipitated with ethanol and LiCl three times following a 2-hour boil in water were more difficult to resuspend in water: the white pellet did not easily go into solution, and flakes of polysaccharide were visible. Using light microscopy, this material contained intensely staining aggregates, filamentous structures and granular particles (Figure 6.6e). The aggregates were more amorphous and less compact than control SmLBs, but stained with a similar intensity, suggesting that the polyglucosan became re-crystallized with precipitation. However, it did not reach the compact, elliptoid form of native SmLBs. In contrast, LBs that were boiled in 30% KOH for 2 hours and then precipitated three times with ethanol and LiCl were very easily resuspended in water and produced no visible precipitate. This material did not contain intensely-staining aggregates with Lugol's solution and only a few small clumps of polyglucosan material were visible (Figure 6.6f). Thus, boiling in KOH instead of water fully disassembles the LBs into polyglucosan, and in these alkaline conditions the polyglucosan is much less prone to re-aggregate upon

172 ethanol/LiCl precipitation. Treatment with heat or the Pflüger treatment was also applied to BrLBs and the product was visualized via light microscopy. As expected, control BrLBs were larger and more irregularly shaped than SmLBs and stained just as intensely with Lugol's iodine (Figure 6.6g). A 30-minute heat treatment in water did not completely convert BrLBs to polyglucosan networks; some BrLBs remained intact but were always surrounded by the lace-like material (Figure 6.6h). As with the SmLBs, Pflüger treatment released a large amount of small granular particles, but more lace-like polyglucosan was still present in these samples than in the Pflüger- treated SmLBs (Figure 6.6i). These samples were also easy to resuspend after precipitation. Image thresholding was used to quantitate LBs, LB-like aggregates and polyglucosan in treated samples. In the recorded micrographs, intact LBs were very dark grey or black, with grey values of 50 or less (Figure 4.6j,k). Polyglucosan networks or particles appeared in lighter shades of grey, with values between 51 and 140 (Figure 6.6k,l). Pixel counts from multiple micrographs showed that the vast majority of intensely-staining LBs were disrupted by sonication and boiling in water for 30 minutes (Figure 6.6m). No intact LBs were found in the SmLB samples that were boiled in water for 2 hours or the Pflüger- treated samples. Interestingly, the high intensity observed in control samples (53% of counted pixels) could be partially restored by ethanol treatment after a 2-hour boil in water (25% of counted pixels). These results clearly demonstrate that crystalline LBs undergo a structural transition to polyglucosan networks in the presence of heat, and they can be fully dispersed with the Pflüger method. Without KOH, ethanol precipitation causes polyglucosan to re-aggregate and form dense, LB-like structures. To capture the appearance of the transition from native LB to lace-like polyglucosan, SmLBs were heated at 95°C for 1 or 5 minutes, immediately flash frozen in liquid nitrogen, thawed, and stained with Lugol's iodine. Light microscopy with a color camera was utilized to capture the stained appearance of the native and treated LBs. Control, native LBs stained bright reddish-brown in Lugol's iodine (Figure 6.7a). Samples that were heated for 1 minute displayed numerous amorphous, red-brown gel-like structures in addition to small fragments of lightly stained lacy polyglucosan (Figure 6.7b). By 5 minutes, only lace-like polyglucosans were visible, and no gel-like structures were observed (Figure 6.7c). These results are strongly reminiscent of gelatinization, the irreversible order-disorder transition that starch undergoes when heated in water characterized by swelling, increased viscosity and a loss of crystallinity (638). These results strongly suggest that like starch, LBs gelatinize at high temperatures, corresponding with a loss of crystallinity.

173

6.2.5 Transmission electron microscopy and lintnerization of LBs

The particles that were dispersed after extensive heating or Pflüger treatment were less than 1 μm in diameter (Figure 6.6d,f,i), but still large enough to observe by light microscopy, which has a resolution of about 100 nm. Glycogen particles are smaller, however, typically 15-35 nm in diameter, and visible only by electron microscopy (73, 337). It is assumed that the basic unit of LBs is an abnormally branched glycogen molecule, but this would be approximately 10-fold smaller than the smallest particles in Figure 5. To further study the structure of LBs, boiled SmLBs and Pfluger-purified polysaccharide from skeletal muscle of Epm2a-/- and WT mice were analyzed by transmission electron microscopy (TEM). Electron micrographs of boiled SmLBs showed very large, round bodies ranging in diameter from less than 20 nm to over 100 nm (Figure 6.8a,b). Smaller particles were also present but were less prominent than the larger bodies. Large spherical particles were not observed in the Pfluger-purified polysaccharide from Epm2a-/- or WT mice (Figure 6.8c,d). Epm2a-/- mice showed only very small particles ranging in diameter from 6-20 nm (Figure 6.8c), while WT particles showed a normal distribution for glycogen particles, ranging from 15-30 nm in diameter (Figure 6.8d). Tagliabracci et al. also reported smaller particles in young Epm2a- /- mice and very similarly large, round aggregates in older mice after purification with the Pflüger method (209). These data strongly suggest that the particles in Figures 6.6 and 6.8a and b are still aggregates of smaller polyglucosan molecules, which are the most basic unit of LB structure. Lintnerization is the process of treating starch with mild hydrochloric acid (HCl) at sub-gelatinzation temperatures for long periods of time in order to study its crystalline structure (143, 639). The amorphous regions of the starch granule hydrolyze relatively early during lintnerization, while the crystalline regions are left intact. This effect is due to the protection of glucose residues and glycosidic linkages within the water-excluding helices, which are not easily penetrated by the hydrogen ions (639, 640). The morphology and properties of the resulting lintners vary with species, amylose content, and starch allomorph (475, 640, 641). BrLBs and SmLBs were treated with 2.2N HCl for 5 days at 36°C and then imaged by TEM. The BrLB and SmLB lintners were very similar in morphology, both appearing as lace-like networks (Figure 6.8e-h). Similar lace-like networks have been reported for lintnerized potato starch, particularly waxy potato starch, which exhibit B-type crystallinity (641). This is in contrast to A-type waxy maize starch or barley starch lintners, which form polyhedral-shaped, platelet nanocrystals, reflective of A-type helix packing (475, 640). The LB lintners appeared to be composed of knobby fibers with a diameter of 5-9 nm and larger 10-20 nm subunits, consistent with the size of potato starch lintners (641). The thickness of the fibers is also consistent with the diameters of the individual particles observed in 7C. These data indicate that the polyglucosan particles associate with one another via intermolecular crystalline helical interactions to

174 form a network or necklace-like pattern. Interestingly, retrograded amylopectin, amlylopectin that has been reprecipitated after gelatinization, takes on a very similar necklace-like appearance under the electron microscope, and over time the molecules associate to form networks (635). This is a spontaneous process and therefore the association and precipitation of polyglucosan particles into networks is similarly likely to be spontaneous.

6.3 Discussion

In this study, we defined the architecture of LBs, the hallmark PGBs of a fatal childhood epilepsy, using techniques previously utilized to define starch structure. LBs from different tissues not only vary in gross morphology, but they contain distinct differences in CLD profile, which may have important pathological implications. Additionally, LBs possess B-type crystallinity, but they lack the hierarchical arrangement of starch and contain randomly arrayed crystalline fibrils. The results highlight that the methods utilized to purify and analyze LBs must be taken into account as the methodology can dramatically alter the polyglucosan structure. Furthermore, the terminology used to describe the purified structure should be carefully considered. We propose a model of their hierarchical structure and formation based on these results. This model and the results are likely highly relevant to understanding PGBs found in other diseases and pathologies.

6.3.1 LB architecture depends on cellular origin

LBs form as a result of aberrant glycogen metabolism, which varies with tissue type. Although glycogen metabolism has been most thoroughly studied in liver and muscle, most cell types have the capacity to make glycogen, and it is regulated and utilized differentially depending on the tissue (113). Liver glycogen is utilized to maintain blood glucose during fasting; in muscle, glycogen utilization fuels muscle contraction during exercise. Glycogen metabolism in the heart is unique; paradoxically, cardiac glycogen is preserved under metabolic stress (374). In the brain, although it exists in low levels, glycogen is a dynamic participant in the sleep-wake cycle, learning and memory, and synaptic plasticity (80, 127, 128). Not only is glycogen differentially regulated, but its structure depends on its cellular origin. TEM studies show that in skeletal muscle, glycogen exist solely as small particles of 42 nm or less in diameter, known as β-particles. In liver, β- particles associate to form rosettes known as α-particles up to 300 nm in diameter and are believed to release glucose more slowly (374). In the heart, glycogen exists in both forms. In the brain, only β-particles have been described under normal conditions although α-particles in the central nervous system have been associated with glycogen storage diseases (354, 355, 642). Phosphate content and average chain length of glycogen also depends on tissue type. Glycogen phosphate is 10-times higher in skeletal muscle than in liver, although the

175 relevance of glycogen phosphate is not yet clear (163, 167). Studies from the mid- 1900s demonstrated that the average chain length of liver glycogen is slightly higher than muscle glycogen (47, 133, 643). Chromatographic techniques for chain length profiling with very high resolution appeared in the 1990s, and it was demonstrated that glycogens from different sources had varying profiles although they were all asymmetrically unimodal (155, 156). Given these observations, it is not surprising that LBs from different tissues have distinctly different morphologies and CLD profiles, and therefore should be considered different LB 'types'. LB shape may simply be a product of the constraints of the cell type where it was formed. However, the differences in CLD underscores the tissue-specific differences in regulating glycogen structure. Indeed, the pathological consequences of these differences are intriguing. Since neurological symptoms dominate in LD patients, the brain is considered to be the most sensitive to glycogen perturbations. However, our results suggest that an additional factor could also contribute to neuronal and astrocytic sensitivity to glycogen perturbations. We found that brain LBs exhibit an exacerbated shift in chain length. Therefore, these LBs may actually be more toxic than heart and skeletal muscle LBs. In contrast, polyglucosan in the muscle, which has a CLD that is closest to normal glycogen, may be the least toxic. Anecdotally, we have observed that native, purified BrLBs clumped more frequently, followed by HtLBs, and then SmLBs, and the denatured polyglucosan from BrLBs was the most prone to precipitation. The severity of the clinical features of LD correlates with the shift in the CLD profile of LBs from the corresponding tissue. Seizures, cognitive deficits and neurodegeneration are the initial and most prominent clinical manifestations of the LD in both mice and humans (167, 326). However, cardiac abnormalities have also been reported in mouse models and patients (595, 596). Muscular complications in LD patients such as ataxia, dysarthria, spasticity, and rigidity are generally attributed to neurological deterioration (326, 644). LD mice only display mild muscle weakness and decreased motor coordination (177, 561). Reports about LBs in liver are also inconsistent: although liver abnormalities and PAS- positive material have been reported in LD patients, they are not generally called LBs and hepatic disease is not typical (526, 645). In mouse models, LBs appear to accumulate more slowly in liver than in brain (177, 211). As we have mentioned, two types of BrLBs have been observed in LD patients and mouse models, often referred to as Van Hoof's type I and type II. It is possible that CLD may vary between the BrLB types, but they are inseparable using our purification method. A recent quandary has arisen regarding the cellular origin of BrLBs. For decades, LBs have been reported to exclusively populate neurons in LD patients (reviewed in (79)). Recently, two independent groups discovered LBs in both astrocytes and neurons in Epm2a-/- and Epm2b-/- mice (324, 325). Augé et al. showed that neuronal and astrocytic LBs were immunologically distinct, and that astrocytic LBs were identical to corpora amylacea. They suggested that Van Hoof's types I and II LBs correspond to astrocytic corpora amylacea and neuronal

176

LBs, respectively. Since we have observed distinct morphologies and CLD in LBs originating from cardiac and skeletal myocytes, it is reasonable to speculate that two distinct morphologies of BrLBs could correspond to their different cellular origins.

6.3.2 Proposed model for LB architecture and formation

We have described three levels of polyglucosan architecture in our study: the native LB, the polyglucosan network, and the polyglucosan particle. The largest entity is the native LB, ranging in diameter from 2-10 μm or more (Figure 6.9a). It is comprised of crystallized polyglucosan particles with B-type crystal packing, where double helices form a hexagonal unit with a water-filled interior channel. It is very likely that some double helices are formed from adjacent chains within the same polyglucosan molecule (intramolecular interactions), while others may be formed from chains from two different polyglucosans (intermolecular interactions). For most LB types, crystalline units are randomly arranged, in stark contrast to the lamellar structure of starch granules. The exception may be the dense core of type II BrLBs, which EM studies suggest are composed of radially arranged fibrils. The second structural entity is the polyglucosan network that emerges when intact LBs are boiled, mechanically agitated, or lintnerized (Figure 6.9b). These networks resemble potato lintners and retrograded (i.e. re-precipitated) amylopectin, both of which exhibit B-type crystallinity (635, 641). Since amylopectin and LD polysaccharide are both infrequently branched polysaccharides with long glucan chains capable of crystallizing, it is not surprising that the two polysaccharides would associate in a similar network formation. The networks are composed of knobby fibers between 5 and 20 nm in diameter. The morphology has been described as resembling a necklace, where the "pearls" are laterally packed helical bundles connected by longer amylopectin chains (635). In the case of LD, it is likely that molecules associate via intermolecular helical interactions to form these necklace-like fibers (Figure 6.9b). The third structural entity is the polyglucosan particle (Figure 6.9c). This particle is presumably an aberrantly branched glycogen β-particle, with such long chains that it crystallizes. We have shown that polyglucosan particles are best dissociated with Pflüger treatment, due to heating in KOH. Indeed, alkaline treatment increases the solubility of both glycogen and starch (646, 647), and certain forms of these polysaccharides (desmoglycogen or 'fixed' glycogen and resistant starch) can only be solubilized with alkali (648, 649). This is likely simply because alkali solutions disrupt hydrogen bonding (650). In starch, the crystalline, double helical conformation of glucan chains prohibits their access by degrading enzymes (139). Presumably, the long chains of polyglucosan particles take on a similar conformation and prevent them from being degraded by the glycogenolytic machinery (188). As a result, polyglucosan particles accumulate, potentially associating into networks inside of cells like the

177 ones we observed in vitro (Figure 6.6). Aggregation and network formation likely occur spontaneously, similar to the natural retrogradation of amylopectin after it is cooked. It is possible that the cell, detecting an unknown and potentially toxic species, sequesters the polyglucosan particles and networks into non-membrane- bound inclusions that become LBs. LBs have been shown to be decorated with a variety of proteins involved in glycogen metabolism, ER stress, ubiquitin, and autophagy and delineated by cytoskeletal elements in vivo (179, 586). Thus, the formation of this structural entity may not be entirely spontaneous; cellular mechanisms may assist in polyglucosan sequestration. We showed that ethanol precipitation of LBs after boiling in water led to some polyglucosan clumps that stained as intensely as native LBs. This treatment may in some way resemble the process of polyglucosan sequestration and compaction that occurs in cells. The shape and size of the LB is probably dictated by the internal constraints of the cell type where it forms, but may also be influenced by the CLD of the constituent polyglucosan and/or effects of surrounding molecules. The random arrangement of fibrils and lack of higher order in LBs is not surprising, given that they are an aggregate that forms in response to a aberrant cellular mechanism. B-type crystallization is favored by longer chains and lower temperatures (651, 652). Even native A-type amylopectin takes on a B-type conformation after retrogradation, and A-type crystallites can convert to B-type with sufficient hydration (653, 654). A previous study showed that enzymatic chain elongation of glycogen produced dendritic particles with B-type crystallinity (655). Electron micrographs suggest that over time, double helical segments formed as a result of intra- and intermolecular entanglement of the extended glucan chains. It has been proposed that in LD, the lack of laforin or malin leads to an imbalance in glycogen synthase and phosphorylase activities, causing aberrant chain elongation (188). Thus, the entanglement and crystallization of artificially elongated glycogen as described by Putaux et al. may resemble what occurs in vivo in LD.

6.3.3 Nomenclature: Polysaccharides, glycogen, polyglucosan, PGBs and LBs

We have endeavored to be precise with our terminology in this study. The term 'polysaccharide' denotes any polymer of sugar molecules, and thus glycogen, amylopectin, amylose, and polyglucosan are all considered polysaccharides. The term 'polyglucan' refers only to polysaccharides containing glucose units attached by glycosidic bonds, and more specifically, α-polyglucans are those with α-1,4 an α-1,6 glycosidic bonds. Thus, all of the polysaccharides that we have discussed in this paper are also α-polyglucans. However, we have abstained from using the term 'polyglucan' due to its similarity to 'polyglucosan' and also since no other types of polysaccharides are discussed. The polysaccharides purified from LD mice via the Pflüger method are frequently called 'glycogen', and LD mice are said to have increased glycogen in

178 their tissues. However, the Pflüger purified polysaccharides are actually a mixture of both normal glycogen and abnormal polysaccharides that we refer to herein as polyglucosan. The polyglucosan of LD is a precipitation-prone, crystalline polysaccharide with long chains. This structure does not fit the definition of glycogen, i.e. a soluble glucose polymer with ~13 glucose units per chain and regular branching (113). Therefore, we propose that the abnormal polysaccharide should be called polyglucosan rather than glycogen. In fact, studies of the polysaccharide content of insoluble vs. soluble 'glycogen' fractions in LD mice show that LD mice display an increase in the insoluble fraction (corresponding to polyglucosan), but the soluble fraction is equivalent to WT mice (178). Thus, LD tissues have approximately the same amount of soluble glycogen as WT mice, and the total increase in polysaccharide, is exclusively due to insoluble polyglucosan. We also want to clarify the difference between 'polyglucosan', 'polyglucosan bodies' (PGBs), and 'LBs'. We use the term polyglucosan to refer to any abnormal glucose polymer containing α-glycosidic linkages found in mammalian tissues, but its exact chemical characteristics vary with its pathological origin. The polyglucosan found in LD tissues is more phosphorylated and has a right-shifted CLD compared to glycogen. 'PGBs' refer to the compacted inclusions of polyglucosan observed in vivo, which can be isolated as intact entities (169, 616) (Chapter 5). Corpora amylacea and LBs are considered PGBs, but not all PGBs are exactly equivalent to LBs or corpora amylacea with respect to CLD and other structural elements. 'LBs' are the PGBs found in LD, which we have shown are crystalline and morphologically distinct based on tissue type.

6.3.4 Similarities of LBs with other PGBs and amyloid and relevance to neurological disease

PGBs of different pathologies are overall very similar but distinguishable by subtle ultrastructural and histochemical qualities. LBs are the only PGB to display a darkly staining central core with radiating fibrils visible by electron microscopy (79). The polyglucosan fibrils in LBs are somewhat thicker and more electron dense than those in the PGBs of GSD IV, a disease caused by a deficiency in glycogen branching enzyme (569). With Lugol's iodine, LBs are brown, corpora amylacea are purplish-brown, and the GSD IV PGBs are distinctly purple, indicating they contain very long glucan chains (616, 624). The PGBs of GSD IV do not stain with toluidine blue, in contrast to corpora amylacea and LBs, suggesting they have a much lower phosphate content (624). Corpora amylacea have even more phosphate than LBs (168). The data presented in the present study on LBs are likely to be very relevant to understanding the architecture and formation of other PGB types. It is likely that the aggregation of polyglucosan into networks and B-type crystallization are common features of PGBs. It will be very interesting to compare their CLD profiles and understand how their chemical

179 compositions affects their architecture and in vitro properties. The different levels of polyglucosan association (PGB, network, and particle) are analogous to the levels of protein aggregation in amyloid disorders. Polyglucosan particles aggregate into networks much like misfolded proteins aggregate into soluble amyloid oligomers. Eventually the polyglucosan compacts into PGBs, and soluble oligomers polymerize into fibrils and plaques, although these higher-orders structures are distinctly different (PGBs contain crystalline α- helices, while amyloid contains β-sheets; polyglucosan fibrils are branched, but amyloid fibrils do not; PGBs are typically not birefringent, in contrast to amyloid; etc.). It is possible that for both types of inclusions, the intermediate aggregates are toxic, and the higher-order structures are reactionary, or even protective. There is a significant amount of evidence supporting the toxicity of soluble oligomers rather than amyloid fibrils (656, 657). Since LBs are often found in apparently healthy neurons, and degenerating neurons usually lack LBs, some have suggested that the sequestration of polyglucosan into LBs may be a protective mechanism in neuronal cells (177, 586). Corpora amylacea, which appear with aging and various neurological conditions, are also considered a protective response to age-related or pathological degeneration (611). It has been suggested that the corpora amylacea are a repository for hazardous cellular waste that can be cleared by the immune system (658). The observation that polyglucosan is a primary constituent of these alleged cellular 'trash bins' further supports their potentially toxic nature. Whether or not higher-order PGBs are themselves pathogenic or reactionary, studies indicate that polyglucosan accumulation is pathogenic at some level. Genetic ablation of glycogen synthase in the brain rescues the neurological phenotype in LD mouse models, and overexpression of glycogen synthase induces polyglucosan accumulation and neurodegeneration in WT flies and mice (182, 563). Similarly, inhibition of glycogen synthase eliminates PGBs and reverses neurological damage in a mouse model of GSD IV (262). Glycogen synthase ablation in WT mice also eliminates the formation of corpora amylacea with aging, and a reduction in glycogen synthesis eliminates polyglucosan deposits and increases lifespan in Drosophila (659). Further investigation will be necessary to understand how polyglucosan is sequestered as PGBs in vivo, why various forms of polyglucosan are pathological, and what can be done to alleviate their toxicity.

180

Figure 6.1 Representative z-stack images of iodine-stained starch and LBs. Starch (a), SmLBs (b), HtLBs (c), and BrLBs (d-e) were visualized by confocal microscopy using 561 nm (TRITC) channel and differential interference contrast (DIC). In (a), scale bar = 10 μm and panels represent 2 μm steps. In (b), scale bar = 2 μm and panels represent 0.4 μm steps. In (c-e), scale bar = 2 μm and panels represent 0.9 μm steps.

181

Figure 6.2 Additional z-stack images. Iodine-stained starch (a,b), SmLBs (c,d), HtLBs (e,f), and BrLBs (g,h). In (a,b), scale bars = 5 μm and panels represent 2.25 μm steps. In (c-h), scale bars = 2 μm and panels represent 0.9 μm steps.

182

) 60

% ( 40

n BrLBs o

i 20

t HtLBs r

o SmLBs

p 4

o r

p 3

v 2

t a

l 1 e

R 0 0 5 10 15 20 25 30 Degree of Polymerization (DP)

Figure 6.3 Chain length distribution of LB types determined by FACE. Degree of polymerization (DP) corresponds to the number of glucose units per linear chain.

183

Figure 6.4 Wide-angle X-ray scattering (WAXS) of LBs and phytoglycogen. (a-c) 2D WAXS diffraction patterns of BrLBs (a) and SmLBs (b) illustrate a B-type ring distribution. Phytoglycogen produces scattering halos without distinct rings (c). The corresponding 1D scattering profile for SmLBs is shown (d); grey arrows designate the B-type reflections.

184

Figure 6.5 Small-angle X-ray scattering (SAXS) of LBs and potato amylopectin starch. (a,b) 2D SAXS diffraction patterns of (a) SmLBs and (b) starch. (c) 1D scattering profile for both samples.

185

186

Figure 6.6 Effects of various treatments on the structure of LBs. SmLBs (a-f) and BrLBs (g-i) were stained with Lugol's iodine after each treatment and visualized by light microscopy. Control SmLBs (a); mildly sonicated SmLBs (b); SmLBs heated in water at 95°C for 30 minutes (c); SmLBs heated in water at 95°C for 2 hours (d); SmLBs heated in water at 95°C for 2 hours, then precipitated 3 times with 2 volumes cold ethanol and LiCl (e); SmLBs treated with the Pflüger method (2 hours at 95°C in 30% KOH) (f). Control BrLBs (g); BrLBs heated in water for 30 minutes (h); BrLBs treated with the Pflüger method (i). Scale bar = 10 μm. Each inset is a 2.5x magnification of the boxed region. Thresholding was used to quantitate LBs (with high intensity staining) and polyglucosan (with low intensity staining) in micrographs. Examples of applied thresholds for (g-i) are shown in (j- l). Pixels with high intensity (grey values 0-50) and low intensity (grey values 51- 140) were counted and shown as a percentage of total counted per micrograph. Pixels with grey values above 141 were considered background and not counted. The values shown are the averages of 9 micrographs ± standard error.

187

Figure 6.7 Structural transition of SmLBs during heating. Control SmLBs (a), SmLBs heated for 1 minute and flash frozen (b), and SmLBs heated for 5 minutes and flash frozen (c). After freezing, samples were thawed, stained with Lugol's iodine, and visualized using a light microscope with a Zeiss 512 color camera. Scale bar = 10 μm.

188

Figure 6.8 Transmission electron micrographs of LBs exposed to various treatments. Boiled SmLBs (a,b), Pflüger-purified polysaccharide from Epm2a-/- mice (c) and WT mice (d), linterized BrLBs (e,g) and lintnerized SmLBs (f,h). Lintnerized samples were incubated in 2.2 N HCl at 36°C for 5 days and then visualized. Panels (a-f) are shown at equal magnification for comparison, while (g,h) are magnified 2.5x for detail. Scale bars = 40 nm.

189

Figure 6.9 A model for LB structure and formation. Blue arrows and text illustrate what was observed in vitro, and red arrows and text illustrate what may occur in vivo.

190

CONCLUDING REMARKS

7.1 Summary

Since its discovery in the 1800s, glycogen metabolism has received an enormous amount of scientific attention. At least four Nobel prizes related to the field of glycogen metabolism were awarded during the 20th century, and the topic occupies at least one chapter in nearly every biochemistry textbook. Since the 1970s, the field has assumed to be an antiquated area of research, one that has been so extensively elucidated that research efforts should be directed elsewhere (103). This assessment is most certainly not the case. The role of glycogen in the brain is only beginning to be understood (80, 127, 128). Additionally, the intricate mechanisms of glycogen regulation are still being defined (137). Basic studies of Pompe and Lafora disease have revealed the importance of neuronal glycogen, glycogen structure and covalent phosphorylation, and glycophagy (167, 275). Novel GSDs are still being described as GSD type XV was only recently discovered (165, 240). Furthermore, with the exception of Pompe disease, therapies for GSDs are severely lacking. Lafora disease (LD), an atypical GSD causing severe epilepsy and early death, is one such disease. Genetic testing has been employed with increasing frequency to diagnose diseases, particularly in the epilepsy field (660). LD is no exception; a genetic diagnosis is essential due to the rarity and severity of the disease. As a result, the resolution with which scientists and clinicians can view the disease is improving. LD is no longer considered a homogenous disorder, but it is understood that genetics can modify or dictate clinical phenotype (328, 484, 488). There are many parallels between the fields of cystic fibrosis and LD, and like CFTR mutations, we have shown that LD missense mutations can be divided into classes with different molecular effects (Chapter 3). Our biochemical results pose a molecular explanation for milder forms of LD, and the pipeline that we have developed will facilitate characterization and classification of novel mutations within a matter of weeks. With LD therapies in preclinical development, personalized diagnoses and patient-specific treatments will become extremely beneficial in the near future. Despite multiple advances in the last decade, many gaps in our understanding of LD remain. Perhaps the most difficult question is why the lack of laforin or malin leads to aberrant glycogen structure and hyperphosphorylation. The relevance of laforin phosphatase activity has been challenged based on mouse models (159, 188, 189). We discovered that the catalytically inactive laforin C/S mutant that is able to rescue the LD phenotype in mouse models dramatically changes the binding properties of laforin. Laforin C/S is a less dynamic protein that enhances the interaction of laforin with both glucan and protein substrates (Chapter 4). Before the work presented herein, the field assumed that the primary cause of LD in EPM2A mutant patients was a lack of laforin phosphatase activity. However, my

work has demonstrated that the primary deficiency of LD-causing laforin missense mutations is a lack of glucan binding and/or deficient protein-protein interactions, and not a defect in phosphatase activity (Chapter 3). While LD-causing mutations lead to destabilization and reduced glucan/protein interaction, the LD-rescuing mutation C/S leads to increased rigidity and enhanced glucan/protein interaction. These results strongly suggest that the primary role of laforin is to serve as an adaptor protein on the glycogen molecule. The similar phenotypes observed by malin- and laforin-deficient patients do indeed suggest that laforin and malin have overlapping functions, and the evidence for a laforin-malin complex is building (188). In fact, many RING-type E3 ubiquitin ligases function as dimers or part of a multi-subunit complex (661). Laforin and malin may function analogously to the Cullin-RING E3 ubiquitin ligase complexes, which consist of an elongated Cullin protein bridging a RING-containing E3 ligase with an adaptor and substrate receptor (662). However, the independent functions of laforin and malin must not be disregarded. The conservation and centrality of glucan phosphorylation to carbohydrate metabolism across multiple kingdoms reinforce the relevance of laforin phosphatase activity (139, 183, 185), as does the observation that even in malin deficiency, aberrantly branched glycogen is hyperphosphorylated (159, 178, 180). Furthermore, malin ubiquitinates glycogen phosphorylase independent of laforin (Sun, Dukhande, Gentry, unpublished data). History can attest to the enormous benefit of basic scientific studies for the development of novel therapies (663, 664). The similarities between LBs and plant starch led to the conception of a targeted fusion enzyme comprised of α-amylase, which naturally degrades starch in the human gut. We showed that this enzyme reduces LB load in muscle, heart and brain of LD mice (Chapter 5). An analogous enzyme designed to treat Pompe disease has shown great promise in phase 1/2 clinical trials (577). Additional lines of therapy for LD are being developed by the same investigators who laid the initial biochemical groundwork in the LD field (328, 564). The concerted efforts of basic scientists, clinicians, and the families of patients have accelerated progress toward therapies and a cure (665). Although LD is a rare disease, the work herein has applications beyond LD. Antibody-fusion therapy has the potential for treating an array of diseases, including most GSDs. Also, accumulated glycogen and/or polyglucosan bodies (PGBs) are found in some cancers and neurological diseases (79, 611, 666, 667). Although PGBs have been studied histologically, their fine structure and molecular composition have not been defined in a level of detail comparable to plant starch (145). We studied LBs using X-ray diffraction, microscopy, and chemical, thermal and mechanical means (Chapter 6). We discovered tissue-specific differences and postulate that both spontaneous precipitation and cellular sequestration contribute to LB formation. This study is broadly applicable since polyglucosan accumulation is pathogenic, and the formation of other types of PGBs may proceed in a similar manner.

192

7.2 Limitations and Future Directions

Although we have defined molecular defect(s) for most of the LD missense mutations analyzed, no defect was detected in 6 of the 27 mutations (E210K, P211L, E224I, K140N, G240S, and P246A) using our current suite of assays. E210, P211 and E224 are surface-exposed residues of the laforin V-loop that may represent a binding surface for additional laforin interaction partner(s). K140 may be a ubiquitination site, suggesting laforin ubiquitination is critical for preventing LD. It is not clear why S25P abolished the interaction of laforin with malin and PTG; however, S25 has been shown to be a site for phosphorylation by AMPK (313). Laforin and malin have also been shown to ubiquitinate the CBM-containing β-subunit of AMPK (303). These data suggest reciprocal regulation between AMPK and the laforin-malin complex. Further study will be required to identify the molecular defects of all missense mutations. A current limitation in studying the laforin-malin interaction in vitro is the inability to produce soluble recombinant malin using E. coli. This issue may be circumvented by alternative expression systems such as insect or mammalian cells. Knock-in mouse models of missense mutations will also be extremely useful for defining and treating mild forms of LD. These models are currently in development using CRISPR gene editing technology (328). We have shown that laforin has a DSP-DSP dimer interface and that this interface is important for cooperative binding to glucan substrates (Chapter 4) (190). Our data suggest that cooperative binding may be specific for long, phosphorylated glucans, which is mimicked by the interaction of phosphate-bound C266S laforin with DP24. Phosphoglucan substrates are not readily available commercially, but they could be prepared by debranching amylopectin or purified LD polyglucosan. Furthermore, it is not clear if and how laforin phosphatase activity and binding to other proteins are coupled. We propose that one subunit may phosphorylate a glucan while the other binds to a protein. We also suggest that dimerization may be required for the malin interaction (Chapter 3). These speculations will need to be tested further. It is not clear if and how laforin binds to branch points in glucans, although the multiple glucan molecules found at both binding sites in the glucan bound structure, as well as the location of W196, suggest a branch point interaction. However, laforin preferentially binds amylopectin-like or LB-like substrates, which have fewer branch points than glycogen. Again, testing this is limited by the availability of branched glucan substrates, which are difficult, although not impossible, to attain. It would also be interesting to monitor the dynamics of the laforin-glucan interaction. Obviously, a single dimer has two primary contact sites on the CBMs. Do these two sites alternate binding and releasing glucans, allowing laforin to "walk" across the surface of a glycogen molecule? Or does the entire dimer associate and dissociate from a glucan as a unit? The observation that laforin always colocalizes tightly with glycogen molecules in cell culture suggests

193 the former. The emerging role(s) of laforin as a scaffolding molecule and/or adaptor protein for glycophagy need to be further defined. Crystal structures of laforin in complex with different glucan substrates and/or binding partners will be tremendously useful in understanding how laforin may assemble with malin other proteins on polyglucans. Again, such studies are limited by the availability of glucan substrates and recombinant proteins. However, structures of glycogen synthase, glycogen debranching enzyme, and the AMPK β subunit and partial structures of p62 have been determined (251, 264, 668-670). The ubiquitination of these proteins by malin may require direct binding by laforin (300, 303, 307, 455). We have shown that VAL-0417 reduces LB load of LD mice. However, the effect of the drug on neurodegeneration and epilepsy is difficult to assess using mouse models. As is the case for many neurodegenerative diseases, LD mouse models incompletely recapitulate the human phenotype. Although they do exhibit behavioral and neurophysiological impairments, LD mice do not have a reduced lifespan, comparable EEG abnormalities, or spontaneous grand mal seizures (177). To detect seizure susceptibility, kindling with convulsive agents is often employed (180, 182, 294, 562). Electrophysiological and behavioral changes are difficult to detect even in older mice (e.g. 6 months) corresponding to a developmental stage beyond the lifespan of most LD patients (30 years). There are canine models of LD that develop spontaneous convulsive seizures, but clinical features often develop later in life and the animals also do not die early (671, 672). Although additional preclinical development can be performed in mice or dogs, such as drug turnover rate, effective dosing range and toxicity levels, there are limitations to both models. Phase 1/2 clinical trials involving another antibody-enzyme fusion therapy for Pompe disease are currently underway, and our preliminary data indicate this drug also reduces LB load in LD mice (328, 577). Therefore, repurposing this drug for LD treatment is also a viable option, particularly since it has already entered the clinical stage. Treatment of LD with antibody-enzyme fusions will require direct injection into the CNS of patients to bypass the blood-brain-barrier. We showed that intracerebroventricular (ICV) injection achieves a better distribution than intrathecal injection of VAL-0417 (Chapter 5). Implanted ICV devices are already utilized for safe, long-term drug administration to the CNS and are favored over repeated lumbar punctures (599). Clinical trials with LD patients would likely involve periodic injections via an implanted ICV device. To prepare for these clinical trials, a natural history study has already been initiated to define the natural course of the disease, establish a primary endpoint, and identify potential biomarkers that could be used to monitor treatment (328).

194

7.3 Conclusion

This work has contributed to the development of personalized diagnoses and therapies for LD and a better understanding of the molecular basis for LD. These studies highlight the importance of basic science research for therapy development and also the necessity for a deeper understanding of glycogen metabolism, particularly in the brain.

195

REFERENCES

1. C. Bernard, Sur une nouvelle fonction du foie chez l'homme et les animaux. (Institut de France, 1850). 2. C. Bernard, Sur le mécanisme physiologique de la formation du sucre dans le foie. C. R. Acad. Sci. (Paris) 44, pp. 578-586 (1857). 3. A. M. Brown, Brain glycogen re-awakened. Journal of neurochemistry 89, pp. 537-552 (2004). 4. F. G. Young, Claude Bernard and the theory of the glycogenic function of the liver. Annals of Science 2, pp. 47-83 (1937). 5. F. G. Young, Claude Bernard and the discovery of glycogen. British Medical Journal 1, pp. 1431 (1957). 6. M. D. Grmek, First steps in Claude Bernard's discovery of the glycogenic function of the liver. Journal of the History of Biology 1, pp. 141-154 (1968). 7. J. Larner, The discovery of glycogen and glycogen today. Claude Bernard and Experimental Medicine, Schenkman Publ Cambridge, Massachusetts, pp. p135 (1967). 8. M. Foster, Claude Bernard. (Unwin, 1899), vol. 6. 9. M. A. Sanson, Note sur la formation physiologique du sucre dans l'economie animale. Comptes rendus des seances de l'Academie des Sciences 44, pp. 1323-1325 (1857). 10. W. B. Carpenter, F. G. Smith, H. P. Meneses, Principles of Human Physiology. (Henry C. Lea, 1876). 11. W. M. Fletcher, F. G. Hopkins, Lactic acid in amphibian muscle 1. The Journal of physiology 35, pp. 247-309 (1907). 12. O. Meyerhof, Die Energieumwandlungen im Muskel. Pflüger's Archiv für die gesamte Physiologie des Menschen und der Tiere 182, pp. 232-283 (1920). 13. O. Meyerhof, II. Das Schicksal der Milchsäure in der Erholungsperiode des Muskels. Pflügers Arch 182, pp. 284 (1920). 14. O. Meyerhof, Kohlenhydrat und Milchsäureumsatz im Froschmuskel. Pflüg. Arch. ges. Physiol 185, pp. 11-20 (1920). 15. D. Nachmanson, S. Ochoa, F. A. Lipmann. (Washington DC: National Academy of Sciences, 1960). 16. E. Pflüger, Meine Methode der quantitativen Analyse des Glykogenes und die Arteigenthümlichkeit der Substanzen des Thierleibes. Archiv für die gesamte Physiologie des Menschen und der Tiere 129, pp. 362-378 (1909). 17. J. V. Passonneau, P. D. Gatfield, D. W. Schulz, O. H. Lowry, An enzymic method for measurement of glycogen. Analytical biochemistry 19, pp. 315- 326 (1967). 18. S. H. Gage, Glycogen in the nervous system of vertebrates. Journal of Comparative Neurology 27, pp. 451-465 (1917). 19. The Lancet. (J. Onwhyn, 1890).

20. E. G. Holmes, B. E. Holmes, Contributions to the study of brain metabolism: Carbohydrate metabolism relationship of glycogen and lactic acid. Biochemical Journal 20, pp. 1196 (1926). 21. C. F. Cori, G. T. Cori, Glycogen formation in the liver from d-and l-lactic acid. Journal of Biological Chemistry 81, pp. 389-403 (1929). 22. C. F. Cori, G. T. Cori, Mechanism of formation of hexosemonophosphate in muscle and isolation of a new phosphate ester. Proceedings of the Society for Experimental Biology and Medicine 34, pp. 702-705 (1936). 23. C. F. Cori, S. P. Colowick, G. T. Cori, The isolation and synthesis of glucose-1-phosphoric acid. J. biol. Chem 121, pp. 465-477 (1937). 24. G. T. Cori, S. P. Colowick, C. F. Cori, The enzymatic conversion of glucose-1-phosphoric ester to 6-ester in tissue extracts. Journal of Biological Chemistry 124, pp. 543-555 (1938). 25. C. F. Cori, G. Schmidt, G. T. Cori, The synthesis of a polysaccharide from glucose-1-phosphate in muscle extract. Science (New York, N.Y.) 89, pp. 464-465 (1939). 26. C. S. Hanes, The reversible formation of starch from glucose-1-phosphate catalysed by potato phosphorylase. Proc. R. Soc. Lond. B 129, pp. 174- 208 (1940). 27. W. Kiessling, Über ein neues Fermentprotein der Hefe und eine reversible enzymatische Synthese des Glykogens. Naturwissenschaften 27, pp. 129- 130 (1939). 28. A. A. Green, G. T. Cori, Crystalline muscle phosphorylase I. Preparation, properties, and molecular weight. Journal of Biological Chemistry 151, pp. 21-29 (1943). 29. G. T. Cori, A. A. Green, Crystalline muscle phosphorylase II. Prosthetic group. Journal of Biological Chemistry 151, pp. 31-38 (1943). 30. C. F. Cori, G. T. Cori, A. A. Green, Crystalline muscle phosphorylase III. Kinetics. Journal of Biological Chemistry 151, pp. 39-55 (1943). 31. G. T. Cori, C. F. Cori, Crystalline muscle phosphorylase IV. Formation of glycogen. Journal of Biological Chemistry 151, pp. 57-63 (1943). 32. D. J. Manners, in Advances in carbohydrate chemistry. (Elsevier, 1963), vol. 17, pp. 371-430. 33. W. J. Whelan, Why the linkage of glycogen to glycogenin was so difficult to determine. Biochemistry and Molecular Biology Education 35, pp. 313- 315 (2007). 34. F. L. Bates, D. French, R. E. Rundle, Amylose and amylopectin content of starches determined by their iodine complex formation. Journal of the American Chemical Society 65, pp. 142-148 (1943). 35. W. N. Haworth, E. G. V. Percival, 324. Polysaccharides. Part XI. Molecular structure of glycogen. Journal of the Chemical Society (Resumed), pp. 2277-2282 (1932). 36. C. E. H. Bawn, E. L. Hirst, G. T. Young, The nature of the bonds in starch. Transactions of the Faraday Society 36, pp. 880-885 (1940). 37. E. M. Montgomery, F. B. Weakley, G. E. Hilbert, Isolation of 6-[α-D- Glucopyranosyl]-D-glucose (Isomaltose) from Enzymic Hydrolyzates of

197

Starch2. Journal of the American Chemical Society 71, pp. 1682-1687 (1949). 38. M. L. Wolfrom, E. N. Lassettre, A. N. O'Neill, Degradation of Glycogen to Isomaltose1. Journal of the American Chemical Society 73, pp. 595-599 (1951). 39. J. Larner, B. Illingworth, G. T. Cori, C. F. Cori, Structure of glycogens and amylopectins II. Analysis by stepwise enzymatic degradation. Journal of Biological Chemistry 199, pp. 641-651 (1952). 40. J. Larner, The action of branching enzymes on outer chains of glycogen. Journal of Biological Chemistry 202, pp. 491-503 (1953). 41. A. J. Parodi, J. Mordoh, C. R. Krisman, L. F. Leloir, In vitro synthesis of particulate glycogen from uridine diphosphate glucose. Archives of biochemistry and biophysics 132, pp. 111-117 (1969). 42. G. T. Cori, J. Larner, Action of amylo-1, 6-glucosidase and phosphorylase on glycogen and amylopectin. Journal of Biological Chemistry 188, pp. 17- 29 (1951). 43. G. J. Walker, W. J. Whelan, The mechanism of carbohydrase action. 8. Structures of the muscle-phosphorylase limit dextrins of glycogen and amylopectin. Biochemical Journal 76, pp. 264 (1960). 44. D. H. Brown, B. Illingworth, The properties of an oligo-1, 4→ 1, 4- glucantransferase from animal tissues. Proceedings of the National Academy of Sciences 48, pp. 1783-1787 (1962). 45. F. Huijing, Glycogen metabolism and glycogen-storage diseases. Physiological reviews 55, pp. 609-658 (1975). 46. K. H. Meyer, P. Bernfeld, Recherches sur l’amidon V. L’arnylopectine. Helv. chim. Acta 23, pp. 875 (1940). 47. Z. Gunja-Smith, J. J. Marshall, E. E. Smith, Enzymatic determination of the unit chain length of glycogen and related polysaccharides. FEBS letters 13, pp. 309-311 (1971). 48. Z. Gunja-Smith, J. J. Marshall, C. Mercier, E. E. Smith, W. J. Whelan, A revision of the Meyer-Bernfeld model of glycogen and amylopectin. FEBS letters 12, pp. 101-104 (1970). 49. M. Somogyi, The solubility and preparation of phosphorus-and nitrogen- free glycogen. Journal of Biological Chemistry 104, pp. 245-253 (1934). 50. R. Willstätter, M. Rohdewald, Über den Zustand des Glykogens in der Leber, im Muskel und in Leukocyten.(Zur Kenntnis der Proteinbindung physiologisch wichtiger Stoffe.). Hoppe-Seyler´ s Zeitschrift für physiologische Chemie 225, pp. 103-124 (1934). 51. A. Lazarow, Particulate glycogen: A submicroscopic component of the guinea pig liver cell; its significance in glycogen storage and the regulation of blood sugar. The Anatomical Record 84, pp. 31-50 (1942). 52. P. C. Calder, Glycogen structure and biogenesis. International journal of biochemistry 23, pp. 1335-1352 (1991). 53. S. E. Kerr, The carbohydrate metabolism of brain VI. Isolation of glycogen. Journal of Biological Chemistry 123, pp. 443-449 (1938).

198

54. S. H. Carter, W. E. Stone, Effect of convulsants on brain glycogen in the mouse. Journal of neurochemistry 7, pp. 16-19 (1961). 55. C. Carr, Practical Physiological Chemistry. By Philip B. Hawk, Bernard L. Oser, and William H. Summerson. The Journal of Physical and Colloid Chemistry 51, pp. 1214-1215 (1947). 56. M. R. A. Chance, D. C. Yaxley, Central Nervous Function and Changes in Brain Metabolite Concentration: I. Glycogen and Lactate in Convulsing Mice. Journal of Experimental Biology 27, pp. 311-323 (1950). 57. F. N. LeBaron, The resynthesis of glycogen by guinea pig cerebral-cortex slices. Biochemical Journal 61, pp. 80 (1955). 58. H. McIlwain, M. A. Tresize, The glucose, glycogen and aerobic glycolysis of isolated cerebral tissues. Biochemical Journal 63, pp. 250 (1956). 59. D. Svorad, Diurnal changes in the brain glycogen. Experientia 14, pp. 452- 452 (1958). 60. R. W. Mowry, J. Bangle, Raymond, Histochemically Demonstrable Glycogen in the Human Heart: With Special Reference to Glycogen Storage Disease and Diabetes Mellitus. The American journal of pathology 27, pp. 611 (1951). 61. R. D. Lillie, Further exploration of the HIO4‐Schiff reaction with remarks on its significance. The Anatomical Record 108, pp. 239-253 (1950). 62. J. F. A. McManus, Histological and histochemical uses of periodic acid. Stain technology 23, pp. 99-108 (1948). 63. J. F. A. McManus, Histological demonstration of mucin after periodic acid. Nature 158, pp. 202 (1946). 64. N. Shimizu, Histochemical studies of glycogen of the area postrema and the allied structures of the mammalian brain. Journal of Comparative Neurology 102, pp. 323-339 (1955). 65. N. Shimizu, T. Kumamoto, Histochemical studies on the glycogen of the mammalian brain. The Anatomical Record 114, pp. 479-497 (1952). 66. N. Shimizu, Y. Hamuro, Deposition of glycogen and changes in some enzymes in brain wounds. Nature 181, pp. 781 (1958). 67. N. Shimizu, Z. Kubo, Histochemical studies on brain glycogen of the guinea pig and its alteration following electric shock. Journal of neuropathology and experimental neurology 16, pp. 40-47 (1957). 68. J. D. Bancroft, M. Gamble, Theory and Practice of Histological Techniques. (Churchill Livingstone, 2008). 69. E. Ruska, M. Knoll, Die magnetische Sammelspule für schnelle Elektronenstrahlen. The magnetic concentrating coil for fast electron beams.) Z. techn. Physik 12, pp. 389-400 (1931). 70. L. Marton, Electron microscopy of biological objects. Physical Review 46, pp. 527 (1934). 71. E. Husemann, u. H. Ruska, Versuche zur Sichtbarmachung von Glykogenmolekülen. Journal für Praktische Chemie 156, pp. 1-10 (1940). 72. J. P. Revel, Electron microscopy of glycogen. Journal of Histochemistry & Cytochemistry 12, pp. 104-114 (1964).

199

73. J. P. Revel, L. Napolitano, D. W. Fawcett, Identification of glycogen in electron micrographs of thin tissue sections. The Journal of cell biology 8, pp. 575-589 (1960). 74. P. Drochmans, Morphologie du glycogène: Etude au microscope électronique de colorations négatives du glycogène particulaire. Journal of ultrastructure research 6, pp. 141-163 (1962). 75. J.-C. Wanson, P. Drochmans, RABBIT SKELETAL MUSCLE GLYCOGEN: A Morphological and Biochemical Study of Glycogen ß- Particles Isolated by the Precipitation-Centrifugation Method. The Journal of cell biology 38, pp. 130-150 (1968). 76. D. S. Maxwell, L. Kruger, The fine structure of astrocytes in the cerebral cortex and their response to focal injury produced by heavy ionizing particles. The Journal of cell biology 25, pp. 141-157 (1965). 77. A. M. Cataldo, R. D. Broadwell, Cytochemical identification of cerebral glycogen and glucose‐6‐phosphatase activity under normal and experimental conditions: I. Neurons and glia. Journal of Electron Microscopy Technique 3, pp. 413-437 (1986). 78. J. Koizumi, Glycogen in the central nervous system. Progress in histochemistry and cytochemistry 6, pp. III-35 (1974). 79. J. B. Cavanagh, Corpora-amylacea and the family of polyglucosan diseases. Brain Res Brain Res Rev 29, pp. 265-295 (1999). 80. J. Duran, J. J. Guinovart, Brain glycogen in health and disease. Mol Aspects Med 46, pp. 70-77 (2015). 81. A. W. Merrick, Encephalic glycogen differences in young and adult rats. The Journal of physiology 158, pp. 476-485 (1961). 82. R. Caputto, L. F. Leloir, C. E. Cardini, A. C. Paladini, Isolation of the coenzyme of the galactose phosphate-glucose phosphate transformation. Journal of Biological Chemistry 184, pp. 333-350 (1950). 83. W. Z. Hassid, E. F. Neufeld, D. S. Feingold, Sugar nucleotides in the interconversion of carbohydrates in higher plants. Proceedings of the National Academy of Sciences 45, pp. 905-915 (1959). 84. L. F. Leloir, C. E. Cardini, BIOSYNTHESIS OF GLYCOGEN FROM URIDINE DIPHOSPHATE GLUCOSE1. Journal of the American Chemical Society 79, pp. 6340-6341 (1957). 85. J. Mordoh, C. R. Krisman, L. F. Leloir, Further studies on high molecular weight liver glycogen. Archives of biochemistry and biophysics 113, pp. 265-272 (1966). 86. A. J. Parodi, C. R. Krisman, L. F. Leloir, J. Mordoh, Properties of synthetic and native liver glycogen. Archives of biochemistry and biophysics 121, pp. 769-778 (1967). 87. C. F. Cori, G. T. Cori, Carbohydrate metabolism. Annual review of biochemistry 15, pp. 193-218 (1946). 88. G. Burnett, E. P. Kennedy, The enzymatic phosphorylation of proteins. Journal of Biological Chemistry 211, pp. 969-980 (1954).

200

89. E. H. Fischer, E. G. Krebs, Conversion of phosphorylase b to phosphorylase a in muscle extracts. Journal of Biological Chemistry 216, pp. 121-132 (1955). 90. E. W. Sutherland, W. D. Wosilait, Inactivation and activation of liver phosphorylase. Nature 175, pp. 169 (1955). 91. P. Cohen, The origins of protein phosphorylation. Nature cell biology 4, pp. E127 (2002). 92. D. L. Friedman, J. Larner, Studies on UDPG-α-glucan transglucosylase. III. Interconversion of two forms of muscle UDPG-α-glucan transglucosylase by a phosphorylation-dephosphorylation reaction sequence. Biochemistry 2, pp. 669-675 (1963). 93. P. J. Roach, Control of glycogen synthase by hierarchal protein phosphorylation. The FASEB Journal 4, pp. 2961-2968 (1990). 94. P. Cohen, S. Frame, The renaissance of GSK3. Nature reviews Molecular cell biology 2, pp. 769 (2001). 95. R. S. Jope, G. V. W. Johnson, The glamour and gloom of glycogen synthase kinase-3. Trends in biochemical sciences 29, pp. 95-102 (2004). 96. C. De Duve, B. Pressman, R. Gianetto, R. Wattiaux, F. Appelmans, Tissue fractionation studies. 6. Intracellular distribution patterns of enzymes in rat-liver tissue. Biochemical Journal 60, pp. 604 (1955). 97. H. G. Hers, α-Glucosidase deficiency in generalized glycogen-storage disease (Pompe's disease). Biochemical Journal 86, pp. 11 (1963). 98. N. Raben, A. Wong, E. Ralston, R. Myerowitz, in American Journal of Medical Genetics Part C: Seminars in Medical Genetics. (Wiley Online Library, 2012), vol. 160, pp. 13-21. 99. M. Fuller, P. J. Meikle, J. J. Hopwood, Epidemiology of lysosomal storage diseases: an overview. (2006). 100. P. L. Jeffrey, D. H. Brown, B. I. Brown, Lysosomal α-glucosidase. I. Purification and properties of the rat liver enzyme. Biochemistry 9, pp. 1403-1415 (1970). 101. P. L. Jeffrey, D. H. Brown, B. I. Brown, Lysosomal α-glucosidase. II. Kinetics of action of the rat liver enzyme. Biochemistry 9, pp. 1416-1422 (1970). 102. B. I. Brown, D. H. Brown, P. L. Jeffrey, Simultaneous absence of α-1, 4- glucosidase and α-1, 6-glucosidase activities (pH 4) in tissues of children with type II glycogen storage disease. Biochemistry 9, pp. 1423-1428 (1970). 103. B. E. Ryman, W. J. Whelan, New aspects of glycogen metabolism. Adv Enzymol Relat Areas Mol Biol 34, pp. 285-443 (1971). 104. M. A. Swanson, C. F. Cori, Studies on the structure of polysaccharides III. Relation of structure to activation of phosphorylases. Journal of Biological Chemistry 172, pp. 815-824 (1948). 105. E. Salsas, J. Larner, Glycogen synthase can use glucose as an acceptor. Journal of Biological Chemistry 250, pp. 1833-1837 (1975). 106. C. R. Krisman, R. Barengo, A precursor of glycogen biosynthesis: α‐1, 4‐ glucan‐protein. European journal of Biochemistry 52, pp. 117-123 (1975).

201

107. N. A. Butler, E. Y. C. Lee, W. J. Whelan, A protein-bound glycogen component of rat liver. Carbohydrate research 55, pp. 73-82 (1977). 108. I. R. Rodriguez, W. J. Whelan, A novel glycosyl-amino acid linkage: rabbit- muscle glycogen is covalently linked to a protein via tyrosine. Biochemical and biophysical research communications 132, pp. 829-836 (1985). 109. J. Pitcher, C. Smythe, P. Cohen, Glycogenin is the priming glucosyltransferase required for the initiation of glycogen biogenesis in rabbit skeletal muscle. European journal of biochemistry 176, pp. 391-395 (1988). 110. J. Lomako, W. M. Lomako, W. J. Whelan, A self-glucosylating protein is the primer for rabbit muscle glycogen biosynthesis. The FASEB journal 2, pp. 3097-3103 (1988). 111. J. Lomako, W. M. Lomako, W. J. Whelan, The biogenesis of glycogen: nature of the carbohydrate in the protein primer. Biochemistry international 21, pp. 251-260 (1990). 112. E. Viskupic, Y. Cao, W. Zhang, C. Cheng, A. A. DePaoli-Roach, P. J. Roach, Rabbit skeletal muscle glycogenin. Molecular cloning and production of fully functional protein in Escherichia coli. Journal of Biological Chemistry 267, pp. 25759-25763 (1992). 113. P. J. Roach, Glycogen and its Metabolism. Current Molecular Medicine 2, pp. 101-120 (2002). 114. P. J. Roach, A. A. Depaoli-Roach, T. D. Hurley, V. S. Tagliabracci, Glycogen and its metabolism: some new developments and old themes. The Biochemical journal 441, pp. 763-787 (2012). 115. C. D’Hulst, Á. Mérida, The priming of storage glucan synthesis from bacteria to plants: current knowledge and new developments. New phytologist 188, pp. 13-21 (2010). 116. D. R. Schneider, B. T. Felt, H. Goldman, Microwave radiation energy: a probe for the neurobiologist. Life sciences 29, pp. 643-653 (1981). 117. E. Marani, Microwave applications in neuromorphology and neurochemistry: safety precautions and techniques. Methods 15, pp. 87- 99 (1998). 118. R. A. Swanson, S. M. Sagar, F. R. Sharp, Regional brain glycogen stores and metabolism during complete global ischaemia. Neurological research 11, pp. 24-28 (1989). 119. S. M. Sagar, F. R. Sharp, R. A. Swanson, The regional distribution of glycogen in rat brain fixed by microwave irradiation. Brain research 417, pp. 172-174 (1987). 120. Y. Oe, O. Baba, H. Ashida, K. C. Nakamura, H. Hirase, Glycogen distribution in the microwave‐fixed mouse brain reveals heterogeneous astrocytic patterns. Glia 64, pp. 1532-1545 (2016). 121. A. M. Salvan, J. Vion-Dury, S. Confort-Gouny, P. Dano, P. J. Cozzone, Increased cerebral glycogen detected by localized 1H-magnetic resonance spectroscopy in a patient with suspected McArdle’s disease. European neurology 37, pp. 251-253 (1997).

202

122. I. Y. Choi, I. Tkáčc, K. Ugurbil, R. Gruetter, Noninvasive measurements of [1‐13C] glycogen concentrations and metabolism in rat brain in vivo. Journal of neurochemistry 73, pp. 1300-1308 (1999). 123. G. Öz, P.-G. Henry, E. R. Seaquist, R. Gruetter, Direct, noninvasive measurement of brain glycogen metabolism in humans. Neurochemistry international 43, pp. 323-329 (2003). 124. G. Öz, E. R. Seaquist, A. Kumar, A. B. Criego, L. E. Benedict, J. P. Rao, P.-G. Henry, P.-F. Van De Moortele, R. Gruetter, Human brain glycogen content and metabolism: implications on its role in brain energy metabolism. American Journal of Physiology-Endocrinology and Metabolism 292, pp. E946-E951 (2007). 125. G. Öz, A. Kumar, J. P. Rao, C. T. Kodl, L. Chow, L. E. Eberly, E. R. Seaquist, Human brain glycogen metabolism during and following hypoglycemia. Diabetes, (2009). 126. L. F. Barros, Metabolic signaling by lactate in the brain. Trends in neurosciences 36, pp. 396-404 (2013). 127. C. M. Alberini, E. Cruz, G. Descalzi, B. Bessières, V. Gao, Astrocyte glycogen and lactate: New insights into learning and memory mechanisms. Glia 66, pp. 1244-1262 (2018). 128. L. K. Bak, A. B. Walls, A. Schousboe, H. S. Waagepetersen, Astrocytic glycogen metabolism in the healthy and diseased brain. The Journal of biological chemistry 293, pp. 7108-7116 (2018). 129. G. A. Dienel, N. F. Cruz, Aerobic glycolysis during brain activation: adrenergic regulation and influence of norepinephrine on astrocytic metabolism. Journal of neurochemistry 138, pp. 14-52 (2016). 130. P. J. Magistretti, I. Allaman, Lactate in the brain: from metabolic end- product to signalling molecule. Nat Rev Neurosci 19, pp. 235-249 (2018). 131. K. Gessler, I. Uson, T. Takaha, N. Krauss, S. M. Smith, S. Okada, G. M. Sheldrick, W. Saenger, V-Amylose at atomic resolution: X-ray structure of a cycloamylose with 26 glucose residues (cyclomaltohexaicosaose). Proceedings of the National Academy of Sciences of the United States of America 96, pp. 4246-4251 (1999). 132. J. M. Berg, J. L. Tymoczko, L. Stryer, Biochemistry. (W.H. Freeman, 2002). 133. B. Illingworth, J. Larner, G. T. Cori, Structure of glycogens and amylopectins. I. Enzymatic determination of chain length. The Journal of biological chemistry 199, pp. 631-640 (1952). 134. E. Melendez-Hevia, T. G. Waddell, E. D. Shelton, Optimization of molecular design in the evolution of metabolism: the glycogen molecule. The Biochemical journal 295 ( Pt 2), pp. 477-483 (1993). 135. G. T. Cori, Glycogen structure and enzyme deficiencies in glycogen storage disease. Harvey lectures 48, pp. 145 (1952). 136. R. Melendez, E. Melendez-Hevia, M. Cascante, How did glycogen structure evolve to satisfy the requirement for rapid mobilization of glucose? A problem of physical constraints in structure building. Journal of molecular evolution 45, pp. 446-455 (1997).

203

137. C. Prats, T. E. Graham, J. Shearer, The dynamic life of the glycogen granule. Journal of Biological Chemistry 293, pp. 7089-7098 (2018). 138. J. Shearer, T. E. Graham, Novel aspects of skeletal muscle glycogen and its regulation during rest and exercise. Exerc Sport Sci Rev 32, pp. 120- 126 (2004). 139. S. Emanuelle, M. K. Brewer, D. A. Meekins, M. S. Gentry, Unique carbohydrate binding platforms employed by the glucan phosphatases. Cellular and molecular life sciences : CMLS 73, pp. 2765-2778 (2016). 140. R. Melendez, E. Melendez-Hevia, F. Mas, J. Mach, M. Cascante, Physical constraints in the synthesis of glycogen that influence its structural homogeneity: a two-dimensional approach. Biophysical journal 75, pp. 106-114 (1998). 141. D. J. Manners, Recent developments in our understanding of amylopectin structure. Carbohydrate Polymers 11, pp. 87-112 (1989). 142. A. Buleon, P. Colonna, V. Planchot, S. Ball, Starch granules: structure and biosynthesis. Int J Biol Macromol 23, pp. 85-112 (1998). 143. E. Bertoft, Understanding Starch Structure: Recent Progress. Agronomy 7, (2017). 144. D. J. Manners, Recent developments in our understanding of glycogen structure. Carbohydrate polymers 16, pp. 37-82 (1991). 145. Y. Nakamura, Ed., Starch: Metabolism and Structure, (Springer, Akita, Japan, 2015). 146. S. Pérez, E. Bertoft, The molecular structures of starch components and their contribution to the architecture of starch granules: A comprehensive review. Starch‐Stärke 62, pp. 389-420 (2010). 147. C. R. Krisman, A method for the colorimetric estimation of glycogen with iodine. Analytical biochemistry 4, pp. 17-23 (1962). 148. M. A. Swanson, Studies on the structure of polysaccharides; relation of the iodine color to the structure. The Journal of biological chemistry 172, pp. 825-837 (1948). 149. W. Banks, C. T. Greenwood, K. M. Khan, The interaction of linear, amylose oligomers with iodine. Carbohydrate Research 17, pp. 25-33 (1971). 150. J. M. Bailey, W. J. Whelan, Physical properties of starch I. Relationship between iodine stain and chain length. Journal of Biological Chemistry 236, pp. 969-973 (1961). 151. A. R. Archibald, I. D. Fleming, A. M. Liddle, D. J. Manners, G. A. Mercer, A. Wright, 232. α-1, 4-Glucosans. Part XI. The absorption spectra of glycogen–and amylopectin–iodine complexes. Journal of the Chemical Society (Resumed), pp. 1183-1190 (1961). 152. C. R. Krisman, C. J. Alfredo, Corn Starch (α1, 4‐α1, 6) Glucopolysaccharides‐Correlation Between Amylose: Amylopectin Ratios and Physical Properties of the Grains. Starch‐Stärke 43, pp. 291-294 (1991). 153. D. L. Morris, Colorimetric determination of glycogen. Disadvantages of the iodine method. Journal of Biological Chemistry 166, pp. 199-203 (1946).

204

154. M. Hirai, T. Hirai, T. Ueki, Effect of branching of amylopectin on complexation with iodine as steric hindrance. Polymer 35, pp. 2222-2225 (1994). 155. M. R. S. Rani, K. Shibanuma, S. Hizukuri, The fine structure of oyster glucogen. Carbohydrate research 227, pp. 183-194 (1992). 156. M. Matsui, M. Kakuta, A. Misaki, Comparison of the unit-chain distributions of glycogens from different biological sources, revealed by anion exchange chromatography. Bioscience, biotechnology, and biochemistry 57, pp. 623-627 (1993). 157. I. Hanashiro, J.-i. Abe, S. Hizukuri, A periodic distribution of the chain length of amylopectin as revealed by high-performance anion-exchange chromatography. Carbohydrate Research 283, pp. 151-159 (1996). 158. F. Nitschke, P. Wang, P. Schmieder, J. M. Girard, D. E. Awrey, T. Wang, J. Israelian, X. Zhao, J. Turnbull, M. Heydenreich, E. Kleinpeter, M. Steup, B. A. Minassian, Hyperphosphorylation of glucosyl c6 carbons and altered structure of glycogen in the neurodegenerative epilepsy lafora disease. Cell metabolism 17, pp. 756-767 (2013). 159. F. Nitschke, M. A. Sullivan, P. Wang, X. Zhao, E. E. Chown, A. M. Perri, L. Israelian, L. Juana-Lopez, P. Bovolenta, S. Rodriguez de Cordoba, M. Steup, B. A. Minassian, Abnormal glycogen chain length pattern, not hyperphosphorylation, is critical in Lafora disease. 9, pp. 906-917 (2017). 160. J. M. Irimia, V. S. Tagliabracci, C. M. Meyer, D. M. Segvich, A. A. DePaoli- Roach, P. J. Roach, Muscle glycogen remodeling and glycogen phosphate metabolism following exhaustive exercise of wild type and laforin knockout mice. The Journal of biological chemistry, (2015). 161. D. J. Manners, J. R. Stark, V. Thambyrajah, Iodine Staining Properties of Branched α-(14)-D-Glucanst. Journal of the Japanese Society of Starch Science 30, pp. 13-18 (1983). 162. J. D. Fontana, The presence of phosphate in glycogen. FEBS letters 109, pp. 85-92 (1980). 163. J. Lomako, W. M. Lomako, B. R. Kirkman, W. J. Whelan, The role of phosphate in muscle glycogen. Biofactors 4, pp. 167-171 (1994). 164. J. Lomako, W. M. Lomako, W. J. Whelan, R. B. Marchase, Glycogen contains phosphodiester groups that can be introduced by UDPglucose: glycogen glucose 1-phosphotransferase. FEBS letters 329, pp. 263-267 (1993). 165. M. M. Adeva-Andany, M. González-Lucán, C. Donapetry-García, C. Fernández-Fernández, E. Ameneiros-Rodríguez, Glycogen metabolism in humans. BBA Clin 5, pp. 85-100 (2016). 166. H. O. Akman, A. Oldfors, S. DiMauro, in Neuromuscular Disorders of Infancy, Childhood, and Adolescence: A Clinician's Approach, B. T. Darras, H. R. Jones, M. M. Ryan, D. C. De Vivo, Eds. (Academic Press, 2015), pp. 735-760. 167. M. S. Gentry, J. J. Guinovart, B. A. Minassian, P. J. Roach, J. M. Serratosa, Lafora disease offers a unique window into neuronal glycogen

205

metabolism. The Journal of biological chemistry 293, pp. 7117-7125 (2018). 168. M. Sakai, J. Austin, F. Witmer, L. Trueb, Studies in myoclonus epilepsy (Lafora body form). II. Polyglucosans in the systemic deposits of myoclonus epilepsy and in corpora amylacea. Neurology 20, pp. 160-176 (1970). 169. S. Yokoi, J. Austin, F. Witmer, M. Sakai, Studies in myoclonus epilepsy (Lafora body form). I. Isolation and preliminary characterization of Lafora bodies in two cases. Arch Neurol 19, pp. 15-33 (1968). 170. B. A. Minassian, J. R. Lee, J. A. Herbrick, J. Huizenga, S. Soder, A. J. Mungall, I. Dunham, R. Gardner, C. Y. Fong, S. Carpenter, L. Jardim, P. Satishchandra, E. Andermann, O. C. Snead, 3rd, I. Lopes-Cendes, L. C. Tsui, A. V. Delgado-Escueta, G. A. Rouleau, S. W. Scherer, Mutations in a gene encoding a novel protein tyrosine phosphatase cause progressive myoclonus epilepsy. Nature genetics 20, pp. 171-174 (1998). 171. E. M. Chan, E. J. Young, L. Ianzano, I. Munteanu, X. Zhao, C. C. Christopoulos, G. Avanzini, M. Elia, C. A. Ackerley, N. J. Jovic, S. Bohlega, E. Andermann, G. A. Rouleau, A. V. Delgado-Escueta, B. A. Minassian, S. W. Scherer, Mutations in NHLRC1 cause progressive myoclonus epilepsy. Nature genetics 35, pp. 125-127 (2003). 172. J. M. Serratosa, P. Gomez-Garre, M. E. Gallardo, B. Anta, D. B. de Bernabe, D. Lindhout, P. B. Augustijn, C. A. Tassinari, R. M. Malafosse, M. Topcu, D. Grid, C. Dravet, S. F. Berkovic, S. R. de Cordoba, A novel protein tyrosine phosphatase gene is mutated in progressive myoclonus epilepsy of the Lafora type (EPM2). Human molecular genetics 8, pp. 345- 352 (1999). 173. C. A. Worby, M. S. Gentry, J. E. Dixon, Laforin: A dual specificity phosphatase that dephosphorylates complex carbohydrates. J. Biol. Chem. 281, pp. 30412-30418 (2006). 174. V. S. Tagliabracci, J. Turnbull, W. Wang, J. M. Girard, X. Zhao, A. V. Skurat, A. V. Delgado-Escueta, B. A. Minassian, A. A. Depaoli-Roach, P. J. Roach, Laforin is a glycogen phosphatase, deficiency of which leads to elevated phosphorylation of glycogen in vivo. Proceedings of the National Academy of Sciences of the United States of America 104, pp. 19262- 19266 (2007). 175. M. S. Gentry, C. A. Worby, J. E. Dixon, Insights into Lafora disease: malin is an E3 ubiquitin ligase that ubiquitinates and promotes the degradation of laforin. Proceedings of the National Academy of Sciences of the United States of America 102, pp. 8501-8506 (2005). 176. H. Lohi, L. Ianzano, X.-C. Zhao, E. M. Chan, J. Turnbull, S. W. Scherer, C. A. Ackerley, B. A. Minassian, Novel glycogen synthase kinase 3 and ubiquitination pathways in progressive myoclonus epilepsy. Hum. Mol. Genet. 14, pp. 2727-2736 (2005). 177. S. Ganesh, A. V. Delgado-Escueta, T. Sakamoto, M. R. Avila, J. Machado-Salas, Y. Hoshii, T. Akagi, H. Gomi, T. Suzuki, K. Amano, K. L. Agarwala, Y. Hasegawa, D. S. Bai, T. Ishihara, T. Hashikawa, S. Itohara,

206

E. M. Cornford, H. Niki, K. Yamakawa, Targeted disruption of the Epm2a gene causes formation of Lafora inclusion bodies, neurodegeneration, ataxia, myoclonus epilepsy and impaired behavioral response in mice. Human molecular genetics 11, pp. 1251-1262 (2002). 178. A. A. DePaoli-Roach, V. S. Tagliabracci, D. M. Segvich, C. M. Meyer, J. M. Irimia, P. J. Roach, Genetic depletion of the malin E3 ubiquitin ligase in mice leads to lafora bodies and the accumulation of insoluble laforin. The Journal of biological chemistry 285, pp. 25372-25381 (2010). 179. O. Criado, C. Aguado, J. Gayarre, L. Duran-Trio, A. M. Garcia-Cabrero, S. Vernia, B. San Millan, M. Heredia, C. Roma-Mateo, S. Mouron, L. Juana- Lopez, M. Dominguez, C. Navarro, J. M. Serratosa, M. Sanchez, P. Sanz, P. Bovolenta, E. Knecht, S. Rodriguez de Cordoba, Lafora bodies and neurological defects in malin-deficient mice correlate with impaired autophagy. Human molecular genetics 21, pp. 1521-1533 (2012). 180. J. Valles-Ortega, J. Duran, M. Garcia-Rocha, C. Bosch, I. Saez, L. Pujadas, A. Serafin, X. Canas, E. Soriano, J. M. Delgado-Garcia, A. Gruart, J. J. Guinovart, Neurodegeneration and functional impairments associated with glycogen synthase accumulation in a mouse model of Lafora disease. EMBO molecular medicine 3, pp. 667-681 (2011). 181. E. Tiberia, J. Turnbull, T. Wang, A. Ruggieri, X. C. Zhao, N. Pencea, J. Israelian, Y. Wang, C. A. Ackerley, P. Wang, Y. Liu, B. A. Minassian, Increased laforin and laforin binding to glycogen underlie Lafora body formation in malin-deficient Lafora disease. The Journal of biological chemistry 287, pp. 25650-25659 (2012). 182. J. Duran, A. Gruart, M. Garcia-Rocha, J. M. Delgado-Garcia, J. J. Guinovart, Glycogen accumulation underlies neurodegeneration and autophagy impairment in Lafora disease. Human molecular genetics 23, pp. 3147-3156 (2014). 183. M. S. Gentry, R. H. Dowen, 3rd, C. A. Worby, S. Mattoo, J. R. Ecker, J. E. Dixon, The phosphatase laforin crosses evolutionary boundaries and links carbohydrate metabolism to neuronal disease. The Journal of cell biology 178, pp. 477-488 (2007). 184. D. A. Meekins, M. Raththagala, K. D. Auger, B. D. Turner, D. Santelia, O. Kotting, M. S. Gentry, C. W. Vander Kooi, Mechanistic Insights into Glucan Phosphatase Activity Against Polyglucan Substrates. The Journal of biological chemistry, (2015). 185. M. S. Gentry, R. M. Pace, Conservation of the glucan phosphatase laforin is linked to rates of molecular evolution and the glycogen metabolism of the organism. BMC evolutionary biology 9, pp. 138 (2009). 186. C. W. Vander Kooi, A. O. Taylor, R. M. Pace, D. A. Meekins, H. F. Guo, Y. Kim, M. S. Gentry, Structural basis for the glucan phosphatase activity of Starch Excess4. Proceedings of the National Academy of Sciences of the United States of America 107, pp. 15379-15384 (2010). 187. D. A. Meekins, H. F. Guo, S. Husodo, B. C. Paasch, T. M. Bridges, D. Santelia, O. Kotting, C. W. Vander Kooi, M. S. Gentry, Structure of the Arabidopsis glucan phosphatase like sex four2 reveals a unique

207

mechanism for starch dephosphorylation. The Plant cell 25, pp. 2302- 2314 (2013). 188. M. A. Sullivan, S. Nitschke, M. Steup, B. A. Minassian, F. Nitschke, Pathogenesis of Lafora Disease: Transition of Soluble Glycogen to Insoluble Polyglucosan. International journal of molecular sciences 18, (2017). 189. J. Gayarre, L. Duran-Trio, O. Criado Garcia, C. Aguado, L. Juana-Lopez, I. Crespo, E. Knecht, P. Bovolenta, S. Rodriguez de Cordoba, The phosphatase activity of laforin is dispensable to rescue Epm2a-/- mice from Lafora disease. Brain : a journal of neurology 137, pp. 806-818 (2014). 190. M. Raththagala, M. K. Brewer, M. W. Parker, A. R. Sherwood, B. K. Wong, S. Hsu, T. M. Bridges, B. C. Paasch, L. M. Hellman, S. Husodo, D. A. Meekins, A. O. Taylor, B. D. Turner, K. D. Auger, V. V. Dukhande, S. Chakravarthy, P. Sanz, V. L. Woods, Jr., S. Li, C. W. Vander Kooi, M. S. Gentry, Structural mechanism of laforin function in glycogen dephosphorylation and lafora disease. Molecular cell 57, pp. 261-272 (2015). 191. B. Levin, E. A. Burgess, P. E. Mortimer, Glycogen storage disease type IV, amylopectinosis. Archives of disease in childhood 43, pp. 548 (1968). 192. R. Mahler, Glycogen storage diseases. Journal of Clinical Pathology. Supplement (Ass. Clin. Path.) 2, pp. 32 (1969). 193. A. Blennow, S. B. Engelsen, Helix-breaking news: fighting crystalline starch energy deposits in the cell. Trends in plant science 15, pp. 236-240 (2010). 194. M. Hejazi, J. Fettke, S. Haebel, C. Edner, O. Paris, C. Frohberg, M. Steup, G. Ritte, Glucan, water dikinase phosphorylates crystalline maltodextrins and thereby initiates solubilization. Plant J 55, pp. 323-334 (2008). 195. O. Kotting, J. Kossmann, S. C. Zeeman, J. R. Lloyd, Regulation of starch metabolism: the age of enlightenment? Curr Opin Plant Biol 13, pp. 321- 329 (2010). 196. A. Blennow, A. M. Bay-Smidt, C. E. Olsen, B. L. Møller, The distribution of covalently bound phosphate in the starch granule in relation to starch crystallinity. International journal of biological macromolecules 27, pp. 211- 218 (2000). 197. Y. Takeda, S. Hizukuri, Location of phosphate groups in potato amylopectin. Carbohydrate Research 102, pp. 321-327 (1982). 198. A. Blennow, A. M. Bay-Smidt, C. E. Olsen, B. L. Møller, Analysis of starch- bound glucose 3-phosphate and glucose 6-phosphate using controlled acid treatment combined with high-performance anion-exchange chromatography. Journal of Chromatography A 829, pp. 385-391 (1998). 199. S. T. Lim, T. Kasemsuwan, J.-L. Jane, Characterization of phosphorus in starch by 31P-nuclear magnetic resonance spectroscopy. Cereal Chem. 71, pp. 488-493 (1994).

208

200. G. Ritte, M. Heydenreich, S. Mahlow, S. Haebel, O. Kotting, M. Steup, Phosphorylation of C6- and C3-positions of glucosyl residues in starch is catalysed by distinct dikinases. FEBS letters 580, pp. 4872-4876 (2006). 201. P. I. Hansen, M. Spraul, P. Dvortsak, F. H. Larsen, A. Blennow, M. S. Motawia, S. B. Engelsen, Starch phosphorylation—Maltosidic restrains upon 3′‐and 6′‐phosphorylation investigated by chemical synthesis, molecular dynamics and NMR spectroscopy. Biopolymers: Original Research on Biomolecules 91, pp. 179-193 (2009). 202. O. Kotting, D. Santelia, C. Edner, S. Eicke, T. Marthaler, M. S. Gentry, S. Comparot-Moss, J. Chen, A. M. Smith, M. Steup, G. Ritte, S. C. Zeeman, STARCH-EXCESS4 Is a Laforin-Like Phosphoglucan Phosphatase Required for Starch Degradation in Arabidopsis thaliana. The Plant cell 21, pp. 334-346 (2009). 203. C. Edner, J. Li, T. Albrecht, S. Mahlow, M. Hejazi, H. Hussain, F. Kaplan, C. Guy, S. M. Smith, M. Steup, G. Ritte, Glucan, water dikinase activity stimulates breakdown of starch granules by plastidial beta-amylases. Plant Physiol 145, pp. 17-28 (2007). 204. D. A. Meekins, C. W. Vander Kooi, M. S. Gentry, Structural mechanisms of plant glucan phosphatases in starch metabolism. The FEBS journal 283, pp. 2427-2447 (2016). 205. M. S. Gentry, M. K. Brewer, C. W. Vander Kooi, Structural biology of glucan phosphatases from humans to plants. Current opinion in structural biology 40, pp. 62-69 (2016). 206. D. Santelia, O. Kotting, D. Seung, M. Schubert, M. Thalmann, S. Bischof, D. A. Meekins, A. Lutz, N. Patron, M. S. Gentry, F. H. Allain, S. C. Zeeman, The phosphoglucan phosphatase like sex Four2 dephosphorylates starch at the C3-position in Arabidopsis. The Plant cell 23, pp. 4096-4111 (2011). 207. D. A. Meekins, M. Raththagala, S. Husodo, C. J. White, H. F. Guo, O. Kotting, Phosphoglucan-bound structure of starch phosphatase Starch Excess4 reveals the mechanism for C6 specificity. 111, pp. 7272-7277 (2014). 208. M. S. Gentry, J. E. Dixon, C. A. Worby, Lafora disease: insights into neurodegeneration from plant metabolism. Trends Biochem Sci 34, pp. 628-639 (2009). 209. V. S. Tagliabracci, J. M. Girard, D. Segvich, C. Meyer, J. Turnbull, X. Zhao, B. A. Minassian, A. A. Depaoli-Roach, P. J. Roach, Abnormal metabolism of glycogen phosphate as a cause for Lafora disease. The Journal of biological chemistry 283, pp. 33816-33825 (2008). 210. A. A. DePaoli-Roach, C. J. Contreras, D. M. Segvich, C. Heiss, M. Ishihara, P. Azadi, P. J. Roach, Glycogen phosphomonoester distribution in mouse models of the progressive myoclonic epilepsy, Lafora disease. The Journal of biological chemistry, (2014). 211. J. Turnbull, P. Wang, J. M. Girard, A. Ruggieri, T. J. Wang, A. G. Draginov, A. P. Kameka, N. Pencea, X. Zhao, C. A. Ackerley, B. A.

209

Minassian, Glycogen hyperphosphorylation underlies lafora body formation. Annals of neurology 68, pp. 925-933 (2010). 212. V. S. Tagliabracci, C. Heiss, C. Karthik, C. J. Contreras, J. Glushka, M. Ishihara, P. Azadi, T. D. Hurley, A. A. DePaoli-Roach, P. J. Roach, Phosphate incorporation during glycogen synthesis and Lafora disease. Cell metabolism 13, pp. 274-282 (2011). 213. C. J. Contreras, D. M. Segvich, K. Mahalingan, V. M. Chikwana, T. L. Kirley, T. D. Hurley, A. A. DePaoli-Roach, P. J. Roach, Incorporation of phosphate into glycogen by glycogen synthase. Archives of biochemistry and biophysics 597, pp. 21-29 (2016). 214. F. Maley, J. F. McGarrahan, R. DelGiacco, Galactosamine: a precursor of glycogen glucosamine. Biochemical and biophysical research communications 23, pp. 85-91 (1966). 215. A. L. Tarentino, F. Maley, Direct evidence that D‐galactosamine incorporation into glycogen occurs via UDP‐glucosamine. FEBS letters 69, pp. 175-178 (1976). 216. P. A. Romero, E. E. Smith, W. J. Whelan, Glucosamine as a substitute for glucose in glycogen metabolism. Biochemistry International 1, pp. 1-9 (1980). 217. B. R. Kirkman, W. J. Whelan, Glucosamine is a normal component of liver glycogen. FEBS letters 194, pp. 6-11 (1986). 218. L. D. Khac, D. Eboué‐bonis, A. M. Chambaut, H. Clauser, The Use of Glucosamine as a Metabolic Probe in the Rat Diaphragm: The Effect of Insulin on Its Metabolic Fate. European journal of biochemistry 31, pp. 86- 94 (1972). 219. A. Tavridou, L. Agius, Phosphorylase regulates the association of glycogen synthase with a proteoglycogen substrate in hepatocytes. FEBS letters 551, pp. 87-91 (2003). 220. G. J. Parker, K. C. Lund, R. P. Taylor, D. A. McClain, Insulin resistance of glycogen synthase mediated by O-linked N-acetylglucosamine. Journal of Biological Chemistry 278, pp. 10022-10027 (2003). 221. R. J. McMahon, S. C. Frost, Glycogen: a carbohydrate source for GLUT-1 glycosylation during glucose deprivation of 3T3-L1 adipocytes. American Journal of Physiology-Endocrinology and Metabolism 270, pp. E640-E645 (1996). 222. N. Ondruskova, T. Honzik, H. Kolarova, Z. Pakanova, J. Mucha, J. Zeman, H. Hansikova, Aberrant apolipoprotein C-III glycosylation in glycogen storage disease type III and IX. Metabolism 82, pp. 135-141 (2018). 223. L. C. Tegtmeyer, S. Rust, M. van Scherpenzeel, B. G. Ng, M.-E. Losfeld, S. Timal, K. Raymond, P. He, M. Ichikawa, J. Veltman, Multiple phenotypes in phosphoglucomutase 1 deficiency. New England Journal of Medicine 370, pp. 533-542 (2014). 224. B. H. Hayee, A. Antonopoulos, E. J. Murphy, F. Z. Rahman, G. Sewell, B. N. Smith, S. McCartney, M. Furman, G. Hall, S. L. Bloom, G6PC3 mutations are associated with a major defect of glycosylation: a novel

210

mechanism for neutrophil dysfunction. Glycobiology 21, pp. 914-924 (2011). 225. J. A. Campbell, G. J. Davies, V. Bulone, B. Henrissat, A classification of nucleotide-diphospho-sugar glycosyltransferases based on amino acid sequence similarities. Biochemical Journal 326, pp. 929 (1997). 226. M. D. Alonso, J. Lomako, W. M. Lomako, W. J. Whelan, A new look at the biogenesis of glycogen. The FASEB journal 9, pp. 1126-1137 (1995). 227. J. Lomako, W. M. Lomako, W. J. Whelan, Glycogenin: the primer for mammalian and yeast glycogen synthesis. Biochimica et Biophysica Acta (BBA)-General Subjects 1673, pp. 45-55 (2004). 228. B. J. Gibbons, P. J. Roach, T. D. Hurley, Crystal structure of the autocatalytic initiator of glycogen biosynthesis, glycogenin. Journal of molecular biology 319, pp. 463-477 (2002). 229. A. Chaikuad, D. S. Froese, G. Berridge, F. von Delft, U. Oppermann, W. W. Yue, Conformational plasticity of glycogenin and its maltosaccharide substrate during glycogen biogenesis. Proceedings of the National Academy of Sciences 108, pp. 21028-21033 (2011). 230. C. Smythe, P. Cohen, in EJB Reviews 1991. (Springer, 1991), pp. 149- 155. 231. M. D. Alonso, J. Lomako, W. M. Lomako, W. J. Whelan, Tyrosine‐194 of glycogenin undergoes autocatalytic glucosylation but is not essential for catalytic function and activity. FEBS letters 342, pp. 38-42 (1994). 232. Y. Cao, A. M. Mahrenholz, A. A. DePaoli-Roach, P. J. Roach, Characterization of rabbit skeletal muscle glycogenin. Tyrosine 194 is essential for function. Journal of Biological Chemistry 268, pp. 14687- 14693 (1993). 233. A. Lin, J. Mu, J. Yang, P. J. Roach, Self-glucosylation of glycogenin, the initiator of glycogen biosynthesis, involves an inter-subunit reaction. Archives of biochemistry and biophysics 363, pp. 163-170 (1999). 234. J. Mu, A. V. Skurat, P. J. Roach, Glycogenin-2, a novel self-glucosylating protein involved in liver glycogen biosynthesis. Journal of Biological Chemistry 272, pp. 27589-27597 (1997). 235. A. V. Skurat, A. D. Dietrich, P. J. Roach, Interaction between glycogenin and glycogen synthase. Archives of biochemistry and biophysics 456, pp. 93-97 (2006). 236. E. Zeqiraj, X. Tang, R. W. Hunter, M. García-Rocha, A. Judd, M. Deak, A. von Wilamowitz-Moellendorff, I. Kurinov, J. J. Guinovart, M. Tyers, Structural basis for the recruitment of glycogen synthase by glycogenin. Proceedings of the National Academy of Sciences, pp. 201402926 (2014). 237. A. V. Skurat, Y. Cao, P. J. Roach, Glucose control of rabbit skeletal muscle glycogenin expressed in COS cells. Journal of Biological Chemistry 268, pp. 14701-14707 (1993). 238. Y. Cao, A. V. Skurat, A. A. DePaoli-Roach, P. J. Roach, Initiation of glycogen synthesis. Control of glycogenin by glycogen phosphorylase. Journal of Biological Chemistry 268, pp. 21717-21721 (1993).

211

239. T. E. Graham, Z. Yuan, A. K. Hill, R. J. Wilson, The regulation of muscle glycogen: the granule and its proteins. Acta physiologica (Oxford, England) 199, pp. 489-498 (2010). 240. E. Malfatti, J. Nilsson, C. Hedberg‐Oldfors, A. Hernandez‐Lain, F. Michel, C. Dominguez‐Gonzalez, G. Viennet, H. O. Akman, C. Kornblum, P. Van den Bergh, A new muscle glycogen storage disease associated with glycogenin‐1 deficiency. Annals of neurology 76, pp. 891-898 (2014). 241. G. Testoni, J. Duran, M. García-Rocha, F. Vilaplana, A. L. Serrano, D. Sebastián, I. López-Soldado, M. A. Sullivan, F. Slebe, M. Vilaseca, Lack of glycogenin causes glycogen accumulation and muscle function impairment. Cell metabolism 26, pp. 256-266. e254 (2017). 242. S. Ros, M. García-Rocha, J. Domínguez, J. C. Ferrer, J. J. Guinovart, Control of liver glycogen synthase activity and intracellular distribution by phosphorylation. Journal of Biological Chemistry 284, pp. 6370-6378 (2009). 243. J. C. Lawrence, P. J. Roach, New insights into the role and mechanism of glycogen synthase activation by insulin. Diabetes 46, pp. 541-547 (1997). 244. B. Pfeiffer-Guglielmi, S. Bröer, A. Bröer, B. Hamprecht, Isozyme pattern of glycogen phosphorylase in the rat nervous system and rat astroglia-rich primary cultures: electrophoretic and polymerase chain reaction studies. Neurochemical research 25, pp. 1485-1491 (2000). 245. O. W. Nadeau, J. D. Fontes, G. M. Carlson, The regulation of glycogenolysis in the brain. The Journal of biological chemistry 293, pp. 7099-7107 (2018). 246. C. Mathieu, I. L. De La Sierra-Gallay, R. Duval, X. Xu, A. Cocaign, T. Léger, G. Woffendin, J.-M. Camadro, C. Etchebest, A. Haouz, Insights into brain glycogen metabolism: the structure of human brain glycogen phosphorylase. Journal of Biological Chemistry, pp. jbc. M116. 738898 (2016). 247. V. L. Rath, M. Ammirati, P. K. LeMotte, K. F. Fennell, M. N. Mansour, D. E. Danley, T. R. Hynes, G. K. Schulte, D. J. Wasilko, J. Pandit, Activation of human liver glycogen phosphorylase by alteration of the secondary structure and packing of the catalytic core. Molecular cell 6, pp. 139-148 (2000). 248. C. M. Lukacs, N. G. Oikonomakos, R. L. Crowther, L. N. Hong, R. U. Kammlott, W. Levin, S. Li, C. M. Liu, D. Lucas-McGady, S. Pietranico, L. Reik, The crystal structure of human muscle glycogen phosphorylase a with bound glucose and AMP: an intermediate conformation with T-state and R-state features. Proteins 63, pp. 1123-1126 (2006). 249. R. Geddes, Glycogen: a metabolic viewpoint. Bioscience reports 6, pp. 415-428 (1986). 250. D. S. Froese, A. Michaeli, T. J. McCorvie, T. Krojer, M. Sasi, E. Melaev, A. Goldblum, M. Zatsepin, A. Lossos, R. Álvarez, Structural basis of glycogen branching enzyme deficiency and pharmacologic rescue by rational peptide design. Human molecular genetics 24, pp. 5667-5676 (2015).

212

251. L. Zhai, L. Feng, L. Xia, H. Yin, S. Xiang, Crystal structure of glycogen debranching enzyme and insights into its catalysis and disease-causing mutations. Nature communications 7, pp. 11229 (2016). 252. S. Janecek, B. Svensson, E. A. MacGregor, Structural and evolutionary aspects of two families of non-catalytic domains present in starch and glycogen binding proteins from microbes, plants and animals. Enzyme and microbial technology 49, pp. 429-440 (2011). 253. W. Verhue, H. G. Hers, A study of the reaction catalysed by the liver branching enzyme. Biochemical Journal 99, pp. 222 (1966). 254. D. H. Brown, B. I. Brown, Action of a muscle branching enzyme on polysaccharides enlarged from UDP [14C] glucose. Biochimica et Biophysica Acta (BBA)-General Subjects 130, pp. 263-266 (1966). 255. W. B. Gibson, B. I. Brown, D. H. Brown, Glycogen branching enzyme. Preparation and properties of α-1, 4-glucan-α-1, 4-glucan 6- glycosyltransferase and its action on the characteristic polysaccharide of the liver of children with type IV glycogen storage disease. Biochemistry 10, pp. 4253-4262 (1971). 256. D. S. Tolmasky, C. R. Krisman, The degree of branching in (α1, 4)‐(α1, 6)‐ linked glucopolysaccharides is dependent on intrinsic properties of the branching enzymes. European journal of biochemistry 168, pp. 393-397 (1987). 257. T. Kuriki, D. C. Stewart, J. Preiss, Construction of chimeric enzymes out of maize endosperm branching enzymes I and II: activity and properties. Journal of Biological Chemistry 272, pp. 28999-29004 (1997). 258. H. Kajiura, R. Kakutani, T. Akiyama, H. Takata, T. Kuriki, A novel enzymatic process for glycogen production. Biocatalysis and Biotransformation 26, pp. 133-140 (2008). 259. H. Kajiura, H. Takata, T. Kuriki, S. Kitamura, Structure and solution properties of enzymatically synthesized glycogen. Carbohydrate research 345, pp. 817-824 (2010). 260. H. Takata, H. Kajiura, T. Furuyashiki, R. Kakutani, T. Kuriki, Fine structural properties of natural and synthetic glycogens. Carbohydrate research 344, pp. 654-659 (2009). 261. N. Raben, M. Danon, N. Lu, E. Lee, L. Shliselfeld, A. V. Skurat, P. J. Roach, J. C. Lawrence, Jr., O. Musumeci, S. Shanske, S. DiMauro, P. Plotz, Surprises of genetic engineering: A possible model of polyglucosan body disease. Neurology 56, pp. 1739-1745 (2001). 262. O. Kakhlon, H. Glickstein, N. Feinstein, Y. Liu, O. Baba, T. Terashima, H. O. Akman, S. Dimauro, A. Lossos, Polyglucosan neurotoxicity caused by glycogen branching enzyme deficiency can be reversed by inhibition of glycogen synthase. Journal of neurochemistry 127, pp. 101-113 (2013). 263. A. Nakayama, K. Yamamoto, S. Tabata, Identification of the catalytic residues of bifunctional glycogen debranching enzyme. Journal of Biological Chemistry 276, pp. 28824-28828 (2001). 264. S. Baskaran, P. J. Roach, A. A. DePaoli-Roach, T. D. Hurley, Structural basis for glucose-6-phosphate activation of glycogen synthase.

213

Proceedings of the National Academy of Sciences 107, pp. 17563-17568 (2010). 265. M. M. Hermans, M. A. Kroos, J. Van Beeumen, B. A. Oostra, A. J. Reuser, Human lysosomal alpha-glucosidase. Characterization of the catalytic site. Journal of Biological Chemistry 266, pp. 13507-13512 (1991). 266. L. H. Hoefsloot, M. Hoogeveen-Westerveld, A. J. J. Reuser, B. A. Oostra, Characterization of the human lysosomal α-glucosidase gene. Biochemical journal 272, pp. 493-497 (1990). 267. H. A. Wisselaar, M. A. Kroos, M. M. Hermans, J. Van Beeumen, A. J. Reuser, Structural and functional changes of lysosomal acid alpha- glucosidase during intracellular transport and maturation. Journal of Biological Chemistry 268, pp. 2223-2231 (1993). 268. T. N. Palmer, The substrate specificity of acid α-glucosidase from rabbit muscle. Biochemical Journal 124, pp. 701-711 (1971). 269. H. Matsui, M. Sasaki, E. Takemasa, T. Kaneta, S. Chiba, Kinetic studies on the substrate specificity and active site of rabbit muscle acid α- glucosidase. The Journal of Biochemistry 96, pp. 993-1004 (1984). 270. G. M. Mancini, C. E. Beerens, F. W. Verheijen, Glucose transport in lysosomal membrane vesicles. Kinetic demonstration of a carrier for neutral hexoses. Journal of Biological Chemistry 265, pp. 12380-12387 (1990). 271. N. Mizushima, M. Komatsu, Autophagy: renovation of cells and tissues. Cell 147, pp. 728-741 (2011). 272. A. M. Cuervo, E. Bergamini, U. T. Brunk, W. Dröge, M. Ffrench, A. Terman, Autophagy and aging: the importance of maintaining" clean" cells. Autophagy 1, pp. 131-140 (2005). 273. J. L. Schneider, A. M. Cuervo, Autophagy and human disease: emerging themes. Current opinion in genetics & development 26, pp. 16-23 (2014). 274. M. Komatsu, S. Waguri, T. Chiba, S. Murata, J.-i. Iwata, I. Tanida, T. Ueno, M. Koike, Y. Uchiyama, E. Kominami, Loss of autophagy in the central nervous system causes neurodegeneration in mice. Nature 441, pp. 880 (2006). 275. H. Zhao, M. Tang, M. Liu, L. Chen, Glycophagy: An emerging target in pathology. Clinica Chimica Acta, (2018). 276. S. Jiang, C. D. Wells, P. J. Roach, Starch-binding domain-containing protein 1 (Stbd1) and glycogen metabolism: Identification of the Atg8 family interacting motif (AIM) in Stbd1 required for interaction with GABARAPL1. Biochemical and biophysical research communications 413, pp. 420-425 (2011). 277. J. B. Cavanagh, H. B. Jones, Glycogenosomes in the aging rat brain: their occurrence in the visual pathways. Acta neuropathologica 99, pp. 496-502 (2000). 278. T. Iwamasa, T. Tsuru, T. Hamada, T. Takeuchi, Physicochemical and ultrastructural studies on glycogenosomes in newborn rat hepatocytes. Pathology-Research and Practice 167, pp. 363-373 (1980).

214

279. T. Iwamasa, S. Fukuda, S. Tokumitsu, N. Ninomiya, I. Matsuda, M. Osame, Myopathy due to glycogen storage disease: pathological and biochemical studies in relation to glycogenosome formation. Experimental and molecular pathology 38, pp. 405-420 (1983). 280. N. Raben, V. Hill, L. Shea, S. Takikita, R. Baum, N. Mizushima, E. Ralston, P. Plotz, Suppression of autophagy in skeletal muscle uncovers the accumulation of ubiquitinated proteins and their potential role in muscle damage in Pompe disease. Human molecular genetics 17, pp. 3897-3908 (2008). 281. J.-A. Lim, H. Zare, R. Puertollano, N. Raben, Atg5flox-derived autophagy- deficient model of Pompe disease: does it tell the whole story? Molecular Therapy-Methods & Clinical Development 7, pp. 11-14 (2017). 282. N. Raben, C. Schreiner, R. Baum, S. Takikita, S. Xu, T. Xie, R. Myerowitz, M. Komatsu, J. H. Van Der Meulen, K. Nagaraju, Suppression of autophagy permits successful enzyme replacement therapy in a lysosomal storage disorder—murine Pompe disease. Autophagy 6, pp. 1078-1089 (2010). 283. H. Nakatogawa, K. Suzuki, Y. Kamada, Y. Ohsumi, Dynamics and diversity in autophagy mechanisms: lessons from yeast. Nature reviews Molecular cell biology 10, pp. 458 (2009). 284. Š. Janeček, A motif of a microbial starch-binding domain found in human genethonin. Bioinformatics 18, pp. 1534-1537 (2002). 285. J. Wang, J. A. Stuckey, M. J. Wishart, J. E. Dixon, A unique carbohydrate binding domain targets the lafora disease phosphatase to glycogen. The Journal of biological chemistry 277, pp. 2377-2380 (2002). 286. S. Jiang, B. Heller, V. S. Tagliabracci, L. Zhai, J. M. Irimia, A. A. DePaoli- Roach, C. D. Wells, A. V. Skurat, P. J. Roach, Starch binding domain containing protein 1/genethonin 1 is a novel participant in glycogen metabolism. Journal of Biological Chemistry, pp. jbc. M110. 150839 (2010). 287. A. Demetriadou, J. Morales-Sanfrutos, M. Nearchou, O. Baba, K. Kyriacou, E. W. Tate, A. Drousiotou, P. P. Petrou, Mouse Stbd1 is N- myristoylated and affects ER-mitochondria association and mitochondrial morphology. Journal of cell science, pp. jcs. 195263 (2017). 288. H. Yi, K. B. Fredrickson, S. Das, P. S. Kishnani, B. Sun, Stbd1 is highly elevated in skeletal muscle of Pompe disease mice but suppression of its expression does not affect lysosomal glycogen accumulation. Molecular genetics and metabolism 109, pp. 312-314 (2013). 289. T. Sun, H. Yi, C. Yang, P. S. Kishnani, B. Sun, Starch binding domain- containing protein 1 plays a dominant role in glycogen transport to lysosomes in liver. Journal of Biological Chemistry 291, pp. 16479-16484 (2016). 290. R. Puri, T. Suzuki, K. Yamakawa, S. Ganesh, Dysfunctions in endosomal- lysosomal and autophagy pathways underlie neuropathology in a mouse model for Lafora disease. Human molecular genetics 21, pp. 175-184 (2012).

215

291. C. Aguado, S. Sarkar, V. I. Korolchuk, O. Criado, S. Vernia, P. Boya, P. Sanz, S. R. de Cordoba, E. Knecht, D. C. Rubinsztein, Laforin, the most common protein mutated in Lafora disease, regulates autophagy. Human molecular genetics 19, pp. 2867-2876 (2010). 292. P. Wang, L. Israelian, Y. Xue, S. Song, L. Attisano, B. A. Minassian, SGK1 (glucose transport), dishevelled2 (wnt signaling), LC3/p62 (autophagy) and p53 (apoptosis) proteins are unaltered in Lafora disease. The all results journals. Biol 7, pp. 28 (2016). 293. B. A. Pederson, J. Turnbull, J. R. Epp, S. A. Weaver, X. Zhao, N. Pencea, P. J. Roach, P. W. Frankland, C. A. Ackerley, B. A. Minassian, Inhibiting glycogen synthesis prevents Lafora disease in a mouse model. Annals of neurology 74, pp. 297-300 (2013). 294. J. Turnbull, J. R. Epp, D. Goldsmith, X. Zhao, N. Pencea, P. Wang, P. W. Frankland, C. A. Ackerley, B. A. Minassian, PTG protein depletion rescues malin-deficient Lafora disease in mouse. Annals of neurology 75, pp. 442- 446 (2014). 295. J. Turnbull, A. A. Depaoli-Roach, X. Zhao, M. A. Cortez, N. Pencea, E. Tiberia, M. Piliguian, P. J. Roach, P. Wang, C. A. Ackerley, B. A. Minassian, PTG Depletion Removes Lafora Bodies and Rescues the Fatal Epilepsy of Lafora Disease. PLoS genetics 7, pp. e1002037 (2011). 296. E. Seranova, K. J. Connolly, M. Zatyka, T. R. Rosenstock, T. Barrett, R. I. Tuxworth, S. Sarkar, Dysregulation of autophagy as a common mechanism in lysosomal storage diseases. Essays in biochemistry 61, pp. 733-749 (2017). 297. A. P. Lieberman, R. Puertollano, N. Raben, S. Slaugenhaupt, S. U. Walkley, A. Ballabio, Autophagy in lysosomal storage disorders. Autophagy 8, pp. 719-730 (2012). 298. C. Munoz-Ballester, A. Berthier, R. Viana, P. Sanz, Homeostasis of the astrocytic glutamate transporter GLT-1 is altered in mouse models of Lafora disease. Biochimica et biophysica acta 1862, pp. 1074-1083 (2016). 299. M. A. Garcia-Gimeno, E. Knecht, P. Sanz, Lafora Disease: A Ubiquitination-Related Pathology. 7, (2018). 300. D. Vilchez, S. Ros, D. Cifuentes, L. Pujadas, J. Valles, B. Garcia-Fojeda, O. Criado-Garcia, E. Fernandez-Sanchez, I. Medrano-Fernandez, J. Dominguez, M. Garcia-Rocha, E. Soriano, S. Rodriguez de Cordoba, J. J. Guinovart, Mechanism suppressing glycogen synthesis in neurons and its demise in progressive myoclonus epilepsy. Nat Neurosci 10, pp. 1407- 1413 (2007). 301. C. A. Worby, M. S. Gentry, J. E. Dixon, Malin decreases glycogen accumulation by promoting the degradation of protein targeting to glycogen (PTG). The Journal of biological chemistry 283, pp. 4069-4076 (2008). 302. C. Rubio-Villena, M. A. Garcia-Gimeno, P. Sanz, Glycogenic activity of R6, a protein phosphatase 1 regulatory subunit, is modulated by the

216

laforin-malin complex. The international journal of biochemistry & cell biology 45, pp. 1479-1488 (2013). 303. D. Moreno, M. C. Towler, D. G. Hardie, E. Knecht, P. Sanz, The laforin- malin complex, involved in Lafora disease, promotes the incorporation of K63-linked ubiquitin chains into AMP-activated protein kinase beta subunits. Molecular biology of the cell 21, pp. 2578-2588 (2010). 304. A. Cheng, M. Zhang, M. S. Gentry, C. A. Worby, J. E. Dixon, A. R. Saltiel, A role for AGL ubiquitination in the glycogen storage disorders of Lafora and Cori's disease. Genes Dev. 21, pp. 2399-2409 (2007). 305. W. Wang, H. Lohi, A. V. Skurat, A. A. Depaoli-Roach, B. A. Minassian, P. J. Roach, Glycogen metabolism in tissues from a mouse model of Lafora disease. Archives of biochemistry and biophysics 457, pp. 264-269 (2007). 306. C. Roma-Mateo, D. Moreno, S. Vernia, T. Rubio, T. M. Bridges, M. S. Gentry, P. Sanz, Lafora disease E3-ubiquitin ligase malin is related to TRIM32 at both the phylogenetic and functional level. BMC evolutionary biology 11, pp. 225 (2011). 307. P. Sanchez-Martin, C. Roma-Mateo, R. Viana, P. Sanz, Ubiquitin conjugating enzyme E2-N and sequestosome-1 (p62) are components of the ubiquitination process mediated by the malin-laforin E3-ubiquitin ligase complex. The international journal of biochemistry & cell biology 69, pp. 204-214 (2015). 308. S. Shaid, C. H. Brandts, H. Serve, I. Dikic, Ubiquitination and selective autophagy. Cell death and differentiation 20, pp. 21 (2013). 309. J. M. M. Tan, E. S. P. Wong, V. L. Dawson, T. Dawson, K.-L. Lim, Lysine 63-linked polyubiquitin potentially partners with p62 to promote the clearance of protein inclusions by autophagy. Autophagy 4, pp. 251-253 (2008). 310. M. S. Gentry, C. Roma-Mateo, P. Sanz, Laforin, a protein with many faces: glucan phosphatase, adapter protein, et alii. The FEBS journal 280, pp. 525-537 (2013). 311. Y.-K. Lee, J.-A. Lee, Role of the mammalian ATG8/LC3 family in autophagy: differential and compensatory roles in the spatiotemporal regulation of autophagy. BMB reports 49, pp. 424 (2016). 312. P. Garyali, D. M. Segvich, A. A. DePaoli-Roach, P. J. Roach, Protein degradation and quality control in cells from laforin and malin knockout mice. The Journal of biological chemistry 289, pp. 20606-20614 (2014). 313. C. Roma-Mateo, C. Solaz-Fuster Mdel, J. V. Gimeno-Alcaniz, V. V. Dukhande, J. Donderis, C. A. Worby, A. Marina, O. Criado, A. Koller, S. Rodriguez De Cordoba, M. S. Gentry, P. Sanz, Laforin, a dual-specificity phosphatase involved in Lafora disease, is phosphorylated at Ser25 by AMP-activated protein kinase. The Biochemical journal 439, pp. 265-275 (2011). 314. E. M. Chan, C. A. Ackerley, H. Lohi, L. Ianzano, M. A. Cortez, P. Shannon, S. W. Scherer, B. A. Minassian, Laforin preferentially binds the neurotoxic starch-like polyglucosans, which form in its absence in

217

progressive myoclonus epilepsy. Human molecular genetics 13, pp. 1117- 1129 (2004). 315. J. Turnbull, J. M. Girard, N. Pencea, X. Zhao, T. E. Graham, P. Wang, C. A. Ackerley, B. A. Minassian, Lafora bodies in skeletal muscle are fiber type specific. Neurology 76, pp. 1674-1676 (2011). 316. D. M. Andrade, C. A. Ackerley, T. S. C. Minett, H. A. G. Teive, S. Bohlega, S. W. Scherer, B. A. Minassian, Skin biopsy in Lafora disease: genotype– phenotype correlations and diagnostic pitfalls. Neurology 61, pp. 1611- 1614 (2003). 317. S. Carpenter, G. Karpati, F. Andermann, J. C. Jacob, E. Andermann, Lafora's disease: peroxisomal storage in skeletal muscle. Neurology 24, pp. 531-538 (1974). 318. B. Quistorff, Gluconeogenesis in periportal and perivenous hepatocytes of rat liver, isolated by a new high-yield digitonin/collagenase perfusion technique. Biochemical Journal 229, pp. 221-226 (1985). 319. M. K. Vøllestad, O. D. D. Vaage, L. Hermansen, Muscle glycogen depletion patterns in type I and subgroups of type II fibres during prolonged severe exercise in man: Glycogen depletion in muscle fibres during exercise. Acta Physiologica Scandinavica 122, pp. 433-441 (1984). 320. W. Montagna, H. B. Chase, J. B. Hamilton, The distribution of glycogen and lipids in human skin. Journal of Investigative Dermatology 17, pp. 147-157 (1951). 321. G. R. Lafora, Uber des Vorkommen amyloider KJrperchen im innern der Ganglienzellen. Virchows Arch. f. Path. Anat. 205, pp. 295 (1911). 322. B. A. Minassian, Lafora's disease: towards a clinical, pathologic, and molecular synthesis. Pediatric neurology 25, pp. 21-29 (2001). 323. I. Saez, J. Duran, C. Sinadinos, A. Beltran, O. Yanes, M. F. Tevy, C. Martinez-Pons, M. Milan, J. J. Guinovart, Neurons have an active glycogen metabolism that contributes to tolerance to hypoxia. Journal of cerebral blood flow and metabolism : official journal of the International Society of Cerebral Blood Flow and Metabolism 34, pp. 945-955 (2014). 324. E. Augé, C. Pelegrí , G. Manich, I. Cabezón, J. J. Guinovart, J. Duran, J. Vilaplana, Astrocytes and neurons produce distinct types of polyglucosan bodies in Lafora Disease. Glia in press, (2018). 325. C. Rubio-Villena, R. Viana, J. Bonet, M. A. Garcia-Gimeno, M. Casado, M. Heredia, P. Sanz, Astrocytes: new players in progressive myoclonus epilepsy of Lafora type. Human molecular genetics, (2018). 326. M. W. Van Heycop Ten Ham, in Handbook of clinical neurology, P. J. Vinken, G. W. Bruyn, Eds. (North Holland Publishing Company, Holland, Amsterdam, 1975), vol. 15, pp. 382-422. 327. G. A. Schwarz, M. Yanoff, Lafora's Disease. Distinct Clinico-Pathologic Form of Unverricht's Syndrome. Archives of neurology 12, pp. 172-188 (1965). 328. M. K. Brewer, T. R. Grossman, T. R. McKnight, Y. P. Goldberg, H. Landy, M. S. Gentry, The 4th International Lafora Epilepsy Workshop: Shifting

218

paradigms, paths to treatment, and hope for patients. Epilepsy Behav 90, pp. 284-286 (2019). 329. O. B. Kotoulas, S. A. Kalamidas, D. J. Kondomerkos, Glycogen autophagy. Microscopy research and technique 64, pp. 10-20 (2004). 330. O. B. Kotoulas, S. A. Kalamidas, D. J. Kondomerkos, Glycogen autophagy in glucose homeostasis. Pathology-Research and Practice 202, pp. 631- 638 (2006). 331. S. Schiaffino, C. Mammucari, M. Sandri, The role of autophagy in neonatal tissues: just a response to amino acid starvation? Autophagy 4, pp. 727- 730 (2008). 332. L. M. D. Delbridge, K. M. Mellor, D. J. R. Taylor, R. A. Gottlieb, Myocardial autophagic energy stress responses—macroautophagy, mitophagy, and glycophagy. American Journal of Physiology-Heart and Circulatory Physiology 308, pp. H1194-H1204 (2015). 333. M. E. Reichelt, K. M. Mellor, C. L. Curl, D. Stapleton, L. M. D. Delbridge, Myocardial glycophagy—a specific glycogen handling response to metabolic stress is accentuated in the female heart. Journal of molecular and cellular cardiology 65, pp. 67-75 (2013). 334. P. Devos, H.-G. Hers, Glycogen in rat adipose tissue: sequential synthesis and random degradation. Biochemical and biophysical research communications 95, pp. 1031-1036 (1980). 335. R. Geddes, G. C. Stratton, The influence of lysosomes on glycogen metabolism. Biochemical Journal 163, pp. 193-200 (1977). 336. N. P. Chee, R. Geddes, P. R. Wills, Metabolic heterogeneity in rabbit brain glycogen. Biochimica et Biophysica Acta (BBA)-General Subjects 756, pp. 9-12 (1983). 337. J. H. Ryu, J. Drain, J. H. Kim, S. McGee, A. Gray-Weale, L. Waddington, G. J. Parker, M. Hargreaves, S. H. Yoo, D. Stapleton, Comparative structural analyses of purified glycogen particles from rat liver, human skeletal muscle and commercial preparations. Int J Biol Macromol 45, pp. 478-482 (2009). 338. R. G. Gilbert, M. A. Sullivan, The molecular size distribution of glycogen and its relevance to diabetes. Australian Journal of Chemistry 67, pp. 538- 543 (2014). 339. J. Shearer, T. E. Graham, New perspectives on the storage and organization of muscle glycogen. Canadian journal of applied physiology 27, pp. 179-203 (2002). 340. M. A. Sullivan, M. J. O’Connor, F. Umana, E. Roura, K. Jack, D. I. Stapleton, R. G. Gilbert, Molecular insights into glycogen α-particle formation. Biomacromolecules 13, pp. 3805-3813 (2012). 341. Q. A. Besford, M. A. Sullivan, L. Zheng, R. G. Gilbert, D. Stapleton, A. Gray-Weale, The structure of cardiac glycogen in healthy mice. International journal of biological macromolecules 51, pp. 887-891 (2012). 342. P. Devos, P. Baudhuin, F. Van Hoof, H.-G. Hers, The alpha particulate liver glycogen. A morphometric approach to the kinetics of its synthesis and degradation. Biochemical Journal 209, pp. 159-165 (1983).

219

343. R. R. Cardell, E. L. Cardell, Heterogeneity of glycogen distribution in hepatocytes. Journal of electron microscopy technique 14, pp. 126-139 (1990). 344. C. Bruni, K. R. Porter, The fine structure of the parenchymal cell of the normal rat liver: I. General observations. The American journal of pathology 46, pp. 691 (1965). 345. L. Napolitano, D. Fawcett, The fine structure of brown adipose tissue in the newborn mouse and rat. The Journal of cell biology 4, pp. 685-692 (1958). 346. K. Hashimoto, R. J. DiBella, Electron microscopic studies of normal and abnormal elastic fibers of the skin. Journal of Investigative Dermatology 48, pp. 405-423 (1967). 347. R. D. Eichner, Adipose tissue glycogen synthesis. International Journal of Biochemistry 16, pp. 257-261 (1984). 348. R. B. Scott, W. J. S. Still, Glycogen in human peripheral blood leukocytes: II. The macromolecular state of leukocyte glycogen. The Journal of clinical investigation 47, pp. 353-359 (1968). 349. E. M. Salmoral, D. S. Tolmasky, C. R. Krisman, Evidence for the presence of glycogen in rat thymus. Cellular and molecular biology 36, pp. 163-174 (1990). 350. A. Okubo, M. Sameshima, K. Unoki, F. Uehara, N. Ohba, Ultracytochemical demonstration of glycogen in cone, but not in rod, photoreceptor cells in the rat retina. Annals of Anatomy-Anatomischer Anzeiger 180, pp. 307-314 (1998). 351. C. H. Phelps, An ultrastructural study of methionine sulphoximine-induced glycogen accumulation in astrocytes of the mouse cerebral cortex. Journal of neurocytology 4, pp. 479-490 (1975). 352. C. Sotelo, S. L. Palay, The fine structure of the lateral vestibular nucleus in the rat: I. Neurons and neuroglial cells. The Journal of cell biology 36, pp. 151-179 (1968). 353. J.-F. Gadisseux, P. Evrard, Glial-neuronal relationship in the developing central nervous system. Developmental neuroscience 7, pp. 12-32 (1985). 354. J. Towfighi, B. S. Yoss, W. W. Wasiewski, R. C. Vannucci, M. S. Bentz, A. Mamourian, Cerebral glycogenosis, alpha particle type: morphologic and biochemical observations in an infant. Human pathology 20, pp. 1210- 1215 (1989). 355. A. Résibois-Grégoire, N. Dourov, Electron microscopic study of a case of cerebral glycogenosis. Acta neuropathologica 6, pp. 70-79 (1966). 356. M. Kornfeld, M. LeBaron, Glycogenosis type VIII. Journal of Neuropathology & Experimental Neurology 43, pp. 568-579 (1984). 357. H. J. Gertz, J. Cervos-Navarro, V. Frydl, F. Schultz, Glycogen accumulation of the aging human brain. Mechanisms of ageing and development 31, pp. 25-35 (1985). 358. D. M. A. Mann, P. Q. Sumpter, C. A. Davies, P. O. Yates, Glycogen accumulations in the cerebral cortex in Alzheimer's disease. Acta neuropathologica 73, pp. 181-184 (1987).

220

359. H. Kajihara, E. Tsutsumi, A. Kinoshita, J. Nakano, K. Takagi, S. Takeo, Activated astrocytes with glycogen accumulation in ischemic penumbra during the early stage of brain infarction: immunohistochemical and electron microscopic studies. Brain research 909, pp. 92-101 (2001). 360. P. Delorme, T. K. Hevor, Glycogen particles in methionine sulfoximine epileptogenic rodent brain and liver after the administration of methionine and actinomycin D. Neuropathology and applied neurobiology 11, pp. 117- 128 (1985). 361. T. Ishikawa, Y. F. Pei, Intramitochondrial glycogen particles in rat retinal receptor cells. The Journal of cell biology 25, pp. 402 (1965). 362. L. M. Buja, V. J. Ferrans, S. Levitsky, Occurrence of intramitochondrial glycogen in canine myocardium after prolonged anoxic cardiac arrest. Journal of molecular and cellular cardiology 4, pp. 237-242 (1972). 363. B. J. Maron, V. J. Ferrans, Intramitochondrial glycogen deposits in hypertrophied human myocardium. Journal of molecular and cellular cardiology 7, pp. 697-698 (1975). 364. M. Alaraj, B. Gajkowska, M. Cholewinski, J. W. Lazarewicz, Hyperglycaemia and intramitochondrial glycogen granules in the brain of mice. Ultrastructural study. Folia neuropathologica 42, pp. 113-118 (2004). 365. I. Marchand, K. Chorneyko, M. Tarnopolsky, S. Hamilton, J. Shearer, J. Potvin, T. E. Graham, Quantification of subcellular glycogen in resting human muscle: granule size, number, and location. Journal of Applied Physiology 93, pp. 1598-1607 (2002). 366. A. V. Skurat, S. S. Lim, P. J. Roach, Glycogen biogenesis in rat 1 fibroblasts expressing rabbit muscle glycogenin. European journal of biochemistry 245, pp. 147-155 (1997). 367. E. Zeqiraj, F. Sicheri, Getting a handle on glycogen synthase–Its interaction with glycogenin. Molecular aspects of medicine 46, pp. 63-69 (2015). 368. R. L. Terjung, K. M. Baldwin, W. W. Winder, J. O. Holloszy, Glycogen repletion in different types of muscle and in liver after exhausting exercise. American Journal of Physiology-Legacy Content 226, pp. 1387-1391 (1974). 369. L. H. Nilsson, E. Hultman, Liver Glycogen in Man–-the Effect of Total Starvation or a Carbohydrate-Poor Diet Followed by Carbohydrate Refeeding. Scandinavian journal of clinical and laboratory investigation 32, pp. 325-330 (1973). 370. T. Matsui, T. Ishikawa, H. Ito, M. Okamoto, K. Inoue, M. c. Lee, T. Fujikawa, Y. Ichitani, K. Kawanaka, H. Soya, Brain glycogen supercompensation following exhaustive exercise. The Journal of physiology 590, pp. 607-616 (2012). 371. W. J. Roesler, R. L. Khandelwal, Diurnal variations in the activities of the glycogen metabolizing enzymes in mouse liver. The International journal of biochemistry 17, pp. 81-85 (1985).

221

372. F. Halberg, P. G. Albrecht, C. P. Barnum Jr, Phase shifting of liver- glycogen rhythm in intact mice. American Journal of Physiology-Legacy Content 199, pp. 400-402 (1960). 373. M. A. Sullivan, S. T. Aroney, S. Li, F. J. Warren, J. S. Joo, K. S. Mak, D. I. Stapleton, K. S. Bell-Anderson, R. G. Gilbert, Changes in glycogen structure over feeding cycle sheds new light on blood-glucose control. Biomacromolecules 15, pp. 660-665 (2014). 374. C. Chandramouli, U. Varma, E. M. Stevens, R. P. Xiao, D. I. Stapleton, K. M. Mellor, L. M. Delbridge, Myocardial glycogen dynamics: new perspectives on disease mechanisms. Clin Exp Pharmacol Physiol 42, pp. 415-425 (2015). 375. Z. Hu, B. Deng, X. Tan, H. Gan, C. Li, S. S. Nada, M. A. Sullivan, J. Li, X. Jiang, E. Li, Diurnal changes of glycogen molecular structure in healthy and diabetic mice. Carbohydrate polymers 185, pp. 145-152 (2018). 376. B. Deng, M. A. Sullivan, J. Li, X. Tan, C. Zhu, B. L. Schulz, R. G. Gilbert, Molecular structure of glycogen in diabetic liver. Glycoconjugate journal 32, pp. 113-118 (2015). 377. S. A. Orrell, E. Bueding, A comparison of products obtained by various procedures used for the extraction of glycogen. Journal of Biological Chemistry 239, pp. 4021-4026 (1964). 378. M. A. Sullivan, F. Vilaplana, R. A. Cave, D. Stapleton, A. A. Gray-Weale, R. G. Gilbert, Nature of α and β particles in glycogen using molecular size distributions. Biomacromolecules 11, pp. 1094-1100 (2010). 379. X. Tan, M. A. Sullivan, S. S. Nada, B. Deng, B. L. Schulz, R. G. Gilbert, Proteomic Investigation of the Binding Agent between Liver Glycogen β Particles. Acs Omega 3, pp. 3640-3645 (2018). 380. B. Deng, M. A. Sullivan, C. Chen, J. Li, P. O. Powell, Z. Hu, R. G. Gilbert, Molecular structure of human-liver glycogen. PloS one 11, pp. e0150540 (2016). 381. D. Stapleton, C. Nelson, K. Parsawar, D. McClain, R. Gilbert-Wilson, E. Barker, B. Rudd, K. Brown, W. Hendrix, P. O'Donnell, G. Parker, Analysis of hepatic glycogen-associated proteins. Proteomics 10, pp. 2320-2329 (2010). 382. K. K. Rybicka, Glycosomes--the organelles of glycogen metabolism. Tissue & cell 28, pp. 253-265 (1996). 383. J. Lomako, W. M. Lomako, W. J. Whelan, R. S. Dombro, J. T. Neary, M. D. Norenberg, Glycogen synthesis in the astrocyte: from glycogenin to proglycogen to glycogen. The FASEB journal 7, pp. 1386-1393 (1993). 384. A. J. M. K. Van Heijningen, A. Kemp, Free and fixed glycogen in rat muscle. Biochemical Journal 59, pp. 487 (1955). 385. J. A. Russell, W. L. Bloom, Extractable and residual glycogen in tissues of the rat. American Journal of Physiology-Legacy Content 183, pp. 345-355 (1955). 386. D. Stetten Jr, M. R. Stetten, Glycogen metabolism. Physiological reviews 40, pp. 505-537 (1960).

222

387. W. J. Whelan, On the origin of primer for glycogen synthesis. Trends in Biochemical Sciences 1, pp. 13-15 (1976). 388. J. A. Curtino, E. R. Lacoste, Two glycogen synthase activities associated with proteoglycogen in retina. Neurochemical research 25, pp. 129-132 (2000). 389. C. Prats, J. A. Cadefau, R. Cussó, K. Qvortrup, J. N. Nielsen, J. F. P. Wojtaszewki, D. G. Hardie, G. Stewart, B. F. Hansen, T. Ploug, Phosphorylation-dependent translocation of glycogen synthase to a novel structure during glycogen re-synthesis. Journal of Biological Chemistry, (2005). 390. C. Prats, J. W. Helge, P. Nordby, K. Qvortrup, T. Ploug, F. Dela, J. F. P. Wojtaszewski, Dual regulation of muscle glycogen synthase during exercise by activation and compartmentalization. Journal of Biological Chemistry 284, pp. 15692-15700 (2009). 391. R. Geddes, G. C. Stratton, Molecular and metabolic heterogeneity of liver glycogen. Carbohydrate research 57, pp. 291-299 (1977). 392. S. Baqué, J. J. Guinovart, J. C. Ferrer, Glycogenin, the primer of glycogen synthesis, binds to actin. FEBS letters 417, pp. 355-359 (1997). 393. J. Pitcher, C. Smythe, D. G. Campbell, P. Cohen, Identification of the 38‐ kDa subunit of rabbit skeletal muscle glycogen synthase as glycogenin. European journal of biochemistry 169, pp. 497-502 (1987). 394. J. N. Nielsen, W. Derave, S. Kristiansen, E. Ralston, T. Ploug, E. A. Richter, Glycogen synthase localization and activity in rat skeletal muscle is strongly dependent on glycogen content. The Journal of Physiology 531, pp. 757-769 (2001). 395. M. García-Rocha, R. Angela, N. de la Iglesia, B. Otto, J. M. Fernández- Novell, J. C. Ferrer, J. J. Guinovart, Intracellular distribution of glycogen synthase and glycogen in primary cultured rat hepatocytes. Biochemical Journal 357, pp. 17-24 (2001). 396. J. M. Fernández-Novell, C. López-Iglesias, J. C. Ferrer, J. J. Guinovart, Zonal distribution of glycogen synthesis in isolated rat hepatocytes. FEBS letters 531, pp. 222-228 (2002). 397. J. C. Ferrer, C. Favre, R. R. Gomis, J. M. Fernández-Novell, M. Garcıá - Rocha, N. de la Iglesia, E. Cid, J. J. Guinovart, Control of glycogen deposition. FEBS letters 546, pp. 127-132 (2003). 398. H. D. Chipps, G. L. Duff, Glycogen infiltration of the liver cell nuclei. The American journal of pathology 18, pp. 645 (1942). 399. H. Sheldon, M. Silverberg, I. Kerner, On the differing appearance of intranuclear and cytoplasmic glycogen in liver cells in glycogen storage disease. The Journal of cell biology 13, pp. 468 (1962). 400. K. Tanikawa, Electron microscopic observation of the intranuclear glycogen of the liver in diabetes mellitus. The Kurume medical journal 11, pp. 177-181 (1965). 401. V. J. Ferrans, B. J. Maron, L. M. Buja, N. Ali, W. C. Roberts, Intranuclear glycogen deposits in human cardiac muscle cells: ultrastructure and

223

cytochemistry. Journal of molecular and cellular cardiology 7, pp. 381-386 (1975). 402. T. Amemiya, INTRANUCLEAR POLYSACCHARIDE SYNTHESIS IN MÜLLER'S CELLS OF THE NORMAL RAT RETINA. Acta histochemica et cytochemica 3, pp. 41-51 (1970). 403. S. Karasaki, Cytoplasmic and nuclear glycogen synthesis in Novikoff ascites hepatoma cells. Journal of ultrastructure research 35, pp. 181-196 (1971). 404. M. Ohyumi, S. Takano, Intranuclear synthesized and native glycogen particles in human gastric cancer: ultrastructure and histochemistry. Histochemistry 50, pp. 239-250 (1977). 405. M. Kopun, C. Granzow, C. R. Krisman, Comparative study of nuclear and cytoplasmic glycogen isolated from mutant HD33 ascites cells. Journal of cellular biochemistry 39, pp. 185-195 (1989). 406. M. F. Binggeli, Abnormal intranuclear and cytoplasmic formations associated with a chemically induced, transplantable chicken sarcoma. The Journal of biophysical and biochemical cytology 5, pp. 143 (1959). 407. E. H. Leduc, J. W. Wilson, An electron microscope study of intranuclear inclusions in mouse liver and hepatoma. The Journal of cell biology 6, pp. 427-430 (1959). 408. F. N. Ghandially, E. W. Parry, Ultrastructure of a human hepatocellular carcinoma and surrounding non‐neoplastic liver. Cancer 19, pp. 1989- 2004 (1966). 409. C. Granzow, M. Kopun, H.-P. Zimmermann, Role of nuclear glycogen synthase and cytoplasmic UDP glucose pyrophosphorylase in the biosynthesis of nuclear glycogen in HD33 Ehrlich-Lettré ascites tumor cells. The Journal of cell biology 89, pp. 475-484 (1981). 410. M. Kopun, H. Spring, C. Granzow, Nuclear glycogen synthase—fact or artifact? FEBS letters 147, pp. 207-210 (1982). 411. T. Miyakawa, R. Kuramoto, A. Shimoji, Y. Higuchi, Fine structure of inclusion body in the nucleus of Alzheimer glia type II in the brain of hepatocerebral degeneration. Acta neuropathologica 56, pp. 315-319 (1982). 412. N. Horita, M. Matsushita, T. Ishii, S. Oyanagi, K. Sakamoto, Ultrastructure of Alzheimer type II glia in hepatocerebral disease. Neuropathology and applied neurobiology 7, pp. 97-102 (1981). 413. M. Lafarga, M. T. Berciano, J. M. Pérez‐Fígares, M. A. Andrés, E. Maquiera, Influence of age on nuclear bodies and nuclear volume in pituicytes of the rat neurohypophysis. The Anatomical Record 230, pp. 319-324 (1991). 414. M. C. Miozzo, C. Maldonado, J. A. Curtino, Cellular and subcellular localization of glycogenin in chicken retina. IUBMB Life 40, pp. 173-180 (1996). 415. J. C. Ferrer, S. Baqué, J. J. Guinovart, Muscle glycogen synthase translocates from the cell nucleus to the cytosol in response to glucose. FEBS letters 415, pp. 249-252 (1997).

224

416. E. Cid, D. Cifuentes, S. Baqué, J. C. Ferrer, J. J. Guinovart, Determinants of the nucleocytoplasmic shuttling of muscle glycogen synthase. The FEBS journal 272, pp. 3197-3213 (2005). 417. A. Díaz, C. Martínez-Pons, I. Fita, J. C. Ferrer, J. J. Guinovart, Processivity and subcellular localization of glycogen synthase depend on a non-catalytic high-affinity glycogen-binding site. Journal of Biological Chemistry, pp. jbc. M111. 236109 (2011). 418. J. M. Fernández-Novell, S. Castel, D. Bellido, J. C. Ferrer, S. Vilaró, J. J. Guinovart, Intracellular distribution of hepatic glucokinase and glucokinase regulatory protein during the fasted to refed transition in rats. FEBS letters 459, pp. 211-214 (1999). 419. A. J. Yáñez, M. Garcia-Rocha, R. Bertinat, C. Droppelmann, I. I. Concha, J. J. Guinovart, J. C. Slebe, Subcellular localization of liver FBPase is modulated by metabolic conditions. FEBS letters 577, pp. 154-158 (2004). 420. M. J. Jurczak, A. M. Danos, V. R. Rehrmann, M. J. Brady, The role of protein translocation in the regulation of glycogen metabolism. Journal of cellular biochemistry 104, pp. 435-443 (2008). 421. P. K. Singh, S. Singh, S. Ganesh, The laforin-malin complex negatively regulates glycogen synthesis by modulating cellular glucose uptake via glucose transporters. Molecular and cellular biology 32, pp. 652-663 (2012). 422. W. A. Wilson, P. J. Roach, M. Montero, E. Baroja-Fernández, F. J. Muñoz, G. Eydallin, A. M. Viale, J. Pozueta-Romero, Regulation of glycogen metabolism in yeast and bacteria. FEMS microbiology reviews 34, pp. 952-985 (2010). 423. W. Hendriks, J. W. Mulders, M. A. Bibby, C. Slingsby, H. Bloemendal, W. W. De Jong, Duck lens epsilon-crystallin and lactate dehydrogenase B4 are identical: a single-copy gene product with two distinct functions. Proceedings of the National Academy of Sciences 85, pp. 7114-7118 (1988). 424. J.-w. Kim, C. V. Dang, Multifaceted roles of glycolytic enzymes. Trends in biochemical sciences 30, pp. 142-150 (2005). 425. A. E. Boukouris, S. D. Zervopoulos, E. D. Michelakis, Metabolic enzymes moonlighting in the nucleus: metabolic regulation of gene transcription. Trends in biochemical sciences 41, pp. 712-730 (2016). 426. A. McBride, S. Ghilagaber, A. Nikolaev, D. G. Hardie, The glycogen- binding domain on the AMPK β subunit allows the kinase to act as a glycogen sensor. Cell metabolism 9, pp. 23-34 (2009). 427. T. Leff. (Portland Press Limited, 2003). 428. K. A. Lamia, U. M. Sachdeva, L. DiTacchio, E. C. Williams, J. G. Alvarez, D. F. Egan, D. S. Vasquez, H. Juguilon, S. Panda, R. J. Shaw, AMPK regulates the circadian clock by cryptochrome phosphorylation and degradation. Science (New York, N.Y.) 326, pp. 437-440 (2009). 429. P. A. M. Michels, F. Bringaud, M. Herman, V. Hannaert, Metabolic functions of glycosomes in trypanosomatids. Biochimica et Biophysica Acta (BBA)-Molecular Cell Research 1763, pp. 1463-1477 (2006).

225

430. R. D. Klausner, J. G. Donaldson, J. Lippincott-Schwartz, Brefeldin A: insights into the control of membrane traffic and organelle structure. The Journal of cell biology 116, pp. 1071-1080 (1992). 431. I. Mellman, G. Warren, The road taken: past and future foundations of membrane traffic. Cell 100, pp. 99-112 (2000). 432. V. J. Ferrans, L. M. Buja, M. Jones, Ultrastructure and cytochemistry of glycogen in cardiac diseases. Recent advances in studies on cardiac structure and metabolism 3, pp. 97 (1973). 433. J. C.-H. Chang, S.-M. Wu, Y.-C. Tseng, Y.-C. Lee, O. Baba, P.-P. Hwang, Regulation of glycogen metabolism in gills and liver of the euryhaline tilapia (Oreochromis mossambicus) during acclimation to seawater. Journal of Experimental Biology 210, pp. 3494-3504 (2007). 434. D. Stapleton, C. Nelson, K. Parsawar, M. Flores-Opazo, D. McClain, G. Parker, The 3T3-L1 adipocyte glycogen proteome. Proteome science 11, pp. 11 (2013). 435. S. Nakamura-Tsuruta, M. Yasuda, T. Nakamura, E. Shinoda, T. Furuyashiki, R. Kakutani, H. Takata, Y. Kato, H. Ashida, Comparative analysis of carbohydrate-binding specificities of two anti-glycogen monoclonal antibodies using ELISA and surface plasmon resonance. Carbohydr Res 350, pp. 49-54 (2012). 436. N. A. Chebotareva, A. V. Meremyanin, V. F. Makeeva, T. B. Eronina, B. I. Kurganov, Glycogen phosphorylase b and phosphorylase kinase binding to glycogen under molecular crowding conditions. Inhibitory effect of FAD. Biochemistry (moscow) 74, pp. 562-568 (2009). 437. M. Bollen, W. Peti, M. J. Ragusa, M. Beullens, The extended PP1 toolkit: designed to create specificity. Trends in biochemical sciences 35, pp. 450- 458 (2010). 438. E. Heroes, B. Lesage, J. Görnemann, M. Beullens, L. Van Meervelt, M. Bollen, The PP1 binding code: a molecular‐lego strategy that governs specificity. The FEBS journal 280, pp. 584-595 (2013). 439. C. B. Newgard, M. J. Brady, R. M. O'Doherty, A. R. Saltiel, Organizing glucose disposal: emerging roles of the glycogen targeting subunits of protein phosphatase-1. Diabetes 49, pp. 1967-1977 (2000). 440. C. Christiansen, M. Abou Hachem, S. Janecek, A. Vikso-Nielsen, A. Blennow, B. Svensson, The carbohydrate-binding module family 20-- diversity, structure, and function. The FEBS journal 276, pp. 5006-5029 (2009). 441. C. C. Greenberg, A. M. Danos, M. J. Brady, Central role for protein targeting to glycogen in the maintenance of cellular glycogen stores in 3T3-L1 adipocytes. Molecular and cellular biology 26, pp. 334-342 (2006). 442. J. A. Printen, M. J. Brady, A. R. Saltiel, PTG, a protein phosphatase 1- binding protein with a role in glycogen metabolism. Science (New York, N.Y.) 275, pp. 1475-1478 (1997). 443. M. Montori-Grau, M. Guitart, C. García-Martínez, A. Orozco, A. M. Gómez-Foix, Differential pattern of glycogen accumulation after protein phosphatase 1 glycogen-targeting subunit PPP1R6 overexpression,

226

compared to PPP1R3C and PPP1R3A, in skeletal muscle cells. BMC biochemistry 12, pp. 57 (2011). 444. R. Gasa, P. B. Jensen, H. K. Berman, M. J. Brady, A. A. DePaoli-Roach, C. B. Newgard, Distinctive regulatory and metabolic properties of glycogen targeting subunits of protein phosphatase-1 (PTG, GL, GM/RGl) expressed in hepatocytes. Journal of Biological Chemistry, (2000). 445. D. G. Hardie, AMPK—sensing energy while talking to other signaling pathways. Cell metabolism 20, pp. 939-952 (2014). 446. G. Polekhina, A. Gupta, B. J. Michell, B. van Denderen, S. Murthy, S. C. Feil, I. G. Jennings, D. J. Campbell, L. A. Witters, M. W. Parker, AMPK β subunit targets metabolic stress sensing to glycogen. Current biology 13, pp. 867-871 (2003). 447. L. Bultot, B. Guigas, A. Von Wilamowitz-Moellendorff, L. Maisin, D. Vertommen, N. Hussain, M. Beullens, J. J. Guinovart, M. Foretz, B. Viollet, K. Sakamoto, L. Hue, M. H. Rider, AMP-activated protein kinase phosphorylates and inactivates liver glycogen synthase. The Biochemical journal 443, pp. 193-203 (2012). 448. D. G. Hardie, K. J. P. Sakamoto, AMPK: a key sensor of fuel and energy status in skeletal muscle. 21, pp. 48-60 (2006). 449. A. Koay, B. Woodcroft, E. J. Petrie, H. Yue, S. Emanuelle, M. Bieri, M. F. Bailey, M. Hargreaves, J. T. Park, K. H. Park, S. Ralph, D. Neumann, D. Stapleton, P. R. Gooley, AMPK beta subunits display isoform specific affinities for carbohydrates. FEBS letters 584, pp. 3499-3503 (2010). 450. G. J. Parker, A. Koay, R. Gilbert-Wilson, L. J. Waddington, D. Stapleton, AMP-activated protein kinase does not associate with glycogen α-particles from rat liver. Biochemical and biophysical research communications 362, pp. 811-815 (2007). 451. C. Vénien-Bryan, S. Jonic, V. Skamnaki, N. Brown, N. Bischler, N. G. Oikonomakos, N. Boisset, L. N. Johnson, The structure of phosphorylase kinase holoenzyme at 9.9 Å resolution and location of the catalytic subunit and the substrate glycogen phosphorylase. Structure 17, pp. 117-127 (2009). 452. M. Berard-Badier, J. F. Pellissier, D. Gambarelli, T. de Barsy, J. Roger, M. Toga, The retina in Lafora disease: light and electron microscopy. Albrecht Von Graefes Arch Klin Exp Ophthalmol 212, pp. 285-294 (1980). 453. P. Sanchez-Martin, M. Raththagala, T. M. Bridges, S. Husodo, M. S. Gentry, P. Sanz, C. Roma-Mateo, Dimerization of the glucan phosphatase laforin requires the participation of cysteine 329. PloS one 8, pp. e69523 (2013). 454. R. S. Sankhala, A. C. Koksal, L. Ho, F. Nitschke, B. A. Minassian, G. Cingolani, Dimeric quaternary structure of human laforin. The Journal of biological chemistry, (2014). 455. M. C. Solaz-Fuster, J. V. Gimeno-Alcaniz, S. Ros, M. E. Fernandez- Sanchez, B. Garcia-Fojeda, O. Criado Garcia, D. Vilchez, J. Dominguez, M. Garcia-Rocha, M. Sanchez-Piris, C. Aguado, E. Knecht, J. Serratosa, J. J. Guinovart, P. Sanz, S. Rodriguez de Cordoba, Regulation of

227

glycogen synthesis by the laforin-malin complex is modulated by the AMP- activated protein kinase pathway. Human molecular genetics 17, pp. 667- 678 (2008). 456. M. A. Garcia-Gimeno, P. N. Rodilla-Ramirez, R. Viana, X. Salas-Puig, M. K. Brewer, M. S. Gentry, P. Sanz, A novel EPM2A mutation yields a slow progression form of Lafora disease. Epilepsy research 145, pp. 169-177 (2018). 457. R. Guerrero, S. Vernia, R. Sanz, I. Abreu-Rodríguez, C. Almaraz, M. García-Hoyos, R. Michelucci, C. A. Tassinari, P. Riguzzi, C. Nobile, A PTG variant contributes to a milder phenotype in Lafora disease. PloS one 6, pp. e21294 (2011). 458. A. R. Sherwood, B. C. Paasch, C. A. Worby, M. S. Gentry, A malachite green-based assay to assess glucan phosphatase activity. Analytical biochemistry 435, pp. 54-56 (2013). 459. S. Hsu, Y. Kim, S. Li, E. S. Durrant, R. M. Pace, V. L. Woods, Jr., M. S. Gentry, Structural insights into glucan phosphatase dynamics using amide hydrogen-deuterium exchange mass spectrometry. Biochemistry 48, pp. 9891-9902 (2009). 460. A. R. Sherwood, Investigating Therapeutic Options for Lafora Disease Using Structural Biology and Translational Methods. (2013). 461. P. Schuck, M. A. Perugini, N. R. Gonzales, G. J. Howlett, D. J. B. j. Schubert, Size-distribution analysis of proteins by analytical ultracentrifugation: strategies and application to model systems. 82, pp. 1096-1111 (2002). 462. L. J. V. Schrodinger, The PyMOL molecular graphics system. 1, pp. 0 (2010). 463. R. Fraczkiewicz, W. Braun, Exact and efficient analytical calculation of the accessible surface areas and their gradients for macromolecules. Journal of computational chemistry 19, pp. 319-333 (1998). 464. E. Krissinel, K. Henrick, Inference of macromolecular assemblies from crystalline state. Journal of molecular biology 372, pp. 774-797 (2007). 465. L. A. Kelley, S. Mezulis, C. M. Yates, M. N. Wass, M. J. J. N. p. Sternberg, The Phyre2 web portal for protein modeling, prediction and analysis. 10, pp. 845 (2015). 466. R. A. Laskowski, M. B. Swindells. (ACS Publications, 2011). 467. L. Holm, C. Sander, Dali: a network tool for protein structure comparison. Trends Biochem Sci 20, pp. 478-480 (1995). 468. N. A. Baker, D. Sept, S. Joseph, M. J. Holst, J. A. J. P. o. t. N. A. o. S. McCammon, Electrostatics of nanosystems: application to microtubules and the ribosome. 98, pp. 10037-10041 (2001). 469. A. A. DePaoli-Roach, D. M. Segvich, C. M. Meyer, Y. Rahimi, C. A. Worby, M. S. Gentry, P. J. Roach, Laforin and malin knockout mice have normal glucose disposal and insulin sensitivity. Human molecular genetics 21, pp. 1604-1610 (2012). 470. V. Jager, K. Bussow, A. Wagner, S. Weber, M. Hust, A. Frenzel, T. Schirrmann, High level transient production of recombinant antibodies and

228

antibody fusion proteins in HEK293 cells. BMC Biotechnol 13, pp. 52 (2013). 471. Y. Oe, O. Baba, H. Ashida, K. C. Nakamura, Glycogen distribution in the microwave-fixed mouse brain reveals heterogeneous astrocytic patterns. 64, pp. 1532-1545 (2016). 472. F. L. Graham, J. Smiley, W. C. Russell, R. Nairn, Characteristics of a human cell line transformed by DNA from human adenovirus type 5. J Gen Virol 36, pp. 59-74 (1977). 473. M. K. Morell, M. S. Samuel, M. G. O'Shea, Analysis of starch structure using fluorophore‐assisted carbohydrate electrophoresis. Electrophoresis 19, pp. 2603-2611 (1998). 474. J. Verbeke, C. Penverne, C. D’Hulst, C. Rolando, N. Szydlowski, Rapid and sensitive quantification of C3-and C6-phosphoesters in starch by fluorescence-assisted capillary electrophoresis. Carbohydrate polymers 152, pp. 784-791 (2016). 475. A. Goldstein, G. Annor, J.-L. Putaux, K. H. Hebelstrup, A. Blennow, E. Bertoft, Impact of full range of amylose contents on the architecture of starch granules. International journal of biological macromolecules 89, pp. 305-318 (2016). 476. L. Ianzano, J. Zhang, E. M. Chan, X. Zhao, H. Lohi, S. W. Scherer, B. A. Minassian, Lafora progressive myoclonus epilepsy mutation database- EPM2A and NHLRC1 (EMP2B) genes. Human mutation 26, pp. 397 (2005). 477. S. Singh, S. Ganesh, Lafora progressive myoclonus epilepsy: a meta- analysis of reported mutations in the first decade following the discovery of the EPM2A and NHLRC1 genes. Human mutation 30, pp. 715-723 (2009). 478. F. Ibrahim, N. Murr, Lafora Disease. (2018). 479. G. Lesca, N. Boutry-Kryza, B. de Toffol, M. Milh, D. Steschenko, M. Lemesle-Martin, L. Maillard, G. Foletti, G. Rudolf, J. E. Nielsen, B. a Rogvi-Hansen, J. Erdal, J. Mancini, C. Thauvin-Robinet, A. M'Rrabet, D. Ville, P. Szepetowski, E. Raffo, E. Hirsch, P. Ryvlin, A. Calender, P. Genton, Novel mutations in EPM2A and NHLRC1 widen the spectrum of Lafora disease. Epilepsia 51, pp. 1691-1698 (2010). 480. Z. Aslam, E. Lee, M. Badshah, M. Naeem, C. Kang, Whole exome sequencing identified a novel missense mutation in EPM2A underlying Lafora disease in a Pakistani family. Seizure-European Journal of Epilepsy 51, pp. 200-203 (2017). 481. H. Lohi, J. Turnbull, X. C. Zhao, S. Pullenayegum, L. Ianzano, M. Yahyaoui, M. A. Mikati, S. Franceschetti, F. Zara, B. A. Minassian, Genetic diagnosis in Lafora disease: Genotype-phenotype correlations and diagnostic pitfalls. Neurology 68, pp. 996-1001 (2007). 482. H. G. Poyrazoğlu, E. Karaca, H. Per, H. Gümüs, H. Onay, M. Canpolat, Ö. Canöz, F. Ozkınay, S. Kumandas, Three patients with lafora disease: Different clinical presentations and a novel mutation. Journal of child neurology 30, pp. 777-781 (2015).

229

483. H. M. Lanoiselee, P. Genton, G. Lesca, F. Brault, B. De Toffol, Are c.436G>A mutations less severe forms of Lafora disease? A case report. Epilepsy Behav Case Rep 2, pp. 19-21 (2014). 484. E. Ferlazzo, L. Canafoglia, R. Michelucci, A. Gambardella, E. Gennaro, E. Pasini, P. Riguzzi, R. Plasmati, L. Volpi, A. Labate, S. Gasparini, F. Villani, M. Casazza, M. Viri, F. Zara, B. A. Minassian, J. Turnbull, J. M. Serratosa, R. Guerrero-Lopez, S. Franceschetti, U. Aguglia, Mild Lafora disease: clinical, neurophysiologic, and genetic findings. Epilepsia 55, pp. e129-133 (2014). 485. S. Gökdemir, H. Çağlayan, M. Kızıltan, N. Karaağaç, C. Leblebici, S. N. Yeni, Presentation of an unusual patient with Lafora disease. Epileptic Disorders 14, pp. 94-98 (2012). 486. P. Couarch, S. Vernia, I. Gourfinkel-An, G. Lesca, S. Gataullina, E. Fedirko, O. Trouillard, C. Depienne, O. Dulac, D. Steschenko, E. Leguern, P. Sanz, S. Baulac, Lafora progressive myoclonus epilepsy: NHLRC1 mutations affect glycogen metabolism. J Mol Med (Berl) 89, pp. 915-925 (2011). 487. B. Baykan, P. Striano, S. Gianotti, N. Bebek, E. Gennaro, C. Gurses, F. Zara, Late-onset and slow-progressing Lafora disease in four siblings with EPM2B mutation. Epilepsia 46, pp. 1695-1697 (2005). 488. A. Jara-Prado, A. Ochoa, M. E. Alonso, G. A. Lima Villeda, F. Fernandez- Valverde, L. Ruano-Calderon, S. Vargas-Canas, R. M. Duron, A. V. Delgado-Escueta, I. E. Martinez-Juarez, Late onset Lafora disease and novel EPM2A mutations: breaking paradigms. Epilepsy research 108, pp. 1501-1510 (2014). 489. D. S. Lynch, N. W. Wood, H. Houlden, Late-onset Lafora disease with prominent parkinsonism due to a rare mutation in EPM2A. Neurol Genet 2, pp. e101 (2016). 490. B. Rabbani, M. Tekin, N. Mahdieh, The promise of whole-exome sequencing in medical genetics. Journal of human genetics 59, pp. 5 (2014). 491. B. M. Kroncke, C. G. Vanoye, J. Meiler, A. L. George, Jr., C. R. Sanders, Personalized biochemistry and biophysics. Biochemistry 54, pp. 2551- 2559 (2015). 492. S. Stefl, H. Nishi, M. Petukh, A. R. Panchenko, E. Alexov, Molecular mechanisms of disease-causing missense mutations. Journal of molecular biology 425, pp. 3919-3936 (2013). 493. T. A. Peterson, E. Doughty, M. G. Kann, Towards precision medicine: advances in computational approaches for the analysis of human variants. Journal of molecular biology 425, pp. 4047-4063 (2013). 494. S. Hicks, D. A. Wheeler, S. E. Plon, M. Kimmel, Prediction of missense mutation functionality depends on both the algorithm and sequence alignment employed. Human mutation 32, pp. 661-668 (2011). 495. V. Potapov, M. Cohen, G. Schreiber, Assessing computational methods for predicting protein stability upon mutation: good on average but not in

230

the details. Protein engineering, design & selection 22, pp. 553-560 (2009). 496. J. Thusberg, A. Olatubosun, M. Vihinen, Performance of mutation pathogenicity prediction methods on missense variants. Human mutation 32, pp. 358-368 (2011). 497. M. A. Hamburg, F. S. Collins, The path to personalized medicine. New England Journal of Medicine 363, pp. 301-304 (2010). 498. J. Zielenski, Genotype and phenotype in cystic fibrosis. Respiration 67, pp. 117-133 (2000). 499. E. Kerem, B. Kerem, Genotype‐phenotype correlations in cystic fibrosis. Pediatric pulmonology 22, pp. 387-395 (1996). 500. M. D. Amaral, Novel personalized therapies for cystic fibrosis: treating the basic defect in all patients. Journal of internal medicine 277, pp. 155-166 (2015). 501. D. L. Masica, P. R. Sosnay, G. R. Cutting, R. Karchin, Phenotype‐ optimized sequence ensembles substantially improve prediction of disease‐causing mutation in cystic fibrosis. Human mutation 33, pp. 1267- 1274 (2012). 502. G. R. Cutting, Cystic fibrosis genetics: from molecular understanding to clinical application. Nature Reviews Genetics 16, pp. 45 (2015). 503. I. A. Adzhubei, S. Schmidt, L. Peshkin, V. E. Ramensky, A. Gerasimova, P. Bork, A. S. Kondrashov, S. R. Sunyaev, A method and server for predicting damaging missense mutations. Nature methods 7, pp. 248 (2010). 504. A. P. Pandurangan, B. Ochoa-Montaño, D. B. Ascher, T. L. Blundell, SDM: a server for predicting effects of mutations on protein stability. Nucleic acids research 45, pp. W229-W235 (2017). 505. V. Parthiban, M. M. Gromiha, D. Schomburg, CUPSAT: prediction of protein stability upon point mutations. Nucleic acids research 34, pp. W239-W242 (2006). 506. K. A. Uniewicz, A. Ori, R. Xu, Y. Ahmed, M. C. Wilkinson, D. G. Fernig, E. A. Yates, Differential scanning fluorimetry measurement of protein stability changes upon binding to glycosaminoglycans: a screening test for binding specificity. Analytical chemistry 82, pp. 3796-3802 (2010). 507. F. H. Niesen, H. Berglund, M. Vedadi, The use of differential scanning fluorimetry to detect ligand interactions that promote protein stability. Nature protocols 2, pp. 2212 (2007). 508. W. Wang, P. J. Roach, Glycogen and related polysaccharides inhibit the laforin dual-specificity protein phosphatase. Biochemical and biophysical research communications 325, pp. 726-730 (2004). 509. S. Ganesh, N. Tsurutani, T. Suzuki, Y. Hoshii, T. Ishihara, A. V. Delgado- Escueta, K. Yamakawa, The carbohydrate-binding domain of Lafora disease protein targets Lafora polyglucosan bodies. Biochemical and biophysical research communications 313, pp. 1101-1109 (2004).

231

510. S. Sharma, C. D. Vander Kooi, M. S. Gentry, C. W. Vander Kooi, Oligomerization and carbohydrate binding of glucan phosphatases. Analytical biochemistry 563, pp. 51-55 (2018). 511. T. A. Wright, J. M. Stewart, R. C. Page, D. Konkolewicz, Extraction of thermodynamic parameters of protein unfolding using parallelized differential scanning Fluorimetry. The journal of physical chemistry letters 8, pp. 553-558 (2017). 512. S. Boivin, S. Kozak, R. J. P. e. Meijers, purification, Optimization of protein purification and characterization using Thermofluor screens. 91, pp. 192- 206 (2013). 513. S. Ganesh, A. V. Delgado-Escueta, T. Suzuki, S. Francheschetti, C. Riggio, G. Avanzini, A. Rabinowicz, S. Bohlega, J. Bailey, M. E. Alonso, A. Rasmussen, A. E. Thomson, A. Ochoa, A. J. Prado, M. T. Medina, K. Yamakawa, Genotype-phenotype correlations for EPM2A mutations in Lafora's progressive myoclonus epilepsy: exon 1 mutations associate with an early-onset cognitive deficit subphenotype. Human molecular genetics 11, pp. 1263-1271 (2002). 514. M. E. Fernandez-Sanchez, O. Criado-Garcia, K. E. Heath, B. Garcia- Fojeda, I. Medrano-Fernandez, P. Gomez-Garre, P. Sanz, J. M. Serratosa, S. Rodriguez de Cordoba, Laforin, the dual-phosphatase responsible for Lafora disease, interacts with R5 (PTG), a regulatory subunit of protein phosphatase-1 that enhances glycogen accumulation. Human molecular genetics 12, pp. 3161-3171 (2003). 515. S. Vernia, M. C. Solaz-Fuster, J. V. Gimeno-Alcaniz, T. Rubio, L. Garcia- Haro, M. Foretz, S. Rodriguez de Cordoba, P. Sanz, AMP-activated protein kinase phosphorylates R5/PTG, the glycogen targeting subunt of the R5/PTG-PP1 holoenzyme and accelerates its downregulation by the laforin-malin complex. The Journal of biological chemistry, (2009). 516. M. C. Good, A. E. Greenstein, T. A. Young, H.-L. Ng, T. Alber, Sensor domain of the Mycobacterium tuberculosis receptor Ser/Thr protein kinase, PknD, forms a highly symmetric β propeller. Journal of molecular biology 339, pp. 459-469 (2004). 517. M. Gabrielsen, L. Buetow, M. A. Nakasone, S. F. Ahmed, G. J. Sibbet, B. O. Smith, W. Zhang, S. S. Sidhu, D. T. Huang, A General Strategy for Discovery of Inhibitors and Activators of RING and U-box E3 Ligases with Ubiquitin Variants. Molecular cell 68, pp. 456-470.e410 (2017). 518. D. M. Dawidziak, J. G. Sanchez, J. M. Wagner, B. K. Ganser-Pornillos, O. Pornillos, Structure and catalytic activation of the TRIM23 RING E3 ubiquitin ligase. Proteins 85, pp. 1957-1961 (2017). 519. S. Mittal, M. Upadhyay, P. K. Singh, R. Parihar, S. Ganesh, Interdependence of laforin and malin proteins for their stability and functions could underlie the molecular basis of locus heterogeneity in Lafora disease. Journal of biosciences 40, pp. 863-871 (2015). 520. B. A. Minassian, L. Ianzano, M. Meloche, E. Andermann, G. A. Rouleau, A. V. Delgado-Escueta, S. W. Scherer, Mutation spectrum and predicted

232

function of laforin in Lafora's progressive myoclonus epilepsy. Neurology 55, pp. 341-346 (2000). 521. S. Franceschetti, A. Gambardella, L. Canafoglia, P. Striano, H. Lohi, E. Gennaro, L. Ianzano, P. Veggiotti, V. Sofia, R. J. E. Biondi, Clinical and genetic findings in 26 Italian patients with Lafora disease. 47, pp. 640-643 (2006). 522. S. Salar, N. Yeni, A. Gunduz, A. Guler, A. Gokcay, S. Velioglu, A. Gundogdu, S. Hande Caglayan, Four novel and two recurrent NHLRC1 (EPM2B) and EPM2A gene mutations leading to Lafora disease in six Turkish families. Epilepsy research 98, pp. 273-276 (2012). 523. S. Singh, P. Satishchandra, S. K. Shankar, S. Ganesh, Lafora disease in the Indian population: EPM2A and NHLRC1 gene mutations and their impact on subcellular localization of laforin and malin. Human mutation 29, pp. E1-12 (2008). 524. R. Viana, P. Lujan, P. Sanz, The laforin/malin E3-ubiquitin ligase complex ubiquitinates pyruvate kinase M1/M2. BMC biochemistry 16, pp. 24 (2015). 525. C. Gomez‐Abad, Z. Afawi, A. D. Korczyn, A. Misk, S. A. Shalev, R. Spiegel, T. Lerman‐Sagie, D. Lev, K. L. Kron, P. Gómez‐Garre, Founder effect with variable age at onset in Arab families with Lafora disease and EPM2A mutation. Epilepsia 48, pp. 1011-1014 (2007). 526. P. Gomez-Garre, E. Gutierrez-Delicado, C. Gomez-Abad, J. Morales- Corraliza, V. E. Villanueva, S. Rodriguez de Cordoba, J. Larrauri, M. Gutierrez, J. Berciano, J. M. Serratosa, Hepatic disease as the first manifestation of progressive myoclonus epilepsy of Lafora. Neurology 68, pp. 1369-1373 (2007). 527. J. Turnbull, S. Kumar, Z.-P. Ren, S. Muralitharan, T. Naranian, C. A. Ackerley, B. A. Minassian, Lafora progressive myoclonus epilepsy: disease course homogeneity in a genetic isolate. Journal of child neurology 23, pp. 240-242 (2008). 528. J. M. Green, J. Hewison, H. L. Bekker, L. D. Bryant, H. S. Cuckle, in NIHR Health Technology Assessment programme: Executive Summaries. (NIHR Journals Library, 2004). 529. P. Gomez-Garre, Y. Sanz, S. R. Rodriguez De Cordoba, J. M. Serratosa, Mutational spectrum of the EPM2A gene in progressive myoclonus epilepsy of Lafora: high degree of allelic heterogeneity and prevalence of deletions. Eur J Hum Genet 8, pp. 946-954 (2000). 530. L. Ianzano, E. J. Young, X. C. Zhao, E. M. Chan, M. T. Rodriguez, M. V. Torrado, S. W. Scherer, B. A. Minassian, Loss of function of the cytoplasmic isoform of the protein laforin (EPM2A) causes Lafora progressive myoclonus epilepsy. Human mutation 23, pp. 170-176 (2004). 531. C. S. Ki, S. Y. Kong, D. W. Seo, S. B. Hong, H. J. Kim, J. W. Kim, Two novel mutations in the EPM2A gene in a Korean patient with Lafora's progressive myoclonus epilepsy. Journal of human genetics 48, pp. 51-54 (2003).

233

532. N. K. Tonks, Protein tyrosine phosphatases: from genes, to function, to disease. Nature reviews. Molecular cell biology 7, pp. 833-846 (2006). 533. K. Kolmodin, J. Aqvist, The catalytic mechanism of protein tyrosine phosphatases revisited. FEBS letters 498, pp. 208-213 (2001). 534. J. M. Denu, J. E. Dixon, Protein tyrosine phosphatases: mechanisms of catalysis and regulation. Current Opinion in Chemical Biology 2, pp. 633 (1998). 535. J. M. Denu, J. A. Stuckey, M. A. Saper, J. E. Dixon, Form and Function in Protein Dephosphorylation. Cell 87, pp. 361 (1996). 536. S. K. Whittier, A. C. Hengge, J. P. Loria, Conformational motions regulate phosphoryl transfer in related protein tyrosine phosphatases. Science (New York, N.Y.) 341, pp. 899-903 (2013). 537. M. J. Begley, G. S. Taylor, M. A. Brock, P. Ghosh, V. L. Woods, J. E. Dixon, Molecular basis for substrate recognition by MTMR2, a myotubularin family phosphoinositide phosphatase. Proceedings of the National Academy of Sciences of the United States of America 103, pp. 927-932 (2006). 538. C. Blanchetot, M. Chagnon, N. Dube, M. Halle, M. L. Tremblay, Substrate- trapping techniques in the identification of cellular PTP targets. Methods 35, pp. 44-53 (2005). 539. D. Barford, A. K. Das, M.-P. Egloff, The structure and mechanism of protein phosphatases: insights into catalysis and regulation. Annual review of biophysics and biomolecular structure 27, pp. 133-164 (1998). 540. A. Alonso, A. Rojas, A. Godzik, T. Mustelin, The dual-specific protein tyrosine phosphatase family. S. Hohmann, Ed., Protein Phosphatases (Springer, Berlin, 2003), vol. 5, pp. 333-358. 541. M. Bouhaddou, M. R. Birtwistle, Dimerization-based control of cooperativity. Molecular bioSystems 10, pp. 1824-1832 (2014). 542. M. I. Stefan, N. Le Novere, Cooperative binding. PLoS computational biology 9, pp. e1003106 (2013). 543. G. K. Ackers, M. L. Doyle, D. Myers, M. A. Daugherty, Molecular code for cooperativity in hemoglobin. Science (New York, N.Y.) 255, pp. 54-63 (1992). 544. A. Neznansky, I. Blus-Kadosh, G. Yerushalmi, E. Banin, Y. J. T. F. J. Opatowsky, The Pseudomonas aeruginosa phosphate transport protein PstS plays a phosphate-independent role in biofilm formation. 28, pp. 5223-5233 (2014). 545. A. J. Hutchison, Oral phosphate binders. Kidney international 75, pp. 906- 914 (2009). 546. T. Mustelin, in Protein phosphatase protocols. (Springer, 2007), pp. 9-22. 547. Y. Shi, Serine/threonine phosphatases: mechanism through structure. Cell 139, pp. 468-484 (2009). 548. A. Seifried, J. Schultz, A. Gohla, Human HAD phosphatases: structure, mechanism, and roles in health and disease. The FEBS journal 280, pp. 549-571 (2013).

234

549. A. Alonso, J. Sasin, N. Bottini, I. Friedberg, I. Friedberg, A. Osterman, A. Godzik, T. Hunter, J. Dixon, T. Mustelin, Protein Tyrosine Phosphatases in the Human Genome. Cell 117, pp. 699 (2004). 550. G. Scapin, S. Patel, V. Patel, B. Kennedy, E. Asante‐Appiah, The structure of apo protein‐tyrosine phosphatase 1B C215S mutant: More than just an S→ O change. Protein Science 10, pp. 1596-1605 (2001). 551. Y. Romsicki, G. Scapin, V. Beaulieu-Audy, S. Patel, J. W. Becker, B. P. Kennedy, E. Asante-Appiah, Functional characterization and crystal structure of the C215D mutant of protein-tyrosine phosphatase-1B. Journal of Biological Chemistry 278, pp. 29009-29015 (2003). 552. X. Hu, C. E. Stebbins, Dynamics of the WPD loop of the Yersinia protein tyrosine phosphatase. Biophysical journal 91, pp. 948-956 (2006). 553. L. J. Juszczak, Z.-Y. Zhang, L. Wu, D. S. Gottfried, D. D. Eads, Rapid loop dynamics of Yersinia protein tyrosine phosphatases. Biochemistry 36, pp. 2227-2236 (1997). 554. M. S. Choy, Y. Li, L. E. S. F. Machado, M. B. A. Kunze, C. R. Connors, X. Wei, K. Lindorff-Larsen, R. Page, W. Peti, Conformational rigidity and protein dynamics at distinct timescales regulate PTP1B activity and allostery. Molecular cell 65, pp. 644-658. e645 (2017). 555. T. A. S. Brandão, S. J. Johnson, A. C. Hengge, The molecular details of WPD-loop movement differ in the protein-tyrosine phosphatases YopH and PTP1B. Archives of biochemistry and biophysics 525, pp. 53-59 (2012). 556. R. A. P. Pádua, Y. Sun, I. Marko, W. Pitsawong, J. B. Stiller, R. Otten, D. Kern, Mechanism of activating mutations and allosteric drug inhibition of the phosphatase SHP2. Nature communications 9, pp. 4507 (2018). 557. F. Wang, W. Li, M. R. Emmett, C. L. Hendrickson, A. G. Marshall, Y.-L. Zhang, L. Wu, Z.-Y. Zhang, Conformational and dynamic changes of Yersinia protein tyrosine phosphatase induced by ligand binding and active site mutation and revealed by H/D exchange and electrospray ionization Fourier transform ion cyclotron resonance mass spectrometry. Biochemistry 37, pp. 15289-15299 (1998). 558. B. Ciani, R. Layfield, J. R. Cavey, P. W. Sheppard, M. S. Searle, Structure of the ubiquitin-associated domain of p62 (SQSTM1) and implications for mutations that cause Paget's disease of bone. The Journal of biological chemistry 278, pp. 37409-37412 (2003). 559. S. Pankiv, T. H. Clausen, T. Lamark, A. Brech, J.-A. Bruun, H. Outzen, A. Øvervatn, G. Bjørkøy, T. Johansen, p62/SQSTM1 binds directly to Atg8/LC3 to facilitate degradation of ubiquitinated protein aggregates by autophagy. Journal of biological chemistry 282, pp. 24131-24145 (2007). 560. A. Shahwan, M. Farrell, N. Delanty, Progressive myoclonic epilepsies: a review of genetic and therapeutic aspects. Lancet Neurol 4, pp. 239-248 (2005). 561. A. M. Garcia-Cabrero, A. Marinas, R. Guerrero, S. R. de Cordoba, J. M. Serratosa, M. P. Sanchez, Laforin and malin deletions in mice produce

235

similar neurologic impairments. J Neuropathol Exp Neurol 71, pp. 413-421 (2012). 562. A. M. Garcia-Cabrero, G. Sanchez-Elexpuru, J. M. Serratosa, M. P. Sanchez, Enhanced sensitivity of laforin- and malin-deficient mice to the convulsant agent pentylenetetrazole. Front Neurosci 8, pp. 291 (2014). 563. J. Duran, M. F. Tevy, M. Garcia-Rocha, J. Calbo, M. Milan, J. J. Guinovart, Deleterious effects of neuronal accumulation of glycogen in flies and mice. EMBO molecular medicine 4, pp. 719-729 (2012). 564. M. K. Brewer, M. S. Gentry, The 3rd International Lafora Epilepsy Workshop: Evidence for a cure. Epilepsy Behav, (2018). 565. C. Klein, W. A. Gahl, Patients with rare diseases: from therapeutic orphans to pioneers of personalized treatments. (2017). 566. A. T. van der Ploeg, P. R. Clemens, D. Corzo, D. M. Escolar, J. Florence, G. J. Groeneveld, S. Herson, P. S. Kishnani, P. Laforet, S. L. Lake, D. J. Lange, R. T. Leshner, J. E. Mayhew, C. Morgan, K. Nozaki, D. J. Park, A. Pestronk, B. Rosenbloom, A. Skrinar, C. I. van Capelle, N. A. van der Beek, M. Wasserstein, S. A. Zivkovic, A randomized study of alglucosidase alfa in late-onset Pompe's disease. The New England journal of medicine 362, pp. 1396-1406 (2010). 567. P. S. Kishnani, D. Corzo, M. Nicolino, B. Byrne, H. Mandel, W. L. Hwu, N. Leslie, J. Levine, C. Spencer, M. McDonald, J. Li, J. Dumontier, M. Halberthal, Y. H. Chien, R. Hopkin, S. Vijayaraghavan, D. Gruskin, D. Bartholomew, A. van der Ploeg, J. P. Clancy, R. Parini, G. Morin, M. Beck, G. S. De la Gastine, M. Jokic, B. Thurberg, S. Richards, D. Bali, M. Davison, M. A. Worden, Y. T. Chen, J. E. Wraith, Recombinant human acid [alpha]-glucosidase: major clinical benefits in infantile-onset Pompe disease. Neurology 68, pp. 99-109 (2007). 568. B. L. Thurberg, C. Lynch Maloney, C. Vaccaro, K. Afonso, A. C. Tsai, E. Bossen, P. S. Kishnani, M. O'Callaghan, Characterization of pre- and post-treatment pathology after enzyme replacement therapy for Pompe disease. Lab Invest 86, pp. 1208-1220 (2006). 569. T. Ishihara, T. Yokota, Y. Yamashita, M. Takahashi, H. Kawano, F. Uchino, T. Kamei, N. Matsumoto, Y. Kusunose, M. Yamada, Comparative study of the intracytoplasmic inclusions in Lafora disease and type IV glycogenosis by electron microscopy. Acta pathologica japonica 37, pp. 1591-1601 (1987). 570. F. Van Hoof, M. Hageman-Bal, Progressive familial myoclonic epilepsy with Lafora bodies. Electron microscopic and histochemical study of a cerebral biopsy. Acta Neuropathol 7, pp. 315-336 (1967). 571. K. Rehman, M. S. Hamid Akash, B. Akhtar, M. Tariq, A. Mahmood, M. Ibrahim, Delivery of Therapeutic Proteins: Challenges and Strategies. Curr Drug Targets 17, pp. 1172-1188 (2016). 572. J. E. Hansen, W. Sohn, C. Kim, S. S. Chang, N. C. Huang, D. G. Santos, G. Chan, R. H. Weisbart, R. N. Nishimura, Antibody-mediated Hsp70 protein therapy. Brain Res 1088, pp. 187-196 (2006).

236

573. J. E. Hansen, R. H. Weisbart, R. N. Nishimura, Antibody mediated transduction of therapeutic proteins into living cells. ScientificWorldJournal 5, pp. 782-788 (2005). 574. R. H. Weisbart, M. Stempniak, S. Harris, D. J. Zack, K. Ferreri, An autoantibody is modified for use as a delivery system to target the cell nucleus: therapeutic implications. J Autoimmun 11, pp. 539-546 (1998). 575. M. W. Lawlor, D. Armstrong, M. G. Viola, J. J. Widrick, H. Meng, R. W. Grange, M. K. Childers, C. P. Hsu, M. O'Callaghan, C. R. Pierson, A. Buj- Bello, A. H. Beggs, Enzyme replacement therapy rescues weakness and improves muscle pathology in mice with X-linked myotubular myopathy. Human molecular genetics 22, pp. 1525-1538 (2013). 576. H. Yi, T. Sun, D. Armstrong, S. Borneman, C. Yang, S. Austin, P. S. Kishnani, B. Sun, Antibody-mediated enzyme replacement therapy targeting both lysosomal and cytoplasmic glycogen in Pompe disease. (2017). 577. L. L. C. Valerion Therapeutics. (Clinicaltrials.gov Identifier NCT02898753, 2016). 578. J. M. Girard, J. Turnbull, N. Ramachandran, B. A. Minassian, Progressive myoclonus epilepsy. Handbook of clinical neurology 113, pp. 1731-1736 (2013). 579. T. Nikaido, J. Austin, H. Stukenbrok, Studies in myoclonus epilepsy. 3. The effects of amylolytic enzymes on the ultrastructure of Lafora bodies. J Histochem Cytochem 19, pp. 382-385 (1971). 580. J. D. Kruse-Jarres, C. Kaiser, J. C. Hafkenscheid, W. Hohenwallner, W. Stein, J. Bohner, G. Klein, W. Poppe, E. Rauscher, Evaluation of a new alpha-amylase assay using 4.6-ethylidene-(G7)-1-4-nitrophenyl-(G1)- alpha-D-maltoheptaoside as substrate. J Clin Chem Clin Biochem 27, pp. 103-113 (1989). 581. Y. S. Jung, B. H. Lee, S. H. Yoo, Physical structure and absorption properties of tailor-made porous starch granules produced by selected amylolytic enzymes. 12, pp. e0181372 (2017). 582. M. J. Sujka, J., Starch granule porosity and its changes by means of amylolysis. International Agrophysics 27, (2007). 583. C. A. K. Good, H.; Somogyi, M., The determination of glycogen. J. Biol. Chem. 100, pp. 485-491 (1933). 584. K. L. Kulp, K., Heat-Moisture Treatment of Starches. I. Physiochemical Properties. Cereal Chemistry 58, pp. 46-48 (1981). 585. G. W. Edgar, PROGRESSIVE MYOCLONUS EPILEPSY AS AN INBORN ERROR OF METABOLISM COMPARABLE TO STORAGE DISEASE. Epilepsia 4, pp. 120-137 (1963). 586. J. Machado-Salas, M. R. Avila-Costa, P. Guevara, J. Guevara, R. M. Duron, D. Bai, M. Tanaka, K. Yamakawa, A. V. Delgado-Escueta, Ontogeny of Lafora bodies and neurocytoskeleton changes in Laforin- deficient mice. Experimental neurology 236, pp. 131-140 (2012). 587. T. Yokota, T. Ishihara, H. Kawano, Y. Yamashita, M. Takahashi, F. Uchino, T. Kamei, Y. Kusunose, M. Yamada, N. Matsumoto,

237

Immunological homogeneity of Lafora body, corpora amylacea, basophilic degeneration in heart, and intracytoplasmic inclusions of liver and heart in type IV glycogenosis. Acta pathologica japonica 37, pp. 941-946 (1987). 588. T. Yokota, T. Ishihara, H. Yoshida, M. Takahashi, F. Uchino, S. Hamanaka, Monoclonal antibody against polyglucosan isolated from the myocardium of a patient with Lafora disease. J Neuropathol Exp Neurol 47, pp. 572-577 (1988). 589. M. W. Van Heycop ten Ham, D. H. Jager, Progressive Myoclonus Epilepsy with Lafora Bodies. Clinical-Pathological Features. Epilepsia 4, pp. 95-119 (1963). 590. J. F. Robyt, D. French, Multiple attach hypothesis of alpha-amylase action: action of porcine pancreatic, human salivary, and Aspergillus oryzae alpha-amylases. Archives of biochemistry and biophysics 122, pp. 8-16 (1967). 591. J. E. Hansen, C. M. Tse, G. Chan, E. R. Heinze, R. N. Nishimura, R. H. Weisbart, Intranuclear protein transduction through a nucleoside salvage pathway. The Journal of biological chemistry 282, pp. 20790-20793 (2007). 592. H. Lu, C. Chen, C. Klaassen, Tissue distribution of concentrative and equilibrative nucleoside transporters in male and female rats and mice. Drug Metab Dispos 32, pp. 1455-1461 (2004). 593. C. R. Crawford, D. H. Patel, C. Naeve, J. A. Belt, Cloning of the human equilibrative, nitrobenzylmercaptopurine riboside (NBMPR)-insensitive nucleoside transporter ei by functional expression in a transport-deficient cell line. The Journal of biological chemistry 273, pp. 5288-5293 (1998). 594. J.-f. Jin, L.-l. Zhu, M. Chen, H.-m. Xu, H.-f. Wang, X.-q. Feng, X.-p. Zhu, Q. Zhou, The optimal choice of medication administration route regarding intravenous, intramuscular, and subcutaneous injection. Patient preference and adherence 9, pp. 923 (2015). 595. M. Villalba-Orero, G. Sanchez-Elexpuru, M. Lopez-Olaneta, O. Campuzano, E. Bello-Arroyo, P. Garcia-Pavia, J. M. Serratosa, R. Brugada, M. P. Sanchez, E. Lara-Pezzi, Lafora Disease Is an Inherited Metabolic Cardiomyopathy. J Am Coll Cardiol 69, pp. 3007-3009 (2017). 596. F. Oksel, H. Tekgul, S. Genc, R. Ozyurek, T. Akalin, S. Tututuncuoglu, A case of Lafora's disease associated with cardiac arrhythmia. Journal of child neurology 14, pp. 745-746 (1999). 597. A. S. de Graaf, E. Ancker, G. S. Rutherfoord, J. J. van der Walt, D. J. Rossouw, Lafora-body disease with optic atrophy, macular degeneration and cardiac failure. Journal of the neurological sciences 93, pp. 69-84 (1989). 598. B. A. Pederson, H. Chen, J. M. Schroeder, W. Shou, A. A. DePaoli-Roach, P. J. Roach, Abnormal cardiac development in the absence of heart glycogen. Molecular and cellular biology 24, pp. 7179-7187 (2004). 599. J. L. Cohen-Pfeffer, S. Gururangan, T. Lester, D. A. Lim, A. J. Shaywitz, M. Westphal, I. Slavc, Intracerebroventricular Delivery as a Safe, Long-

238

Term Route of Drug Administration. Pediatric neurology 67, pp. 23-35 (2017). 600. P. J. Carter, G. A. Lazar, Next generation antibody drugs: pursuit of the 'high-hanging fruit'. Nature reviews. Drug discovery 17, pp. 197-223 (2018). 601. A. Beck, L. Goetsch, C. Dumontet, N. Corvaia, Strategies and challenges for the next generation of antibody-drug conjugates. Nature reviews. Drug discovery 16, pp. 315-337 (2017). 602. P. A. Baeuerle, C. Reinhardt, Bispecific T-cell engaging antibodies for cancer therapy. Cancer Res 69, pp. 4941-4944 (2009). 603. F. L. Moolten, S. R. Cooperband, Selective destruction of target cells by diphtheria toxin conjugated to antibody directed against antigens on the cells. Science (New York, N.Y.) 169, pp. 68-70 (1970). 604. B. Illingworth, G. T. Cori, Structure of glycogens and amylopectins. III. Normal and abnormal human glycogen. The Journal of biological chemistry 199, pp. 653-660 (1952). 605. Z. Jian, M. R. Alley, J. Cayzer, G. R. Swinney, Lafora's disease in an epileptic Basset hound. New Zealand veterinary journal 38, pp. 75-79 (1990). 606. J. M. Young, E. Weser, The metabolism of circulating maltose in man. J Clin Invest 50, pp. 986-991 (1971). 607. D. S. Leong, J. G. Tan, C. L. Chin, S. Y. Mak, Y. S. Ho, S. K. Ng, Evaluation and use of disaccharides as energy source in protein-free mammalian cell cultures. Sci Rep 7, pp. 45216 (2017). 608. W. Loscher, D. Schmidt, Modern antiepileptic drug development has failed to deliver: ways out of the current dilemma. Epilepsia 52, pp. 657-678 (2011). 609. H. E. Scharfman, Neuroscience. Metabolic control of epilepsy. Science (New York, N.Y.) 347, pp. 1312-1313 (2015). 610. H. Ozen, Glycogen storage diseases: new perspectives. World J Gastroenterol 13, pp. 2541-2553 (2007). 611. T. T. Rohn, Corpora Amylacea in Neurodegenerative Diseases: Cause or Effect? Int J Neurol Neurother 2, (2015). 612. R. Virchow, Zur Cellulose-Frage. Virchows Arch. 6, pp. 416-426 (1854). 613. R. A. Kyle, Amyloidosis: a convoluted story. Br J Haematol 114, pp. 529- 538 (2001). 614. J. D. Sipe, A. S. Cohen, Review: history of the amyloid fibril. J Struct Biol 130, pp. 88-98 (2000). 615. M. S. Cafferty, R. E. Lovelace, A. P. Hays, S. Servidei, S. Dimauro, L. P. Rowland, Polyglucosan body disease. Muscle Nerve 14, pp. 102-107 (1991). 616. M. Sakai, J. Austin, F. Witmer, L. Trueb, Studies of corpora amylacea. I. Isolation and preliminary characterization by chemical and histochemical techniques. Arch Neurol 21, pp. 526-544 (1969). 617. T. Niittyla, S. Comparot-Moss, W.-L. Lue, G. Messerli, M. Trevisan, M. D. J. Seymour, J. A. Gatehouse, D. Villadsen, S. M. Smith, J. Chen, S. C.

239

Zeeman, A. M. Smith, Similar protein phosphatases control starch metabolism in plants and glycogen metabolism in mammals. J. Biol. Chem. 281, pp. 11815-11818 (2006). 618. W. S. Ratnayake, D. S. Jackson, Starch gelatinization. Adv Food Nutr Res 55, pp. 221-268 (2009). 619. J. Wikman, A. Blennow, E. Bertoft, Effect of amylose deposition on potato tuber starch granule architecture and dynamics as studied by lintnerization. Biopolymers 99, pp. 73-83 (2013). 620. C. Gerard, V. Planchot, P. Colonna, E. Bertoft, Relationship between branching density and crystalline structure of A- and B-type maize mutant starches. Carbohydr Res 326, pp. 130-144 (2000). 621. S. Srichuwong, N. Isono, T. Mishima, M. Hisamatsu, Structure of lintnerized starch is related to X-ray diffraction pattern and susceptibility to acid and enzyme hydrolysis of starch granules. International journal of biological macromolecules 37, pp. 115-121 (2005). 622. R. E. Rundle, J. F. Foster, R. R. Baldwin, On the nature of the starch— iodine complex. Journal of the American Chemical Society 66, pp. 2116- 2120 (1944). 623. M. K. Herrick, J. L. Twiss, G. D. Vladutiu, G. F. Glasscock, D. S. Horoupian, Concomitant branching enzyme and phosphorylase deficiencies. An unusual glycogenosis with extensive neuronal polyglucosan storage. J Neuropathol Exp Neurol 53, pp. 239-246 (1994). 624. G. B. Reed Jr, J. F. Dixon, J. B. Neustein, G. N. Donnell, B. H. Landing, Type IV glycogenosis. Patient with absence of a branching enzyme alpha- 1, 4-glucan: alpha-1, 4-glucan 6-glycosyl transferase. Laboratory investigation; a journal of technical methods and pathology 19, pp. 546 (1968). 625. M. Ovecka, A. Bahaji, F. J. Munoz, G. Almagro, I. Ezquer, E. Baroja- Fernandez, J. Li, J. Pozueta-Romero, A sensitive method for confocal fluorescence microscopic visualization of starch granules in iodine stained samples. Plant Signal Behav 7, pp. 1146-1150 (2012). 626. A. Bahaji, J. Li, M. Ovecka, I. Ezquer, F. J. Muñoz, E. Baroja-Fernández, J. M. Romero, G. Almagro, M. Montero, M. Hidalgo, Arabidopsis thaliana mutants lacking ADP-glucose pyrophosphorylase accumulate starch and wild-type ADP-glucose content: further evidence for the occurrence of important sources, other than ADP-glucose pyrophosphorylase, of ADP- glucose linked to leaf starch biosynthesis. Plant and cell physiology 52, pp. 1162-1176 (2011). 627. A. K. Donald, K. L. Kato, P. A. Perry, T. A. Waigh, Scattering Studies of the Internal Structure of Starch Granules. Starch 53, pp. 504-512 (2001). 628. J. Blazek, E. P. Gilbert, Application of small-angle X-ray and neutron scattering techniques to the characterisation of starch structure: A review. Carbohydrate Polymers 85, pp. 281-293 (2011). 629. S. C. Zeeman, J. Kossmann, A. M. Smith, Starch: Its Metabolism, Evolution, and Biotechnological Modification in Plants. Annu Rev Plant Biol 61, pp. 209-234 (2010).

240

630. P. Colonna, A. Buleon, M. Lemaguer, C. Mercier, Pisum sativum and vicia faba carbohydrates: Part IV — Granular structure of wrinkled pea starch. Carbohydrate Polymers 2, pp. 43-59 (1982). 631. T. Shimonaga, M. Konishi, Y. Oyama, S. Fujiwara, A. Satoh, N. Fujita, C. Colleoni, A. Buleon, J. L. Putaux, S. G. Ball, A. Yokoyama, Y. Hara, Y. Nakamura, M. Tsuzuki, Variation in storage alpha-glucans of the Porphyridiales (Rhodophyta). Plant Cell Physiol 49, pp. 103-116 (2008). 632. P. Adhikary, K. N. Tiwari, R. P. Singh, Synthesis, Characterization, and Flocculation Characteristics of Polyacrylamide-Grafted Glycogen. Journal of Applied Polymer Science 103, pp. 773-778 (2006). 633. K. Svegmark, K. Helmersson, G. Nilsson, P. O. Nilsson, R. Andersson, E. Svensson, Comparison of potato amylopectin starches and potato starches—influence of year and variety. Carbohydrate Polymers 47, pp. 331-340 (2002). 634. R. E. Cameron, A. M. Donald, A small-angle X-ray scattering study of the annealing and gelatinization of starch. Polymer 33, pp. 2628-2635 (1992). 635. J.-L. Putaux, A. Buleon, H. Chanzy, Network formation in dilute amylose and amylopectin studied by TEM. Macromolecules 33, pp. 6416-6422 (2000). 636. R. E. Cameron, A. M. Donald, in Food polymers, gels and colloids. (The Royal Society of Chemistry Cambridge, 1991), pp. 301-309. 637. M. Majzoobi, N. Seifzadeh, A. Farahnaky, G. Mesbahi, Effects of sonication on physical properties of native and cross‐linked wheat starches. Journal of Texture Studies 46, pp. 105-112 (2015). 638. W. S. Ratnayake, D. S. Jackson, Gelatinization and solubility of corn starch during heating in excess water: new insights. Journal of Agricultural and Food Chemistry 54, pp. 3712-3716 (2006). 639. H. Jacobs, R. C. Eerlingen, N. Rouseu, P. Colonna, J. A. Delcour, Acid hydrolysis of native and annealed wheat, potato and pea starches—DSC melting features and chain length distributions of lintnerised starches. Carbohydrate Research 308, pp. 359-371 (1998). 640. J.-L. Putaux, S. Molina-Boisseau, T. Momaur, A. Dufresne, Platelet nanocrystals resulting from the disruption of waxy maize starch granules by acid hydrolysis. Biomacromolecules 4, pp. 1198-1202 (2003). 641. J. Wikman, A. Blennow, A. Buleon, J. L. Putaux, S. Perez, K. Seetharaman, E. Bertoft, Influence of amylopectin structure and degree of phosphorylation on the molecular composition of potato starch lintners. Biopolymers 101, pp. 257-271 (2014). 642. M. Kornfeld, M. LeBaron, Glycogenosis type VIII. J Neuropathol Exp Neurol 43, pp. 568-579 (1984). 643. T. Iwamasa, Physico-chemical properties of glycogen from the rat liver and muscle. Acta Histochemica et Cytochemica 5, pp. 106-114 (1972). 644. S. DiMauro, C. Lamperti, Muscle glycogenoses. Muscle Nerve 24, pp. 984-999 (2001).

241

645. R. Nishimura, K. Ishak, R. Reddick, S. James, R. Porter, J. A. Barranger, Lafora's disease; diagnosis by liver biopsy. Trans Am Neurol Assoc 104, pp. 119-120 (1979). 646. S. Wang, H. Luo, J. Zhang, Y. Zhang, Z. He, S. Wang, Alkali-induced changes in functional properties and in vitro digestibility of wheat starch: the role of surface proteins and lipids. Journal of agricultural and food chemistry 62, pp. 3636-3643 (2014). 647. M. Kerly, The solubility of glycogen. Biochemical Journal 24, pp. 67 (1930). 648. A. Perera, V. Meda, R. T. Tyler, Resistant starch: A review of analytical protocols for determining resistant starch and of factors affecting the resistant starch content of foods. Food Research International 43, pp. 1959-1974 (2010). 649. M. R. Stetten, H. M. Katzen, D. Stetten, A comparison of the glycogens isolated by acid and alkaline procedures. Journal of Biological Chemistry 232, pp. 475-488 (1958). 650. R. C. S. Thys, H. Westfahl Jr, C. P. Z. Noreña, L. D. F. Marczak, N. P. Silveira, M. B. Cardoso, Effect of the alkaline treatment on the ultrastructure of C-type starch granules. Biomacromolecules 9, pp. 1894- 1901 (2008). 651. M. J. Gidley, P. V. Bulpin, Crystallisation of malto-oligosaccharides as models of the crystalline forms of starch: minimum chain-length requirement for the formation of double helices. Carbohydrate Research 161, pp. 291-300 (1987). 652. L. Cai, Y.-C. Shi, Self-assembly of short linear chains to A-and B-type starch spherulites and their enzymatic digestibility. Journal of agricultural and food chemistry 61, pp. 10787-10797 (2013). 653. R. Hoover, T. Hughes, H. J. Chung, Q. Liu, Composition, molecular structure, properties, and modification of pulse starches: A review. Food research international 43, pp. 399-413 (2010). 654. D. Qiao, B. Zhang, J. Huang, F. Xie, D. K. Wang, F. Jiang, S. Zhao, J. Zhu, Hydration-induced crystalline transformation of starch polymer under ambient conditions. International journal of biological macromolecules 103, pp. 152-157 (2017). 655. J.-L. Putaux, G. Potocki-Véronese, M. Remaud-Simeon, A. Buleon, α-D- Glucan-based dendritic nanoparticles prepared by in vitro enzymatic chain extension of glycogen. Biomacromolecules 7, pp. 1720-1728 (2006). 656. M. Fändrich, Oligomeric intermediates in amyloid formation: structure determination and mechanisms of toxicity. Journal of molecular biology 421, pp. 427-440 (2012). 657. C. G. Glabe, Common mechanisms of amyloid oligomer pathogenesis in degenerative disease. Neurobiology of aging 27, pp. 570-575 (2006). 658. E. Auge, I. Cabezon, C. Pelegri, J. Vilaplana, New perspectives on corpora amylacea in the human brain. Sci Rep 7, pp. 41807 (2017). 659. C. Sinadinos, J. Valles-Ortega, L. Boulan, E. Solsona, M. F. Tevy, M. Marquez, J. Duran, C. Lopez-Iglesias, J. Calbo, E. Blasco, M. Pumarola,

242

M. Milan, J. J. Guinovart, Neuronal glycogen synthesis contributes to physiological aging. Aging Cell 13, pp. 935-945 (2014). 660. I. E. Scheffer, Genetic Testing in Epilepsy: What Should you be Doing? Genetic Testing in Epilepsy. Epilepsy currents 11, pp. 107-111 (2011). 661. M. B. Metzger, V. A. Hristova, A. M. Weissman, HECT and RING finger families of E3 ubiquitin ligases at a glance. Journal of cell science 125, pp. 531-537 (2012). 662. E. Bulatov, A. Ciulli, Targeting Cullin-RING E3 ubiquitin ligases for drug discovery: structure, assembly and small-molecule modulation. The Biochemical journal 467, pp. 365-386 (2015). 663. N. F. Schor, Why our patients (and we) need basic science research. Neurology 80, pp. 2070-2075 (2013). 664. F. S. Collins. (American Association for the Advancement of Science, 2012). 665. K. Rice, M. S. Gentry, in ASBMB Today. (American Society for Biochemistry and Molecular Biology, Rockville, MD, 2017), vol. May 2017. 666. Z. Zhou, C. J. Kinslow, H. Hibshoosh, H. Guo, S. K. Cheng, C. He, M. S. Gentry, R. C. Sun, Clinical Features, Survival and Prognostic Factors of Glycogen-Rich Clear Cell Carcinoma (GRCC) of the Breast in the US Population. Journal of clinical medicine 8, pp. 246 (2019). 667. J. C. Dodge, C. M. Treleaven, J. A. Fidler, T. J. Tamsett, C. Bao, M. Searles, T. V. Taksir, K. Misra, R. L. Sidman, S. H. Cheng, Metabolic signatures of amyotrophic lateral sclerosis reveal insights into disease pathogenesis. Proceedings of the National Academy of Sciences 110, pp. 10812-10817 (2013). 668. G. Polekhina, A. Gupta, B. J. van Denderen, S. C. Feil, B. E. Kemp, D. Stapleton, M. W. Parker, Structural basis for glycogen recognition by AMP-activated protein kinase. Structure 13, pp. 1453-1462 (2005). 669. J. Long, T. R. Gallagher, J. R. Cavey, P. W. Sheppard, S. H. Ralston, R. Layfield, M. S. Searle, Ubiquitin recognition by the ubiquitin-associated domain of p62 involves a novel conformational switch. The Journal of biological chemistry 283, pp. 5427-5440 (2008). 670. D. H. Kwon, O. H. Park, L. Kim, Insights into degradation mechanism of N- end rule substrates by p62/SQSTM1 autophagy adapter. 9, pp. 3291 (2018). 671. L. Swain, G. Key, A. Tauro, S. Ahonen, P. Wang, C. Ackerley, B. A. Minassian, C. Rusbridge, Lafora disease in miniature Wirehaired Dachshunds. PloS one 12, pp. e0182024 (2017). 672. H. Gredal, M. Berendt, P. S. Leifsson, Progressive myoclonus epilepsy in a beagle. J Small Anim Pract 44, pp. 511-514 (2003).

243

VITA

Mary Kathryn (Katy) Brewer

Education Bachelor of Science in Biology, Minor in Mathematics May 2013 University of Kentucky, Lexington, KY

Honors Barcelona Institute of Science and Technology 2019-2022 3-year PROBIST postdoctoral fellowship Department of Biochemistry Spring Research Conference 2018 First place in graduate student poster competition International Conference on Brain Energy Metabolism (ICBEM) 2018 Travel award, $1100 First place in poster competition UK College of Medicine Student Travel Award, $800 2017 Philanthropic Educational Organization (P.E.O.) Scholar Award 2017 $15,000 Department of Biochemistry Spring Research Conference 2017 First place in graduate student poster competition Bluegrass Society for Neuroscience Spring Neuroscience Day 2017 Outstanding poster award National Institute of Neurological Disorders and Stroke 2016-2019 F31 Predoctoral Fellowship Kentucky Opportunity Fellowship, $15,000 2015 Northern Kentucky/Greater Cincinnati UK Alumni Club Fellowship 2015 $5,000 ASBMB Graduate/Postdoctoral Travel Award, $1,000 2015

Publications 1. Brewer MK, Brizzee, C, Raththagala M, Wayne, J, Hellman, L, Viana R, García-Gimeno MA, Sanz P, Li S, Vander Kooi CW, Gentry MS. The catalytic activity and dynamics of the human glycogen phosphatase laforin provide insights into Lafora disease. In preparation. 2. Brewer MK, Viana R, García-Gimeno MA, Raththagala M, Sternbach S, Wayne J, Li S, Sanz P, Vander Kooi CW, and Gentry MS. Personalized biochemistry of Lafora disease patient mutations. In preparation. 3. Brewer MK, Putaux JL, Uittenbogaard A, Szydlowski N, Helle S, Gentry MS. Structural characterization of the polyglucosan bodies of Lafora disease. In preparation. 4. Brewer MK, Uittenbogaard A, Austin G, McCarthy JJ, Segvich DM, DePaoli-Roach A, Roach PJ, Hodges BL, Zeller J, Pauly JR, McKnight T, Armstrong D, Gentry MS. Targeting pathogenic Lafora bodies in Lafora disease using an antibody-enzyme fusion. Reviewed at Cell Metabolism

and currently addressing the reviews. 5. Brewer MK, Grossman TR, McKnight T, Goldberg PY, Landy H, Gentry MS. The 4th International Lafora Workshop: shifting paradigms, paths to treatment, and hope for patients. Editorial. Epilepsy & Behavior 2019, 90:284-286. 6. Brewer MK, Gentry MS. Brain glycogen structure and its associated proteins: past present and future. Advances in Neurobiology: Brain Glycogen Metabolism (invited book chapter). Springer. In press. 7. García-Gimeno MA, Rodilla-Ramirez PN, Viana R, Salas-Puig X, Brewer MK, Gentry MS, Sanz P. A novel EPM2A mutation yields a slow progression form of Lafora disease. Epilepsy Research 2018, 145:169-177. 8. Brewer MK, Gentry MS. The 3rd International Lafora Workshop: Evidence for a cure. Editorial. Epilepsy & Behavior. 2018, 81:125-127. 9. Brewer MK, Sherwood AR, Dukhande VV, and Gentry MS. Function and action of the glucan phosphatase laforin. In Encyclopedia of Signaling Molecules. Springer. 2017. 10. Gentry MS, Brewer MK, Vander Kooi CW. Structural Biology of Glucan Phosphatases from Humans to Plants. Review. Current Opinion in Structural Biology 2016, 40:62-69. 11. Emanuelle S#, Brewer MK#, Meekins DA, Gentry MS. Unique binding platforms among the glucan phosphatases. Review. Cellular and Molecular Life Sciences 2016, 73(14):2765-78. #These authors contributed equally to this work. 12. Raththagala M, Brewer MK, Parker MW, Sherwood AR, Wong BK, Hsu S, Bridges TM, Paasch BC, Hellman LM, Husodo S, Meekins DA, Taylor AO, Turner BD, Auger KD, Dukhande VV, Chakravarthy S, Sanz P, Woods VL Jr, Li S, Vander Kooi CW, Gentry MS. Structural Mechanism of Laforin Function in Glycogen Dephosphorylation and Lafora Disease. Molecular Cell 2015, 57(2):261-72. 13. Singh R, Brewer MK, Mashburn CB, Lou D, Bondada V, Graham B, Geddes JW. Calpain 5 is highly expressed in the CNS, carries dual nuclear localization signals, and is associated with nuclear Promyelocytic Leukemia Protein bodies. Journal of Biological Chemistry 2014, 289(28):19383-94. 14. Brewer MK, Husodo S, Dukhande VV, Johnson MB, Gentry MS. Expression, purification and characterization of soluble red rooster laforin as a fusion protein in Escherichia coli. BMC Biochemistry 2014, 15:8.

My artwork was featured on the cover of the 2018 JBC thematic minireview series, "Brain Glycogen Metabolism."

245