Targeting PTEN for Therapy in Cancer and Ptenopathies

University of South Florida Scholar Commons

Graduate Theses and Dissertations Graduate School

July 2019

Targeting PTEN for Therapy in Cancer and PTENopathies

Emily Palumbo University of South Florida, [email protected]

Follow this and additional works at: https://scholarcommons.usf.edu/etd

Part of the Cell Biology Commons, Medicine and Health Sciences Commons, and the Molecular Biology Commons

Scholar Commons Citation Palumbo, Emily, "Targeting PTEN for Therapy in Cancer and PTENopathies" (2019). Graduate Theses and Dissertations. https://scholarcommons.usf.edu/etd/8399

This Dissertation is brought to you for free and open access by the Graduate School at Scholar Commons. It has been accepted for inclusion in Graduate Theses and Dissertations by an authorized administrator of Scholar Commons. For more information, please contact [email protected].

Targeting PTEN for Therapy in Cancer and PTENopathies

Emily Palumbo

A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Pathology and Cell Biology Morsani College of Medicine University of South Florida

Major Professor: Vrushank Davé, Ph.D. Meera Nanjundan, Ph.D. Yu Chen, Ph.D. Subhra Mohapatra, Ph.D. Patricia Kruk, Ph.D.

Date of Approval: June 25th, 2019

Keywords: Phosphatase and Tensin Homolog, γ-AAPeptides, Allosteric activation, PTEN restoration therapy, PTEN protein aggregation

Acknowledgments

I have a multitude of people to thank for their support over the past several years of my graduate studies. First and foremost, I would like to acknowledge my advisor, Dr. Vrushank Davé. Dr. Davé embodies the true spirit of a mentor and advisor. His advocacy for the importance of student growth, both personally and professionally, is infallible. Dr. Davé interviewed me for admission to the PhD program at

USF, and his passion for science and discovery was tangible from that first conversation. His eagerness to train and mentor students to become productive scientists drives his success as an advisor and has personally inspired me to push myself in all aspects of my graduate school training. I can thank him endlessly for his writing critique, advice on designing experiments, and development of professional skills to be carried over into my future career. But most of all, I will remember his guidance in designing clean and succinct

PowerPoint slides (where we spent countless hours debating colors, font sizes, punctuation and word/figure placement), an exhausting but necessary process for each one of my presentations. I hope to adopt his kindness and good-willed nature in my future professional endeavors.

Dr. Davé created a lab environment which fostered fun, collaboration and scientific exploration.

His past students, Dr. Jaymin Kathiriya and Dr. Prerna Malaney, are exemplary models and I strived to live up to the expectations they set as students and scientists. They answered my endless questions as a juvenile graduate student and helped me navigate my project when it was just starting to form. They were eager to teach me wet-bench work and all tricks of the trade to excel in any lab. I am grateful for their wisdom and thankful for them creating such a fun, refreshing and productive lab environment. There have been a number of undergraduate volunteers who I have mentored, as well as master’s student Jacob Wilson. Each extra pair of hands contributed to the success and completion of my studies. I am grateful for the teaching opportunity and their dedication to and interest in my project.

I extend my gratitude to my committee members Dr. Meera Nanjundan, Dr. Yu Chen, Dr. Subhra

Mohapatra, and Dr. Patricia Kruk for their guidance in structuring my thesis work. Their scientific input helped me refine my critical thinking skills and approach each problem from different vantage points. I also thank my mentors Dr. Robert Deschenes, Dr. Vladimir Uversky for their advice on my project, critique of my publications and overall collaborations. Of course, I would like to thank my collaborators throughout

USF: the labs of Dr. Diane Allen-Gipson, Dr. Jianfeng Cai, Dr. H. Lee Woodcock and Dr. Bin Xue. I extend my gratitude to Michael Kemp and Fiona Kearns, two talented, bright graduate students. Furthermore, a well-deserved thank you must go to program administrators Shonique Edwards, Meagan Eastman and Dr.

Micheal Teng for their roles which contributed to my graduation and professional development.

From endless conversations about life, bingeing on comfort food, coffee dates, move nights, and nights out on the town, my friends inside and outside the program have never failed to provide the love and support I needed to keep going. Special thank you to Emma Griffith for listening to me rant (often) and

Lauren Terraglio for bringing me coffee when I was too busy to take off from work. My friends from the

2015 cohort, as well as those from other years, understand the most what we all encounter in graduate school. On my toughest days they made me laugh the most. We will always be able to relate to each other and build each other up in ways that our friends outside of this experience can’t, and I am forever grateful for these friendships.

One of the most important acknowledgements that I have to give is to my family. My grandfather’s fighting spirit provided me with the courage that I needed on my hardest days. He passed away from lung cancer in early 2018, but I remember him in every moment I need strength. My parents, despite knowing little about cell and molecular biology, were my greatest supporters. Words cannot express how lucky I feel to have the guidance that they graciously give me every day. Thank you to my mom, for listening to me during our weekly Sunday phone calls, and to my dad for sending light-hearted Facebook posts about cars and dogs to make me smile. I have cried, laughed, smiled and celebrated with them throughout this journey, they are the best, best friends in my life.

Table of Contents

List of Tables v

List of Figures vi

List of Abbreviations viii

Abstract xviii

Chapter 1: Introduction 1 Introduction to Phosphatase and Tensin Homolog, PTEN 1 PTEN in Cancer 3 PTEN and the Continuum Model of Tumor Suppression 5 PTENopathies 6 PTEN Structure 8 The N-terminus 9 The Phosphatase Domain 10 The C2 Domain 11 The C-terminus 12 PTEN Dimer Formation 13 PTEN Isoforms 14 Alternative Translational Proteoforms 15 Alternative Splice Proteoforms 15 PTEN Enzymatic Mechanism 16 PTEN Functions 17 Lipid Phosphatase Dependent Functions 17 Cell Proliferation and Survival 18 Cell Migration, Invasion and Polarity 18 Cell Metabolism 19 Cancer Stem Cells 19 Protein Phosphatase Dependent Functions 20 Nuclear functions 22 PTEN Nuclear Transport 22 Role of PTEN in Genome Stability and Cell Cycle 22 Role of PTEN in Chromatin Integrity and Transcriptional Regulation 23 Other Phosphatase Independent Functions 24 PTEN Regulation 24 Genomic Mechanisms 24 Mutations 25 Epigenetic Regulation 26 Transcriptional Regulation 27 Non-genomic Mechanisms 27 Post-transcriptional Regulation 28 Post-translational Regulation 28

i Oxidation 29 S-Nitrosylation 30 Acetylation 30 Methylation 31 SUMOylation 31 Ubiquitination 32 Phosphorylation 32 PTEN C-tail Phosphorylation 32 PTEN Phosphorylation of the Core Domains 34 Clinical Significance of PTEN Phopshorylation 35 PTEN Restoration Therapy 35 Summary 36

Chapter 2: Activating PTEN via Small Molecules 38 Background 38 Materials and Methods 39 FTMap Analysis 39 γ-AAPeptide Synthesis 39 Lipid Phosphatase Assays 40 Microtite-plate Screening 40 Enzyme Kinetics 40 Cancer Cell Culture Maintenance 41 Protein Isolation and Immuno-blotting 42 Cell Proliferation Assay 43 Cell Migration Assay via ECIS 43 Cell Cycle Analysis via Flow Cytometry 44 Results 44 Identifying a PTEN Primary Binding Hotspot 44 γ-AAPeptides Enhance PTEN Lipid Phosphatase Activity 47 A Select γ-AAPeptide Reduces Oncogenic Signaling in Cancer Cells 50 A Select γ-AAPeptide Inhibits Tumorigenic Features in Lung Cancer Cells 51 Conclusion 52

Chapter 3: Elucidation of γ-AAPeptide Mediated Enhancement of PTEN Activity 54 Background 54 Materials and Methods 55 Molecular Docking 55 Rigid Docking with DOCK 55 Induced Fit Docking with Schrodinger 55 Intrinsic Disorder Analysis 56 Molecular Dynamics Simulations 57 Results 58 Select γ-AAPeptides Target Unique Residues in the PD/C2D Interface 58 Exploratory Study via Rigid Docking 58 Validation via Induced Fit Docking 59 γ-AAPeptide Binding Alters PIP3 Binding Affinity and Orientation 60 Identifying Regions of PTEN that Induce Allostery 63 γ-AAPeptide Binding Alters Active Site Conformation 64 γ-AAPeptide Binding Stabilizes PTEN Protein Structure 65 Conformational Changes in R130 Dictate PIP3 Active Site Binding and Orientation 70

ii Conclusion 71

Chapter 4: Defining the Role of PTEN Aggregation Propensity in Diseases 72 Background 72 Materials and Methods 74 Curation of Clinically Relevant PTEN Mutations 74 TANGO Analysis of Wild-Type and Clinically Relevant PTEN Mutations 75 Hydrophobicity Analysis of Wild-Type and Clinically Relevant PTEN Mutations 76 Secondary Structure Analysis of Wild-Type and Clinically Relevant PTEN Mutations 76 Intrinsic Disorder Analysis of Wild-Type and Clinically Relevant PTEN Mutations 77 Statistical Analysis 77 Results 78 PTEN Contains Five Aggregation-Prone Regions 78 Alternative Translational PTEN Proteoforms Have a Distinct Aggregation-Prone Region 79 Curation of Clinically Relevant PTEN Mutations and Their Aggregation Propensities 80 Mutations that Increase Hydrophobicity Enhance Aggregation Propensity of PTEN are Frequent 81 Frameshift Mutations in PTEN Severely Affect its Aggregation Potential 83 Aggregation Propensity is Associated with Highly Ordered Structures 84 Contribution of Structural and Chemical Properties to PTEN Conformer Aggregation 85 Conclusion 87

Chapter 5: Discussion 88

References 93

Appendix A: Supplementary Data 109

Appendix B: Copyright Permissions 118

Appendix C: Published Manuscripts 120

iii

List of Tables

Table 1.1: Table of Diseases with Compromised PTEN Function 3

Table 2.1: List of Primary Antibodies 41

Table 2.2: Structures of Select γ-AAPeptides which Enhanced PTEN Lipid Phosphatase Activity 49

Table 3.1: Mutations in PTEN 1D5R PD/C2D Interface Alters γ-AAPeptide Binding Affinity 62

Table 3.2: Mutations in PTEN 1D5R PD/C2D Interface Disrupts γ-AAPeptide-induced Effects on PIP3 Binding 62

Table 4.1: PTEN Mutations that Change Aggregation Propensity Compared to PTEN Wild-Type 80

Table A1: HPLC purities and retention time of compounds 109

List of Figures

Figure 1.1: PTEN Regulates the PI3K Signaling Pathway 2

Figure 1.2: PTEN Follows the Continuum Model of Tumor Suppression 6

Figure 1.3: PTEN Structure 8

Figure 1.4: PTEN Proteoforms Comprising of Variants Generated by Alternative Translation and Splicing from the Same ORF 14

Figure 1.5: PTEN is Extensively Modified Post-translationally 29

Figure 1.6: Mechanisms of PTEN Inactivation 37

Figure 2.1: Overview of PTEN Structure 45

Figure 2.2: PTEN PD/C2D Interface is a Small Molecule Binding Hotspot 46

Figure 2.3: PTEN PD/C2D Interface is Suitable for Targeting with Small Molecules 46

Figure 2.4: γ-AAPeptide Synthesis 47

Figure 2.5: Select γ-AAPeptides Enhanced PTEN Lipid Phosphatase Activity 48

Figure 2.6: γ-AAPeptide #43 Enhances PTEN Activity 50

Figure 2.7: A Select γ-AAPeptide Antagonized PI3K Pathway Signaling in Cancer Cell Lines 51

Figure 2.8: A Select γ-AAPeptide Inhibits Oncogenic Potential of A549 Lung Cancer Cells 52

Figure 3.1: Rigid Docking Demonstrates Potential γ-AAPeptide #43-PTEN Interaction 58

Figure 3.2: Adamantyl Containing γ-AAPeptides Have a High Binding Affinity for the PD/C2D Interface 59

Figure 3.3: γ-AAPeptide Binding Influences PIP3 Active Site Binding and Orientation 61

Figure 3.4: γ-AAPeptides Bind PTEN at Regions Adjacent to Intrinsically Disordered Loops 63

Figure 3.5: Binding of γ-AAPeptides to the PD/C2D interface Changes Active Site Residue Conformations 64

Figure 3.6: γ-AAPeptide Binding to the PD/C2D Interface Stabilizes the Domain Interface and Active Site Pocket 66

v Figure 3.7: Overall Conformational Changes in PTEN Active Site Residues Following γ- AAPeptide Binding 67

Figure 3.8: Select γ-AAPeptide Binding Induces Selective Conformational Changes in Active Site Residues 68

Figure 3.9: R130 Contributes to Productive PIP3 Binding and Orientation 70

Figure 4.1: Pathway to Protein Aggregation 73

Figure 4.2: p53 and PTEN Properties Significantly Overlap 74

Figure 4.3: Putative Aggregation-Prone Regions of PTEN 78

Figure 4.4: PTEN Alternative Translational Proteoforms Have an Aggregation-Prone Region 80

Figure 4.5: Select Mutations in PTEN Wild-Type Protein Increase Aggregation Propensity 81

Figure 4.6: Increased PTEN Aggregation Propensity is Dominated by Hydrophobic Mutations 82

Figure 4.7: PTEN Frameshift Mutations Drastically Increase Aggregation Propensity 83

Figure 4.8: PTEN Mutants that Increase Aggregation Propensity Favor Ordered Structures 85

Figure 4.9: Both Hydrophobicity and Secondary Structure Influence the Degree of Aggregation Propensity 86

Figure A1: 1H-NMR data of select γ-AAPeptides 110

Figure A2: γ-AAPeptide #42 Inhibits PTEN Lipid Phosphatase Activity 115

Figure A3: γ-AAPeptide Binding Does Not Induce Conformational Changes in Select Active Site Residues 116

List of Abbreviations

4EBP1 Eukaryotic translation initiation factor 4E Binding Protein 1

AD Alzheimer's Disease

AIB1 Amplified in Breast Cancer 1

AKT Protein Kinase B

Ala (A) Alanine

ALOX5 5-lipoxygenase

AMPK AMP-activated Protein Kinase

APC/C Anaphase Promoting Complex aPKC Atypical Protein Kinase C

APSSP The Advanced Secondary Structure Prediction

Arg (R) Arginine

ASD Autism Spectrum Disorder

Asn (N) Asparagine

Asp (D) Aspartate asRNA antisense RNA

BAD Bcl-2 Associated Death Promoter

BMI-1 B cell-specific Moloney murine leukemia virus integration site 1

vii BRRS Bannayan Riley-Ruvacalba Syndrome

BS Binding Site

C2D C2 Domain

Ca2+ Calcium

CAV1 Calveolin 1

CBF1 Centromere Binding Protein 1 cBio Portal cBio Portal for Cancer Genomics

CBP CREB Binding Protein

Cdc42 Cell Division Cycle 42

CHIP C-terminus of Hsc70-interacting protein

CK2 Casein Kinase 2

COSMIC Catalog of Somatic Mutations in Cancer

COX-2 Cyclooxygenase-2

CREB CAMP Responsive Element Binding Protein 1

CS Cowden Syndrome

C-tail Carboxy-terminal tail

Cys (C) Cysteine

DNA Deoxyribonucleic acid

DTT Dithiothreitol

DUSP Dual Specificity Phosphatase

viii E2F1 E2F Transcription Factor 1

ECIS Electric Cell Substrate Impedance Sensing Measurement

ECM Extracellular Matrix

EGFR Epidermal Growth Factor Receptor

EGR1 Early Growth Response 1

ELM Eukaryotic Linear Motif

EMT Epithelial Mesenchymal Transition eNOS endothelial NOS

ENTPD5 Ectonucleoside Triphosphate Diphosphohydrolase 5

ERK Extracellular Signaling Related Kinases

EVI1 Ecotropic Virus Integration Site 1

FGFR2 Fibroblast Growth Factor Receptor 2

FOXO Forkhead Box

G1 Growth Phase 1

G2 Growth Phase 2

GlideSP Glide Standard Precision

GlideXP Glide eXtra Precision

Gln (Q) Glutamine

Glu (E) Glutamate

ix GLUT4 Glucose Transporter Type 4

Gly (G) Glycine

GSK3β Glycogen Synthase Kinase 3 Beta

HD Huntington's Disease

HDAC Histone Deacetylase

HEX Healthy Exosomes

HGMD Human Gene Mutation Database

His (H) Histidine

HPLC High Pressure Liquid Chromatography hy hypomorphic

ID1 Inhibitor of DNA Binding 1

IDP Intrinsically Disordered Protein

IDR Intrinsically Disordered Region

IFD Induced Fit Docking

IGF Insulin Growth Factor

Ile (I) Isoleucine

IRS1 Insulin Receptor Substrate 1

JARID1B Jumonji, AT Rich Interactive Domain 1B

KD Kyte-Doolittle

KIs Kinase Inhibitors

x KLF4 Krüppel-like Factor 4

KRAS Kirsten Rat Sarcoma Viral Oncogene Homolog

LDD Lhermitte Duclos-Disease

Leu (L) Leucine

Lys (K) Lysine

M Mitosis

MAPK Mitogen Activated Protein Kinase

MCM2 Minichromosome Maintenance Complex

MD Molecular Dynamics

MDM2 Mouse Double Minute 2 homolog

MEFs Mouse Embryonic Fibroblasts

Met (M) Methionine

MG Malachite Green

Mi-2/NuRD Nucleosome Remodeling Deacetylase miRNA microRNA

MKK4 MAP Kinase Kinase 4

MMAC1 Mutated in Multiple Advanced Cancers 1

MoRE Molecular Recognition Element

MoRF Molecular Recognition Feature

MSP58 58-kDa microspherule protein

xi MTF1 Metal-responsive Transcription Factor-1 mTOR Mammalian Target of Rapamycin

MVP Major Vault Protein

Ndfip1/2 NEDD4 family interacting protein 1/2

NDRG2 N-myc Downstream Regulated Gene 2

NDRG2 N-myc downstream regulated gene 2

NEDD4-1 Neural precursor cell expressed developmentally down-regulated protein 4-1

NEP Neutral Endopeptidase

NF-Kβ Nuclear Factor Kappa-light-chain-enhancer of Activated B Cells

NHERF1 Na+/H+ Exchanger Regulatory Factor

NMR Nuclear Magnetic Resonance

NOS1 Nitric Oxide Synthase 1

NSCLC Non-small Cell Lung Cancer

OTUD3 OTU Deubiquitinase 3 p53 Tumor Suppressor p53 p70S6K p70S6 Kinase

PAGE Polyacrylamide Gel Electrophoresis p-Akt Phosphorylated/phospho-AKT

PARK2 Parkinson's Disease 2

PBM PIP2 Binding Module

xii PBS Phosphate Buffered Saline

PCAF p300/CBP-associated factor

PD Phosphatase Domain

PD Parkinson's Disease

PDB Protein Data Bank

PDK1 Phosphoinositide-dependent Kinase 1

PDZ Post synaptic density protein (PSD95), Drosophila disc large tumor suppressor,

(Dlg1), and Zonula occludens-1 protein (ZO-1)

PEST Proline (P), Glutamic Acid (E), Serine (S), Threonine (T)

PFKFB3 6-phosphofructo-2-kinase/fructose-2,6-biphosphatase 3

Phe (F) Phenylalanine

PHTS PTEN Hamartoma Tumor Syndrome

PI Propidium Iodide

PI3K Phosphatidylinositol 3 Kinase

PICS PTEN-loss Induced Cellular Senescence

PICT-1 Protein Interacting with Carboxyl Terminus 1

PIP2 Phosphatidylinositol-4,5-phosphate

PIP3 Phosphatidylinositol-3,4,5-phosphate

PKC Protein Kinase C

Plk1/3 Polo-like Kinase 1/3

xiii PMSF Phenylmethysulfonyl Flouride

PONDR Predictor of Natural Disordered Regions

PP2A Protein Phosphatase 2A

PPARγ Peroxisome proliferator-activated receptor gamma

PPI Protein-Protein Interaction

PRAS40 Proline-rich Akt substrate of 40 kDa

PRMT6 Protein Arginine Methyltransferase 6

PS Proteus Syndrome

PTEN Phosphatase and Tensin Homolog

PTEN-L PTEN-Long

PTK6 Protein Tyrosine Kinase 6

PTM Post-translational Modification

PTP Protein-Tyrosine Phosphatase

R State Relaxed State

RAC1 Rac Family Small GTPase-1

RHEB Ras Homolog Enriched in Brain

RMSD Root Mean Square Deviation

RNA Ribonucleic acid

ROCK Rho-A associated Kinase

ROCK Rho-A associated kinase

xiv RTK Receptor Tyrosine Kinase

S Synthesis Phase

S6K Ribosomal protein S6 kinase beta-1

SALL4 Spalt Like Transcription Factor 4

SDS Sodium dodecyl sulfate

SDS-PAGE Sodium Dodecyl Sulfate-Polyacrylamide Gel Electrophoresis

SE Standard Error

Ser (S) Serine

SIRT1 Sirtuin 1

SLiM Short Linear Motif

SMYD2 SET And MYND Domain Containing 2

SNO-PTEN S-nitrosylated PTEN

SOX2 SRY (sex determining region Y)-box 2

Sp1 Specificity Protein 1

SPRY2 Sprouty RTK Signaling Antagonist 2

SREBP1C Sterol Regulatory Element Binding Protein 1C

SRF Serum Response Factor

SUMO Small Ubiquitin-related Modifier

T State Tense State

TCGA The Cancer Genome Atlas

xv TET1 Ten-eleven translocation methylcytosine dioxygenase 1

TGF-β Transforming Growth Factor β

Thr (T) Threonine

TNF-ɑ Tumor Necrosis Factor ɑ

Trp (W) Tryptophan

TSC Tuberous Sclerosis Complex

Tyr (Y) Tyrosine

USP13 Ubiquitin Specific Peptidase 13

UTR Untranslated Region

Val (V) Valine

VHR Vaccinia H1-related

VSMC Vascular Smooth Muscle Cell

WT Wild-type

WWP2 WW Domain Containing E3 Ubiquitin Protein Ligase 2

XIAP X-Linked Inhibitor of Apoptosis

γH2AX gamma Histone 2AX

xvi

Abstract

PTEN, a dual protein and lipid phosphatase, regulates a myriad of cellular functions including PI3K pathway signaling, cell migration, proliferation, invasion and apoptosis. PTEN mutations often lead to multiple malignancies, including prostate, breast, endometrial, skin and brain cancers, associated with hyperactive PI3K signaling. PTEN mutations have also been associated with a variety of other diseases, classified as PTEN Hamartoma Tumor Syndromes (PHTS). In addition, compromised function or reduced expression of PTEN due to non-genomic mechanisms are associated with many types of hyperproliferative diseases, such as restenosis and neoplastic diseases, including melanoma, lung, breast, prostate and colon cancers. Although PI3K pathway inhibitors are the first line of targeted therapy in these cancers, off- target effects, toxicities, and chemoresistance persists in the clinic, warranting development of alternative therapeutic paradigms. Consequently, restoration of PTEN expression and/or function has increasingly become an attractive option. Cell-based and pre-clinical studies have demonstrated that reintroduction of

PTEN attenuates oncogenic signaling and tumorigenic potential. Unfortunately, restoration methods involve expression of a gene, mRNA or protein as therapy, which are difficult to implement in the clinic.

My thesis attempts to fill this therapy gap by developing a novel small-molecule compound that can easily be administered to patients. I have identified small molecules that enhance endogenous PTEN activity and reduce oncogenic potential of cancer cells. This approach has engendered a novel avenue for PTEN restoration therapy that will likely serve as an adjunctive treatment for patients with PTEN-related diseases.

To begin to define molecules that can enhance endogenous PTEN activity, I first identified a likely binding hotspot for small molecules on PTEN between its two core domains, and determined that targeting of this hotspot via an α-helical peptide can modulate PTEN lipid phosphatase activity. Furthermore, in vitro biochemical assays determined a lead compound to enhance lipid phosphatase activity. This compound

xvii suppressed the canonical PI3K signaling cascade and inhibited cell proliferation, migration and cell cycle progression of A549 lung cancer cells. Computational docking and molecular dynamics simulations revealed that binding of this small molecule can induce conformational changes in the active site that enhance binding affinity and orientation of PTEN substrate, PIP3, in the catalytic pocket. Thus, I have identified and discerned the likely mechanism of the first known small molecule that directly binds and enhances PTEN enzymatic activity in vitro and asserts its biological functions in cancer cells.

Mutations which compromise PTEN phosphatase activity have been well-studied; however, little is known about other clinically relevant PTEN mutations which alter its structure and function. Indeed,

PTEN has overlapping clinical significance to tumor suppressor p53 and both molecules form multimers in vivo. Since select p53 mutant proteins are inactivated via an aggregation mechanism, I hypothesized that likewise, select clinically relevant PTEN mutations may aggregate and inactivate PTEN function, driving disease pathogenesis. To this end, I curated 1,523 known PTEN mutations in cancers and PHTS that were derived from various patient databases and assessed their potential to alter chemical properties, secondary protein structure, and aggregation propensity. I identified >250 mutations that increase the aggregation propensity of PTEN wild-type protein and determined that β-sheet structure and increased hydrophobicity greatly enhanced the potential of PTEN to become aggregation-prone. Validation of aggregation of these mutants in vivo will aid in the development of small molecule-based therapies that can target specific aggregation-prone PTEN regions, re-activating endogenous PTEN. In summary, my work implicates small molecule-based therapy that enhances endogenous PTEN activity as a potentially beneficial alternative or adjunctive therapy for patients suffering from diseases associated with hyperactive PI3K signaling, either due to PI3K pathway aberrations or compromised PTEN function.

xviii

Chapter 1: Introduction

Introduction to Phosphatase and Tensin Homolog, PTEN

In 1997, several groups identified human chromosome 10, region q23-24 to be highly deleted/mutated across various types of cancer [23-25]. It comprised of a gene with an open reading frame, which was named PTEN (Phosphatase and Tensin Homolog) due to its homology to the superfamily of

PTPs (Protein Tyrosine Phosphatases) and chicken tensin [23-26]. PTEN has DUSP (dual specificity protein phosphatase) functions, with a preference for highly acidic substrates, which also allows for its additional function as a lipid phosphatase targeting the PI3K (Phosphatidylinositol 3 Kinase) pathway [27-

29]. PTEN dephosphorylates PIP3 (phosphatidylinositol-3,4,5-phosphate) and converts it to PIP2

(phosphatidylinositol-4,5-phosphate), thereby modulating the PI3K signaling pathway (Figure 1.1) [27, 28,

30, 31]. Thus, PTEN as a lipid phosphatase inhibits PI3K signaling, reducing activation of downstream effector kinases AKT (Protein Kinase B/PKB), mTOR (Mammalian Target of Rapamycin), and p70S6K

(p70S6 Kinase) (Figure 1.1) [4, 32]. Apart from its dual phosphatase functions, PTEN also has phosphatase- independent tumor suppressor functions [32, 33].

Compromised PTEN activity is not only associated with cancers, but also with other PTEN-related diseases, coined PTENopathies. These diseases include PHTS (PTEN Hamartoma Tumor Syndromes), like

CS (Cowden Syndrome), and ASD (Autism Spectrum Disorder) with associated macrocephaly [1, 2, 9, 13].

PTEN knockout mouse models demonstrated its role in embryonic development and disease progression in multiple organs. PTEN-/- mice are embryonically lethal, whereas PTEN+/- mice develop intestinal polyps, lymphoma/leukemia, and neoplasia in several organs, including thyroid, prostate, endometrium, and skin

[5, 6, 34]. These phenotypes recapitulate disease progression in humans with PHTS, associated with multiple cancers. The PTEN-/- mice have brain enlargement due to increased soma size in neurons and

Figure 1.1: PTEN Regulates the PI3K Signaling Pathway. Select growth factors via their membrane-bound growth factor receptors induce phosphorylation and activation of PI3 Kinase. Activated PI3K phosphorylates its substrate, PIP2, generating PIP3. PIP3, an oncogenic signaling molecule, stimulates the phosphorylation and activation of downstream effector kinases. PTEN directly antagonizes this pathway by dephosphorylating PIP3 back to PIP2, resulting in suppression of cell growth, cell proliferation and sensitization to apoptosis. astrocytes. Additionally, these astrocytes are hyperproliferative [35]. PTEN haploinsufficiency in stem/progenitor cells increases migration, invasion and decreases susceptibility to apoptosis [36]. Complete

PTEN deletion in stem/progenitor cells also confers increased cell proliferation [37]. Overall, systemic, organ or cell specific in vivo reduction or loss of PTEN contributes to multiple PTENopathies.

Compromised PTEN function is implicated in many diseases, therefore, I will first review the various facets of PTEN structure and function to understand how functions can be compromised. I will then illustrate the widespread inactivation of PTEN in disease, and current attempts in developing effective therapies and their limitations. Thus, the overall goal of my thesis work is to elucidate mechanisms by which activation of endogenous intracellular PTEN can be enhanced in a targeted manner in vivo for therapeutic purpose.

2 PTEN in Cancer

Because PTEN is a highly mutated and/or deleted tumor suppressor protein, its aberrant expression or compromised function continues to be elucidated in various malignancies (Table 1.1). Although PTEN mutations are found in virtually all types of sporadic cancers, only endometrial, central nervous system, skin and prostate cancers are associated with high PTEN mutation propensity in the range of ~14-38% [4].

Table 1.1: Table of Diseases with Compromised PTEN Function.

Disease References

Cancer

Prostate [3-6]

Breast [3, 4, 8]

Endometrial [4, 8]

Glioma/CNS [4, 8]

Skin [4, 5, 8]

Colon [4, 5, 8]

Lung [4, 8, 10]

Lymphoma/Leukemia [6]

Thyroid [6, 8]

Pancreatic [4, 8]

Ovarian [4]

Esophageal [4]

Urinary Tract [4]

Stomach [4, 11]

Liver [4, 8]

3 Table 1.1 (continued).

Disease References

PTEN Hamartoma Tumor Syndromes

Cowden Syndrome [1, 2]

Bannayan Riley-Ruvacalba Syndrome [2, 9]

Lhermitte-Duclos Disease [12]

Autism Spectrum Disorder [13]

Other PTEN-related Diseases

Ischemia [14, 15]

Restenosis [16]

Alzheimer’s Disease [17]

Parkinson’s Disease [18]

Huntington’s Disease [19, 20]

Metabolic disorders (Diabetes/Obesity) [21, 22]

Conversely, many cancers demonstrate reduced PTEN expression, which is not attributed to a genomic aberration. For example, lung cancers have a signature loss of PTEN function or expression despite low frequency of mutation or allelic deletion [10, 38-44]. In a recent review, it was highlighted that many cancers such as breast, thyroid, endometrial, colorectal, lung, bladder, liver, pancreatic as well as melanomas, and gliomas are PTEN deficient due to promoter methylation, loss of heterozygosity or other unknown mechanisms [8]. Understanding the degree to which PTEN is down-regulated by non-genomic mechanisms will serve to better understand its role in disease initiation and progression.

The sheer frequency of reduced PTEN expression found across cancers suggests that quantitative

PTEN status may serve as a relevant prognostic marker. Indeed, loss of PTEN is correlated with late stage

4 disease and aggressive cancers, while PTEN positive cancers have a more favorable prognosis and higher survival rates [11, 45-47]. Many of these studies only address PTEN as a prognostic factor in the context of PTEN-positive and PTEN-negative patient samples; however, recent evidence suggests that the degree of PTEN function and its expression levels define the type and severity of a disease spectrum. This novel concept is explored in the following section.

PTEN and the Continuum Model of Tumor Suppression

Since PTEN loss of heterozygosity is more common than bi-allelic inactivation, and heterozygous

PTEN knockout models demonstrate disease phenotypes, PTEN haploinsufficiency is clinically relevant

[5, 6, 34, 48]. The concept of dose-dependent PTEN inactivation was further explored in PTEN hy

(hypomorphic) mouse models, whereby PTEN+/+ = 100% expression, PTENhy/+ = ~80% expression,

PTEN+/- = 50% expression and PTENhy/- = ~30% expression. As PTEN levels decreased, a corresponding increase in p-AKT (phosphorylated/activated AKT) was observed, as was the development of breast, uterine and prostate cancers (Figure 1.2) [3]. Interestingly, PTEN hy/- mice rescued embryonic lethality, albeit with incomplete penetrance, indicating that reduction of PTEN expression beyond 50% is lethal in some cases [3, 49]. This evidence led to the proposal of the continuum model of tumor suppression, which incorporates haploinsufficiency and quasi-insufficiency as a driver for disease, rather than discrete changes in expression (i.e. 100%, 50%, 0%) [50]. Furthermore, reduction in PTEN expression has been detected in patient samples, of which a high proportion of samples lacked DNA/RNA aberrations [51, 52]. Therefore, loss of heterozygosity due to allelic deletion is not the only factor which dictates PTEN expression levels.

It is thought that regulation of PTEN protein expression can occur through PTMs (post-translational modifications) which modulate PTEN degradation, localization, structure and function [32, 33, 38, 51]. The mechanisms of regulation at the genomic and non-genomic levels will be discussed in subsequent sections.

Figure 1.2: PTEN Follows the Continuum Model of Tumor Suppression. PTEN has been determined to assert its cellular functions as a haplo- and quasi-insufficient protein, where graded changes in PTEN expression levels along a continuum dictates disease type and severity in specific tissues. Small decreases in levels of PTEN expression in breast tissue (20% reduction) can drive breast malignancies, where more severe reduction, such as 50% or 70% reduction are minimally required to drive lung/uterine and prostate malignancies, respectively.

PTENopathies

In addition to sporadic PTEN mutations found in a wide variety of malignancies, PTEN germline mutations are associated with PHTS (Table 1.1). The most common PHTS include CS, BRSS (Bannayan-

Riley-Ruvacalba Syndrome), LDD (Lhermitte Duclos-Disease), PS (Proteus Syndrome) and Proteus-like

Syndrome [12, 13, 52]. Irrespective of specific clinical presentations, diagnosed germline mutations in

PTEN are classified as a PHTS. It is thought that up to 85% of patients with CS display PTEN germline mutations, whereas 60% of BRRS patients, 7-20% of PS patients and ~60% of Proteus-like Syndrome patients harbor germline PTEN mutations [2, 9, 12]. Occasionally, germline mutations within the canonical

PI3K pathway are observed in patients with these heritable syndromes when there are no detectable PTEN mutations, highlighting the importance that the PI3K pathway may play in disease initiation and progression.

6 Phenotypically, these diseases are characterized by the presence of hamartomas, or benign tumors and tissue overgrowth throughout the body, and neurological conditions like ASD or macrocephaly [12].

Also of note are the disease risks associated with these syndromes. For example, patients with PTEN germline mutations presenting as CS have an increased lifetime risk of developing organ specific cancers, like breast (85%), thyroid (35%) and endometrial (28%) as compared to the general population [12].

Furthermore, the risk of multiple neoplasia is increased dramatically, with greater than 90% of patients harboring PTEN germline mutations at risk of developing intestinal polyps [12, 53]. Additionally, it is suggested that ~10-20% of ASD cases with macrocephaly are positive for PTEN germline mutations [12,

54]. The presence of PTEN germline mutations in these diseases is not surprising, considering that PTEN heterozygous or conditional PTEN-null homozygous mice recapitulate similar disease phenotypes of various cancers, neoplastic lesions, metabolic disorders and cognitive impairments [5, 6, 12, 34, 35, 54].

The importance of normal PTEN functions have also been demonstrated in non PTEN-related diseases like AD (Alzheimer’s Disease), PD (Parkinson’s Disease), HD (Huntington’s Disease), ischemia, restenosis and select metabolic disorders [14-20, 22] (Table 1.1). Inhibition of PTEN restores cognitive function in AD mouse models or can serve as a neuroprotective event in PD [17, 18]. Upregulation of PTEN expression through p53 in medium spiny neurons is neuroprotective in early HD; however, increased PTEN expression in indirect pathway striatal projection neurons confers decreased long-term potentiation, resulting in unwanted motor movements [19, 20]. Thus, activation or inhibition of PTEN in specific cells types in HD models may influence disease progression. Inhibition of PTEN, as observed in mouse models, may serve as a protective mechanism against ischemic damage in patients [14, 15, 55]. Conversely, in cases of restenosis, or narrowing of arterial walls by VSMCs (vascular smooth muscle cells), PTEN overexpression was able to suppress cell proliferation and migration and induce apoptosis in VSMCs [16].

Whole body overexpression of PTEN in mice confers decreased body weight and lipogenesis, increased energy expenditure, and maintenance of insulin sensitivity throughout ageing, indicating that upregulation of PTEN may be therapeutic for metabolic disorders like obesity and diabetes [21, 22, 56]. Taken together,

7 intended inhibition or upregulation of PTEN may serve as promising disease therapies in the above- mentioned diseases associated with aberrant PI3K pathway signaling or PTEN function.

PTEN Structure

The PTEN protein is 403 amino acids in length and is composed of two core domains, the PD

(Phosphatase Domain) and C2D (C2 Domain), which are flanked by flexible N-terminal and C-terminal tails (Figure 1.3A) [26]. Sequence homology analysis identified regions in PTEN that have high homology to PTPs and DUSPs, mostly in the active site residues. Additionally, a majority of the PD has high sequence homology to chicken tensin and bovine auxillin, which aided in naming PTEN [26].

Figure 1.3: PTEN Structure. (A) Representation of full-length PTEN wild-type protein and its four domains. 1) PIP2 Binding Module (PBM), 2) Phosphatase Domain (PD), 3) C2 Domain (C2D) and 4) the C-terminal tail containing a PDZ Binding Domain (PDZ-BD). (B) 1D5R PTEN crystal structure, residues 7-353 Δ286-309 with the C-terminal tail added to the crystal structure. The active site comprised of the P-loop, TI-loop and WPD-loop. PTEN monomers can dimerize using a domain-swapping mechanism, where each monomer associates at the PDs and the C-tail interacts with the PD of the other monomer.

8 A few years after the discovery of PTEN as an important tumor suppressor protein, most of its structure was determined by X-ray crystallography methods. Since the N- and C-tail regions are highly intrinsically disordered, crystalizing PTEN with these tail regions remains challenging [57]. Additionally, the C2D contains a 24 amino acid intrinsically disordered loop which had to be removed to obtain PTEN crystals [26]. Thus, the solved crystal structure of PTEN comprises of amino acid residues 7-353 Δ286-309

(Figure 1.3B). This truncated PTEN protein structure was determined to have phosphatase activity similar to that of wild-type PTEN. Structural homology analysis using the PTEN crystal structure revealed that the

PD has structural similarities to the DUSP VHR. The C2D was found to be structurally similar to

Ca2+-dependent C2 Domains, which aids in membrane recruitment for signaling molecules like PI3K and

PKC; however, PTEN does not utilize Ca2+ binding ligands for its membrane recruitment [26]. The following subsections will cover a detailed overview of PTEN protein domain structures.

The N-terminus

The N-terminus of PTEN has been shown to play a vital role in PTEN lipid phosphatase activity at the membrane. PTEN residues 6-15 are a proposed PIP2 binding motif which targets PTEN to the membrane to dephosphorylate substrate PIP3 (Figure 1.3A) [58]. Indeed, several studies have shown that in Dictyostelium discoideum, PTEN binds PIP2 at the rear of the cell during chemotaxis, modulating directional sensing [59-61]. Interestingly, PIP2 has been shown to be an allosteric activator of PTEN, enhancing PTEN phosphatase activity against PIP3 with residues Lys13, Arg14 and Arg15 being important for this interaction [58, 62-64]. Furthermore, circular dichroism studies identified increased α-helical content upon PIP2 binding, which enhanced lipid phosphatase activity, confirming that PIP2 does activate

PTEN via an allosteric mechanism [65].

The PTEN N-terminus also contains multiple localization sequences which alter PTEN subcellular localization. Residues 20-25 (GFDLDL) are suggested to play a role in PTEN cytoplasmic localization, as

9 clinically-relevant mutations in these residues cause nuclear import [66]. These mutants retain lipid phosphatase activity but cannot antagonize PI3K pathway signaling because of their nuclear targeting, thus they lack growth suppression functions as compared to PTEN wild-type protein. Alternatively, deletion of residues 13-16 (LRRY) or mutation of Tyr26, Tyr27 and Tyr29 prevent PTEN nuclear accumulation, indicating that these two small N-terminal regions are likely nuclear localization sequences [67].

Interestingly, clinically relevant PTEN K13E mutant impedes nuclear import by abolishing monoubiquitination [68]. The role of monoubiquitination in PTEN subcellular localization is covered later in this chapter. Furthermore, PTEN cytoplasmic and nuclear functions, which are both tumor suppressive in nature, will be reviewed in the subsequent sections. Lastly, PTEN has been shown to be exported from the cell in exosomes, likely through monoubiquitination of Lys13 via NEDD4-1 (neural precursor cell expressed developmentally down-regulated protein 4-1) and Ndfip1 [69]. Decreased p-AKT levels and cell proliferation was observed in cells which were exosome recipients [69]. Taken together, the N-terminus of

PTEN plays a crucial role in PTEN phosphatase activity and localization.

The Phosphatase Domain

The PTEN PD is comprised of amino acid residues 7-185 and is highly homologous to other PTP and DUSPs, as it contains conserved active site residues [26]. The active site pocket is made up of three walls: the P-loop (residues 123-130), the WPD-loop (residues 91-94) and the TI-loop (residues 161-169)

(Figure 1.3B). The P-loop mimics the catalytic motif seen in PTPs and DUSPs, with a sequence

123-HCXXGXXR-130, where X is any amino acid residue [26]. The PTEN active site is distinct from its other related phosphatases because of a four amino acid residue insertion in the TI-loop (residues 163-166) which extends the active site, allowing for binding of its bulkier lipid substrate PIP3, in addition to phospho- serine, -threonine and -tyrosine residues. Positively charged residues in the TI-loop are also proposed to play a role in non-specific binding to the plasma membrane, and may serve as a nuclear exclusion motif

[67, 70]. The shape of the pocket, as well as the overall basic charge of the pocket, due to the presence of

10 His123 and Lys125, accommodates protein and lipid substrates. Lys128 and Arg130 attract the negatively charged substrates [26]. The catalytic mechanism of PTEN will be discussed later.

The C2 Domain

The PTEN C2D lies C-terminal to the PD, from residues 186-351 (Figure 1.3B) [26]. A majority of this domain is made of β-strands with an orientation that demonstrates structural homology to

Phospholipase Cδ1, Phosphokinase Cδ and Phospholipase A2 [26]. These latter three enzymes bind the plasma membrane in a Ca2+ dependent manner; however, PTEN lacks the residues necessary for

Ca2+-dependent binding. Instead, the PTEN CBR3 loop and cɑ2 helix have an abundance of charged residues which associate with phosphatidylserines in the plasma membrane [71]. Mutation of these regions abrogated membrane binding, and expression in PTEN-null U87MG glioblastoma cells increased proliferation and contact-independent cell growth, indicating the important contribution of the C2D to tumor suppressor function [26, 72].

The C2D also modulates PTEN nuclear import utilizing select residues. These residues interact with MVP (Major Vault Protein), which likely acts as a cytoplasmic-nuclear transporter [26, 72-74].

Furthermore, mutations in amino acids 265-269 (C2D) and residues 160-164 (PD) or 233-237 (C2D) abrogated PTEN’s interaction with MVP, resulting in nuclear localization defects [73]. Additionally, clinically relevant K289E PTEN mutant is excluded from the nucleus, likely by preventing monoubiquitination necessary for nuclear targeting [68]. In contrast, mutation to proposed nuclear localization sequences in a PTEN 1-375 mutant background, which has enhanced nuclear localization, increased nuclear accumulation, indicating that these sequences may contribute to PTEN nuclear export

[67]. These regions include R233/R234/K237, K263/M264/L265/K266/K267/K269 in the CBR3 loop, and

K327/N329/K330/K332/ R335 in the cɑ2 helix [67]. Thus, the C2D is necessary for plasma membrane and nuclear targeting.

11 Despite the C2D lacking a catalytic motif, select clinically relevant mutants abolish PTEN lipid phosphatase activity in vitro and are more susceptible to degradation in cells [75]. Notably, a truncated

PTEN mutant harboring the N-terminus, PD and part of the C2D did not display lipid phosphatase activity, indicating that the C2D is integral to lipid phosphatase activity [75]. Thus, select C2D mutations likely affect the conformation of the C2D or multiple PTEN domains, which in turn abolishes lipid phosphatase activity.

The C-terminus

The PTEN C-terminus is comprised of the PTEN C-tail from residues 351-403, which includes a

PDZ (Post synaptic density protein (PSD95), Drosophila disc large tumor suppressor, (Dlg1), and Zonula occludens-1 protein (ZO-1)) binding motif from residues 401-403 (Figure 1.3A) [26]. Extensive computational analysis confirmed that the C-tail is highly intrinsically disordered region (IDR), classifying

PTEN as an IDP (intrinsically disordered protein) [44, 57]. The PTEN C-tail IDR provides structural plasticity and has high preponderance of MoREs/MoRFs (molecular recognition elements and features) and

SLiMs/ELMs (short linear motifs and Eukaryotic linear motifs), which facilitates PPIs (protein-protein interactions). For example, approximately 400 PPIs were identified for PTEN and 60% of those interactions occur at the C-tail, and analyses of the primary/secondary interactome identify these PPIs in a myriad of signaling pathways [44, 57]. The PDZ binding motif alone facilitates important PPIs which contribute to

PTEN membrane translocation and stability [76, 77]. For example, the PTEN PDZ binding motif interacts with the PDZ domain in MAGI2/3 (membrane-associated guanylate kinase 2 and 3) to reduce AKT activation at the plasma membrane [76, 77]. Furthermore, PTEN utilizes its PDZ binding motif to build higher order PTEN-associated complexes, which have a variety of functions such as cell signaling, homeostasis, cell polarity and invasion, and neuronal regeneration [78-82]. In addition to the regulatory roles of these scaffolds, they also contribute to PTEN protein stability [81].

12 Participation of PTEN in scaffold structures greatly enhances its stability. Additionally, PTEN contains two putative PEST sequences within amino acid residues 350-375 and 379-396 [75]. Typically, these sequences target a protein for proteolytic degradation; however, deletion of these sequences decreases

PTEN expression levels in cells [75]. Thus, partial or whole loss of the C-tail reduces protein stability, likely by altering folding. Both the PTEN C-tail and PDZ binding motif are subject to PTMs which regulate intramolecular and intermolecular protein interactions. All PTEN PTMs are covered in later sections and select PTMs which alter domain interactions will be covered more extensively.

PTEN Dimer Formation

In 2014, PTEN was discovered to exist as a dimer [83]. Notably, the dimeric complex is catalytically more active than a PTEN monomer; however, heterodimerization between PTEN wild-type and mutants C124S, G129E, or R130G, which are catalytically dead against phosphoinositides, exerted a dominant negative effect and suppressed wild-type PTEN activity in the heterodimer [83]. Structural and computational studies on PTEN dimer formation revealed association of the PDs of two PTEN monomers, generating a dimer with coplanar membrane binding sites [84]. Furthermore, the membrane binding affinity of the dimer was greater than that of the monomer [83, 84]. Additional evidence suggested that phosphorylation inhibited dimer formation, implicating the role of a flexible C-tail in dimer formation [83,

84]. It is proposed that the flexible, unphosphorylated C-tail associates with the C2D of the other monomer, stabilizing dimer formation through a domain swapping mechanisms [84]. Unphosphorylated PTEN monomers were also shown to undergo oligomerization; however, the mechanism and structures have not been defined [83]. Thus, perturbations in PTEN protein structure via mutations or structural modifications via PTMs can disrupt dimer formation, resulting in loss of enzymatic activity.

13 PTEN Isoforms

PTEN has multiple isoforms, which are now considered as a subset of PTEN proteoforms [38].

They arise from alternative translation initiation or alternative splicing, with functions that range from tumor suppressive to oncogenic. Several non-canonical translation initiation codons exist 5’ upstream to the canonical AUG start site, which result in the translation of PTEN-L, -M, -N and -O protein products

(Figure 1.4A) [85]. Additionally, alternative splicing of PTEN mRNA generates ten distinct isoforms:

PTEN-delta, PTEN-B, DelE5, DelE6, PTEN-3a, -3b, -3c and PTEN-5a, -5b, -5c (Figure 1.4B) [86, 87].

This section focuses on the structures and functions of these naturally occurring alternative translational and alternative splice variants.

Figure 1.4: PTEN Proteoforms Comprising of Variants Generated by Alternative Translation and Splicing from the Same ORF. (A) PTEN has four alternative translational variants (L-O) which are generated from upstream alternative initiation codons. (B) PTEN has eight alternative splice variants which are truncated forms of PTEN wild-type and are differentially expressed in disease. The role of these PTEN variants in normal cellular homeostasis and disease is an emerging area of research.

14 Alternative Translational Proteoforms

All alternative translational isoforms are N-terminally extended and retain PTEN lipid phosphatase activity (Figure 1.4A) [85]. While PTEN-L, -M and -N are endogenously expressed in several cell lines, with PTEN-M being the most abundant, endogenous expression was not detected for PTEN-O [85]. Despite

PTEN-M being the most highly expressed of the alternative translational variants, PTEN-L (also known as

PTEN-Long) has unique structural and functional properties; hence the most well studied. Apart from

PTEN-L retaining the canonical lipid phosphatase activity, it has additional localization functions not found in PTEN wild-type protein. PTEN-L contains a sequence motif that confers its extracellular secretory function. Likewise, PTEN-L also has a sequence motif in its N-terminal extension that confers a cell penetrating function. Indeed, PTEN-L has been shown to be secreted and to re-enter the cells, as shown both in cell-based assays and in vivo [88]. Additional studies have demonstrated the ability of PTEN-L to localize to the inner membrane of the mitochondria, maintain mitochondria structure and function, and synergize with PTEN wild-type to increase cytochrome c oxidase activity, aiding in oxidative phosphorylation [89]. The cell-penetrating function of PTEN-L, i.e. its N-terminal signaling sequences, makes it a potent tumor suppressor that can be exploited for tumor suppressor restoration therapy in cancers with low PTEN expression or compromised function.

Alternative Splice Proteoforms

All ten PTEN splice variants result in a truncated PTEN protein (Figure 1.4B). PTEN-Δ and PTEN-

B include intronic sequences after exons 8 and 5, respectively, resulting in premature truncations only a few residues after the alternative splice site [87]. Interestingly, PTEN-Δ expression is observed in renal cell carcinoma, where its expression correlates with extended survival. Stable overexpression of this variant in renal carcinoma cell lines dampened AKT and MAPK (Mitogen Activated Protein Kinase) signaling and decreased cell migration capabilities, indicating that PTEN-Δ has similar functions to PTEN wild-type [90].

15 While DelE5 lacks part of exon 5, DelE6 lacks all of exon 6 [86]. Variants -3a, -3b, and -3c incorporate parts of intron 3 immediately 3’ of exon 3, while variants -5a, -5b, and -5c incorporate parts of intron 5 immediately 3’ of exon 5 [86]. Functional assessment of these variants determined that only PTEN-5a expression was able to decrease p-AKT levels and inhibit cyclin D1 promoter activity. Conversely, PTEN

-5b and -5c overexpression enhanced cyclin D1 promoter activity, indicating a potential oncogenic effect.

In patients with CS and BRRS, PTEN-5b is typically overexpressed and variants -3a and -3b are relatively under expressed, as compared to controls [91]. Notably, there is high turnover of these variants due to rapid protein degradation. By virtue of differential expression in various normal tissues, the role of these splice variants in disease has not been well characterized [86].

PTEN Enzymatic Mechanism

Shortly after the discovery of PTEN as a likely tumor suppressor protein that has homology to

PTPs, PTEN was discovered to dephosphorylate serine, threonine and tyrosine substrates, as well as phosphoinositide substrates [23, 27, 28, 86]. Although both protein and lipid phosphatase activities of

PTEN are tumor suppressive in nature, the canonical tumor suppressor function of PTEN is through regulation of the oncogenic PI3K signaling pathway modulated by its lipid phosphatase activity [29]. PTEN was discovered as a phosphatase for the D3 position of inositol rings, with its preferred substrate as

PI(3,4,5)P3 [28]. However, PTEN can dephosphorylate PI(3)P, PI(3,4)P2, and inositol(1,3,4,5)P4, albeit with a lesser efficiency [31].

The PTEN catalytic mechanism used to dephosphorylate its protein and lipid substrates is thought to be the same two step mechanism employed by PTPs. Upon stabilization of the anionic substrate by positively charged active site residues, like Lys125, Lys128 and Arg130 in the PTEN P-loop, the catalytic cysteine residue (Cys124 in PTEN) engages in nucleophilic attack on the D3 position phosphate to be removed [26, 92, 93]. An aspartate residue in the WPD-loop of the PTPs/DUSPs acts as a general acid and

16 allows cleavage of the phosphate group from the substrate/leaving group [93]. This general acid is thought to be Asp92 in PTEN; however, one study has shown that Asp92 is dispensable for phosphatase activity, implying that a different residue acts as the general acid [26, 94]. This first step creates a phospho-enzyme intermediate by transferring the phosphate to the thiol group on the cysteine [93]. In the second step, a water molecular hydrolyzes the phospho-enzyme bond, freeing inorganic phosphate to leave the active site, which is regenerated for another round of catalysis [93].

Since the main tumor suppressor function of PTEN is through its lipid phosphatase activity, it is important to recognize that catalysis is affected by plasma membrane binding. Specifically, binding of the

PIP2 binding module to PIP2 orients the active site for PIP3 hydrolysis [63]. As mentioned above, PIP2 binding also acts as an allosteric activator, changing PTEN from a less active Tense (T) state to a more active Relaxed (R) state [38]. Taken together, combination of a basic active site pocket, which has increased affinity for highly anionic substrates, combined with allosteric modulation via PIP2 binding at the plasma membrane confers preferential catalysis of PIP3 substrate.

PTEN Functions

Lipid Phosphatase Dependent Functions

The lipid phosphatase activity of PTEN is the most well-studied of all its functions. PTEN uses its

PIP2 binding module at its N-terminus and its C2D to correctly position itself at the plasma membrane to dephosphorylate PIP3 [26, 58, 64, 71]. Recent evidence suggests that the diffuse cytoplasmic distribution of PTEN is attributed to C2D mediated interaction with cytoplasmic endosomes, where PTEN exerts its lipid phosphatase activity [95]. The following subsections describe a variety of physiological processes that are regulated by PTEN lipid phosphatase activity.

17 Cell Proliferation and Survival

Dephosphorylation of lipid substrate PIP3 to PIP2 directly counteracts the function of class I PI3K

[28]. Growth factor interaction with cognate RTKs (receptor tyrosine kinases) induces RTK dimerization and recruits PI3K to the membrane where the p85 subunit interacts with the receptor or adaptor substrates and the catalytic p110 subunit phosphorylates PIP2 [96]. PIP3 generation recruits serine/threonine kinase

AKT to the plasma membrane where its phosphorylated at Thr308 by PDK1, then at Ser473 by mTORC2 for full activation [96, 97]. Activated AKT can phosphorylate TSC2, an inhibitor of RHEB, thereby allowing RHEB to stimulate mTOR [33, 98]. Alternatively, AKT exerts its kinase activity to inhibit

PRAS40, a negative regulator of mTORC1 [33, 99]. mTORC1 activation subsequently phosphorylates endpoint pathway effectors kinase p70S6K and 4EBP1 to stimulate protein translation and cell growth

(Figure 1.1) [33, 98, 100]. Interestingly, mTOR kinase, found in both mTOR complexes, can induce cellular senescence after PTEN loss via phosphorylating and stabilizing p53. Enhancing p53 stability results in p21Cip1/WAF1 upregulation and renders cells more susceptible to senescence rather than oncogenesis, a process named PICS (PTEN-loss induced cellular senescence) [33, 98, 100, 101]. Additional AKT substrates include FOXO proteins, GSK3β (Glycogen Synthase Kinase 3β), BAD, MDM2, p27Kip1, and p21Cip1/WAF1 to promote cell survival and cell proliferation [33, 100]. Therefore, dephosphorylation of PIP3 by PTEN antagonizes signaling outputs downstream of PI3K [32, 33, 102-104].

Cell Migration, Invasion and Polarity

Regulation of the PIP3 levels and the PI3K pathway plays critical roles in cell migration, invasion and polarity. PTEN deletion in mouse fibroblasts upregulates the functions of Rac1 and Cdc42 in modulating the actin cytoskeleton, while reintroduction of PTEN with lipid phosphatase activity inhibited cell migration [105]. Transcriptional repression of PTEN by BMI-1 enhances cell migration and EMT

(epithelial mesenchymal transition) via the activation of PI3K/AKT/GSK3β/Snail axis [106]. PTEN also

18 plays a significant role in both mammalian and Drosophila epithelial morphogenesis and polarity during organ development by localizing to and maintaining PIP2 levels at the apical plasma membrane [82, 107].

In mammalian cells, PTEN localization recruits Anax2, Cdc42 and aPKC to the apical membrane and PTEN loss results in improper lumen development [107].

Cell Metabolism

The PI3K/AKT pathway plays a crucial role in insulin-mediated glucose metabolism. Insulin stimulation activates the pathway and drives lipogenesis and the Warburg effect [32, 33]. Notably, PTEN loss also conveys similar features. For example, PTEN loss in hepatocytes induces PPARγ and SREBP1C transcription, likely by loss of regulator MAF1, leading to inflammation, fat deposition and hepatocellular carcinoma [108, 109]. SREBP target genes promote lipid biosynthesis and its upregulation is dependent on

AKT/mTORC1 signaling [110]. AKT activation also drives the Warburg effect by stimulating GLUT4 plasma membrane translocation for increased glucose uptake, and by upregulating ENTPD5, a regulator of aerobic glycolysis [111-114]. Interestingly, several studies have generated mice containing additional copies of the Pten gene, resulting in heightened PTEN expression levels. These mice were found to be cancer resistant and exhibited an overall tumor suppressive metabolic state, as established by small body size due to decreased cell proliferation, increased energy expenditure, and decreased lipid synthesis [21,

56]. Thus, one of the tumor suppression function of PTEN is to assert an anti-Warburg effect in cells.

Cancer Stem Cells

PTEN has also been shown to play a role in stem cell maintenance. For example, PTEN deletion in hematopoietic, neural, skin and intestinal stem cells enhances cell proliferation via progression through the cell cycle G0/G1 phases [37, 115-118]. In some of these cases, deletion of PTEN alone can contribute to disease pathology, such as epithelial hyperplasia in the skin, respiratory tract and lungs or formation of

19 polyps in the intestine [116, 117, 119, 120]. Hematopoietic cells depleted of PTEN eventually undergo exhaustion; however, this phenomenon is not observed in neural or skin stem cells, likely because they share similar cell origins [37, 115, 117]. PTEN deletion in the prostate confers an expansion of Sca1+ cells, which are associated with prostate stem cells phenotypes [121]. Although the frequency of PTEN mutations/deletions is low (~10%) in the lung epithelium, PTEN downregulation or phosphorylation- mediated inactivation in lung cancers is common [10]. Furthermore, conditional deletion of PTEN in the lungs displayed an increase in bronchioalveolar stem cells and expansion of neuroepithelial bodies comprising of pulmonary neuroendocrine cells, considered to harbor stemness properties that may contribute to tumorigenesis [119, 122-124]. Similar outcomes after PTEN deletion have been observed in other cell types such as breast and germ cells [117]. Understanding the role of PTEN in stem cells, and expansion of cancer stem cells in particular, is important as loss of PTEN may drive cancer initiation and progression through dysregulation of normal stem cell homeostasis.

Protein Phosphatase Dependent Functions

Although canonical function of PTEN is attributed to it’s the lipid phosphatase activity, its protein phosphatase activity also contributes to its tumor suppressive properties. The first protein substrate of PTEN was found to be FAK (Focal Adhesion Kinase) [30]. Dephosphorylation of FAK by PTEN can suppress cell migration and invasion, as well as sensitize cells to apoptosis after detachment from the ECM

(extracellular matrix) [30, 125, 126]. While dephosphorylation of FAK inhibits directional cell migration,

PTEN dephosphorylation of Shc, a player in the MAPK pathway, has been shown to attenuate random cell migration movements [127-129]. Since Shc is upstream in the Ras/MAPK pathway, these studies suggest that PTEN also acts as a regulator of the MAPK pathway from Shc to endpoint effector ERK (Extracellular

Signaling Related Kinases) [128]. PTEN also inhibits cell migration and invasion phenotypes in PTEN-null cancer cell lines by targeting protein substrates PTK6 and even itself at phosphorylation sites located in the

20 C2D [130-132]. In other instances, PTEN autodephosphorylation may regulate neuronal spine density, suggesting that autodephosphorylation has cell type specific functions [133].

PTEN has various other protein substrates which regulate processes like cell proliferation, cell structure formation, cell differentiation, growth factor and apoptotic related signaling. Cell proliferation was shown to be suppressed through PTEN-mediated dephosphorylation of c-Src [134]. Conditional loss of PTEN in the lung epithelia in a mouse model demonstrated ~3 fold increase in cyclin D1 and Ki-67 staining, both hallmarks of increased cell proliferation [119]. Motile cilia formation in mice was modulated through PTEN dephosphorylation of Dishevelled [135]. Dephosphorylation of transcription factor KLF4 regulates the differentiation of VSMCs [136]. PTEN also targets MTF-1 to upregulate its function in response to metal and oxidative stress response [137]. PTEN can suppress insulin-mediated glucose metabolism signaling by dephosphorylating IRS1 (insulin receptor substrate-1) or ERK [138]. Interestingly, dephosphorylation of ERK kinases by PTEN resulted in reduced downstream phosphorylation of ETS-2.

Depletion of PTEN in mouse mammary stromal fibroblasts causes an upregulation in ETS-2 phosphorylation, which was associated with ECM remodeling, angiogenesis and innate immune infiltration

[139, 140]. Thus, the protein phosphatase activity of PTEN can regulate the MAPK pathway to modulate growth signaling and the tumor microenvironment [138, 140]. Furthermore, PTEN protein phosphatase activity regulates EGFR (epidermal growth factor receptor) signaling by dephosphorylating Rab7, promoting endosomal trafficking of EGFR into lysosomes, and apoptotic signaling by releasing calcium via dephosphorylation of inositol 1,4,5-trisphosphate receptors at the endoplasmic reticulum and mitochondria-associated membranes [141-143]. Taken together, the protein phosphatase activity has a variety of tumor suppressor functions, which, while critically important, remains less explored.

21 Nuclear Functions

PTEN Nuclear Transport

PTEN functions in the nucleus in a phosphatase dependent and independent fashion. PTEN may enter the nucleus as proposed by: passive diffusion, Ran GTPase/importin-11 mediated shuttling, and MVP mediated import [67, 73, 144, 145]. PTEN ubiquitination and phosphorylation heavily modulate nuclear/cytoplasmic shuttling as discussed below under “Post-translational regulation.” In brief, monoubiquitination of PTEN augments nuclear import, while phosphorylation of the PTEN C-tail results in cytoplasmic retention [10, 68, 146-148]. Nuclear exclusion of PTEN is seen in proliferating cells, while quiescent cells have abundant nuclear PTEN, indicating that aggressive malignancies may promote PTEN nuclear exclusion to drive oncogenesis, as demonstrated by our lab and others [10, 33, 149, 150]. Nuclear export of PTEN is thought to be driven by acid ceramidase/sphingosine-1 phosphate/AKT signaling and mediated by Exportin 1, also known as CRM, however this area remains less understood [151]. The nuclear functions discussed below are integral tumor suppressor functions of PTEN.

Role of PTEN in Genome Stability and Cell Cycle

Nuclear PTEN exerts tumor suppression, mostly in a lipid phosphatase independent manner. PTEN re-expression in PTEN-null U87MG cells decreases the PIP3 pool at the plasma membrane, but not within the nucleus, indicating that PTEN promotes tumor suppression by other mechanisms [152]. Anchorage- independent growth is mediated by a nuclear PTEN pool that downregulates p70S6K through AMPK (AMP activated protein kinase) [153]. Protein phosphatase active mutant PTEN G129E had a higher nuclear distribution compared to wild-type and catalytically dead C124S [67]. Nuclear PTEN utilizes its protein phosphatase activity to halt replication fork progression, inhibit insulin signaling and prevent cell cycle progression [154, 155]. PTEN dephosphorylates CREB to inhibit cell proliferation and attenuate neuronal stem cell differentiation [156, 157]. Conversely, nuclear PTEN needs its lipid phosphatase activity to

22 sensitize cells to TNF-ɑ induced apoptosis [67]. Taken together, lipid and protein phosphatase functions likely have independent tumor suppressive activities in the nucleus.

Notably, PTEN has a wide variety of phosphatase independent functions in the nucleus through its

C-terminus [158]. PTEN binds to the E3 ligase APC/C and augments its interaction with CDH1 to target mitotic cyclins and kinases or glutaminase and PFKFB3 for degradation, promoting growth suppression and a tumor suppressive metabolic state [56, 159]. PTEN’s interaction with topoisomerase II stabilizes the protein so it can decatenate DNA in G2/M cell cycle phases and PTEN/CENP-C binding regulates centromere stability [160]. Furthermore, genotoxic stress stimulates PTEN to colocalize with DNA damage repair proteins Rad52 and γH2AX or activate AMPK-mediated autophagy [161, 162]. Thus, nuclear PTEN utilizes PPIs via its C-terminus to assert its tumor suppressive functions such as maintenance of genome stability and regulation of the cell cycle.

Role of PTEN in Chromatin Integrity and Transcriptional Regulation

Nuclear PTEN, via its PPIs, also influences gene transcription. PTEN interacts with E2F1 to target and inhibit E2F1-mediated gene promoter activity, including that of E2F1, cyclin E1 and cyclin D1, which suppresses cell proliferation [10]. PTEN, via binding to activated p53, stabilizes it, leading to enhanced p53 binding and transcription of p53-dependent genes, associates with BMI1 to derepress transcription of cell cycle modulators p16INK4 and p14ARF, binds SRF to modulate smooth muscle cell contractile gene transcription and interacts with AIB1 to promote its degradation, suppressing transcription of its oncogenic targets [163-167]. PTEN can directly interact with chromatin and act as a transcription factor, as it does to induce Rad51 expression, augmenting double stranded break repair, or repress the expression of genes involved in transcription [160, 168]. Additionally, PTEN interactions with chromatin or chromatin- associated histone proteins modulates chromatin condensation and heterochromatin stability [169, 170].

23 These PTEN functions at the chromatin attenuate the transcription of critical genes necessary for proliferation of cancer cells.

Other Phosphatase Independent Functions

PTEN has few other phosphatase-independent functions, the most prominent being its involvement in high molecular weight scaffolds that increase its stability and localize to the plasma membrane to augment lipid phosphatase activity [80]. These scaffold interactions include proteins hDLG, MAGI-1,

NHERF1, and NEP and are mediated through the PTEN PDZ binding motif located at the C-terminus [80,

171, 172]. PTEN also participates in individual PPIs in a phosphatase-independent manner. Binding of

MSP58 inhibits MSP58-driven oncogenic transformation in mouse embryonic fibroblasts (MEFs) [173].

Furthermore, PTEN binds to CAV1 and prevents its interaction with β-catenin. PTEN loss frees CAV1 to bind to β-catenin, which translocates to the nucleus to suppress the transcription of cell cycle inhibitor p16INK4A [174]. Taken together, these scaffolding functions contribute to tumor suppression outside of canonical PTEN phosphatase activities.

PTEN Regulation

There are several mechanisms of regulation which positively and negatively modulate PTEN function (Figure 1.5). In the coming sections I will explore these mechanisms in detail.

Genomic Mechanisms

PTEN function and expression can be regulated at the genomic level. PTEN mutations can alter its phosphatase activity, stability, subcellular localization or protein-protein interactions [32, 33]. Furthermore,

PTEN expression levels can be regulated via epigenetic mechanisms and association of transcriptional

24 activators or repressors with regulatory gene elements [32, 33]. The next two subsections explore genomic mechanisms of PTEN regulation in more detail.

Mutations

PTEN was initially discovered as a gene frequently mutated or deleted in advanced cancers [23-

25]. It is now known that sporadic PTEN mutations occur at high incidence in cancers of the breast, prostate, endometrium, skin (melanoma) and brain (glioma) [175-177]. Additionally, PHTS diseases were shown to have a high incidence of germline PTEN mutations, as described above in PTENopathies. An in-depth analysis of the mutations which manifest at the protein level demonstrates that all nine exons of PTEN are targeted for mutation. However, the C-tail of PTEN, encoded by exon nine, has a much lower susceptibility for mutations in cancer, whereas exons five, seven and eight have the highest incidence of mutations across multiple disease phenotypes [32, 176, 178, 179]. Furthermore, these highly susceptible regions also tend to be highly conserved, like the active site in exon five [177]. Thus, mutations like catalytically dead C124S, lipid phosphatase dead G129E/R and phosphatase inactive R130X are highly implicated in cancer and

PHTS. Furthermore, most clinically relevant PTEN mutations in the P-loop, TI-loop and WPD-loop are functionally inactive [180].

Pathogenic PTEN mutations range from missense, which can either totally or partially inactivate

PTEN function, to nonsense, insertion and deletion mutants which produce truncated protein products with decreased expression or function [32, 175, 176]. For instance, missense mutations in the catalytic motif often alter the active site pocket in shape or chemical environment, resulting in ablation of phosphatase activity [175, 176]. Select missense mutants arising from monoallelic mutations contain dominant negative properties by dimerizing with and inactivating PTEN wild-type monomers [83]. Furthermore, mutants like

K13E abrogate PTM sites, contributing to aberrant subcellular localization [176]. Alternatively, nonsense mutations in the PD or C2D often produce truncated PTEN with the active site intact; however, the stability

25 of the protein product is too low to be functionally relevant in the cell [32, 75]. Interestingly, mutations observed in patients with ASD typically do not ablate PTEN lipid phosphatase activity, while mutations observed in cancer and PHTS patients often show PTEN lipid phosphatase activity [180]. Thus, specific mutations contribute to different disease phenotypes through their regulation of PTEN function, expression and subcellular localization.

Mutations within intronic sequences or at splice sites have been identified in CS and BRRS patients, which disrupt normal PTEN mRNA splicing, resulting in exon skipping, inclusion of intronic sequences or premature truncation [181-183]. These various mutants have different functions, subcellular localization and stability [181-183]. For example, skipping of exon four severely reduces lipid phosphatase activity and stability at the protein level, while skipping of exon three significantly reduces protein phosphatase activity but enhanced the nuclear localization of the protein product [181]. Taken together, defects in splicing due to mutations can generate functionally compromised PTEN variants that contribute to pathogenicity in

PHTS.

Epigenetic Regulation

PTEN mRNA expression, and thus protein expression, can be regulated by epigenetic mechanisms like promoter methylation. Hypermethylation of the PTEN promoter can downregulate PTEN transcription and mRNA levels, as demonstrated in several types of cancers, including breast, thyroid, and lung cancer

[33, 184-186]. Furthermore, promoter methylation may serve as a prognostic factor for patient survival in cancers [187]. Alternatively, JARID1B is a histone demethylase that targets histones associated with active genes. JARID1B activity suppresses PTEN transcription and drives proliferation of cancer cells [188].

Additionally, PTEN transcriptional repression due to histone deacetylation likely serves as another mechanism of epigenetic regulation. Transcription factor SALL4 can associate with the histone deacetylase complex Mi-2/NuRD to bind and deacetylate the PTEN promoter, suppressing transcription [189]. Taken

26 together, epigenetic mechanisms like promoter methylation and histone deacetylation are upregulated in many cancers to repress PTEN transcription.

Transcriptional Regulation

Additional transcriptional regulation of PTEN is modulated via transcriptional activators/repressors, or small molecules, which activate or repress PTEN transcription. PPARγ, p53,

EGR1, IGF2, MAF1, CBF1, TET1, SPRY2 transcription factors have demonstrated the ability to enhance

PTEN transcription, in addition to the cytokine, resistin [190-198]. Select compounds, which include resveratrol, found in red wine, indole-3-carbinol, found in cruciferous vegetables, and quercetin, found in fruits and vegetables induce PTEN transcription through indirect mechanisms [199, 200]. PML/RARα,

BMI1, c-JUN, SLUG, SOX2, HES1, EVI1, ID1, SNAIL, NF-Κβ (induced by MKK4, IGF1 and

17β-estradiol), SMAD transcription factors (induced by TGF-β) and Sp1 (induced by OCT4) inhibit PTEN gene transcription [106, 189, 201-212]. Notably, TNF-α stimulated NF-Kβ signaling can augment or suppress PTEN gene transcription depending on the cellular environment [213, 214]. These mechanisms provide insight into the negative and positive regulations of PTEN expression, specifically in a disease context which lacks allelic loss or mutation of PTEN but exhibits deregulation of one or more of the aforementioned proteins [32].

Non-genomic Mechanisms

In addition to genomic mechanisms, PTEN expression, function, and subcellular localization can be modulated after transcription at the mRNA, as well as after translation at the protein level. These mechanisms will be explored in the next two subsections.

27 Post-transcriptional Regulation

Negative regulation of PTEN expression is observed at the level of PTEN mRNA, post- transcription. An abundance of small non-coding microRNAs (miRNAs) have been shown to target the

PTEN mRNA 3’ UTR (untranslated region) [32, 33]. Furthermore, miRNA targeting of the PTEN 3’ UTR is differentially regulated in various diseases like cancer, PHTS and metabolic diseases [215, 216]. For example, miR-19a/b and miR-21 target and deplete PTEN mRNA in multiple cancers [32, 215-220].

Additionally, there is evidence to suggest that oncogenes can upregulate miRNA expression. For instance,

Myc oncogene can induce the miR-17~92 cluster and miR-200c PTEN-targeting miRNAs in lymphoma or nasopharyngeal carcinoma, respectively [221, 222]. Thus, PTEN post-transcriptional regulation via miRNAs can be disease and tissue specific.

Interestingly, a PTEN pseudogene which has over 95% sequence identity to PTEN, denoted as

PTENP1, does not get translated and acts as a competing endogenous RNA [223]. Since the PTENP1 3’

UTR has high homology to PTEN, it can act as a sponge for miRNAs that target PTEN. Therefore, expression of PTENP1 can increase PTEN levels and activity by acting as a decoy for PTEN-targeting miRNAs [223]. Loss of PTENP1 correlates with decreased PTEN protein expression [32]. The PTENP1 gene locus also expresses PTENP1 asRNA (antisense RNA) transcripts α and β. PTENP1 asRNA α can directly bind to and regulate the PTEN promoter, thus modulating PTEN transcription [224]. PTENP1 asRNA β can sequester the PTENP1 transcript, thus regulating PTEN protein expression indirectly [224].

Taken together, transcription of PTENP1 can positively or negatively regulate PTEN mRNA expression levels.

Post-translational Regulation

PTEN is abundantly post-translationally modified, with modification sites spanning all four domains (Figure 1.5) [32, 33, 215]. Eight distinct types of PTMs have been shown to target PTEN. PTMs

28 affect PTEN structure, function, localization, and stability. Since these modifications can modulate PTEN function and expression, several of these posttranslational regulation mechanisms contribute to PTEN haploinsufficiency in disease (Figure 1.6) [32, 33, 215]. The following subsections will explore all known

PTEN post-translational modifications and their roles in PTEN protein modulation.

Figure 1.5: PTEN is Extensively Modified Post-translationally. Known PTEN modifications by eight distinct mechanisms: Phosphorylation, Ubiquitination, SUMOylation, Oxidation, Acetylation, S-Nitrosylation, Methylation and Ribosylation. PTEN PTMs occur across the full length of the protein, although most PTMs occur within the C2D and the C-tail. It is important to understand the breadth by which PTEN is post-translationally modified in disease, since these PTMs regulate PTEN enzymatic activity, structure, PPIs, and subcellular localization, resulting in overall modulation of PTEN function and stability.

Oxidation

Oxidation of PTEN forms a disulfide bond at the catalytic Cys124 residue with the adjacent Cys71 residue, inhibiting phosphatase activity (Figure 1.6) [225-231]. Inactivation via oxidation directly results in phosphorylation and activation of AKT and downstream signaling effectors. Interestingly, thioredoxin-1, which reduces and reactivates PTEN, has been shown to form disulfide bond with the C2D of PTEN, which prevents its interaction with the plasma membrane and suppresses lipid phosphatase activity, adding a layer of complexity to thioredoxin-mediated PTEN regulation [232, 233]. PTEN interaction with peroxiredoxin

I prevents PTEN oxidation [234]. Metabolism of arachidonic acid by enzymes COX-2 or ALOX5 drives

29 AKT pathway stimulation through oxidation and inactivation of PTEN, although the oxidative species targeting the cysteine residues is unknown [235]. Altogether, these studies determined that generation of reactive oxygen species can enhance AKT pathway signaling via inactivation of PTEN through oxidation of active site residues.

S-Nitrosylation

PTEN is S-nitrosylated at Cys83, with Cys71 and Cys124 (Figure 1.6) [236-239]. S-nitrosylation of PTEN (SNO-PTEN) reduced its lipid phosphatase activity in vitro and decreased PTEN stability in cell- based assays via enhanced ubiquitin-mediated proteasomal degradation [236, 240]. Thus, formation of

SNO-PTEN is associated with enhanced p-AKT levels and cell survival. SNO-PTEN has been detected during brain ischemia and during an early phase of AD known as mild cognitive impairment. Furthermore,

S-nitrosylation of PTEN has been implicated as an inactivation mechanism in cancer. Depletion of PARK2 or activation of AMPK increases SNO-PTEN levels leading to enhanced AKT signaling and tumorigenesis

[241]. SNO-PTEN generation through NOS1 also prevents excessive autophagy and promotes survival through the PI3K/AKT pathway in nasopharyngeal carcinoma cells [240]. Taken together, S-nitrosylation of PTEN is an inactivation mechanism that promotes cell survival via AKT/mTOR signaling in multiple cell types.

Acetylation

Acetylation of PTEN plays multiple roles depending on the acetylation site. For example, PCAF

(P300/CBP-associated factor) driven acetylation at Lys125 and Lys128 within the catalytic cleft inactivates

PTEN phosphatase activity and upregulates AKT phosphorylation (Figure 1.6) [242, 243]. On the contrary, acetylation at Lys163 enhances PTEN function via membrane translocation (Figure 1.6) [244]. CREB binding protein (CBP) acetylates PTEN at lysine 402, augmenting interactions with proteins containing

PDZ domains like MAGI-2 and hDLG, which are known modulators of cell structure and polarity (Figure

30 1.6) [245]. Indeed, PTEN binds MAGI/Par-3 complex and participates in the maintenance of cell shape and polarity [82]. Therefore, it is likely that PTEN acetylation may directly regulate cell shape, behavior and polarity. Histone deacetylases like SIRT1, HDAC6, HDAC I and HDAC II counteract PTEN acetylation

[82, 242-244, 246]. Thus, modulation of PTEN acetylation directly affects enzymatic activity, subcellular localization and PDZ binding motif-mediated PPIs.

Methylation

Recently, PTEN has been shown to be methylated at lysine and arginine residues with distinct functional consequences. Methylation by SMYD2 at Lys313 enhances PTEN phosphorylation at Ser380, causing a closed/inactive conformation (Figure 1.6) [247]. On the contrary, methylation by PRMT6 at

Arg159 enhanced PTEN-mediated suppression of the PI3K pathway (Figure 1.6) [248]. PTEN contains many conserved arginine residues which are often mutated in cancer; therefore, methylation at these arginine sites may uncover novel mechanisms of modulating PTEN functions [248].

SUMOylation

PTEN has three known sites which are modified by small ubiquitin-related modifier, or SUMO.

Modification at Lys266 and Lys289 aid in tethering PTEN to the plasma membrane, where it can dephosphorylate PIP3 substrate and suppress the PI3K pathway (Figure 1.6) [249, 250]. Alternatively,

SUMO modification at Lys254 targets PTEN to the nucleus (Figure 1.6) [251]. Therefore, SUMOylation of PTEN dictates its subcellular localization.

31 Ubiquitination

The role of PTEN ubiquitination at lysine residues 13, 66 and 289 regulate PTEN stability and localization (Figure 1.6) [68, 236]. Polyubiquitination of PTEN by E3 ligases XIAP, CHIP, WWP2 and

NEDD4-1 target PTEN for proteasomal degradation [252-255]. Although the role of NEDD4-1 remains controversial, as PTEN expression in NEDD4-1 positive or negative MEFs showed no significant difference and PTEN was not found to interact with NEDD4-1 [256]. PTEN expression levels are inversely correlated with high E3 ligase expression in cancer, indicating that these ligases negatively regulate PTEN expression to promote disease progression [252, 257]. Interestingly, polyubiquitination by E3 ligase RFP does not affect PTEN stability nor localization but suppresses its phosphatase activity [258]. PTEN monoubiquitination of Lys289 and Lys13 modulates its subcellular localization [68]. Mutation of the

Lys289 and Lys13 residues prevents nuclear accumulation, suggesting that monoubiquitination plays a role in PTEN nuclear import [68]. Furthermore, molecules which are not E3 ligases can regulate PTEN ubiquitination. NEDD4-1 activators Ndfip1 and Ndfip2, as well as PTEN ribosylation via tankyrases enhance PTEN ubiquitination [259, 260]. Contrary to E3 ligases, deubiquitylases USP13 and OTUD3 directly bind to and de-polyubiquitinate PTEN to enhance its stability, while HAUSP/USP7 de-monoubiquitinates PTEN to decrease the nuclear PTEN pool [261-263]. In brief, polyubiquitination of

PTEN promotes its proteasomal degradation, regulating PTEN protein expression levels, while monoubiquitination promotes PTEN nuclear import, thus augmenting PTEN nuclear functions.

Phosphorylation

PTEN C-tail Phosphorylation:

Phosphorylation, the most extensive PTM of PTEN, directly regulates PTEN function, conformation, stability and subcellular localization (Figure 1.6). PTEN is heavily phosphorylated on its

C-tail (residues 351-403). Phosphorylation of Ser380, Thr382, Thr383 and/or Ser385 by CK2 (casein kinase

32 2) causes a conformational change in which the C-tail folds back on the core PD and C2D [146, 264-268].

This conformation enhances PTEN stability by occluding ubiquitination sites which target PTEN for proteasomal degradation and proteolytic digestion by caspase-3 [146, 269]. However, this conformation also occludes sites necessary for membrane-binding in the C2D and possibly the PD, causing cytoplasmic retention and loss of lipid phosphatase activity at the plasma membrane [146, 265, 266, 268]. The intramolecular C-tail/C2D interaction also prevents interactions of the PTEN PDZ binding motif with its constituents in the PTEN-associated complex, like MAGI-2 proteins [147, 270]. Disruption of this supramolecular complex prevents PTEN membrane recruitment and subsequent p-AKT inhibition [147].

Thus, phosphorylation at the C-tail cluster enhances stability but inactivates PTEN via cytoplasmic retention and abrogation of PPIs.

Multiple kinases can phosphorylate one or more residues in the 380/383/383/385 cluster to regulate

PTEN function. For example, p70S6K phosphorylates PTEN at Ser380 under oxidative stress, resulting in

PTEN deubiquitylation and nuclear export, affecting PTEN nuclear functions [148]. Conversely, Ser380 phosphorylation by Plk1 (Polo-like kinase 1) augments PTEN chromatin binding during mitosis, resulting in mitotic exit [271]. Interestingly, PTEN C-terminal interacting protein PICT-1 regulates Ser380 phosphorylation, where PICT-1 knockout significantly reduced Ser380 phosphorylation levels [272].

Ser385 is an additional Plk1 phosphorylation site on PTEN, which enhances PTEN stability by inhibiting polyubiquitination, but excludes it from the nucleus by inhibiting PTEN monoubiquitination [273]. Plk1- mediated Ser385 phosphorylation also results in PI3K pathway upregulation associated with a metabolic state that drives the Warburg effect and tumorigenesis [273]. Taken together, phosphorylation at this cluster via alternative kinases regulates nuclear and cytoplasmic functions.

To counteract phosphorylation at the C-tail cluster, PTEN interacting protein NDRG2 recruits

PP2A to dephosphorylate PTEN, which drives membrane recruitment and decreased p-AKT levels [274].

PTEN also has auto-dephosphorylation capabilities, which regulate its function and stability through targeting its C-tail phosphorylated amino acid cluster [133]. Notably, use of a phosphorylation deficient

33 mutant, known as PTEN-4A, which contains alanine mutations at Ser380, Thr382, Thr383 and Ser385, has enhanced lipid phosphatase activity and preferentially localizes to the nucleus as compared to PTEN wild- type protein [10, 146, 147]. Thus, this cluster of residues modulates stability and subcellular localization.

It is also thought that phosphorylation of the C-tail at this cluster prevents PTEN dimer formation, thereby suppressing the most active form of PTEN wild-type [83]. Additional C-tail phosphorylation sites which regulate stability include Ser362 by GSK3β, and Ser370 by CK2 or Plk3, which can prime Thr366 phosphorylation by GSK3β or Plk3 [275-278]. Taken together, regulation of the phosphorylation of all

C-tail phosphorylation sites dictates PTEN function, localization and stability.

PTEN Phosphorylation of the Core Domains:

Phosphorylation of PTEN occurs at various sites in the C2D. Phosphorylation at Thr336 by Rak enhances PTEN stability by abrogating polyubiquitination and subsequent proteasomal degradation by

NEDD4-1, while Tyr27 and Tyr174 phosphorylation via insulin receptor activation destabilizes PTEN

[279, 280]. PTEN phosphorylation at Thr223, Ser229, T319 and T321 by ROCK during leukocyte chemotaxis targets PTEN to the posterior plasma membrane, contributing to directional cell migration via its lipid phosphatase activity [281]. Tyr240 phosphorylation, possibly by FGFR2 after ionizing radiation, promoted PTEN interaction with Ki-67 and chromatin, recruiting DNA repair molecules like Rad51 [282].

Conversely, Tyr240 phosphorylation, as well as Tyr336 and Tyr315 phosphorylation by Src destabilizes

PTEN and upregulates the PI3K pathway [283]. Although phosphorylation at Tyr240 does not affect PTEN lipid phosphatase activity, it is associated with EGFR inhibitor resistance and shortened survival in glioblastoma [284]. Like the C-tail phosphorylation sites, phosphorylation at select sites in the C2D modulates stability and activity of PTEN and may serve as prognostic markers for resistance and overall survival in patients.

34 Clinical Significance of PTEN Phosphorylation:

Since phosphorylation of specific PTEN residues dictates its functional status and localization,

PTEN phosphorylation may serve as a relevant clinical marker in diseases like cancer [10, 44, 285]. For example, PTEN phosphorylation status is associated with shorter overall survival in patients with acute myeloid leukemia [285]. Increased phosphorylation at the C-tail cluster in >200 lung cancer patients directly correlated with disease severity (normal, low and high grade), implicating a critical functional role of C-tail phosphorylation in cancer progression [10, 44]. Concomitant to increased phosphorylation, there was decreased nuclear staining of total PTEN, indicating that increased nuclear exclusion also strongly correlated with progressively higher grade non-small lung cell lung cancer [10, 44]. A similar trend in nuclear PTEN expression was observed in melanoma [150]. Thus, disease progression to high grade cancers is driven by increased nuclear export or selective cytoplasmic retention, resulting in reduced PTEN function. These results suggest that merely analyzing the PTEN gene for mutations or staining for PTEN wild-type in patient biopsies is not representative of PTEN functional status in disease.

PTEN Restoration Therapy

Since PI3K pathway hyperactivation is widespread across most cancers, KIs (kinase inhibitors) which target one or more effector kinases in the pathway have been used for treatment [286, 287].

Unfortunately, these therapies have met with therapy-resistance, toxicities and off-target effects [286, 287].

Therefore, various alternative approaches that “restore” PTEN function by ectopic overexpression or indirect activation have been attempted.

Restoring PTEN can be implemented directly through gene, mRNA, or protein therapy, or indirectly by compounds which aid in upregulating PTEN transcription. Gene therapy has been extensively studied in multiple cancers such as small cell lung cancer, glioma and prostate, bladder, and colorectal cancer [288-295]. These studies exemplified the ability of active PTEN to be expressed through gene therapy in PTEN-null, PTEN-mutant or PTEN wild-type cell lines and in vivo animal models. Furthermore,

35 reintroduction of wild-type PTEN suppressed and reduced tumor burden in vivo or sensitized the tumors to additional treatment modalities [288-295]. Similar studies have been done with PTEN mRNA packaged into nanoparticles for delivery into cells [296]. Recently, PTEN-L, a PTEN proteoform that has the ability to be secreted and uptaken by cells has been used to validate PTEN restoration therapy. Intra-peritoneal injection of PTEN-L in PTEN U87MG glioblastoma, MDA-MB-468 breast carcinoma, or 786-0 renal cell carcinoma xenografted tumors showed promising results [88, 297, 298]. Genetically engineered tumor- homing neural stem cells with modified PTEN-L for enhanced secretion and cell entry was also effective in treating glioblastoma [299]. These studies validate PTEN-L therapy as a bona fide biological therapy.

Finally, natural compounds with suggested anticancer properties have been shown to upregulate PTEN expression levels [199, 200, 300]. Unfortunately, these mechanisms of PTEN reintroduction rely on gene therapy or ectopic and transient PTEN expression, which are difficult to implement in the clinic [286, 297,

299, 301]. Recently, the naturally occurring indole-3-carbinol was shown to be a potent inhibitor of WWP1, an E3 ligase which negatively regulates PTEN function by suppressing dimerization and membrane recruitment [302]. Although indole-3-carbinol was able to reactivate PTEN in PTEN heterozygous mice, this method remains an indirect means of reactivation. Considering these therapy gaps, novel PTEN restoration therapy approaches that can easily be implemented in the clinic remains imperative. Therefore, my thesis will focus on identifying potential restoration therapies suitable for patients suffering from diseases resulting from compromised PTEN function or low levels of PTEN protein expression.

Summary

Since compromised PTEN function or hyperactivation of the PI3K signaling pathway upstream of

PTEN is common in multiple disease pathologies, PTEN is an attractive target for therapy aimed at suppressing oncogenic PI3K pathway signaling [303]. PTEN restoration therapy has already proved promising, as demonstrated in the section above; however, more simple modes of treatment conducive to clinical implementation are warranted [286, 297, 301]. In the next two chapters, I explore the potential of

36 small molecule compounds to directly target and activate PTEN lipid phosphatase activity, enhancing its canonical tumor suppressor function. I use cancer as a disease model because it is the second leading cause of death in the United States, with an estimate of 1.7 million new cases and more than 600,000 deaths in

2019 [304]. Approximately 17 million new cases will be reported and 9.5 million deaths will occur due to cancer in 2019, resulting in an enormous health burden estimated to reach >$158 billion in the US alone in

2020 [304]. These data warrants research aimed at identifying effective therapies to alleviate patient suffering and social cost.

In chapters 2 and 3, I identify and characterize promising small molecule compounds that directly enhance PTEN enzymatic activity. These novel small molecules are promising therapeutic molecules in diseases driven by compromised function or low levels of PTEN protein. In chapter 4, I have interrogated clinically relevant PTEN mutations derived from patients with PTEN-related diseases, including cancer.

My work has discerned a novel mechanism of PTEN inactivation via protein aggregation, which will aid in targeting inactive PTEN aggregates for better treatments. Therefore, my thesis work presented here focuses on defining potential therapies for patients who suffer from diseases due to genomic or non-genomic PTEN inactivation and/or a myriad of proliferative diseases associated with hyperactive PI3K pathway signaling.

Figure 1.6: Mechanisms of PTEN Inactivation. PTEN has a multitude of inactivation mechanisms that alter PTEN structure, function, stability, expression or subcellular localization. These mechanisms include somatic/germline mutations, histone acetylation, generation of splice variants, PTMs, transcriptional repression, promoter methylation, RNA regulation by long noncoding RNAs or micro RNAs and PPIs. In principle, therapeutic strategies to abrogate these PTEN inactivating mechanisms must be sought to enhance PTEN function.

Chapter 2: Activating PTEN via Small Molecules

Background

Current therapies to block growth signaling in various cancers include the use of KIs that target one or more oncogenic pathway kinases [96]. The PI3K pathway is commonly upregulated in multiple cancers due to overexpression of growth factors/receptors, gain-of-function mutation/hyperactivation of pathway kinases or dysregulation of its pathway regulator PTEN [96]. Although KIs are effective in the clinic, they are often met with off-target effects, causing adverse clinical outcomes in patients; therefore, there is an unmet need for alternative approaches to develop novel therapeutic options to target the oncogenic PI3K pathway [287].

It is now well-accepted that, unlike the Knudson’s two-hit model, PTEN follows the continuum model of tumor suppression, whereby the extent of loss of function or expression of PTEN differs between types of cancer (Figure 1.2) [50]. PTEN somatic or germline mutations were earlier identified as major drivers of oncogenesis [23-25]. However, accumulating evidence over the last decade indicates that alteration in normal PTEN protein levels or function, including PTEN protein-inactivation, is considered to drive many cancers, which is independent of its mutational status [10, 32, 33, 44, 215]. Therefore, the paradigm of PTEN restoration therapy via overexpression of PTEN protein in tumors has been attempted.

While this approach has proved effective in principle, limiting oncogenic potential in vitro and in vivo, implementation of gene and protein therapy in the clinic remains challenging [294, 295, 297, 301]. The possibility of a second approach of directly targeting and enhancing endogenous PTEN activity via small molecules exists, which can reduce the oncogenic effects of PI3K signaling in tumors. Unfortunately, there are no reported direct activators of PTEN. Therefore, in this chapter I use the second approach as a major

38 premise, wherein I identified and elucidated a binding hotspot on PTEN protein that can be exploited for its targeted activation. Utilizing this information, I have also identified and characterized select small molecule compounds that targeted this hotspot, modulated PTEN lipid phosphatase activity and reduced oncogenic potential in cancer cells.

Materials and Methods

FTMap Analysis

The FTMap computational map server (http://ftmap.bu.edu, Boston University) was utilized to perform FTMap analysis for PTEN tumor suppressor protein (PBD 1D5R). The PTEN structure was uploaded to the servers and analyzed per the documentation [305]. Utilizing the visualization software

PyMol (V1.7.4, Shrödinger), the output was inspected for potential ligand binding hot spots.

γ-AApeptide Synthesis

Fifty non-natural peptidomimetics mimicking bioactive peptides were synthesized from N- acylated-N-aminoethyl amino acid building blocks, synthesized as previously described [7, 306, 307]. The building block harboring the R2 side chain was attached to a chloro-trityl resin. The sequential addition of

R3 and R1 side chains occurred after deprotection steps. After cleavage of this intermediate from the solid phase, the R4 group was added to the C-terminus, completing synthesis (Patent Application No. 62/460,324

(USPTO); USF Ref. No. 17A011PR). γ-AAPeptides were purified as previously described [7]. Purity of γ-

AAPeptides was verified by High Pressure Liquid Chromatography (HPLC) and Nuclear Magnetic

Resonance (NMR) spectroscopy analysis [7].

39 Lipid Phosphatase Assays

Microtiter-plate Screening

PTEN lipid phosphatase activity was measured using a Malachite Green (MG) screening assay

[308]. Pure PTEN protein produced in SF9 insect cells (Cayman Chemical, Ann Arbor, MI) was used because it maintains all post-translational modifications; therefore, it is a physiologically and functionally more relevant protein that mimics human PTEN functions in vivo. For the MG titration assays with

γ-AAPeptides, 16.67ng of pure recombinant PTEN protein was incubated with 50mM Tris HCl, pH 8.0,

100mM NaCl, 5mM DTT and increasing amount of indicated γ-AAPeptide (0, 0.25, 0.50, 1, 2, and 5μM) for 1 hour at 37°C in a 96-well microtiter plate. Then, water soluble diC8-PI(3,4,5)P3 substrate (Echelon

Biosciences, Salt Lake City, UT) was added in each well to a final concentration of 54nM and the reaction continued for additional 60 minutes at 37°C. To stop the reaction, 195μL MG reagent (Echelon Biosciences,

Salt Lake City, UT) was added to each well and allowed to incubate for additional 20 min at room temperature to allow for green color development. Absorbance was measured at 620nm and phosphate released was calculated from a standard curve. Assays were performed in quadruplets, at least twice, and reported as mean ± S.E.

Enzyme Kinetics

Similar to the γ-AAPeptide titration assay, a MG dose-response assay was performed using PTEN substrate PIP3 in the absence or presence of γ-AAPeptide. 16.67ng of pure recombinant PTEN protein was incubated with 50mM Tris HCl, pH 8.0, 100mM NaCl, 5mM DTT and 1μM γ-AAPeptide #43 for 1 hour at 37°C in a 96-well microtiter plate. Then, diC8-PI(3,4,5)P3 substrate was added yielding a select range of final concentrations (0, 20, 40, 60, 80, 100, and 120μM). The reaction continued for additional 60 minutes at 37°C. To stop the reaction, 195μL MG reagent was added to each well and allowed to incubate for additional 20 min at room temperature to allow for green color development. Absorbance was measured at

40 620nm and phosphate released was calculated from a standard curve. Assays were performed in triplicates, performed six times, and reported as mean ± S.E. The standard curve was generated by titrating the phosphate standard (Echelon Biosciences, Salt Lake City, UT) in the sub saturating range of

0-2000ρmol/well. Kinetic parameters such as Vmax, Khalf, and h values were calculated using the Allosteric

Sigmoidal function in GraphPad Prism 6 (San Diego, California).

Cancer Cell Culture Maintenance

A549 cells (ATCC® CRM-CCl-185TM, Manassas, VA) and PC3 cells (ATCC® CRL-1435TM,

Manassas, VA) were cultured in F12K media (Thermo Fisher, Waltham, MA) supplemented with 10%

Fetal Bovine Serum (FBS), Antibiotic-antimycotic solution (final concentration 200 units/mL penicillin G,

Table 2.1: List of Primary Antibodies

Primary Antibody Catalog # Concentration Company

p-AKT Ser473 (D9E) XP Cell Signaling Technologies, Danvers, 4060 1:2000 (rabbit, monoclonal) MA

Cell Signaling Technologies, Danvers, AKT (rabbit, polyclonal) 9272 1:1000 MA

p-P70S6K Thr389 (108D2) Cell Signaling Technologies, Danvers, 9234 1:2000 (rabbit, monoclonal) MA

P70S6K (49D7) (rabbit, Cell Signaling Technologies, Danvers, 2708 1:1000 monoclonal) MA

p-PTEN Cell Signaling Technologies, Danvers, Ser380/Thr382/383 (rabbit, 9554 1:2000 MA polyclonal)

Cell Signaling Technologies, Danvers, PTEN (rabbit, monoclonal) 9559 1:1000 MA

β-actin (mouse, A2228 1:1000 Sigma-Aldrich, St. Louis, MO monoclonal)

41 200 µg/mL streptomycin sulfate and 0.5 µg/mL amphotericin B, Sigma Aldrich, St. Louis, MO) and

Plasmocin (1.25 µg/mL, Invitrogen, San Diego, CA). Cells were maintained in a 5% CO2 humidified incubator at 37°C. OVCAR3 cells (ATCC® HTB-161TM, Manassas, VA) were cultured in RPMI-1640 media supplemented with 10% FBS, Antibiotic-antimycotic solution (final concentration 200 units/mL penicillin G, 200 µg/mL streptomycin sulfate and 0.5 µg/mL amphotericin B, Sigma Aldrich, St. Louis,

MO) and Plasmocin (1.25 µg/mL, Invitrogen, San Diego, CA). Cells were maintained in a 5% CO2 humidified incubator at 37°C.

Protein Isolation and Immuno-blotting

For treatment with γ-AAPeptides, cells were grown to 50-60% confluency in 6-well plates, treated with indicated amount of γ-AAPeptide and allowed to incubate for 6 hours at 5% CO2 and 37°C. After 6 hours, media was aspirated and cells were lysed in ice cold Lysis Buffer (25 mM Tris•HCl pH 7.4, 150 mM

NaCl, 1% NP-40, 1 mM EDTA, 5% glycerol) supplemented with 100X protease inhibitor cocktail (Sigma

Aldrich, St. Louis, MO), phosphatase inhibitor cocktail (final concentration 10mM sodium azide, 10mM

NaF, 4mM sodium orthovanadate, 10mM sodium-molybdate dehydrate, 4mM sodium-tartrate dibasic dehydrate, 5mM EDTA disodium dehydrate, 2mM imidazole), 1mM DTT and 1mM PMSF. Samples were sonicated at 10% amplitude using a digital sonifier (Branson Ultrasonics, Dabury, CT), centrifuged at

14,000 rpm for 10 minutes and the supernatant was collected. Protein concentration was determined using

Bradford Dye Reagent (BioRad, Hercules, CA). The remaining supernatant was boiled in 100μL of 2X or

4X Laemmli Buffer containing 5% β-mercaptoethanol (Sigma Aldrich, St. Louis, MO) and subjected to

SDS-PAGE gel analysis, followed by immunoblotting. 20μg of total cellular proteins obtained from cell extract were separated on a 12% SDS-PAGE gel and electro-blotted to nitrocellulose membranes (0.45 μm;

GE Healthcare, Pittsburgh, PA). Blots were blocked with 5% nonfat dry milk or bovine serum albumin in

TBST buffer (10 mM Tris, pH 8,150 mM NaCl, 0.1% Tween 20) and incubated with indicated primary antibodies (Table 2.1). Peroxidase conjugated mouse or rabbit secondary antibodies were used, dependent

42 upon the primary antibody, at dilution of 1:10,000 (Jackson ImmunoResearch, West Grove, PA). Blots were developed by chemiluminescence (Thermo Fisher Scientific, Waltham, MA) and developed on a

Chemidoc MP System (BioRad, Hercules, CA). Quantitative Densitometric analysis on the immunoblots are reported as mean ± S.E. (*p-value of ≤0.05 was considered to be significant).

Cell Proliferation Assay

Cell proliferation studies were performed using A549 cells and the Cell Counting Kit-8 (Dojindo

Molecular Technologies, Rockville, MD). Cells were plated in a 96-well plate with a cell density of 9000 cells/well and cultured in F12K medium as described above. Cells were treated with indicated amount of

γ-AAPeptide, diluted in F12K media, for 24 hours. After 24 hours, final cell numbers were assessed as a function of absorbance at 450nm of reduced WST-8 (2-(2-methoxy-4-nitrophenyl)-3-(4- nitrophenyl)-5-

(2,4-disulfophenyl)-2H-tetrazolium, monosodium salt) [14]. Values were reported as the mean fold change in the number of cells ± S.E. (*p-value of ≤0.05 was considered as significant; n=6).

Cell Migration Assay via ECIS

Migration assays were performed on the ECIS (Electric Cell Substrate Impedance Sensing) Z instrument (Applied Biophysics, Troy, NY). Briefly, A549 cells were grown on electric cell substrate impedance sensing 8-well plate arrays (8W1E; Applied Biophysics, Troy, NY) in F12K media as described above. Confluent cells were treated with media containing γ-AAPeptide at desired concentration for 12 hours prior to wounding. Cells were wounded using an elevated field pulse of 1400 mA at 32,000 Hz applied for 20 seconds, producing a uniform circular lesion of 250 mm in size, and wound closure was tracked until healing. The impedance (Z) was measured at 32000 Hz, normalized to its value at the initiation of data acquisition, and plotted as a function of time. Assays were performed in triplicate and reported as the mean normalized resistance ± S.E. (*p-value of ≤0.05 was considered as significant).

43 Cell Cycle Analysis via Flow Cytometry

Cell cycle analysis was performed on A549 cells treated with indicated amount of γ-AAPeptide at

50-60% confluency for 24 hours. Subsequently, cells were trypsinized and centrifuged at 2,000 rpm for 10 minutes. Media and trypsin were aspirated and cells were washed twice with PBS and spun at 2,000 rpm for 10 minutes at 4°C after each wash. Cells were then stained with trypan blue (Thermo Fisher Scientific,

Waltham, MA) and counted using a hemocytometer. Cells were resuspended in Phosphate Buffered Saline

(PBS) at a density of 2 X 106 cells/mL. Cells suspended in 1 mL PBS were fixed in 3 mL ice cold absolute ethanol for 1 hour at 4°C. After fixing, cells were washed twice with PBS and spun at 3,000 rpm for 10 minutes at 4°C after each wash. 1 mL of Propidium Iodide (PI) staining solution (3.8mM sodium citrate,

40μg/mL PI and PBS to volume) was added to each cell pellet, along with 50μL of RNase A (Qiagen,

Hilden, Germany) at a concentration of 10μg/mL. Cells were incubated at 37°C for 20 minutes and stored at 4°C until analyzed by flow cytometry. Assays were performed in triplicate, performed twice, and reported as mean ± S.E. (p value of ≤0.05 was considered to be significant).

Results

Identifying a PTEN Primary Binding Hotspot

Lee et al. crystallized the PTEN Phosphatase Domain (PD) and C2 Domain (C2D) by removing the intrinsically disordered N-terminal and C-terminal tails, as well as an internal 24 amino acid loop in the

C2D (PDB 1D5R) (Figure 2.1) [26]. Notably, this truncated protein retains lipid phosphatase activity, similar to the PTEN wild-type protein [26]. To define where small molecules may likely target PTEN, I performed computational mapping of most probable small molecule binding surfaces on the available

PTEN crystal structure, utilizing FTMap analysis [305]. The primary binding hot spot was revealed to be at the interface of the PD and C2D (Figure 2.2), specifically at the pα6 helix (amino acid residues 170-185) highlighted in purple in Figure 2.1. A secondary consensus sight was located at the cα1 helix (amino acid

44 residues 275-281) (Figure 2.2), highlighted in blue in Figure 2.1. This PD/C2D interface consists of mostly hydrophobic and aromatic residues, supplemented with a network of nine hydrogen bonds that preserve the structural integrity and function of the interface [26]. Therefore, targeting the PD/C2D interface has a high likelihood of influencing PTEN lipid phosphatase activity.

To test the above concept, a synthetic ɑ-helical peptide mimicking the pɑ6 helix (amino acid residues 170-184), which is known to engage in productive and dynamic interaction at the PD/C2D interface (Figure 2.3A), was generated to ascertain its influence on PTEN activity utilizing the in vitro MG phosphatase assay. Indeed, titration with a synthetic ɑ-helical peptide mimicking the pɑ6 helix enhanced full length PTEN wild-type enzyme activity as detected by the MG assay (Figure 2.3B).

Figure 2.1: Overview of PTEN Structure. (A) Cartoon representation of domains in PTEN. PBM: PIP2 Binding Module; PD: Phosphatase Domain: C2D: C2 Domain; C-tail: flexible C-terminal; PDZ-BD: PDZ Binding Domain. (B) PTEN crystal structure ID5R with domains and structures.

Figure 2.2: PTEN PD/C2D Interface is a Small Molecule Binding Hotspot. FTMap analysis revealed a binding hot spot on PTEN crystal structure PDB 1D5R to be located at the interface between the PD and C2D (pα6 helix: amino acids 170-185 and cα1 helix: amino acids 275-281). All 16 standard probes aggregated near the PD/C2D interface.

Figure 2.3: PTEN PD/C2D Interface is Suitable for Targeting with Small Molecules. (A) Representation of the location of the pɑ6 helix and the synthesized amino acid sequence. (B) The synthetic pɑ6 helix demonstrated dose-dependent activation of PTEN lipid phosphatase function in a Malachite Green assay, n=4. Values reported as mean ± SE.

46 γ-AAPeptides Enhance PTEN Lipid Phosphatase Activity

With a primary binding hotspot identified, 51 non-natural peptidomimetics, known as

γ-AAPeptides, were synthesized from the previously described N-acylated-N-aminoethyl amino acid building block [7, 306]. γ-AAPeptides have high propensity for increased chemical diversification, bioactivity, prolonged intracellular retention, and ability to abrogate protein-protein interactions, making them ideal candidates to target the PD/C2D interface [309]. Briefly, the building block, already containing the R2 side chain, underwent de-protection before addition of the R3 and R1 side chains. After cleavage of this intermediate from the solid phase, the R4 group was appended to the C-terminus, completing synthesis

(Figure 2.4). Select γ-AAPeptides described below were purified and verified by HPLC and NMR spectroscopy, respectively (Table A1 and Figure A1) [7].

Figure 2.4: γ-AAPeptide Synthesis. γ-AAPeptides were synthesized directly from the N-acylated-N-aminoethyl amino acid building block harboring the R2 side chain. Structurally and chemically diverse R1, R3 and R4 side chains are appended to create the final structure.

The MG assay was used to screen 51 γ-AAPeptides at a 1μM concentration and found that 33

γ-AAPeptides increased PTEN activity (Figure 2.5A). Strikingly, 19 of 51 γ-AAPeptides contained an adamantyl side group at position R3, of which 18 enhanced PTEN activity (Figure 2.5B). Further analysis identified 10 γ-AAPeptides to be promising in a MG titration assay, with γ-AAPeptide #43 as the most potent activator (Figure 2.5C and Table 2.2). Therefore, all further analysis was performed with

γ-AAPeptide #43.

Figure 2.5: Select γ-AAPeptides Enhanced PTEN Lipid Phosphatase Activity. (A) 51 γ-AAPeptides were screened for changes in PTEN lipid phosphatase activity at 1μM using Malachite Green assay, n=4. Values reported as mean ± SE. (B) Flow chart representing γ-AAPeptide structure function relationship analysis. From 51 γ-AAPeptides, 19 contained an adamantyl side chain and 18 were shown to enhance PTEN activity (66% of all activators). (C) 10 select γ-AAPeptides were titrated and measured for changes in PTEN lipid phosphatase activity using Malachite Green assay, n=4. Values reported as mean ± SE.

48 Table 2.2: Structures of Select γ-AAPeptides which Enhanced PTEN Lipid Phosphatase Activity.

49 Since these γ-AAPeptides were designed to bind at the PD/C2D interface, which is a site that is distinct from the active binding site of PTEN, γ-AAPeptide #43 likely enhances PTEN lipid phosphatase activity via indirect conformational modulation in PTEN. To this point, I utilized the MG assay to generate a PI(3,4,5)P3 substrate dose-response curve in the absence or presence of 1μM γ-AAPeptide #43. PTEN activity in the absence of activator demonstrated a sigmoidal curve with increased PI(3,4,5)P3 substrate

(Figure 2.6), suggesting inherent allostery. Indeed, the Hill coefficient was h=1.852, Vmax was calculated as 61.57 pmol phosphate released/min, and Khalf was calculated to be 55.78μM. Addition of 1μM

γ-AAPeptide #43 shifted the curve to the left and significantly increased the Vmax of PTEN to 70.34pmol phosphate/min, reflecting the effect of γ-AAPeptide #43 binding to the PD/C2D interface (Figure 2.6).

Although Khalf (54.18μM) and h (1.717) changed slightly, the values were not calculated to be significantly different from the PTEN alone curve.

Figure 2.6: γ-AAPeptide #43 Enhances PTEN Activity. Increasing PIP3 concentration in the presence of PTEN alone produces a sigmoidal curve with a calculated Vmax= 61.57 pmol phosphate/min, Khalf= 55.78μM, and h=1.852. PTEN protein pre-incubated with γ-AAPeptide #43 prior to PIP3 addition also produced a sigmoidal curve which is shifted to the right. The Vmax of the curve significantly increased to 70.34 pmol phosphate/min (t-test, *p-value≤0.05), with a Khalf= 54.18μM and h=1.717.

A Select γ-AAPeptide Reduces Oncogenic Signaling in Cancer Cells

Since γ-AAPeptide #43 caused the most prominent increase in PTEN lipid phosphatase activity in vitro, I tested it’s effects in attenuating oncogenic potential of a lung cancer cell line, A549, which has normal PTEN protein expression but hyperactive KRAS signaling due to KRAS mutation. Treatment of

A549 cells with γ-AAPeptide #43 for 6 hours demonstrated a significant decrease in the phosphorylated,

50 active form of AKT (pAKT Ser473) and P70S6K, (p-P70S6K Thr389), indicating inhibition of canonical

PI3K signaling (Figure 2.7A). In contrast, γ-AAPeptide #43 did not significantly alter the phosphorylation at the PTEN C-tail amino acid cluster (Ser380, Thr382, Thr383, Ser385) (Figure 2.7A). PI3K pathway inhibition after γ-AAPeptide #43 treatment was observed in another cell line with normal PTEN protein expression: ovarian cancer cell line OVCAR3. Treatment with γ-AAPeptide #43 significantly decreased both p-AKT and p-S6K levels (Figure 2.7B).

Figure 2.7: A Select γ-AAPeptide Antagonized PI3K Pathway Signaling in Cancer Cell Lines. (A) Immuno-blot analysis revealed a significant decrease in phosphorylated and activated PI3K/AKT/S6K pathway effectors, p-AKT (Ser473) and p-P70 S6K (Thr389), in A549 cells treated with 40μM #43 for 6 hours. Quantitative densitometric analyses of blots for each component is shown below, n=4 (*p-value ≤0.05; n.s.= not significant). (B) Immuno-blot analysis revealed a significant decrease in phosphorylated and activated PI3K/AKT/S6K pathway effectors, p-AKT (Ser473), p-P70 S6K (Thr389) and p-PTEN (Ser380, Thr382, Thr383) in OVCAR3 cells treated with 40μM γ-AA Peptide #43 for 6 hours. Quantitative densitometric analyses of blots for each component is shown below, n=3 (*p-value ≤0.05).

A Select γ-AAPeptide Inhibits Tumorigenic Features in Lung Cancer Cells

Enhancing PTEN expression or function in cancer cells is known to suppress cell proliferation and cell migration, and inhibit cell cycle progression [32, 33]. Since γ-AAPeptide #43 can augment endogenous

PTEN function in two distinct cancer cell lines to dampen PI3K pathway signaling, I hypothesized that #43

51 would likely modulate other oncogenic properties of cancer cells, including cell proliferation, migration and cell cycle progression. To test this hypothesis, I treated A549 cells with γ-AAPeptide #43 and assessed its effects on cell proliferation, migration and alteration in cell cycle. Indeed, 24 hours of treatment with

γ-AAPeptide #43 suppressed cell proliferation in a dose-dependent manner (Figure 2.8A). Likewise, cell migration studies using the ECIS method demonstrated that γ-AAPeptide #43 significantly inhibited the rate of A549 cell migration (Figure 2.8B). PTEN inhibits cell proliferation by inducing cell cycle arrest at the G1 and G2/M transition phases [310]. Treatment of A549 cells with γ-AAPeptide #43 for 24 hours arrested cells in the G1 phase, as determined by the increase in cell population in the G1 phase and decrease in cell population in the S phase (Figure 2.8C). Notably, γ-AAPeptide #43 had a negligible increase in the sub-G1 cell population (Figure 2.8C), suggesting that it is not cytotoxic at the concentration of

γ-AAPeptide #43 assessed in this assay [311]. Taken together, γ-AAPeptide #43 inhibited well-established oncogenic properties of A549 cells, including PI3K signaling, cell proliferation, cell migration and cell cycle progression.

Figure 2.8: A Select γ-AAPeptide Inhibits Oncogenic Potential of A549 Lung Cancer Cells. (A) Dose-dependent inhibition of cell proliferation was observed in cells treated with γ-AAPeptide #43 for 24 hours. Cells were plated at 9000 cells/well, n=6 (*p-value ≤0.05). (B) Continuous impedance sensing measurements revealed a significant decrease in rate of wound closure for cells treated with 30μM γ-AAPeptide #43, 12 hours prior to wounding, as compared to cells treated with vehicle, n=3 (p-value of ≤0.05 was considered to be significant). (C) Cell cycle analysis of cells treated with 40μM γ-AAPeptide #43 for 24 hours increased G1 cell population and decreased S phase cell population, n=3 (*p-value ≤0.05).

Conclusion

The lipid phosphatase activity of PTEN is crucial for its function as a potent tumor suppressor [31].

As the master regulator of the PI3K pathway, reduction in PTEN expression or function, or hyperactivation

52 of the PI3K pathway, heightens oncogenic potential of cells and drive tumorigenesis [96, 303, 312]. Since therapeutic resistance and toxicities are common in patients undergoing chemotherapy or KI treatment, the need for novel and alternative adjunctive therapies can reduce the doses of standard therapies, and thus morbidity observed in the clinic [96, 286, 287, 303]. PTEN restoration therapy has proved promising; however, direct and targeted activation of PTEN has not been attempted. In this chapter, I have demonstrated targeted activation of PTEN lipid phosphatase activity using a class of peptidomimetics known as γ-AAPeptides. Their cell penetration efficiency and high intra-cellular retention makes them efficient therapeutic molecules. I was able to identify a single lead γ-AAPeptide that increased PTEN activity in vitro, likely binding to the PD/C2D interface, modulating PTEN structure and enhancing its activity. This γ-AAPeptide likely activated endogenous cellular PTEN, reducing the PI3K signaling and decreasing cell proliferation, migration and cell cycle progression in lung cancer cells. Since this

γ-AAPeptide also decreased PI3K pathway activation in another PTEN containing cell line, OVCAR3, these results suggest that γ-AAPeptides have the potential to be used as a treatment for cancer patients that harbor varying amounts of PTEN wild-type protein. In the next chapter, I will elucidate the binding interactions of this promising γ-AAPeptide with PTEN protein to determine its allosteric mechanism of activation.

CHAPTER 3: Elucidation of γ-AAPeptide Mediated Enhancement of PTEN Activity

Background

Both PTEN lipid and protein phosphatase activities are dependent upon the active catalytic site located entirely within the PD [26]. As detailed in Chapter 1, the pɑ4 helix contains the catalytic cysteine residue responsible for dephosphorylation of lipid and protein substrates. The remainder of the active site is composed of the WPD-loop, which stabilizes substrate binding for dephosphorylation, and the TI-loop, which is necessary to accommodate large phospholipid substrates [26]. γ-AAPeptides were identified to target the PD/C2D interface, which is directly adjacent to the active site, to influence PTEN activity.

Notably, PTEN enzymatic activity is dependent upon the integrity of the PD/C2D interface, and I have shown that competitive binding at this interface can enhance PTEN activity in vitro. Since I have identified

γ-AAPeptides with a common adamantyl functional group, which have the ability to either activate or inhibit PTEN function, it is likely that differences in binding affinity and orientation modulates PTEN enzymatic activity. Therefore, in this chapter, I have performed molecular docking and molecular dynamic simulation studies to elucidate how γ-AAPeptides may bind to the PTEN PD/C2D interface and influence

PIP3 catalysis in the active site.

54 Materials and Methods

Molecular Docking

Rigid Docking with DOCK

PTEN crystal structure 1D5R was used for rigid body docking to screen for potential ligand binding using the program DOCK 3.5.54 [313, 314]. The binding pocket was resolved from FTMap analysis and was designated as amino acid residues 170-185, 278 and 279 of PTEN. These residues were used to generate matching spheres, which were chemically labeled for matching based on ionization states and hydrogen bonding of protein residues in the nearby vicinity. The test set of γ-AAPeptides were docked into the binding pocket, ranked based upon score, and visually inspected with the visualization software PyMol for chemical complementarity and spatial positioning.

Induced Fit Docking with Schrodinger

All flexible ligand/flexible receptor docking studies were performed using Schrödinger’s Induced

Fit Docking module [315, 316]. In our first set of docking studies, γ-AAPeptide ligands #43 (activator) and

#42 (inhibitor) were prepared using LigPrep 2.3 to produce all 3D structures with corresponding stereochemistry, and Epik, which generated protonation states at a pH of 7.0 ± 2. Additionally, Epik was used to prepare the protonation states of PTEN 1D5R at pH 7.0 ± 2 [317]. Propka was used to optimize all hydrogen bonds within the crystals and OPLS3 force field was used for atom minimization [318]. The prepared γ-AAPeptide ligands were docked into our proposed binding site, as determined by FTMap analysis (residues 170-185, 278, 279). γ-AAPeptides #43 and #42 were initially docked into the binding site using GlideSP (Glide Standard Precision) [319, 320]. This binding site, as well as relevant PTEN active site residues (91-94, 123-130, 163-166), were “trimmed” (mutated to alanine, then later back-mutated to their original identities) to promote induced flexibility. Furthermore, residues within 10Å of the binding site were subject to energy minimization. After initial placement, Schrödinger’s Prime side-chain prediction

55 is used to predict new orientations for all “trimmed” residues. Finally, GlideXP (Glide eXtra Precision) is then used to “re-place” the ligand in its initially predicted location and determine the final top scoring binding modes [321].

The top scoring PTEN-complexed γ-AAPeptide #43 and #42 structures were then used as starting structures in the second set of docking studies, where PTEN substrate PIP3 was docked into the active site of PTEN (residues 91-94, 123-130, 163-166). To simulate PIP3 docking into an apo-PTEN active site, a

“dummy” water molecule was first docked at the γ-AAPeptide binding site using the same above-mentioned protocol. This was to ensure that apo-PTEN underwent the same trimming/refinement process as

γ-AAPeptide-bound PTEN structures prior to PIP3 active site docking. All residues within 10Å of the active site, with the exception of the γ-AAPeptide binding site residues (170-185, 278, 279) were subject to structural and energetic refinement. In order to preserve, any conformational changes in the active site structure produced by γ-AAPeptide or water molecule binding, trimming was not employed in these dockings. Glide XP was used to generate the top scoring binding modes of PIP3 in the active site.

Intrinsic Disorder Analysis

Intrinsic disorder profiles for the full-length human PTEN (UniProt ID: P60484) were generated by PONDR® VLXT [322], PONDR® VL3 [323], PONDR® VSL2 [323, 324], PONDR® FIT [325],

IUPred_short and IUPred_long [326]. PONDR® VLXT is one of the first-generation disorder predictors.

Although its accuracy is not too high, this predictor is known to have high sensitivity to local sequence peculiarities and can be used for identifying disorder-based interaction sites [322, 327-329]. In fact, because it was trained on the limited set of then available proteins with experimentally validated intrinsically disordered status, PONDR® VLXT may underestimate the occurrence of long disordered regions in proteins. PONDR® VL3 was specifically designed to accurately predict long disordered regions [323].

PONDR® VSL2 is one of the more accurate stand-alone disorder predictors [330, 331]. PONDR® FIT is

56 a metapredictor of intrinsic disorder that is known to be moderately more accurate than each of the component predictors [325]. After obtaining an average disorder score by each predictor, all predictor- specific average scores were averaged again to generate an average per-protein intrinsic disorder score. Use of consensus for evaluation of intrinsic disorder is motivated by empirical observations that this approach usually increases the predictive performance compared to the use of a single predictor [330-332].

Molecular Dynamics Simulations

PTEN crystals structure 1D5R and all simulations for γ-AAPeptide #43, #42 and apo complexes were run using the CHARMM (Chemistry at HARvard Macromolecular Mechanics) biomolecular modeling program [333]. The online CHARMM interface, CHARMMing, was used to prepare all structures for simulation, as explained here [334]. 1D5R protein structure was treated with CHARMM (C36) force field and all ligand topology and parameters were determined using the CHARMM generalized forcefield

(CGenFF) available through the ParamChem web-interface [335]. Each protein-ligand complex (herein referred to as protein system) was solvated in a 91.565 Å x 91.565 Å x 91.565 Å cubic box using 23,029

TIP3P water molecules [336], long range conditions were modeled using periodic boundary conditions.

Loose/short (50 step) Steepest Descent (SD) minimization protocol was used to minimize bad initial contacts. The systems were subsequentially neutralized with the addition of 0.15M KCl salt. Following neutralization, 10 steps of SD minimization and 25 steps of Adopted Basis Newton-Raphson (ABNR) minimization was performed to alleviate steric clashing between protein, solvent, and ion atoms. Domain

Decomposition (DOMDEC) and Graphical Processing Unit (GPU) compatible CHARMM versions were used to heat each system from 110.5K to 310K over a 400 picosecond (ps) timeframe, with a time step of

1 femtosecond (fs), and then allowed to equilibrate under isothermal-isobaric conditions (NpT ensemble) for 20 nanosecond (ns). After initial equilibration, each system was simulated under NpT conditions for

140ns. The SHAKE algorithm constrained all covalent bonds to hydrogen atoms. Long range electrostatic interactions were evaluated using Particle Mesh Ewald (PME) (k=0.34, order of B-spline interpolation=6)

57 and dispersion-type interactions were modeled using Lennard-Jones potential with a switching function between 8-12Å from the central atom and a 12Å cut-off, both of these methods were used in conjunction with periodic boundary conditions [337].

Results

Select γ-AAPeptides Target Unique Residues in the PD/C2D Interface

Exploratory Study via Rigid Docking

To delineate the molecular mechanism by which γ-AAPeptides may bind and activate PTEN, molecular docking studies were performed to determine the binding of γ-AAPeptide #43 with PTEN protein. DOCK 3.5.54 [313] software was used, targeting amino acid residues identified within the FTMap analysis binding hotspot (170-185, 278 and 279) on PTEN crystal structure PDB 1D5R. γ-AAPeptide #43 oriented productively within the PD/C2D interface, where the R3 adamantyl side chain inserted into the cleft, which contains multiple aromatic residues (Y176, Y177, Y180, F279) (Figure 3.1). This non-polar binding pocket also accommodated aromatic side chains R2 and R4, while the aliphatic R1 side chains protruded out of the binding pocket towards bulk solvent (Figure 3.1). Overall, rigid docking suggests that activator γ-AAPeptide #43 likely makes significant and favorable contacts with the suspected PTEN

PD/C2D binding interface.

Figure 3.1: Rigid Docking Demonstrates Potential γ-AAPeptide #43-PTEN Interaction. γ-AAPeptide #43 docked into PTEN 1D5R using DOCK 3.5.54. Docking pose is in a plausible orientation within the hydrophobic/nonpolar PD/C2D interface. Visualized in PyMol.

58 Validation via Induced Fit Docking

Since proteins utilize their inherent flexibility to achieve best binding, I utilized the induced fit approach, which accounts for flexibility in the binding pocket upon γ-AAPeptide binding. Shrödinger’s

Glide module was employed to dock γ-AAPeptide #43 to PTEN 1D5R. Additionally, the screening analysis discussed in Chapter 2 highlighted that only one γ-AAPeptide containing an adamantyl ring decreases

PTEN enzymatic activity, γ-AAPeptide #42. Titration of γ-AAPeptide #42 in a MG assay shows a dose- dependent decrease in PTEN lipid phosphatase activity (Figure A2). Since this inhibitor is highly similar in structure to the adamantyl containing activator γ-AAPeptide #43, I posited that γ-AAPeptide #42 makes distinct interactions at the PD/C2D interface from that of γ-AAPeptide #43, which may accounts for the functional differences (i.e. inhibition of PTEN activity by #42 binding) in PTEN lipid phosphatase activity between the two γ-AAPeptides.

As outlined in the methods, the γ-AAPeptide binding site residues (170-185, 278, 279) were trimmed to open the binding pocket, and all residues within 10Å of the binding site were subject to

Figure 3.2: Adamantyl Containing γ-AAPeptides Have a High Binding Affinity for the PD/C2D Interface. Induced Fit Docking (IFD) was utilized to simulate protein flexibility upon γ-AAPeptide binding and obtain predicted binding energies of each interaction. (A) γ-AAPeptide #43 docked into the PTEN PD/C2D interface in 1D5R with a binding energy of -13.427kcal/mol. Hydrogen bonding interactions are observed between the γ-AAPeptide R1 aliphatic group and D324 and T319 in the C2D. Additionally, the γ-AAPeptide backbone interacts with R173 in the PD. (B) γ-AAPeptide #42 docked into the PTEN PD/C2D interface in 1D5R with a binding energy of -11.596kcal/mol. Hydrogen bonding occurs between the aliphatic R1 group and residues D324 and T319 in the C2D. A potential base stacking interaction exists between the hexafluorobenzene and T180.

59 refinement. Docking of γ-AAPeptide #43 and #42 to PTEN structure 1D5R revealed low energy binding modes of -13.427kcal/mol and -11.596kcal/mol, respectively (Figure 3.2). Both binding poses demonstrate extensive interactions of the aromatic ring structures of the R2-4 side chains with aromatic PTEN residues

Y176, Y180, F278 and F279, yielding overall favorable binding modes. γ-AAPeptide #43 makes hydrogen bonding interactions with residues in the PD and C2D (R173, D324 and T319), while γ-AAPeptide #42 makes hydrogen bonds only within the C2D (D324, T319 and N276) (Figure 3.2). Additionally, there is a potential base stacking interaction between Y180 and the hexafluorobenzene group at the R4 position of

γ-AAPeptide #42 (Figure 3.2B). Interestingly, the R3 adamantyl groups on each γ-AAPeptide are located at different locations within the binding pocket (Figure 3.2). Thus, each γ-AAPeptide has a distinct binding pose within the PD/C2D interface.

γ-AAPeptide Binding Alters PIP3 Binding Affinity and Orientation

Since γ-AAPeptide #43 and #42 have opposing effects on PTEN lipid phosphatase activity in vitro, but are predicted to bind to the same interface, I posited that binding of each γ-AAPeptide has differential effects on PIP3 binding in the PTEN active site. To test this hypothesis, PIP3 was docked into the active site of apo-, #43 bound-, and #42 bound-PTEN 1D5R. Indeed, PIP3 binding affinity and orientation varied based on the bound γ-AAPeptide (Figure 3.3).

When PIP3 was docked into the active site of PTEN 1D5R, it gave a binding score of

-9.972kcal/mol (Figure 3.3A). Furthermore, the 3’ phosphate is 5.2Å from the catalytic C124 residue, therefore the binding pose is within nucleophilic attack distance, conducive for catalysis (Figure 3.3A).

Strikingly, when PIP3 is docked into the active site with γ-AAPeptide #43 bound to the PD/C2D interface, the PIP3 active site binding affinity increased, with a binding score of -14.180kcal/mol, and the 3’ phosphate group is closer to the catalytic C124 residue by more than 1Å, at a distance of 4.1Å (Figure

3.3B). In contrast, when PIP3 is docked into the active site with γ-AAPeptide #42 bound to the PD/C2D

60 interface, the PIP3 active site binding affinity is decreased by 2.56 kcal/mol as compared to the #43-bound

PTEN. Furthermore, 3’ phosphate group is not within range of the catalytic C124 residue, therefore is not in range for hydrolysis (Figure 3.3C). These finding indicate that γ-AAPeptides #43 and #42 binding differentially modulates PIP3 active site binding affinity, orientation and catalysis.

Figure 3.3: γ-AAPeptide Binding Influences PIP3 Active Site Binding and Orientation. (A) PIP3 docked into the apo-PTEN 1D5R active site with a binding energy of -9.972kcal/mol. The 3’ phosphate group is positioned 5.2Å away from the catalytic C124. (B) PIP3 docked into γ-AAPeptide #43-bound PTEN 1D5R with a binding energy of -14.180kcal/mol, a higher binding affinity than PIP3 docked into apo-PTEN. The 3’ phosphate group is positioned 4.1Å away from the catalytic C124, closer than the PIP3 3’ phosphate in apo-PTEN. (C) PIP3 docked into γ-AAPeptide #42-bound PTEN 1D5R with a binding energy of -11.621kcal/mol, a higher binding affinity than PIP3 docked into apo-PTEN, but lower than PIP3 docked into #43-bound PTEN. The 3’ phosphate group is positioned 11.6Å away from the catalytic C124, a position not within the nucleophilic attack distance. To validate that PD/C2D is indeed the major site from where the γ-AAPeptides modulate PTEN catalytic activity, I systematically mutated critical amino acids lining the PD/C2D interface. Interestingly, these mutations drastically altered γ-AAPeptide #43 and PIP3 active site binding. Alanine mutations were generated within the interface at residues Y176, Y177, Y180, F279, D324 and γ-AAPeptide #43 was docked into the interface. Each individual mutation decreased the binding affinity, as determined by an increase in the binding score (Table 3.1). Furthermore, PIP3 was docked into the active site of these mutant PTEN structures in the absence or presence of γ-AAPeptide #43. In the absence of γ-AAPeptide #43 (apo-PTEN), the F278A mutation allowed PIP3 binding in the active site, albeit in an improper orientation so the 3’ phosphatase is not within distance of the catalytic C124 residue, indicating that this residue may be necessary for productive PIP3 binding in wild-type PTEN (Table 3.2). All other alanine mutations allowed for proper PIP3 binding and orientation in apo-PTEN (Table 3.2). Interestingly, docking of PIP3 into the active site of γ-AAPeptide #43-bound mutant PTEN either abrogated PIP3 binding or resulted in an

61 unproductive binding orientation where the 3’ phosphate was not in range of the catalytic cysteine residue

(Table 3.2).

Table 3.1: Mutations in PTEN 1D5R PD/C2D Interface Alters γ-AAPeptide Binding Affinity. Mutations in PD/C2D interface residues decrease the binding affinity (an increase in binding score) of γ-AAPeptide #43 for the interface.

PTEN (1D5R) #43 binding (kcal/mol)

WT -13.427

Y176A -13.023

Y177A -13.404

Y180A -13.134

F278A -11.704

D324A -12.653

Table 3.2: Mutations in PTEN 1D5R PD/C2D Interface Disrupts γ-AAPeptide-induced Effects on PIP3 Binding. Mutations at the PD/C2D interface alter the binding affinity for PIP3 in the active site in the presence or absence of γ-AAPeptide. Notably, the F278A mutations does not produce a binding pose where the 3’ phosphatase is within attacking distance of the catalytic C124 residue. Docking γ-AAPeptide #43 to the mutated PD/C2D interface disrupts effective PIP3 active site binding by producing an incorrect active site orientation (3’displaced: 3’ phosphatase is not within range of the catalytic C124 residue) or abolishing PIP3 binding entirely (No binding).

PTEN (1D5R) PIP3 – apo (kcal/mol) PIP3 - #43 (kcal/mol)

WT -9.972 -14.180

Y180A -11.18 3’ displaced

F278A 3’ displaced 3’displaced

Y177A -10.412 No binding

Y176A -12.1 No binding

D324A -11.252 3’ displaced

62 Identifying Regions of PTEN that Induce Allostery

Although these γ-AAPeptides bind at a distant region from the catalytic site, they alter PTEN activity significantly. It is plausible that such binding induces an allosteric effect, which can, via positive coupling, augment the function of the catalytic site of PTEN, particularly facilitated by intrinsically disordered regions (IDRs) in enzymes [338]. Figure 3.4 represents computational analysis that identifies intrinsic disorder in human PTEN wild-type. The positions of α-helices pα6 and cα1, where these

γ-AAPeptides likely bind, are indicated by the shaded areas. Although both helices are located within mostly ordered regions of the protein due to the high content of aromatic residues, they occupy a unique position in close proximity to the highly disordered loops, such as the TI-loop (amino acid residues 161-

169) and the non-crystallizable loop in the C2D (amino acid residues 286-309), being positioned right at the edge of the sharp "disorder-to-order" or "order-to-disorder" transitions. It is likely that such positioning of binding sites should significantly affect mobility of these neighboring loops when a γ-AAPeptide is bound. Furthermore, it is also likely that the γ-AAPeptides binding can affect the dynamics of the P-loop indirectly, via the TI-loop, which may influence PIP3 active site binding and orientation.

Figure 3.4: γ-AAPeptides Bind PTEN at Regions Adjacent to Intrinsically Disordered Loops. Intrinsic disorder propensity of PTEN evaluated by several well-established disorder predictors. Cyan dash-dot-dotted lines show the mean disorder propensity calculated by averaging disorder profiles of individual predictors. Intrinsic disorder propensity above 0.5 indicates intrinsically disordered residues. The gray and green shaded areas are where γ-AAPeptides are predicted to bind. The green shaded area is directly adjacent to the intrinsically disordered loop in the C2D that is deleted from the crystal structure. The peak of intrinsic disorder next to the gray shaded region is the TI-loop within the active site of PTEN.

63 γ-AAPeptide Binding Alters Active Site Conformation

Since both γ-AAPeptides #43 and #42 stabilize the PD/C2D interface and the active site, but have differential effects on PTEN lipid phosphatase activity, it is plausible that each γ-AAPeptide stabilizes the active site residues in different conformations to regulate PIP3 binding and orientation. To test this hypothesis, I overlaid the three active sites (apo-PTEN, #43-bound PTEN and #42-bound PTEN) after PIP3 binding via the IFD protocol. Comparison of the orientation of residues across the three active sites of

PTEN revealed a set of select residues in distinct positions (Figure 3.5). For example, residues C124, K125,

K163, I168 and P169 exhibit similar conformations in the absence and presence of γ-AAPeptide binding

(Figure 3.5). Conversely, K128, V166 and T167 positions show variation in the absence or presence of each

γ-AAPeptides (Figure 3.5). D92 and R130 orientations in apo-PTEN and γ-AAPeptide #43-bound PTEN are in the same conformation, while the R130 and D92 orientations in γ-AAPeptide #42-bound PTEN differ

(Figure 3.5A). Alternatively, γ-AAPeptide #43 or #42 binding locked residues H93, R161, D162 and K164 in the same conformation, differing from apo-PTEN (Figure 3.5B). Thus, these overlaid structures show

Figure 3.5: Binding of γ-AAPeptides to the PD/C2D Interface Changes Active Site Residue Conformations. After PIP3 was docked into the active site using the IFD protocol, the PIP3 molecules were removed and the active sites of apo-PTEN, γ-AAPeptide #43-bound PTEN and γ-AAPeptide #42-bound PTEN were overlaid in PyMol. (A) Overlaid PTEN structures highlighting the conformations of select residues from the P-loop (C124, K125 and K128) and the WPD-loop (D92 and H93). (B) Overlaid PTEN structures highlighting the conformations of select residues from the TI-loop. Select active site residues do not have significant changes in their conformations upon γ-AAPeptide binding, such as C124, K125 and K163. Conversely, the remainder of the highlighted residues seem to adopt alternative conformations in comparison to apo-PTEN, likely due to γ-AAPeptide binding.

64 differential displacement of active site residue conformations in the presence or absence of these

γ-AAPeptides.

γ-AAPeptide Binding Stabilizes PTEN Protein Structure

To determine the effects of γ-AAPeptide binding on PTEN structure and active site residue orientation, molecular dynamics (MD) simulations were performed over a 140ns timeframe utilizing PTEN

1D5R structure. The γ-AAPeptide binding site (BS), the PIP3 BS and the TI-loop in apo-PTEN are highly flexible with RMSD values between ~15-22Å (Figure 3.6A). However, upon γ-AAPeptide #43 or #42 binding, the γ-AAPeptide BS, PIP3 BS and the TI-loop alone all become much more rigid structures, decreasing the RMSD values by 10-15Å (Figure 3.6B-C).

Since γ-AAPeptides #43 and #42 have differential effects on PTEN lipid phosphatase activity in vitro and stabilize PTEN as determined by MD analysis, it is possible that γ-AAPeptide binding induces conformational changes in the active site. In order to determine if γ-AAPeptide #43 or #42 binding can stabilize the active site residues in distinct orientations, the dihedral angles of the active site binding pocket were analyzed (residues 91-94, 123-130 and 161-169). γ-AAPeptide #43 or #42 binding showed no distinct differences in the dihedral angle orientations and fluctuations in residues E91, D92, H93, H123, T160,

R161, K163, V166, and I168 as compared to apo-PTEN protein (Figure A3). In contrast, both γ-AAPeptides induce similar dihedral angle orientations and fluctuations in residues K128, G129, D162 and K164 as compared to apo-PTEN, likely indicating a structural change induced by general γ-AAPeptide binding

(Figure 3.7). Individually, each γ-AAPeptide can produce distinct residue conformations as compared to

γ-AAPeptide-bound or apo-PTEN. For example, binding of γ-AAPeptide #43 results in similar dihedral angle orientations and fluctuations in N94 and K125 as compared to apo-PTEN; however, γ-AAPeptide

#42 binding produces different dihedral angles in the select bonds of these residues (Figure 3.8A-B). In contrast, γ-AAPeptide #42 binding does not alter the dihedral angles of residue T167 as compared to apo-

65 PTEN, while γ-AAPeptide #43 binding produces a distinct and altered dihedral angle signature (Figure

3.8C). Lastly, dihedral angle analysis of residues C124, A126, G127, R130 and G165 shows distinct signatures between apo-PTEN, γ-AAPeptide #43-bound PTEN and γ-AAPeptide #42-bound PTEN (Figure

3.8D-H).

Figure 3.6: γ-AAPeptide Binding to the PD/C2D Interface Stabilizes the Domain Interface and Active Site Pocket. Molecular dynamic simulations were run over a 140ns timeframe. (A) The PTEN 1D5R γ-AAPeptide binding site (BS), PIP3 BS and TI-loop fluctuate with an RMSD value between 15-20Å over the course of the simulation. (B) γ-AAPeptide #43 binding significantly decreases the RMSD values of the γ-AAPeptide binding site (BS), PIP3 BS and TI-loop in PTEN 1D5R. The RMSD values are below 5Å for all sites over the course of the simulation, a decrease in ≥10Å compared to apo-PTEN. (C) γ-AAPeptide #42 binding significantly decreases the RMSD values of the γ-AAPeptide binding site (BS), PIP3 BS and TI-loop in PTEN 1D5R. The RMSD values are below 5Å for all sites over the course of the simulation, a decrease in ≥10Å compared to apo-PTEN.

Figure 3.7: Overall Conformational Changes in PTEN Active Site Residues Following γ-AAPeptide Binding. Dihedral angle analysis of the active site residues in apo-PTEN, γ-AAPeptide #43-bound PTEN and γ-AAPeptide #42-bound PTEN revealed that despite differences in overall γ-AAPeptide structure, they can induce similar changes in specific dihedral angles of select residues as compared to apo-PTEN. (A) The Psi angle in K128 is mostly stabilized at 50° in γ-AAPeptide bound structures (i, iii) while it flips between 150° and -150° in apo-PTEN (ii). (B) Both Phi and Psi angles in G129 are stabilized in similar positions in γ-AAPeptide bound PTEN structures (i, iii) while the angles consistently fluctuate in apo-PTEN (ii). (C) The Chi1 angles in γ-AAPeptide bound PTEN structures flip between 50° and -50° in D162 (i, iii) while the angle is stabilized at -50° in apo-PTEN (ii). (D) The Psi angle in K164 flips between approximately -25° and 125° in γ-AAPeptide bound PTEN structures (i, iii) while it is mostly stabilized at 125° in apo-PTEN (ii). All other dihedral angles and their fluctuations in A-D remain consistent throughout all three PTEN structures. (Phi: black, Psi: blue, Chi1: green, Chi2: magenta, Chi3: red, Chi4, cyan).

Figure 3.8: Select γ-AAPeptide Binding Induces Selective Conformational Changes in Active Site Residues. (A) The fluctuations in Phi, Psi and Chi1 angles in N94 are similar between apo-PTEN and γ-AAPeptide bound PTEN structures. In γ-AAPeptide #42-bound PTEN, the Chi2 angle fluctuates between 0-100° (iii) while the fluctuations in apo-PTEN and γ-AAPeptide #43-bound PTEN range from -50-0° (i, iii). (B) While Psi flips between -50° and 150° in K125 in apo-PTEN and γ-AAPeptide #43-bound PTEN structures, Psi is stabilized at approximately -50° in γ-AAPeptide #42 bound PTEN (iii). (C) Psi in T167 is mostly stabilized at approximately -25° in apo-PTEN and γ-AAPeptide #42-bound PTEN structures (ii, iii) while it flips between -25° and 125° in γ-AAPeptide #43-bound PTEN (i). Additionally, Chi1 in γ-AAPeptide #43-bound PTEN fluctuates more than in apo-PTEN and γ-AAPeptide #42-bound PTEN. (D) The Phi and Psi angles of C124 between γ-AAPeptide #43-bound PTEN and apo-PTEN are similar in comparison to γ-AAPeptide #42-bound PTEN, while the Chi1 angles are mostly similar in their fluctuations between γ-AAPeptide #43-bound PTEN and γ-AAPeptide #42-bound PTEN. (E) The Phi and Psi angles in A126 are stabilized in different positions between all three structures. (F) The G127 Psi angles differ between all three structures. Phi in apo-PTEN flips between -75° and 75° (ii) while it is stabilized at 75° and -75°, respectively, in γ-AAPeptide #43- and γ-AAPeptide #42-bound PTEN (i, iii). (G) Chi3 in R130 apo-PTEN flips between 50° and -75° (ii) while γ-AAPeptide #43- and γ-AAPeptide #42-bound Chi3 is stabilized at -75° and approximately 50°, respectively (i, iii). (H) The Phi angle in G165 is similar between γ-AAPeptide #43-bound PTEN and γ-AAPeptide #42- bound PTEN (i, iii) as compared to control (ii). Psi fluctuates greatly between all three structures, with no identifiable pattern between any of the PTEN structures. All other dihedral angles and their fluctuations in A-H remain consistent throughout all three PTEN structures. (Phi: black, Psi: blue, Chi1: green, Chi2: magenta, Chi3: red, Chi4, cyan, Chi5: yellow).

Figure 3.8 (continued). (A) The fluctuations in Phi, Psi and Chi1 angles in N94 are similar between apo-PTEN and γ- AAPeptide bound PTEN structures. In γ-AAPeptide #42-bound PTEN, the Chi2 angle fluctuates between 0-100° (iii) while the fluctuations in apo-PTEN and γ-AAPeptide #43-bound PTEN range from -50-0° (i, iii). (B) While Psi flips between -50° and 150° in K125 in apo-PTEN and γ-AAPeptide #43-bound PTEN structures, Psi is stabilized at approximately -50° in γ- AAPeptide #42 bound PTEN (iii). (C) Psi in T167 is mostly stabilized at approximately -25° in apo-PTEN and γ-AAPeptide #42-bound PTEN structures (ii, iii) while it flips between -25° and 125° in γ-AAPeptide #43-bound PTEN (i). Additionally, Chi1 in γ-AAPeptide #43-bound PTEN fluctuates more than in apo-PTEN and γ-AAPeptide #42-bound PTEN. (D) The Phi and Psi angles of C124 between γ-AAPeptide #43-bound PTEN and apo-PTEN are similar in comparison to γ-AAPeptide #42-bound PTEN, while the Chi1 angles are mostly similar in their fluctuations between γ-AAPeptide #43- bound PTEN and γ-AAPeptide #42-bound PTEN. (E) The Phi and Psi angles in A126 are stabilized in different positions between all three structures. (F) The G127 Psi angles differ between all three structures. Phi in apo-PTEN flips between -75° and 75° (ii) while it is stabilized at 75° and -75°, respectively, in γ-AAPeptide #43- and γ-AAPeptide #42- bound PTEN (i, iii). (G) Chi3 in R130 apo-PTEN flips between 50° and -75° (ii) while γ-AAPeptide #43- and γ-AAPeptide #42-bound Chi3 is stabilized at -75° and approximately 50°, respectively (i, iii). (H) The Phi angle in G165 is similar between γ-AAPeptide #43-bound PTEN and γ-AAPeptide #42-bound PTEN (i, iii) as compared to control (ii). Psi fluctuates greatly between all three structures, with no identifiable pattern between any of the PTEN structures. All other dihedral angles and their fluctuations in A-H remain consistent throughout all three PTEN structures. (Phi: black, Psi: blue, Chi1: green, Chi2: magenta, Chi3: red, Chi4, cyan, Chi5: yellow).

69 Conformational Changes in R130 Dictate PIP3 Active Site Binding and Orientation

Interestingly, amongst of all the active site residues, R130 has a distinct profile in the Chi3 dihedral angle. In apo-PTEN, the Chi3 dihedral angle flips between 50° and -50° while the γ-AAPeptide #43-bound

PTEN stabilizes Chi3 around -50° and γ-AAPeptide #42-bound PTEN stabilizes Chi3 around 50° (Figure

3.8). Since each γ-AAPeptide stabilizes Chi3 at two distinct angles, where apo-PTEN fluctuates between the two angles, I posited that the R130 orientation may dictate proper PIP3 binding in the active site. To test this hypothesis, I first utilized the IFD γ-AAPeptide #43-bound PTEN structure after PIP3 was docked into the active site and changed the orientation of R130 to mimic the γ-AAPeptide #42-bound PTEN structure. PIP3 was docked into the altered active site with a lower binding affinity (indicated by an increased binding score) and produced an improper orientation where the 3’ phosphate group was not within attacking distance from C124 residue (Figure 3.9A). Conversely, I utilized the IFD γ-AAPeptide #42-bound

PTEN structure after PIP3 was docked into the active site and changed the orientation of R130 to mimic the γ-AAPeptide #43-bound PTEN structure. PIP3 was docked into the altered active site with a higher binding affinity (indicated by a lower binding score). However, despite this maneuver, the orientation was

Figure 3.9: R130 Contributes to Productive PIP3 Binding and Orientation. (A) After PIP3 was docked into the active site of γ-AAPeptide #43-bound PTEN via the IFD protocol (binding score: -14.180kcal/mol), the PIP3 molecule was removed and R130 was placed in an alternate conformation which mimicked the conformation found in γ-AAPeptide #42-bound PTEN. PIP3 was then docked into the active site using the IFD protocol and resulted in a binding score of -12.448kcal/mol in an orientation not conducive for catalysis. (B) After PIP3 was docked into the active site of γ-AAPeptide #42-bound PTEN via the IFD protocol (binding score: -11.621kcal/mol), the PIP3 molecule was removed and R130 was placed in an alternate conformation which mimicked the conformation found in γ-AAPeptide #43-bound PTEN. PIP3 was then docked into the active site using the IFD protocol and resulted in a binding score of -13.237kcal/mol, also in an orientation not conducive for catalysis.

70 not conducive for dephosphorylation, as the 3’ phosphate was positioned away from the catalytic C124 residue (Figure 3.9B).

Conclusion

In this chapter, I employed molecular docking to support the concept that the PD/C2D interface is a suitable region for γ-AAPeptide binding. Both γ-AAPeptide #43 and #42 make hydrogen bonding contacts and utilize their aromatic side chains to fit into the mostly non-polar PD/C2D interface. Furthermore, IFD simulations suggest that PIP3 binding affinity and orientation within the active site can be modulated via

γ-AAPeptide binding. Docking of PIP3 into the PTEN active site after γ -AAPeptide #43 PD/C2D interface binding enhanced the PIP3 binding affinity and positioned the 3’ phosphate group in a more suitable position for catalysis as compared to apo-PTEN, suggesting that γ-AAPeptide #43 binding enhances PIP3 catalysis. In contrast, PIP3 binding after γ-AAPeptide #42 PD/C2D interface binding resulted in the loss of a productive PIP3 binding pose, which corroborates with its inhibitory effect on PIP3 catalysis observed in vitro. Furthermore, analysis of the IFD simulations and MD simulations suggest that γ-AAPeptide binding can induce conformational changes in active site residues, which likely accounts for differential PIP3 binding affinity and orientation in apo-PTEN, γ-AAPeptide #43-bound PTEN and γ-AAPeptide #42-bound

PTEN. Taken together, my findings suggest that γ-AAPeptide #43 and #42 can act as an allosteric activator and an inhibitor, respectively.

Chapter 4: Defining the Role of PTEN Aggregation Propensity in Diseases

Note to Readers: This chapter have been adopted from my accepted manuscript, “Analyzing Aggregation

Propensities of Clinically Relevant PTEN Mutants: A New Culprit in Pathogenesis of Cancer and Other

PTENopathies” published by Taylor & Francis in The Journal of Biomolecular Structure and Dynamics on June 22,

2019, available online: https://www.tandfonline.com/doi/full/10.1080/07391102.2019.1630005

Background

It is now clear that all proteins have the propensity to aggregate and form fibrils under varying cellular milieus, favoring partially unfolded intermediate states (Figure 4.1) [339-341]. Consequently, proteins which are not traditionally defined as amyloidogenic do undergo aggregation as evidenced in certain cancer pathologies [342, 343]. Recently, tumor suppressor protein p53, and its mutant conformers, were shown to undergo protein aggregation, exacerbating the cancer phenotype [343-351]. These findings raise the possibility that inactivation of tumor suppressor proteins via aggregation may participate in cancer and other disease pathologies.

Since aggregation of select clinically relevant p53 mutants have been demonstrated to inactivate its function as a tumor suppressor, I posited that PTEN protein aggregation may serve as an additional mechanism of tumor suppressor inactivation. Indeed, PTEN has structural and functional similarities to p53, including high mutational propensity, a myriad of PTMs, multiple PPIs, IDRs and oligomerization properties (Figure 4.2) [32, 33, 352]. PTEN is highly mutated in multiple diseases, where somatic/germline mutations or PTEN deficiency are seen in cancer patients and patients with PHTS or other PTENopathies

[175, 353-359]. Because PTEN inhibits the oncogenic PI3K pathway, suppresses cell proliferation and

72 migration, and induces apoptosis, PTEN mutations or deficiency drives disease pathology [27, 28, 32, 38,

50, 360]. PTEN mutant C124S not only ablates its function, but also exerts dominant negative effects by dimerizing to PTEN wild-type, similar to select p53 mutants [83, 84]. Therefore, it is indeed possible that mutations in PTEN may disrupt normal dimerization interfaces or cause abnormal PTMs, which exacerbate aggregation propensity rather than the normal physiological dimerization. PTEN also acts as a scaffold protein, nucleating a myriad of proteins via multiple PPIs to regulate cell structure and polarity, as well as participating in many cellular functions like DNA repair, genomic stability, and metabolism [33, 38, 105,

107, 286]. Since domain swapping, PTMs and oligomerization/PPIs are common characteristics which augment amyloid fibril formation, I hypothesized that PTEN has inherent aggregation potential in its wild- type state, which gets exacerbated following mutations/deletions encompassing select amino acids in certain clinical conditions [361-363]. In this chapter, I investigate the aggregation propensity of PTEN wild- type protein and its clinically relevant mutants derived from patients with cancer, PHTS and associated

ASD/macrocephaly phenotypes.

Figure 4.1: Pathway to Protein Aggregation. Destabilization of a natively folded protein monomer, or incomplete folding of a protein after synthesis yields a non-native, partially folded protein monomer which can take several paths. Cellular machinery can either utilize Unfolded Protein Response (UPR), where chaperone proteins aid in folding the protein to a native state, or activate the Endoplasmic Reticulum Associated protein Degradation (ERAD) response, which will degrade non-native protein structures. When both responses fail, the aberrantly folded monomer are left to associate, yielding nascent oligomeric aggregates. Amorphous aggregates lack a specific morphology, while oligomeric and annular aggregates adopt a more defined structure and are the precursors to fibrillary aggregates. Association of oligomeric/annular aggregates generates protofilaments, which align to generate the final amyloid fibril which is ~6-12nm in diameter and 100nm in length. All aggregate structures contain high β-sheet content in comparison to their native protein counterparts, with amyloid fibrils having a signature cross-β pattern. The differences in morphology and β-sheet content allow each type of aggregate to be distinguished by staining and microscopy techniques.

Figure 4.2: p53 and PTEN Properties Significantly Overlap. Venn diagram comparison of p53 and PTEN reveals that although these two proteins have distinct functions, they are the two most important tumor suppressor proteins whose aberrant expression/function contribute to disease pathologies. Additionally, both proteins are involved in mechanisms that drive aggregation processes, such as the presence of IDRs, myriad of mutations, multimerization, and localization to phase separating membrane-less organelles. They both undergo a high-level of PTMs which affects their structure, function and overall stability. p53 has a confirmed aggregation prone region in the wild-type monomer and several mutants have been confirmed to induce wild-type and mutant conformer aggregation. Multiple similar features of p53 are mirrored by PTEN.

Materials and Methods

Curation of Clinically Relevant PTEN Mutations

Known somatic/germline PTEN protein mutations were curated from several publicly available, clinically relevant patient databases: The Catalog of Somatic Mutations in Cancer (COSMIC: https://cancer.sanger.ac.uk/cosmic), The Cancer Genome Atlas (TCGA) (TCGA Research Network: http://cancergenome.nih.gov/.), Simons Foundation Autism Research Initiative (Sfari) Gene

(https://gene.sfari.org/), cBio Portal for Cancer Genomics (https://www.cbioportal.org/), The Human Gene

Mutation Database (HGMD: http://www.hgmd.cf.ac.uk/ac/index.php) and Healthy Exosomes (HEX:

74 https://www.alzforum.org/exomes/hex) [364-369]. Curation was done as follows: missense/nonsense mutations were from all of the aforementioned databases, insertion/deletion mutations were from COSMIC,

TCGA, cBio Portal, and frameshift mutations were from COSMIC and Sfari Gene. Once compiled, the duplications were removed leaving 1,523 distinct mutations, at the cDNA level, for analysis.

TANGO Analysis of Wild-Type and Clinically Relevant PTEN Mutations

Each mutated PTEN sequence along with wild-type PTEN was formatted in a plain text document as an input file and run through the TANGO algorithm [370]. The website for TANGO (http://tango.crg.es/) provides explicit instructions on input file format, where information for the analysis of each sequence was specified as follows: tango Name nt="N" ct="N" ph="7" te="310.15" io="0.1" tf="0" seq="sequence”).

The specifications nt and ct represent capping of the N-terminus or C-terminus, respectively (Y= Yes,

N= No). The pH (ph), temperature (te; in Kelvin), ionic concentrations (io; in Molar units) and organic solvent (tf; in Molar units) were selected to approximate cellular conditions. The PTEN wild-type or PTEN mutant sequences were uploaded to TANGO in full. Multiple sequences were provided in the input files, with one sequence and its corresponding running conditions to a line. A single output file in plain text formatting was provided for each mutant after the TANGO run, which were compiled into excel spreadsheets for further analysis. A short linear region is considered to be aggregation-prone if five or more consecutive amino acid residues have an aggregation score (also known as aggregation propensity or

TANGO score) of ≥5%. Thus, an overall aggregation score was calculated as the average value of the per residue scores of the entire aggregation-prone region. The data, output and analysis that support the findings of this study are openly available in Mendeley Data at reserved

DOI: http://doi.org/10.17632/3dxhn32c78.2.

75 Hydrophobicity Analysis of Wild-Type and Clinically Relevant PTEN Mutations

Mutant PTEN protein sequences that either increased or decreased aggregation propensity of PTEN wild-type regions, as well as PTEN wild-type protein alone, were run through the Expasy Protscale module, with a sliding window of three amino acids, to determine the hydrophobicity of each residue in the sequence based on the Kyte-Doolittle Scale [371]. Each sequence in full was analyzed by the Expasy Protscale module and the output was downloaded into the excel spreadsheets already created for each respective sequence. The hydrophobicity score for each aggregation-prone region was calculated as the average hydrophobicity score of each residue in the aggregation-prone region. The data, output and analysis that support the findings of this study are openly available in Mendeley Data at reserved

DOI: http://doi.org/10.17632/3dxhn32c78.2.

Secondary Structure Analysis of Wild-Type and Clinically Relevant PTEN Mutations

The Advanced Protein Secondary Structure Prediction algorithm (APSSP; http:// imtech.res.in/raghava/apssp/) was used to predict β-sheet (E), ɑ-helix (H) and coil (C) propensity of each residue in the sequences. Each sequence in full was analyzed by the algorithm and the output was downloaded into an excel spreadsheet created for each respective sequence. The number of residues per predicted structures (E, H or C) was counted for each aggregation-prone region and divided by the total number of residues in the region to determine the contribution of a secondary structure to each aggregation- prone region. Change in secondary structure propensity was calculated by subtracting the propensity of a structure (E, H or C) in the mutant from the propensity of the same structure in wild-type. Positive values indicated a structural gain, while negative values indicated a structural loss. All changes were reported as the percent frequency of gained or lost structures. The data, output and analysis that support the findings of this study are openly available in Mendeley Data at reserved DOI: http://doi.org/10.17632/3dxhn32c78.2.

76 Intrinsic Disorder Analysis of Wild-Type and Clinically Relevant PTEN Mutations

The intrinsic disorder meta-analysis predictor DANN was used to determine the intrinsic disorder propensity of each residue in the sequences [372]. Sequences were run through DANN in FASTA format and a plain text output file was generated for each sequence. The data was respectively compiled into the spreadsheets created for each sequence. The intrinsic disorder score for an aggregation-prone region was calculated as the average intrinsic disorder score of each residue in the region. A score >0 represents disorder while scores less than <0 are considered ordered. The data, output and analysis that support the findings of this study are openly available in Mendeley Data at reserved

DOI: http://doi.org/10.17632/3dxhn32c78.2.

Statistical Analysis

In order to determine changes in hydrophobicity scores in high or low scoring aggregation-prone regions dominated by β-sheet, ɑ-helical or coil structure, hydrophobicity values were grouped into five categories: 1. β-sheet structure (average E score of >0.5) and high aggregation potential (≥50%), n=80; 2.

β-sheet structure (average E score of >0.5) and low aggregation potential (≤50%), n=76; 3. ɑ-helical structure (average H score of >0.5) and high aggregation potential (≥50%), n=66; 4. ɑ-helical structure

(average H score of >0.5) and low aggregation potential (≤50%), n=53; 5. Coil structure (average C score of >0.5) for all aggregation prone regions, n=17. The Shapiro-Wilk normality test revealed that three of the five categories were not normally distributed. Therefore, the non-parametric Kruskal-Wallis test was used to determine if there were significant differences between the mean ranks differences of hydrophobicity scores in each category, *p≤0.05. The data, output and analysis that support the findings of this study are openly available in Mendeley Data at reserved DOI: http://doi.org/10.17632/3dxhn32c78.2.

77 Results

PTEN Contains Five Aggregation-Prone Regions

The aggregation predictor TANGO successfully predicted an aggregation-prone region present in wild-type p53, validating its utility for analysis of PTEN protein and its mutants [343, 350]. Full-length human PTEN wild-type protein was run through the TANGO algorithm under conditions approximating the cellular milieu. Five regions in PTEN protein were predicted to be prone to aggregation, including the

Figure 4.3: Putative Aggregation-Prone Regions of PTEN. (A) Schematic of full-length PTEN wild-type protein with the colors depicting the various domains and critical amino acid sequence stretches. PTEN has an N-terminal PIP2 Binding Module (aa residues 1-6), a Phosphatase Domain (aa residues 7-185), a C2 Domain (aa residues 186-350) and an intrinsically disordered C-terminal tail (aa residues 351-403). (B) Aggregation propensity was determined using the algorithm TANGO, providing physiological conditions (pH 7, 0.1M ionic strength, temperature: 310.15 Kelvin, 0% organic solvent). Five aggregation prone regions were revealed in PTEN. The regions are differentially highlighted in the PDB 1D5R crystal structure. Accessibility of the aggregation-prone regions are shown for PTEN as a monomer and as a dimer. Under dephosphorylated conditions, multiple monomers of PTEN associate at the surfaces of the Phosphatase Domains, while the C-tails of each monomer interact with the opposing monomers Phosphatase Domain, creating a dimer. This mechanism is known as domain swapping, which is thought to precede aggregation as it enhances intermolecular interactions between monomers. Since PTEN dimers also have exposed aggregation-prone regions, aggregation of PTEN wild-type/mutant homo- and hetero-dimers is possible. (C) Protein aggregation analysis utilizing TANGO revealed five regions of PTEN to have high aggregation propensity. The cβ1 strand has the lowest aggregation propensity of the 5 regions, at 22%. The residues of the pɑ4 helix (magenta) and the cβ6/cɑ1residues (red) have a peak propensity of ~40-55%, while the pɑ6 helix (orange) has a peak aggregation propensity of ~55-65%. The cβ7 strand (purple) has the highest peak aggregation propensity between 80-85%.

78 pɑ4 helix (residues 133-140), pɑ6 helix (residues 174-182), cβ1 strand (residues 191-195), cβ6 strand/cɑ1 helix (residues 271-280) and cβ7 strand (residues 315-321) (Figure 4.3). The PTEN cβ7 strand had the highest aggregation propensity of ~85%, whereas the pɑ4, pɑ6 and cβ6 strand/cɑ1 helix had a predicted aggregation propensity range between ~50-60% and the cβ1 strand had a predicted aggregation of ~22%

(Figure 4.3C). These aggregation-prone regions are rich in hydrophobic residues (33-100%), poor in polar

(20-66%) or charged residues (0-10%) and are exclusively located within the structured Phosphatase

Domain (PD) and C2 Domain (C2D) of PTEN. Notably, the pɑ4 helix is buried within the core of the PD, while the other four aggregation sites are exposed to the surface (pɑ6 helix, cβ1 strand, cβ7 strand and cβ6 strand/cɑ1 helix). In summary, PTEN has five putative aggregation-prone regions with distinct physico- chemical properties.

Alternative Translational PTEN Proteoforms Have a Distinct Aggregation-Prone Region

The four additional PTEN proteoforms generated via alternative translation were identified to have increased aggregation propensity compared to wild-type PTEN protein in my analysis (Figure 4.4). PTEN translational proteoforms L, M, N and O are all produced from alternative initiation codons upstream of the canonical initiation codon for PTEN wild-type [85]. These N-terminally extended PTEN conformers retain catalytic phosphatase activity and are diffuse in the cytoplasm, like PTEN wild-type [85]. TANGO analysis revealed that each translational variant contains a peptide sequence, “LSSFFF,” twelve residues upstream of the canonical first residue in PTEN, which has a predicted aggregation propensity of ~7-8% (Figure 4.4).

Figure 4.4: PTEN Alternative Translational Proteoforms Have an Aggregation-Prone Region. Four PTEN proteoforms (PTEN-L, -M, -N, -O) are generated by alternative translational initiation codons upstream of the canonical AUG start. Each N-terminal extension contains a unique sequence “LSSFFF” located 12 amino acids upstream of the first amino acid in PTEN wild-type protein is also aggregation prone.

Curation of Clinically Relevant PTEN Mutations and Their Aggregation Propensities

To examine whether mutations in PTEN influences its aggregation potential, 1,523 known somatic/germline PTEN protein mutations were identified for curation from several publicly available, clinically relevant patient databases. Together, these databases comprehensively covered patients with

PTEN-related diseases, including several malignancies/cancer, PHTS, autism spectrum disorders (ASD) and other neurological disorders/macrocephaly. The databases include COSMIC, TCGA (TCGA Research

Network: http://cancergenome.nih.gov/.), Sfari Gene, cBio Portal, HGMD and HEX [364-369]. Of the

1,523 PTEN mutations analyzed, 91 were insertions/deletions, 880 were missense/nonsense and 601 were frameshift mutations. Interestingly, 49 of the 601 frameshift mutations produced truncated PTEN protein with no de novo frame-shifted sequence, like the product of a nonsense mutation. Herein, I redefined these

Table 4.1: PTEN Mutations that Change Aggregation Propensity Compared to PTEN Wild-Type.

80 49 mutants and all nonsense mutants as truncations (Table 4.1). Of the 1,523 collective mutations, 274 mutants were predicted to increase aggregation propensity (increased aggregation score and/or increased length of the aggregation-prone region) and 119 were predicted to decrease aggregation propensity

(decreased aggregation score and/or decreased length of aggregation-prone region) of PTEN wild-type protein at the mutated residue and/or at the surrounding residues (Table 4.1 and Supplement S2-S3: DOI: http://doi.org/10.17632/3dxhn32c78.2). Notably, 448 mutants removed one or more aggregation-prone regions of PTEN in the resulting protein, but did not increase or decrease the aggregation propensity at the location of the mutation (Supplement S4: DOI: http://doi.org/10.17632/3dxhn32c78.2). In my subsequent analysis, I focused on mutants that are predicted to increase or decrease the aggregation propensity at/around the location of the mutation.

Figure 4.5: Select Mutations in PTEN Wild-Type Protein Increase Aggregation Propensity. (A) Representation of the distribution of the 89 missense/truncation mutations that increase aggregation propensity in PTEN wild-type protein. (B) Distribution of the 11 insertion/deletion mutations that increase PTEN wild-type aggregation propensity. Boxed amino acids represent the wild-type residue, numbers represent the amino acid location within wild-type PTEN, and colored amino acids represent the mutation. Deletion= Del, Insertion= Ins, Duplication= Dup, truncations represented by a red asterisk. Figure not drawn to scale. Mutations that Increase Hydrophobicity Enhance Aggregation Propensity of PTEN are Frequent

Mutation of a polar or charged residue to a more hydrophobic residue within or adjacent to aggregation-prone regions, or highly hydrophobic regions, drastically increased the local predicted

81 aggregation propensity of aggregation-prone and non-aggregation-prone regions in PTEN. Of the PTEN mutants that increased aggregation propensity, approximately 75% of the missense mutants had replaced the wild-type residue to a hydrophobic residue, while only ~18% and ~7% were mutated to a polar or charged residue, respectively (Figure 4.5A). All deletion mutations removed one or more charged residue(s) from the PTEN sequence, and three out of the four insertion mutations inserted one or more hydrophobic residues into the PTEN sequence (Figure 4.5B).

The frequency difference in various amino acid mutations in PTEN mutants that increased or decreased aggregation propensity revealed a mutational preference of chemical character. Indeed, missense/insertion mutations to a hydrophobic residue were more prevalent for mutants that increase aggregation propensity, as compared to mutants that decrease aggregation propensity (Figure 4.6). Whereas mutation to a polar or charged residue was more specific to mutants that decreased aggregation propensity

(Figure 4.6). Notably, mutation to lysine or arginine was selective only to mutants that decreased aggregation propensity, while mutation to tryptophan, phenylalanine or valine was selective only to mutants that increased aggregation propensity (Figure 4.6).

Figure 4.6: Increased PTEN Aggregation Propensity is Dominated by Hydrophobic Mutations. The frequency of mutation to each amino acid residue was calculated for missense/truncation/insertion/deletion mutants that increased aggregation propensity (n=91) and decreased aggregation propensity (n=68). The frequency values for mutants that decrease aggregation propensity was subtracted from that of mutants which increase aggregation propensity. Negative values indicate a higher mutation frequency in mutants that decrease aggregation propensity, while positive values represent higher mutation frequency in mutants that increase aggregation propensity. Blue asterisks indicate amino acids that are specific only to mutants that decrease aggregation (Lysine [K], Arginine [R]) while red asterisks indicate amino acids that are specific only to mutants that increase aggregation propensity (Tryptophan [W], Phenylalanine [F], Valine [V]). Mutations to hydrophobic amino acids are much more frequent among PTEN mutants that increase aggregation propensity of PTEN wild-type, demonstrating the strong influence of hydrophobic character on aggregation potential.

82 Frameshift Mutations in PTEN Severely Affect its Aggregation Potential

Of the 553 PTEN frameshift mutations, 174 mutants exhibited increased aggregation propensity

(Table 4.1), with three mutants demonstrating two aggregation-prone regions in the novel frame-shifted sequence. Interestingly, 70 aggregation-prone regions occurred at the junction between the wild-type and frame-shifted sequence, while 37 PTEN frameshift mutations only enhanced the aggregation propensity of an already aggregation-prone region of PTEN wild-type. The other 108 mutants increased aggregation propensity after the start of the frameshift, within the novel sequence. Approximately 52% of the aggregation-prone frameshift mutations have an aggregation propensity of ≥80% (Figure 4.7), indicating that these frameshift mutations are likely to severely effect PTEN structure and induce aggregation.

Conversely, only 20 frameshift mutants decreased aggregation propensity of an aggregation-prone region in PTEN wild-type protein (Table 4.1).

Figure 4.7: PTEN Frameshift Mutations Drastically Increase Aggregation Propensity. Of the 552 PTEN frameshift mutations curated, 174 revealed increase aggregation propensity at or downstream of the frameshift mutation. Each frameshift mutation is represented as a node. The color of the node indicates the average percent aggregation propensity of 5 consecutive residues with the highest propensity within the total aggregating region (green: 5-20%; beige: 20-40%; yellow: 40-60%; orange 60-80%; red: 80-100%). A majority of the frameshift mutations (52%) have an aggregation propensity of 80%, indicating that these mutants likely impact PTEN structure and aggregation propensity.

83 Aggregation Propensity is Associated with Highly Ordered Structures

Herein, I interrogated the structural propensities of the PTEN mutants that changed aggregation propensity. First, it was noted that mutants that increased or decreased aggregation propensity of PTEN wild-type protein had distinct changes in their secondary structures. Mutations that increase aggregation propensity tend to increase β-sheet character, as compared to mutants that decrease aggregation propensity, which predominantly gain coil character (Figure 4.8A). Conversely, when one structure is gained, another is lost. Thus, mutants that increased aggregation propensity predominantly lost their coil secondary structure, while mutants that decreased aggregation propensity mostly lost β-sheet or ɑ-helical character

(Figure 4.8A). Overall, β-sheet structure is more frequently observed in mutants that increased aggregation propensity, while mutants that decreased aggregation propensity lose more ordered structures (ɑ-helix and

β-sheet) to become less structured coils.

Second, a highly accurate intrinsic disorder meta-predictor, DANN, was used to generate the intrinsic disorder scores for each residue of each PTEN mutant conformer and the average disorder score was calculated for the aggregation-prone region [372]. DANN predictions revealed that PTEN mutants which increased aggregation propensity are highly structured, with only 1.4% (4 out of 277 aggregation- prone regions) of aggregation-prone regions predicted to be disordered (Figure 4.8B). Conversely, select mutations which ablated aggregation potential in aggregation-prone regions demonstrated intrinsic disorder propensity, while mutations causing decreased, albeit prevalent, aggregation propensity remained highly structured (Figure 4.8B).

84 Figure 4.8: PTEN Mutants that Increase Aggregation Propensity Favor Ordered Structures. (A) Of the 274 PTEN mutants with increased aggregation propensity, 136 changed the aggregation propensity of one or more wild-type residues. The difference in structural propensity of β-sheets (E), α-helices (H) and coiled-coils (C) was calculated (mutant – wild-type) and the regions which gained or lost a structure was determined. In total, there were 45 different structural losses among the mutants (one mutant can lose one or more of the three structures), yielding 43 structural gains (one mutant can gain one or more of the three structures). There is a high tendency for mutants which increase aggregation propensity to cause a shift from a coiled-coil structure (lost) to a β-sheet (gained). Of the 119 mutants that decreased aggregation propensity of PTEN wild-type, there were 66 structural losses and 68 structural gains. Mutants that decrease aggregation propensity have a much higher tendency to lose more ordered structures like β-sheets and α-helices and adopt a less ordered coiled-coil structure. (B) Intrinsic disorder analysis using the algorithm DANN shows that mutation-induced increase in aggregation potential falls in highly structured regions of PTEN. Only 1.4% of aggregation-prone regions (4/277) are intrinsically disordered. n=277. PTEN mutants which decrease aggregation propensity of PTEN wild-type can increase intrinsic disorder propensity. Only select mutants which have 0% aggregation propensity are intrinsically disordered, while aggregation-prone regions in mutants that retain some degree of aggregation potential remain ordered structures. n= 119. DANN scores and TANGO scores were calculated as the average score of all residues in the aggregation prone region. DANN score values greater than 0 indicate disordered regions, values less than 0 indicate ordered regions.

Contribution of Structural and Chemical Properties to PTEN Conformer Aggregation

Since structural and chemical properties have a great influence over protein aggregation, PTEN mutants that increase or decrease aggregation potential were analyzed for changes in two critical factors that drive protein aggregation, i.e. 3D structural alterations and chemical characteristics such as

85 hydrophobicity. To define the physico-chemical properties that render a select number of PTEN mutants aggregation-prone to varying degrees (TANGO score above 5%), the degree of hydrophobicity within three distinct secondary structures was analyzed. Advanced Protein Secondary Structure Prediction was utilized to predict structural propensity, which included extended β-sheet (E), ɑ-helical (H) or coil (C) regions.

Notably, PTEN aggregation-prone regions in pathogenic mutants are fairly split between β-sheet and ɑ- helical structures, with 156 and 151 aggregation-prone structures, respectively. Only 17 mutants were dominated by a coil structure. Therefore, mutants dominated by β-sheets and ɑ-helices were then split into high scoring (≥50% aggregation score) and low scoring aggregation (<50% aggregation score) and all coil structures were grouped together. Analysis of the distribution of hydrophobicity scores (using the Kyte-

Doolittle [KD] prediction available through ProtScale) revealed that regions with high aggregation potential

Figure 4.9: Both Hydrophobicity and Secondary Structure Influence the Degree of Aggregation Propensity. Distribution of hydrophobicity scores of high and low scoring aggregating regions in PTEN mutants, based on their dominant structure (average structural propensity of >0.5). Mutants that increased and decreased PTEN wild-type aggregation propensity were considered in the analysis, under the condition that they contained an aggregation-prone region (5 consecutive residues of 5% or more aggregation score). β-sheet structures are shown in dark and light red, ɑ-helical structures are shown in dark and light blue and coiled-coil structures are shown in green. Since only several aggregation prone-regions were dominated by coiled-coil structure, these aggregation-prone regions were not separated into high/low aggregating categories in order to keep a larger sample size suitable for statistical analysis. A Kruskal-Wallis test revealed that PTEN mutants with high aggregation propensity and dominant β-sheet characteristic have a significantly higher hydrophobicity score, as compared to low aggregation scoring mutants with β-sheet characteristics and all aggregating mutants with ɑ-helical or coiled-coil structure (*p<0.05). High scoring β-sheet structures, n=80; low scoring β-sheet structures, n=76; high scoring ɑ-helical structures, n=66; low scoring ɑ-helical structures, n=53; coiled-coil structures, n=17. Hydrophobicity score is based off the Kyte-Doolittle index.

86 and β-sheet character had a higher hydrophobicity score, as compared to β-sheet structures with low aggregation potential and all ɑ-helical and coil aggregation-prone regions (Figure 4.9).

Conclusion

The role of protein aggregation in cancer as a means to inactivate tumor suppressor functions has recently emerged, with p53 protein as the most well-known example. Like p53, PTEN is a highly mutated tumor suppressor protein with multiple intrinsic properties which can augment aggregation when altered

[352]. However, until now, the aggregation potential of PTEN has not been interrogated. In this chapter, I elucidated five putative aggregation-prone regions in full-length human PTEN wild-type protein.

Additionally, I identified >270 mutants which increase the aggregation propensity of PTEN wild-type.

Consistent with studies on bona fide amyloidogenic proteins, PTEN mutants with high aggregation potential demonstrated tendencies for high hydrophobicity and ordered regions of the protein, with a tendency for β-sheet formation. My analysis suggests that PTEN mutants found in cancer, and other

PTENopathies, may drive disease progression via aggregation. Ultimately, validation of PTEN aggregation in vivo will likely catalyze the development of novel therapeutics targeting PTEN protein aggregation in multiple diseases.

Chapter 5: Discussion

Current therapies for diseases, including cancer, associated with hyperactive PI3K pathway signaling largely rely on a myriad of kinase inhibitors that target multiple kinases driving the PI3K cascade.

However, these kinase inhibitor therapies cause off-target effects associated with varying toxicities and therapy-resistance, warranting the need for novel therapy approaches [287]. To fill this therapy gap, enhancing intracellular PTEN activity represents a novel alternative therapeutic paradigm to mitigate pathogenic PI3K signaling or compromised PTEN levels or function. [303]. PTEN mutations driving multiple diseases and cancer is well established [4, 8, 33]. However, it is increasingly becoming clear that reduction in PTEN function or expression due to non-genomic PTEN inactivation (i.e. transcriptional, post-transcriptional and post-translational mechanisms) is a major driver in disease initiation and progression across multiple tissue types [32, 33]. Thus, my thesis is based on the premise that enhancing the intracellular enzymatic activity and function of residual PTEN protein in such diseases is potentially a novel therapeutic strategy.

Recently, PTEN restoration therapy has proved to successfully suppress tumorigenic features in cancer cell lines and preclinical studies, limiting or resolving tumor burden in these models [88, 288-295,

297, 298]. Thus, re-activation or enhancement of endogenous PTEN activity/function is an attractive alternative or adjunctive therapy to existing standard chemotherapies and kinase inhibitors. While restoration of endogenous intracellular PTEN level holds promise as a future clinical therapy; delivery modes for PTEN gene or protein for therapy remain challenging in the clinic. To this end, my work has focused on developing small molecules that enhance PTEN activity in cells with residual PTEN expression and hyperactive PI3K pathway signaling. I have successfully identified one lead small molecule that directly targets PTEN protein and enhances its lipid phosphatase activity in biochemical and cell-based

88 assays. To my knowledge, this strategy is the first of its kind that reduces oncogenic potential of cancer cells by directly enhancing PTEN enzymatic activity.

To develop small molecules that enhance PTEN activity, I first utilized FTMap analysis to identify a small molecule binding hotspot on PTEN at the interface of the PD/C2D. Targeting this region in vitro using a synthesized peptide which competitively binds to the interface enhanced PTEN lipid phosphatase activity in a biochemical in vitro microtiter plate assay. I then identified a select small molecule peptidomimetic, a γ-AAPeptide subclass containing an adamantyl ring, which enhanced PTEN enzymatic activity in vitro and suppressed PI3K pathway signaling in A549 lung cancer and OVCAR3 ovarian cancer cells. Moreover, treatment with this promising peptidomimetic, γ-AAPeptide #43, decreased cell proliferation, migration and cell cycle progression, inhibiting the oncogenic potential of lung cancer cells harboring wild-type PTEN protein. Taken together, these results indicate that γ-AAPeptide #43 directly enhances PTEN lipid phosphatase activity in vitro and likely targets endogenous PTEN in cell-based assays.

Notably, all γ-AAPeptides containing an adamantyl side chain were able to enhance PTEN lipid phosphatase activity in vitro, with the exception of γ-AAPeptide #42. These in vitro studies suggest that these select γ-AAPeptides rely on an adamantyl containing side chain for specific binding while additional functional groups modulate their effects on PTEN function.

Since both FTMap analysis and titration with a synthetic helical peptide mimicking the pɑ6 helix implicated the PD/C2D interface as a probable small molecule binding hotspot, I hypothesized that

γ-AAPeptides target the PD/C2D interface and elicit their modulatory effects on PTEN activity. To test this hypothesis, molecular docking studies were employed to determine the binding affinity of γ-AAPeptides for this interface. IFD predicted favorable binding of both activator γ-AAPeptide #43 and inhibitor

γ-AAPeptide #42, suggesting that γ-AAPeptides harboring aromatic and adamantyl side chain(s) are suitable to directly target this interdomain region. Secondary IFD docking analysis using the PTEN substrate, PIP3, revealed that prior binding of activator γ-AAPeptide #43 at the PD/C2D interface could enhance PIP3 active site binding, with a productive orientation that would allow for catalytic 3’ phosphate

89 dephosphorylation. In contrast, prior binding with inhibitor γ-AAPeptide #42 resulted in increased affinity for PIP3 active site binding; however, the binding did not place the 3’ phosphate within an ideal distance for nucleophilic attack by the catalytic cysteine residue. These docking results corroborate well with my in vitro studies, highlighting the differential effects of γ-AAPeptides on PIP3 hydrolysis.

Furthermore, MD simulations suggest that select γ-AAPeptide binding can stabilize PTEN active site residues, which likely accounts for differential PIP3 binding and orientation in the active site. IFD analysis of γ-AAPeptide #43 and #42 PD/C2D interface binding demonstrated that each γ-AAPeptide makes distinct contacts within the PD/C2D domain interface; therefore, it is likely that these differential interactions modulate their enhancement or inhibitory effects on PTEN lipid phosphatase activity. Taken together, γ-AAPeptides are suitable small molecule peptidomimetics to target PTEN protein for therapeutic purposes. Furthermore, I have identified the first small molecule which directly regulates PTEN lipid phosphatase activity, likely through allosteric regulation of the PTEN active site conformation.

PTEN enzymatic activity has been showed to be modulated by allosteric mechanisms in multiple fashions. A recent study has demonstrated the ability of select mutations, especially in the PD and C2D, to alter PTEN core structure, translating to the active site and functionally abrogating enzymatic activity [373].

In addition, PTEN is allosterically activated by binding to PIP2 at the N-terminal PIP2 binding module [58,

62-64]. PIP2 binding at the PIP2 binding motif, distinct from the active site, results in conformation changes in PTEN and enhanced PIP3 hydrolysis [65]. My PTEN enzyme kinetics studies are consistent with the notion of allostery. Thus, the use of small molecules, like γ-AAPeptide #43, to induce conformational changes in PTEN which allosterically enhance phosphatase activity is plausible. Herein, my work exemplifies the benefits of targeting PTEN wild-type to enhance catalytic activity as a means to counteract

PI3K pathway hyperactivation.

In addition to non-genomic mechanisms which suppress PTEN wild-type function, mutation- induced PTEN inactivation is observed in almost all types of cancers and with high frequency in prostate, brain, skin and endometrial cancers, as well as in patients with PHTS [4, 12, 13, 52]. Since PTEN is a

90 frequently mutated tumor suppressor protein like p53, with shared structural commonalities like IDRs,

PTMs, and oligomerization properties [32, 33, 352], I hypothesized that similar mechanisms that drive p53 inactivation observed in diseases may exist for certain select PTEN mutants. To this end, I derived more than 1,500 clinically relevant PTEN mutants from various patient databases and interrogated PTEN inactivation mechanisms, focusing to aggregation, as seen with select p53 mutants. These p53 mutants are observed to aggregate in vitro, in cancer cell lines and in patient biopsies, ultimately inactivating endogenous p53 tumor suppressor functions, accelerating the cancer phenotype [343-351]. Since the aggregation propensity of PTEN wild-type or mutant PTEN has not been characterized, I tested the hypothesis that select clinically relevant PTEN mutations found in patients may functionally inactive endogenous intracellular PTEN by driving aggregation.

Indeed, my studies revealed that protein aggregation is a novel potential mechanism of inactivation for PTEN mutants that may drive disease pathogenesis in cancer and PHTS. I searched publicly available patient databases for known PTEN missense, nonsense, insertion/deletion and frameshift mutations and used online algorithms to predict each mutant’s aggregation propensity by TANGO, hydrophobicity by

Expasy’s Protscale KD index and secondary structure of PTEN and its mutants utilizing APSSP. I first identified five regions of predicted aggregation in the PTEN wild-type core domains. Furthermore, my analysis revealed >270 distinct PTEN mutations in patients with increased aggregation propensity compared to PTEN wild-type protein. Aggregation prone regions induced by mutations are largely evenly split between β-sheet and α-helical structures; however, there is a significant trend towards increased hydrophobicity following a mutation in PTEN. Further analysis revealed that high aggregation scoring mutants with dominant β-sheet characteristics have an increased hydrophobicity score as compared to other high/low aggregation scoring structures. Thus, my work indicates that future verification of PTEN mutant aggregation in vivo is warranted to define aggregation as a bona fide inactivation mechanism that may be pathogenic for select PTEN mutants. Furthermore, my analysis will also aid in the generation of small molecules which can target the specific physico-chemical properties of PTEN mutants, resulting in the

91 dissolution of PTEN aggregates, freeing endogenous PTEN for functional re-activation as a disease treatment.

In summary, my work highlights the critical utility of small molecules in PTEN restoration therapy.

Enhancing endogenous PTEN wild-type activity via γ-AAPeptide #43 antagonized PI3K pathway signaling and suppressed oncogenic features of tumor progression. This therapy would be most beneficial for cancer patients with hyperactive PI3K pathway signaling resulting from an aberration in PI3K pathway kinases, or reduced PTEN expression driven by non-genomic mechanisms. Additionally, targeted re-activation of

PTEN mutants remains an important avenue of therapy in patients with cancers or PHTS, where a PTEN mutation drives disease pathology. Future assessment of mutant PTEN aggregation-induced functional inactivation will aid in the development of clinically relevant small molecules that target pathogenic PTEN aggregates for dissolution, generating free, active PTEN molecules for restoration of PTEN function.

Notably, γ-AAPeptides have been shown to disrupt Aβ amyloid found in the brains of patients with

Alzheimer’s, indicating γ-AAPeptides can be generated to target protein aggregates for dissolution [309].

Thus, small molecule based PTEN restoration therapy provides a novel alternative or adjunctive therapy that can be used to develop a personalized clinical rationale for treating multiple diseases, including a myriad of malignancies driven by compromised PTEN activity/function and/or hyperactive PI3K pathway signaling.

References

1. Alimonti, A., et al., Subtle variations in Pten dose determine cancer susceptibility. Nat Genet, 2010. 42(5): p. 454-8. 2. Chalhoub, N. and S.J. Baker, PTEN and the PI3-kinase pathway in cancer. Annu Rev Pathol, 2009. 4: p. 127-50. 3. Di Cristofano, A., et al., Pten is essential for embryonic development and tumour suppression. Nat Genet, 1998. 19(4): p. 348-55. 4. Podsypanina, K., et al., Mutation of Pten/Mmac1 in mice causes neoplasia in multiple organ systems. Proc Natl Acad Sci U S A, 1999. 96(4): p. 1563-8. 5. Teng, P., et al., Small Antimicrobial Agents Based on Acylated Reduced Amide Scaffold. J Med Chem, 2016. 59(17): p. 7877-87. 6. Liaw, D., et al., Germline mutations of the PTEN gene in Cowden disease, an inherited breast and thyroid cancer syndrome. Nat Genet, 1997. 16(1): p. 64-7. 7. Marsh, D.J., et al., Mutation spectrum and genotype-phenotype analyses in Cowden disease and Bannayan-Zonana syndrome, two hamartoma syndromes with germline PTEN mutation. Hum Mol Genet, 1998. 7(3): p. 507-15. 8. Milella, M., et al., PTEN: Multiple Functions in Human Malignant Tumors. Front Oncol, 2015. 5: p. 24. 9. Marsh, D.J., et al., PTEN mutation spectrum and genotype-phenotype correlations in Bannayan- Riley-Ruvalcaba syndrome suggest a single entity with Cowden syndrome. Hum Mol Genet, 1999. 8(8): p. 1461-72. 10. Malaney, P., et al., PTEN Physically Interacts with and Regulates E2F1-mediated Transcription in Lung Cancer. Cell Cycle, 2018. 17(8): p. 947-962. 11. Zhu, X., et al., Loss and reduced expression of PTEN correlate with advanced-stage gastric carcinoma. Exp Ther Med, 2013. 5(1): p. 57-64. 12. Yehia, L. and C. Eng, 65 YEARS OF THE DOUBLE HELIX: One gene, many endocrine and metabolic syndromes: PTEN-opathies and precision medicine. Endocr Relat Cancer, 2018. 25(8): p. T121-t140. 13. Mester, J. and C. Eng, When overgrowth bumps into cancer: the PTEN-opathies. Am J Med Genet C Semin Med Genet, 2013. 163c(2): p. 114-21. 14. Cai, Z. and G.L. Semenza, PTEN activity is modulated during ischemia and reperfusion: involvement in the induction and decay of preconditioning. Circ Res, 2005. 97(12): p. 1351-9. 15. Li, W., et al., PTEN degradation after ischemic stroke: a double-edged sword. Neuroscience, 2014. 274: p. 153-61. 16. Huang, J. and C.D. Kontos, Inhibition of vascular smooth muscle cell proliferation, migration, and survival by the tumor suppressor protein PTEN. Arterioscler Thromb Vasc Biol, 2002. 22(5): p. 745-51. 17. Knafo, S., et al., PTEN recruitment controls synaptic and cognitive function in Alzheimer's models. Nat Neurosci, 2016. 19(3): p. 443-53. 18. Domanskyi, A., et al., Pten ablation in adult dopaminergic neurons is neuroprotective in Parkinson's disease models. FASEB J, 2011. 25(9): p. 2898-910. 19. Kreiner, G., et al., A neuroprotective phase precedes striatal degeneration upon nucleolar stress. Cell Death Differ, 2013. 20(11): p. 1455-64.

93 20. Plotkin, J.L. and D.J. Surmeier, Impaired striatal function in Huntington's disease is due to aberrant p75NTR signaling. Rare Dis, 2014. 2(1): p. e968482. 21. Ortega-Molina, A., et al., Pten positively regulates brown adipose function, energy expenditure, and longevity. Cell Metab, 2012. 15(3): p. 382-94. 22. Ortega-Molina, A. and M. Serrano, PTEN in cancer, metabolism, and aging. Trends Endocrinol Metab, 2013. 24(4): p. 184-9. 23. Li, D.M. and H. Sun, TEP1, encoded by a candidate tumor suppressor locus, is a novel protein tyrosine phosphatase regulated by transforming growth factor beta. Cancer Res, 1997. 57(11): p. 2124-9. 24. Steck, P.A., et al., Identification of a candidate tumour suppressor gene, MMAC1, at chromosome 10q23.3 that is mutated in multiple advanced cancers. Nat Genet, 1997. 15(4): p. 356-62. 25. Li, J., et al., PTEN, a putative protein tyrosine phosphatase gene mutated in human brain, breast, and prostate cancer. Science, 1997. 275(5308): p. 1943-7. 26. Lee, J.O., et al., Crystal structure of the PTEN tumor suppressor: implications for its phosphoinositide phosphatase activity and membrane association. Cell, 1999. 99(3): p. 323-34. 27. Myers, M.P., et al., P-TEN, the tumor suppressor from human chromosome 10q23, is a dual- specificity phosphatase. Proc Natl Acad Sci U S A, 1997. 94(17): p. 9052-7. 28. Maehama, T. and J.E. Dixon, The tumor suppressor, PTEN/MMAC1, dephosphorylates the lipid second messenger, phosphatidylinositol 3,4,5-trisphosphate. J Biol Chem, 1998. 273(22): p. 13375-8. 29. Maehama, T. and J.E. Dixon, PTEN: a tumour suppressor that functions as a phospholipid phosphatase. Trends Cell Biol, 1999. 9(4): p. 125-8. 30. Tamura, M., et al., Inhibition of cell migration, spreading, and focal adhesions by tumor suppressor PTEN. Science, 1998. 280(5369): p. 1614-7. 31. Myers, M.P., et al., The lipid phosphatase activity of PTEN is critical for its tumor supressor function. Proc Natl Acad Sci U S A, 1998. 95(23): p. 13513-8. 32. Song, M.S., L. Salmena, and P.P. Pandolfi, The functions and regulation of the PTEN tumour suppressor. Nat Rev Mol Cell Biol, 2012. 13(5): p. 283-96. 33. Lee, Y.R., M. Chen, and P.P. Pandolfi, The functions and regulation of the PTEN tumour suppressor: new modes and prospects. Nat Rev Mol Cell Biol, 2018. 19(9): p. 547-562. 34. Suzuki, A., et al., High cancer susceptibility and embryonic lethality associated with mutation of the PTEN tumor suppressor gene in mice. Curr Biol, 1998. 8(21): p. 1169-78. 35. Fraser, M.M., et al., Pten loss causes hypertrophy and increased proliferation of astrocytes in vivo. Cancer Res, 2004. 64(21): p. 7773-9. 36. Li, L., F. Liu, and A.H. Ross, PTEN regulation of neural development and CNS stem cells. J Cell Biochem, 2003. 88(1): p. 24-8. 37. Groszer, M., et al., Negative regulation of neural stem/progenitor cell proliferation by the Pten tumor suppressor gene in vivo. Science, 2001. 294(5549): p. 2186-9. 38. Malaney, P., V.N. Uversky, and V. Dave, PTEN proteoforms in biology and disease. Cell Mol Life Sci, 2017. 74(15): p. 2783-2794. 39. Marsit, C.J., et al., PTEN expression in non-small-cell lung cancer: evaluating its relation to tumor characteristics, allelic loss, and epigenetic alteration. Hum Pathol, 2005. 36(7): p. 768-76. 40. Tang, J.M., et al., Phosphorylated Akt overexpression and loss of PTEN expression in non-small cell lung cancer confers poor prognosis. Lung Cancer, 2006. 51(2): p. 181-91. 41. Xiao, J., et al., PTEN expression is a prognostic marker for patients with non-small cell lung cancer: a systematic review and meta-analysis of the literature. Oncotarget, 2016. 7(36): p. 57832- 57840. 42. Yanagawa, N., et al., Loss of phosphatase and tensin homolog protein expression is an independent poor prognostic marker in lung adenocarcinoma. J Thorac Oncol, 2012. 7(10): p. 1513-21. 43. Yokomizo, A., et al., PTEN/MMAC1 mutations identified in small cell, but not in non-small cell lung cancers. Oncogene, 1998. 17(4): p. 475-9.

94 44. Malaney, P., Significance of PTEN Phosphorylation and its Nuclear Function in Lung Cancer. 2016, University of South Florida. 45. Jamaspishvili, T., et al., Clinical implications of PTEN loss in prostate cancer. Nat Rev Urol, 2018. 15(4): p. 222-234. 46. Jayakumar, R., S. Lanjewar, and C.A. Axiotis, Loss of PTEN and Increased pAKT Expression Distinguishes Aggressive Low-grade Neuroendocrine Tumors. Ann Clin Lab Sci, 2018. 48(5): p. 565-572. 47. Tachibana, M., et al., Expression and prognostic significance of PTEN product protein in patients with esophageal squamous cell carcinoma. Cancer, 2002. 94(7): p. 1955-60. 48. Kwabi-Addo, B., et al., Haploinsufficiency of the Pten tumor suppressor gene promotes prostate cancer progression. Proc Natl Acad Sci U S A, 2001. 98(20): p. 11563-8. 49. Trotman, L.C., et al., Pten dose dictates cancer progression in the prostate. PLoS Biol, 2003. 1(3): p. E59. 50. Berger, A.H., A.G. Knudson, and P.P. Pandolfi, A continuum model for tumour suppression. Nature, 2011. 476(7359): p. 163-9. 51. Naguib, A. and L.C. Trotman, PTEN plasticity: how the taming of a lethal gene can go too far. Trends Cell Biol, 2013. 23(8): p. 374-9. 52. Leslie, N.R. and M. Foti, Non-genomic loss of PTEN function in cancer: not in my genes. Trends Pharmacol Sci, 2011. 32(3): p. 131-40. 53. Heald, B., et al., Frequent gastrointestinal polyps and colorectal adenocarcinomas in a prospective series of PTEN mutation carriers. Gastroenterology, 2010. 139(6): p. 1927-33. 54. Butler, M.G., et al., Subset of individuals with autism spectrum disorders and extreme macrocephaly associated with germline PTEN tumour suppressor gene mutations. J Med Genet, 2005. 42(4): p. 318-21. 55. Mao, L., et al., Delayed administration of a PTEN inhibitor BPV improves functional recovery after experimental stroke. Neuroscience, 2013. 231: p. 272-81. 56. Garcia-Cao, I., et al., Systemic elevation of PTEN induces a tumor-suppressive metabolic state. Cell, 2012. 149(1): p. 49-62. 57. Malaney, P., V.N. Uversky, and V. Dave, Identification of intrinsically disordered regions in PTEN and delineation of its function via a network approach. Methods, 2015. 77-78: p. 69-74. 58. Walker, S.M., et al., The tumour-suppressor function of PTEN requires an N-terminal lipid-binding motif. Biochem J, 2004. 379(Pt 2): p. 301-7. 59. Funamoto, S., et al., Spatial and temporal regulation of 3-phosphoinositides by PI 3-kinase and PTEN mediates chemotaxis. Cell, 2002. 109(5): p. 611-23. 60. Iijima, M. and P. Devreotes, Tumor suppressor PTEN mediates sensing of chemoattractant gradients. Cell, 2002. 109(5): p. 599-610. 61. Iijima, M., et al., Novel mechanism of PTEN regulation by its phosphatidylinositol 4,5- bisphosphate binding motif is critical for chemotaxis. J Biol Chem, 2004. 279(16): p. 16606-13. 62. Campbell, R.B., F. Liu, and A.H. Ross, Allosteric activation of PTEN phosphatase by phosphatidylinositol 4,5-bisphosphate. J Biol Chem, 2003. 278(36): p. 33617-20. 63. McConnachie, G., et al., Interfacial kinetic analysis of the tumour suppressor phosphatase, PTEN: evidence for activation by anionic phospholipids. Biochem J, 2003. 371(Pt 3): p. 947-55. 64. Wei, Y., et al., Phospholipid-binding sites of phosphatase and tensin homolog (PTEN): exploring the mechanism of phosphatidylinositol 4,5-bisphosphate activation. J Biol Chem, 2015. 290(3): p. 1592-606. 65. Redfern, R.E., et al., PTEN phosphatase selectively binds phosphoinositides and undergoes structural changes. Biochemistry, 2008. 47(7): p. 2162-71. 66. Denning, G., et al., A short N-terminal sequence of PTEN controls cytoplasmic localization and is required for suppression of cell growth. Oncogene, 2007. 26(27): p. 3930-40.

95 67. Gil, A., et al., Nuclear localization of PTEN by a Ran-dependent mechanism enhances apoptosis: Involvement of an N-terminal nuclear localization domain and multiple nuclear exclusion motifs. Mol Biol Cell, 2006. 17(9): p. 4002-13. 68. Trotman, L.C., et al., Ubiquitination regulates PTEN nuclear import and tumor suppression. Cell, 2007. 128(1): p. 141-56. 69. Putz, U., et al., The tumor suppressor PTEN is exported in exosomes and has phosphatase activity in recipient cells. Sci Signal, 2012. 5(243): p. ra70. 70. Das, S., J.E. Dixon, and W. Cho, Membrane-binding and activation mechanism of PTEN. Proc Natl Acad Sci U S A, 2003. 100(13): p. 7491-6. 71. Shenoy, S.S., H. Nanda, and M. Losche, Membrane association of the PTEN tumor suppressor: electrostatic interaction with phosphatidylserine-containing bilayers and regulatory role of the C- terminal tail. J Struct Biol, 2012. 180(3): p. 394-408. 72. Raftopoulou, M., et al., Regulation of cell migration by the C2 domain of the tumor suppressor PTEN. Science, 2004. 303(5661): p. 1179-81. 73. Chung, J.H., M.E. Ginn-Pease, and C. Eng, Phosphatase and tensin homologue deleted on chromosome 10 (PTEN) has nuclear localization signal-like sequences for nuclear import mediated by major vault protein. Cancer Res, 2005. 65(10): p. 4108-16. 74. Yu, Z., et al., PTEN associates with the vault particles in HeLa cells. J Biol Chem, 2002. 277(43): p. 40247-52. 75. Georgescu, M.M., et al., The tumor-suppressor activity of PTEN is regulated by its carboxyl- terminal region. Proc Natl Acad Sci U S A, 1999. 96(18): p. 10182-7. 76. Wu, X., et al., Evidence for regulation of the PTEN tumor suppressor by a membrane-localized multi-PDZ domain containing scaffold protein MAGI-2. Proc Natl Acad Sci U S A, 2000. 97(8): p. 4233-8. 77. Wu, Y., et al., Interaction of the tumor suppressor PTEN/MMAC with a PDZ domain of MAGI3, a novel membrane-associated guanylate kinase. J Biol Chem, 2000. 275(28): p. 21477-85. 78. Gericke, A., M. Munson, and A.H. Ross, Regulation of the PTEN phosphatase. Gene, 2006. 374: p. 1-9. 79. Kotelevets, L., et al., Implication of the MAGI-1b/PTEN signalosome in stabilization of adherens junctions and suppression of invasiveness. Faseb j, 2005. 19(1): p. 115-7. 80. Mosessian, S. and H. Wu, PTEN-Associated Complexes: An Overview. Curr Top Biochem Res, 2010. 12(1): p. 37-42. 81. Sotelo, N.S., et al., PTEN-PDZ domain interactions: binding of PTEN to PDZ domains of PTPN13. Methods, 2015. 77-78: p. 147-56. 82. von Stein, W., et al., Direct association of Bazooka/PAR-3 with the lipid phosphatase PTEN reveals a link between the PAR/aPKC complex and phosphoinositide signaling. Development, 2005. 132(7): p. 1675-86. 83. Papa, A., et al., Cancer-associated PTEN mutants act in a dominant-negative manner to suppress PTEN protein function. Cell, 2014. 157(3): p. 595-610. 84. Heinrich, F., et al., The PTEN Tumor Suppressor Forms Homodimers in Solution. Structure, 2015. 23(10): p. 1952-1957. 85. Tzani, I., et al., Systematic analysis of the PTEN 5' leader identifies a major AUU initiated proteoform. Open Biol, 2016. 6(5). 86. Agrawal, S. and C. Eng, Differential expression of novel naturally occurring splice variants of PTEN and their functional consequences in Cowden syndrome and sporadic breast cancer. Hum Mol Genet, 2006. 15(5): p. 777-87. 87. Sharrard, R.M. and N.J. Maitland, Alternative splicing of the human PTEN/MMAC1/TEP1 gene. Biochim Biophys Acta, 2000. 1494(3): p. 282-5. 88. Hopkins, B.D., et al., A secreted PTEN phosphatase that enters cells to alter signaling and survival. Science, 2013. 341(6144): p. 399-402.

96 89. Liang, H., et al., PTENalpha, a PTEN isoform translated through alternative initiation, regulates mitochondrial function and energy metabolism. Cell Metab, 2014. 19(5): p. 836-48. 90. Breuksch, I., et al., In renal cell carcinoma the PTEN splice variant PTEN-Delta shows similar function as the tumor suppressor PTEN itself. Cell Commun Signal, 2018. 16(1): p. 35. 91. Sarquis, M.S., et al., Distinct expression profiles for PTEN transcript and its splice variants in Cowden syndrome and Bannayan-Riley-Ruvalcaba syndrome. Am J Hum Genet, 2006. 79(1): p. 23-30. 92. Jackson, M.D. and J.M. Denu, Molecular reactions of protein phosphatases--insights from structure and chemistry. Chem Rev, 2001. 101(8): p. 2313-40. 93. Zhang, Z.Y., Protein-tyrosine phosphatases: biological function, structural characteristics, and mechanism of catalysis. Crit Rev Biochem Mol Biol, 1998. 33(1): p. 1-52. 94. Xiao, Y., et al., PTEN catalysis of phospholipid dephosphorylation reaction follows a two-step mechanism in which the conserved aspartate-92 does not function as the general acid--mechanistic analysis of a familial Cowden disease-associated PTEN mutation. Cell Signal, 2007. 19(7): p. 1434-45. 95. Naguib, A., et al., PTEN functions by recruitment to cytoplasmic vesicles. Mol Cell, 2015. 58(2): p. 255-68. 96. Liu, P., et al., Targeting the phosphoinositide 3-kinase pathway in cancer. Nat Rev Drug Discov, 2009. 8(8): p. 627-44. 97. Hemmings, B.A. and D.F. Restuccia, PI3K-PKB/Akt pathway. Cold Spring Harb Perspect Biol, 2012. 4(9): p. a011189. 98. Guertin, D.A. and D.M. Sabatini, Defining the role of mTOR in cancer. Cancer Cell, 2007. 12(1): p. 9-22. 99. Vander Haar, E., et al., Insulin signalling to mTOR mediated by the Akt/PKB substrate PRAS40. Nat Cell Biol, 2007. 9(3): p. 316-23. 100. Manning, B.D. and L.C. Cantley, AKT/PKB signaling: navigating downstream. Cell, 2007. 129(7): p. 1261-74. 101. Jung, S.H., et al., mTOR kinase leads to PTEN-loss-induced cellular senescence by phosphorylating p53. 2019. 38(10): p. 1639-1650. 102. Radu, A., et al., PTEN induces cell cycle arrest by decreasing the level and nuclear localization of cyclin D1. Mol Cell Biol, 2003. 23(17): p. 6139-49. 103. Stambolic, V., et al., Negative regulation of PKB/Akt-dependent cell survival by the tumor suppressor PTEN. Cell, 1998. 95(1): p. 29-39. 104. Sun, H., et al., PTEN modulates cell cycle progression and cell survival by regulating phosphatidylinositol 3,4,5,-trisphosphate and Akt/protein kinase B signaling pathway. Proc Natl Acad Sci U S A, 1999. 96(11): p. 6199-204. 105. Liliental, J., et al., Genetic deletion of the Pten tumor suppressor gene promotes cell motility by activation of Rac1 and Cdc42 GTPases. Curr Biol, 2000. 10(7): p. 401-4. 106. Song, L.B., et al., The polycomb group protein Bmi-1 represses the tumor suppressor PTEN and induces epithelial-mesenchymal transition in human nasopharyngeal epithelial cells. J Clin Invest, 2009. 119(12): p. 3626-36. 107. Martin-Belmonte, F., et al., PTEN-mediated apical segregation of phosphoinositides controls epithelial morphogenesis through Cdc42. Cell, 2007. 128(2): p. 383-97. 108. Horie, Y., et al., Hepatocyte-specific Pten deficiency results in steatohepatitis and hepatocellular carcinomas. J Clin Invest, 2004. 113(12): p. 1774-83. 109. Palian, B.M., et al., Maf1 is a novel target of PTEN and PI3K signaling that negatively regulates oncogenesis and lipid metabolism. PLoS Genet, 2014. 10(12): p. e1004789. 110. Porstmann, T., et al., SREBP activity is regulated by mTORC1 and contributes to Akt-dependent cell growth. Cell Metab, 2008. 8(3): p. 224-36. 111. Eguez, L., et al., Full intracellular retention of GLUT4 requires AS160 Rab GTPase activating protein. Cell Metab, 2005. 2(4): p. 263-72.

97 112. Fang, M., et al., The ER UDPase ENTPD5 promotes protein N-glycosylation, the Warburg effect, and proliferation in the PTEN pathway. Cell, 2010. 143(5): p. 711-24. 113. Nakashima, N., et al., The tumor suppressor PTEN negatively regulates insulin signaling in 3T3- L1 adipocytes. J Biol Chem, 2000. 275(17): p. 12889-95. 114. Wong, J.T., et al., Pten (phosphatase and tensin homologue gene) haploinsufficiency promotes insulin hypersensitivity. Diabetologia, 2007. 50(2): p. 395-403. 115. Groszer, M., et al., PTEN negatively regulates neural stem cell self-renewal by modulating G0-G1 cell cycle entry. Proc Natl Acad Sci U S A, 2006. 103(1): p. 111-6. 116. He, X.C., et al., PTEN-deficient intestinal stem cells initiate intestinal polyposis. Nat Genet, 2007. 39(2): p. 189-98. 117. Hill, R. and H. Wu, PTEN, stem cells, and cancer stem cells. J Biol Chem, 2009. 284(18): p. 11755- 9. 118. Yilmaz, O.H., et al., Pten dependence distinguishes haematopoietic stem cells from leukaemia- initiating cells. Nature, 2006. 441(7092): p. 475-82. 119. Davé, V., et al., Conditional deletion of Pten causes bronchiolar hyperplasia. Am J Respir Cell Mol Biol, 2008. 38(3): p. 337-45. 120. Suzuki, A., et al., Keratinocyte-specific Pten deficiency results in epidermal hyperplasia, accelerated hair follicle morphogenesis and tumor formation. Cancer Res, 2003. 63(3): p. 674-81. 121. Wang, S., et al., Pten deletion leads to the expansion of a prostatic stem/progenitor cell subpopulation and tumor initiation. Proc Natl Acad Sci U S A, 2006. 103(5): p. 1480-5. 122. Song, H., et al., Functional characterization of pulmonary neuroendocrine cells in lung development, injury, and tumorigenesis. Proc Natl Acad Sci U S A, 2012. 109(43): p. 17531-6. 123. Tiozzo, C., et al., Deletion of Pten expands lung epithelial progenitor pools and confers resistance to airway injury. Am J Respir Crit Care Med, 2009. 180(8): p. 701-12. 124. Yanagi, S., et al., Pten controls lung morphogenesis, bronchioalveolar stem cells, and onset of lung adenocarcinomas in mice. J Clin Invest, 2007. 117(10): p. 2929-40. 125. Haier, J. and G.L. Nicolson, PTEN regulates tumor cell adhesion of colon carcinoma cells under dynamic conditions of fluid flow. Oncogene, 2002. 21(9): p. 1450-60. 126. Tamura, M., et al., PTEN interactions with focal adhesion kinase and suppression of the extracellular matrix-dependent phosphatidylinositol 3-kinase/Akt cell survival pathway. J Biol Chem, 1999. 274(29): p. 20693-703. 127. Gu, J., et al., Shc and FAK differentially regulate cell motility and directionality modulated by PTEN. J Cell Biol, 1999. 146(2): p. 389-403. 128. Gu, J., M. Tamura, and K.M. Yamada, Tumor suppressor PTEN inhibits integrin- and growth factor-mediated mitogen-activated protein (MAP) kinase signaling pathways. J Cell Biol, 1998. 143(5): p. 1375-83. 129. Schneider, E., et al., Migration of renal tumor cells depends on dephosphorylation of Shc by PTEN. Int J Oncol, 2011. 38(3): p. 823-31. 130. Tibarewal, P., et al., PTEN protein phosphatase activity correlates with control of gene expression and invasion, a tumor-suppressing phenotype, but not with AKT activity. Sci Signal, 2012. 5(213): p. ra18. 131. Tibarewal, P., et al., PTEN Protein Phosphatase Activity Correlates with Control of Gene Expression and Invasion, a Tumor-Suppressing Phenotype, But Not with AKT Activity. Science Signaling, 2012. 5(213): p. ra18-ra18. 132. Wozniak, D.J., et al., PTEN is a protein phosphatase that targets active PTK6 and inhibits PTK6 oncogenic signaling in prostate cancer. Nat Commun, 2017. 8(1): p. 1508. 133. Zhang, X.C., et al., Functional analysis of the protein phosphatase activity of PTEN. Biochem J, 2012. 444(3): p. 457-64. 134. González-Sánchez, A., et al., Connexin43 recruits PTEN and Csk to inhibit c-Src activity in glioma cells and astrocytes. Oncotarget, 2016. 7(31): p. 49819-49833. 135. Shnitsar, I., et al., PTEN regulates cilia through Dishevelled. Nat Commun, 2015. 6: p. 8388.

98 136. He, M., et al., KLF4 mediates the link between TGF-beta1-induced gene transcription and H3 acetylation in vascular smooth muscle cells. Faseb j, 2015. 29(9): p. 4059-70. 137. Lin, M.C., et al., PTEN interacts with metal-responsive transcription factor 1 and stimulates its transcriptional activity. Biochem J, 2012. 441(1): p. 367-77. 138. Weng, L.P., et al., PTEN blocks insulin-mediated ETS-2 phosphorylation through MAP kinase, independently of the phosphoinositide 3-kinase pathway. Hum Mol Genet, 2002. 11(15): p. 1687- 96. 139. Bronisz, A., et al., Reprogramming of the tumour microenvironment by stromal PTEN-regulated miR-320. Nat Cell Biol, 2012. 14(2): p. 159-67. 140. Trimboli, A.J., et al., Pten in stromal fibroblasts suppresses mammary epithelial tumours. Nature, 2009. 461(7267): p. 1084-91. 141. Bononi, A., et al., Identification of PTEN at the ER and MAMs and its regulation of Ca(2+) signaling and apoptosis in a protein phosphatase-dependent manner. Cell Death Differ, 2013. 20(12): p. 1631-43. 142. Shi, Y., et al., PTEN is a protein tyrosine phosphatase for IRS1. Nat Struct Mol Biol, 2014. 21(6): p. 522-7. 143. Shinde, S.R. and S. Maddika, PTEN modulates EGFR late endocytic trafficking and degradation by dephosphorylating Rab7. Nat Commun, 2016. 7: p. 10689. 144. Chen, M., et al., The nuclear transport receptor Importin-11 is a tumor suppressor that maintains PTEN protein. J Cell Biol, 2017. 216(3): p. 641-656. 145. Liu, F., et al., PTEN enters the nucleus by diffusion. J Cell Biochem, 2005. 96(2): p. 221-34. 146. Vazquez, F., et al., Phosphorylation of the PTEN tail acts as an inhibitory switch by preventing its recruitment into a protein complex. J Biol Chem, 2001. 276(52): p. 48627-30. 147. Vazquez, F., et al., Phosphorylation of the PTEN tail regulates protein stability and function. Mol Cell Biol, 2000. 20(14): p. 5010-8. 148. Wu, Y., et al., PTEN phosphorylation and nuclear export mediate free fatty acid-induced oxidative stress. Antioxid Redox Signal, 2014. 20(9): p. 1382-95. 149. Planchon, S.M., K.A. Waite, and C. Eng, The nuclear affairs of PTEN. J Cell Sci, 2008. 121(Pt 3): p. 249-53. 150. Whiteman, D.C., et al., Nuclear PTEN expression and clinicopathologic features in a population- based series of primary cutaneous melanoma. Int J Cancer, 2002. 99(1): p. 63-7. 151. Beckham, T.H., et al., Acid ceramidase promotes nuclear export of PTEN through sphingosine 1- phosphate mediated Akt signaling. PLoS One, 2013. 8(10): p. e76593. 152. Lindsay, Y., et al., Localization of agonist-sensitive PtdIns(3,4,5)P3 reveals a nuclear pool that is insensitive to PTEN expression. J Cell Sci, 2006. 119(Pt 24): p. 5160-8. 153. Liu, J.L., et al., Nuclear PTEN-mediated growth suppression is independent of Akt downregulation. Mol Cell Biol, 2005. 25(14): p. 6211-24. 154. Chung, J.H., et al., The ERK1/2 pathway modulates nuclear PTEN-mediated cell cycle arrest by cyclin D1 transcriptional regulation. Hum Mol Genet, 2006. 15(17): p. 2553-9. 155. Feng, J., et al., PTEN Controls the DNA Replication Process through MCM2 in Response to Replicative Stress. Cell Rep, 2015. 13(7): p. 1295-1303. 156. Gu, T., et al., CREB is a novel nuclear target of PTEN phosphatase. Cancer Res, 2011. 71(8): p. 2821-5. 157. Lyu, J., et al., The protein phosphatase activity of PTEN is essential for regulating neural stem cell differentiation. Mol Brain, 2015. 8: p. 26. 158. Sun, Z., et al., PTEN C-terminal deletion causes genomic instability and tumor development. Cell Rep, 2014. 6(5): p. 844-54. 159. Song, M.S., et al., Nuclear PTEN regulates the APC-CDH1 tumor-suppressive complex in a phosphatase-independent manner. Cell, 2011. 144(2): p. 187-99. 160. Shen, W.H., et al., Essential role for nuclear PTEN in maintaining chromosomal integrity. Cell, 2007. 128(1): p. 157-70.

99 161. Chen, J.H., et al., ATM-mediated PTEN phosphorylation promotes PTEN nuclear translocation and autophagy in response to DNA-damaging agents in cancer cells. Autophagy, 2015. 11(2): p. 239-52. 162. Choi, B.H., Y. Chen, and W. Dai, Chromatin PTEN is involved in DNA damage response partly through regulating Rad52 sumoylation. Cell Cycle, 2013. 12(21): p. 3442-7. 163. Fan, C., et al., PTEN inhibits BMI1 function independently of its phosphatase activity. Mol Cancer, 2009. 8: p. 98. 164. Freeman, D.J., et al., PTEN tumor suppressor regulates p53 protein levels and activity through phosphatase-dependent and -independent mechanisms. Cancer Cell, 2003. 3(2): p. 117-30. 165. Horita, H., et al., Nuclear PTEN functions as an essential regulator of SRF-dependent transcription to control smooth muscle differentiation. Nat Commun, 2016. 7: p. 10830. 166. Li, A.G., et al., Mechanistic insights into maintenance of high p53 acetylation by PTEN. Mol Cell, 2006. 23(4): p. 575-87. 167. Yang, C., et al., PTEN suppresses the oncogenic function of AIB1 through decreasing its protein stability via mechanism involving Fbw7 alpha. Mol Cancer, 2013. 12: p. 21. 168. Steinbach, N., et al., PTEN interacts with the transcription machinery on chromatin and regulates RNA polymerase II-mediated transcription. Nucleic Acids Res, 2019. 169. Chen, Z.H., et al., PTEN interacts with histone H1 and controls chromatin condensation. Cell Rep, 2014. 8(6): p. 2003-2014. 170. Gong, L., et al., Nuclear PTEN tumor-suppressor functions through maintaining heterochromatin structure. Cell Cycle, 2015. 14(14): p. 2323-32. 171. Sumitomo, M., et al., Synergy in tumor suppression by direct interaction of neutral endopeptidase with PTEN. Cancer Cell, 2004. 5(1): p. 67-78. 172. Takahashi, Y., et al., PTEN tumor suppressor associates with NHERF proteins to attenuate PDGF receptor signaling. Embo j, 2006. 25(4): p. 910-20. 173. Okumura, K., et al., Cellular transformation by the MSP58 oncogene is inhibited by its physical interaction with the PTEN tumor suppressor. Proc Natl Acad Sci U S A, 2005. 102(8): p. 2703-6. 174. Conde-Perez, A., et al., A caveolin-dependent and PI3K/AKT-independent role of PTEN in beta- catenin transcriptional activity. Nat Commun, 2015. 6: p. 8093. 175. Eng, C., PTEN: one gene, many syndromes. Hum Mutat, 2003. 22(3): p. 183-98. 176. Hollander, M.C., G.M. Blumenthal, and P.A. Dennis, PTEN loss in the continuum of common cancers, rare syndromes and mouse models. Nat Rev Cancer, 2011. 11(4): p. 289-301. 177. Smith, I.N. and J.M. Briggs, Structural mutation analysis of PTEN and its genotype-phenotype correlations in endometriosis and cancer. Proteins, 2016. 84(11): p. 1625-1643. 178. Bonneau, D. and M. Longy, Mutations of the human PTEN gene. Hum Mutat, 2000. 16(2): p. 109- 22. 179. Malaney, P., et al., Intrinsic disorder in PTEN and its interactome confers structural plasticity and functional versatility. Sci Rep, 2013. 3: p. 2035. 180. Rodriguez-Escudero, I., et al., A comprehensive functional analysis of PTEN mutations: implications in tumor- and autism-related syndromes. Hum Mol Genet, 2011. 20(21): p. 4132-42. 181. Agrawal, S., R. Pilarski, and C. Eng, Different splicing defects lead to differential effects downstream of the lipid and protein phosphatase activities of PTEN. Hum Mol Genet, 2005. 14(16): p. 2459-68. 182. Celebi, J.T., et al., Association of splicing defects in PTEN leading to exon skipping or partial intron retention in Cowden syndrome. Hum Genet, 2000. 107(3): p. 234-8. 183. Chen, H.J., et al., Characterization of cryptic splicing in germline PTEN intronic variants in Cowden syndrome. 2017. 38(10): p. 1372-1377. 184. Alvarez-Nunez, F., et al., PTEN promoter methylation in sporadic thyroid carcinomas. Thyroid, 2006. 16(1): p. 17-23. 185. Garcia, J.M., et al., Promoter methylation of the PTEN gene is a common molecular change in breast cancer. Genes Chromosomes Cancer, 2004. 41(2): p. 117-24.

100 186. Soria, J.C., et al., Lack of PTEN expression in non-small cell lung cancer could be related to promoter methylation. Clin Cancer Res, 2002. 8(5): p. 1178-84. 187. Roh, M.R., et al., Promoter Methylation of PTEN Is a Significant Prognostic Factor in Melanoma Survival. J Invest Dermatol, 2016. 136(5): p. 1002-1011. 188. Tang, B., et al., JARID1B promotes metastasis and epithelial-mesenchymal transition via PTEN/AKT signaling in hepatocellular carcinoma cells. Oncotarget, 2015. 6(14): p. 12723-39. 189. Lu, J., et al., Stem cell factor SALL4 represses the transcriptions of PTEN and SALL1 through an epigenetic repressor complex. PLoS One, 2009. 4(5): p. e5577. 190. Hettinger, K., et al., c-Jun promotes cellular survival by suppression of PTEN. Cell Death Differ, 2007. 14(2): p. 218-29. 191. Li, Y., et al., MAF1 suppresses AKT-mTOR signaling and liver cancer through activation of PTEN transcription. Hepatology, 2016. 63(6): p. 1928-42. 192. Moorehead, R.A., et al., Insulin-like growth factor-II regulates PTEN expression in the mammary gland. J Biol Chem, 2003. 278(50): p. 50422-7. 193. Patel, L., et al., Tumor suppressor and anti-inflammatory actions of PPARgamma agonists are mediated via upregulation of PTEN. Curr Biol, 2001. 11(10): p. 764-8. 194. Shen, Y.H., et al., Up-regulation of PTEN (phosphatase and tensin homolog deleted on chromosome ten) mediates p38 MAPK stress signal-induced inhibition of insulin signaling. A cross-talk between stress signaling and insulin signaling in resistin-treated human endothelial cells. J Biol Chem, 2006. 281(12): p. 7727-36. 195. Stambolic, V., et al., Regulation of PTEN transcription by p53. Mol Cell, 2001. 8(2): p. 317-25. 196. Virolle, T., et al., The Egr-1 transcription factor directly activates PTEN during irradiation- induced signalling. Nat Cell Biol, 2001. 3(12): p. 1124-8. 197. Whelan, J.T., S.L. Forbes, and F.E. Bertrand, CBF-1 (RBP-J kappa) binds to the PTEN promoter and regulates PTEN gene expression. Cell Cycle, 2007. 6(1): p. 80-4. 198. Pei, Y.F., et al., TET1 inhibits gastric cancer growth and metastasis by PTEN demethylation and re-expression. Oncotarget, 2016. 7(21): p. 31322-35. 199. Fagone, E., et al., Resveratrol inhibits transforming growth factor-beta-induced proliferation and differentiation of ex vivo human lung fibroblasts into myofibroblasts through ERK/Akt inhibition and PTEN restoration. Exp Lung Res, 2011. 37(3): p. 162-74. 200. Tamguney, T. and D. Stokoe, New insights into PTEN. J Cell Sci, 2007. 120(Pt 23): p. 4071-9. 201. Chou, H.T., et al., HES-Mediated Repression of Pten in Caenorhabditis elegans. G3 (Bethesda), 2015. 5(12): p. 2619-28. 202. Chow, J.Y., et al., RAS/ERK modulates TGFbeta-regulated PTEN expression in human pancreatic adenocarcinoma cells. Carcinogenesis, 2007. 28(11): p. 2321-7. 203. Escriva, M., et al., Repression of PTEN phosphatase by Snail1 transcriptional factor during gamma radiation-induced apoptosis. Mol Cell Biol, 2008. 28(5): p. 1528-40. 204. Lee, J.Y., et al., Id-1 activates Akt-mediated Wnt signaling and p27(Kip1) phosphorylation through PTEN inhibition. Oncogene, 2009. 28(6): p. 824-31. 205. Lin, Y., T. Kang, and B.P. Zhou, Doxorubicin enhances Snail/LSD1-mediated PTEN suppression in a PARP1-dependent manner. Cell Cycle, 2014. 13(11): p. 1708-16. 206. Noguera, N.I., et al., PML/RARa inhibits PTEN expression in hematopoietic cells by competing with PU.1 transcriptional activity. Oncotarget, 2016. 7(41): p. 66386-66397. 207. Tang, Y.A., et al., Global Oct4 target gene analysis reveals novel downstream PTEN and TNC genes required for drug-resistance and metastasis in lung cancer. Nucleic Acids Res, 2015. 43(3): p. 1593-608. 208. Uygur, B., et al., SLUG is a direct transcriptional repressor of PTEN tumor suppressor. Prostate, 2015. 75(9): p. 907-16. 209. Wang, S., et al., SOX2, a predictor of survival in gastric cancer, inhibits cell proliferation and metastasis by regulating PTEN. Cancer Lett, 2015. 358(2): p. 210-219.

101 210. Xia, D., et al., Mitogen-activated protein kinase kinase-4 promotes cell survival by decreasing PTEN expression through an NF kappa B-dependent pathway. J Biol Chem, 2007. 282(6): p. 3507- 19. 211. Yoshimi, A., et al., Evi1 represses PTEN expression and activates PI3K/AKT/mTOR via interactions with polycomb proteins. Blood, 2011. 117(13): p. 3617-28. 212. Zhang, H., et al., 17betaE2 promotes cell proliferation in endometriosis by decreasing PTEN via NFkappaB-dependent pathway. Mol Cell Endocrinol, 2010. 317(1-2): p. 31-43. 213. Lee, Y.R., et al., TNF-alpha upregulates PTEN via NF-kappaB signaling pathways in human leukemic cells. Exp Mol Med, 2007. 39(1): p. 121-7. 214. Vasudevan, K.M., S. Gurumurthy, and V.M. Rangnekar, Suppression of PTEN expression by NF- kappa B prevents apoptosis. Mol Cell Biol, 2004. 24(3): p. 1007-21. 215. Bermudez Brito, M., E. Goulielmaki, and E.A. Papakonstanti, Focus on PTEN Regulation. Front Oncol, 2015. 5: p. 166. 216. Tian, L., et al., Four microRNAs promote prostate cell proliferation with regulation of PTEN and its downstream signals in vitro. PLoS One, 2013. 8(9): p. e75885. 217. Liu, Z.L., et al., MicroRNA-21 (miR-21) expression promotes growth, metastasis, and chemo- or radioresistance in non-small cell lung cancer cells by targeting PTEN. Mol Cell Biochem, 2013. 372(1-2): p. 35-45. 218. Meng, F., et al., MicroRNA-21 regulates expression of the PTEN tumor suppressor gene in human hepatocellular cancer. Gastroenterology, 2007. 133(2): p. 647-58. 219. Wang, F., et al., MicroRNA-19a/b regulates multidrug resistance in human gastric cancer cells by targeting PTEN. Biochem Biophys Res Commun, 2013. 434(3): p. 688-94. 220. Zhang, B.G., et al., microRNA-21 promotes tumor proliferation and invasion in gastric cancer by targeting PTEN. Oncol Rep, 2012. 27(4): p. 1019-26. 221. Chen, P., et al., MiR-200c is a cMyc-activated miRNA that promotes nasopharyngeal carcinoma by downregulating PTEN. Oncotarget, 2017. 8(3): p. 5206-5218. 222. Mu, P., et al., Genetic dissection of the miR-17~92 cluster of microRNAs in Myc-induced B-cell lymphomas. Genes Dev, 2009. 23(24): p. 2806-11. 223. Poliseno, L., et al., A coding-independent function of gene and pseudogene mRNAs regulates tumour biology. Nature, 2010. 465(7301): p. 1033-8. 224. Johnsson, P., et al., A pseudogene long-noncoding-RNA network regulates PTEN transcription and translation in human cells. Nat Struct Mol Biol, 2013. 20(4): p. 440-6. 225. Connor, K.M., et al., Mitochondrial H2O2 regulates the angiogenic phenotype via PTEN oxidation. J Biol Chem, 2005. 280(17): p. 16916-24. 226. Han, S.J., et al., Redox regulation of the tumor suppressor PTEN by the thioredoxin system and cumene hydroperoxide. Free Radic Biol Med, 2017. 112: p. 277-286. 227. Kim, Y., et al., Redox regulation of the tumor suppressor PTEN by glutathione. FEBS Lett, 2010. 584(16): p. 3550-6. 228. Kwon, J., et al., Reversible oxidation and inactivation of the tumor suppressor PTEN in cells stimulated with peptide growth factors. Proc Natl Acad Sci U S A, 2004. 101(47): p. 16419-24. 229. Lee, C.U., et al., Redox Modulation of PTEN Phosphatase Activity by Hydrogen Peroxide and Bisperoxidovanadium Complexes. Angew Chem Int Ed Engl, 2015. 54(46): p. 13796-800. 230. Lee, S.R., et al., Reversible inactivation of the tumor suppressor PTEN by H2O2. J Biol Chem, 2002. 277(23): p. 20336-42. 231. Leslie, N.R., et al., Redox regulation of PI 3-kinase signalling via inactivation of PTEN. Embo j, 2003. 22(20): p. 5501-10. 232. Meuillet, E.J., et al., Thioredoxin-1 binds to the C2 domain of PTEN inhibiting PTEN's lipid phosphatase activity and membrane binding: a mechanism for the functional loss of PTEN's tumor suppressor activity. Arch Biochem Biophys, 2004. 429(2): p. 123-33. 233. Schwertassek, U., et al., Reactivation of oxidized PTP1B and PTEN by thioredoxin 1. Febs j, 2014. 281(16): p. 3545-58.

102 234. Cao, J., et al., Prdx1 inhibits tumorigenesis via regulating PTEN/AKT activity. Embo j, 2009. 28(10): p. 1505-17. 235. Covey, T.M., K. Edes, and F.A. Fitzpatrick, Akt activation by arachidonic acid metabolism occurs via oxidation and inactivation of PTEN tumor suppressor. Oncogene, 2007. 26(39): p. 5784-92. 236. Gupta, A. and N.R. Leslie, Controlling PTEN (Phosphatase and Tensin Homolog) Stability: A DOMINANT ROLE FOR LYSINE 66. 2016. 291(35): p. 18465-73. 237. Kwak, Y.D., et al., NO signaling and S-nitrosylation regulate PTEN inhibition in neurodegeneration. Mol Neurodegener, 2010. 5: p. 49. 238. Lim, S. and M.V. Clement, Phosphorylation of the survival kinase Akt by superoxide is dependent on an ascorbate-reversible oxidation of PTEN. Free Radic Biol Med, 2007. 42(8): p. 1178-92. 239. Numajiri, N., et al., On-off system for PI3-kinase-Akt signaling through S-nitrosylation of phosphatase with sequence homology to tensin (PTEN). Proc Natl Acad Sci U S A, 2011. 108(25): p. 10349-54. 240. Zhu, L., et al., NOS1 S-nitrosylates PTEN and inhibits autophagy in nasopharyngeal carcinoma cells. Cell Death Discov, 2017. 3: p. 17011. 241. Gupta, A., et al., PARK2 Depletion Connects Energy and Oxidative Stress to PI3K/Akt Activation via PTEN S-Nitrosylation. Mol Cell, 2017. 65(6): p. 999-1013.e7. 242. Dhar, S., et al., Resveratrol regulates PTEN/Akt pathway through inhibition of MTA1/HDAC unit of the NuRD complex in prostate cancer. Biochim Biophys Acta, 2015. 1853(2): p. 265-75. 243. Okumura, K., et al., PCAF modulates PTEN activity. J Biol Chem, 2006. 281(36): p. 26562-8. 244. Meng, Z., L.F. Jia, and Y.H. Gan, PTEN activation through K163 acetylation by inhibiting HDAC6 contributes to tumour inhibition. Oncogene, 2016. 35(18): p. 2333-44. 245. Ikenoue, T., et al., PTEN acetylation modulates its interaction with PDZ domain. Cancer Res, 2008. 68(17): p. 6908-12. 246. Chae, H.D. and H.E. Broxmeyer, SIRT1 deficiency downregulates PTEN/JNK/FOXO1 pathway to block reactive oxygen species-induced apoptosis in mouse embryonic stem cells. Stem Cells Dev, 2011. 20(7): p. 1277-85. 247. Nakakido, M., et al., Dysregulation of AKT Pathway by SMYD2-Mediated Lysine Methylation on PTEN. Neoplasia, 2015. 17(4): p. 367-73. 248. Feng, J., et al., PTEN arginine methylation by PRMT6 suppresses PI3K-AKT signaling and modulates pre-mRNA splicing. Proc Natl Acad Sci U S A, 2019. 116(14): p. 6868-6877. 249. Gonzalez-Santamaria, J., et al., Regulation of the tumor suppressor PTEN by SUMO. Cell Death Dis, 2012. 3: p. e393. 250. Huang, J., et al., SUMO1 modification of PTEN regulates tumorigenesis by controlling its association with the plasma membrane. Nat Commun, 2012. 3: p. 911. 251. Bassi, C., et al., Nuclear PTEN controls DNA repair and sensitivity to genotoxic stress. Science, 2013. 341(6144): p. 395-9. 252. Ahmed, S.F., et al., The chaperone-assisted E3 ligase C terminus of Hsc70-interacting protein (CHIP) targets PTEN for proteasomal degradation. J Biol Chem, 2012. 287(19): p. 15996-6006. 253. Maddika, S., et al., WWP2 is an E3 ubiquitin ligase for PTEN. Nat Cell Biol, 2011. 13(6): p. 728- 33. 254. Van Themsche, C., et al., X-linked inhibitor of apoptosis protein (XIAP) regulates PTEN ubiquitination, content, and compartmentalization. J Biol Chem, 2009. 284(31): p. 20462-6. 255. Wang, X., et al., NEDD4-1 is a proto-oncogenic ubiquitin ligase for PTEN. Cell, 2007. 128(1): p. 129-39. 256. Fouladkou, F., et al., The ubiquitin ligase Nedd4-1 is dispensable for the regulation of PTEN stability and localization. Proc Natl Acad Sci U S A, 2008. 105(25): p. 8585-90. 257. Amodio, N., et al., Oncogenic role of the E3 ubiquitin ligase NEDD4-1, a PTEN negative regulator, in non-small-cell lung carcinomas. Am J Pathol, 2010. 177(5): p. 2622-34. 258. Lee, J.T., et al., RFP-mediated ubiquitination of PTEN modulates its effect on AKT activation. Cell Res, 2013. 23(4): p. 552-64.

103 259. Li, N., et al., Poly-ADP ribosylation of PTEN by tankyrases promotes PTEN degradation and tumor growth. Genes Dev, 2015. 29(2): p. 157-70. 260. Mund, T. and H.R. Pelham, Regulation of PTEN/Akt and MAP kinase signaling pathways by the ubiquitin ligase activators Ndfip1 and Ndfip2. Proc Natl Acad Sci U S A, 2010. 107(25): p. 11429- 34. 261. Song, M.S., et al., The deubiquitinylation and localization of PTEN are regulated by a HAUSP- PML network. Nature, 2008. 455(7214): p. 813-7. 262. Yuan, L., et al., Deubiquitylase OTUD3 regulates PTEN stability and suppresses tumorigenesis. Nat Cell Biol, 2015. 17(9): p. 1169-81. 263. Zhang, J., et al., Deubiquitylation and stabilization of PTEN by USP13. Nat Cell Biol, 2013. 15(12): p. 1486-1494. 264. Bolduc, D., et al., Phosphorylation-mediated PTEN conformational closure and deactivation revealed with protein semisynthesis. Elife, 2013. 2: p. e00691. 265. Odriozola, L., et al., Regulation of PTEN activity by its carboxyl-terminal autoinhibitory domain. J Biol Chem, 2007. 282(32): p. 23306-15. 266. Rahdar, M., et al., A phosphorylation-dependent intramolecular interaction regulates the membrane association and activity of the tumor suppressor PTEN. Proc Natl Acad Sci U S A, 2009. 106(2): p. 480-5. 267. Torres, J. and R. Pulido, The tumor suppressor PTEN is phosphorylated by the protein kinase CK2 at its C terminus. Implications for PTEN stability to proteasome-mediated degradation. J Biol Chem, 2001. 276(2): p. 993-8. 268. Chen, Z., et al., Molecular Features of Phosphatase and Tensin Homolog (PTEN) Regulation by C-terminal Phosphorylation. J Biol Chem, 2016. 291(27): p. 14160-9. 269. Torres, J., et al., Phosphorylation-regulated cleavage of the tumor suppressor PTEN by caspase- 3: implications for the control of protein stability and PTEN-protein interactions. J Biol Chem, 2003. 278(33): p. 30652-60. 270. Tolkacheva, T., et al., Regulation of PTEN binding to MAGI-2 by two putative phosphorylation sites at threonine 382 and 383. Cancer Res, 2001. 61(13): p. 4985-9. 271. Choi, B.H., M. Pagano, and W. Dai, Plk1 protein phosphorylates phosphatase and tensin homolog (PTEN) and regulates its mitotic activity during the cell cycle. J Biol Chem, 2014. 289(20): p. 14066-74. 272. Okahara, F., et al., Regulation of PTEN phosphorylation and stability by a tumor suppressor candidate protein. J Biol Chem, 2004. 279(44): p. 45300-3. 273. Li, Z., et al., Plk1 phosphorylation of PTEN causes a tumor-promoting metabolic state. Mol Cell Biol, 2014. 34(19): p. 3642-61. 274. Nakahata, S., et al., Loss of NDRG2 expression activates PI3K-AKT signalling via PTEN phosphorylation in ATLL and other cancers. Nat Commun, 2014. 5: p. 3393. 275. Al-Khouri, A.M., et al., Cooperative phosphorylation of the tumor suppressor phosphatase and tensin homologue (PTEN) by casein kinases and glycogen synthase kinase 3beta. J Biol Chem, 2005. 280(42): p. 35195-202. 276. Maccario, H., et al., PTEN is destabilized by phosphorylation on Thr366. Biochem J, 2007. 405(3): p. 439-44. 277. Miller, S.J., et al., Direct identification of PTEN phosphorylation sites. FEBS Lett, 2002. 528(1-3): p. 145-53. 278. Xu, D., et al., Regulation of PTEN stability and activity by Plk3. J Biol Chem, 2010. 285(51): p. 39935-42. 279. Yim, E.K., et al., Rak functions as a tumor suppressor by regulating PTEN protein stability and function. Cancer Cell, 2009. 15(4): p. 304-14. 280. Liu, J., et al., Insulin activates the insulin receptor to downregulate the PTEN tumour suppressor. Oncogene, 2014. 33(29): p. 3878-85. 281. Li, Z., et al., Regulation of PTEN by Rho small GTPases. Nat Cell Biol, 2005. 7(4): p. 399-404.

104 282. Ferrarelli, L.K., FGFR inhibitors get to nuclear PTEN. Science Signaling, 2019. 12(577): p. eaax6492. 283. Lu, Y., et al., Src family protein-tyrosine kinases alter the function of PTEN to regulate phosphatidylinositol 3-kinase/AKT cascades. J Biol Chem, 2003. 278(41): p. 40057-66. 284. Fenton, T.R., et al., Resistance to EGF receptor inhibitors in glioblastoma mediated by phosphorylation of the PTEN tumor suppressor at tyrosine 240. Proc Natl Acad Sci U S A, 2012. 109(35): p. 14164-9. 285. Cheong, J.W., et al., Phosphatase and tensin homologue phosphorylation in the C-terminal regulatory domain is frequently observed in acute myeloid leukaemia and associated with poor clinical outcome. Br J Haematol, 2003. 122(3): p. 454-6. 286. Malaney, P., V.N. Uversky, and V. Dave, The PTEN Long N-tail is intrinsically disordered: increased viability for PTEN therapy. Mol Biosyst, 2013. 9(11): p. 2877-88. 287. McCubrey, J.A., et al., Ras/Raf/MEK/ERK and PI3K/PTEN/Akt/mTOR cascade inhibitors: how mutations can result in therapy resistance and how to overcome resistance. Oncotarget, 2012. 3(10): p. 1068-111. 288. Anai, S., et al., Combination of PTEN gene therapy and radiation inhibits the growth of human prostate cancer xenografts. Hum Gene Ther, 2006. 17(10): p. 975-84. 289. Chen, H., et al., PTEN restoration and PIK3CB knockdown synergistically suppress glioblastoma growth in vitro and in xenografts. J Neurooncol, 2011. 104(1): p. 155-67. 290. Cheney, I.W., et al., Suppression of tumorigenicity of glioblastoma cells by adenovirus-mediated MMAC1/PTEN gene transfer. Cancer Res, 1998. 58(11): p. 2331-4. 291. Li, D., et al., Enhanced tumor suppression by adenoviral PTEN gene therapy combined with cisplatin chemotherapy in small-cell lung cancer. Cancer Gene Ther, 2013. 20(4): p. 251-9. 292. Saito, Y., et al., Adenovirus-mediated transfer of the PTEN gene inhibits human colorectal cancer growth in vitro and in vivo. Gene Ther, 2003. 10(23): p. 1961-9. 293. She, Q.B., et al., Resistance to gefitinib in PTEN-null HER-overexpressing tumor cells can be overcome through restoration of PTEN function or pharmacologic modulation of constitutive phosphatidylinositol 3'-kinase/Akt pathway signaling. Clin Cancer Res, 2003. 9(12): p. 4340-6. 294. Tanaka, M. and H.B. Grossman, In vivo gene therapy of human bladder cancer with PTEN suppresses tumor growth, downregulates phosphorylated Akt, and increases sensitivity to doxorubicin. Gene Ther, 2003. 10(19): p. 1636-42. 295. Tanaka, M., C.J. Rosser, and H.B. Grossman, PTEN gene therapy induces growth inhibition and increases efficacy of chemotherapy in prostate cancer. Cancer Detect Prev, 2005. 29(2): p. 170-4. 296. Islam, M.A., et al., Restoration of tumour-growth suppression in vivo via systemic nanoparticle- mediated delivery of PTEN mRNA. Nat Biomed Eng, 2018. 2(11): p. 850-864. 297. Hopkins, B.D., et al., PTEN function: the long and the short of it. Trends Biochem Sci, 2014. 39(4): p. 183-90. 298. Wang, H., et al., Relevance and therapeutic possibility of PTEN-long in renal cell carcinoma. PLoS One, 2015. 10(2): p. e114250. 299. Lavictoire, S.J., et al., Engineering PTEN-L for Cell-Mediated Delivery. Mol Ther Methods Clin Dev, 2018. 9: p. 12-22. 300. Waite, K.A., M.R. Sinden, and C. Eng, Phytoestrogen exposure elevates PTEN levels. Hum Mol Genet, 2005. 14(11): p. 1457-63. 301. Hopkins, B.D. and R.E. Parsons, Molecular pathways: intercellular PTEN and the potential of PTEN restoration therapy. Clin Cancer Res, 2014. 20(21): p. 5379-83. 302. Lee, Y.-R., et al., Reactivation of PTEN tumor suppressor for cancer treatment through inhibition of a MYC-WWP1 inhibitory pathway. Science, 2019. 364(6441): p. eaau0159. 303. Yuan, T.L. and L.C. Cantley, PI3K pathway alterations in cancer: variations on a theme. Oncogene, 2008. 27(41): p. 5497-510. 304. Siegel, R.L., K.D. Miller, and A. Jemal, Cancer statistics, 2019. CA: A Cancer Journal for Clinicians, 2019. 69(1): p. 7-34.

105 305. Kozakov, D., et al., The FTMap family of web servers for determining and characterizing ligand- binding hot spots of proteins. Nat Protoc, 2015. 10(5): p. 733-55. 306. Niu, Y., et al., Identification of gamma-AApeptides with potent and broad-spectrum antimicrobial activity. Chem Commun (Camb), 2011. 47(44): p. 12197-9. 307. Niu Y, H.Y., Li X, Chen J, Cai J, γ-AApeptides: design, synthesis and evaluation. New J. Chem, 2011. 35. 308. Itaya, K. and M. Ui, A new micromethod for the colorimetric determination of inorganic phosphate. Clin Chim Acta, 1966. 14(3): p. 361-6. 309. Bolarinwa, O., et al., Structure and Function of AApeptides. Biochemistry, 2017. 56(3): p. 445- 457. 310. Brandmaier, A., S.Q. Hou, and W.H. Shen, Cell Cycle Control by PTEN. J Mol Biol, 2017. 311. Riccardi, C. and I. Nicoletti, Analysis of apoptosis by propidium iodide staining and flow cytometry. Nat Protoc, 2006. 1(3): p. 1458-61. 312. Liu, W., et al., PTEN mutation: many birds with one stone in tumorigenesis. Anticancer Res, 2008. 28(6a): p. 3613-9. 313. Irwin, J.J., et al., Automated docking screens: a feasibility study. J Med Chem, 2009. 52(18): p. 5712-20. 314. Lorber, D.M. and B.K. Shoichet, Hierarchical docking of databases of multiple ligand conformations. Curr Top Med Chem, 2005. 5(8): p. 739-49. 315. Sherman, W., H.S. Beard, and R. Farid, Use of an induced fit receptor structure in virtual screening. Chem Biol Drug Des, 2006. 67(1): p. 83-4. 316. Sherman, W., et al., Novel procedure for modeling ligand/receptor induced fit effects. J Med Chem, 2006. 49(2): p. 534-53. 317. Shelley, J.C., et al., Epik: a software program for pK( a ) prediction and protonation state generation for drug-like molecules. J Comput Aided Mol Des, 2007. 21(12): p. 681-91. 318. Harder, E., et al., OPLS3: A Force Field Providing Broad Coverage of Drug-like Small Molecules and Proteins. J Chem Theory Comput, 2016. 12(1): p. 281-96. 319. Friesner, R.A., et al., Glide: a new approach for rapid, accurate docking and scoring. 1. Method and assessment of docking accuracy. J Med Chem, 2004. 47(7): p. 1739-49. 320. Halgren, T.A., et al., Glide: a new approach for rapid, accurate docking and scoring. 2. Enrichment factors in database screening. J Med Chem, 2004. 47(7): p. 1750-9. 321. Friesner, R.A., et al., Extra precision glide: docking and scoring incorporating a model of hydrophobic enclosure for protein-ligand complexes. J Med Chem, 2006. 49(21): p. 6177-96. 322. Romero, P., et al., Sequence complexity of disordered protein. Proteins, 2001. 42(1): p. 38-48. 323. Peng, K., et al., Length-dependent prediction of protein intrinsic disorder. BMC Bioinformatics, 2006. 7: p. 208. 324. Obradovic, Z., et al., Exploiting heterogeneous sequence properties improves prediction of protein disorder. Proteins, 2005. 61 Suppl 7: p. 176-82. 325. Xue, B., et al., PONDR-FIT: a meta-predictor of intrinsically disordered amino acids. Biochim Biophys Acta, 2010. 1804(4): p. 996-1010. 326. Dosztanyi, Z., et al., IUPred: web server for the prediction of intrinsically unstructured regions of proteins based on estimated energy content. Bioinformatics, 2005. 21(16): p. 3433-4. 327. Cheng, Y., et al., Mining alpha-helix-forming molecular recognition features with cross species sequence alignments. Biochemistry, 2007. 46(47): p. 13468-77. 328. Dunker, A.K., et al., Intrinsically disordered protein. J Mol Graph Model, 2001. 19(1): p. 26-59. 329. Oldfield, C.J., et al., Coupled folding and binding with alpha-helix-forming molecular recognition elements. Biochemistry, 2005. 44(37): p. 12454-70. 330. Fan, X. and L. Kurgan, Accurate prediction of disorder in protein chains with a comprehensive and empirically designed consensus. J Biomol Struct Dyn, 2014. 32(3): p. 448-64. 331. Peng, Z. and L. Kurgan, On the complementarity of the consensus-based disorder prediction. Pac Symp Biocomput, 2012: p. 176-87.

106 332. Walsh, I., et al., Comprehensive large-scale assessment of intrinsic protein disorder. Bioinformatics, 2015. 31(2): p. 201-8. 333. Brooks, B.R., et al., CHARMM: the biomolecular simulation program. J Comput Chem, 2009. 30(10): p. 1545-614. 334. Miller, B.T., et al., CHARMMing: a new, flexible web portal for CHARMM. J Chem Inf Model, 2008. 48(9): p. 1920-9. 335. Vanommeslaeghe, K., et al., CHARMM general force field: A force field for drug-like molecules compatible with the CHARMM all-atom additive biological force fields. J Comput Chem, 2010. 31(4): p. 671-90. 336. Jorgensen, W.L., et al., Comparison of simple potential functions for simulating liquid water. The Journal of Chemical Physics, 1983. 79(2): p. 926-935. 337. Cheatham, T.E., III, et al., Molecular Dynamics Simulations on Solvated Biomolecular Systems: The Particle Mesh Ewald Method Leads to Stable Trajectories of DNA, RNA, and Proteins. Journal of the American Chemical Society, 1995. 117(14): p. 4193-4194. 338. Motlagh, H.N., et al., The ensemble nature of allostery. Nature, 2014. 508(7496): p. 331-9. 339. Dobson, C.M., Protein folding and misfolding. Nature, 2003. 426(6968): p. 884-90. 340. Dobson, C.M., Principles of protein folding, misfolding and aggregation. Semin Cell Dev Biol, 2004. 15(1): p. 3-16. 341. Stefani, M. and C.M. Dobson, Protein aggregation and aggregate toxicity: new insights into protein folding, misfolding diseases and biological evolution. J Mol Med (Berl), 2003. 81(11): p. 678-99. 342. Suzuki, K., Diagnosis and treatment of multiple myeloma and AL amyloidosis with focus on improvement of renal lesion. Clin Exp Nephrol, 2012. 16(5): p. 659-71. 343. Xu, J., et al., Gain of function of mutant p53 by coaggregation with multiple tumor suppressors. Nat Chem Biol, 2011. 7(5): p. 285-95. 344. Ano Bom, A.P., et al., Mutant p53 aggregates into prion-like amyloid oligomers and fibrils: implications for cancer. J Biol Chem, 2012. 287(33): p. 28152-62. 345. de Oliveira, G.A., et al., Misfolding, Aggregation, and Disordered Segments in c-Abl and p53 in Human Cancer. Front Oncol, 2015. 5: p. 97. 346. Higashimoto, Y., et al., Unfolding, aggregation, and amyloid formation by the tetramerization domain from mutant p53 associated with lung cancer. Biochemistry, 2006. 45(6): p. 1608-19. 347. Lasagna-Reeves, C.A., et al., Dual role of p53 amyloid formation in cancer; loss of function and gain of toxicity. Biochem Biophys Res Commun, 2013. 430(3): p. 963-8. 348. Ng, J.W., et al., R248Q mutation--Beyond p53-DNA binding. Proteins, 2015. 83(12): p. 2240-50. 349. Rangel, L.P., et al., The aggregation of mutant p53 produces prion-like properties in cancer. Prion, 2014. 8(1): p. 75-84. 350. Silva, J.L., et al., Prion-like aggregation of mutant p53 in cancer. Trends Biochem Sci, 2014. 39(6): p. 260-7. 351. Willis, A., et al., Mutant p53 exerts a dominant negative effect by preventing wild-type p53 from binding to the promoter of its target genes. Oncogene, 2004. 23(13): p. 2330-8. 352. Uversky, V.N., p53 Proteoforms and Intrinsic Disorder: An Illustration of the Protein Structure- Function Continuum Concept. Int J Mol Sci, 2016. 17(11). 353. Backman, S.A., et al., Deletion of Pten in mouse brain causes seizures, ataxia and defects in soma size resembling Lhermitte-Duclos disease. Nat Genet, 2001. 29(4): p. 396-403. 354. Cantley, L.C. and B.G. Neel, New insights into tumor suppression: PTEN suppresses tumor formation by restraining the phosphoinositide 3-kinase/AKT pathway. Proc Natl Acad Sci U S A, 1999. 96(8): p. 4240-5. 355. Cupolillo, D., E. Hoxha, and A. Faralli, Autistic-Like Traits and Cerebellar Dysfunction in Purkinje Cell PTEN Knock-Out Mice. 2016. 41(6): p. 1457-66. 356. Igarashi, A., et al., Nuclear PTEN deficiency causes microcephaly with decreased neuronal soma size and increased seizure susceptibility. J Biol Chem, 2018. 293(24): p. 9292-9300.

107 357. Kwon, C.H., et al., Pten regulates neuronal arborization and social interaction in mice. Neuron, 2006. 50(3): p. 377-88. 358. Kwon, C.H., et al., Pten regulates neuronal soma size: a mouse model of Lhermitte-Duclos disease. Nat Genet, 2001. 29(4): p. 404-11. 359. Yin, Y. and W.H. Shen, PTEN: a new guardian of the genome. Oncogene, 2008. 27(41): p. 5443- 53. 360. Nguyen, H.N., A New Class of Cancer-Associated PTEN Mutations Defined by Membrane. 2015. 34(28): p. 3737-43. 361. Chiti, F. and C.M. Dobson, Protein misfolding, functional amyloid, and human disease. Annu Rev Biochem, 2006. 75: p. 333-66. 362. Kirilyuk, A., et al., An intrinsically disordered region of the acetyltransferase p300 with similarity to prion-like domains plays a role in aggregation. PLoS One, 2012. 7(11): p. e48243. 363. Mascarenhas, N.M. and S. Gosavi, Understanding protein domain-swapping using structure-based models of protein folding. Prog Biophys Mol Biol, 2017. 128: p. 113-120. 364. Banerjee-Basu, S. and A. Packer, SFARI Gene: an evolving database for the autism research community. Dis Model Mech, 2010. 3(3-4): p. 133-5. 365. Cerami, E., et al., The cBio cancer genomics portal: an open platform for exploring multidimensional cancer genomics data. Cancer Discov, 2012. 2(5): p. 401-4. 366. Forbes, S., et al., The Catalogue of Somatic Mutations in Cancer (COSMIC). Curr Protoc Hum Genet, 2008. CHAPTER: p. Unit-10 11. 367. Gao, J., et al., Integrative analysis of complex cancer genomics and clinical profiles using the cBioPortal. Sci Signal, 2013. 6(269): p. pl1. 368. Guerreiro, R., et al., A comprehensive assessment of benign genetic variability for neurodegenerative disorders. bioRxiv, 2018. 369. Stenson, P.D., et al., Human Gene Mutation Database (HGMD): 2003 update. Hum Mutat, 2003. 21(6): p. 577-81. 370. Fernandez-Escamilla, A.M., et al., Prediction of sequence-dependent and mutational effects on the aggregation of peptides and proteins. Nat Biotechnol, 2004. 22(10): p. 1302-6. 371. Kyte, J. and R.F. Doolittle, A simple method for displaying the hydropathic character of a protein. J Mol Biol, 1982. 157(1): p. 105-32. 372. Zhao, B. and B. Xue, Decision-Tree Based Meta-Strategy Improved Accuracy of Disorder Prediction and Identified Novel Disordered Residues Inside Binding Motifs. Int J Mol Sci, 2018. 19(10). 373. Smith, I.N., et al., Conformational Dynamics and Allosteric Regulation Landscapes of Germline PTEN Mutations Associated with Autism Compared to Those Associated with Cancer. Am J Hum Genet, 2019. 104(5): p. 861-878.

108

Appendix A: Supplementary Data

Table A1: HPLC purities and retention time of compounds. γ-AAPeptides 11, 12, 14, 39, and 43 were analyzed following our previous report [7].

Compound Name Purity (based on analytical HPLC) (%) Retention Time (min) 5 98.04 23.59 8 97.92 24.55 9 98.16 26.11 10 99.75 29.39 40 96.01 27.40

109

Figure A1: 1H-NMR data of select γ-AAPeptides. (A) γ-AAPeptide #5. 1H-NMR (500 MHz, d6-DMSO) δ 10.17 (d, J = 19.0 Hz, 0.4H), 9.96 (d, J = 19.0 Hz, 0.6H), 8.60 (d, J = 8.5 Hz, 0.4H), 8.49 (d, J = 8.5 Hz, 0.3H), 8.37 (t, J = 8.0 Hz, 0.4H), 8.01‒8.13 (m, 4H), 7.92‒7.96 (m, 1H), 7.76–7.83 (m, 4H), 7.63–7.67 (m, 0.5H), 7.48–7.58 (m, 4H), 7.15‒7.33 (m, 5H), 4.11–4.52 (m, 3H), 3.51–3.80 (m, 4.2H), 3.19‒3.29 (m, 0.8H), 2.98 (d, J = 13.5 Hz, 0.6H), 2.72–2.88 (m, 2.4H), 2.53–2.68 (m, 2.7H), 2.18‒2.27 (m, 0.7H), 2.08‒2.14 (m, 0.8H), 1.86–2.01 (m, 3.6H), 1.51‒1.69 (m, 14.5H), 1.15–1.45 (m, 3.3H), 0.89‒0.98 (m, 0.5H), 0.72‒0.78 (m, 0.4H). HRMS (ESI) C39H52N5O3 [M+H]+ calc’d = 638.4065; found = 638.4086. (B) γ-AAPeptide #8. 1H-NMR (500 MHz, d6- DMSO) δ 8.63 (d, J = 8.0 Hz, 0.7H), 8.54 (t, J = 5.0 Hz, 0.4H), 8.36 (d, J = 8.5 Hz, 1.7H), 8.28 (t, J = 5.0 Hz, 0.6H), 8.10‒8.14 (m, 1.6H), 7.99‒8.05 (m, 0.8H), 7.93 (d, J = 6.5 Hz, 1H), 7.67‒7.72 (m, 1H), 7.51‒7.62 (m, 4.4H), 7.16‒7.30 (m, 5.8H), 6.52 (s, 0.8H), 5.27‒6.36 (m, 1.5H), 4.10‒4.18 (m, 1H), 4.01 (dd, J = 14.5, 7.5 Hz, 1H), 3.94 (t, J = 8.5 Hz, 1H), 3.71‒3.79 (m, 0.7H), 3.58‒3.67 (m, 1.2H), 2.88‒3.16 (m, 2.7H), 2.67‒2.76 (m, 4H), 2.62 (s, 0.5H), 2.35 (s, 0.5H), 2.14‒2.17 (m, 0.6H), 1.98 (s, 1.8H), 1.87 (brs, 2H), 1.80 (brs, 2H), 1.72‒1.77 (m, 1H), 1.44‒1.66 (m, 16H), 1.23 (br, 1.8H), 1.16 (t, J = 7.0 Hz, 2.5H), HRMS (ESI) C44H56N5O3 [M+H]+ calc’d = 702.4378; found = 702.4396. (C) γ-AAPeptide #9. 1H-NMR (500 MHz, d6-DMSO) δ 8.48 (d, J = 8.5 Hz, 0.4H), 8.25 (d, J = 8.5 Hz, 0.5H), 8.15 (d, J = 8.5 Hz, 0.5H), 8.05 (d, J = 8.5 Hz, 0.6 H), 7.93–7.99 (m, 4.8H), 7.83–7.90 (m, 3H), 7.69–7.72 (m, 3H), 7.48–7.57 (m, 4.2H), 4.36–7.43 (m, 3.6H), 7.25–7.28 (m, 1H), 7.07–7.21 (m, 5H), 4.91¬–5.31 (m, 4.3H), 4.14–4.37 (m, 2.5H), 3.97–4.06 (m, 1H), 3.27–3.34 (m, 1.5H), 2.91 (td, J = 14.0, 4.0 Hz, 0.4H), 2.70–2.82 (m, 1.5H), 2.62– 2.68 (m, 1.4H), 2.54–2.58 (m, 1.4H), 2.33 (d, J = 13.0 Hz, 0.6H), 1.90 (s, 1.7H), 1.70–1.87 (m, 3H), 1.56–1.66 (m, 8.4H), 1.49– 1.52 (m, 0.9H), 1.38–1.46 (m, 6.1H), 1.26–1.34 (m, 1.5H), 1.13–1.24 (m, 1.3H), 1.02–1.09 (m, 0.5H), 0.94–1.00 (m, 0.5H), 0.81– 0.92 (m, 0.2H), 0.66–0.73 (m, 0.2H). HRMS (ESI) C51H62N5O3 [M+H]+ calc’d = 792.4847; found = 792.4878. (D) γ-AAPeptide #10. 1H-NMR (500 MHz, d6-DMSO) δ 8.48 (dd, J = 15.5, 8.5 Hz, 0.7H), 8.29 (d, J = 8.0 Hz, 0.3H), 8.24 (d, J = 8.0 Hz, 0.3H), 8.13 (d, J = 8.5 Hz, 0.3H), 7.99–8.02 (m, 2H), 7.94–7.97 (m, 2.5H), 7.87–7.92 (m, 1.5H), 7.80–7.84 (m, 2.2H), 7.68–7.73 (m, 3.3H), 7.55–7.60 (m, 1H), 7.50–7.53 (m, 0.8H), 7.44–7.50 (m, 0.6H), 7.32–7.41 (m, 1.2H), 7.27–7.30 (m, 2H), 7.20–7.24 (m, 1.6H), 7.08–7.18 (m, 1.5H), 5.06–5.22 (m, 1.7H), 4.73–4.96 (m, 1.8H), 4.62 (d, J = 15.5 Hz, 0.3H), 4.40 (d, J = 16.0 Hz, 0.3H), 4.30 (td, J = 18.5, 6.0 Hz, 0.8H), 3.99–4.24 (m, 2H), 3.27–3.35 (m, 2H), 3.07–3.15 (m, 0.6H), 2.87–2.91 (m, 0.4H), 2.57–2.83 (m, 4H), 2.28–2.36 (m, 1H), 1.03 (t, J = 6.0 Hz, 1H), 1.88 (brd, 1H), 1.71–1.82 (m, 2.3H), 1.54–1.67 (m, 10H), 1.13–1.48 (m, 5H), 0.96–1.10 (m, 1.3H). HRMS (ESI) C49H58F6N5O3 [M+H]+ calc’d = 878.4438; found = 878.4469. (E) γ-AAPeptide #40. 1H- NMR (500 MHz, d6-DMSO) δ 10.89 (d, J = 2.0 Hz, 0.5H), 10.75 (d, J = 2.0 Hz, 0.3H), 8.61 (t, J = 5.5 Hz, 0.4H), 8.48 (d, J = 8.5Hz, 0.4H), 8.30 (t, J = 6.0 Hz, 0.5H), 8.24 (d, J = 8.5 Hz, 0.3H), 8.05–8.10 (m, 2.7H), 2.93–7.96 (m, 1.8H), 7.82–7.88 (m, 1.2H), 7.58–7.68 (m, 3H), 7.50–7.57 (m, 2.4H), 7.40–7.47 (m, 2H), 7.30–7.37 (m, 1H), 7.18–7.22 (m, 0.6H), 7.02–7.10 (m, 1H), 6.97– 6.70 (m, 0.6H), 6.93 (t, J = 7.5 Hz, 0.5H), 4.76 (d, J = 5.5 Hz, 0.6H), 4.72 (d, J = 6.0 Hz, 1H), 4.27–4.33 (m, 0.7H), 4.07–4.16 (m, 1H), 3.87 (d, J = 16.0 Hz, 0.6H), 3.62–3.67 (m, 1.7H), 3.07–3.15 (m, 1.8H), 2.96 (dd, J = 14.5, 4.0 Hz, 1H), 2.81 (d, J = 7.5 Hz, 1H), 2.60–2.64 (m, 1.4H), 2.35 (t, J = 2.0 Hz, 0.5H), 1.84–1.93 (m, 3H), 1.67–1.76 (m, 3H), 1.38–1.62 (m, 12.8H), 1.22–1.33 (m, 4H), 1.09–1.17 (m, 1.6H), 0.97–1.05 (m, 0.9H), 0.76–0.95 (m, 0.8H). HRMS (ESI) C42H55N6O3 [M+H]+ calc’d = 691.4330; found = 691.4357.

110

Figure A1 (continued). (A) γ-AAPeptide #5. 1H-NMR (500 MHz, d6-DMSO) δ 10.17 (d, J = 19.0 Hz, 0.4H), 9.96 (d, J = 19.0 Hz, 0.6H), 8.60 (d, J = 8.5 Hz, 0.4H), 8.49 (d, J = 8.5 Hz, 0.3H), 8.37 (t, J = 8.0 Hz, 0.4H), 8.01‒8.13 (m, 4H), 7.92‒7.96 (m, 1H), 7.76–7.83 (m, 4H), 7.63–7.67 (m, 0.5H), 7.48–7.58 (m, 4H), 7.15‒7.33 (m, 5H), 4.11–4.52 (m, 3H), 3.51–3.80 (m, 4.2H), 3.19‒ 3.29 (m, 0.8H), 2.98 (d, J = 13.5 Hz, 0.6H), 2.72–2.88 (m, 2.4H), 2.53–2.68 (m, 2.7H), 2.18‒2.27 (m, 0.7H), 2.08‒2.14 (m, 0.8H), 1.86–2.01 (m, 3.6H), 1.51‒1.69 (m, 14.5H), 1.15–1.45 (m, 3.3H), 0.89‒0.98 (m, 0.5H), 0.72‒0.78 (m, 0.4H). HRMS (ESI) C39H52N5O3 [M+H]+ calc’d = 638.4065; found = 638.4086. (B) γ-AAPeptide #8. 1H-NMR (500 MHz, d6-DMSO) δ 8.63 (d, J = 8.0 Hz, 0.7H), 8.54 (t, J = 5.0 Hz, 0.4H), 8.36 (d, J = 8.5 Hz, 1.7H), 8.28 (t, J = 5.0 Hz, 0.6H), 8.10‒8.14 (m, 1.6H), 7.99‒8.05 (m, 0.8H), 7.93 (d, J = 6.5 Hz, 1H), 7.67‒7.72 (m, 1H), 7.51‒7.62 (m, 4.4H), 7.16‒7.30 (m, 5.8H), 6.52 (s, 0.8H), 5.27‒6.36 (m, 1.5H), 4.10‒4.18 (m, 1H), 4.01 (dd, J = 14.5, 7.5 Hz, 1H), 3.94 (t, J = 8.5 Hz, 1H), 3.71‒3.79 (m, 0.7H), 3.58‒3.67 (m, 1.2H), 2.88‒3.16 (m, 2.7H), 2.67‒2.76 (m, 4H), 2.62 (s, 0.5H), 2.35 (s, 0.5H), 2.14‒2.17 (m, 0.6H), 1.98 (s, 1.8H), 1.87 (brs, 2H), 1.80 (brs, 2H), 1.72‒1.77 (m, 1H), 1.44‒1.66 (m, 16H), 1.23 (br, 1.8H), 1.16 (t, J = 7.0 Hz, 2.5H), HRMS (ESI) C44H56N5O3 [M+H]+ calc’d = 702.4378; found = 702.4396. (C) γ-AAPeptide #9. 1H-NMR (500 MHz, d6-DMSO) δ 8.48 (d, J = 8.5 Hz, 0.4H), 8.25 (d, J = 8.5 Hz, 0.5H), 8.15 (d, J = 8.5 Hz, 0.5H), 8.05 (d, J = 8.5 Hz, 0.6 H), 7.93–7.99 (m, 4.8H), 7.83–7.90 (m, 3H), 7.69–7.72 (m, 3H), 7.48–7.57 (m, 4.2H), 4.36–7.43 (m, 3.6H), 7.25–7.28 (m, 1H), 7.07–7.21 (m, 5H), 4.91¬–5.31 (m, 4.3H), 4.14–4.37 (m, 2.5H), 3.97–4.06 (m, 1H), 3.27–3.34 (m, 1.5H), 2.91 (td, J = 14.0, 4.0 Hz, 0.4H), 2.70–2.82 (m, 1.5H), 2.62–2.68 (m, 1.4H), 2.54– 2.58 (m, 1.4H), 2.33 (d, J = 13.0 Hz, 0.6H), 1.90 (s, 1.7H), 1.70–1.87 (m, 3H), 1.56–1.66 (m, 8.4H), 1.49–1.52 (m, 0.9H), 1.38– 1.46 (m, 6.1H), 1.26–1.34 (m, 1.5H), 1.13–1.24 (m, 1.3H), 1.02–1.09 (m, 0.5H), 0.94–1.00 (m, 0.5H), 0.81–0.92 (m, 0.2H), 0.66– 0.73 (m, 0.2H). HRMS (ESI) C51H62N5O3 [M+H]+ calc’d = 792.4847; found = 792.4878. (D) γ-AAPeptide #10. 1H-NMR (500 MHz, d6-DMSO) δ 8.48 (dd, J = 15.5, 8.5 Hz, 0.7H), 8.29 (d, J = 8.0 Hz, 0.3H), 8.24 (d, J = 8.0 Hz, 0.3H), 8.13 (d, J = 8.5 Hz, 0.3H), 7.99–8.02 (m, 2H), 7.94–7.97 (m, 2.5H), 7.87–7.92 (m, 1.5H), 7.80–7.84 (m, 2.2H), 7.68–7.73 (m, 3.3H), 7.55–7.60 (m, 1H), 7.50–7.53 (m, 0.8H), 7.44–7.50 (m, 0.6H), 7.32–7.41 (m, 1.2H), 7.27–7.30 (m, 2H), 7.20–7.24 (m, 1.6H), 7.08–7.18 (m, 1.5H), 5.06–5.22 (m, 1.7H), 4.73–4.96 (m, 1.8H), 4.62 (d, J = 15.5 Hz, 0.3H), 4.40 (d, J = 16.0 Hz, 0.3H), 4.30 (td, J = 18.5, 6.0 Hz, 0.8H), 3.99–4.24 (m, 2H), 3.27–3.35 (m, 2H), 3.07–3.15 (m, 0.6H), 2.87–2.91 (m, 0.4H), 2.57–2.83 (m, 4H), 2.28–2.36 (m, 1H), 1.03 (t, J = 6.0 Hz, 1H), 1.88 (brd, 1H), 1.71–1.82 (m, 2.3H), 1.54–1.67 (m, 10H), 1.13–1.48 (m, 5H), 0.96–1.10 (m, 1.3H). HRMS (ESI) C49H58F6N5O3 [M+H]+ calc’d = 878.4438; found = 878.4469. (E) γ-AAPeptide #40. 1H-NMR (500 MHz, d6- DMSO) δ 10.89 (d, J = 2.0 Hz, 0.5H), 10.75 (d, J = 2.0 Hz, 0.3H), 8.61 (t, J = 5.5 Hz, 0.4H), 8.48 (d, J = 8.5Hz, 0.4H), 8.30 (t, J = 6.0 Hz, 0.5H), 8.24 (d, J = 8.5 Hz, 0.3H), 8.05–8.10 (m, 2.7H), 2.93–7.96 (m, 1.8H), 7.82–7.88 (m, 1.2H), 7.58–7.68 (m, 3H), 7.50–7.57 (m, 2.4H), 7.40–7.47 (m, 2H), 7.30–7.37 (m, 1H), 7.18–7.22 (m, 0.6H), 7.02–7.10 (m, 1H), 6.97–6.70 (m, 0.6H), 6.93 (t, J = 7.5 Hz, 0.5H), 4.76 (d, J = 5.5 Hz, 0.6H), 4.72 (d, J = 6.0 Hz, 1H), 4.27–4.33 (m, 0.7H), 4.07–4.16 (m, 1H), 3.87 (d, J = 16.0 Hz, 0.6H), 3.62–3.67 (m, 1.7H), 3.07–3.15 (m, 1.8H), 2.96 (dd, J = 14.5, 4.0 Hz, 1H), 2.81 (d, J = 7.5 Hz, 1H), 2.60–2.64 (m, 1.4H), 2.35 (t, J = 2.0 Hz, 0.5H), 1.84–1.93 (m, 3H), 1.67–1.76 (m, 3H), 1.38–1.62 (m, 12.8H), 1.22–1.33 (m, 4H), 1.09–1.17 (m, 1.6H), 0.97–1.05 (m, 0.9H), 0.76–0.95 (m, 0.8H). HRMS (ESI) C42H55N6O3 [M+H]+ calc’d = 691.4330; found = 691.4357.

111

Figure A1 (continued). (A) γ-AAPeptide #5. 1H-NMR (500 MHz, d6-DMSO) δ 10.17 (d, J = 19.0 Hz, 0.4H), 9.96 (d, J = 19.0 Hz, 0.6H), 8.60 (d, J = 8.5 Hz, 0.4H), 8.49 (d, J = 8.5 Hz, 0.3H), 8.37 (t, J = 8.0 Hz, 0.4H), 8.01‒8.13 (m, 4H), 7.92‒7.96 (m, 1H), 7.76–7.83 (m, 4H), 7.63–7.67 (m, 0.5H), 7.48–7.58 (m, 4H), 7.15‒7.33 (m, 5H), 4.11–4.52 (m, 3H), 3.51–3.80 (m, 4.2H), 3.19‒3.29 (m, 0.8H), 2.98 (d, J = 13.5 Hz, 0.6H), 2.72–2.88 (m, 2.4H), 2.53–2.68 (m, 2.7H), 2.18‒2.27 (m, 0.7H), 2.08‒2.14 (m, 0.8H), 1.86–2.01 (m, 3.6H), 1.51‒1.69 (m, 14.5H), 1.15–1.45 (m, 3.3H), 0.89‒0.98 (m, 0.5H), 0.72‒0.78 (m, 0.4H). HRMS (ESI) C39H52N5O3 [M+H]+ calc’d = 638.4065; found = 638.4086. (B) γ-AAPeptide #8. 1H-NMR (500 MHz, d6-DMSO) δ 8.63 (d, J = 8.0 Hz, 0.7H), 8.54 (t, J = 5.0 Hz, 0.4H), 8.36 (d, J = 8.5 Hz, 1.7H), 8.28 (t, J = 5.0 Hz, 0.6H), 8.10‒8.14 (m, 1.6H), 7.99‒8.05 (m, 0.8H), 7.93 (d, J = 6.5 Hz, 1H), 7.67‒7.72 (m, 1H), 7.51‒7.62 (m, 4.4H), 7.16‒7.30 (m, 5.8H), 6.52 (s, 0.8H), 5.27‒6.36 (m, 1.5H), 4.10‒4.18 (m, 1H), 4.01 (dd, J = 14.5, 7.5 Hz, 1H), 3.94 (t, J = 8.5 Hz, 1H), 3.71‒3.79 (m, 0.7H), 3.58‒3.67 (m, 1.2H), 2.88‒3.16 (m, 2.7H), 2.67‒2.76 (m, 4H), 2.62 (s, 0.5H), 2.35 (s, 0.5H), 2.14‒2.17 (m, 0.6H), 1.98 (s, 1.8H), 1.87 (brs, 2H), 1.80 (brs, 2H), 1.72‒1.77 (m, 1H), 1.44‒1.66 (m, 16H), 1.23 (br, 1.8H), 1.16 (t, J = 7.0 Hz, 2.5H), HRMS (ESI) C44H56N5O3 [M+H]+ calc’d = 702.4378; found = 702.4396. (C) γ-AAPeptide #9. 1H-NMR (500 MHz, d6-DMSO) δ 8.48 (d, J = 8.5 Hz, 0.4H), 8.25 (d, J = 8.5 Hz, 0.5H), 8.15 (d, J = 8.5 Hz, 0.5H), 8.05 (d, J = 8.5 Hz, 0.6 H), 7.93–7.99 (m, 4.8H), 7.83– 7.90 (m, 3H), 7.69–7.72 (m, 3H), 7.48–7.57 (m, 4.2H), 4.36–7.43 (m, 3.6H), 7.25–7.28 (m, 1H), 7.07–7.21 (m, 5H), 4.91¬–5.31 (m, 4.3H), 4.14–4.37 (m, 2.5H), 3.97–4.06 (m, 1H), 3.27–3.34 (m, 1.5H), 2.91 (td, J = 14.0, 4.0 Hz, 0.4H), 2.70–2.82 (m, 1.5H), 2.62–2.68 (m, 1.4H), 2.54–2.58 (m, 1.4H), 2.33 (d, J = 13.0 Hz, 0.6H), 1.90 (s, 1.7H), 1.70–1.87 (m, 3H), 1.56–1.66 (m, 8.4H), 1.49–1.52 (m, 0.9H), 1.38–1.46 (m, 6.1H), 1.26–1.34 (m, 1.5H), 1.13–1.24 (m, 1.3H), 1.02–1.09 (m, 0.5H), 0.94–1.00 (m, 0.5H), 0.81–0.92 (m, 0.2H), 0.66–0.73 (m, 0.2H). HRMS (ESI) C51H62N5O3 [M+H]+ calc’d = 792.4847; found = 792.4878. (D) γ- AAPeptide #10. 1H-NMR (500 MHz, d6-DMSO) δ 8.48 (dd, J = 15.5, 8.5 Hz, 0.7H), 8.29 (d, J = 8.0 Hz, 0.3H), 8.24 (d, J = 8.0 Hz, 0.3H), 8.13 (d, J = 8.5 Hz, 0.3H), 7.99–8.02 (m, 2H), 7.94–7.97 (m, 2.5H), 7.87–7.92 (m, 1.5H), 7.80–7.84 (m, 2.2H), 7.68– 7.73 (m, 3.3H), 7.55–7.60 (m, 1H), 7.50–7.53 (m, 0.8H), 7.44–7.50 (m, 0.6H), 7.32–7.41 (m, 1.2H), 7.27–7.30 (m, 2H), 7.20– 7.24 (m, 1.6H), 7.08–7.18 (m, 1.5H), 5.06–5.22 (m, 1.7H), 4.73–4.96 (m, 1.8H), 4.62 (d, J = 15.5 Hz, 0.3H), 4.40 (d, J = 16.0 Hz, 0.3H), 4.30 (td, J = 18.5, 6.0 Hz, 0.8H), 3.99–4.24 (m, 2H), 3.27–3.35 (m, 2H), 3.07–3.15 (m, 0.6H), 2.87–2.91 (m, 0.4H), 2.57–2.83 (m, 4H), 2.28–2.36 (m, 1H), 1.03 (t, J = 6.0 Hz, 1H), 1.88 (brd, 1H), 1.71–1.82 (m, 2.3H), 1.54–1.67 (m, 10H), 1.13– 1.48 (m, 5H), 0.96–1.10 (m, 1.3H). HRMS (ESI) C49H58F6N5O3 [M+H]+ calc’d = 878.4438; found = 878.4469. (E) γ- AAPeptide #40. 1H-NMR (500 MHz, d6-DMSO) δ 10.89 (d, J = 2.0 Hz, 0.5H), 10.75 (d, J = 2.0 Hz, 0.3H), 8.61 (t, J = 5.5 Hz, 0.4H), 8.48 (d, J = 8.5Hz, 0.4H), 8.30 (t, J = 6.0 Hz, 0.5H), 8.24 (d, J = 8.5 Hz, 0.3H), 8.05–8.10 (m, 2.7H), 2.93–7.96 (m, 1.8H), 7.82–7.88 (m, 1.2H), 7.58–7.68 (m, 3H), 7.50–7.57 (m, 2.4H), 7.40–7.47 (m, 2H), 7.30–7.37 (m, 1H), 7.18–7.22 (m, 0.6H), 7.02–7.10 (m, 1H), 6.97–6.70 (m, 0.6H), 6.93 (t, J = 7.5 Hz, 0.5H), 4.76 (d, J = 5.5 Hz, 0.6H), 4.72 (d, J = 6.0 Hz, 1H), 4.27– 4.33 (m, 0.7H), 4.07–4.16 (m, 1H), 3.87 (d, J = 16.0 Hz, 0.6H), 3.62–3.67 (m, 1.7H), 3.07–3.15 (m, 1.8H), 2.96 (dd, J = 14.5, 4.0 Hz, 1H), 2.81 (d, J = 7.5 Hz, 1H), 2.60–2.64 (m, 1.4H), 2.35 (t, J = 2.0 Hz, 0.5H), 1.84–1.93 (m, 3H), 1.67–1.76 (m, 3H), 1.38– 1.62 (m, 12.8H), 1.22–1.33 (m, 4H), 1.09–1.17 (m, 1.6H), 0.97–1.05 (m, 0.9H), 0.76–0.95 (m, 0.8H). HRMS (ESI) C42H55N6O3 [M+H]+ calc’d = 691.4330; found = 691.4357.

112

113

114

Figure A2: γ-AAPeptide #42 Inhibits PTEN Lipid Phosphatase Activity. γ-AAPeptide #42 was titrated and measured for changes in PTEN lipid phosphatase activity using Malachite Green assay, n=4. Values reported as mean ± SE.

115

Figure A3: γ-AAPeptide Binding Does Not Induce Conformational Changes in Select Active Site Residues. (A-I) Molecular dynamics simulations reveal that residues E91, D92, H93, H123, T160, R161, K163, V166, and I168 have similar fluctuations and orientations of their dihedral angles between apo-PTEN and γ-AAPeptide-bound PTEN. (Phi: black, Psi: blue, Chi1: green, Chi2: magenta, Chi3: red, Chi4, cyan, Chi5: yellow).

116

Figure A3 (continued). (A-I) Molecular dynamics simulations reveal that residues E91, D92, H93, H123, T160, R161, K163, V166, and I168 have similar fluctuations and orientations of their dihedral angles between apo-PTEN and γ-AAPeptide-bound PTEN. (Phi: black, Psi: blue, Chi1: green, Chi2: magenta, Chi3: red, Chi4, cyan, Chi5: yellow).

117

Appendix B: Copyright Permissions

B1: Copyright permission for: Palumbo, E., Zhao, B., Xue, B., Uversky, V., Davé, V., Analyzing the

Aggregation Propensities of Clinically Relevant PTEN Mutants: A New Culprit in Pathogenesis of

Cancer and Other PTENopathies, Journal of Biomolecular Structure and Dynamics, 2019

118 B1: Copyright permission for: Palumbo, E., Zhao, B., Xue, B., Uversky, V., Davé, V., Analyzing the

Aggregation Propensities of Clinically Relevant PTEN Mutants: A New Culprit in Pathogenesis of

Cancer and Other PTENopathies, Journal of Biomolecular Structure and Dynamics, 2019, DOI:

10.1080/07391102.2019.1630005

119

Appendix C: Published Manuscripts

This appendix contains the first page of the original publication for: Palumbo, E., Zhao, B., Xue, B.,

Uversky, V., Davé, V., Analyzing the Aggregation Propensities of Clinically Relevant PTEN Mutants: A

New Culprit in Pathogenesis of Cancer and Other PTENopathies, Journal of Biomolecular Structure and

Dynamics, 2019, DOI: 10.1080/07391102.2019.1630005

120