THE DISCOVERY OF NOVEL GSK3 SUBSTRATES AND THEIR ROLE IN THE BRAIN

James Robinson

A DISSERTATION IN FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

St. Vincent’s Clinical School Facility of Medicine The University of New South Wales

May 2015

ORIGINALITY STATEMENT

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

Signed ………………………………………………

Date …………………………………………………

ii |Page

Abstract

Bipolar Disorder (BD) is a debilitating disease that dramatically impairs people’s lives and severe cases can lead to exclusion from society and suicide. There are no clear genetic or environmental causes, and current treatments suffer from limiting side effects. Therefore, alternative intervention strategies are urgently required. Our approach is to determine mechanisms of action of current drug therapies, in the hope that they will lead to discovery of next generation therapeutic targets. Lithium has been the mainstay treatment for BD for over 50 years, although its mechanism of action is not yet clear. A major target of lithium (and other mood stabilizers) is Glycogen Synthase Kinase-3 (GSK-3), a Ser/Thr protein kinase that is dysregulated in BD. Pathogenic targets of GSK3 could become novel therapeutic targets for improved treatment of BD, although these are not yet known. Identifying these targets is the primary goal of this project.

Our lab used a combination of bioinformatics and phosphoproteomics to discover 45 novel substrates of GSK3 involved in vesicular trafficking events. Here, I focus on two promising trafficking proteins; the lipid kinase phosphatidylinositol 4-kinase II alpha (PI4KII) and the AP-2 kinase adaptin associated kinase-1 (AAK1).

PI4KIIα regulates cell-surface expression of AMPA receptors in neurons and is therefore likely to affect neurotransmission in the brain. Phosphorylation regulates this process by promoting binding of PI4KIIα to Adaptin  of the AP-3 complex for trafficking to the lysosome to be degraded. Depletion of PI4KIIα in neurons of Drosophila increased their locomotor activity, consistent with hyperactivity exhibited by BD patients in their manic phase, and this was prevented with lithium treatment. Separately, I demonstrate that AAK1 regulates trafficking of the AP-2 complex to recycling endosomes in cells. Surprisingly, depletion of AAK1 increases autophagy flux, implicating it as a potential target in autophagy-related Parkinson’s disease (PD). Accordingly, depletion of AAK1 in Drosophila neurons increased their susceptibility to PD and autophagy-related death.

This project identifies two novel substrates of GSK3 that are linked to debilitating neurological disorders. This provides a valuable basis for future research investigating the therapeutic potential of these and other novel GSK3 substrates discovered in our lab.

iii |Page

Acknowledgements

I would like to thank my primary supervisor and mentor Dr. Adam Cole for taking me on as a student as part of the Neurosignalling and Mood Disorders group at the Garvan Institute of Medical Research. Thank you for all your support and guidance throughout my PhD. Thank you for you endless help on daily tasks, refining experimental design, reviewing presentations and thesis drafts and the continual encouragement needed to make me a better scientist. Your office door was always open for educative discussions and I appreciate all the time you made available for me. I had a great PhD experience and I have learnt so much.

I would like to thank Hovik Farghaian, a key member of the Neurosignalling and Mood Disorders group, and my very good friend. Your support, expertise and friendship where invaluable throughout my time at the Garvan Institute. Thank you for being such a great friend.

I would also like to thank my associate supervisors Prof. Herbert Herzog and Dr. Greg Neely for all your support and wisdom during laboratory meetings and presentations and for your critical evaluation of my work. Thank you to all members, past and present, of the Functional Genomic group for all your collaboration and support over the years.

Thank you to our collaborators Dr Vladimir Sytnyk and Iryna Leshchyns’ka at The University of New South Wales (UNSW) for their hippocampal neuron work and to Dr William Hughes for his help and support with microscopy.

I would like to thank my family and friends for their endless support and expressing an interest in my work. Special thanks to my beautiful wife Jessica Robinson for all your love, support and understanding during this time. I couldn’t have asked for a better partner in life.

Without their combined efforts, the completion of this project would not have been possible.

iv |Page

Contents

ABSTRACT III

ACKNOWLEDGEMENTS IV

LIST OF TABLES VIII

LIST OF FIGURES IX

ABBREVIATIONS XII

CHAPTER 1 – INTRODUCTION AND LITERATURE REVIEW 1 1.1 CHARACTERISATION OF GSK3 ...... 1 1.1.1 GSK3 GENES ...... 1 1.1.2 GSK3 PROTEINS ...... 2 1.1.3 GSK3 EXPRESSION ...... 3 1.1.4 GSK3 KINASE ACTIVITY ...... 3 1.2 SIGNALLING PATHWAYS REGULATING GSK3 ACTIVITY ...... 6 1.2.1 THE GROWTH FACTOR SIGNALLING PATHWAY ...... 6 1.2.2 THE WNT PATHWAY ...... 8 1.2.3 THE HEDGEHOG PATHWAY ...... 10 1.2.4 THE NOTCH PATHWAY ...... 12 1.3 PHYSIOLOGY OF GSK3 MUTANT MICE ...... 14 1.3.1 GSK3 MUTANT MICE ...... 14 1.3.2 GSK3 MUTANT MICE ...... 15 1.4 GSK3 SUBSTRATES ...... 16 1.4.1 GSK3 AND ITS SUBSTRATES INVOLVED IN APOPTOSIS ...... 16 1.4.1.1 The role of GSK3 in pro-apoptotic signalling ...... 16 1.4.1.2 The role of GSK3 in cell survival signalling ...... 19 1.4.1.3 Summary: GSK3 is a key regulator of cell apoptosis ...... 21 1.4.2 THE ROLE OF GSK3 IN THE REGULATION OF IMMUNE RESPONSES ...... 21 1.4.2.1 Innate immunity ...... 21 1.4.2.2 Adaptive immunity ...... 22 1.4.2.3 Immune transcription factors ...... 23 1.4.3 GSK3’S ROLE IN STEM CELL PROLIFERATION VS. DIFFERENTIATION ...... 24 1.4.3.1 Signalling pathways in stem cells ...... 24 1.4.3.2 GSK3 mediates the regulation of gene expression downstream of transcription factors ...... 27 1.4.3.3 Summary: GSK3 activity is important in the regulation of cell-fate ...... 28 1.4.4 GSK3 REGULATION OF NEUROGENESIS ...... 29 1.4.4.1 GSK3 is a critical regulator of neurogenesis downstream of signalling pathways ...... 29 1.4.4.2 Summary: GSK3’s dynamic control of neurogenesis ...... 31 1.4.5 GSK3 AND NEURONAL MORPHOLOGY ...... 32 1.4.5.1 GS3 is a key regulator of axon growth and branching ...... 32 1.4.5.2 GSK3-mediated growth factor signalling in neuronal development ...... 34 1.4.5.3 Downstream substrates ...... 35 1.4.5.4 Summary: Modulation of GSK3 activity regulates neuronal morphology ...... 39 1.4.6 GSK3 AND NEUROTRANSMISSION ...... 39 1.4.6.1 The role of GSK3 in LTP/LTD ...... 40 1.4.6.2 Neurotransmitter receptors ...... 41 1.4.6.3 GSK3 regulates synaptic proteins ...... 43 1.4.6.4 Summary: GSK3 is a central regulator of learning and memory formation ...... 44 1.5 GSK3 AND MOOD DISORDERS ...... 45 1.5.1 GSK3 IS A MAIN TARGET IN BIPOLAR DISORDER ...... 45

v |Page

1.5.2 GSK3 ACTIVITY IN THE DEVELOPMENT OF SCHIZOPHRENIA ...... 50 1.5.3 CIRCADIAN PROTEINS ...... 51 1.5.4 SUMMARY: GSK3 IS IMPORTANT IN THE REGULATION OF MOOD DISORDERS/SCHIZOPHRENIA ...... 52 1.6 PROJECT HYPOTHESIS AND AIM ...... 54 1.6.1 RATIONALE ...... 54 1.6.2 HYPOTHESIS ...... 54 1.6.3 AIM AND SIGNIFICANCE...... 56

CHAPTER 2: EXPERIMENTAL PROCEDURES 57 2.1 MATERIALS ...... 57 2.1.1 DNA CONSTRUCTS ...... 57 2.1.2 PHOSPHOSPECIFIC ANTIBODIES ...... 57 2.1.3 REAGENTS ...... 58 2.2 METHODS ...... 58 2.2.1 BIOCHEMICAL VALIDATION OF GSK3 SUBSTRATES ...... 58 2.2.2 CELL CULTURE ...... 59 2.2.3 SINGLE-ROUND ENDOCYTOSIS AND RECYCLING OF TRANSFERRIN ...... 59 2.2.4 IMMUNOFLUORESCENCE MICROSCOPY ...... 60 2.2.5 WESTERN BLOTTING ...... 61 2.2.6 IN VITRO KINASE ASSAY ...... 61 2.2.7 IN VITRO PHOSPHATASE ASSAY ...... 61 2.2.8 SOURCE AND MAINTENANCE OF FLIES ...... 62 2.2.9 LOCOMOTOR ACTIVITY AND SLEEP ANALYSIS ...... 63 2.2.10 BODY WEIGHT AND TRIACYGLYCERIDE (TAG) ANALYSIS ...... 63 2.2.11 CAPILLARY FEEDER (CAFE) ASSAY ...... 64 2.2.12 STARVATION AND NICOTINE TOXICITY ASSAY ...... 64 2.2.13 STARTLE-INDUCED NEGATIVE GEOTAXIS ASSAY (CLIMBING ASSAY) ...... 64 2.2.14 TOTAL RNA EXTRACTION AND CDNA SYNTHESIS ...... 65 2.2.15 QPCR ...... 65 2.2.16 STATISTICAL ANALYSIS ...... 66

CHAPTER 3: PI4KIIΑ PHOSPHORYLATION BY GSK3 DIRECTS VESICULAR TRAFFICKING TO LYSOSOMES 67 3.1 INTRODUCTION ...... 67 3.1.1 PI4-KINASES ...... 67 3.1.2 ROLE OF PI4KIIΑ IN VESICULAR TRAFFICKING ...... 68 3.1.3 PI4KIIΑ ROLE IN NEUROTRANSMISSION ...... 69 3.1.4 AIMS ...... 70 3.2 RESULTS ...... 71 3.2.1 MAPPING AND CHARACTERISATION OF PHOSPHOSITES IN PI4KIIΑ ...... 71 3.2.1.1 Relative resistance of PI4KIIα phosphosites to phosphatases...... 74 3.2.2 PHOSPHORYLATION OF PI4KIIΑ BY GSK3 REGULATES VESICULAR TRAFFICKING ...... 78 3.2.2.1 Phosphorylation of PI4KIIα slows transferrin recycling in HeLa cells ...... 78 3.2.2.2 Phosphorylation of PI4KIIα promotes trafficking of AMPA receptors away from the cell surface in neurons ...... 78 3.2.2.3 Phosphorylation of PI4KIIα has no effect on neuronal morphology ...... 79 3.2.3 PHOSPHORYLATION OF PI4KIIΑ BY GSK3 PROMOTES TRAFFICKING TO LYSOSOMES BY THE AP-3 COMPLEX ...... 83 3.2.3.1 GSK3 phosphorylation of PI4KIIα targets it for degradation by the lysosome ...... 83 3.2.3.2 Phosphorylation of PI4KIIα by GSK3 promotes binding to adaptin  ...... 83 3.2.3.3 Phosphorylation of PI4KIIα promotes binding to AP-3 for trafficking to the lysosome to be degraded ...... 88 3.2.3.4 The N-terminal region of PI4KIIα regulates binding of the dileucine motif to the AP-3 complex 88 3.2.4 INVESTIGATING THE PHYSIOLOGICAL FUNCTION OF PI4KIIΑ IN DROSOPHILA MELANOGASTER ...... 93 3.2.4.1 Depletion of PI4KIIα produces hyperactivity in Drosophila ...... 93

vi |Page

3.2.4.2 Lithium reduces hyperactivity of PI4KIIα-depleted flies ...... 96 3.2.4.3 Effect of lithium on Drosophila sleep behaviour ...... 100 3.2.4.4 PI4KIIα depletion has no effect on feeding behaviour or locomotor function ...... 100 3.2.5 QPCR VALIDATION OF PI4KIIΑ DEPLETION IN DROSOPHILA ...... 104 3.2.5.1 Determining gene amplification efficiency ...... 104 3.2.5.2 Calculating relative gene expression using the 2-CT (Livak) method ...... 104 3.3 DISCUSSION ...... 106

CHAPTER 4: AAK1 PHOSPHORYLATION BY GSK3 PROMOTES AP-2-MEDIATED AUTOPHAGIC CLEARANCE. 111 4.1 INTRODUCTION ...... 111 4.1.1 ADAPTOR-ASSOCIATED PROTEIN KINASE 1 (AAK1) ...... 113 4.1.2 AAK1 KINASE ACTIVITY COORDINATES THE ASSEMBLY OF ENDOCYTIC MACHINERY ...... 114 4.1.3 AAK1 ROLE IN NEURODEVELOPMENT AND FUNCTION ...... 114 4.2 RESULTS ...... 116 4.2.1 MAPPING AND CHARACTERISATION OF GSK3 PHOSPHOSITES IN AAK1 ...... 116 4.2.1.1 AAK1 phosphosite Thr620 is resistant to dephosphorylation by phosphatases ...... 118 4.2.1.2 GSK3 phosphorylation of AAK1 increases its stability in cells ...... 121 4.2.2 PHOSPHORYLATION OF AAK1 BY GSK3 PROMOTES PHOSPHORYLATION OF ADAPTIN µ AND DISSOCIATION FROM THE AP-2 COMPLEX ...... 122 4.2.2.1 GSK3 phosphorylation of AAK1 promotes dissociation from AP-2 complex ...... 122 4.2.2.2 Phosphorylation by GSK3 promotes AAK1-mediated phosphorylation of adaptin µ ...... 125 4.2.3 PHOSPHORYLATION OF AAK1 BY GSK3 REGULATES VESICULAR TRAFFICKING ...... 128 4.2.3.1 Phosphorylation of AAK1 promotes Transferrin recycling in HeLa cells ...... 128 4.2.3.2 AAK1 dissociation with the AP-2 complex increases recycling ...... 132 4.2.4 INVESTIGATING THE PHYSIOLOGICAL FUNCTION OF AAK1 IN DROSOPHILA MELANOGASTER ...... 135 4.2.4.1 Depletion of AAK1 in Drosophila produce no behaviours that correlate with BD ...... 135 4.2.4.2 AAK1 depleted flies are susceptible to starvation-induced death ...... 138 4.2.5 AAK1 DEPLETION INCREASES AUTOPHAGY PROCESSES IN CELLS ...... 141 4.2.5.1 HeLa cells stably expressing shAAK1 ...... 141 4.2.5.2 AAK1 depletion reduces LC3-II staining in HeLa cells ...... 142 4.2.5.3 AAK1 depletion increases autophagy flux ...... 142 4.2.6 A ROLE OF AAK1 IN PD PATHOGENESIS ...... 145 4.2.6.1 Depletion of AAK1 produces PD phenotype in flies ...... 145 4.2.6.2 Model of GSK3-mediated phosphorylation of AAK1 increasing autophagy and PD pathogenesis ...... 147 4.3 DISCUSSION ...... 149

CHAPTER 5: CONCLUSIONS, PERSPECTIVES AND FUTURE DIRECTIONS 155 5.1 PROJECT OUTCOMES ...... 155 5.1.1 PI4KIIΑ REGULATES AP-3-MEDIATED VESICULAR TRAFFICKING IN BD ...... 156 5.1.2 AAK1 REGULATES AP-2-MEDIATED AUTOPHAGOSOME FORMATION IN PD ...... 157 5.2 RESEARCH IMPLICATIONS ...... 158 5.2.1 GSK3’S ROLE IN NEUROTRANSMISSION AND THE IMPLICATION FOR BD PATHOGENICITY ...... 158 5.2.2 GSK3’S ROLE IN AUTOPHAGY AND THE IMPLICATION FOR PD PATHOGENICITY ...... 161 5.3 PROJECT LIMITATIONS ...... 165 5.4 RECOMMENDATIONS FOR FUTURE RESEARCH ...... 168 5.4.1 VALIDATING THE ROLE OF PI4KIIΑ AND AAK1 IN MOOD AND NEURODEGENERATIVE DISORDERS OF THE BRAIN...... 168 5.4.2 SCREENING OTHER GSK3 CANDIDATE SUBSTRATES USING THE DAMS ASSAY FOR BEHAVIOURS ASSOCIATED WITH MOOD DISORDERS ...... 170 5.5 OVERALL CONCLUSION ...... 175

REFERENCES 176

vii |Page

List of Tables

2.1 DNA constructs: primers and amplicons size (bp) ……………………………………….. 57 2.2 Peptide antigens for antibody production ………………………………………………… 58

2.3 Primer sequences for qPCR, melting Tm and amplicon sizes ……………………………. 66

viii |Page

List of Figures

1.1 Schematic representation of mammalian GSK3 and GSK3 …………………………… 2 1.2 GSK3 substrates specificity …………………………………………………………...... 5 1.3 Schematic illustration of the PI3K/Akt signalling pathway regulating GSK3 activity and influencing glycogen synthesis ….………………………………………………………… 7 1.4 Schematic illustration of the Wnt/β-catenin signalling pathway ………………………….. 9 1.5 Schematic illustration of the Hedgehog signalling pathway ……………………………… 11 1.6 Signalling pathways regulating cell fate target different subsets of GSK3 substrates ……. 13 1.7 GSK3 promotes mitochondrial apoptosis signalling ……………………………………… 18

1.8 GSK3 positively regulates NF-B transcriptional activity directly and indirectly ………. 21 1.9 Increased GSK3 activity promotes innate inflammatory response ……………………….. 22 1.10 GSK3 regulates T cell response to maintain adaptive immunity homeostasis …………… 23 1.11 GSK3 also targets other transcription factors and cell cycle regulators for Fbw7 ubiquitination and degradation by the proteasome ……………………………………….. 28 1.12 GSK3 signalling pathways regulate stem cell-fate ……………………………………….. 29 1.13 Schematic illustration of GSK3 dynamic control of NPC homeostasis ………………….. 32 1.14 GSK3 is a key regulator of axon growth and branching …………………………………. 34 1.15 GSK3 activity is important for regulating LTP/LTD during memory formation / reconsolidation …………………………………………………………………………… 42 1.16 Summary of the involvement of GSK3 and monoaminergic signalling in mood disorders 47 1.17 Candidate GSK3 substrates involved in vesicular trafficking ……………………………. 55

3.1 PI4KIIα is a physiological substrate of GSK3 ……………………………………………. 72 3.2 Mapping and comparison of PI4KIIα phosphosites ………………………………………. 73 3.3 PI4KIIα phosphosites are relatively resistant to dephosphorylation by phosphatases ……. 75 3.4 Kinetics of dephosphorylation of PI4KIIα in intact cells …………………………………. 76 3.5 Cdk5 phosphorylates PI4KIIα, priming for subsequent GSK3 phosphorylation …………. 77 3.6 Phosphorylation of PI4KIIα by GSK3 regulates Tfn trafficking in HeLa cells …………... 80 3.7 Phosphorylation of PI4KIIα by GSK3 promotes GluA1 trafficking away from the cell surface …………………………………………………………………………………….. 81 3.8 PI4KIIα does not affect neurite outgrowth in cortical neurons …………………………… 82 3.9 Phosphorylation regulates the abundance of PI4KIIα …………………………………….. 85 3.10 PI4KIIα colocalises with lysosomes ……………………………………………………… 86 3.11 Phosphorylation by GSK3 regulates PI4KIIα binding to AP-3 …………………………... 87 3.12 Phosphorylation promotes PI4KIIα binding to AP-3 and trafficking to the lysosome for degradation ……………………………………………………………………………….. 89

ix |Page

3.13 The N-terminal region of PI4KIIα restricts access of the AP-3 complex to the dileucine motif ……………………………………………………………………………………… 90 3.14 Model of GSK3-mediated regulation of PI4KIIα binding to AP-3 and trafficking to the lysosome ………………………………………………………………………………….. 92 3.15 PI4KIIα heterozygous mutant exhibit normal motor function by negative geotaxis assay 94 3.16 Actogram of PI4KIIα heterozygous mutant and neuron-specific RNAi knockdown flies 94 3.17 Schematic representation of the UAS-GAL4 system …………………………………….. 95 3.18 Proposed model of the relationship between high GSK3 activity, depletion of PI4KIIα and hyperactivity in BD ………………………………………………………………………. 96 3.19 Sleep analysis of PI4KIIα-depleted flies …………………………………………………. 97 3.20 Actogram of PI4KIIα-depleted flies treated with lithium ………………………………… 98 3.21 Sleep analysis following lithium treatment ……………………………………………….. 101 3.22 PI4KIIα depletion has no effect on feeding behaviour or locomotor function …………… 103 3.23 qPCR validation of reduced PI4KIIα gene expression in Drosophila ……………………. 105

4.1 Schematic representation of clathrin-mediated endocytosis depicting the role of AAK1 and AP-2 complexes …………………………………………………………………………... 112 4.2 Schematic representation of the domain structure of the short and long isoform of human AAK1 ……………………………………………………………………………………... 113 4.3 AAK1 is a physiological substrate of GSK3 ……………………………………………... 117 4.4 Sequence alignment of substrates containing a conserved GSK3 phosphorylation consensus sequence and phosphatase-resistant GSK3 sites …………………………………………. 118 4.5 Phosphorylation of AAK1 by GSK3 ……………………………………………………... 119 4.6 Kinetics of dephosphorylation of AAK1 …………………………………………………. 120 4.7 Phosphorylation regulates the abundance of AAK1 ……………………………………… 121 4.8 AAK1 binds to adaptin α and β of the AP-2 complex ……………………………………. 123 4.9 Phosphorylation by GSK3 regulates AAK1 binding to the AP-2 complex ………………. 124 4.10 AAK1-S624A reduces phosphorylation of adaptin µ …………………………………….. 126 4.11 AAK1-S624A has increased kinase activity ……………………………………………… 127 4.12 Phosphorylation of AAK1 by GSK3 is required for efficient trafficking of Tfn in cells .... 130 4.13 AAK1-F698A has reduced binding to adaptin α (AP-2) and reduced stability in the cell ... 131 4.14 AAK1 and its ability to bind AP-2 is critical for Tfn sorting functions in cells ………….. 133 4.15 Model of GSK3-mediated phosphorylation of AAK1 promoting the recycling of Tfn ….. 134 4.16 AAK1 depletion does not affect activity or anhedonia in flies …………………………... 137 4.17 AAK1 depleted flies are susceptible to starvation-induced death ………………………... 139 4.18 HeLa cells expressing shAAK1 knockdown ……………………………………………... 141 4.19 Serum starvation of AAK1-knockdown HeLa cells reduces LC3 staining ………………. 143 4.20 HeLa cells stably expressing AAK1-shRNA knockdown have increased autophagy flux .. 144

x |Page

4.21 Simultaneous RNAi-mediated knockdown of PINK1 and AAK1 in neurons exacerbates PD phenotypes in Drosophila …………………………………………………………….. 146 4.22 Model of GSK3-mediated phosphorylation of AAK1 in autophagosome production and PD pathogenesis ………………………………………………………………………….. 148

5.1 Schematic of the CRISPR/Cas9 system to disrupt or modified the genomic target ……... 169 5.2 Screening other novel GSK3 substrates potentially important in human neurological disease ……………………………………………………………………………...... 171 5.3 Actogram of GSK3 gene candidates - neural-specific RNAi knockdown flies ………….. 174

xi |Page

Abbreviations

AAK1 Adaptin associated kinase-1 AAK1L AAK1 long AAP Amyolid-β precursor protein AAV Adeno-associated viral vectors AD Alzheimer's disease ADBE Activity-dependent bulk endocytosis ALS Amyotrophic lateral sclerosis AMPA α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid AMPK AMP-activated protein kinase ANOVA Analysis of variance AP180 Clathrin coat assembly protein 180 (also known as SNAP91) APC Adenomatosis polyposis coli APC Adenomatous polyposis coli aPKC Atypical protein kinase C APs Adaptin proteins Arg Arginine ATGs Autophagy-related proteins ATP Adenosine triphosphate BAX Bcl2-associated X protein Bcl B-cell lymphoma BD Bipolar disorder BDNF Brain-derived neurotrophic factor BIPPs Bioinformatic prediction of phosphorylated substrate BMAL1 Brain and Muscle ARNT-Like 1 BSA Bovine serum albumin βTrCP Beta-transducin repeats-containing protein CACNA1c Calcium Channel, Voltage-Dependent, L Type, Alpha 1c CAFÉ Capillary Feeder CaMK1 Calmodulin-dependent protein kinase 1 Cas CRISPR associated endonuclease CBD Clathrin-binding domain CBP CREB-binding protein CBR-250 Coomassie Brilliant Blue 250 CCVs Clathrin-coated vesicles CDKs Cyclin-dependent protein kinases cDNA Complementary DNA CK Casein kinases CLASP2 CLIP-associated protein 2 CLOCK Circadian Locomotor Output Cycles Kaput CMV Cytomegalovirus CNV Copy number variation Cos2 Costal 2 CREB cAMP response element-binding protein CRISPR Clustered Regularly Interspaced Short Palindromic Repeats

xii |Page

CRMPs Collapsin response mediator proteins CRY Cryptochrome CT Cycle threshold C-terminal Carboxyl-terminal DA Dopamine DAMs Drosophila activity monitor system DAPI 4',6-Diamidino-2-Phenylindole dcr2 Dicer-2 ddc Dopa decarboxylase DISC Death-inducing complex DISC1 Disrupted in schizophrenia 1 DIV day(s) in vitro DLL Dileucine motif DMEM Dulbecco’s modified Eagle's medium DNA Deoxyribonucleic acid DRG Dorsal root ganglia Dsh Dishevelled DYRKs Dual specificity tyrosine-phosphorylation-regulated kinases E17 Embryonic day 17 EDTA Ethylene diamine tetraacetic acid EEA1 Early endosome antigen 1 EFR3B EFR3 homolog B EGF Epidermal growth factor EGFR Epidermal growth factor receptor EGTA Ethylene glycol tetraacetic acid Eps15 Epidermal growth factor receptor substrate 15 ErbB4 V-Erb-B2 avian erythroblastic leukemia viral oncogene homolog 4 ERK Extracellular signal-regulated kinase ES cells Embryonic stem cells Esrrb Estrogen-related receptor beta F1 First filial generation FADD Fas-associated death domain proteins Fbw7α F-box/WD repeat-containing protein 7 alpha FGF Fibroblast growth factor FTDP-17 Frontotemporal dementia with parkinsonism-17 Fu Fused Fz Frizzled G/C Guanine/Cytosine GABA Gamma-aminobutyric acid GAK Cyclin G-associated kinase GAPDH Glyceraldehyde 3-phosphate dehydrogenase GBA β-glucocerebrosidase GDNF Glial cell line-derived neurotrophic factor GFP Green fluorescent protein GGAs Golgi-associated, gamma adaptin ear containing, ARF binding protein 1 Glu Glutamic acid

xiii |Page

GluA1 Glutamate receptor A1 GMR Glass multiple reporter GPR37 G protein-coupled receptor 37 gRNA Guide-RNA GSK3 Glycogen synthase kinase-3 HA Influenza virus hemagglutinin protein HCN4 Hyperpolarization activated cyclic nucleotide-gated potassium channel 4 HD Huntington’s disease HDR Homology Directed Repair HEK Human embryonic kidney Hes Hairy and enhancer of split HET Heterozygous Hh Hedgehog Hsc70 70-kDa heat shock cognate protein HSF-1 Heat shock factor 1 HSP Hereditary spastic paraplegia HSP70 Heat shock protein 70 IFN- Interferon gamma IGF1 Insulin-like growth factor 1 IL Interleukin IMPase Inositol-monophosphatase IP-3 1,4,5-inositol triphosphate IPPase Inositol polyphosphate 1-phosphatase iPS cell Induced pluripotent stem cell IRS Insulin receptor substrate JAZF1 Juxtaposed with another zinc finger protein 1 kDa Kilodalton Klfs Krüppel-like Factor Lamp1 Lysosome-associated membrane protein 1 LC3-II Microtubule-associated protein light chain 3-II LD Light:dark LDRP Low-density lipoprotein receptor-related protein LEF Lymphoid enhancing factor Leu Leucine LIF Leukemia inhibitory factor LIMP2 Lysosome membrane protein 2 LPS Lipopolysaccharides LRP Low-density-lipoprotein-related protein LRRK2 Leucine-rich repeat kinase 2 LTD Long-term depression LTP Long-term potentiation MAO Monoamine oxidase MAPK Mitogen-activated protein kinase MAPs Microtubule-associated proteins MBP Myelin basic protein MCL-1 Myeloid cell leukemia 1 MCP-1 Monocyte chemoattractant protein 1

xiv |Page

MDD Major depressive disorder MDM2 Mouse double minute 2 Mef2D Myocyte-specific enhancer factor 2D mEPSC Miniature excitatory postsynaptic current MOMP Mitochondrial outer membrane permeabilisation mRNA Messenger ribonucleic acid mTOR Mammalian target of rapamycin Myc V-Myc Avian Myelocytomatosis Viral Oncogene Homolog NFTs Neurofibrillary tangles NF-κB Nuclear factor kappa-light-chain-enhancer of activated B cells NGF Nerve growth factor NHEJ Homologous End Joining NICD Notch intracellular domain NMDA N-methyl-D-aspartate NPCs Neuronal precursor cells NR1D1 Nuclear receptor subfamily 1, group D, member 1 NRG1 Neuregulin 1 nSyb Synaptobrevin N-terminal Amino-terminal Oct4 Octamer-binding transcription factor 4 OE Overexpression PAGE Polyacrylamide gel electrophoresis PBS Phosphate-buffered saline PC pheochromocytoma PCR Polymerase chain reaction PCTK1 PCTAIRE protein kinase 1 PD Parkinson's disease PDGF Platelet-derived growth factor PDK1 3-phosphoinositide-dependent protein kinase 1 PER Period PH pleckstrin homology PI phosphatidylinositol PI3K Phosphatidylinositol 3-kinase PI4KIIα Phosphatidylinositol 4-kinase II alpha PINK1 PTEN-induced putative kinase 1 PIP5K Phosphatidylinositol 4-phosphate 5-kinase PIPP Proline-rich inositol polyphosphate 5-phosphatase PKA Protein kinase A PKB Protein kinase B, also known as Akt PKC Protein kinase C PP1 Protein phosphatase 1 PP2A Protein phosphatase 2 pS/T/Y Phospho-serine/threonine/tyrosine PSD-95 Postsynaptic density protein 95 P-site Phosphotransfer site Ptc Patched

xv |Page

PTEN Phosphatase and tensin homolog PUMA p53 upregulated modulator of apoptosis qPCR Quantitative PCR Rabs Ras-related proteins RBP-J Recombination signal binding protein for immunoglobulin kappa J region RNA Ribonucleic acid RNAi RNA interference RSK Ribosomal s6 kinase RTKs Receptor tyrosine kinases S.E.M. Standard Error of the Mean 5-HT Seotonin (5-hydroxytryptamine) SCF Skp/Cullin,/F-box protein SCN Suprachiasmatic nucleus SDS Sodium dodecyl sulfate Ser/Thr Serine/Threonine SGG Shaggy (Drosophila homologue of GSK3) SGIP1 SH3-containing GRB2-like protein 3-interacting protein 1 shRNA Short hairpin RNA siRNA Small interfering RNA Smo Smoothened SNARE SNAP (soluble NSF attachment protein) receptor SNP Single nucleotide polymorphism SOD1 Superoxide dismutase 1 Sox2 SRY (sex determining region Y) box 2 SSRIs Selective serotonin reuptake inhbitors STAT Signal transducer and activator of transcription Sufu Suppressor of fused TAG Triacyglyceride TBI Traumatic brain injury TCF T-cell factor Tfn Transferrin TfnR Transferrin receptor TGN Trans-Golgi network TIM Timeless TIP60 HIV-1 Tat interactive protein 60 T-loop Telomere loop TLRs Toll-like receptors TNF Tumour necrosis factor TPH2 Tryptophan hydroxylase 2 Tris-HCL Tris-Hydrochloride TrkB Tropomyosin receptor kinase B TSC2 Tuberous sclerosis complex subunit 2 TTBK2 Tau tubulin kinase 2 TTBS Tris-Buffered Saline and Tween 20 Tyr Tyrosine UAS Upstream Activation Sequence

xvi |Page

ULK1 Unc-51 like autophagy activating kinase 1 UNSW The University of New South Wales UTR Untranslated region VAMP3 Vesicle-associated membrane protein 3 VDAC1 Voltage-dependent anion channel 1 VEGF Vascular endothelial growth factor VSVG Vesicular stomatitis virus G protein ZEB Zinc finger E-box-binding homeobox ZT Zeitgeber time

xvii |Page

CHAPTER 1 – Introduction and literature review

Glycogen synthase kinase 3 (GSK3) is a serine/threonine (Ser/Thr) protein kinase that is expressed ubiquitously throughout the body (Leroy & Brion, 1999a; Takahashi et al., 1994; Woodgett, 1990a). It was first identified as an enzyme capable of phosphorylating and inactivating the enzyme glycogen synthase, the final enzyme in glycogen biosynthesis (Embi et al., 1980). Shortly afterwards, it was separately identified as a kinase that phosphorylates the brain cytoskeletal protein Tau (Ishiguro et al., 1993), which was later shown to contribute to the formation of neurofibrillary tangles, a characteristic lesion of Alzheimer’s Disease (AD) (Jackson et al., 2002; Lucas et al., 2001; Spillantini & Goedert, 1998). Since then, it has been shown to have additional roles in numerous physiological functions including brain function, oncogenesis, embryonic development, apoptosis, immunity, cell cycle regulation and proliferation. These functions are mediated by phosphorylation of a wide range of substrates, including metabolic enzymes, signalling proteins, transcription factors and cytoskeletal proteins (Cole, 2012; Medina & Wandosell, 2011; Sutherland, 2011). Deregulation of some of these GSK3-assoicated signalling pathways have been implicated in the development of human diseases, including diabetes, AD, mood disorders and cancer. Therefore, GSK3 is an important enzyme to study in order to fully understand basic signalling mechanisms underlying several physiological processes and human diseases. 1.1 Characterisation of GSK3

1.1.1 GSK3 genes

GSK3 is conserved in all eukaryotes, including plants, fungi and animals (Itoh et al., 1995; Kim et al., 1999; Ruel et al., 1993). In mammals, there are three GSK3 isoforms encoded by two distinct genes; GSK3α located on chromosome 19q13.2 and comprising of 12 exons, and GSK3β located on chromosome 3q13.3 comprising of 11 exons. GSK3 activity can be modulated at the transcriptional level; in particular the promoter regions of both gene contain several CCAAT boxes (Lau et al., 1999b) that are frequently found in promoters of metabolic enzymes and are important to control tissue-specific and inducible expression during development. The GSK3 promoter was shown to have differential activity depending on cell lines used. For example, SH-SY5Y neuronal cell lines display higher promoter driven expression than COS-7 kidney cell lines (Lau et al.,

1 |Page

Sec. 1.1. Characterisation of GSK3

1999b). However, GSK3 mRNA levels can vary significantly from its protein levels, suggesting that GSK3 is also controlled at the post-transcriptional or translational level (Lau et al., 1999a; Schaffer et al., 2003). For example, studies have shown that several features in the 5’ UTR of GSK3, such as longer length and a high G/C content, are associated with secondary structure that can affect gene translation (Schaffer et al., 2003). Therefore, changes in transcription and translation influence GSK3 abundance and function temporally and spatially in different tissues compartments (Doble & Woodgett, 2003).

1.1.2 GSK3 proteins

The two GSK3 genes encode proteins of molecular weights of 51 kDa (GSK3α) and 47 kDa (GSK3β), respectively. These isoforms are highly homologous within their kinase catalytic domains (98% identical), however differ significantly from one another outside this region. The major difference in size is due to an extended 63 residue glycine- rich N-terminal tail in GSK3α (Woodgett, 1990b) (Fig.1.1 – Domain structure of GSK3 isoforms). An alternatively spliced isoform of GSK3β expressed exclusively in the nervous system (GSK3β2) has a 13-residue insert on the activation loop of the kinase domain (Mukai et al., 2002). The difference in function and activity of this isoform is only beginning to be investigated, but seems to be largely similar to the major GSK3β isoform (e.g. Soutar et al., 2011).

Gly-rich Kinase domain

α 51 kDa ∧ ∧ S21 Y279 β 47 kDa ∧ ∧ Y216 S9

Figure. 1.1. Schematic representation of mammalian GSK3 and GSK3. The conserved kinase domain shared by both isoforms and the glycine-rich N-terminal domain unique to GSK3 are highlighted.

2 |Page

Sec. 1.1. Characterisation of GSK3

1.1.3 GSK3 expression

GSK3 is expressed at relatively high levels in the cytoplasm, nucleus and mitochondria of most cell types throughout the body (if not all). See reviews: (Doble & Woodgett, 2003; Medina & Wandosell, 2011). Its transcript and protein levels are generally higher in fetal tissues than in corresponding adult tissues, although expression remains high throughout adulthood (Lau et al., 1999a). For example in rat brains, GSK3 expression is high just before birth, peaks at 8 days postpartum then decreases thereafter (Takahashi et al., 1994). Although GSK3α and GSK3β are expressed from two separate genes, their expression patterns are usually overlapping, with only minor differences being reported. For example, in the brain, GSK3α is abundant in cerebellar glomeruli of rat brains, whereas GSK3β is more abundant in the Purkinje cells (Takahashi et al., 1994). In neurons, all three GSK3 isoforms have been detected in neurites and growth cones as well as the cell body and nucleus. GSK3α and GSK3β are also expressed in glial cells, but not the alternatively spliced GSK32 isoform (Wood-Kaczmar et al., 2009). Overall GSK3 and GSK3 are widely expressed with largely overlapping expression patterns, although some differences between isoforms suggest there is potential for GSK3α and GSK3β to have different physiological functions in some cell types and/or tissues that remain to be identified.

1.1.4 GSK3 kinase activity

GSK3 is a member of the CMGC family of protein kinases, which is comprised of mostly proline directed Ser/Thr kinases (49 out of 51). Accordingly, GSK3 is also a Ser/Thr kinase with a preference for pro-directed sites, although it can efficiently phosphorylate non-proline-directed Ser/Thr sites as well. Phylogenetically, GSK3 is most closely related to the cyclin-dependent protein kinases (CDKs), such as CDK1 and CDK2. The crystal structure of GSK3 provides insight into both its regulation and its affinity towards its substrates. GSK3 has the typical two-domain kinase fold (Hanks & Quinn, 1991) with a -strand domain (residues 25-138) at the N-terminal end and an - helical domain (residues 139-343) at the C-terminal end. The ATP-binding site is at the interface of the α-helical and β-strand domain and the activation loop (residues 200-226) runs along the surface of the substrate-binding groove. GSK3 contains two conserved residues that function to preserve the catalytic activity of GSK3. Arg96 ensures alignment of the α-helical and β-strand domains and Glu97 residue located in the active site mediates

3 |Page

Sec. 1.1. Characterisation of GSK3

catalysis (ter Haar et al., 2001). Like many of it closely related kinases (e.g. CDKs), GSK3 requires phosphorylation of tyrosine residues Tyr279 (GSK3α) or Tyr216 (GSK3β) in its activation loop (T-loop) as a prerequisite for activity. GSK3 auto-phosphorylates soon after protein translation and during folding to allow the activation-loop tyrosine to be phosphorylated (Lochhead et al., 2006). Phosphorylation of these tyrosine residues are necessary for opening of the substrate-binding groove to facilitate substrate binding and phosphorylation (Dajani et al., 2001; ter Haar et al., 2001). Protein kinases related to GSK3, such as CDK2 and extracellular signal-regulated kinase 2 (ERK2), have an equivalent phosphothreonine in their activation loop.

Of the 500 kinases in the human kinome, GSK3 is one of the most unusual for three main reasons:

Firstly, the substrate specificity of GSK3 is unique in that it requires a prior ‘priming’ phosphate (phosphorylated Ser or Thr residue) located 4 residues C-terminal to the GSK3 target site. This must be phosphorylated by another kinase before the substrate can be efficiently recognised and phosphorylated by GSK3 (Frame et al., 2001; Kennelly & Krebs, 1991). For example, glycogen synthase must first be phosphorylated by casein kinase 2 (CK2) at Ser656 before it can be efficiently phosphorylated by GSK3 at Ser652/648/644/640 (Frame et al., 2001) (Fig.1.2). The priming phospho-serine (pS) or threonine (pT) residues aligns the two domains of GSK3 for optimal catalytic activity (ter Haar et al., 2001). This priming mechanism is rare amongst kinases, with casein kinase 1 (CK1; pS/T at P-3 or 4) (Flotow et al., 1990) and Tau tubulin kinase 2 (TTBK2; pY at P+2 position) (Michale et al., 2011) being the only other kinases known to use a phospho- priming site (although their priming phosphosites are in different positions to that of GSK3 substrates).

4 |Page

Sec. 1.1. Characterisation of GSK3

Figure 1.2. GSK3 substrate specificity. Phosphorylation of Glycogen synthase at Ser656 by CK2 ‘primes’ for GSK3 to phosphorylate Ser652. This in turn primes for subsequent phosphorylation at Ser648 and so on, until five serine residues become phosphorylated. (Figure adapted from (Cohen & Frame, 2001)).

Secondly, unlike most other kinases, GSK3 is highly active under basal conditions. This is in part due to constitutive auto-phosphorylation of conserved tyrosine residues on the activation loop of the kinase domain (Tyr279 on GSK3α, Tyr216 on GSK3β). Phosphorylation of these residues is equivalent to that of the phosphotyrosine residue that is critical for mitogen-activated protein kinase (MAPK) activity (Hughes et al., 1993). GSK3 auto-phosphorylates soon after protein translation and during folding to allow the activation-loop tyrosine to be phosphorylated (Lochhead et al., 2006). Phosphorylation of these residues is absolutely necessary for GSK3 kinase activity and is most likely not regulatable. (Cole et al., 2004a; Lochhead et al., 2006).

Thirdly, activation of upstream signalling pathways in response to certain cellular signals reduces GSK3 activity, rather than increasing it as found for most other kinases. Furthermore, GSK3 activity is an important regulator of numerous signal transduction pathways. Recent gene expression profiling studies have revealed the importance of GSK3 activity on gene expression patterns including target genes of phosphatidylinositol 3-kinase (PI3K) and Wnt/β-catenin signalling pathways (Bartman et al., 2014). Signalling pathways that reduce GSK3 activity are discussed below.

5 |Page

Sec. 1.2. Signalling pathways regulating GSK3 activity

1.2 Signalling pathways regulating GSK3 activity

1.2.1 The growth factor signalling pathway

The PI3K signalling pathway is a major regulator of GSK3 activity in cells (Brazil & Hemmings, 2001; Brunet et al., 2001; Datta et al., 1999) (Fig.1.3). This pathway is commonly activated by growth factors, such as insulin-like growth factor-1 (IGF-1), brain-derived neurotrophic factor (BDNF), and insulin for controlling the growth, proliferation and/or survival of cells. Ligand binding to receptors at the plasma membrane (e.g. insulin binding to the insulin receptor) induces conformational changes in the receptors, resulting in auto-phosphorylation of a number of key tyrosine residues in its cytoplasmic domain. This induces recruitment and activation of the insulin receptor substrate (IRS) adaptor proteins. IRS activation stimulates recruitment of PI3K to the plasma membrane where it catalyses phosphorylation of phosphatidylinositol lipid

PI(4,5)P2 to generate PI(3,4,5)P3. The elevated abundance of PIP3 induces recruitment of protein kinase B (PKB/Akt) to the plasma membrane via its pleckstrin homology (PH) domain binding to PIP3. There, it is phosphorylated at two key residues by PDK1 (3- phosphoinositide-dependent protein kinase 1) (Thr308 and Ser473) which activates its kinase activity. Activated Akt subsequently phosphorylates GSK3 (Ser21 for GSK3, Ser9 for GSK3), reducing its kinase activity. The phosphorylated N-terminus acts like a pseudo-substrate, which competes with the priming phosphate for substrate binding. (Cross et al., 1995; Sutherland et al., 1993). This impairs GSK3’s ability to bind and phosphorylate its substrates. In addition to Akt, inhibitory phosphorylation at these sites can also be mediated by other members of the AGC family of kinases, including protein kinase A (PKA) (Fang et al., 2000), protein kinase C (PKC) (Goode et al., 1992) and ribosomal s6 kinases, p70RSK (Armstrong et al., 2001) and p90RSK (Saito et al., 1994; Shaw & Cohen, 1999).

6 |Page

Sec. 1.2. Signalling pathways regulating GSK3 activity

Figure 1.3. Schematic illustration of the PI3K/Akt signalling pathway regulating GSK3 activity and influencing glycogen synthesis. Growth factor binding to its receptor activates intrinsic protein tyrosine kinase activity and induces the recruitment and activation of IRS. This stimulates further recruitment of PI3K to the plasma membrane where it catalyses the phosphorylation of

PIP2 to generate PIP3. This induces recruitment and activation of PDK1, which in turn phosphorylates and activates Akt/PKB. Akt/PKB, in turn, phosphorylate and inhibits GSK3, resulting in the dephosphorylation of substrates of GSK3, including glycogen synthesis. (Figure adapted from (Cohen & Frame, 2001)).

RSK’s are Ser/Thr kinases that are activated by the MAPK cascade. Growth factors such as epidermal growth factor (EGF) and platelet-derived growth factor (PDGF) activate the MAPK cascade and stimulate GSK3 phosphorylation by p90RSK (Saito et al., 1994; Shaw & Cohen, 1999). GSK3 has been shown in vitro to be phosphorylated by p70RSK in response to stimulation by amino acids in human myocytes (Armstrong et al., 2001). These provide other routes for inhibition of GSK3 by growth factors and other signals that activate the MAPK pathway.

7 |Page

Sec. 1.2. Signalling pathways regulating GSK3 activity

MAPK signalling has also been reported to regulate the phosphorylation of two other regulatory sites on GSK3β that are claimed to decrease GSK3 catalytic activity. Thr43 was reported to be phosphorylated by Erk (Ding et al., 2005), while Ser389/Thr390 was reported to be phosphorylated by p38 MAPK (Thornton et al., 2008). However these studies have not been replicated and require further evidence to prove the functional relevance of phosphorylation at these sites.

1.2.2 The Wnt pathway

GSK3 has a central role in regulating the Canonical Wnt signalling pathway (Fig.1.4), which influences cell growth, differentiation, migration and survival of cells. This is independent of the GSK3-isoform, since both GSK3 and GSK3 participate and are function redundant (Doble et al., 2007). The Wnts are a family of secreted, cysteine- rich glycoprotein ligands with at least 19 different isoforms existing in mammals (Miller, 2002). In unstimulated cells, GSK3 forms part of the cytoplasmic β-catenin destruction complex. Axin acts as a scaffold, as it directly interacts with all other key components of the complex; β-catenin, the tumor suppressor protein APC (adenomatosis polyposis coli), and the kinases GSK3 and Caseine kinase 1 (CK1; γ and  isoforms) (Davidson et al., 2005; Zeng et al., 2005). This multi-protein complex facilitates sequential phosphorylation of -catenin at Ser45 by CK1 (Hagen et al., 2002; Liu et al., 2002; Sakanaka, 2002) followed by phosphorylation of Ser33/37/Thr41 by GSK3 (Hagen et al., 2002; Hagen & Vidal-Puig, 2002). Phosphorylated β-catenin is recognised by βTrCP, a substrate-targeting subunit of the SCF (Skp1/Cull/F-box protein) E3 ubiquitin ligase, targeting it for rapid destruction by the proteasome (Aberle et al., 1997). However, upon Wnt ligand binding and activation of the Frizzled (Fz)/LRP-5/6 (low-density lipoprotein receptor-related protein 5/6) coreceptor complex, a cytoplasmic protein called Dishevelled (Dsh) mediates the dissociation of the destruction complex. This prevents β- catenin phosphorylation by CK1/GSK3, thus evading ubiquitylation-depended destruction (See reviews: (Hur & Zhou, 2010; Song et al., 2014)). As a consequence β- catenin accumulates in the cell cytoplasm, then translocates to the nucleus, where it binds to members of the TCF/LEF (T-cell factor/lymphoid enhancing factor) family of transcription factors. This promotes the transcription of genes that regulate cell fate (Behrens et al., 1996; Huber et al., 1996; Kühl & Wedlich, 1997), such as c-Myc (He et al., 1998), NeuroD1 (Kuwabara et al., 2009), cyclin D1 (Shtutman et al., 1999; Tetsu & McCormick, 1999).

8 |Page

Sec. 1.2. Signalling pathways regulating GSK3 activity

Figure 1.4. Schematic illustration of the Wnt/β-catenin signalling pathway. In the absence of Wnt, GSK3 forms part of the cytoplasmic β-catenin destruction complex. Axin acts as the scaffold as it directly interacts with all other components of the complex- β-catenin, APC, and kinases, GSK3 and CK1. This multi-protein complex facilitates sequential phosphorylation, tagging - catenin for proteasomal degradation. Wnt promotes the association of Fz, LRP and Dsh. This induces translocation of the destruction complex to the plasma membrane where GSK3 and CK1 mediate phosphorylation of the cytoplasmic tail of LRP. This causes release of docked Axin and dissociation of the destruction complex. This allows β-catenin to accumulate and localize to the nucleus and subsequently initiate TCF-mediated gene transcription. (Figure adapted from (Hur & Zhou, 2010)).

9 |Page

Sec. 1.2. Signalling pathways regulating GSK3 activity

1.2.3 The Hedgehog pathway

The Hedgehog (Hh) pathway is important for regulating cell fate and patterning of embryonic structures during development, as well as proliferation of adult neural stem cell (Lai et al., 2003). In mammals, the Gli zinc finger proteins (Gli1, Gli2 and Gli3) are transcription factors that transduce Hh signalling. In the absence of Hh ligands, Gli proteins associated with a scaffolding complex containing Costal 2 (Cos2), the Ser/Thr kinase Fused (Fu) and Suppressor of fused (Sufu). This complex facilitates phosphorylation of Gli by a cascade of GSK3, CK1 and PKA phosphorylation, where PKA phosphorylates priming sites for subsequent phosphorylation by GSK3/CK1 (Tempé et al., 2006). This promotes direct binding and ubiquitylation by the ubiquitin ligase complex SCFβTrCP for processing and proteolysis, generating a truncated repressor lacking the C-terminal activation domains (Jia et al., 2002; Pan et al., 2006; Tempé et al., 2006). In the presence of Hh signalling, ligand binding to its transmembrane receptor Patched (Ptc), disrupts this complex via activation of the Smoothened (Smo) transmembrane protein (see review: (Villavicencio et al., 2000)). This inhibits phosphorylation and processing of Gli, leading to nuclear translocation and transcriptional activation (Fig.1.5). Overall, constitutive activity of GSK3 is require to antagonise Hedgehog signalling by mediating proteolysis and processing of Gli proteins.

10 |Page

Sec. 1.2. Signalling pathways regulating GSK3 activity

Figure 1.5. Schematic illustration of the Hedgehog signalling pathway. In the absence of Hh ligands, Gli proteins associated with a scaffolding complex containing Cos2, Fu and Sufu. This multi-protein complex facilitates sequential phosphorylation by kinases GSK3, CK1 and PKA, promoting cleavage of Gli and generating a truncated repressor lacking the C-terminal activation domains. In the presence of Hh signalling, ligand binding to its transmembrane receptor Ptc, disrupts this complex via activation of the Smo transmembrane protein. This inhibits phosphorylation and processing of Gli, leading to nuclear translocation and transcriptional activation. (Figure adapted from (Ingham & Placzek, 2006)).

11 |Page

Sec. 1.2. Signalling pathways regulating GSK3 activity

1.2.4 The Notch pathway

The Notch signalling pathway promotes proliferation and halts cell differentiation signalling during embryogenesis and has more recently been implicated in adult neurogenesis and memory processing (see review: (Alberi et al., 2013)). As such, Notch1 expression is high in embryonic mouse and human brain and much lower in adult brains (Berezovska et al., 1998). Notch signalling also promotes neurite outgrowth (Redmond et al., 2000) and development of the vasculature (Morrow et al., 2008). The activation of the Notch family of transmembrane proteins triggers proteolytic cleavage and induces the release of Notch intracellular domain (NICD) for translocation to the nucleus. Here it binds the transcription factor RBP-J, inducing transcription of factors that inhibit differentiation and enhance proliferation (see reviews: (Baron, 2003; Kopan & Ilagan, 2009; Mumm & Kopan, 2000)). GSK3β has been shown to bind and phosphorylate the NICD, attenuating transcriptional activation of Notch target genes (Espinosa et al., 2003). Meanwhile, there are conflicting reports on the role of GSK3β in regulating Notch transcriptional activity, in which inhibition of GSK3 using pharmacological inhibitor or genetic tools decrease NICD levels in neuroblastoma cells (Foltz et al., 2002) and vascular smooth muscle cells (Guha et al., 2011). While a consensus has not yet been meet, the majority of studies agree on the translocation and enrichment of NICD levels in the nucleus as required for the transcriptional activation of Notch signalling following GSK3 inhibition (Espinosa et al., 2003; Jin et al., 2009; Kim et al., 2009b; Shimizu et al., 2008). Although, it is possible that the regulatory effect of GSK3 phosphorylation on Notch signalling may be context- and cell type-specific.

Overall, activation of particular signalling pathways selectively affects different subsets of GSK3 substrates in determining cell fate. Ligand-mediated activation of the Wnt, Hedgehog and Notch pathways reduces phosphorylation of key effector substrates (i.e. β-catenin for Wnt, Gli for Hedgehog and NICD for Notch), disrupting their respective signalling/scaffolding complexes without directly inhibiting GSK3 activity. In contrast, growth factor signalling directly inhibits GSK3 kinase activity via phosphorylation of the N-terminal serine residues reducing phosphorylation of a wider range of GSK3 substrates (Fig.1.6).

12 |Page

Sec. 1.2. Signalling pathways regulating GSK3 activity

Figure 1.6. Signalling pathways regulating cell fate target different subsets of GSK3 substrates. Ligand-mediated activation of the Wnt, Hedgehog and Notch pathways prevents phosphorylation of a single substrate by disrupting multi-subunit signalling/scaffolding complexes required to mediate their phosphorylation. In contrast, growth factor-mediated inhibition of GSK3 via phosphorylation of N-terminal serine residues reduces phosphorylation of a wider range of GSK3 substrates. (Figure adapted from (Cole, 2013b)).

13 |Page

Sec. 1.3. Physiology of GSK3 mutant mice

1.3 Physiology of GSK3 mutant mice

Genetically modified mice provide a powerful means to study and understand the physiological function of GSK3. These mice have been generated using a variety of genetic approaches, including conventional knockouts and knockins, conditional knockouts (tissue-specific) and transgenic (over-expression mice).

1.3.1 GSK3 mutant mice

GSK3 homozygous knockout is lethal in mice (Hoeflich et al., 2000). These animals die late in development from hepatic apoptosis. This suggests there may be a GSK3β isoform-specific role for preventing liver cell apoptosis during development that cannot be compensated for by GSK3α, although the precise mechanisms are yet to be determined. In contrast, GSK3 heterozygous mice are viable and morphologically relatively normal (Hoeflich et al., 2000). However, they do exhibit a number of neurological phenotypes, exemplifying the important role of GSK3 in the brain. These include reduced exploratory behaviour, reduced aggressive behaviour and increased anxiety (Beaulieu et al., 2008; Bersudsky et al., 2008). This overall anti-depressive-like state is similar to control mice treated with lithium, suggesting that reducing GSK3 activity has a sedating affect (O'Brien et al., 2004). Conversely, transgenic mice overexpressing constitutively active GSK3β (S9A) exhibit locomotor hyperactivity and an associative manic phenotype (Prickaerts et al., 2006). These mice are a useful model for studying hyperactivity and hyper-reactivity behaviours that are similar to those observed in the manic phase of Bipolar Disorder (BD) patients.

More recently GSK3 has been implicated in the formation of long-term memories. GSK3β heterozygous mice were shown to have retrograde amnesia, impaired associative memory and impaired memory reconsolidation (Kaidanovich-Beilin et al., 2009; Kimura et al., 2008), while transgenic mice overexpressing GSK3β display impaired spatial learning (Hernández et al., 2002). They also have increased Tau phosphorylation, apoptotic neuronal death and reactive astrocytosis, consistent with a role for GSK3 in the development of AD (Lucas et al., 2001). Memory dysfunction is a characteristic symptom of AD and GSK3 is known to be a key regulator of long-term potentiation (LTP) / long- term depression (LTD) which underlies memory formation (discussed further below). Therefore, GSK3 may contribute and/or be an important therapeutic target for memory dysfunction during normal aging and in neurodegenerative conditions.

14 |Page

Sec. 1.3. Physiology of GSK3 mutant mice

1.3.2 GSK3 mutant mice

In contrast to GSK3 null mice, animals completely lacking GSK3 are viable and initially appear morphological normal (MacAulay et al., 2007). However, they exhibit decreased exploratory activity and reduced aggressive behaviours similar to GSK3 HET mice further supporting the sedating effect of reduced GSK3 activity. They also exhibit decreased locomotor activity and coordination, increased sensitivity to environmental cues, impaired associative memory and decreased number of purkinje cells in the cerebellum (Kaidanovich-Beilin et al., 2009). Furthermore, GSK3 null mice have accelerated age-related pathology including, disruption and degradation of cardiac and skeletal muscle, and the development of associated cardiac hypertrophy and contractile dysfunction. These are classic characteristics of aging (Zhou et al., 2013). These mice also developed large tubular aggregates in skeletal muscle, consistent with impaired clearance of insoluble cellular debris. It was discovered that mTOR, a key enzyme involved in stimulating autophagy, was unrestrained in the absence of GSK3 regulation, and resulted in the dysregulation of autophagy (Zhou et al., 2013). Therefore, it is possible that knockdown of GSK3 may indicate impairment in learning and memory acquisition and accelerated age-related pathologies as a result of impaired autophagy.

Meanwhile, mice overexpressing constitutively active GSK3 (S21A knockin mice) exhibit locomotor hyperactivity, heightened response to new environmental cues, mild anxiety, impaired social behaviour and increased curiosity (Ackermann et al., 2010; Polter et al., 2010). This suggests that consistent with GSK3β, GSK3α also plays a role in regulating mood functions.

In summary, increased or decreased expression of GSK3 in mice predominantly causes neurological and behavioural phenotypes. Impaired memory acquisition and/or memory reconsolidation may be associated with GSK3’s role in regulating neurotransmission (see section 1.4.6), while the mechanism underlying its role in mood regulation is less clear, but consistent with the sedating/anti-depressant effects of GSK3 inhibitors. Furthermore, the use of GSK3 knockout mice may enable us to predict the long-term effect of the usage of GSK3 inhibitors such as lithium treatment in patients. The comparative analysis of different GSK3 mutant animals provides a powerful means to study and understand GSK3 function in a physiological system.

15 |Page

Sec. 1.4. GSK3 substrates

1.4 GSK3 substrates

GSK3 mediates its functional effects via phosphorylation of a specific subset of substrates. This is the primary function of this kinase. Over 70 substrates have so far been identified for GSK3, including a number of cytoskeletal, signalling, and DNA-binding proteins. Below we discuss groups/examples of substrates phosphorylated by GSK3 and their roles in a number of important physiological processes.

1.4.1 GSK3 and its substrates involved in Apoptosis

GSK3 has an unusual role in the regulation of cell death as it has been shown to both promote and inhibit apoptosis. In particular, GSK3 is reported to promote the mitochondrial intrinsic apoptotic pathway (cell death), yet inhibit receptor-mediated extrinsic apoptotic signalling (survival) (Beurel & Jope, 2006). These seemingly opposing functions can largely be explained by the cell type and developmental stage of the cell/organism, as discussed below.

1.4.1.1 The role of GSK3 in pro-apoptotic signalling

It was first shown that overexpression of GSK3 in pheochromocytoma (PC12) and Rat-1 fibroblast cell lines was sufficient to induce apoptosis (Pap & Cooper, 1998). Overexpression of a catalytically inactive mutant of GSK3 on the other hand protected cells from apoptosis, indicating that the kinase activity of GSK3 was required for cell death. Since then, numerous studies have shown that pharmacological inhibitors of GSK3 (e.g. lithium) protect many cell types from a wide range of stressful stimuli, including heat shock, DNA damage (Watcharasit et al., 2002), hypoxia (Loberg et al., 2002), endoplasmic reticulum stress (Song et al., 2002), Traumatic brain injury (TBI) (Dash et al., 2011; Yu et al., 2012) and polyglutamine toxicity associated with Huntington Disease (Carmichael et al., 2002).

In particular, primary cerebellar, cortical and hippocampal neurons are protected against glutamate-induced excitotoxicity by pre-treatment with therapeutic concentrations of lithium (Hashimoto et al., 2002; Nonaka et al., 1998). Treatment with GSK3 inhibitors also stimulates expression of cell survival factors, such as BDNF/tropomyosin receptor kinase B (TrkB), anti-apoptotic B-cell lymphoma 2 (Bcl-2) and heat shock protein 70 (HSP70) (Chuang et al., 2011).

16 |Page

Sec. 1.4. GSK3 substrates

Similar neuroprotective effects have been observed in vivo, whereby pretreatment with lithium reduces neurological deficits and decreases the size of the damaged brain region after ischemic shock in rats (Nonaka & Chuang, 1998). In another study using controlled cortical impact to simulate traumatic brain injury, early lithium treatment post- injury increased GSK3 Ser9 phosphorylation and reduced neuronal loss in the hippocampus, as well as decreased hippocampal-dependent deficits in learning and memory (Dash et al., 2011; Yu et al., 2012). These observations demonstrate the neuroprotective role of GSK3 inhibition in vivo, and suggest that pharmacological inhibitors of GSK3 could be used therapeutically as neuroprotective agents. These could be especially beneficial for maintaining neuron survival in the immediate hours/days following stroke, when secondary death of neurons in the penumbral region exacerbates brain damage and reduces functional recovery in patients. Indeed, lithium treatment has been shown to be highly neuroprotective in preventing neuronal apoptosis in several rat/mouse models of stroke/brain injury (For reviews: (Chuang, 2004; Chuang et al., 2011; Leeds et al., 2014).

Mitochondrial apoptotic signalling

The PI3K signalling pathway, causing inhibition of GSK3, antagonises mitochondrial apoptotic signalling by preventing mitochondrial release of cytochrome c and blocking subsequent activation of the caspase cascade (Pap and Cooper, 1998), thus preventing apoptosis (Watcharasit et al., 2002).

This is largely achieved by targeting several members of the BCL-2 family of proteins. GSK3β has been shown to directly phosphorylate the pro-apoptotic Bcl2- associated X (BAX) protein that promotes mitochondrial outer membrane permeabilisation (MOMP), a critical event in triggering apoptosis (see review: (David, 2012)). Phosphorylation at Ser163 promotes translocation of Bax to the mitochondrial outer membrane where it forms a pore to release cytochrome c, inducing apoptosis (Linseman et al., 2004; Wei et al., 2001). However, activation of the PI3K pathway reduces this, helping to protect cells. VDAC1 is a voltage-dependent anion channel that also mediates cytochrome c release from mitochondria. It is also phosphorylated by GSK3, although the affect this has on VDAC1 function and cytochrome c release is not yet clear (Pastorino et al., 2005). Conversely, GSK3 phosphorylates several proteins that antagonise BAX and cytochrome c release from mitochondria, including MCL-1, BCL-

17 |Page

Sec. 1.4. GSK3 substrates

2 and BCL-3 (Cheng et al., 2001; Viatour et al., 2004) (Fig.1.7). For example, interleukin (IL)-3 signalling through the PI3K pathway protects against apoptosis by promoting the stability of MCL-1. GSK3 normally phosphorylates MCL-1 at Ser155 and Ser159, which targets it for degradation by the ubiquitin-proteasome pathway, facilitating cytochrome c release and apoptosis. However, activation of the PI3K pathway by IL-3 in cells reduces GSK3 activity, resulting in stabilisation of MCL-1 (Maurer et al., 2006). In summary, activation of the PI3K signalling pathways and inhibition of GSK3 reduces apoptosis by promoting the stability of the BAX antagonists MCL-1, BCL-2, and BCL-3.

Figure 1.7. GSK3 promotes mitochondrial apoptosis signalling. GSK3 promotes apoptosis by phosphorylates several proteins that antagonise pro-apoptotic BAX and cytochrome c release from mitochondria, including MCL-1, BCL-2 and BCL-3.

Apoptotic transcription factors

Interestingly, apoptotic stimuli have been shown to induce nuclear accumulation of GSK3β, facilitating increased interactions with nuclear substrates early in the apoptotic process and preceding activation of caspases (Bijur & Jope, 2000). Some reported nuclear targets include the prosurvival transcription factors, cAMP response element-binding protein (CREB), heat shock factor-1 (HSF-1) and -catenin (Bijur & Jope, 2000; Grimes & Jope, 2001b). For example, in the canonical Wnt signalling pathway, over-expression of -catenin and its transcriptional activation reduces apoptosis similar to the protection provided by treatment with GSK3 inhibitors (Chen et al., 2001; Yuan et al., 2005). This

18 |Page

Sec. 1.4. GSK3 substrates

suggests that the action of GSK3 on -catenin prevents the initiation of cell pro-survival transcription and instead promotes apoptosis.

GSK3 activity also simultaneously activates proapoptotic transcription factors. Of particular interest is its ability to directly phosphorylate the proapoptotic transcription factor and tumour suppressor protein p53 (Pluquet et al., 2005; Turenne & Price, 2001; Watcharasit et al., 2003). Phosphorylation by GSK3 is reported to promote p53 nuclear localisation and transcriptional activation of proapoptotic genes, such as BAX, PUMA (p53 upregulated modulator of apoptosis) and Noxa (Kruse & Gu, 2009). GSK3 also regulates p53 activity indirectly by phosphorylating the p53-specific E3 ubiquitin ligase, mouse double minute 2 (MDM2) (Kulikov et al., 2005). GSK3 phosphorylation of MDM2 disrupts the ubiquitylation of p53, preventing its degradation by the proteasome (Thotala et al., 2011). The cooperative interactions and activation of GSK3 with p53 and MDM2 results in up-regulation of pro-apoptotic genes that contribute to intrinsic apoptotic signalling in the cell.

Conversely, GSK3 has been reported to have an inhibitory role on p53, whereby GSK3 mediated phosphorylation activates MDM2 leading to p53 degradation and cell survival (Kulikov et al., 2005). While the majority of studies suggest a proapoptotic role for GSK3 in p53 signalling, this may depend on cell type and/or circumstance, which requires greater understanding.

Overall, numerous studies have clearly demonstrated a proapoptotic role for GSK3 in many cell types and tissues, with the upstream growth factor signalling pathway antagonising its action to promote survival.

1.4.1.2 The role of GSK3 in cell survival signalling

Despite ample evidence for a substantial proapoptotic role of GSK3, it also appears to perform pro-survival roles. For example, the viability of mouse neuroblastoma cells (Neuro-2a) requires the activity of GSK3, since inhibition of GSK3 using the inhibitor SB415286 promotes apoptosis (Dickey et al., 2011). Importantly, deletion of GSK3β in knockout mice is lethal due to massive hepatocyte apoptosis during development (Hoeflich et al., 2000). This clearly demonstrates a pro-survival role for GSK3β activity, at least for hepatocyte apoptosis during embryonic development.

19 |Page

Sec. 1.4. GSK3 substrates

GSK3 negatively regulates extrinsic apoptotic signalling

The protective role of GSK3 in developing hepatocytes is likely associated with downstream of the tumour necrosis factor (TNF) signalling pathway and its target transcription factor nuclear factor kappa-light-chain-enhancer of activated B cells (NF- B). Ligand binding (e.g. TNFα) to the TNF family of receptors initiates an apoptotic signalling pathway, whereby the cytoplasmic proteins FADD (Fas-associated death domain proteins) and procaspase-8 (or procaspase-10) are recruited to form the death- inducing complex (DISC) (Wajant et al., 2003). DISC formation triggers activation of effector caspases, leading to apoptosis (Chinnaiyan et al., 1996; Peter & Krammer, 2003). The anti-apoptotic effect of GSK3 in TNF receptor-stimulated apoptosis is believed to act through the NF-B transcription factor. NF-B is a signalling protein and transcription factor that mediates expression of genes that promote cell survival and supress TNF- induced apoptosis.

GSK3 positively regulates NF-B transcription factor

Activation of NF-B is essential for suppression of TNF-induced apoptosis. GSK3 promotes NF-B transcriptional activity and cell survival in two ways: Firstly, GSK3 directly phosphorylates NF-B, promoting its transcriptional activity (Beurel & Jope, 2006; Hoeflich et al., 2000). However, it only promotes transcription of a subset of NF-B target genes, such as IL-6 and monocyte chemoattractant protein 1 (MCP-1) (Steinbrecher et al., 2005).

Secondly, GSK3 phosphorylates the NF-B subunit and inhibitor, p100 (Busino et al., 2012). p100 normally binds to NF-B in the nucleus and inhibits its transcriptional activity (Senftleben et al., 2001). However, phosphorylation by GSK3 targets it for ubiquitination by the F-box protein Fbw7, a substrate-targeting subunit of the SCF ubiquitin ligase complex (Busino et al., 2012; Skowyra et al., 1997). This targets p100 for proteolysis and degradation by the proteasome, relieving its inhibitory action of NF- B. Thus GSK3 promotes cell survival by increasing NF-B transcriptional activity directly and indirectly via p100 (Fig.1.8).

20 |Page

Sec. 1.4. GSK3 substrates

Figure 1.8. GSK3 positively regulates NF-B transcriptional activity directly and indirectly. Firstly, GSK3 directly phosphorylates NF-B, promoting transcription of a subset of target genes. Secondly, GSK3 phosphorylates the NF-B subunit and inhibitor, p100, for subsequent ubiquitination and degradation by the proteasome.

1.4.1.3 Summary: GSK3 is a key regulator of cell apoptosis

In summary, GSK3 has been shown to both promote and inhibit cell apoptosis. It promotes apoptosis via the mitochondrial intrinsic pathway by targeting the BCL-2 family of proteins, and its action can be suppressed by the growth factor survival pathway (N-terminal phosphorylation of GSK3). In contrast, it promotes survival when the extrinsic apoptosis pathway is activated by promoting NF-B transcriptional activity. This distinction is important to remember when considering GSK3 inhibitors as potential therapeutic agents, such as neuroprotectors in stroke treatment or other purposes.

1.4.2 The role of GSK3 in the regulation of immune responses

The innate and adaptive immune system is critical for sustaining life, but the resulting inflammatory response can contribute to a range of debilitating diseases. GSK3 has been identified as a regulator of inflammation and immune responses by the body, either promoting or inhibiting these processes through the expression of cytokines. (Beurel et al., 2010; Cortés-Vieyra et al., 2012).

1.4.2.1 Innate immunity

GSK3 activity was found to be necessary for pro-inflammatory cytokine production following stimulation of Toll-like receptors (TLRs) (Martin et al., 2005). TLRs are expressed mainly on antigen-presenting cells such as monocytes, macrophages and dendritic cells. These have the ability to discriminate among distinct molecular patterns to recognise and respond to foreign microbial components. Recognition of

21 |Page

Sec. 1.4. GSK3 substrates

foreign antigens by TLRs leads to signal transduction pathways that regulate an inflammatory immune response. Upon stimulation of TLRs with LPS, GSK3 inhibition was found to significantly increase the production of the anti-inflammatory cytokine IL- 10 (Martin et al., 2005), while simultaneously reducing production of pro-inflammatory cytokines, including IL-1β, IL-6, TNF, IL-12 and IFN-γ. GSK3 has also been shown to negatively regulate the level of the anti-inflammatory cytokine IL-1 receptor antagonist (IL-1Ra), a member of the IL-1 family that binds to IL-1 receptors but does not induce any intracellular response (Arend et al., 1998), while concurrently increasing the levels of IL-1β in LPS-stimulated human monocytes. Thus, GSK3 inhibition can differentially modulate TLR signalling to suppress pro-inflammatory cytokine production while enhancing the production of anti-inflammatory cytokines. This demonstrates the broad ability of GSK3 inhibition to attenuate the inflammatory response after TLR stimulation (Martin et al., 2005), whereas increased GSK3 activity acts to promote the inflammatory process. (Cortés-Vieyra et al., 2012) (Fig.1.9).

Figure 1.9. Increased GSK3 activity promotes innate inflammatory response. GSK3 activity, mediated through transcription factors such as CREB, suppress pro-inflammatory cytokines while enhancing the production of anti-inflammatory cytokines to promote TLR inflammatory response.

1.4.2.2 Adaptive immunity

GSK3 also play an important role in regulating adaptive immunity. T cell activation and the stimulation of co-receptors, such as CD28, provide maximal and sustained activation and proliferation of T cells. CD28 activation inhibits GSK3 activity, mediated via the activation of the PI3K signalling pathway (Wood et al., 2006). This inhibition of GSK3 is critical for CD4+ and CD8+ T cell survival and proliferation, whereas GSK3 activation is required for T cell death after clonal expansion (Sengupta et al., 2007). GSK3 is believed to mediate this process through regulation of immune transcription factors, such as NF-B and CREB, important for T cell and immune system function (Wen et al., 2010). Together, this suggests that while GSK3 inhibition is necessary for T cell proliferation, differentiation and survival, its activation is important

22 |Page

Sec. 1.4. GSK3 substrates

to downregulate T cell response, processes that are necessary for effective homeostasis of the adaptive immune system (Fig.1.10).

Figure 1.10. GSK3 regulates T cell response to maintain adaptive immunity homeostasis.

1.4.2.3 Immune transcription factors

The ability of GSK3 to regulate cytokine production and immune responses is largely mediated by influencing the activation of downstream transcription factors, of which many are important for immune function. In particular, GSK3 can regulate the NF- B and CREB transcription factors that mediate expression of several pro-inflammatory cytokines (Martin et al., 2005; Tsai et al., 2009). Optimal transcriptional activity of NF- B is mediated by its association with the nuclear coactivator CREB-binding protein (CBP) (Sheppard et al., 1999). GSK3 inhibition following TLR-mediated signalling was found to up regulate the activation of CREB by enhancing its association with CBP while reducing the interaction between NF-B and CBP. This demonstrates that GSK3 regulation of CREB activity following TLR activation is required to suppress the production of pro- inflammatory cytokines while concurrently enhancing anti- inflammatory cytokines production (Martin et al., 2005). However under certain condition, GSK3 can also negatively regulate NF-B, reducing NF-B transcriptional activity (Buss et al., 2004). This opposing effect of GSK3 on pro- and anti-inflammatory cytokines is partially mediated by competition between NF-B and CREB for the co- activator CBP (Martin et al., 2005). GSK3 regulation of NF-B is therefore context- specific and provides insight into the opposing stimulatory or inhibitory effect on NF-B mediated expression of pro-inflammatory cytokines.

Overall, GSK3 is an important regulator of innate and adaptive immunity by mediating the activation of downstream transcription factors (such as NF-B and CREB) that are important for immune function. GSK3 signalling is necessary for pro- inflammatory cytokine production following stimulation of TLRs of the innate immune

23 |Page

Sec. 1.4. GSK3 substrates

system and important for efficient and sustained activation and proliferation of T cells during adaptive immune responses.

1.4.3 GSK3’s role in stem cell proliferation vs. differentiation

GSK3 and its substrates play an important role in regulating stem cell homeostasis. It was first observed that treatment with pharmacological inhibitors of GSK3, including lithium (Sato et al., 2004) and the highly specific GSK3 inhibitor CT99021 (Ying et al., 2008), promotes proliferation and pluripotency of stem cells while reducing differentiation. Meanwhile, withdrawal of GSK3 inhibitors promotes ES cell differentiation into multiple cell lineages (Sato et al., 2004). In fact, the combination of only CT99021 and a MAP-kinase inhibitor is sufficient to maintain self-renewal of ES cells in vitro (Ying et al., 2008), highlighting the importance of low GSK3 activity for promoting pluripotency.

Consistent with these pharmacological studies, genetic depletion of GSK3 activity using RNAi-mediated knockdown (Huang et al., 2009) or gene knockout (Kim et al., 2009b) technologies also promotes proliferation and inhibits differentiation of ES cells. This is accompanied by increased -catenin, Hedgehog and Notch signalling, together with increased expression of the proliferation markers Nanog, Sox2 and Oct4 (Doble et al., 2007; Trowbridge et al., 2006). Consistently, expression of constitutively active GSK3 (S21A) and GSK3 (S9A) impairs proliferation of neuronal precursor cells (NPCs) in the hippocampus of adult mice (Eom & Jope, 2009; McManus et al., 2005). Together, these observations clearly show that, low GSK3 activity favours proliferation and maintenance of pluripotency of ES cells, while high levels of GSK3 favours differentiation.

1.4.3.1 Signalling pathways in stem cells

Wnt signalling

Wnt signalling promotes the maintenance/proliferative capacity of ES cells and NPC’s and is down regulated upon differentiation. Inhibition of GSK3 by Wnt signalling prevents GSK3-mediated phosphorylation of β-catenin, promoting its stabilisation and translocation to the nucleus. There it directly interacts with members of the TCF/LEF family of transcription factors (Behrens et al., 1996; Huber et al., 1996; Kühl & Wedlich, 1997) to promote expression of genes important for maintaining pluripotency (Wray et

24 |Page

Sec. 1.4. GSK3 substrates

al., 2011), including c-Myc (He et al., 1998), cyclin D1 (Shtutman et al., 1999; Tetsu & McCormick, 1999), Esrrb (Martello et al., 2012) and Nanog (Kim et al., 2011). β-catenin also interacts with the transcription factor Oct-4 to promote expression of pluripotency genes, such as Nanog (Kelly et al., 2011; Takao et al., 2007b) and Klfs, principally Klf2 (Hall et al., 2009). In summary, Wnt signalling attenuates GSK3 activity promoting the stabilisation of β-catenin and the transcription of multiple pluripotency targets that promote ES cell proliferation while suppressing their differentiation.

DISC1 and GSK3 signalling

DISC1 (Disrupted In Schizophrenia 1) is a scaffolding protein that is highly expressed in NPCs and is required for their proliferation. It is mutated in some schizophrenic and BD patients, demonstrating its importance in healthy brain development and function (Mao et al., 2009). DISC1 directly binds to GSK3 and inhibits its phosphorylation of -catenin, promoting stabilisation of -catenin and activation of TCF/LEF family of transcription factors. Mice expressing a DISC1 loss-of-function mutant specifically in the brain are viable but display increased GSK3 activity, lower - catenin levels, and reduced proliferation of NPC’s (Mao et al., 2009). Therefore, increased GSK3 activity due to loss of DISC1 antagonises NPC proliferation. This may also impair neurogenesis in the adult brain and may underlie some of the symptoms experienced by schizophrenic and BD patients. The discovery of DISC1 mutants further supports a role for hyperactive GSK3 in psychiatric disorders.

Growth factor signalling

Growth factors such as IGF-1, insulin, EGF and neurotrophins (reviewed in (Wada, 2009)) induce activation of the PI3K-Akt pathway that results in inhibition of GSK3 mediated via phosphorylation of the N-terminal serine residue (Ser9 in GSK3 and Ser21 in GSK3). This promotes pluripotency in ES cells by stabilising a number of transcription factors downstream of GSK3, including -catenin, c-Myc, c-Jun, Klf5, and Mef2D (Cole, 2012; Ying et al., 2008). Consistently, withdrawal of growth factors promotes an increase in GSK3 activity levels and promotes the differentiation of ES cells (Sato et al., 2004). For example, increased GSK3 activity in the absence of LIF signalling, antagonises the transcriptional regulator STAT3 to promote expression of pluripotency targets, including Klf5 and c-Myc (Cartwright et al., 2005; Li et al., 2005; Niwa et al., 2009). These are among only a few transcription factors that have been used to induce

25 |Page

Sec. 1.4. GSK3 substrates

pluripotency in an iPS cell system (Hall et al., 2009; Jiang et al., 2008), exemplifying the importance of GSK3 substrates in ES cells.

Notch signalling

Notch signalling is critical to the maintenance of early neuronal progenitor cells (radial progenitors) during development of the nervous system. Generally, Notch signalling promotes proliferation of NPCs and decreases during differentiation into intermediate neuronal progenitors. GSK3 attenuates Notch signalling both directly and indirectly. It binds and phosphorylates Notch2, tagging it for proteolysis (Espinosa et al., 2003). Indirectly, GSK3 antagonises the Wnt signalling pathway that enhances transcription of Notch1 (Jin et al., 2009). Notch signalling is increased in GSK3 knockout brains, as are its downstream signalling repressors Hes1 and Hes5 (Kim et al., 2009c). These proteins repress transcription of pro-neural genes, thus inhibiting differentiation into neurons and maintaining proliferation of NPC’s. Therefore, when GSK3 levels are low, increased Notch signalling is associated with suppressed differentiation of NPCs into mature neurons.

Hedgehog signalling

The hedgehog signalling pathways is critical to determine cell fate during development and for maintaining proliferation of stem cells (Lai et al., 2003). Gli family of proteins are critical effectors of Hh signalling where they can regulate the transcription of genes important in regulating progenitor proliferation, such as cyclin D and E (Duman- Scheel et al., 2002; Kenney & Rowitch, 2000; Lai et al., 2003). In the absence of Hh signalling, Gli is phosphorylated by GSK3 and CK1 (following PKA priming), which target it for ubiquitination and proteolysis, generating a truncated repressor lacking the C-terminal activation domains (Pan et al., 2006; Tempé et al., 2006). In the presence of Hh signalling, this signalling is disrupted, inhibiting the phosphorylation and processing of Gli and leading to the active full-length protein in the nucleus and transcription of positive regulators of proliferation (Lai et al., 2003). Therefore, when GSK3 activity is high, this promotes proteolysis and formation of truncated repressor forms of Gli, facilitating differentiation of NPCs. Conversely, when GSK3 activity is low, full length Gli maintains proliferation of NPCs.

26 |Page

Sec. 1.4. GSK3 substrates

1.4.3.2 GSK3 mediates the regulation of gene expression downstream of transcription factors

Myc transcription factors

C-Myc and its related family members N-Myc and L-Myc are critical effectors of stem cell proliferation and for delaying differentiation (Canelles et al., 1997; Knoepfler et al., 2002). Activation of c-Myc transcriptional activity drives expression of other genes promoting cell cycle progression, including cyclin D1, cyclin A, cyclin E and increased CDK function (Dang, 1999; He et al., 1998). Transcription of the c-Myc gene is activated by the of Wnt/-catenin and LIF/STAT3 self-renewal pathways (Cartwright et al., 2005; He et al., 1998), previously implicated in maintaining cell-renewal of Murine ES cells (Matsuda et al., 1999; Niwa et al., 1998). However the stability of the c-Myc protein is controlled by GSK3 downstream of the growth factor-PI3K signalling pathway. Under basal conditions, GSK3 activity is relatively high and it phosphorylates c-Myc at Thr58 following priming phosphorylation of Ser62 by dual-specificity tyrosine-regulated kinase 2 (DYRK2) (Taira et al., 2012; Welcker et al., 2004). This creates a binding site on c- Myc for Fbw7-mediated ubiquitination, followed by degradation by the proteasome. Upon ligand binding and activation of the PI3K pathway, GSK3 activity is reduced, thus decreasing c-Myc phosphorylation and ubiquitination, therefore stabilising the protein and increasing its abundance.

Therefore, c-Myc is negatively regulated by GSK3 in two ways: firstly, GSK3 controls its transcription by antagonising the Wnt/-catenin signalling pathway, and secondly, targets it for ubiquitin-mediated degradation.

Other GSK3 transcriptional targets mediated by Fbw7 degradation

GSK3 also targets other transcription factors and cell cycle regulators in a similar fashion (Fig.1.11). Phosphorylation of N-myc (Sjostrom et al., 2005), Klf5 (Jiang et al., 2008; Liu et al., 2010; Luan & Wang, 2014), c-Jun, Notch1 (Hoeck et al., 2010), and cyclin E (Welcker et al., 2003) targets them for ubiquitination by Fbw7 and degradation by the proteasome. Indeed, phosphorylation of -catenin also targets it for similar destruction, although ubiquitination in this case is mediated by the E3 ubiquitin ligase SCFβTrCP (Aberle et al., 1997). Interestingly, the GSK3 priming phosphorylation site in c- Jun (Ser243) is mutated to a non-phosphorylatable phenylalanine in a viral oncogenic form (v-Jun) (Wei et al., 2005). This prevents phosphorylation by GSK3 at Ser239 and

27 |Page

Sec. 1.4. GSK3 substrates

subsequent ubiquitination, stabilising the protein and driving uncontrolled proliferation in tumorigenesis. Similarly, the GSK3 phosphorylation site is mutated in -catenin in a form of hereditary colon cancer (Johnson et al., 2005). These mutations highlight the importance of GSK3 in regulating stem cell maintenance, cell cycle progression and cancer.

Figure 1.11. GSK3 also targets other transcription factors and cell cycle regulators for Fbw7 ubiquitination and degradation by the proteasome. (Figure adapted from (Welcker & Clurman, 2008)).

1.4.3.3 Summary: GSK3 activity is important in the regulation of cell-fate

GSK3 is therefore an important target of several signalling pathways controlling stem cell fate (Fig.1.12). Many of these pathways can be activated simultaneously and GSK3 appears to have a central role to negotiate these simultaneous inputs to determine cellular outcome. Importantly, there is very little cross-talk between these pathways i.e. inhibiting one pathways usually doesn’t affect other pathways. This is because, Wnt, Notch and Hedgehog signalling don’t inhibit GSK3 activity directly, but rather they inhibit phosphorylation of particular substrates by disrupting complexes. Typically, under basal conditions where GSK3 activity is high, GSK3 antagonises Wnt, Notch, Hedgehog and growth factor signalling to facilitate differentiation of ES cells. Upon ligand stimulation and activation of these pathways, GSK3 activity is reduced, promoting up- regulation of transcription factors that contribute to ES cell proliferation and pluripotency.

28 |Page

Sec. 1.4. GSK3 substrates

Figure 1.12 GSK3 signalling pathways regulate stem cell-fate. High GSK3 activity antagonises Wnt, Notch, Hedgehog and growth factor signalling to facilitate differentiation of ES cells. Upon ligand stimulation and activation of these pathways, GSK3 activity is reduced, promoting up- regulation of transcription factors that contribute to ES cell proliferation and pluripotency.

1.4.4 GSK3 regulation of neurogenesis

1.4.4.1 GSK3 is a critical regulator of neurogenesis downstream of signalling pathways

Consistent with its role in regulating proliferation and differentiation of stem cells, GSK3 is also a critical regulator of neurogenesis in both the developing and adult brain (Hur & Zhou, 2010; Lange et al., 2011). This is clearly demonstrated in the brains of genetically modified mice, whereby deletion of both GSK3/ results in hyper- proliferation of neuronal progenitors in the developing brain. This is accompanied by suppression of neuronal maturation/differentiation (Kim et al., 2009b) and enlargement ventral brain areas due to reduced numbers of mature neurons. Wnt, Notch, Hedgehog and FGF signalling is enhanced, together with strong Sox2 and c-Myc positive staining, indicative of expanded neural progenitor pools (Kim et al., 2009b).

29 |Page

Sec. 1.4. GSK3 substrates

Meanwhile, knock-in mice expressing GSK3 that is insensitive to growth factor inhibition (Ser21/9 mutated to alanine in GSK3 and , respectively) are viable and appear to develop normally, although they display reduced neurogenesis (Eom & Jope, 2009; Spittaels et al., 2002), despite normal rates of proliferation (Eom & Jope, 2009; McManus et al., 2005). This suggests a defect in differentiation/maturation or survival of NPC’s in these mice. Together, these observation suggest that the Wnt, Notch and Hedgehog pathways reduce GSK3 activity to promote proliferation and pluripotency of NPC’s, while growth factor mediated inhibition of GSK3 promotes differentiation and/or survival of NPC’s to form post-mitotic neurons. GSK3 activity during neurogenesis is an intricate balance between appropriate attenuation of GSK3 activity to promote survival and proliferation, while periods of increased GSK3 activity are required for progenitor differentiation. Therefore, finely-tuned balance of GSK3 activity is critical during neurogenesis and for healthy brain development.

For example, the Hh transcriptional activators, Gli1 and Gli2 expression levels are increased in conditional brain-specific GSK3 knockout mice (Kim et al., 2009b). Gli1 in particular has been identified as a direct transcriptional target of Gli2 in neural systems (Ding et al., 1998). Therefore, Gli1 knocked down using small interfering RNA (siRNA) in GSK3 null mice was associated with a significant decrease in progenitor proliferation compared with control mice (Kim et al., 2009b). Furthermore, N-myc expression, downstream of Hh signalling, positivity regulates NPC proliferation while suppressing differentiation during neurogenesis (Knoepfler & Kenney, 2006; Sjostrom et al., 2005). In the presence of downstream Hedgehog and growth factor signalling, GSK3 activity is reduced, stabilising N-myc levels and promoting NPC proliferation. Together, this demonstrates the importance of Hh signalling downstream of GSK3 for maintaining/supporting NPC survival/proliferation and suppression of neurogenesis.

In addition, FGF-PI3K growth factor signalling is important to promote neural progenitor differentiation and neurogenesis. For example, in culture, progenitor proliferation was assessed in control and GSK3-deficent cortical neurons after treatment with FGF2 and/or the PI3K inhibitor, LY294002 (Kim et al., 2009c). Treatment with growth factor only increased progenitor proliferation in control cultures, but addition of LY294002 decreased proliferation in FGF2-treated control cells by more than 50%. However, the inhibitory effects of LY294002 were completely suppressed in GSK3- deficient cells. Elsewhere, stimulation of the FGF-PI3K pathway increases c-Myc levels

30 |Page

Sec. 1.4. GSK3 substrates

in cultured neurons and this was shown to be directly mediated by reduced GSK3 activity (Kim et al., 2009b). Together, these observations suggest that GSK3 is a critical downstream target of FGF-PI3K signalling that drives proliferation by promoting expression of pluripotency genes such as c-Myc.

GSK3 activity is also implicated in the decrease of neurotrophin growth factor signalling during neurogenesis. VEGF is a neurotrophin-like growth factor well established to promote neurogenesis (Cao et al., 2004; Jin et al., 2002). In hippocampal extracts from mice expressing constitutively active GSK3, VEGF levels were significantly decreased, indicating that low GSK3 activity supports VEGF signalling during neurogenesis. This supports the idea that the loss of inhibitory control of GSK3 impairs neurogenesis via decreasing signalling from factors such as VEGF (Eom & Jope, 2009).

1.4.4.2 Summary: GSK3’s dynamic control of neurogenesis

Overall, the dynamic control of GSK3’s activity in the brain is important in the regulation of NPC homeostasis (Fig.1.13). GSK3 inhibition removes homeostatic control on NPC, shifting the balance towards self-renewal and away from differentiation and neurogenesis (Kim et al., 2009b). However, the loss of inhibitory control over GSK3, brought about from mutations in upstream regulators such as DISC1 can also impair neurogenesis by not supporting the proliferative and survival pathways necessary for NPC survival during neurogenesis. This demonstrates the importance of GSK3 in healthy brain function and signalling, and the critical balance of GSK3 activity in the regulation of neurogenesis in vivo.

31 |Page

Sec. 1.4. GSK3 substrates

Figure 1.13. Schematic illustration of GSK3’s dynamic control of NPC homeostasis. The decrease in GSK3 activity shifts the balance towards self-renewal and away from differentiation and neurogenesis. Whereas, increase in GSK3 activity promotes differentiation shifting the balance away from self-renewal and towards neuritogenesis.

1.4.5 GSK3 and neuronal morphology

1.4.5.1 GS3 is a key regulator of axon growth and branching

GSK3 plays a central role in the organisation of neuronal morphology. In particular, signalling pathways that reduce GSK3 activity increase stabilisation of microtubules, promoting neurite elongation, branching, pathfinding and synapse formation (see reviews: (Goold & Gordon-Weeks, 2004; Gordon‐Weeks, 2004; Zhou & Snider, 2005)).

GSK3α/ are highly expressed during axonogenesis and are functionally redundant with respect to neurite outgrowth (Kim et al., 2006; Leroy & Brion, 1999a). Using highly specific pharmacological GSK3 inhibitors as well as shRNA-mediated knockdown of GSK3, inhibition of GSK3 reduces axon outgrowth, increased spreading of growth cones and thickening of axon shafts of mouse embryonic dorsal root ganglia (DRG) neurons cultured in vitro (Kim et al., 2006). It also induces formation of multiple axons in rat Hippocampal cultures (Jiang et al., 2005). Interestingly, treatment with GSK3 peptide inhibitors or lithium at relatively low concentrations (weaker inhibition of GSK3),

32 |Page

Sec. 1.4. GSK3 substrates

increases neurite/axon outgrowth, including increased elongation, sprouting and branching (Kim et al., 2006; Lucas et al., 1998; Shah et al., 2013; Takahashi et al., 1999). As the concentration of lithium increases, microtubules become increasingly stabilised, yet disorganised in structure. GSK3 activity is therefore important in driving microtubule reorganisation during neuritogenesis (Shah et al., 2013). Altogether, this demonstrates that weak inhibition of GSK3 enhances neurite outgrowth and branching by increasing the dynamics of microtubule assembly, especially at the growth cone. However, under stronger inhibition, neurite outgrowth is reduced due to excessive microtubule stability in neurites.

It has also been reported that inactive pools of GSK3 and GSK3 (phosphorylated at Ser21/9, respectively) are enriched at the leading edge of growth cones in DRG neuron cultures. A study by Eickholt et al. showed that treatment with the inhibitory guidance molecule Semaphorin 3A induced a rapid increase in active GSK3 at the leading edge of growth cones. Also, collapse of growth cones could be prevented by treatment with GSK3 inhibitors including lithium (Eickholt et al., 2002). This suggests that GSK3 inhibition maybe important for stabilising/organising microtubules at the leading edge of growth cones to promote neuronal polarity and extension. This also suggest that the level and/or localisation of GSK3 activity in different parts/regions of neurons is important for neurite outgrowth, growth cone dynamics and response to guidance cues.

GSK3 overexpression (OE) in vivo negatively regulates granule neuron morphology and connectivity during adult hippocampal neurogenesis. The conditional GSK3-OE in mice produced less branching in dendritic tree morphology and reduced the number of functional synapses (Llorens-Martin et al., 2013). Interestingly, these effects could be reverted in part by inhibiting the transgene, increasing hippocampal plasticity (Fuster-Matanzo et al., 2013; Llorens-Martin et al., 2013). Interestingly, phospho-GSK3 (inactive) was preferential localised to axon tips, while active GSK3 was found to be significantly higher in the dendrites. Indeed, expression of constitutively active GSK3 (GSK3-S9A) prevents axon formation (Jiang et al., 2005) and leads to the shrinkage of dendrites (Rui et al., 2013). Together this is consistent with the idea that GSK3 inhibition is important for neurite outgrowth at the leading edge of growth cones.

33 |Page

Sec. 1.4. GSK3 substrates

In summary, high GSK3 activity, impairs neurite extension and neuronal connectivity during neurogenesis, evident by decreased neurite branching and suppression of dendrite formation. GSK3 inhibition is critical, particularly at the leading edge of growth cones, in establishing neuronal polarity and axon formation. This is likely mediated by stabilisation of microtubules necessary for tubulin polymerisation and neurite elongation. This also suggests that application of weak GSK3 inhibitors (e.g. lithium) may be a novel approach to promote generation of new axons and neuronal connections after neural injuries (Fig.1.14).

Figure 1.14. GSK3 is a key regulator of axon growth and branching. High GSK3 activity impairs microtubule polymerisation, preventing neurite extension and reduces branching, whereas weak inhibition of GSK3 enhances neurite outgrowth and branching by increasing the dynamics of microtubule assembly, especially at the growth cone. However, under strong inhibition neurite outgrowth is reduced due to excessive microtubule stability in neurites.

1.4.5.2 GSK3-mediated growth factor signalling in neuronal development

Consistent with the use of GSK3 inhibitors, activation of signalling pathways that reduce GSK3 activity also stimulate growth cone enlargement, spreading and axon growth (Valerio et al., 2006; Zhou & Snider, 2006). Neurotrophins such as NGF, IGF-1, and GDNF reduce GSK3 activity ~50%, similar to treatment with low doses of lithium. They induce neurite formation and axon growth (Zhou & Snider, 2006) by activation of receptor tyrosine kinases (RTKs) triggering ERK1/2 signalling in the MAPK pathway (Cowley et al., 1994; Pang et al., 1995) and/or PI3K pathway activation (Zhou et al., 2004), both of which downregulate GSK3 activity. For example, NGF activation of the PI3K pathway promotes the GSK3 N-terminal serine phosphorylation and inhibition, enhancing axon growth (Zhou et al., 2004). Consistently, the inhibition of the MAPK pathway using pharmacological inhibitors, increases GSK3 activation, and consequently reduces axon growth (Goold & Gordon-Weeks, 2005). This further supports weak GSK3

34 |Page

Sec. 1.4. GSK3 substrates

inhibition downstream of growth factor signalling as critical to enhance neurite outgrowth.

Wnt signalling has also been reported to play an important role during neurite outgrowth, especially during development. For example, Wnt-7a signalling in developing granule neurons stimulates axon elongation and branching (Lucas et al., 1998) and has been shown to increase synapsin I expression, a presynaptic protein involved in synapse formation and function (Chin et al., 1995; Hilfiker et al., 1999; Rosahl et al., 1995). Wnt signalling induces clustering of synapsin I in granule neurons and mossy fiber axons in culture (Hall et al., 2000; Lucas & Salinas, 1997). Consistent with this, treatment of developing neurons with lithium also induces synapsin I clustering, implicating GSK3 in this process (Hall et al., 2002). Because lithium directly mimics the effect of WNT-7a signaling, it has been suggested that mossy fiber axonal branching and enhanced presynaptic differentiation is mediated through inhibition of GSK3 (Hall et al., 2000).

Together, these results suggest that growth factor/Wnt signalling facilitates neurite outgrowth and synapse formation and that GSK3 inhibition is pivotal in mediating these effects.

1.4.5.3 Downstream substrates

Several microtubule-associated proteins are directly phosphorylated by GSK3, including MAP1B, MAP2, tau, CRMP2, CLASP2 and APC (Cole, 2012; Sutherland, 2011). In general, phosphorylation of these proteins causes dissociation from tubulin/microtubules, resulting in more dynamic/less stable microtubules that encourages growth cone dynamics and neurite outgrowth (Sergeant et al., 2008).

Tau

Tau is a tubulin-binding protein that promotes microtubule assembly and stability in neurons. Tau’s ability to bind and stabilise microtubules is regulated by site-specific phosphorylation (Cho & Johnson, 2004). GSK3 phosphorylates numerous sites on tau, with the main sites identified being, Ser396/400 and Thr231 (Cho & Johnson, 2003; Llorens-Marítin et al., 2014). Phosphorylation at these sites decreases tau’s affinity for microtubules, destabilising their structure. (Cho & Johnson, 2004; Sengupta et al., 2006). For example, co-expression of GSK3 and tau in cultured neurons or transgenic mice increases tau phosphorylation and decreases it’s binding to microtubules (Nuydens et al.,

35 |Page

Sec. 1.4. GSK3 substrates

2002; Spittaels et al., 2000). Conversely, treatment with lithium and other GSK3 inhibitors reduces GSK3-dependent phosphorylation of tau and increases its binding to microtubules (Takahashi et al., 1999).

Hyperphosphorylation of tau by GSK3 and other kinases (e.g. Cdk5, DYRK, MAPK) promotes aggregation of tau protein in the cytoplasm of neurons, forming neurofibrillary tangles (NFTs), a characteristic lesion of AD (Luna-Muñoz et al., 2007). Mice overexpressing GSK3 display pretangle-like somatodendritic localisation of tau and associated reactive astrocytosis and microgliosis, indicative of neuronal stress and death (Lucas et al., 2001). Furthermore, transgenic mice overexpressing human tau develop hyper-phosphorylated tau aggregates in neuronal cell bodies and dendrites (Andorfer et al., 2003), although, treatment with the GSK3 inhibitor lithium, reduces the number of tau tangles and decreases associated axonal degeneration (Nakashima et al., 2005; Noble et al., 2005; Pérez et al., 2003). This strongly suggests that GSK3-mediated hyperphosphorylation of tau promotes tauopathy progression and inhibiting GSK3 may be an effective therapeutic strategy for reducing tau aggregation and NFT formation. Encouragingly, in human trials assessing the therapeutic effects of lithium in AD, chronic treatment with lithium reduced patients decline in cognition and memory function, suggesting that long-term treatment with lithium may at least delay the progression of AD (Forlenza et al., 2011; Nunes et al., 2007).

Microtubule-associated proteins (MAPs)

MAPs are microtubule assembly-promoting proteins that are important for the polymerisation, stability and arrangement of microtubules (Sanchez et al., 2000). MAP1B is important in the development of neurons with expression highest in differentiating neurons and in adult neurons that display neuronal plasticity (Gonzalez-Billault et al., 2004). It regulates axon growth and growth cone structure by controlling microtubule stability (Goold & Gordon-Weeks, 2004). GSK3 phosphorylates MAP1B (Lucas et al., 1998) at Ser1260 and Thr1264 in vivo (Trivedi et al., 2005), which decreases its ability to associate with and stabilise microtubules (Goold et al., 1999; Trivedi et al., 2005). This results in unstable microtubules that promote dynamic growth cones at the expense of neurite outgrowth (Goold et al., 1999; Hall et al., 2000).

MAP2 is another GSK3 substrate implicated in neuronal outgrowth and polarity (Craig & Banker, 1994; Matus, 1994; Weisshaar et al., 1992). It is primarily expressed in

36 |Page

Sec. 1.4. GSK3 substrates

the nervous system and is essential during the first steps of neurite outgrowth. For example, no neurite growth occurs when MAP2 expression is suppressed in neuronal cultures (Caceres et al., 1992; Sharma et al., 1994). GSK3 phosphorylates MAP2 at Thr- 1620 and Thr-1623 (Sánchez et al., 1998; Sánchez et al., 1996), reducing its affinity for microtubules and decreasing microtubule bundling (Sánchez et al., 2000). Taken together, GSK3-mediated phosphorylation of MAP1B and MAP2 induces microtubule destabilisation, increasing microtubule dynamics. This is likely important for effective growth cone dynamics and pathfinding, while negatively regulating microtubule polymerisation and axon elongation (Owen & Gordon-Weeks, 2003).

GSK3 role also extends to the regulation of transcriptional factors that facilitate the expression of actin/tubulin-binding proteins involved in axonal outgrowth. For example, serum response factor (SRF) is a stimulus-dependent transcription factor that plays a critical role in promoting axon growth in the brain. GSK3 has recently been identified to directly phosphorylate and activate SRF at Ser224 (Li et al., 2014), promoting the expression of actin/tubulin-binding protein such as, vinculin and MAP1B in neurons and facilitating axon growth.

CRMP family

Collapsin response mediator proteins (CRMP1-5) are a family of microtubule- associated proteins, enriched in neurons that regulate neurite outgrowth and growth cone dynamics. CRMPs selectively bind to tubulin  and  heterodimers to promote microtubule polymerization. CRMP1, CRMP2 and CRMP4 have the same conserved GSK3 phosphorylation sites at Ser518, Thr514 and Thr509 (Cole et al., 2006). Phosphorylation at these sites is dependent on prior priming phosphorylation at Ser522 by Cdk5 for CRMP1/2, or DYRK for CRMP4 (Cole et al., 2006). Phosphorylation reduces CRMP2 binding to tubulin dimers, decreasing microtubule polymerization. This alters CRMP2-induced axon elongation, although different studies have reported contrasting effects (Cole et al., 2004b; Fukata et al., 2002; Yoshimura et al., 2005). CRMP4 has also been shown to promote axon elongation, albeit to a lesser extent than CRMP2 (Cole et al., 2006). Therefore, CRMP2 regulates neurite outgrowth by its ability to bind and stabilise the assembly of tubulin heterodimers in both neurite shafts and growth cones. The other CRMP isoforms are also predicted to perform similar roles, although they are less well studied and this requires confirmation.

37 |Page

Sec. 1.4. GSK3 substrates

Similar to tau, CRMP2 is also hyperphosphorylated in mouse models of AD and human post-mortem brain tissue from AD patients. In fact, this occurs at a very early age, well before the appearance of plaques or tangles (Cole et al., 2007), but does not occur in other closely related forms of dementias that also feature hyperphosphorylated tau and NFTs (FTDP-17 and Pick’s disease) (Williamson et al., 2011). This suggests that hyperphosphorylated CRMP2 might be useful as a biomarker for early and specific detection of AD.

Hyperphosphorylation of CRMP2 and tau is most likely not caused by overt changes to GSK3 activity or the Cdk5 priming kinase, since their activities are not dramatically altered in AD (Cole et al., 2006; Cole et al., 2007; Pei et al., 1997; Tandon et al., 2003; Williamson et al., 2011). Rather, the Cdk5 priming site on these 2 proteins were found to be remarkably resistant to dephosphorylation by phosphatases (Cole et al., 2008). Therefore, theoretically, small changes in activity of the Cdk5 priming kinase could be amplified to cause a relatively large increase in priming phosphorylation, leading to increased phosphorylation of CRMP2 and tau by GSK3. This might explain why large changes to GSK3 and Cdk5 activity have not been detected in AD, despite large increases in phosphorylation of their substrates.

CLASP2

Another microtubule-binding substrate of GSK3 is CLIP-associated protein 2 (CLASP2). GSK3 phosphorylates CLASP2 at two motifs (Ser594/598/602/606/610 and a partial duplication of this region, Ser569/572/576) (Kumar et al., 2009). This phosphorylation promotes CLASP2 dissociation from microtubules, impairing axon growth. However, under strong suppression of GSK3 activity, this promotes high affinity of CLASP2 for microtubule, inducing over stabilisation of microtubule structure, looping in the growth cone and attenuating axon growth (Hur et al., 2011). CLASP2 can therefore promote and inhibit axon growth depending on its interaction with microtubules. This dynamic interaction is dependent on the level of phosphorylation by GSK3 and further suggests that a delicate balance between activation/inactivation of GSK3 signalling is required to regulate CLASP2 for effective growth cone dynamics.

Actin-associated proteins

Some neuronal GSK3 substrates have been identified that regulate the actin cytoskeleton. -adducin is an actin-capping protein that recruits spectrin to the fast-

38 |Page

Sec. 1.4. GSK3 substrates

growing ends of actin filaments. It has been shown to be important for dynamic disassembly and reassembly of synapses (Bednarek & Caroni, 2011). GSK3 phosphorylates -adducin’s C-terminal domain at three distinct sites (Ser697/693, Ser613, Ser600/596), following priming by Cdk5 at Ser701/617/604, respectively (Farghaian et al., 2011). Transfection of wild type -adducin but not a non- phosphorylatable mutant into primary cortical neurons increased axon elongation, branching and dendritic growth (Farghaian et al., 2011). Furthermore, GSK3-mediated phosphorylation induces localisation of -adducin to actin/spectrin junctions at the cell membrane, suggesting that phosphorylation of -adducin is necessary to mediate its functional effects. Importantly, -adducin has been shown to be required for assembly of new synapses and important for the precision of memory and learnt behaviour (Bednarek & Caroni, 2011; Ruediger et al., 2011). It will be interesting to see if GSK3-mediated phosphorylation of -adducin also influences synapse formation and/or stability.

1.4.5.4 Summary: Modulation of GSK3 activity regulates neuronal morphology

In summary, the level and/or localisation of GSK3 activity is critical for the regulation of axon and growth cone morphology. In general, neurotrophin and growth factor signalling that weakly inhibit GSK3 activity, promote neurite outgrowth by reducing phosphorylation of downstream substrates, promoting binding to tubulin/microtubules and their polymerisation. However, evidence suggests that an optimal level of GSK3 activity is required for efficient axon elongation and growth cone dynamics. Under such conditions, the localisation of GSK3 activity may be important to ensure that some substrates are dephosphorylated to promote microtubule extension, whereas others are phosphorylated to maintain microtubule dynamics. This delicate balance between GSK3 activation and inactivation acts to support coordination of growth cone pathfinding and its dynamic response to guidance cues, while promoting efficient axon elongation.

1.4.6 GSK3 and neurotransmission

GSK3’s role in the brain is especially important for regulating synaptic plasticity. In particular, the balance between LTP and LTD. GSK3 kinase activity is essential for activation of NMDA (N-methyl-D-aspartate)-mediated LTD and has been shown to be important for internalisation and trafficking of NMDA and AMPA (-amino-3-hydroxyl- 5-methyl-4-isoxazole-propionate) receptors. It has also been implicated in the trafficking

39 |Page

Sec. 1.4. GSK3 substrates

and internalisation of GABA (gamma-aminobutyric acid) receptors. Through its regulation of synaptic plasticity, GSK3 has an important role in learning and memory formation/reconsolidation. This may explain how dysregulation of GSK3 activity contributes to psychiatric and neurodegenerative diseases.

1.4.6.1 The role of GSK3 in LTP/LTD

A study by Peineau et al. investigated the potential involvement of 58 Ser/Thr protein kinases in NMDA dependent LTD in hippocampal brain slices. Only one kinase, GSK3, was confirmed to have an essential role in this process (Peineau et al., 2009). Induction of LTD reduces Ser21/9 phosphorylation of GSK3, increasing its kinase activity. Furthermore, activation of GSK3 has been shown to reduce presynaptic glutamate release (Zhu et al., 2007). These observations indicate that high GSK3 activity favours LTD. Conversely, induction of LTP reduces GSK3 activity via increased phosphorylation at its inhibitory N-terminal residues (Ser21/9) (Chen et al., 2006; Costa et al., 2005; Wang et al., 2004). Indeed, abnormally high GSK3 activity antagonises induction of LTP (Peineau et al., 2007; Zhu et al., 2007). Therefore, high GSK3 activity promotes LTD, while low GSK3 activity favours LTP.

LTP is closely associated with learning and forming new memories, while LTD is seen to antagonise this process. However, LTD is believed to be important for memory reconsolidation (i.e. reinforcing prior potentiated circuits) by a process called competitive synaptic maintenance. This involves stabilising previously formed memories while supressing the acquisition of new memories (Diamond et al., 2005; Kimura et al., 2008). According to this model, low GSK3 activity favours LTP and new memory formation, while high GSK3 activity favours LTD and memory reconsolidation at the expense of new memory formation (at individual synapses) (Fig.1.15). Indeed, transgenic mice overexpressing GSK3β in the brain (high GSK3 activity) exhibit spatial learning (Liu et al., 2003) and memory defects (Engel et al., 2006; Hernández et al., 2002; Hooper et al., 2007) that can be reversed by reducing GSK3 activity in these animals using GSK3 inhibitors or silencing the transgene to enhance synaptic plasticity (Hooper et al., 2007; Zhu et al., 2007). Conversely, adult mice with impaired GSK3 expression (low GSK3 activity), either by pharmacological inhibition or by heterozygous knockdown (GSK3+/- ), exhibit problems reconsolidating memories and retrograde amnesia, indicating impairment in long-term memory formation (Kimura et al., 2008). This suggests that

40 |Page

Sec. 1.4. GSK3 substrates

GSK3 activity maybe critical for LTD and memory reconsolidation of learnt information. Furthermore, this is consistent with upstream regulators/inhibitors of GSK3, such as BDNF, that promote initial learning and memory, but is not required for reconsolidation of long-term memories (Lee et al., 2004a). Finally, the unusually high GSK3 activity and low BDNF expression correlates with long-term memory defects often experienced by BD patients (Cunha et al., 2006; Ferrier & Thompson, 2002; Nicol Ferrier et al., 2004; Scott et al., 2000).

1.4.6.2 Neurotransmitter receptors

LTP and LTD are both triggered by synaptic activation of one class of glutamate receptor, the NMDA receptor. Alterations in synaptic efficiency, including changes in neurotransmitter release and changes in the function of postsynaptic receptors, can shift the balance between NMDA triggered LTP/LTD (see review: (Bradley et al., 2012)). NMDA receptors are excitatory ligand-gated ion channels implicated in multiple neuronal functions including, synaptic plasticity, learning and memory. Inhibition of GSK3 using potent pharmacological inhibitors stimulates rapid internalisation of NMDA receptors, reducing the post-synaptic surface expression of NR1 and NR2B subunits (Chen et al., 2007). Indeed, treatment of cortical pyramidal neurons with GSK3 inhibitors or RNAi- mediated knockdown causes long-lasting reduction of LTD dependent NMDA synaptic currents (Chen et al., 2007; Diamond et al., 2005; Kimura et al., 2008). Therefore, antagonising GSK3 activity reduces NMDAR-mediated synaptic transmission as a result of receptor internalisation. Possible mechanisms involve Rab5/clathrin-mediated endocytosis. For example, expression of dominant-negative Rab5 blocks the effect of GSK3 inhibitors to down-regulate NMDA currents, while constitutively active Rab5 mimics GSK3 inhibitors by reducing basal NMDAR currents (Chen et al., 2007). Similarly, pharmacological or RNAi-mediated GSK3 inhibition activates Rab5 and reduced expression and cluster density of AMPA receptors GluA1/2/3 at the synaptic membrane. Consistently, this was accompanied by decreased mEPSC current amplitude (Wei et al., 2010). These observations suggest that under-basal conditions, GSK3 activity may antagonise Rab5/clathrin-mediated endocytosis, promoting retention of NMDA/AMPA receptors at the cell/synapse surface. Recently, increased GSK3 activity in Alzheimer’s disease mice has been shown to impair Rab5’s ability to promote NMDAR internalisation, resulting in NMDA-induced excitotoxicity (Deng et al., 2014). This may suggest a mechanism by which GSK3 inhibitors offer neuroprotective qualities from

41 |Page

Sec. 1.4. GSK3 substrates

glutamate excitotoxicity, whereas excessive GSK3 activity may be instrumental in neurodegeneration.

However, since high GSK3 activity promotes LTD and because activation of NMDA receptors and endocytosis of AMPA receptors are important for the induction and expression of LTD (Collingridge et al., 2004; Hanley, 2010a; Kessels & Malinow, 2009; Malinow & Malenka, 2002), it would be expected that high GSK3 activity promotes the internalisation of AMPA receptors at the post-synaptic surface. Indeed, Peineau et al. describes a complex formation between functionally active GSK3 and AMPA receptors. It was suggest that GSK3 activity may regulate the trafficking or function of AMPA receptors, thereby affecting the expression of LTD (Bradley et al., 2012; Peineau et al., 2007). Consistent with this, GSK3 inhibition has been shown to prevent NMDA- induced AMPA receptor endocytosis during LTD (Du et al., 2010). This suggests that during NMDA stimulated LTD when GSK3 activity is high, GSK3 may promote endocytosis of AMPA receptors, important for the induction and expression of LTD.

GSK3 may also regulate GABAergic signalling, since it was shown to phosphorylate and decrease post-synaptic expression of the GABAergic scaffolding molecule gephyrin at Ser270, likely targeting it for proteolytic degradation (Rui et al., 2013; Tyagarajan et al., 2011). Consequently, this is associated with a decrease in GABAergic inhibitory transmission, resulting in neuronal hyperexcitability (Petrini et al., 2014; Rui et al., 2013). Treatment with pharmacological GSK3 inhibitors increases GABA receptors expression at the surface of cultured cortical neurons (Wei et al., 2010). These findings suggest that GSK3 negatively regulates GABAergic transmission.

Figure 1.15. GSK3 activity is important for regulating LTP/LTD during memory formation / reconsolidation. High GSK3 activity promotes AMPA receptor internalisation, favouring LTD

42 |Page

Sec. 1.4. GSK3 substrates

and memory reconsolidation. Whereas, low GSK3 activity promotes AMPA receptor expression, favouring LTP and memory encoding.

In summary, GSK3 activity is important for the trafficking and expression of glutamatergic receptors (NMDA and AMPA receptors) at the post-synapse, while negatively regulating the expression of GABA receptors. NMDA/AMPA (glutamate) transmission is stimulatory, while GABA transmission is inhibitory. Therefore, high levels of GSK3 activity in neurons promotes LTD, whereas low GSK3 activity favours LTP (Hooper et al., 2007; Peineau et al., 2009; Peineau et al., 2007).

1.4.6.3 GSK3 regulates synaptic proteins

GSK3 regulates neurotransmission via a number of synaptic and structure-related proteins. GSK3 phosphorylates the postsynaptic density protein, PSD-95, at Thr-19. PSD- 95 is a major scaffold protein of the PSD that promotes maturation and strength of excitatory synapses. Phosphorylation by GSK3 destabilises PSD-95, promoting AMPA receptor internalisation and LTD (Nelson et al., 2013). Interestingly, the Thr-19 site in PSD-95 does not appear to have a priming site for GSK3 (i.e. Ser/Thr at position P+4). However, a Glutamic acid at this position could act as a priming site. If this is confirmed, it could greatly expand the number of potential substrates phosphorylated by GSK3.

At the presynapse GSK3 activity has been shown to inhibit vesicle fusion events and exocytosis in response to membrane depolarisation. GSK3β has been shown to phosphorylate and inhibit P/Q-type calcium channels, suppressing Ca2+ influx and interfering with the Ca2+ -dependent SNARE (SNAP receptor) complex responsible for synaptic vesicle fusion events in the synapse (Zhu et al., 2010). Interestingly, the GSK3 substrate CRMP2 interacts with the N-type calcium channel, CaV2.2, positively regulating its surface expression and enhancing Ca2+ currents (Brittain et al., 2011). Since GSK3 has been previously shown to antagonise CRMP2 functions (Cole et al., 2004b; Yoshimura et al., 2006; Yoshimura et al., 2005), its activity may also negatively regulate CaV2.2 surface expression and Ca2+ currents, hence down regulating neurotransmission, although this is yet to be determined. Furthermore, GSK3 activity is required for the retrieval of synaptic vesicles at the presynapse by activity-dependent bulk endocytosis (ADBE). Dynamin 1 is a GTPase that facilities re-uptake of neurotransmitters at the presynapse. GSK3 phosphorylates Dynamin 1 at Ser774 (following priming by Cdk5 at Ser778) and is required to activate Dynamin 1 to facilitate ADBE during elevated neural

43 |Page

Sec. 1.4. GSK3 substrates

activity (Clayton et al., 2010). Accordingly, inhibition of GSK3 in hippocampal slices was shown to relieve the depression of neurotransmission during high frequency stimulation. Taken together, GSK3 activity promotes presynaptic re-uptake of synaptic vesicles during times of elevated neuronal stimulation. The activation of ADBE may reduce the availability of synaptic vesicles in the short-term and hence down regulate neurotransmission. This may contribute to GSK3’s ability to antagonise LTP (Zhu et al., 2007).

Together, GSK3 activity down regulates neurotransmitter release from the presynapse and decreases postsynaptic response, restricting LTP. This may be associated with encoding and consolidation of learnt information during LTP (Peineau et al., 2009; Peineau et al., 2007).

1.4.6.4 Summary: GSK3 is a central regulator of learning and memory formation

Overall, GSK3 activity promotes and maintains LTD while negatively regulating LTP, implicating GSK3 in learning/memory formation. Reduced GSK3 activity (LTP) is critical for learning and memory formation/acquisition, while GSK3 activation (LTD) is required for memory reconsolidation. Interestingly, this may suggest that the dysregulation of GSK3 activity and disruption of LTP/LTD may contribute to deficits in learning and memory behaviour in neurological disorders such as mood disorders and the early stages of Alzheimer’s disease, before the later neurodegenerative aspects of the disease become dominant (Giese, 2009).

44 |Page

Sec. 1.5. GSK3 and mood disorders

1.5 GSK3 and mood disorders

GSK3 has been implicated in the development of BD. This debilitating psychiatric condition is characterised by recurring periods of manic and depressive episodes that can last days, weeks or even months and are accompanied by changes in activity/energy levels and associated with characteristic cognitive, physical and behavioural symptoms. Manic episodes is defined by severe and sustained elevation of mood characterised by abnormal excitability, energy, or irritability. Whereas, periods of depression are usually more common and longer lasting than elevated mood, and contributes most to overall patient morbidity (see review: (Anderson et al., 2012)). It affects approximately 1% of the Australian population (Slade et al., 2007), and is among the leading causes of disability in young patients, with an average age onset of 25 years (Merikangas et al., 2011). Deficits in emotion regulation underlie the neurobiology of BD, whereby it is hypothesised that abnormal activation of the limbic regions of the brain, important in the processing of memory, attention, motivation and emotional reactions, coupled with decreased activation of prefrontal circuitry, important in decision making and inhibition of a prepotent responses, are central to BD (Chen et al., 2011a; Hafeman et al., 2014). There are no clear genetic and environmental causes identified for mood disorders, although several genetic loci and environmental influences have been shown to confer relatively weak risk and susceptibility to this disease. Therefore, alternative approaches are required to elucidate the mechanism underlying these diseases and to develop improved treatment options for use in the clinic.

1.5.1 GSK3 is a main target in Bipolar Disorder

GSK3 has been linked to the pathogenesis of BD for seven main reasons:

1) GSK3 is a target of lithium, which has been the mainstay treatment for BD for over 50 years. It was initially discovered as a mood stabiliser by Australian psychiatrist John Cade in 1949 (Cade, 1949) and has since been shown to be especially effective at preventing manic episodes (Jope, 2003; Zhang et al., 2003) and suicide. Lithium inhibits GSK3 kinase activity both directly (Klein & Melton, 1996; Stambolic et al., 1996) and indirectly by promoting N-terminal serine phosphorylation of the enzyme (Chalecka- Franaszek & Chuang, 1999).

45 |Page

Sec. 1.5. GSK3 and mood disorders

2) Other more specific GSK3 inhibitors have been shown to replicate the behavioural effects of lithium in mice, further implicating GSK3 as a key pathogenic target in BD (Gould et al., 2004).

3) GSK3 is a common target of other antipsychotic and antidepressant drugs used to treat BD and schizophrenia patients (see reviews: (Beaulieu et al., 2009; Kalinichev & Dawson, 2011). This is likely mediated via the regulation of monoaminergic signalling in the brain (Fig.1.16). Indeed, GSK3 has been identified as a downstream target of dopaminergic (Beaulieu et al., 2007; Beaulieu et al., 2005; Beaulieu et al., 2004) and serotonergic signalling (Li et al., 2004) in neurons. Dopamine (DA) signalling is reported to stimulate the formation of a protein complex involving the dopamine receptor, β- arrestin, the protein phosphatase PP2A and Akt (Beaulieu et al., 2005). This complex facilitates dephosphorylation and inactivation of Akt by PP2A, thus reducing inhibitory phosphorylation of the GSK3 N-terminal domain (Ser21/9), increasing its kinase activity (Beaulieu et al., 2005). Therefore, high dopaminergic signalling in schizophrenia or the manic phase of BD (Yatham, 2003) promotes high GSK3 activity. Many antipsychotic drugs (e.g. haloperidol, risperidone, olanzapine, quetiapine, clozapine, and ziprasidone) antagonise dopaminergic signalling by blocking dopamine receptors (Beaulieu et al., 2009; Kapur et al., 1999; Meltzer et al., 1988), leading to decreased GSK3 activity (Alimohamad et al., 2005; Kang et al., 2004; Li et al., 2007). For example, mice with elevated DA signalling (either from administration of amphetamine (DA-receptor agonist) or from the lack of the DA transporter) exhibit reduced Akt activation and increased GSK3 activity, and this is associated with locomotor hyperactivity and stereotypic movements in these mice (Beaulieu et al., 2005; Beaulieu et al., 2004). Furthermore, the DA-dependent hyperactivity of these mice was reversed by inhibition of GSK3 with lithium and other specific GSK-3 inhibitors (Beaulieu et al., 2005; Beaulieu et al., 2004). Together this demonstrates that elevated GSK-3 activity is involved in the expression of DA-associated neuropsychiatric conditions, such as BD and schizophrenia, and that GSK3 is likely to be an important target of antipsychotic drugs.

Similarly, GSK3 activity is regulated by serotonin signalling. Mice expressing loss of function tryptophan hydroxylase 2 (TPH2), an enzyme for neuronal serotonin (5-HT) synthesis, exhibited ‘anxiety-like’ behaviours accompanied by a twofold increase in GSK3β activity in the brains of these mice (Beaulieu et al., 2009; Beaulieu et al., 2008). These behaviours were restored to normal level by genetic or pharmacological inhibition

46 |Page

Sec. 1.5. GSK3 and mood disorders

of GSK3β. Similarly, reduction of GSK3β signalling in 5-HT-depeleted mice also reduced associated aggressive behaviours and enhanced social investigation (Beaulieu et al., 2008). Several antidepressant drugs, such as selective serotonin reuptake inhibitors (SSRIs), monoamine oxidase (MAO) inhibitors and tricyclic antidepressants, elevate serotonin signalling in the brain and reduce GSK3 kinase activity (Li et al., 2004). This further supports the idea that reducing GSK3 activity is an effective treatment of mood disorders (Benedetti et al., 2011).

Figure 1.16. Summary of the involvement of GSK3 and monoaminergic signalling in mood disorders. In depression, deficiencies in signals that maintain inhibition such as serotonin signalling, causes up-regulation of GSK3 activity, promoting susceptibility to depression. Excessive dopaminergic signalling in mania, induces activation of GSK3. Therapeutic actions of antidepressant and mood stabilisers can directly and indirectly inhibit GSK3 activity. (Figure adapted from: (Jope, 2011)).

4) Genetic and pharmacological manipulation of GSK3 activity in mice produce behaviours that correlate with mood disorders and schizophrenia, such as anxiety and decreased social interactions. For example, transgenic mice overexpressing a constitutively active mutant form of GSK3β (S9A) in the brain, display increased locomotor activity and reactivity, whereas habituation was decreased (Prickaerts et al.,

47 |Page

Sec. 1.5. GSK3 and mood disorders

2006). This behaviour is consistent with hyperactivity observed in the manic phase of patients with BD. Consistently, mice with reduced GSK3 activity (GSK3β+/-; GSK3α-/- ) (Beaulieu et al., 2004; Beaulieu et al., 2008; Kaidanovich-Beilin et al., 2009) or when GSK3β is inhibited in the forebrain (Latapy et al., 2012), display behaviours that reflect chronic lithium treatment (Gould et al., 2004). The use of genetic and pharmacological therapies for manipulating GSK3 activity in the brains of mice demonstrate strongly that inhibition of GSK3 significantly reduces manic behaviours associated with BD (Li & Jope, 2010).

5) GSK3β gene variants have been associated with BD and schizophrenia (Benedetti et al., 2012). In particular, a single nucleotide polymorphism (SNP) (rs334558) identified in the GSK3β promoter region that reduces gene transcription is associated with later onset of BD and improved response to lithium and antidepressant treatments (Benedetti et al., 2004a; Benedetti et al., 2004b; Benedetti et al., 2005; Tsai et al., 2008). This same GSK3β variant was later found to influence brain structure and to be protective against temporal lobe grey matter loss in schizophrenia patients (Benedetti et al., 2010). Furthermore, an increase in copy number variation (CNV) of the GSK3β gene has been associated with patients with BD (Lachman et al., 2007). Together, these findings suggest that sequence variations in the GSK3β gene could influence expression that may have a functional consequence for susceptibility of individuals to BD and schizophrenia.

6) Upstream regulators of GSK3 activity (Akt1, DISC1, BDNF) are genetically associated with mood disorders and schizophrenia (Karege et al., 2010; Lipina et al., 2012; Mao et al., 2009; Neves-Pereira et al., 2002; Schumacher et al., 2005).

Akt: Akt1 protein levels are significantly reduced in post-mortem brain tissue from patients with schizophrenia compared to controls, along with a corresponding increase in GSK3 activity (Emamian et al, 2004). Furthermore, a two-SNP variant in the Akt1 locus was found to be associated with individuals suffering from schizophrenia (Emamian et al, 2004). This haplotype may be associated with the observed decrease in Akt1 levels in the brain and resulting reduction of GSK3β phosphorylation.

Disrupted-In-Schizophrenia-1 (DISC1): DISC1 is a strong genetic risk factor associated with major depression disorder (MDD), BD and schizophrenia (Lipina et al., 2012; Mao et al., 2009). DISC1 was originally associated with schizophrenia when a translocation mutation was detected that co-segregated with major psychiatric illness in

48 |Page

Sec. 1.5. GSK3 and mood disorders

a large Scottish family (Millar et al., 2000). Since then several SNPs in the DISC1 gene have also been associated with BD and schizophrenia (Chubb et al., 2007). The DISC1 protein is a scaffold that binds and inhibits the kinase activity of GSK3, preventing phosphorylation and degradation of β-catenin, and instead, promote its transcriptional activity in the cell nucleus (Mao et al., 2009). Loss of DISC1 reduced neural progenitor proliferation and produced hyperactivity and depressive behaviours in mice. These behavioural phenotypes were normalised by inhibiting GSK3β chemically (Mao et al., 2009). Other mouse models expressing truncated DISC1 mutants in the brain were found to have behavioural and cognitive disturbances, similar to schizophrenia, which were associated with altered glutamatergic and GABAergic neurotransmission (Holley et al., 2013). In summary, current evidence suggests that mutation/loss of DISC1 function predisposes individuals to increased risk of psychiatric illness by disrupting NPC differentiation, shifting the balance towards proliferation.

Brain Derived Neurotropic Factor (BDNF): Polymorphisms within the BDNF gene have been associated with BD, MDD and schizophrenia (Neves-Pereira et al., 2002; Schumacher et al., 2005). They are also associated with a decrease in prefrontal cortex grey matter volume (Pezawas et al., 2004), consistent with reduced grey matter volume in mood disorder and schizophrenic patients. BDNF protein levels are reduced in the hippocampus of depressed patients (Dwivedi et al., 2003) and elevated by treatment with antidepressants (Brunoni et al., 2008; Matrisciano et al., 2009; Shimizu et al., 2003). Finally, elevation of BDNF by infusion of recombinant human BDNF, produce an antidepressant-like effect in rat models of depression (Siuciak et al., 1997). Together, these observations demonstrate that low levels of BDNF are associated with psychiatric disorders.

BDNF may also act downstream of GSK3 since treatment with lithium and other selective GSK3 inhibitors increases BDNF levels in cultured neurons (Yasuda et al., 2007), rats (Fukumoto et al., 2001) and humans (Leyhe et al., 2009). This is believed to be mediated by CREB, a known regulator of BDNF transcription (Tao et al., 1998) and direct substrate of GSK3 (Grimes & Jope, 2001a). Indeed, reduced CREB phosphorylation has been observed in post-mortem brains of mood disorder patients treated with lithium (Young et al., 2004). Furthermore, reduced BDNF signalling in schizophrenic patients is accompanied by down regulation of Akt and up regulation of GSK3 activity (Emamian et al., 2004). Together, this suggests that elevated GSK3

49 |Page

Sec. 1.5. GSK3 and mood disorders

activity in BD and schizophrenic patients may be contributed by reduced BDNF signalling, further implicating BDNF-GSK3 signalling in the pathogeneses of BD and schizophrenia.

7) Elevated GSK3 kinase activity has been detected in human BD patients, without changes in GSK3 expression or abundance. For example, N-terminal inhibitory phosphorylation of GSK3 is reduced in BD patients and mouse-models of BD (Polter et al., 2010).

All together, these observations strongly implicate high levels of GSK3 activity in the pathogenesis of mood disorders.

1.5.2 GSK3 activity in the development of schizophrenia

GSK3 has also been implicated in the development of schizophrenia. GSK3β kinase activity was found to be increased in post-mortem brain tissue from schizophrenic patients (Emamian et al., 2004) and polymorphisms in the GSK3β promoter region have been reported to be associated with schizophrenia (Tang et al., 2013). Interestingly, a polymorphism in the promoter region that decreases GSK3 gene transcription is associated with reduced risk of developing schizophrenia (Benedetti et al., 2010). Furthermore, SNPs identified in DISC1 (Chubb et al., 2007), likely affects its expression and subsequent regulation of GSK3 activity in this disease (Lipina et al., 2012; Mao et al., 2009). Altogether, these observations suggest GSK3 is hyperactivated in schizophrenia, similar to BD. However, schizophrenic patients do not respond to lithium treatment. The reason for this is not clear, although it suggests that regulating GSK3 level of activity may not be critical to normalising psychosis. Instead, it is possible that increased GSK3 activity during brain development in adolescence may increase the risk of developing schizophrenia. Indeed, as described earlier (sections 1.4.4 - 1.4.5), GSK3 has key regulatory roles during neurogenesis, abnormal migration and positioning of developing neurons, disruptions of which may contribute to the etiology of schizophrenia (Austin et al., 2004; Kim et al., 2009a). This supports the hypothesis that impaired/dysregulated GSK3 signalling during neurodevelopment could increase susceptibility to schizophrenia later in life. Overall, further research is required to elucidate the precise role of GSK3 in the development of this debilitating disease.

50 |Page

Sec. 1.5. GSK3 and mood disorders

1.5.3 Circadian proteins

Mood disorders and schizophrenia are commonly associated with sleep impairments, including difficulty falling asleep, fragmentation of sleep, decreased total sleep and decreased sleep efficiency (Harvey et al., 2006). For example, BD patients typically require less sleep but display higher activity levels and appetite during their manic phase (Geller et al., 1998). Normalisation of the sleep/wake cycles is often essential for effective mood stabilisation, while irregular sleep timing and reduction in total sleep can trigger manic episodes (Boivin, 2000; Wehr, 1992). Similarly, schizophrenic patients often suffer from sleep-onset latency and problems maintaining sleep (Cohrs, 2008), while abnormal sleep and sleep deprivation can trigger or exacerbate psychosis (Manoach & Stickgold, 2009). Therefore, sleep and circadian rhythm abnormalities are important symptom/risk factors of mood disorders and schizophrenia.

The circadian clock is regulated by transcriptional activation of CLOCK/BMAL1 and repression by Period (PER)/Cryptochrome (CRY) protein complexes. These protein complexes form a cyclic feedback loop that acts as the primary circadian pacemaker in the suprachiasmatic nucleus (SCN) of the hypothalamus (King & Takahashi, 2000). Genetic studies have identified associations between circadian gene polymorphisms and BD or schizophrenia (see review: (Lamont et al., 2010)). For example, polymorphisms in the clock, Per3 and mammalian Timeless (Tim) genes are associated with schizophrenia (Mansour et al., 2006; Takao et al., 2007a), while polymorphisms in Bmal1, Tim and clock are associated with BD (Benedetti et al., 2007; Benedetti et al., 2003; Mansour et al., 2006; Serretti et al., 2003). Disruption of clock in mice induces manic-like behaviours, such as hyperactivity, decreased sleep and antidepressive effects (Roybal et al., 2007). Interestingly, mood stabilizers and antidepressants increase clock and Bmal1 expression in the hippocampus of humans (Manev & Uz, 2006). These observations suggest that circadian genes may be important in the development and treatment of mood disorders and schizophrenia.

GSK3 has been reported to interact and phosphorylate nearly all core clock proteins in Drosophila, mice and humans. For example, phosphorylation of CLOCK, BMAL1 and CRY2 by GSK3 decreases their stability and promotes their degradation (Harada et al., 2005; Sahar et al., 2010; Spengler et al., 2009), whereas phosphorylation of PER promotes its translocation to the nucleus (Ko et al., 2010). GSK3β has also been

51 |Page

Sec. 1.5. GSK3 and mood disorders

shown to phosphorylate and stabilize Rev-Erbα (NR1D1), a negative regulator of Bmal1 transcription and circadian rhythm (Yin et al., 2006). Transgenic mice overexpressing GSK3 exhibit severe fragmentation of sleep-wake cycle during both the day and night periods (Ahnaou & Drinkenburg, 2011), whereas haploinsufficiency of GSK3 in mice lengthens the circadian locomotor period of these mice (Lavoie et al., 2013). Furthermore, tissue-specific over expression of shaggy (sgg, homologue of GSK3) shortens the period of the Drosophila circadian rhythm (Martinek et al., 2001). Conversely, Drosophila sgg mutants lacking activity display lengthening of the circadian period (Martinek et al., 2001).

Lithium treatment lengthens the circadian period in Drosophila, mice and humans (Dokucu et al., 2005). Over-expression of GSK3 in clock cells of the Drosophila brain exacerbates lithium’s effect, lengthening the circadian period two-fold (Dokucu et al., 2005). Furthermore, chronic lithium treatment reduces manic-like behaviour in Clock mutant mice (Roybal et al., 2007). Therefore, GSK3 is a critical intrinsic regulator of the circadian clock and plays an important role in regulating circadian period in response to lithium treatment.

Overall, sleep and circadian rhythm abnormalities play an important role in the development of mood disorders and schizophrenia. GSK3-mediated phosphorylation regulates the expression of core clock components and its activity has been associated with shortened circadian period. Furthermore, common mood stabilisers, such as lithium and antidepressants that reduce GSK3 activity, up-regulate the abundance of circadian genes and lengthens the circadian period, further supporting a role for GSK3 activity in the etiology of mood disorders and schizophrenia.

1.5.4 Summary: GSK3 is important in the regulation of mood disorders/schizophrenia

Genetic and pharmacological evidence support a central role for GSK3 in the development of mood disorders and schizophrenia. Common psychiatric drugs, such as anti-depressants, mood stabilizers and antipsychotics have been shown to reduce GSK3 activity, implicating it as a therapeutic target. However, it is important to remember that it’s the downstream substrates of GSK3 that mediate its functional effects. Therefore, these substrates must be identified in order to fully elucidate the mechanisms underlying the pathogenesis of mood disorders and schizophrenia. This could in turn reveal new

52 |Page

Sec. 1.5. GSK3 and mood disorders

therapeutic targets downstream of GSK3, potentially providing greater specificity with fewer side effects than current treatments affecting GSK3 activity.

53 |Page

Sec. 1.6. Project hypothesis and aim

1.6 Project hypothesis and aim

1.6.1 Rationale

GSK3 has been implicated in several psychiatric diseases, including BD and schizophrenia. These debilitating diseases severely impair people’s lives and in severe cases lead to exclusion from society and suicide. At present there are no clear genetic or environmental causes known for these conditions, so other strategies are urgently required to develop improved treatments. Lithium has been the mainstay treatment for BD for over 50 years and has been especially effective at reducing suicidal tendencies. However, it has side-effects that limit its use and reduce compliance (adherence). Therefore, identifying downstream substrates of GSK3 and targeting them therapeutically could improve specificity for pathogenic effects while reducing side-effects.

One approach is to elucidate the mechanisms of action of current drug therapies and to use this as a basis from which novel therapeutic targets can be developed. Lithium and other mood stabilizers used for treating BD have been identified as inhibitors of the Ser/Thr kinase GSK3 in the brain. GSK3 expression is highest in the brain, particularly in regions that demonstrate highest neuronal plasticity, such as the hippocampus and cortex (Leroy & Brion, 1999a; Takahashi et al., 1994; Woodgett, 1990b; Yao et al., 2002). Accordingly, GSK3 regulate aspects of neuroplasticity and neurotransmission, providing a potential link between therapeutics and brain function/dysfunction in psychiatric illnesses (see reviews: (Cole, 2012, 2013a; Peineau et al., 2008)). However the pathogenic targets downstream of GSK3 are not yet known. We propose that pathogenic substrates of GSK3 could become novel therapeutic targets for improved treatment of mood disorders, potentially providing greater specificity and fewer side effects than current treatments today.

1.6.2 Hypothesis

Using bioinformatics, biochemical analysis and proteomics, our group discovered an enrichment of novel substrates of GSK3 involved in vesicular trafficking events (Fig.1.17). Given the importance of trafficking in synaptic transmission and also the role of GSK3 in regulating this process, we hypothesised that this could be an important class of protein regulated by GSK3 that is dysregulated in mood disorders. In particular, we have focused on two novel signalling targets of GSK3, the lipid kinase

54 |Page

Sec. 1.6. Project hypothesis and aim

phosphatidylinositol 4-kinase II alpha (PI4KII) and the AP2 kinase adaptin associated kinase-1 (AAK1). PI4KII and AAK1 have both been implicated in vesicle trafficking in the brain and at synapses, including vesicle formation, transport/delivery and sorting functions that are important during neurotransmission in the brain. PI4KII is highly expressed in neurons and synapses and has been implicated in vesicle trafficking between Golgi apparatus, endosomes and lysosomes (Craige et al., 2008; Guo et al., 2003; Salazar et al., 2005). Whereas, AAK1 has been shown to inhibit neuregulin/ErbB4 signalling which is genetically associated with mood disorders and schizophrenia (Jaaro-Peled et al., 2009; Kuai et al., 2011), as well as activating Notch signalling during clathrin- mediated endocytosis (Gupta-Rossi et al., 2011). Therefore, we hypothesise that these novel downstream targets of GSK3 could be important mediators of neurotransmission in healthy and diseased brains.

Figure 1.17. Candidate GSK3 substrates involved in vesicular trafficking. Candidate substrates of GSK3 were identified using a combination of bioinformatics analysis and phosphoproteomics. Human candidates were identified and their functional categories are shown as a pie graph. 45 candidates have previously been associated with endocytosis/vesicular trafficking. 28 contain the GSK3 consensus sequence [S/T]PPx[S/T]P, while a further 17 contain the slightly more relaxed consensus sequence [S/T]Pxx[S/T]P. (Analysis performed by Adam Cole, Neurological signalling and mood disorders group, Garvan Institute).

55 |Page

Sec. 1.6. Project hypothesis and aim

1.6.3 Aim and significance

The aim of this project is to characterise the molecular mechanisms controlling GSK3 mediated phosphorylation of PI4KII and AAK1, and to elucidate its physiological function in neurons. This will provide valuable information on the basic function of these kinases in the brain and identify potential links between GSK3, synaptic transmission, and impaired brain function in mood disorders. Identifying pathogenic substrates of GSK3 that are dysregulated in mood disorders would lead to more targeted medications while reducing adverse side-effects of current treatment, reducing multifactorial approaches of current treatments options and creating more tailored treatments for patients.

Mood disorders and psychiatric illnesses cost the Australian health system approximately 2.4 billion annually in direct care costs and loss of productivity (Ageing, 2013). These debilitating diseases affect patients, family, relationships and employment and in severe cases lead to suicide. Mental diseases, such as BD, can be complex and may result in delayed diagnosis and treatment, or inappropriate treatment that can have a negative impact on the course and severity of the disorder. Understanding the role of GSK3 in neurotransmission may help determine the pathology of mood disorders and potentially other neurological diseases in which GSK3 is dysregulated. Ultimately, this research has the potential to aid in the development of more effective treatment options for patients suffering from mood disorders. This could help improved patient outcomes, allowing them to lead more normal lives, while also helping to reduce costs on public health services.

56 |Page

CHAPTER 2: Experimental procedures

2.1 Materials

2.1.1 DNA constructs

The cDNA encoding the full-length mouse PI4KII (UniProt Q2TBE6) and AAK1 (UniProt Q2M2I8) was amplified by PCR from Image clone 5346917 (PI4KII) and Kazusa clone ORK00722 (AAK1) using designed primers (Table 2.1).

Table 2.1 – DNA constructs: primers and amplicons size (bp)

Target Primer sequence (5'-3')

PI4KIIα full-length (F) GAA TTC GCC ACC ATG GAC GAG ACG AGC CC (mouse) (R) GCC CTT CTT TTC ATG GTG GGA CTA CAA GGA

PI4KIIα G24 (F) GAA TTC GCC ACC ATG GGA GCT CAC TTT CCG CAA GTA CC (Truncation mutant)

PI4KIIα G52 (F) GAA TTC GCC ACC ATG GGC CAC GAC CGG GAG CG (Truncation mutant)

AAK1 full-length (F) GAA TTC GCC ACC ATG GAC TAC AAG GAC GAC (human) (R) GGC GAA TTC TTA AAT AGC CTT GGC

The PCR products were subcloned into a pRK5 (CMV promoter) for mammalian expression with a custom-generated 3’ (C-terminal) FLAG-tag, or into pEGFP with a 3’ (C-terminal) GFP-tag. All mutant forms of PI4KIIα and AAK1 were either synthesized by DNA2.0 and GeneArt (Life Technologies) or generated using QuikChange II site- directed mutagenesis kit (Stratagene) according to the manufacturer’s instructions. All constructs were verified by DNA sequencing. The shRNA sequence targeted to mouse and rat PI4KIIα (5’-GATTTGATTCTTCCAAAGA) was synthesized and cloned into pSUPER vector. The shRNA expression construct targeting human AAK1 was a kind gift from Dr. Neetu Gupta-Rossi, Institut Pasteur, Paris.

2.1.2 Phosphospecific antibodies

Phosphospecific antibodies were generated by injection of rabbits with synthetic peptides listed in Table 2.2, conjugated to keyhole limpet hemocyanin. Antisera from

57 |Page

Sec. 2.1. Materials

bleeds 2 (week 9) and 3 (week 11) were collected (performed by Pacific Immunology (Ramona, CA, USA)). Antibodies were affinity-purified on a phosphopeptide antigen- agarose column, eluted in 50 mM Glycine, pH 2.5, then immediately neutralised using 1 M Tris-HCL, pH 8. Immunoblotting and immunofluorescence analysis using purified phosphospecific antibodies were routinely performed in the presence of 1μM de- phosphopeptide to reduce nonspecific binding to de-phosphorylated PI4KII/AAK1 (or vice versa for the de-phosphospecific antibodies).

Table 2.2: Peptide antigens for antibody production Protein Peptide antigens for antibody production (where pSer is phosphoserine; pThr is phosphothreonine) PI4KIIα (mouse) pSer5-CMDETpSPLVS pSer9-CSPLVpSPERA pSer47-AGSGPpSPPC pSer51-SPPCpSPGHD de-pSer5-CMDETSPLVS de-pSer47-AGSGPSPPC AAK1 (human) pThr620-CKVGSLpTPPSSP pThr674-CKSATTpTPSGSP Peptides were synthesised by Mimotopes (Melbourne, Australia)

2.1.3 Reagents

Anti-FLAG monoclonal antibody, anti-FLAG agarose, anisomycin, cycloheximide, lithium and Bafilomycin A were purchased from Sigma. CT99021 was purchased from Selleck, MG132 from Calbiochem and 488-labelled transferrin (Tfn) from Life Technologies. Actin, GAPDH, Lamp1 polyclonal antibodies were purchased from Cell Signalling, GluA1 antibody from Millipore and adaptins α, β, γ, , EEA1 and clathrin antibodies were from Transduction Laboratories. Fluorescent secondary antibodies were supplied by Li-COR (for Western blotting) and Jackson ImmunoResearch Laboratories (for immunofluorescence microscopy).

2.2 Methods

2.2.1 Biochemical validation of GSK3 substrates

The BIPPS assay (Bioinformatic prediction of phosphorylated substrates) was used to validate candidate GSK3 substrates experimentally. Protein sequences tagged with a FLAG peptide were cloned into a mammalian expression vector (pRK5), expressed in HEK293 cells in the absence or presence of the highly specific GSK3 inhibitor

58 |Page

Sec. 2.2. Methods

CT99021 (2µM, 16h), and then immunoprecipitated from cell lysates using 10µl of anti- FLAG agarose. Purified substrates were washed and separately subjected to in vitro kinase assays with recombinant GSK3β (75 milliunits; Millipore) in kinase buffer containing 50 mM Tris-HCL, pH 7.5, 0.03% (v/v) Brij-35, 0.1% (v/v) β-mercaptoethanol and radiolabeled [γ-32P]ATP (PerkinElmer) (30°C, 0.5 h). Reactions were terminated by addition of SDS, subjected to SDS-PAGE and stained with Coomassie Brilliant Blue (CBR-250). Radiolabeled bands were visualized and the amount of 32P incorporated was quantitated using an FLA5000 Fuji Phosphorimager. This approach was previously published by our group (Farghaian et al., 2011).

2.2.2 Cell culture

HEK293 and HeLa cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% (v/v) heat-inactivated fetal bovine serum, 2 mM GlutaMAX, penicillin (50 units/ml), and streptomycin (100 units/ml) (all from Life

Technologies) at 37C in 5% CO2. HEK293 cells were transfected using DharmaFECT 1 lipsome transfection reagent (Dharmacon), while HeLa cells were transfected using Effectene (Qiagen) which was more effective at transfecting these cells (typically have lower transfection efficiency).

Primary cortical neurons were isolated from embryonic day 17 (E17) Sprague- Dawley rats. They were plated onto glass coverslips coated with high molecular weight poly-D-lysine (Millipore) and incubated at 37C, 5% CO2 in Neurobasal medium containing 2% (v/v) B27 serum replacement (Life Technologies), 2 mM GlutaMAX, penicillin (50 units/ml), and streptomycin (100 units/ml). Cortical neurons were transfected at 1 day in vitro (DIV) using calcium phosphate precipitation and harvested at 3 DIV. Primary hippocampal neurons from P2 C57BL6 mice were transfected before plating by electroportation using the Neon transfection system (Life Technologies) and maintained in Neurobasal medium containing 2% B27 serum replacement and 2 mM GlutaMAX. This hippocampal work was performed in collaboration with Dr Vladimir Sytnyk and his group at The University of New South Wales (UNSW).

2.2.3 Single-round endocytosis and recycling of Transferrin

HeLa cells were plated on glass coverslips coated with high molecular weight poly-D-lysine (50 ug/mL) and grown overnight until half-confluent. Cells were incubated

59 |Page

Sec. 2.2. Methods

in serum-free DMEM to deplete endogenous Tfn (3 h, 37C). The media was replaced with ice-cold serum-free DMEM, 1.0% (w/v) BSA containing 5 ug/ml Tfn conjugated with Alexa Fluor 488 (Life Technologies) to allow ligand binding to endogenous Tfn receptors (TfnR; 4C, 45 min). Excess Tfn was removed by washing in cold serum-free DMEM, then cells were incubated in pre-warmed serum-free DMEM at 37C to initiate trafficking. Endocytosis, trafficking and recycling of Tfn was stopped by washing cells in cold phosphate-buffered saline (PBS) at the indicated times, followed by fixing in 4% (w/v) paraformaldehyde (Henderson & Conner, 2007). To determine the amount of Tfn recycled we utilised fluorescence microscopy and manually scored the localisation of Tfn under blinded conditions (n=150 cells from randomly chosen fields; at least 3 independent experiments). Scoring was based on a 5 stage grading scale designed to describe the distribution of Tfn and the quantity remaining in individual cells after a set time point. Cells count towards the recycling pool if the majority (70%) of Tfn vesicles were located on the cell surface or cell extremities and missing from the cell body. Therefore, the greater the amount of Tfn localised to the cell surface and missing from the cell body, the higher the recycling score for that cell. The recycling pool is calculated by averaging the recycling score of 150 cells for each test group from randomly chosen fields and from at least 3 independent experiments. Statistical analyses were performed using paired Student’s t test, and results were considered significant when p < 0.05.

2.2.4 Immunofluorescence microscopy

HeLa cells and cortical neurons were fixed in 4% (w/v) paraformaldehyde, permeabilised with 0.5% (v/v) Triton X-100, blocked with 2% (w/v) BSA, and incubated with primary antibodies for 2 h at room temperature. Following washing, cells were incubated with fluorescent secondary antibodies and DAPI (4',6-Diamidino-2- Phenylindole; 1:500) for 1 h at room temperature. Image acquisition was performed on a Zeiss Axiocam mRm microscope (Zeiss, Germany) using a 63x objective. Image analysis was performed using Image J software (National Institute of Health).

Hippocampal neurons (12 DIV) were fixed with 4% (w/v) paraformaldehyde and treated for 20 min with blocking solution (PBS containing 0.3% BSA and 10% horse serum). GluA1 surface expression was detected by incubating with rabbit polyclonal GluA1 antibody (30 min) (Alomone Labs) and fluorochrome-coupled secondary antibodies (30 min) at room temperature under nonpermeabilised conditions. Images

60 |Page

Sec. 2.2. Methods

were acquired using a confocal laser-scanning microscope C1si, NIS Elements software and oil Plan Apo VC 60x objective (numerical aperture 1.4; all from Nikon). Image analyses were performed using Image J software.

2.2.5 Western blotting

Cultured HEK293 cells were rinsed in cold PBS and lysed in buffer containing 1% (v/v) Triton X-100, 50 mM Tris-HCL, pH 7.5, 0.27 M sucrose, 1 mM EDTA, 0.1 mM EGTA, 1 mM sodium orthovanadate, 50 mM Sodium fluoride, 5 mM sodium pyrophosphate, 0.1% (v/v) -mercaptoethanol and Complete protease inhibitor tablets (Roche Applied Science) (4C). Lysates were centrifuged to remove insoluble material and supernatants collected. Protein concentrations were determined using the Bradford assay (Sigma). Samples were mixed with SDS loading buffer. Equal concentrations were subjected to SDS-PAGE (150 Vh), then transferred to nitrocellulose membrane (Hybond- C Extra, Amersham Bioscience) using the XCell II blot module (52.5 Vh; Invitrogen). Membranes were blocked in 5% (w/v) skim milk powder in PBS for 1 h, then incubated with primary antibody overnight at 4C (phosphospecific antibodies at 1 g/ml in the presence of 1 M dephosphopeptide (vice versa for de-phosphospecfic antibodies); anti- FLAG antibody 1 g/ml). Following washing in Tris-buffered saline and 0.25% (w/v) Tween 20 (TTBS), membranes were incubated with fluorescent secondary antibodies (LI- COR) (in 1% w/v skim milk in TTBS), washed twice in TTBS, and visualized using a LI- COR Odyssey infrared imaging system. Densitometry analysis was performed using LI- COR Odyssey software or Image J software.

2.2.6 In vitro kinase assay

FLAG-tagged PI4KIIα protein was isolated from HEK293 cell lysate using anti- FLAG agarose. Following washing, PI4KIIα was subjected to in vitro kinase assays with recombinant GSK3 and/or Cdk5 (2.5 milliunits/μl) in kinase buffer containing 50 mM Tris-HCl, pH 7.5, 0.03% (v/v) Brij-35, 0.1% (v/v) β-mercaptoethanol, and unlabeled ATP (30C, 30 min). Reactions were terminated by addition of SDS loading buffer and proteins were visualized by Western blotting.

2.2.7 In vitro phosphatase assay

HEK293 cells were lysed and harvested in lysis buffer without phosphatase inhibitors (i.e. 1% (v/v) Triton X-100, 50mM Tris-HCL, pH7.5, 0.27M sucrose, 1mM

61 |Page

Sec. 2.2. Methods

EDTA, 0.1 mM EGTA, 0.1% (v/v) -mercaptoethanol, and Complete protease inhibitor tablets (4C)). Following centrifugation to remove insoluble material, supernatants were collected, and protein concentrations determined using the Bradford assay. Lysates, containing a mixture of endogenous phosphatases were incubated at 30C for up to 4 h in the presence of 10 mM MgCl2 or CaCl2 to activate endogenous phosphatases. Reactions were terminated by the addition of SDS loading buffer and proteins were visualized by Western blotting. Alternatively, FLAG-tagged protein was isolated from HEK293 cell lysate using anti-FLAG agarose and resuspended in the buffer described above (without phosphatase inhibitors) followed by the addition of different amounts of recombinant protein phosphatase 1 (PP1; New England Biolabs; 30C, 30 min) and its cofactor 10 mM

MnCl2. Reactions were terminated by addition of SDS loading buffer and proteins were visualized by Western blotting.

2.2.8 Source and maintenance of flies

Flies were cultured at 25C, 70% humidity, 12 h:12 h light:dark (LD) cycle in vials containing normal chow food (6.5% molasses (w/v), 4.1% cornmeal (w/v), 1.3% yeast (w/v), 0.85% agar (w/v), 1.2% nipagin (w/v) (H5501, Sigma), 0.6% propoic acid (w/v) (402907, Sigma)). All assays started 2 h before dark cycle (zeitgeber time 10 h (ZT10)). PI4KII (CG2929) mutant flies [Mi(ET1)PI4KII: (29058); PBac(PB)PI4KII: (10833)] were obtained from the Bloomington Drosophila Stock Center. PI4KII transgenic RNAi (110687 and 25459) and AAK1 (Nak; CG10637) transgenic RNAi (109507) fly lines were obtained from the Vienna Drosophila RNAi Center (VDRC) (Dietzl et al., 2007). AAK1 (NAKJ35) RNAi fly was a kind gift from Dr. Cheng-Ting Chien, Institute of Molecular Biology, Academia Sinica, Taipei. PINK1 transgenic RNAi fly was a kind gift from Dr Greg Neely, Garvan Institute. UAS-RNAi transgenic flies targeting genes: synaptojanin (CG6562), PTEN (CG5671), PCTK1 (CG10579), SGIP1 (CG8176), EFR3B (CG8739), intersectin-2 (CG1099), CACNA1c (CG4894), Ube4B (CG9934), JAZF1 (CG12054), Bcl-11a (CG9650) and ZEB1/2 (CG1322) were obtained from the VDRC. Pan-neuronal expression of RNAi transgenes used the GAL4/UAS system under the control of the neuronal-specific synaptobrevin (nSyb-Gal4) driver with UAS-Dicer-2 to enhance RNAi target degradation (dcr2/yhh;;n-Syb-Gal4). Eye-specific expression of RNAi transgenes used the glass multiple reporter (GMR-Gal4) driver and for dopaminergic expression the dopa decarboxylase (ddc-Gal4) driver. All fly

62 |Page

Sec. 2.2. Methods

maintenance and experiments were performed in Dr Greg Neely’s lab in the Neuroscience Division, Garvan Institute.

2.2.9 Locomotor activity and sleep analysis

Male flies 3-5 d old were individually placed into clear plastic vials (25 mm x 95 mm) in the Drosophila activity monitor system (DAMs, Trikinetics) in LD conditions at constant temperature and humidity. Vials contain sugar-agar food (4% sucrose (w/v), 2% agar (w/v)) approximately 15 mm deep and the food end is sealed with a plastic cap to prevent desiccation. Once flies are loaded, the open end of the vial is plugged with cotton. Following one full day of adaptation, activity and sleep data was analysed for 6 days of baseline LD. Individual fly waveforms were calculated from the 6 d of analysis and normalized to 24 h; group waveforms were then calculated by averaging those individual waveforms. Locomotor activity was measured as the number of counts in 30 min bins. Sleep was defined as any period of uninterrupted behavioural immobility (0 counts) lasting >5min (Huber et al., 2004). The duration of sleep episodes was calculated by counting the number of consecutive 5 min periods of sleep. All behavioural experiments were reproduced at least three times and averaged.

2.2.10 Body weight and triacyglyceride (TAG) analysis

Male flies 4 d old were collected in 1.5 mL eppendorf tubes and weighed using an analytical balance (ATX224, Shimadzu). Flies were euthanased/snap-frozen in liquid nitrogen and temporarily stored at -80C. Files were placed into a 96-well plate (Eppendorf) on ice (10 files per well) and 200 uL of cold Milli-Q water was added to each well. Flies were homogenised using a Pellet Pestle Motor (Kontes) followed by sonication using a probe sonicator (5 kHz, 10 sec; Heat Systems). Additional cold Milli-Q water (800 uL) was added, the plate sealed and mixed by inversion. 50 uL of each sample (1:19 dilution), blank (water) control and a glycerol calibrator of known amount (22.9 nmol/L) were added to a new 96-well plate (in triplicate). 300 uL of Triglyceride reagent GPO- PAP (TG; Roche) was added to each sample (including controls and blanks), mixed by inversion and incubated at 37C, 30 min (mixed by inversion at 15 min interval). The plate was centrifuged (3 min), 100 uL of each sample was transferred to a new 96 well plate and the absorbance was measured at 505 nm (FLUOstar Omega spectrophotometer; BMG Lab Tech). TAG levels were calculated by comparing against the glycerol calibrator.

63 |Page

Sec. 2.2. Methods

2.2.11 Capillary Feeder (CAFE) assay

The CAFE assays were performed as previously described with modifications (William et al., 2007). Male flies 3-5 d old were collected and placed in vials with wet tissue paper as a water source and kept in LD conditions at constant temperature and humidity. Vials were sealed with a rubber stopper (LH-3178-1, Wiltronics) containing one trimmed 200 uL pipette tip to hold the glass capillary (708707, Brand). One capillary containing 5 uL of liquid food (5% yeast extract (w/v), 10% sucrose (w/v)) was provided per CAFE chamber and replaced every 12 h. For each line tested, 3-5 CAFE vials containing 7 files each were used and cumulative food intake data was recorded every 12 h for two days (48 h).

Alternatively, male files were provided with two capillaries per CAFE chamber containing 5 uL of liquid food (5% yeast extract (w/v), 0.25 M Sorbitol), supplemented with a choice of either high (2.5% (w/v) sucralose) or low (0.1% (w/v) sucralose) sucralose concentration to adjust the intensity of sweetness. Both food solutions contain the same caloric content with variation in sweetness using the non-caloric sucralose. Capillaries were replaced every 12 h and for each line test, 3-5 CAFE vials containing 10 flies each were used. Cumulative food intake data was recorded every 12 h for two days (48 h).

2.2.12 Starvation and nicotine toxicity assay

Male flies 3-5 d old were placed in vials (10-15 flies per vial; n=3) with wet tissue paper (water source) and no food source to test longevity during starvation stress, or tissue paper soaked in 1 mL of 6.16mM Nicotine-S (N3876, Sigma), 50mM sucrose solution to test resistance to neurotoxicity (sucrose provides an energy source and also masks the bitter taste of the nicotine). Experiments were performed under LD conditions at constant temperature and humidity. The number of deaths were recorded every 4 h for the first 12 h and then every 12 h until all flies had perished.

2.2.13 Startle-induced negative geotaxis assay (climbing assay)

Male flies 3-5 d old were collected and placed in vials (< 12 flies per vial; n=3). The vial was placed at eye-level and a height of 5 cm was marked on a vertical surface directly behind the vial to allow accurate height measurement. Flies were gently tapped to the bottom of the vial and the number of flies that climbed above 5 cm was counted

64 |Page

Sec. 2.2. Methods

after 10 s of climbing. This experiment was reproduced at least three times and averaged. Flies have a natural tendency to climb to the top of the vial and an inability to do this within 10 s is a measure of locomotor dysfunction.

2.2.14 Total RNA extraction and cDNA synthesis

Male flies 3-5 d old were placed in 1.5 mL tubes (~45 files per tube; n=2), euthanased/snap-frozen in liquid nitrogen and temporarily stored at -80C. RNA isolation, cDNA synthesis, and quantitative PCR were performed on whole flies or from dissected fly heads. Flies were homogenised in 1.5 mL RNase-free tubes containing TRIzol reagent (Life Technologies) using a 1.5 mL pestle (Sigma) and incubated at room temperature for 5 min. Chloroform (200 uL per mL of TRIzol reagent used) was added, mixed vigorously by shaking for 15 sec and incubated for a further 3 min at room temperature. Samples were centrifuged at 12,000 x g (15 min, 4C). The aqueous phase containing RNA was carefully removed and placed into a new 1.5 mL tube. 100% isopropanol (0.5 mL per 1 ml of TRIzol reagent used) was added to the aqueous phase, incubated for 10 min at room temperature. RNA was then pelleted by centrifugation at 12,000 x g (10 min, 4C), supernatant discarded and the RNA pellet was washed twice with 75% ethanol. Total RNA was resuspended in 20 uL water and stored (-20C) before processing. The quality of the RNA was assessed with a Nanodrop ND-1000 spectrophotometer (Nanodrop Technologies). Samples with a 260/280 ratio superior to 1.8 and a 260/230 ratio superior to 2.0 were considered appropriate. Total RNA samples (5 ug) were used to synthesise cDNA (Oligo(dT) 15 primer) using GoScript Reverse Transcription System (Promega). Reverse transcription reactions were performed using a Veriti Thermal Cycler (Applied Biosystems) in 15 uL total volumes under the following sequential conditions: annealing (25C, 5 min), extension (42C, 1 h) and a final reverse transcriptase inactivation step (70C, 15 min). cDNA was stored at -20C until used.

2.2.15 qPCR

Real-time quantitative PCR (qPCR) was performed using a Power SYBR Green PCR Master Mix (Life Technologies) in a Roche LightCycler 480 thermal cycler (Roche). Amplification reactions were performed in 10 uL total volumes in 384-well optical plates (Roche) under the following sequential conditions: 95C for 10 min, followed by 45 cycles of denaturing (95C, 15 s), annealing (58C, 30 s) and extension (72C, 30 s).

65 |Page

Sec. 2.2. Methods

Primer sequences are shown in Table 2.3. Relative quantification was calculated using the comparative CT (cycle threshold) method as previously described (Livak & Schmittgen, 2001).

To do this, CT is calculated for both test sample and calibrator sample according to the equation:

CT = CT(target) - CT(reference)

The CT values of the test is then normalised to the CT of the calibrator according to the equation:

CT = CT(test) - CT(calibrator)

Finally, expression ratio is calculated:

2-CT = Normalised expression ratio

Target gene expression is normalised to the relative expression of Drosophila actin42A as the reference control. qPCR efficiency was determined for each gene and calculated according to the equation: E=(10[-1/slope]-1) x 100 (Radonić et al., 2004). Sizes of PCR products were checked with gel electrophoresis to confirm identities.

Table 2.3 – Primer sequences for qPCR, melting Tm and amplicon sizes.

Target Primer sequence (5'-3') Melting Tm. Amplicon (Nearest size neighbour)

PI4KIIα (CG2929) F GAG ATC AAC ACC GTC GGC AT 58C 202 bp (NM_169063.4) R CCA AGT TAT CCA GTT GCG GT 56C

Actin42A (CG12051) F AAG CTG CAA CCT CTT CGT CA 58C 138 bp (NM_078901.3) R GCG TCG GTC AAT TCA ATC TT 55C

2.2.16 Statistical analysis

All quantitative data was generated from at least three biological replicates. Data was analysed by unpaired, two-tailed t-test or two-way ANOVA. All data shown are means  S.E.M.

66 |Page

CHAPTER 3: PI4KIIα phosphorylation by GSK3 directs vesicular trafficking to lysosomes

3.1 Introduction

Vesicular and organelle membranes are defined by unique lipid and protein compositions which determine vesicle formation, transport/sorting and delivery and are therefore essential in the regulation of neurotransmission in the brain (Salazar et al., 2009). Phosphatidylinositol (PI) lipids located at the plasma and internal membranes are important signalling and second messenger molecules for regulating vesicular transport in cells (Antonietta De Matteis et al., 2005; Guo et al., 2003). Distinct organelles/cellular subdomains contain different sets of PI lipids and this distribution is finely controlled by different PI kinases and phosphatases (Krauß & Haucke, 2007). Among these are PI4- kinases (PI4Ks) that phosphorylate the D-4 position of the inositol ring (Chu et al., 2010; Minogue et al., 2010). PI4P is predominantly located in the membranes of the Golgi apparatus where it plays important roles in regulating membrane biogenesis, lipid dynamics and vesicle budding (Antonietta De Matteis et al., 2005; Graham & Burd, 2011; Guo et al., 2003). PI4P is also an important precursor for synthesis of other PIs, including

PI(4,5)P2, PI(3,4)P2 and PI(3,4,5)P3 (Antonietta De Matteis et al., 2005). These are also important second messengers for regulating protein interactions with the lipid bilayer in vesicle budding, cytoskeletal organisation and signalling (Antonietta De Matteis et al., 2005; Guo et al., 2003). In particular, it was recently shown that the pool of PI4P in the

Golgi is important for supplying PI(4,5)P2 at the plasma membrane, necessary for subsequent endocytosis and trafficking functions (Dickson et al., 2014).

3.1.1 PI4-kinases

There are two types of PI4Ks in mammalian cells (Type II and III) and each type is comprised of two isoforms, α and β (therefore four in total). Only type III PI4Ks are sensitive to the lipid kinase inhibitor Wortmannin; the basis by which type II and III PI4Ks are distinguished. Type II PI4Ks are smaller proteins (~54kDa) that contain a highly conserved catalytic domain from yeast to humans (Barylko et al., 2001; Barylko et al., 2002). A conserved cysteine-rich domain is located within the catalytic domain that is palmitoylated, keeping the protein strongly attached to membranes near its PI substrate (Balla & Balla, 2006; Barylko et al., 2001; Barylko et al., 2009). Despite their similar

67 |Page

Sec. 3.1. Introduction

structure and catalytic activity, type II PI4Ks have different localisations in cells and perform distinct functions (Chu et al., 2010; Minogue et al., 2010). PI4KIIα is ubiquitously expressed in mammals, with higher levels in the brain, kidney, stomach and lung tissues (Li et al., 2010) and accounts for the majority of PI4K activity in the brain (Guo et al., 2003). PI4KIIα has been shown to be tightly bound to membranes of the Golgi complex and endocytic vesicles. Recently, analysis of its crystal structure revealed an additional putative membrane binding pocket that may contribute to the binding of PI4KIIα to membranes. If so, this supports a novel mechanism that may control PI4KIIα’s recruitment to membranes and correspondingly its function (Baumlova et al., 2014). In contrast, PI4KIIβ is distributed evenly between peripheral membranes and the cell cytosol. Cytosolic PI4KIIβ is catalytically inactive, but becomes activated upon recruitment to membranes (i.e. plasma membrane, endoplasmic reticulum, and the Golgi) (Antonietta De Matteis et al., 2005; Jung et al., 2011; Jung et al., 2008; Wei et al., 2002).

3.1.2 Role of PI4KIIα in vesicular trafficking

Relatively little is known about the function of PI4KIIα. It has been implicated in diverse roles including, Wnt/β-catenin signalling (Pan et al., 2008b; Qin et al., 2009), cell survival (Chu et al., 2010; Minogue et al., 2010) and more recently, tumor progression (Li et al., 2010). It preferentially localises to Golgi and endocytic compartments, including late endosomes and vesicles rich in adaptor complex-3 (AP-3) (and to a lesser extent with adaptor complex AP-1) (Craige et al., 2008; Li et al., 2010; Salazar et al., 2005). It binds to clathrin-coated vesicles (CCVs) and vesicle-associated membrane protein 3 (VAMP3), which are important mediators of endosomal recycling and retrograde transport (Jović et al., 2014).

Knockdown of PI4KIIα decreases Golgi PI4P levels, which prevents recruitment of the AP-1 adaptor complex to the Golgi (Salazar et al., 2005; Wang et al., 2003) and disrupts Golgi structure and secretory function (Di Paolo & De Camilli, 2006; Wang et al., 2003). For example, it reduces export of the constitutively secreted proteins, HA (influenza virus hemagglutinin protein) and VSVG (vesicular stomatitis virus G protein) from the trans-Golgi network (TGN) in a PIP2 dependent manner (Wang et al., 2003). Interestingly, RNAi mediated knockdown of PI4KIIα selectively reduces clathrin levels at the Golgi (Wang et al., 2003) and vice versa (Salazar et al., 2009). Together, these observations support an essential role for PI4KIIα in the synthesis of Golgi PI4P and PIP2

68 |Page

Sec. 3.1. Introduction

lipids that are required for efficient Golgi secretory function (Krauß & Haucke, 2007; Martin et al., 1997; Wang et al., 2003).

PI4KIIα has been shown to regulate membrane recruitment and sorting functions of the AP-3 adaptor protein complex. AP-3 transports membrane proteins from endosomes to lysosomes or lysosome related organelles (Salazar et al., 2009). siRNA knockdown of PI4KIIα in PC12 cells redistributes AP-3 complexes from membranes to cytosol (Craige et al., 2008). Meanwhile, PI4KIIα abundance at the perinucleus, peripheral organelles (e.g. the Golgi complex) and nerve terminals is significantly reduced in fibroblasts from AP-3 deficient mocha mice (Craige et al., 2008; Salazar et al., 2009). These observations suggest that PI4KIIα and AP-3 influence each other’s abundance and subcellular distribution. PI4KIIα contains a dilucine sorting motif (ExxxLL) that mediates binding to AP-3 (Craige et al., 2008). This binding promotes the trafficking of Tfn to late endosomes, as well as endosomal trafficking and degradation of EGF receptor by the lysosomal pathway (Balla et al., 2002; Minogue et al., 2006; Salazar et al., 2009). Furthermore, PI4KIIα kinase activity has been shown to regulate delivery of AP-3-dependent cargo molecules to lysosomal compartments. For example, in PC12 cells and PI4KIIα kinase-dead HEK293 cells, PI4KIIα kinase activity enhances AP-3 dependent recruitment and abundance of ZnT3 cargo in AP-3 derived vesicles (Craige et al., 2008; Salazar et al., 2005). Together, these observations show that PI4KIIα resides in organelles containing AP-3 and it promotes membrane recruitment, trafficking and sorting of AP-3 complexes.

3.1.3 PI4KIIα role in neurotransmission

As well as the peri-nuclear/endosomal compartment, PI4KIIα is enriched at synapses in the brain, where it accounts for the bulk of PI4K activity (Guo et al., 2003). Pharmacological inhibition of PI4K activity attenuates glutamate release from the pre- synapse (Wiedemann et al., 1998). Its PI product, PIP2 is likely to be important for this process, since it is concentrated at the plasma / synaptic membrane and is important for synaptic vesicle endocytosis and recycling (Guo et al., 2003). Interestingly, down regulation of PI4KII has been associated with defective axonal transport. PI4KII knockout mice present an age dependent phenotype similar to human hereditary spastic paraplegia (HSP), which is characterized by degradation of long ascending and descending spinal cord axons, consistent with defective axonal transport (Simons et al.,

69 |Page

Sec. 3.1. Introduction

2009). Together these results suggest an important role for PI4KII activity at synapses and in axons.

3.1.4 Aims

Overall, PI4KII and its PI products have been identified as important regulators of vesicular trafficking. This process is particularly important in the brain, since it regulates neurotransmitter release at pre-synapses and expression of neurotransmitter receptors at the surface of post-synapses (Guo et al., 2003). Therefore, PI4KII is a potential regulator of neurotransmission in the brain. Here we demonstrate that PI4KII is a novel substrate of GSK3 in the brain and investigate the physiological function of its phosphorylation in neurons.

The following biochemical and cell culture assays (Fig.3.1-3.13) were recently published. Robinson, J. W., Leshchyns'ka, I., Farghaian, H., Hughes, W. E., Sytnyk, V., Neely, G. G., & Cole, A. R. (2014). PI4KIIalpha phosphorylation by GSK3 directs vesicular trafficking to lysosomes. Biochem J, 464(1), 145-156.

70 |Page

Sec. 3.2. Results

3.2 Results

3.2.1 Mapping and characterisation of phosphosites in PI4KIIα

PI4KIIα contains two conserved GSK3 phosphorylation consensus sequences in its N-terminal region, whereby Ser5 and Ser47 are the putative GSK3 target sites, while Ser9 and Ser51 are their respective priming sites (Fig.3.1A). To determine which of these sites are targeted by GSK3 in cells, we subjected full-length and truncated forms of PI4KIIα lacking the N-terminal 24 amino acid residues (removing the Ser5-9 site but not the Ser47-51 site), to the BIPPs assay previously developed in our lab for identifying novel GSK3 substrates (see Methods, section 2.2.1) (Farghaian et al., 2011). In brief, PI4KIIα was cloned into a mammalian expression vector and expressed in HEK293 cells in the presence or absence of the highly-specific GSK3 inhibitor CT99021 (Bain et al., 2007). It was then pulled-down via its C-terminal GFP-tag and subjected to in vitro kinase assays with recombinant GSK3β and radiolabelled ATP. If it is a physiological target of GSK3, transfection into HEK293 cells should result in phosphorylation by endogenous GSK3. This will be blocked by the GSK3 inhibitor, leaving the GSK3 target sites vacant but any required priming events intact. In the subsequent in vitro kinase assay, recombinant GSK3 should be able to incorporate more radiolabelled phosphate into PI4KIIα isolated from inhibitor-treated cells compared to untreated cells. If so, this indicates PI4KIIα is a good substrate for GSK3 in vitro and in cells. Indeed, significantly more radiolabelled phosphate was incorporated into full-length PI4KIIα isolated from inhibitor-treated (Fig.3.1B, C; lane 2) than non-treated cells (lane 1), indicating it is a bona fide GSK3 substrate. This was blocked in the truncated form of PI4KIIα (G24) (lane 3), indicating the majority of phosphorylation occurred at the Ser5-9 site, not the Ser47-51 site.

Phosphospecific antibodies were generated to all four putative phosphosites (Ser5, 9, 47 and 51) and their specificity was validated by Western blotting of wild type and phospho-deficient forms of PI4KIIα (Fig.3.2A). Ser5 was confirmed as a GSK3 target site, since phosphorylation was reduced by treatment with CT99021 (Fig.3.2A), lithium (Fig.3.2B) and mutation of the priming site Ser9 to non-phosphorylatable alanine (Fig.3.2A). Phosphorylation of Ser47 was more complicated. While it was inhibited by mutation of Ser51, therefore priming-dependent and consistent with being a GSK3 target site, it was not affected by CT99021 (Fig.3.2A, C) or lithium treatments (Fig.3.2B). A

71 |Page

Sec. 3.2. Results

possible explanation is that pSer47 is relatively resistant to de-phosphorylation by phosphatases, as found previously for other GSK3 substrates (e.g. β-adducin (Farghaian et al., 2011); CRMP2 (Cole et al., 2008)).

Figure 3.1. PI4KIIα is a physiological substrate of GSK3. (A) Sequence alignment of putative GSK3 target sites in the N-terminal region of PI4KIIα from various species. GSK3 sites and the priming sites at the +4 position are numbered and underlined. (B) PI4KIIα was expressed in HEK293 cells that were untreated (lane 1), treated with CT99021 (lane 2) or expressed a truncated form of PI4KIIα lacking the N-terminal 24 amino acids (ΔG24; lane 3). Full-length and truncated PI4KIIα was immunoprecipitated via its C-terminal GFP-tag and subjected to an in vitro kinase assay with recombinant GSK3 and radiolabelled ATP. (C) The stoichiometry of phosphate incorporation is shown as a graph (Mean ± S.E.M.,* = p<0.05; (T-test), n=3).

72 |Page

Sec. 3.2. Results

Figure 3.2. Mapping and comparison of PI4KIIα phosphosites. (A) HEK293 cells transfected with PI4KIIα wild type and phosphosite mutants, immunoprecipitated and subjected to Western blot analysis using custom phosphospecific antibodies, as well as an antibody recognizing the C- terminal GFP-tag as a loading control. Also, wild type PI4KIIα was expressed with/without CT99021 (2 μM). (B) HEK293 cells were transfected with GFP or PI4KIIα wild type, then treated with 20 mM NaCl (control) or 20 mM LiCl for 4 h. Lysates were subjected to Western blotting using the pSer5, pSer47, FLAG and actin antibodies. (C) Primary rat cortical neurons were treated with 2 μM CT99021 for 0, 1, 4, 24 h and subjected to Western blotting. Endogenous PI4KIIα was detected using the pSer5 and pSer47 antibodies.

73 |Page

Sec. 3.2. Results

3.2.1.1 Relative resistance of PI4KIIα phosphosites to phosphatases.

To determine if GSK3 phosphosites in PI4KIIα are resistant to dephosphorylation by phosphatases, HEK293 cells were transfected with PI4KIIα and harvested in lysis buffer without phosphatase inhibitors. Lysates were incubated at 30°C for up to 4 h with added CaCl2 or MgCl2 to activate endogenous phosphatases. Relative rates of dephosphorylation at each site were measured using Western blotting. HEK293 cells were also transfected with another GSK3 substrate called Cdc42EP4 (GSK3 phosphorylates Ser136) as a positive control, and lysates were subjected to the same phosphatase assay. Fig.3.3A shows that although there was rapid and complete dephosphorylation of pSer136 in Cdc42EP4, there was no change in phosphorylation of Ser47 and relatively minor dephosphorylation of pSer5 (~10%). In a separate experiment, FLAG-tagged PI4KIIα immunoprecipitated from HEK293 cell lysate was resuspended in lysis buffer without phosphatase inhibitors, followed by addition of different amounts of recombinant protein phosphatase 1 (PP1) (Fig.3.3B). Both Ser5 and Ser47 GSK3 phosphosites display relative resistance to dephosphorylation by PP1, whereas for comparison, the GSK3 phosphosite on Cdc42EP4 was completely dephosphorylated by 1 unit of PP1. Together, these observations show that pSer5 and pSer47, are both relatively resistance to dephosphorylation by phosphatases in vitro.

To further explore dephosphorylation of Ser5 and Ser47 in intact cells, a dose- response assay using increasing concentrations of CT99021 on PI4KIIα-transfected HEK293 cells was performed. Low concentrations of CT99021 were sufficient to reduce phosphorylation of Ser5, but not Ser47 (Fig.3.4A, B). This indicates that Ser47 is relatively more resistant to dephosphorylation than Ser5 in intact cells. It would therefore be expected that under basal conditions, Ser47 has a higher stoichiometry of phosphorylation than Ser5. Antibodies were generated that specifically recognise Ser5 and Ser47 when non-phosphorylated. Mutation of Ser9 and 51 priming sites completely blocks phosphorylation of Ser5 and Ser47, respectively (Fig.3.2A), but causes a much larger increase in staining using the non-pSer47 antibody compared to non-pSer5 (Fig.3.4C). This indicates the stoichiometry of phosphorylation is higher for Ser47 compared to Ser5 due to its relatively high resistance to dephosphorylation by phosphatases. We attempted to identify the endogenous phosphatases targeting pSer5 using pharmacological inhibitors of different phosphatases, as previously performed for CRMP2 (Cole et al., 2008). However, due to the relatively slow turnover of

74 |Page

Sec. 3.2. Results

phosphorylation at this site and the rapid toxicity of these inhibitors (<45 min), this experiment was not possible.

In summary, both Ser5 and Ser47 are relatively resistant to phosphatases in vitro, while only Ser5 is more efficiently dephosphorylated inside cells. The reason for the latter is not yet clear. It’s possible that an unknown binding protein or association with lipid membranes facilitates dephosphorylation by phosphatases, although this remains to be proven.

Figure 3.3. PI4KIIα phosphosites are relatively resistant to dephosphorylation by phosphatases. (A) HEK293 cells were transfected with PI4KIIα or Cdc42EP4 and lysates (without phosphatase inhibitors) were incubated at 30°C for up to 4 h in the presence of 10 mM MgCl2 or CaCl2 to activate endogenous phosphatases. Dephosphorylation by endogenous phosphatases was determined by Western blot analysis using phosphospecific antibodies (pSer5, pSer47 for PI4KIIα; pSer136 for Cdc42EP4), as well as an antibody recognising the C-terminal FLAG-tag as a loading control. (B) HEK293 cells were transfected with PI4KIIα or Cdc42EP4, immunoprecipitated via their C-terminal FLAG-tags, resuspended in lysis buffer (without phosphatase inhibitors) and incubated with different amounts of recombinant PP1 plus its cofactor

75 |Page

Sec. 3.2. Results

Mn2+ (30°C, 30 min). Dephosphorylation was determined by Western blotting using phosphospecific antibodies (pSer5, pSer47 for PI4KIIα; pSer136 for Cdc42EP4), as well as an antibody recognising the C-terminal FLAG-tag as a loading control.

Figure 3.4. Kinetics of dephosphorylation of PI4KIIα in intact cells. (A) HEK293 cells transfected with wild type PI4KIIα were treated with various concentrations of CT99021 for 4 h. Lysates were subjected to Western blotting using pSer5, pSer47 and FLAG antibodies. (B) Relative dephosphorylation induced by CT99021 treatment in A is presented as a graph (Mean ± S.E.M.,* = p<0.05; (T-test), n=3). (C) HEK293 cells transfected with PI4KIIα wild type or S9/51A were subjected to Western blotting using antibodies that recognise Ser5 and Ser47 when non-phosphorylated, as well as an antibody for the C-terminal GFP tag as a loading control.

76 |Page

Sec. 3.2. Results

The priming phosphosites at Ser9 and Ser51 both fit the consensus sequence of Cdk5 (i.e. proline-directed sites with a basic residue at P + 3). Cdk5 is known to prime other GSK3 substrates (e.g. CRMPs (Cole et al., 2006) and β-adducin (Farghaian et al., 2011)). Therefore, we investigated Cdk5’s ability to prime PI4KIIα for subsequent phosphorylation by GSK3. PI4KIIα was immunoprecipitated from HEK293 cells via its C-terminal FLAG-tag and was incubated with Cdk5 and/or GSK3β (30°C, 30 min). Western blot analysis using phosphospecific antibodies (pSer5, pSer47 for PI4KIIα) revealed phosphorylation of both Ser5 and Ser47 was highest when incubated with Cdk5 plus GSK3β, compared to either kinase alone (Fig.3.5). Cdk5 alone did not increase phosphorylation of Ser5 or Ser47 (lane 3), while GSK3 alone causes only a small increase in phosphorylation (lane 2). This indicates that Cdk5 can phosphorylate PI4KIIα and increase the efficiency of subsequent phosphorylation by GSK3 in vitro, although this remains to be demonstrated in vivo and therefore is only a preliminary observation.

Figure 3.5. Cdk5 phosphorylates PI4KIIα, priming for subsequent GSK3 phosphorylation. PI4KIIα, immunoprecipitated via C-terminal FLAG-tag, was incubated alone (-) or with Cdk5, GSK3β, or Cdk5 plus GSK3β in the presence of unlabeled ATP (30°C, 30 min). Phosphorylation was determined by Western blot analysis using the pSer5, pSer47 and FLAG antibodies.

77 |Page

Sec. 3.2. Results

In summary, GSK3 phosphorylates two sites in the N-terminal region of PI4KIIα (Ser5 and Ser47), which is dependent upon prior priming phosphorylation at Ser9 and Ser51, respectively. Cdk5 can phosphorylate these sites and prime for subsequent phosphorylation in vitro although this remains to be shown in vivo. Also, the Ser47 site is relatively more resistant to dephosphorylation by phosphatases than Ser5, suggesting that the Ser5/9 site might be more regulatable/dynamic than the Ser47/51 site in cells.

3.2.2 Phosphorylation of PI4KIIα by GSK3 regulates vesicular trafficking

3.2.2.1 Phosphorylation of PI4KIIα slows transferrin recycling in HeLa cells

Transport of fluorescently-labelled Tfn through HeLa cells is a well-established model of vesicular trafficking in cells. Tfn is an iron-binding protein that facilitates iron uptake by cells. It binds to its receptor (TfnR) at the surface of cells, becomes internalised via clathrin-mediated endocytosis and traffics to early endosomes, where it is separated from its receptor and iron cargo for recycling back to the cell surface. To determine whether PI4KIIα phosphorylation by GSK3 regulates vesicular trafficking, HeLa cells were transfected with wild type or phospho-mutant PI4KIIα, then incubated with fluorescently-labelled Tfn for 45 min at 4°C. Unbound Tfn was removed by washing, followed by incubation at 37°C for up to 30 min. Cells were fixed at different time points and the localization of Tfn was determined using fluorescence microscopy (Fig.3.6A). The rate of Tfn internalisation was similar in cells transfected with wild type, de-phospho (S9/51A) or phospho-mimetic (S5/9/47/51D; Quad D) forms of PI4KIIα (Fig.3.6B). However, recycling back to the cell surface was significantly increased by the presence of non-phosphorylatable-PI4KIIα (S9/51A) and decreased by phospho-mimetic PI4KIIα (Quad-D) compared to wild type (Fig.3.6C). This demonstrates that phosphorylation of PI4KIIα by GSK3 slows recycling of Tfn back to the surface of HeLa cells.

3.2.2.2 Phosphorylation of PI4KIIα promotes trafficking of AMPA receptors away from the cell surface in neurons

To determine if PI4KIIα phosphorylation also regulates trafficking in neurons, we measured cell surface expression levels of the AMPA receptor subunit GluA1 in transfected primary mouse hippocampal neurons. Endogenous PI4KIIα was knocked down using an shRNA construct and the surface expression of GluA1 was determined by immunofluorescence microscopy using an antibody recognising the extracellular region of GluA1 under non- permeabilising conditions (Sytnyk et al., 2002). PI4KIIα

78 |Page

Sec. 3.2. Results

knockdown significantly increased GluA1 at the surface of hippocampal neurons compared to controls (Fig.3.7A, B). This was rescued back to control levels by co- expression of an shRNA-resistant form of wild type PI4KIIα, but not the PI4KIIα-S9/51A mutant. Validation of the knockdown and shRNA-resistant constructs is shown in Fig.3.7C. This data indicates that GluA1 is a cargo protein of PI4KIIα that is normally directed away from the cell surface by PI4KIIα. Phosphorylation by GSK3 is required for this activity, since the non-phosphorylated form (S9/51A) was unable to restore normal surface expression of GluA1. These observations are consistent with the Tfn trafficking assays in HeLa cells (Fig.3.6A-C), in that phosphorylation by GSK3 promotes trafficking of PI4KIIα and its cargo proteins away from the cell surface.

3.2.2.3 Phosphorylation of PI4KIIα has no effect on neuronal morphology

We have previously shown GSK3-mediated phosphorylation of substrates affects neurite outgrowth in cultured neurons (CRMP2 and CRMP4 (Cole et al., 2006; Cole et al., 2004b); β-Adducin (Farghaian et al., 2011)). To determine if PI4KIIα phosphorylation affects neuron morphology, primary cortical neurons isolated from embryonic (E17) Sprague-Dawley rats were transfected with wild type or phospho-mutants forms of PI4KIIα. Images were acquired for transfected neurons (GFP, green) using fluorescence microscopy and neuronal morphology analysed using Image J software. Transfected neurons were analysed for axon length, axon branch number, axon branch length, dendrite number, dendrite length and growth cone number. Overall, expression of wild type or non-phosphorylatable (S9/51A) forms of PI4KIIα had no effect on the morphology of cultured cortical neurons (Fig.3.8).

79 |Page

Sec. 3.2. Results

*

*

Figure 3.6. Phosphorylation of PI4KIIα by GSK3 regulates Tfn trafficking in HeLa cells. (A) HeLa cells transfected with PI4KIIα wild type, S9/51A or Quad-D mutants were pulse-labelled with fluorescent Tfn, then incubated in Tfn-free medium for 0, 6, 15 or 30 min. Cells were analysed using fluorescence microscopy for Tfn (green), transfected cells (mCherry, red) and cell nuclei (blue). Images are representative of the 15 min time point. (B) Relative rates of Tfn internalisation were scored and results are presented as a graph (blinded, average of 3 independent experiments, n=150 cells). (C) Relative rates of Tfn recycling were scored and results are presented as a graph (*p<0.05, Two-way ANOVA).

80 |Page

Sec. 3.2. Results

Figure 3.7. Phosphorylation of PI4KIIα by GSK3 promotes GluA1 trafficking away from the cell surface. (A) Phosphorylation of PI4KIIα by GSK3 regulates GluA1 trafficking in neurons. Primary mouse hippocampal neurons (12 DIV) co-transfected with scrambled or PI4KIIα shRNA constructs, as well as shRNA-resistant wild type and mutant PI4KIIα as labeled, were fixed and incubated with antibodies to GluA1 and synaptophysin under nonpermeabilising conditions to label surface-exposed proteins. (B) Quantitation of GluA1 surface expression from (A) is presented as a graph (***p<0.005, Student’s T-test). (C) Confirmation of effective knockdown by shRNA. Wild type and S9/51A forms of PI4KIIα, with or without synonymous mutations rendering them resistant to shRNA-mediated knockdown, were co-transfected into HEK293 cells with scrambled or PI4KIIα shRNA. Lysates were subjected to Western blotting for FLAG and actin.

81 |Page

Sec. 3.2. Results

Figure 3.8. PI4KIIα does not affect neurite outgrowth in cortical neurons. (A) Cortical neurons were co-transfected with GFP and either empty vector, PI4KIIα wild type or PI4KIIα-S9/51A. Neurons were fixed and visualized using fluorescence microscopy. The morphology of transfected neurons was quantitated using Image J software and is presented as graphs: (B) average axon length. (C) number of axon branches. (D) axon branch length. (E) dendrite number. (F) dendrite length. (G) growth cone number. (GFP control n=161, PI4KIIα wild type n=141, PI4KIIα-S9/51A n=111; average ± SEM; n.s., not significant; Student’s t-test).

82 |Page

Sec. 3.2. Results

3.2.3 Phosphorylation of PI4KIIα by GSK3 promotes trafficking to lysosomes by the AP-3 complex

3.2.3.1 GSK3 phosphorylation of PI4KIIα targets it for degradation by the lysosome

We next investigated the molecular mechanisms by which GSK3 regulates PI4KIIα function. Differences in abundance levels were consistently observed between wild type and phospho-mutant forms of PI4KIIα, suggesting that phosphorylation by GSK3 might regulate the stability of PI4KIIα. Indeed, treatment with CT99021 or mutation of the GSK3 phosphosites (Ser5/47) or priming sites (Ser9/51) to alanine increased the abundance of PI4KIIα (Fig.3.9A, B). Treatment with the translation inhibitors anisomycin or cycloheximide demonstrated that wild type PI4KIIα is more rapidly degraded than PI4KIIα-S9/51A (Fig.3.9C, D). Degradation is mediated by the lysosome and not the proteasome, since pharmacological inhibition of the lysosome using Bafilomycin A increased PI4KIIα levels, while treatment with the proteasome inhibitor MG132 had no significant effect (Fig.3.9.E, F). Consistent with this, wild type and PI4KIIα-S9/51A clustered with Lamp1 staining of lysosomes in HeLa cells (Fig.3.10A). Interestingly, wild type PI4KIIα and Lamp1 were more tightly clustered in the peri- nuclear region compared to PI4KIIα-S9/51A, which was more diffusely spread throughout the cell (Fig.3.10B). Together, these observations show that phosphorylation by GSK3 promotes degradation of PI4KIIα by lysosomes.

3.2.3.2 Phosphorylation of PI4KIIα by GSK3 promotes binding to adaptin 

We next investigated binding partners of PI4KIIα that could mediate its trafficking to lysosomes. Wild type and S9/51A forms of PI4KIIα were expressed in HEK293 cells, immunoprecipitated and subjected to Western blotting using antibodies to various trafficking proteins. PI4KIIα bound to adaptin δ of the AP-3 complex, but not adaptins α, β (AP-2 complex), γ (AP-1 complex), clathrin or EEA1 (Fig.3.11A). Binding of adaptin δ was decreased for PI4KIIα-S9/51A compared to wild type and by treatments with GSK3 inhibitors (Fig.3.11B, C). Mutation of each site individually also reduced the interaction with adaptin δ (i.e. S9A and S51A; Fig.3.11D, E). PI4KIIα co-localised with endogenous adaptin δ in HeLa cells (Fig.3.11F), with the wild type form more tightly clustered with adaptin δ in the peri-nuclear region than PI4KIIα-S9/51A, similar to

83 |Page

Sec. 3.2. Results

Lamp1 (Fig.3.10A). Together, these observations indicate the phosphorylation of PI4KIIα by GSK3 promotes binding to the AP-3 complex.

84 |Page

Sec. 3.2. Results

) E

= S.E.M.,* ± (Mean

transfectedHEK293 cells ) C

PI4KIIα, phosphomutant or type wild with transfected cells HEK293 from Lysates ) A (

( graph. a as presented is treatment anisomycin of h 2 after actin to FLAG of ratio The ) D

graph. a as presented is A Bafilomycin and MG132 with treatment after actin to FLAG of ratio The ) F

Phosphorylation regulates the abundance of PI4KIIα. of abundance the regulates Phosphorylation , not significant; T-test).

) Relative abundance of PI4KIIα was quantitated as the ratio between FLAG and actin and is presented as a graph. ( graph. a as presented is actinand FLAGand between ratio the quantitatedas was PI4KIIα abundanceofRelative ) B n.s.

Figure 3.9. Figure loading a as actin and tag C-terminalFLAG the forblotting Western to weresubjected h), 16 µM, CT99021(2 GSK3inhibitor the withtreated or untreated ( control. with PI4KIIα wild type or S9/51A were treated with 10 µg/ml anisomycin or cycloheximide for the times indicated. Lysates were subjected to Western ( control. blotting loading a as actin and tag FLAG C-terminal the for Western to subjected were Lysates h. 2 for A Bafilomycin nM 50 or MG132 µM 10 with treated were PI4KIIα type wild with transfected cells HEK293 ( actin. and FLAG for blotting p<0.05;

85 |Page

Sec. 3.2. Results

) Localization) of Lamp1 a in focused peri- B

) HeLa cells werecellstransfectedS9/51AGFP with HeLatypeor wild) (control),PI4KIIα (FLAG-tagged, green).

A (

PI4KIIαwith lysosomes.colocalises

Figure 3.10. Figure Lysosomes were visualized using an antibody to Lamp1 (red), while nuclei were visualized using DAPI (blue). ( position,nuclei polarised to of side one the cell was scored and results presented are graph as (n=150a cells, S.E.M., Mean ± *p<0.05, T-Test).

86 |Page

Sec. 3.2. Results

Figure 3.11. Phosphorylation by GSK3 regulates PI4KIIα binding to AP-3. (A) HEK293 cells transfected with PI4KIIα wild type or S9/51A were immunoprecipitated via their C-terminal FLAG-tag and subjected to Western blotting for various trafficking proteins, as well as FLAG as a loading control. (B) Western blots of immunoprecipitates of PI4KIIα wild type, phosphomutants and GSK3 inhibitor-treated wild type PI4KIIα for adaptin  and FLAG. (C) Relative amounts of adaptin  binding to PI4KIIα is shown as a graph (Mean ± S.E.M.,* = p<0.05; (T-test)). (D) Western blots of immunoprecipitates of PI4KIIα wild type and phosphomutants for adaptin  and FLAG. (E) Relative amounts of adaptin  binding to PI4KIIα is shown as a graph (Mean ± S.E.M.,* = p<0.05; T-test). (F) HeLa cells were transfected with wild type or S9/51A PI4KIIα (FLAG-tagged, green) and stained for adaptin  (red) and nuclei using DAPI (blue).

87 |Page

Sec. 3.2. Results

3.2.3.3 Phosphorylation of PI4KIIα promotes binding to AP-3 for trafficking to the lysosome to be degraded

We speculated that GSK3 may decrease the stability of PI4KIIα (Fig.3.9) by promoting binding to the AP-3 complex (Fig.3.11) for trafficking to the lysosome to be degraded (Fig.3.9E, F). To test this, mutants of PI4KIIα were generated that could potentially block its interaction with adaptin δ (L60/61A (Craige et al., 2008) and P16/Y18A (Mössinger et al., 2012)). The L60/61A mutant, but not the P16/Y18A mutant, successfully blocked binding to adaptin δ in co-immunoprecipitation assays (Fig.3.12A). This was accompanied by a dramatic increase in PI4KIIα abundance levels (Fig.3.12B). Inhibition of lysosomes using Bafilomycin A increased the abundance of wild type PI4KIIα, but not the L60/61A or S9/51A mutants (Fig.3.12C, D). Meanwhile, reducing phosphorylation of PI4KIIα using the GSK3 inhibitor CT99021 increased the abundance of wild type PI4KIIα, but not L60/61A or S9/51A mutants (Fig.3.12E, F). Together, these observations support the hypothesis that phosphorylation by GSK3 promotes PI4KIIα binding to AP-3 for trafficking to the lysosome to be degraded. However, the influence of other unknown binding partners contributing to this process cannot be ruled out.

3.2.3.4 The N-terminal region of PI4KIIα regulates binding of the dileucine motif to the AP-3 complex

The GSK3 phosphosites are located adjacent to the L60/61 dileucine motif on the flexible N-terminal domain of PI4KIIα (Fig.3.13A). We postulated that the N-terminal region might be able to fold back on itself to restrict access of the AP-3 complex to the dileucine motif. To test this, PI4KIIα truncation mutants were generated lacking the N- terminal 24 or (ΔG24) or 52 (ΔG52) amino acids and access to the dileucine motif was measured by relative binding to adaptin δ in co-immunoprecipitation experiments. Fig.3.13B and C shows that binding of the truncation mutants to adaptin δ is dramatically increased compared to full-length PI4KIIα. This is accompanied by decreased abundance levels of PI4KIIα. This supports the idea that the N-terminal region blocks the interaction of the dileucine motif with the AP-3 complex for trafficking to the lysosome.

88 |Page

Sec. 3.2. Results

Figure 3.12. Phosphorylation promotes PI4KIIα binding to AP-3 and trafficking to the lysosome for degradation. (A) HEK293 cells transfected with GFP control, PI4KIIα wild type, L60/61A or P16/Y18A mutants were immunoprecipitated via their C-terminal FLAG-tags and subjected to Western blotting for adaptin δ and FLAG. (B) HEK293 cell lysates transfected with PI4KIIα wild type, L60/61A or P16/Y18A were subjected to Western blotting for FLAG and actin. (C) HEK293 cell lysates transfected with PI4KIIα wild type, S9/51A or L60/61A and treated without/with Bafilomycin A were subjected to Western blotting for FLAG. (D) The ratio of FLAG to actin after Bafilomycin A treatment in (C) is presented as a graph. (E) HEK293 cells transfected with PI4KIIα wild type, L60/61A or S9/51A and treated without/with CT99021 were subjected to Western blotting for FLAG and actin. (F) The ratio of FLAG to actin after CT99021 treatment in (E) is presented as a graph. (Mean ± S.E.M.,* = p<0.05; n.s., not significant; T-test).

89 |Page

Sec. 3.2. Results

) HEK293 transfectedcells ) B

), except the four residues mutated to D ) Schematic representation of the domain the of representation Schematic )

A (

) Same as ( E

) HEK293) cells transfected with GFP control, PI4KIIα wild ortype

F

) Relative amounts of adaptin δ binding to PI4KIIα full length and truncation mutants is shown as a graph (Mean ± (Mean graph a as shown is mutants truncation and length full PI4KIIα to binding δ adaptin of amounts Relative )

C ) Amino acid sequences surrounding the dileucine motif and Ser5-9 phosphosites (numbered) at the N-terminus of PI4KIIα. of N-terminus the at (numbered) phosphosites Ser5-9 and motif dileucine the surrounding sequences acid Amino )

D

The N-terminal region of PI4KIIα restricts access of the AP-3 complex to the dileucine motif. dileucine the to complex AP-3 the of access restricts PI4KIIα of region N-terminal The

Figure 3.13. Figure PI4KIIα,ofpositionof showingphosphorylationstructure the the N-terminus.andsites, dileucine at truncationsites motif the ( with GFP control, PI4KIIα full-length, ΔG24 or ΔG52 truncation mutants were immunoprecipitated via their C-terminal FLAG-tags and subjected to Western ( FLAG. and δ adaptin for blotting ( T-test). p<0.05; = S.E.M.,* Charged (underlined) and hydrophobic (italics) residues predicted to form an interaction site are shown. ( alanine PI4KIIα-Quadin disrupt A to the potential interaction site shown are bold.in (

90 |Page

Sec. 3.2. Results

) Relative amounts )

G

Quad A mutant were immunoprecipitated via their C-terminal FLAG-tags and subjected to Western blotting for adaptin δ and FLAG.( and adaptinδ blotting Western for to FLAG-tags andC-terminalsubjected wereimmunoprecipitated their via mutant A Quad adaptinof δ binding to PI4KIIα wild type and QuadA mutant shownis graph as a (Mean S.E.M.,* p<0.05;± = T-test).

91 |Page

Sec. 3.2. Results

Analysis of the amino acid sequences surrounding the dileucine motif and Ser5-9 phosphosites revealed a series of complementary charged and hydrophobic residues that could constitute an interaction site between these regions (Fig.3.13D, underlined and italics). To test this, four residues were mutated to alanine to reduce this interaction, exposing the dileucine motif for binding to the AP-3 complex (Quad A; Fig.3.13E). Co- immunoprecipitation experiments revealed increased binding of adaptin δ to PI4KIIα- Quad A compared to wild type, together with decreased abundance levels of the mutant form (Fig.3.13F, G). These observations identify a novel binding site between the N- terminus and dileucine regions that restricts access of the AP-3 complex to the dileucine motif. Altogether, we propose a model whereby the N-terminal region restricts access to the dileucine motif (DLL), but phosphorylation by GSK3 exposes this site for binding AP-3 and subsequent trafficking to the lysosome to be degraded (Fig.3.14).

Figure 3.14. Model of GSK3-mediated regulation of PI4KIIα binding to AP-3 and trafficking to the lysosome. Phosphorylation of PI4KIIα by GSK3 induces structural rearrangement of its N- terminal region, exposing the DLL motif for binding to the AP-3 complex and subsequent trafficking to the lysosome to be degraded. In its non-phosphorylated form, the N-terminal region blocks access to the DLL motif, reducing binding of PI4KIIα to the AP-3 complex and trafficking to the lysosome. Instead, PI4KIIα abundance increases and is trafficked to recycling and secretory endosomes.

92 |Page

Sec. 3.2. Results

3.2.4 Investigating the physiological function of PI4KIIα in Drosophila melanogaster

3.2.4.1 Depletion of PI4KIIα produces hyperactivity in Drosophila

Next, we investigated the physiological function of PI4KIIα and potential links with symptoms of BD using the fruit fly Drosophila melanogaster. Previous work in flies (and mice) showed that locomotor activity (especially hyperactivity) and sleep are regulated by GSK3 and are lithium sensitive (Ackermann et al., 2010; Beaulieu et al., 2004; Chan et al., 2012; Dokucu et al., 2005; Martinek et al., 2001; O'Brien et al., 2004; Polter et al., 2010; Prickaerts et al., 2006; Roybal et al., 2007). These two behaviours are closely associated with symptoms of human BD (Anderson et al., 2012). Therefore, we investigated locomotor activity and sleep patterns in PI4KIIα-depleted flies. Control and PI4KIIα-mutant flies (equivalent to knockouts in mice) (Males, 3-5 d) were subjected to the Drosophila Activity Monitoring System (DAMS) assay to measure overall activity of the flies (12:12 LD, 25°C, 75% humidity). Specically designed monitors measure the locomoter movement of Drosophila, using the interruption of a beam of infrared light to record the locomoter activity of indiviual flies contained inside small vials. Activity is the measurement of the number of beam interruption (counts) in 30 min bins. Individual fly waveforms are calculated from 6 d of analysis and normalized to 24 h; group waveforms were then calculated by averaging those individual waveforms. PI4KIIα homozygous mutant flies exhibit strong motor issues precluding use in this system, however heterozygous flies exhibit normal motor function by negative geotaxis assay (Fig.3.15), therefore are suitable for DAMs analysis. These flies exhibited extensive hyperactivity compared to isogenic controls (Fig.3.16A), and this was particularly pronounced during expected periods of peak arousal around dawn and dusk (ZT 0-2 and ZT 8-12, respectively) (Fig.3.16A, B).

93 |Page

Sec. 3.2. Results

Figure 3.15. PI4KIIα heterozygous mutant exhibit normal motor function by negative geotaxis assay. Negative geotaxis (climbing assay; 5 cm, 10s) was performed on PI4KIIα heterozygous mutant (29058) and W1118 (control) and results are presented as a graph. (Mean ± SEM; n.s., not significant; Student’s t-test).

Figure 3.16. Actogram of PI4KIIα heterozygous mutant and neuron-specific RNAi knockdown flies. (A) Locomotor activity (count/30 min) of PI4KIIα heterozygous mutant (blue line) and W1118 (control, red line) averaged over 6 d. (B) Quantitation of activity (sum of counts) between ZT 8 and 12, equivalent to 4pm-8pm (dotted lines) from (A) is presented as a graph (Mean ± S.E.M.,* = p<0.05; T-test). (C) Locomotor activity (count/30 min) of PI4KIIα-RNAi (#1) knockdown (dcr2;PI4KIIα-RNAi(110687);nSyb-Gal4) (blue line), driver only control (W1118,dcr2;;nSyb-Gal4, red line) and RNAi only control (W1118;PI4KIIα-RNAi(110687), green line) averaged over 6 d. (D) Quantitation of activity (sum of counts) between ZT 8 and 12,

94 |Page

Sec. 3.2. Results

equivalent to 4pm-8pm (dotted lines) from (C) is presented as a graph (Mean ± S.E.M.,* = p<0.05; T-test).

To determine if this phenotype is due to reduced PI4KIIα function in the brain, its expression was specifically knocked down in neurons of flies using the UAS-Gal4 system. The UAS-Gal4 system utilises the yeast transcription activator protein GAL4, and UAS (Upstream Activation Sequence), an enhancer element to which GAL4 specifically binds for transcriptional activation of GAL4-regulated genes (Duffy, 2002). Using this system, transgenic flies harbouring a UAS-PI4KIIα RNAi transgene were mated with flies containing the neural specific nSyb-Gal4 driver (Fig.3.17). Progeny containing both elements of the system are produced, selectively expressing PI4KIIα- RNAi in the nervous system of these flies and knocking down PI4KIIα in the brain.

Figure 3.17. Schematic representation of the UAS-GAL4 system. Male transgenic flies harbouring a UAS-PI4KIIα-RNAi transgene are crossed with virgin female nSyb-Gal4 transgenic flies (pan- neuronal driver line) with UAS-Dicer-2 (dcr2) to enhance RNAi target degradation (dcr2/yhh;;nSyb-Gal4). First filial generation (F1) male flies (PI4KIIα knockdown) are tested for phenotypes associated with BD.

Flies with neural specific knockdown of PI4KIIα exhibit similar hyperactivity (Fig.3.16C) to the PI4KIIα heterozygous mutant line, confirming it is primarily a neurological effect. This increase is most pronounced 4 hours before dusk (8-12 ZT; Fig.3.16C, D). Hyperactivity of PI4KIIα-depleted flies is consistent with hyperactivity of human BD patients during their manic phase, as well as hyperactivity exhibited by

95 |Page

Sec. 3.2. Results

transgenic mice with elevated levels of GSK3 (Ackermann et al., 2010; Prickaerts et al., 2006). Our model is high GSK3 activity in BD promotes degradation of PI4KIIα by the lysosome (similar to PI4KIIα-knockdown), and low levels of PI4KIIα contributes to the hyperactivity phenotype of this disease (Fig.3.18).

Figure. 3.18. Proposed model of the relationship between high GSK3 activity, depletion of PI4KIIα and hyperactivity in BD.

BD patients typically have disruptions in circadian function with irregular sleep timings and reduction in total sleep time (Geller et al., 1998; Harvey et al., 2006). These irregularities can sometimes trigger manic episodes (Boivin, 2000; Wehr, 1992). The DAMS assay can also be used to evaluate fly sleep patterns (Shaw et al., 2000; Ueno et al., 2012). Sleep in flies is defined as any period of uninterrupted behavioural immobility lasting longer than 5 min (Huber et al., 2004). Specifically, we looked for differences in total amount of sleep (Fig.3.19A, B) and the number of speed bouts (sleep episodes; Fig.3.19C, D) in 24 h. Overall, PI4KIIα-knockdown flies exhibit no significant difference in sleep patterns compared with control flies (Fig.3.19).

3.2.4.2 Lithium reduces hyperactivity of PI4KIIα-depleted flies

Lithium is particularly effective at reducing mania in BD patients (Jope, 2003; Zhang et al., 2003). Also, PI4KIIα phosphorylation is reduced by lithium (Fig.3.2B) and its abundance increased (Fig.3.9). This led us to investigate the effect lithium treatment has on the hyperactivity phenotype observed in our PI4KIIα-depleted flies. To test this, PI4KIIα-depleted flies and controls were subjected to the DAMs assay in vials with food containing added lithium (20 mM LiCl) or salt (20 mM NaCl, control) and their overall activity was monitored. In addition to the original PI4KIIα-RNAi knockdown line and heterozygous mutant line, a second PI4KIIα-RNAi hairpin was also tested (PI4KIIα- RNAi #2). Both neuron-specific PI4KIIα knockdown lines displayed a reduction of locomotor hyperactivity when treated with lithium, especially during the peak afternoon arousal period (8-12 ZT) (Fig.3.20A-D). This was statistically significant for the original RNAi line (RNAi #1) (Fig.3.20A, B), while the second line (RNAi #2, Fig.3.20C, D) and

96 |Page

Sec. 3.2. Results

the heterozygous mutant line (Fig.3.20E, F) displayed a strong trend towards significance. Meanwhile, treatment with lithium had no effect on driver only controls (Fig.3.20A-F) and only a small reducing effect on the activity of RNAi only control flies (Fig.3.20B, D). This is likely due to the genetic background of these flies (i.e. these flies are more susceptible to the sedating effects of lithium (Matsagas et al., 2009) at the chosen concentration). Together, this strongly demonstrates that the hyperactivity phenotype observed in PI4KIIα-depleted flies can be reduced by lithium treatment to levels indistinguishable from control flies.

Figure 3.19. Sleep analysis of PI4KIIα-depleted flies. (A) Average percentage of sleep (24 h, 6d) of PI4KIIα-RNAi (#1) knockdown (dcr2;PI4KIIα-RNAi(110687);nSyb-Gal4), driver only control (W1118,dcr2;;nSyb-Gal4) and RNAi only control (W1118;PI4KIIα-RNAi(110687)) is shown as a graph. (B) Average percentage of sleep (24 h, 6d) of PI4KIIα heterozygous mutant (29058) and W1118 (control) is shown as a graph. (C) Average sleep bout number (24 h, 6d) of PI4KIIα-RNAi knockdown (dcr2;PI4KIIα-RNAi(110687);nSyb-Gal4), driver only control (W1118,dcr2;;nSyb-Gal4) and RNAi only control (W1118;PI4KIIα-RNAi(110687)) is shown as a graph. (D) Average sleep bout number (24 h, 6d) of PI4KIIα heterozygous mutant (29058) and W1118 (control) is shown as a graph. (Mean ± SEM; n.s., not significant; Student’s t-test).

97 |Page

Sec. 3.2. Results

) Locomotor) activity (count/30min) of PI4KIIα-RNAi #1 (110687) knockdown A (

Actogram PI4KIIα-depleted of flies treated with lithium.

Figure 3.20. Figure

98 |Page

Sec. 3.2. Results

mM 20 (+ ) dcr2;PI4KIIα-

) 20(+ mM NaCl, green line; + W1118,dcr2;;nSyb-Gal4 mM 20 + orange; NaCl, mM 20 (+ )

) Quantitation of activity (sum of counts) between F

W1118,dcr2;;nSyb-Gal4 ) ) (+ 20 mM NaCl, orange; + 20 mM LiCl, light blue) (Mean and driver only control ( control only driver and

W1118;PI4KIIα-RNAi(110687) and driver only control (

= p=0.095; T-test). # ) Quantitation of activity (sum of counts) between ZT 8 and 12, equivalent to 4pm-8pmto equivalent 12, and 8 ZT between counts) of (sum activity of Quantitation ) B W1118;PI4KIIα-RNAi(25459) Lcmtr ciiy cut3mn o P4IαRA # (55) ncdw ( knockdown (25459) #2 PI4KIIα-RNAi of (count/30min) activity Locomotor ) C ) Quantitation of activity (sum of counts) between ZT 8 and 12, equivalent to 4pm-8pm (dotted lines) from from lines) (dotted 4pm-8pm to equivalent 12, and 8 ZT between counts) of (sum activity of Quantitation ) D ) (+ 20mM NaCl, blue line, + 20mM LiCl, red line) red LiCl, 20mM + line, blue NaCl, 20mM (+ ) ) Locomotor activity (count/30 min) of PI4KIIα heterozygous mutant (29058) (+ 20 mM NaCl, blue line; + 20 mM LiCl, mM 20 + line; blue NaCl, mM 20 (+ (29058) mutant heterozygous PI4KIIα of min) (count/30 activityLocomotor ) E ) is presented ) a graphas (Mean S.E.M.,± E ) ) (+ 20mM NaCl, blue line; + 20mM LiCl, red line) ) is presented as a graph and the addition of RNAi only control ( control only RNAi of addition the and graph a as presented is ) A (control) (+ 20 mM NaCl, green line; + 20 mM LiCl, purple line) averaged over 6 d. ( W1118 = p=0.099; T-test). ( T-test). p=0.099; = # ) ) is presented as a graph and the addition of RNAi only control ( dcr2;PI4KIIα-RNAi(110687);nSyb-Gal4 C ( ( d. 6 over averaged line) purple LiCl, mM 20 + line; green NaCl, ( from lines) (dotted ( T-test). p<0.05; = S.E.M.,* ± (Mean blue) light LiCl, RNAi(25459);nSyb-Gal4 ( d. 6 over averaged line) purple LiCl, mM 20 ( S.E.M., ± red line) and 4pm-8pm (dotted lines) from (

99 |Page

Sec. 3.2. Results

3.2.4.3 Effect of lithium on Drosophila sleep behaviour

Lithium has previously been shown to lengthen the period of circadian rhythm and increase arrhythmicity of fly sleep patterns in Drosophila (Dokucu et al., 2005). We therefore investigated if lithium affects sleep behaviour in PI4KIIα-depleted flies. Treatment with 20 mM LiCl did not significantly affect sleep amount (Fig.3.21A and C) or number of sleep episodes (bout number; Fig.3.21B and D) in either PI4KIIα-RNAi lines. However, the PI4KIIα heterozygous mutant line was an exception. Fig.3.21E shows there was a small but significant decrease in sleep amount of the PI4KIIα mutant when exposed to lithium. This also occurred in the control flies, suggesting the effect is due to the genetic background of these flies, rather than depletion of PI4KIIα. There was also a small increase in the number of sleep episodes of the PI4KIIα heterozygous mutant and control lines (sleep bout number; Fig.3.21F), although these did not reach significance. Together, these results indicate that lithium has a strong effect on reducing locomotor hyperactivity in PI4KIIα-depleted flies while having minimal effect on sleep arrhythmicity in Drosophila. This further supports the idea that PI4KIIα may be a therapeutic target of lithium in BD.

3.2.4.4 PI4KIIα depletion has no effect on feeding behaviour or locomotor function

We next wanted to determine if PI4KIIα depletion had any further effects on fly homeostasis control. In particular, fly feeding behaviour, body weight and fat (TAG) levels. PI4KIIα-knockdown flies exhibit no significant differences in food intake (Fig.3.22A), average body weight (Fig.3.22B) or average TAG levels (Fig.3.22C) compared to control flies. Finally, to ensure that PI4KIIα-depleted flies exhibit normal motor function, we tested their climbing ability in the negative geotaxis assay. Briefly, male flies 3-5 d old were placed in clear vials. Flies were gently tapped to the bottom of the vial and the number of flies that climbed above 5 cm was counted after 10 s of climbing. Flies have a natural tendency to climb to the top of the vial and an inability to do this within 10 s is a measure of locomotor dysfunction. Overall, PI4KIIα-depletion had no effect on locomotor function or climbing ability compared to controls (Fig.3.22D, E). Together, this data indicates that locomotor hyperactivity induced by PI4KIIα depletion occurs independent of changes in feeding behaviour and locomotor function.

100 |Page

Sec. 3.2. Results

Figure 3.21. Sleep analysis following lithium treatment. Sleep analysis of PI4KIIα-depleted flies (Males, 3-5 d) exposed to food with added LiCl (red) or NaCl (control, blue) is presented as graphs: (A) Average percentage of sleep (24 h, 6 d) of PI4KIIα-RNAi #1 (110687) knockdown (dcr2;PI4KIIα-RNAi(110687);nSyb-Gal4), driver only control (W1118,dcr2;;nSyb-Gal4) and RNAi only control (W1118;PI4KIIα-RNAi(110687)). (B) Average bout number (24 h, 6 d) of PI4KIIα-RNAi #1 (110687) knockdown (dcr2;PI4KIIα-RNAi(110687);nSyb-Gal4), driver only control (W1118,dcr2;;nSyb-Gal4) and RNAi only control (W1118xPI4KIIα-RNAi(110687)). (C) Average percentage of sleep (24 h, 6 d) of PI4KIIα-RNAi #2 (25459) knockdown (dcr2;PI4KIIα- RNAi(25459);nSyb-Gal4), driver only control (W1118,dcr2;;nSyb-Gal4) and RNAi only control (W1118;PI4KIIα-RNAi(25459)). (D) Average bout number (24 h, 6 d) of PI4KIIα-RNAi #2

101 |Page

Sec. 3.2. Results

(25459) knockdown (dcr2;PI4KIIα-RNAi(25459);nSyb-Gal4), driver only control (W1118,dcr2;;nSyb-Gal4) and RNAi only control (W1118;PI4KIIα-RNAi(25459)). (E) Average percentage of sleep (24 h, 6 d) of PI4KIIα heterozygous mutant (29058) and W1118 (control). (F) Average bout number (24 h, 6d) of PI4KIIα heterozygous mutant (29058) and W1118 (control). (Mean ± SEM; n.s., not significant; * = p<0.05; Student’s t-test).

102 |Page

Sec. 3.2. Results

Figure 3.22. PI4KIIα depletion has no effect on feeding behaviour or locomotor function. (A) Neural-specific knockdown of PI4KIIα (dcr2;PI4KIIα-RNAi(110687);nSyb-Gal4) and driver only control (W1118,dcr2;;nSyb-Gal4) were placed in vials and subjected to the café assay for 48 h. Average cumulative food consumption was recorded at 16, 24, 32 and 48 h and presented as a graph. (B) Average body weight (per fly, mg) of PI4KIIα-RNAi knockdown (dcr2;PI4KIIα- RNAi(110687);nSyb-Gal4) and driver only control (W1118,dcr2;;nSyb-Gal4) is presented as a graph. (C) TAG colourimetric assays were performed on PI4KIIα-RNAi knockdown (dcr2;PI4KIIα-RNAi(110687);nSyb-Gal4) and driver only control (W1118,dcr2;;nSyb-Gal4) and results are presented as a graph. (D) Negative geotaxis (climbing assay; 5 cm, 10s) was performed on PI4KIIα-RNAi knockdown (dcr2;PI4KIIα-RNAi(110687);nSyb-Gal4) and driver only control (W1118,dcr2;;nSyb-Gal4) and results are presented as a graph. (Mean ± SEM; n.s., not significant; Student’s t-test).

103 |Page

Sec. 3.2. Results

3.2.5 qPCR validation of PI4KIIα depletion in Drosophila

Next we performed qPCR analysis to validate PI4KIIα-depletion, in both the neural-specific PI4KIIα-RNAi knockdown and heterozygous mutant flies. To test this, we first isolated RNA from 3-5 d old males, either from whole flies (for heterozygous mutant flies) or from dissected fly heads (for flies with neural-specific PI4KIIα- knockdown). RNA samples were then reverse transcribed using an Oligo(dT)15 primer to synthesise cDNA from mRNA template. Primers were designed to bind the N-terminus of the Drosophila PI4KIIα gene (gene id, CG2929) spanning the junction between exon 2 and exon 3 of the gene to avoid amplification of genomic DNA. Relative quantification of PI4KIIα expression (target gene) was normalised to the expression of Drosophila Actin42A gene (gene id, CG12051) as a reference control.

3.2.5.1 Determining gene amplification efficiency

We first determined amplification efficiency for each gene. Ideally, the amount of PCR product will perfectly double during each cycle of exponential amplification i.e. a 2-fold increase in the number of copies with each cycle. Therefore, qPCR efficiency was determined according to the equation: E=(10[-1/slope]-1) x 100 (Radonić et al., 2004). Fig.3.23 shows qPCR efficiency is above 89% when tested using cDNA from control flies (W1118) (Fig.3.23A) and above 83% when tested using cDNA from PI4KIIα heterozygous mutants (Fig.3.23B).

3.2.5.2 Calculating relative gene expression using the 2-CT (Livak) method

The 2-CT method (Livak & Schmittgen, 2001) was used to calculate the relative gene expression of PI4KIIα in the neural-specific PI4KIIα-RNAi knockdown and heterozygous mutant flies. Normalised expression ratio (2-CT value) was calculated for each qPCR run and averaged to determine relative abundance change / PI4KIIα depletion. The heterozygous mutant was only found to have a 15.3% reduction in PI4KIIα expression levels (Fig.3.23A), however, the neural-specific PI4KIIα-RNAi knockdown flies (lines #1 and #2) had reduction of 51.7% (statistical significant) and 54.2% (trend towards significance), respectively (Fig.3.23B). A possible explanation for the relatively low reduction of PI4KIIα levels in the heterozygous mutant is that the remaining wild type allele is sufficient to generate the majority of PI4KIIα transcripts. However, it also suggests that even a small reduction in PI4KIIα expression is sufficient to produce

104 |Page

Sec. 3.2. Results

significant hyperactivity phenotype. This is consistent with growing evidence from the field of developmental genetics supporting the hypothesis that small differences in the time of gene activation or in the level of activity of even a single gene, can result in changes in regulatory interactions and the developmental processes and, thus, organismal form and function (Garfield & Wray, 2010; Harrison et al., 2012).

Figure 3.23. qPCR validation of reduced PI4KIIα gene expression in Drosophila. (A) Amplification efficiency of PI4KIIα and Actin42A genes amplified from control line (W1118) is presented as a graph. (B) Amplification efficiency of PI4KIIα and Actin42A genes amplified from PI4KIIα heterozygous mutant (25059) is presented as a graph. (C) Averaged normalised gene expression ratio (2-CT) of PI4KIIα to Actin42A was calculated for the PI4KIIα heterozygous mutant (25059) and presented as a graph. (D) Averaged normalised gene expression ratio (2-CT) of neuronal specific PI4KIIα to Actin42A was calculated for PI4KIIα-RNAi #1 (110687) (dcr2;PI4KIIα RNAi(110687);nSyb-Gal4), PI4KIIα-RNAi #2 (25459) (dcr2;PI4KIIα- RNAi(25459);nSyb-Gal4) knockdown lines and driver only control (W1118,dcr2;;nSyb-Gal4) and presented as a graph. (Mean ± S.E.M., ** = p<0.006; * = p<0.05; # = p=0.10; Student’s t- test).

105 |Page

Sec. 3.3. Discussion

3.3 Discussion

In this study we identified the trafficking protein PI4KIIα as a novel substrate of GSK3, strengthening the case that GSK3 may be an important regulator of vesicular trafficking in cells. The PI lipid products of PI4KIIα are known to be important for endocytosis and trafficking. For example, PI4P is an important precursors for the synthesis of PI(4,5)P2 and PI(4,5)P2, concentrated at the plasma / synaptic membrane, is important for synaptic vesicle endocytosis and recycling (Dickson et al., 2014; Guo et al., 2003). The lipid phosphatase PTEN has previously been identified as a GSK3 substrate (Maccario et al., 2007), while our group used in silico analyses to identify two other lipid phosphatases (Synaptojanin and PIPP (proline-rich inositol polyphosphate 5- phosphatase)) and a lipid kinase (PIP5K) as likely GSK3 substrates. PIPP was confirmed as a novel substrate using the biochemical validation assay described here for PI4KIIα (Farghaian et al., 2011). Therefore, it will be interesting to determine if GSK3 directly regulates PI4P, PI(4,5)P2 and other PI lipid levels at the plasma and intracellular membranes.

PI4KIIα is highly expressed in the brain and synapses, where it accounts for the majority of 4’ PI phosphorylation activity (Guo et al., 2003). It also localises to the TGN/endosomal membranes (Balla et al., 2002; Wang et al., 2003) via palmitoylation (Barylko et al., 2009; Lu et al., 2012) and has been shown to bind to the AP-3 complex (Salazar et al., 2005). GSK3 phosphorylates two sites in the N-terminal region of PI4KIIα (Ser5 and Ser47), following obligatory priming phosphorylation at Ser9 and Ser51. The identity of the priming kinase(s) is not yet known, although both sites fit the consensus sequences of Cdk5 and DYRK (i.e. proline-directed sites with a basic residue at P+3). Both of these kinases are known to prime other GSK3 substrates (e.g. CRMPs (Cole et al., 2006)) and Cdk5 primed for GSK3 phosphorylation at Ser5 and Ser47 in vitro (Fig.3.5), although this remains to be proven in vivo. The Ser47/Ser51 phosphosites appeared later in evolution than the Ser5/Ser9 sites (Fig.3.1A). These sites are relatively resistant to phosphatases (Fig.3.2A–C, Fig.3.3 and Fig.3.4A, B) and are highly phosphorylated in cells (Fig.3.4C), suggesting they may not be regulatable. Therefore, it is likely that phosphorylation of Ser5/9 is the primary mechanism for dynamically regulating PI4KIIα function. It is possible that constitutive phosphorylation of Ser47/Ser51 evolved later to enhance this, although its precise function is not yet clear.

106 |Page

Sec. 3.3. Discussion

Inhibition of phosphorylation increases the abundance of PI4KIIα by reducing its degradation by the lysosome (Fig.3.9). It also reduces binding to adaptin δ of the AP-3 complex (Fig.3.11). These functions are linked, whereby a reduction in phosphorylation decreases PI4KIIα binding to the AP-3 complex, reducing its trafficking to the lysosome for degradation and resulting in increased PI4KIIα abundance (Fig.3.12). In addition, disruption of the PI4KIIα-AP-3 interaction by expression of S9/51A mutant caused redistribution of PI4KIIα, adaptin δ and lysosomes (Lamp1) from a polarised peri-nuclear position to a more scattered pattern throughout the cytoplasm and nuclear circumference (Fig.3.10). Similar results for PI4KIIα knockdown and L60/61A mutant were previously reported (Daboussi et al., 2012). Together, these observations demonstrate that PI4KIIα and its phosphorylation by GSK3 are required for effective organisation of the trafficking pathway from endosomes to lysosomes in an AP-3-dependent manner.

Overall, phosphorylation of PI4KIIα promotes binding to the AP-3 complex for trafficking to the lysosome to be degraded. We identified a novel binding site between the N-terminus and dileucine regions that restricts access of the AP-3 complex to the dileucine motif (Fig.3.13D). Since Leu60-Leu61 in the primary sequence of PI4KIIα is permanent/non-modifiable, it is likely that the nearby phosphosites evolved to dynamically regulate the interaction of PI4KIIα with the AP-3 complex. Therefore, we propose a model whereby the N-terminal region restricts access to the dileucine motif, but phosphorylation by GSK3 exposes this site for binding to AP-3 and subsequent trafficking to the lysosome to be degraded (Fig.3.14). Interestingly, conserved adaptin- binding sequences are located adjacent to predicted GSK3 phosphosites in other GSK3 substrates such as CRMP2, AAK1 and AP180 suggesting that phosphorylation by GSK3 could be a common mechanism for regulating the function of several trafficking proteins. Elsewhere, it has been suggested that ubiquitination of PI4KIIα promoted by a PPxY motif in its N-terminal region directs it towards the degradative endosomal pathway (Mössinger et al., 2012). However, when we mutated this site, we did not observe any effect on its binding to the AP-3 complex or any change in its abundance (Fig.3.12B), implying that the dileucine motif at Leu60-Leu61 is the predominant motif on PI4KIIα for regulating its abundance.

In addition to regulating itself, PI4KIIα regulates trafficking of several cargo- proteins, including Tfn (Fig.3.6), GluA1 (Fig.3.7), epidermal growth factor receptor (EGFR) (Minogue et al., 2006), lysosome membrane protein 2 (LIMP2), β-

107 |Page

Sec. 3.3. Discussion

glucocerebrosidase (GBA) and VAMP3 (Jović et al., 2014; Jović et al., 2012). Since PI4KIIα acts as an integral membrane protein due to its palmitoylation (Barylko et al., 2009; Lu et al., 2012), it probably links the AP-3 complex to lipid bilayer vesicles containing these and other cargo proteins. All together this data fits with a model (Jović et al., 2014; Jović et al., 2012) whereby PI4KIIα and AP-3 mediate transport of cargo proteins from early/sorting endosomes to late endosomes/lysosomes. Disruption of the interaction between PI4KIIα and AP-3 reduces transport of cargo proteins to late endosomes/lysosomes for degradation. Instead, they may accumulate and are subsequently trafficked via a default pathway to recycling and secretory endosomes. This model is supported by accelerated recycling of internalized Tfn back to the cell surface (Fig.3.6) and a previous report showing reduced degradation of endocytosed EGFR in cells depleted of PI4KIIα (Minogue et al., 2006). It is possible that its lipid kinase activity is important for trafficking, since PI4P production has been implicated in recruitment of TGN trafficking proteins GGAs, VAMP3 and the AP-1 complex to the Golgi, promoting Golgi-to-endosomal transport (Daboussi et al., 2012; Jović et al., 2014; Jović et al., 2012). GSK3-mediated phosphorylation may indirectly regulate PI4P production and trafficking by controlling the abundance of PI4KIIα. Interestingly, a population of GSK3 located at the Golgi was previously shown to promote cargo transport from the Golgi to pre- lysosomal compartments (Adachi et al., 2010). Therefore, PI4KIIα may be a major target of GSK3 for maintaining a transport route from Golgi/endosomes to lysosomes for degradation of cargo proteins.

Given the importance of trafficking in synaptic transmission and also the role of GSK3 in regulating this process, we hypothesised that PI4KIIα could be an important target of GSK3 that is dysregulated in mood disorders. Indeed, depletion of PI4KIIα (Fig.3.23C, D) induces hyperactivity in flies (Fig.3.16), independent of changes in feeding behaviour and locomotor function (Fig.3.22). This is consistent with hyperactivity exhibited by human BD patients (during their manic phase) and GSK3- overexpressing mice (Ackermann et al., 2010; Prickaerts et al., 2006). In BD, elevated GSK3 activity is expected to increase phosphorylation of PI4KIIα, leading to increased binding to AP-3 and trafficking to lysosomes for degradation, thus decreased PI4KIIα abundance. Therefore, our PI4KIIα-depleted flies exhibiting hyperactivity is entirely consistent with elevated GSK3 activity and hyperactivity in BD. Lithium treatment is particularly effective at reducing mania in BD patients (Jope, 2003; Zhang et al., 2003).

108 |Page

Sec. 3.3. Discussion

Also, PI4KIIα phosphorylation is reduced by lithium (Fig.3.2B). Investigating the effect of lithium treatment on PI4KIIα-depleted flies, we found that it reduced hyperactivity of these flies to levels indistinguishable from control flies (Fig.3.20) and with only minimal effect on sleep arrhythmicity (Fig.3.21). This was statistically significant for the original RNAi line (RNAi #1) (Fig.3.20A, B), while the second line (RNAi #2, Fig.3.20C, D) and the heterozygous mutant line (Fig.3.20E, F) displayed a strong trend towards significance. This is consistent with our model, whereby lithium would reduce phosphorylation of remaining PI4KIIα in depleted flies, stabilising its protein levels and thus reducing hyperactivity. It is possible that increased expression of AMPA receptors at the cell surface (Fig.3.7) may underlie this hyperactivity, since elevated GluA1 expression has previously been associated with enhanced synaptic activity with wakefulness (versus sleep; (Vyazovskiy et al., 2008)), locomotor hyperactivity (Yamamoto & Zahniser, 2012; Zhang et al., 2008) and emotion-enhanced learning and memory (Hu et al., 2007).

High surface expression of AMPA receptors at synapses promotes LTP, whereas its removal promotes LTD (Hanley, 2010a; Kerchner & Nicoll, 2008; Kessels & Malinow, 2009). Here, PI4KIIα knockdown and the S9/51A mutant form of PI4KIIα increased surface expression of GluA1 in hippocampal neurons, favouring LTP. Therefore, it would be expected that inhibition of GSK3 would stabilize GluA1 levels and increase its surface expression, promoting LTP. Indeed, low GSK3 activity promotes LTP, whereas high GSK3 activity promotes LTD (Hooper et al., 2007; Peineau et al., 2009; Peineau et al., 2007). Surprisingly, inhibition of GSK3 was previously shown to reduce surface expression of AMPA (Wei et al., 2010) and NMDA (Chen et al., 2007) receptors on cortical neurons. This discrepancy may reflect differences in synaptic transmission between hippocampal and cortical neurons. Alternatively, it may reflect difficulties in interpreting results from global inhibition of a multifunctional kinase, like GSK3. Inhibition of particular pools of GSK3 in different subcellular compartments (especially in neurons) or inhibiting phosphorylation of individual substrates, as performed in the present study, may be a more precise approach for delineating the role of GSK3 and its substrates in neurotransmission.

In summary, we have discovered a novel substrate of GSK3 that regulates cell- surface expression of the AMPA receptor GluA1 and is therefore likely to affect neurotransmission in the brain. Accordingly, deregulation of PI4KIIα activity is associated with a hyperactive phenotype in flies and corrected when treated with lithium,

109 |Page

Sec. 3.3. Discussion

the mainstay treatment for BD in the clinic. Together this suggests that deregulation of PI4KIIα activity is consistent with a role in the pathogenesis and/or treatment of BD, in which GSK3 activity is dysregulated.

110 |Page

CHAPTER 4: AAK1 phosphorylation by GSK3 promotes AP-2- mediated autophagic clearance.

4.1 Introduction

Receptor-mediated endocytosis from the plasma membrane via CCV is important for nutrient uptake, immune response, the activation of second messenger signalling and recycling of synaptic vesicles after neurotransmission. Clathrin and adaptor protein-2 complex (AP-2) are important coat proteins that assemble into a flexible lattice surrounding these endocytic vesicles (see reviews: (Kirchhausen, 1999, 2000). AP-2 is a multi-subunit complex that functions to link membrane cargoes to the clathrin framework (Ricotta et al., 2002). It is composed of two 100 kD subunits, the α and β2 adaptins, the µ2 subunit of 50 kD, and a small 17-kD 2 subunit (Kirchhausen, 1999). The α and β2 subunits mediate protein interactions with clathrin and other endocytic machinery components, such as AP180, dynamin, auxilin and Eps15 (Goodman & Keen, 1995; Hao et al., 1999; Shih et al., 1995), while the µ2 subunit recognises sorting signals present in the cytoplasmic tails of receptors, together mediating cargo recruitment during coated pit formation (Aguilar et al., 1997; Ohno et al., 1995; Owen & Evans, 1998). Although the function of 2 subunit is still not clear, it’s thought to be important for stabilisation of the entire complex (Collins et al., 2002).

Clathrin-mediated endocytosis is initiated by the recruitment of AP-2 to the plasma membrane, directing receptor/ligand recruitment into coated pits and coordinating the assembly of clathrin (Schmid, 1997). The assembly of the clathrin lattice, potentiated by its interactions with endocytic components including, AP180, epsin and amphiphysin (see review: (Owen et al., 2004)), induces invagination of the plasma membrane formating vesicles (Musacchio et al., 1999). Following clathrin-coating and curvature of the vesicle, the membrane protein amphiphysin recruits dynamin, which polymerises into a helical collar around the vesicle neck and mediates vesicle scission from the plasma membrane (Fournier et al., 2003; Roux et al., 2006). The formed CCV containing its cargo then undergoes coat disassembly mediated by accessory proteins including, auxilin and the ATPase Hsc70 (70-kDa heat shock cognate protein) (Hannan et al., 1998; Massol et al., 2006; Newmyer & Schmid, 2001), which releases components for recycling back to the cell surface. The uncoated vesicles fuse with early/sorting endosomes for trafficking

111 |Page

Sec. 4.1. Introduction

to recycling endosomes and back to the cell surface or towards late endosomes and lysosomes for degradation (see review: (Traub, 2009)) (Fig.4.1).

Phosphorylation plays a key role in regulating clathrin-mediated endocytosis, especially in selecting certain receptors and cargo proteins for inclusion in CCVs (Olusanya et al., 2001). For example, phosphorylation of the µ2 subunit of AP-2 is thought to be a key regulatory step in cargo recognition, whereby it enhances binding affinity of AP-2 for tyrosine-based (Yxx) sorting motifs found on many internalised membrane proteins, such as TfnR (Fingerhut et al., 2001; Ohno et al., 1995; Ricotta et al., 2002). AAK1 is a Ser/Thr kinase that binds AP-2 in vivo and was identified as the kinase responsible for phosphorylating the µ2 subunit (Conner & Schmid, 2002) (Fig.4.1). Furthermore, it has been implicated in receptor sorting functions after internalisation to coordinate recycling (Henderson & Conner, 2007). Loss of AAK1 activity and reduced µ2 phosphorylation prevents efficient binding of AP-2 to sorting signals of receptor proteins, reducing internalisation into the cell (Ricotta et al., 2002). Therefore, AAK1 appears to be a relatively new and understudied regulatory of receptor endocytosis and trafficking.

Figure 4.1. Schematic representation of clathrin-mediated endocytosis depicting the role of AAK1 and AP-2 complexes. Clathrin-mediated endocytosis is initiated by the recruitment of AP- 2 to the plasma membrane, directing receptor/ligand recruitment into coated pits and coordinating the assembly of clathrin. AAK1 phosphorylation of the µ2 subunit of AP-2 enhances binding affinity for receptor proteins, stimulating sequestration/recruitment into CCVs. (Figure adapted from (Sebastian et al., 2006)).

112 |Page

Sec. 4.1. Introduction

4.1.1 Adaptor-associated protein kinase 1 (AAK1)

AAK1 is a member of the Ark1/Prk1 family of Ser/Thr kinases that functions during receptor-mediated endocytosis. Two splice variants have been discovered: a short form encoding an 863-amino acid protein (93 kDa), and a long form (AAK1L) encoding a 960 amino acid protein (~140 kDa) that contains an additional C-terminal clathrin- binding domain (CBD) (Henderson & Conner, 2007) (Fig.4.2). These isoforms are functionally similar, with no difference in their ability to phosphorylate AP-2 (Henderson & Conner, 2007).

Figure 4.2. Schematic representation of the domain structure of the short and long isoform of human AAK1. The conserved N-terminal kinase domain and middle domain enriched in glutamine, proline and alanine (QPA) are shared by both isoforms. However the long isoform has an additional CBD at its C-terminus.

AAK1 is ubiquitously expressed but is enriched in the brain at presynaptic terminals in neurons, where it colocalises with clathrin and AP-2 in clathrin-coated pits at the leading edge of cells and also found on trafficking vesicles/organelles in the peri- nuclear region (Conner & Schmid, 2002; Henderson & Conner, 2007). AAK1 directly binds adaptin α of the AP-2 complex and phosphorylates the µ2 subunit at a single threonine residue (Thr156) (Ricotta et al., 2002). Although phosphorylation at this site is not obligatory for receptor uptake, it was found to enhance AP-2 function during receptor- mediated endocytosis both in vitro and in vivo (Motley et al., 2006; Olusanya et al., 2001; Ricotta et al., 2002). Insights obtained from the resolved crystal structure of AP-2 has revealed that µ2 phosphorylation likely induces a conformational change that exposes the YxxΦ binding motif for AP-2 binding, promoting its recruitment to membrane receptors (Collins et al., 2002; Henderson & Conner, 2007). Interestingly, AAK1 kinase activity towards µ2 was found to be stimulated by the assembly of clathrin, suggesting that AAK1 is activated in clathrin-coated pits to enhance cargo recruitment and efficient internalisation (Conner et al., 2003; Jackson et al., 2003). However, AAK1 overexpression in vitro was shown to reduce AP-2-dependent internalisation of TfnR and

113 |Page

Sec. 4.1. Introduction

LRP into clathrin-coated vesicles by sequestering AP-2 away from the cell surface and into the cytosol (Conner & Schmid, 2002). In a later study, AAK1 was demonstrated to be important for TfnR recycling from early/sorting endosomes back to the cell surface, since AAK1 knockdown by siRNA impaired this process (Henderson & Conner, 2007). Therefore, it is possible that in the absence of clathrin, AAK1 binds to the AP-2 complex and inhibits cargo recruitment. However, in the presence of clathrin, its kinase activity is increased, leading to phosphorylation of AP-2, promoting cargo recruitment. Alternatively, AAK1 may be important for clathrin coat disassembly following vesicle internalisation. Clearly, the precise role of AAK1 in endocytosis is not well understood and required further investigation.

4.1.2 AAK1 kinase activity coordinates the assembly of endocytic machinery

AAK1 has also been implicated in phosphorylating and coordinating the endocytic adaptor protein Numb, believed to function at multiple steps during clathrin- coated pit assembly and endocytosis. While its precise function is still not completely clear, Numb’s role in receptor-mediated endocytosis is supported by its interactions and colocalisation with AP-2. AAK1 has been shown to bind to and phosphorylate Numb at Thr102 (Sorensen & Conner, 2008), mediating redistribution of Numb to perinuclear endosomes and away from the plasma membrane (Sorensen & Conner, 2008). Phosphorylation by AAK1 may be an important priming step that is a necessary for subsequent Numb phosphorylation by additional kinases including, atypical protein kinase C (aPKC) or calmodulin-dependent protein kinase 1 (CaMK1) (Sorensen & Conner, 2008), further potentiating Numbs release from clathrin-coated pits into the cytoplasm (Nishimura & Kaibuchi, 2007; Smith et al., 2007; Tokumitsu et al., 2005; Tokumitsu et al., 2006). Together, AAK1-mediated phosphorylation of Numb regulates its interaction with endocytic components during coated-pit assembly and may be critical in promoting clathrin-coated pit maturation through redistribution/trafficking of Numb.

4.1.3 AAK1 role in neurodevelopment and function

Notch signalling is critical for neurogenesis during development, neural differentiation and mature brain function (see review: (Alberi et al., 2013)). Numb antagonises Notch when bound to AP-2 in CCVs. AAK1’s phosphorylation of Numb redistributes it to recycling endosomes, stabilising Notch activity. Contrary to AAK1, Numb inhibits Notch by targeting it for polyubiquitination and degradation via

114 |Page

Sec. 4.1. Introduction

proteolysis (Gupta-Rossi et al., 2011). Therefore, AAK1 phosphorylation of Numb is likely important for positively regulating Notch signalling and enhancing endocytosis through CCV assembly. This occurs not through direct phosphorylation of the Notch receptor itself, but through its interaction with other endocytic partners, especially Numb. This potentially implicates AAK1 as have an important role in regulating Notch signalling in the brain.

AAK1 has also been shown to negatively regulate Neuregulin-ErbB4 signalling, a crucial pathway in brain development that is implicated in BD and schizophrenia (Goes et al., 2011; Pan et al., 2011; Yokley et al., 2012). Neuregulin-ErbB4 signalling regulates neuronal migration, axon guidance and synapse formation during neurodevelopment (Mei & Xiong, 2008), as well as synaptic plasticity and survival of mature neurons (Mei & Xiong, 2008). RNAi mediated knockdown or pharmacologic inhibition of AAK1 (using the non-selective kinase inhibitor K252a) in PC12 cells resulted in elevation of ErbB4 receptor abundance and redistribution to the plasma membrane, enhancing neuregulin-1 (Nrg1) driven neuritogenesis (Kuai et al., 2011). It is possible that loss-of-function of AAK1 may reduce internalisation of ErbB4, increasing its abundance at the cell surface. Alternatively, AAK1 inhibition may redirect ErbB4-containing vesicles back to the plasma membrane, instead of towards the degradation pathway i.e. lysosomes (Kuai et al., 2011). In any case, reducing AAK1 activity causes sustained activation of ErbB4 signalling that leads to enhanced Nrg1-dependent neuritogenesis.

Overall, AAK1 has been implicated in the endocytic and early vesicle trafficking pathways in cells. Given that AAK1 regulates Notch and ErbB4/neuregulin signalling, both of which have been implicated in BD and schizophrenia, we speculate that AAK1 may also be involved in these mood disorders. In addition, these signalling pathways are critical for brain development, healthy brain function and neuron survival. Indeed, a genome-wide association study has identified polymorphisms in the AAK1 gene as a potential risk factor for the age of onset of Parkinson’s disease (PD) (Latourelle et al., 2009). Furthermore, dysfunction of AAK1 has recently been implicated in amyotrophic lateral sclerosis (ALS) pathology (Shi et al., 2014). Together, this suggests a potentially important role for AAK1 in several neurological disorders, warranting further investigation.

Here we demonstrate that AAK1 is phosphorylated by GSK3 in the brain and investigate the physiological function of its phosphorylation in neurons.

115 |Page

Sec. 4.2. Results

4.2 Results

4.2.1 Mapping and characterisation of GSK3 phosphosites in AAK1

AAK1 contains two conserved GSK3 phosphorylation consensus sequences in its central region. Thr620 and Thr674 are the putative GSK3 target sites, while Ser624 and Ser678 are their respective priming sites (Fig.4.3A). To determine which of these sites are targeted by GSK3 in cells, we subjected AAK1 to the BIPPS assay, which was previously developed in our lab for identifying novel GSK3 substrates (see Methods, section 2.2.1) (Farghaian et al., 2011). Briefly, AAK1 was cloned into a mammalian expression vector (with a C-terminal FLAG-tag) and expressed in HEK293 cells in the presence or absence of the highly-specific GSK3 inhibitor CT99021 (Bain et al., 2007). It was pulled-down via its C-terminal FLAG-tag and subjected to an in vitro kinase assay with recombinant GSK3β and radiolabelled ATP. If AAK1 is a physiological target of GSK3, transfection into HEK293 cells should result in phosphorylation by endogenous GSK3. This will be blocked by the GSK3 inhibitor, leaving the GSK3 target sites vacant but any required priming events intact. In the subsequent in vitro kinase assay, recombinant GSK3 should be able to incorporate more radiolabelled phosphate into AAK1 isolated from inhibitor-treated cells compared to untreated cells. If so, this indicates it is a good substrate for GSK3 in vitro and in cells.

Significantly more radiolabelled phosphate was incorporated into AAK1 isolated from inhibitor-treated (Fig.4.3B, C; lane 2) than non-treated cells (lane 1), indicating it is a bona fide GSK3 substrate. Mutating the putative priming site S678 to a non- phosphorylatable alanine (S678A), completely blocked phosphate incorporation into AAK1 (lane 3), while mutation of Ser624 (S624A) had no effect. This indicates that the majority of radiolabelled phosphate was incorporated into S674, but not Thr620. At first glance, this indicates that Thr620 is not a physiological target site for GSK3. However, we have previously found that genuine GSK3 substrates can give negative results in the BIPPs assay due to their phosphorylation sites being relatively resistant to dephosphorylation by phosphatases (e.g. β-adducin (Farghaian et al., 2011); CRMP2 (Cole et al., 2008); PI4KIIα (Robinson et al., 2014)). That is, while addition of the inhibitor CT99021 successfully blocks phosphorylation of the substrate by GSK3, the site may be resistant to dephosphorylation by phosphatases, and so remains phosphorylated. Indeed, the primary sequence surrounding Thr620 in AAK1 is very similar to the

116 |Page

Sec. 4.2. Results

phosphatase-resistant GSK3 sites in β-adducin, CRMP2 and PI4KIIα (Fig.4.4). Therefore, we investigated the possibility that Thr620 might be a phosphatase-resistant target site for GSK3.

Figure 4.3. AAK1 is a physiological substrate of GSK3. (A) Sequence alignment of putative GSK3 target sites bordering the QPA-enriched domain and the CBD1 domain of AAK1 from various species. GSK3 sites and the priming sites at the +4 position are numbered and underlined. (B) AAK1 was expressed in HEK293 cells that were untreated (lane 1), treated with CT99021 (lane 2), or expressed priming site mutant forms of AAK1, S678A (lane 3), S624A (lane 4). AAK1 was immunoprecipitated via its C-terminal Flag-tag and subjected to an in vitro kinase assay with recombinant GSK3 and radiolabelled ATP. (C) The stoichiometry of phosphate incorporation in (B) is shown as a graph (Mean ± S.E.M., * = p<0.05; (T-test), n=3).

117 |Page

Sec. 4.2. Results

Figure 4.4. Sequence alignment of substrates containing a conserved GSK3 phosphorylation consensus sequence and phosphatase-resistant GSK3 sites.

4.2.1.1 AAK1 phosphosite Thr620 is resistant to dephosphorylation by phosphatases

Phosphospecific antibodies were generated to both putative GSK3 phosphosites (Thr620 and Thr674) and their specificity was validated by Western blotting of wild type and phospho-mutant forms of AAK1 (Fig.4.5A). Thr674 was confirmed as a GSK3 target site, since phosphorylation was inhibited by mutation of the priming site Ser678 to a non- phosphorylatable alanine (Fig.4.5A), which blocks subsequent phosphorylation of Thr674 by GSK3. Treatment with CT99021 and lithium also reduced phosphorylation of Thr674 (Fig.4.5B). Meanwhile, phosphorylation of Thr620 was reduced by mutation of Ser624, indicating it is priming-dependent, consistent with it being a GSK3 target site. However, it was not affected by CT99021 or lithium treatments. A likely explanation for these observations is that in the former experiment, mutation of the priming site reduced the efficiency of phosphorylation of Thr620 by GSK3 in the first place, independent of phosphatase activity. While in the latter experiment, GSK3 inhibitors reduced the activity of GSK3, but the Thr620 site was not able to be removed by phosphatases, so the stoichiometry of phosphorylation remained unchanged.

The relative resistance of pThr620 to phosphatases was further investigated by transfection of HEK293 cells with AAK1, followed by harvesting in lysis buffer without phosphatase inhibitors. Lysates were incubated at 30°C for up 60 min with added MgCl2 to activate endogenous phosphatases and relative rates of dephosphorylation at each site were measured using Western blotting. Efficient dephosphorylation of pThr674 was observed over time with endogenous phosphatases removing the majority of phosphate after 60 min (Fig4.6A; graphed in B (dotted-line)). However, substantially less phosphate

118 |Page

Sec. 4.2. Results

was removed at Thr620 (Fig.4.6A; graphed in B (solid line)), indicating that this site is substantially more resistant to endogenous phosphatases than Thr674.

Furthermore, a dose-response assay using increasing concentrations of CT99021 on AAK1-transfected HEK293 cells revealed that low concentrations of CT99021 were sufficient to reduce phosphorylation of Thr674 (Fig.4.6C; graphed in D (dotted-line)), but not Thr620 (solid line). Even high concentrations of CT99021 did not significantly reduce phosphorylation of Thr620 (Fig.4.6C; graphed in D (solid line)). Altogether, these observations demonstrate that Thr620 is a bona fide GSK3 target site that is relatively resistant to dephosphorylation by phosphatases both in vitro and in cells.

Figure 4.5. Phosphorylation of AAK1 by GSK3. (A) HEK293 lysates transfected with AAK1 wild type and phosphosite mutants were subjected to Western blot analysis using custom phosphospecific antibodies, as well as an antibody recognising the C-terminal FLAG-tag as a loading control. (B) GFP or AAK1 wild type was expressed in HEK293 cells that were untreated (lane 1), treated with CT99021 (2 µM) (lane 2) or treated with 20 mM LiCl (lane 3) for 4 h. Lysates were subjected to Western blotting using the pThr620, pThr674 and FLAG antibodies.

119 |Page

Sec. 4.2. Results

Figure 4.6. Kinetics of dephosphorylation of AAK1. (A) HEK293 cell were transfected with AAK1 and cell lysates (without phosphatase inhibitors) were incubated at 30°C for up to 60 min in the presence of 10 mM MgCl2 to activate endogenous phosphatases. Dephosphorylation by endogenous phosphatases was determined by Western blotting using phosphospecific antibodies (pThr620, pThr674), as well as an antibody recognising the C-terminal FLAG-tag as a loading control. (B) Relative dephosphorylation by endogenous phosphatases in (A) is presented as a graph. (C) HEK293 cells transfected with wild type AAK1 were treated with various concentrations of CT99021 for 4 h. Lysates were subjected to Western blotting using pThr620, pThr674 and FLAG antibodies. (D) Relative dephosphorylation induced by CT99021 treatment in (C) is presented as a graph. (Mean ± S.E.M, n=3).

120 |Page

Sec. 4.2. Results

4.2.1.2 GSK3 phosphorylation of AAK1 increases its stability in cells

Next, we investigated the effect that phosphorylation has on AAK1 stability in the cell. Mutation of the putative priming site Ser624 to a non-phosphorylatable alanine significantly reduced AAK1 abundance in the cell, while dephospho-mutant S678A was also reduced, although it did not quite reach significance (p=0.1; Fig.4.7A). Treatment with the translation inhibitor anisomycin demonstrated that dephospho-mutant forms of AAK1 are more rapidly degraded in the cell than wild type, particularly for the S624A mutant (Fig.4.7B). Together this suggests that AAK1 phosphorylation by GSK3 increases its stability in cells.

Figure 4.7. Phosphorylation regulates the abundance of AAK1. (A) Lysates from HEK293 cells transfected with wild type or phosphomutant AAK1 were subjected to Western blotting for the C-terminal FLAG tag and actin as a loading control. Relative abundance of AAK1 was quantitated as the ratio between FLAG and actin and is presented as a graph. (B) HEK293 cells transfected with wild type or phosphomutant AAK1 were treated with 10 µg/ml anisomycin for 2 h. The ratio of FLAG to actin after anisomycin treatment is presented as a graph. (Mean ± S.E.M, * = p<0.05; *** =p<0.005; n.s., not significant; T-test).

121 |Page

Sec. 4.2. Results

4.2.2 Phosphorylation of AAK1 by GSK3 promotes phosphorylation of Adaptin µ and dissociation from the AP-2 complex

4.2.2.1 GSK3 phosphorylation of AAK1 promotes dissociation from AP-2 complex

Next, we investigated binding partners of AAK1 and the role that phosphorylation by GSK3 may have on this process. Wild type and dephospho-mutant forms of AAK1 were expressed in HEK293 cells, immunoprecipitated and subjected to Western blotting using antibodies to various trafficking proteins. AAK1 bound to adaptin α and β of the AP-2 complex, but not adaptin  (AP-3 complex), γ (AP-1 complex), clathrin or EEA1 (Fig.4.8A). In HeLa cells, endogenous AAK1 localises to the same cytoplasmic and peri- nuclear regions as endogenous adaptin α and β (Fig.4.8B). Together, this suggests that AAK1 is indeed associated with the AP-2 complex as previously reported (Conner & Schmid, 2002).

Mutation of AAK1 phosphosites to a non-phosphorylatable alanine individually increased binding to adaptin α and β in immunoprecipitation experiments compared to wild type AAK1 (i.e. S624A and S678A; Fig.4.9A). Similarly, treatment with the GSK3 inhibitor CT99021 also increased binding to adaptin α (Fig.4.9D, E). Interestingly, AAK1-S624A mutant had a greater binding affinity for adaptin α and β than the S678A mutant, suggesting that phosphorylation at Thr620 by GSK3 has a greater effect on disrupting AP-2 binding (Fig.4.9B, C). Together, these observations show that phosphorylation of AAK1 by GSK3 reduces binding to the AP-2 complex.

122 |Page

Sec. 4.2. Results

) HeLa cells were stained forHeLa) cells B ) HEK293 cells transfected with AAK1 wild type were immunoprecipitated via its C- its via immunoprecipitated were type wild AAK1 with transfected cells HEK293 ) A ( AAK1 binds to adaptin α and β of the AP-2 complex. AP-2 the of β and α adaptin to binds AAK1

Figure 4.8. Figure loadingFLAGcontrol. various as a ( FLAG-tagasWesternproteins, as well to trafficking for terminal andblotting subjected endogenous AAK1 (red), adaptin α, β (green) and nuclei DAPI (blue).

123 |Page

Sec. 4.2. Results

Figure 4.9. Phosphorylation by GSK3 regulates AAK1 binding to the AP-2 complex. (A) Western blots of immunoprecipitates of AAK1 wild type and phosphomutants (S624A and S678A) for adaptin α, β (AP-2 complex) and FLAG. (B) Relative amount of adaptin α binding to AAK1 is shown as a graph. (C) Relative amount of adaptin β binding to AAK1 is shown as a graph. (D) Western blots of immunoprecipitates of GSK3 inhibitor-treated wild type AAK1 for adaptin α and FLAG. (E) Relative amounts of adaptin α binding to AAK1 is shown as a graph. (Mean ± S.E.M, * = p<0.05; *** =p<0.005; T-test).

124 |Page

Sec. 4.2. Results

4.2.2.2 Phosphorylation by GSK3 promotes AAK1-mediated phosphorylation of adaptin µ

AAK1 has previously been identified as the kinase responsible for phosphorylating the µ2 subunit of AP-2 (Ricotta et al., 2002). This was found to enhance binding to sorting signals during receptor-mediated endocytosis (Motley et al., 2006; Olusanya et al., 2001; Ricotta et al., 2002). Therefore, we speculated that GSK3 phosphorylation of AAK1 reduces its binding to the AP-2 complex, thus reducing phosphorylation of adaptin µ. To investigate this, wild type and dephospho-mutant forms of AAK1 were co-immunoprecipitated from lysates and blotted for adaptin µ, phospho- adaptin µ (pT156-adaptin µ) and FLAG. Both dephospho-mutant forms of AAK1 displayed increased binding to adaptin µ compared to wild type (Fig.4.10A, B). Furthermore, AAK1-S624A bound with stronger affinity than AAK1-S678A (Fig.4.10B), consistent with its higher affinity for adaptin α and β. Surprisingly, phosphorylation of adaptin µ was decreased in immunoprecipitates with AAK1-S624A compared to wild type (Fig.4.10C). A possible explanation for this is the S624A mutation inhibits the kinase activity of AAK1. To test this, wild type and AAK1-S624A (as well as S678A) were immunoprecipitated from HEK293 cell lysates and subjected to an in vitro kinase assay with the generic kinase substrate myelin basic protein (MBP) and radiolabelled ATP. Wild type and S678A mutant AAK1 displayed identical kinase activity, while the S624A mutant displayed increased activity (Fig.4.11). Therefore, the S624A mutation does not reduce the kinase activity of AAK1, instead it actually increases it. The reason for this surprising result remains unresolved and requires further investigation.

125 |Page

Sec. 4.2. Results

) Relative amounts of p-adaptin µ binding tobinding µ p-adaptin of amounts Relative )

C

for phosphomutants and type wild AAK1 of immunoprecipitates of blots Western ) A (

<0.005; T-test). p

<0.05; *** = p

graph.( a as shown is AAK1 to binding µ adaptin ofamounts Relative ) B

µ. adaptin of phosphorylation reduces AAK1-S624A

Figure 4.10. Figure ( FLAG. and µ p-adaptin µ, adaptin AAK1 shown a is graph.as (Mean S.E.M, ± * = 126 |Page

Sec. 4.2. Results

) B

in vitro ) ) is shown A

A1S2A a icesd kinase increased has AAK1-S624A AK wl tp, 64 ad S678A and S624A type, wild AAK1 )

(Mean S.E.M,± n=3). A (

4.11. Figure activity. were immunoprecipitated and subjected to an ( ATP. radiolabelled and MBP with assay kinase The relative phosphate incorporation in ( a graph.as

127 |Page

Sec. 4.2. Results

4.2.3 Phosphorylation of AAK1 by GSK3 regulates vesicular trafficking

4.2.3.1 Phosphorylation of AAK1 promotes Transferrin recycling in HeLa cells

To determine if AAK1 phosphorylation by GSK3 regulates vesicular trafficking in cells, HeLa cells were transfected with wild type or phospho-mutant AAK1, then incubated with fluorescently-labelled Tfn for 45 min at 4°C. Unbound Tfn was removed by washing, followed by incubation at 37°C for 15 min. Cells were fixed and localisation of Tfn was determined using fluorescence microscopy (Fig.4.12A). There was a small increase in the rate of Tfn internalisation in cells transfected with AAK1 wild type or phospho-mutant S624A forms compared to empty vector, although this did not reach significance (Fig.4.12B). However, recycling back to the cell surface was significantly reduced at 15 min in the presence of AAK1 phospho-mutant S624A compared to empty vector, wild type and phospho-mutant S678A (Fig.4.12C).

In a separate experiment, we investigated the role of AAK1 phosphorylation on the trafficking of TfnR to degradative pathways. TfnR is a well characterised plasma membrane protein that travels between the plasma membrane and internal membrane compartments. While the majority of TfnR is recycled back to the cell surface through recycling endosomes, a proportion of TfnR is constitutively trafficked via the degradation pathway to lysosomes (Matsui et al., 2011). Lysate from HEK293 cells co-transfected with GFP-tagged TfnR and either empty vector, AAK1 wild type or phosphomutant AAK1, were subjected to Western blotting using a GFP antibody to determine the relative abundance of TfnR in cells. Interestingly, co-transfection with AAK1 wild type or phospho-mutant S678A increased the total abundance of TfnR, whereas the AAK1- S624A phospho-mutant had no significant effect on TfnR levels compared with empty- vector control (Fig4.12D and quantitated in E). This suggests that AAK1 might favour trafficking towards recycling endosomes and the cell surface (for itself and cargo e.g. TnfR). Phosphorylation of Thr620/Ser624 is required for this, since dephospho-S624A reduced the rate of recycling (Fig.4.12A, C) and reduced the abundance of TnfR compared to wild type AAK1. This might suggest that instead of trafficking through recycling vesicles back to the cell surface, AAK1-S624A favours the degradative pathways towards the lysosomes. This is consistent with reduced cellular abundance of AAK1-S624A (Fig.4.7). One way to test this would be to treat cells with the lysosome inhibitor Bafilomycin A, since blocking the lysosome would have different effects on the

128 |Page

Sec. 4.2. Results

abundance of wild type and phospho-mutant forms of AAK1, depending on whether they were preferentially trafficked toward the lysosome or cell surface. These experiments were attempted, but were not possible due to the relatively long half-life of AAK1 (>8 h; data not shown) and the fact that Bafilomycin A treatment was toxic to cells for time periods longer than this.

Instead, we generated a new mutant form of AAK1 that has reduced binding to AP-2 (F698A mutant) (Jha et al., 2004). AAK1-F698A successfully reduced binding to adaptin α in co-immunoprecipitation assays (Fig.4.13A, B) and was associated with a small but significant decrease in total abundance compared with wild type AAK1 (Fig.4.13C, D). This may suggest that AAK1-F698A has reduced stability in the cell, however, whether this is the result of increased trafficking to the lysosome is yet to be determined.

129 |Page

Sec. 4.2. Results

Figure 4.12. Phosphorylation of AAK1 by GSK3 is required for efficient trafficking of Tfn in cells. (A) HeLa cells transfected with empty vector, AAK1 wild type, S624A or S678A were pulse- labelled with fluorescent Tfn, then incubated in Tfn-free medium for 15 min. Cells were analysed using fluorescence microscopy for Tfn (green), transfected cells (mCherry, red) and cell nuclei (DAPI, blue). Images are representative of the 15 min time point. (B) Relative rates of Tfn internalisation were scored and results are presented as a graph (blinded, average of 3 independent experiments, n=150 cells for each group). (C) Relative rate of Tfn recycling back to the cell

130 |Page

Sec. 4.2. Results

surface at 15 min were scored and results are presented as a graph. (D) Lysates from HEK293 cells co-transfected with GFP-tagged TfnR and either empty vector, AAK1 wild type or phosphomutant AAK1, were subjected to Western blotting for GFP and actin as a loading control. (E) Relative abundance of TfnR was quantitated as the ratio between GFP and actin and presented as a graph. (Mean ± S.E.M, * = p<0.05; ** = p<0.01; *** =p<0.005; n.s., not significant; T-test).

Figure 4.13. AAK1-F698A has reduced binding to adaptin α (AP-2) and reduced stability in the cell. (A) HEK293 cells transfected with GFP control, AAK1-WT or F698A mutant were immunoprecipitated via their C-terminal FLAG-tags and subjected to Western blotting for adaptin α and FLAG. (B) The ratio of adaptin α to FLAG in (A) is presented as a graph. (C) AAK1-WT or F698A mutant lysates were subjected to Western blotting for C-terminal FLAG and GAPDH as a loading control. (D) The ratio of FLAG to GAPDH in (C) is presented as a graph. (Mean ± S.E.M, * =p<0.05; *** =p<0.005; T-test).

131 |Page

Sec. 4.2. Results

4.2.3.2 AAK1 dissociation with the AP-2 complex increases recycling

We further investigated the role of AAK1’s binding affinity for the AP-2 complex in regulating trafficking and recycling in cells. To investigate this, stable HeLa cell lines were generated expressing shAAK1 against human AAK1 (see section 4.2.5.1 for delineation of the shAAK1 HeLa cells used in this experiment). HeLa cells expressing shAAK1 were subjected to our single-round Tfn recycling assay (37°C for 15 min). Recycling of Tfn back to the cell surface was increased approximately 50% at 15 min in shAAK1 cells compared to scrambled control (Fig.4.14A). This contrasts the previous finding of Henderson and Conner (2007) who found that AAK1 knockdown reduces Tfn recycling. The reason for this discrepancy is not yet clear. In a separate experiment, AAK1 mutants resistant to shAAK1 knockdown were generated that blocked interaction with AP-2 (F698A) (Jha et al., 2004) in combination with phosphosite mutation (F698A- S624A and F698A-S678A). Validation of the knockdown and shRNA-resistant constructs are shown in Fig.4.14C. The transfection of F698A mutants in shAAK1 cells had a dramatic increase in Tfn recycling. However, AAK1-F698A plus GSK3 phosphosites mutation had no further effect on Tfn recycling (Fig.4.14B). This suggests that mutation of F698A surpasses any affect GSK3 phosphorylation may have on trafficking. Furthermore, the transfection of wild type AAK1 into shAAK1 cells did not rescue the increased Tfn recycling observed in these knockdown cells (Fig.4.14B). It’s possible that high overexpression of AAK1, above normal endogenous levels, may disrupt the fine balance of AP-2 trafficking to/from the cell surface.

Our observations suggest that AAK1 dissociation from AP-2 promotes Tfn recycling back to the cell surface. AAK1-S624A has increased binding to AP-2 and decreases recycling. Whereas, AAK1-F698A does the opposite; reduced binding to AP- 2 and increased Tfn recycling. This suggests that phosphorylation at Thr620/Ser624 is important for dissociation with AP-2 and required for efficient Tfn recycling in cells. Therefore, our results are consistent with a model whereby AAK1 antagonises Tfn recycling back to the cell surface (Fig.4.15). This is dependent on AAK1 binding to the AP-2 complex. It is possible that AAK1 sequesters AP-2 in the peri-nuclear region, although this is yet to be confirmed. Phosphorylation by GSK3 reduces the affinity of AAK1 with AP-2, releasing it back to the cell surface to promote Tfn recycling. Therefore, the primary role of phosphorylation by GSK3 is likely to be regulating the binding affinity of AAK1 to AP-2. It is therefore possible that the discrepancy between

132 |Page

Sec. 4.2. Results

our results and those of Henderson and Conner (2007) is due to differing levels of GSK3- mediated phosphorylation of AAK1 between our cells.

Figure 4.14. AAK1 and its ability to bind AP-2 is critical for Tfn sorting functions in cells. (A) Relative rates of Tfn recycling at 15 min in HeLa cell stably expressing shAAK1 or scrambled control (scram) were scored and results are presented as a graph. (B) Relative rates of Tfn recycling at 15 mins in shAAK1 cells transfected with empty vector, AAK1 wild type, F698A, F698-S624A or F698A-S678A mutants were scored and results are presented as a graph. (Blinded, average of 3 independent experiments, n=150) (Mean ± S.E.M, * = p<0.05; *** =p<0.005; n.s., not significant; T-test). (C) Confirmation of effective knockdown of AAK1 by shRNA (shAAK1) and the resistance of shRNA-resistant AAK1 (AAK1-Res). Wild type AAK1 or shRNA resistant AAK1 was co-transfected into HEK293 cells with scrambled or AAK1- shRNA. Lysates were subjected to Western blotting for FLAG and GAPDH.

133 |Page

Sec. 4.2. Results

Figure 4.15. Model of GSK3-mediated phosphorylation of AAK1 promoting the recycling of Tfn. Phosphorylation of AAK1 promotes the activation and release of AP-2 for subsequent trafficking of itself and cargoes (e.g. Tfn) to recycling and secretory endosomes. In its non-phosphorylated form, AAK1 binds with high affinity to the AP-2 complex, sequestering vesicles inside cells and towards sites of degradation.

134 |Page

Sec. 4.2. Results

4.2.4 Investigating the physiological function of AAK1 in Drosophila melanogaster

4.2.4.1 Depletion of AAK1 in Drosophila produce no behaviours that correlate with BD

Next, we investigated the physiological function of AAK1 in the fruit fly Drosophila melanogaster, screening for behaviours that correlate with symptoms of BD. To test this, AAK1 gene expression was specifically knocked down in neurons of flies using the UAS-Gal4 system. This was accomplished by crossing transgenic flies harbouring a UAS-AAK1-RNAi transgene with the neural specific nSyb-Gal4 driver. Flies were then subjected to DAMs analysis to screen for changes in locomotor activity or circadian rhythm. Neural-specific AAK1 knockdown in flies had no obvious effect on fly locomotor activity levels or circadian rhythm (Fig.4.16A). In addition, we explored reward seeking behaviour in our flies to assess for anhedonia. Anhedonia is a major symptom of mood disorders (especially the depressive disorders), where patients have reduced ability to experience pleasure (see review: (Treadway & Zald, 2011)). Rewards serve to elicit approach and consummatory behaviours, increase the frequency and intensity of the behaviours and induce feelings of pleasure or positive emotional states (see reviews: (Gorwood, 2008; Koob, 1996)). For example, depressed patients have higher sweet taste perception thresholds and reduced reward responsiveness (Berlin et al., 1998). Therefore, reward responsiveness is a good measure for assessing anhedonia in animal models such as Drosophila (Stafford et al., 2012). To test for anhedonia, AAK1- RNAi flies were subjected to the Café assay to measure their preference towards sweetened food (reward). Control and neural-specific AAK1-RNAi flies (Males, 3-5 days) were placed in vials (n=10) with a choice of two capillaries containing food solutions of either high (5% Yeast/0.25M Sorbitol+2.5% sucralose) or low (5% Yeast/0.25M Sorbitol+0.1% sucralose) intensity of sweetness for 48 h. Drosophila preferentially consume high caloric sugar sources, independently of taste input (Dus et al., 2011). Therefore, in order to test reward behaviour associated with sweetness only, both food solutions contain the same caloric content with variation in sweetness using the non-caloric sweetener sucralose. Both control and AAK1-RNAi flies exhibited a strong and consistent preference for sweetened food (Fig.4.16B), without any significant difference between them. Together with the DAMs locomotor and circadian rhythm

135 |Page

Sec. 4.2. Results

results, these observations suggest that AAK1 is unlikely to be associated with behaviours/symptoms related to BD.

136 |Page

Sec. 4.2. Results

dcr2;AAK1-

Peeec o AAK1-RNAi of Preference ) B

( d. 6 over averaged line) red ,

Locomotor) activity (count/30 ofmin) AAK1-RNAi knockdown ( A ( W1118,dcr2;;nSyb-Gal4 (control) given the choice between two non-caloric sugars of different sweetness (0.1% sucralose, red; 2.5% sucralose, blue). sucralose, 2.5% red; sucralose, (0.1% sweetness different of sugars non-caloric two between choice the given (control) (le ie ad rvr ny oto ( control only driver and line) (blue ) W1118 AAK1 depletion does affectnot or activity anhedonia in flies.

Figure 4.16. Figure RNAi(109507);nSyb-Gal4 and flies knockdown (Mean S.E.M.,*± = p<0.05; T-test).

137 |Page

Sec. 4.2. Results

4.2.4.2 AAK1 depleted flies are susceptible to starvation-induced death

In order to identify a physiological function for AAK1, we screened the AAK1- RNAi flies for a number of other behavioural and physiological phenotypes. Flies with neural-specific AAK1-knockdown exhibit no significant differences in food intake (Fig.4.17A) or average body weight (Fig4.17B) compared to control flies. Furthermore, they displayed normal locomotor function and climbing ability in the negative geotaxis assay (Fig.4.17D). Next, AAK1-RNAi flies were subjected to survival assays under starvation conditions (no food source) or neurotoxic stress (exposed to nicotine). Interestingly, AAK1-depleted flies displayed increased susceptibility to death by starvation (also reproduced using a second AAK1-RNAi line, Fig.4.17E and F), but not nicotine excitotoxicity (Fig.4.17G). In addition, AAK1-RNAi flies had decreased fat mass (Fig.4.17C). One possible explanation for these observations is that the decreased fat mass of the AAK1-RNAi flies reduces their survival time in the starvation assay due to lower energy reserves. Another explanation is that AAK1 has a role in regulating autophagy, since it is a physiologically important process in regulating cell survival during starvation (Li et al., 2013) and ketosis (Finn & Dice, 2005). Moreover, AP-2 is required for autophagosome production, which is an established binding partner of AAK1 (Rong et al., 2012; Tian et al., 2013; Tian et al., 2014). Therefore, we further investigated a potential role of AAK1 in autophagy (below). This process is implicated in several neurodegenerative diseases and is particularly evident in the brain of PD patients and in animal models of PD (Cheung & Ip, 2009; Xiong et al., 2013). Interestingly, AAK1 has previously been genetically associated with the age of onset of PD (Latourelle et al., 2009). It is possible that due to its role in vesicle trafficking events (above) and their close relationship to autophagy, that AAK1 might also be involved in autophagy. This could explain the increased susceptibility of AAK1-RNAi flies to starvation-induced death. Therefore, a potential role for AAK1 in PD was investigated further.

138 |Page

Sec. 4.2. Results

)

D dcr2;AAK1- ) ) and onlydriver control dcr2;AAK1-RNAi(109507);nSyb- ) ) and results are presented as a graph. ( ) TAG colourimetric assays were performed on performed were assays colourimetric TAG ) C dcr2;AAK1-RNAi(109507);nSyb-Gal4 W1118,dcr2;;nSyb-Gal4 ) Neural-specific knockdown of AAK1 ( AAK1 of knockdown Neural-specific ) A ( ) is presented as a graph. ( graph. a as presented is ) Aeae oy egt pr l, g o AK-Ni ncdw ( knockdown AAK1-RNAi of mg) fly, (per weight body Average ) B ) and driver only control ( ) were) placed vialsin and subjected to the café assay48 for h. Average cumulative food consumption W1118,dcr2;;nSyb-Gal4 W1118,dcr2;;nSyb-Gal4 dcr2;AAK1-RNAi(109507);nSyb-Gal4 ) and driver only control ( control only driver and ) AAK1 depleted flies are susceptible to starvation-induced death. starvation-induced to susceptible are flies depleted AAK1

) and) driver only control ( Figure 4.17. Figure ( graph. a as presented and Gal4 h 48 and 32 24, 16, at recorded was RNAi(109507);nSyb-Gal4 AAK1-RNAi knockdown ( geotaxisNegative (climbing 5 assay; cm, 10s) was performed on AAK1-RNAi knockdown (

139 |Page

Sec. 4.2. Results

) G (

W1118,dcr2;;nSyb- (solid line) (up to 60 h,60 to (up line) (solid ) ) Starvation survival curve

E (dotted line) and driver only driver and line) (dotted <0.05, Two-way ANOVA). Two-way <0.05,

p

<0.001, * = * <0.001,

p W1118,dcr2;;nSyb-Gal4) <0.05; Student’s t-test). (

p *** = = *** (dotted line) and driver only control ( controlonly driver and line) (dotted

dcr2;NAKJ35-RNAi;nSyb-Gal4) ± SEM; ±

(Mean

, not significant; * = n.s.

(dotted line) and driver only control ( control only driver and line) (dotted

dcr2;AAK1-RNAi(109507);nSyb-Gal4)

) Starvation survival curve of AAK1-RNAi #2 knockdown ( knockdown #2 AAK1-RNAi of curve survival Starvation ) F (

dcr2;AAK1-RNAi(109507);nSyb-Gal4) (solid line) (up to 72 h, 25°C) is presented as a graph a as presented is 25°C) h, 72 to (up line) (solid ) and results are presented as a graph. (Mean ± SEM; yw,dcr2;;nSyb-Gal4) (solid line) (up60 to presentedh, 25°C) graph.is as a W1118,dcr2;;nSyb-Gal4 ( ( knockdownAAK1-RNAi of graph. a as presented is 25°C) ( control ( knockdownAAK1-RNAi of curve excitotoxicity Nicotine Gal4)

140 |Page

Sec. 4.2. Results

4.2.5 AAK1 depletion increases autophagy processes in cells

4.2.5.1 HeLa cells stably expressing shAAK1

To investigate AAK1’s role in autophagy, stable HeLa cell lines were generated expressing shAAK1 against human AAK1. To do this, an shAAK1 vector containing a neomycin selectable marker was transiently transfected into HeLa cells. G418 (Geneticin; Life Technologies) was used as a positive selection marker for cells containing the G418 resistance gene (neo) as well as shAAK1. Colonies were picked to generate a homogenous clonal population of HeLa cells stably expressing shAAK1. Substantial reduction in AAK1 protein (~50%) was observed for both transfected (Fig.4.18A) and endogenous (Fig.4.18B) AAK1 in cells stably expressing shAAK1 compared with scrambled control. Two splice variants of AAK1 exist in cells. A short form (93 kDa), used as the template for our AAK1 constructs and a second longer form (AAK1L, ~140 kDa), that differs only by an extended C-terminus containing additional clathrin binding motifs (Henderson & Conner, 2007). Both variants have been reported to have similar function and activity (Henderson & Conner, 2007). In HeLa cells only the AAK1L variant was detectable suggesting this is likely the predominant variant found in these cells. Therefore, the shAAK1 depletes both transfected and endogenous AAK1 proteins equally well.

Figure 4.18. HeLa cells expressing shAAK1 knockdown. (A) Stable HeLa cell line expressing shAAK1 or scrambled control (scram) were transiently transfected with wild type AAK1 then subjected to Western blotting for FLAG tag and GAPDH as a loading control. Relative abundance of AAK1 was quantitated as the ratio between FLAG and GAPDH and is presented as a graph. (B) Lysates from HeLa cells stably expressing shAAK1 or scrambled control (scram) were subjected to Western blotting for endogenous AAK1L and GAPDH as a loading control. Relative

141 |Page

Sec. 4.2. Results

abundance of AAK1L was quantitated as the ratio between AAK1L and GAPDH and is presented as a graph. (Mean ± S.E.M, ** = p<0.01; *** =p<0.005; T-test).

4.2.5.2 AAK1 depletion reduces LC3-II staining in HeLa cells

Autophagy is an essential process for neuronal cell homeostasis, and its dysfunction has been directly linking to neurodegenerative diseases, such as AD and PD (see review: (Wong & Cuervo, 2010)). To investigate a role for AAK1 in autophagy, endogenous AAK1 was knocked down using an shRNA construct in HeLa cells and subjected to 24 h serum starvation to induce autophagy. Cells were fixed and stained for the autophagy marker, microtubule-associated protein light chain 3-II (LC3-II), and subjected to immunofluorescence microscopy analysis. HeLa cells transfected with the AAK1-shRNA construct had reduced LC3-II staining compared with cell transfected with scrambled control (Fig.4.19A). This was rescued back to control levels by co-expression of an shRNA-resistant form of AAK1. Validation of the knockdown and shRNA-resistant constructs is shown in Fig.4.19B. This data appears to indicate that AAK1-depletion disrupts and reduces autophagy in these cells. Furthermore, we observed a similar decrease in LC3-II staining in stable shAAK1 cells under basal conditions (Fig.4.20A), which was quantitated via Western blotting using an antibody recognising LC3-II (Fig.4.20B).

4.2.5.3 AAK1 depletion increases autophagy flux

LC3-II itself is degraded by autophagy, and therefore the amount of LC3-II at a certain time point alone is not an accurate indication of autophagic flux. Therefore, it is important to measure the amount of LC3-II delivered to lysosomes by comparing LC3-II levels in the presence/absence of lysosomal inhibitors such as Bafilomycin A. To investigate autophagy flux, shAAK1 and scrambled control cells were serum starved for up to 6 h, to induce autophagy, and treated with/without Bafilomycin A to inhibit LC3-II degradation. Lysates were then blotted for LC3-II and GAPDH as a loading control. Interestingly, shAAK1 HeLa cells had a significant increase in the accumulation of LC3- II protein when treated with Bafilomycin A compared to scrambled control cell and this observation was further exacerbated overtime. This is indicative of increased autophagy flux (shown in Fig.4.20C and presented as a graph in D). Therefore, it’s likely the decreased levels of LC3-II staining in cells depleted of AAK1 (Fig.4.19A and Fig.4.20A)

142 |Page

Sec. 4.2. Results

is the result of increased autophagic flux as LC3-II is constitutively degraded inside autolysosomes. This indicates that AAK1 antagonises autophagic flux.

Cnimto o efcie ncdw o AK by AAK1 of knockdown effective of Confirmation ) B

Serum starvation of AAK1-knockdown HeLa cells reduces LC3 staining.

) HeLa cells transfected with scrambled, AAK1-shRNA, or AAK1-shRNA plus AAK1-shRNA or AAK1-shRNA, scrambled, with transfected cells HeLa ) A Figure 4.19. ( AAK1 constructs (green) and subjected to 24 h serum starvation. Yellow arrows point ( (red). staining LC3II to or scrambled with cells HEK293 into co-transfected was AAK1 type Wild shRNA. AAK1-shRNA. Lysates were subjected to blotting Western FLAG for and GAPDH.

143 |Page

Sec. 4.2. Results

Figure 4.20. HeLa cells stably expressing AAK1-shRNA knockdown have increased autophagy flux. (A) HeLa cells stably expressing shAAK1 or scrambled shRNA (scram) under basal conditions and stained for LC3 (red), adaptin β (green) and nuclei using DAPI (blue). (B) Western blots of scram or shAAK1 HeLa cells under basal condition blotted for LC3-II and GAPDH as a loading control. Relative abundance of LC3-II was quantitated as the ratio between LC3-II and GAPDH and is presented as a graph (Mean ± S.E.M, *** =p<0.005; T-test). (C) shAAK1 or scram control HeLa cells were treated with/without 50 nM Bafilomycin A for the times indicated. (D) The ratio of LC3-II to GAPDH after Bafilomycin A treatment is normalised to untreated and presented as a graph (* = p<0.05, Two-way ANOVA).

144 |Page

Sec. 4.2. Results

4.2.6 A role of AAK1 in PD pathogenesis

4.2.6.1 Depletion of AAK1 produces PD phenotype in flies

AAK1 has previously been identified as potential risk gene in the age of onset of PD (Latourelle et al., 2009). Therefore, we wanted to determine if AAK1 is associated with symptoms of PD using a Drosophila model of PD. PD is caused by degeneration of the midbrain dopaminergic system of the brain. PTEN-induced putative kinase 1 (PINK1), encodes a protein kinase identified to be responsible for an autosomal recessive form of PD (Silvestri et al., 2005; Valente et al., 2004). Loss of PINK1 (RNAi knockdown) in Drosophila results in degeneration of mitochondria and is successfully used to mimic the degenerative processes seen in PD (Clark et al., 2006; Park et al., 2006; Todd & Staveley, 2008; Yang et al., 2006). To investigate a potential role of AAK1 in neural degeneration, its expression was specifically knocked down in the Drosophila eye (GMR-Gal4 driver), a tissue rich in neurons. GMR-Gal4 is a well describe pan-neuronal promoter used in Drosophila to selectively overexpress/knockdown expression of proteins in photoreceptor neurons of the Drosophila compound eye. The compound eye of the Drosophila melanogaster is composed of 800 unit eyes called ommatidia, which are composed of eight neural photoreceptor cells and 12 supporting cells. All these cells are generated from an epithelial sheet called the eye imaginal disc (Şahin & Çelik, 2013). The GMR promoter drives expression specifically in the cells posterior to the morphogenetic furrow in the eye discs (Freeman, 1996). Therefore, expression occurs in all photoreceptor cells and supporting cells of the ommatidia.” Flies with simultaneous RNAi knockdown of PINK1 and AAK1 in neurons of the eye had increased degeneration compared to knockdown of PINK1 or AAK1 alone (5 d old, 25°C). This is evident by increased discolouration, disorganisation and fusion of ommatidea (represents degeneration of eye neurons) (Fig.4.21A; represented by arrows). In a separate experiment, simultaneous knockdown of PINK1 and AAK1 expression in dopaminergic neurons (ddc-Gal4 driver) was associated with an increased occurrence of flies exhibiting a drooped wing phenotype, a characteristic of dopaminergic degeneration and considered a PD phenotype in flies (Cha et al., 2005; Greene et al., 2003; Pesah et al., 2004) (Fig.4.21B). Together, these results demonstrate that down-regulation of AAK1 increases susceptibility to PD pathogenesis in Drosophila and provides the first functional evidence of an association between AAK1 and PD. However, this is a preliminary investigation and further work is required to validate this observation.

145 |Page

Sec. 4.2. Results

was

ddc-Gal4)

) Simultaneous RNAiSimultaneous ) A (

<0.05; T-test). p

( neurons dopaminergic in AAK1 and PINK1 of knockdown RNAi Simultaneous ) B Simultaneous RNAi-mediated knockdown of PINK1 and AAK1 in neurons exacerbates PD phenotypes in Drosophila. in phenotypes PD exacerbates neurons in AAK1 and PINK1 of knockdownRNAi-mediated Simultaneous Figure 4.21. Figure knockdown of PINK1had neurons AAK1 eye and increasedcompared degeneration the in to knockdown of dPINK1 orArrows (5 25°C). AAK1 old, alone point to areas of increased discolouration, disorganisation and fusion of ommatidea (represents degeneration of eye neurons). Experiment performed by Carla ( Institute. Garvan group, genomic Functional Gentile, dopaminergic of degeneration (represents alone PINK1 of knockdown with compared defects wing with flies of occurrence increased an with associated neurons; d,4-5 25°C). Presented as a graph (Mean S.E.M, ± * =

146 |Page

Sec. 4.2. Results

4.2.6.2 Model of GSK3-mediated phosphorylation of AAK1 increasing autophagy and PD pathogenesis

Dysfunction in autophagy is becoming increasingly implicated in the pathogenesis of PD, whereby excessive autophagy can result in cell death and neurodegeneration (Cheung & Ip, 2009; Xiong et al., 2013). The plasma membrane and activation of clathrin-mediated endocytosis are contributors to the formation of pre- autophagosomal structures (Ravikumar et al., 2010). AP-2 has an important role in the production of CCV’s and autophagosomes (Motley et al., 2006; Olusanya et al., 2001; Ricotta et al., 2002; Tian et al., 2013; Tian et al., 2014). Therefore the abundance of AP- 2 at the plasma membrane is likely a contributing factor to the formation of autophagosomes in cells. Overall, we propose a model whereby GSK3-mediated phosphorylation of AAK1 in PD promotes release of AP-2 for subsequent trafficking to the cell surface. This increases AP-2 abundance at the plasma membrane and induces overproduction of autophagosomes. In times of cellular stress, such as those experienced in PD, autophagy is activated. But in the AAK1-depleted cells, with increased autophagic flux, it is exhausted too quickly, resulting in premature cell death (Fig.4.22).

147 |Page

Sec. 4.2. Results

Figure 4.22. Model of GSK3-mediated phosphorylation of AAK1 in autophagosome production and PD pathogenesis. Phosphorylation of AAK1 by GSK3 leads to release of AP-2 for subsequent trafficking back to the cell surface, promoting autophagosome production. Excessive autophagocytosis results in cell death and neurodegeneration of dopaminergic neurons, enhancing PD pathogenesis.

148 |Page

Sec. 4.3. Discussion

4.3 Discussion

AAK1 has been identified as a regulator of clathrin-mediated endocytosis and trafficking of CCVs (Conner & Schmid, 2002; Jackson et al., 2003). However, AAK1 is relatively understudied and its precise role in these and other processes required further investigation. So far, it has been reported to phosphorylate the µ2 subunit of AP-2 (Ricotta et al., 2002), which enhances AP-2 binding to surface receptors and recruitment to internalisation machinery (Motley et al., 2006; Olusanya et al., 2001; Ricotta et al., 2002) (e.g. Numb; (Sorensen & Conner, 2008)). Cargo proteins downstream of AAK1 and AP- 2, include Notch and ErbB4. These pathways are important during neurodevelopment and for synaptic plasticity of the brain throughout life. Since GSK3 also has been identified as an important regulator of Notch signalling (Espinosa et al., 2003; Guha et al., 2011) and its activity can be attenuated by ErbB/neuregulin signalling (Beaulieu et al., 2005), we hypothesised that GSK3 phosphorylation of AAK1 may be important for receptor transport and neurotransmission. Interestingly, neuregulin-1 and ErbB4 signalling have been identified as genetic risk factors for BD and schizophrenia (Goes et al., 2011; Mei & Xiong, 2008; Pan et al., 2011; Yokley et al., 2012). Together, this suggests that AAK1 may have an important role for regulating synaptic plasticity downstream of GSK3, dysregulation of which, may lead to pathogenic processes and the development of mood disorders.

In this study, we identified the trafficking protein AAK1 as a novel substrate of GSK3. GSK3 phosphorylates AAK1 at two sites (Thr620 and Thr674), following priming phosphorylation at Ser624 and Ser678. The pThr620 was relatively resistant to de- phosphorylation both in vitro and in cells (Fig.4.6A-D), consistent with phosphorylation sites on PI4KIIα (section 3.2.1.1), CRMP2 (Cole et al., 2008), and β-adducin (Farghaian et al., 2011). This, at least in part, is due to the presence of proline residues C-terminal to the phosphorylation site, which are known to confer phosphatases resistance, as well as basic residues that may act to stabilise the negatively charged phospho-group (Cole et al., 2008). However, the reason why these sites are relatively resistant to phosphatases is not yet clear. It might demarcate two separate pools of substrates to perform separate functions. Alternatively, it might permanently target these proteins for degradation by the lysosome. That is, protect AAK1 from degradation, but when dephosphorylated, AAK1 is trafficked to the lysosome to be degraded, as evidenced by the low abundance and stability of the AAK1-S624A mutant (Fig.4.7A, B). The opposite was demonstrated in

149 |Page

Sec. 4.3. Discussion

Chapter 3 for PI4KIIα. Our lab previously demonstrated that phosphatase-resistant sites on CRMP2 and Tau are highly susceptible to hyperphosphorylation in AD (Cole et al., 2007). Net phosphorylation is always a balance between the activity of the kinase and the phosphatase. However, in these cases, the substrates are relatively resistant to phosphatases, implying that theoretically small changes in kinase activity could lead to large changes in the phosphorylation levels of substrates (i.e. hyperphosphorylation). Therefore, these sites may be vulnerable to dysregulation in neurodegenerative diseases, possibly rendering them as potential ‘weak spots’ that promote pathogenesis. Given our recent discovery that AAK1 is functionally involved in autophagy and possibly the pathogenesis of PD, it will be interesting to see if AAK1, and especially the Thr620/Ser624 site, is hyperphosphorylated in the brains of PD patients.

AAK1 associates with adaptin α and β of the AP-2 complex and localises to clathrin-rich regions in cells (Fig.4.8A, B). These interactions are important for AP-2 recruitment and to stimulate AAK1’s phosphorylation of the μ2 subunit, implicating it as an important regulator of early endosomal trafficking in cells. AAK1 regulates the trafficking of several cargo-proteins, including Tfn (Fig.4.12), LDRP (low-density lipoprotein receptor-related protein) (Conner & Schmid, 2003) and ErbB4 (Kuai et al., 2011). Inhibition of phosphorylation at Thr620/Ser624 significantly slowed the rate of Tfn (ligand) recycling in HeLa cells (Fig.4.12A and C) and may suggest that Tfn/TfnR is redirected towards degradative pathways instead of the cell surface, consistent with the precedent set for phosphorylation of PI4KIIα (Chapter 3). Indeed, overexpression AAK1- S624A decreased the abundance of TfnR in cells compared to wild type AAK1 (Fig.4.12E). Interestingly, increased recycling of Tfn was dependent on AAK1’s dissociation from the AP-2 complex since HeLa cells expressing AAK1-F698A (reducing AAK1 interaction with AP-2) or RNAi-mediated depletion of AAK1 had a significant increase in Tfn recycling back to the cell surface. This is consistent with previous research, whereby ErbB4 accumulated at the plasma membrane in cells depleted of AAK1 (Kuai et al., 2011). Indeed, phosphorylation by GSK3 induces AAK1’s dissociation from AP-2, promoting Tfn recycling back to the cell surface (Fig.4.9 and Fig.4.12A, C). Also, the decrease in abundance of dephospho-forms of AAK1 seems to be proportional to the amount these dephospho mutants bind to AP-2 (Fig.4.7A and Fig.4.9B, C). Together, this suggests these functions are linked. That is, phosphorylation of AAK1 by GSK3 decreases binding to AP-2 for increased trafficking to recycling endosomes and possibly

150 |Page

Sec. 4.3. Discussion

back to the cell surface. Instead, it is probably trafficked (via an unknown mechanism) to the degradation pathway for destruction by the lysosome, consistent with decreased abundance of AAK1-S624A (and to a lesser extent S678A) compared to wild type, as well as decreased abundance of TfnR.

Phosphorylation of AAK1 at Thr620 by GSK3 promotes phosphorylation of the AP-2 subunit adaptin µ (Fig.4.10C). Since AAK1 phosphorylation of AP-2 promotes recruitment of cargo proteins into CCVs, we expected that dephosphorylation of AAK1 would disrupt CCV assembly and internalisation of Tfn. However, expression of wild type or phosphomutant forms of AAK1 did not significantly affect the rate of Tfn internalisation (Fig.4.12B). Instead, expression of AAK1-S624A dramatically reduced recycling of Tfn back to the cell surface. Consistent with this, phosphorylation of µ2, although important for efficient receptor uptake, is not absolutely required (Motley et al., 2006). This suggests that AAK1 phosphorylation at Thr620 may be more pertinent during AP-2 sorting/recycling processes, rather than recruitment of cargoes, such as Tfn, into CCVs. Similarly, it has been shown that overexpression of kinase-inactive AAK1 (K74A) or silencing AAK1 using siRNA do not significant effect µ2 phosphorylation in cells (Conner & Schmid, 2003). This may reflect functional redundancies with other kinases of the Ark1/Prk1 family. Indeed, GAK (Ark1/Prk1 family member) is known to also be associated with CCVs and to phosphorylate µ2 in vitro (Korolchuk & Banting, 2002; Lee et al., 2005). Alternatively, it is possible that phosphorylation of AAK1 is required for higher order functions that influence the efficiency and/or fidelity of receptor endocytosis. Overall, GSK3-mediated phosphorylation of AAK1 at Thr620 is important for efficient Tfn recycling, however is not essential for AP-2-dependent Tfn internalisation. Therefore, we propose that AAK1 is required for endosomal sorting functions, directing the transport of cargoes away from the cell surface/recycling pathways and instead possibly towards degradative pathways. Phosphorylation by GSK3 regulates this process by inducing AAK1’s dissociation from AP-2, leading to the redirection of cargoes (Tfn) towards recycling/secretory pathways.

Since AAK1 is important for endosomal sorting and has previously been implicated in ErbB4/neuregulin signalling, which is genetically associated with BD and schizophrenia, we hypothesised that AAK1 may be another link between GSK3 and mood disorders (similar to PI4KIIα in Chapter 3). However, unlike the hyperactivity observed in PI4KIIα depleted flies, the neural-specific knockdown of AAK1 was not associated

151 |Page

Sec. 4.3. Discussion

with changes in activity level, circadian rhythm or reward seeking behaviour (Fig.4.16A, B), suggesting that AAK1 might not be associated with behaviours related to BD. Interestingly, neural-specific depletion of AAK1 in flies increased susceptibility to death by starvation, whereas survival was unchanged when exposed to nicotine excitotoxicity. Given the similarities between CCV trafficking and autophagy, we speculate that AAK1 might be involved in starvation-induced autophagy. Coincidently, AAK1 has been identified as a potential risk factor for age of onset of PD; a neurodegenerative condition involving impaired autophagy (Latourelle et al., 2009). AAK1 protein levels have also been reported to be decreased in ALS patients (Shi et al., 2014). Therefore, we explored the possibility that AAK1 might be involved in regulating autophagy in PD. Indeed, flies with simultaneous knockdown of AAK1 and PINK1 (an established PD-related gene in humans) had increased degeneration of neurons in the fly eye as well as increased dopaminergic neuron degeneration, evident by the drooped wing phenotype (Fig.4.21A, B). Therefore, down-regulation of AAK1 increases susceptibility to PD pathogenesis in Drosophila and provides the first functional evidence of an association between AAK1 and PD.

It is expected that autophagy and endocytosis partly share the same machinery since these pathways are both implicated in the degradation of cellular materials (see review: (Lamb et al., 2013)). Autophagy is a sequential and complex process through which eukaryotic cells degrade and recycle long-lived proteins, misfolded proteins and impaired cytoplasmic organelles (Klionsky & Emr, 2000). Although autophagy is typically associated with neuroprotective function in the brain, excessive autophagy can result in cell death and neurodegeneration and is becoming increasingly implicated in the pathogenesis of PD (Cheung & Ip, 2009; Xiong et al., 2013). Interestingly, we observed an extensive reduction of the autophagy marker, LC3-II, in HeLa cells depleted of AAK1 (Fig.4.19A and Fig.4.20A, B). Also, LC3-II could be rescued back to control levels by co-expressing shRNA-resistant AAK1 (Fig.4.19A), demonstrating an AAK1 specific function. This suggests either a down-regulation of autophagosome formation, or an increase in the clearance of autophagic compartments in these cells, since LC3 located in the inner membrane of autophagosomes is also degraded when these compartments fuse to lysosome. To accurately measure autophagy flux, we measured the amount of LC3-II delivery to lysosome by comparing LC3-II levels in the presence/absence of lysosomal inhibition. Interestingly, cells depleted of AAK1 have increased autophagic flux

152 |Page

Sec. 4.3. Discussion

(Fig.4.20C, D), consistent with reduced staining for LC3-II in these cells as it is constitutively degraded inside autolysosomes. Our results are similar to recent studies investigating familial PD, whereby the leucine-rich repeat kinase 2 (LRRK2) mutation, G2019S (the most common genetic cause of familial PD) has been linked to excessive autophagy clearance in G2019S-expressing cells and dopaminergic neurons derived from PD patient fibroblasts carrying this mutation (Bravo-San Pedro et al., 2013; Ramonet et al., 2011; Su & Qi, 2013). Furthermore, these studies also observed a concomitant occurrence of mitochondrial dysfunction and fragmentation and suggest a link between mitochondrial fission and autophagy in PD. Since mitochondrial fission is a prerequisite for autophagy (Twig et al., 2008a; Twig et al., 2008b), it would be interesting to investigate mitochondrial function in cells depleted of AAK1 or expressing phosphomutant forms of AAK1. Overall, these observations indicate that AAK1 antagonises autophagy and reduces autophagy flux. AAK1 depletion in cells and flies increases susceptibility to starvation and neural degeneration, consistent with excessive autophagy in PD pathogenesis.

Since polymorphisms in GSK3 are associated with PD susceptibility and that GSK3 inhibitors have been shown to offer neuroprotection against PD (see reviews: (Morales-García et al., 2013; Yuan et al., 2013)), it’s likely that GSK3 activity and its phosphorylation of downstream targets, such as AAK1, is a means by which it regulates trafficking pathways that controls autophagy. Our results suggest that the mode of action of GSK3 phosphorylation of trafficking proteins may be to either enhance (PI4KIIα, chapter 3) or disrupt (AAK1) associations with adaptor protein complexes, which are key regulators of protein sorting functions in cells. The abundance of AP-2 at the cell surface is likely important in the formation of autophagosomes, since the activation of clathrin- mediated endocytosis and internalisation of CCVs from the plasma membrane contribute to the formation of pre-autophagosomal structures (Ravikumar et al., 2010). Given that AAK1 dissociation from the AP-2 complex increases recycling and that phosphorylation of AAK1 at Thr620/Thr674 is required for this, it is likely to increase AP-2 abundance at the plasma membrane and hence promote autophagic processes. Overall, the disruption of normal sorting in cells, particular in neurons, is likely to affect autophagy processes that could lead to PD pathogenicity. Further research is required to validate the role of phosphorylation of AAK1 and increased autophagy flux in neurodegenerative disorders,

153 |Page

Sec. 4.3. Discussion

and therefore it would be interesting to measure the phosphorylation levels of AAK1 in the brains of PD and AD patients.

In summary, AAK1 is a novel substrate of GSK3 that regulates AP-2/cargo protein trafficking and is important for antagonising autophagic processes in cells. Here we demonstrate an important role of GSK3-mediated phosphorylation of AAK1 to promote phosphorylation and release of AP-2 for subsequent trafficking to sorting/recycling endosomes in cells. Meanwhile, depletion of AAK1 was associated with a constitutive increase in autophagy flux, implicating AAK1 as a potential target in autophagic-related PD pathogenesis. Accordingly, AAK1 deregulation in Drosophila is associated with decreased survival during starvation and neurodegenerative phenotypes related to PD. Together, this is consistent with abnormal autophagy observed in neurodegenerative diseases and PD pathogenesis.

154 |Page

CHAPTER 5: Conclusions, perspectives and future directions

5.1 Project outcomes

GSK3 is a Ser/Thr kinase that is highly expressed in the brain and at synapses and deregulated in psychiatric and neurodegenerative diseases, including MDD, BD, schizophrenia and AD (see reviews: (Jope & Roh, 2006; Salcedo-Tello et al., 2011)). These debilitating diseases severely impair people’s lives and in severe cases lead to exclusion from society and suicide. GSK3 expression is highest in regions of the brain that demonstrate enhanced neuroplasticity, such as the hippocampus and cortex (Leroy & Brion, 1999b; Takahashi et al., 1994; Yao et al., 2002), where it is an important regulator of neuroplasticity and neurotransmission (see reviews: (Cole, 2012, 2013a; Peineau et al., 2008)). These functions are mediated by phosphorylation of a wide range of signalling proteins, transcription factors and cytoskeletal proteins (Cole, 2012; Medina & Wandosell, 2011; Sutherland, 2011). Also, deregulation of signalling pathways upstream of GSK3 have been implicated in the development of these neurological diseases, including the Wnt, Notch, Hedgehog and growth factor pathways (see review: (Grimes & Jope, 2001b; Jope & Roh, 2006)). Therefore, GSK3 is an important enzyme to study in order to fully understand basic signalling mechanisms controlling neurotransmission and neuroplasticity in healthy and diseased brains.

At present, the pathogenic targets downstream of GSK3 in mood disorders and neurodegenerative diseases are not yet known. Antidepressants and mood stabilizers (including lithium) target GSK3 in the brain (Beaulieu, 2007; Beaulieu et al., 2009; Chalecka-Franaszek & Chuang, 1999; Kalinichev & Dawson, 2011) and are especially effective at reducing suicidal tendencies. Also, GSK3 inhibitors are neuroprotective (Beaulieu, 2007; Chuang, 2004; Chuang et al., 2011; Leeds et al., 2014; Nonaka & Chuang, 1998; Nonaka et al., 1998; Yazlovitskaya et al., 2006; Yu et al., 2012) and may act as a therapeutic target for treatment of AD and PD, as well as stroke (Alvarez et al., 2002; Hong et al., 1997; Leyhe et al., 2009; Morales-García et al., 2013; Nakashima et al., 2005; Noble et al., 2005; Nunes et al., 2007; Pérez et al., 2003). However, treatment with GSK3 inhibitors has limiting side effects. For example, lithium induces weight gain, which reduces compliance in many patients. Therefore, identifying pathogenic substrates downstream of GSK3 and targeting them therapeutically could improve the specificity of

155 |Page

Sec. 5.1. Project outcomes

treatment and reduce side-effects, thus improving compliance and better patient outcomes.

In order to identify as many new substrates of GSK3 as possible, our group used a combination of bioinformatics, biochemical analysis and mass spectrometry-based phosphoproteomics. This revealed a surprising enrichment of substrates involved in vesicular trafficking events. Given the importance of trafficking in synaptic transmission and also the role of GSK3 in regulating this process, this suggested that trafficking- associated substrates of GSK3 may be particularly vulnerable to change in mood and/or neurodegenerative diseases. In this project, I focused on two of the most promising trafficking substrates; the lipid kinase PI4KII and the Ser/Thr protein kinase AAK1. PI4KIIα and AAK1 are both highly expressed in neurons and at synapses where they have been implicated in vesicle formation, transport/delivery and sorting functions that are important for efficient neurotransmission. My work here describes the mechanisms by which they are linked to BD and PD, respectively.

5.1.1 PI4KIIα regulates AP-3-mediated vesicular trafficking in BD

Our work has established PI4KIIα as a regulator of synaptic vesicle trafficking in the brain and demonstrated physiological behaviours associated with its dysregulation and BD pathogenicity. PI4KIIα is enriched at synapses, where it accounts for the bulk of PI4K activity in the brain (Guo et al., 2003). It has established roles in vesicle trafficking between Golgi apparatus, endosomes and lysosomes (Craige et al., 2008; Guo et al., 2003; Salazar et al., 2005), at least partly via its interaction with the AP-3 adaptor protein complex. GSK3 phosphorylates PI4KII at two sites in the N-terminal region (Ser5 and Ser47), which promotes binding to the AP-3 complex and transport from early/sorting endosomes to late endosomes/lysosomes to be degraded. This also applies to cargo proteins of PI4KIIα and is the basis upon which GSK3-mediated phosphorylation of PI4KII regulates expression of the AMPA receptor GluA1 at the surface of neurons. Since induction of LTP is effected by the number of AMPA receptors at the post-synaptic surface (Hanley, 2010b; Kerchner & Nicoll, 2008; Kessels & Malinow, 2009), it is likely that PI4KII affects synaptic transmission. This could be investigated by future electrophysiological experiments, preferably using neurons from PI4KIIα-knockout mice.

156 |Page

Sec. 5.1. Project outcomes

At a behavioural level, deregulation of PI4KII in neurons of Drosophila increased their locomotor activity, consistent with hyperactivity exhibited by human BD patients in their manic phase. Moreover, this was corrected with lithium treatment. Together, this establishes GSK3-mediated phosphorylation of PI4KII as a regulator of AMPA receptors, vesicle trafficking and behaviours associated with BD.

5.1.2 AAK1 regulates AP-2-mediated autophagosome formation in PD

My second area of focus was the Ser/Thr kinase AAK1. This protein has been implicated in clathrin-mediated endocytosis and coordinating trafficking of CCV’s (Conner & Schmid, 2002; Jackson et al., 2003). However, its precise role and mechanism underlying this are poorly understood. Here, I established a new role for AAK1 in regulating autophagy in cells and presented new evidence supporting its deregulation that contributes to neurodegeneration and potentially PD pathogenicity. It is highly expressed in the brain and at synapses of neurons and has been shown to inhibit neuregulin/ErbB4 signalling as well as activating Notch signalling during clathrin-mediated endocytosis (Gupta-Rossi et al., 2011). These pathways are important during neurodevelopment and for synaptic plasticity. Also, AAK1 is a genetic risk factor for age of onset of PD, and has been reported to have decreased protein levels in ALS (Latourelle et al., 2009; Shi et al., 2014), supporting its role in neurodegenerative pathogenesis. Here, I demonstrate that GSK3-mediated phosphorylation of AAK1 at Thr620 and Thr674 promotes dissociation of AAK1/AP-2 complex for subsequent trafficking to sorting/recycling endosomes in cells and back towards the cell surface. Surprisingly, depletion of AAK1 increases autophagy flux, implicating it as a potential target in autophagy-related PD. Accordingly, AAK1 deregulation in Drosophila is associated with decreased survival during starvation and increased neurodegenerative phenotypes related to PD. Together, this work establishes AAK1 as a key target of GSK3 for regulating recycling/trafficking pathways that are important in cellular autophagy where dysregulation is consistent with abnormal autophagy processes observed in PD pathogenesis.

157 |Page

Sec. 5.2. Research implications

5.2 Research implications

Mood disorders and neurodegenerative diseases are debilitating illnesses that affect patients, family, relationships and employment and are a major cause of disability and mortality in society, costing the Australian health system billions annually in direct care costs and loss of productivity. Incidences of neurodegenerative diseases are increasing due to an ageing population, while psychiatric illnesses, such a mood disorders, are often poorly managed using current therapies. These complex diseases can be difficult to diagnose, with no earlier genetic markers or biomarkers available, resulting in delayed diagnosis and treatment or inappropriate treatment that can negatively impact the course and severity of the disease. The aim of this work was to further the knowledge of GSK3 function in the brain and to identify links between GSK3, synaptic transmission, and impaired brain function. This work hopes to help delineate the pathology of psychiatric and neurodegenerative diseases and lay the early foundations for development of earlier diagnostic markers and/or more effective treatment options for patients.

5.2.1 GSK3’s role in neurotransmission and the implication for BD pathogenicity

Our research has identified a novel subset of trafficking proteins downstream of GSK3 that we believe are important for neurotransmission and synaptic plasticity. Trafficking is especially critical in the brain, since it is essential for efficient synaptic transmission. Neurotransmission (e.g. LTP, LTD) is regulated in vivo by growth factors, such as BDNF and IGF-1 that are physiological inhibitors of GSK3 activity. Therefore, by elucidating the substrates of GSK3 associated with vesicular trafficking, this will help to delineate the mechanisms linking growth factors to neurotransmission. Both PI4KIIα and AAK1 have functional roles in vesicle formation, transport/delivery and sorting functions that are important for efficient neurotransmission of the brain. For the first time we have demonstrated that GSK3-mediated phosphorylation of these targets regulate their binding to adaptor complexes (AP-3 for PI4KIIα; AP-2 for AAK1). The binding of PI4KIIα and AAK1 to these adaptor complexes was found to be important for trafficking cargoes away from the cell surface and towards degradative pathways. However, when dissociated from their respective complex, this favoured trafficking back to the cell surface. This may represent a mechanism by which GSK3 phosphorylation of trafficking proteins regulates the transport of neurotransmitters in neurons, their expression and hence synaptic plasticity in the brain.

158 |Page

Sec. 5.2. Research implications

Indeed, abnormalities in multiple neurotransmitter systems, such as monoaminergic (serotonin, dopamine and norepinephrine) and GABAergic systems (see reviews: (Manji et al., 2003; Shi et al., 2008)), as well as physiological processes have been found in complex disorders such as BD. For example, patients with mood disorders show abnormalities of morphology or morphometry in many visceromotor network structures of the brain, in particular reduced grey matter volumes in the medial prefrontal cortex and other closely related regions (Botteron et al., 2002; Coryell et al., 2005; Drevets et al., 1997; Lyoo et al., 2004; Nugent et al., 2006; Ongur et al., 2003). Signal transduction is critical in the central nervous system for balanced transmission in the brain and therefore undoubtedly important in regulating diverse functions such as mood, appetite, and wakefulness; abnormalities of which are intertwined in the pathophysiology of BD. Accordingly, the molecular targets underlying lithium’s action to stabilise mood states is strongly associated with abnormal signalling pathways in the brain. For example, lithium’s action to directly/indirect inhibit GSK3 are associated with decreased DA signalling, elevation of serotonin signalling (Beaulieu et al., 2009; Beaulieu et al., 2005; Beaulieu et al., 2004; Beaulieu et al., 2008), as well as increasing brain BDNF levels (Fukumoto et al., 2001; Leyhe et al., 2009). Our work has identified PI4KIIα as a key target of lithium and GSK3 that contributes to signalling in the brain, dysfunction of which likely contributes to BD pathogenicity. Both the knockdown of PI4KIIα and the expression of dephosphorylated mutant PI4KIIα promoted the trafficking of the AMPA receptor GluA1 towards the cell surface in hippocampal neurons, favouring LTP. It is possible that increased expression of AMPA receptors at the cell surface may underlie hyperactivity observed in our PI4KIIα-depleted flies, since elevated GluA1 expression has previously been associated with enhanced synaptic activity, wakefulness (versus sleep; (Vyazovskiy et al., 2008)) and locomotor hyperactivity (Yamamoto & Zahniser, 2012; Zhang et al., 2008). Interestingly, the antidepressant imipramine, which can provoke mania, increased synaptic expression of GluA1 in the hippocampus of rat brains (Du et al., 2004a). In contrast, GluA1 knockout mice demonstrate mania-related phenotypes, such as hyperactivity, anxiety and increased novelty-seeking or risk-taking behaviours (Fitzgerald et al., 2010), further demonstrating a link between GluA1 expression and BD. Therefore, overactivity of GSK3 in BD may promote trafficking of PI4KIIα towards the lysosome for degradation, reducing its levels at the synaptic membrane. This may contribute to the manic phase of BD through increased abundance of AMPA receptors at the synapse and promoting LTP. In addition, other neurotransmitter

159 |Page

Sec. 5.2. Research implications

receptors have been implicated in BD. For example, altered sensitivity of adrenergic receptors, α2 and β2 have been implicated in the pathophysiology and treatment of mood disorders (see reviews: (Dremencov et al., 2009; Manji et al., 2003)) and a reduction in

5-HT1A and 5-HT2A receptors have been observed in brains of BD patients (Drevets et al., 1999; Drevets et al., 2007; López-Figueroa et al., 2004; Sargent et al., 2000). Furthermore, several cell surface receptors are genetically associated with mood disorders, such as CACNA1c (CaV1.2, a voltage-gated calcium channel) (Ferreira et al.,

2008), HCN4 (Kelmendi et al., 2011), and GABAA receptors (Craddock et al., 2008). Therefore, future studies could investigate if GSK3 and PI4KIIα also regulate the cell surface expression of these receptors as well.

It is widely believed that the balance of LTP/LTD in the brain plays an important role in regulating mood. The overactivity of GSK3 has been widely implicated in mood disorders and may affect the fine balance of LTP/LTD through the interplay between NMDA and AMPA receptors expression and activation. Indeed, mood stabilising drugs and inhibitors of GSK3, such as lithium, modulate synaptic plasticity by regulating AMPA-receptor trafficking (see review: (Du et al., 2004b)) as well as stabilising synaptic protein levels in neurons, such as BDNF, PSD-95, neuroligin 1, β-neurexin and synaptophysin (Park et al., 2014). Moreover, GSK3 has been shown to phosphorylate PSD-95 (Chen et al., 2007; Nelson et al., 2013) and dynamin-1 (Clayton et al., 2010), important for receptor trafficking and reuptake in neurons. In addition, GSK3 has been shown to complex with and phosphorylate 5-HT1B receptor, enhancing its trafficking to the cell surface and promoting receptor activation (Chen et al., 2009; Chen et al., 2011b). Therefore, GSK3 undoubtedly regulates multiple downstream targets that affect synaptic transmission. Dysregulation of this fine balance of transmission may be the basis of both depressive and manic phases of BD. It will be interesting to investigate whether GSK3 is important in regulating the trafficking of other cargoes associated with BD (see review: (Cole, 2013a)). For example, does PI4KIIα also regulate the trafficking of cargoes, such as the NMDA receptor, GABAA receptors, CaV1.2, Nrg1 or its receptor ErbB4, BDNF or its receptor TrkB or monoamine receptors? Alternatively, GSK3 may regulate other trafficking protein responsible for transport of these cargoes. Indeed, AAK1 has previously been shown to negatively regulate neuregulin-ErbB4 trafficking, likely as a result of increased internalisation of ErbB4 or decreased trafficking of ErbB4-containing vesicles back to the plasma membrane (Kuai et al., 2011). This is consistent with our

160 |Page

Sec. 5.2. Research implications

work here, whereby phosphorylation of AAK1 by GSK3 decreases binding to AP-2 for increased trafficking to recycling endosomes and back to the cell surface. Therefore, it is worth further investigating the role of AAK1 phosphorylation in regulating ErbB4 surface expression. Together, our research provides further evidence of the importance of GSK3 in the brain and the subsequent effect it has on key downstream substrates involved in vesicle trafficking, receptor transport and neurotransmission, dysregulation of which may contribute to the development of mood disorders, such as BD.

5.2.2 GSK3’s role in autophagy and the implication for PD pathogenicity

GSK3 dysregulation is also implicated in the pathogenesis of neurodegenerative diseases, particular AD and PD (Engel et al., 2006; Hernandez et al., 2013; Lei et al., 2011; Llorens-Marítin et al., 2014; Morales-García et al., 2013; Yuan et al., 2013). Although the etiology of these neurodegenerative diseases remains uncertain, abnormal or misfolded proteins, when aggregated in cytoplasmic, nuclear and extracellular inclusions, cause organelle damage and synaptic dysfunction in the nervous system and play a central role in neurodegenerative disorders, such as AD, PD, Huntington’s disease (HD) and ALS (see reviews: (Ghavami et al., 2014; Kesidou et al., 2013; Nixon, 2013)). Our work here has demonstrated for the first time that down-regulation of AAK1 increased susceptibility to neurodegeneration in a fly model of PD. Also, increased autophagy flux was observed in cells depleted of AAK1. These observations are likely linked, since autophagy is an essential process of neuronal homeostasis, the digestions of long-live/misfolded proteins and damaged organelles, and impaired in neurodegenerative diseases (see reviews:(Jiang & Mizushima, 2014; Kesidou et al., 2013; Schneider & Cuervo, 2014; Tan et al., 2014), particularly evident in the brain of PD patients and in animal models of PD (Cheung & Ip, 2009; Xiong et al., 2013). Therefore, this implicates AAK1 as a potential regulator of autophagy in neurodegeneration. Indeed, AAK1 has previously been implicated in both PD and ALS pathogenesis (Latourelle et al., 2009; Shi et al., 2014). Since, autophagy and endocytosis partly share the same machinery and both pathways are implicated in the degradation of cellular materials (see review: (Lamb et al., 2013)), it is possible that the role of GSK3-mediated phosphorylation of AAK1 in vesicle trafficking events is important for initiating autophagy processes of the cell. Although autophagy is typically associated with neuroprotective functions in the brain, more evidence is emerging suggesting that excessive autophagy leads to cell death and neurodegeneration (Cheung & Ip, 2009; Xiong et al., 2013).

161 |Page

Sec. 5.2. Research implications

The role of GSK3 in autophagy is still unresolved. It has been suggested that GSK3 inhibition promotes lysosomal biogenesis and autophagy degradation of Amyolid- β Precursor Protein (AAP), implying a role in the negative regulation of autophagy that may attribute to the progression of intracellular neurofibrillary tangles of aggregated proteins and neurodegenerative disease progression (Komatsu et al., 2006; Parr et al., 2012). However, other research has implicated the activation of GSK3 in the absence of growth factors as a pathway to elicit increased autophagy (Lin et al., 2012a; Yu et al., 2008). Indeed, it has been well established that deprivation of growth factors induces autophagic processes in cells (Li et al., 2013), and since GSK3 activity is increased in the absence of growth factor signalling, it is reasonable to link increased GSK3 activity with autophagy and cell survival/death. Indeed, following insulin withdrawal (inducing activation of GSK3), adult hippocampal neural stem cells undergo autophagic cell death (Yu et al., 2008). This was found to be a consequence of signalling downstream of growth factor deprivation that deactivates mTOR, a negative regulator of autophagy. Interestingly, GSK3 has been implicated in the downstream regulation of mTOR (Ka et al., 2014). For example, GSK3 has been shown to inhibit mTOR signalling by phosphorylating tuberous sclerosis complex subunit 2 (TSC2) (Buller et al., 2008; Inoki et al., 2006), a component of the mTOR repressor complex. mTOR activity is increased in the absence of GSK3 regulation in knockout mice, resulting in decreased autophagy (Zhou et al., 2013). Therefore, mTOR pathways may play an important role in transducing GSK3-mediated autophagy signals in neural cells. In addition, GSK3 activation in the absence of growth factors has been shown to phosphorylate and activate the acetyltransferase TIP60 at Ser86 (Lin et al., 2012b). This induces the acetylation and activation of the protein kinase ULK1, which leads to the initiation of autophagy. Indeed, ULK1-/- mouse fibroblasts fail to undergo serum deprivation-induced autophagy (Lin et al., 2012b). Furthermore, siRNA knockdown of TIP60 in cells and expression of the non- phosphorylatable alanine mutant (TIP60-S86A), similarly impairs autophagy induction (Lin et al., 2012b). Together, these works suggest that increased GSK3 signalling promotes autophagy in cells.

Our work here is in agreement with this idea and suggests that GSK3 hyperactivity positively regulates autophagy flux via the phosphorylation of AAK1. AP-2 is an established binding partner of AAK1 and required for autophagosome production (Rong et al., 2012; Tian et al., 2013; Tian et al., 2014). Our results suggest that increased

162 |Page

Sec. 5.2. Research implications

phosphorylation of AAK1 promotes dissociation of AAK1/AP-2 for increased trafficking of AP-2 towards the cell surface, potentially enhancing CCV internalisation (Motley et al., 2006; Olusanya et al., 2001; Ricotta et al., 2002) that is required for autophagy initiation (Tian et al., 2013; Tian et al., 2014). Indeed, if GSK3-mediated dissociation of AAK1/AP-2 is important in promoting autophagy clearance, this would demonstrate another mechanism, in addition to TIP60 activation and mTOR inhibition, which links GSK3 activity to increased autophagy flux in cells. Our results suggest that when GSK3 activity is low, AAK1 antagonises autophagy to reduce autophagy flux. This may suggest a role for lithium in reducing GSK3 activity and to limit excessive autophagy flux that could lead to neural death in PD patients and possibly other neurodegenerative diseases.

Indeed, lithium has an established role as an autophagy modulator (see review: (Pasquali et al., 2009). As an inhibitor of GSK3, lithium would negatively regulate autophagy. However, lithium has also been found to induce autophagy by inhibiting inositol-monophosphatase (IMPase), decreasing the autophagy antagonist, 1,4,5-inositol triphosphate (IP-3), independent of GSK3 (Sarkar et al., 2005). Since, lithium is known to affect multiple (and often opposite) pathways, the final effect of its action critically depends on the dose or duration of exposure. Therefore, it is not surprising that lithium has a neuronal protective function by promoting increased autophagy and facilitating the clearance of proteins that form toxic aggregates, including, α-synuclein (Lewy bodies), Tau (NFTs), AAP, SOD1 and huntingtin (Fornai et al., 2008; Leroy et al., 2010; Parr et al., 2012; Sarkar et al., 2005). This may suggest, at least in part, that lithium’s therapeutic action in neuroprotection extends beyond GSK3 inhibition.

Depending on the context of stresses in neurons, autophagy can promote both cell survival as well as cell death. Furthermore, both autophagy inducers and inhibitors have been proven to serve as neuroprotectors against PD (Li et al., 2011; Pan et al., 2008a; Xiong et al., 2011). It is widely agreed that autophagy can be extremely useful for the clearance of toxic aggregate proteins in post-mitotic cells, like neurons and is impaired in many neurodegenerative diseases. However, excessive autophagy activation has also been associated with neuronal loss (see review: (Bredesen et al., 2006)). Indeed, abnormal presence of autophagic vacuoles are evident in the brains of PD patients, in contrast to the rare detection of autophagosomes in normal healthy brains (Anglade et al., 1997; Zhu et al., 2003). This could represent aberrant activation of autophagy, although defective clearance of the autophagic vacuoles could also account for this observation.

163 |Page

Sec. 5.2. Research implications

Furthermore, overexpression of proteins that aggregate in PD Lewy bodies, including mutant α-synuclein (Xilouri et al., 2009) and GPR37 (Marazziti et al., 2009), have shown to induce autophagy and promote autophagic cell death. Indeed, recent studies investigating familial PD have demonstrated that dopaminergic neurons derived from PD patient fibroblast expressing the PD-related LRRK2 mutation, G2019S, undergo excessive autophagy clearance (Bravo-San Pedro et al., 2013; Ramonet et al., 2011; Su & Qi, 2013). Overall, autophagy is a highly regulated catabolic pathway that requires balanced rates of induction and degradation for effective neural homeostasis. Although the activation of autophagy may serve a protective function in the brain, excessive activation could result in neuronal death and hence also contribute to neuron loss in PD. Understanding the mechanisms by which activated GSK3 regulates autophagy homeostasis will offer important insights for PD and other neurological diseases. In particular, identifying downstream substrates of GSK3 that transduce GSK3-dependent signals in autophagy may be beneficial for developing future therapeutics to combat the harmful levels of autophagy stress in PD. Indeed, our results suggest that GSK3-mediated phosphorylation of AAK1 promotes dissociation of AAK1/AP-2 complex, likely by regulating AAK1 phosphorylation of the AP-2 subunit µ2. If this is the case, targeting AAK1’s kinase activity may represent a potential therapeutic target in PD, whereby its inhibition may be an effective means to sequester AP-2 away from the cell surface and reduce excessive autophagic flux in cells. However, this has yet to be tested and requires further investigation.

Overall, this research advances the knowledge of GSK3’s role in vesicle trafficking/transport. We have identified and investigated two novel GSK3 substrates, PI4KIIα and AAK1, and their role in vesicle trafficking in the brain. This work sets up future research into better understanding how GSK3 regulates trafficking proteins in healthy and diseased brains and help delineate the affect this has on neurotransmission, synaptic plasticity and neurodegeneration as well as the potential implications this may have for BD and PD pathogenicity. Ultimately, this knowledge could aid the development of novel more effective, targeted medications while reducing adverse side-effects of current treatments available in the clinic today. This would lead to improve patient outcomes, allowing them to lead more normal lives, while also helping to reduce costs on public health services.

164 |Page

Sec. 5.3. Project limitations

5.3 Project limitations

The complex nature of neurological diseases and the established complexity of GSK3’s role in healthy and diseased brains signifies the need for ongoing research to identify novel substrates and to investigate their role in brain function. Indeed, our study has only focused on only two of the extensive number of target proteins that are functionally dependent on the action of GSK3 phosphorylation, either directly or indirectly, that may have pathogenic effects on neurological signalling, neuronal morphology or neurogenesis in the brain – and therefore limited in scope. Given this, PI4KIIα and AAK1 are contributing factors to neurological function/dysfunction in the brain, however, the extent to which other factors contribute is still unknown.

Due to limited time and/or lack of resources, there are a number of areas of this project that require further investigation to better understand the role of GSK3 phosphorylation of PI4KIIα and AAK1 in the brain.

Firstly, it’s important to clarify the role of these substrates in human neurological diseases. While we have successfully identified behavioural or physiological phenotypes from dysregulation of these substrates in flies, these associations can only be considered consistent with, but not representative of, human disease. Further work is required in higher organisms, such as the mouse, in order to establish their relevance in human disease neurophysiology. This would include generating knockout and/or phospho- mutant knockin mice for PI4KIIα and AAK1. This would provide tools to better explore complex and physiological relevant behaviours, such as anxiety, depression and learning/memory function, reminiscent of human disease states, as well as elucidate the mechanism of phosphorylation to help validate these as targets in neurological diseases.

Secondly, lithium treatment of the fly reduced hyperactivity observed in PI4KIIα knockdown flies. This may be indicative of lithium directly reducing phosphorylation of endogenous PI4KIIα in the fly brain and consistent with the observed reduction of PI4KIIα phosphorylation in vitro. However, it’s possible that lithium may reduce hyperactivity independent of regulating PI4KIIα, instead affecting other proteins or signalling pathways that normalise activity. Therefore, further work is required to characterise phosphorylation of PI4KIIα in the fly. This can be done by generating a PI4KIIα phospho-mutant line to compare and contrast how phosphorylation affects hyperactivity in the fly. Alternatively, generating phospho-specific antibodies to explore

165 |Page

Sec. 5.3. Project limitations

phosphorylation of endogenous PI4KIIα in the fly and the effect lithium has on this process.

Thirdly, our focus was on understanding the role of GSK3-mediated phosphorylation and as such we did not thoroughly investigate the enzymatic activity of

PI4KIIα on its PI substrate. Indeed, its PI lipid products (e.g. PI4P, PI(4,5)P2 and

PI(3,4,5)P3) are known to be important for endocytosis and trafficking. In particular, PI4P production has been demonstrated to promote Golgi-to-endosomal transport (Daboussi et al., 2012; Jović et al., 2012). We expect that GSK3-mediated phosphorylation may indirectly regulate PI4P production and trafficking by controlling the abundance of PI4KIIα itself rather than directly regulating its kinase activity. Evidence suggests that elevated phosphoinositol signalling may be an underlying cause of BD and that the down- regulation of phosphoinositide levels following BD drug treatment may underlie improved patient outcomes (see review: (Ludtmann et al., 2011)). For example, lithium attenuates inositol recycling through inhibition of IMPase and inositol polyphosphate 1- phosphatase (IPPase) (Hallcher & Sherman, 1980) and hence down-regulates phosphoinositol signalling (King et al., 2009). Interestingly, IMPase polymorphisms have also been associated with increased risk of BD (Ohnishi et al., 2007; Sjøholt et al., 2004; Sjøholt et al., 2000). Together, this implicates reduced phosphoinositide-mediated signalling as a potential target for treatment of BD. Given this, it is possible that targeting PI4KIIα’s kinase activity may offer a therapeutic alternative in regulating PI4P production and cargo transport from Golgi/endosomes to lysosomes. Therefore, it would be interesting to determine whether GSK3 directly regulates PI lipids levels at the plasma or intracellular membranes. For example, does the expression of non-phosphorylatable mutant PI4KIIα affect the abundance of PI4P or PIP2 at membranes? Recent advances in mass spectrometry has made it possible to measure different lipid classes in cells (Ivanova et al., 2009; Kielkowska et al., 2014; Postle et al., 2007) and provides a sensitive and accurate method to explore the role of GSK3 in phosphoinositide signalling.

Finally, we have established AAK1 as having an important role in regulating AP- 2 mediated autophagic clearance in cells and contributing to neurodegeneration in our fly PD model. To better characterise the role of AAK1 in cell death, a cell viability assay could be conducted using a stable knockdown cell line plus rescue experiment. Unfortunately, our stable AAK1 knockdown cells had limited reduction of AAK1 protein levels (~50%) and are therefore not ideal. Unlike the approach we used to achieve gene

166 |Page

Sec. 5.3. Project limitations

knockdown using shRNA interference, a different means of silencing genes identified in prokaryotic adaptive immunity, has recently been adapted for genome editing in eukaryotic cells involving loci called ‘Clustered Regularly Interspaced Short Palindromic Repeats’ or CRISPRs. The CRISPR/Cas9 system is comprised of two components: the CRISPR associated (Cas) endonuclease, Cas9 and sequence specific guide-RNA (gRNA). This technique allows for sequence-specific induction of double-strand breaks in genomic DNA of individual cells, including primary neurons, effectively resulting in knock-out of targeted genes (Straub et al., 2014). This approach would be ideal for generating stable AAK1 knockout cells as well as knockout mice.

AAK1 knockout cells would be useful for further investigating the role of AAK1 in autophagy. In particular, staining for other subcellular markers, including endosomal and autophagosome-like structures, such as Rabs and autophagy-related (ATGs) proteins, to help characterise which step along the endocytosis/trafficking pathway that AAK1 depletion affects autophagy clearance. Furthermore, performing rescue experiments by expressing AAK1-phosphomutants to establish how exactly GSK3 phosphorylation affects vesicle trafficking and promotes autophagy flux. For example, does the phosphorylation of AAK1 affect the trafficking of other endocytic/autophagy related proteins, such as ATGs and components of the AMP-activated protein kinase (AMPK) signalling pathway, important for inducing autophagy processes? (Lee et al., 2014). Additionally, does GSK3-AAK1 regulate the trafficking of AMPA receptors (similar to the role of PI4KIIα in chapter 3) or other synaptic proteins important for neurotransmission in mood disorders? Together, these experiments are important to help determine the specific function of GSK3-mediated phosphorylation of AAK1 and the effect on neurotransmission.

167 |Page

Sec. 5.4. Recommendations for future research

5.4 Recommendations for future research

5.4.1 Validating the role of PI4KIIα and AAK1 in mood and neurodegenerative disorders of the brain.

This study provides impetus for future investigation required to validate and further explore the roles of PI4KIIα and AAK1 in vesicular trafficking and their likely roles in neurotransmission in the brain. This would include further work in more physiological systems to confirm their functional effects on synaptic transmission and further characterise their role in mood and neurodegenerative disorders in which GSK3 activity is dysregulated.

It is recommended that future studies utilise available technologies for manipulating target genes by generating knockout or phospho-mutant knockin mice for PI4KIIα and AAK1. This would provide valuable tools to further investigate their physiological role in early-stage animal testing and help to validate these targets as potential therapeutic intervention points. An ideal approach would be to use the CRISPR/Cas9 systems and sequence specific gRNA to either disrupt the target gene (via insertions/deletion) using the Homologous End Joining (NHEJ) DNA repair pathway or edit/modify the endogenous genome, using a specific repair template, via Homology Directed Repair (HDR) (see review: (Sander & Joung, 2014)) (Fig.5.1). The CRISPR components can be introduced to the brain of mouse embryos in-utero or adult brain in vivo via adeno-associated viral vectors (AAV) or electroporation transfections as described previously (Straub et al., 2014; Swiech et al., 2015). Alternatively, microinjections of transcribe CRISPR RNA components into the cytoplasm of mouse zygotes is an effective method for generating knockout/knockin mice (Horii et al., 2014; Wang et al., 2013). This technology is becoming widely-used because of its ease of use and providing an efficient and complete means for genetic manipulation of genes. Therefore, this technique would be ideal for creating mouse models for investigating the roles of targets genes on neuronal function in vivo.

These knockout/knockin mouse models would enable the assessment and validation of behavioural and physiological phenotypes in higher organisms that resemble human disease states. For example, looking for behaviours that are consistent with mood disorders, such as changes in locomotor activity, sleep behaviour, circadian rhythm, anxiety and feeding patterns. Whereas, investigating changes in neurophysiology include

168 |Page

Sec. 5.4. Recommendations for future research

assessing for neurodegeneration, measuring neuronal proliferation, differentiation and incorporation of newborn neurons into existing circuits. It will be interesting to determine whether crossing AAK1 knockdown mice with mice that exhibit PD pathology will exacerbate this phenotype, similar to what we observed in our Drosophila model of PD. Finally, phospho-mutant knockin models would help elucidate how phosphorylation of PI4KIIα and AAK1 affects interaction with other binding partners or downstream proteins in vivo, their role in vesicle trafficking and how this may contribute to disease state.

Figure 5.1 Schematic of the CRISPR/Cas9 system to disrupt or modified the genomic target. The CRISPR/Cas9 systems and sequence specific gRNA can be used to either disrupt the target gene (via insertions/deletion (InDels)) using the Homologous End Joining (NHEJ) pathway or edit/modify the endogenous genome, using a specific repair template, via Homology Directed Repair (HDR). (Figure adapted from (Addgene.org, 2014)).

169 |Page

Sec. 5.4. Recommendations for future research

It is also important to investigate PI4KIIα and AAK1 functional effects on synaptic transmission by performing electrophysiological measurement of action potentials in rodent brain slices. Using brains from knockout and knockin models, generated using the CRISPR/Cas9 system, will provide valuable information on pre/post synaptic response and elucidate the affect target expression has on modulating synaptic transmission at either side of the synapse. Since PI4KIIα regulates AMPA receptor levels at the cell surface, it would be interesting to validate the effect this has on synaptic transmission. For example, it is expected that PI4KIIα knockdown and expression of de- phosphomutant PI4KIIα would increase synaptic currents across the synaptic cleft and promote LTP as a result of increased AMPA receptor expression. Furthermore, conducting electrophysiological measurements at different life stages could help characterise onset of disease and disease progression throughout life.

Ultimately, this work will further characterise the role of PI4KIIα and AAK1 in the brain and confirm their role in neurological disease. This could potentially validate these substrates as novel therapeutic targets for improved treatment of BD and/or PD.

5.4.2 Screening other GSK3 candidate substrates using the DAMs assay for behaviours associated with mood disorders

Using bioinformatics and mass spectrometry phosphoproteomics, our lab has identified many novel GSK3 substrates. Although this project focuses on only two of these, many other substrates offer promise as potentially important targets in human neurological diseases (Fig.5.2). Therefore, we screened a number of other candidate genes identified in our labs as promising GSK3 targets in mood disorders. Interestingly, two of these are lipid phosphatases, PTEN (Chu & Tarnawski, 2004; Hlobilková et al., 2003) and synaptojanin (Haffner et al., 1997; Lee et al., 2004b), consistent with PI4KIIα and phospholipids being important for trafficking. Furthermore, SGIP1 (Dergai et al., 2010; Uezu et al., 2007), intersectin-2 (Pucharcos et al., 2000) and EFR3B (Bojjireddy et al., 2015) are associated with vesicular trafficking. Therefore, these substrates may also have functional roles in synaptic plasticity, trafficking and neurotransmission at synapses of the brain and are worthy of further investigation.

170 |Page

Sec. 5.4. Recommendations for future research

Figure 5.2. Screening other novel GSK3 substrates potentially important in human neurological disease. Substrates of interest were identified by a combination of bioinformatics and mass spec phosphoproteomics and selected candidates were screened for behaviours associated with BD.

To test this, candidate gene expression was specifically knocked down in neurons of flies using the UAS-Gal4 system and flies were then subjected to DAMs analysis to screen for changes in locomotor activity or circadian rhythm that correlate with symptoms of mood disorders in humans. A few candidate genes tested in this system produced similar hyperactivity phenotypes as observed for PI4KIIα-depleted flies. Furthermore, the increase in locomotor activity was most pronounced 4 hours before the dark phase (8-12 ZT), again consistent with PI4KIIα-depleted flies. Accordingly, many of these positive hits are endocytic membrane proteins, phosphatases or voltage-gated ion channels that have functional roles in trafficking and neurotransmission at synapses of the brain (Catterall et al., 2005; Dergai et al., 2010; Haffner et al., 1997; Lee et al., 2004b; Pucharcos et al., 2000; Uezu et al., 2007). These include: CACNA1c, PTEN (Fig.5.3B),

171 |Page

Sec. 5.4. Recommendations for future research

synaptojanin (Fig.5.3D), and SGIP1 (Fig.5.3F). In particular CACNA1c, a voltage- dependent calcium channel, has been associated with psychiatric illness including BD, MDD (Liu et al., 2011; Sklar et al., 2011) and schizophrenia (Bigos et al., 2010; Ripke et al., 2011). Whereas, polymorphisms in synaptojanin has been associated with early-onset PD (Krebs et al., 2013; Quadri et al., 2013). Together with PI4KIIα, PTEN and synaptojanin are lipid enzymes (phosphatases) (Chu & Tarnawski, 2004; Haffner et al., 1997; Hlobilková et al., 2003; Lee et al., 2004b), further implicating phospholipid signalling as an important target of GSK3 in neurons. Phospholipid signalling pathways are critical in the brain and are modulated by the actions of lipid kinases and phosphatases. Dysregulated phospholipid signalling has been implicated in BD and a target of lithium and other mood stabilizers and anti-psychotic medications (King et al., 2009; Ludtmann et al., 2011; Machado‐Vieira et al., 2009; Williams et al., 2002). Interestingly, decreased phosphorylation of lipid second messengers as a result of dysregulation of PTEN phosphatase activity has previously been observed in the brains of depressed and suicide victims (Karege et al., 2011). Furthermore, polymorphisms in synaptojanin have also been linked to BD (Saito et al., 2001; Stopkova et al., 2004). Together, this suggests that modulation of phospholipid second messenger pathways may be an underlying cause and treatment for patients suffering mood disorders. It is possible that the combined dysregulation of GSK3-mediated phosphorylation of PI4KIIα, PTEN and synaptojanin may offer further insights into how GSK3 modulates mood. Therefore, phospholipid signalling might be a site of convergence for both GSK3-dependent and independent actions of current drug treatments. Since PI4KIIα, PTEN and synaptojanin are lipid enzymes, it is possible that one or more of these could become alternative therapeutic targets in the regulation of lipid messengers for improving mood state in BD. Further exploring the role of GSK3 phosphorylation on lipid signalling in the brain offers an exciting area for future research.

Together, this supports the idea that GSK3 directly regulates a number of trafficking proteins in the brain. Dysregulation of these may affect the fine balance of these trafficking proteins that are critical for efficient transmission and synaptic plasticity in the brain. Potentially, this may contribute to the development of mood disorders and requires further investigation.

172 |Page

Sec. 5.4. Recommendations for future research

In addition, a number of substrates identified here may represent interesting targets in other human diseases. For example, Bc1-11a, ZEB1/2 and JAZF1 are zinc- finger transcription factors that are genetically associated with cancer (Koontz et al., 2001; Liu et al., 2014; Satterwhite et al., 2001; Weniger et al., 2006) and type-2 diabetes (Langberg et al., 2012; Stevens et al., 2010; Voight et al., 2010). Also, Ube4b, a ubiquitin ligase, is an important regulator of the tumour suppressor p53 in human cancers (Zhang et al., 2014). Similar to the role of phosphorylation, ubiquitin modification allows recognition and association of cargoes with endosomally associated protein complexes for sorting of proteins destined to be either degraded by lysosomes or recycled. Since Ube4b has been shown to ubiquitinate EGFR (Sirisaengtaksin et al., 2014) and the MDM2/p53 complex (Wu & Leng, 2011; Wu et al., 2011), both of which are implicated in tumour growth, it is possible that GSK3 phosphorylation of Ube4b may be a mechanism to control the cellular levels of EGFR and MDM2/p53. The affect this has on downstream signalling may be important for regulating tumour growth. Together, this presents many new areas to investigate the role of GSK3 signalling in human diseases and provides great promise for future work.

173 |Page

Sec. 5.4. Recommendations for future research

Figure 5.3. Actogram of GSK3 gene candidates - neural-specific RNAi knockdown flies. (A) Locomotor activity (count/30 min) of Ube4B-RNAi (blue line) knockdown and driver only control (W1118,dcr2;;nSyb-Gal4, green line). (B) Locomotor activity (count/30 min) of CACNA1c-RNAi (blue line), PTEN-RNAi (red line) knockdown and driver only control (W1118,dcr2;;nSyb-Gal4, green line). (C) Locomotor activity (count/30 min) of JAZF1-RNAi (blue line), PCTK1-RNAi (red line) knockdown and driver only control (W1118,dcr2;;nSyb- Gal4, green line). (D) Locomotor activity (count/30 min) of Synaptojanin-RNAi (blue line) knockdown and driver only control (W1118,dcr2;;nSyb-Gal4, green line). (E) Locomotor activity (count/30 min) of Bcl-11a-RNAi (blue line), EFR3B-RNAi (red line) knockdown and driver only control (W1118,dcr2;;nSyb-Gal4, green line). (F) Locomotor activity (count/30 min) of intersectin-2-RNAi (purple line), SGIP1-RNAi (blue line), ZEB1/2-RNAi (red line) knockdown and driver only control (W1118,dcr2;;nSyb-Gal4, green line). All averages over 6 d.

174 |Page

Sec. 5.5. Overall conclusion

5.5 Overall conclusion

GSK3 is an important enzyme in the brain for regulating neuroplasticity, neurotransmission and morphology and is found dysregulated in many neurological diseases. It is an important kinase to study in order to fully understand basic signalling mechanisms underlying several physiological processes in healthy and diseased brains. Antipsychotic, antidepressants and other mood stabilisers target GSK3 activity in the brain, implicating a role for GSK3 in disease pathogenicity. However, current treatment options have side-effects that limit their use and reduce adherence. Therefore, identifying downstream substrates of GSK3 and targeting them therapeutically could improve specificity for pathogenic effects while reducing side-effects. Our group discovered an enrichment of novel substrates of GSK3 involved in vesicular trafficking and synaptic transmission, suggesting an important class of protein that is dysregulated in diseased brains. This project has focused on two promising trafficking proteins; the lipid kinase PI4KIIα and the AP-2 kinase AAK1. Both of these enzymes are highly expressed in the brain and have been implicated in trafficking events, including vesicle formation, transport/delivery and sorting functions, all important during neurotransmission and regulating synaptic plasticity. This work attempts to elucidate the molecular mechanisms by which GSK3-mediated phosphorylation can regulate synaptic vesicle trafficking in neurons and the implications this may have for neurological function. We have established a role for PI4KIIα in regulating cell-surface expression of the AMPA receptor, GluA1 in primary hippocampal neurons and presented evidence for a new role of AAK1 in regulating AP-2-mediated autophagic processes. Finally, we have demonstrated that deregulation of PI4KIIα and AAK1 activity is linked with behavioural and physiological phenotypes that are consistent with a role in the pathogenesis and/or treatment of BD and PD, respectively. Overall, this research provides new information on the role of GSK3 phosphorylation in neurons and provides impetus for future discovery of novel GSK3 targets involved in neurological signalling in healthy and diseased brains. This will aid the development of more effective and targeted medication with the aim to improve outcomes for patients that suffer from debilitating neurological conditions.

175 |Page

References

Aberle, H., Bauer, A., Stappert, J., Kispert, A., & Kemler, R. (1997). β‐catenin is a target for the ubiquitin–proteasome pathway. The EMBO Journal, 16(13), 3797- 3804. doi: 10.1093/emboj/16.13.3797 Ackermann, T. F., Kempe, D. S., Lang, F., & Lang, U. E. (2010). Hyperactivity and enhanced curiosity of mice expressing PKB/SGK-resistant glycogen synthase kinase-3 (GSK-3). Cellular Physiology and Biochemistry, 25(6), 775-786. doi: 10.1159/000315097 Adachi, A., Kano, F., Tsuboi, T., Fujita, M., Maeda, Y., & Murata, M. (2010). Golgi- associated GSK3β regulates the sorting process of post-Golgi membrane trafficking. Journal of Cell Science, 123(19), 3215-3225. doi: 10.1242/jcs.063941 Addgene.org. (2014). CRISPR in the Lab: A Practical Guide. Retrieved 12 Feburary, 2015, from https://www.addgene.org/CRISPR/guide/ Department of Health and Ageing (2013). National Mental Health Report 2013: Tracking progress of mental health reform in Australia 1993 – 2011. Commonwealth of Australia, Canberra. Aguilar, R. C., Ohno, H., Roche, K. W., & Bonifacino, J. S. (1997). Functional domain mapping of the clathrin-associated adaptor medium chains μ1 and μ2. Journal of Biological Chemistry, 272(43), 27160-27166. doi: 10.1074/jbc.272.43.27160 Ahnaou, A., & Drinkenburg, W. (2011). Disruption of glycogen synthase kinase-3-beta activity leads to abnormalities in physiological measures in mice. Behavioural Brain Research, 221(1), 246-252. doi: 10.1016/j.bbr.2011.03.004 Alberi, L., Hoey, S. E., Brai, E., Scotti, A. L., & Marathe, S. (2013). Notch signaling in the brain: In good and bad times. Ageing Research Reviews, 12(3), 801-814. doi: 10.1016/j.arr.2013.03.004 Alimohamad, H., Rajakumar, N., Seah, Y.-H., & Rushlow, W. (2005). Antipsychotics alter the protein expression levels of β-catenin and GSK-3 in the rat medial prefrontal cortex and striatum. Biological Psychiatry, 57(5), 533-542. doi: 10.1016/j.biopsych.2004.11.036 Alvarez, G., Muñoz‐Montaño, J. R., Satrústegui, J., Avila, J., Bogónez, E., & Díaz‐ Nido, J. (2002). Regulation of tau phosphorylation and protection against β‐ amyloid‐induced neurodegeneration by lithium. Possible implications for Alzheimer's disease. Bipolar Disorders, 4(3), 153-165. doi: 10.1034/j.1399- 5618.2002.01150.x Anderson, I. M., Haddad, P. M., & Scott, J. (2012). Bipolar disorder. The BMJ, 345, e8508. doi: 10.1136/bmj.e8508 Andorfer, C., Kress, Y., Espinoza, M., De Silva, R., Tucker, K. L., Barde, Y. A., . . . Davies, P. (2003). Hyperphosphorylation and aggregation of tau in mice expressing normal human tau isoforms. Journal of Neurochemistry, 86(3), 582- 590. doi: 10.1046/j.1471-4159.2003.01879.x Anglade, P., Vyas, S., Javoy-Agid, F., Herrero, M. T., Michel, P. P., Marquez, J., . . . Agid, Y. (1997). Apoptosis and autophagy in nigral neurons of patients with Parkinson's disease. Histology and Histopathology, 12(1), 25-31. Antonietta De Matteis, M., Di Campli, A., & Godi, A. (2005). The role of the phosphoinositides at the Golgi complex. Biochimica et Biophysica Acta - Molecular Cell Research, 1744(3), 396-405. doi: 10.1016/j.bbamcr.2005.04.013

176 |Page

Arend, W. P., Malyak, M., Guthridge, C. J., & Gabay, C. (1998). Interleukin-1 receptor antagonist: Role in biology. Annual Review of Immunology, 16(1), 27-55. doi: 10.1146/annurev.immunol.16.1.27 Armstrong, J. L., Bonavaud, S. M., Toole, B. J., & Yeaman, S. J. (2001). Regulation of glycogen synthesis by amino acids in cultured human muscle cells. Journal of Biological Chemistry, 276(2), 952-956. doi: 10.1074/jbc.M004812200 Austin, C. P., Ky, B., Ma, L., Morris, J. A., & Shughrue, P. J. (2004). Expression of Disrupted-In-Schizophrenia-1, a schizophrenia-associated gene, is prominent in the mouse hippocampus throughout brain development. Neuroscience, 124(1), 3-10. doi: 10.1016/j.neuroscience.2003.11.010 Bain, J., Plater, L., Elliott, M., Shpiro, N., Hastie, C. J., McLauchlan, H., . . . Cohen, P. (2007). The selectivity of protein kinase inhibitors: A further update. Biochemical Journal, 408(3), 297-315. doi: 10.1042/bj20070797 Balla, A., & Balla, T. (2006). Phosphatidylinositol 4-kinases: Old enzymes with emerging functions. Trends in Cell Biology, 16(7), 351-361. doi: 10.1016/j.tcb.2006.05.003 Balla, A., Tuymetova, G., Barshishat, M., Geiszt, M., & Balla, T. (2002). Characterization of type II phosphatidylinositol 4-kinase isoforms reveals association of the enzymes with endosomal vesicular compartments. Journal of Biological Chemistry, 277(22), 20041-20050. doi: 10.1074/jbc.M111807200 Baron, M. (2003). An overview of the Notch signalling pathway. Seminars in Cell & Developmental Biology, 14(2), 113-119. doi: 10.1016/S1084-9521(02)00179-9 Bartman, C. M., Egelston, J., Kattula, S., Zeidner, L. C., D’Ippolito, A., Doble, B. W., & Phiel, C. J. (2014). Gene expression profiling in mouse embryonic stem cells reveals glycogen synthase kinase-3-dependent targets of phosphatidylinositol 3- kinase and Wnt/β-catenin signaling pathways. Frontiers in Endocrinology, 5, 133. doi: 10.3389/fendo.2014.00133 Barylko, B., Gerber, S. H., Binns, D. D., Grichine, N., Khvotchev, M., Südhof, T. C., & Albanesi, J. P. (2001). A novel family of phosphatidylinositol 4-kinases conserved from yeast to humans. Journal of Biological Chemistry, 276(11), 7705-7708. doi: 10.1074/jbc.C000861200 Barylko, B., Mao, Y. S., Wlodarski, P., Jung, G., Binns, D. D., Sun, H.-Q., . . . Albanesi, J. P. (2009). Palmitoylation controls the catalytic activity and subcellular distribution of phosphatidylinositol 4-kinase IIα. Journal of Biological Chemistry, 284(15), 9994-10003. doi: 10.1074/jbc.M900724200 Barylko, B., Wlodarski, P., Binns, D. D., Gerber, S. H., Earnest, S., Sudhof, T. C., . . . Albanesi, J. P. (2002). Analysis of the catalytic domain of phosphatidylinositol 4-kinase type II. Journal of Biological Chemistry, 277(46), 44366-44375. doi: 10.1074/jbc.M203241200 Baumlova, A., Chalupska, D., Róźycki, B., Jovic, M., Wisniewski, E., Klima, M., . . . Balla, T. (2014). The crystal structure of the phosphatidylinositol 4‐kinase IIα. EMBO Reports, 15(10), 1085-1092. doi: 10.15252/embr.201438841 Beaulieu, J.-M. (2007). Not only lithium: Regulation of glycogen synthase kinase-3 by antipsychotics and serotonergic drugs. The International Journal of Neuropsychopharmacology, 10(01), 3-6. doi: 10.1017/S1461145706006857 Beaulieu, J.-M., Gainetdinov, R. R., & Caron, M. G. (2007). The Akt-GSK-3 signaling cascade in the actions of dopamine. Trends in Pharmacological Sciences, 28(4), 166-172. doi: 10.1016/j.tips.2007.02.006

177 |Page

Beaulieu, J.-M., Gainetdinov, R. R., & Caron, M. G. (2009). Akt/GSK3 signaling in the action of psychotropic drugs. Annual Review of Pharmacology and Toxicology, 49, 327-347. doi: 10.1146/annurev.pharmtox.011008.145634 Beaulieu, J.-M., Sotnikova, T. D., Marion, S., Lefkowitz, R. J., Gainetdinov, R. R., & Caron, M. G. (2005). An Akt/β-arrestin 2/PP2A signaling complex mediates dopaminergic neurotransmission and behavior. Cell, 122(2), 261-273. doi: 10.1016/j.cell.2005.05.012 Beaulieu, J.-M., Sotnikova, T. D., Yao, W.-D., Kockeritz, L., Woodgett, J. R., Gainetdinov, R. R., & Caron, M. G. (2004). Lithium antagonizes dopamine- dependent behaviors mediated by an AKT/glycogen synthase kinase 3 signaling cascade. Proceedings of the National Academy of Sciences, 101(14), 5099-5104. doi: 10.1073/pnas.0307921101 Beaulieu, J.-M., Zhang, X., Rodriguiz, R. M., Sotnikova, T. D., Cools, M. J., Wetsel, W. C., . . . Caron, M. G. (2008). Role of GSK3β in behavioral abnormalities induced by serotonin deficiency. Proceedings of the National Academy of Sciences, 105(4), 1333-1338. doi: 10.1073/pnas.0711496105 Bednarek, E., & Caroni, P. (2011). β-Adducin is required for stable assembly of new synapses and improved memory upon environmental enrichment. Neuron, 69(6), 1132-1146. doi: 10.1016/j.neuron.2011.02.034 Behrens, J., von Kries, J. P., Kühl, M., Bruhn, L., Wedlich, D., Grosschedl, R., & Birchmeier, W. (1996). Functional interaction of β-catenin with the transcription factor LEF-1. Nature, 382(6592), 638-642. doi: 10.1038/382638a0 Benedetti, F., Bernasconi, A., Lorenzi, C., Pontiggia, A., Serretti, A., Colombo, C., & Smeraldi, E. (2004a). A single nucleotide polymorphism in glycogen synthase kinase 3-B promoter gene influences onset of illness in patients affected by bipolar disorder. Neuroscience Letters, 355(1), 37-40. doi: 10.1016/j.neulet.2003.10.021 Benedetti, F., Bollettini, I., Barberi, I., Radaelli, D., Poletti, S., Locatelli, C., . . . Colombo, C. (2012). Lithium and GSK3B Promoter Gene Variants Influence White Matter Microstructure in Bipolar Disorder. Neuropsychopharmacology, 38(2), 313-327. doi: 10.1038/npp.2012.172 Benedetti, F., Dallaspezia, S., Fulgosi, M. C., Lorenzi, C., Serretti, A., Barbini, B., . . . Smeraldi, E. (2007). Actimetric evidence that CLOCK 3111 T/C SNP influences sleep and activity patterns in patients affected by bipolar depression. American Journal of Medical Genetics Part B: Neuropsychiatric Genetics, 144(5), 631- 635. doi: 10.1002/ajmg.b.30475 Benedetti, F., Dallaspezia, S., Lorenzi, C., Pirovano, A., Radaelli, D., Locatelli, C., . . . Smeraldi, E. (2011). Gene-ene interaction of glycogen synthase kinase 3-β and serotonin transporter on human antidepressant response to sleep deprivation. Journal of Affective Disorders, 136(3), 514-519. doi: 10.1016/j.jad.2011.10.039 Benedetti, F., Poletti, S., Radaelli, D., Bernasconi, A., Cavallaro, R., Falini, A., . . . Locatelli, C. (2010). Temporal lobe grey matter volume in schizophrenia is associated with a genetic polymorphism influencing glycogen synthase kinase 3- B activity. Genes, Brain and Behavior, 9(4), 365-371. doi: 10.1038/npp.2012.172 Benedetti, F., Serretti, A., Colombo, C., Barbini, B., Lorenzi, C., Campori, E., & Smeraldi, E. (2003). Influence of CLOCK gene polymorphism on circadian mood fluctuation and illness recurrence in bipolar depression. American Journal of Medical Genetics Part B: Neuropsychiatric Genetics, 123(1), 23-26. doi: 10.1002/ajmg.b.20038

178 |Page

Benedetti, F., Serretti, A., Colombo, C., Lorenzi, C., Tubazio, V., & Smeraldi, E. (2004b). A glycogen synthase kinase 3-β promoter gene single nucleotide polymorphism is associated with age at onset and response to total sleep deprivation in bipolar depression. Neuroscience Letters, 368(2), 123-126. doi: 10.1016/j.neulet.2004.06.050 Benedetti, F., Serretti, A., Pontiggia, A., Bernasconi, A., Lorenzi, C., Colombo, C., & Smeraldi, E. (2005). Long-term response to lithium salts in bipolar illness is influenced by the glycogen synthase kinase 3-β− 50 T/C SNP. Neuroscience Letters, 376(1), 51-55. doi: 10.1016/j.neulet.2004.11.022 Berezovska, O., Xia, M., & Hyman, B. (1998). Notch is expressed in adult brain, is coexpressed with presenilin-1, and is altered in Alzheimer disease. Journal of Neuropathology & Experimental Neurology, 57(8), 738-745. Berlin, I., Givry-Steiner, L., Lecrubier, Y., & Puech, A. J. (1998). Measures of anhedonia and hedonic responses to sucrose in depressive and schizophrenic patients in comparison with healthy subjects. European Psychiatry, 13(6), 303- 309. doi: 10.1016/s0924-9338(98)80048-5 Bersudsky, Y., Shaldubina, A., Kozlovsky, N., Woodgett, J. R., Agam, G., & Belmaker, R. H. (2008). Glycogen synthase kinase-3β heterozygote knockout mice as a model of findings in postmortem schizophrenia brain or as a model of behaviors mimicking lithium action: Negative results. Behavioural Pharmacology, 19(3), 217-224. doi: 10.1097/FBP.0b013e3282feb099 Beurel, E., & Jope, R. S. (2006). The paradoxical pro-and anti-apoptotic actions of GSK3 in the intrinsic and extrinsic apoptosis signaling pathways. Progress in Neurobiology, 79(4), 173-189. doi: 10.1016/j.pneurobio.2006.07.006 Beurel, E., Michalek, S. M., & Jope, R. S. (2010). Innate and adaptive immune responses regulated by glycogen synthase kinase-3 (GSK3). Trends in Immunology, 31(1), 24-31. doi: 10.1016/j.it.2009.09.007 Bigos, K. L., Mattay, V. S., Callicott, J. H., Straub, R. E., Vakkalanka, R., Kolachana, B., . . . Weinberger, D. R. (2010). Genetic variation in CACNA1C affects brain circuitries related to mental illness. Archives of General Psychiatry, 67(9), 939- 945. doi: 10.1001/archgenpsychiatry.2010.96 Bijur, G. N., & Jope, R. S. (2000). Opposing Actions of Phosphatidylinositol 3‐Kinase and Glycogen Synthase Kinase‐3β in the Regulation of HSF‐1 Activity. Journal of Neurochemistry, 75(6), 2401-2408. doi: 10.1046/j.1471-4159.2000.0752401.x Boivin, D. B. (2000). Influence of sleep-wake and circadian rhythm disturbances in psychiatric disorders. Journal of Psychiatry and Neuroscience, 25(5), 446-158. Bojjireddy, N., Guzman-Hernandez, M. L., Reinhard, N. R., Jovic, M., & Balla, T. (2015). EFR3s are palmitoylated plasma membrane proteins that control responsiveness to G-protein-coupled receptors. Journal of Cell Science, 128(1), 118-128. doi: 10.1242/jcs.157495 Botteron, K. N., Raichle, M. E., Drevets, W. C., Heath, A. C., & Todd, R. D. (2002). Volumetric reduction in left subgenual prefrontal cortex in early onset depression. Biological Psychiatry, 51(4), 342-344. doi: 10.1016/S0006- 3223(01)01280-x Bradley, C. A., Peineau, S., Taghibiglou, C., Nicolas, C. S., Whitcomb, D. J., Bortolotto, Z. A., . . . Collingridge, G. L. (2012). A pivotal role of GSK-3 in synaptic plasticity. Frontiers in Molecular Neuroscience, 5, 13. doi: 10.3389/fnmol.2012.00013 Bravo-San Pedro, J. M., Niso-Santano, M., Gómez-Sánchez, R., Pizarro-Estrella, E., Aiastui-Pujana, A., Gorostidi, A., . . . de Munain, A. L. (2013). The LRRK2

179 |Page

G2019S mutant exacerbates basal autophagy through activation of the MEK/ERK pathway. Cellular and Molecular Life Sciences, 70(1), 121-136. doi: 10.1007/s00018-012-1061-y Brazil, D. P., & Hemmings, B. A. (2001). Ten years of protein kinase B signalling: A hard Akt to follow. Trends in Biochemical Sciences, 26(11), 657-664. doi: 10.1016/S0968-0004(01)01958-2 Bredesen, D. E., Rao, R. V., & Mehlen, P. (2006). Cell death in the nervous system. Nature, 443(7113), 796-802. doi: 10.1038/nature05293 Brittain, J. M., Duarte, D. B., Wilson, S. M., Zhu, W., Ballard, C., Johnson, P. L., . . . Wang, Y. (2011). Suppression of inflammatory and neuropathic pain by uncoupling CRMP-2 from the presynaptic Ca2+ channel complex. Nature Medicine, 17(7), 822-829. doi: 10.1038/nm.2345 Brunet, A., Datta, S. R., & Greenberg, M. E. (2001). Transcription-dependent and- independent control of neuronal survival by the PI3K–Akt signaling pathway. Current Opinion in Neurobiology, 11(3), 297-305. doi: 10.1016/S0959- 4388(00)00211-7 Brunoni, A. R., Lopes, M., & Fregni, F. (2008). A systematic review and meta-analysis of clinical studies on major depression and BDNF levels: implications for the role of neuroplasticity in depression. The International Journal of Neuropsychopharmacology, 11(08), 1169-1180. doi: 10.1017/S1461145708009309 Buller, C. L., Loberg, R. D., Fan, M.-H., Zhu, Q., Park, J. L., Vesely, E., . . . Brosius, F. C. (2008). A GSK-3/TSC2/mTOR pathway regulates glucose uptake and GLUT1 glucose transporter expression. American Journal of Physiology-Cell Physiology, 295(3), 836-843. doi: 10.1152/ajpcell.00554.2007 Busino, L., Millman, S. E., Scotto, L., Kyratsous, C. A., Basrur, V., O’Connor, O., . . . Pagano, M. (2012). Fbxw7α-and GSK3-mediated degradation of p100 is a pro- survival mechanism in multiple myeloma. Nature Cell Biology, 14(4), 375-385. doi: 10.1038/ncb2463 Buss, H., Dörrie, A., Schmitz, M. L., Frank, R., Livingstone, M., Resch, K., & Kracht, M. (2004). Phosphorylation of serine 468 by GSK-3β negatively regulates basal p65 NF-κB activity. Journal of Biological Chemistry, 279(48), 49571-49574. doi: 10.1074/jbc.C400442200 Caceres, A., Mautino, J., & Kosik, K. S. (1992). Suppression of MAP2 in cultured cerebellar macroneurons inhibits minor neurite formation. Neuron, 9(4), 607- 618. doi: 10.1016/0896-6273(92)90025-9 Cade, J. F. J. (1949). Lithium salts in the treatment of psychotic excitement. The Medical Journal of Australia, 2, 518-520. Canelles, M., Delgado, M. D., Hyland, K. M., Lerga, A., Richard, C., Dang, C. V., & León, J. (1997). Max and inhibitory c-Myc mutants induce erythroid differentiation and resistance to apoptosis in human myeloid leukemia cells. Oncogene, 14(11), 1315-1327. doi: 10.1038/sj.onc.1200948 Cao, L., Jiao, X., Zuzga, D. S., Liu, Y., Fong, D. M., Young, D., & During, M. J. (2004). VEGF links hippocampal activity with neurogenesis, learning and memory. Nature Genetics, 36(8), 827-835. doi: 10.1038/ng1395 Carmichael, J., Sugars, K. L., Bao, Y. P., & Rubinsztein, D. C. (2002). Glycogen synthase kinase-3β inhibitors prevent cellular polyglutamine toxicity caused by the Huntington's disease mutation. Journal of Biological Chemistry, 277(37), 33791-33798. doi: 10.1074/jbc.M204861200

180 |Page

Cartwright, P., McLean, C., Sheppard, A., Rivett, D., Jones, K., & Dalton, S. (2005). LIF/STAT3 controls ES cell self-renewal and pluripotency by a Myc-dependent mechanism. Development, 132(5), 885-896. doi: 10.1242/dev.01670 Catterall, W. A., Perez-Reyes, E., Snutch, T. P., & Striessnig, J. (2005). International Union of Pharmacology. XLVIII. Nomenclature and structure-function relationships of voltage-gated calcium channels. Pharmacological Reviews, 57(4), 411-425. doi: 10.1124/pr.57.4.5 Cha, G.-H., Kim, S., Park, J., Lee, E., Kim, M., Lee, S. B., . . . Cho, K. S. (2005). Parkin negatively regulates JNK pathway in the dopaminergic neurons of Drosophila. Proceedings of the National Academy of Sciences, 102(29), 10345-10350. doi: 10.1073/pnas.0500346102 Chalecka-Franaszek, E., & Chuang, D.-M. (1999). Lithium activates the serine/threonine kinase Akt-1 and suppresses glutamate-induced inhibition of Akt-1 activity in neurons. Proceedings of the National Academy of Sciences, 96(15), 8745-8750. doi: 10.1073/pnas.96.15.8745 Chan, M.-H., Chiu, P.-H., Lin, C.-Y., & Chen, H.-H. (2012). Inhibition of glycogen synthase kinase-3 attenuates psychotomimetic effects of ketamine. Schizophrenia Research, 136(1), 96-103. doi: 10.1016/j.schres.2012.01.024 Chen, C. H., Suckling, J., Lennox, B. R., Ooi, C., & Bullmore, E. T. (2011a). A quantitative meta-analysis of fMRI studies in bipolar disorder. Bipolar Disorders, 13(1), 1-15. doi: 10.1111/j.1399-5618.2011.00893.x Chen, J., Park, C. S., & Tang, S.-J. (2006). Activity-dependent synaptic Wnt release regulates hippocampal long term potentiation. Journal of Biological Chemistry, 281(17), 11910-11916. doi: 10.1074/jbc.M511920200 Chen, L., Salinas, G. D., & Li, X. (2009). Regulation of serotonin 1B receptor by glycogen synthase kinase-3. Molecular Pharmacology, 76(6), 1150-1161. doi: 10.1124/mol.109.056994 Chen, L., Zhou, W., Chen, P. C., Gaisina, I., Yang, S., & Li, X. (2011b). Glycogen synthase kinase-3beta is a functional modulator of serotonin-1B receptors. Molecular Pharmacology, 79(6), 974-986. doi: 10.1124/mol.111.071092 Chen, P., Gu, Z., Liu, W., & Yan, Z. (2007). Glycogen synthase kinase 3 regulates N- methyl-D-aspartate receptor channel trafficking and function in cortical neurons. Molecular Pharmacology, 72(1), 40-51. doi: 10.1124/mol.107.034942 Chen, S., Guttridge, D. C., You, Z., Zhang, Z., Fribley, A., Mayo, M. W., . . . Wang, C.- Y. (2001). Wnt-1 signaling inhibits apoptosis by activating β-catenin/T cell factor–mediated transcription. The Journal of Cell Biology, 152(1), 87-96. doi: 10.1083/jcb.152.1.87 Cheng, E. H.-Y., Wei, M. C., Weiler, S., Flavell, R. A., Mak, T. W., Lindsten, T., & Korsmeyer, S. J. (2001). BCL-2, BCL-X(L) Sequester BH3 Domain-Only Molecules Preventing BAX-and BAK-Mediated Mitochondrial Apoptosis. Molecular Cell, 8(3), 705-711. doi: 10.1016/S1097-2765(01)00320-3 Cheung, Z. H., & Ip, N. Y. (2009). The emerging role of autophagy in Parkinson’s disease. Molecular Brain, 2(1), 29. doi: 10.1186/1756-6606-2-29 Chin, L.-S., Li, L., Ferreira, A., Kosik, K. S., & Greengard, P. (1995). Impairment of axonal development and of synaptogenesis in hippocampal neurons of synapsin I-deficient mice. Proceedings of the National Academy of Sciences, 92(20), 9230-9234. doi: 10.1073/pnas.92.20.9230 Chinnaiyan, A. M., Tepper, C. G., Seldin, M. F., O'Rourke, K., Kischkel, F. C., Hellbardt, S., . . . Dixit, V. M. (1996). FADD/MORT1 is a common mediator of CD95 (Fas/APO-1) and tumor necrosis factor receptor-induced apoptosis.

181 |Page

Journal of Biological Chemistry, 271(9), 4961-4965. doi: 10.1074/jbc.271.9.4961 Cho, J.-H., & Johnson, G. V. W. (2003). Glycogen Synthase Kinase 3β Phosphorylates Tau at Both Primed and Unprimed Sites. Differential impact on microtubule binding. Journal of Biological Chemistry, 278(1), 187-193. doi: 10.1074/jbc.M206236200 Cho, J. H., & Johnson, G. V. W. (2004). Primed phosphorylation of tau at Thr231 by glycogen synthase kinase 3β (GSK3β) plays a critical role in regulating tau's ability to bind and stabilize microtubules. Journal of Neurochemistry, 88(2), 349-358. doi: 10.1111/j.1471-4159.2004.02155.x Chu, E. C., & Tarnawski, A. S. (2004). PTEN regulatory functions in tumor suppression and cell biology. Medical Science Monitor, 10(10), 235-241. Chu, K., Minogue, S., Hsuan, J., & Waugh, M. (2010). Differential effects of the phosphatidylinositol 4-kinases, PI4KIIα and PI4KIIIβ, on Akt activation and apoptosis. Cell Death & Disease, 1(12), e106. doi: 10.1038/cddis.2010.84 Chuang, D.-M. (2004). Neuroprotective and neurotrophic actions of the mood stabilizer lithium: Can it be used to treat neurodegenerative diseases? Critical Reviews in Neurobiology, 16(1-2), 83-90. doi: 10.1615/CritRevNeurobiol.v16.i12.90 Chuang, D.-M., Wang, Z., & Chiu, C.-T. (2011). GSK-3 as a target for lithium-induced neuroprotection against excitotoxicity in neuronal cultures and animal models of ischemic stroke. Frontiers in Molecular Meuroscience, 4. doi: 10.3389/fnmol.2011.00015 Chubb, J., Bradshaw, N., Soares, D., Porteous, D., & Millar, J. (2007). The DISC locus in psychiatric illness. Molecular Psychiatry, 13(1), 36-64. doi: 10.1038/sj.mp.4002106 Clark, I. E., Dodson, M. W., Jiang, C., Cao, J. H., Huh, J. R., Seol, J. H., . . . Guo, M. (2006). Drosophila pink1 is required for mitochondrial function and interacts genetically with parkin. Nature, 441(7097), 1162-1166. doi: 10.1038/nature04779 Clayton, E. L., Sue, N., Smillie, K. J., O'Leary, T., Bache, N., Cheung, G., . . . Robinson, P. J. (2010). Dynamin I phosphorylation by GSK3 controls activity- dependent bulk endocytosis of synaptic vesicles. Nature Neuroscience, 13(7), 845-851. doi: 10.1038/nn.2571 Cohen, P., & Frame, S. (2001). The renaissance of GSK3. Nature Reviews Molecular Cell Biology, 2(10), 769-776. doi: 10.1038/35096075 Cohrs, S. (2008). Sleep disturbances in patients with schizophrenia: Impact and effect of antipsychotics. CNS Drugs, 22(11), 939-962. doi: 10.2165/00023210- 200822110-00004 Cole, A., Frame, S., & Cohen, P. (2004a). Further evidence that the tyrosine phosphorylation of glycogen synthase kinase-3 (GSK3) in mammalian cells is an autophosphorylation event. Biochemical Journal, 377, 249-255. doi: 10.1042/BJ20031259 Cole, A. R. (2012). GSK3 as a sensor determining cell fate in the brain. Frontiers in Molecular Neuroscience, 5, 4. doi: 10.3389/fnmol.2012.00004 Cole, A. R. (2013a). Glycogen synthase kinase 3 substrates in mood disorders and schizophrenia. FEBS Journal, 280(21), 5213-5227. doi: 10.1111/febs.1240 Cole, A. R. (2013b). Regulation of Cell Fate in the Brain by GSK3,Trends in Cell Signaling Pathways in Neuronal Fate Decision, Dr Sabine Wislet-Gendebien (Ed.): InTech. Retrieved from http://www.intechopen.com/books/trends-in-cell-

182 |Page

signaling-pathways-in-neuronal-fate-decision/regulation-of-cell-fate-in-the- brain-by-gsk3. doi: 10.5772/55180 Cole, A. R., Causeret, F., Yadirgi, G., Hastie, C. J., McLauchlan, H., McManus, E. J., . . . Sutherland, C. (2006). Distinct priming kinases contribute to differential regulation of collapsin response mediator proteins by glycogen synthase kinase- 3 in vivo. Journal of Biological Chemistry, 281(24), 16591-16598. doi: 10.1074/jbc.M513344200 Cole, A. R., Knebel, A., Morrice, N. A., Robertson, L. A., Irving, A. J., Connolly, C. N., & Sutherland, C. (2004b). GSK-3 phosphorylation of the Alzheimer epitope within collapsin response mediator proteins regulates axon elongation in primary neurons. Journal of Biological Chemistry, 279(48), 50176-50180. doi: 10.1074/jbc.C400412200 Cole, A. R., Noble, W., Aalten, L. v., Plattner, F., Meimaridou, R., Hogan, D., . . . Verkhratsky, A. (2007). Collapsin response mediator protein‐2 hyperphosphorylation is an early event in Alzheimer’s disease progression. Journal of Neurochemistry, 103(3), 1132-1144. doi: 10.1111/j.1471- 4159.2007.04829.x Cole, A. R., Soutar, M. P. M., Rembutsu, M., Van Aalten, L., Hastie, C. J., McLauchlan, H., . . . Sutherland, C. (2008). Relative resistance of Cdk5- phosphorylated CRMP2 to dephosphorylation. Journal of Biological Chemistry, 283(26), 18227-18237. doi: 10.1074/jbc.M801645200 Collingridge, G. L., Isaac, J. T., & Wang, Y. T. (2004). Receptor trafficking and synaptic plasticity. Nature Reviews Neuroscience, 5(12), 952-962. doi: 10.1038/nrn1556 Collins, B. M., McCoy, A. J., Kent, H. M., Evans, P. R., & Owen, D. J. (2002). Molecular architecture and functional model of the endocytic AP2 complex. Cell, 109(4), 523-535. doi: 10.1016/S0092-8674(02)00735-3 Conner, S. D., & Schmid, S. L. (2002). Identification of an adaptor-associated kinase, AAK1, as a regulator of clathrin-mediated endocytosis. The Journal of Cell Biology, 156(5), 921-929. doi: 10.1083/jcb.200108123 Conner, S. D., & Schmid, S. L. (2003). Differential requirements for AP-2 in clathrin- mediated endocytosis. The Journal of Cell Biology, 162(5), 773-780. doi: 10.1083/jcb.200304069 Conner, S. D., Schröter, T., & Schmid, S. L. (2003). AAK1‐Mediated μ2 Phosphorylation is Stimulated by Assembled Clathrin. Traffic, 4(12), 885-890. doi: 10.1046/j.1398-9219.2003.0142.x Cortés-Vieyra, R., Bravo-Patiño, A., Valdez-Alarcón, J. J., Juárez, M. C., Finlay, B. B., & Baizabal-Aguirre, V. M. (2012). Role of glycogen synthase kinase-3 beta in the inflammatory response caused by bacterial pathogens. Journal of Inflammation, 9(1), 23-23. doi: 10.1186/1476-9255-9-23 Coryell, W., Nopoulos, P., Drevets, W., Wilson, T., & Andreasen, N. C. (2005). Subgenual prefrontal cortex volumes in major depressive disorder and schizophrenia: Diagnostic specificity and prognostic implications. The American Journal of Psychiatry, 162(9), 1706-1712. doi: 10.1176/appi.ajp.162.9.1706 Costa, R. M., Drew, C., & Silva, A. J. (2005). Notch to remember. Trends in Neurosciences, 28(8), 429-435. doi: 10.1016/j.tins.2005.05.003 Cowley, S., Paterson, H., Kemp, P., & Marshall, C. J. (1994). Activation of MAP kinase kinase is necessary and sufficient for PC12 differentiation and for transformation of NIH 3T3 cells. Cell, 77(6), 841-852. doi: 10.1016/0092- 8674(94)90133-3

183 |Page

Craddock, N., Jones, L., Jones, I. R., Kirov, G., Green, E. K., Grozeva, D., . . . O'Donovan, M. C. (2008). Strong genetic evidence for a selective influence of GABAA receptors on a component of the bipolar disorder phenotype. Molecular Psychiatry, 15(2), 146-153. doi: 10.1038/mp.2008.66 Craig, A. M., & Banker, G. (1994). Neuronal polarity. Annual Review of Neuroscience, 17(1), 267-310. doi: 10.1146/annurev.ne.17.030194.001411 Craige, B., Salazar, G., & Faundez, V. (2008). Phosphatidylinositol-4-Kinase Type II alpha contains an AP-3-sorting motif and a kinase domain that are both required for endosome traffic. Molecular Biology of the Cell, 19(4), 1415-1426. doi: 10.1091/mbc.E07-12-1239 Cross, D. A., Alessi, D. R., Cohen, P., Andjelkovich, M., & Hemmings, B. A. (1995). Inhibition of glycogen synthase kinase-3 by insulin mediated by protein kinase B. Nature, 378(6559), 785-789. doi: 10.1038/378785a0 Cunha, A., Frey, B. N., Andreazza, A. C., Goi, J. D., Rosa, A. R., Gonçalves, C. A., . . . Kapczinski, F. (2006). Serum brain-derived neurotrophic factor is decreased in bipolar disorder during depressive and manic episodes. Neuroscience Letters, 398(3), 215-219. doi: 10.1016/j.neulet.2005.12.085 Daboussi, L., Costaguta, G., & Payne, G. S. (2012). Phosphoinositide-mediated clathrin adaptor progression at the trans-Golgi network. Nature Cell Biology, 14(3), 239- 248. doi: 10.1038/ncb2427 Dajani, R., Fraser, E., Roe, S. M., Young, N., Good, V., Dale, T. C., & Pearl, L. H. (2001). Crystal structure of glycogen synthase kinase 3β: Structural basis for phosphate-primed substrate specificity and autoinhibition. Cell, 105(6), 721-732. doi: 10.1016/S0092-8674(01)00374-9 Dang, C. V. (1999). c-Myc target genes involved in cell growth, apoptosis, and metabolism. Molecular and Cellular Biology, 19(1), 1-11. Dash, P. K., Johnson, D., Clark, J., Orsi, S. A., Zhang, M., Zhao, J., . . . Pati, S. (2011). Involvement of the glycogen synthase kinase-3 signaling pathway in TBI pathology and neurocognitive outcome. PLOS ONE, 6(9), e24648. doi: 10.1371/journal.pone.0024648 Datta, S. R., Brunet, A., & Greenberg, M. E. (1999). Cellular survival: A play in three Akts. Genes & Development, 13(22), 2905-2927. doi: 10.1101/gad.13.22.2905 David, R. (2012). Apoptosis: A lipid trigger of MOMP. Nature Reviews Molecular Cell Biology, 13(4), 208-209. doi: 10.1038/nrm3316 Davidson, G., Wu, W., Shen, J., Bilic, J., Fenger, U., Stannek, P., . . . Niehrs, C. (2005). Casein kinase 1 γ couples Wnt receptor activation to cytoplasmic signal transduction. Nature, 438(7069), 867-872. doi: 10.1038/nature04170 Deng, Y., Xiong, Z., Chen, P., Wei, J., Chen, S., & Yan, Z. (2014). β-Amyloid impairs the regulation of N-methyl-D-aspartate receptors by glycogen synthase kinase 3. Neurobiology of Aging, 35(3), 449-459. doi: 10.1016/j.neurobiolaging.2013.08.031 Dergai, O., Novokhatska, O., Dergai, M., Skrypkina, I., Tsyba, L., Moreau, J., & Rynditch, A. (2010). Intersectin 1 forms complexes with SGIP1 and Reps1 in clathrin-coated pits. Biochemical and Biophysical Research Communications, 402(2), 408-413. doi: 10.1016/j.bbrc.2010.10.045 Di Paolo, G., & De Camilli, P. (2006). Phosphoinositides in cell regulation and membrane dynamics. Nature, 443(7112), 651-657. doi: 10.1038/nature05185 Diamond, D. M., Park, C. R., Campbell, A. M., & Woodson, J. C. (2005). Competitive interactions between endogenous LTD and LTP in the hippocampus underlie the

184 |Page

storage of emotional memories and stress‐induced amnesia. Hippocampus, 15(8), 1006-1025. doi: 10.1002/hipo.20107 Dickey, A., Schleicher, S., Leahy, K., Hu, R., Hallahan, D., & Thotala, D. K. (2011). GSK-3β inhibition promotes cell death, apoptosis, and in vivo tumor growth delay in neuroblastoma Neuro-2A cell line. Journal of Neuro-Oncology, 104(1), 145-153. doi: 10.1007/s11060-010-0491-3 Dickson, E. J., Jensen, J. B., & Hille, B. (2014). Golgi and plasma membrane pools of PI(4)P contribute to plasma membrane PI(4,5)P2 and maintenance of KCNQ2/3 ion channel current. Proceedings of the National Academy of Sciences, 111(22), E2281-E2290. doi: 10.1073/pnas.1407133111 Dietzl, G., Chen, D., Schnorrer, F., Su, K. C., Barinova, Y., Fellner, M., . . . Dickson, B. J. (2007). A genome-wide transgenic RNAi library for conditional gene inactivation in Drosophila. Nature, 448(7150), 151-156. doi: 10.1038/nature05954 Ding, Q., Motoyama, J., Gasca, S., Mo, R., Sasaki, H., Rossant, J., & Hui, C. C. (1998). Diminished Sonic hedgehog signaling and lack of floor plate differentiation in Gli2 mutant mice. Development, 125(14), 2533-2543. Ding, Q., Xia, W., Liu, J.-C., Yang, J.-Y., Lee, D.-F., Xia, J., . . . Li, Z. (2005). Erk associates with and primes GSK-3β for its inactivation resulting in upregulation of β-catenin. Molecular Cell, 19(2), 159-170. doi: 10.1016/j.molcel.2005.06.009 Doble, B. W., Patel, S., Wood, G. A., Kockeritz, L. K., & Woodgett, J. R. (2007). Functional redundancy of GSK-3α and GSK-3β in Wnt/β-catenin signaling shown by using an allelic series of embryonic stem cell lines. Developmental Cell, 12(6), 957-971. doi: 10.1016/j.devcel.2007.04.001 Doble, B. W., & Woodgett, J. R. (2003). GSK-3: Tricks of the trade for a multi-tasking kinase. Journal of Cell Science, 116(7), 1175-1186. doi: 10.1242/jcs.00384 Dokucu, M. E., Yu, L., & Taghert, P. H. (2005). Lithium-and valproate-induced alterations in circadian locomotor behavior in Drosophila. Neuropsychopharmacology, 30(12), 2216-2224. doi: 10.1038/sj.npp.1300764 Dremencov, E., el Mansari, M., & Blier, P. (2009). Brain norepinephrine system as a target for antidepressant and mood stabilizing medications. Current Drug Targets, 10(11), 1061-1068. doi: 10.2174/138945009789735165 Drevets, W. C., Frank, E., Price, J. C., Kupfer, D. J., Holt, D., Greer, P. J., . . . Mathis, C. (1999). PET imaging of serotonin 1A receptor binding in depression. Biological Psychiatry, 46(10), 1375-1387. doi: 10.1016/S0006-3223(99)00189-4 Drevets, W. C., Price, J. L., Simpson, J. R., Jr., Todd, R. D., Reich, T., Vannier, M., & Raichle, M. E. (1997). Subgenual prefrontal cortex abnormalities in mood disorders. Nature, 386(6627), 824-827. doi: 10.1038/386824a0 Drevets, W. C., Thase, M. E., Moses-Kolko, E. L., Price, J., Frank, E., Kupfer, D. J., & Mathis, C. (2007). Serotonin-1A receptor imaging in recurrent depression: replication and literature review. Nuclear Medicine and Biology, 34(7), 865-877. doi: 10.1016/j.nucmedbio.2007.06.008 Du, J., Gray, N. A., Falke, C. A., Chen, W., Yuan, P., Szabo, S. T., . . . Manji, H. K. (2004a). Modulation of synaptic plasticity by antimanic agents: The role of AMPA glutamate receptor subunit 1 synaptic expression. The Journal of Neuroscience, 24(29), 6578-6589. doi: 10.1523/jneurosci.1258-04.2004 Du, J., Quiroz, J., Yuan, P., Zarate, C., & Manji, H. K. (2004b). Bipolar disorder: Involvement of signaling cascades and AMPA receptor trafficking at synapses. Neuron Glia Biology, 1(3), 231-243. doi: 10.1017/S1740925X05000098

185 |Page

Du, J., Wei, Y., Liu, L., Wang, Y., Khairova, R., Blumenthal, R., . . . Drevets, W. (2010). A kinesin signaling complex mediates the ability of GSK-3β to affect mood-associated behaviors. Proceedings of the National Academy of Sciences, 107(25), 11573-11578. doi: 10.1073/pnas.0913138107 Duffy, J. B. (2002). GAL4 system in Drosophila: A fly geneticist's Swiss army knife. Genesis, 34(1‐2), 1-15. doi: 10.1002/gene.10150 Duman-Scheel, M., Weng, L., Xin, S., & Du, W. (2002). Hedgehog regulates cell growth and proliferation by inducing Cyclin D and Cyclin E. Nature, 417(6886), 299-304. doi: 10.1038/417299a Dus, M., Min, S., Keene, A. C., Lee, G. Y., & Suh, G. S. (2011). Taste-independent detection of the caloric content of sugar in Drosophila. Proceedings of the National Academy of Sciences, 108(28), 11644-11649. doi: 10.1073/pnas.1017096108 Dwivedi, Y., Rizavi, H. S., Conley, R. R., Roberts, R. C., Tamminga, C. A., & Pandey, G. N. (2003). Altered gene expression of brain-derived neurotrophic factor and receptor tyrosine kinase B in postmortem brain of suicide subjects. Archives of General Psychiatry, 60(8), 804-815. doi: 10.1001/archpsyc.60.8.804 Eickholt, B. J., Walsh, F. S., & Doherty, P. (2002). An inactive pool of GSK-3 at the leading edge of growth cones is implicated in Semaphorin 3A signaling. The Journal of Cell Biology, 157(2), 211-217. doi: 10.1083/jcb.200201098 Emamian, E. S., Hall, D., Birnbaum, M. J., Karayiorgou, M., & Gogos, J. A. (2004). Convergent evidence for impaired AKT1-GSK3β signaling in schizophrenia. Nature Genetics, 36(2), 131-137. doi: 10.1038/ng1296 Embi, N., Rylatt, D. B., & Cohen, P. (1980). Glycogen synthase kinase-3 from rabbit skeletal muscle. Separation from cyclic-AMP-dependent protein kinase and phosphorylase kinase. European Journal of Biochemistry, 107(2), 519-527. doi: 10.1111/j.1432-1033.1980.tb06059.x Engel, T., Hernández, F., Avila, J., & Lucas, J. J. (2006). Full reversal of Alzheimer's disease-like phenotype in a mouse model with conditional overexpression of glycogen synthase kinase-3. The Journal of Neuroscience, 26(19), 5083-5090. doi: 10.1523/JNEUROSCI.0604-06.2006 Eom, T.-Y., & Jope, R. S. (2009). Blocked inhibitory serine-phosphorylation of glycogen synthase kinase-3α/β impairs in vivo neural precursor cell proliferation. Biological Psychiatry, 66(5), 494-502. doi: 10.1016/j.biopsych.2009.04.015 Espinosa, L., Inglés-Esteve, J., Aguilera, C., & Bigas, A. (2003). Phosphorylation by glycogen synthase kinase-3β down-regulates Notch activity, a link for Notch and Wnt pathways. Journal of Biological Chemistry, 278(34), 32227-32235. doi: 10.1074/jbc.M304001200 Fang, X., Yu, S. X., Lu, Y., Bast, R. C., Jr., Woodgett, J. R., & Mills, G. B. (2000). Phosphorylation and inactivation of glycogen synthase kinase 3 by protein kinase A. Proceedings of the National Academy of Sciences, 97(22), 11960- 11965. doi: 10.1073/pnas.220413597 Farghaian, H., Turnley, A. M., Sutherland, C., & Cole, A. R. (2011). Bioinformatic prediction and confirmation of β-adducin as a novel substrate of glycogen synthase kinase 3. Journal of Biological Chemistry, 286(28), 25274-25283. doi: 10.1074/jbc.M111.251629 Ferreira, M. A., O'Donovan, M. C., Meng, Y. A., Jones, I. R., Ruderfer, D. M., Jones, L., . . . Craddock, N. (2008). Collaborative genome-wide association analysis

186 |Page

supports a role for ANK3 and CACNA1C in bipolar disorder. Nature Genetics, 40(9), 1056-1058. doi: 10.1038/ng.209 Ferrier, I. N., & Thompson, J. M. (2002). Cognitive impairment in bipolar affective disorder: Implications for the bipolar diathesis. The British Journal of Psychiatry, 180(4), 293-295. doi: 10.1192/bjp.180.4.293 Fingerhut, A., von Figura, K., & Höning, S. (2001). Binding of AP2 to sorting signals is modulated by AP2 phosphorylation. Journal of Biological Chemistry, 276(8), 5476-5482. doi: 10.1074/jbc.M009516200 Finn, P. F., & Dice, J. F. (2005). Ketone bodies stimulate chaperone-mediated autophagy. Journal of Biological Chemistry, 280(27), 25864-25870. doi: 10.1074/jbc.M502456200 Fitzgerald, P. J., Barkus, C., Feyder, M., Wiedholz, L. M., Chen, Y.-C., Karlsson, R.- M., . . . Holmes, A. (2010). Does gene deletion of AMPA GluA1 phenocopy features of schizoaffective disorder? Neurobiology of Disease, 40(3), 608-621. doi: 10.1016/j.nbd.2010.08.005 Flotow, H., Graves, P. R., Wang, A. Q., Fiol, C. J., Roeske, R. W., & Roach, P. J. (1990). Phosphate groups as substrate determinants for casein kinase I action. Journal of Biological Chemistry, 265(24), 14264-14269. Foltz, D. R., Santiago, M. C., Berechid, B. E., & Nye, J. S. (2002). Glycogen synthase kinase-3β modulates notch signaling and stability. Current Biology, 12(12), 1006-1011. doi: 10.1016/S0960-9822(02)00888-6 Forlenza, O. V., Diniz, B. S., Radanovic, M., Santos, F. S., Talib, L. L., & Gattaz, W. F. (2011). Disease-modifying properties of long-term lithium treatment for amnestic mild cognitive impairment: randomised controlled trial. The British Journal of Psychiatry, 198(5), 351-356. doi: 10.1192/bjp.bp.110.080044 Fornai, F., Longone, P., Cafaro, L., Kastsiuchenka, O., Ferrucci, M., Manca, M. L., . . . Lenzi, P. (2008). Lithium delays progression of amyotrophic lateral sclerosis. Proceedings of the National Academy of Sciences, 105(6), 2052-2057. doi: 10.1073/pnas.0708022105 Fournier, J. B., Dommersnes, P. G., & Galatola, P. (2003). Dynamin recruitment by clathrin coats: A physical step? Comptes Rendus Biologies, 326(5), 467-476. doi: 10.1016/S1631-0691(03)00096-9 Frame, S., Cohen, P., & Biondi, R. M. (2001). A common phosphate binding site explains the unique substrate specificity of GSK3 and its inactivation by phosphorylation. Molecular Cell, 7(6), 1321-1327. doi: 10.1016/S1097- 2765(01)00253-2 Freeman, M. (1996). Reiterative use of the EGF receptor triggers differentiation of all cell types in the Drosophila eye. Cell, 87(4), 651-660. doi: 10.1016/S0092- 8674(00)81385-9 Fukata, Y., Itoh, T. J., Kimura, T., Ménager, C., Nishimura, T., Shiromizu, T., . . . Hotani, H. (2002). CRMP-2 binds to tubulin heterodimers to promote microtubule assembly. Nature Cell Biology, 4(8), 583-591. doi: 10.1038/ncb825 Fukumoto, T., Morinobu, S., Okamoto, Y., Kagaya, A., & Yamawaki, S. (2001). Chronic lithium treatment increases the expression of brain-derived neurotrophic factor in the rat brain. Psychopharmacology, 158(1), 100-106. doi: 10.1007/s002130100871 Fuster-Matanzo, A., Llorens-Martín, M., Sirerol-Piquer, M. S., García-Verdugo, J. M., Avila, J., & Hernández, F. (2013). Dual effects of increased glycogen synthase kinase-3β activity on adult neurogenesis. Human Molecular Genetics, 22(7), 1300-1315. doi: 10.1093/hmg/dds53

187 |Page

Garfield, D. A., & Wray, G. A. (2010). The evolution of gene regulatory interactions. BioScience, 60(1), 15-23. doi: 10.1525/bio.2010.60.1.6 Geller, B., Williams, M., Zimerman, B., Frazier, J., Beringer, L., & Warner, K. L. (1998). Prepubertal and early adolescent bipolarity differentiate from ADHD by manic symptoms, grandiose delusions, ultra-rapid or ultradian cycling. Journal of Affective Disorders, 51(2), 81-91. doi: 10.1016/S0165-0327(98)00175-X Ghavami, S., Shojaei, S., Yeganeh, B., Ande, S. R., Jangamreddy, J. R., Mehrpour, M., . . . Kashani, H. H. (2014). Autophagy and apoptosis dysfunction in neurodegenerative disorders. Progress in Neurobiology, 112, 24-49. doi: 10.1016/j.pneurobio.2013.10.004 Giese, K. P. (2009). GSK‐3: A key player in neurodegeneration and memory. IUBMB life, 61(5), 516-521. doi: 10.1002/iub.187 Goes, F. S., Rongione, M., Chen, Y. C., Karchin, R., Elhaik, E., & Potash, J. B. (2011). Exonic DNA sequencing of ERBB4 in bipolar disorder. PLOS ONE, 6(5), e20242. doi: 10.1371/journal.pone.0020242 Gonzalez‐Billault, C., Jimenez‐Mateos, E. M., Caceres, A., Diaz‐Nido, J., Wandosell, F., & Avila, J. (2004). Microtubule‐associated protein 1B function during normal development, regeneration, and pathological conditions in the nervous system. Journal of Neurobiology, 58(1), 48-59. doi: 10.1002/neu.10283 Goode, N., Hughes, K., Woodgett, J., & Parker, P. (1992). Differential regulation of glycogen synthase kinase-3 beta by protein kinase C isotypes. Journal of Biological Chemistry, 267(24), 16878-16882. Goodman, O. B., & Keen, J. H. (1995). The α chain of the AP-2 adaptor is a clathrin binding subunit. Journal of Biological Chemistry, 270(40), 23768-23773. doi: 10.1074/jbc.270.40.23768 Goold, R. G., & Gordon-Weeks, P. R. (2004). Glycogen synthase kinase 3beta and the regulation of axon growth. Biochemical Society Transactions, 32(5), 809-811. doi: 10.1042/BST0320809 Goold, R. G., & Gordon-Weeks, P. R. (2005). The MAP kinase pathway is upstream of the activation of GSK3β that enables it to phosphorylate MAP1B and contributes to the stimulation of axon growth. Molecular and Cellular Neuroscience, 28(3), 524-534. doi: 10.1016/j.mcn.2004.11.005 Goold, R. G., Owen, R., & Gordon-Weeks, P. R. (1999). Glycogen synthase kinase 3beta phosphorylation of microtubule-associated protein 1B regulates the stability of microtubules in growth cones. Journal of Cell Science, 112(19), 3373-3384. Gordon‐Weeks, P. R. (2004). Microtubules and growth cone function. Journal of Neurobiology, 58(1), 70-83. doi: 10.1002/neu.10266 Gorwood, P. (2008). Neurobiological mechanisms of anhedonia. Dialogues in Clinical Neuroscience, 10(3), 291-299. Gould, T. D., Einat, H., Bhat, R., & Manji, H. K. (2004). AR-A014418, a selective GSK-3 inhibitor, produces antidepressant-like effects in the forced swim test. The International Journal of Neuropsychopharmacology, 7(4), 387-390. doi: 10.1017/S1461145704004535 Graham, T. R., & Burd, C. G. (2011). Coordination of Golgi functions by phosphatidylinositol 4-kinases. Trends in Cell Biology, 21(2), 113-121. doi: 10.1016/j.tcb.2010.10.002 Greene, J. C., Whitworth, A. J., Kuo, I., Andrews, L. A., Feany, M. B., & Pallanck, L. J. (2003). Mitochondrial pathology and apoptotic muscle degeneration in

188 |Page

Drosophila parkin mutants. Proceedings of the National Academy of Sciences, 100(7), 4078-4083. doi: 10.1073/pnas.0737556100 Grimes, C. A., & Jope, R. S. (2001a). CREB DNA binding activity is inhibited by glycogen synthase kinase‐3β and facilitated by lithium. Journal of Neurochemistry, 78(6), 1219-1232. doi: 10.1046/j.1471-4159.2001.00495.x Grimes, C. A., & Jope, R. S. (2001b). The multifaceted roles of glycogen synthase kinase 3β in cellular signaling. Progress in Neurobiology, 65(4), 391-426. doi: 10.1016/S0301-0082(01)00011-9 Guha, S., Cullen, J. P., Morrow, D., Colombo, A., Lally, C., Walls, D., . . . Cahill, P. A. (2011). Glycogen synthase kinase 3 beta positively regulates Notch signaling in vascular smooth muscle cells: Role in cell proliferation and survival. Basic Research in Cardiology, 106(5), 773-785. doi: 10.1007/s00395-011-0189-5 Guo, J., Wenk, M. R., Pellegrini, L., Onofri, F., Benfenati, F., & De Camilli, P. (2003). Phosphatidylinositol 4-kinase type IIα is responsible for the phosphatidylinositol 4-kinase activity associated with synaptic vesicles. Proceedings of the National Academy of Sciences, 100(7), 3995-4000. doi: 10.1073/pnas.0230488100 Gupta-Rossi, N., Ortica, S., Meas-Yedid, V., Heuss, S., Moretti, J., Olivo-Marin, J.-C., & Israël, A. (2011). The adaptor-associated kinase 1, AAK1, is a positive regulator of the Notch pathway. Journal of Biological Chemistry, 286(21), 18720-18730. doi: 10.1074/jbc.M110.190769 Hafeman, D. M., Bebko, G., Bertocci, M. A., Fournier, J. C., Bonar, L., Perlman, S. B., . . . Phillips, M. L. (2014). Abnormal deactivation of the inferior frontal gyrus during implicit emotion processing in youth with bipolar disorder: Attenuated by medication. Journal of Psychiatric Research, 58, 129-136. doi: 10.1016/j.jpsychires.2014.07.023 Haffner, C., Takei, K., Chen, H., Ringstad, N., Hudson, A., Butler, M. H., . . . De Camilli, P. (1997). Synaptojanin 1: Localization on coated endocytic intermediates in nerve terminals and interaction of its 170 kDa isoform with Eps15. FEBS Letters, 419(2-3), 175-180. doi: 10.1016/S0014-5793(97)01451-8 Hagen, T., Di Daniel, E., Culbert, A. A., & Reith, A. D. (2002). Expression and characterization of GSK-3 mutants and their effect on β-catenin phosphorylation in intact cells. Journal of Biological Chemistry, 277(26), 23330-23335. doi: 10.1074/jbc.M201364200 Hagen, T., & Vidal-Puig, A. (2002). Characterisation of the phosphorylation of β- catenin at the GSK-3 priming site Ser45. Biochemical and Biophysical Research Communications, 294(2), 324-328. doi: 10.1016/S0006-291X(02)00485-0 Hall, A. C., Brennan, A., Goold, R. G., Cleverley, K., Lucas, F. R., Gordon-Weeks, P. R., & Salinas, P. C. (2002). Valproate regulates GSK-3-mediated axonal remodeling and synapsin I clustering in developing neurons. Molecular and Cellular Neuroscience, 20(2), 257-270. doi: 10.1006/mcne.2002.1117 Hall, A. C., Lucas, F. R., & Salinas, P. C. (2000). Axonal remodeling and synaptic differentiation in the cerebellum is regulated by WNT-7a signaling. Cell, 100(5), 525-535. doi: 10.1016/S0092-8674(00)80689-3 Hall, J., Guo, G., Wray, J., Eyres, I., Nichols, J., Grotewold, L., . . . Walker, R. (2009). Oct4 and LIF/Stat3 additively induce Krüppel factors to sustain embryonic stem cell self-renewal. Cell Stem Cell, 5(6), 597-609. doi: 10.1016/j.stem.2009.11.003 Hallcher, L. M., & Sherman, W. R. (1980). The effects of lithium ion and other agents on the activity of myo-inositol-1-phosphatase from bovine brain. Journal of Biological Chemistry, 255(22), 10896-10901.

189 |Page

Hanks, S. K., & Quinn, A. M. (1991). Protein kinase catalytic domain sequence database: Identification of conserved features of primary structure and classification of family members. Methods in Enzymology, 200, 38-62. doi: 10.1016/0076-6879(91)00126-h Hanley, J. (2010a). Endosomal sorting of AMPA receptors in hippocampal neurons. Biochemical Society Transactions, 38(2), 460-465. Hanley, J. G. (2010b). Endosomal sorting of AMPA receptors in hippocampal neurons. Biochemical Society Transactions, 38(2), 460-465. doi: 10.1042/bst0380460 Hannan, L. A., Newmyer, S. L., & Schmid, S. L. (1998). ATP- and cytosol-dependent release of adaptor proteins from clathrin-coated vesicles: A dual role for Hsc70. Molecular Biology of the Cell, 9(8), 2217-2229. doi: 10.1091/mbc.9.8.2217 Hao, W., Luo, Z., Zheng, L., Prasad, K., & Lafer, E. M. (1999). AP180 and AP-2 interact directly in a complex that cooperatively assembles clathrin. Journal of Biological Chemistry, 274(32), 22785-22794. doi: 10.1074/jbc.274.32.22785 Harada, Y., Sakai, M., Kurabayashi, N., Hirota, T., & Fukada, Y. (2005). Ser-557- phosphorylated mCRY2 is degraded upon synergistic phosphorylation by glycogen synthase kinase-3β. Journal of Biological Chemistry, 280(36), 31714- 31721. doi: 10.1074/jbc.M506225200 Harrison, P. W., Wright, A. E., & Mank, J. E. (2012). The evolution of gene expression and the transcriptome-phenotype relationship. Seminars in Cell & Developmental Biology, 23(2), 222-229. doi: 10.1016/j.semcdb.2011.12.004 Harvey, A. G., Mullin, B. C., & Hinshaw, S. P. (2006). Sleep and circadian rhythms in children and adolescents with bipolar disorder. Development and Psychopathology, 18(04), 1147-1168. doi: 10.1017/S095457940606055X Hashimoto, R., Hough, C., Nakazawa, T., Yamamoto, T., & Chuang, D. M. (2002). Lithium protection against glutamate excitotoxicity in rat cerebral cortical neurons: Involvement of NMDA receptor inhibition possibly by decreasing NR2B tyrosine phosphorylation. Journal of Neurochemistry, 80(4), 589-597. doi: 10.1046/j.0022-3042.2001.00728.x He, T.-C., Sparks, A. B., Rago, C., Hermeking, H., Zawel, L., da Costa, L. T., . . . Kinzler, K. W. (1998). Identification of c-MYC as a target of the APC pathway. Science, 281(5382), 1509-1512. doi: 10.1126/science.281.5382.1509 Henderson, D. M., & Conner, S. D. (2007). A novel AAK1 splice variant functions at multiple steps of the endocytic pathway. Molecular Biology of the Cell, 18(7), 2698-2706. doi: 10.1091/mbc.E06-09-0831 Hernández, F., Borrell, J., Guaza, C., Avila, J., & Lucas, J. J. (2002). Spatial learning deficit in transgenic mice that conditionally over‐express GSK‐3β in the brain but do not form tau filaments. Journal of Neurochemistry, 83(6), 1529-1533. doi: 10.1046/j.1471-4159.2002.01269.x Hernandez, F., Lucas, J. J., & Avila, J. (2013). GSK3 and tau: Two convergence points in Alzheimer's disease. Journal of Alzheimer's Disease, 33, 141-144. doi: 10.3233/JAD-2012-129025 Hilfiker, S., Pieribone, V. A., Czernik, A. J., Kao, H.-T., Augustine, G. J., & Greengard, P. (1999). Synapsins as regulators of neurotransmitter release. Philosophical Transactions of the Royal Society B: Biological Sciences, 354(1381), 269-279. doi: 10.1098/rstb.1999.0378 Hlobilková, A., Knillová, J., Bártek, J., Lukas, J., & Kolar, Z. (2003). The mechanism of action of the tumour suppressor gene PTEN. Biomedical Papers of the Medical Faculty of the University Palacky, Olomouc, Czechoslovakia, 147(1), 19-25. doi: 10.5507/bp.2003.003

190 |Page

Hoeck, J. D., Jandke, A., Blake, S. M., Nye, E., Spencer-Dene, B., Brandner, S., & Behrens, A. (2010). Fbw7 controls neural stem cell differentiation and progenitor apoptosis via Notch and c-Jun. Nature Neuroscience, 13(11), 1365- 1372. doi: 10.1038/nn.2644 Hoeflich, K. P., Luo, J., Rubie, E. A., Tsao, M.-S., Jin, O., & Woodgett, J. R. (2000). Requirement for glycogen synthase kinase-3β in cell survival and NF-κB activation. Nature, 406(6791), 86-90. doi: 10.1038/35017574 Holley, S. M., Wang, E. A., Cepeda, C., Jentsch, J. D., Ross, C. A., Pletnikov, M. V., & Levine, M. S. (2013). Frontal cortical synaptic communication is abnormal in Disc1 genetic mouse models of schizophrenia. Schizophrenia Research, 4(13), 00099-00096. doi: 10.1016/j.schres.2013.02.007 Hong, M., Chen, D. C. R., Klein, P. S., & Lee, V. M. Y. (1997). Lithium reduces tau phosphorylation by inhibition of glycogen synthase kinase-3. Journal of Biological Chemistry, 272(40), 25326-25332. doi: 10.1074/jbc.272.40.25326 Hooper, C., Markevich, V., Plattner, F., Killick, R., Schofield, E., Engel, T., . . . Bliss, T. (2007). Glycogen synthase kinase‐3 inhibition is integral to long‐term potentiation. European Journal of Neuroscience, 25(1), 81-86. doi: 10.1111/j.1460-9568.2006.05245.x Horii, T., Arai, Y., Yamazaki, M., Morita, S., Kimura, M., Itoh, M., . . . Hatada, I. (2014). Validation of microinjection methods for generating knockout mice by CRISPR/Cas-mediated genome engineering. Scientific Reports, 4, 4513. doi: 10.1038/srep04513 Hu, H., Real, E., Takamiya, K., Kang, M. G., Ledoux, J., Huganir, R. L., & Malinow, R. (2007). Emotion enhances learning via norepinephrine regulation of AMPA- receptor trafficking. Cell, 131(1), 160-173. doi: S0092-8674(07)01205-6 [pii] 10.1016/j.cell.2007.09.017 Huang, J., Zhang, Y., Bersenev, A., O’Brien, W. T., Tong, W., Emerson, S. G., & Klein, P. S. (2009). Pivotal role for glycogen synthase kinase–3 in hematopoietic stem cell homeostasis in mice. The Journal of Clinical Investigation, 119(12), 3519-3529. doi: 10.1172/JCI40572 Huber, O., Korn, R., McLaughlin, J., Ohsugi, M., Herrmann, B. G., & Kemler, R. (1996). Nuclear localization of β-catenin by interaction with transcription factor LEF-1. Mechanisms of Development, 59(1), 3-10. doi: 10.1016/0925- 4773(96)00597-7 Huber, R., Hill, S. L., Holladay, C., Biesiadecki, M., Tononi, G., & Cirelli, C. (2004). Sleep homeostasis in Drosophila melanogaster. Sleep, 27(4), 628-639. Hughes, K., Nikolakaki, E., Plyte, S. E., Totty, N. F., & Woodgett, J. R. (1993). Modulation of the glycogen synthase kinase-3 family by tyrosine phosphorylation. The EMBO Journal, 12(2), 803. Hur, E.-M., Lee, B. D., Kim, S.-J., Xu, W.-L., & Zhou, F.-Q. (2011). GSK3 controls axon growth via CLASP-mediated regulation of growth cone microtubules. Genes & Development, 25(18), 1968-1981. doi: 10.1101/gad.17015911 Hur, E.-M., & Zhou, F.-Q. (2010). GSK3 signalling in neural development. Nature Reviews Neuroscience, 11(8), 539-551. doi: 10.1038/nrn2870 Ingham, P. W., & Placzek, M. (2006). Orchestrating ontogenesis: Variations on a theme by sonic hedgehog. Nature Reviews Genetics, 7(11), 841-850. doi: 10.1038/nrg1969 Inoki, K., Ouyang, H., Zhu, T., Lindvall, C., Wang, Y., Zhang, X., . . . Stankunas, K. (2006). TSC2 integrates Wnt and energy signals via a coordinated

191 |Page

phosphorylation by AMPK and GSK3 to regulate cell growth. Cell, 126(5), 955- 968. doi: 10.1016/j.cell.2006.06.055 Ishiguro, K., Shiratsuchi, A., Sato, S., Omori, A., Arioka, M., Kobayashi, S., . . . Imahori, K. (1993). Glycogen synthase kinase 3β is identical to tau protein kinase I generating several epitopes of paired helical filaments. FEBS Letters, 325(3), 167-172. doi: 10.1016/0014-5793(93)81066-9 Itoh, K., Tang, T. L., Neel, B. G., & Sokol, S. Y. (1995). Specific modulation of ectodermal cell fates in Xenopus embryos by glycogen synthase kinase. Development, 121(12), 3979-3988. Ivanova, P. T., Milne, S. B., Myers, D. S., & Brown, H. A. (2009). Lipidomics: A mass spectrometry based systems level analysis of cellular lipids. Current Opinion in Chemical Biology, 13(5-6), 526-531. doi: 10.1016/j.cbpa.2009.08.011 Jaaro-Peled, H., Hayashi-Takagi, A., Seshadri, S., Kamiya, A., Brandon, N. J., & Sawa, A. (2009). Neurodevelopmental mechanisms of schizophrenia: understanding disturbed postnatal brain maturation through neuregulin-1–ErbB4 and DISC1. Trends in Neurosciences, 32(9), 485-495. doi: 10.1016/j.tins.2009.05.007 Jackson, A. P., Flett, A., Smythe, C., Hufton, L., Wettey, F. R., & Smythe, E. (2003). Clathrin promotes incorporation of cargo into coated pits by activation of the AP2 adaptor μ2 kinase. The Journal of Cell Biology, 163(2), 231-236. doi: 10.1083/jcb.200304079 Jackson, G. R., Wiedau-Pazos, M., Sang, T.-K., Wagle, N., Brown, C. A., Massachi, S., & Geschwind, D. H. (2002). Human Wild-Type Tau Interacts with wingless Pathway Components and Produces Neurofibrillary Pathology in Drosophila. Neuron, 34(4), 509-519. doi: 10.1016/S0896-6273(02)00706-7 Jha, A., Agostinelli, N. R., Mishra, S. K., Keyel, P. A., Hawryluk, M. J., & Traub, L. M. (2004). A novel AP-2 adaptor interaction motif initially identified in the long- splice isoform of synaptojanin 1, SJ170. Journal of Biological Chemistry, 279(3), 2281-2290. doi: 10.1074/jbc.M305644200 Jia, J., Amanai, K., Wang, G., Tang, J., Wang, B., & Jiang, J. (2002). Shaggy/GSK3 antagonizes Hedgehog signalling by regulating Cubitus interruptus. Nature, 416(6880), 548-552. doi: 10.1038/nature733 Jiang, H., Guo, W., Liang, X., & Rao, Y. (2005). Both the establishment and the maintenance of neuronal polarity require active mechanisms: Critical roles of GSK-3β and its upstream regulators. Cell, 120(1), 123-135. doi: 10.1016/j.cell.2004.12.033 Jiang, J., Chan, Y.-S., Loh, Y.-H., Cai, J., Tong, G.-Q., Lim, C.-A., . . . Ng, H.-H. (2008). A core Klf circuitry regulates self-renewal of embryonic stem cells. Nature Cell Biology, 10(3), 353-360. doi: 10.1038/ncb1698 Jiang, P., & Mizushima, N. (2014). Autophagy and human diseases. Cell Research, 24(1), 69-79. doi: 10.1038/cr.2013.161 Jin, K., Zhu, Y., Sun, Y., Mao, X. O., Xie, L., & Greenberg, D. A. (2002). Vascular endothelial growth factor (VEGF) stimulates neurogenesis in vitro and in vivo. Proceedings of the National Academy of Sciences, 99(18), 11946-11950. doi: 10.1073/pnas.182296499 Jin, Y. H., Kim, H., Ki, H., Yang, I., Yang, N., Lee, K. Y., . . . Kim, K. (2009). Beta- catenin modulates the level and transcriptional activity of Notch1/NICD through its direct interaction. Biochimica et Biophysica Acta - Molecular Cell Research, 1793(2), 290-299. doi: 10.1016/j.bbamcr.2008.10.002 Johnson, V., Volikos, E., Halford, S. E., Sadat, E. T. E., Popat, S., Talbot, I., . . . Houlston, R. (2005). Exon 3 β-catenin mutations are specifically associated with

192 |Page

colorectal carcinomas in hereditary non-polyposis colorectal cancer syndrome. Gut, 54(2), 264-267. doi: 10.1136/gut.2004.048132 Jope, R. S. (2003). Lithium and GSK-3: One inhibitor, two inhibitory actions, multiple outcomes. Trends in Pharmacological Sciences, 24(9), 441-443. doi: 10.1016/S0165-6147(03)00206-2 Jope, R. S. (2011). Glycogen synthase kinase-3 in the etiology and treatment of mood disorders. Frontiers in Molecular Neuroscience, 4, 16. doi: 10.3389/fnmol.2011.00016 Jope, R. S., & Roh, M. S. (2006). Glycogen synthase kinase-3 (GSK3) in psychiatric diseases and therapeutic interventions. Current Drug Targets, 7(11), 1421-1434. doi: 10.2174/1389450110607011421 Jović, M., Kean, M. J., Dubankova, A., Boura, E., Gingras, A.-C., Brill, J. A., & Balla, T. (2014). Endosomal sorting of VAMP3 is regulated by PI4K2A. Journal of Cell Science, 127(17), 3745-3756. doi: 10.1242/jcs.148809 Jović, M., Kean, M. J., Szentpetery, Z., Polevoy, G., Gingras, A.-C., Brill, J. A., & Balla, T. (2012). Two phosphatidylinositol 4-kinases control lysosomal delivery of the Gaucher disease enzyme, β-glucocerebrosidase. Molecular Biology of the Cell, 23(8), 1533-1545. doi: 10.1091/mbc.E11-06-0553 Jung, G., Barylko, B., Lu, D., Shu, H., Yin, H., & Albanesi, J. P. (2011). Stabilization of phosphatidylinositol 4-kinase type IIβ by interaction with Hsp90. Journal of Biological Chemistry, 286(14), 12775-12784. doi: 10.1074/jbc.M110.178616 Jung, G., Wang, J., Wlodarski, P., Barylko, B., Binns, D., Shu, H., . . . Albanesi, J. (2008). Molecular determinants of activation and membrane targeting of phosphoinositol 4-kinase IIbeta. Biochemical Journal, 409, 501-509. doi: 10.1042/BJ20070821 Ka, M., Condorelli, G., Woodgett, J. R., & Kim, W.-Y. (2014). mTOR regulates brain morphogenesis by mediating GSK3 signaling. Development, 141(21), 4076- 4086. doi: 10.1242/dev.108282 Kaidanovich-Beilin, O., Lipina, T. V., Takao, K., van Eede, M., Hattori, S., Laliberté, C., . . . Fletcher, P. J. (2009). Abnormalities in brain structure and behavior in GSK-3alpha mutant mice. Molecular Brain, 2(1), 35. doi: 10.1186/1756-6606-2- 35 Kalinichev, M., & Dawson, L. A. (2011). Evidence for antimanic efficacy of glycogen synthase kinase-3 (GSK3) inhibitors in a strain-specific model of acute mania. The International Journal of Neuropsychopharmacology, 14(8), 1051-1067. doi: 10.1017/S1461145710001495 Kang, U. G., Seo, M. S., Roh, M.-S., Kim, Y., Yoon, S. C., & Kim, Y. S. (2004). The effects of clozapine on the GSK-3-mediated signaling pathway. FEBS letters, 560(1-3), 115-119. doi: 10.1016/S0014-5793(04)00082-1 Kapur, S., Zipursky, R. B., & Remington, G. (1999). Clinical and theoretical implications of 5-HT2 and D2 receptor occupancy of clozapine, risperidone, and olanzapine in schizophrenia. American Journal of Psychiatry, 156(2), 286-293. doi: 10.1176/ajp.156.2.286 Karege, F., Perroud, N., Burkhardt, S., Fernandez, R., Ballmann, E., La Harpe, R., & Malafosse, A. (2011). Alterations in phosphatidylinositol 3-kinase activity and PTEN phosphatase in the prefrontal cortex of depressed suicide victims. Neuropsychobiology, 63(4), 224-231. doi: 10.1159/000322145 Karege, F., Perroud, N., Schürhoff, F., Meary, A., Marillier, G., Burkhardt, S., . . . Leboyer, M. (2010). Association of AKT1 gene variants and protein expression

193 |Page

in both schizophrenia and bipolar disorder. Genes, Brain and Behavior, 9(5), 503-511. doi: 10.1111/j.1601-183X.2010.00578.x Kelly, K. F., Ng, D. Y., Jayakumaran, G., Wood, G. A., Koide, H., & Doble, B. W. (2011). β-catenin enhances Oct-4 activity and reinforces pluripotency through a TCF-independent mechanism. Cell Stem Cell, 8(2), 214-227. doi: 10.1016/j.stem.2010.12.010 Kelmendi, B., Holsbach-Beltrame, M., McIntosh, A. M., Hilt, L., George, E. D., Kitchen, R. R., . . . Simen, A. A. (2011). Association of polymorphisms in HCN4 with mood disorders and obsessive compulsive disorder. Neuroscience Letters, 496(3), 195-199. doi: 10.1016/j.neulet.2011.04.026 Kennelly, P. J., & Krebs, E. G. (1991). Consensus sequences as substrate specificity determinants for protein kinases and protein phosphatases. Journal of Biological Chemistry, 266(15), 15555-15558. Kenney, A. M., & Rowitch, D. H. (2000). Sonic hedgehog promotes G(1) cyclin expression and sustained cell cycle progression in mammalian neuronal precursors. Molecular and Cellular Biology, 20(23), 9055-9067. doi: 10.1128/MCB.20.23.9055-9067.2000 Kerchner, G. A., & Nicoll, R. A. (2008). Silent synapses and the emergence of a postsynaptic mechanism for LTP. Nature Reviews Neuroscience, 9(11), 813- 825. doi: 10.1038/nrn2501 Kesidou, E., Lagoudaki, R., Touloumi, O., Poulatsidou, K.-N., & Simeonidou, C. (2013). Autophagy and neurodegenerative disorders. Neural Regeneration Research, 8(24), 2275. doi: 10.3969/j.issn.1673-5374.2013.24.007 Kessels, H. W., & Malinow, R. (2009). Synaptic AMPA receptor plasticity and behavior. Neuron, 61(3), 340-350. doi: 10.1016/j.neuron.2009.01.015 Kielkowska, A., Niewczas, I., Anderson, K. E., Durrant, T. N., Clark, J., Stephens, L. R., & Hawkins, P. T. (2014). A new approach to measuring phosphoinositides in cells by mass spectrometry. Advances in Biological Regulation, 54, 131-141. doi: 10.1016/j.jbior.2013.09.001 Kim, C. G., Chung, I.-Y., Lim, Y., Lee, Y. H., & Shin, S. Y. (2011). A Tcf/Lef element within the enhancer region of the human ANOG gene plays a role in promoter activation. Biochemical and Biophysical Research Communications, 410(3), 637-642. doi: 10.1016/j.bbrc.2011.06.044 Kim, J. Y., Duan, X., Liu, C. Y., Jang, M.-H., Guo, J. U., Pow-anpongkul, N., . . . Ming, G.-l. (2009a). DISC1 regulates new neuron development in the adult brain via modulation of AKT-mTOR signaling through KIAA1212. Neuron, 63(6), 761- 773. doi: 10.1016/j.neuron.2009.08.008 Kim, L., Liu, J., & Kimmel, A. R. (1999). The novel tyrosine kinase ZAK1 activates GSK3 to direct cell fate specification. Cell, 99(4), 399-408. doi: 10.1016/S0092- 8674(00)81526-3 Kim, W.-Y., Wang, X., Wu, Y., Doble, B. W., Patel, S., Woodgett, J. R., & Snider, W. D. (2009b). GSK-3 is a master regulator of neural progenitor homeostasis. Nature Neuroscience, 12(11), 1390-1397. doi: 10.1038/nn.2408 Kim, W.-Y., Zhou, F.-Q., Zhou, J., Yokota, Y., Wang, Y.-M., Yoshimura, T., . . . Snider, W. D. (2006). Essential roles for GSK-3s and GSK-3-primed substrates in neurotrophin-induced and hippocampal axon growth. Neuron, 52(6), 981-996. doi: 10.1016/j.neuron.2006.10.031 Kim, W. Y., Wang, X., Wu, Y., Doble, B. W., Patel, S., Woodgett, J. R., & Snider, W. D. (2009c). GSK-3 is a master regulator of neural progenitor homeostasis. Nat Neurosci, 12(11), 1390-1397. doi: 10.1038/nn.2408

194 |Page

Kimura, T., Yamashita, S., Nakao, S., Park, J.-M., Murayama, M., Mizoroki, T., . . . Takashima, A. (2008). GSK-3β is required for memory reconsolidation in adult brain. PLOS ONE, 3(10), e3540. doi: 10.1371/journal.pone.0003540 King, D. P., & Takahashi, J. S. (2000). Molecular genetics of circadian rhythms in mammals. Annual Review of Neuroscience, 23(1), 713-742. doi: 10.1146/annurev.neuro.23.1.713 King, J. S., Teo, R., Ryves, J., Reddy, J. V., Peters, O., Orabi, B., . . . Harwood, A. J. (2009). The mood stabiliser lithium suppresses PIP3 signalling in Dictyostelium and human cells. Disease Models & Mechanisms, 2(5-6), 306-312. doi: 10.1242/dmm.001271 Kirchhausen, T. (1999). Adaptors for clathrin-mediated traffic. Annual Review of Cell and Developmental Biology, 15, 705-732. doi: 10.1146/annurev.cellbio.15.1.705 Kirchhausen, T. (2000). Clathrin. Annual Review of Biochemistry, 69, 699-727. doi: 10.1146/annurev.biochem.69.1.699 Klein, P. S., & Melton, D. A. (1996). A molecular mechanism for the effect of lithium on development. Proceedings of the National Academy of Sciences, 93(16), 8455-8459. doi: 10.1073/pnas.93.16.8455 Klionsky, D. J., & Emr, S. D. (2000). Autophagy as a regulated pathway of cellular degradation. Science, 290(5497), 1717-1721. doi: 10.1126/science.290.5497.1717 Knoepfler, P. S., Cheng, P. F., & Eisenman, R. N. (2002). N-myc is essential during neurogenesis for the rapid expansion of progenitor cell populations and the inhibition of neuronal differentiation. Genes & Development, 16(20), 2699- 2712. doi: 10.1101/gad.1021202 Knoepfler, P. S., & Kenney, A. M. (2006). Neural precursor cycling at sonic speed: N- Myc pedals, GSK-3 brakes. Cell Cycle, 5(1), 47-52. doi: 10.4161/cc.5.1.2292 Ko, H. W., Kim, E. Y., Chiu, J., Vanselow, J. T., Kramer, A., & Edery, I. (2010). A hierarchical phosphorylation cascade that regulates the timing of PERIOD nuclear entry reveals novel roles for proline-directed kinases and GSK-3β/SGG in circadian clocks. The Journal of Neuroscience, 30(38), 12664-12675. doi: 10.1523/JNEUROSCI.1586-10.2010 Komatsu, M., Waguri, S., Chiba, T., Murata, S., Iwata, J., Tanida, I., . . . Tanaka, K. (2006). Loss of autophagy in the central nervous system causes neurodegeneration in mice. Nature, 441(7095), 880-884. doi: 10.1038/nature04723 Koob, G. F. (1996). Hedonic valence, dopamine and motivation. Molecular Psychiatry, 1(3), 186-189. Koontz, J. I., Soreng, A. L., Nucci, M., Kuo, F. C., Pauwels, P., van Den Berghe, H., . . . Sklar, J. (2001). Frequent fusion of the JAZF1 and JJAZ1 genes in endometrial stromal tumors. Proceedings of the National Academy of Sciences, 98(11), 6348- 6353. doi: 10.1073/pnas.101132598 Kopan, R., & Ilagan, M. X. (2009). The canonical Notch signaling pathway: Unfolding the activation mechanism. Cell, 137(2), 216-233. doi: 10.1016/j.cell.2009.03.045 Korolchuk, V. I., & Banting, G. (2002). CK2 and GAK/auxilin2 are major protein kinases in clathrin-coated vesicles. Traffic, 3(6), 428-439. doi: 10.1034/j.1600- 0854.2002.30606.x Krauß, M., & Haucke, V. (2007). Phosphoinositide-metabolizing enzymes at the interface between membrane traffic and cell signalling. EMBO Reports, 8(3), 241-246. doi: 10.1038/sj.embor.7400919

195 |Page

Krebs, C. E., Karkheiran, S., Powell, J. C., Cao, M., Makarov, V., Darvish, H., . . . Paisan-Ruiz, C. (2013). The Sac1 domain of SYNJ1 identified mutated in a family with early-onset progressive Parkinsonism with generalized seizures. Human Mutation, 34(9), 1200-1207. doi: 10.1002/humu.22372 Kruse, J.-P., & Gu, W. (2009). Modes of p53 regulation. Cell, 137(4), 609-622. doi: 10.1016/j.cell.2009.04.050 Kuai, L., Ong, S.-E., Madison, J. M., Wang, X., Duvall, J. R., Lewis, T. A., . . . Wood, J. L. (2011). AAK1 identified as an inhibitor of neuregulin-1/ErbB4-dependent neurotrophic factor signaling using integrative chemical genomics and proteomics. Chemistry & Biology, 18(7), 891-906. doi: 10.1016/j.chembiol.2011.03.017 Kühl, M., & Wedlich, D. (1997). Wnt signalling goes nuclear. Bioessays, 19(2), 101- 104. doi: 10.1002/bies.950190204 Kulikov, R., Boehme, K. A., & Blattner, C. (2005). Glycogen synthase kinase 3- dependent phosphorylation of Mdm2 regulates p53 abundance. Molecular and Cellular Biology, 25(16), 7170-7180. doi: 10.1128/MCB.25.16.7170-7180.2005 Kumar, P., Lyle, K. S., Gierke, S., Matov, A., Danuser, G., & Wittmann, T. (2009). GSK3β phosphorylation modulates CLASP–microtubule association and lamella microtubule attachment. The Journal of Cell Biology, 184(6), 895-908. doi: 10.1083/jcb.200901042 Kuwabara, T., Hsieh, J., Muotri, A., Yeo, G., Warashina, M., Lie, D. C., . . . Gage, F. H. (2009). Wnt-mediated activation of NeuroD1 and retro-elements during adult neurogenesis. Nature Neuroscience, 12(9), 1097-1105. doi: 10.1038/nn.2360 Lachman, H. M., Pedrosa, E., Petruolo, O. A., Cockerham, M., Papolos, A., Novak, T., . . . Stopkova, P. (2007). Increase in GSK3B gene copy number variation in bipolar disorder. American Journal of Medical Genetics Part B: Neuropsychiatric Genetics, 144(3), 259-265. doi: 10.1002/ajmg.b.30498 Lai, K., Kaspar, B. K., Gage, F. H., & Schaffer, D. V. (2003). Sonic hedgehog regulates adult neural progenitor proliferation in vitro and in vivo. Nature Neuroscience, 6(1), 21-27. doi: 10.1038/nn983 Lamb, C. A., Dooley, H. C., & Tooze, S. A. (2013). Endocytosis and autophagy: Shared machinery for degradation. Bioessays, 35(1), 34-45. doi: 10.1002/bies.201200130 Lamont, E. W., Coutu, D. L., Cermakian, N., & Boivin, D. B. (2010). Circadian rhythms and clock genes in psychotic disorders. Israel Journal of Psychiatry and Related Sciences, 47(1), 27-35. Langberg, K. A., Ma, L., Sharma, N. K., Hanis, C. L., Elbein, S. C., Hasstedt, S. J., & Das, S. K. (2012). Single nucleotide polymorphisms in JAZF1 and BCL11A gene are nominally associated with type 2 diabetes in African-American families from the GENNID study. Journal of Human Genetics, 57(1), 57-61. doi: 10.1038/jhg.2011.133 Lange, C., Mix, E., Frahm, J., Glass, Ä., Müller, J., Schmitt, O., . . . Hübner, R. (2011). Small molecule GSK-3 inhibitors increase neurogenesis of human neural progenitor cells. Neuroscience Letters, 488(1), 36-40. doi: 10.1016/j.neulet.2010.10.076 Latapy, C., Rioux, V., Guitton, M. J., & Beaulieu, J.-M. (2012). Selective deletion of forebrain glycogen synthase kinase 3β reveals a central role in serotonin- sensitive anxiety and social behaviour. Philosophical Transactions of the Royal Society B: Biological Sciences, 367(1601), 2460-2474. doi: 10.1098/rstb.2012.0094

196 |Page

Latourelle, J. C., Pankratz, N., Dumitriu, A., Wilk, J. B., Goldwurm, S., Pezzoli, G., . . . Gusella, J. F. (2009). Genomewide association study for onset age in Parkinson disease. BMC Medical Genetics, 10, 98. doi: 10.1186/1471-2350-10-98 Lau, K.-F., Miller, C. C., Anderton, B. H., & Shaw, P.-C. (1999a). Expression analysis of glycogen synthase kinase‐3 in human tissues. The Journal of Peptide Research, 54(1), 85-91. doi: 10.1034/j.1399-3011.1999.00083.x Lau, K.-F., Miller, C. C., Anderton, B. H., & Shaw, P.-C. (1999b). Molecular cloning and characterization of the human glycogen synthase kinase-3β promoter. Genomics, 60(2), 121-128. doi: 10.1006/geno.1999.5875 Lavoie, J., Hébert, M., & Beaulieu, J.-M. (2013). Glycogen synthase kinase-3β haploinsufficiency lengthens the circadian locomotor activity period in mice. Behavioural Brain Research, 253, 262-265. doi: 10.1016/j.bbr.2013.08.001 Lee, D. W., Zhao, X., Zhang, F., Eisenberg, E., & Greene, L. E. (2005). Depletion of GAK/auxilin 2 inhibits receptor-mediated endocytosis and recruitment of both clathrin and clathrin adaptors. Journal of Cell Science, 118(Pt 18), 4311-4321. doi: 10.1242/jcs.02548 Lee, J. H., Jeong, J. K., & Park, S. Y. (2014). Sulforaphane-induced autophagy flux prevents prion protein-mediated neurotoxicity through AMPK pathway. Neuroscience, 278, 31-39. doi: 10.1016/j.neuroscience.2014.07.072 Lee, J. L., Everitt, B. J., & Thomas, K. L. (2004a). Independent cellular processes for hippocampal memory consolidation and reconsolidation. Science, 304(5672), 839-843. doi: 10.1126/science.1095760 Lee, S. Y., Wenk, M. R., Kim, Y., Nairn, A. C., & De Camilli, P. (2004b). Regulation of synaptojanin 1 by cyclin-dependent kinase 5 at synapses. Proceedings of the National Academy of Sciences, 101(2), 546-551. doi: 10.1073/pnas.0307813100 Leeds, P. R., Yu, F., Wang, Z., Chiu, C.-T., Zhang, Y., Leng, Y., . . . Chuang, D.-M. (2014). A new avenue for lithium: Intervention in traumatic brain injury. ACS Chemical Neuroscience, 5(6), 422-433. doi: 10.1021/cn500040g Lei, P., Ayton, S., Bush, A. I., & Adlard, P. A. (2011). GSK-3 in neurodegenerative diseases. International Journal of Alzheimer's Disease, 2011, 189246. doi: 10.4061/2011/189246 Leroy, K., Ando, K., Héraud, C., Yilmaz, Z., Authelet, M., Boeynaems, J.-M., . . . Brion, J.-P. (2010). Lithium treatment arrests the development of neurofibrillary tangles in mutant tau transgenic mice with advanced neurofibrillary pathology. Journal of Alzheimer's Disease, 19(2), 705-719. doi: 10.3233/JAD-2010-1276 Leroy, K., & Brion, J.-P. (1999a). Developmental expression and localization of glycogen synthase kinase-3β in rat brain. Journal of Chemical Neuroanatomy, 16(4), 279-293. doi: 10.1016/S0891-0618(99)00012-5 Leroy, K., & Brion, J. P. (1999b). Developmental expression and localization of glycogen synthase kinase-3 [beta] in rat brain. Journal of Chemical Neuroanatomy, 16(4), 279-293. Leyhe, T., Eschweiler, G. W., Stransky, E., Gasser, T., Annas, P., Basun, H., & Laske, C. (2009). Increase of BDNF serum concentration in lithium treated patients with early Alzheimer's disease. Journal of Alzheimer's Disease, 16(3), 649-656. doi: 10.3233/JAD-2009-1004 Li, C. L., Sathyamurthy, A., Oldenborg, A., Tank, D., & Ramanan, N. (2014). SRF phosphorylation by glycogen synthase kinase-3 promotes axon growth in hippocampal neurons. The Journal of Neuroscience, 34(11), 4027-4042. doi: 10.1523/JNEUROSCI.4677-12.2014

197 |Page

Li, J., Lu, Y., Zhang, J., Kang, H., Qin, Z., & Chen, C. (2010). PI4KIIA is a novel regulator of tumor growth by its action on angiogenesis and HIF-1α regulation. Oncogene, 29(17), 2550-2559. doi: 10.1038/onc.2010.14 Li, L., Chen, Y., & Gibson, S. B. (2013). Starvation-induced autophagy is regulated by mitochondrial reactive oxygen species leading to AMPK activation. Cellular Signalling, 25(1), 50-65. doi: 10.1016/j.cellsig.2012.09.020 Li, L., Wang, X., Fei, X., Xia, L., Qin, Z., & Liang, Z. (2011). Parkinson's disease involves autophagy and abnormal distribution of cathepsin L. Neuroscience Letters, 489(1), 62-67. doi: 10.1016/j.neulet.2010.11.068 Li, X., & Jope, R. S. (2010). Is glycogen synthase kinase-3 a central modulator in mood regulation? Neuropsychopharmacology, 35(11), 2143-2154. doi: 10.1038/npp.2010.105 Li, X., Rosborough, K. M., Friedman, A. B., Zhu, W., & Roth, K. A. (2007). Regulation of mouse brain glycogen synthase kinase-3 by atypical antipsychotics. The International Journal of Neuropsychopharmacology, 10(01), 7-19. doi: 10.1017/S1461145706006547 Li, X., Zhu, W., Roh, M.-S., Friedman, A. B., Rosborough, K., & Jope, R. S. (2004). In vivo regulation of glycogen synthase kinase-3β (GSK3β) by serotonergic activity in mouse brain. Neuropsychopharmacology, 29(8), 1426. doi: 10.1038/sj.npp.1300439 Li, Y., McClintick, J., Zhong, L., Edenberg, H. J., Yoder, M. C., & Chan, R. J. (2005). Murine embryonic stem cell differentiation is promoted by SOCS-3 and inhibited by the zinc finger transcription factor Klf4. Blood, 105(2), 635-637. doi: 10.1182/blood-2004-07-2681 Lin, S.-Y., Li, T. Y., Liu, Q., Zhang, C., Li, X., Chen, Y., . . . Ruan, K. (2012a). Protein phosphorylation-acetylation cascade connects growth factor deprivation to autophagy. Autophagy, 8(9), 1385-1386. doi: 10.4161/auto.20959 Lin, S. Y., Li, T. Y., Liu, Q., Zhang, C., Li, X., Chen, Y., . . . Lin, S. C. (2012b). GSK3- TIP60-ULK1 signaling pathway links growth factor deprivation to autophagy. Science, 336(6080), 477-481. doi: 10.1126/science.1217032 Linseman, D. A., Butts, B. D., Precht, T. A., Phelps, R. A., Le, S. S., Laessig, T. A., . . . Heidenreich, K. A. (2004). Glycogen synthase kinase-3β phosphorylates Bax and promotes its mitochondrial localization during neuronal apoptosis. The Journal of Neuroscience, 24(44), 9993-10002. doi: 10.1523/JNEUROSCI.2057- 04.2004 Lipina, T. V., Wang, M., Liu, F., & Roder, J. C. (2012). Synergistic interactions between PDE4B and GSK-3: DISC1 mutant mice. Neuropharmacology, 62(3), 1252-1262. doi: 10.1016/j.neuropharm.2011.02.020 Liu, C., Li, Y., Semenov, M., Han, C., Baeg, G.-H., Tan, Y., . . . He, X. (2002). Control of β-catenin phosphorylation/degradation by a dual-kinase mechanism. Cell, 108(6), 837-847. doi: 10.1016/S0092-8674(02)00685-2 Liu, N., Li, H., Li, S., Shen, M., Xiao, N., Chen, Y., . . . Wang, Q. (2010). The Fbw7/human CDC4 tumor suppressor targets proproliferative factor KLF5 for ubiquitination and degradation through multiple phosphodegron motifs. Journal of Biological Chemistry, 285(24), 18858-18867. doi: 10.1074/jbc.M109.099440 Liu, S. J., Zhang, A. H., Li, H. L., Wang, Q., Deng, H. M., Netzer, W. J., . . . Wang, J. Z. (2003). Overactivation of glycogen synthase kinase‐3 by inhibition of phosphoinositol‐3 kinase and protein kinase C leads to hyperphosphorylation of tau and impairment of spatial memory. Journal of Neurochemistry, 87(6), 1333- 1344. doi: 10.1046/j.1471-4159.2003.02070.x

198 |Page

Liu, W., Huang, Y.-J., Liu, C., Yang, Y.-Y., Liu, H., Cui, J.-G., . . . Li, B.-L. (2014). Inhibition of TBK1 attenuates radiation-induced epithelial-mesenchymal transition of A549 human lung cancer cells via activation of GSK-3β and repression of ZEB1. Laboratory Investigation, 94(4), 362-370. doi: 10.1038/labinvest.2013.153 Liu, Y., Blackwood, D. H., Caesar, S., de Geus, E. J., Farmer, A., Ferreira, M. A., . . . Sullivan, P. F. (2011). Meta-analysis of genome-wide association data of bipolar disorder and major depressive disorder. Molecular Psychiatry, 16(1), 2-4. doi: 10.1038/mp.2009.107 Livak, K. J., & Schmittgen, T. D. (2001). Analysis of Relative Gene Expression Data Using Real-Time Quantitative PCR and the 2-ΔΔCT Method. Methods, 25(4), 402-408. doi: 10.1006/meth.2001.1262 Llorens-Marítin, M., Jurado, J., Hernández, F., & Ávila, J. (2014). GSK-3β, a pivotal kinase in Alzheimer disease. Frontiers in Molecular Neuroscience, 7, 46. doi: 10.3389/fnmol.2014.00046 Llorens-Martin, M., Fuster-Matanzo, A., Teixeira, C. M., Jurado-Arjona, J., & Ulloa, F. (2013). GSK-3β overexpression causes reversible alterations on postsynaptic densities and dendritic morphology of hippocampal granule neurons in vivo. Molecular Psychiatry, 18(4), 451-460. doi: 10.1038/mp.2013.4 Loberg, R. D., Vesely, E., & Brosius, F. C. (2002). Enhanced glycogen synthase kinase- 3β activity mediates hypoxia-induced apoptosis of vascular smooth muscle cells and is prevented by glucose transport and metabolism. Journal of Biological Chemistry, 277(44), 41667-41673. doi: 10.1074/jbc.M206405200 Lochhead, P. A., Kinstrie, R., Sibbet, G., Rawjee, T., Morrice, N., & Cleghon, V. (2006). A chaperone-dependent GSK3β transitional intermediate mediates activation-loop autophosphorylation. Molecular Cell, 24(4), 627-633. doi: 10.1016/j.molcel.2006.10.009 López-Figueroa, A. L., Norton, C. S., López-Figueroa, M. O., Armellini-Dodel, D., Burke, S., Akil, H., . . . Watson, S. J. (2004). Serotonin 5-HT1A, 5-HT1B, and 5-HT2A receptor mRNA expression in subjects with major depression, bipolar disorder, and schizophrenia. Biological Psychiatry, 55(3), 225-233. doi: 10.1016/j.biopsych.2003.09.017 Lu, D., Sun, H.-q., Wang, H., Barylko, B., Fukata, Y., Fukata, M., . . . Yin, H. L. (2012). Phosphatidylinositol 4-kinase IIα is palmitoylated by Golgi-localized palmitoyltransferases in cholesterol-dependent manner. Journal of Biological Chemistry, 287(26), 21856-21865. doi: 10.1074/jbc.M112.348094 Luan, Y., & Wang, P. (2014). FBW7-mediated ubiquitination and degradation of KLF5. World Journal of Biological Chemistry, 5(2), 216-223. doi: 10.4331/wjbc.v5.i2.216 Lucas, F. R., Goold, R. G., Gordon-Weeks, P. R., & Salinas, P. C. (1998). Inhibition of GSK-3beta leading to the loss of phosphorylated MAP-1B is an early event in axonal remodelling induced by WNT-7a or lithium. Journal of Cell Science, 111(10), 1351-1361. Lucas, F. R., & Salinas, P. C. (1997). WNT-7a induces axonal remodeling and increases synapsin I levels in cerebellar neurons. Developmental Biology, 192(1), 31-44. doi: 10.1006/dbio.1997.8734 Lucas, J. J., Hernández, F., Gómez‐Ramos, P., Morán, M. A., Hen, R., & Avila, J. (2001). Decreased nuclear β‐catenin, tau hyperphosphorylation and neurodegeneration in GSK‐3β conditional transgenic mice. The EMBO Journal, 20(1-2), 27-39. doi: 10.1093/emboj/20.1.27

199 |Page

Ludtmann, M. H., Boeckeler, K., & Williams, R. S. (2011). Molecular pharmacology in a simple model system: Implicating MAP kinase and phosphoinositide signalling in bipolar disorder. Seminars in Cell & Developmental Biology, 22(1), 105-113. doi: 10.1016/j.semcdb.2010.11.002 Luna-Muñoz, J., Chávez-Macías, L., García-Sierra, F., & Mena, R. (2007). Earliest stages of tau conformational changes are related to the appearance of a sequence of specific phospho-dependent tau epitopes in Alzheimer's disease. Journal of Alzheimer's Disease, 12(4), 365-375. Lyoo, I. K., Kim, M. J., Stoll, A. L., Demopulos, C. M., Parow, A. M., Dager, S. R., . . . Renshaw, P. F. (2004). Frontal lobe gray matter density decreases in bipolar I disorder. Biological Psychiatry, 55(6), 648-651. doi: 10.1016/j.biopsych.2003.10.017 MacAulay, K., Doble, B. W., Patel, S., Hansotia, T., Sinclair, E. M., Drucker, D. J., . . . Woodgett, J. R. (2007). Glycogen synthase kinase 3α-specific regulation of murine hepatic glycogen metabolism. Cell Metabolism, 6(4), 329-337. doi: 10.1016/j.cmet.2007.08.013 Maccario, H., Perera, N., Davidson, L., Downes, C., & Leslie, N. (2007). PTEN is destabilized by phosphorylation on Thr366. Biochemical Journal, 405(3), 439- 444. doi: 10.1042/BJ20061837 Machado‐Vieira, R., Manji, H. K., & Zarate Jr, C. A. (2009). The role of lithium in the treatment of bipolar disorder: Convergent evidence for neurotrophic effects as a unifying hypothesis. Bipolar Disorders, 11(s2), 92-109. doi: 10.1111/j.1399- 5618.2009.00714.x Malinow, R., & Malenka, R. C. (2002). AMPA receptor trafficking and synaptic plasticity. Annual Review of Neuroscience, 25, 103-126. doi: 10.1146/annurev.neuro.25.112701.142758 Manev, H., & Uz, T. (2006). Clock genes: Influencing and being influenced by psychoactive drugs. Trends in Pharmacological Sciences, 27(4), 186-189. doi: 10.1016/j.tips.2006.02.003 Manji, H. K., Quiroz, J. A., Payne, J. L., Singh, J., Lopes, B. P., Viegas, J. S., & Zarate, C. A. (2003). The underlying neurobiology of bipolar disorder. World Psychiatry, 2(3), 136-146. Manoach, D. S., & Stickgold, R. (2009). Does abnormal sleep impair memory consolidation in schizophrenia? Frontiers in Human Neuroscience, 3, 21. doi: 10.3389/neuro.09.021.2009 Mansour, H. A., Wood, J., Logue, T., Chowdari, K. V., Dayal, M., Kupfer, D. J., . . . Nimgaonkar, V. L. (2006). Association study of eight circadian genes with bipolar I disorder, schizoaffective disorder and schizophrenia. Genes, Brain and Behavior, 5(2), 150-157. doi: 10.1111/j.1601-183X.2005.00147.x Mao, Y., Ge, X., Frank, C. L., Madison, J. M., Koehler, A. N., Doud, M. K., . . . Singh, K. K. (2009). Disrupted in schizophrenia 1 regulates neuronal progenitor proliferation via modulation of GSK3β/β-catenin signaling. Cell, 136(6), 1017- 1031. doi: 10.1016/j.cell.2008.12.044 Marazziti, D., Di Pietro, C., Golini, E., Mandillo, S., Matteoni, R., & Tocchini- Valentini, G. P. (2009). Induction of macroautophagy by overexpression of the Parkinson's disease-associated GPR37 receptor. FASEB Journal, 23(6), 1978- 1987. doi: 10.1096/fj.08-121210 Martello, G., Sugimoto, T., Diamanti, E., Joshi, A., Hannah, R., Ohtsuka, S., . . . Smith, A. (2012). Esrrb is a pivotal target of the Gsk3/Tcf3 axis regulating embryonic

200 |Page

stem cell self-renewal. Cell Stem Cell, 11(4), 491-504. doi: 10.1016/j.stem.2012.06.008 Martin, M., Rehani, K., Jope, R. S., & Michalek, S. M. (2005). Toll-like receptor– mediated cytokine production is differentially regulated by glycogen synthase kinase 3. Nature Immunology, 6(8), 777-784. doi: 10.1038/ni1221 Martin, T., Loyet, K., Barry, V., & Kowalchyk, J. (1997). The role of PtdIns (4,5)P2 in exocytotic membrane fusion. Biochemical Society Transactions, 25(4), 1137- 1141. Martinek, S., Inonog, S., Manoukian, A. S., & Young, M. W. (2001). A role for the segment polarity gene shaggy/GSK-3 in the Drosophila circadian clock. Cell, 105(6), 769-779. doi: 10.1016/S0092-8674(01)00383-X Massol, R. H., Boll, W., Griffin, A. M., & Kirchhausen, T. (2006). A burst of auxilin recruitment determines the onset of clathrin-coated vesicle uncoating. Proceedings of the National Academy of Sciences, 103(27), 10265-10270. doi: 10.1073/pnas.0603369103 Matrisciano, F., Bonaccorso, S., Ricciardi, A., Scaccianoce, S., Panaccione, I., Wang, L., . . . Girardi, P. (2009). Changes in BDNF serum levels in patients with major depression disorder (MDD) after 6 months treatment with sertraline, escitalopram, or venlafaxine. Journal of Psychiatric Research, 43(3), 247-254. doi: 10.1016/j.jpsychires.2008.03.014 Matsagas, K., Lim, D. B., Horwitz, M., Rizza, C. L., Mueller, L. D., Villeponteau, B., & Rose, M. R. (2009). Long-term functional side-effects of stimulants and sedatives in Drosophila melanogaster. PLOS ONE, 4(8), e6578. doi: 10.1371/journal.pone.0006578 Matsuda, T., Nakamura, T., Nakao, K., Arai, T., Katsuki, M., Heike, T., & Yokota, T. (1999). STAT3 activation is sufficient to maintain an undifferentiated state of mouse embryonic stem cells. The EMBO Journal, 18(15), 4261-4269. doi: 10.1093/emboj/18.15.4261 Matsui, T., Itoh, T., & Fukuda, M. (2011). Small GTPase Rab12 regulates constitutive degradation of transferrin receptor. Traffic, 12(10), 1432-1443. doi: 10.1111/j.1600-0854.2011.01240.x Matus, A. (1994). Stiff microtubules and neuronal morphology. Trends in Neurosciences, 17(1), 19-22. doi: 10.1016/0166-2236(94)90030-2 Maurer, U., Charvet, C., Wagman, A. S., Dejardin, E., & Green, D. R. (2006). Glycogen synthase kinase-3 regulates mitochondrial outer membrane permeabilization and apoptosis by destabilization of MCL-1. Molecular Cell, 21(6), 749-760. doi: 10.1016/j.molcel.2006.02.009 McManus, E. J., Sakamoto, K., Armit, L. J., Ronaldson, L., Shpiro, N., Marquez, R., & Alessi, D. R. (2005). Role that phosphorylation of GSK3 plays in insulin and Wnt signalling defined by knockin analysis. The EMBO Journal, 24(8), 1571- 1583. doi: 10.1038/sj.emboj.7600633 Medina, M., & Wandosell, F. (2011). Deconstructing GSK-3: The fine regulation of its activity. International Journal of Alzheimer's Disease, 2011, 479249. doi: 10.4061/2011/479249 Mei, L., & Xiong, W. C. (2008). Neuregulin 1 in neural development, synaptic plasticity and schizophrenia. Nature Reviews Neuroscience, 9(6), 437-452. doi: 10.1038/nrn2392 Meltzer, H. Y., Matsubara, S., & Lee, J. C. (1988). The ratios of serotonin2 and dopamine2 affinities differentiate atypical and typical antipsychotic drugs. Psychopharmacology Bulletin, 25(3), 390-392.

201 |Page

Merikangas, K. R., Jin, R., He, J.-P., Kessler, R. C., Lee, S., Sampson, N. A., . . . Karam, E. G. (2011). Prevalence and correlates of bipolar spectrum disorder in the world mental health survey initiative. Archives of General Psychiatry, 68(3), 241-251. doi: 10.1001/archgenpsychiatry.2011.12 Michale, B., Noor, E., Laurie, G., Emily, H. F., Maria, D., Michael, J. B., . . . Dario, R. A. (2011). TTBK2 kinase substrate specificity and the impact of spinocerebellar-ataxia-causing mutations on expression, activity, localization and development. Biochemical Journal, 437(1), 157-167. doi: 10.1042/BJ20110276 Millar, J. K., Wilson-Annan, J. C., Anderson, S., Christie, S., Taylor, M. S., Semple, C. A., . . . Blackwood, D. H. (2000). Disruption of two novel genes by a translocation co-segregating with schizophrenia. Human Molecular Genetics, 9(9), 1415-1423. doi: 10.1093/hmg/9.9.1415 Miller, J. R. (2002). The Wnts. Genome Biology, 3(1), 1-15. Minogue, S., Chu, K. M. E., Westover, E. J., Covey, D. F., Hsuan, J. J., & Waugh, M. G. (2010). Relationship between phosphatidylinositol 4-phosphate synthesis, membrane organization, and lateral diffusion of PI4KIIα at the trans-Golgi network. Journal of Lipid Research, 51(8), 2314-2324. doi: 10.1194/jlr.M005751 Minogue, S., Waugh, M. G., De Matteis, M. A., Stephens, D. J., Berditchevski, F., & Hsuan, J. J. (2006). Phosphatidylinositol 4-kinase is required for endosomal trafficking and degradation of the EGF receptor. Journal of Cell Science, 119(3), 571-581. doi: 10.1242/jcs.02752 Morales-García, J. A., Susı́n, C., Alonso-Gil, S., Pérez, D. I., Palomo, V., Perez, C., . . . Martinez, A. (2013). Glycogen synthase kinase-3 inhibitors as potent therapeutic agents for the treatment of Parkinson disease. ACS Chemical Neuroscience, 4(2), 350-360. doi: 10.1021/cn300182g Morrow, D., Guha, S., Sweeney, C., Birney, Y., Walshe, T., O’Brien, C., . . . Cahill, P. A. (2008). Notch and vascular smooth muscle cell phenotype. Circulation Research, 103(12), 1370-1382. doi: 10.1161/circresaha.108.187534 Mössinger, J., Wieffer, M., Krause, E., Freund, C., Gerth, F., Krauss, M., & Haucke, V. (2012). Phosphatidylinositol 4‐kinase IIα function at endosomes is regulated by the ubiquitin ligase Itch. EMBO Reports, 13(12), 1087-1094. doi: 10.1038/embor.2012.164 Motley, A. M., Berg, N., Taylor, M. J., Sahlender, D. A., Hirst, J., Owen, D. J., & Robinson, M. S. (2006). Functional analysis of AP-2 alpha and mu2 subunits. Molecular Biology of the Cell, 17(12), 5298-5308. doi: 10.1091/mbc.E06-05- 0452 Mukai, F., Ishiguro, K., Sano, Y., & Fujita, S. C. (2002). Alternative splicing isoform of tau protein kinase I/glycogen synthase kinase 3β. Journal of Neurochemistry, 81(5), 1073-1083. doi: 10.1046/j.1471-4159.2002.00918.x Mumm, J. S., & Kopan, R. (2000). Notch signaling: From the outside in. Developmental Biology, 228(2), 151-165. doi: 10.1006/dbio.2000.9960 Musacchio, A., Smith, C. J., Roseman, A. M., Harrison, S. C., Kirchhausen, T., & Pearse, B. M. (1999). Functional organization of clathrin in coats: Combining electron cryomicroscopy and X-ray crystallography. Molecular Cell, 3(6), 761- 770. doi: 10.1016/S1097-2765(01)80008-3 Nakashima, H., Ishihara, T., Suguimoto, P., Yokota, O., Oshima, E., Kugo, A., . . . Lee, V. M. Y. (2005). Chronic lithium treatment decreases tau lesions by promoting

202 |Page

ubiquitination in a mouse model of tauopathies. Acta Neuropathologica, 110(6), 547-556. doi: 10.1007/s00401-005-1087-4 Nelson, C. D., Kim, M. J., Hsin, H., Chen, Y., & Sheng, M. (2013). Phosphorylation of threonine-19 of PSD-95 by GSK-3β is required for PSD-95 mobilization and long-term depression. The Journal of Neuroscience, 33(29), 12122-12135. doi: 10.1523/jneurosci.0131-13.2013 Neves-Pereira, M., Mundo, E., Muglia, P., King, N., Macciardi, F., & Kennedy, J. L. (2002). The brain-derived neurotrophic factor gene confers susceptibility to bipolar disorder: Evidence from a family-based association study. The American Journal of Human Genetics, 71(3), 651-655. doi: 10.1086/342288 Newmyer, S. L., & Schmid, S. L. (2001). Dominant-interfering Hsc70 mutants disrupt multiple stages of the clathrin-coated vesicle cycle in vivo. The Journal of Cell Biology, 152(3), 607-620. doi: 10.1083/jcb.152.3.607 Nicol Ferrier, I., Chowdhury, R., Thompson, J. M., Watson, S., & Young, A. H. (2004). Neurocognitive function in unaffected first‐degree relatives of patients with bipolar disorder: A preliminary report. Bipolar Disorders, 6(4), 319-322. doi: 10.1111/j.1399-5618.2004.00122.x Nishimura, T., & Kaibuchi, K. (2007). Numb controls integrin endocytosis for directional cell migration with aPKC and PAR-3. Developmental Cell, 13(1), 15-28. doi: 10.1016/j.devcel.2007.05.003 Niwa, H., Burdon, T., Chambers, I., & Smith, A. (1998). Self-renewal of pluripotent embryonic stem cells is mediated via activation of STAT3. Genes & Development, 12(13), 2048-2060. doi: 10.1101/gad.12.13.2048 Niwa, H., Ogawa, K., Shimosato, D., & Adachi, K. (2009). A parallel circuit of LIF signalling pathways maintains pluripotency of mouse ES cells. Nature, 460(7251), 118-122. doi: 10.1038/nature08113 Nixon, R. A. (2013). The role of autophagy in neurodegenerative disease. Nature Medicine, 19(8), 983-997. doi: 10.1038/nm.3232 Noble, W., Planel, E., Zehr, C., Olm, V., Meyerson, J., Suleman, F., . . . Feinstein, B. (2005). Inhibition of glycogen synthase kinase-3 by lithium correlates with reduced tauopathy and degeneration in vivo. Proceedings of the National Academy of Sciences, 102(19), 6990-6995. doi: 10.1073/pnas.0500466102 Nonaka, S., & Chuang, D.-M. (1998). Neuroprotective effects of chronic lithium on focal cerebral ischemia in rats. Neuroreport, 9(9), 2081-2084. doi: 10.1097/00001756-199806220-00031 Nonaka, S., Hough, C. J., & Chuang, D.-M. (1998). Chronic lithium treatment robustly protects neurons in the central nervous system against excitotoxicity by inhibiting N-methyl-D-aspartate receptor-mediated calcium influx. Proceedings of the National Academy of Sciences, 95(5), 2642-2647. doi: 10.1073/pnas.95.5.2642 Nugent, A. C., Milham, M. P., Bain, E. E., Mah, L., Cannon, D. M., Marrett, S., . . . Drevets, W. C. (2006). Cortical abnormalities in bipolar disorder investigated with MRI and voxel-based morphometry. NeuroImage, 30(2), 485-497. doi: 10.1016/j.neuroimage.2005.09.029 Nunes, P. V., Forlenza, O. V., & Gattaz, W. F. (2007). Lithium and risk for Alzheimer's disease in elderly patients with bipolar disorder. The British Journal of Psychiatry, 190(4), 359-360. doi: 10.1192/bjp.bp.106.029868 Nuydens, R., Van Den Kieboom, G., Nolten, C., Verhulst, C., Van Osta, P., Spittaels, K., . . . Van Leuven, F. (2002). Coexpression of GSK-3β corrects phenotypic aberrations of dorsal root ganglion cells, cultured from adult transgenic mice

203 |Page

overexpressing human protein tau. Neurobiology of Disease, 9(1), 38-48. doi: 10.1006/nbdi.2001.0454 O'Brien, W. T., Harper, A. D., Jové, F., Woodgett, J. R., Maretto, S., Piccolo, S., & Klein, P. S. (2004). Glycogen synthase kinase-3β haploinsufficiency mimics the behavioral and molecular effects of lithium. The Journal of Neuroscience, 24(30), 6791-6798. doi: 10.1523/jneurosci.4753-03.2004 Ohnishi, T., Yamada, K., Ohba, H., Iwayama, Y., Toyota, T., Hattori, E., . . . Ozaki, N. (2007). A promoter haplotype of the inositol monophosphatase 2 gene (IMPA2) at 18p11.2 confers a possible risk for bipolar disorder by enhancing transcription. Neuropsychopharmacology, 32(8), 1727-1737. doi: 10.1038/sj.npp.1301307 Ohno, H., Stewart, J., Fournier, M. C., Bosshart, H., Rhee, I., Miyatake, S., . . . Bonifacino, J. S. (1995). Interaction of tyrosine-based sorting signals with clathrin-associated proteins. Science, 269(5232), 1872-1875. doi: 10.1126/science.7569928 Olusanya, O., Andrews, P. D., Swedlow, J. R., & Smythe, E. (2001). Phosphorylation of threonine 156 of the μ2 subunit of the AP2 complex is essential for endocytosis in vitro and in vivo. Current Biology, 11(11), 896-900. doi: 10.1016/S0960- 9822(01)00240-8 Ongur, D., Ferry, A. T., & Price, J. L. (2003). Architectonic subdivision of the human orbital and medial prefrontal cortex. The Journal of Comparative Neurology, 460(3), 425-449. doi: 10.1002/cne.10609 Owen, D. J., Collins, B. M., & Evans, P. R. (2004). Adaptors for clathrin coats: Structure and function. Annual Review of Cell and Developmental Biology, 20, 153-191. doi: 10.1146/annurev.cellbio.20.010403.104543 Owen, D. J., & Evans, P. R. (1998). A structural explanation for the recognition of tyrosine-based endocytotic signals. Science, 282(5392), 1327-1332. doi: 10.1126/science.282.5392.1327 Owen, R., & Gordon-Weeks, P. R. (2003). Inhibition of glycogen synthase kinase 3β in sensory neurons in culture alters filopodia dynamics and microtubule distribution in growth cones. Molecular and Cellular Neuroscience, 23(4), 626- 637. doi: 10.1016/S1044-7431(03)00095-2 Pan, B., Huang, X.-F., & Deng, C. (2011). Antipsychotic treatment and neuregulin 1– ErbB4 signalling in schizophrenia. Progress in Neuro-Psychopharmacology and Biological Psychiatry, 35(4), 924-930. doi: 10.1016/j.pnpbp.2011.04.002 Pan, T., Kondo, S., Zhu, W., Xie, W., Jankovic, J., & Le, W. (2008a). Neuroprotection of rapamycin in lactacystin-induced neurodegeneration via autophagy enhancement. Neurobiology of Disease, 32(1), 16-25. doi: 10.1016/j.nbd.2008.06.003 Pan, W., Choi, S. C., Wang, H., Qin, Y., Volpicelli-Daley, L., Swan, L., . . . Li, L. (2008b). Wnt3a-mediated formation of phosphatidylinositol 4, 5-bisphosphate regulates LRP6 phosphorylation. Science Signalling, 321(5894), 1350-1353. doi: 10.1126/science.1160741 Pan, Y., Bai, C. B., Joyner, A. L., & Wang, B. (2006). Sonic hedgehog signaling regulates Gli2 transcriptional activity by suppressing its processing and degradation. Molecular and Cellular Biology, 26(9), 3365-3377. doi: 10.1128/MCB.26.9.3365-3377.2006 Pang, L., Sawada, T., Decker, S. J., & Saltiel, A. R. (1995). Inhibition of MAP kinase kinase blocks the differentiation of PC-12 cells induced by nerve growth factor.

204 |Page

Journal of Biological Chemistry, 270(23), 13585-13588. doi: 10.1074/jbc.270.23.13585 Pap, M., & Cooper, G. M. (1998). Role of glycogen synthase kinase-3 in the phosphatidylinositol 3-kinase/Akt cell survival pathway. Journal of Biological Chemistry, 273(32), 19929-19932. doi: 10.1074/jbc.273.32.19929 Park, J., Lee, S. B., Lee, S., Kim, Y., Song, S., Kim, S., . . . Chung, J. (2006). Mitochondrial dysfunction in Drosophila PINK1 mutants is complemented by parkin. Nature, 441(7097), 1157-1161. doi: 10.1038/nature04788 Park, S. W., Lee, J. G., Seo, M. K., Cho, H. Y., Lee, C. H., Lee, J. H., . . . Kim, Y. H. (2014). Effects of mood‐stabilizing drugs on dendritic outgrowth and synaptic protein levels in primary hippocampal neurons. Bipolar Disorders. doi: 10.1111/bdi.12262 Parr, C., Carzaniga, R., Gentleman, S. M., Van Leuven, F., Walter, J., & Sastre, M. (2012). Glycogen synthase kinase 3 inhibition promotes lysosomal biogenesis and autophagic degradation of the amyloid-β precursor protein. Molecular and Cellular Biology, 32(21), 4410-4418. doi: 10.1128/MCB.00930-12 Pasquali, L., Longone, P., Isidoro, C., Ruggieri, S., Paparelli, A., & Fornai, F. (2009). Autophagy, lithium, and amyotrophic lateral sclerosis. Muscle & Nerve, 40(2), 173-194. doi: 10.1002/mus.21423 Pastorino, J. G., Hoek, J. B., & Shulga, N. (2005). Activation of glycogen synthase kinase 3β disrupts the binding of hexokinase II to mitochondria by phosphorylating voltage-dependent anion channel and potentiates chemotherapy-induced cytotoxicity. Cancer Research, 65(22), 10545-10554. doi: 10.1158/0008-5472.can-05-1925 Pei, J.-J., Tanaka, T., Tung, Y.-C., Braak, E., Iqbal, K., & Grundke-Iqbal, I. (1997). Distribution, levels, and activity of glycogen synthase kinase-3 in the Alzheimer disease brain. Journal of Neuropathology & Experimental Neurology, 56(1), 70- 78. Peineau, S., Bradley, C., Taghibiglou, C., Doherty, A., Bortolotto, Z. A., Wang, Y. T., & Collingridge, G. L. (2008). The role of GSK‐3 in synaptic plasticity. British Journal of Pharmacology, 153(S1), S428-S437. doi: 10.1038/bjp.2008.2 Peineau, S., Nicolas, C. S., Bortolotto, Z. A., Bhat, R. V., Ryves, W. J., Harwood, A. J., . . . Collingridge, G. L. (2009). A systematic investigation of the protein kinases involved in NMDA receptor-dependent LTD: Evidence for a role of GSK-3 but not other serine/threonine kinases. Molecular Brain, 2, 22. doi: 10.1186/1756- 6606-2-22 Peineau, S., Taghibiglou, C., Bradley, C., Wong, T. P., Liu, L., Lu, J., . . . Bouschet, T. (2007). LTP inhibits LTD in the hippocampus via regulation of GSK3β. Neuron, 53(5), 703-717. doi: 10.1016/j.neuron.2007.01.029 Pérez, M., Hernández, F., Lim, F., Díaz-Nido, J., & Avila, J. (2003). Chronic lithium treatment decreases mutant tau protein aggregation in a transgenic mouse model. Journal of Alzheimer's Disease, 5(4), 301-308. Pesah, Y., Pham, T., Burgess, H., Middlebrooks, B., Verstreken, P., Zhou, Y., . . . Mardon, G. (2004). Drosophila parkin mutants have decreased mass and cell size and increased sensitivity to oxygen radical stress. Development, 131(9), 2183-2194. doi: 10.1242/dev.01095 Peter, M. E., & Krammer, P. (2003). The CD95 (APO-1/Fas) DISC and beyond. Cell Death & Differentiation, 10(1), 26-35. doi: 10.1038/sj.cdd.4401186 Petrini, E. M., Ravasenga, T., Hausrat, T. J., Iurilli, G., Olcese, U., Racine, V., . . . Benfenati, F. (2014). Synaptic recruitment of gephyrin regulates surface

205 |Page

GABAA receptor dynamics for the expression of inhibitory LTP. Nature Communications, 5, 3921. doi: 10.1038/ncomms4921 Pezawas, L., Verchinski, B. A., Mattay, V. S., Callicott, J. H., Kolachana, B. S., Straub, R. E., . . . Weinberger, D. R. (2004). The brain-derived neurotrophic factor val66met polymorphism and variation in human cortical morphology. The Journal of Neuroscience, 24(45), 10099-10102. doi: 10.1523/JNEUROSCI. 2680-04.2004 Pluquet, O., Qu, L.-K., Baltzis, D., & Koromilas, A. E. (2005). Endoplasmic reticulum stress accelerates p53 degradation by the cooperative actions of Hdm2 and glycogen synthase kinase 3β. Molecular and Cellular Biology, 25(21), 9392- 9405. doi: 10.1128/MCB.25.21.9392-9405.2005 Polter, A., Beurel, E., Yang, S., Garner, R., Song, L., Miller, C. A., . . . Li, X. (2010). Deficiency in the inhibitory serine-phosphorylation of glycogen synthase kinase- 3 increases sensitivity to mood disturbances. Neuropsychopharmacology, 35(8), 1761-1774. doi: 10.1038/npp.2010.43 Postle, A. D., Wilton, D. C., Hunt, A. N., & Attard, G. S. (2007). Probing phospholipid dynamics by electrospray ionisation mass spectrometry. Progress in Lipid Research, 46(3-4), 200-224. doi: 10.1016/j.plipres.2007.04.001 Prickaerts, J., Moechars, D., Cryns, K., Lenaerts, I., van Craenendonck, H., Goris, I., . . . Steckler, T. (2006). Transgenic mice overexpressing glycogen synthase kinase 3β: A putative model of hyperactivity and mania. The Journal of Neuroscience, 26(35), 9022-9029. doi: 10.1523/jneurosci.5216-05.2006 Pucharcos, C., Estivill, X., & de la Luna, S. (2000). Intersectin 2, a new multimodular protein involved in clathrin-mediated endocytosis. FEBS Letters, 478(1-2), 43- 51. doi: 10.1016/S0014-5793(00)01793-2 Qin, Y., Li, L., Pan, W., & Wu, D. (2009). Regulation of phosphatidylinositol kinases and metabolism by Wnt3a and Dvl. Journal of Biological Chemistry, 284(34), 22544-22548. doi: 10.1074/jbc.M109.014399 Quadri, M., Fang, M., Picillo, M., Olgiati, S., Breedveld, G. J., Graafland, J., . . . Bonifati, V. (2013). Mutation in the SYNJ1 gene associated with autosomal recessive, early-onset Parkinsonism. Human Mutation, 34(9), 1208-1215. doi: 10.1002/humu.22373 Radonić, A., Thulke, S., Mackay, I. M., Landt, O., Siegert, W., & Nitsche, A. (2004). Guideline to reference gene selection for quantitative real-time PCR. Biochemical and Biophysical Research Communications, 313(4), 856-862. doi: 10.1016/j.bbrc.2003.11.177 Ramonet, D., Daher, J. P. L., Lin, B. M., Stafa, K., Kim, J., Banerjee, R., . . . Yang, L. (2011). Dopaminergic neuronal loss, reduced neurite complexity and autophagic abnormalities in transgenic mice expressing G2019S mutant LRRK2. PLOS ONE, 6(4), e18568. doi: 10.1371/journal.pone.0018568 Ravikumar, B., Moreau, K., Jahreiss, L., Puri, C., & Rubinsztein, D. C. (2010). Plasma membrane contributes to the formation of pre-autophagosomal structures. Nature Cell Biology, 12(8), 747-757. doi: 10.1038/ncb2078 Redmond, L., Oh, S.-R., Hicks, C., Weinmaster, G., & Ghosh, A. (2000). Nuclear Notch1 signaling and the regulation of dendritic development. Nature Neuroscience, 3(1), 30-40. doi: 10.1038/71104 Ricotta, D., Conner, S. D., Schmid, S. L., von Figura, K., & Höning, S. (2002). Phosphorylation of the AP2 μ subunit by AAK1 mediates high affinity binding to membrane protein sorting signals. The Journal of Cell Biology, 156(5), 791- 795. doi: 10.1083/jcb.200111068

206 |Page

Ripke, S., Sanders, A. R., Kendler, K. S., Levinson, D. F., Sklar, P., Holmans, P. A., . . . Andreassen, O. A. (2011). Genome-wide association study identifies five new schizophrenia loci. Nature Genetics, 43(10), 969-976. doi: 10.1038/ng.940 Robinson, J. W., Leshchyns'ka, I., Farghaian, H., Hughes, W. E., Sytnyk, V., Neely, G. G., & Cole, A. R. (2014). PI4KIIalpha phosphorylation by GSK3 directs vesicular trafficking to lysosomes. Biochemical Journal, 464(1), 145-156. doi: 10.1042/bj20140497 Rong, Y., Liu, M., Ma, L., Du, W., Zhang, H., Tian, Y., . . . Yu, L. (2012). Clathrin and phosphatidylinositol-4,5-bisphosphate regulate autophagic lysosome reformation. Nature Cell Biology, 14(9), 924-934. doi: 10.1038/ncb2557 Rosahl, T. W., Spillane, D., Missler, M., Herz, J., Selig, D. K., Wolff, J. R., . . . Südhof, T. C. (1995). Essential functions of synapsins I and II in synaptic vesicle regulation. Nature, 375(6531), 488-493. doi: 10.1038/375488a0 Roux, A., Uyhazi, K., Frost, A., & De Camilli, P. (2006). GTP-dependent twisting of dynamin implicates constriction and tension in membrane fission. Nature, 441(7092), 528-531. doi: 10.1038/nature04718 Roybal, K., Theobold, D., Graham, A., DiNieri, J. A., Russo, S. J., Krishnan, V., . . . Birnbaum, S. (2007). Mania-like behavior induced by disruption of CLOCK. Proceedings of the National Academy of Sciences, 104(15), 6406-6411. doi: 10.1073/pnas.0609625104 Ruediger, S., Vittori, C., Bednarek, E., Genoud, C., Strata, P., Sacchetti, B., & Caroni, P. (2011). Learning-related feedforward inhibitory connectivity growth required for memory precision. Nature, 473(7348), 514-518. doi: 10.1038/nature09946 Ruel, L., Pantesco, V., Lutz, Y., Simpson, P., & Bourouis, M. (1993). Functional significance of a family of protein kinases encoded at the shaggy locus in Drosophila. The EMBO Journal, 12(4), 1657-1669. Rui, Y., Myers, K. R., Yu, K., Wise, A., De Blas, A. L., Hartzell, H. C., & Zheng, J. Q. (2013). Activity-dependent regulation of dendritic growth and maintenance by glycogen synthase kinase 3β. Nature Communications, 4, 2628. doi: 10.1038/ncomms3628 Sahar, S., Zocchi, L., Kinoshita, C., Borrelli, E., & Sassone-Corsi, P. (2010). Regulation of BMAL1 protein stability and circadian function by GSK3β-mediated phosphorylation. PLOS ONE, 5(1), e8561. doi: 10.1371/journal.pone.0008561 Şahin, H. B., & Çelik, A. (2013). Drosophila Eye Development and Photoreceptor Specification. eLS. doi: 10.1002/9780470015902.a0001147.pub2 Saito, T., Guan, F., Papolos, D. F., Lau, S., Klein, M., Fann, C. S., & Lachman, H. M. (2001). Mutation analysis of SYNJ1: A possible candidate gene for chromosome 21q22-linked bipolar disorder. Molecular Psychiatry, 6(4), 387-395. doi: 10.1038/sj.mp.4000871 Saito, Y., Vandenheede, J., & Cohen, P. (1994). The mechanism by which epidermal growth factor inhibits glycogen synthase kinase 3 in A431 cells. Biochemical Journal, 303, 27-31. Sakanaka, C. (2002). Phosphorylation and regulation of β-catenin by casein kinase I epsilon. Journal of Biochemistry, 132(5), 697-703. doi: 10.1093/oxfordjournals.jbchem.a003276 Salazar, G., Craige, B., Wainer, B. H., Guo, J., De Camilli, P., & Faundez, V. (2005). Phosphatidylinositol-4-kinase type II α is a component of adaptor protein-3- derived vesicles. Molecular Biology of the Cell, 16(8), 3692-3704. doi: 10.1091/mbc.E05-01-0020

207 |Page

Salazar, G., Zlatic, S., Craige, B., Peden, A. A., Pohl, J., & Faundez, V. (2009). Hermansky-Pudlak syndrome protein complexes associate with phosphatidylinositol 4-kinase type II α in neuronal and non-neuronal cells. Journal of Biological Chemistry, 284(3), 1790-1802. doi: 10.1074/jbc.M805991200 Salcedo-Tello, P., Ortiz-Matamoros, A., & Arias, C. (2011). GSK3 function in the brain during development, neuronal plasticity, and neurodegeneration. International Journal of Alzheimer’s Disease, 2011, 189728. doi: 10.4061/2011/189728 Sanchez, C., Dıaz-Nido, J., & Avila, J. (2000). Phosphorylation of microtubule- associated protein 2 (MAP2) and its relevance for the regulation of the neuronal cytoskeleton function. Progress in Neurobiology, 61(2), 133-168. doi: 10.1016/S0301-0082(99)00046-5 Sánchez, C., Dı́az-Nido, J., & Avila, J. (1998). Regulation of a site-specific phosphorylation of the microtubule-associated protein 2 during the development of cultured neurons. Neuroscience, 87(4), 861-870. doi: 10.1016/S0306- 4522(98)00195-x Sánchez, C., Pérez, M., & Avila, J. (2000). GSK3β-mediated phosphorylation of the microtubule-associated protein 2C (MAP2C) prevents microtubule bundling. European Journal of Cell Biology, 79(4), 252-260. doi: 10.1078/S0171- 9335(04)70028-x Sánchez, C., Tompa, P., Szücs, K., Friedrich, P., & Avila, J. (1996). Phosphorylation and dephosphorylation in the proline‐rich C‐terminal domain of microtubule‐ associated protein 2. European Journal of Biochemistry, 241(3), 765-771. doi: 10.1111/j.1432-1033.1996.00765.x Sander, J. D., & Joung, J. K. (2014). CRISPR-Cas systems for editing, regulating and targeting genomes. Nature Biotechnology, 32(4), 347-355. doi: 10.1038/nbt.2842 Sargent, P. A., Kjaer, K. H., Bench, C. J., Rabiner, E. A., Messa, C., Meyer, J., . . . Cowen, P. J. (2000). Brain serotonin1A receptor binding measured by positron emission tomography with [11C]WAY-100635: effects of depression and antidepressant treatment. Archives of General Psychiatry, 57(2), 174-180. doi: 10.1001/archpsyc.57.2.174 Sarkar, S., Floto, R. A., Berger, Z., Imarisio, S., Cordenier, A., Pasco, M., . . . Rubinsztein, D. C. (2005). Lithium induces autophagy by inhibiting inositol monophosphatase. The Journal of Cell Biology, 170(7), 1101-1111. doi: 10.1083/jcb.200504035 Sato, N., Meijer, L., Skaltsounis, L., Greengard, P., & Brivanlou, A. H. (2004). Maintenance of pluripotency in human and mouse embryonic stem cells through activation of Wnt signaling by a pharmacological GSK-3-specific inhibitor. Nature Medicine, 10, 55-63. doi: 10.1038/nm979 Satterwhite, E., Sonoki, T., Willis, T. G., Harder, L., Nowak, R., Arriola, E. L., . . . Dyer, M. J. (2001). The BCL11 gene family: Involvement of BCL11A in lymphoid malignancies. Blood, 98(12), 3413-3420. doi: 10.1182/blood.V98.12.3413 Schaffer, B., Wiedau-Pazos, M., & Geschwind, D. (2003). Gene structure and alternative splicing of glycogen synthase kinase 3 beta (GSK-3β) in neural and non-neural tissues. Gene, 302(1-2), 73-81. doi: 10.1016/S0378-1119(02)01092-2 Schmid, S. L. (1997). Clathrin-coated vesicle formation and protein sorting: an integrated process. Annual Review of Biochemistry, 66, 511-548. doi: 10.1146/annurev.biochem.66.1.511

208 |Page

Schneider, J. L., & Cuervo, A. M. (2014). Autophagy and human disease: Emerging themes. Current Opinion in Genetics & Development, 26, 16-23. doi: 10.1016/j.gde.2014.04.003 Schumacher, J., Jamra, R. A., Becker, T., Ohlraun, S., Klopp, N., Binder, E. B., . . . Höfels, S. (2005). Evidence for a relationship between genetic variants at the brain-derived neurotrophic factor (BDNF) locus and major depression. Biological Psychiatry, 58(4), 307-314. doi: 10.1016/j.biopsych.2005.04.006 Scott, J., Stanton, B., Garland, A., & Ferrier, I. (2000). Cognitive vulnerability in patients with bipolar disorder. Psychological Medicine, 30(2), 467-472. doi: 10.1017/S0033291799008879 Sebastian, R., Díaz, E., Ayala, G., Díaz, M. E., Zoncu, R., & Toomre, D. (2006). Studying endocytosis in space and time by means of temporal Boolean models. Pattern Recognition, 39(11), 2175-2185. doi: 10.1016/j.patcog.2006.04.030 Senftleben, U., Cao, Y., Xiao, G., Greten, F. R., Krähn, G., Bonizzi, G., . . . Sun, S.-C. (2001). Activation by IKKα of a second, evolutionary conserved, NF-κB signaling pathway. Science, 293(5534), 1495-1499. doi: 10.1126/science.1062677 Sengupta, A., Novak, M., Grundke-Iqbal, I., & Iqbal, K. (2006). Regulation of phosphorylation of tau by cyclin-dependent kinase 5 and glycogen synthase kinase-3 at substrate level. FEBS Letters, 580(25), 5925-5933. doi: 10.1016/j.febslet.2006.09.060 Sengupta, S., Jayaraman, P., Chilton, P. M., Casella, C. R., & Mitchell, T. C. (2007). Unrestrained glycogen synthase kinase-3β activity leads to activated T cell death and can be inhibited by natural adjuvant. The Journal of Immunology, 178(10), 6083-6091. doi: 10.4049/jimmunol.178.10.6083 Sergeant, N., Bretteville, A., Hamdane, M., Caillet-Boudin, M.-L., Grognet, P., Bombois, S., . . . Vanmechelen, E. (2008). Biochemistry of Tau in Alzheimer's disease and related neurological disorders. Expert Review of Proteomics, 5(2), 207-224. doi: 10.1586/14789450.5.2.207 Serretti, A., Benedetti, F., Mandelli, L., Lorenzi, C., Pirovano, A., Colombo, C., & Smeraldi, E. (2003). Genetic dissection of psychopathological symptoms: Insomnia in mood disorders and CLOCK gene polymorphism. American Journal of Medical Genetics Part B: Neuropsychiatric Genetics, 121(1), 35-38. doi: 10.1002/ajmg.b.20053 Shah, S. M., Patel, C. H., Feng, A. S., & Kollmar, R. (2013). Lithium alters the morphology of neurites regenerating from cultured adult spiral ganglion neurons. Hearing Research, 304, 137-144. doi: 10.1016/j.heares.2013.07.001 Sharma, N., Kress, Y., & Shafit‐Zagardo, B. (1994). Antisense MAP‐2 oligonucleotides induce changes in microtubule assembly and neuritic elongation in pre‐existing neurites of rat cortical neurons. Cell Motility and the Cytoskeleton, 27(3), 234- 247. doi: 10.1002/cm.970270305 Shaw, M., & Cohen, P. (1999). Role of protein kinase B and the MAP kinase cascade in mediating the EGF-dependent inhibition of glycogen synthase kinase 3 in Swiss 3T3 cells. FEBS Letters, 461(1), 120-124. doi: 10.1016/S0014-5793(99)01434-9 Shaw, P. J., Cirelli, C., Greenspan, R. J., & Tononi, G. (2000). Correlates of sleep and waking in Drosophila melanogaster. Science, 287(5459), 1834-1837. doi: 10.1126/science.287.5459.1834 Sheppard, K.-A., Rose, D. W., Haque, Z. K., Kurokawa, R., McInerney, E., Westin, S., . . . Collins, T. (1999). Transcriptional activation by NF-κB requires multiple coactivators. Molecular and Cellular Biology, 19(9), 6367-6378.

209 |Page

Shi, B., Conner, S. D., & Liu, J. (2014). Dysfunction of endocytic kinase AAK1 in ALS. International Journal of Molecular Sciences, 15(12), 22918-22932. doi: 10.3390/ijms151222918 Shi, J., Badner, J. A., Hattori, E., Potash, J. B., Willour, V. L., McMahon, F. J., . . . Liu, C. (2008). Neurotransmission and Bipolar Disorder: A Systematic Family-based Association Study. American Journal of Medical Genetics Part B: Neuropsychiatric Genetics, 147B(7), 1270-1277. doi: 10.1002/ajmg.b.30769 Shih, W., Gallusser, A., & Kirchhausen, T. (1995). A clathrin-binding site in the hinge of the 2 chain of mammalian AP-2 complexes. Journal of Biological Chemistry, 270(52), 31083-31090. doi: 10.1074/jbc.270.52.31083 Shimizu, E., Hashimoto, K., Okamura, N., Koike, K., Komatsu, N., Kumakiri, C., . . . Okada, S.-i. (2003). Alterations of serum levels of brain-derived neurotrophic factor (BDNF) in depressed patients with or without antidepressants. Biological Psychiatry, 54(1), 70-75. doi: 10.1016/S0006-3223(03)00181-1 Shimizu, T., Kagawa, T., Inoue, T., Nonaka, A., Takada, S., Aburatani, H., & Taga, T. (2008). Stabilized β-catenin functions through TCF/LEF proteins and the Notch/RBP-Jκ complex to promote proliferation and suppress differentiation of neural precursor cells. Molecular and Cellular Biology, 28(24), 7427-7441. doi: 10.1128/MCB.01962-07 Shtutman, M., Zhurinsky, J., Simcha, I., Albanese, C., D’Amico, M., Pestell, R., & Ben- Ze’ev, A. (1999). The cyclin D1 gene is a target of the β-catenin/LEF-1 pathway. Proceedings of the National Academy of Sciences, 96(10), 5522-5527. doi: 10.1073/pnas.96.10.5522 Silvestri, L., Caputo, V., Bellacchio, E., Atorino, L., Dallapiccola, B., Valente, E. M., & Casari, G. (2005). Mitochondrial import and enzymatic activity of PINK1 mutants associated to recessive parkinsonism. Human Molecular Genetics, 14(22), 3477-3492. doi: 10.1093/hmg/ddi377 Simons, J. P., Al-Shawi, R., Minogue, S., Waugh, M. G., Wiedemann, C., Evangelou, S., . . . Warner, T. T. (2009). Loss of phosphatidylinositol 4-kinase 2α activity causes late onset degeneration of spinal cord axons. Proceedings of the National Academy of Sciences, 106(28), 11535-11539. doi: 10.1073/pnas.0903011106 Sirisaengtaksin, N., Gireud, M., Yan, Q., Kubota, Y., Meza, D., Waymire, J. C., . . . Bean, A. J. (2014). UBE4B protein couples ubiquitination and sorting machineries to enable epidermal growth factor receptor (EGFR) degradation. Journal of Biological Chemistry, 289(5), 3026-3039. doi: 10.1074/jbc.M113.495671 Siuciak, J. A., Lewis, D. R., Wiegand, S. J., & Lindsay, R. M. (1997). Antidepressant- like effect of brain-derived neurotrophic factor (BDNF). Pharmacology Biochemistry and Behavior, 56(1), 131-137. doi: 10.1016/S0091- 3057(96)00169-4 Sjøholt, G., Ebstein, R., Lie, R., Berle, J. Ø., Mallet, J., Deleuze, J., . . . Bannoura, I. (2004). Examination of IMPA1 and IMPA2 genes in manic-depressive patients: Association between IMPA2 promoter polymorphisms and bipolar disorder. Molecular Psychiatry, 9(6), 621-629. doi: 10.1038/sj.mp.4001460 Sjøholt, G., Gulbrandsen, A., Løvlie, R., Berle, J., Molven, A., & Steen, V. (2000). A human myo-inositol monophosphatase gene (IMPA2) localized in a putative susceptibility region for bipolar disorder on chromosome 18p11. 2: Genomic structure and polymorphism screening in manic-depressive patients. Molecular Psychiatry, 5(2), 172-180. doi: 10.1038/sj.mp.4000681

210 |Page

Sjostrom, S. K., Finn, G., Hahn, W. C., Rowitch, D. H., & Kenney, A. M. (2005). The Cdk1 complex plays a prime role in regulating N-myc phosphorylation and turnover in neural precursors. Developmental Cell, 9(3), 327-338. doi: 10.1016/j.devcel.2005.07.014 Sklar, P., Ripke, S., Scott, L. J., Andreassen, O. A., Cichon, S., Craddock, N., . . . Blackwood, D. (2011). Large-scale genome-wide association analysis of bipolar disorder identifies a new susceptibility locus near ODZ4. Nature Genetics, 43(10), 977-983. doi: 10.1038/ng.943 Skowyra, D., Craig, K. L., Tyers, M., Elledge, S. J., & Harper, J. W. (1997). F-box proteins are receptors that recruit phosphorylated substrates to the SCF ubiquitin-ligase complex. Cell, 91(2), 209-219. doi: 10.1016/S0092- 8674(00)80403-1 Department of Health (2007). The mental health of Australians 2. Commonwealth of Australia, Canberra. Smith, C. A., Lau, K. M., Rahmani, Z., Dho, S. E., Brothers, G., She, Y. M., . . . Schweisguth, F. (2007). aPKC‐mediated phosphorylation regulates asymmetric membrane localization of the cell fate determinant Numb. The EMBO Journal, 26(2), 468-480. doi: 10.1038/sj.emboj.7601495 Song, L., De Sarno, P., & Jope, R. S. (2002). Central role of glycogen synthase kinase- 3β in endoplasmic reticulum stress-induced caspase-3 activation. Journal of Biological Chemistry, 277(47), 44701-44708. doi: 10.1074/jbc.M206047200 Song, X., Wang, S., & Li, L. (2014). New insights into the regulation of Axin function in canonical Wnt signaling pathway. Protein & Cell, 5(3), 186-193. doi: 10.1007/s13238-014-0019-2 Sorensen, E. B., & Conner, S. D. (2008). AAK1 regulates Numb function at an early step in clathrin‐mediated endocytosis. Traffic, 9(10), 1791-1800. doi: 10.1111/j.1600-0854.2008.00790.x Spengler, M. L., Kuropatwinski, K. K., Schumer, M., & Antoch, M. (2009). A serine cluster mediates BMAL1-dependent CLOCK phosphorylation and degradation. Cell Cycle, 8(24), 4138-4146. doi: 10.4161/cc.8.24.10273 Spillantini, M. G., & Goedert, M. (1998). Tau protein pathology in neurodegenerative diseases. Trends in Neurosciences, 21(10), 428-433. doi: 10.1016/S0166- 2236(98)01337-x Spittaels, K., Van den Haute, C., Van Dorpe, J., Geerts, H., Mercken, M., Bruynseels, K., . . . Boon, T. (2000). Glycogen synthase kinase-3β phosphorylates protein tau and rescues the axonopathy in the central nervous system of human four- repeat tau transgenic mice. Journal of Biological Chemistry, 275(52), 41340- 41349. doi: 10.1074/jbc.M006219200 Spittaels, K., Van Den Haute, C., Van Dorpe, J., Terwel, D., Vandezande, K., Lasrado, R., . . . Van Lint, J. (2002). Neonatal neuronal overexpression of glycogen synthase kinase-3β reduces brain size in transgenic mice. Neuroscience, 113(4), 797-808. doi: 10.1016/S0306-4522(02)00236-1 Stafford, J. W., Lynd, K. M., Jung, A. Y., & Gordon, M. D. (2012). Integration of taste and calorie sensing in Drosophila. The Journal of Neuroscience, 32(42), 14767- 14774. doi: 10.1523/jneurosci.1887-12.2012 Stambolic, V., Ruel, L., & Woodgett, J. R. (1996). Lithium inhibits glycogen synthase kinase-3 activity and mimics wingless signalling in intact cells. Current Biology, 6(12), 1664-1669. doi: 10.1016/S0960-9822(02)70790-2 Steinbrecher, K. A., Wilson, W., Cogswell, P. C., & Baldwin, A. S. (2005). Glycogen synthase kinase 3β functions to specify gene-specific, NF-κB-dependent

211 |Page

transcription. Molecular and Cellular Biology, 25(19), 8444-8455. doi: 10.1128/MCB.25.19.8444-8455.2005 Stevens, V. L., Ahn, J., Sun, J., Jacobs, E. J., Moore, S. C., Patel, A. V., . . . Hayes, R. B. (2010). HNF1B and JAZF1 genes, diabetes, and prostate cancer risk. Prostate, 70(6), 601-607. doi: 10.1002/pros.21094 Stopkova, P., Vevera, J., Paclt, I., Zukov, I., & Lachman, H. M. (2004). Analysis of SYNJ1, a candidate gene for 21q22 linked bipolar disorder: a replication study. Psychiatry Research, 127(1-2), 157-161. doi: 10.1016/j.psychres.2004.03.003 Straub, C., Granger, A. J., Saulnier, J. L., & Sabatini, B. L. (2014). CRISPR/Cas9- Mediated Gene Knock-Down in Post-Mitotic Neurons. PLOS ONE, 9(8), e105584. doi: 10.1371/journal.pone.0105584 Su, Y.-C., & Qi, X. (2013). Inhibition of excessive mitochondrial fission reduced aberrant autophagy and neuronal damage caused by LRRK2 G2019S mutation. Human Molecular Genetics, 22(22), 4545-4561. doi: 10.1093/hmg/ddt301 Sutherland, C. (2011). What are the bona fide GSK3 substrates? International Journal of Alzheimer's Disease, 2011, 505607. doi: 10.4061/2011/505607 Sutherland, C., Leighton, I. A., & Cohen, P. (1993). Inactivation of glycogen synthase kinase-3 beta by phosphorylation: new kinase connections in insulin and growth-factor signalling. Biochemical Journal, 296, 15-19. Swiech, L., Heidenreich, M., Banerjee, A., Habib, N., Li, Y., Trombetta, J., . . . Zhang, F. (2015). In vivo interrogation of gene function in the mammalian brain using CRISPR-Cas9. Nature Biotechnology, 33(1), 102-106. doi: 10.1038/nbt.3055 Sytnyk, V., Leshchyns' ka, I., Delling, M., Dityateva, G., Dityatev, A., & Schachner, M. (2002). Neural cell adhesion molecule promotes accumulation of TGN organelles at sites of neuron-to-neuron contacts. The Journal of Cell Biology, 159(4), 649-661. doi: 10.1083/jcb.200205098 Taira, N., Mimoto, R., Kurata, M., Yamaguchi, T., Kitagawa, M., Miki, Y., & Yoshida, K. (2012). DYRK2 priming phosphorylation of c-Jun and c-Myc modulates cell cycle progression in human cancer cells. The Journal of Clinical Investigation, 122(3), 859-872. doi: 10.1172/JCI60818 Takahashi, M., Tomizawa, K., Kato, R., Sato, K., Uchida, T., Fujita, S. C., & Imahori, K. (1994). Localization and developmental changes of τ protein kinase I/glycogen synthase kinase‐3β in rat brain. Journal of Neurochemistry, 63(1), 245-255. doi: 10.1046/j.1471-4159.1994.63010245.x Takahashi, M., Yasutake, K., & Tomizawa, K. (1999). Lithium inhibits neurite growth and tau protein kinase I/glycogen synthase kinase-3β-dependent phosphorylation of juvenile tau in cultured hippocampal neurons. Journal of Neurochemistry, 73(5), 2073-2083. doi: 10.1046/j.1471-4159.1999.02073.x Takao, T., Tachikawa, H., Kawanishi, Y., Mizukami, K., & Asada, T. (2007a). CLOCK gene T3111C polymorphism is associated with Japanese schizophrenics: A preliminary study. European Neuropsychopharmacology, 17(4), 273-276. doi: 10.1016/j.euroneuro.2006.09.002 Takao, Y., Yokota, T., & Koide, H. (2007b). β-Catenin up-regulates Nanog expression through interaction with Oct-3/4 in embryonic stem cells. Biochemical and Biophysical Research Communications, 353(3), 699-705. doi: 10.1016/j.bbrc.2006.12.072 Tan, C.-C., Yu, J.-T., Tan, M.-S., Jiang, T., Zhu, X.-C., & Tan, L. (2014). Autophagy in aging and neurodegenerative diseases: Implications for pathogenesis and therapy. Neurobiology of Aging, 35(5), 941-957. doi: 10.1016/j.neurobiolaging.2013.11.019

212 |Page

Tandon, A., Yu, H., Wang, L., Rogaeva, E., Sato, C., Chishti, M. A., . . . Davies, P. (2003). Brain levels of CDK5 activator p25 are not increased in Alzheimer's or other neurodegenerative diseases with neurofibrillary tangles. Journal of Neurochemistry, 86(3), 572-581. doi: 10.1046/j.1471-4159.2003.01865.x Tang, H., Shen, N., Jin, H., Liu, D., Miao, X., & Zhu, L.-Q. (2013). GSK-3β polymorphism discriminates bipolar disorder and schizophrenia: A systematic meta-analysis. Molecular Neurobiology, 48(3), 404-411. doi: 10.1007/s12035- 013-8414-x Tao, X., Finkbeiner, S., Arnold, D. B., Shaywitz, A. J., & Greenberg, M. E. (1998). Ca2+ influx regulates BDNF transcription by a CREB family transcription factor-dependent mechanism. Neuron, 20(4), 709-726. doi: 10.1016/S0896- 6273(00)81010-7 Tempé, D., Casas, M., Karaz, S., Blanchet-Tournier, M.-F., & Concordet, J.-P. (2006). Multisite protein kinase A and glycogen synthase kinase 3β phosphorylation leads to Gli3 ubiquitination by SCFβTrCP. Molecular and Cellular Biology, 26(11), 4316-4326. doi: 10.1128/MCB.02183-05 ter Haar, E., Coll, J. T., Austen, D. A., Hsiao, H.-M., Swenson, L., & Jain, J. (2001). Structure of GSK3β reveals a primed phosphorylation mechanism. Nature Structural & Molecular Biology, 8(7), 593-596. doi: 10.1038/89624 Tetsu, O., & McCormick, F. (1999). β-catenin regulates expression of cyclin D1 in colon carcinoma cells. Nature, 398(6726), 422-426. doi: 10.1038/18884 Thornton, T. M., Pedraza-Alva, G., Deng, B., Wood, C. D., Aronshtam, A., Clements, J. L., . . . Doble, B. (2008). Phosphorylation by p38 MAPK as an alternative pathway for GSK3β inactivation. Science, 320(5876), 667-670. doi: 10.1126/science.1156037 Thotala, D. K., Hallahan, D. E., & Yazlovitskaya, E. M. (2011). Glycogen synthase kinase 3β inhibitors protect hippocampal neurons from radiation-induced apoptosis by regulating MDM2-p53 pathway. Cell Death & Differentiation, 19(3), 387-396. doi: 10.1038/cdd.2011.94 Tian, Y., Chang, J. C., Fan, E. Y., Flajolet, M., & Greengard, P. (2013). Adaptor complex AP2/PICALM, through interaction with LC3, targets Alzheimer's APP- CTF for terminal degradation via autophagy. Proceedings of the National Academy of Sciences, 110(42), 17071-17076. doi: 10.1073/pnas.1315110110 Tian, Y., Chang, J. C., Greengard, P., & Flajolet, M. (2014). The convergence of endosomal and autophagosomal pathways: implications for APP-CTF degradation. Autophagy, 10(4), 694-696. doi: 10.4161/auto.27802 Todd, A., & Staveley, B. (2008). Pink1 suppresses alpha-synuclein-induced phenotypes in a Drosophila model of Parkinson's disease. Genome, 51(12), 1040-1046. doi: 10.1139/G08-085 Tokumitsu, H., Hatano, N., Inuzuka, H., Sueyoshi, Y., Yokokura, S., Ichimura, T., . . . Kobayashi, R. (2005). Phosphorylation of Numb family proteins possible involvement of Ca2+/calmodulin-dependent protein kinases. Journal of Biological Chemistry, 280(42), 35108-35118. doi: 10.1074/jbc.M503912200 Tokumitsu, H., Hatano, N., Yokokura, S., Sueyoshi, Y., Nozaki, N., & Kobayashi, R. (2006). Phosphorylation of Numb regulates its interaction with the clathrin- associated adaptor AP-2. FEBS Letters, 580(24), 5797-5801. doi: 10.1016/j.febslet.2006.09.043 Traub, L. M. (2009). Tickets to ride: Selecting cargo for clathrin-regulated internalization. Nature Reviews Molecular Cell Biology, 10(9), 583-596. doi: 10.1038/nrm2751

213 |Page

Treadway, M. T., & Zald, D. H. (2011). Reconsidering anhedonia in depression: Lessons from translational neuroscience. Neuroscience and Biobehavioral Reviews, 35(3), 537-555. doi: 10.1016/j.neubiorev.2010.06.006 Trivedi, N., Marsh, P., Goold, R. G., Wood-Kaczmar, A., & Gordon-Weeks, P. R. (2005). Glycogen synthase kinase-3β phosphorylation of MAP1B at Ser1260 and Thr1265 is spatially restricted to growing axons. Journal of Cell Science, 118(5), 993-1005. doi: 10.1242/jcs.01697 Trowbridge, J. J., Xenocostas, A., Moon, R. T., & Bhatia, M. (2006). Glycogen synthase kinase-3 is an in vivo regulator of hematopoietic stem cell repopulation. Nature Medicine, 12(1), 89-98. doi: 10.1038/nm1339 Tsai, C.-C., Kai, J.-I., Huang, W.-C., Wang, C.-Y., Wang, Y., Chen, C.-L., . . . Chen, S.-H. (2009). Glycogen synthase kinase-3β facilitates IFN-γ-induced STAT1 activation by regulating Src homology-2 domain-containing phosphatase 2. The Journal of Immunology, 183(2), 856-864. doi: 10.4049/jimmunol.0804033 Tsai, S. J., Liou, Y. J., Hong, C. J., Yu, Y. W., & Chen, T. J. (2008). Glycogen synthase kinase-3β gene is associated with antidepressant treatment response in Chinese major depressive disorder. The Pharmacogenomics Journal, 8(6), 384-390. doi: 10.1038/sj.tpj.6500486 Turenne, G. A., & Price, B. D. (2001). Glycogen synthase kinase3 beta phosphorylates serine 33 of p53 and activates p53's transcriptional activity. BMC Cell Biology, 2, 12. doi: 10.1186/1471-2121-2-12 Twig, G., Elorza, A., Molina, A. J., Mohamed, H., Wikstrom, J. D., Walzer, G., . . . Shirihai, O. S. (2008a). Fission and selective fusion govern mitochondrial segregation and elimination by autophagy. The EMBO Journal, 27(2), 433-446. doi: 10.1038/sj.emboj.7601963 Twig, G., Hyde, B., & Shirihai, O. S. (2008b). Mitochondrial fusion, fission and autophagy as a quality control axis: The bioenergetic view. Biochimica et Biophysica Acta, 1777(9), 1092-1097. doi: 10.1016/j.bbabio.2008.05.001 Tyagarajan, S. K., Ghosh, H., Yévenes, G. E., Nikonenko, I., Ebeling, C., Schwerdel, C., . . . Muller, D. (2011). Regulation of GABAergic synapse formation and plasticity by GSK3β-dependent phosphorylation of gephyrin. Proceedings of the National Academy of Sciences, 108(1), 379-384. doi: 10.1073/pnas.1011824108 Ueno, T., Tomita, J., Tanimoto, H., Endo, K., Ito, K., Kume, S., & Kume, K. (2012). Identification of a dopamine pathway that regulates sleep and arousal in Drosophila. Nature Neuroscience, 15(11), 1516-1523. doi: 10.1038/nn.3238 Uezu, A., Horiuchi, A., Kanda, K., Kikuchi, N., Umeda, K., Tsujita, K., . . . Nakanishi, H. (2007). SGIP1alpha is an endocytic protein that directly interacts with phospholipids and Eps15. Journal of Biological Chemistry, 282(36), 26481- 26489. doi: 10.1074/jbc.M703815200 Valente, E. M., Abou-Sleiman, P. M., Caputo, V., Muqit, M. M., Harvey, K., Gispert, S., . . . Healy, D. G. (2004). Hereditary early-onset Parkinson's disease caused by mutations in PINK1. Science, 304(5674), 1158-1160. doi: 10.1126/science.1096284 Valerio, A., Ghisi, V., Dossena, M., Tonello, C., Giordano, A., Frontini, A., . . . Carruba, M. O. (2006). Leptin increases axonal growth cone size in developing mouse cortical neurons by convergent signals inactivating glycogen synthase kinase-3β. Journal of Biological Chemistry, 281(18), 12950-12958. doi: 10.1074/jbc.M508691200 Viatour, P., Dejardin, E., Warnier, M., Lair, F., Claudio, E., Bureau, F., . . . Green, D. (2004). GSK3-mediated BCL-3 phosphorylation modulates its degradation and

214 |Page

its oncogenicity. Molecular Cell, 16(1), 35-45. doi: 10.1016/j.molcel.2004.09.004 Villavicencio, E. H., Walterhouse, D. O., & Iannaccone, P. M. (2000). The sonic hedgehog–patched–gli pathway in human development and disease. The American Journal of Human Genetics, 67(5), 1047-1054. doi: 10.1086/321187 Voight, B. F., Scott, L. J., Steinthorsdottir, V., Morris, A. P., Dina, C., Welch, R. P., . . . McCarthy, M. I. (2010). Twelve type 2 diabetes susceptibility loci identified through large-scale association analysis. Nature Genetics, 42(7), 579-589. doi: 10.1038/ng.609 Vyazovskiy, V. V., Cirelli, C., Pfister-Genskow, M., Faraguna, U., & Tononi, G. (2008). Molecular and electrophysiological evidence for net synaptic potentiation in wake and depression in sleep. Nat Neurosci, 11(2), 200-208. doi: nn2035 [pii] 10.1038/nn2035 Wada, A. (2009). Lithium and neuropsychiatric therapeutics: Neuroplasticity via glycogen synthase kinase-3beta, beta-catenin, and neurotrophin cascades. Journal of Pharmacological Sciences, 110(1), 14-28. doi: 10.1254/jphs.09r02cr Wajant, H., Pfizenmaier, K., & Scheurich, P. (2003). Tumor necrosis factor signaling. Cell Death & Differentiation, 10(1), 45-65. doi: 10.1038/sj.cdd.4401189 Wang, H., Yang, H., Shivalila, C. S., Dawlaty, M. M., Cheng, A. W., Zhang, F., & Jaenisch, R. (2013). One-Step Generation of Mice Carrying Mutations in Multiple Genes by CRISPR/Cas-Mediated Genome Engineering. Cell, 153(4), 910-918. doi: 10.1016/j.cell.2013.04.025 Wang, Y., Chan, S. L., Miele, L., Yao, P. J., Mackes, J., Ingram, D. K., . . . Furukawa, K. (2004). Involvement of Notch signaling in hippocampal synaptic plasticity. Proceedings of the National Academy of Sciences, 101(25), 9458-9462. doi: 10.1073/pnas.0308126101 Wang, Y. J., Wang, J., Sun, H. Q., Martinez, M., Sun, Y. X., Macia, E., . . . Yin, H. L. (2003). Phosphatidylinositol 4 phosphate regulates targeting of clathrin adaptor AP-1 complexes to the Golgi. Cell, 114(3), 299-310. doi: 10.1016/S0092- 8674(03)00603-2 Watcharasit, P., Bijur, G. N., Song, L., Zhu, J., Chen, X., & Jope, R. S. (2003). Glycogen synthase kinase-3β (GSK3β) binds to and promotes the actions of p53. Journal of Biological Chemistry, 278(49), 48872-48879. doi: 10.1074/jbc.M305870200 Watcharasit, P., Bijur, G. N., Zmijewski, J. W., Song, L., Zmijewska, A., Chen, X., . . . Jope, R. S. (2002). Direct, activating interaction between glycogen synthase kinase-3β and p53 after DNA damage. Proceedings of the National Academy of Sciences, 99(12), 7951-7955. doi: 10.1073/pnas.122062299 Wehr, T. A. (1992). Improvement of depression and triggering of mania by sleep deprivation. JAMA, 267(4), 548-551. doi: 10.1001/jama.267.4.548 Wei, J., Liu, W., & Yan, Z. (2010). Regulation of AMPA receptor trafficking and function by glycogen synthase kinase 3. Journal of Biological Chemistry, 285(34), 26369-26376. doi: 10.1074/jbc.M110.121376 Wei, M. C., Zong, W.-X., Cheng, E. H.-Y., Lindsten, T., Panoutsakopoulou, V., Ross, A. J., . . . Korsmeyer, S. J. (2001). Proapoptotic BAX and BAK: A requisite gateway to mitochondrial dysfunction and death. Science, 292(5517), 727-730. doi: 10.1126/science.1059108 Wei, W., Jin, J., Schlisio, S., Harper, J. W., & Kaelin Jr, W. G. (2005). The v-Jun point mutation allows c-Jun to escape GSK3-dependent recognition and destruction by

215 |Page

the Fbw7 ubiquitin ligase. Cancer Cell, 8(1), 25-33. doi: 10.1016/j.ccr.2005.06.005 Wei, Y. J., Sun, H. Q., Yamamoto, M., Wlodarski, P., Kunii, K., Martinez, M., . . . Yin, H. L. (2002). Type II phosphatidylinositol 4-kinase β is a cytosolic and peripheral membrane protein that is recruited to the plasma membrane and activated by Rac-GTP. Journal of Biological Chemistry, 277(48), 46586-46593. doi: 10.1074/jbc.M206860200 Weisshaar, B., Doll, T., & Matus, A. (1992). Reorganisation of the microtubular cytoskeleton by embryonic microtubule-associated protein 2 (MAP2c). Development, 116(4), 1151-1161. Welcker, M., & Clurman, B. E. (2008). FBW7 ubiquitin ligase: A tumour suppressor at the crossroads of cell division, growth and differentiation. Nature Reviews Cancer, 8(2), 83-93. doi: 10.1038/nrc2290 Welcker, M., Orian, A., Jin, J., Grim, J. A., Harper, J. W., Eisenman, R. N., & Clurman, B. E. (2004). The Fbw7 tumor suppressor regulates glycogen synthase kinase 3 phosphorylation-dependent c-Myc protein degradation. Proceedings of the National Academy of Sciences, 101(24), 9085-9090. doi: 10.1073/pnas.0402770101 Welcker, M., Singer, J., Loeb, K. R., Grim, J., Bloecher, A., Gurien-West, M., . . . Roberts, J. M. (2003). Multisite phosphorylation by Cdk2 and GSK3 controls cyclin E degradation. Molecular Cell, 12(2), 381-392. doi: 10.1016/S1097- 2765(03)00287-9 Wen, A. Y., Sakamoto, K. M., & Miller, L. S. (2010). The role of the transcription factor CREB in immune function. The Journal of Immunology, 185(11), 6413- 6419. doi: 10.4049/jimmunol.1001829 Weniger, M. A., Pulford, K., Gesk, S., Ehrlich, S., Banham, A. H., Lyne, L., . . . Barth, T. F. (2006). Gains of the proto-oncogene BCL11A and nuclear accumulation of BCL11A(XL) protein are frequent in primary mediastinal B-cell lymphoma. Leukemia, 20(10), 1880-1882. doi: 10.1038/sj.leu.2404324 Wiedemann, C., Schäfer, T., Burger, M. M., & Sihra, T. S. (1998). An essential role for a small synaptic vesicle-associated phosphatidylinositol 4-kinase in neurotransmitter release. The Journal of Neuroscience, 18(15), 5594-5602. William, W. J., Carvalho, G. B., Mak, E. M., Noelle, N., Fang, A. Y., Liong, J. C., . . . Benzer, S. (2007). Prandiology of Drosophila and the CAFE assay. Proceedings of the National Academy of Sciences, 104(20), 8253-8256. doi: 10.1073/pnas.0702726104 Williams, R. S., Cheng, L., Mudge, A. W., & Harwood, A. J. (2002). A common mechanism of action for three mood-stabilizing drugs. Nature, 417(6886), 292- 295. doi: 10.1038/417292a Williamson, R., van Aalten, L., Mann, D. M. A., Platt, B., Plattner, F., Bedford, L., . . . Sutherland, C. (2011). CRMP2 hyperphosphorylation is characteristic of Alzheimer's disease and not a feature common to other neurodegenerative diseases. Journal of Alzheimer's Disease, 27(3), 615-625. doi: 10.3233/jad- 2011-110617 Wong, E., & Cuervo, A. M. (2010). Autophagy gone awry in neurodegenerative diseases. Nature Neuroscience, 13(7), 805-811. doi: 10.1038/nn.2575 Wood-Kaczmar, A., Kraus, M., Ishiguro, K., Philpott, K. L., & Gordon-Weeks, P. R. (2009). An alternatively spliced form of glycogen synthase kinase-3β is targeted to growing neurites and growth cones. Molecular and Cellular Neuroscience, 42(3), 184-194. doi: 10.1016/j.mcn.2009.07.002

216 |Page

Wood, J. E., Schneider, H., & Rudd, C. E. (2006). TcR and TcR-CD28 engagement of protein kinase B (PKB/AKT) and glycogen synthase kinase-3 (GSK-3) operates independently of guanine nucleotide exchange factor VAV-1. Journal of Biological Chemistry, 281(43), 32385-32394. doi: 10.1074/jbc.M604878200 Woodgett, J. R. (1990a). Molecular cloning and expression of glycogen synthase kinase-3/factor A. The EMBO journal, 9(8), 2431. Woodgett, J. R. (1990b). Molecular cloning and expression of glycogen synthase kinase-3/factor A. The EMBO Journal, 9(8), 2431-2438. Wray, J., Kalkan, T., Gomez-Lopez, S., Eckardt, D., Cook, A., Kemler, R., & Smith, A. (2011). Inhibition of glycogen synthase kinase-3 alleviates Tcf3 repression of the pluripotency network and increases embryonic stem cell resistance to differentiation. Nature Cell Biology, 13(7), 838-845. doi: 10.1038/ncb2267 Wu, H., & Leng, R. P. (2011). UBE4B, a ubiquitin chain assembly factor, is required for MDM2-mediated p53 polyubiquitination and degradation. Cell Cycle, 10(12), 1912-1915. doi: 10.4161/cc.10.12.15882 Wu, H., Pomeroy, S. L., Ferreira, M., Teider, N., Mariani, J., Nakayama, K. I., . . . Leng, R. P. (2011). UBE4B promotes Hdm2-mediated degradation of the tumor suppressor p53. Nature Medicine, 17(3), 347-355. doi: 10.1038/nm.2283 Xilouri, M., Vogiatzi, T., Vekrellis, K., Park, D., & Stefanis, L. (2009). Abberant α- Synuclein Confers Toxicity to Neurons in Part through Inhibition of Chaperone- Mediated Autophagy. PLOS ONE, 4(5), e5515. doi: 10.1371/journal.pone.0005515 Xiong, N., Jia, M., Chen, C., Xiong, J., Zhang, Z., Huang, J., . . . Liang, Z. (2011). Potential autophagy enhancers attenuate rotenone-induced toxicity in SH-SY5Y. Neuroscience, 199, 292-302. doi: 10.1016/j.neuroscience.2011.10.031 Xiong, N., Xiong, J., Jia, M., Liu, L., Zhang, X., Chen, Z., . . . Luo, Z. (2013). The role of autophagy in Parkinson's disease: Rotenone-based modeling. Behavioral and Brain Functions, 9, 13. doi: 10.1186/1744-9081-9-13 Yamamoto, D. J., & Zahniser, N. R. (2012). Differences in rat dorsal striatal NMDA and AMPA receptors following acute and repeated cocaine-induced locomotor activation. PLoS One, 7(5), e37673. doi: 10.1371/journal.pone.0037673 PONE-D-12-01835 [pii] Yang, Y., Gehrke, S., Imai, Y., Huang, Z., Ouyang, Y., Wang, J. W., . . . Lu, B. (2006). Mitochondrial pathology and muscle and dopaminergic neuron degeneration caused by inactivation of Drosophila Pink1 is rescued by Parkin. Proceedings of the National Academy of Sciences, 103(28), 10793-10798. doi: 10.1073/pnas.0602493103 Yao, H.-B., Shaw, P.-C., Wong, C.-C., & Wan, D. C.-C. (2002). Expression of glycogen synthase kinase-3 isoforms in mouse tissues and their transcription in the brain. Journal of Chemical Neuroanatomy, 23(4), 291-297. doi: 10.1016/S0891-0618(02)00014-5 Yasuda, S., Liang, M., Marinova, Z., Yahyavi, A., & Chuang, D. (2007). The mood stabilizers lithium and valproate selectively activate the promoter IV of brain- derived neurotrophic factor in neurons. Molecular Psychiatry, 14(1), 51-59. doi: 10.1038/sj.mp.4002099 Yatham, L. N. (2003). Acute and maintenance treatment of bipolar mania: The role of atypical antipsychotics. Bipolar Disorders, 5(s2), 7-19. doi: 10.1111/j.1399- 2406.2003.00060.x Yazlovitskaya, E. M., Edwards, E., Thotala, D., Fu, A., Osusky, K. L., Whetsell, W. O., . . . Hallahan, D. E. (2006). Lithium treatment prevents neurocognitive deficit

217 |Page

resulting from cranial irradiation. Cancer Research, 66(23), 11179-11186. doi: 10.1158/0008-5472.can-06-2740 Yin, L., Wang, J., Klein, P. S., & Lazar, M. A. (2006). Nuclear receptor Rev-erbα is a critical lithium-sensitive component of the circadian clock. Science, 311(5763), 1002-1005. doi: 10.1126/science.1121613 Ying, Q.-L., Wray, J., Nichols, J., Batlle-Morera, L., Doble, B., Woodgett, J., . . . Smith, A. (2008). The ground state of embryonic stem cell self-renewal. Nature, 453(7194), 519-523. doi: 10.1038/nature06968 Yokley, J. L., Prasad, K. M., Chowdari, K. V., Talkowski, M. E., Wood, J., Gur, R. C., . . . Pogue-Geile, M. F. (2012). Genetic associations between neuregulin-1 SNPs and neurocognitive function in multigenerational, multiplex schizophrenia families. Psychiatric Genetics, 22(2), 70. doi: 10.1097/ypg.0b013e32834f352c Yoshimura, T., Arimura, N., Kawano, Y., Kawabata, S., Wang, S., & Kaibuchi, K. (2006). Ras regulates neuronal polarity via the PI3-kinase/Akt/GSK-3β/CRMP-2 pathway. Biochemical and Biophysical Research Communications, 340(1), 62- 68. doi: 10.1016/j.bbrc.2005.11.147 Yoshimura, T., Kawano, Y., Arimura, N., Kawabata, S., Kikuchi, A., & Kaibuchi, K. (2005). GSK-3β regulates phosphorylation of CRMP-2 and neuronal polarity. Cell, 120(1), 137-149. doi: 10.1016/j.cell.2004.11.012 Young, L. T., Bezchlibnyk, Y. B., Chen, B., Wang, J.-F., & MacQueen, G. M. (2004). Amygdala cyclic adenosine monophosphate response element binding protein phosphorylation in patients with mood disorders: Effects of diagnosis, suicide, and drug treatment. Biological Psychiatry, 55(6), 570-577. doi: 10.1016/j.biopsych.2003.10.023 Yu, F., Wang, Z., Tchantchou, F., Chiu, C.-T., Zhang, Y., & Chuang, D.-M. (2012). Lithium ameliorates neurodegeneration, suppresses neuroinflammation, and improves behavioral performance in a mouse model of traumatic brain injury. Journal of Neurotrauma, 29(2), 362-374. doi: 10.1089/neu.2011.1942 Yu, S. W., Baek, S. H., Brennan, R. T., Bradley, C. J., Park, S. K., Lee, Y. S., . . . Kim, S. W. (2008). Autophagic death of adult hippocampal neural stem cells following insulin withdrawal. Stem Cells, 26(10), 2602-2610. doi: 10.1634/stemcells.2008-0153 Yuan, J., Zhang, J., Wong, B., Si, X., Wong, J., Yang, D., & Luo, H. (2005). Inhibition of glycogen synthase kinase 3β suppresses coxsackievirus-induced cytopathic effect and apoptosis via stabilization of β-catenin. Cell Death & Differentiation, 12(8), 1097-1106. doi: 10.1038/sj.cdd.4401652 Yuan, Y., Tong, Q., Zhou, X., Zhang, R., Qi, Z., & Zhang, K. (2013). The association between glycogen synthase kinase 3 beta polymorphisms and Parkinson's disease susceptibility: A meta-analysis. Gene, 524(2), 133-138. doi: 10.1016/j.gene.2013.04.045 Zeng, X., Tamai, K., Doble, B., Li, S., Huang, H., Habas, R., . . . He, X. (2005). A dual- kinase mechanism for Wnt co-receptor phosphorylation and activation. Nature, 438(7069), 873-877. doi: 10.1038/nature04185 Zhang, F., Phiel, C. J., Spece, L., Gurvich, N., & Klein, P. S. (2003). Inhibitory phosphorylation of glycogen synthase kinase-3 (GSK-3) in response to lithium. Evidence for autoregulation of GSK-3. Journal of Biological Chemistry, 278(35), 33067-33077. doi: 10.1074/jbc.M212635200 Zhang, L., Schessl, J., Werner, M., Bonnemann, C., Xiong, G., Mojsilovic-Petrovic, J., . . . Kalb, R. (2008). Role of GluR1 in activity-dependent motor system development. J Neurosci, 28(40), 9953-9968. doi: 28/40/9953 [pii]

218 |Page

10.1523/JNEUROSCI.0880-08.2008 Zhang, Y., Lv, Y., Zhang, Y., & Gao, H. (2014). Regulation of p53 level by UBE4B in breast cancer. PLOS ONE, 9(2), e90154. doi: 10.1371/journal.pone.0090154 Zhou, F.-Q., & Snider, W. D. (2005). GSK-3β and microtubule assembly in axons. Science, 308(5719), 211-214. doi: 10.1126/science.1110301 Zhou, F.-Q., & Snider, W. D. (2006). Intracellular control of developmental and regenerative axon growth. Philosophical Transactions of the Royal Society B: Biological Sciences, 361(1473), 1575-1592. doi: 10.1098/rstb.2006.1882 Zhou, F.-Q., Zhou, J., Dedhar, S., Wu, Y.-H., & Snider, W. D. (2004). NGF-induced axon growth is mediated by localized inactivation of GSK-3β and functions of the microtubule plus end binding protein APC. Neuron, 42(6), 897-912. doi: 10.1016/j.neuron.2004.05.011 Zhou, J., Freeman, T. A., Ahmad, F., Shang, X., Mangano, E., Gao, E., . . . Woodgett, J. (2013). GSK-3α is a central regulator of age-related pathologies in mice. The Journal of Clinical Investigation, 123(4), 1821-1832. doi: 10.1172/JCI64398 Zhu, J. H., Guo, F., Shelburne, J., Watkins, S., & Chu, C. T. (2003). Localization of phosphorylated ERK/MAP kinases to mitochondria and autophagosomes in Lewy body diseases. Brain Pathology, 13(4), 473-481. doi: 10.1111/j.1750- 3639.2003.tb00478.x Zhu, L.-Q., Liu, D., Hu, J., Cheng, J., Wang, S.-H., Wang, Q., . . . Wang, J.-Z. (2010). GSK-3β inhibits presynaptic vesicle exocytosis by phosphorylating P/Q-type calcium channel and interrupting snare complex formation. The Journal of Neuroscience, 30(10), 3624-3633. doi: 10.1523/jneurosci.5223-09.2010 Zhu, L.-Q., Wang, S.-H., Liu, D., Yin, Y.-Y., Tian, Q., Wang, X.-C., . . . Wang, J.-Z. (2007). Activation of glycogen synthase kinase-3 inhibits long-term potentiation with synapse-associated impairments. The Journal of Neuroscience, 27(45), 12211-12220. doi: 10.1523/jneurosci.3321-07.2007

219 |Page