Characterization of a Non-Canonical Function for Threonyl-Trna Synthetase in Angiogenesis Adam Christopher Mirando University of Vermont

University of Vermont ScholarWorks @ UVM

Graduate College Dissertations and Theses Dissertations and Theses

2015 Characterization Of A Non-Canonical Function For Threonyl-Trna Synthetase In Angiogenesis Adam Christopher Mirando University of Vermont

Follow this and additional works at: https://scholarworks.uvm.edu/graddis Part of the Biochemistry Commons, Cell Biology Commons, and the Pharmacology Commons

Recommended Citation Mirando, Adam Christopher, "Characterization Of A Non-Canonical Function For Threonyl-Trna Synthetase In Angiogenesis" (2015). Graduate College Dissertations and Theses. 523. https://scholarworks.uvm.edu/graddis/523

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

CHARACTERIZATION OF A NON-CANONICAL FUNCTION FOR THREONYL- tRNA SYNTHETASE IN ANGIOGENESIS

A Dissertation Presented

Adam C. Mirando

The Faculty of the Graduate College

The University of Vermont

In Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy Specializing in Biochemistry

May, 2015

Defense Date: March 23, 2015 Dissertation Examination Committee:

Christopher S. Francklyn, Ph.D., Advisor Karen M. Lounsbury, Ph.D., Chairperson Robert J. Kelm, Ph.D. Robert J. Hondal, Ph.D. Paula B. Tracy, Ph.D. Cynthia J. Forehand, Dean of the Graduate College

ABSTRACT

In addition to its canonical role in aminoacylation, threonyl-tRNA synthetase (TARS) possesses pro-angiogenic activity that is susceptible to the TARS-specific antibiotic borrelidin. However, the therapeutic benefit of borrelidin is offset by its strong toxicity to living cells. The removal of a single methylene group from the parent borrelidin generates BC194, a modified compound with significantly reduced toxicity but comparable anti-angiogenic potential. Biochemical analyses revealed that the difference in toxicities was due to borrelidin’s stimulation of amino acid starvation at ten-fold lower concentrations than BC194. However, both compounds were found to inhibit in vitro and in vivo models of angiogenesis at sub-toxic concentrations, suggesting a similar mechanism that is distinct from the toxic responses. Crystal structures of TARS in complex with each compound indicated that the decreased contacts in the BC194 structure may render it more susceptible to competition with the canonical substrates and permit sufficient aminoacylation activity over a wider concentration of inhibitor. Conversely, both borrelidin and BC194 induce identical conformational changes in TARS, providing a rationale for their comparable effects on angiogenesis. The mechanisms of TARS and borrelidin-based compounds on angiogenesis were subsequently tested using zebrafish and cell-based models. These data revealed ectopic branching, non-functional vessels, and increased cell-cell contracts following BC194- treatment or knockdown of TARS expression, suggesting a role for the enzyme in the maturation and guidance of nascent vasculature. Using various TARS constructs this function was found to be dependent on two interactions or activities associated with the TARS enzyme that are distinct from its canonical aminoacylation activity. Furthermore, observations that TARS may influence VEGF expression and purinergic signaling suggest the possibility for a receptor-mediated response. Taken together, the results presented here demonstrate a clear role for TARS in angiogenesis, independent of its primary function in translation. Although the exact molecular mechanisms through which TARS and borrelidin regulate this activity remain to be determined, these data provide a foundation for future investigations of TARS’s function in vascular biology and its use as a target for angiogenesis-based therapy.

CITATIONS

Material from this dissertation has been published in the following form:

Mirando, A.C., Francklyn, C.S., Lounsbury, K.M.. (2014) Regulation of angiogenesis by aminoacyl-tRNA synthetases. International Journal of Molecular Sciences, 15, 23725-48.

Novoa, E. M., Camacho, N., Tor, A., Wilkinson, B., Moss, S., Marín-García, P., Azcárate, I. G., Bautista, J., Mirando, A. C., Francklyn, C. S., Cortés, A., Ribas de Pouplana, L.. (2014) Analogs of natural aminoacyl-tRNA synthetase inhibitors clear malaria in vivo. Proceedings of the National Academy of Science (USA), 111, E5508- 551.

Williams, T.F., Mirando, A.C., Wilkinson, B., Francklyn, C.S., Lounsbury, K.M.. (2013) Secreted Threonyl-tRNA synthetase stimulates endothelial cell migration and angiogenesis. Scientific Reports, 3, 1317

AND

Material from this dissertation has been submitted for publication to Scientific Reports on February 13, 2015 in the following form:

Mirando, A. C., Fang, P., Williams, T. F., Baldor, L. C., Howe, A. K., Ebert, A. M., Wilkinson, B., Lounsbury, K. M., Guo, M., Francklyn, C. S.. Aminoacyl-tRNA synthetase dependent angiogenesis revealed by a bioengineered macrolide inhibitor. Scientific Reports.

ii DEDICATION

I would like to dedicate this work to my beloved parents, Gerald and Janice

Mirando, whose unerring support throughout my life has laid the foundation for this accomplishment.

iii ACKNOWLEDGEMENTS

First and foremost I would like to express my sincerest gratitude to my advisor,

Christopher Francklyn, whose guidance, patience, and curiosity have undoubtedly improved my skills and passion as a scientist. I am also very grateful for sharing my graduate experience with past and present members of the Francklyn Lab and the

Department of Biochemistry, especially Adam, Astrid, Jamie, and Susan for their advice, assistance, and friendship have greatly enriched my studies in both fun and productive ways. I could not have asked for a better mentor and lab group.

Also, a special thanks to the lab of Karen Lounsbury, specifically Karen and

Terry, for their kindness and collaboration from the beginning of this project. Karen has been like a second mentor throughout the worlds of cell culture and angiogenesis and

Terry has always been willing to lend her skill and advice when needed. Additionally, I would like to acknowledge Drs. Jason Botton, Alicia Ebert, Alan Howe, and Tamara

Williams for providing assistance, supplies, and guidance in unfamiliar assays. Finally, I would like to thank the members of my committee, Drs. Robert Hondal, Robert Kelm, and Paula Tracy, for their continued support and interest in my developing career as a scientist.

iv TABLE OF CONTENTS

Page

CITATIONS ...... ii

DEDICATION ...... iii

ACKNOWLEDGEMENTS ...... iv

LIST OF TABLES ...... xiii

LIST OF FIGURES ...... xiv

LIST OF ABBREVIATIONS ...... xvi

CHAPTER 1: INTRODUCTION ...... 1

1.1 Functions of Aminoacyl-tRNA Synthetases ...... 1 1.1.1 Aminoacyl-tRNA synthetases translate the genetic code ...... 1 1.1.2 Maintenance of aminoacylation fidelity ...... 2

1.2 Aminoacyl-tRNA synthetases as regulators of angiogenesis ...... 5 1.2.1 Aminoacyl-tRNA synthetases as nutrient sensors ...... 6 1.2.2 Basics of angiogenic signaling pathways ...... 9

1.3 Angiogenic signaling by extra-cellular ARSs ...... 12 1.3.1 Tyrosyl-tRNA Synthetase ...... 12 1.3.2 Tryptophanyl-tRNA Synthetase ...... 16

1.4 Transcriptional and translational regulation of angiogenesis by ARSs ...... 19 1.4.1 Seryl-tRNA Synthetase ...... 19 1.4.2 Glutamyl-Prolyl-tRNA Synthetase ...... 24

v 1.5 Regulation of angiogenesis by ARS-associated proteins...... 27 1.5.1 Aminoacyl-tRNA synthetase interacting multifunctional protein 1 ...... 27

1.6 General characteristics of angiogenesis-associated aminoacyl-tRNA synthetases ...... 30 1.6.1 Aminoacyl-tRNA synthetases may connect angiogenesis to nutrient sensing. 30 1.6.2 Mechanistic separation of angiogenesis-associated ARSs...... 31 1.6.3 Associations of angiogenesis-related ARSs to inflammation ...... 32 1.6.4 The role of appended domains in ARS-mediated angiogenesis...... 33 1.6.5 ARSs as targets for anti-angiogenic therapy...... 34 1.6.6 Concluding remarks for aminoacyl-tRNA synthetase functions in angiogenesis ...... 35

1.7 Threonyl-tRNA Synthetase ...... 36 1.7.1 Introduction to threonyl-tRNA synthetase ...... 36 1.7.2 Non-canonical functions of threonyl-tRNA synthetase ...... 37

1.8 References...... 41

CHAPTER 2: CHARACTERIZING THE INHIBITION OF HUMAN THREONYL- tRNA SYNTHETASE BY BORRELIDIN AND DERIVATIVES ...... 53

2.1 Introduction...... 53

2.2 Results and Discussion ...... 56 2.2.1 Discussion of borrelidin’s slow, tight-binding kinetics of enzymes ...... 56 2.2.2 Inhibition of human threonyl-tRNA synthetase by borrelidin and derivatives 60 2.2.3 Determination of borrelidin’s mode of inhibition...... 63 2.2.4 Characterizing the differences in borrelidin and BC194 binding ...... 69 2.2.5 Development of a Borrelidin-resistant point mutation ...... 73

2.3 Experimental Procedures ...... 76

vi 2.3.1 Threonyl-tRNA Synthetase Preparation ...... 76 2.3.2 tRNAThr Preparation...... 77 2.3.3 Slow-Tight Binding Assay ...... 78 2.3.4 Steady-State Aminoacylation Assay ...... 79 app 2.3.5 Ki Determination Assays ...... 79 2.3.6 Crystallization and Structure Determination...... 80 2.3.7 In Vitro Translation Assay ...... 80

2.4 References...... 82

CHAPTER 3: SEPARATION OF THE TOXICITIY AND ANTI-ANGIOGENIC EFFECTS OF BORRELIDIN AND BC194 ...... 86

3.1 Introduction...... 86

3.2 Results ...... 89 3.2.1 Borrelidin mediated amino acid starvation elicits cell cycle arrest ...... 89 3.2.2 Borrelidin and BC194 exhibit comparable inhibition of angiogenesis at sub-toxic levels...... 94 3.2.3 BC194 induces specific ectopic blood vessel formation in zebrafish without the developmental toxicity associated with borrelidin...... 98 3.2.4 The effects of borrelidin and derivatives on gene expression of endothelial cell migration associated genes...... 104

3.3 Discussion ...... 105 3.3.1 The greater toxicity exhibited by borrelidin relative to BC194 is due to its decreased susceptibility to threonine competition...... 106 3.3.2 Borrelidin compounds influence blood vessel maturation and migration in zebrafish...... 108

3.4 Experimental Procedures ...... 110 3.4.1 Cell Culture and Reagents ...... 110

vii 3.4.2 Flow Cytometry ...... 110 3.4.3 Endothelial Cell Proliferation Assay ...... 111 3.4.4 Western Blotting ...... 111 3.4.5 In vitro Tube Formation Assay ...... 112 3.4.6 Chick Chorioallantoic Membrane Assay ...... 112 3.4.7 Cell Migration Assay ...... 113 3.4.8 Zebrafish Handling and Treatments ...... 113 3.4.9 Quantitative Assessment of Toxicity ...... 114 3.4.10 Confocal Imaging ...... 115 3.4.11 Live-Imaging of Zebrafish Blood Flow ...... 115 3.4.12 RNA Expression in Zebrafish ...... 115 3.4.13 Statistical Analysis ...... 116

3.5 References...... 117

CHAPTER 4: INVESTIGATING THE MECHANISMS OF THREONYL-tRNA SYNTHETASE’S ANGIOGENIC FUNCTIONS ...... 124

4.1. Introduction...... 124

4.2 Results ...... 126 4.2.1 Knockdown of threonyl-tRNA synthetase expression in zebrafish morphants produces a vascular phenotype ...... 126 4.2.2 Necessary components of threonyl-tRNA synthetase-mediated angiogenesis ...... 129 4.2.3 TARS-mediated angiogenesis is independent of aminoacylation activity .... 132 4.2.4 TARS-mediated angiogenesis requires full-length constructs ...... 134 4.2.5 Mutants that mimic the borrelidin-bound conformation eliminate TARS-angiogenic activity ...... 136

4.3 Discussion ...... 138

viii 4.3.1 Borrelidin regulates angiogenesis in zebrafish specifically though interactions with TARS ...... 138 4.3.2 Defining the TARS components required for the regulation of angiogenesis ...... 139 4.3.3 Models for TARS-mediated angiogenesis and their disruption by borrelidin derivatives ...... 142

4.4 Experimental Procedures ...... 145 4.4.1 Cell Culture and Reagents ...... 145 4.4.2 Protein Preparation ...... 145 4.4.3 In Vitro Tube Formation Assay ...... 146 4.4.4 Chick Chorioallantoic Membrane Assay ...... 147 4.4.5 Zebrafish Handling and Treatments ...... 147 4.4.6 Whole Fish and Confocal Imaging ...... 147 4.4.7 Statistical Analysis ...... 148

4.5 References...... 149

CHAPTER 5: CHARACTERIZING THE NON-CANONICAL, CATALYTIC FUNCTIONS OF HUMAN THREONYL-tRNA SYNTHETASE ...... 152

5.1 Introduction...... 152

5.2 Results ...... 155 5.2.1 TARS possesses GTPase activity ...... 155 5.2.2 TARS-mediated catalysis of Diadenosine Tetraphospate (Ap4A) ...... 158 5.2.3 TARS Catalyzes the Formation of ADP from ATP ...... 164 5.2.4 Effects of purinergic inhibitors on TARS-mediated angiogenesis ...... 169

5.3 Discussion ...... 170 5.3.1 Correlations between non-canonical, catalytic reactions and TARS- mediated angiogenesis ...... 170

ix 5.3.2 Nucleotide preference in non-canonical, catalytic reactions ...... 174 5.3.3 ADP-formation and GTPase activity as nutrient and energy sensors ...... 175 5.3.4 The effect of tRNAThr on ADP formation and Ap4A ...... 176

5.4 Experimental Procedures ...... 178 5.4.1 Preparation of TARS Constructs and tRNAThr ...... 178 5.4.2 Ap4A, AMP, and ADP Formation assays ...... 178 5.4.3 Ap4G synthesis and GTPase Assays ...... 179 5.4.4 Cell Culture and Reagents ...... 179 5.4.5 Tube Formation Assay with MRS2179 and Ip5I ...... 180

5.5 References...... 181

CHAPTER 6: DISCUSSION AND FUTURE PERSPECTIVES ...... 184

6.1 Mechanisms of TARS-mediated angiogenesis ...... 184 6.1.1 Borrelidin’s inhibition of angiogenesis is mediated through direct interactions with TARS ...... 184 6.1.2 Investigating the angiogenic potential of various TARS constructs in the zebrafish model ...... 185 6.1.3 TARS interacts with endothelial cell-derived extracellular matrices ...... 187 6.1.4 Identification of extracellular TARS binding partners ...... 191 6.1.5 Identification of downstream signaling ...... 192

6.2 The nucleo-cytoplasmic distribution of TARS ...... 194 6.2.1 TARS possesses an N-terminal region of unknown function ...... 194 6.2.2 TARS’s N-terminal region possess a possible nuclear localization signal .... 196

6.3 TARS secretion by exosomes ...... 202

6.4 Further investigation of non-canonical, catalytic functions of TARS ...... 204

x 6.5 TARS is associated with ovarian and prostate tumor progression ...... 206 6.5.1 Angiogenesis is an essential process in tumorigenesis ...... 206 6.5.2 The intracellular levels of TARS increases with advancing stages of ovarian and prostate cancers ...... 207

6.6 Selective inhibition of the malarial parasite Plasmodium falciparum ...... 210

6.7 Final Remarks ...... 211

6.8 References...... 212

APPENDIX I: IDENTIFICATION AND CHARACTERIZATION OF THREONYL- tRNA SYNTHETASE INTERACTION PARTNERS ...... 219

A1.1 Introduction ...... 219

A1.2 Results ...... 221 A1.2.1 Characterizing the TAR-VHL interaction ...... 221 A1.2.2 Identifying the TARS interactome through mass-spectrometry ...... 228 A1.2.3 Characterizing TARS’s interaction with PARP1 ...... 230 A1.2.4 Characterizing TARS’s interaction with eEF1A1 ...... 233

A1.3 Discussion ...... 236 A1.3.1 Possible contributions of the TARS-VHL interaction to the regulation of angiogenesis...... 236 A1.3.2 Possible contributions of the TARS-eEF1α and TARS-PARP1 interactions...... 237

A1.4 Experimental Procedures ...... 238 A1.4.1 Affinity and Immunoprecipitation Procedures ...... 238 A1.4.2 Mass spectrometry analysis of affinity purified TARS from cell lysate .... 239

A1.5 References ...... 241

xi COMPREHENSIVE BIBLIOGRAPHY ...... 246

xii LIST OF TABLES

Table Page

Table 1. Angiogenesis-associated activities of aminoacyl-tRNA synthetases ...... 11 app Table 2. Ki values of borrelidin derivatives on TARS constructs ...... 61 Table 3. Aminoacylation activity data for various protein constructs ...... 132 Table 4. Activity data for borrelidin conformation mimetic mutations ...... 138 Table 5. Rates of GTPase activity by various TARS constructs ...... 157 Table 6. The effects of tRNAThr on the rates of WT TARS GTPase activity ...... 158 Table 7. Rates of Ap4A synthesis by various TARS constructs...... 161 Table 8. The effects of tRNAThr on the rates of WT TARS Ap4A synthesis ...... 162 Table 9. Rates of Ap4G synthesis by various TARS constructs...... 163 Table 10. Rates of ADP formation by various TARS constructs ...... 165 Table 11. The effects of tRNAThr on the rates of WT TARS ADP formation ...... 168 Table 12. The functions of various TARS constructs ...... 171 Table 13. Interactions for affinity purified TARS identified by mass spectrometry ...... 229 Table 14. eEF1A1 and TARS GTPase activity ...... 236

xiii LIST OF FIGURES

Figure Page

Figure 1. Mechanisms of angiogenesis by extracellularly acting ARS...... 14 Figure 2. Mechanisms of angiogenesis by intracellularly acting ARS...... 21 Figure 3. The biphasic responses of aminoacyl-tRNA synthetase interacting multifunctional protein 1 (AIMP1) and endothelial monocyte activating polypeptide II (EMAP II) on angiogenesis...... 28 Figure 4. Basic structural components of TARS...... 39 Figure 5. Structures of macrolides used in this study...... 55 Figure 6. Borrelidin exhibits slow-binding kinetics...... 58 app Figure 7. Determination of Ki for borrelidin and related compounds against human TARS...... 62 Figure 8. TARS-borrelidin binding interactions...... 65 Figure 9. Borrelidin induces conformational changes in TARS and disrupts substrate binding...... 68 Figure 10. Structure of TARS-BC194 complex...... 72 app Figure 11. L567V TARS exhibits a Ki value greater than WT TARS...... 75 Figure 12. Borrelidin elicits toxic responses related to cell cycle progression in endothelial cells at lower concentrations than BC194...... 90 Figure 13. The cytotoxicity of borrelidin is linked to the induction of the amino acid starvation response...... 92 Figure 14. BC220 does not induce amino acid starvation toxicity responses...... 94 Figure 15. Borrelidin and BC194 exhibit comparable inhibition of angiogenesis at sub- toxic levels in HUVEC tube formation assays...... 95 Figure 16. Borrelidin and BC194 exhibit comparable inhibition of angiogenesis at sub- toxic levels in chorioallantoic membrane assays...... 96 Figure 17. Borrelidin and BC194 treatment influences cell migration and cell-cell contacts...... 97 Figure 18. Borrelidin and BC194 exhibit different effects on zebrafish toxicity...... 100 Figure 19. Exposure to Borrelidin and BC194 results in vascular defects and mis- patterning...... 101 Figure 20. The vascular defects and mis-patterning resulting from BC194 treatment is not due to the induction of amino acid starvation...... 102

xiv Figure 21. Treatment with BC194 inhibits blood-flow through zebrafish ISVs...... 103 Figure 22. Borrelidin and BC194 treatment affects gene expression...... 104 Figure 23. Depletion of TARS elicits intermediate toxicity relative to borrelidin and BC194 inhibtion...... 127 Figure 24. TARS is involved in vascular development in the zebrafish...... 128 Figure 25. Linear schematics of TARS protein constructs...... 130 Figure 26. Characterizing the angiogenic potential of TARS and related protein constructs...... 131 Figure 27. Characterizing the angiogenic potential of TARS and related protein constructs in chorioallantoic membrane assays...... 134 Figure 28. The effects of borrelidin-conformation mimetic mutations on TARS- mediated angiogensis...... 137 Figure 29. Kinetic characterization of TARS GTPase activity...... 156 Figure 30. Ap4A reaction schematics...... 159 Figure 31. Kinetic characterization of TARS Ap4A synthesis activity...... 160 Figure 32. TARS catalyzes the synthesis of the hybrid dinucleotide tetraphosphate Ap4G...... 164 Figure 33. Kinetic characterization of TARS ADP formation...... 166 Figure 34. The effects of purinergic receptor inhibitors MRS2179 and Ip5I on TARS-mediated angiogenesis...... 170 Figure 35. Proposed models for TARS-stimulation and borrelidin-inhibition of angiogenesis...... 186 Figure 36. TARS interacts with endothelial cell derived fibronectin...... 190 Figure 37. BC194 and VEGF treatments influence the cellular distribution of TARS. . 198 Figure 38. Effects of BC194, TNFα, and VEGFA on the nucleo-cytoplasmic distribution of TARS...... 200 Figure 39. Transmission electron migrographs of PC-3-derived exosomes ...... 203 Figure 40. TARS may regulate VHL E3-ligase function...... 223 Figure 41. TARS interacts with VHL through its N1-domain...... 225 Figure 42. TARS interacts with VHL by displacing elongin B...... 227 Figure 43. TARS interacts with poly(ADP-ribose) polymerase 1 (PARP1)...... 232 Figure 44. TARS interacts with eukaryotic elongation factor 1α1 (eEF1A1)...... 234

xv LIST OF ABBREVIATIONS

AARS – Alanyl-tRNA synthetase AAV – Aortic arch vessels AIMP1 - Aminoacyl-tRNA synthetase interacting multifunctional protein 1 AKT – Protein kinase B ARS – Aminoacyl-tRNA synthetase ASK-1 – Apoptosis signal-regulating kinase 1 ATF4 – Activating transcription factor 4 BN – Borrelidin CAM – Chorioallantoic membrane CCL22 – C-C motif ligand factor-A ChIP – Chromatin immunoprecipitation CHOP – (C/EβP) homology protein Cp – Ceruloplasmin CXCR – CXC-receptor DAPK – Death associated protein kinase DMSO – Dimethyl sulfoxide DNA-PKcs – DNA-dependent protein kinases ECM – Extracellular Matrix Ec ThrRS – Threonyl-tRNA synthetase from Escherichia coli EC – Endothelial cell eEF1A1 – Eukaryotic elongation factor 1A1 EGFR – Epidermal growth factor receptor eIF2α – Eukaryotic initiation factor 1α ELR – Glutamate-leucine-arginine sequence motif EMAP II – Endothelial monocyte activating polypeptide II eNOS – Nitric oxide synthase EPRS – glutamyl-prolyl-tRNA synthetase

xvi ERK – Extracellular-signal-regulated kinase FARS – Phenylalanyl-tRNA synthetase FGF – Fibroblast growth factor GAIT – Interferon--activated inhibitor of translation GAPDH – Glyceraldehyde 3-phosphate dehydrogenase GCN2 – General control nonderepressible 2 HARS – Histidyl-tRNA synthetase HAT – Histone acetyltransferase HEK293 – Human embryonic kidney cells HIFα – Hypoxia inducible factor α hpf – Hours post-fertilization HUVEC – Human umbilical vein endothelial cell IARS – Isoleucyl-tRNA synthetase IFN – Interferon- ISV – Intersegmental vessels JNK – c-Jun N-terminal kinase KARS – Lysyl-tRNA synthetase LARS – Leucyl-tRNA synthetase MAPK – Mitogen-activated protein kinase MARS – Methionyl-tRNA synthetase mTOR – Mammalian target of rapamycin mTORC – Mammalian target of rapamycin complex MSC – Multi-synthetase complex NF-B – Nuclear factor -light-chain-enhancer of activated B cells NLS – Nuclear localization signal NO – Nitric oxide PARP1 – Poly-(ADP-ribose) polymerase 1 PC-3 – PC-3 human prostate cancer cell line PDGF – Platelet derived growth factor

xvii PDMS – Polydimethylsiloxane PHD – Prolyl hydroxylases PKC – Protein kinase C POODLE – Prediction of order and disorder by machine learning PSMA7 – Proteasome subunit alpha type-7 PTI – Prostate tumor inducer QARS – Glutaminyl-tRNA synthetase RRL – Rabbit reticulocyte lysate SARS – Seryl-tRNA synthetase SKOV – SKOV3 ovarian carcinoma cell line Src – Proto-oncogene tyrosine-protein kinase Src STAT3 – Signal transducer and activator of transcription 3 TARS – Eukaryotic threonyl-tRNA synthetase TGS – a unique domain named for TARS, GTPases, and SpoT in which it is found TLR – Toll-like receptor TNFα – Tumor necrosis factor α TNFR – Tumor necrosis factor α receptor WARS – Tryptophanyl-tRNA synthetase UBL – Ubiquitin-like protein VE-cadherin – Vascular endothelial cadherin VEGF – Vascular endothelial growth factor VEGFR – Vascular endothelial growth factor receptor VHL – Von Hippel Lindau tumor suppressor WHEP-domain – a unique domain named or its association with tryptophanyl-, histidyl-, and glutamyl-prolyl-tRNA synthetases XC – Cross-correlation value YARS – tyrosyl-tRNA synthetase ZIPK – Zipper-interacting protein kinase.

xviii CHAPTER 1: INTRODUCTION

1.1 Functions of Aminoacyl-tRNA Synthetases

1.1.1 Aminoacyl-tRNA synthetases translate the genetic code

The aminoacyl-tRNA synthetases (ARSs) are a family of enzymes responsible for the covalent attachment of amino acids to their corresponding tRNA through a process known as aminoacylation. They were first discovered in 1954 by their activation of amino acids with AMP and subsequently found to transfer the amino acid to its corresponding tRNA molecule (1,2). ARSs are thought to be among the first proteins to evolve and their ubiquitous distribution throughout all cellular life can be ascribed to their critical roles in translation (3,4). In order to ensure translational fidelity each amino acid is associated with a specific ARS. These enzymes are highly discriminatory of the amino acids and tRNA molecules used as substrates. Despite their profound structural differences, all ARS catalyze the core aminoacylation reaction through a two-step mechanism described below:

1) AA + ATP  AA~AMP + PPi

2) AA-AMP + tRNAAA  AA-tRNAAA + AMP

The first step, or adenylation, activates the amino acid (AA) through condensation with ATP to form an enzyme-bound adenylate intermediate (AA-AMP) along with the release of inorganic pyrophosphate (PPi). Subsequently, the activated amino acid is

1 transferred to the 2’- or 3’-hydroxyl of the terminal adenosine on the acceptor arm of the tRNA molecule, releasing charged-tRNA and AMP as products.

The ARSs can be subdivided into two, evolutionarily distinct classes, referred to as Class I or Class II, each containing about half the available species. In general, the

ARS for each AA/tRNA pairing fall into one of the two categories, with the notable exception of lysyl-tRNA synthetase (KARS) which has varieties associated with either class. These classifications were originally categorized on the basis of conserved sequence and tertiary structure motifs but have expanded to include substrate binding and quaternary structure as well. Class I ARSs are mostly monomeric and characterized by a

Rossman dinucleotide binding domain and consensus HIGH and KMSKS sequences.

Conversely, Class II ARSs are multimeric, ususually homodimers, and defined by a conserved fold of antiparallel β-strands and three complex sequence motifs. These structural differences also contribute to varying biochemical properties between the two classes. The Rossman fold of Class I enzymes binds ATP in an extended conformation in contrast to the bent conformation adopted in Class II. Additionally, amino acids are transferred to the 2’-hydroxyl of terminal adenosine moiety in tRNA in Class I ARS while Class II varieties transfer to the 3’-hydroxyl group (with the exception of the Class

II phenylalanyl-tRNA synthetase (FARS) which also uses the 2’-hydroxyl) (5,6).

1.1.2 Maintenance of aminoacylation fidelity

High fidelity translation is an essential component for life as misincorporation of amino acids can be detrimental to cell survival. Consequently, errors in translation are rare under normal circumstances, occurring at a rate of 1 for every 103 or 104 translated

2 codons (7,8). A major contributing factor to this fidelity is ARS substrate specificity for both amino acids and cognate tRNA. Discrimination between tRNA molecules is achieved through mutually exclusive interactions with specific tRNA bases, referred to as identity elements, and the respective ARS (reviewed in (9)). Errors in tRNA specificity are quite rare under normal circumstances, with some reports suggesting a 1 in 107 misidentification events, and likely represent a minor factor in overall translation error rate (10). Conversely, specific binding of the much smaller amino acid provides a greater challenge.

Using structurally or chemically restrictive active sites most ARSs are able to distinguish their target amino acid from the majority of others within the cell. However, the close physiochemical properties of certain amino acids makes it nearly impossible for complete discrimination without further processing. In example, several pairs of amino acids differ only by a single methyl group, such as threonine and serine, valine and isoleucine, or alanine and glycine. The smaller molecule of each of these pairs should be able to fit into the binding pocket available to the other with only a modest reduction of binding affinity. Specifically, the loss of hydrophobic interactions from a methyl group is estimated to decrease binding affinity by 14.2 kJ/mol on average, which would reduce discrimination, at best, to 1 in 200 (11). Therefore, to obtain the discrimination actually observed for ARSs, secondary editing functions would have to evolve to reduce the misincorporation of false cognate amino acids.

Editing by ARSs is often described through the use of the classical “double- sieve” model proposed by Fersht and Dingwall in 1979 (12). In this mechanism, the active site of the ARS functions as the “course-sieve” by permitting the binding and

3 catalysis of the cognate and near-cognate amino acids but precluding interactions with most others through steric or incompatible chemical contributions (such as the steric occlusion of leucine observed with isoleucyl-tRNA synthetase (IARS) (13)). The subsequent and specific hydrolysis of false cognate products or intermediates relative to the intended amino acid then constitutes the “fine-sieve”. These “fine-sieve” processes can be further divided into one of two categories, termed pre-transfer or post-transfer editing, based on when the editing event occurs in the reaction scheme. While some

ARSs show a preference for one of these methods, others will switch between the two under certain conditions, indicating the possibility of redundant fidelity mechanisms (14-

16). In pre-transfer editing aminoacyl-adenylates generated from reactions with false cognates are hydrolyzed prior to transfer onto awaiting tRNA molecules as is demonstrated in the editing of homocysteine by methionyl-tRNA synthetase (MARS)

(17). Cleavage of the adenylate intermediate in these situations appears to be kinetically driven and occurs under conditions of reduced aminoacylation activity (14,16).

In contrast, post-transfer editing involves the cleavage of false cognates from tRNA subsequent to the aminoacyl-transfer reaction. Unlike pre-transfer editing, this reaction occurs within an active site separate from the one involved in aminoacylation. In

Class I ARSs this site is composed of a conserved CP-1 domain associated with the invariant Rossman fold (6). In contrast, Class II ARSs do not share a common editing domain structure (18-20). Following aminoacylation, the charged-tRNAs are immediately transferred to a second active site dedicated to editing. Transfer RNA charged with the cognate amino acid lack the physiochemical properties required for binding to this site and are subsequently released. False cognates, however, are readily admissible and

4 cleaved, generating uncharged-tRNA and free amino acid. Features that facilitate this discrimination can utilize the false cognates’ chemical properties, like the Zn2+ atom in alanyl-tRNA synthetase (AARS) which coordinates serine’s hydroxyl group, or size, as demonstrated by steric occlusion of isoleucine but not valine, by IARS’s editing site

(13,21).

The importance of these editing activities is emphasized by the detrimental effects associated with their loss or dysfunction. In cell culture, the increased burden of mistranslation can lead to the activation of cellular stress responses and apoptosis (22).

Conversely, the effects in whole organisms are more complicated. Editing disrupting mutations in AARS have been associated with the accumulation of misfolded proteins in mice, which leads to the loss of Purkinje cells (23). In contrast, while mutations in human

ARSs have been associated with diseases, none have been specifically linked to editing domain mutations. In example, the peripheral neuropathy Charcot Marie Tooth syndrome has been associated with mutations in several ARSs. However, these mutations seem to be unrelated to the aminoacylation and editing functions of the proteins (24). Instead these effects appear to be related to yet unknown noncanonical functions of these ARSs.

Interestingly, the field of ARS secondary functions has expanded rapidly in recent times and have been associated with a variety of different processes, one of which serves as the basis for this work.

1.2 Aminoacyl-tRNA synthetases as regulators of angiogenesis

In addition to their canonical functions in aminoacylation, a growing body of evidence has identified several secondary activities associated with ARSs. Among the

5 earlier reported non-translational functions were effects on RNA splicing. Mitochondrial varieties of tyrosyl-tRNA synthetase (YARS) and leucyl-tRNA synthetase (LARS) possess splice factor-like activities and assist in the excision of group I introns (25,26).

Transcriptional and translational regulatory properties have also been identified in ARSs through association with transcription factors or the binding of tRNA-like structures within mRNA promoter regions (27,28). Furthermore, ARSs in higher-order eukaryotes have evolved roles in both internal and external cellular signaling responses, including apoptosis and inflammation (29,30).

1.2.1 Aminoacyl-tRNA synthetases as nutrient sensors

As catalysts for the aminoacylation reaction, ARS are sensitive to intracellular concentrations of amino acids and ATP. This feature opens opportunities for ARSs to be used as nutritional and energy sensors within the cell. One of the best examples of this concept was observed in recent investigations of LARS, in which the synthetase regulated the activities of the mammalian target of rapamycin (mTOR) signaling pathway. The mTOR pathway serves to transduce a variety of internal and external signaling events into the regulation of translation, cell growth, metabolism and autophagy

(reviewed in (31)). Notably, mTOR is required for the initiation of translation in eukaryotic cells by phosphorylating 4E-BP1 and preventing its inhibition of eIF-4E’s interaction with the 5’- cap of mRNA (32). These responses are primarily mediated by two multi-protein complexes termed mTOR complexes 1 (mTORC1) and 2 (mTORC2).

Of particular interest, the functions of mTORC1 are highly responsive to cellular nutrient levels, specifically amino acids. High concentrations of amino acids facilitate the

6 association of the complex to the lysosomal membrane through activation of the Rag family GTPases (33,34). This translocation permits mTORC1 to interact with its primary activator, Rheb, which is attached to the lysosome membrane, thereby initiating downstream signaling events (35). Interestingly, mTORC1 activation is particularly sensitive to the essential amino acid leucine (36). Subsequent investigations revealed that the mechanism through which leucine regulates mTORC1 was dependent on LARS

(37,38). When leucine is plentiful, LARS activates the GTPase activity of the mTOR regulating protein RagD. Activated RagD joins with the mTORC1 and stimulates its relocation to the lysosomal membrane. Conversely, leucine depletion deactivates mTOR, which arrests translation and restores amino acid homeostasis through the degradation of cellular proteins.

In addition to their roles as protein building blocks some amino acids are required as precursors in other biosynthetic pathways. In particular, glutamine is an important component in nitrogen metabolism and a major source of malic acid used in NADPH synthesis (39,40). Likely owing to its intimate role in basic metabolic pathways, high levels of glutamine contributes to anti-apoptotic signaling in cells (30,41,42). One of these mechanisms involves the glutamine-dependent association of glutaminyl-tRNA synthetase (QARS) and the apoptosis signal-regulating kinase 1 (ASK-1). ASK-1 is activated in response to FAS-ligand signaling and oxidative stress, which eventually leads to apoptosis through the c-jun N-terminal kinase (JNK) associated pathways

(43,44). High levels of glutamine, however, can attenuate this effect by encouraging the binding of QARS to ASK1 and inhibiting its pro-apoptotic signaling. Interestingly, this

7 QARS-ASK1 interaction requires the binding of glutamine to the QARS active site and emphasizes the role of the ARS as a nutrient sensor (30).

Similar to LARS and QARS, tryptophanyl-tRNA synthetase (WARS) also possesses non-canonical signaling functions sensitive to the levels of its cognate amino acid. This effect is dependent on the conformation of an N-terminal WHEP domain.

These helix-turn-helix domains are exclusive to ARSs and are named for tryptophanyl

(W), histidyl (H), and glutamlyl-prolyl (EP) ARSs in which they are found. When the active site of WARS is occupied by its canonical substrates, the WHEP domain closes over it like a cap. However, when this active site is unoccupied (apo-WARS), this domain is unstructured and remains available for mediating protein-protein interactions.

Since WARS is a functional homo-dimer this allows for two separate WHEP interactions.

Specifically, the exposed WHEP domains of apo-WARS dimers form the central component of a complex with the poly-(ADP-ribose) polymerase 1 (PARP1) and DNA- dependent protein kinases (DNA-PKcs) following stimulation by interferon γ (IFNγ).

Formation of this complex encourages the poly-ADP ribosylation of DNA-PKcs and activates its kinase activity against p53. As such, WARS appears to possess a function in the regulation of p53 activity (45). Previously, IFNγ has been shown to activate p53- dependent pathways involved in cellular senescence and apoptosis, suggesting that

WARS’s association with p53 may also be another nutritional stress response (46,47).

However, the full extent of this signaling remains to be determined.

In addition to activating responses related to senescence and apoptosis, amino acid concentrations can be linked to other signaling events as well. For all ARSs, the deprivation of amino acids can activate the general control nonderepressible 2 (GCN2)

8 amino acid starvation and ER-stress pathways through the accumulation of uncharged tRNA molecules (a pathway that will be discussed in Chapter 3 at greater detail) (48). Of particular interest is the dysregulation of mTOR signaling or induction of ER-stress by nutrient deprivation that has been observed to influence the expression of vascular endothelial growth factor A (VEGFA, originally vascular permeability factor) through the regulation of the hypoxia-inducible factor α (HIFα) transcription factor (see Appendix

I for more details on HIFα signaling) (49,50). Since VEGFA is a key signaling molecule in the regulation of angiogenesis (see below), these observations provide an indirect mechanism for ARS-regulation of angiogenesis. More relevant to this dissertation, however, is the observation of direct angiogenic regulation by specific ARSs.

1.2.2 Basics of angiogenic signaling pathways

Angiogenesis, or the growth of new blood vessels from preexisting ones, is a multifaceted process governed by the balance of pro-angiogenic and quiescent signals.

The factors mediating these effects consist of various extracellular matrix (ECM) components, soluble growth factors, cell-cell adhesion proteins, and associated receptors

(reviewed in (51,52)). Under normal conditions vessels maintain a quiescent state through the release of angiopoetin-1 and other survival signals from mural cells (53).

However, the spatially and temporally regulated expression of other factors can tip the balance positively or negatively with respect to proliferation, migration, and maturation of nascent vasculature. Many of these factors, such as fibroblast growth factor (FGF) and platelet-derived growth factor (PDGF), are functionally redundant and associated with a variety of biological functions (53). Conversely, VEGFA and its corresponding receptor

9 VEGFR2 (also referred to as the kinase insert domain receptor (KDR) or fetal liver kinase 1 (Flk1)) are indispensable for the process of angiogenesis and are highly specific to endothelial cell signaling (54). VEGFA is released from tissues in response to hypoxia, inflammation, or wound healing and activates VEGFR2’s receptor tyrosine kinase activity in endothelial cells. Consequently, proliferation, migration, and survival are enhanced by the activation of downstream signaling targets, including ERK1/2, p38-

MAPK, and AKT (55). While these various kinases may elicit overlapping effects, their differences allow for the fine tuning of the angiogenesis process by favoring specific responses, such as migration or proliferation.

To date, angiogenesis-associated activities have been observed in tyrosyl-tRNA synthetase (YARS), tryptophanyl-tRNA synthetase (WARS), seryl-tRNA synthetase

(SARS), glutamyl-prolyl-tRNA synthetase (EPRS) and one closely associated protein, aminoacyl-tRNA synthetase interacting multifunctional protein 1 (AIMP1, also known as p43) (Table 1). Similar to growth factors, YARS, WARS, and AIMP1 function extracellularly and are thought to be secreted. Once outside the cell, receptor interactions vary with each ARS and elicit different responses related to inflammation, proliferation, migration, and cell-cell contacts. Alternatively, SARS and EPRS regulate angiogenesis intracellularly through interactions with DNA and mRNA associated with the transcription and translation of VEGF and other angiogenic genes (56,57). While this

10 Table 1. Angiogenesis-associated activities of aminoacyl-tRNA synthetases Aminoacyl-tRNA synthetases (ARSs) are subdivided based on whether they elicit angiogenesis- based responses outside the cell, inside the cell, or are ARS-associated proteins. Effects on angiogenesis are reported as “Pro” for pro-angiogenic responses (i.e., stimulate new blood vessel growth), “static” for angiostatic responses (i.e., inhibit new blood vessel growth), and “Anti” for anti-angiogenic responses (i.e., reduce current number of vessels). Abbreviations used in Table 1: AIMP1 - Aminoacyl-tRNA synthetase interacting multifunctional protein 1; ARS – Aminoacyl-tRNA synthetases; EC – Endothelial cell; ELR – Glutamate-leucine-arginine sequence motif; EMAP II – Endothelial monocyte activating polypeptide II; EPRS – glutamyl-prolyl-tRNA synthetase; GAIT – Interferon--activated inhibitor of translation; IFN – Interferon-; MSC – Multi-synthetase complex; NLS – Nuclear localization signal; SARS – Seryl-tRNA synthetase; WARS – Tryptophanyl-tRNA synthetase; WHEP-domain – a unique domain named or its association with tryptophanyl-, histidyl-, and glutamyl-prolyl-tRNA synthetases; YARS – tyrosyl-tRNA synthetase.

ARS or Location of Effect on Associated Associated Angiogenic Details Angiogenesis Domains Factor Function

Externally Acting ARS Secreted and cleaved into N and C ELR motif, cytokine fragments; N-terminal ELR YARS Extracellular Pro EMAP II- stimulates angiogenesis; C-terminus may like (58,59) possess EMAP II-like signaling (58- 60,63) Secreted and WHEP-domain removed WHEP by secretion or alternative splicing; loss WARS Extracellular Static domain of WHEP-domain allows for (70) interactions with E-cadherin (67-70)

Internally Acting ARS Directed to nucleus by NLS; Binds to NLS SARS Nucleus Static vegfaa promoter and disrupts cMyc (56,74,78) induction of vegfaa mRNA. (56,74,78) IFNγ stimulates the release of EPRS Three from the MSC and its association with WHEP EPRS Cytoplasm Static the GAIT complex; complex binds to domains mRNA 3' elementsvariety and inhibits of mechanisms (73) translation (57,73,81-85) complicates our understanding of ARS secondary function, it also opens up opportunities ARS Associated Factors Released from MSC in response to for the development of drugs targetingPro specific aspectsapoptotic of angiogenesis. signals; secreted Therefore, as full-length (low conc.) EMAP II AIMP1 or EMAP II; stimulates AIMP1 Extracellular further progress in this field wouldAnti improve our(86) knowledgeangiogensis of mammalian at low concentrations physiology but (high conc.) induces EC apoptosis at high and provide new therapeutic targets for vascular-basedconcentrations therapies. (86-91,95-97)

11 1.3 Angiogenic signaling by extra-cellular ARSs

The process of angiogenesis requires the collaborative effort of numerous cells and cell-types. Coordination of their activities relies heavily on the secretion of various chemokines and growth factors to disseminate information throughout the local microenvironment. The balance of these factors determines the overall stimulatory or inhibitory nature of the signals (53). Two of the angiogenesis-associated ARSs, YARS and WARS, utilize a similar mechanism in which the enzymes are secreted from the cell in response to certain stimuli to engage in autocrine and paracrine type signaling (Figure

1).

1.3.1 Tyrosyl-tRNA Synthetase

The angiogenic properties of YARS are mediated by pro-inflammatory features that were incorporated into the protein during the evolution of higher eukaryotes. Early sequencing of YARS revealed an appended C- terminal domain in eukaryotes that shares

40% homology with endothelial monocyte activating polypeptide II (EMAP II), a pro- cytokine that is released from cells in response to apoptotic signals (58). Under similar conditions, YARS is secreted into the extracellular space. While no signaling activity is observed for the full length protein, cleavage by extracellular polymorphonuclear leukocyte (PMN) elastase generates two separate fragments: an N-terminal catalytic domain fragment, termed “mini-YARS”, and a C-terminal EMAP II-like domain. Similar to EMAP II, the C-terminal domain promotes chemotaxis in immune cells and stimulates the release of tumor necrosis factor-α (TNFα) and tissue factor cytokines. While no direct connection between angiogenesis and the YARS C-terminal domain has been reported,

12 EMAP II itself originates from an ARS associated protein with ties to angiogenesis and will be discussed in greater detail below. Interestingly, exposure to the mini-YARS fragment also elicits chemotactic responses in leukocytes similar to α-chemokines (Cys-

X-Cys (CXC)-chemokines), suggesting that it possesses cytokine activity as well (59).

Therefore, the rest of this section will focus on the mini-YARS fragment.

Like other members of the CXC-chemokine family, mini-YARS possesses a highly conserved Glu-Leu-Arg (ELR) motif essential for receptor interactions. Scatchard analysis of 125I-labeled mini-YARS binding to PMNs revealed saturable binding, indicative of a specific receptor. This interaction is not observed in mutants with a Glu-

Leu-Gln (ELQ) sequence suggesting that the association is dependent on an intact ELR motif. Furthermore, this binding is susceptible to competition by increasing amounts of the CXC-chemokine interleukin-8 (IL-8), suggesting that the IL-8 receptor type A

(CXCR1) is the likely mini-YARS receptor. These results are supported by the observation that mini-YARS bound to cells transfected with the gene for CXCR1 but not those with the CXCR2 gene (59). Although these studies focused on YARS signaling in immune cells, the ELR-motif bearing CXC chemokines also exhibit pro-angiogenic activities (60).

The angiogenic potential of mini-YARS has been investigated through a battery of in vitro and in vivo studies. Similar to other CXC-chemokines, mini-YARS induces

Figure 1. Mechanisms of angiogenesis by extracellularly acting ARS. Left-Full-length tyrosyl-tRNA synthetase (YARS) is secreted from the cell by an unknown mechanism. Outside the cell it is cleaved by polymorphonuclear leukocyte (PMN)-elastase or other protease molecules into N- and C-terminal fragments which stimulate angiogenesis or immune responses respectively. Right-Full-length tryptophanyl-tRNA synthetase (WARS) or mini-WARS are generated from alternative splicing of WARS mRNA and are subsequently secreted into the extracellular space by unknown mechanisms. Full-length WARS is further processed by PMN-elastase or other proteases to form fragments like minipro-- WARS. The WARS fragments disrupt endothelial cell–cell contacts and angiogenic angiogenicsignaling molecules, responses eliciting related anto angiendothelialostatic effect. cell migration and proliferation. Transwell migration assays indicate a significant increase in the migration of endothelial cells when mini-YARS is present in the lower chamber (60,61). Similarly, endothelial cell cultures treated with mini-YARS show increased migration of cells to scratch sites in wound migration assays. Together, these studies suggest that YARS invokes chemotactic responses in endothelial cells similar to those observed with immune cells. Additionally,

14 treatment of endothelial cells with mini-YARS stimulates proliferation and organization of vessel networks according to commercial dye and tube formation assays respectively.

Interestingly, the extent of these angiogenic effects by mini-YARS is comparable to that demonstrated by VEGF, indicating a potent response (61). These observations are supported by several in vivo models as well. Exposure to mini-YARS increases basal vessel formation in chorioallantoic membrane (CAM) and mouse matrigel models of angiogenesis (60).

The pro-angiogenic responses observed for the migration, tube-formation, and

CAM assays are dependent on an intact ELR motif as mutation of any of these residues in mini-YARS appears to inhibit these processes. Given the importance of the ELR for the binding of mini-YARS to CXCR1, these results implicate this receptor as the mediator of angiogenesis. While there was originally some doubt of this due to the lack of a murine

CXCR1 homolog, subsequent studies have since identified a potential rodent CXCR1 candidate (62), suggesting that this receptor pathway may still be viable in mouse models. In addition, the similar angiogenic effects demonstrated by mini-YARS and

VEGF led investigators to examine the involvement of another receptor, VEGFR2.

Treatment with mini-YARS stimulates phosphorylation of Y1054, activating the receptor. Because VEGF expression is not stimulated by exposure to mini-YARS, the authors believe that signaling by the ARS activates VEGFR2 through a trans-activation mechanism (61).

Investigation of down-stream mini-YARS signaling cascades revealed phosphorylation of ERK, Src, and AKT which have been previously demonstrated to initiate pro-angiogenic responses including endothelial cell migration, proliferation, and

15 survival (reviewed in (63)). To further investigate the importance of ERK signaling, the downstream kinase MEK was inhibited using the compound U0126. Interestingly, this treatment blocks mini-YARS induced migration, suggesting that the ERK signaling cascade is important for mini-YARS mediated angiogenesis. In addition to kinases, mini-

YARS activates endothelial nitric oxide synthase (eNOS) through phosphorylation of

Ser-1179, leading to increased nitric oxide (NO) production (61). Previous studies have demonstrated significant effects by NO on vascular permeability and other angiogenic responses, suggesting that it could be another contributing factor to YARS’s pro- angiogenic activity (64). Overall, the combination of these various signaling cascades establishes a clear angiogenic mechanism for extra-cellular YARS that is characteristic of its similarities to the CXC-chemokine family.

1.3.2 Tryptophanyl-tRNA Synthetase

Aspects of WARS secondary functions were observed as early as 1969 with the appearance of alternatively spliced fragments in preparations from bovine pancreas extracts (65). Later, the presence of WARS within exocrine cells provided evidence of secretion (66). Subsequently, a truncated, alternative splice-form of WARS, termed mini-

WARS, was discovered. Interestingly, the structure of the WARS catalytic core closely resembles that of YARS and the discovery of pro-angiogenic functions by fragments of

YARS indicates that WARS fragments may also possess possible biological functions separate from aminoacylation (67). Some of the general mechanisms governing WARS’s secondary functions appear similar to YARS; however, the two proteins elicit opposing responses on angiogenesis. Expression of both mini-WARS and the full-length protein is

16 regulated by interferon-γ (IFNγ), a cytokine that stimulates the expression of angiostatic chemokines and suppresses the expression of the pro-angiogenic cytokine IL-8. This correlation between the structures of WARS and YARS coupled with the biological connection with IFNγ led to investigations of potential angiogenic effects of WARS.

Like YARS, full-length WARS has no influence on endothelial cell migration and proliferation assays prior to the cleavage of the appended WHEP domain. This can be accomplished through alternative splicing, as in mini-WARS, or through limited proteolysis, such as by PMN elastase, which generates T1 and T2 WARS fragments.

However, whereas the active YARS fragments are pro-angiogenic, WARS fragments inhibit both VEGF- and YARS-induced migration and proliferation in endothelial cells, suggesting an angiostatic function (60,67). Similar attenuation of angiogenic responses are observed in the CAM, mouse matrigel plug, and mouse retinal models indicating that these effects can be recapitulated under in vivo conditions.

To better understand the inhibitory effects of WARS, the effect of the T2-WARS fragment on downstream, angiogenic kinase cascades was investigated. The presence of extracellular, T2-WARS inhibits the VEGF- and shear stress-mediated phosphorylation of ERK, AKT, and eNOS (68,69). The localization of T2-WARS to the cell periphery

(68) coupled with the susceptibility of treated endothelial cells to shear induced detachment (69) led to the hypothesis that WARS fragments interacted with cell-cell contacts. Subsequently it was determined that knockdown of vascular endothelial cadherin (VE-cadherin) disrupts the interaction between WARS and endothelial cells, implicating it as the putative extracellular mediator of WARS’s non-canonical activities

(68). Molecular modeling studies suggest that the basis of this interaction was the

17 insertion of Trp2 and Trp4 from the N-terminal EC1 domain of VE-cadherin into the active site of the WARS core. This is supported by mutational analyses in which substitution of Trp2 and Trp4 on EC1 or essential binding residues in WARS abolish the interaction between the proteins. Interestingly, the binding of VE-cadherin to full-length

WARS is occluded by the presence of the N-terminally appended WHEP domain in full- length WARS. The loss of this domain in WARS fragments allows VE-cadherin to access the WARS active site, providing a mechanism for the cleavage-dependence of WARS’s angiostatic activity (70).

As with both YARS and TARS (see below), the mechanism of WARS exocytosis remains unknown. The fact that WARS-secretion is not affected by brefeldin A or

A23187 indicates that it is not dependent on Golgi- or calcium-dependent transport respectively (71). The lack of a canonical export sequence led to the hypothesis that secretion may be dependent on protein-protein interactions. Yeast two-hybrid and co- immunoprecipitation experiments revealed an interaction between WARS and p11 (also referred to as S100A10), a protein member of the annexin II complex (71). Previously p11 had been found to be important for the cellular retention of the TASK-1 transcription factor (72). Similarly, the connection of this complex to WARS’s retention was confirmed by demonstrating that knockdown of either p11 or annexin II increases the secretion of WARS (71). Taken together, these data suggest a model in which angiostatic signals stimulate the release of WARS from p11 allowing for its subsequent secretion from the cell through unknown mechanisms. Once in the extra-cellular milieu WARS can then be processed into active fragments that inhibit angiogenesis by disrupting endothelial cell-cell contacts.

1.4 Transcriptional and translational regulation of angiogenesis by ARSs

As part of their role in aminoacylation, ARSs possess features necessary for interactions with nucleic acids. Although normally limited to tRNA, adaptations of this feature can allow for interactions with other molecules bearing the necessary recognition motifs. In particular, seryl-tRNA synthetase is able to bind to the vegfaa gene promoter and reduce its transcription while glutamyl-prolyl-tRNA synthetase is able to bind to the

3’ UTR of VEGFA mRNA and inhibit its translation (73,74). In contrast to the inter- cellular communication propogated by the secreted ARSs, SARS and EPRS mediate intracellular responses to angiogenesis through regulation of canonical angiogenic factors at the transcriptional and translational level (Figure 2).

1.4.1 Seryl-tRNA Synthetase

The connection between SARS and angiogenesis was first noticed based on two independent investigations of vascular deficient zebrafish mutants related to the adrasteia

(adr) gene (75) and ko095 locus (56). These mutations were associated with the dilation of aortic arch vessels (AAV) and the presence of ectopic branching in the cranial and intersegmental vessels (ISV). These effects appeared after 60 and 72 hours post fertilization (hpf) for the ko095 and adr mutants respectively. Genetic mapping using select SSLP markers subsequently identified two of the adr (76) and one of the ko095

(56) mutations within the sars gene. All three mutations correspond to single nucleotide changes with the ko095 (56) and one of the adr mutants resulting in a premature stop codon and the second adr mutant converting a conserved phenylalanine (Phe383) to a

19 valine (76). In further corroboration of the importance of SARS in vascular development, a third study reported circulation restriction in zebrafish containing sars retroviral insertional mutants (77). The fact that these separately reported mutations in sars present with similar deficiencies in vascular development supports the hypothesis that sars is the likely source of this phenotype.

Although the above mutations fall within the catalytic core of the enzyme, the effects on the vasculature appear to be unrelated to the canonical aminoacylation function. Knockdown of SARS in endothelial cells using siRNA results in concentration- dependent cell death after 48 hours. This effect is attributed to loss of SARS aminoacylation activity and, thus, is non-specific towards cell type. In contrast, SARS knockdown in tube-formation experiments at earlier time points results in increased, albeit disorganized, branching of the tube networks, indicating a endothelial cell-specific function (56). These results are supported by investigations using the catalytically inactive T429A SARS mutant. Despite its deficiency in aminoacylation injections of

T429A SARS mRNA into ko095 mutant fish completely rescues the aberrant vessel phenotype, thereby separating the two functions (56).

To date, the mechanism through which SARS elicits angiogenesis remains to be elucidated; however, it appears to be dependent on signaling by the VEGFA-VEGFR2 and VEGFC-VEGFR3 pathways. Inhibition of VEGFR2 and 3 with the inhibitors

SU5416 or MAZ51 respectively blocks the dilation of AAVs and ISV ectopic branches.

Figure 2. Mechanisms of angiogenesis by intracellularly acting ARS. (1) seryl-tRNA synthetase (SARS) is transported into the nucleus via a canonical nuclear localization sequence (NLS) sequence and binds to the vascular endothelial growth factor A gene (vegfaa) promoter. The binding of SARS inhibits c-Myc-mediated expression of vascular endothelial growth factor A (VEGFA) through modification of vegfaa epigenetics. This effect is angiostatic; (2) glutamyl-prolyl-tRNA synthetase (EPRS) is phosphorylated upon treatment of the cell with IFNγ, releasing it from the multisynthetase complex (MSC). Phospho-EPRS binds with L13a, GAPDH, and NSAP1 to form the interferon-γ-activated inhibitor of Similar translation (GAIT) complex. This complex then binds to mRNA containing stem-loop, 3' UTR effectsGAIT elements are not observedvia EPRS withWHEP the domains, EGFR- specificthereby blockinginhibitor their AF1478, translation. indicating As VEGFA a VEGF - contains a GAIT element, this effect is angiostatic. mediated response (76). Likewise, knockdown of VEGFA and VEGFC or their respective receptors, VEGFR2 and VEGFR3, in zebrafish by morpholino injection prevents the development of ectopic branches as well (56).

21 Despite this link to the VEGF receptor family, the necessity of the VEGF ligands for SARS-mediated angiogenesis indicates that the ARS is unlikely to be a direct activator of VEGFR2 or 3. Investigation of mRNA expression changes in vascular- related genes as a result of the ko095 mutation provides some clarification. Similar to

TARS, the ko095 SARS mutant induces no changes in Notch gene expression while vegfaa shows significant increases. Interestingly, the augmented vegfaa expression is attenuated by the injection of the aminoacylation-deficient T429A SARS mRNA, suggesting that this effect is due to SARS’s non-canonical function. These findings led to the conclusion that SARS mediates angiogenesis by suppressing VEGFA expression with the dependence on the VEGFC-VEGFR3 pathway explained simply due to the fact that it is essential for angiogenesis in general (56).

Similar to what was observed for YARS, SARSs secondary activity is found to depend on a C-terminal, appended domain that arose during vertebrate evolution. This domain, termed UNE-S, encompasses residues 470 to 514 and contains a functional nuclear localization signal (NLS) (78). Interestingly, both the ko095 and adr non-sense mutations completely abolish the presence of this domain. Similarly, while not located in

UNE-S, the vascular-deficient F383V SARS mutant does not localize to the nucleus while the catalytically inactive, but vascular-functional, T429A SARS mutant possesses a clear nuclear distribution. The importance of SARS localization on VEGFA mRNA expression in zebrafish was investigated by knocking down endogenous zebrafish SARS with anti-SARS morpholinos and complementing with mRNA for various SARS constructs. Interestingly, F383V and Δ482-514 SARS (SARS lacking UNE-S domain) exhibit threefold increased expression of vegfaa compared to wildtype human SARS and

22 all other constructs with a functional NLS (78). These results were later recapitulated in both endothelial cells and human embryonic kidney (HEK293) cells in which endogenous SARS was knocked down with shRNA and replaced by WT SARS or SARS lacking the NLS sequence (ΔNLS-SARS) (74). Similar to zebrafish studies, vegfaa expression is increased two-fold in cells treated with SARS shRNA relative to cells treated with control shRNA. Furthermore, SARS knockdown cells complemented with expression of ΔNLS-SARS exhibit a four-fold increase in VEGFA mRNA levels when compared to those complemented with wildtype SARS. As a functional consequence of the increased VEGFA expression, complementation of sh-SARS-treated endothelial cells with ΔNLS-SARS promotes angiogenesis in an endothelial tube-formation assay in contrast to complementation with wildtype SARS. Together, these data suggest a model in which nuclear SARS suppresses expression of vegfaa and downstream angiogenesis through a system that is conserved from zebrafish to humans.

Recent investigations have presented a new mechanism for SARS-mediated regulation of vegfaa expression (74). Affinity precipitation of purified SARS with calf- thymus DNA demonstrates that the synthetase is capable of direct interactions with DNA.

Chromatin immunoprecipitation (ChIP) and DNA foot-printing analyses narrows this interaction to the vegfaa promoter region within -62 to -38 bp from the start site. Of note, this site overlaps with the c-Myc binding site in the vegfaa promoter, suggesting that

SARS may inhibit the pro-angiogenic activities of c-Myc. Specifically, c-Myc recruits histone acetyltransferase (HAT) proteins to acetylate histones in the vegfaa locus, generating transcriptionally active euchromatin (79). Direct competition between SARS and c-Myc for binding to the vegfaa promoter was subsequently demonstrated using an

23 electrophoretic mobility shift assay (74). Once bound to the DNA, SARS facilitates the formation of heterochromatin through the recruitment of the histone deacetylase SIRT2, thereby counteracting c-Myc-mediated acetylation and inhibiting vegfaa transcription.

1.4.2 Glutamyl-Prolyl-tRNA Synthetase

Glutamyl-prolyl-tRNA synthetase (EPRS) is unique among the ARSs due to its bifunctional nature, a characteristic that first originated in higher eukaryotes. The enzyme possesses a modular structure consisting of an N-terminal elongation factor 1β-like domain, a glutamyl-tRNA synthetase catalytic domain (EARS), a linker region composed of three tandem WHEP domains, and a C-terminal prolyl-tRNA synthetase catalytic domain (PARS) (73). Additionally, EPRS is associated with the aminoacyl multisynthetase complex (MSC), a heterocomplex composed of nine ARSs (D, EP, I, K,

L, M, Q, and R) and three non-catalytic scaffolding proteins (AIMP1-3). Both the

WHEP-linker region and association with the MSC contribute to EPRS’s non-canonical functions.

The connection between EPRS and angiogenesis arose from investigations of a phenomenon in which prolonged exposure of cells to IFNγ induces a marked reduction of translation in specific genes. Termed the interferon-gamma-activated inhibitor of translation (GAIT), its mechanism involves the binding of a proteinaceous complex

(termed GAIT complex) to stem-loop structures, or GAIT-elements, in the 3’- untranslated region (UTR) of a specific mRNA, thereby preventing its translation (80,81).

Yeast three-hybrid and RNA-affinity purification experiments using GAIT element sequences as bait identified the four comprising members of the GAIT complex:

24 ribosomal protein L13a (L13a), EPRS, NS1-associated protein-1 (NSAP1), and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) (82,83). Stimulation of cells by

IFNγ results in the phosphorylation and release of EPRS and L13a from the MSC and large ribosomal subunit, respectively, and initiates the formation of the active GAIT complex (83). Binding of the complex to mRNA GAIT elements is primarily mediated by the three tandem WHEP domains in EPRS, emphasizing the ARS as a critical component of the GAIT system (73).

The mechanism used by the GAIT complex to inhibit translation was initially determined in studies of the ceruloplasmin (Cp) protein expression and then expanded to transcripts associated with other proteins, including VEGF. In Cp, treatment of monocytic cells with INFγ for more than 24 hours results in the attenuation of Cp translation. Since the reduction of protein levels are not associated with a concurrent drop in mRNA expression, regulation must occur at the translational level (80). Binding of the

GAIT complex to GAIT element in mRNA 3’ UTR inhibits the assembly of the 43S ribosome, preventing the translation process (81,82,84). Using the pattern analysis algorithm software PatSearch several other genes have since been identified with similar predicted 3’ UTR GAIT elements, including death associated protein kinase (DAPK), zipper-interacting protein kinase (ZIPK), C-C motif ligand factor-A (CCL22), and

VEGFA (57). In particular, the presence of GAIT elements in VEGFA mRNA suggested the possibility that the GAIT complex could be important for the regulation of angiogenic processes.

To better understand the possible role of the GAIT complex in angiogenesis, the effects of IFNγ-treatment on VEGFA expression and downstream angiogenic activities

25 were investigated with respect to the putative VEGF mRNA GAIT element. The interaction between the GAIT complex and VEGFA mRNA was confirmed using electrophoretic mobility shift assays of 32P labeled VEGF riboprobes containing the

VEGFA GAIT element. U937 monocyte lysates treated with IFNγ for 16 or 24 hours exhibit interactions with all four components of the GAIT complex. Furthermore, the effect of this interaction on VEGF translation was investigated using a luciferase reporter assay. Incubation of luciferase mRNA containing the VEGFA 3’UTR GAIT element with

U937 lysate treated with IFNγ for 24 hours show a significant reduction in luciferase activity relative to untreated lysate or mRNA lacking the GAIT element. However, translation can be rescued regardless of IFNγ-treatment through the immunodepletion of

EPRS, confirming that the suppression was mediated by the GAIT complex. The effect of reduced VEGF levels on angiogenesis was subsequently investigated using endothelial cell proliferation and tube formation assays (85). Exposure of endothelial cells to conditioned media from U937 monocytes treated with IFNγ for 8 or 12 hours exhibits a marked increase in proliferation whereas only basal tube formation is observed with treatments of 16 or 24 hours. Similarly, tube formation of endothelial cells cultured on matrigel is enhanced 2.5 fold relative to basal levels in U937 conditioned media treated for 8 hours with IFNγ while no change is observed in media treated for 16 or 24 hours.

These changes in angiogenic activities are consistent with the kinetics for GAIT complex activation and support a role for the complex in the time-dependent attenuation of VEGF- signaling by IFNγ following the initial induction period.

26 1.5 Regulation of angiogenesis by ARS-associated proteins.

1.5.1 Aminoacyl-tRNA synthetase interacting multifunctional protein 1

The aminoacyl-tRNA synthetase interacting multifunctional protein 1 (AIMP1; also known as p43) is one of three auxiliary proteins that combines with nine ARSs to make up the MSC heterocomplex. Isolation of the AIMP1 gene from Chinese hamster ovary

(CHO) cDNA led to the discovery that the exact sequence for EMAP II, a proinflammatory cytokine released from apoptotic cells, is contained within the C-terminal region, suggesting that it may serve as the EMAP II precursor molecule and possess potent signaling activity (86). As indicated above, an EMAP II-like domain is found in a

C-terminal fragment of YARS and may possess similar signaling functions. During apoptosis, activated caspase 7 catalyzes the cleavage of the C-terminal EMAP II domain of AIMP1 and subsequently releases it from the MSC complex (Figure 3) (87). Once freed, EMAP II is secreted from the cell and stimulates the recruitment of macrophages and initiates other signaling cascades associated with the regulation of inflammation and, of particular interest, angiogenesis.

Both EMAP II and full-length AIMP1 have been observed to regulate angiogenesis in a variety of assays. However, the nature of this regulation is not straightforward as they exhibit both dose-dependent and biphasic characteristics (88). Low concentrations

(0.25 to 1 nM) of either EMAP II or AIMP1 induce endothelial cell migration in transwell migration assays and AIMP1 stimulates angiogenesis in the CAM assay (the effects of EMAP II in CAM assays remain to be investigated). Conversely, concentrations of 10 nM and above exhibit dose-dependent inhibition in the same assay and appear to be related to the induction of endothelial cell specific apoptosis (88).

Figure 3. The biphasic responses of aminoacyl-tRNA synthetase interacting multifunctional protein 1 (AIMP1) and endothelial monocyte activating polypeptide II (EMAP II) on angiogenesis. (1) Caspase 7 cleaves the C-terminal EMAP II peptide from AIMP1, releasing EMAP II from the multi-synthetase complex; (2) Alternatively, AIMP1 is released by unknown processes; (3) Both EMAP II and AIMP1 are secreted from the cell by mechanisms that remain to be determined; (4) At low concentrations (perhaps corresponding to early portions of the AIMP1/EMAP II signaling responses in vivo) EMAPII and AIMP1 bind to surface receptors, activating various kinase cascades; (5) These kinase cascades induce the expression of tumor necrosis factor α (TNFα); (6) TNFα is secreted from the cell and mediates angiogenesis through canonical TNFα pathways; (7) At high concentrations (perhaps corresponding to prolonged exposure in vivo), AIMP1 and EMAP II induce the expression of the TNFα-receptor (TNFR); and (8) Over-stimulation of TNFR by TNFα induces apoptosis through signaling by the TNFR death-domain, resulting in an anti- angiogenic response.

Additional studies with EMAP II revealed similar anti-angiogenic responses in both tube formation and rat aortic ring assays at concentrations of 20 µg/ml (0.6 µM) and greater

(89). The cause of this biphasic signaling may be linked to the varying sensitivities of specific kinase cascades (88,90). Interestingly, full-length AIMP1 appears to be a more

28 potent inducer of migration and apoptosis than the processed EMAP II peptide (88) suggesting a role for the whole protein in signaling events.

Several lines of evidence implicate TNFα in the AIMP1 angiogenic signaling pathway. AIMP1 elicits a striking increase in the expression of TNFα, a potent regulator of both pro- and anti-angiogenic activities (91). In terms of pro-angiogenic activities,

TNFα increases expression of VEGF via NF-κB activation (92). Conversely, TNFα can also induce apoptosis in endothelial cells, lending itself to an anti-angiogenic role (93).

AIMP1-stimulated secretion of TNFα initiates the downstream phosphorylation and activation of ERK1/2, a well-known angiogenic signal. The necessity of ERK1/2 activation for AIMP1 regulation of TNFα was subsequently confirmed using the MEK inhibitor U0126, which was shown to attenuate the induction of TNFα (91). As such, the

TNFα activation of ERK1/2 likely contributes to the initial, pro-angiogenic phase of

AIMP1 and EMAP II signaling. However, higher concentrations or sustained secretion of

AIMP1 or EMAPII enhances the surface representation of TNF receptor 1 (TNFR1) on endothelial cells, thereby facilitating TNFα-induced apoptosis through the TNFR1 death domain (94). Therefore, TNFα-signaling may contribute to the second, anti-angiogenic portion of AIMP1 and EMAPII’s biphasic response.

In addition to its role in inflammatory signaling, several different studies have revealed that EMAP II is capable of effecting canonical angiogenic pathways. Treatment with EMAP II inhibits endothelial cell migration by disrupting interactions between fibronectin and alpha5 beta1 integrins (95). ELISA-based assays also determined that

EMAP II directly interacts with VEGFR1 and 2 and inhibits the binding of VEGF.

Functionally, EMAP II inhibition of VEGFRs prevents the phosphorylation of

29 downstream signal mediators, including AKT, ERK1/2, p38MAPK, and Raf as well as

VEGF-dependent endothelial cell migration (96). Furthermore, over expression of EMAP

II in cells increases the proteasome subunit alpha type-7 (PSMA7)-dependent proteasomal breakdown of HIF1α even under hypoxic conditions and, consequently, reduces hypoxia-induced tube formation in endothelial cells at concentrations of 20 µM

(97). It should be noted, however, that most of the above functions were investigated at

EMAP II concentrations that would induce the inhibitory portion of EMAP II’s biphasic response. Therefore, under normal conditions, it is possible that these inhibitory effects would be delayed relative to the initial induction of angiogenesis associated with VEGF- and hypoxia-based signaling.

1.6 General characteristics of angiogenesis-associated aminoacyl-tRNA

synthetases

1.6.1 Aminoacyl-tRNA synthetases may connect angiogenesis to nutrient sensing.

As indicated in section 1.2.1 above, the dependence of the aminoacylation reaction on ATP and amino acid concentrations opens opportunities for ARSs to function as nutrient sensors within the cell. Perhaps the best example of this was demonstrated in recent studies of leucyl-tRNA synthetase (LARS) in which LARS was found to regulate mTOR-mediated autophagy in a leucine-dependent manner (37,38). Alternatively, the accumulation of uncharged-tRNA due to amino acid depletion or inhibition of any ARS can activate the GCN2-mediated amino acid starvation response, leading to translational arrest and apoptosis (48). Since nutritionally starved cells would benefit from the increased vasculature, the acquisition of angiogenesis-related functions by these enzymes

30 would complement their roles as intracellular sensory mechanisms. While no direct nutrient sensing has been observed in angiogenesis-related ARSs, substrate-induced conformational changes are well documented among ARSs in general (98,99). These changes may regulate important interactions responsible for the transport of the ARSs and down-stream signaling events essential for their non-canonical activities and would warrant further study.

1.6.2 Mechanistic separation of angiogenesis-associated ARSs.

Relevant ARSs use a wide variety of mechanisms in their regulation of angiogenesis. However, these enzymes can be subdivided into two broad categories based on the location of their actions. The first, consisting of YARS, WARS, and

AIMP1, function extracellularly and must be secreted from the cell. While little is known of the secretion mechanisms these events are able to accommodate full-length proteins.

Since the canonical function of ARSs in translation is catalyzed in the cytoplasm, the export of these proteins likely serves as an important regulatory step in their non- canonical functions. The identification of this system is critical to understanding these secondary functions and their regulatory process.

Conversely, SARS and EPRS regulate angiogenesis intracellularly through interactions with DNA or mRNA (73,74). Unlike their secreted counterparts, the regulation methods of the intracellular-acting ARSs are better understood. EPRS interacts with mRNA in the cytoplasm and does not require a transport mechanism. Instead, phosphorylation initiates its dissociation from the MSC, thereby acting as a switch between the primary and secondary functions. SARS, on the other hand, interacts with

31 nuclear DNA and must undergo nuclear transport. While the processes regulating this event are not completely known, a nuclear localization signal has been identified within the C-terminal domain of SARS that is essential for its regulation of VEGFA transcription. Taken together, these findings would suggest that all of the angiogenic

ARSs appear to separate their canonical and non-canonical functions by sequestration.

Therefore, undesired angiogenic activity can be limited by regulating the translocation of the ARSs from the cytoplasm or their dissociation from a protein complex.

1.6.3 Associations of angiogenesis-related ARSs to inflammation

Except for SARS, all angiogenesis-associated ARSs possess some connection to inflammation. Both YARS and AIMP1 are associated with the pro-inflammatory EMAP

II or ELR domains which activate receptors from the CXC-receptor (CXCR) family.

These receptors mediate chemokine signaling pathways important for the coordinated regulation of inflammatory and angiogenic responses during wound healing (reviewed in

(100)). Consistent with other ELR containing chemokines, mini-YARS stimulates pro- angiogenic signaling through interactions with CXCR1 expressed by endothelial cells.

Interestingly, while all other known ELR chemokines also stimulate CXCR2, knockdown of the receptor has no effect on angiogenesis by mini-YARS, suggesting that mini-YARS may be specific to CXCR1 or possesses yet unknown activities associated with CXCR2.

Similarly, EMAPII, as suspected for other ELR-lacking chemokines, regulates angiogenesis through CXCR3. Knockdown of CXCR3 inhibits EMAP II-mediated migration by endothelial cells, a response that would be anticipated for an angiogenic molecule. However, these results appear atypical since CXCR3 is primarily a mediator

32 of angiostasis. Given EMAP II’s biphasic response, it is may be possible that EMAP II’s angiogenic and angiostatic activities are both mediated through CXCR3 but are differentially regulated. Additionally, the EMAP II precursor, AIMP1, regulates production of TNFα and its receptor, TNFR, through interations with CD23. Connections of TNFα to angiogenesis and inflammation are well established in the literature (101). Of particular interest, TNFα exhibits a dose-response effect, stimulating angiogenesis at lower concentrations and inhibiting it at higher (102), consistent with the biphasic response observed for EMAP II. Given the sequence homology between EMAP II, similar connections to TNFα may exist with the C-terminal domain of YARS and may play an important role in YARS-mediated angiogenesis.

The non-canonical functions of WARS and EPRS appear to be regulated by signals from inflammatory factors rather than initiating them like YARS and AIMP1.

Treatment of cells with IFNγ stimulates the secretion of WARS and releases EPRS from the MSC, initiating translational control by the GAIT complex. In contrast to other inflammatory signals, like TNFα, IFNγ is anti-angiogenic, consistent with its association with WARS and EPRS signaling. Overall, the connection to inflammation underlies a complex association of ARSs with responses to cellular stress separate from their better known, house-keeping roles in translation.

1.6.4 The role of appended domains in ARS-mediated angiogenesis.

The angiogenesis regulatory mechanisms exhibited by the ARSs described above are often dependent on appended domains or sequences not found in the prokaryotic enzymes (reviewed in (103)). These features include EMAP II-domains in YARS and

33 AIMP1, an ELR motif in YARS, WHEP domains in WARS and EPRS, and an NLS in

SARS. The WHEP domain is exclusive to the ARS family of enzymes and is associated with RNA interactions. The RNA binding affinity of the WHEP domain increases with the number of tandem repeats and is essential for the non-canonical functions associated with EPRS (104). Conversely, the single WHEP domain of WARS does not have a known RNA-binding function but instead regulates angiogenesis by blocking access of

N-terminal E-cadherin to the WARS active site. Interestingly, the ELR and EMAP II domains appear to have arisen in insects, whose blood diffuses throughout the body cavity instead of within defined vessels (105). Likewise, WHEP domains have been observed in organisms as early as choanoflagellates, the closest unicellular ancestor to animals (106). Therefore, it is likely that these domains developed for a separate purpose and expanded to include angiogenesis following the development of a complex vascular system.

1.6.5 ARSs as targets for anti-angiogenic therapy.

Given the importance of angiogenesis in the growth and metastatic progression of several solid tumors, many of the most prominent angiogenic factors, such as VEGF, have been investigated as potential targets for anti-metastatic therapies. Unfortunately, the benefits of these approaches are often temporary owing to adaptive responses by the target tissue (107). Therefore, better understanding of ARS angiogenic activities may allow for the development of new small molecule inhibitors for use in combinatorial cancer therapies.

34 Several natural and synthetic ARS inhibitors have been reported that exhibit anti- microbial or immunosuppressant activity by inhibiting canonical functions of aminoacyl- tRNA synthetases (108-111). More relevant to anti-cancer treatments, disruption of an interaction between lysyl-tRNA synthetase (KARS) and the laminin receptor by a small molecule inhibitor was found to inhibit the metastatic progression of a mouse breast cancer model (112). Considering that chemical intervention of KARS non-canonical functions influenced cancer progression it would not be surprising if similar results could be obtained through the manipulation of other angiogenesis-associated ARSs. Therefore, these efforts can be enhanced through the better understanding of regulation and mechanistic processes.

1.6.6 Concluding remarks for aminoacyl-tRNA synthetase functions in angiogenesis

The above observations provide clear evidence for a role of ARSs in angiogenesis. Their very different mechanisms suggest that these functions developed separately for each enzyme over the course of evolution. This is not completely surprising as ARSs are among the most ancient and essential enzymes to come into existence and have had a long evolutionary window to acquire secondary functions.

Furthermore, their connection to the nutritional and translational status of the cell makes them prime candidates for the development of homeostatic functions, like angiogenesis, that are essential for the survival of complex, multicellular organisms. Nonetheless, many features of the molecular mechanisms associated with ARS-mediated angiogenesis remain to be determined. With the exception of EPRS, little is known about the regulatory signals separating the canonical and non-canonical functions. Notably, the lack

35 of a clear secretory mechanism for the extracellularly active ARSs has confounded investigations focused on the initiation of their angiogenesis-associated activities.

Furthermore, future drug design would benefit from improvements in knowledge of downstream signaling pathways and structural studies of ARS-partner interactions.

Overall, the connection between ARSs and angiogenesis emphasizes the importance of their characterization to the field of vascular biology and opens up new and exciting areas for future research.

1.7 Threonyl-tRNA Synthetase

1.7.1 Introduction to threonyl-tRNA synthetase

Threonyl-tRNA synthetase (TARS) is an 83 kDa (human, monomer; 76 kDa E. coli) class II ARS that catalyzes the attachment of threonine to its corresponding tRNA for use in translation. However, like other class II ARSs, the protein is typically found as a homodimer (6). The structure of TARS is modular with four core domains that have been conserved from prokaryotes to eukaryotes (Figure 4A) (18). The most N-terminal of these is the N1-domain (residues 83-144 human, 1-62 E. coli), an ubiquitin-like domain of unknown function. This domain is adjacent to the N2 (or editing) domain (residues

145-303 human, 63-224 E. coli), which functions in the post-transfer editing of tRNAThr aminoacylated with the false cognate serine (113). This structure is connected to the aminoacylation (or catalytic) domain (residues 319-607 human, 240-528 E. coli) by a large linker helix (residues 304-321 human, 225-242 E. coli).

The aminoacylation domain contains all the conserved structural elements characteristic of class II ARSs as well as the three substrate binding sites (5,6). While 36 numerous residues contribute to substrate binding, a Class II conserved arginine, R442 in human TARS (R383 in E. coli), is perhaps the most important (Figure 4B). R442 interacts with both the α-phosphate of ATP and the carboxyl group of threonine and stabilizes transition state for the adenylation reaction (114). Additionally, a Zn2+ is coordinated within the active site by two histidines and a cysteine. Crystallographic evidence suggests that this ion is important for interacting with the threonine substrate’s hydroxyl and removing it results in loss of TARS aminoacylation activity (18). The role of the Zn2+ atom in discrimination of amino acids with hydroxyl groups is further supported by kinetic evidence of alanyl-tRNA synthetase (AARS) editing, which appears to use a Zn2+ atom to preferentially bind serine (21). The most C-terminal domain is referred to as the anti-codon binding domain (residues 614-723 human, 535-642 E. coli), which selects and binds anticodons corresponding to the substrate tRNAThr (18).

Together, the aminoacylation and anticodon binding domains form the dimer interface. In addition to these core domains, eukaryotic TARS species possess an unstructured, N- terminal extension (residues 1-81 in humans). While the length of this region can vary greatly between organisms some sequence features, such as a lysine rich region, appear to be conserved amongst most species. The function of this region, however, remains unknown.

1.7.2 Non-canonical functions of threonyl-tRNA synthetase

Prior to the work presented herein, the connection between TARS and angiogenesis had not been well established. However, evidence of its connection to vascular biology was identified concurrently with studies of the TARS-specific inhibitor

37 borrelidin, a naturally occurring polyketide anti-microbial, and anti-fungal compound from Streptomyces rochei. Characterizations of borrelidin-resistant bacterial strains and

Chinese hamster ovary cells with selective amplification of tars genes established TARS as the principal target of borrelidin (115,116). Subsequent investigations revealed that borrelidin inhibits tube formation of rat aortic ring fragments with an IC50 of 0.4 ng/ml

(0.8 nM). This value was 15-fold lower than the IC50 for endothelial cell growth inhibition (6 ng/ml), suggesting a specific anti-angiogenic activity for the compound

(117). Subsequent investigations of borrelidin’s effects also demonstrated that treatment with the compound reduced the occurrence of lung metastases in B16-BL6 mouse models of melanoma at concentrations known to inhibit angiogenesis (118). The fact that no alternative targets to TARS can be identified coupled with the threonine-dependent nature of some of the anti-angiogenic effects strongly implicates TARS as the mediator of these effects.

In addition to the angiogenesis connections alluded to by borrelidin, TARS is the target of the PL-7 autoantibody in the polymyositis and dermatomyositis autoimmune disorders (119). In addition to TARS, auto-antibodies have been discovered against five other ARSs which lead to similar complications. Etiological characterizations associated these conditions with auto-reactive T-cells in muscle tissue mediated through two major inflammatory pathways: NF-κB and TNFα signaling (120). Both of these signaling pathways share connections with inflammation and angiogenesis through regulation of downstream kinases or hypoxia inducible factor 1α stabilization (121,122). Of note,

Figure 4. Basic structural components of TARS. (A) The monomeric structure of TARS (PDB: 1QF6) with domains highlighted as follows: N1-domain (purple), N2/editing domain (yellow), linker helix (pink), aminoacylation domain (cyan), and anticodon-binding domain (blue). (B) Ribbon diagram of the TARS active site (PDB:1QF6) with specific features highlighted. AMP, A76 of the tRNAThr, and the conserved R442 are shown as stick models; the bound Zn2+ is shown as a grey sphere.

39 muscle biopsies from patients in the early phases of polymyositis and dermatomyositis have demonstrated increased expression of VEGF, possibly due to reduction in capillaries within the affected tissue (123). In addition, these disorders have been associated with an increased risk of several varieties of cancer (124). However, it is currently unclear as to whether or not the secretion of TARS in angiogenesis has any connection to these diseases or if these autoimmune disorders contribute to neovascularization in tumorigenesis.

The ability of the TARS inhibitor, borrelidin, to disrupt angiogenesis and metastasis strongly implicated the compound as a possible cancer therapeutic.

Additionally, the association of other ARSs with angiogenesis and the lack of any other targets for borrelidin suggested that TARS may possess secondary functions in aminoacylation. Despite the extensive literature on the topic angiogenesis remains a complicated process that is incompletely understood. Further investigation of TARS and its inhibitors will help clarify another aspect of angiogenesis and may open up new opportunities for pharmacological manipulation of vascular biology. The data presented herein provide evidence for a role of extracellular TARS in the stimulation of vessel migration and maturation in both in vitro and in vivo systems. Treatment with borrelidin or its derivatives are found to inhibit TARS’s angiogenic properties through the induction of conformational changes within the enzyme. Contrary to previous reports these effects are demonstrated to be separate from the toxicity exhibited by the compounds and rectifies incorrect assumptions of inhibitor mechanism. Together these data provide a foundation for future investigations TARS’s secondary function in angiogenesis and the mechanism of its inhibitor, borrelidin.

40 1.8 References

1. Hoagland, M. B., Stephenson, M. L., Scott, J. F., Hecht, L. I., and Zamecnik, P. C. (1958) A soluble ribonucleic acid intermediate in protein synthesis. J Biol Chem 231, 241-257

2. Zamecnik, P. C., and Keller, E. B. (1954) Relation between phosphate energy donors and incorporation of labeled amino acids into proteins. J Biol Chem 209, 337-354

3. Woese, C. R., Olsen, G. J., Ibba, M., and Soll, D. (2000) Aminoacyl-tRNA synthetases, the genetic code, and the evolutionary process. Microbiology and molecular biology reviews : MMBR 64, 202-236

4. Wolf, Y. I., Aravind, L., Grishin, N. V., and Koonin, E. V. (1999) Evolution of aminoacyl-tRNA synthetases--analysis of unique domain architectures and phylogenetic trees reveals a complex history of horizontal gene transfer events. Genome research 9, 689-710

5. Carter, C. W., Jr. (1993) Cognition, mechanism, and evolutionary relationships in aminoacyl-tRNA synthetases. Annu Rev Biochem 62, 715-748

6. Ibba, M., and Soll, D. (2000) Aminoacyl-tRNA synthesis. Annu Rev Biochem 69, 617-650

7. Loftfield, R. B., and Vanderjagt, D. (1972) The frequency of errors in protein biosynthesis. Biochem J 128, 1353-1356

8. Parker, J. (1989) Errors and alternatives in reading the universal genetic code. Microbiological reviews 53, 273-298

9. Giege, R., Sissler, M., and Florentz, C. (1998) Universal rules and idiosyncratic features in tRNA identity. Nucleic Acids Res 26, 5017-5035

10. Jakubowski, H., and Goldman, E. (1992) Editing of errors in selection of amino acids for protein synthesis. Microbiological reviews 56, 412-429

11. Fersht, A. R., Shindler, J. S., and Tsui, W. C. (1980) Probing the limits of protein- amino acid side chain recognition with the aminoacyl-tRNA synthetases. Discrimination against phenylalanine by tyrosyl-tRNA synthetases. Biochemistry 19, 5520-5524

12. Fersht, A. R., and Dingwall, C. (1979) Evidence for the double-sieve editing mechanism in protein synthesis. Steric exclusion of isoleucine by valyl-tRNA synthetases. Biochemistry 18, 2627-2631

41 13. Nureki, O., Vassylyev, D. G., Tateno, M., Shimada, A., Nakama, T., Fukai, S., Konno, M., Hendrickson, T. L., Schimmel, P., and Yokoyama, S. (1998) Enzyme structure with two catalytic sites for double-sieve selection of substrate. Science 280, 578-582

14. Dulic, M., Cvetesic, N., Perona, J. J., and Gruic-Sovulj, I. (2010) Partitioning of tRNA-dependent editing between pre- and post-transfer pathways in class I aminoacyl-tRNA synthetases. J Biol Chem 285, 23799-23809

15. Martinis, S. A., and Boniecki, M. T. (2010) The balance between pre- and post- transfer editing in tRNA synthetases. FEBS Lett 584, 455-459

16. Minajigi, A., and Francklyn, C. S. (2010) Aminoacyl transfer rate dictates choice of editing pathway in threonyl-tRNA synthetase. J Biol Chem 285, 23810-23817

17. Jakubowski, H., and Fersht, A. R. (1981) Alternative pathways for editing non- cognate amino acids by aminoacyl-tRNA synthetases. Nucleic Acids Res 9, 3105- 3117

18. Sankaranarayanan, R., Dock-Bregeon, A. C., Romby, P., Caillet, J., Springer, M., Rees, B., Ehresmann, C., Ehresmann, B., and Moras, D. (1999) The structure of threonyl-tRNA synthetase-tRNA(Thr) complex enlightens its repressor activity and reveals an essential zinc ion in the active site. Cell 97, 371-381

19. Sasaki, H. M., Sekine, S., Sengoku, T., Fukunaga, R., Hattori, M., Utsunomiya, Y., Kuroishi, C., Kuramitsu, S., Shirouzu, M., and Yokoyama, S. (2006) Structural and mutational studies of the amino acid-editing domain from archaeal/eukaryal phenylalanyl-tRNA synthetase. Proc Natl Acad Sci U S A 103, 14744-14749

20. Sokabe, M., Ose, T., Nakamura, A., Tokunaga, K., Nureki, O., Yao, M., and Tanaka, I. (2009) The structure of alanyl-tRNA synthetase with editing domain. Proc Natl Acad Sci U S A 106, 11028-11033

21. Pasman, Z., Robey-Bond, S., Mirando, A. C., Smith, G. J., Lague, A., and Francklyn, C. S. (2011) Substrate specificity and catalysis by the editing active site of Alanyl-tRNA synthetase from Escherichia coli. Biochemistry 50, 1474- 1482

22. Nangle, L. A., Motta, C. M., and Schimmel, P. (2006) Global effects of mistranslation from an editing defect in mammalian cells. Chem Biol 13, 1091- 1100

23. Lee, J. W., Beebe, K., Nangle, L. A., Jang, J., Longo-Guess, C. M., Cook, S. A., Davisson, M. T., Sundberg, J. P., Schimmel, P., and Ackerman, S. L. (2006) Editing-defective tRNA synthetase causes protein misfolding and neurodegeneration. Nature 443, 50-55 42 24. Yao, P., and Fox, P. L. (2013) Aminoacyl-tRNA synthetases in medicine and disease. EMBO Mol Med 5, 332-343

25. Li, G. Y., Becam, A. M., Slonimski, P. P., and Herbert, C. J. (1996) In vitro mutagenesis of the mitochondrial leucyl tRNA synthetase of Saccharomyces cerevisiae shows that the suppressor activity of the mutant proteins is related to the splicing function of the wild-type protein. Molecular & general genetics : MGG 252, 667-675

26. Mohr, G., Rennard, R., Cherniack, A. D., Stryker, J., and Lambowitz, A. M. (2001) Function of the Neurospora crassa mitochondrial tyrosyl-tRNA synthetase in RNA splicing. Role of the idiosyncratic N-terminal extension and different modes of interaction with different group I introns. J Mol Biol 307, 75-92

27. Lee, Y. N., Nechushtan, H., Figov, N., and Razin, E. (2004) The function of lysyl- tRNA synthetase and Ap4A as signaling regulators of MITF activity in FcepsilonRI-activated mast cells. Immunity 20, 145-151

28. Putney, S. D., and Schimmel, P. (1981) An aminoacyl tRNA synthetase binds to a specific DNA sequence and regulates its gene transcription. Nature 291, 632-635

29. Howard, O. M., Dong, H. F., Yang, D., Raben, N., Nagaraju, K., Rosen, A., Casciola-Rosen, L., Hartlein, M., Kron, M., Yang, D., Yiadom, K., Dwivedi, S., Plotz, P. H., and Oppenheim, J. J. (2002) Histidyl-tRNA synthetase and asparaginyl-tRNA synthetase, autoantigens in myositis, activate chemokine receptors on T lymphocytes and immature dendritic cells. The Journal of experimental medicine 196, 781-791

30. Ko, Y. G., Kim, E. Y., Kim, T., Park, H., Park, H. S., Choi, E. J., and Kim, S. (2001) Glutamine-dependent antiapoptotic interaction of human glutaminyl-tRNA synthetase with apoptosis signal-regulating kinase 1. J Biol Chem 276, 6030-6036

31. Kim, J., and Guan, K. L. (2011) Amino acid signaling in TOR activation. Annu Rev Biochem 80, 1001-1032

32. Gingras, A. C., Raught, B., and Sonenberg, N. (2001) Regulation of translation initiation by FRAP/mTOR. Genes Dev 15, 807-826

33. Sancak, Y., Bar-Peled, L., Zoncu, R., Markhard, A. L., Nada, S., and Sabatini, D. M. (2010) Ragulator-Rag complex targets mTORC1 to the lysosomal surface and is necessary for its activation by amino acids. Cell 141, 290-303

34. Sancak, Y., Peterson, T. R., Shaul, Y. D., Lindquist, R. A., Thoreen, C. C., Bar- Peled, L., and Sabatini, D. M. (2008) The Rag GTPases bind raptor and mediate amino acid signaling to mTORC1. Science 320, 1496-1501

43 35. Long, X., Lin, Y., Ortiz-Vega, S., Yonezawa, K., and Avruch, J. (2005) Rheb binds and regulates the mTOR kinase. Current biology : CB 15, 702-713

36. Cota, D., Proulx, K., Smith, K. A., Kozma, S. C., Thomas, G., Woods, S. C., and Seeley, R. J. (2006) Hypothalamic mTOR signaling regulates food intake. Science 312, 927-930

37. Han, J. M., Jeong, S. J., Park, M. C., Kim, G., Kwon, N. H., Kim, H. K., Ha, S. H., Ryu, S. H., and Kim, S. (2012) Leucyl-tRNA synthetase is an intracellular leucine sensor for the mTORC1-signaling pathway. Cell 149, 410-424

38. Segev, N., and Hay, N. (2012) Hijacking leucyl-tRNA synthetase for amino acid- dependent regulation of TORC1. Mol Cell 46, 4-6

39. DeBerardinis, R. J., Mancuso, A., Daikhin, E., Nissim, I., Yudkoff, M., Wehrli, S., and Thompson, C. B. (2007) Beyond aerobic glycolysis: transformed cells can engage in glutamine metabolism that exceeds the requirement for protein and nucleotide synthesis. Proc Natl Acad Sci U S A 104, 19345-19350

40. Newsholme, P., Procopio, J., Lima, M. M., Pithon-Curi, T. C., and Curi, R. (2003) Glutamine and glutamate--their central role in cell metabolism and function. Cell biochemistry and function 21, 1-9

41. Evans, M. E., Jones, D. P., and Ziegler, T. R. (2003) Glutamine prevents cytokine-induced apoptosis in human colonic epithelial cells. The Journal of nutrition 133, 3065-3071

42. Ropeleski, M. J., Riehm, J., Baer, K. A., Musch, M. W., and Chang, E. B. (2005) Anti-apoptotic effects of L-glutamine-mediated transcriptional modulation of the heat shock protein 72 during heat shock. Gastroenterology 129, 170-184

43. Noguchi, T., Ishii, K., Fukutomi, H., Naguro, I., Matsuzawa, A., Takeda, K., and Ichijo, H. (2008) Requirement of reactive oxygen species-dependent activation of ASK1-p38 MAPK pathway for extracellular ATP-induced apoptosis in macrophage. J Biol Chem 283, 7657-7665

44. Tobiume, K., Matsuzawa, A., Takahashi, T., Nishitoh, H., Morita, K., Takeda, K., Minowa, O., Miyazono, K., Noda, T., and Ichijo, H. (2001) ASK1 is required for sustained activations of JNK/p38 MAP kinases and apoptosis. EMBO reports 2, 222-228

45. Sajish, M., Zhou, Q., Kishi, S., Valdez, D. M., Jr., Kapoor, M., Guo, M., Lee, S., Kim, S., Yang, X. L., and Schimmel, P. (2012) Trp-tRNA synthetase bridges DNA-PKcs to PARP-1 to link IFN-gamma and p53 signaling. Nat Chem Biol 8, 547-554

44 46. Kim, K. S., Kang, K. W., Seu, Y. B., Baek, S. H., and Kim, J. R. (2009) Interferon-gamma induces cellular senescence through p53-dependent DNA damage signaling in human endothelial cells. Mechanisms of ageing and development 130, 179-188

47. Luth, S., Schrader, J., Zander, S., Carambia, A., Buchkremer, J., Huber, S., Reifenberg, K., Yamamura, K., Schirmacher, P., Lohse, A. W., and Herkel, J. (2011) Chronic inflammatory IFN-gamma signaling suppresses hepatocarcinogenesis in mice by sensitizing hepatocytes for apoptosis. Cancer Res 71, 3763-3771

48. Wek, R. C., Jiang, H. Y., and Anthony, T. G. (2006) Coping with stress: eIF2 kinases and translational control. Biochem Soc Trans 34, 7-11

49. Ghosh, R., Lipson, K. L., Sargent, K. E., Mercurio, A. M., Hunt, J. S., Ron, D., and Urano, F. (2010) Transcriptional regulation of VEGF-A by the unfolded protein response pathway. PLoS One 5, e9575

50. Guba, M., von Breitenbuch, P., Steinbauer, M., Koehl, G., Flegel, S., Hornung, M., Bruns, C. J., Zuelke, C., Farkas, S., Anthuber, M., Jauch, K. W., and Geissler, E. K. (2002) Rapamycin inhibits primary and metastatic tumor growth by antiangiogenesis: involvement of vascular endothelial growth factor. Nature medicine 8, 128-135

51. Adams, R. H., and Alitalo, K. (2007) Molecular regulation of angiogenesis and lymphangiogenesis. Nat Rev Mol Cell Biol 8, 464-478

52. Sottile, J. (2004) Regulation of angiogenesis by extracellular matrix. Biochim Biophys Acta 1654, 13-22

53. Carmeliet, P., and Jain, R. K. (2011) Molecular mechanisms and clinical applications of angiogenesis. Nature 473, 298-307

54. Hoeben, A., Landuyt, B., Highley, M. S., Wildiers, H., Van Oosterom, A. T., and De Bruijn, E. A. (2004) Vascular endothelial growth factor and angiogenesis. Pharmacological reviews 56, 549-580

55. Cross, M. J., and Claesson-Welsh, L. (2001) FGF and VEGF function in angiogenesis: signalling pathways, biological responses and therapeutic inhibition. Trends in pharmacological sciences 22, 201-207

56. Fukui, H., Hanaoka, R., and Kawahara, A. (2009) Noncanonical activity of seryl- tRNA synthetase is involved in vascular development. Circ Res 104, 1253-1259

57. Mukhopadhyay, R., Jia, J., Arif, A., Ray, P. S., and Fox, P. L. (2009) The GAIT system: a gatekeeper of inflammatory gene expression. Trends Biochem Sci 34, 324-331 45 58. Kleeman, T. A., Wei, D., Simpson, K. L., and First, E. A. (1997) Human tyrosyl- tRNA synthetase shares amino acid sequence homology with a putative cytokine. J Biol Chem 272, 14420-14425

59. Wakasugi, K., and Schimmel, P. (1999) Two distinct cytokines released from a human aminoacyl-tRNA synthetase. Science 284, 147-151

60. Wakasugi, K., Slike, B. M., Hood, J., Ewalt, K. L., Cheresh, D. A., and Schimmel, P. (2002) Induction of angiogenesis by a fragment of human tyrosyl- tRNA synthetase. J Biol Chem 277, 20124-20126

61. Greenberg, Y., King, M., Kiosses, W. B., Ewalt, K., Yang, X., Schimmel, P., Reader, J. S., and Tzima, E. (2008) The novel fragment of tyrosyl tRNA synthetase, mini-TyrRS, is secreted to induce an angiogenic response in endothelial cells. FASEB J 22, 1597-1605

62. Fan, X., Patera, A. C., Pong-Kennedy, A., Deno, G., Gonsiorek, W., Manfra, D. J., Vassileva, G., Zeng, M., Jackson, C., Sullivan, L., Sharif-Rodriguez, W., Opdenakker, G., Van Damme, J., Hedrick, J. A., Lundell, D., Lira, S. A., and Hipkin, R. W. (2007) Murine CXCR1 is a functional receptor for GCP-2/CXCL6 and interleukin-8/CXCL8. J Biol Chem 282, 11658-11666

63. Munoz-Chapuli, R., Quesada, A. R., and Angel Medina, M. (2004) Angiogenesis and signal transduction in endothelial cells. Cellular and molecular life sciences : CMLS 61, 2224-2243

64. Fukumura, D., Gohongi, T., Kadambi, A., Izumi, Y., Ang, J., Yun, C. O., Buerk, D. G., Huang, P. L., and Jain, R. K. (2001) Predominant role of endothelial nitric oxide synthase in vascular endothelial growth factor-induced angiogenesis and vascular permeability. Proc Natl Acad Sci U S A 98, 2604-2609

65. Lemaire, G., Dorizzi, M., Spotorno, G., and Labouesse, B. (1969) [Purification of tryptophan tRNA synthetase from beef pancreas]. Bulletin de la Societe de chimie biologique 51, 495-510

66. Beresten, S. F., Zargarova, T. A., Favorova, O. O., Rubikaite, B. I., Ryazanov, A. G., and Kisselev, L. L. (1989) Molecular and cellular studies of tryptophanyl- tRNA synthetase using monoclonal antibodies. Evaluation of a common antigenic determinant in eukaryotic, prokaryotic and archaebacterial enzymes which maps outside the catalytic domain. Eur J Biochem 184, 575-581

67. Wakasugi, K., Slike, B. M., Hood, J., Otani, A., Ewalt, K. L., Friedlander, M., Cheresh, D. A., and Schimmel, P. (2002) A human aminoacyl-tRNA synthetase as a regulator of angiogenesis. Proc Natl Acad Sci U S A 99, 173-177

46 68. Tzima, E., Reader, J. S., Irani-Tehrani, M., Ewalt, K. L., Schwartz, M. A., and Schimmel, P. (2005) VE-cadherin links tRNA synthetase cytokine to anti- angiogenic function. J Biol Chem 280, 2405-2408

69. Tzima, E., Reader, J. S., Irani-Tehrani, M., Ewalt, K. L., Schwartz, M. A., and Schimmel, P. (2003) Biologically active fragment of a human tRNA synthetase inhibits fluid shear stress-activated responses of endothelial cells. Proc Natl Acad Sci U S A 100, 14903-14907

70. Zhou, Q., Kapoor, M., Guo, M., Belani, R., Xu, X., Kiosses, W. B., Hanan, M., Park, C., Armour, E., Do, M. H., Nangle, L. A., Schimmel, P., and Yang, X. L. (2010) Orthogonal use of a human tRNA synthetase active site to achieve multifunctionality. Nature structural & molecular biology 17, 57-61

71. Kapoor, M., Zhou, Q., Otero, F., Myers, C. A., Bates, A., Belani, R., Liu, J., Luo, J. K., Tzima, E., Zhang, D. E., Yang, X. L., and Schimmel, P. (2008) Evidence for annexin II-S100A10 complex and plasmin in mobilization of cytokine activity of human TrpRS. J Biol Chem 283, 2070-2077

72. Renigunta, V., Yuan, H., Zuzarte, M., Rinne, S., Koch, A., Wischmeyer, E., Schlichthorl, G., Gao, Y., Karschin, A., Jacob, R., Schwappach, B., Daut, J., and Preisig-Muller, R. (2006) The retention factor p11 confers an endoplasmic reticulum-localization signal to the potassium channel TASK-1. Traffic 7, 168- 181

73. Jia, J., Arif, A., Ray, P. S., and Fox, P. L. (2008) WHEP domains direct noncanonical function of glutamyl-Prolyl tRNA synthetase in translational control of gene expression. Mol Cell 29, 679-690

74. Shi, Y., Xu, X., Zhang, Q., Fu, G., Mo, Z., Wang, G. S., Kishi, S., and Yang, X. L. (2014) tRNA synthetase counteracts c-Myc to develop functional vasculature. eLife 3, e02349

75. Jin, S. W., Herzog, W., Santoro, M. M., Mitchell, T. S., Frantsve, J., Jungblut, B., Beis, D., Scott, I. C., D'Amico, L. A., Ober, E. A., Verkade, H., Field, H. A., Chi, N. C., Wehman, A. M., Baier, H., and Stainier, D. Y. (2007) A transgene-assisted genetic screen identifies essential regulators of vascular development in vertebrate embryos. Developmental biology 307, 29-42

76. Herzog, W., Muller, K., Huisken, J., and Stainier, D. Y. (2009) Genetic evidence for a noncanonical function of seryl-tRNA synthetase in vascular development. Circ Res 104, 1260-1266

77. Amsterdam, A., Nissen, R. M., Sun, Z., Swindell, E. C., Farrington, S., and Hopkins, N. (2004) Identification of 315 genes essential for early zebrafish development. Proc Natl Acad Sci U S A 101, 12792-12797

47 78. Xu, X., Shi, Y., Zhang, H. M., Swindell, E. C., Marshall, A. G., Guo, M., Kishi, S., and Yang, X. L. (2012) Unique domain appended to vertebrate tRNA synthetase is essential for vascular development. Nature communications 3, 681

79. Grandori, C., Cowley, S. M., James, L. P., and Eisenman, R. N. (2000) The Myc/Max/Mad network and the transcriptional control of cell behavior. Annual review of cell and developmental biology 16, 653-699

80. Mazumder, B., and Fox, P. L. (1999) Delayed translational silencing of ceruloplasmin transcript in gamma interferon-activated U937 monocytic cells: role of the 3' untranslated region. Mol Cell Biol 19, 6898-6905

81. Sampath, P., Mazumder, B., Seshadri, V., and Fox, P. L. (2003) Transcript- selective translational silencing by gamma interferon is directed by a novel structural element in the ceruloplasmin mRNA 3' untranslated region. Mol Cell Biol 23, 1509-1519

82. Mazumder, B., Sampath, P., Seshadri, V., Maitra, R. K., DiCorleto, P. E., and Fox, P. L. (2003) Regulated release of L13a from the 60S ribosomal subunit as a mechanism of transcript-specific translational control. Cell 115, 187-198

83. Sampath, P., Mazumder, B., Seshadri, V., Gerber, C. A., Chavatte, L., Kinter, M., Ting, S. M., Dignam, J. D., Kim, S., Driscoll, D. M., and Fox, P. L. (2004) Noncanonical function of glutamyl-prolyl-tRNA synthetase: gene-specific silencing of translation. Cell 119, 195-208

84. Kapasi, P., Chaudhuri, S., Vyas, K., Baus, D., Komar, A. A., Fox, P. L., Merrick, W. C., and Mazumder, B. (2007) L13a blocks 48S assembly: role of a general initiation factor in mRNA-specific translational control. Mol Cell 25, 113-126

85. Ray, P. S., and Fox, P. L. (2007) A post-transcriptional pathway represses monocyte VEGF-A expression and angiogenic activity. EMBO J 26, 3360-3372

86. Quevillon, S., Agou, F., Robinson, J. C., and Mirande, M. (1997) The p43 component of the mammalian multi-synthetase complex is likely to be the precursor of the endothelial monocyte-activating polypeptide II cytokine. J Biol Chem 272, 32573-32579

87. Shalak, V., Kaminska, M., Mitnacht-Kraus, R., Vandenabeele, P., Clauss, M., and Mirande, M. (2001) The EMAPII cytokine is released from the mammalian multisynthetase complex after cleavage of its p43/proEMAPII component. J Biol Chem 276, 23769-23776

88. Park, S. G., Kang, Y. S., Ahn, Y. H., Lee, S. H., Kim, K. R., Kim, K. W., Koh, G. Y., Ko, Y. G., and Kim, S. (2002) Dose-dependent biphasic activity of tRNA synthetase-associating factor, p43, in angiogenesis. J Biol Chem 277, 45243- 45248 48 89. Berger, A. C., Alexander, H. R., Tang, G., Wu, P. S., Hewitt, S. M., Turner, E., Kruger, E., Figg, W. D., Grove, A., Kohn, E., Stern, D., and Libutti, S. K. (2000) Endothelial monocyte activating polypeptide II induces endothelial cell apoptosis and may inhibit tumor angiogenesis. Microvascular research 60, 70-80

90. Ko, Y. G., Park, H., Kim, T., Lee, J. W., Park, S. G., Seol, W., Kim, J. E., Lee, W. H., Kim, S. H., Park, J. E., and Kim, S. (2001) A cofactor of tRNA synthetase, p43, is secreted to up-regulate proinflammatory genes. J Biol Chem 276, 23028- 23033

91. Kwon, H. S., Park, M. C., Kim, D. G., Cho, K., Park, Y. W., Han, J. M., and Kim, S. (2012) Identification of CD23 as a functional receptor for the proinflammatory cytokine AIMP1/p43. Journal of cell science 125, 4620-4629

92. Kawai, T., and Akira, S. (2006) TLR signaling. Cell Death Differ 13, 816-825

93. Polunovsky, V. A., Wendt, C. H., Ingbar, D. H., Peterson, M. S., and Bitterman, P. B. (1994) Induction of endothelial cell apoptosis by TNF alpha: modulation by inhibitors of protein synthesis. Exp Cell Res 214, 584-594

94. van Horssen, R., Rens, J. A., Schipper, D., Eggermont, A. M., and ten Hagen, T. L. (2006) EMAP-II facilitates TNF-R1 apoptotic signalling in endothelial cells and induces TRADD mobilization. Apoptosis : an international journal on programmed cell death 11, 2137-2145

95. Schwarz, M. A., Zheng, H., Liu, J., Corbett, S., and Schwarz, R. E. (2005) Endothelial-monocyte activating polypeptide II alters fibronectin based endothelial cell adhesion and matrix assembly via alpha5 beta1 integrin. Exp Cell Res 311, 229-239

96. Awasthi, N., Schwarz, M. A., Verma, V., Cappiello, C., and Schwarz, R. E. (2009) Endothelial monocyte activating polypeptide II interferes with VEGF- induced proangiogenic signaling. Laboratory investigation; a journal of technical methods and pathology 89, 38-46

97. Tandle, A. T., Calvani, M., Uranchimeg, B., Zahavi, D., Melillo, G., and Libutti, S. K. (2009) Endothelial monocyte activating polypeptide-II modulates endothelial cell responses by degrading hypoxia-inducible factor-1alpha through interaction with PSMA7, a component of the proteasome. Exp Cell Res 315, 1850-1859

98. Qin, X., Hao, Z., Tian, Q., Zhang, Z., Zhou, C., and Xie, W. (2014) Cocrystal structures of glycyl-tRNA synthetase in complex with tRNA suggest multiple conformational states in glycylation. J Biol Chem 289, 20359-20369

99. Son, J., Lee, E. H., Park, M., Kim, J. H., Kim, J., Kim, S., Jeon, Y. H., and Hwang, K. Y. (2013) Conformational changes in human prolyl-tRNA synthetase 49 upon binding of the substrates proline and ATP and the inhibitor halofuginone. Acta crystallographica. Section D, Biological crystallography 69, 2136-2145

100. Romagnani, P., Lasagni, L., Annunziato, F., Serio, M., and Romagnani, S. (2004) CXC chemokines: the regulatory link between inflammation and angiogenesis. Trends in immunology 25, 201-209

101. Sethi, G., Sung, B., and Aggarwal, B. B. (2008) TNF: a master switch for inflammation to cancer. Frontiers in bioscience : a journal and virtual library 13, 5094-5107

102. Fajardo, L. F., Kwan, H. H., Kowalski, J., Prionas, S. D., and Allison, A. C. (1992) Dual role of tumor necrosis factor-alpha in angiogenesis. The American journal of pathology 140, 539-544

103. Guo, M., Schimmel, P., and Yang, X. L. (2010) Functional expansion of human tRNA synthetases achieved by structural inventions. FEBS Lett 584, 434-442

104. Cahuzac, B., Berthonneau, E., Birlirakis, N., Guittet, E., and Mirande, M. (2000) A recurrent RNA-binding domain is appended to eukaryotic aminoacyl-tRNA synthetases. EMBO J 19, 445-452

105. Guo, M., Yang, X. L., and Schimmel, P. (2010) New functions of aminoacyl- tRNA synthetases beyond translation. Nat Rev Mol Cell Biol 11, 668-674

106. Ray, P. S., and Fox, P. L. (2014) Origin and evolution of glutamyl-prolyl tRNA synthetase WHEP domains reveal evolutionary relationships within Holozoa. PLoS One 9, e98493

107. Bergers, G., and Hanahan, D. (2008) Modes of resistance to anti-angiogenic therapy. Nat Rev Cancer 8, 592-603

108. Rock, F. L., Mao, W., Yaremchuk, A., Tukalo, M., Crepin, T., Zhou, H., Zhang, Y. K., Hernandez, V., Akama, T., Baker, S. J., Plattner, J. J., Shapiro, L., Martinis, S. A., Benkovic, S. J., Cusack, S., and Alley, M. R. (2007) An antifungal agent inhibits an aminoacyl-tRNA synthetase by trapping tRNA in the editing site. Science 316, 1759-1761

109. Sundrud, M. S., Koralov, S. B., Feuerer, M., Calado, D. P., Kozhaya, A. E., Rhule-Smith, A., Lefebvre, R. E., Unutmaz, D., Mazitschek, R., Waldner, H., Whitman, M., Keller, T., and Rao, A. (2009) Halofuginone inhibits TH17 cell differentiation by activating the amino acid starvation response. Science 324, 1334-1338

110. Thomas, C. M., Hothersall, J., Willis, C. L., and Simpson, T. J. (2010) Resistance to and synthesis of the antibiotic mupirocin. Nature reviews. Microbiology 8, 281- 289 50 111. Van de Vijver, P., Ostrowski, T., Sproat, B., Goebels, J., Rutgeerts, O., Van Aerschot, A., Waer, M., and Herdewijn, P. (2008) Aminoacyl-tRNA synthetase inhibitors as potent and synergistic immunosuppressants. Journal of medicinal chemistry 51, 3020-3029

112. Kim, D. G., Lee, J. Y., Kwon, N. H., Fang, P., Zhang, Q., Wang, J., Young, N. L., Guo, M., Cho, H. Y., Mushtaq, A. U., Jeon, Y. H., Choi, J. W., Han, J. M., Kang, H. W., Joo, J. E., Hur, Y., Kang, W., Yang, H., Nam, D. H., Lee, M. S., Lee, J. W., Kim, E. S., Moon, A., Kim, K., Kim, D., Kang, E. J., Moon, Y., Rhee, K. H., Han, B. W., Yang, J. S., Han, G., Yang, W. S., Lee, C., Wang, M. W., and Kim, S. (2014) Chemical inhibition of prometastatic lysyl-tRNA synthetase-laminin receptor interaction. Nat Chem Biol 10, 29-34

113. Dock-Bregeon, A., Sankaranarayanan, R., Romby, P., Caillet, J., Springer, M., Rees, B., Francklyn, C. S., Ehresmann, C., and Moras, D. (2000) Transfer RNA- mediated editing in threonyl-tRNA synthetase. The class II solution to the double discrimination problem. Cell 103, 877-884

114. Torres-Larios, A., Sankaranarayanan, R., Rees, B., Dock-Bregeon, A. C., and Moras, D. (2003) Conformational movements and cooperativity upon amino acid, ATP and tRNA binding in threonyl-tRNA synthetase. J Mol Biol 331, 201-211

115. Paetz, W., and Nass, G. (1973) Biochemical and immunological characterization of threonyl-tRNA synthetase of two borrelidin-resistant mutants of Escherichia coli K12. Eur J Biochem 35, 331-337

116. Gantt, J. S., Bennett, C. A., and Arfin, S. M. (1981) Increased levels of threonyl- tRNA synthetase in a borrelidin-resistant Chinese hamster ovary cell line. Proc Natl Acad Sci U S A 78, 5367-5370

117. Wakabayashi, T., Kageyama, R., Naruse, N., Tsukahara, N., Funahashi, Y., Kitoh, K., and Watanabe, Y. (1997) Borrelidin is an angiogenesis inhibitor; disruption of angiogenic capillary vessels in a rat aorta matrix culture model. J Antibiot (Tokyo) 50, 671-676

118. Funahashi, Y., Wakabayashi, T., Semba, T., Sonoda, J., Kitoh, K., and Yoshimatsu, K. (1999) Establishment of a quantitative mouse dorsal air sac model and its application to evaluate a new angiogenesis inhibitor. Oncol Res 11, 319- 329

119. Labirua, A., and Lundberg, I. E. (2010) Interstitial lung disease and idiopathic inflammatory myopathies: progress and pitfalls. Current opinion in rheumatology 22, 633-638

120. Suber, T. L., Casciola-Rosen, L., and Rosen, A. (2008) Mechanisms of disease: autoantigens as clues to the pathogenesis of myositis. Nature clinical practice. Rheumatology 4, 201-209 51 121. Gorlach, A., and Bonello, S. (2008) The cross-talk between NF-kappaB and HIF- 1: further evidence for a significant liaison. Biochem J 412, e17-19

122. Karin, M., and Greten, F. R. (2005) NF-kappaB: linking inflammation and immunity to cancer development and progression. Nat Rev Immunol 5, 749-759

123. Grundtman, C., Tham, E., Ulfgren, A. K., and Lundberg, I. E. (2008) Vascular endothelial growth factor is highly expressed in muscle tissue of patients with polymyositis and patients with dermatomyositis. Arthritis and rheumatism 58, 3224-3238

124. Hill, C. L., Zhang, Y., Sigurgeirsson, B., Pukkala, E., Mellemkjaer, L., Airio, A., Evans, S. R., and Felson, D. T. (2001) Frequency of specific cancer types in dermatomyositis and polymyositis: a population-based study. Lancet 357, 96-100

52 CHAPTER 2: CHARACTERIZING THE INHIBITION OF HUMAN

THREONYL-tRNA SYNTHETASE BY BORRELIDIN AND DERIVATIVES

2.1 Introduction

Borrelidin is a polyketide, macrolide antibiotic that was first discovered in isolates from Streptomyces rocheii in 1949 and later identified in other Streptomyces species (1-

3). It was named for its potent anti-microbial effects against the fever-inducing spirochetes from the Borrelia genus. However, its functions have expanded greatly since it discovery. In addition to Borrelia, borrelidin was found to inhibit pathogenic spirochetes of genus Treponema, the malarial parasite Plasmodium falciparum, viral replication, and pathogenic fungi (4-7). These effects have been linked to the molecule’s specific inhibition of TARS and, through it, the overall process of translation (8).

TARS exhibits high sequence conservation between bacteria and eukaryotes.

Consequently, borrelidin has been demonstrated to inhibit functions in a variety of organisms. Studies of borrelidin-resistant E. coli K12 strains and Chinese hamster ovary cells reveal modified or increased expression of ThrRS, the only known target of the compound (9,10). These observations support an interaction between TARS and borrelidin and suggests that this interaction is maintained through bacteria to mammals.

Similarly, borrelidin-treatment was demonstrated to influence endothelial cell processes in both human umbilical vein endothelial cells (HUVECs) and mouse models, indirectly insinuating that borrelidin is capable of inhibiting several varieties of mammalian TARS

(11,12). However, these hypotheses remain circumstantial without support from more direct approaches.

53 The interactions between borrelidin and several bacterial ThrRSs were confirmed in assays utilizing isolated enzymes from the representative bacterial sources E. coli, S. solfataricus, and Halobacterium sp. Calculated values of Ki were obtained for both E. coli and S. solfataricus and were found to be nearly identical at 3.7 nM and 4.5 nM respectively. Conversely, representative archaeal species, M. jannaschii and A. fulgidus, exhibited no significant inhibition by borrelidin, consistent with the observed low sequence conservation in archaeal TARS compared to the other domains of life (13).

Despite this suggestive evidence that borrelidin is capable of inhibiting the mammalian

TARS, direct proof of the enzyme-inhibitor interaction has not been reported.

In addition to its inhibition of aminoacylation, borrelidin was found to disrupt angiogenesis in the mouse aortic ring fragment model (12). This feature led to subsequent investigations interested in the development of borrelidin as an anti-metastatic agent.

Unfortunately, the potential therapeutic benefits of borrelidin are limited by its acute toxicity to living cells (14). Efforts have been made to improve the clinical usefulness of borrelidin by generating derivative molecules with a wider therapeutic index. In order to synthesize these derivative molecules, the complete borrelidin biosynthetic operon from

Streptomyces parvelus Tu4055 was cloned and sequenced. By selective gene mutation and “priming” the biosynthesis of borrelidin with alternative dicarboxylic starter compounds novel variants of borrelidin were readily achieved. While this technique could theoretically be used to modify a variety of locations within the borrelidin molecule, modifications to the macrolide backbone and nitrile group completely eliminated the compound’s functionality. In contrast, modifications to the C17 pendent group produced derivative compounds with varying affinities and properties (15,16).

54 Therefore, all borrelidin derivatives used in recent studies vary exclusively at the C17 site.

Figure 5. Structures of macrolides used in this study. Structures of borrelidin (top left), BC194 (top right), BC196 (bottom left), and BC220 (bottom right). Carbons C1 and C17 for the macrolide component of each compound have been labeled.

Semisynthetic generation of borrelidin derivatives has produced a battery of compounds (Figure 5). The most promising of these molecules for use in anti-angiogenic treatments is BC194, which substitutes a cyclobutane carboxylic acid for the C17 cyclopentane carboxylic acid group of borrelidin. Surprisingly, this seemingly minor modification has been demonstrated to reduce the toxicity of the compound by nearly

1000-fold as determined by growth inhibition of human umbilical vein endothelial cells

(HUVEC) (16). Despite this change in toxicity, however, BC194 retains potent anti-

55 angiogenic activity and will be discussed in greater detail within later chapters.

Additionally, the derivatives BC196 and BC220, consisting of 2-methylbutanoic acid and cyclopentane carboxylic acid 3-ethyl piperidine ester C17 substitutions respectively, represent lower affinity derivatives and have been included in these studies as well.

Originally, these derivatives were investigated for their high specificity against the malarial parasite Plasmodium falciparum relative to the human host, however, they are used here to draw structural comparisons between the other compounds (17).

In this work we demonstrate that the C17 modification of BC194 does not significantly alter its affinity for human TARS from that of borrelidin in enzymatic assays. However, in accordance with recent crystallographic data, BC194 is found to be more susceptible than borrelidin to competition by threonine in biological assays. Taking these observations into consideration with the conformational changes presented by the structural data and the slow, tight-binding kinetics of the compound leads to the presentation of a two-step, induced-fit model of binding. This model justifies the apparent incongruence of the competitive structural data with the noncompetitive kinetics and will facilitate future understanding of borrelidin and its derivatives. Finally, point mutations are generated in human TARS in order to develop borrelidin-resistant mutants that can be used to investigate the TARS specificity for borrelidin-related functions in biology.

2.2 Results and Discussion

2.2.1 Discussion of borrelidin’s slow, tight-binding kinetics of enzymes

Inhibitors are often described by an equilibrium dissociation constant (Ki) for the formation of the enzyme-inhibitor complex. The true Ki value accounts for the inhibitor’s

56 modality (ie. competitive, noncompetitive, etc.) and how this affinity is influenced by changes in substrate with respect to the reactions Km. Therefore, this value is independent of the reaction conditions. However, in absence of detailed knowledge of

app inhibitor modality or kinetic constants an apparent Ki value (Ki ) can used for relative comparisons of inhibitor affinity so long as nearly identical conditions are maintained.

app Commonly, calculations for Ki are based on reversible, steady-state measurements of initial velocity (v) and must satisfy several conditions. In particular, these methods assume that the formation and dissociation of the enzyme inhibitor complex (EI) is fast relative to the rate of enzymatic turnover and that the concentration of EI is negligible to the concentration of free inhibitor. However, early investigations of borrelidin binding modality with E. coli ThrRS using intrinsic tryptophan fluorescence

6 -1 -1 exhibited kon and koff values of 5*10 s and <0.1 s respectively. From these data, the lower limit of the TARS-borrelidin complex half-life can be determined by using the following equation (Derivation found in (18)):

0.6931 t1/ 2  koff

-1 By fitting koff at <0.1 s , a half-life of at least 7 minutes is obtained, indicating that the rapid binding assumption is not met. As a notable consequence, the degree of inhibition for these slow-binding inhibitors appears to change over time as demonstrated in Figure

6. In absence of inhibitor and at a concentration of 10 nM E. coli TARS (Ec ThrRS;

ThrRS designates the prokaryotic enzyme), the steady-state aminoacylation reaction proceeds in a linear fashion until it has exhausted its available substrates. Similar isotherms would be predicted for rapid binding inhibitors except that the initial slopes

57 would be reduced as a function of inhibitor concentration. In contrast, reactions containing 20 nM borrelidin exhibit clearly visible time dependence for inhibition. In absence of pre-incubation, the reaction rates appear similar to the uninhibited reaction at earlier time points but level off within 10 minutes. Given that the substrate concentrations are identical in all reactions, this plateau is directly related to inhibitor binding and not the exhaustion of substrates. Conversely, pre-incubation of Ec ThrRS and borrelidin prior to the addition of substrates bypasses the slow-binding step and establishes a linear inhibition-isotherm. The fact that the rate of the pre-incubated reaction does not seem to increase over time suggests the rate of dissociation is so slow as to appear irreversible.

Therefore, Ki values will be determined from initial velocity values of reactions pre- incubated with borrelidin to account for the slow-binding character of the compound.

Figure 6. Borrelidin exhibits slow-binding kinetics. The aminoacylation activity of Ec ThrRS treated with the DMSO vehicle (Uninhibited, blue) or borrelidin for either 30 minutes before (pre-incubation, green) or simultaneously with (No incubation, red) the canonical substrates. Activity was determined by the production of Thr- tRNAThr per active site per minute; mean ± SD, n=3.

58 Initial investigations of borrelidin’s effects on Ec ThrRS revealed significant inhibition when ThrRS and borrelidin were present at similar concentrations, indicating that borrelidin is also a tight-binding inhibitor. This property, like slow-binding, is accompanied by its own set of complications that must be addressed for proper kinetic

app analysis. For Ec ThrRS, the Ki was calculated from the IC50 (the concentration of inhibitor that inhibits 50% of enzymatic activity under the given conditions) under the assumption of excess inhibitor and later adjusted for tight-binding character. Often the

app IC50 value of weak-binding inhibitors is a sufficient approximation for Ki . This value can be determined by plotting the fractional velocity (vi/v0, where vi and v0 are the initial velocities of the inhibited and uninhibited reactions respectively) against the total concentration of inhibitor in the reaction ([I]T) and fitting the resulting data to the following Langmuir isotherm equation (19,20):

v 1 i  v [I] 0 1 T IC50

However, the derivation of this equation depends on the assumption that [I]T is in great excess of the total enzyme concentration ([E]T). That is to say [I]T >> [E]T. Therefore, the concentration of inhibitor in complex with enzyme is negligible compared to the free inhibitor concentration ([I]f) and [I]T can be used as a close approximation for [I]f (ie. [I]T

≈ [I]f). In contrast, borrelidin significantly reduces Ec ThrRS activity at nearly equimolar concentrations to the enzyme, indicating that bound borrelidin represents a significant proportion of the total concentration. To address this concern, early borrelidin studies 59 app converted IC50 into Ki values using an equation that accounts for borrelidin’s tight-

app binding properties by associating IC50, Ki , and [E]T using the following equation derived by Easson and Stedman in 1933 (21):

1 IC  K app  [E] 50 i 2 T

app While this method could also be used for the determination of Ki with human TARS, the work presented in this dissertation utilized the more robust Morrison quadratic equation developed in 1969 (22,23).

v [E]  [I]  K  ([E]  [I]  K ) 2  4[E] [I] i  1 T T i T T i T T v0 2[E]T

Unlike the IC50 method, the Morrison equation requires no assumptions related to the depletion of inhibitor concentration due to complex formation with the enzyme and can

app directly calculate Ki from a plot of fractional velocity versus inhibitor concentration.

2.2.2 Inhibition of human threonyl-tRNA synthetase by borrelidin and derivatives

The ability of borrelidin and related-compounds to inhibit human TARS was investigated kinetically using in vitro enzymatic assays similar to the early methods used with prokaryotic versions of the enzyme. These procedures utilized human tRNAThr and

TARS constructs recombinantly expressed in E. coli cells and monitored the production of Thr-tRNAThr as a measurement of aminoacylation activity. By plotting the relative 60 velocity of the inhibited reaction (vi/v0) versus the concentration of inhibitor, a curve is

app obtained from which Ki can be determined using the Morrison quadratic equation above. Care was taken to maintain conditions as close to those reported for E. coli ThrRS

app to keep comparisons between Ki values as accurate as possible. These procedures were repeated for the various derivatives of borrelidin as well. All data have been included in

Table 2.

app Table 2. Ki values of borrelidin derivatives on TARS constructs Construct Borrelidin BC194 BC196 BC220 E. coli ThrRS 3.7a 0.006b -- -- Human TARS 5.44 4.09 208.2 >10,000 L567V TARS -- 32.2 -- -- a data from Ruan et al. (13) b data from Anand Minajigi dissertation – Note these data have been difficult to replicate. This may be due to the use of mislabeled stock samples and should be repeated.

Recombinant human TARS was found to be active with a rate of 3.325 ± 0.2083 min-1 (Figure 7A). Consistent with its high sequence conservation to Ec ThrRS, this

TARS construct exhibited tight-binding interactions with borrelidin. Likewise, a pre- incubation step between the enzyme and inhibitor was used to account for the time

app dependence of the borrelidin-TARS interaction. Borrelidin inhibited TARS with a Ki of 5.44 nM (Figure 7B) nearly identical to the 3.7 nM value obtained for ThrRS. Despite its significantly reduced toxicity, BC194 also formed a comparably tight interaction with

app app TARS, presenting with a Ki of 4.09 nM (Figure 7C). Conversely, the Ki values for

BC196 (Figure 7D) and BC220 (Figure 7E) were significantly higher at 208.20 nM and

>10 µM respectively.

app Taken together, the Ki values can provide some mechanistic insights on their own. The structural variability for all the tested compounds is limited to modifications in 61

app Figure 7. Determination of Ki for borrelidin and related compounds against human TARS. (A) Aminoacylation activity of purified TARS as determined by formation of Thr-tRNAThr per active site per minute; mean ± S.D.; n=3. (B-E) Plots showing the fractional initial velocity of TARS (10 nM) as a function of borrelidin (B), BC194 (C), BC196 (D), or BC220 (E) concentration; mean ± S.D.; n=3.

62 the pendant side chain, demonstrating the importance of this group to the binding each compound. Specifically, the two strongest binders, borrelidin and BC194, possess closed ring structures; whereas BC196, which resembles BC194 with an open carboxyl-

app cyclobutyl ring, exhibits a significantly higher Ki . Most likely, this difference is due to the higher degrees of freedom present in the open structure compared to the closed.

Conversely, while BC220 possesses a closed cyclopentyl ring like borrelidin, the esterification of the carboxyl group severely reduces the compounds affinity for TARS.

However, it is uncertain as to whether or not this modification reduces binding through steric interactions or through the loss of the negatively charged carboxyl group. In contrast, the macrolide backbone is identical for all compounds and is predicted to make the majority of the compounds’ contact with the enzyme. Therefore, the difference in affinity due to the C17 group may not be enough to counteract the strong contributions of

app the macrolide ring in borrelidin and BC194, explaining the nearly identical Ki values.

Conversely, the variations in the BC196 and BC220 side chains may be disruptive enough to inhibit strong interactions with the enzyme. However, structural information must be considered to truly understand borrelidin’s binding mode.

2.2.3 Determination of borrelidin’s mode of inhibition.

As the high resolution crystal structure of the borrelidin-Ec ThrRS complex was not reported until recently, early investigations of borrelidin’s binding mode were limited to kinetic and mutational data. Naturally occurring mutations that conferred borrelidin resistance were found to correspond to a hydrophobic pocket adjacent to the Ec ThrRS active site implicating this region as the likely borrelidin binding site (13). This was not

63 completely surprising as the non-polar macrolide ring would form higher affinity interactions with hydrophobic residues. Other studies also demonstrated that the binding of borrelidin induced a dose-dependent quenching of an intrinsically fluorescencent tryptophan residue (13). Previously, this fluorescence had been associated with Trp434 and Trp478 in the E. coli ThrRS (Trp512 and H556 in humans) positioned near the active site and were quenched in response to conformational changes induced by the binding of

ATP and threonine (24). Together these data would likewise suggest that borrelidin induces conformational changes in Ec ThrRS upon binding. More importantly, however, the phenomenon occurred in absence of the canonical substrates and eliminated the possibility of borrelidin binding using an uncompetitive mode. Furthermore, despite this location near the active site, borrelidin’s inhibitory properties were unaffected by the addition of threonine and ATP ranging in concentrations from 0.01 to 1 mM and 0.02 to

10 mM respectively, suggesting that borrelidin does not possess a competitive or mixed- type inhibition mode as well (13). Therefore, these early reports suggested that borrelidin must bind noncompetitively. However, the true modality of borrelidin is more complicated than originally predicated owing to recent crystallographic data of the

TARS-borrelidin complex solved by our collaborator Min Guo (25).

Recently, the structure of borrelidin co-crystallized with a fragment of human

TARS (ΔN1N2-TARS) comprising the aminoacylation and anticodon-binding domains

(residues 318-723) was solved to a resolution of 2.6 Å. The resulting structure challenges the conclusion of a noncompetitive binding mode suggested by the kinetic data.

Consistent with the previous investigations, borrelidin was found to bind within the hydrophobic pocket of the aminoacylation domain active site of the protein. Specifically,

Figure 8. TARS-borrelidin binding interactions. (A, B) Residues involved in the borrelidin-TARS interaction for the macrolide (A) and pendent group (B) components. The borrelidin structure is shown in yellow and binding resides are shown in magenta. Structure provided by Min Guo. the macrolide component forms hydrophobic interactions with residues L567, S386,

H388, Y540, D564, Q562, H590, Y392, H391, and F539 (human numbering; Figure 8A).

The pendent side chain further contributes interactions with residues T560, R442, M411,

65 C413, Q460, and A592 (Figure 8B). This large number of contacts likely explains the tight-binding character of the compound. Establishing all of these interactions, however, requires structural changes within the enzyme as demonstrated by comparison of the borrelidin-TARS complex to other TARS structures.

The borrelidin induced conformational changes can be easily observed when compared to the structures of earlier Ec ThrRS complexes. These Ec ThrRS structures are either empty (apo Ec ThrRS; PDB: 1EVK) or in complex with the threonyl-adenylate analog, threonyl sulfamoyl adenylate (PDB: 1NYQ), or AMP (PDB: 1QF6). Although the borrelidin-TARS complex described above refers to the human enzyme, the structure for the aminoacylation and anti-codon binding domains of the Ec ThrRS ortholog was also solved. Both structures superimpose with an r.m.s.d (root mean square distance) of

0.644 Å.for all 279 Cα carbons, indicating that the overall folds are highly similar.

Superposition of E. coli ΔN1N2-ThrRS with the various substrate bound structures revealed nearly identical conformations with the exception of one region adjacent to the aminoacylation active site (Figure 9A). As demonstrated by the threonine and AMP bound structures, substrate binding induces the closing of a specific helix and loop

(residues 492-547 for humans and 415-469 in E. coli) into the active site as has been previously observed (26). By comparison, borrelidin binding extrudes this helix by 14° relative to the substrate bound conformations and opens the TARS active site even more than the apo structure. These structural changes are consistent with the earlier intrinsic tryptophan fluorescence data for both borrelidin and substrate binding mentioned above since Trp434 (Trp512 in humans) is located within this region. In addition to these conformational changes, superposition of the borrelidin-ThrRS structure with the other

66 complexes also revealed substantial overlap of borrelidin with all three substrate binding pockets, suggesting that borrelidin functions as a competitive inhibitor (Figure 9B, C).

This result is surprising, since borrelidin has been previously reported as a non- competitive inhibitor. Nonetheless, this observation can be explained.

The structural evidence suggesting that borrelidin binds in a competitive fashion seems to contradict earlier kinetic data indicating a non-competitive binding mode (13).

However, the misinterpretation of a competitive-binding mode as non-competitive is common with slow, tight-binding inhibitors. Several mechanisms can contribute to the slow-binding of a compound (reviewed in (18)). In the case of small molecule inhibitors, these scenarios typically fit one of three descriptions: 1) the binding of the compound is inherently slow; 2) the compound binds only one isomer of an enzyme that slowly isomerizes in solution; or 3) the compound forms a weak initial complex and subsequently induces a slow conformational change in the enzyme for full accommodation. The dose-dependent effect of borrelidin, ATP, or threonine on the intrinsic tryptophan fluorescence of E. coli ThrRS as well as the observed structural changes within TARS upon the binding of borrelidin or substrates clearly indicate that conformational changes are induced upon substrate or compound binding (13,24).

Therefore, model three appears to be the most relevant.

This model would suggest that borrelidin and related compounds only weakly associate with TARS upon first encounter. In this initial phase, binding occurs under rapid equilibrium conditions and is competitive with respect to substrates. Subsequently, the enzyme undergoes a slow isomerization into a final, tightly-bound complex. This new conformation precludes the binding of substrate molecules and would appear to behave

Figure 9. Borrelidin induces conformational changes in TARS and disrupts substrate binding. (A) Superposition of the ΔN1N2-ThrRS-borrelidin complex (yellow) with apo- (cyan; PDB:1EVK), threonine sulfamoyl adenylate (AMS-Thr; magenta; PDB: 1NYQ), and AMP (PDB:1QF6) Ec ThrRS structures with focus on the substrate binding helix and loop (residues 492-547 for humans and 415-469 in E. coli). (B) Superposition of the ΔN1N2-TARS-borrelidin and Ec ThrRS-AMS-Thr (PDB: 1NYQ) complexes revealing steric clashing between borrelidin (green) and AMS-Thr (magenta). The true space occupancy of the borrelidin side chain is indicated by dots. (C) Superposition of the ΔN1N2-TARS-borrelidin and Ec ThrRS- A76- tRNAThr (PDB: 1QF6) complexes revealing steric clashing between borrelidin (green) and A76- of tRNAThr (magenta). Borrelidin’s true space occupancy is indicated by dots.

68 app non-competitively (27). Owing to the pre-incubation step used in the Ki determination experiments described above and in previous reports, TARS-borrelidin complexes would be predominantly in the second binding phase at the start of these reactions. Effectively, once in this tightly-bound phase, borrelidin titrates away free TARS, resulting in an apparent reduction of [E]T (18). Kinetically, a decrease in [E]T manifests as a lower value for Vmax. However, [E]T is unrelated to substrate binding affinity and would not change the value of Km for any of the substrates. Classically, inhibitors that reduce Vmax without affecting the Km would be characterized kinetically as non-competitive. Therefore, borrelidin appears to be a competitive inhibitor that exhibits non-competitive characteristics.

2.2.4 Characterizing the differences in borrelidin and BC194 binding

As mentioned above, the great toxicity of borrelidin to living cells led to the development of less toxic derivatives, the most successful of which appears to be BC194.

Previous reports clearly indicate that BC194 exhibits significantly reduced toxicity when compared to borrelidin (16). The physiological cause of this toxicity is linked to the initiation of amino acid starvation responses in response to the compounds’ inhibition of

TARS aminoacylation function and will be addressed in more detail in Chapter 3.

Interestingly, borrelidin toxicity can be rescued by point mutations in the TARS enzyme, suggesting that the observed effects of the compounds are mediated exclusively through

TARS (13). This suggests two different hypotheses to account for the differences in toxicity between borrelidin and BC194: 1) the two compounds induce distinct conformational changes, only one of which is associated with the toxic effects; or 2) the

69 compounds have altered interactions with TARS, allowing for varying degrees of competition with the canonical substrates. Crystallographic data would eventually confirm the second model.

Recently the structure of the human ΔN1N2-TARS fragment was co-crystallized with BC194 and solved by out collaborator, Min Guo, to a resolution of 2.8 Å.

Superposition of the borrelidin and BC194 TARS complexes reveals nearly identical structure (0.62 Å r.m.s.d.) over 402 Cα’s including the retention of all macrolide interactions (Figure 10A, B). In a global structural sense, borrelidin and BC194 act to stabilize the same conformational state for TARS, which indicates that conformational changes are unlikely to be responsible for variance in toxicity between the two compounds. The loss of the methylene group from BC194’s C17 cyclobutane group lengthens the interaction between the C17 carboxylate and the amide nitrogen of Q460 by

0.9 Å relative to the corresponding interaction in the borrelidin complex (Figure 10C).

Likewise, the Van der Waals interactions between the pendant ring and A592 are completely lost in the BC194 structure. Interestingly, examination of the Ec ThrRS structure co-crystallized with threonine reveals that Q460 and A592 make key hydrogen- bond and hydrophobic interactions with L-threonine (Figure 10D). The loss of these interactions in BC194 may render its interaction with TARS more susceptible to competition with threonine and permit translation to continue over wider range of concentrations. An important prediction of these structural studies is that borrelidin and

BC194 should exhibit differential sensitivity with respect to the inhibition of overall translation. This hypothesis was tested by comparing the formation of luciferase in a cell free protein synthesis system at varying concentrations of borrelidin and BC194 and

70 app app determining the Ki (Figure 10E, F). Borrelidin exhibited a Ki of 100 nM, which is more than half the value of 209 nM recorded with BC194. Therefore, BC194 appears to be more susceptible to competition with threonine, allowing for translation to occur over a broader range of the compound.

The structural and biochemical differences between borrelidin and BC194

app discussed above appear to be in striking contrast to its nearly identical Ki obtained for both compounds in enzymatic assays (Section 2.2.2). However, the two-phase binding model described in Section 2.2.3 provides an explanation. Once again, the formation of the initial encounter-complex in the first phase of binding occurs under rapid equilibrium conditions and would be readily susceptible to competition by the canonical substrates

(27). The C17 side-chain likely contributes to the formation of this initial complex and would be highly responsive to the structural differences between borreldin and BC194.

This would also hold true with BC196 and BC220 as well since these molecules differ even further from borrelidin than BC194. Isomerization of the enzyme then occurs, creating a complex with an extremely low off rate that is substantially more resistant to

app substrate competition. This model also accounts for reported Ki values since TARS is pre-incubated with the inhibitors prior to initiating the reaction in these enzymatic assays and would be predominantly tied up in the stable, tight-binding mode (13,17,28). Since this tight-binding phase consists primarily of interactions with the macrolide backbone the affinities of borrelidin and BC194 would likely be identical at this stage. Conversely, pre-incubation is not practical in the more complex biological approaches, such as in vitro translation or cellular based assays, due to constant availability of substrates or translation of new TARS. Therefore, a significant population of TARS would remain in

Figure 10. Structure of TARS-BC194 complex. (A) Structure superimposition of TARS-BC194 and TARS-borrelidin complexes. The protein is shown in ribbon cartoon representation, and the bound BC194 and borrelidin are shown as pink and blue sticks, respectively. (B) Two-dimensional scheme of TARS-BC194 interactions. H- bonding residues are shown as sticks. Hydrophobic interacting residues are shown in grey. (C) Close up view of BC194 and borrelidin binding site residues. The five shared H-bonds are shown as black dash lines. The 2 borrelidin-specific interactions are shown as blue dash lines, while the corresponding distances in BC194 structure are indicated in pink. (D) Close up view of threonine binding interactions. Interactions are shown as dashed lines. (E, F) The effects of borrrelidin and BC194 treatment on a cell-free translation system. Rabbit reticulocyte lysate (RRL) was incubated with 0.02 mg/ml luciferase mRNA and translation of luciferase enzyme was quantified in a luminescence assay. Serial diluted BC194 (E) and borrelidin (F) (2.5 nM - 25 µM) was added to inhibit the translation of luciferase mRNA; mean ± SEM, n=3. Data in parts (E) and (F) were kindly provided by Min Guo, Scripps, Fl. 72 the rapid-equilibrium phase and be subject to competition by threonine. Under these conditions, borrelidin’s greater affinity would facilitate the progression of the TARS- borrelidin complexes to the tight-binding phase relative to the BC194-containing complexes.

2.2.5 Development of a Borrelidin-resistant point mutation

Early studies of borrelidin inhibition identified several naturally occurring point mutations in E. coli ThrRS that conferred varying degrees of resistance to the compound

(13). Since the majority of borrelidin-binding interactions are maintained from E. coli to humans it seems reasonable that similar resistance-conferring mutations could be generated in the human enzyme. Of particular interest, the successful generation of these mutations would be useful in subsequent investigations of borrelidin’s anti-angiogenic mechanism. At the beginning of this dissertation work TARS was not thought to have a role in angiogenesis; as such borrelidin’s anti-angiogenic properties were thought to be possibly mediated through secondary targets of the compound, namely the splicing associated factor FBP21 and the cyclin dependent kinase Cdc28/Cln2 (29-31). However, the effects for both compounds were observed in cells at low μM concentrations (0.5-5

μM for FBP21 and about 10-50 μM for Cdc/Cln2), which is well above low nM toxicity threshold observed for the compound and seem to be unlikely candidates for the anti- angiogenic effect of the compound (16). Therefore, if a point mutation in TARS were to confer resistance to borrelidin’s effects it would confirm the TARS dependent nature of the interaction.

73 In early efforts to identify mutations that confer borrelidin resistance, libraries of randomly mutagenized E. coli were generated and screened for colonies resistant to borrelidin treatment. Subsequent analyses revealed that the majority of these mutations were associated with L489M substitution in the Ec ThrRS gene (thrS). Similar substitutions were also generated using site-directed mutagenesis in order to further investigate the effects of mutations within the borrelidin binding site of TARS. Notably, the conversion of L489 to Trp conferred the greatest resistance to borrelidin, increasing the Ki value nearly 1500-fold relative to the wildtype enzyme. Using a combination of structural and sequence alignments L489 was found to be conserved as Leu567 in the human enzyme. Structural data indicate that this leucine residue forms Van der Waals contacts with the C12 nitrile of the borrelidin compound in both human and E. coli enzymes (Figure 11A), suggesting that mutations at this site should produce similar effects related to borrelidin binding (25). Therefore, site-directed mutagenesis was used to generate the equivalent L567W mutation as well as the more conservative L567V.

app In order to verify the resistance provided by the point mutations, the Ki values were determined. Initial kinetic characterizations revealed that both mutations possessed aminoacylation activity, albeit at very different rates. L567V TARS turned over at a rate of 4.429 min-1, close to the 3.325 min-1 observed for the wildtype enzyme (Figure 11B).

app Using the same method as the wildtype TARS above, L567V TARS possessed a Ki of

32.2 nM for BC194 (Figure 11C), approximately 8-fold higher than the 4.09 nM obtained for the wildtype enzyme (Table 2). Conversely, L567W TARS catalyzed aminoacylation at a rate of 0.2970 min-1. Owing to this much slower rate, L567W TARS activity data

app Figure 11. L567V TARS exhibits a Ki value greater than WT TARS. (A) Close up of the L567 and C12 nitrile interaction in the borrelidin (yellow)-human TARS complex. (B) Aminoacylation activity of purified WT TARS (blue), L567V TARS (red), and L567W TARS (green) as determined by formation of Thr-tRNAThr per active site per minute; mean ± S.D.; n=3. (C) Plots showing the fractional initial velocity of WT TARS (blue) and L567V TARS (red) (10 nM) as a function of BC194 concentration; mean ± S.D.; n=3.

were difficult to distinguish from background radioactivity counts after the addition of

app BC194 and was not used for subsequent Ki investigations.

2.3 Experimental Procedures

2.3.1 Threonyl-tRNA Synthetase Preparation

N-terminal His6-tagged human TARS was expressed and purified from E. coli

RossettaTM 2(DE3)pLysS competent cells (EMD) transformed with the plasmid pET28a hctThrRS (provided by Dieter Söll, Yale). L567V TARS was generated using identical procedures following QuickChange (Stratagene) site-directed mutagenesis with olio- deoxynucleotide primers from Operon. Transformant cultures were grown in terrific broth supplemented with 100 mg/ml kanamycin and 100 mg/ml chloramphenicol at 37°C to a cell density of A600 = 0.6. Expression of HsThrRS was induced with 1 mM isopropyl

1-thio-β-D-galactoside overnight at 15°C. The bacterial pellet was lysed by sonication in sonication buffer (20 mM potassium phosphate buffer pH 8.0, 100 mM KCl, 35 mM imidazole, and 5 mM β-mercaptoethanol) and cleared by centrifugation at 17050 x g for

30 minutes. Nucleic acids were precipitated by the addition of protamine sulfate to a final concentration of 0.3% followed be centrifugation. The supernatant was loaded onto a

HisTrapTM FF column (GE Healthcare) equilibrated with sonication buffer and eluted by an imidazole gradient of 35-250 mM over 20 column volumes. TARS containing fractions were identified by SDS-PAGE and GelCodeTM Blue (Thermo Scientific), pooled, and dialyzed into hydroxyapatite binding buffer (100 mM potassium phosphate buffer pH 6.8 and 5 mM β-mercaptoethanol). The sample was loaded onto a CHT-

Tricorn hydroxyapatite column and eluted over 20 column volumes by using a gradient of hydroxyapatite binding buffer to 500 mM potassium phosphate pH 8.0 and 5 mM β- mercaptoethanol. ThrRS containing fractions were determined by SDS-PAGE, dialyzed

76 into 20 mM HEPES pH 8.0, 100 mM KCl, 5 mM β-mercaptoethanol, and 40% glycerol, and stored at -20°C. The concentration of active sites was determined as previously described (32).

Full-length, N-terminal 6xHis-tagged E. coli ThrRS (Ec ThrRS) was expressed and purified from IBPC 6881 E. coli cells containing the pTetthrS plasmid through identical procedures to those used for human TARS with minor exceptions (33).

Kanamycin and chloramphenicol were replaced by 10 μg/ml tetracycline. The protein was purified using only the HisTrapTM FF column (GE Healthcare).

For crystallization, the N-terminal His-tagged human ΔN1N2-TARS (residues

322-723) was constructed in a PET28a vector. The protein was expressed in strain BL21

(DE3) and induced with 0.2 mM IPTG for 20 h at 16 °C. The cell pellet (from 4-8 liters) was lysed in NTA-wash buffer (500 mM NaCl, 20 mM Tris-HCl pH 8.0, 30 mM imidazole), loaded onto a Ni-HiTrap column and washed with NTA-wash buffer. The protein was eluted with NTA-elution buffer (500 mM NaCl, 20 mM Tris-HCl pH 8.0,

250 mM imidazole). The eluted protein was further purified by a QAE anion exchange column with NaCl gradient. The peak fractions of the protein were then concentrated for crystallization.

2.3.2 tRNAThr Preparation

E. coli XL-1 Blue (Stratagene) cells expressing the plasmid pUC18hcttRNAThr-TGT

(provided by Dieter Söll, Yale), coding for human tRNAThr, were cultured in LB broth supplemented with 100 mg/ml ampicillin at 37°C to a cell density of A600 = 0.5.

Expression of tRNAThr was induced with 1 mM isopropyl 1-thio-β-D-galactoside for 14

77 hours. Bacterial pellets were resuspended in 20 mM Tris pH 7.5, 20 mM MgCl2, and 10 mM β-mercaptoethanol and extracted with water saturated phenol (pH = 4.8). The aqueous fractions were ethanol precipitated, resuspended in 10 mM HEPES and 50% formamide, and resolved using 8M Urea/10% polyacrylamide gel electrophoresis. The tRNA was identified by UV shadowing, electroeluted, and ethanol precipitated. The tRNA pellet was resuspended in 10 mM HEPES and stored at -20°C. The active concentration of tRNAThr was determined by plateau charging assay (32).

2.3.3 Slow-Tight Binding Assay

For slow-tight binding assays reaction mixtures consisted of 1 mM ATP, 2 U/µl pyrophosphatase (New England Biosciences), 80 µM threonine, 20 µM [14C]-threonine, and 10 µM of human tRNAThr in Buffer A. In addition untreated reactions and non- incubated borrelidin reactions contained DMSO or 10 μM borrelidin respectively, with the concentration of DMSO normalized to the amount present in borrelidin reactions

(about 0.00001%). These reactions were initiated with the addition of 2.5 nM Ec ThrRS and run at 37°C. For pre-incubated borrelidin reactions, 175 nM Ec ThrRS was incubated with 700 nM borrelidin for 30 minutes on ice. The reaction was then initiated by adding 1 μl sample of the Ec ThrRS-borrelidin to 69 μl of reaction buffer, resulting in a 70-fold dilution. Aliquots were taken at varying time points and spotted onto

Whatmann 3MM paper filters pre-soaked in 5% trichloroacetic acid (TCA). Filters were washed 3 times in excess TCA, once in 95% ethanol, and dried under a heating lamp. The formation of Thr-tRNAThr was detected by scintillation counter and the activity determined by linear regression of threonyl-tRNAThr formed per active site per unit time.

2.3.4 Steady-State Aminoacylation Assay

The aminoacylation activities of the various TARS constructs were determined using modifications to established steady state aminoacylation procedures (32). Briefly, reaction mixtures consisted of 2 mM ATP, 2 U/µl pyrophosphatase (New England

Biosciences), 160 µM threonine, 40 µM [14C]-threonine, and 10 µM of human tRNAThr in Buffer A (20 mM HEPES pH 8.0, 100 mM KCl, 10 mM MgCl2, 5 mM β- mercaptoethanol). Reactions were initiated with the addition of 100 nM TARS and run at

37°C. Aliquots were taken at varying time points and spotted onto Whatmann 3MM paper filters pre-soaked in 5% trichloroacetic acid (TCA). Filters were washed 3 times in excess TCA, once in 95% ethanol, and dried under a heating lamp. The formation of Thr- tRNAThr was detected by scintillation counter and the activity determined by linear regression of threonyl-tRNAThr formed per active site per unit time.

app 2.3.5 Ki Determination Assays

app For Ki determination, reaction mixtures consisting of 2 mM ATP, 10 µM tRNA, 10 nM

TARS and various concentrations of BC194 (0 - 35 nM) in Buffer A were incubated for 3 minutes at 37°C and initiated with the addition of radio-labeled threonine (20 µM [14C]- threonine and 180 µM threonine). Data collection procedures were otherwise identical to

app those described for steady-state aminoacylation. Values for Ki were obtained by plotting fractional velocity as a function of BC194 concentration and fitting the data to

Morrison’s quadratic equation for tight binding inhibitors (18):

79 v [E]  [I]  K  ([E]  [I]  K ) 2  4[E] [I] i  1 T T i T T i T T v0 2[E]T

2.3.6 Crystallization and Structure Determination.

Crystallization was done by the sitting drop method. To crystallize human

TARS – BC194 complex, protein solution (10 mg/ mL) was pre-mixed with 2 mM

BC194 and 5mM L-Thr at 4 °C. The protein was then crystallized by mixing 0.5 μL of protein solution with 0.5 μL of precipitant solution, containing 0.1 M Tris pH 8.0, and

8% PEG8000. After incubation at 18 °C for 3-7 days, the crystals were flash-frozen in liquid nitrogen for data collection with a cryo solution consisting of 0.08M calcium acetate, 6.4% PEG4000, and 20% glycerol. Diffraction data were obtained from the Ls- cat 21-ID-G beamline at the Advanced Photon Source (APS, Argonne, IL).

A dataset was integrated and scaled with HKL2000 (34). The structures were determined by molecular replacement based on the human TARS structure (pdb: 4P3N) in program Molrep (35). After corrections for bulk solvent and overall B values, data were refined by iterative cycles of positional refinement and TLS refinement with

PHENIX (36) and model building with COOT (37). All current models have good geometry and no residues are in the disallowed region of the Ramachandran plot.

2.3.7 In Vitro Translation Assay

The effects of BC194 and borrelidin on cell-free protein translation were assayed in a rabbit reticulocyte lysate (RRL) according to the manufacturer’s instructions

(Promega), with the exceptions that no extra amino acid mix was added and 0.02 mg/ml

80 firefly luciferase mRNA was used in the assays. Experiments were performed in triplicate and fit to the Gaddum/Schild EC50 shift equation in order to determine the IC50 value:

Top  Bottom Activity  Bottom 110(LogIC50X )*HillSlope

81 2.4 References

1. Berger, J., Jampolsky, L. M., and Goldberg, M. W. (1949) Borrelidin, a new antibiotic with antiborrelia activity and penicillin enhancement properties. Arch Biochem 22, 476-478

2. Lumb, M., Macey, P. E., Spyvee, J., Whitmarsh, J. M., and Wright, R. D. (1965) Isolation of vivomycin and borrelidin, two antibiotics with anti-viral activity, from a species of Streptomyces (C2989). Nature 206, 263-265

3. Maehr, H., and Evans, R. H. (1987) Identity of borrelidin with treponemycin. J Antibiot (Tokyo) 40, 1455-1456

4. Dickinson, L., Griffiths, A. J., Mason, C. G., and Mills, R. F. (1965) Anti-viral activity of two antibiotics isolated from a species of Streptomyces. Nature 206, 265-268

5. Gao, Y. M., Wang, X. J., Zhang, J., Li, M., Liu, C. X., An, J., Jiang, L., and Xiang, W. S. (2012) Borrelidin, a potent antifungal agent: insight into the antifungal mechanism against Phytophthora sojae. Journal of agricultural and food chemistry 60, 9874-9881

6. Otoguro, K., Ui, H., Ishiyama, A., Kobayashi, M., Togashi, H., Takahashi, Y., Masuma, R., Tanaka, H., Tomoda, H., Yamada, H., and Omura, S. (2003) In vitro and in vivo antimalarial activities of a non-glycosidic 18-membered macrolide antibiotic, borrelidin, against drug-resistant strains of Plasmodia. J Antibiot (Tokyo) 56, 727-729

7. Singh, S. K., Gurusiddaiah, S., and Whalen, J. W. (1985) Treponemycin, a nitrile antibiotic active against Treponema hyodysenteriae. Antimicrobial agents and chemotherapy 27, 239-245

8. Hutter, R., Poralla, K., Zachau, H. G., and Zahner, H. (1966) [Metabolic products of microorganisms. 5l. On the mechanism of action of borrelidin-inhibition of the threonine incorporation in sRNA]. Biochem Z 344, 190-196

9. Gantt, J. S., Bennett, C. A., and Arfin, S. M. (1981) Increased levels of threonyl- tRNA synthetase in a borrelidin-resistant Chinese hamster ovary cell line. Proc Natl Acad Sci U S A 78, 5367-5370

10. Paetz, W., and Nass, G. (1973) Biochemical and immunological characterization of threonyl-tRNA synthetase of two borrelidin-resistant mutants of Escherichia coli K12. Eur J Biochem 35, 331-337

11. Funahashi, Y., Wakabayashi, T., Semba, T., Sonoda, J., Kitoh, K., and Yoshimatsu, K. (1999) Establishment of a quantitative mouse dorsal air sac model 82 and its application to evaluate a new angiogenesis inhibitor. Oncol Res 11, 319- 329

12. Wakabayashi, T., Kageyama, R., Naruse, N., Tsukahara, N., Funahashi, Y., Kitoh, K., and Watanabe, Y. (1997) Borrelidin is an angiogenesis inhibitor; disruption of angiogenic capillary vessels in a rat aorta matrix culture model. J Antibiot (Tokyo) 50, 671-676

13. Ruan, B., Bovee, M. L., Sacher, M., Stathopoulos, C., Poralla, K., Francklyn, C. S., and Soll, D. (2005) A unique hydrophobic cluster near the active site contributes to differences in borrelidin inhibition among threonyl-tRNA synthetases. J Biol Chem 280, 571-577

14. Gafiuc, D., Weiss, M., Mylonas, I., and Bruning, A. (2013) Borrelidin has limited anti-cancer effects in bcl-2 overexpressing breast cancer and leukemia cells and reveals toxicity in non-malignant breast epithelial cells. J Appl Toxicol

15. Moss, S. J., Carletti, I., Olano, C., Sheridan, R. M., Ward, M., Math, V., Nur, E. A. M., Brana, A. F., Zhang, M. Q., Leadlay, P. F., Mendez, C., Salas, J. A., and Wilkinson, B. (2006) Biosynthesis of the angiogenesis inhibitor borrelidin: directed biosynthesis of novel analogues. Chem Commun (Camb), 2341-2343

16. Wilkinson, B., Gregory, M. A., Moss, S. J., Carletti, I., Sheridan, R. M., Kaja, A., Ward, M., Olano, C., Mendez, C., Salas, J. A., Leadlay, P. F., vanGinckel, R., and Zhang, M. Q. (2006) Separation of anti-angiogenic and cytotoxic activities of borrelidin by modification at the C17 side chain. Bioorg Med Chem Lett 16, 5814- 5817

17. Novoa, E. M., Camacho, N., Tor, A., Wilkinson, B., Moss, S., Marin-Garcia, P., Azcarate, I. G., Bautista, J. M., Mirando, A. C., Francklyn, C. S., Varon, S., Royo, M., Cortes, A., and Ribas de Pouplana, L. (2014) Analogs of natural aminoacyl- tRNA synthetase inhibitors clear malaria in vivo. Proc Natl Acad Sci U S A 111, E5508-5517

18. Copeland, R. A. (2005) Evaluation of enzyme inhibitors in drug discovery : a guide for medicinal chemists and pharmacologists, Wiley-Interscience, Hoboken, N.J.

19. Clark, A. J. (1937) General pharmacology, handbook of experimental pharmacology, Springer Verlag, Berlin

20. Langmuir, I. (1916) The constituion and fundamental properties of solids and liquids. J Am Chem Soc 38, 2221-2295

21. Easson, L. H., and Stedman, E. (1936) The Absolute Activity of Choline-Esterase. Proc R Soc London Series B 121, 142-164

83 22. Morrison, J. F., and Walsh, C. T. (2006) The Behavior and Significance of Slow- Binding Enzyme Inhibitors. in Advances in Enzymology and Related Areas of Molecular Biology, John Wiley & Sons, Inc. pp 201-301

23. Williams, J. W., and Morrison, J. F. (1979) The kinetics of reversible tight- binding inhibition. Methods in enzymology 63, 437-467

24. Bovee, M. L., Pierce, M. A., and Francklyn, C. S. (2003) Induced fit and kinetic mechanism of adenylation catalyzed by Escherichia coli threonyl-tRNA synthetase. Biochemistry 42, 15102-15113

25. Fang, P., Yu, X., Jae Jeong, S., Mirando, A. C., Chen, K., Chen, X., Sunghoon, K., Francklyn, C. S., and Guo, M. Structural basis for full-spectrum inhibition of translational functions on a tRNA synthetase. Nature communications In press, 1- 11

26. Torres-Larios, A., Sankaranarayanan, R., Rees, B., Dock-Bregeon, A. C., and Moras, D. (2003) Conformational movements and cooperativity upon amino acid, ATP and tRNA binding in threonyl-tRNA synthetase. J Mol Biol 331, 201-211

27. Copeland, R. A. (2013) Evaluation of enzyme inhibitors in drug discovery : a guide for medicinal chemists and pharmacologists, 2nd ed., Wiley, Hoboken, N.J.

28. Williams, T. F., Mirando, A. C., Wilkinson, B., Francklyn, C. S., and Lounsbury, K. M. (2013) Secreted Threonyl-tRNA synthetase stimulates endothelial cell migration and angiogenesis. Sci Rep 3, 1317

29. Kawamura, T., Liu, D., Towle, M. J., Kageyama, R., Tsukahara, N., Wakabayashi, T., and Littlefield, B. A. (2003) Anti-angiogenesis effects of borrelidin are mediated through distinct pathways: threonyl-tRNA synthetase and caspases are independently involved in suppression of proliferation and induction of apoptosis in endothelial cells. J Antibiot (Tokyo) 56, 709-715

30. Tsuchiya, E., Yukawa, M., Miyakawa, T., Kimura, K. I., and Takahashi, H. (2001) Borrelidin inhibits a cyclin-dependent kinase (CDK), Cdc28/Cln2, of Saccharomyces cerevisiae. J Antibiot (Tokyo) 54, 84-90

31. Woolard, J., Vousden, W., Moss, S. J., Krishnakumar, A., Gammons, M. V., Nowak, D. G., Dixon, N., Micklefield, J., Spannhoff, A., Bedford, M. T., Gregory, M. A., Martin, C. J., Leadlay, P. F., Zhang, M. Q., Harper, S. J., Bates, D. O., and Wilkinson, B. (2011) Borrelidin modulates the alternative splicing of VEGF in favour of anti-angiogenic isoforms. Chemical science 2011, 273-278

32. Francklyn, C. S., First, E. A., Perona, J. J., and Hou, Y. M. (2008) Methods for kinetic and thermodynamic analysis of aminoacyl-tRNA synthetases. Methods 44, 100-118

84 33. Minajigi, A., and Francklyn, C. S. (2008) RNA-assisted catalysis in a protein enzyme: The 2'-hydroxyl of tRNA(Thr) A76 promotes aminoacylation by threonyl-tRNA synthetase. Proc Natl Acad Sci U S A 105, 17748-17753

34. Otwinowski, Z., and Minor, W. (1997) Processing of X-ray Diffraction Data Collected in Oscillation Mode. Methods in Enzymology 276, 307-326

35. Vagin, A., and Teplyakov, A. (2010) Molecular replacement with MOLREP. Acta Crystallogr D Biol Crystallogr 66, 22-25

36. Adams, P. D., Afonine, P. V., Bunkoczi, G., Chen, V. B., Davis, I. W., Echols, N., Headd, J. J., Hung, L. W., Kapral, G. J., Grosse-Kunstleve, R. W., McCoy, A. J., Moriarty, N. W., Oeffner, R., Read, R. J., Richardson, D. C., Richardson, J. S., Terwilliger, T. C., and Zwart, P. H. (2010) PHENIX: a comprehensive Python- based system for macromolecular structure solution. Acta Crystallogr D Biol Crystallogr 66, 213-221

37. Emsley, P., Lohkamp, B., Scott, W. G., and Cowtan, K. (2010) Features and development of Coot. Acta Crystallogr D Biol Crystallogr 66, 486-501

85 CHAPTER 3: SEPARATION OF THE TOXICITIY AND ANTI-ANGIOGENIC

EFFECTS OF BORRELIDIN AND BC194

3.1 Introduction

Aminoacyl-tRNA synthetases (ARSs) attach amino acids to their corresponding tRNA adaptors with high specificity in an essential and early reaction of protein synthesis

(1,2). In addition to this canonical function, ARSs and ARS-related proteins exhibit diverse alternative activities including RNA splicing, translational regulation, immune system modulation, and angiogenesis (3-7). Recent genetic evidence has served to link

ARSs to a variety of human and murine diseases associated with the brain and the nervous system, including Charcot-Marie Tooth disease (8,9), Type III Usher Syndrome

(10), and various encephalopathies (11,12). In several cases, these associations appear to be linked to secondary ARS functions, potentially those most closely tied to signaling.

One secondary function with particular significance for human physiology is angiogenesis, where multiple ARS play a variety of stimulatory and inhibitory modes.

As indicated in Chapter 1, tyrosyl-tRNA and tryptophanyl-tRNA synthetase are secreted in response to the inflammatory cytokines TNF-α and interferon γ, respectively (6,13,14).

Fragments or splice variants of these ARSs exert opposite effects, with the YARS fragment stimulating angiogenesis and WARS inhibiting angiogenesis. While there is evidence to suggest that the angiostatic properties of WARS depend on direct interactions with VE-cadherin, a role for ARSs in well-established angiogenic signaling pathways, such as those associated with vascular endothelial growth factor (VEGF), has not been

86 defined. In zebrafish, a mutations in the SARS gene encoding seryl-tRNA synthetase are associated with altered vascular development (15,16).

The class II threonyl-tRNA synthetase (TARS) is the most recent ARS for whom an angiogenic function has been identified, and exhibits some features that are distinct from YARS and WARS. TARS is secreted from endothelial cells in response to

TNF-α and VEGF, and potently stimulates angiogenesis in the human umbilical vein endothelial cell (HUVEC) tube formation and chicken chorioallantoic membrane assays

(17). The recent report of a strong association between TARS expression and advancing stage of ovarian cancer provides evidence that the pro-angiogenic function of TARS in angiogenesis is significant, at least in a pathophysiological context (18). An open question is whether this pro-angiogenic function is a directly linked to the canonical aminoacylation function, or is instead an as yet undetermined secondary function.

Moreover, the role of TARS in normal vascular development in metazoans, if any, remains to be determined.

Insights into the pro-angiogenic properties of TARS are provided by the characterization of its potent inhibitor borrelidin. Previously, several point mutations within the borrelidin binding site of TARS were shown to confer resistance to the compound, with the greatest difference resulting from the substitution of L489 (E. coli numbering) with a tryptophan (19). The crystallographic data presented in Chapter 2 provide an explanation for this dramatic change in affinity as this modification would disrupt a key interaction with borrelidin’s essential nitrile group (20). L489 is conserved from E. coli to humans in the form of L567 indicating that similar mutations could be generated in the human enzyme (see Chapter 2, section 2.2.5). Interestingly, the

87 substitution of L567 with valine in the human TARS enzyme was found to block the angiogenesis inhibiting properties of BC194. These data provide strong evidence that borrelidin and related compounds inhibit angiogenesis specifically through interaction with TARS in in vitro assays (21).

The anti-angiogenic properties of borrelidin provide an opportunity for its use as an anti-metastatic agent (22). However, borrelidin has yet to be exploited clinically, owing to its cytotoxicity to normal epithelial cells at nanomolar concentrations (23).

Several derivatives of borrelidin have since been developed through the biosynthetic engineering of bor gene cluster in attempt to alleviate this toxicity (20). Of particular interest is the derivative BC194 (Figure 5, Chapter 2) which possesses equipotent inhibition of TARS relative to borrelidin in enzymatic assays (Ki of 4.09 nM for aminoacylation) and a nearly identical structure to borrelidin save for the elimination of a single methylene group in the C17 pendent chain (21). Despite its similar properties,

BC194 exhibits greatly reduced toxicity relative to borrelidin while still maintaining function as an anti-angiogenic compound (20).

Recent work has also demonstrated that the toxic effects are due to the induction of the general control nonderepressible 2 kinase (GCN2) based amino acid starvation and is mechanistically distinct from the anti-angiogenic effects (21). The reduction of aminoacylation activity resulting from ARS inhibitors or amino acid deprivation increases the intracellular concentration of uncharged-tRNA. These molecules then associate with GCN2, stimulating its dimerization and auto-phosphorylation activity.

Active GCN2 then phosphorylates the guanine nucleotide cycling protein, eukaryotic translational initiation factor 2α (eIF2α), locking it in its GDP-bound conformation

88 (GDP-eIF2α) (24). Similar phosphorylation of eIF2α is also observed following the activation of protein kinase R (PKR)-like endoplasmic reticulum kinase (PERK) by ER stress and the unfolded protein response (25,26). GDP-eIF2α inhibits translation of mRNA by blocking Met-tRNA initiator tRNA with the exception of select transcripts involved in stress response pathways (27). Specifically, GDP-eIF2α encourages the translation of activating transcription factor 4 (ATF4), a transcription factor that up- regulates the expression of cell survival genes, such as VEGFA, (C/EβP) homology protein (CHOP), and other autophagy associated genes (28-30).

The differences in toxicity between BC194 and borrelidin are shown to originate from their varying ability to elicit the amino acid starvation response, yet both compounds exhibit equal potency in blocking angiogenesis. Characterization of the molecular basis of this difference opens the door to potential exploitation of these compounds in a clinical setting, such as in the therapeutic inhibition of angiogenesis.

Here, we demonstrate that the differences in toxicity between BC194 and borrelidin are shown to originate from their varying ability to elicit the amino acid starvation response, while their inhibition of angiogenesis occurs through a yet undetermined mechanism.

Finally, we compare the effects of borrelidin derivatives on vascular development in the zebrafish, and provide direct evidence for the role of TARS in angiogenesis in vivo.

3.2 Results

3.2.1 Borrelidin mediated amino acid starvation elicits cell cycle arrest

Borrelidin and BC194 were previously shown to exhibit significant differences in their toxicity to endothelial cells (20). However, the molecular basis for these effects has

Figure 12. Borrelidin elicits toxic responses related to cell cycle progression in endothelial cells at lower concentrations than BC194. (A) Cell-cycle analysis of synchronized HUVECs treated with serum containing borrelidin or BC194 using flow cytometry. (B) Percentage of cells in G2/M phase following treatment with borrelidin or BC194. HUVEC cells were treated with 10 nM borrelidin or BC194, harvested at 0, 8, 16, and 24 h, and stained with propidium iodide. The cells in G2/M were determined by gating using the 16h serum control; mean ± SEM, n=3, *p<0.05 (one-way ANOVA, Tukey Test). (C) The effects of borrelidin and BC194 on HUVEC proliferation. Cells were exposed to the indicated concentrations of compounds for 48 h and the relative proliferation measured using an alamarBlue® assay. not been established. To better understand the physiological consequences of borrelidin and BC194, we compared their effects on HUVEC cell-cycle progression using flow cytometry (Figure 12A, B). Treatment with 10 nM borrelidin arrested cells at the G0 boundary, delaying progression into G2/M up to 24 h. In contrast, cells treated with 10 nM BC194 did not significantly deviate from serum treated controls, and entered the

90 G2/M phase 16 h following serum exposure (Figure 12B). The effects of borrelidin and

BC194 on HUVEC proliferation were subsequently investigated using an alamarBlue® based assay (Figure 12C). Reductions in cell proliferation were observed at concentrations as low as 10 nM for borrelidin, whereas a 10-fold higher concentration of

BC194 was required for a comparable decrease in cell proliferation. These observations suggest that a major component of borrelidin’s toxicity is its ability to block progression through the cell cycle.

The observed differences in the ability of two compounds to induce cell cycle arrest may arise from variations in their ability to induce amino acid starvation. The amino acid starvation (AAS) response is typically detected as increased phosphorylation of the translational initiation factor eIF2α, which results from activation of the GCN2 translational control kinase (31). Previously, borrelidin treatment was demonstrated to de- repress expression of amino acid biosynthetic genes in yeast, a signature of GCN2 activity (32). Particularly when coupled to the unfolded protein response, amino acid starvation leads to cell cycle arrest and, in severe cases, the initiation of apoptosis

(24,33,34). To confirm that the anti-proliferative effects of borrelidin are linked to AAS pathways in animal cells, cultured HUVECs were exposed to increasing amounts of borrelidin or BC194, and then lysates from these cultures were probed for the starvation and apoptosis markers phospo-eIF2α and cleaved-caspase 3, respectively (Figure 13A,

D). The phosphorylation of eIF2α was induced at concentrations as low as 10 nM borrelidin (Figure 13B) but at least 10-fold higher concentrations of BC194 were required to generate an equivalent response (Figure 13E). Likewise, the appearance of cleaved- caspase 3 at 100 nM borrelidin (Figure 13C) compared to 1000 nM for BC194 (Figure

Figure 13. The cytotoxicity of borrelidin is linked to the induction of the amino acid starvation response. (A-F) HUVEC cells grown in full serum media were exposed to the indicated concentrations of borrelidin (A-C) or BC194 (D-F). Western blots of cell extracts were analyzed using antibodies recognizing phospo-eIF2α and cleaved-caspase 3 with β-tubulin as a loading control. Quantification of phospho-eIF2α and cleaved-caspase 3 for borrelidin (B, C) and BC194 (E, F) relative to β-tubulin; mean ± SEM, n≥3, *p<0.05 relative to 0 nM (one-way ANOVA, Tukey Test). #, Owing to the high variance at 1000 nM borrelidin resulting from extensive apoptosis and cell death, the extent of eIF2α phosphorylation could not be reliably quantitated. (G, H) Quantification of eif2α amounts relative to β-tubulin for both borrelidin- and BC194-treated cells (100 nM) in the presence of various threonine concentrations (0-10 mM); mean ± SEM, n≥3, *p<0.05 relative to 0 mM threonine (one-way ANOVA, Tukey Test), #p<0.01 between compounds at same threonine concentration (two-way ANOVA, Sidak Test).

13F) demonstrates the greater potency of borrelidin relative to BC194 in inducing apoptosis. These experiments underscore the existence of a critical concentration in the range of 10-100 nM where borrelidin elicits AAS, and BC194 does not.

92 An important prediction of the structural and cell free protein synthesis studies in

Chapter 2, section 2.2.4, is that borrelidin and BC194 should exhibit differential sensitivity with respect to the ability of threonine to rescue compound-induced amino acid starvation. Significantly, when threonine was added at concentrations between 2.5-

10 mM to HUVECs treated with 100 nM BC194, the phosphorylation of eIF2α was decreased to untreated levels (Figure 13G, H). By contrast, cells treated with 100 nM borrelidin still exhibited substantial levels of phospho-eIF2α over this same range of threonine concentrations. Thus, the combined structural and biochemical analysis suggests that the reduced ability of BC194 to inhibit protein synthesis attenuates its induction of amino acid starvation in HUVECs, thereby reducing its toxicity in growth and proliferation assays.

As an additional control to demonstrate that borrelidin induced amino acid starvation response is directly linked to TARS inhibition, HUVECs were incubated with a third compound, BC220, which is structurally distinguished from borrelidin by a 3- ethyl piperidine group that is esterified to the cyclopentane carboxylic acid moiety

(Chapter 2, section 2.2.2). The Ki for inhibition of TARS aminoacylation by BC220 is greater than 10 μM (35). HUVECs exposed to BC220 showed no induction of eIF2α phosphorylation or cleavage of caspase 3 (Figure 14A-C), indicating that the lack of physiological response is highly correlated with its failure to inhibit TARS. In summary, the combined results with all three compounds demonstrate that the cytotoxicity of borrelidin activation is linked to the activation of the amino acid starvation response and resultant apoptosis, which originates from inhibition of TARS.

Figure 14. BC220 does not induce amino acid starvation toxicity responses. (A-C) Western blots of cell extracts exposed to BC220 at the indicated concentrations using antibodies against phospho-eif2α and cleaved-caspase 3 (A). Thapsigargin (Th; 1 µg/ml) was used as a positive control for amino acid starvation. Protein values for phospho-eif2α (B) and cleaved-caspase 3 (C) were quantified relative to the β-tubulin loading control; mean ± SEM, n≥3, *p<0.0001 relative to 0 nM (one-way ANOVA, Tukey Test).

3.2.2 Borrelidin and BC194 exhibit comparable inhibition of angiogenesis at sub- toxic levels.

Previous studies indicate that borrelidin and BC194 are both potent inhibitors of angiogenesis (20,22,36). One potential explanation of the anti-angiogenic effect is that it is a direct consequence of the induction of the amino acid starvation response in endothelial cells, leading to their self-destruction through apoptosis. To test this hypothesis, we compared the ability of all three macrolides to inhibit HUVEC branching at sub-toxic concentrations in the endothelial tube formation assay (Figure 15A, B).

Significant decreases in branch formation were observed for both borrelidin and BC194 relative to uninhibited conditions at concentrations as low as 1 nM, which is 10- and 100- fold below the lowest borrelidin and BC194 concentrations demonstrated to induce eIF2α phosphorylation. In contrast, exposure to BC220 at these same concentrations had no effect on tube formation (Figure 15A). Borrelidin and BC194 thus exhibit comparable potencies in inhibiting angiogenesis, despite the fact that 10-100-fold higher

94 concentrations of BC194 are required to produce the levels of cellular toxicity seen with borrelidin.

Figure 15. Borrelidin and BC194 exhibit comparable inhibition of angiogenesis at sub- toxic levels in HUVEC tube formation assays. (A) Quantification of HUVEC branching at varying concentrations of borrelidin, BC194, and BC220. HUVECs were plated on Matrigel in full EGM-2 (2% FBS) media and exposed to the indicated concentrations of compounds. Cells were fixed after 4-8 h and stained with Oregon Green 488 Phalloidin. Numbers represent percentage of control (0 nM) samples, mean ± SEM, n=3 *p<0.05 relative to 0 nM (one-way ANOVA, Tukey Test for each compounds). (B) Representative images of HUVEC branches for section A.

95 We extended these results by using the chorioallantoic membrane (CAM) assay

(Figure 16A, B), which monitors basal vessel formation in the fertilized chicken egg and thus more closely approximates in vivo angiogenesis, relative to the simple HUVEC tube formation assay. As shown in Figure 4A, recurring administration of borrelidin or BC194 to CAMs at concentrations of 10 nM and 1 µM over a 72 h period inhibited basal vessel formation relative to the PBS control. Similar to the HUVEC tube formation assays, borrelidin and BC194 exhibited comparable potencies at concentrations in the range of 1-

10 nM. At substantially higher concentrations (100 µM), necrosis was clearly evident in the borrelidin treated CAMs, but not those treated with BC194 (Figure 16B). These observations demonstrate that BC194 appears to retain all of the potency of borrelidin in the inhibition of angiogenesis, while exhibiting substantially less cytotoxicity. We

Figure 16. Borrelidin and BC194 exhibit comparable inhibition of angiogenesis at sub-toxic levels in chorioallantoic membrane assays. (A) The effects of borrelidin (BN) and BC194 on in vivo angiogenesis in a CAM assay. Fertilized chicken embryos were cultured ex-ova for 10 days after which compounds at the indicated concentrations were applied to gelform sponges on the CAM. Images were taken daily over 72 h and quantified as the change in vascularity score over this entire period; mean ± SEM, n≥12, *p<0.05 relative to PBS (one-way ANOVA, Tukey Test). (B) Representative images of CAM over time for BN and BC194 at 100 µM.

96 conclude that the inhibition of angiogenesis by borrelidin is therefore not dependent on the induction of apoptosis in HUVECs, and is more likely to be linked to a TARS- specific property that is distinct from induction of the AAS.

Figure 17. Borrelidin and BC194 treatment influences cell migration and cell-cell contacts. (A) Representative images of HUVEC cells in the donut migration assay. Upper panels- Images of Hoechst 33342-labeled HUVEC cells 0, 5, and 24 h after removal of PDMS gasket. Lower panels - Images of cells outside the original PDMS gasket boundary after 5 and 24 h using cell subtraction. (B) Quantification of migrated cells exposed to 5 or 25 nM BC194 after 5 and 24 h in the donut assay. HUVEC monolayers were cultured overnight in a 2 mm diameter, circular gasket. The gasket was removed and the cells treated with media containing the indicated concentrations of compound. Images were taken after 5 and 24 h and the number of cells outside the original 2 mm diameter region were counted; mean ± SEM, n≥6, *p<0.01 relative to DMSO of same time point (one-way ANOVA, Tukey Test). (C) Image of cells at the leading edge of the migratory boundary after 24 h exposure to DMSO (left) or 25 nM BC194 (right).

97 In a previous study, exposure of HUVEC cultures to exogenous TARS in transwell migration assays was found to increase cell migration, an important component in angiogenesis (17). These observations complemented an earlier study showing that borrelidin inhibits tumor cell migration (37). To explore this observation further, the effect of BC194 on cell migration was examined using a donut migration assay (Figure

17A-C) (38). In this assay, endothelial cells are confined by a cloning ring on a fibronectin-coated dish, and then cells that migrate after ring removal are quantified by image subtraction (Figure 17A). The number of migratory cells was significantly lower in cultures treated with 5 nM BC194 for 5 h and 25 nM BC194 for either 5 or 24 h, relative to untreated cells (Figure 17B). Interestingly, close inspection of the cells at the leading edge of each sample reveals a marked decrease in cell-cell contacts in samples exposed to DMSO relative to 25 nM BC194 (Figure 17C). These data suggest that the slow migration may be, in part, due to the retention of cell-cell adhesion molecules in

BC194-treated cells. These observations demonstrate that for borrelidin and BC194, the inhibition of angiogenesis and the inhibition of migration are correlated to each other, but not to the ability to induce amino acid starvation and apoptosis.

3.2.3 BC194 induces specific ectopic blood vessel formation in zebrafish without the developmental toxicity associated with borrelidin.

Close inspection of the surface vessels proximal to the borrelidin or BC194 applied to the CAM assay revealed several instances of abrupt directional changes during vessel migration (Figure 18A). Moreover, these directional changes often came in sets of three, potentially corresponding to three cycles of compound addition and clearance over

98 the 72-hour period. An attractive hypothesis is that the macrolide compounds elicit their anti-angiogenic effects by confounding TARS-dependent vessel guidance signals, resulting in an irregular vascular architecture.

In order to investigate the effects of anti-TARS macrolides on vessel development in a whole animal model, we exposed transgenic zebrafish embryos derived from the

Tg(flk: eGFP) or Tg(flk: dsRed) lines at 13-15 h post fertilization (hpf) to a range of concentrations of borrelidin, BC194, and BC220. After 30 h of compound exposure (~48 hpf), significant changes in gross morphology were observed in embryos exposed to borrelidin (left column) but not BC194 (middle column) or BC220 (right column) (Figure

18B). A more quantitative assessment of treatment toxicity was obtained through measurements of body length (Figure 18C) and heart rate (Figure 18D). Significant intra- group changes were observed for fish treated with either borrelidin or BC194, but not

BC220. However, inter-group analyses revealed that reductions in heart rate and body length were more characteristic of borrelidin-treated than BC194-treated embryos.

To assess the angiogenesis-related effects of these compounds, we screened for aberrancies in the development of trunk inter-segmental vessels (ISVs) by fluorescence microscopy following exposure to 5 µM concentrations of the macrolides (Figure 19A).

Significant patterning defects were observed in response to either borrelidin or BC194 treatment. Ectopic branches were readily formed in the ISVs of embryos exposed to 5

µM BC194 for 48 h (72 hpf), but not in those of borrelidin-treated embryos (Figure 19B).

By contrast, the ISVs in borrelidin-exposed fish were often truncated or incomplete at both 48 and 72 hpf (Figure 19C). Significantly, these effects were not observed when

Figure 18. Borrelidin and BC194 exhibit different effects on zebrafish toxicity. (A) Representative CAM images showing rapid changes in vessel directionality, indicated by black arrows, before and after 72 hour treatment with 1 µM borrelidin and BC194. (B) Whole body images of ~48 hpf zebrafish embryos treated with various concentrations of borrelidin, BC194, and BC220 for 24 h. The DMSO vehicle was standardized for each sample. (C, D) Histograms showing the effects of borrelidin, BC194, and BC220 exposure on zebrafish body length (C) and heart rate (D). Dechorionated embryos were incubated for 24 hours in egg water containing the indicated compounds. Body length measurements were obtained by analysis of bright-field images using Spot v 5.1 software; mean ± SEM, n=11, *p<0.05 relative to 0 nM (one-way ANOVA, Tukey Test). Heart rates were determined using bright-field microscopy to count heart beats over 15 second intervals; mean ± SEM, n=16, *p<0.05 relative to 0 nM (one-way ANOVA, Tukey Test).

100

Figure 19. Exposure to Borrelidin and BC194 results in vascular defects and mis- patterning. (A) Representative confocal images (20x magnification) of zebrafish ISVs at 48 and 72 hpf. Zebrafish embryos were incubated in egg water at 25 °C until 8-12 somites of age. The embryos were then manually dechorionated and incubated in egg water containing the indicated compounds at 28.5°C until the images were taken. Calibration bar represents 100 µm. (B, C) Quantification of ectopic branching (B) and aberrant patterning (C) of ISVs. Embryos were incubated with the indicated compounds (5 µM) or DMSO until images were taken by fluorescent microscopy at 48 and 72 hpf. Aberrant structures were then counted within a region encompassing 5 ISVs anterior and posterior from the end of the yolk extension; mean ± SEM, n≥8, #*p<0.0001 relative to DMSO at 48 and 72 hpf respectively (one-way ANOVA, Tukey Test). § Borrelidin-treated fish rarely survived to 72 hpf.

101 embryos were treated with either DMSO or BC220 (Figure 19A-C).

One explanation for the diminished toxicity of BC194 relative to borrelidin is that, as seen with in animal cell culture, BC194 has a reduced ability to induce amino acid starvation in the context of a whole animal model. To test this hypothesis, quantitative real time PCR was performed on total RNA prepared from embryos treated with equivalent (5 µM) concentrations of borrelidin, BC194, and BC220, and the DMSO vehicle. The ATF4 target genes asns, gpt2, and eif4ebp1 served as AAS markers in this experiment (24). Consistent with our model, borrelidin caused a significant increase in the expression of the markers asns and gpt2 but not BC194 (Figure 20). This experiment demonstrates BC194’s induction of ectopic ISV formation in the zebrafish is not a direct consequence of the induction of the amino acid starvation response.

Figure 20. The vascular defects and mis-patterning resulting from BC194 treatment is not due to the induction of amino acid starvation. RT-qPCR values for the expression amino acid starvation response genes asns, gpt2, and eif4ebp1. Dechorionated embryos (15 hpf) were incubated in egg water containing the indicated compounds for 24 hours. Total RNA was extracted by trizol/chloroform and the expression levels relative to actb were determined using the ΔΔCT method; mean ± SEM, n=3, ****p<0.0001 relative to DMSO (two-way ANOVA, Tukey Test).

102 In addition to faulty vessel patterning, a reduction in vessel lumen formation was also observed in embryos treated with either 5 µM borrelidin or BC194 for 24 h (48 hpf) and remained partially incomplete even at 72 hpf. This observation was confirmed by live videos of blood flow in trunk vessels of 72 hpf embryos treated with BC194 or DMSO.

No restriction of blood flow was observed in ISVs of DMSO-treated fish (Video 1) consistent with fluorescent images showing well-defined lumens (Figure 21A). In contrast, several ISVs in embryos treated with 5 µM BC194 were non-functional in terms of blood flow (Video 2) and appeared to correspond to vessels associated with ectopic branches and poorly-defined lumens (Figure 21B).

Figure 21. Treatment with BC194 inhibits blood-flow through zebrafish ISVs. (A, B) Still frame image from videos of trunk blood flow in 72 hpf zebrafish embryos (left) and fluorescence image of blood vessels (right) treated with DMSO (A) or 5 µM BC194 (B) for 48 h. Non-functioning lumens and ectopic branches are designated with black arrows and asterisks respectively. See also associated Videos 1 and 2.

103 3.2.4 The effects of borrelidin and derivatives on gene expression of endothelial cell migration associated genes.

The results of the prior experiments argue that the faulty patterning and reduction in vessel lumen formation seen with BC194 in the zebrafish originate from a TARS function distinct from the canonical role in protein synthesis. To obtain further insights into the role of TARS in angiogenesis, we investigated the effects of borrelidin, BC194, and BC220 on the expression of tars, vegfaa, and the downstream Notch-reporter genes ephrinb2a and heyL, in the zebrafish that are either known to be or hypothesized to be

Figure 22. Borrelidin and BC194 treatment affects gene expression. (A-D) RT-qPCR values for the expression of tars (A), vegfaa (B), ephrinb2a (C), and heyL (D) at 48 and 72 hpf. Dechorionated embryos (15 hpf) were incubated in egg water containing the indicated compounds. Total RNA was extracted by trizol/chloroform and the expression levels relative to actb were determined using the ΔΔCT method; mean ± SEM, n=3, #*p<0.05 relative to DMSO at 48 and 72 hpf respectively (one-way ANOVA, Tukey Test). § denotes that borrelidin-treated embryos rarely survived to 72 hpf and were not included in this figure.

104 related to vascular development and branching. Quantitative real time PCR (RT-qPCR) of total embryo RNA revealed significant increases in the expression of both tars and vegfaa (Figure 22A, B) mRNA after exposure for 24 hours with borrelidin (48 hpf) or 48 hours with BC194 (72 hpf). No changes in gene expression were observed for embryos treated with either DMSO or BC220, consistent with our TARS-dependent nature of our current model for borrelidin and BC194 action. Interestingly, the notch reporter genes showed no change over any of the tested conditions (Figure 22C, D).

3.3 Discussion

Natural product and synthetic inhibitors have been identified for many ARSs, and readily exploited to generate anti-microbial infectives (39,40), immunosuppressive agents

(24,41), and potential modulators of cancer metastasis (42). However, many natural compounds that target ARSs have a very narrow therapeutic index, dramatically limiting their clinical usefulness. The toxicity seen in many cases may reflect the acute sensitivity of animal cells to the inhibition of protein synthesis, which can arrest the cell cycle and elicit apoptosis (24,33).

Borrelidin’s pleiotropic effects on mammalian cell physiology have been appreciated for decades, but the molecular basis of its action or that of related compounds has been poorly understood (20). Here, a chemical and structural biology approach was used to compare the effects of borrelidin and two different derivatives, and gain insights into their ability to inhibit angiogenesis. Significantly, the cytotoxicity of borrelidin was found to originate from its induction of amino acid starvation, which leads to cell cycle arrest and apoptosis. The ability of borrelidin to induce amino acid starvation in yeast

105 was reported earlier (32). In animal cells, amino acid starvation induces the expression of the CCAAT/enhancer binding protein homology binding protein (CHOP), a transcription factor that stimulates apoptosis after prolonged nutritional deprivation (43). Not surprisingly, borrelidin is a potent inducer of apoptosis in acute lymphoblastic leukemia cells (44). These studies therefore support a strong linkage between the ability of a given aminoacyl-tRNA synthetase inhibitor to induce the amino acid starvation response and its potency with respect to apoptosis and cell killing. Conversely, the inhibition of angiogenesis was observed at sub-toxic concentrations for both borrelidin and BC194, suggesting that the compounds’ inhibition of angiogenesis is mediated through a mechanism distinct from the induction of amino acid starvation.

3.3.1 The greater toxicity exhibited by borrelidin relative to BC194 is due to its decreased susceptibility to threonine competition.

The major structural difference between the two TARS macrolide complexes is that significantly weaker interactions with critical residues constituting the threonine binding pocket (45,46) were observed in the BC194 complex (Chapter 2). Despite this structural change, BC194’s potency as an inhibitor of angiogenesis is undiminished. One hypothesis is that the inhibition of angiogenesis depends on the ability of borrelidin and

BC194 to stabilize a TARS conformation that is distinct from unliganded and substrate bound states (Chapter 2, Figure 9). Such a model would be consistent with the observation that the stimulation of angiogenesis depends on the structural integrity of the catalytic domain, and its potential to associate with additional protein partners (see

Appendix I for investigations of TARS binding partners). Initially, however, this model

106 may appear contradictory, given that BC194’s weaker interaction with the TARS enzyme would also be expected to disrupt its ability to inhibit angiogenesis as well. The resistance conferred by L567V TARS against BC194’s anti-angiogenic effects in CAM assays argues against a second borrelidin binding site within the TARS enzyme (21).

However, the two-part binding mechanism of borrelidin and its related compounds presented in Chapter 2 may provide an alternative explanation.

Upon first encounter BC194 forms a weaker complex with TARS relative to borrelidin, accounting for its increased susceptibility to threonine competition (Figure 13

G, H). However, as demonstrated by enzymatic assays (Chapter 2, section 2.2.2), the binding affinity for the second, tighter interaction between TARS and either inhibitor is nearly identical. As such, compound toxicity appears to be associated with the weak encounter complex while the anti-angiogenic effects are related to the tight-binding mode. Considering also that TARS appears to mediate angiogenesis extracellularly while performing aminoacylation within the cell, the concentrations of threonine must be considered in both compartments (21). Notably, intracellular concentrations of threonine are significantly higher than those outside the cell, indicating that competition between the amino acid and the compounds would be reduced in the regulation of angiogenesis in comparison to aminoacylation (47). As an alternative mechanism, full incorporation of borrelidin or BC194 into the TARS active site is characterized by a very slow off rate

(19). Moreover, despite the increased susceptibility of BC194 to threonine competition relative to borrelidin in the compound’s initial encounter with TARS, borrelidin and

app BC194 exhibit nearly identical final values for Ki . Therefore, both inhibitors may eventually proceed to a tightly bound complex but the greater activation energy

107 associated with BC194 binding at equimolar concentrations of threonine will reduce its rate of complex formation relative to that of borrelidin (48). Under these circumstances, compensatory mechanisms, such as the synthesis of new TARS, may be able to rescue the cell from the toxicity associated with the slower-binding BC194 but not borrelidin.

However, as both compounds eventually develop a full set of interactions with the TARS active site, they would still be able to induce the necessary changes to inhibit angiogenesis.

3.3.2 Borrelidin compounds influence blood vessel maturation and migration in zebrafish.

A major finding of this study is that the reduction of TARS activity via borrelidin or BC194 administration produced phenotypes associated with incomplete vessel development, increased ectopic branching, and decreased vessel lumens (Figure 19, 21

Videos 1, 2). This represents the first demonstration of a role for TARS as a pro- angiogenic factor in a whole organism model. Yet treatment with BC194 also increased expression of TARS (Figure 22A), an observation reminiscent of the response of CHO cells to persistent exposure to borrelidin (49). In view of the relatively high macrolide concentrations employed (i.e. several orders of magnitude over the apparent Ki), the increased TARS protein levels resulting from BC194 administration might effectively be masked by the stoichiometric formation of TARS-BC194 complexes, which are characterized by very slow off rates (19). Alternatively, the disruption in TARS levels resulting from BC194 treatment could alter a TARS-dependent concentration gradient

108 that is necessary for proper guidance and patterning of the developing vasculature

(17,50,51).

Along these lines, endothelial cells treated with BC194 showed a higher degree of cell-cell contacts relative to controls, even at the leading edge of a migratory boundary

(Figure 17C). Weakened cell-cell contacts have been proposed to be an important component of angiogenesis (52). Inhibition of tyrosine phosphorylation on VE-cadherin, which alters cell-cell interactions, decreases VEGF-mediated migration and angiogenesis in cell culture (53). Furthermore, certain models of lumenogenesis involve the strictly regulated disruption of contacts between endothelial cells of the developing vessel (54).

Borrelidin and related macrolide compounds may act to disrupt TARS-mediated regulation of cell-adhesion proteins, thereby preventing both migration and proper vascular development (37). The availability of specific inhibitory compounds and TARS variants that ablate its angiogenic function should prove to be valuable tools to critically test this hypothesis. The actions of borrelidin and BC194 can be ablated by single residue substitutions in the E. coli ThrRS (19), and TARS active sites (17), respectively (17,19), arguing against the previous suggestion that borrelidin binding to other secondary targets is physiologically significant in the inhibition of angiogenesis (36). Structural analysis of the ThrRS-borrelidin complex (see Chapter 2, section 2.2.3) suggests that the inhibition mechanisms of borrelidin and BC194 involve occlusion of the binding sites for one or more of the canonical aminoacylation substrates (55). Borrelidin is a slow tight binding inhibitor that competes with amino acid substrate binding, accounting for the potent stimulation of amino acid starvation (19).

109 Unfortunately, the above observations do not provide a direct mechanism through which TARS and borrelidin-based compounds influence angiogenesis. Early reports suggested that apoptosis of endothelial cells could contribute to the reduction in angiogenesis (56). While the impact of this phenomenon cannot be completely ruled out, it is complicated by a lack of specificity to a single cell type (23,57). Furthermore, the above data clearly indicate that anti-angiogenic effects of the compounds are observable at concentrations that do not induce amino acid starvation, suggesting that another mechanism is required to explain the relationship between TARS and borrelidin in angiogenesis.

3.4 Experimental Procedures

3.4.1 Cell Culture and Reagents

Human umbilical endothelial cells (HUVECs) were maintained in Clonetics

EGM®-2 complete media (Lonza) and used between passages 5 to 8. Borrelidin and

BC194 were obtained from Biotica (now Isomerase Therapeutics Ltd (Cambridge, UK)).

3.4.2 Flow Cytometry

HUVECs were plated onto 10 cm dishes and serum starved in 0.5% FBS EGM®-

2 media for 24 h. Cells were pre-treated with 10 nM borrelidin or BC194 for 6 h. To assess the serum-induced cell cycle entry, full serum media (EGM®-2 complete media) was added at staggered times, such that all samples were collected simultaneously for the

8 h, 16 h and 24 h time points. Cells were trypsinized, washed and fixed in 80% ethanol, washed in 2% BSA and then incubated with a solution containing 10 μg/ml propidium

110 iodide (PI) and 250 μg/ml RNAse A for 16 h as in (58). Analysis of the PI signal (620 nm) was performed on an EPICS XL four-color analytical flow cytometer coupled with

EXPO32 ADC software.

3.4.3 Endothelial Cell Proliferation Assay

The MTT-based alamarBlue® (Invitrogen) reagent was used to assess cell proliferation (59). HUVECs were seeded in a 96-well dish (1x103 cells/well) and grown for 48 h in EGM®-2 media. Cells were incubated with borrelidin or BC194 (0.1-1000 nM) as indicated; media alone served as the control. After 48 h in culture, 10 µl/well premixed alamarBlue® was added and after 3 h at 37ºC the amount of reduced alamarBlue® was quantified by fluorescence (excitation at 530 nm, emission at 590 nm) on a microplate reader (Synergy™ HT, BioTek). Values were graphed as % Control compared with media alone.

3.4.4 Western Blotting

HUVEC cells were seeded into 6-well dishes and incubated at 37 °C in EGM-2 media with 2% FBS. When cells had reached 95-100% confluence the media was exchanged and the cultures were incubated with various concentrations of borrelidin,

BC194, BC220, and DMSO for 24 h. For threonine rescue experiments, 2.5 to 10 mM threonine was also added to the cultures. Cell monolayers were washed with PBS and harvested using CellLyticTM M cell lysis reagent (Sigma) supplemented with complete

Mini protease inhibitor tablets (Roche) according to the manufacturer’s instructions. For the thapsigargin control, the media was exchanged at the same time as for the borrelidin

111 compounds, but 1 µg/ml thapsigargin (Sigma) was not added until 1 h before harvesting.

Lysate protein concentrations were determined using Quick StartTM Bradford Reagent

(Bio-Rad) and stored at -80 °C until used. Lysates were resolved by 10% SDS-PAGE and transferred to nitrocellulose membranes for Western blot analysis. Primary antibodies for phospo-eif2α, cleaved-caspase 3, and β-tubulin were purchased from Cell Signaling and used at 1:1000. The secondary antibody, HRP-goat-anti-rabbit IgG (Santa Cruz), was used at 1:5000. Bands were detected by X-ray film or the VersaDoc 4000 MP (BioRad) using Pierce ECL blotting substrate or Supersignal West Femto reagent (Pierce) as needed. Densitometric values were determined using ImageJ software lane analysis tools

(http://imagej.nih.gov/ij/; W. Rasband, NIH).

3.4.5 In vitro Tube Formation Assay

Tube formation assays were performed as previously described (17,60). Briefly

HUVEC cells were cultured until 95-100% confluent and seeded in to 48-well plates coated with MatrigelTM (BD Biosciences). Cells were incubated at 37 °C in EGM®-2 media (Lonza) with 2% FBS and various concentrations of borrelidin, BC194, or BC220 for 4 to 8 h. Cells were fixed with 10% neutral buffered formalin, stained with Oregon

Green 488 phalloidin (Molecular Probes), and imaged via fluorescence microscopy. Tube structures were quantified using the Simple Neurite Tracer plug-in on ImageJ software

(NIH).

3.4.6 Chick Chorioallantoic Membrane Assay

112 CAM assays were performed as previously described (17). Briefly, embryos from fertilized chicken eggs (Sunrise Farms, Catskill, NY) were cultured ex-ova in 10 cm2 tissue culture dishes at 37°C until developmental day 10. Surgical sponges (Surgifoam) approximately 1 mm3 were then arranged within the outer third of the membrane avoiding major vessels. Compounds in 10 µl volumes were applied directly to the sponges at 24 h intervals over the course of 72 h. Images were taken daily using a Leica

MZ6 stereomicroscope and scored using modified Intensity Scoring described previously

(61).

3.4.7 Cell Migration Assay

The donut cell migration assay was performed as previously described (38).

Briefly, polydimethylsiloxane (PDMS) donut-shaped gaskets were placed on fibronectin

(BD Biosciences) -coated glass coverslips to form a 2mm-diameter culture well into which 8x103 HUVECs were seeded and allowed to adhere overnight. After removal of the gaskets, the circular HUVEC monolayers were washed twice with complete media to remove non-adherent cells, nuclei were labeled by incubation with Hoechst 33342 for 15 min, and the monolayers were re-fed with complete media containing vehicle (DMSO) or

BC194 at the indicated concentrations. Images of the circular monolayers were acquired after re-feeding and at the indicated times and image pairs were analyzed using a custom

ImageJ macro to calculate the number of cells at each time point that have migrated beyond the original monolayer.

3.4.8 Zebrafish Handling and Treatments

113 This project utilized the transgenic lines Tg(flk1:eGFP) and Tg(flk1:dsRed) in which eGFP and dsRed expression respectively are controlled by the flk1 promoter. The

Tg(flk1:eGFP) or Tg(flk1:dsRed) fish lines exhibit endothelial specific fluorescence starting at 48 and 72 hpf respectively. Due to its earlier expression the Tg(flk1:eGFP) line was used for all 48 hpf fluorescence images; all other applications used the

Tg(flk1:dsRed) line. Zebrafish were bred and maintained in accordance with approved

University of Vermont IACUC standards for vertebrate animals and staged as previously described (62). Fertilized embryos were transferred to glass dishes and allowed to develop at 25°C until 8 to 12 somites in age. The fish were manually dechorionated and transferred to 24 well dishes containing 1 ml of egg water treated with 0.003% 1-phenyl

2-thiourea to inhibit pigment, and various concentrations of borrelidin, BC194, BC220, or DMSO (vehicle control) and incubated at 28.5 °C for downstream analyses. In all conditions DMSO was standardized to match the highest concentration for each compound. Dead and dying embryos were removed and the media exchanged for fresh egg water and compounds every 24 h.

3.4.9 Quantitative Assessment of Toxicity

Heart rate and body length data were acquired after the first 24 h of compound exposure using the Tg(flk1:dsRed) line and bright field microscopy. Heart rates were counted over 15 sec and converted to beats per min. For body length, images were taken using a Spot camera attached to a Nikon SMZ800 dissecting microscope at a 3X objective and analyzed with Spot v 5.1 software.

114 3.4.10 Confocal Imaging

Embryos were live imaged with a 20x objective using step sizes of 2 µm after mounting in 0.5% low melt agarose on glass bottom dishes (Mat-Tek) on a Nikon Eclipse

Ti inverted microscope. Stacks were subjected to a kalman stack filter in ImageJ and are presented as maximum intensity projections. Images were processed with Adobe

Photoshop for brightness and contrast. For quantification of ISV ectopic branches, images were taken using an Olympus IX71 inverted fluorescence microscope. Abnormal branches and vessel defects associated with ISVs were counted within a region encompassing five ISVs anterior and posterior to the end of the yolk extension.

3.4.11 Live-Imaging of Zebrafish Blood Flow

Zebrafish embryos (72 hpf) of the Tg(flk:dsRed) line were treated for 48 hours with DMSO or 5 µM BC194 as described above. Fluorescence images were acquired on a Nikon TE2000-PFS epifluorescence microscope through a 20X Plan-Fluor objective using wavelength-appropriate filters (Chroma Technologies) with a Clara CCD camera

(Andor, Inc.). Acquisition was controlled using Nikon Elements software and images were adjusted for presentation using Elements or Fiji/ImageJ (http://imagej.nih.gov/ij/;

W. Rasband, NIH).

3.4.12 RNA Expression in Zebrafish

Following manual dechorionation embryos (~8 somites; Tg(flk:dsRed)) were incubated for 24 or 48 h in egg water containing varying concentrations of compounds at

28.5 °C. The fish were subsequently transferred in minimum volume to 1.5 mL

115 microfuge tubes and then total RNA was obtained using Trizol:chloroform (Invitrogen).

First strand cDNA was synthesized using Superscript II RT-PCR technology (Invitrogen) according to manufacturer’s instructions. RT-qPCR was performed using an ABI prism

7700 Sequence Detection System (Applied Biosystems) and TaqMan® Assay primer pairs (Applied Biosystems) for tars, vegfaa, heyL, ephrinb2a, and act1a. Relative mRNA expression was determined using a comparative CT (ΔΔCT) method with act1b as an endogenous control.

3.4.13 Statistical Analysis

All experiments were repeated at least 3 times; n values are reported in the corresponding figure legend. Values are presented as means ± SEM. Data were analyzed by one- or two-way ANOVA and Tukey or Sidak test as indicated using GraphPad Prism v6 software. Significance was defined as P < 0.05.

116 3.5 References

1. Ibba, M., and Soll, D. (2000) Aminoacyl-tRNA synthesis. Ann. Rev. Biochem. 69, 617-650

2. Yao, P., and Fox, P. L. (2013) Aminoacyl-tRNA synthetases in medicine and disease. EMBO Mol Med 5, 332-343

3. Wakasugi, K., and Schimmel, P. (1999) Two distinct cytokines released from a human aminoacyl-tRNA synthetase. Science 284, 147-151

4. Sampath, P., Mazumder, B., Seshadri, V., Gerber, C. A., Chavatte, L., Kinter, M., Ting, S. M., Dignam, J. D., Kim, S., Driscoll, D. M., and Fox, P. L. (2004) Noncanonical function of glutamyl-prolyl-tRNA synthetase: gene-specific silencing of translation. Cell 119, 195-208

5. Park, S. G., Kim, H. J., Min, Y. H., Choi, E. C., Shin, Y. K., Park, B. J., Lee, S. W., and Kim, S. (2005) Human lysyl-tRNA synthetase is secreted to trigger proinflammatory response. Proc Natl Acad Sci U S A 102, 6356-6361

6. Greenberg, Y., King, M., Kiosses, W. B., Ewalt, K., Yang, X., Schimmel, P., Reader, J. S., and Tzima, E. (2008) The novel fragment of tyrosyl tRNA synthetase, mini-TyrRS, is secreted to induce an angiogenic response in endothelial cells. FASEB J 22, 1597-1605

7. Guo, M., and Schimmel, P. (2013) Essential nontranslational functions of tRNA synthetases. Nat Chem Biol 9, 145-153

8. Antonellis, A., Ellsworth, R. E., Sambuughin, N., Puls, I., Abel, A., Lee-Lin, S. Q., Jordanova, A., Kremensky, I., Christodoulou, K., Middleton, L. T., Sivakumar, K., Ionasescu, V., Funalot, B., Vance, J. M., Goldfarb, L. G., Fischbeck, K. H., and Green, E. D. (2003) Glycyl tRNA Synthetase Mutations in Charcot-Marie-Tooth Disease Type 2D and Distal Spinal Muscular Atrophy Type V. Am J Hum Genet 72, 1293-1299

9. Jordanova, A., Irobi, J., Thomas, F. P., Van Dijck, P., Meerschaert, K., Dewil, M., Dierick, I., Jacobs, A., De Vriendt, E., Guergueltcheva, V., Rao, C. V., Tournev, I., Gondim, F. A., D'Hooghe, M., Van Gerwen, V., Callaerts, P., Van Den Bosch, L., Timmermans, J. P., Robberecht, W., Gettemans, J., Thevelein, J. M., De Jonghe, P., Kremensky, I., and Timmerman, V. (2006) Disrupted function and axonal distribution of mutant tyrosyl-tRNA synthetase in dominant intermediate Charcot-Marie-Tooth neuropathy. Nat Genet 38, 197-202

10. Puffenberger, E. G., Jinks, R. N., Sougnez, C., Cibulskis, K., Willert, R. A., Achilly, N. P., Cassidy, R. P., Fiorentini, C. J., Heiken, K. F., Lawrence, J. J., Mahoney, M. H., Miller, C. J., Nair, D. T., Politi, K. A., Worcester, K. N., Setton, 117 R. A., Dipiazza, R., Sherman, E. A., Eastman, J. T., Francklyn, C., Robey-Bond, S., Rider, N. L., Gabriel, S., Morton, D. H., and Strauss, K. A. (2012) Genetic mapping and exome sequencing identify variants associated with five novel diseases. PLoS One 7, e28936

11. Scheper, G. C., van der Klok, T., van Andel, R. J., van Berkel, C. G., Sissler, M., Smet, J., Muravina, T. I., Serkov, S. V., Uziel, G., Bugiani, M., Schiffmann, R., Krageloh-Mann, I., Smeitink, J. A., Florentz, C., Van Coster, R., Pronk, J. C., and van der Knaap, M. S. (2007) Mitochondrial aspartyl-tRNA synthetase deficiency causes leukoencephalopathy with brain stem and spinal cord involvement and lactate elevation. Nature Genetics 39, 534-539

12. Kodera, H., Osaka, H., Iai, M., Aida, N., Yamashita, A., Tsurusaki, Y., Nakashima, M., Miyake, N., Saitsu, H., and Matsumoto, N. (2014) Mutations in the glutaminyl-tRNA synthetase gene cause early-onset epileptic encephalopathy. J Hum Genet

13. Wakasugi, K., Slike, B. M., Hood, J., Ewalt, K. L., Cheresh, D. A., and Schimmel, P. (2002) Induction of angiogenesis by a fragment of human tyrosyl- tRNA synthetase. J Biol Chem 277, 20124-20126

14. Wakasugi, K., Slike, B. M., Hood, J., Otani, A., Ewalt, K. L., Friedlander, M., Cheresh, D. A., and Schimmel, P. (2002) A human aminoacyl-tRNA synthetase as a regulator of angiogenesis. Proc Natl Acad Sci U S A 99, 173-177

15. Fukui, H., Hanaoka, R., and Kawahara, A. (2009) Noncanonical activity of seryl- tRNA synthetase is involved in vascular development. Circ Res 104, 1253-1259

16. Herzog, W., Muller, K., Huisken, J., and Stainier, D. Y. (2009) Genetic evidence for a noncanonical function of seryl-tRNA synthetase in vascular development. Circ Res 104, 1260-1266

17. Williams, T. F., Mirando, A. C., Wilkinson, B., Francklyn, C. S., and Lounsbury, K. M. (2013) Secreted Threonyl-tRNA synthetase stimulates endothelial cell migration and angiogenesis. Sci Rep 3, 1317

18. Wellman, T. L., Eckenstein, M., Wong, C., Rincon, M., Ashikaga, T., Mount, S. L., Francklyn, C. S., and Lounsbury, K. M. (2014) Threonyl-tRNA synthetase overexpression correlates with angiogenic markers and progression of human ovarian cancer. BMC Cancer 14, 620

19. Ruan, B., Bovee, M. L., Sacher, M., Stathopoulos, C., Poralla, K., Francklyn, C. S., and Soll, D. (2005) A unique hydrophobic cluster near the active site contributes to differences in borrelidin inhibition among threonyl-tRNA synthetases. J Biol Chem 280, 571-577

118 20. Wilkinson, B., Gregory, M. A., Moss, S. J., Carletti, I., Sheridan, R. M., Kaja, A., Ward, M., Olano, C., Mendez, C., Salas, J. A., Leadlay, P. F., vanGinckel, R., and Zhang, M. Q. (2006) Separation of anti-angiogenic and cytotoxic activities of borrelidin by modification at the C17 side chain. Bioorg Med Chem Lett 16, 5814- 5817

21. Williams, T. F., Mirando, A. C., Wilkinson, B., Francklyn, C. S., and Lounsbury, K. M. (2013) Secreted Threonyl-tRNA synthetase stimulates endothelial cell migration and angiogenesis. Sci Rep 3, 1317

22. Funahashi, Y., Wakabayashi, T., Semba, T., Sonoda, J., Kitoh, K., and Yoshimatsu, K. (1999) Establishment of a quantitative mouse dorsal air sac model and its application to evaluate a new angiogenesis inhibitor. Oncol Res 11, 319- 329

23. Gafiuc, D., Weiss, M., Mylonas, I., and Bruning, A. (2013) Borrelidin has limited anti-cancer effects in bcl-2 overexpressing breast cancer and leukemia cells and reveals toxicity in non-malignant breast epithelial cells. J Appl Toxicol

24. Sundrud, M. S., Koralov, S. B., Feuerer, M., Calado, D. P., Kozhaya, A. E., Rhule-Smith, A., Lefebvre, R. E., Unutmaz, D., Mazitschek, R., Waldner, H., Whitman, M., Keller, T., and Rao, A. (2009) Halofuginone inhibits TH17 cell differentiation by activating the amino acid starvation response. Science 324, 1334-1338

25. Hamanaka, R. B., Bennett, B. S., Cullinan, S. B., and Diehl, J. A. (2005) PERK and GCN2 contribute to eIF2alpha phosphorylation and cell cycle arrest after activation of the unfolded protein response pathway. Mol Biol Cell 16, 5493-5501

26. Teske, B. F., Wek, S. A., Bunpo, P., Cundiff, J. K., McClintick, J. N., Anthony, T. G., and Wek, R. C. (2011) The eIF2 kinase PERK and the integrated stress response facilitate activation of ATF6 during endoplasmic reticulum stress. Mol Biol Cell 22, 4390-4405

27. Wek, R. C., Jiang, H. Y., and Anthony, T. G. (2006) Coping with stress: eIF2 kinases and translational control. Biochem Soc Trans 34, 7-11

28. B'Chir, W., Maurin, A. C., Carraro, V., Averous, J., Jousse, C., Muranishi, Y., Parry, L., Stepien, G., Fafournoux, P., and Bruhat, A. (2013) The eIF2alpha/ATF4 pathway is essential for stress-induced autophagy gene expression. Nucleic Acids Res 41, 7683-7699

29. Dey, S., Savant, S., Teske, B. F., Hatzoglou, M., Calkhoven, C. F., and Wek, R. C. (2012) Transcriptional repression of ATF4 gene by CCAAT/enhancer-binding protein beta (C/EBPbeta) differentially regulates integrated stress response. J Biol Chem 287, 21936-21949

119 30. Ghosh, R., Lipson, K. L., Sargent, K. E., Mercurio, A. M., Hunt, J. S., Ron, D., and Urano, F. (2010) Transcriptional regulation of VEGF-A by the unfolded protein response pathway. PLoS One 5, e9575

31. Wek, R. C., Jiang, H. Y., and Anthony, T. G. (2006) Coping with stress: eIF2 kinases and translational control. Biochem Soc Trans 34, 7-11

32. Eastwood, E. L., and Schaus, S. E. (2003) Borrelidin induces the transcription of amino acid biosynthetic enzymes via a GCN4-dependent pathway. Bioorganic & Medicinal Chemistry Letters 13, 2235-2237

33. Brooks, R. F. (1977) Continuous protein synthesis is required to maintain the probability of entry into S phase. Cell 12, 311-317

34. Hamanaka, R. B., Bennett, B. S., Cullinan, S. B., and Diehl, J. A. (2005) PERK and GCN2 contribute to eIF2alpha phosphorylation and cell cycle arrest after activation of the unfolded protein response pathway. Mol Biol Cell 16, 5493-5501

35. Novoa, E. M., Camacho, N., Tor, A., Wilkinson, B., Moss, S., Marin-Garcia, P., Azcarate, I. G., Bautista, J. M., Mirando, A. C., Francklyn, C. S., Varon, S., Royo, M., Cortes, A., and Ribas de Pouplana, L. (2014) Analogs of natural aminoacyl- tRNA synthetase inhibitors clear malaria in vivo. Proc Natl Acad Sci U S A 111, E5508-5517

36. Kawamura, T., Liu, D., Towle, M. J., Kageyama, R., Tsukahara, N., Wakabayashi, T., and Littlefield, B. A. (2003) Anti-angiogenesis effects of borrelidin are mediated through distinct pathways: threonyl-tRNA synthetase and caspases are independently involved in suppression of proliferation and induction of apoptosis in endothelial cells. J Antibiot (Tokyo) 56, 709-715

37. Harisi, R., Kenessey, I., Olah, J. N., Timar, F., Babo, I., Pogany, G., Paku, S., and Jeney, A. (2009) Differential inhibition of single and cluster type tumor cell migration. Anticancer Res 29, 2981-2985

38. McKenzie, A. J., Campbell, S. L., and Howe, A. K. Protein kinase A activity and anchoring are required for ovarian cancer cell migration and invasion. PLoS One 6, e26552

39. Thomas, C. M., Hothersall, J., Willis, C. L., and Simpson, T. J. (2010) Resistance to and synthesis of the antibiotic mupirocin. Nature reviews. Microbiology 8, 281- 289

40. Rock, F. L., Mao, W., Yaremchuk, A., Tukalo, M., Crepin, T., Zhou, H., Zhang, Y. K., Hernandez, V., Akama, T., Baker, S. J., Plattner, J. J., Shapiro, L., Martinis, S. A., Benkovic, S. J., Cusack, S., and Alley, M. R. (2007) An antifungal agent inhibits an aminoacyl-tRNA synthetase by trapping tRNA in the editing site. Science 316, 1759-1761 120 41. Van de Vijver, P., Ostrowski, T., Sproat, B., Goebels, J., Rutgeerts, O., Van Aerschot, A., Waer, M., and Herdewijn, P. (2008) Aminoacyl-tRNA synthetase inhibitors as potent and synergistic immunosuppressants. Journal of medicinal chemistry 51, 3020-3029

42. Kim, D. G., Lee, J. Y., Kwon, N. H., Fang, P., Zhang, Q., Wang, J., Young, N. L., Guo, M., Cho, H. Y., Mushtaq, A. U., Jeon, Y. H., Choi, J. W., Han, J. M., Kang, H. W., Joo, J. E., Hur, Y., Kang, W., Yang, H., Nam, D. H., Lee, M. S., Lee, J. W., Kim, E. S., Moon, A., Kim, K., Kim, D., Kang, E. J., Moon, Y., Rhee, K. H., Han, B. W., Yang, J. S., Han, G., Yang, W. S., Lee, C., Wang, M. W., and Kim, S. (2013) Chemical inhibition of prometastatic lysyl-tRNA synthetase-laminin receptor interaction. Nat Chem Biol 10, 29-34

43. B'Chir, W., Chaveroux, C., Carraro, V., Averous, J., Maurin, A. C., Jousse, C., Muranishi, Y., Parry, L., Fafournoux, P., and Bruhat, A. (2014) Dual role for CHOP in the crosstalk between autophagy and apoptosis to determine cell fate in response to amino acid deprivation. Cellular signalling 26, 1385-1391

44. Habibi, D., Ogloff, N., Jalili, R. B., Yost, A., Weng, A. P., Ghahary, A., and Ong, C. J. (2011) Borrelidin, a small molecule nitrile-containing macrolide inhibitor of threonyl-tRNA synthetase, is a potent inducer of apoptosis in acute lymphoblastic leukemia. Invest New Drugs 30, 1361-1370

45. Sankaranarayanan, R., Dock-Bregeon, A. C., Rees, B., Bovee, M., Caillet, J., Romby, P., Francklyn, C. S., and Moras, D. (2000) Zinc ion mediated amino acid discrimination by threonyl-tRNA synthetase.[see comment]. Nature Structural Biology 7, 461-465

46. Torres-Larios, A., Sankaranarayanan, R., Rees, B., Dock-Bregeon, A. C., and Moras, D. (2003) Conformational movements and cooperativity upon amino acid, ATP and tRNA binding in threonyl-tRNA synthetase. J Mol Biol 331, 201-211

47. Bergstrom, J., Furst, P., Noree, L. O., and Vinnars, E. (1974) Intracellular free amino acid concentration in human muscle tissue. Journal of applied physiology 36, 693-697

48. Copeland, R. A. (2005) Evaluation of enzyme inhibitors in drug discovery : a guide for medicinal chemists and pharmacologists, Wiley-Interscience, Hoboken, N.J.

49. Gantt, J. S., Bennett, C. A., and Arfin, S. M. (1981) Increased levels of threonyl- tRNA synthetase in a borrelidin-resistant Chinese hamster ovary cell line. Proc Natl Acad Sci U S A 78, 5367-5370

50. Ruhrberg, C., Gerhardt, H., Golding, M., Watson, R., Ioannidou, S., Fujisawa, H., Betsholtz, C., and Shima, D. T. (2002) Spatially restricted patterning cues

121 provided by heparin-binding VEGF-A control blood vessel branching morphogenesis. Genes Dev 16, 2684-2698

51. Adams, R. H., and Alitalo, K. (2007) Molecular regulation of angiogenesis and lymphangiogenesis. Nat Rev Mol Cell Biol 8, 464-478

52. Vestweber, D. (2008) VE-cadherin: the major endothelial adhesion molecule controlling cellular junctions and blood vessel formation. Arterioscler Thromb Vasc Biol 28, 223-232

53. Lin, M. T., Yen, M. L., Lin, C. Y., and Kuo, M. L. (2003) Inhibition of vascular endothelial growth factor-induced angiogenesis by resveratrol through interruption of Src-dependent vascular endothelial cadherin tyrosine phosphorylation. Mol Pharmacol 64, 1029-1036

54. Strilic, B., Kucera, T., Eglinger, J., Hughes, M. R., McNagny, K. M., Tsukita, S., Dejana, E., Ferrara, N., and Lammert, E. (2009) The molecular basis of vascular lumen formation in the developing mouse aorta. Dev Cell 17, 505-515

55. Fang, P., Yu, X., Jae Jeong, S., Mirando, A. C., Chen, K., Chen, X., Sunghoon, K., Francklyn, C. S., and Guo, M. Structural basis for full-spectrum inhibition of translational functions on a tRNA synthetase. Nature communications In press, 1- 11

56. Kawamura, T., Liu, D., Towle, M. J., Kageyama, R., Tsukahara, N., Wakabayashi, T., and Littlefield, B. A. (2003) Anti-angiogenesis effects of borrelidin are mediated through distinct pathways: threonyl-tRNA synthetase and caspases are independently involved in suppression of proliferation and induction of apoptosis in endothelial cells. J Antibiot (Tokyo) 56, 709-715

57. Habibi, D., Ogloff, N., Jalili, R. B., Yost, A., Weng, A. P., Ghahary, A., and Ong, C. J. (2011) Borrelidin, a small molecule nitrile-containing macrolide inhibitor of threonyl-tRNA synthetase, is a potent inducer of apoptosis in acute lymphoblastic leukemia. Invest New Drugs 30, 1361-1370

58. Wellman, T. L., Jenkins, J., Penar, P. L., Tranmer, B., Zahr, R., and Lounsbury, K. M. (2004) Nitric oxide and reactive oxygen species exert opposing effects on the stability of hypoxia-inducible factor-1alpha (HIF-1alpha) in explants of human pial arteries. FASEB J 18, 379-381

59. Ahmed, S. A., Gogal, R. M., Jr., and Walsh, J. E. (1994) A new rapid and simple non-radioactive assay to monitor and determine the proliferation of lymphocytes: an alternative to [3H]thymidine incorporation assay. J Immunol Methods 170, 211-224

60. Arnaoutova, I., and Kleinman, H. K. (2010) In vitro angiogenesis: endothelial cell tube formation on gelled basement membrane extract. Nat Protoc 5, 628-635 122 61. Ribatti, D., Nico, B., Vacca, A., and Presta, M. (2006) The gelatin sponge- chorioallantoic membrane assay. Nat Protoc 1, 85-91

62. Kimmel, C. B., Ballard, W. W., Kimmel, S. R., Ullmann, B., and Schilling, T. F. (1995) Stages of embryonic development of the zebrafish. Dev Dyn 203, 253-310

123 CHAPTER 4: INVESTIGATING THE MECHANISMS OF THREONYL-tRNA

SYNTHETASE’S ANGIOGENIC FUNCTIONS

4.1. Introduction

Angiogenesis is a multifaceted process critical to the formation and metastatic progression of many human cancers (1). Many of the most prominent angiogenic factors, such as vascular endothelial growth factor (VEGF), have been investigated as potential targets for anti-metastatic therapeutics with several drugs already approved for use (2,3).

Unfortunately, the benefits of these therapies are often temporary owing to adaptive responses by the target tissue (4). Therefore, the identification of novel angiogenic factors and small molecule inhibitors would be beneficial for the development of new diagnostics and combinatorial cancer therapies.

The pro-angiogenic functions of TARS were recently demonstrated through studies led by Tamara Williams in the lab of Karen Lounsbury for which we provided essential enzymes and compounds for testing (5). The addition of exogenous TARS was found to stimulate human umbilical vein endothelial cell (HUVEC) tube formation and vascular growth in chorioallantoic membrane (CAM) assays. As angiogenesis is associated with both proliferation and migration of endothelial cells, the effect of TARS on these two aspects were subsequently investigated. Specifically, the addition of TARS to the lower chamber in a transwell migration assay enhanced the movement endothelial cells from the upper level relative to media alone; however, the enzyme had no effect on the proliferation of these cells as determined by an alamarBlue® assay. In all cases, the angiogenesis-associated responses were subject to inhibition by subtoxic doses of the

124 borrelidin derivative, BC194. Interestingly, the borrelidin-resistant TARS mutant, L567V

TARS (See Chapter 2, 2.2.5) also exhibited comparable angiogenic activity to the wildtype enzyme but did not respond to BC194 inhibition, suggesting that the effects of

BC194 were mediated directly through the compound’s interactions with TARS. In contrast, exposure of CAM assays to the control ARS, leucyl-tRNA synthetase (LARS), had no effect on vessel formation, which indicates that this secondary function is not shared by ARSs in general. Given that all of the above observations were in response to extracellular TARS, the ability of HUVEC cultures to secrete the ARS was also examined. Notably, the usually cytoplasmic TARS was identified in conditioned media from endothelial cells treated with TNFα or VEGF (5). This mechanism is reminiscent of the angiogenesis-associated ARSs tyrosyl- (YARS) and tryptophanyl- (WARS) tRNA synthetases, which are also secreted from the cell in response to apoptotic or interferon-γ signaling respectively (6,7). However, the exact mechanism through which extracellular

TARS regulates this process remains unclear.

Investigations of the angiogenesis-inhibitory effects of borrelidin provide some clues to the mechanism of TARS-mediated angiogenesis. Early reports attributed borrelidin’s inhibition of endothelial cells to the induction of apoptosis (8). However, similar mechanisms have also been proposed for the apparent anti-tumor activities of borrelidin observed in both breast cancer and leukemia cells (9,10). Therefore, these apoptotic effects appear to be related to the general borrelidin toxicity responses demonstrated in Chapter 3. Alternatively, the anti-angiogenic properties could be specifically associated with defects in vessel patterning and maturation as demonstrated by the zebrafish studies. The observation that the L567V point mutation is able to confer

125 resistance to the compound in both aminoacylation (see Chapter 2, section 2.2.5) and angiogenesis procedures implicates TARS as the target for macrolide induced anti- angiogenic responses (5). Therefore, better understanding of TARS’s specific effect on angiogenesis would provide opportunities for its use as a target of angiogenic based therapies.

The data presented in Chapter 3 revealed that treatment of zebrafish with BC194 induces defects in vessel patterning that are distinct from the toxicity responses associated with borrelidin exposure. The lack of alternative targets for borrelidin coupled with the resistance conferred by the L567V TARS mutation in in vitro assays suggests that the compounds mediate these effects through direct inhibition of TARS’s secondary function (5). Knockdown of TARS in zebrafish using a splice-altering morpholino produces similar vascular defects to borrelidin and BC194 inhibition, confirming a role for TARS in vessel guidance in the developing zebrafish embryo. However, the mechanisms that control these effects remain a mystery. Using various point and truncation mutations, TARS’s stimulation of angiogenesis is shown to be independent of its aminoacylation function but requires a full-length protein. Finally, mutations that mimic borrelidin-induced conformational changes are shown to eliminate TARS’s angiogenic potential and strongly implicate the importance of protein-protein interactions in this secondary function.

4.2 Results

4.2.1 Knockdown of threonyl-tRNA synthetase expression in zebrafish morphants produces a vascular phenotype

126

Figure 23. Depletion of TARS elicits intermediate toxicity relative to borrelidin and BC194 inhibtion. (A) RT-PCR to validate the reduction of TARS levels by morpholino injection. Whole fish mRNA (72 hpf) was trizol/chloroform extracted and mRNA sequences amplified by RT-PCR with gene specific primers. The TARS (338 bp) and EF1α control amplicons were run on 1.5% agarose to show differences in relative intensities between uninjected controls (UIC) and morphant fish. (B) Whole body images of UIC and TARS morphant fish at 48 and 72 hpf.

The zebrafish data from Chapter 3 clearly demonstrate distinct toxicity and anti- angiogenic effects in response to borrelidin and BC194 treatment. Furthermore, the angiogenesis specific results appear to be related to the induction of abnormal vessel

127

Figure 24. TARS is involved in vascular development in the zebrafish. (A) Representative confocal images (20x magnification) of control and TARS morphant zebrafish ISVs at 48 and 72 hpf. Embryos were injected with a TARS morpholino (1.5 µM) at the one to four cell stage. Fish were manually dechorionated at 24 hpf and imaged at 48 and 72 hpf. Arrows and asterisks denote the location of ectopic branches and missing/incomplete vessels respectively. (B, C) Quantification of ectopic branching (B) and missing/incomplete ISVs (C) from control and morphant zebrafish within a region encompassing five ISVs anterior and posterior to the yolk extension ; mean ± SEM, n≥29, #*p<0.0001 relative to uninjected controls at 48 and 72 hpf respectively (one-way ANOVA, Tukey Test). patterning. However, the mechanism behind these patterning defects remains to be determined. Specifically, TARS has been shown to be the target of borrelidin in in vitro assays but has not been proven to be responsible for the effects in zebrafish (5). To confirm that the vessel patterning effects of borrelidin and BC194 in zebrafish in Chapter

3 are a direct consequence of action on TARS, the levels of functional TARS were reduced using a splice altering, antisense morpholino oligonucleotide (MO) (Figure 23A).

128 Consistent with the hypothesis that TARS was responsible for the vascular effects of the borrelidin compounds, morphant fish developed with significantly more ectopic branches and incomplete vessels than the uninjected controls (Figure 24A-C). The severe morphological defects associated with borrelidin toxicity (Figure 23B) were not seen in the TARS morphant fish, but their phenotypes exhibited characteristics that were intermediate with respect to borrelidin and BC194 treatments. These results demonstrate that vascular development in the zebrafish is highly sensitive to alterations in the activity of TARS, induced either as a result of chemical inhibition or altered gene expression.

Additionally, these data provide strong evidence that the vascular effects of the borrelidin and BC194 are mediated specifically through the inhibition of TARS.

4.2.2 Necessary components of threonyl-tRNA synthetase-mediated angiogenesis

Consistent with previous reports, Chapter 3 demonstrated a clear role of TARS in angiogenesis (5). This function was further characterized in investigations of borrelidin and its derivatives. Specifically, the fact that the L567V point mutation in TARS was found to confer resistance to these compounds coupled with the recapitulation of borrelidin and BC194 treatment effects in fish through knockdown of TARS indicates that the effects of these compounds on angiogenesis must occur through modification of

TARS’s functions (5). Furthermore, these effects can be separated from amino acid starvation, and would suggest that TARS functions through a more direct mechanism.

However, the specific structural features of TARS involved in angiogenesis remain to be determined.

129

Figure 25. Linear schematics of TARS protein constructs. Linear schematics showing the TARS domains as they appear in the various protein constructs, specifically TGS (N1) (Purple), Editing (N2) (Red), catalytic (green), and anticodon binding (ACBD) (Blue) domains. Amino acid numbers and point mutation locations are indicated in the figure.

To better understand the features of TARS involved in its secondary function, the angiogenic potentials of various TARS constructs (Figure 25) were investigated using

HUVEC tube-formation assays. Unless noted otherwise, the catalytic activity for all constructs were determined by the aminoacylation assay and the resulting rates reported in Tables 3 and 4. As a baseline the pro-angiogenic TARS and non-angiogenic LARS were tested (Figures 26A-C). Similar to earlier reports, WT TARS significantly increased

130

Figure 26. Characterizing the angiogenic potential of TARS and related protein constructs. (A) Quantification of HUVEC branching in response to WT TARS, R442A TARS, Ec ThrRS, N1-TARS, ΔN1N2-TARS, and LARS exposure. HUVECs were plated on Matrigel in 10% EGM-2 (0.2% FBS) media and exposed to 100 nM of the indicated constructs. Cells were fixed after 4-8 h and stained with Oregon Green 488 Phalloidin. Numbers represent percentage of control (0 nM) samples, mean ± SEM, n=3 *,**, and **** correspond to p<0.05, 0.01, and 0.0001 respectively relative to the media control (one-way ANOVA, Tukey Test). (B-D) Activity data for WT TARS, Ec ThrRS, R442A TARS, LARS and buffer control. WT TARS (B, D), Ec ThrRS (B) and R442A TARS (C) were determined as formation of Thr-tRNAThr per active site per minute; LARS was determined as AMP per LARS per minute. WT TARS was included in both graphs for comparative purposes; mean ± S.D.; n=3 (n=1 for Ec ThrRS). tube-formation of endothelial cells relative to untreated cultures (Figure 26A).

Surprisingly, the control ARS, LARS, exhibited a minor reduction of tube formation, although the cause of this effect remains unknown. It should also be noted that LARS activity was determined as the rate of adenylate formation (using AMP formation as the 131 readout) due to the lack of a viable tRNALeu purification protocol. Effectively this reaction measures the rate of formation and subsequent breakdown of the leucyl- adenylate intermediate, which occurs slowly in absence of tRNALeu. While the acquired rate data cannot be directly compared to the aminoacylation rate data of the other constructs they clearly demonstrate that the negative control enzyme used in the tube- formation assays was active for aminoacylation, its primary function.

Table 3. Aminoacylation activity data for various protein constructs Construct Activity (min-1) % Activity* WT TARS 3.325 ± 0.2083 100 R442A TARS Below detectable limit -- Ec ThrRS 55.30 ± 5.594 1663 LARS‡ 0.1140 ± 0.007845 n/a *Relative to WT TARS ‡ LARS activity was measured as AMP formation and is not directly comparable to aminoacylation data

4.2.3 TARS-mediated angiogenesis is independent of aminoacylation activity

Chapter 3 demonstrated that the toxicity effects of borrelidin and related compounds were the result of amino acid starvation from loss of aminoacylation activity.

Conversely, the effects of angiogenesis were observed at sub-toxic concentrations of the drug. Together these data appear to indicate that TARS’s angiogenesis function is not dependent on aminoacylation. However, there is still a possibility that angiogenesis is more sensitive to a reduction of aminoacylation or that aminoacylation is one of several contributing factors. To confirm that angiogenesis was separate from the aminoacylation function of TARS, HUVEC tube-formation assays were performed using the R442A

TARS. Arginine 442 (R363 in E. coli) forms key stabilization interactions with the

132 threonyl-adenylate and its mutations would be predicted to eliminate all aspects of TARS aminoacylation function (11). Consistent with this hypothesis, R442A TARS exhibited no adenylate formation (data not shown) or aminoacylation activity (Figure 26D; Table

3). Despite this lack of catalytic activity, treatment of HUVEC cells with R442A TARS was found to significantly increase tube formation relative to the buffer control (Figure

26A). Moreover, these results were comparable to those received for WT TARS.

The angiogenic potential of R442A TARS was further supported through investigations by Dr. Tamara Williams using the in vivo chorioallantoic membrane assay.

The addition of 1 μM R442A TARS soaked sponges to the membrane surface of chicken embryos stimulated the growth of surface vessels towards the sponge significantly more than PBS controls, much like has been previously observed with WT TARS (Figure 27).

Similar results were not obtained with LARS, again suggesting this effect was not a general function of ARSs. The retention of angiogenic activity in R442A TARS clearly demonstrates that TARS aminoacylation function is not required for angiogenesis. As a side note, co-treatment of cells with 1 μM R442A TARS and 1 μM BC194 (a non-toxic concentration in CAM assays) attenuated R442A TARS-mediated angiogenesis. These data further support the findings in Chapter 3 that indicate separate inhibitory mechanisms for borrelidin-derivatives against TARS aminoacylation and angiogenesis activities. Therefore, these findings implicate borrelidin-mediated conformational changes as the cause of its anti-angiogenic activity and will be further investigated in

Section 4.2.6.

133

Figure 27. Characterizing the angiogenic potential of TARS and related protein constructs in chorioallantoic membrane assays. The effects of 1 μM R442A TARS with and without 1 μM BC194 and LARS on in vivo angiogenesis in a CAM assay. Fertilized chicken embryos were cultured ex-ova for 10 days after which compounds at the indicated concentrations were applied to gelform sponges on the CAM. Images were taken daily over 72 h and quantified as the change in vascularity score over this entire period; mean ± SEM, n≥12, ****p<0.0001 relative to PBS (one-way ANOVA, Tukey Test). Data were kindly provided by Drs. Tamara Williams and Karen Lounsbury.

4.2.4 TARS-mediated angiogenesis requires full-length constructs

As indicated in Chapter 1, section 1.6.4, all other ARSs possessing angiogenic functions do so through mechanisms requiring an appended domain or sequence

(reviewed in Chapter 1). While no specific domain has been directly associated with

TARS, it does contain an N-terminal 82 amino acid extension predicted to be unstructured followed by an N-terminal ubiquitin like domain of unknown function, known as the TGS domain. The TGS domain structure is conserved from bacteria to humans despite very different sequences. Conversely, the amino acid extension appears to have originated during eukaryotic evolution and is not found in prokaryotic versions of the enzyme. As such, it may be possible that this N-terminal region, known as the N1- 134 domain, is important for the TARS’s angiogenic function. Therefore, the importance of both domains was investigated using the HUVEC tube formation assay with the isolated

N1-TARS (TGS domain and the 82 amino acid extension), the ΔN1N2-TARS (isolated aminoacylation and anti-codon binding domains), and the Ec ThrRS constructs.

As the N-terminal 82 amino acid extension in only found in eukaryotic versions of the enzyme, the necessity of this domain was investigated using Ec ThrRS. Like other constructs, Ec ThrRS’s aminoacylation activity was investigated (Figure 26B).

Interestingly, treatment of HUVEC cultures with the E. coli enzyme exhibited comparable stimulation of angiogenesis to human TARS (Figure 26A), indicating that the

N-terminal extension is not required for TARS-mediated angiogenesis (at least once extracellular). Conversely, exposure of HUVECS to N1-TARS had no significant effect on tube-formation relative to the negative, buffer control but was significantly lower in comparison to full-length TARS (Figure 26A). As such these data would suggest that the

N1-domain alone is not capable of stimulating angiogenesis. This is not completely surprising as Chapter 1 clearly demonstrated that borrelidin interacts with the TARS catalytic domain likely implicating that this domain is essential for angiogenic function.

Therefore, the ability of ΔN1N2-TARS, which only possesses the isolated catalytic and anticodon-binding domains, to induce angiogenesis was investigated. Interestingly, treatment with the ΔN1N2-TARS construct produced significantly less tube structures than both the negative control and full-length TARS (Figure 26A), suggesting that

ΔN1N2-TARS may even be inhibitory towards angiogenesis. Together, these data suggest that the entire TARS protein is required for the extracellular stimulation of angiogenesis with the exception of the eukaryotic, N-terminal extension.

135

4.2.5 Mutants that mimic the borrelidin-bound conformation eliminate TARS- angiogenic activity

The clear separation of aminoacylation and angiogenesis functions in TARS by experiments in Chapter 3 and with R442A TARS above led to the hypothesis that the anti-angiogenic properties of borrelidin-related compounds are due to their ability to stabilize a non-permissive conformation of the enzyme. In support of this mechanism, intrinsic tryptophan fluorescence studies demonstrate that the binding of borrelidin induces conformational changes within the TARS structure (also shown in (12)) suggesting that the anti-angiogenic properties of the compound could function through the disruption of TARS-protein interactions. Additionally, our crystallographic data (see

Chapter 2) show the extrusion of a surface helix in the TARS aminoacylation domain by

14° in both the borrelidin and BC194 bound complexes and agrees with our tube formation and CAM data that suggest similar anti-angiogenic potential for both compounds. To confirm the importance of this helix for TARS-mediated angiogenesis, the point mutations Q566W TARS and A538W TARS were generated because they were predicted by molecular modeling programs to mimic the borrelidin-bound conformation

(Figure 28A) (13). Initial activity investigations revealed that Q566W TARS possessed aminoacylation activity greater than WT TARS (Figure 28B; Table 4) while A538W

TARS was inactive in active site titration assays and was not investigated for aminoacylation activity. Interestingly, neither mutant increased tube formation relative to the buffer control, suggesting that this conformational change alone is enough to inhibit angiogenesis (Figure 28C).

136

Figure 28. The effects of borrelidin-conformation mimetic mutations on TARS-mediated angiogensis. (A) Schematic of the borrelidin-induced conformational change in the TARS active site. Borrelidin (BN, orange) binding opens up a region at the edge of the aminoacylation active site consisting of residues 492-547 (designated by arrow). Similar conformational changes are observed with Q566W and A538W mutations (positions in active site cleft designated by blue lines). Inset is a superposition of the TARS-borrelidin structure (TARS-BN; TARS shown in teal and BN in orange) superimposed with a computational model of the Q566W TARS (blue) showing nearly identical structures. (B) Quantification of HUVEC branching in response to WT TARS, Q566W TARS, and A538W TARS. HUVECs were plated on Matrigel in 10% EGM-2 (0.2% FBS) media and exposed to 100 nM of the indicated constructs. Cells were fixed after 4- 8 h and stained with Oregon Green 488 Phalloidin. Numbers represent percentage of control (0 nM) samples, mean ± SEM, n=3 * p<0.05 and # p<0.01 relative to the buffer control and WT TARS respectively (one-way ANOVA, Tukey Test). (C) Aminoacylation activity data for WT TARS and Q566W TARS as determined by formation of Thr-tRNAThr per active site per minute; mean ± S.D.; n=3. Diagrams in (A) were adapted from images kindly provided by our collaborator Min Guo.

137 Table 4. Activity data for borrelidin conformation mimetic mutations Construct Activity (min-1) % Activity* WT TARS 3.325 ± 0.2083 100 Q566 TARS 9.151 ± 0.3729 275 A538W TARS‡ Below detectable limit -- * Relative to WT TARS ‡ As A538W TARS showed no adenylate formation in active site assays it was not tested for aminoacylation activity.

4.3 Discussion

4.3.1 Borrelidin regulates angiogenesis in zebrafish specifically though interactions with TARS

Early reports suggested that apoptosis of endothelial cells could contribute to the reduction in angiogenesis (8). While the impact of this phenomenon cannot be completely ruled out, it is complicated by a lack of specificity to a single cell type (9,10).

Furthermore, the data from Chapter 3 clearly indicate that anti-angiogenic effects are observable at sub-toxic levels suggesting that another mechanism is required to explain the relationship between TARS and borrelidin in angiogenesis. The actions of borrelidin and BC194 can be blocked by single residue substitutions in the E. coli ThrRS, and

TARS active sites, respectively (5,12), arguing against the previous suggestion that borrelidin binding to other secondary targets is physiologically significant in the inhibition of angiogenesis (14). This observation was shown to extend into whole animal models as demonstrated by the similar phenotypes exhibited between BC194 and borrelidin exposure compared to the knockdown of TARS through treatment with a

TARS-specific morpholino. As such, the effects of compound treatment on zebrafish in

Chapter 3 can be attributed to the direct inhibition of TARS. 138

4.3.2 Defining the TARS components required for the regulation of angiogenesis

Structural analysis of the ThrRS-borrelidin complex suggests that the inhibition mechanisms of borrelidin and BC194 involve occlusion of the binding sites for one or more of the canonical aminoacylation substrates. Borrelidin is a slow tight binding inhibitor that competes with amino acid substrate binding, accounting for the potent stimulation of toxic responses related to amino acid starvation (12). This toxicity response, however, has been demonstrated to be completely separate from the angiogenic specific effects. Conversely, the angiogenic effects exhibited by the various TARS constructs help to characterize the mechanism of TARS-mediated angiogenesis. Since Ec

ThrRS was capable of stimulating angiogenesis it would seem that the N-terminal ~82 amino acid extension found only in higher organism is not required for angiogenesis.

However, it’s important to note that the tube formation assay only looks at the effects of constructs outside of the cell and would not provide any evidence for intracellular events.

As such, the N-terminal extension maybe serve an important regulatory role or function in the currently unknown TARS secretory mechanism that is unnecessary in E. coli. The lack of structure in this region suggests a potential role as a protein-protein interaction domain since intrinsically disordered regions allow for versatile interactions with interaction partners (15,16). However, the actual function of this domain, if any, remains to be established.

The loss of angiogenesis activity by Q566W TARS and A538W TARS clearly demonstrates the importance of borrelidin’s conformational change to the mechanism of the compound. Since Q566W TARS retains its aminoacylation activity it seems unlikely

139 that this effect is due to large scale structural effects or local unfolding linked to the mutant substitution. Moreover, the fact that this conformational change does not affect the normal function of the protein raises the possibility that its inhibition of angiogenesis is though the disruption of some currently unknown protein-protein interaction. This concept is further supported by the R442A TARS tube formation assays since these data emphasize that TARS’s aminoacylation function is not required for angiogenesis.

The TARS-mediated mechanism of angiogenesis can be further characterized by the data from the N1-domain and ΔN1N2-TARS constructs. Since neither of these constructs appear to stimulate angiogenesis, these data suggest that the full-length protein is required. The use of ΔN1N2-TARS for crystallization further indicates that this construct retains its three dimensional structure and has not lost its angiogenic activity owing to partial denaturation. Interestingly, the Q566 mutation and its associated conformational change are located within the aminoacylation domain, suggesting that this construct retains the ability to engage in the protein-protein interactions that we hypothesize borrelidin disrupts. Consistent with this interpretation ΔN1N2-TARS reduced the number of tube like structures that grew in HUVEC culture, suggesting that the truncated construct likely titrates potential interactions made by the full length protein. Moreover, the necessity of the full length protein indicates that a second interaction with the N-terminus is likely essential for TARS angiogenic function. As described above, the ability of Ec ThrRS to stimulate angiogenesis suggests that this interaction is not mediated by the N-terminal extension and is more likely to be associated with the N1 (TGS) and N2 (Editing) domains of the protein.

140 Unlike the ΔN1N2-TARS construct, the N1 domain alone did not inhibit angiogenesis in the tube-formation assay. However, the data for this construct were associated with a high standard error of mean (Figure 26A), which may indicate that this domain has some influence in the assay and its involvement in angiogenesis should not be completely ruled out. As the N1-domain does not possess a known active site, its contribution, if any, would be limited to the formation of interactions between TARS and other molecules. In this model TARS would be able to bridge two molecules together through the formation of concurrent interactions (further developed in section 4.3.3).

Conversely, the N2 domain contains the editing active site and could theoretically function in both the formation of interactions and non-canonical catalytic activities

(alternative catalytic functions of TARS are discussed in Chapter 5) (17). However, unlike the other domains of TARS, the contributions of the isolated N2 domain on angiogenesis have yet to be investigated.

One important caveat of these data is that neither the N1-domain nor the ΔN1N2-

TARS constructs contain the N2 editing domain. This apparent oversight is due to the fact that both constructs were designed with other purposes in mind: the N1-domain was designed for protein-protein interaction studies (see Appendix I) and the ΔN1N2-TARS was designed for crystallization. Given the proven importance of the aminoacylation domain to the angiogenic mechanism it would seem unlikely that a construct consisting of both the N1 and N2 domains would have any pro-angiogenic activity. Conversely, if the N2 domain is the site of the second interaction, then a ΔN1-TARS (containing only the editing, aminoacylation, and anti-codon binding domains) could possibly exhibit angiogenic activity. Due to technical difficulties in the creation of the ΔN1-TARS

141 construct, however, these studies have yet to be performed. Nonetheless, the conclusion that two interactions are necessary would become very important in directing future investigations of this mechanism.

4.3.3 Models for TARS-mediated angiogenesis and their disruption by borrelidin derivatives

Together, the zebrafish morpholino and TARS construct data support a mechanism in which TARS regulates angiogenesis by stimulating proper vessel patterning and migration independent of TARS’s aminoacylation function. This process appears to be based on the formation of two separate protein-protein interactions, one of which is located within the TARS catalytic domain. Possible identities for these interactions have been investigated using mass-spectrometry approaches and the results of these experiments can be observed in Appendix I. The methods used in those interaction experiments, however, focused primarily on intra-cellular partners and may not fully account for the apparent extracellular nature of TARS’s secondary function (5).

In contrast, the models presented below will focus on possible extracellular binding partners for TARS.

One possible explanation for this two-binding site model is that extra-cellular

TARS establishes a spatially regulated gradient through interactions with the extracellular matrix. The second site is then able to mediate a contact with a cell surface receptor involved in vessel migration and guidance. In support of this hypothesis, interactions between TARS and fibronectin 1 (FN1), a major component of endothelial cell ECM, have been reported in the literature. This possible mechanism of TARS-

142 mediated angiogenesis is currently being investigated and will be discussed with more detail in Chapter 6.

In this model, the conformational change of Q566W TARS or treatment with borrelidin would disrupt either TARS’s interaction with the ECM or the cellular receptor.

In the case of the ECM interaction, the TARS gradient would become more diffuse and harder for vessels to follow. Likewise, disruption of the TARS-receptor interaction would inhibit the perception of TARS by cells within a growing vessel. In either event, the vessels would not receive the proper guidance cues, leading to the appearance of ectopic vessels. This may explain the directional confusion displayed by in vivo assays, such as the ISV patterning and CAM vessel formation, following borrelidin and BC194 treatment since the diffuse TARS gradient would not allow for proper polarization of endothelial cells. This explanation is also compatible with our cell-based assays since each has a matrix component associated with the system. Previously, other angiogenesis-associated factors, such as VEGF, may elicit different signaling depending on if they are free or bound to the ECM. Therefore, disruption of TARS-ECM or receptor interactions could have profound effects on the regulation and specifics of its signaling activity. Currently, attempts to identify TARS-responsive receptors and borrelidin-susceptible interactions are underway.

While the substantial cross-talk between receptors linked to angiogenesis makes it difficult to verify a particular pathway, circumstantial evidence implicates the toll like receptors (TLR) of the innate immune system as possible effectors. As mentioned in

Chapter 1, TARS is the second most common of six ARS associated with anti-synthetase syndrome, a rare autoimmune disorder that includes polymyositis and dermatomyositis

143 resulting from the production of the anti-ARS autoantibodies. Recently, the most common ARS involved in this disorder, histidyl-tRNA synthetase (HARS), was demonstrated to induce muscle inflammation in mice by a TLR2- and TLR4-dependent mechanism (18). Activation of the TLRs has been shown to contribute to the process of angiogenesis, in part, through MyD88-mediated activation of the NF-κB transcription factor (19). This pathway up-regulates the production of TNFα and, consequently, VEGF

(20). As indicated earlier, TNFα and VEGF signaling stimulate the secretion of TARS from endothelial cells (5). Furthermore, TLR-induced expression of IL-10 in macrophages has been associated with the activation of STAT3, which binds to a specific site within the promoter region of the tars gene (21). Taken together these observations would describe a scenario in which TARS-mediated stimulation of TLRs would initiate pro-angiogenic responses and propagate the signal through the expression and secretion of additional TARS.

Other possible receptor identities include focal adhesion complexes and purinergic signaling. Focal adhesion complexes are dynamic, multi-protein structures that mediate contacts between a cell and the surrounding ECM and possess well documented roles in migration (22). Interactions with ECM components are established by integrins and can signal into the cell via complex associated kinases (23,24). The co-clustering of growth factor associated receptors with integrins have also been observed and may collaborate with integrin signaling by interacting with ECM bound ligands, in this case

TARS. Alternatively, the purinergic system responds to extracellular nucleotides and became of interest upon the discovery of non-canonical, catalytic functions in TARS. The binding of nucleotides to these purinergic receptors has been demonstrated to elicit

144 functions associated with angiogenesis, including migration and proliferation of endothelial cells (25-27). The catalysis of these reactions by TARS and the possible involvement of this receptor system will be discussed thoroughly in the next chapter.

4.4 Experimental Procedures

4.4.1 Cell Culture and Reagents

Human umbilical endothelial cells (HUVECs) were maintained in Clonetics

EGM®-2 complete media (Lonza) and used between passages 5 to 8. Borrelidin and

BC194 were obtained from Biotica (now Isomerase Therapeutics Ltd (Cambridge, UK)).

The TARS morpholino was purchased from GeneTools.

4.4.2 Protein Preparation

Purification of human TARS, Ec TARS, and ΔN1N2-TARS was performed identical to the procedures described in Chapter 2, section 2.3.1. Single amino acid mutants were generated using identical procedures following QuickChange (Stratagene) site-directed mutagenesis with olio-deoxynucleotide primers from Operon.

Preparation of human cytoplasmic LARS was adapted from previously described methods (28,29). A plasmid encoding for N-terminal His6-tagged human cytoplasmic LARS in the pPROEX hTb expression vector were transformed into E. coli

B21(DE3) codon plus competent cells (Stratagene) and cultured in terrific broth supplemented with 100 μg/ml ampicillin 37°C to a cell density of A600 = 0.6. Expression of LARS was induced with 1 mM isopropyl 1-thio-β-D-galactoside overnight at room temperature. The bacterial pellet was lysed by sonication in buffer A (20 mM potassium

145 phosphate buffer pH 8.0, 150 mM KCl, and 5 mM β-mercaptoethanol) and cleared by centrifugation at 17050 x g for 30 minutes. Nucleic acids were precipitated by the addition of protamine sulfate to a final concentration of 0.2% followed by centrifugation.

The supernatant was loaded onto a HisTrapTM FF column (GE Healthcare) in buffer A, washed in 15 mM imidazole, and eluted by an imidazole gradient from 15 to 500 mM over 20 column volumes. LARS containing fractions were identified by SDS-PAGE and

GelCodeTM Blue (Thermo Scientific), pooled, and dialyzed into buffer B (20 mM HEPES pH 8.0, 50 mM KCl, 5 mM β-mercaptoethanol, and 10% glycerol). The sample was loaded onto a HiTrap Q FF column (Amersham Pharmacia Biotech) and eluted over 20 column volumes by KCl gradient from 50 to 1000 mM. LARS containing fractions were determined by SDS-PAGE, dialyzed into storage buffer (100 mM potassium phosphate pH 6.8, 10 mM β-mercaptoethanol, and 50% glycerol), and stored at -20°C.

4.4.3 In Vitro Tube Formation Assay

Tube formation assays were performed as previously described (5,30). Briefly

HUVEC cells were cultured until 95-100% confluent and seeded in to 48-well plates coated with MatrigelTM (BD Biosciences). Cells were incubated at 37 °C in 10 %

EGM®-2 media in EBM®-2 media (Lonza) (ie. 0.2% FBS and and 10% normal supplement concentrations of EGM®-2). Various protein constructs were then added for

4 to 8 h. Cells were fixed with 10% neutral buffered formalin, stained with Oregon Green

488 phalloidin (Molecular Probes), and imaged via fluorescence microscopy. Tube structures were quantified using the Simple Neurite Tracer plug-in on ImageJ software

(NIH).

146

4.4.4 Chick Chorioallantoic Membrane Assay

CAM assays were performed as described in Chapter 3, section 3.4.6.

4.4.5 Zebrafish Handling and Treatments

This project utilized the transgenic line Tg(flk1:eGFP) in which eGFP is expressed in 48 hpf and controlled by the flk1 promoter. Zebrafish were bred and maintained in accordance with approved University of Vermont IACUC standards for vertebrate animals and staged as previously described (31). An anti-TARS morpholino

(5’-TAATTTAGCCACTAACCTGATCCCA-3’) was designed targeting the splice donor site within exon 3 of TARS mRNA. Blocking this site is predicted to join exon 2 to exon

4, introducing a frameshift and producing a stop codon five codons into exon 4. TARS morpholinos (1.5 mM) were injected into the yolk sac of Tg(flk1:eGFP) embryos at the one to four cell stage and incubated for 24 hours at 28.5 °C. The embryos were then dechorionated, transferred to fish egg water containing 0.003% PTU, and replaced at 28.5

°C. The successful knockdown of functional TARS was confirmed by RT-PCR using the following primers:

Exon 2, Forward: 5’-CAAGAATGCTGCTGGAGATG-3’

Exon 4, Reverse: 5’-CTTGTGCCTCTTCATCGTCA-3’

Imaging and aberrant vessel quantification were performed at 48 and 72 hpf as described below.

4.4.6 Whole Fish and Confocal Imaging

147 Whole body and confocal images were obtained using procedures identical to those described in Chapter 3, sections 3.4.10 and 3.4.11.

4.4.7 Statistical Analysis

148 4.5 References

1. Carmeliet, P., and Jain, R. K. (2000) Angiogenesis in cancer and other diseases. Nature 407, 249-257

2. Lien, S., and Lowman, H. B. (2008) Therapeutic anti-VEGF antibodies. Handbook of experimental pharmacology, 131-150

3. Morabito, A., De Maio, E., Di Maio, M., Normanno, N., and Perrone, F. (2006) Tyrosine kinase inhibitors of vascular endothelial growth factor receptors in clinical trials: current status and future directions. The oncologist 11, 753-764

4. Bergers, G., and Hanahan, D. (2008) Modes of resistance to anti-angiogenic therapy. Nat Rev Cancer 8, 592-603

5. Williams, T. F., Mirando, A. C., Wilkinson, B., Francklyn, C. S., and Lounsbury, K. M. (2013) Secreted Threonyl-tRNA synthetase stimulates endothelial cell migration and angiogenesis. Sci Rep 3, 1317

6. Wakasugi, K., Slike, B. M., Hood, J., Ewalt, K. L., Cheresh, D. A., and Schimmel, P. (2002) Induction of angiogenesis by a fragment of human tyrosyl- tRNA synthetase. J Biol Chem 277, 20124-20126

7. Wakasugi, K., Slike, B. M., Hood, J., Otani, A., Ewalt, K. L., Friedlander, M., Cheresh, D. A., and Schimmel, P. (2002) A human aminoacyl-tRNA synthetase as a regulator of angiogenesis. Proc Natl Acad Sci U S A 99, 173-177

8. Kawamura, T., Liu, D., Towle, M. J., Kageyama, R., Tsukahara, N., Wakabayashi, T., and Littlefield, B. A. (2003) Anti-angiogenesis effects of borrelidin are mediated through distinct pathways: threonyl-tRNA synthetase and caspases are independently involved in suppression of proliferation and induction of apoptosis in endothelial cells. J Antibiot (Tokyo) 56, 709-715

9. Gafiuc, D., Weiss, M., Mylonas, I., and Bruning, A. (2013) Borrelidin has limited anti-cancer effects in bcl-2 overexpressing breast cancer and leukemia cells and reveals toxicity in non-malignant breast epithelial cells. J Appl Toxicol

10. Habibi, D., Ogloff, N., Jalili, R. B., Yost, A., Weng, A. P., Ghahary, A., and Ong, C. J. (2011) Borrelidin, a small molecule nitrile-containing macrolide inhibitor of threonyl-tRNA synthetase, is a potent inducer of apoptosis in acute lymphoblastic leukemia. Invest New Drugs 30, 1361-1370

11. Torres-Larios, A., Sankaranarayanan, R., Rees, B., Dock-Bregeon, A. C., and Moras, D. (2003) Conformational movements and cooperativity upon amino acid, ATP and tRNA binding in threonyl-tRNA synthetase. J Mol Biol 331, 201-211

149 12. Ruan, B., Bovee, M. L., Sacher, M., Stathopoulos, C., Poralla, K., Francklyn, C. S., and Soll, D. (2005) A unique hydrophobic cluster near the active site contributes to differences in borrelidin inhibition among threonyl-tRNA synthetases. J Biol Chem 280, 571-577

13. Fang, P., Yu, X., Jae Jeong, S., Mirando, A. C., Chen, K., Chen, X., Sunghoon, K., Francklyn, C. S., and Guo, M. Structural basis for full-spectrum inhibition of translational functions on a tRNA synthetase. Nature communications In press, 1- 11

14. Kawamura, T., Liu, D., Towle, M. J., Kageyama, R., Tsukahara, N., Wakabayashi, T., and Littlefield, B. A. (2003) Anti-angiogenesis effects of borrelidin are mediated through distinct pathways: threonyl-tRNA synthetase and caspases are independently involved in suppression of proliferation and induction of apoptosis in endothelial cells. Journal of Antibiotics 56, 709-715

15. Dyson, H. J., and Wright, P. E. (2002) Coupling of folding and binding for unstructured proteins. Current opinion in structural biology 12, 54-60

16. Dyson, H. J., and Wright, P. E. (2005) Intrinsically unstructured proteins and their functions. Nat Rev Mol Cell Biol 6, 197-208

17. Dock-Bregeon, A., Sankaranarayanan, R., Romby, P., Caillet, J., Springer, M., Rees, B., Francklyn, C. S., Ehresmann, C., and Moras, D. (2000) Transfer RNA- mediated editing in threonyl-tRNA synthetase. The class II solution to the double discrimination problem. Cell 103, 877-884

18. Fernandez, I., Harlow, L., Zang, Y., Liu-Bryan, R., Ridgway, W. M., Clemens, P. R., and Ascherman, D. P. (2013) Functional redundancy of MyD88-dependent signaling pathways in a murine model of histidyl-transfer RNA synthetase- induced myositis. J Immunol 191, 1865-1872

19. Kawai, T., and Akira, S. (2006) TLR signaling. Cell Death Differ 13, 816-825

20. Aggarwal, B. B. (2003) Signalling pathways of the TNF superfamily: a double- edged sword. Nat Rev Immunol 3, 745-756

21. Hu, X., Chen, J., Wang, L., and Ivashkiv, L. B. (2007) Crosstalk among Jak- STAT, Toll-like receptor, and ITAM-dependent pathways in macrophage activation. J Leukoc Biol 82, 237-243

22. Sulzmaier, F. J., Jean, C., and Schlaepfer, D. D. (2014) FAK in cancer: mechanistic findings and clinical applications. Nat Rev Cancer 14, 598-610

23. Porter, J. C., and Hogg, N. (1998) Integrins take partners: cross-talk between integrins and other membrane receptors. Trends Cell Biol 8, 390-396

150 24. Somanath, P. R., Malinin, N. L., and Byzova, T. V. (2009) Cooperation between integrin alphavbeta3 and VEGFR2 in angiogenesis. Angiogenesis 12, 177-185

25. Burnstock, G. (2002) Purinergic signaling and vascular cell proliferation and death. Arterioscler Thromb Vasc Biol 22, 364-373

26. Corriden, R., and Insel, P. A. (2012) New insights regarding the regulation of chemotaxis by nucleotides, adenosine, and their receptors. Purinergic signalling 8, 587-598

27. Shen, J., and DiCorleto, P. E. (2008) ADP stimulates human endothelial cell migration via P2Y1 nucleotide receptor-mediated mitogen-activated protein kinase pathways. Circ Res 102, 448-456

28. Pang, Y. L., and Martinis, S. A. (2009) A paradigm shift for the amino acid editing mechanism of human cytoplasmic leucyl-tRNA synthetase. Biochemistry 48, 8958-8964

29. Seiradake, E., Mao, W., Hernandez, V., Baker, S. J., Plattner, J. J., Alley, M. R., and Cusack, S. (2009) Crystal structures of the human and fungal cytosolic Leucyl-tRNA synthetase editing domains: A structural basis for the rational design of antifungal benzoxaboroles. J Mol Biol 390, 196-207

30. Arnaoutova, I., and Kleinman, H. K. In vitro angiogenesis: endothelial cell tube formation on gelled basement membrane extract. Nat Protoc 5, 628-635

31. Kimmel, C. B., Ballard, W. W., Kimmel, S. R., Ullmann, B., and Schilling, T. F. (1995) Stages of embryonic development of the zebrafish. Dev Dyn 203, 253-310

151 CHAPTER 5: CHARACTERIZING THE NON-CANONICAL, CATALYTIC

FUNCTIONS OF HUMAN THREONYL-tRNA SYNTHETASE

5.1 Introduction

The canonical aminoacylation reaction proceeds through a two-step mechanism.

First, the amino acid is activated through condensation with ATP, forming an aminoacyl- adenylate intermediate. This is followed by the transfer of the activated amino acid to the

2’ or 3’-OH of the tRNA to generate charged-tRNA and AMP. Interestingly, several

ARSs have developed alterations to this core mechanism for the formation of dinucleoside polyphosphates. Excluding glycyl-tRNA synthetase, the majority of these reactions progress normally through adenylate formation but deviate in the transfer of adenosine phosphate from the adenylate to a second nucleotide triphosphate (1-3). The identity of this second nucleotide determines the identity of the dinucleotides (ie ATP 

Ap4A; GTP  Ap4G) (4).

The first reported synthesis of Ap4A by an ARS was in 1966 during investigations with lysyl-tRNA synthetase (KARS) (5). The biological significance of this find was debated over the next several decades, with potential roles in cellular proliferation and signaling events. More specific to ARS secondary functions is the connection between Ap4A synthesis and a non-canonical function of KARS in the regulation of the micropthalmia transcription factor (MITF), an important transcription factor in mast cells, osteoclasts, and melanocytes. Normally, MITF’s transcription factor activity is inhibited by the suppressor protein Hint (previously known as protein kinase C interacting protein 1; PKCI) but can be restored by the binding of diadenosine

152 oligophosphates, including Ap4A (6). Immune activation of mast cells leads to the phosphorylation and translocation of KARS into the nucleus, which binds to MITF and begins local production of Ap4A. This increase in Ap4A releases Hint from MITF, thereby allowing for the transcription of MITF-associated genes (7). Although this function of Ap4A is not directly related to TARS, it does demonstrate that Ap4A generation by ARSs is more than an accidental side reaction.

While most well-known for its role in the non-canonical function of KARS there are reports that provide possible connections of Ap4A synthesis to secondary functions in

TARS. An early report of Ap4A synthesis by TARS published in 1989 and revealed rates comparable, if not slightly faster, than those observed for KARS. Even more interesting, the phosphorylation of unidentified serine residues by cAMP-dependent kinase was shown to enhance the rate of this reaction by nearly 8-fold (1). The fact that this side reaction is influenced by a common regulatory mechanism like phosphorylation emphasizes the possibility that it may be involved in an important biological function.

Previously, extracellular TARS has been demonstrated to induce angiogenesis

(8). However, the mechanism behind this process remains to be determined. Interestingly, extracellular nucleotides have been associated with angiogenesis through receptor- mediated signaling cascades known as purinergic signaling. This raises the possibility that one component of the ability of TARS to stimulate angiogenesis is related to its ability to catalyze the interconversion of nucleotides outside of the cell, including the synthesis of Ap4A from ATP. While most extracellular nucleotides are able to elicit some form of purinergic signaling the actual responses are specific for each nucleotide. This specificity is achieved by the presence of numerous receptors with varying degrees of

153 affinity for different nucleotide species. For Ap4A, signaling events are primarily mediated by the P2X ligand-gated ion channel family. Activation of P2X leads to an influx of Na+ and Ca2+ and stimulates the arterial vasoconstriction (9). Alternatively,

P2Y1 activation by ADP leads to the release of intracellular Ca2+ stores through IP3- mediated pathways and stimulates the proliferation of vascular smooth muscles cells through ERK1/2 pathways (10). In endothelial cells, ADP has been demonstrated to induce proliferation and migration (11,12). Additionally, activation of P2Y receptors by

ADP on platelets was found to stimulate the release of VEGF and endostatin, which suggests the possibility of synergistic effects between purinergic signaling in platelets and endothelial cells on endothelial cell migration in vivo (13). Interestingly, Ap4A exhibits antagonistic activities against platelet P2Y1, emphasizing a complex interplay between various purinergic ligands.

Section 1.6.4 of Chapter 1 introduced the observation that, with the exception of

TARS, all angiogenesis-related functions of ARSs were dependent in some way on an appended domain. However, there are domains of unknown function contained within the

TARS protein. One in particular is an N-terminal, ubiquitin like domain that is conserved from bacteria to humans. The domain is called the TGS domain for its appearance in

TARS, GTPase, and SpoT proteins (14). The recurrence of this domain in GTPases is particularly interesting, since proteins, such as elongation factors, often use the catalysis of GTP to GDP to cycle between “on” and “off” states (14-16). Additionally, the protein

SpoT catalyzes the formation of guanosine 5’-(tri)diphosphate, 3’-diphosphate

((p)ppGpp), which is involved in the regulation of some bacterial genes (17,18).

Therefore, the identification of secondary guanine nucleotide activities in TARS might

154 lead to the discovery of a switch between the canonical and non-canonical functions or otherwise provide insights into their mechanisms.

The data in this chapter present several non-canonical, catalytic functions of

TARS, some of which have not been reported elsewhere. These reactions include a cryptic GTPase function, conversion of ATP to ADP, and the formation of Ap4A. They have been included in this dissertation work because of possible connections to the TARS pro-angiogenic mechanism. The rates of these reactions have been investigated using various TARS constructs in the presence and absence of borrelidin derivatives in order to draw possible connections with TARS-mediated angiogenesis. These connections will be further validated by investigating the involvement of the ADP and Ap4A receptors P2Y1 and P2X using specific inhibitors to these receptors. Together, the characterization of these various aspects of TARS non-canonical functions provides context and foundation for future investigations into their contribution to angiogenesis.

5.2 Results

5.2.1 TARS possesses GTPase activity

The conversion of GTP to GDP by TARS was monitored over time using autoradiography and thin layer chromatography to resolve the products and substrates

(Figure 29A). Rates are reported as molecules of GDP produced per active site (or per monomer in the case of R442A TARS) per minute, which results in units of min-1 (Table

5). With GTP as the only substrate, the base rate of GTPase activity for WT TARS was calculated to be 0.3033 min-1 (Figure 29B). While this rate may appear slow, it should be noted that GTPases often possess inherently low catalytic rates in the absence of GTPase

155

Figure 29. Kinetic characterization of TARS GTPase activity. (A) Representative autoradiographic images of TARS GTPase reactions under untreated (left) and 2 mM ATP (right) conditions using [α32P]-labeled GTP as a tracer. Reaction components were separated by thin-layer chromatography on polyethyleneimine-cellulose in 750 mM potassium phosphate buffer, pH 3.5. (B-E) Rate of GDP formation by purified TARS constructs under various conditions. (B) GTPase activity of 5 μM WT TARS (blue), R442A TARS (red), or Ec ThrRS (green). (C) GTPase activity of WT TARS in the presence of 0 μM (blue), 1 μM (red), 5 μM (green), or 10 μM (purple) tRNAThr. (D) WT TARS GTPase activity untreated (blue), with 2 mM threonine (red), 2 mM ATP (green), both 2 mM threonine and 2 mM ATP (adenylation, purple), or adenylation with 10 μM tRNAThr (aminoacylation, orange). (E) WT TARS GTPase activity uninhibited (blue), 10 μM BC194 (red), or 10 μM borrelidin (green). All data are presented as mean ± S.D, n=3.

156 activating proteins (GAP). Notably, Ec ThrRS exhibited no significant GTPase activity

(0.01576 min-1) suggesting that this effect is not found in bacteria. Moreover, the lack of activity in the Ec ThrRS construct demonstrates that the observed rates for GDP production in WT and R442A TARS were not simply the result of solution hydrolysis.

Table 5. Rates of GTPase activity by various TARS constructs Construct Condition Activity (min-1) % Activity* WT TARS: Uninhibited 0.3033 ± 0.01541 100 2 mM Threonine 0.3825 ± 0.03592 126 2 mM ATP 0.07097 ± 0.01488 23 Adenylation 0.03080 ± 0.01268 10 Aminoacylation 0.03434 ± 0.01167 11 10 μM Borrelidin 0.3082 ± 0.02850 102 10 μM BC194 0.3067 ± 0.02155 101 R442A TARS: Uninhibited 0.2137 ± 0.01704 70 Ec ThrRS: Uninhibited 0.01576 ± 0.01062 5 *Relative to WT TARS, uninhibited

Subsequently, the effects of the canonical substrates on the GTPase activity were investigated (Figure 29C, D). The addition of 2 mM threonine provided a minor enhancement to the rate, increasing it to 0.3825 min-1 (Figure 29C). In contrast, ATP in any combination with other canonical substrates, significantly decreased the GTPase activity: alone, 2 mM ATP dropped the rate to 0.07097 min-1, 2 mM ATP and 2 mM threonine (adenylation conditions) dropped the rate to 0.03080 min-1, and 2 mM ATP with 2 mM threonine and 10 μM tRNAThr (aminoacylation conditions) dropped the rate to 0.03434 min-1 (Figure 29D). In contrast to the other two substrates, tRNAThr had no effect when added alone at 1 μM, 5 μM, and 10 μM, corresponding to the rates 0.3485 min-1, 0.3231 min-1, and 0.3284 min-1 (Table 6).

157 Table 6. The effects of tRNAThr on the rates of WT TARS GTPase activity Construct Condition Activity (min-1) % Activity* WT TARS: 0 μM tRNAThr 0.3033 ± 0.01541 100 1 μM tRNAThr 0.3485 ± 0.02269 115 5 μM tRNAThr 0.3231 ± 0.01755 107 10 μM tRNAThr 0.3284 ± 0.01779 108 *Relative to WT TARS, uninhibited

Owing to the potential connection between the GTPase function and TARS’s secondary activities in angiogenesis, the effects of borrelidin and BC194 were also tested

(Figure 29E). Both compounds were added at concentrations of 10 μM, which would saturate the 5 μM enzyme concentration. Interestingly, neither borrelidin nor BC194 had any effect on the rate of the GTPase reaction, exhibiting rates of 0.3067 min-1 for borrelidin and 0.3083 min-1 for BC194 relative to 0.3033 min-1 for the uninhibited reaction.

5.2.2 TARS-mediated catalysis of Diadenosine Tetraphospate (Ap4A)

Two mechanisms have been reported for the formation of Ap4A. Most commonly, the reaction progresses through the canonical adenylation step of aminoacylation, forming an aminoacyl-adenylate intermediate. However, instead of transferring the amino acid to an awaiting tRNA molecule, the adenosine moiety is transferred to a second nucleotide triphosphate instead, generating a dinucleotide tetraphosphate (Figure 30A). As such, this mechanism would require that the corresponding amino acid is present to proceed. Alternatively, glycyl-tRNA synthetase has been observed to form Ap4A through an amino acid independent method. In this mechanism an oxygen of the γ-phosphate of an ATP molecule initiates a nucleophilic

158 attack on the α-phosphate of another nucleotide, producing Ap4A (Figure 30B). Most often the second nucleotide in these reactions is ATP, resulting in Ap4A; however, other combinations of nucleotides are also possible so long as at least one nucleotide is adenosine (ie. Ap4N, where N is any nucleotide) (4). Despite the previously reported observations of TARS Ap4A synthesis activity the mechanism through which this reaction proceeds has yet to be determined.

Figure 30. Ap4A reaction schematics. (A) Threonine-dependent reaction schematic requiring the formation of a threonyl- adenylate intermediated. (B) Threonine-independent reaction schematic with direct formation of Ap4A from ATP.

In order to determine the mechanisms used by TARS to synthesize Ap4A, WT

TARS Ap4A synthetase reactions were monitored in the presence and absence of threonine. The rate of Ap4A synthesis was determined using autoradiography of α- labeled ATP tracers resolved by thin layer chromatography on PEI-cellulose (Figure

159

Figure 31. Kinetic characterization of TARS Ap4A synthesis activity. (A) Representative autoradiographic images used for quantifying TARS Ap4A synthesis under untreated or 10 μM BC194 conditions. Reactions were monitored using [α32P]- labeled ATP as a tracer and each time point was spotted onto a polyethyleneimine- cellulose thin-layer chromatography plate and resolved using 3M ammonium sulfate and 2% EDTA. (B-E) Rate of Ap4A synthesis by purified TARS constructs under various conditions (B-E). (B) Ap4A formation by 5 μM WT TARS (blue), R442A TARS (red), Ec ThrRS (green), or WT TARS in absence of threonine (purple). (C) Ap4A synthesis by WT TARS in the presence of 0 μM (blue), 1 μM (red), 5 μM (green), or 10 μM (purple) tRNAThr. (D) WT TARS Ap4A formation uninhibited (blue), with 10 μM BC194 (red), or with 10 μM borrelidin (green). (E) Ap4A synthesis by WT TARS in the presence of 0 μM (blue), 1 μM (red), 5 μM (green), or 10 μM (purple) tRNAThr and 10 μM BC194. All data presented as mean ± SD, n=3.

160 31A). Rates are presented as the number of Ap4A molecules generated per enzyme per minute (Table 7). Where possible, the rates have been adjusted to represent the active concentration of enzyme.

The reactions without threonine did not deviate significantly from background, in contrast to the 0.1391 min-1 received for the threonine containing reactions (Figure

31B). These data indicated that the reaction likely proceeded though an adenylate intermediate and that all subsequent reactions should be performed in the presence of threonine. To confirm that the threonine was not simply required for structural purposes threonine-containing reactions were investigated using the adenylation deficient R442A

TARS mutant. Like the reactions lacking threonine, no significant activity was observed.

Conversely, Ec ThrRS readily catalyzed the reaction at a rate of 0.5052 min-1, which is more than three times the rate observed for the human enzyme.

Table 7. Rates of Ap4A synthesis by various TARS constructs Construct Condition Activity (min-1) % Activity* WT TARS: Uninhibited 0.1391 ± 0.007845 100 10 μM Borrelidin 0.05652 ± 0.005461 40.6 10 μM BC194 0.09595 ± 0.002774 69 No Threonine 0.0008236 ± 0.0003371 0.5 R442A TARS: Uninhibited 0.001343 ± 0.0001381 1 Ec ThrRS: Uninhibited 0.5052 ± 0.04095 363 *Relative to WT TARS, uninhibited

In addition to threonine, the effect of the tRNAThr on the reaction rate was also investigated (Figure 31C; Table 8). The presence of uncharged tRNA would be expected to inhibit Ap4A synthesis through the adenylate intermediate mechanism by allowing for the resolution of the adenylate through the canonical transfer reaction and reducing its

161 availability for Ap4A formation. Due to low concentrations of tRNAThr used in the reactions (1 μM to 10 μM) relative to the enzyme concentration of 5 μM and the long reaction duration, the presence of tRNAThr was only expected to exert minor effects, if any, as most would be converted to aminoacylated product early in the reaction’s progression. Surprisingly, the addition of 1 μM, 5 μM, and 10 μM of tRNAThr produced a dose-responsive increase in Ap4A formation rates of 0.1690 min-1, 0.1793 min-1, and

0.2528 min-1 (corresponding to 21%, 29%, and 82% increases) relative to 0.1391 min-1 without tRNAThr. The reason for this increase, however, remains unknown.

Table 8. The effects of tRNAThr on the rates of WT TARS Ap4A synthesis Construct Condition Activity (min-1) % Activity* WT TARS: 0 μM tRNAThr 0.1391 ± 0.007845 100 1 μM tRNAThr 0.1690 ± 0.008302 121 5 μM tRNAThr 0.1793 ± 0.01076 129 10 μM tRNAThr 0.2528 ± 0.01468 182 0 μM tRNAThr; 10 μM BC194 0.09595 ± 0.002774 69 1 μM tRNAThr; 10 μM BC194 0.08143 ± 0.004001 59 5 μM tRNAThr; 10 μM BC194 0.1171 ± 0.01041 84 10 μM tRNAThr; 10 μM BC194 0.1206 ± 0.01165 87 *Relative to WT TARS, uninhibited.

Previous work has demonstrated that borrelidin mediates its anti-angiogenic effects through TARS. Since Ap4A-signaling represents a possible mechanism for

TARS-mediated angiogenesis it was important to investigate the effects of borrelidin on rate of Ap4A synthesis. Like with the GTPase activity, borrelidin or BC194 were added at saturating conditions (10 μM compound compared to 5 μM enzyme) (Figure 31D). The presence of the two compounds reduced the rate of Ap4A formation to 40.5% (0.05652 min-1) and 69% (0.09595 min-1) respectively. This was not surprising as borrelidin and

BC194 are known to inhibit TARS adenylate formation (19). The effect of BC194 162 inhibition on the tRNA stimulation effect was also tested (Figure 31E). Unlike with the uninhibited reaction, the addition of 1 μM, 5 μM, and 10 μM of tRNAThr had less of an effect on Ap4A formation with rates of 0.08143 min-1, 0.1171 min-1, 0.1206 min-1

(corresponding to a 15% decrease at 1 μM tRNAThr and a 22% and 26% increase at 5 μM and 10 μM tRNAThr relative to the reaction treated with BC194 alone).

Table 9. Rates of Ap4G synthesis by various TARS constructs Construct Condition Activity (min-1) % Activity* WT TARS: Uninhibited 0.1639 ± 0.01224 100 10 μM BC194 0.07065 ± 0.009176 43 No ATP 0.01278 ± 0.005864 8 R442A TARS: Uninhibited 0.01062 ± 0.004659 6 Ec ThrRS: Uninhibited 0.5612 ± 0.06121 342 *Relative to WT TARS, uninhibited

In addition to Ap4A synthesis, TARS was also found to catalyze the formation of Ap4G. Ap4G formation was monitored using the same method as Ap4A except that α- phosphate labeled GTP was used as a tracer and final calculated value was not cut in half because, unlike with Ap4A, only one labeled phosphate would be included per molecule

(Table 9). WT TARS catalyzed the formation of Ap4G at a rate of 0.1639 min-1, nearly identical to the 0.1391 min-1 obtained for Ap4A formation (Figure 32A). As proof of the products identity as Ap4G and not Gp4G, the reaction was run in the absence of ATP and with the aminoacylation deficient R442A mutant, which yielded rates of 0.01278 min-1 and 0.01062 min-1 respectively. Together these data indicated that Ap4G synthesis was dependent on the formation of the threonyl-adenylate before the adenosine moiety is transferred to the second GTP molecule. Consistent with Ap4A formation, treatment with

163

Figure 32. TARS catalyzes the synthesis of the hybrid dinucleotide tetraphosphate Ap4G. (A) Ap4G formation by 5 μM WT TARS (blue), R442A TARS (red), Ec ThrRS (green), or WT TARS in absence of threonine (purple). (B) WT TARS Ap4G formation uninhibited (blue) or with 10 μM BC194 (red). All data are presented as mean ± SD, n=3.

BC194 reduced the rate of Ap4G synthesis to 0.07065 min-1, 43% of the uninhibited reaction (Figure 32B).

5.2.3 TARS Catalyzes the Formation of ADP from ATP

The conversion of ATP to ADP by TARS was discovered accidently when a band corresponding to ADP was identified on a thin layer chromatography plate originally monitoring the formation of the threonyl-adenylate intermediated. While initially dismissed as a substrate contaminant, its formation appeared to increase over the course of the reaction, suggesting that it was a product of the reaction. Like with the GTPase and

Ap4A activities above, this find was particularly interesting due to the recognized connection between extracellular nucleotides and the regulation of angiogenesis through purinergic signaling. Notably, extracellular ADP is well documented to regulate endothelial cell migration, much like TARS. Specifically, ADP activates the G-protein coupled receptor P2Y1, which is highly expressed within endothelial cells (20).

164 Stimulation of this receptor in endothelial cells leads to the phosphorylation and activation of several key mitogen-activated protein kinases involved in endothelial cell migration, including ERK1/2, JNK, and p38 MAPK (12). Additionally, activation of the

P2Y1 receptor has been demonstrated to transactivate VEGFR2, providing an indirect method for the regulation of angiogenesis (21).

Table 10. Rates of ADP formation by various TARS constructs Construct Condition Activity (min-1) % Activity* WT TARS: Uninhibited 0.1417 ± 0.01409 100 10 μM Borrelidin 0.2777 ± 0.03141 182 10 μM BC194 0.2738 ± 0.02056 193 No Threonine 0.3925 ± 0.03888 277 R442A TARS: Uninhibited 0.1059 ± 0.01121 75 Ec ThrRS: Uninhibited 0.01330 ± 0.02283 9 *Relative to WT TARS, uninhibited

ADP formation was measured from same reaction samples used for Ap4A except that they were separated by TLC using a potassium phosphate buffer instead of the ammonium sulfate/EDTA mix (Figure 33A). As with Ap4A, the amount of ADP at each time point was determined as a fraction of the total nucleotides in the sample and compared to the initial concentration of ATP added to the reaction. Rates are presented as the number of ADP molecules generated per active site (or total enzyme concentration for R442A) per minute (Table 10). Because the measurements of ADP were taken from

Ap4A reactions retrospectively, they are limited by the conditions used for Ap4A synthesis. As such, nearly all measurements contain threonine.

Full kinetic investigations have been performed on WT TARS, WT ThrRS, and the aminoacylation deficient R442A TARS (Figure 33B). Both WT and R442A TARS

165

Figure 33. Kinetic characterization of TARS ADP formation. (A) Representative autoradiographic images used for quantifying TARS ADP formation under untreated or 10 μM BC194 conditions. Reactions were monitored using [α32P]-labeled ATP as a tracer and each time point was spotted onto a polyethyleneimine-cellulose thin-layer chromatography plate and resolved using 750 mM potassium phosphate, pH 3.5 buffer. (B-E) Rate of ADP formation by purified TARS constructs under various conditions. (B) ADP formation activity of 5 μM WT TARS (blue), R442A TARS (red), Ec ThrRS (green), or WT TARS in absence of threonine (purple). (C) ADP formation rates of WT TARS in the presence of 0 μM (blue), 1 μM (red), 5 μM (green), or 10 μM (purple) tRNAThr. (D) WT TARS ADP formation rates uninhibited (blue), 10 μM BC194 (red), or 10 μM borrelidin (green). (E) ADP formation rates of WT TARS in the presence of 0 μM (blue), 1 μM (red), 5 μM (green), or 10 μM (purple) tRNAThr with 10 μM BC194. All data are presented as mean ± SD, n=3.

166 formed ADP from ATP, with rates of 0.1417 min-1 and 0.1059 min-1 respectively. Since

R442A TARS is unable to catalyze the first, adenylation step of aminoacylation, these data would suggest that ADP formation is separate from the canonical reaction of the enzyme. It may also be worthwhile to note that, due to loss of adenylation activity, the active concentration of R442A could not be determined before these assays requiring that the total protein concentration be used instead. Therefore, it is possible that the rates of

WT and R442A TARS would be closer if R442A TARS could be adjusted for the active concentration. In contrast to the human constructs, Ec ThrRS exhibited no activity above background with a rate of 0.01330 min-1, about 9% of the activity observed with the human enzyme. Therefore, like GTPase activity, ADP formation appeared to be a human phenomenon (although it remains to be tested in other organisms).

When the reaction was run in absence of threonine, WT TARS exhibited a rate of

0.3925 min-1, about 2.77 fold higher than the reaction with threonine (Figure 33B). While this reaction was not completed in triplicate like the others (this reaction was not necessary for the Ap4A studies) it does indicate that the reaction does not require threonine and suggests that, while ADP formation is a separate reaction from aminoacylation, it may still respond to aminoacylation substrates. Surprisingly, while threonine appeared to negatively impact the rate of ADP formation, the opposite was true for tRNAThr additions to the reaction (Table 11). The addition of 1 μM, 5 μM, and 10 μM tRNAThr increased the reaction rate from 0.1417 min-1 to 0.2804 min-1, 0.4063 min-1, and

0.6561 min-1 respectively, which correspond to increases of 98%, 187%, and 363%

(Figure 33C).

167 Table 11. The effects of tRNAThr on the rates of WT TARS ADP formation Construct Condition Activity (min-1) % Activity* WT TARS: 0 μM tRNAThr 0.1417 ± 0.01409 100 1 μM tRNAThr 0.2804 ± 0.008072 198 5 μM tRNAThr 0.3993 ± 0.03835 287 10 μM tRNAThr 0.6479 ± 0.04162 463 0 μM tRNAThr; 10 μM BC194 0.2738 ± 0.02056 193 1 μM tRNAThr; 10 μM BC194 0.3766 ± 0.03478 266 5 μM tRNAThr; 10 μM BC194 0.4463 ± 0.06611 304 10 μM tRNAThr; 10 μM BC194 0.5870 ± 0.04631 413 *Relative to WT TARS, uninhibited

The connection between the ADP and angiogenesis opened the possibility of this ADP reaction as a potential mechanism for TARS’s effects on endothelial cell migration. Like with Ap4A, this hypothesis would also suggest that borrelidin and related compounds would be able to inhibit angiogenesis by altering the rate of this reaction.

Therefore, the effects of Borrelidin and BC194 on this reaction were subsequently investigated (Figure 33D). At saturating amounts (10 μM relative to the 5 μM TARS) both compounds appeared to enhance the rate of the reaction: borrelidin increased the rate by 82% (0.2578 min-1) and BC194 by 92% (0.2738 min-1). This is very different from the potent inhibitory effects exhibited by the compounds against the canonical aminoacylation reaction. The effect of tRNAThr on the BC194-treated reaction was also

Thr investigated to check if the activity increase by tRNA is lost, enhanced, or unaffected by BC194 (Figure 33E). At concentrations of 1 μM, 5 μM, and 10 μM tRNAThr the reactions increased to 0.3766 min-1, 0.4304 min-1, and 0.5859 min-1 respectively. This corresponds to enhancements of 38%, 57%, and 114% relative to the reaction treated with BC194 alone and 93%, 166%, and 313% relative to the untreated reaction. Given

168 the similarities in these values when compared to the untreated, WT TARS reaction, it appears that the rate increases by BC194 are masked by those provided by the tRNA.

While kinetic rates for ADP formation have not be thoroughly investigated with the other TARS constructs used in this work, the ability of these enzymes to catalyze this reaction can be identified qualitatively from observations of other experiments.

Specifically, the TLC plates used in the determination of active site concentrations can be used to assess for the absence or presence of ADP. However, due to the very different assay conditions, direct comparisons concerning rates cannot be made. Using this method

ADP formation has been observed with the borrelidin-resistant L567V TARS, the angiogenesis but aminoacylation active Q566W TARS, and the angiogenesis and aminoacylation deficient A538W TARS (Data not shown).

5.2.4 Effects of purinergic inhibitors on TARS-mediated angiogenesis

Previously ADP and Ap4A have been shown to regulate angiogenesis by stimulating the purinergic receptors P2Y1 and P2X respectively. Therefore, the effects of purinergic signaling by TARS-generated ADP and Ap4A on angiogenesis were investigated using an endothelial tube formation assay and inhibitors of the P2Y1 and

P2X receptors (Figure 34A, B). As previously described, the presence of TARS at 100 nM enhanced tube formation by around 50%. Interestingly, the inclusion of 10 μM of the

P2Y1 receptor inhibitor MRS2179 completely inhibited TARS-mediated angiogenesis back to baseline levels (Figure 34A). Conversely, no significant inhibitory effects were observed for the P2X inhibitor Ip5I (Figure 34B). However, the Ip5I treated cultures appear to visually trend to a reduction in tube-formation and may warrant additional

169 replications to see if statistical significance can be obtained. Neither MRS2179 nor Ip5I exhibited significant effects on baseline tube formation in absence of TARS.

Figure 34. The effects of purinergic receptor inhibitors MRS2179 and Ip5I on TARS- mediated angiogenesis. (A, B) Quantification of HUVEC branching in control of 100 nM WT TARS-treated cultures with and without MRS2179 (A) or Ip5I (B). HUVECs were plated on Matrigel in 10% EGM-2 (0.2% FBS) media and exposed to various combinations of buffer, 100 nM TARS, and 10 μM of MRS2179 or Ip5I. Cells were fixed after 4-8 h and stained with Oregon Green 488 Phalloidin. Numbers represent percentage of control (0 nM) samples, mean ± SEM, n=3, ** and **** correspond to p<0.01 and 0.0001 respectively relative to the TARS-treated sample (one-way ANOVA, Tukey Test).

5.3 Discussion

5.3.1 Correlations between non-canonical, catalytic reactions and TARS-mediated angiogenesis

The data presented above demonstrate that TARS is capable of catalyzing several different reactions in addition to its canonical role in aminoacylation. For comparative purposes the various constructs and their ability to perform canonical and

170 non-canonical functions are listed in Table 12. Unfortunately, no simple conclusions can be drawn between the angiogenic potential of the various TARS constructs and their participation in non-canonical catalytic activities. Primarily, the difficulty arises with either R442A TARS or Ec ThrRS. Both of these constructs have been demonstrated to stimulate angiogenic responses in endothelial tube formation assays. However, since Ec

ThrRS lacks GTPase activity and ADP formation and R442A TARS is unable to synthesize Ap4A or Ap4G, none of these reactions can be clearly attributed to the initiation of angiogenesis. Nonetheless, some explanations can be developed.

Table 12. The functions of various TARS constructs Amino- Angio- ATP to Ap4A/ Construct Description GTPase acylation genesis ADP Ap4G WT TARS Human enzyme Yes Yes Yes Yes Yes WT ThrRS E. coli enzyme Yes Yes No No Yes Aminoacylation R442A TARS No Yes Yes Yes No deficient L567V TARS Borrelidin resistant Yes Yes Yes Yes Yes Most N-terminal N1-Domain domain; no known No No Unlikely* No No function Catalytic and anti- ΔN1N2- codon binding Likely‡ No Unknown Unknown Likely# Domain domains only Borrelidin Q566W TARS Yes No Yes Likely* Likely# conformation mimic Borrelidin A538W TARS No No Yes Likely* Unlikely# conformation mimic * When tested, all constructs with ATP to ADP or GTPase activity have exhibited the other as well. ‡ ΔN1N2-TARS is capable of rescuing cells from loss of WT TARS, supporting aminoacylation function. # So far, all constructs able to perform aminoacylation appear capable of Ap4A and Ap4G synthesis.

The discovery of GTPase activity in TARS opened up the possibility that this reaction could function as a switch between TARS’s canonical and non-canonical functions. Since TARS is normally found in the cytoplasm, it seems likely that the

171 GTPase would initiate this switch event intracellularly. However, TARS’s angiogenic function requires that the protein be secreted outside the cell. Since the test substances are added extracellularly during the tube formation assay, any potential importance of the

GTPase function would have been bypassed and, thus, would not have affected Ec

ThrRS’s ability to simulate angiogenesis.

Of the three non-canonical reactions (GTPase, Ap4A/Ap4G synthesis, and ADP formation) the formation of ADP appears to be the strongest candidate as a mechanism for TARS-mediated angiogenesis. First, ADP is already known to stimulate endothelial cell migration, a function that is also attributed to TARS, through the stimulation of the

P2Y1 receptor. The fact that the P2Y1 specific inhibitor, MRS2179, appears to attenuate

TARS-mediated angiogenesis suggests that ADP purinergic signaling is important for this process as well. Furthermore, while Ec ThrRS is unable to generate ADP, it does catalyze the formation of Ap4A at a faster rate than human TARS. Since Ip5I did not seem to affect TARS-mediated angiogenesis, it seems unlikely that Ap4A is instigating its own purinergic signaling through the P2X receptor. However, Ap4A can be broken down into two ADP molecules by extracellular proteins, providing an alternative method of synthesizing ADP. Therefore, all angiogenic constructs appear to possess the ability to synthesize ADP to some degree. Furthermore, the rate of this hydrolysis appears to be slower than the breakdown of other extracellular adenosine phosphate nucleotides to adenosine, suggesting that Ap4A may also serve to prolong the purinergic response (22).

This model, however, does not provide a clear mechanism for borrelidin and BC194’s inhibition of angiogenesis.

172 Treatment of TARS with borrelidin and BC194 appears to stimulate the production of ADP instead of inhibiting it. One possible explanation for this may be connected to the assay conditions used in the measurement of ADP formation. Because

ADP formation was measured retrospectively using time points from the Ap4A assays, most of the reactions contained threonine. Interestingly, the one reaction that did not contain threonine exhibited the fastest measured rate of ADP formation. While threonine did not fully block the reaction it may have stabilized a less efficient conformation of the

TARS active site and slowed the reaction rate. This observation suggests that the rate enhancement exhibited by borrelidin and BC194 may be due to their competition with threonine in the reaction, a hypothesis that would be easily confirmed by investigating the effects of borrelidin and BC194 on ADP formation in absence of threonine. Therefore, it seems unlikely that borrelidin and BC194 would influence the ADP-mediated angiogenesis through modification of the ADP formation rate.

Since borrelidin and BC194 do not actually appear to affect the rate of ADP formation, they would have to inhibit ADP-mediated angiogenesis through a different mechanism. Previously, the borrelidin and BC194 induced conformational changes have been demonstrated to be important for inhibiting TARS-mediated angiogenesis as demonstrated by the conformational mimetic mutations Q566W TARS and A538W

TARS (Chapter 4, section 4.2.6). Interestingly, both of these constructs possess the ability to catalyze the conversion of ATP to ADP, providing further evidence that inhibition of ADP formation cannot be responsible for borrelidin and BC194’s inhibition of angiogenesis. As an alternative it may be that, as part of its angiogenic mechanism, secreted TARS forms protein complexes that help localize it to specific extracellular

173 regions. This spatial arrangement may bring it close to the P2Y1 receptor and, thus, would enhance its activation. In this model, borrelidin and BC194 would then induce conformational changes in TARS and disrupt its interaction with the protein complexes.

It is important to note, however, that this discussion of how TARS can stimulate angiogenesis through ADP formation and its disruption by borrelidin and BC194 is only to establish it as a possible model. Therefore, while all the acquired data would support

ADP as a possible mechanism for TARS-mediated angiogenesis, more data are needed to affirm it.

5.3.2 Nucleotide preference in non-canonical, catalytic reactions

One trend that can be observed in Table 12 is that all enzymes that possess

GTPase activity also possess the ability to convert ATP to ADP. Furthermore, the addition of ATP clearly inhibits the GTPase activity in all conditions. Therefore, it may be possible that both functions compete for the same active site and, based on the rate data, there is a preference for adenosine nucleotides. Without further data it is difficult to say if the use of GTP in this reaction is an unintended side reaction or if it has a specific purpose. Phosphorylation has been previously shown to enhance the rate of Ap4A synthesis in TARS, therefore, it may be that some similar regulatory event might be able to increase the sites affinity for GTP over ATP. Alternatively, it may be that this GTPase activity functions as a switch between the canonical and non-canonical functions of the

TARS proteins.

While the conversion of nucleotide triphosphates to nucleotide diphosphates by

TARS appears to prefer adenosine-based molecules over those with guanosine, the

174 opposite appears to be true for the synthesis of dinucleotide tetraphosphates. The synthesis of both Ap4A and Ap4G required that ATP be present for the formation of the threonyl-adenylate intermediated. As such, the synthesis of Ap4G should be in competition with that of Ap4A, assuming that both reactions are occurring in the same active site (which is reasonable given the dependence on the threonyl-adenylate). Under these conditions the rate of Ap4G would be expected to be slower than Ap4A due to this competition. However, Ap4G formed at rates comparable to Ap4A when ATP was the only nucleotide present. Therefore, it appears as if the formation of Ap4G is preferred over Ap4A. The affinity of other nucleotides for this reaction remain to be determined.

5.3.3 ADP-formation and GTPase activity as nutrient and energy sensors

As presented in Chapter 1, sections 1.2.1 and 1.6.1 a growing body of evidence implicates the ARSs as cellular nutrient and energy sensors. As the rates of ADP formation and GTPase activity in TARS are influenced by the levels of threonine and

ATP respectively, these reaction may also serve a similar function. In absence of threonine, the rate of ADP generation increased 277%. The elevated levels of ADP associated with these rates could then improve TARS’s stimulation of angiogenesis through the P2Y1 receptor or by activating some other, yet unknown response (11,12).

Moreover, the extracellular concentrations of threonine are thought to be lower than those inside the cell, which would further enhance ADP production following TARS secretion

(23). In contrast to ADP formation, the presence of ATP almost completely inhibits

TARS’s GTPase activity. As such, this activity would be limited to conditions of low energy within the cell. Since GTPase activities in other proteins have been demonstrated

175 to modify their functions, this property could serve to switch TARS between its canonical and non-canonical functions in response to cellular energetic levels (15). However, it remains to be determined if ADP, which would be abundant under low energy conditions, has any influence on the rate of GTPase activity as well.

5.3.4 The effect of tRNAThr on ADP formation and Ap4A

The observation that tRNAThr stimulates both ADP and Ap4A formation is interesting but its function and mechanism remain unknown. In both these reactions threonine and ATP are present in great excess, suggesting that any tRNAThr present should be fully charged by the first time point (this seems to be a safe assumption as

TARS is present at 5 μM and would only require two turnovers in 30 minutes to aminoacylate 10 μM tRNAThr, the highest concentration used in any of the reactions).

Curiously, the stimulatory effect of tRNAThr on ADP formation is unaffected by the presence of BC194 while it is reduced in Ap4A synthesis. However, this difference may be due to BC194’s inhibition of the adenylation reaction, which would disrupt Ap4A synthesis and compete with the positive effects of tRNAThr. Conversely, ADP formation appears independent of aminoacylation and would be unaffected.

As a possible explanation for the positive effects of tRNAThr on ADP-formation, the accumulation of charged-tRNAThr in the reaction may block the canonical aminoacylation reaction through product inhibition. As the GTPase reactions above did not include threonine, no Thr-tRNAThr was generated, consistent with the lack of tRNA- induced kinetic effects. Conversely, since threonine was demonstrated to reduce ADP formation, product inhibition by Thr-tRNAThr may enhance ADP formation by blocking

176 threonine binding to its original substrate pocket. Likewise, BC194 also blocks threonine binding (Chapter 2, section 2.2.3) and may explain its enhancement to ADP-formation rates as well. Because the binding of tRNAThr and BC194 are mutually exclusive and would both depend on the exclusion of threonine to induce rate enhancements in this model, synergistic effects would not be expected, consistent with the above results.

However, since the rate of ADP generation in absence of threonine (0.3925 min-1) is less than that observed at 10 μM tRNAThr with or without BC194 (0.6479 and 0.5870 min-1) it seems likely that threonine exclusion may not be the only effect.

As an alternative or complementary mechanism, the binding of Thr-tRNAThr to one active site of a TARS dimer may alter the non-canonical activities within the same or other monomer. Previously, pre-steady state kinetics have demonstrated that the presence of tRNA reduces the rate of amino acid activation by Ec ThrRS (24). For another dimeric

ARS, E. coli Histidyl-tRNA synthetase (HisRS), the adenylation reaction appears to proceed asymmetrically between the two monomers in absence of tRNA. Specifically, these investigations reveal that the second site is able to bind ATP and amino acid only after the first generates a bound adenylate intermediate. Even at this point, however, the generation of the adenylate intermediate in the second site appears to occur at a much slower rate (25). Interestingly, the addition of uncharged-tRNA appears to equalize the rates of both sites, which suggests that tRNA may be important for coupling the reactions in both sites (26). Therefore, it may be that product inhibition by charged-tRNA also alters the individual activity or coupling of the active site monomers, albeit in favor of non-canonical reactions. For ADP, the binding of charged-tRNA may induce conformational changes within either active site that facilitates ADP-formation. In the

177 case of Ap4A, the binding of charged tRNA in one site may encourage adenylate formation in the other, which, without free-tRNA to continue with aminoacylation, is resolved through the formation of Ap4A.

5.4 Experimental Procedures

5.4.1 Preparation of TARS Constructs and tRNAThr

The preparation of TARS constructs and tRNAThr was performed using procedures identical to those described in Chapter 2, sections 2.3.1 and 2.3.2.

5.4.2 Ap4A, AMP, and ADP Formation assays

Ap4A synthesis was investigated using adaptations of previously reported methods (3). Reaction mixtures consisted of 20 mM HEPES, pH 8.0, 100 mM KCl, 10 mM MgCl2, 5 mM β-mercaptoethanol, 5 mM threonine (unless otherwise noted), 2.5 U pyrophosphatase (Roche), and 2.5, 5, or 20 µM E. coli ThrRS, wildtype human TARS, or

R442A TARS respectively. Where necessary, reactions were pre-incubated with 10 μM

BC194 or Borrelidin in DMSO or a comparable amount (0.05% V/V) of DMSO vehicle for 15 minutes on ice prior to initiation. Reactions were then incubated at 37 °C for 3 minutes and initiated with the addition of 2 mM ATP. Aliquots were taken at varying time points and quenched in 3 volumes of 400 mM sodium acetate, pH 4.5, and 0.1% sodium dodecyl sulfate. Products were resolved by polyethyleneimine-cellulose TLC plates with 3M (NH4)2SO4 and 2% EDTA as mobile phase and quantified by phosphorimaging. The identity of Ap4A was confirmed by UV shadowing of an un- labeled standard. Ap4A production was calculated as a percentage of total counts and

178 compared to the original amount of ATP present in the reaction. To account for the incorporation of two ATP molecules per Ap4A product the final values were multiplied by 0.5. ADP was measured using identical reactions except that products were resolved using TLC with 750 mM potassium phosphate buffer, pH 3.5.

5.4.3 Ap4G synthesis and GTPase Assays

Ap4G and GTPase assays were performed as Ap4A synthesis above with the following exceptions. Ap4G reactions always included 2 mM ATP and 2 mM threonine and 10 µM BC194 where indicated in Figure 32B. For GTPase assays ATP and threonine were not present in all reactions but were included in combination with 10 µM tRNAThr,

10 µM BC194, and 10 µM borrelidin according to Figure 29 C and D. Adenylation conditions consisted of 2 mM ATP and 2 mM threonine and aminoacylation conditions further included 10 μM tRNAThr. Reactions were initiated using 2 mM GTP with trace amounts of [α-32P]-GTP. The components were resolved by TLC using the mobile phases

750 mM KH2PO4, pH 3.5 for GTPase data and 3 M (NH4)2SO4 and 2% EDTA for Ap4G.

Unlike with Ap4A, Ap4G involves only the incorporation of one labeled nucleotide and does not require the 0.5 correction factor.

5.4.4 Cell Culture and Reagents

Human umbilical endothelial cells (HUVECS) were maintained in Clonetics EGM-2 complete media (Lonza) and used between passages 5 to 8. MRS2179 and Ip5I were obtained from Abcam® and Sigma-Aldich® respectively.

179 5.4.5 Tube Formation Assay with MRS2179 and Ip5I

Tube formation assays were performed as previously described (8,27). Briefly

HUVEC cells were cultured on 0.1% gelatin coated plates until 95-100% confluent and then seeded in to 48-well plates coated with MatrigelTM (BD biosciences). Cells were incubated at 37 °C in EGM®-2 media with 2% FBS and 10 μM MRS2179 or Ip5I where necessary for 4 to 8 h. Cellular structures were fixed with 10% neutral buffered formalin, stained with Oregon Green 488 phalloidin (Molecular Probes), and imaged via fluorescence microscopy. Tube structures were quantified using the Simple Neurite

Tracer plug-in on ImageJ software (NIH).

180 5.5 References

1. Dang, C. V., and Traugh, J. A. (1989) Phosphorylation of threonyl- and seryl- tRNA synthetase by cAMP-dependent protein kinase. A possible role in the regulation of P1, P4-bis(5'-adenosyl)-tetraphosphate (Ap4A) synthesis. J Biol Chem 264, 5861-5865

2. Goerlich, O., Foeckler, R., and Holler, E. (1982) Mechanism of synthesis of adenosine(5')tetraphospho(5')adenosine (AppppA) by aminoacyl-tRNA synthetases. Eur J Biochem 126, 135-142

3. Guo, R. T., Chong, Y. E., Guo, M., and Yang, X. L. (2009) Crystal structures and biochemical analyses suggest a unique mechanism and role for human glycyl- tRNA synthetase in Ap4A homeostasis. J Biol Chem 284, 28968-28976

4. Traut, T. W. (1987) Synthesis of hybrid bisnucleoside 5',5"'-P1,P4- tetraphosphates by aminoacyl-tRNA synthetases. Molecular and cellular biochemistry 75, 15-21

5. Zamecnik, P. C., Stephenson, M. L., Janeway, C. M., and Randerath, K. (1966) Enzymatic synthesis of diadenosine tetraphosphate and diadenosine triphosphate with a purified lysyl-sRNA synthetase. Biochem Biophys Res Commun 24, 91-97

6. Lee, Y. N., Nechushtan, H., Figov, N., and Razin, E. (2004) The function of lysyl- tRNA synthetase and Ap4A as signaling regulators of MITF activity in FcepsilonRI-activated mast cells. Immunity 20, 145-151

7. Ofir-Birin, Y., Fang, P., Bennett, S. P., Zhang, H. M., Wang, J., Rachmin, I., Shapiro, R., Song, J., Dagan, A., Pozo, J., Kim, S., Marshall, A. G., Schimmel, P., Yang, X. L., Nechushtan, H., Razin, E., and Guo, M. (2013) Structural switch of lysyl-tRNA synthetase between translation and transcription. Mol Cell 49, 30-42

8. Williams, T. F., Mirando, A. C., Wilkinson, B., Francklyn, C. S., and Lounsbury, K. M. (2013) Secreted Threonyl-tRNA synthetase stimulates endothelial cell migration and angiogenesis. Sci Rep 3, 1317

9. van der Giet, M., Jankowski, J., Schluter, H., Zidek, W., and Tepel, M. (1998) Mediation of the vasoactive properties of diadenosine tetraphosphate via various purinoceptors. Journal of hypertension 16, 1939-1943

10. Bobbert, P., Schluter, H., Schultheiss, H. P., and Reusch, H. P. (2008) Diadenosine polyphosphates Ap3A and Ap4A, but not Ap5A or Ap6A, induce proliferation of vascular smooth muscle cells. Biochem Pharmacol 75, 1966-1973

11. McAuslan, B. R., Reilly, W. G., Hannan, G. N., and Gole, G. A. (1983) Angiogenic factors and their assay: activity of formyl methionyl leucyl 181 phenylalanine, adenosine diphosphate, heparin, copper, and bovine endothelium stimulating factor. Microvascular research 26, 323-338

12. Shen, J., and DiCorleto, P. E. (2008) ADP stimulates human endothelial cell migration via P2Y1 nucleotide receptor-mediated mitogen-activated protein kinase pathways. Circ Res 102, 448-456

13. Bambace, N. M., Levis, J. E., and Holmes, C. E. (2010) The effect of P2Y- mediated platelet activation on the release of VEGF and endostatin from platelets. Platelets 21, 85-93

14. Wolf, Y. I., Aravind, L., Grishin, N. V., and Koonin, E. V. (1999) Evolution of aminoacyl-tRNA synthetases--analysis of unique domain architectures and phylogenetic trees reveals a complex history of horizontal gene transfer events. Genome research 9, 689-710

15. Bourne, H. R., Sanders, D. A., and McCormick, F. (1991) The GTPase superfamily: conserved structure and molecular mechanism. Nature 349, 117-127

16. Teplyakov, A., Obmolova, G., Chu, S. Y., Toedt, J., Eisenstein, E., Howard, A. J., and Gilliland, G. L. (2003) Crystal structure of the YchF protein reveals binding sites for GTP and nucleic acid. Journal of bacteriology 185, 4031-4037

17. Magnusson, L. U., Farewell, A., and Nystrom, T. (2005) ppGpp: a global regulator in Escherichia coli. Trends in microbiology 13, 236-242

18. Schreiber, G., Ron, E. Z., and Glaser, G. (1995) ppGpp-mediated regulation of DNA replication and cell division in Escherichia coli. Current microbiology 30, 27-32

20. Wang, L., Karlsson, L., Moses, S., Hultgardh-Nilsson, A., Andersson, M., Borna, C., Gudbjartsson, T., Jern, S., and Erlinge, D. (2002) P2 receptor expression profiles in human vascular smooth muscle and endothelial cells. Journal of cardiovascular pharmacology 40, 841-853

21. Rumjahn, S. M., Yokdang, N., Baldwin, K. A., Thai, J., and Buxton, I. L. (2009) Purinergic regulation of vascular endothelial growth factor signaling in angiogenesis. Br J Cancer 100, 1465-1470

22. Luthje, J., and Ogilvie, A. (1988) Catabolism of Ap4A and Ap3A in whole blood. The dinucleotides are long-lived signal molecules in the blood ending up as intracellular ATP in the erythrocytes. Eur J Biochem 173, 241-245 182 23. Bergstrom, J., Furst, P., Noree, L. O., and Vinnars, E. (1974) Intracellular free amino acid concentration in human muscle tissue. Journal of applied physiology 36, 693-697

24. Minajigi, A., and Francklyn, C. S. (2008) RNA-assisted catalysis in a protein enzyme: The 2'-hydroxyl of tRNA(Thr) A76 promotes aminoacylation by threonyl-tRNA synthetase. Proc Natl Acad Sci U S A 105, 17748-17753

25. Guth, E., Farris, M., Bovee, M., and Francklyn, C. S. (2009) Asymmetric amino acid activation by class II histidyl-tRNA synthetase from Escherichia coli. J Biol Chem 284, 20753-20762

26. Guth, E. C., and Francklyn, C. S. (2007) Kinetic discrimination of tRNA identity by the conserved motif 2 loop of a class II aminoacyl-tRNA synthetase. Mol Cell 25, 531-542

27. Arnaoutova, I., and Kleinman, H. K. (2010) In vitro angiogenesis: endothelial cell tube formation on gelled basement membrane extract. Nat Protoc 5, 628-635

183 CHAPTER 6: DISCUSSION AND FUTURE PERSPECTIVES

6.1 Mechanisms of TARS-mediated angiogenesis

6.1.1 Borrelidin’s inhibition of angiogenesis is mediated through direct interactions with TARS

Initial connections between TARS and angiogenesis were based on the anti- angiogenic properties of its potent inhibitor borrelidin. The fact that TARS is the only known target of borrelidin with a Ki value in the low nM range coupled with the observation of angiogenesis-related activities in other ARSs opens the possibility that

TARS may possess its own secondary functions in vascular biology. This relationship was subsequently confirmed when extracellular TARS was found to stimulate vessel formation and endothelial cell migration in several in vitro assays but was susceptible to inhibition by the borrelidin-derivative BC194. Moreover, BC194’s inhibition of TARS- mediated angiogenesis could be attenuated through a single TARS point mutation,

L567V, which indicates that the compound functions directly through TARS.

Further investigations of borrelidin and BC194 revealed that their inhibition of angiogenesis could be separated from toxic effects on cells and more related to their association with TARS’s function in vessel patterning and maturation. Specifically, treatment of zebrafish with BC194 or knockdown of TARS expression with an anti-

TARS morpholino was shown to induce ectopic vessel growth and poor lumen development. The similar responses exhibited by both the inhibitors and the knockdown of TARS expression further supports TARS as the only target of borrelidin inhibition.

Furthermore, these experiments represent the first time that TARS’s secondary function

184 in angiogenesis has been investigated in the context of a whole animal model. This connection between TARS and vessel guidance indicated that TARS is likely to form a regulated gradient that is followed by the migrating endothelial cells. Based on these observations a model for the induction of angiogenesis by extracellular-TARS was generated (Figure 35).

6.1.2 Investigating the angiogenic potential of various TARS constructs in the zebrafish model

The investigations of TARS and its inhibitors on zebrafish vascular biology provide critical insights into the mechanisms of TARS-mediated angiogenesis. Moreover, the successful knockdown of TARS expression by the anti-TARS morpholino offers an opportunity to investigate the effects of the various TARS constructs on vessel patterning in the context of a whole organism model. Following morpholino treatment capped- mRNA corresponding to the various TARS constructs characterized in Chapter 4 can be injected into developing zebrafish embryos to complement the reduction of endogenous

TARS. Of particular interest, L567V TARS can be used to investigate the ability of this construct to resist the effects of borrelidin treatment and provide additional confirmation to the TARS-specific nature of these observations. Alternatively, Q566W TARS can be injected to investigate the effect of the borrelidin-conformation mimic on zebrafish vessel development. Since Q566W TARS retains aminoacylation function, this complementation would be expected to produce the ectopic branching associated with the angiogenesis specific effects of TARS without the toxicity associated with amino acid starvation. Finally, the ΔN1N2-TARS construct can be investigated since it still possesses

185

Figure 35. Proposed models for TARS-stimulation and borrelidin-inhibition of angiogenesis. (Left) Step by step proposal for stimulation of angiogenesis by TARS: 1) Activation of immune cells and hypoxic responses in cancer cells stimulate the secretion of TNFα and VEGF respectively. 2) TNFα and VEGF bind to receptors on the surface of endothelial cells. 3) Activation of unknown signaling pathways stimulates the secretion of TARS from endothelial cells. 4) Stimulation of hypoxic cancer cells with TNFα also stimulates secretion of TARS (1). 5) Extra-cellular TARS forms a free or ECM-bound gradient that guides the migration of nascent vessels. (Right) Proposed mechanisms of borrelidin’s inhibition of angiogenesis. Mechanism 1 – binding of Bor elicits a conformational change in TARS that precludes its binding to an unknown receptor. Mechanism 2 – A borrelidin-mediated conformational change disrupts spatial regulation of secreted TARS by blocking interactions with the extra-cellular matrix. Mechanism 3 – Inhibition of TARS aminoacylation activity by borrelidin results in the accumulation of uncharged-tRNA which activates GCN2-mediated amino acid starvation responses. aminoacylation activity (although not trans-editing function) but lacks the ability to stimulate angiogenesis in tube-formation assays. Therefore, the two binding site model could be investigated in the context of a whole organism.

186

6.1.3 TARS interacts with endothelial cell-derived extracellular matrices

The observation of an angiogenic factor mediating two simultaneous interactions is not uncommon. Many of these molecules, notably platelet-derived growth factor (PDGF) and vascular endothelial growth factor (VEGF), associate with both a cell surface receptor as well as components of the extracellular matrix (2-4). This mechanism allows for the establishment of a spatially regulated gradient that assists in proper signaling to growing blood vessels. Disruption of the gradient’s spatial arrangement or its ability to interact with cellular receptors would, therefore, influence the navigation ability of migrating cells and produce defects in vessel patterning.

PDGF released from vessels is important for guiding the migration of pericytes, blood vessel support cells, to newly formed vasculature. The targeting of the pericytes to these vessels is dependent on the binding of PDGF to the ECM adjacent to the endothelial cells. Loss of ECM binding by PDGF results in the accumulation of pericytes away from existing vessels as PDGF diffuses into the extracellular space (2).

Alternatively, remodeling of the ECM by matrix-metalloproteases (MMPs) during angiogenesis can expose trapped soluble factors concealed within the matrix. For instance, MMP cleavage can separate the receptor binding portion of VEGFA from the

ECM interaction domain, effectively generating a soluble VEGFA population. While both ECM-bound and soluble VEGFA appear to be bioactive, the two forms elicit differential effects on endothelial cells. Specifically, soluble VEGFA initiates more proliferative functions while bound isoforms induce a more migratory state (5,6). The

187 cause of this differential signaling is the direct association of the VEGFA receptor

(VEGFR2) and a class of ECM-recognizing receptors known as integrins.

Endothelial cells are sensitive to the condition of the ECM through the heterodimeric integrins which are activated by interactions with matrix proteins like fibronectin and vitronectin (7). Even when anchored to the ECM, VEGF remains bioactive and induces responses distinct from soluble VEGF (4). Specifically, bound

VEGF appears to stimulate a population of VEGFR2 in complex with β1 integrin, which subsequently relocates to multimeric focal adhesion complexes involved in endothelial cell migration. In comparison to soluble VEGF, the cooperation between integrin and

VEGF receptors exhibits greater activation of the pro-migratory p38-mitogen-activated protein kinase (p38-MAPK) (4). Moreover, other receptor tyrosine kinases (RTKs) than

VEGFR2 have been demonstrated to interact with integrins and underlie a complex interaction between angiogenic signaling and the ECM that is too extensive to be explained here in full detail (8). However, these observations do emphasize the potential importance of TARS-ECM interactions for the mechanism of the ARS’s function in angiogenesis.

Similar to VEGF and PDGF, observations in Chapter 4 suggest that TARS forms two separate interactions in the process of stimulating angiogenesis. Combining the various TARS constructs with the HUVEC tube formation assay (Chapter 4) allowed the structural components and activities important for TARS-mediated angiogenesis to be determined. The lack of angiogenic activity associated with the ΔN1N2-TARS truncation

(containing only the catalytic and anti-codon binding domains) demonstrates that the full- length protein is necessary for TARS’s vascular functions. Moreover, the addition of this

188 construct appeared to inhibit tube formation, suggesting that it might titrate away and interaction between the full-length protein and its target. When considered with the loss of secondary activity observed with the borrelidin-conformation mimic mutation Q566W

TARS, these data suggest that two simultaneous protein-protein interactions are required for TARS-mediated angiogenesis. This hypothesis is also supported by the lack of angiogenesis exhibited by the N1-domain alone as well. However, this experiment should be reinvestigated using a fragment consisting of both the N1 and N2 domains so that all domains of TARS can be fully investigated.

In this model, the first interaction, putatively between TARS and the ECM, establishes a gradient for nascent blood vessels essential for proper patterning. The bound

TARS then makes a second interaction with a cellular surface receptor or related protein complex to mediate the necessary intracellular signaling pathways for migration.

Borrelidin would be able to disrupt either the formation of the TARS gradient by inhibiting the TARS-ECM interaction or the intracellular signaling by blocking the

TARS-receptor interaction. Recent preliminary data by Hannah Poulette and Karen

Lounsbury provide some support for this model. Immunofluorescence experiments reveal that extracellular TARS readily binds to HUVEC-derived extracellular matrix (Figure

36A, B). Treatment with BC194, however, has no noticeable effect on TARS binding

(Figure 36B). Similar to in vitro angiogenesis assays, equivalent interactions are not observed with the non-specific ARS control leucyl-tRNA synthetase (LARS) (Figure

36B) ((9) and Chapter 4), suggesting that the interaction is not general among the ARS family.

189

Figure 36. TARS interacts with endothelial cell derived fibronectin. Endothelial cells (HUVECs) were plated on gelatin and incubated for five days to allow for the accumulation of extracellular matrix (ECM). In some cases (B) cells were lysed using 20 mM NH4OH and 0.5% Triton-X in PBS, which leaves the ECM intact. The cultures were then treated for one hour under the indicated conditions and LARS, TARS, fibronectin (FN1), nuclei (DAPI) detected by immunofluorescence. (A) Images of HUVEC-derived matrix with intact cells treated with 100 nM TARS showing FN1 (left), TARS (middle), and merged with DAPI (right). (B) Merged images of HUVEC-derived ECM with lysed cells treated with PBS (Control) or as indicated in the figure. Data were kindly provided by Hannah Poulette and Karen Lounsbury.

190 The above observations are not the first evidence of TARS interaction with the

ECM. Affinity-purification mass spectrometry experiments interested in focal adhesion complexes have reported interactions by TARS with fibronectin 1 (FN1) and integrins

(10). As mentioned above, the connection between focal adhesions complexes and endothelial cell migration has been well established and is very important to the process of angiogenesis. The interaction between integrin receptors and FN1 is strongly dependent on an RGDS amino acid sequence and can disrupted through exposure to the equivalent peptide (11,12). Interestingly, treatment with the RGDS-peptide appears to disrupt the TARS-ECM interaction as determined by immunofluorescence experiments

(Figure 35B). However, the effect is specific to regions that appear to have been directly beneath or adjacent to cells prior to dissolution and ineffective against TARS bound to more distantly deposited FN1. A possible explanation for this difference is the status of the fibronectin matrix itself. Soluble FN1 is typically found in a dimeric structure.

However, in a cell and integrin-dependent process these FN1 dimers can be reorganized into multimeric, fibrillar networks through fibrillogenesis. Interestingly, the process of generating matrix FN1 can also expose new interaction sites in the FN1 molecules

(13,14). Since the formation of the fibrillar form is dependent on cellular-based integrins, the RGDS peptide may disrupt normal fibrillogenesis and negatively impact the binding of TARS to these sites (15). Therefore, TARS-mediated angiogenesis may be dependent on specific structure attributes of the FN1 matrix. However, the exact mechanistic details of how these various forms of FN1 affect this process remain to be determined.

6.1.4 Identification of extracellular TARS binding partners

191 Chapter 4 revealed that the angiogenesis-inhibitory properties of borrelidin and

BC194 are mediated through a specific conformational change within the TARS aminoacylation domain. This observation suggests that TARS protein-protein interactions are important for the ARS’s angiogenic function. As described in Appendix

I, affinity purification of TARS followed by mass spectrometry analysis revealed the major cellular interaction partners of TARS. However, these experiments focused on interactions within cellular lysates despite the fact that TARS appears to stimulate angiogenesis extracellularly. As such, future experiments would benefit from interaction studies focusing on extracellular and cellular surface interactions. Specifically, fluorescently labeled TARS constructs can be used to determine if the enzyme interacts with the cell surface and the specific domains involved. Furthermore, putative receptors or extracellular interaction partners could be isolated by affinity purification of TARS- associated membrane complexes and identified by mass-spectrometry using modifications to established procedures (10). The successful identification of these interactions partners would allow for the better understanding of TARS-mediated angiogenesis and would facilitate subsequent investigations of downstream signaling pathways.

6.1.5 Identification of downstream signaling

In addition to establishing a gradient through associations with the ECM, the two interaction model assumes that a second interaction mediates endothelial cell migration through the activation of downstream signaling pathways. Several common pathways involved in this process have already been investigated in the lab of Karen

192 Lounsbury, including ERK1/2, p38-MAPK, and NF-κB (data not shown). Unfortunately, the results so far have been inconclusive. Treatment of HUVECs with 100 nM TARS or

10 nM BC194 have exhibited no effects on the phosphorylation of p38-MAPK and NF-

κB. Conversely, both treatments have been observed to induce phosphorylation of

ERK1/2 within five minutes. However, this result is often inconsistent. Therefore, other pathways should be investigated.

Several RTKs are involved in the regulation of angiogenesis. As mentioned in

Chapter 4, section 4.3.3, TLR 2 and 4 exhibit interesting characteristics that implicate these proteins as possible TARS receptors. While these receptors often activate NF-κB, which does not appear to be related to TARS-mediated angiogenesis, reported associations between TLRs and integrins can alter the TLR-signaling, much like ECM- bound VEGF (4,16,17). Furthermore, connections between TLR-signaling and activation of signal transducer and activator of transcription 3 (STAT3) signaling have been observed (18). STAT3 is activated through phosphorylation, which allows for the protein’s homodimerization, nuclear translocation, and subsequent upregulation of target genes though interactions with STAT binding elements (SBE) in the DNA. Interestingly,

SBEs have been identified upstream of both VEGF and TARS (TARS site identified by

DECODE, Decipherment of DNA elements database; SABiosciences), allowing for the induction of angiogenesis as well as a potential positive feedback mechanism (19). This potential connection between TARS and STAT3 signaling can be easily investigated through Western blot analysis against phospho-STAT3. Nonetheless, no direct evidence exists for TLRs as the putative TARS receptor, indicating that other targets should also be considered.

193 The connection between growth factor receptors and angiogenesis has been well- established in the literature. The most important of these, of course, is the activation of

VEGFR2 by VEGFA. However, EGFR, PDGFR, and FGFR can strongly influence this process in response to other ligands as well (20). Furthermore, many of these receptors have reported interactions with integrins, which can alter ligand-mediated signaling or initial ligand-independent activation (21). Despite the varying affinities of these receptors to different ligands the down-stream signaling cascades often overlap. The connections between several of these common factors to TARS-mediated angiogenesis, such as ERK1/2 and p38-MAPK, have already been investigated without clear results.

However, the zebrafish data presented in Chapters 3 and 4 implicate that TARS is likely involved in the regulation of endothelial cell migratory mechanism. One potential candidate for further investigation is Src kinase, which is can be activated through phosphorylation by growth factor or integrin receptor stimulation (22,23).

Phosphorylated Src is able to activate focal adhesion kinase (FAK) which can subsequently activate a variety of responses connected to migration, invasion, survival, and growth (24). Furthermore, Src signaling has been demonstrated to activate STAT3

(25,26). As with STAT3 above, the connection between Src and TARS-mediated angiogenesis can be easily investigated by Western blot against phosphorylated Src.

6.2 The nucleo-cytoplasmic distribution of TARS

6.2.1 TARS possesses an N-terminal region of unknown function

In addition to TARS four other ARSs have been associated with effects on angiogenesis (described in Chapter 1). The secondary functions of each of these is

194 dependent on domains acquired over the course of evolution not associated with aminoacylation. While no appended domain has been directly connected to TARS’s secondary function, the N-terminal region of TARS consists of the ubiquitin-like TGS domain (residues 83-144) and an unstructured 82 amino acid extension (residues 1-82) of unknown function. The N-terminal extension appears to have arisen in eukaryotic cells and can be observed in even the most ancient of protists, such as Giardia lamblia.

Conversely, the TGS domain is retained from bacteria to humans. As indicated in

Chapter 4, E. coli TARS is capable of stimulating angiogenesis in tube formation assays, suggesting that the N-terminal extension is dispensable for TARS-mediated angiogenesis once TARS is extracellular (the possibility of intracellular regulation for TARS’s angiogenic function by this extension remains to be determined). Moreover, since the

ΔN1N2-TARS fragment (consisting of the catalytic and anticodon-binding domains) was unable to induce angiogenesis, the TGS domain may be essential for mediating a second extracellular interaction important for angiogenesis. However, additional investigations with the N2 editing domain are required to confirm this. The possible roles for the N- terminal extension, if any, have yet to be investigated.

Structural data and sequence alignments demonstrate that prokaryotic TARS begins at the ubiquitin-like TGS domain while the eukaryotic homologs possess an N- terminal extension that varies between species. In humans this extension consists of 82 amino acids not visible by crystallography or NMR. Additionally, analysis of this region using Prediction of Order and Disorder by machine LEarning (POODLE) software confirms that the first 52 amino acids of the protein are disordered (27). Regions of intrinsic disorder are often sites of protein-protein interactions and are covered in greater

195 detail in Appendix I. Therefore, it is possible that the N-terminal extensions of eukaryotic

TARS mediate interactions required for the regulation of the protein’s secondary functions.

6.2.2 TARS’s N-terminal region possess a possible nuclear localization signal

Alignments of eukaryotic N-terminal regions reveal that the sequences of these extensions can vary greatly among different organisms. This is not surprising, however, as mutations in obligate disordered regions are common owing to the fact that sequence conservation is not as critical to maintaining disordered regions as it is with ordered ones

(28). Despite this sequence variance, however, some features appear to be maintained.

Notably, animals and other eukaryotes, such as Saccharomyces cerevisiae, possess a

KKXK sequence (KKKNK in humans) which is similar to other sequences demonstrated to function as nuclear localization signals (NLSs) in other proteins (29). Of particular interest, an NLS is essential for SARS’s control of angiogenesis through the regulation of

VEGF expression (30). As shown in Chapter 3, pharmacological inhibition of TARS can also regulate VEGF expression, however, there are currently no reports that the KKKNK sequence functions as an NLS, let alone is involved in the regulation of VEGF expression.

In support of TARS’s putative NLS, preliminary data demonstrate that TARS possesses a variable nucleo-cytoplasmic distribution. Immunofluorescence studies reveal the presence of punctate TARS containing structures within the nucleus in both human umbilical vein endothelial cells (HUVEC) (Figure 37; data by Hannah Naughton) and

SKOV3 ovarian carcinoma cells (SKOV) (Karen Lounsbury and Jessica Cassavaugh,

196 unpublished data). Exposure of HUVEC cell cultures to 50 ng/μl VEGF further increased this nuclear population of TARS (Figure 37, right panels). In contrast, treatment with 10 nM BC194 induced a clear peri-nuclear localization as demonstrated by the appearance of a halo-like appearance of TARS immediately around the nucleus (Figure 37, left panels). In combination, the VEGF and BC194 treatments increased both the nuclear and peri-nuclear distribution of TARS. As a possible explanation for the peri-nuclear distribution, the induction of amino acid starvation by BC194-inhibition of TARS may induce endoplasmic reticulum-associated degradation pathways by activating responses similar to the unfolded protein response (9,31,32). This nuclear distribution is partially corroborated by Western blots comparing the presence of TARS in the cytoplasm and nucleus.

To further investigate the subcellular distribution of TARS in endothelial cells, nuclear and cytoplasmic fractions from cultures treated with BC194 and VEGF were obtained and the presence of TARS determined by Western blot (Figure 38A, B). While full-length TARS was not observed in endothelial cell nuclear extracts treated with control, BC194, VEGF, or BC194/VEGF combination conditions, a band corresponding to a ~35 kDa fragment was observed under all conditions and appeared to increase in

BC194 and VEGF treated samples. Furthermore, the effects of BC194 and VEGF on the appearance of the band were additive as the greatest intensity was observed in cultures treated with both. Of particular interest, this band is the same size as would be expect for the combined N-terminal extension, TGS-domain, and N2 editing domain (N1N2)

(Figure 38A). The TARS antibody that was used in these Western blot experiments targets amino acids 1-98, which would fall in this region. Conversely, the antibody used

197

Figure 37. BC194 and VEGF treatments influence the cellular distribution of TARS. Immunofluorescence images of TARS (red) with DAPI nuclear stain (blue) following treatment with DMSO vehicle control (top left), 10 nM BC194 (bottom left), 50 ng/μl VEGF (top right), and 10 nM BC194 and 50 ng/μl VEGF (bottom right). HUVEC cell cultures were grown to near confluency in full EGM-2 media (2% FBS) and then treated as indicated for 24 hours in 10% EGM-2 media (0.2%). Images are shown in duplicates with and without DAPI stain for better visualization of nuclear TARS. Areas occupied by the nucleus are outlined by dotted lines. Images kindly provided by Hannah Naughton. 198 in the immunofluorescence experiments targets the catalytic domain. Since target size is undeterminable with immunofluorescence the exact form of TARS in the different locations remain unknown, which makes it difficult to fully consolidate these data with those acquired by Western blot. Nonetheless, cleavage of the TARS N-terminal domains may be important for the enzyme’s nuclear localization and warrants further investigation. Another similar N-terminal fragment has been previously observed in nuclear fractions from the PC-3 prostate carcinoma cell line following treatment with 10 ng/μl TNFα (Figure 38C). Titration of the BC194 compound demonstrated a visual increase in the appearance of a band corresponding to ~25-37 kDa within the nuclear fraction, which is the approximate location expected for the N1N2 fragment, however, difficulties visualizing the lamin loading control prevented accurate quantification of the nuclear fraction. Full-length TARS was also observed in these fractions as well, but its appearance did not correlate with treatment. A caveat of nuclear fractionation is that components of the peri-nuclear region may also remain associated with the nucleus. As such, it may be that some of the observed full-length TARS is an artifact of the fractionation process.

The possible redistribution of an N-terminal TARS fragment is of particular interest due to its interactions with the von Hippel-Lindau tumor suppressor, a master regulator of the pro-angiogenic transcription factor 1α (HIF1α) (see Appendix I for more information) (33). The interaction with VHL appeared to be greater with N1-TARS relative to the full-length protein, suggesting that liberation of the N1-domain could have downstream biological functions through modification of normal VHL function.

Specifically, the nucleo-cytoplasmic distribution of VHL is thought to contribute to the

199

Figure 38. Effects of BC194, TNFα, and VEGFA on the nucleo-cytoplasmic distribution of TARS. (A, B) Nuclear (A) and cytoplasmic (B) fractions from HUVECs blotted for TARS and loading controls lamin A/C (nuclear fractions) and β-tubulin (cytoplasmic fractions). Please note the differences in molecular weights for nuclear and cytoplasmic TARS. (C) Nuclear and cytoplasmic fractions of PC-3 prostate carcinoma cells blotted for TARS in the presence or absence of 10 ng/μl TNFα. Full length TARS and the putative TARS truncation are indicated. (D, E) Western blots against HA-tag (TARS), HIF1α, and β-tubulin in nuclear and cytoplasmic lysates from HEK293 cells transfected with over-expression plasmids encoding WT TARS (D) or L567V TARS (E) exposed to the indicated concentrations of BC194. * indicates that the corresponding nuclear loading control (Histone 3) did not yield an interpretable blot. Data in parts A and B were kindly provided by Hannah Naughton. 200 regulation of its functions (34-36). When over-expressed in HEK293 cells, full-length

WT TARS was found to redistribute to the nucleus in a BC194 dose-dependent response

(Figure 38D). Unexpectedly, L567V TARS was also found to relocate to the nucleus upon treatment with BC194 (Figure 38E) despite its resistance to the compound’s effects in angiogenesis based assays (9). Possibly, this relocation may be due to a BC194- activated stress response whose activation is normally involved in the nucleo-cytoplasmic distribution of TARS or is simply an artifact associated with the over-expression system.

Interestingly, HIF1α protein levels appeared to stabilize at higher concentrations of

BC194. However, other experiments demonstrate a reduction in HIF1α levels following

BC194 treatment (see Figure 40 in Appendix I), suggesting a complex relationship between TARS and HIF1α that warrants further investigation.

Altogether the above findings suggest that subsequent investigations related to

TARS’s nucleo-cytoplasmic distribution and intracellular fragmentation may be beneficial for complete understanding of the ARS’s angiogenic functions. The importance of the KKKNK sequence for angiogenic and other biological functions can be investigated by generating mRNA for human TARS containing a deletion of this sequence. This mRNA can then be injected into zebrafish embryos along with an anti-

TARS MO to see if any vascular or morphological defects occur. Furthermore, the identity of the putative N-terminal TARS fragment found in the nuclear fractions should be verified. This could be accomplished using affinity purification of TARS from

HUVEC cell cultures with an N-terminal biotin tag followed by SDS-PAGE separation and mass-spectrometry identification of the resulting peptides.

201 6.3 TARS secretion by exosomes

One of the most important questions that remains for the extracellularly acting

ARSs in angiogenesis, YARS, WARS, and TARS, is the mechanism through which these proteins are secreted. The identification of these pathways will likely be critical to the understanding the regulation of these secondary functions. For each of the three ARSs, secretion is induced by specific signaling molecules, such as interferon γ, TNFα, or

VEGF, in cell culture experiments (9,37,38). Additionally, in all cases the cells appear to be remain intact throughout the secretory process, indicating that secretion is not simply the release of cellular components from dying cells. More in depth investigations of

WARS eliminated the classical golgi and Ca2+-dependent secretory pathways (39).

While this observation does not eliminate these canonical pathways as secretion mechanisms for YARS and TARS it does indicate that other mechanisms exist that are capable of secreting entire ARS molecules. As an alternative secretory mechanism,

TARS has been identified in exosomes, protein and RNA rich vesicles ranging in 20 to

100 nm in diameter (Figure 39) ((40,41) and David Lyden direct communication).

Exosomes form by budding into the endosomal compartment, forming multi- vesicular bodies and acquiring portions of the cytoplasm during the process. Fusion of the multi-vesicular bodies with the plasma membrane releases the exosomes and their cargo into the extracellular space. Of particular interest for TARS’s secondary function, exosomes are associated with the establishment of the pre-metastatic niche of melanoma to bone marrow progenitor cells (42). Similarly, TARS is associated with cancer progression and its inhibition by borrelidin has been shown to inhibit the progression of melanoma metastases (1,43). However, one potential caveat with the exosomal secretion

202

Figure 39. Transmission electron migrographs of PC-3-derived exosomes Electron micrographs of exosomes isolated from PC-3 conditioned media. Exosomes were isolated using ExoQuickTM reagent. model concerns the release of the components contained within the vesicles once outside the cell. Interestingly, activation of P2X purinergic receptors have been shown to stimulate the release of exosomes from dendritic cells. While not directly related to endothelial cells, this observation reveals a connection between exosomal and purinergic signaling events (44). More importantly, the nucleotides produced through non-canonical catalytic functions of TARS observed in Chapter 5 can be transported out of the exosomes more readily than the TARS protein itself, suggesting that they could mediate

TARS’s extracellular function. Nonetheless, further experimentation is required to truly establish the role of exosomes in the secretion of angiogenesis-inducing TARS. 203

6.4 Further investigation of non-canonical, catalytic functions of TARS

The identification of non-canonical, catalytic functions (ie. GTPase, ADP formation, and Ap4A synthesis) of TARS provides an alternative mechanism through which TARS can influence receptor-mediated angiogenic signals independent of direct protein-protein interactions. The importance of these reactions is emphasized by the ability of the P2Y1 specific inhibitor MRS2179 to inhibit TARS-mediated angiogenesis.

Specifically, P2Y1 responds to extracellular ADP, which is produced by all of the angiogenic constructs tested either directly or indirectly through the breakdown of Ap4A

(45,46). Although borrelidin and BC194 did not exhibit any significant effects on these reactions, the conformational changes associated with the binding of the compounds may prevent interactions essential for localizing TARS near the purinergic receptors.

Previously, the binding of lysyl-tRNA synthetase to the MITF transcriptional regulator was shown to stimulate MITF’s release from its inhibitor HINT through the local generation of Ap4A (47). Moreover, the Ap4A production by human TARS can be enhanced by TARS phosphorylation suggesting that these TARS secondary reactions may serve a biological purpose (48). Despite these observations, however, the connections between these non-canonical activities remain speculative and could benefit from further investigations.

In order for the various nucleotide products of TARS to contribute to angiogenesis TARS would have to be brought into close contact with the necessary purinergic receptors. The most likely candidate receptor would be P2Y1 based on the

MRS2179 inhibition data and the connection between P2Y1 and endothelial cell

204 migration (46). This putative connection between TARS and P2Y1 could be verified through co-localization studies using immunofluorescence. Additionally, P2Y1 functions by activating the downstream Gq signaling pathways, such as inositol 1,4,5-trisphosphate

(IP3) and diacyl glycerol (DAG). These pathways subsequently activate calcium mobilization within the cell and protein kinase C (PKC) association with the cell membrane (reviewed in (49)). Evidence for P2Y1 activation by TARS can be monitored by Ca2+-responsive, UV-fluorescent dye or Western blotting of membrane fractions for activated PKC (50,51).

In addition to establishing the connection between TARS and purinergic receptors other investigations should focus on the reactions themselves. Based on previous studies of other ARSs, the Ap4A reaction of TARS likely occurs within the

TARS catalytic site (52,53). The GTPase and ADP formation activities of TARS, however, have not been previously reported and occur in absence of the essential R442 residue in the TARS aminoacylation active site. As such, these reactions could occur within a different site of the protein, specifically the editing domain. Therefore, characterization of these reactions would benefit from investigations of their activities in the ΔN1N2-TARS construct. The presence of activity in this construct would suggest that the reaction is catalyzed in the aminoacylation domain while absence, seeing as the isolated N1-domain has already been shown to lack GTPase activity, would indicate the editing domain. Once the proper domain has been established efforts could focus on the generation of point mutations that inhibit these effects for subsequent testing of their angiogenic potential through tube formation or CAM assays. Since Ap4A may also

205 contribute to the generation of ADP, these mutations would have to be paired with the

R442A mutation to eliminate both ADP and Ap4A production.

6.5 TARS is associated with ovarian and prostate tumor progression

6.5.1 Angiogenesis is an essential process in tumorigenesis

The metastatic progression of many tumors is tightly coupled to their angiogenic potential (reviewed in (54,55)). Early in the tumorigenic process the size of these tumors is limited despite the presence of clear proliferative activities. The cause of this restriction appears to be connected to apoptosis initiated by the hypoxic and nutrient deprived state of the tissue. The hyperplastic nature of these tumor cells allows them to quickly grow beyond the support of the available vasculature, limiting their access to essential nutrients and oxygen. Eventually, however, small populations of cells within the tumor acquire the ability to induce the growth of new capillaries within the surrounding tissue and relieve the former constraint. This sudden increase of angiogenic potential by the tumor mass as a whole is referred to as the “angiogenic switch”.

In addition to re-supplying the tumor tissue with nutrients and oxygen from the blood, the angiogenic switch is also necessary for the tumor’s metastatic progression.

Newly formed blood vessels within tumor tissue are often immature and hyperpermeable, allowing metastatic cells direct access to the blood stream (56). Furthermore, the activation of HIF1α, a major tumor-associated inducer of angiogenesis, is known to enhance the expression of the transcriptional repressors, such as SNAIL1, ZEB2, and

E47, which subsequently down-regulate the cellular contact proteins E-cadherin, occludin, and claudin (57,58). Up-regulation of VEGF also disrupts pericyte coverage of

206 nascent vessels through its suppression of platelet-derived growth factor (PDGF) signaling (59). Normally, pericytes function to stabilize mature vessels through the release of anigopoietins and PDGF. Furthermore, the activation of extracellular matrix metalloproteinases (MMPS) facilitate the migratory nature of endothelial cells by breaking down components of the ECM (20). Overall, these processes contribute to the increased permeability associated with immature blood vessels in angiogenesis.

Interestingly, this same reduction in cell-cell contacts and matrix integrity that permits new vasculature to work its way between the tissues also facilitates the epithelial to mesenchymal transition (EMT) associated with metastatic tumors (60,61). Specifically, the increase vessel permeability permits the invasion and dissemination of tumor cells to distal regions of the body. Therefore, the identification of new factors involved in this angiogenic switch would be beneficial the development of novel therapeutic techniques.

6.5.2 The intracellular levels of TARS increases with advancing stages of ovarian and prostate cancers

The recent identification of angiogenic activity in TARS implicates the ARS as a potential factor involved in the progression of tumor development. In support of this hypothesis, data mined from the NIH Cancer Gene Anatomy Project (CGAP) project and published DNA array analyses reveal that TARS is upregulated relative to other ARSs in several types of metastatic cancer (62,63). Of particular interest, metastatic ovarian and prostate cancers are highly angiogenic and exhibit increased expression of TARS according to the above databases (64-66). Moreover, TARS is secreted by cells as part of its angiogenic function, suggesting that serum levels of TARS could be developed as a

207 potential diagnostic for the development and metastatic progression of ovarian and prostate cancers.

The possibility of TARS as a diagnostic marker for prostate and ovarian cancer have been recently investigated by the labs of Christopher Francklyn and Karen

Lounsbury respectively. In these studies the levels of TARS in patient tissue and serum samples collected by the University of Vermont Medical Center were determined using immunohistochemistry and ELISA procedures. The association between TARS and ovarian cancer was recently published (9). As hypothesized, TARS staining intensity was found to increase with advancing stages of ovarian cancer. Furthermore, this TARS staining was found to co-localize with VEGF in tumor sites close to nascent vasculature.

However, while TARS was readily observed in all serum samples, there was no significant differences between TARS levels in lower and high stages of the disease.

Therefore, further investigation is required if TARS is to be established as a diagnostic marker of ovarian cancer.

Currently, diagnosis and risk assessment of prostate cancer relies heavily on the expression of prostate serum antigen (PSA). Elevated levels of PSA in the blood following surgical or radiation therapy, known as PSA biochemical failure, is associated with a high risk of recurrence by the cancer and poor prognosis (67). However, the accuracy of PSA-based diagnoses can be improved though the inclusion of other markers, leading to the current investigations of TARS’s connection to prostate cancer. In these studies TARS protein levels were compared to the Gleason scores of prostate cancer samples at various stages of the disease. This Gleason score ranges from 2 to 10 and is a measurement of a prostate tumors’s likelihood to spread as determined by tissue

208 histology. Effectively, Gleason score increases as cellular morphology deviates from that of healthy cells. Immunohistochemical staining of patient prostate samples revealed that

TARS stain intensity increased significantly as the tissue progressed from a primary

Gleason score of 3 to 4. To further investigate the use of TARS as a prognostic marker, the Gleason scores and TARS levels were compared to the incidence of biochemical failure in the samples. Like with TARS, increasing Gleason scores correlated with a higher risk of PSA biochemical failure. Moreover, calculations of the ratio of the product of TARS stain intensity and the percentage of the tumor at a particular Gleason score over the product of the average TARS intensity and Gleason score percentage revealed a correlation between risk of biochemical failure and elevated TARS levels in Gleason 4 and 5 tumors. Together the above data implicate TARS as a potential indicator of prostate cancer prognostic outcome and warrants further investigation.

A possible mechanistic connection between TARS and prostate cancer was demonstrated by Jon Ramsey when knockdown of TARS in a DU145 cancer cell line using anti-TARS siRNA disrupted the migratory ability of these cells in a wound healing assay. When considered with the vessel patterning defects observed with knockdown of

TARS in zebrafish (Chapter 4, section 4.2.1) these data appear to indicate a role for

TARS in cellular migration that is not limited to vascular cells. As such, the potential roles for TARS as a diagnostic marker in prostate cancer and as a general component of cellular migration warrant further investigation.

209 6.6 Selective inhibition of the malarial parasite Plasmodium falciparum

Perhaps one of the greatest findings of this work is the mechanistic rationale behind the attenuation of borrelidin associated toxicity responses resulting from seemingly minor modifications to the compounds’ structures. As discussed in Chapter 3, the difference between intracellular and extracellular threonine pools may account for the seemingly contradictory observations that BC194 is more susceptible to threonine competition than borrelidin but remains equally potent at inhibiting angiogenesis. A similar mechanism may also be able to explain the high selectivity recently reported for

BC196 and BC220 towards the malarial parasite Plasmodium falciparum relative to humans (68). The modifications associated with these derivatives likely increases their susceptibility to threonine competition to a level that renders them ineffective at practical concentrations for inhibiting aminoacylation in human cells. However, as part of its life- cycle the malarial parasite must infect host erythrocytes and obtain many of its amino acids from break-down of hemoglobin (69). Interestingly, P. falciparum lacks the traditional GCN-2 mediated amino acid starvation response and responds to nutrient deprivation by entering a hibernatory-like state in which the organisms metabolism and growth is significantly reduced (70). Furthermore, measurements of amino acid concentrations in isolated samples of the related avian malarial parasite, Plasmodium lophurae, revealed very low free levels of several amino acid species. Specifically, threonine levels were determined to be <20 μM (71). Together, these observations suggest that erythrocytic malarial parasites are consistently subject to low levels of free amino acid concentrations. Therefore, the reduction in threonine competition within P.

210 falciparum relative to human cells may account for the increased efficacy of BC196 and

BC220 against the malarial parasite.

6.7 Final Remarks

The findings demonstrated by this dissertation work provide clear evidence for the role of TARS in angiogenesis both in vitro and in vivo. The ability to disrupt these effects using TARS specific inhibitors underlies the potential use of TARS as a therapeutic target in angiogenesis associated diseases and, recently, parasite pathogenesis.

However, like with other members of the ARS family, the efficacy of these treatments is negatively impacted by the associated toxicity stemming from their inhibition of the canonical aminoacylation reaction. Using a combination of biochemical, cellular, and organismal based methods we were able to successfully separate these toxicity-related responses from the desired effect on angiogenesis. Moreover, the application of similar techniques may allow for the development of other ARS inhibitors previously limited by their unintended side effects. Together, these findings unveil a novel role for TARS in normal vascular development and provide opportunities for the selective modulation of these effects by TARS-specific inhibitors.

211 6.8 References

1. Wellman, T. L., Eckenstein, M., Wong, C., Rincon, M., Ashikaga, T., Mount, S. L., Francklyn, C. S., and Lounsbury, K. M. (2014) Threonyl-tRNA synthetase overexpression correlates with angiogenic markers and progression of human ovarian cancer. BMC Cancer 14, 620

2. Abramsson, A., Lindblom, P., and Betsholtz, C. (2003) Endothelial and nonendothelial sources of PDGF-B regulate pericyte recruitment and influence vascular pattern formation in tumors. J Clin Invest 112, 1142-1151

3. Bashkin, P., Doctrow, S., Klagsbrun, M., Svahn, C. M., Folkman, J., and Vlodavsky, I. (1989) Basic fibroblast growth factor binds to subendothelial extracellular matrix and is released by heparitinase and heparin-like molecules. Biochemistry 28, 1737-1743

4. Chen, T. T., Luque, A., Lee, S., Anderson, S. M., Segura, T., and Iruela-Arispe, M. L. (2010) Anchorage of VEGF to the extracellular matrix conveys differential signaling responses to endothelial cells. J Cell Biol 188, 595-609

5. Lee, S., Jilani, S. M., Nikolova, G. V., Carpizo, D., and Iruela-Arispe, M. L. (2005) Processing of VEGF-A by matrix metalloproteinases regulates bioavailability and vascular patterning in tumors. J Cell Biol 169, 681-691

6. Stupack, D. G., and Cheresh, D. A. (2002) ECM remodeling regulates angiogenesis: endothelial integrins look for new ligands. Sci STKE 2002, pe7

7. Weis, S. M., and Cheresh, D. A. (2011) alphaV integrins in angiogenesis and cancer. Cold Spring Harbor perspectives in medicine 1, a006478

8. Somanath, P. R., Malinin, N. L., and Byzova, T. V. (2009) Cooperation between integrin alphavbeta3 and VEGFR2 in angiogenesis. Angiogenesis 12, 177-185

9. Williams, T. F., Mirando, A. C., Wilkinson, B., Francklyn, C. S., and Lounsbury, K. M. (2013) Secreted Threonyl-tRNA synthetase stimulates endothelial cell migration and angiogenesis. Sci Rep 3, 1317

10. Humphries, J. D., Byron, A., Bass, M. D., Craig, S. E., Pinney, J. W., Knight, D., and Humphries, M. J. (2009) Proteomic analysis of integrin-associated complexes identifies RCC2 as a dual regulator of Rac1 and Arf6. Science signaling 2, ra51

11. Charo, I. F., Nannizzi, L., Smith, J. W., and Cheresh, D. A. (1990) The vitronectin receptor alpha v beta 3 binds fibronectin and acts in concert with alpha 5 beta 1 in promoting cellular attachment and spreading on fibronectin. J Cell Biol 111, 2795-2800

212 12. Pierschbacher, M. D., and Ruoslahti, E. (1984) Cell attachment activity of fibronectin can be duplicated by small synthetic fragments of the molecule. Nature 309, 30-33

13. Hocking, D. C., and Kowalski, K. (2002) A cryptic fragment from fibronectin's III1 module localizes to lipid rafts and stimulates cell growth and contractility. J Cell Biol 158, 175-184

14. Sottile, J., and Hocking, D. C. (2002) Fibronectin polymerization regulates the composition and stability of extracellular matrix fibrils and cell-matrix adhesions. Mol Biol Cell 13, 3546-3559

15. Schwarz, M. A., Zheng, H., Liu, J., Corbett, S., and Schwarz, R. E. (2005) Endothelial-monocyte activating polypeptide II alters fibronectin based endothelial cell adhesion and matrix assembly via alpha5 beta1 integrin. Exp Cell Res 311, 229-239

16. Gianni, T., Leoni, V., Chesnokova, L. S., Hutt-Fletcher, L. M., and Campadelli- Fiume, G. (2012) alphavbeta3-integrin is a major sensor and activator of innate immunity to herpes simplex virus-1. Proc Natl Acad Sci U S A 109, 19792-19797

17. Kawai, T., and Akira, S. (2006) TLR signaling. Cell Death Differ 13, 816-825

18. Hu, X., Chen, J., Wang, L., and Ivashkiv, L. B. (2007) Crosstalk among Jak- STAT, Toll-like receptor, and ITAM-dependent pathways in macrophage activation. J Leukoc Biol 82, 237-243

19. Chen, Z., and Han, Z. C. (2008) STAT3: a critical transcription activator in angiogenesis. Medicinal research reviews 28, 185-200

20. Carmeliet, P., and Jain, R. K. (2011) Molecular mechanisms and clinical applications of angiogenesis. Nature 473, 298-307

21. Yamada, K. M., and Even-Ram, S. (2002) Integrin regulation of growth factor receptors. Nat Cell Biol 4, E75-76

22. Arias-Salgado, E. G., Lizano, S., Sarkar, S., Brugge, J. S., Ginsberg, M. H., and Shattil, S. J. (2003) Src kinase activation by direct interaction with the integrin beta cytoplasmic domain. Proc Natl Acad Sci U S A 100, 13298-13302

23. Bromann, P. A., Korkaya, H., and Courtneidge, S. A. (2004) The interplay between Src family kinases and receptor tyrosine kinases. Oncogene 23, 7957- 7968

24. Lechertier, T., and Hodivala-Dilke, K. (2012) Focal adhesion kinase and tumour angiogenesis. J Pathol 226, 404-412

213 25. Cao, X., Tay, A., Guy, G. R., and Tan, Y. H. (1996) Activation and association of Stat3 with Src in v-Src-transformed cell lines. Mol Cell Biol 16, 1595-1603

26. Turkson, J., Bowman, T., Garcia, R., Caldenhoven, E., De Groot, R. P., and Jove, R. (1998) Stat3 activation by Src induces specific gene regulation and is required for cell transformation. Mol Cell Biol 18, 2545-2552

27. Hirose, S., Shimizu, K., Kanai, S., Kuroda, Y., and Noguchi, T. (2007) POODLE- L: a two-level SVM prediction system for reliably predicting long disordered regions. Bioinformatics 23, 2046-2053

28. Brown, C. J., Takayama, S., Campen, A. M., Vise, P., Marshall, T. W., Oldfield, C. J., Williams, C. J., and Dunker, A. K. (2002) Evolutionary rate heterogeneity in proteins with long disordered regions. Journal of molecular evolution 55, 104- 110

29. Kosugi, S., Hasebe, M., Matsumura, N., Takashima, H., Miyamoto-Sato, E., Tomita, M., and Yanagawa, H. (2009) Six classes of nuclear localization signals specific to different binding grooves of importin alpha. J Biol Chem 284, 478-485

30. Shi, Y., Xu, X., Zhang, Q., Fu, G., Mo, Z., Wang, G. S., Kishi, S., and Yang, X. L. (2014) tRNA synthetase counteracts c-Myc to develop functional vasculature. eLife 3, e02349

31. Teske, B. F., Wek, S. A., Bunpo, P., Cundiff, J. K., McClintick, J. N., Anthony, T. G., and Wek, R. C. (2011) The eIF2 kinase PERK and the integrated stress response facilitate activation of ATF6 during endoplasmic reticulum stress. Mol Biol Cell 22, 4390-4405

32. Tsai, Y. C., and Weissman, A. M. (2010) The Unfolded Protein Response, Degradation from Endoplasmic Reticulum and Cancer. Genes & cancer 1, 764- 778

33. Maxwell, P. H., Wiesener, M. S., Chang, G. W., Clifford, S. C., Vaux, E. C., Cockman, M. E., Wykoff, C. C., Pugh, C. W., Maher, E. R., and Ratcliffe, P. J. (1999) The tumour suppressor protein VHL targets hypoxia-inducible factors for oxygen-dependent proteolysis. Nature 399, 271-275

34. Cai, Q., and Robertson, E. S. Ubiquitin/SUMO modification regulates VHL protein stability and nucleocytoplasmic localization. PLoS One 5

35. Khacho, M., Mekhail, K., Pilon-Larose, K., Pause, A., Cote, J., and Lee, S. (2008) eEF1A is a novel component of the mammalian nuclear protein export machinery. Mol Biol Cell 19, 5296-5308

214 36. Lewis, M. D., and Roberts, B. J. (2003) Role of nuclear and cytoplasmic localization in the tumour-suppressor activity of the von Hippel-Lindau protein. Oncogene 22, 3992-3997

37. Beresten, S. F., Zargarova, T. A., Favorova, O. O., Rubikaite, B. I., Ryazanov, A. G., and Kisselev, L. L. (1989) Molecular and cellular studies of tryptophanyl- tRNA synthetase using monoclonal antibodies. Evaluation of a common antigenic determinant in eukaryotic, prokaryotic and archaebacterial enzymes which maps outside the catalytic domain. Eur J Biochem 184, 575-581

38. Wakasugi, K., and Schimmel, P. (1999) Two distinct cytokines released from a human aminoacyl-tRNA synthetase. Science 284, 147-151

39. Kapoor, M., Zhou, Q., Otero, F., Myers, C. A., Bates, A., Belani, R., Liu, J., Luo, J. K., Tzima, E., Zhang, D. E., Yang, X. L., and Schimmel, P. (2008) Evidence for annexin II-S100A10 complex and plasmin in mobilization of cytokine activity of human TrpRS. J Biol Chem 283, 2070-2077

40. Buschow, S. I., van Balkom, B. W., Aalberts, M., Heck, A. J., Wauben, M., and Stoorvogel, W. (2010) MHC class II-associated proteins in B-cell exosomes and potential functional implications for exosome biogenesis. Immunology and cell biology 88, 851-856

41. Welton, J. L., Khanna, S., Giles, P. J., Brennan, P., Brewis, I. A., Staffurth, J., Mason, M. D., and Clayton, A. (2010) Proteomics analysis of bladder cancer exosomes. Mol Cell Proteomics 9, 1324-1338

42. Peinado, H., Aleckovic, M., Lavotshkin, S., Matei, I., Costa-Silva, B., Moreno- Bueno, G., Hergueta-Redondo, M., Williams, C., Garcia-Santos, G., Ghajar, C., Nitadori-Hoshino, A., Hoffman, C., Badal, K., Garcia, B. A., Callahan, M. K., Yuan, J., Martins, V. R., Skog, J., Kaplan, R. N., Brady, M. S., Wolchok, J. D., Chapman, P. B., Kang, Y., Bromberg, J., and Lyden, D. (2012) Melanoma exosomes educate bone marrow progenitor cells toward a pro-metastatic phenotype through MET. Nature medicine 18, 883-891

43. Funahashi, Y., Wakabayashi, T., Semba, T., Sonoda, J., Kitoh, K., and Yoshimatsu, K. (1999) Establishment of a quantitative mouse dorsal air sac model and its application to evaluate a new angiogenesis inhibitor. Oncol Res 11, 319- 329

44. Pizzirani, C., Ferrari, D., Chiozzi, P., Adinolfi, E., Sandona, D., Savaglio, E., and Di Virgilio, F. (2007) Stimulation of P2 receptors causes release of IL-1beta- loaded microvesicles from human dendritic cells. Blood 109, 3856-3864

45. Luthje, J., and Ogilvie, A. (1988) Catabolism of Ap4A and Ap3A in whole blood. The dinucleotides are long-lived signal molecules in the blood ending up as intracellular ATP in the erythrocytes. Eur J Biochem 173, 241-245 215 46. Shen, J., and DiCorleto, P. E. (2008) ADP stimulates human endothelial cell migration via P2Y1 nucleotide receptor-mediated mitogen-activated protein kinase pathways. Circ Res 102, 448-456

47. Lee, Y. N., Nechushtan, H., Figov, N., and Razin, E. (2004) The function of lysyl- tRNA synthetase and Ap4A as signaling regulators of MITF activity in FcepsilonRI-activated mast cells. Immunity 20, 145-151

48. Dang, C. V., and Traugh, J. A. (1989) Phosphorylation of threonyl- and seryl- tRNA synthetase by cAMP-dependent protein kinase. A possible role in the regulation of P1, P4-bis(5'-adenosyl)-tetraphosphate (Ap4A) synthesis. J Biol Chem 264, 5861-5865

49. Newton, A. C. (1995) Protein kinase C: structure, function, and regulation. J Biol Chem 270, 28495-28498

50. Miranti, C. K., Ohno, S., and Brugge, J. S. (1999) Protein kinase C regulates integrin-induced activation of the extracellular regulated kinase pathway upstream of Shc. J Biol Chem 274, 10571-10581

51. Takahashi, A., Camacho, P., Lechleiter, J. D., and Herman, B. (1999) Measurement of intracellular calcium. Physiological reviews 79, 1089-1125

52. Belrhali, H., Yaremchuk, A., Tukalo, M., Berthet-Colominas, C., Rasmussen, B., Bosecke, P., Diat, O., and Cusack, S. (1995) The structural basis for seryl- adenylate and Ap4A synthesis by seryl-tRNA synthetase. Structure 3, 341-352

53. Guo, R. T., Chong, Y. E., Guo, M., and Yang, X. L. (2009) Crystal structures and biochemical analyses suggest a unique mechanism and role for human glycyl- tRNA synthetase in Ap4A homeostasis. J Biol Chem 284, 28968-28976

54. Bergers, G., and Benjamin, L. E. (2003) Tumorigenesis and the angiogenic switch. Nat Rev Cancer 3, 401-410

55. Carmeliet, P., and Jain, R. K. (2000) Angiogenesis in cancer and other diseases. Nature 407, 249-257

56. Zetter, B. R. (1998) Angiogenesis and tumor metastasis. Annu Rev Med 49, 407- 424

57. Esteban, M. A., Tran, M. G., Harten, S. K., Hill, P., Castellanos, M. C., Chandra, A., Raval, R., O'Brien T, S., and Maxwell, P. H. (2006) Regulation of E-cadherin expression by VHL and hypoxia-inducible factor. Cancer Res 66, 3567-3575

58. Harten, S. K., Shukla, D., Barod, R., Hergovich, A., Balda, M. S., Matter, K., Esteban, M. A., and Maxwell, P. H. (2009) Regulation of renal epithelial tight

216 junctions by the von Hippel-Lindau tumor suppressor gene involves occludin and claudin 1 and is independent of E-cadherin. Mol Biol Cell 20, 1089-1101

59. Greenberg, J. I., Shields, D. J., Barillas, S. G., Acevedo, L. M., Murphy, E., Huang, J., Scheppke, L., Stockmann, C., Johnson, R. S., Angle, N., and Cheresh, D. A. (2008) A role for VEGF as a negative regulator of pericyte function and vessel maturation. Nature 456, 809-813

60. Garcia-Roman, J., and Zentella-Dehesa, A. (2013) Vascular permeability changes involved in tumor metastasis. Cancer letters 335, 259-269

61. Huber, M. A., Kraut, N., and Beug, H. (2005) Molecular requirements for epithelial-mesenchymal transition during tumor progression. Curr Opin Cell Biol 17, 548-558

62. Chandran, U. R., Ma, C., Dhir, R., Bisceglia, M., Lyons-Weiler, M., Liang, W., Michalopoulos, G., Becich, M., and Monzon, F. A. (2007) Gene expression profiles of prostate cancer reveal involvement of multiple molecular pathways in the metastatic process. BMC Cancer 7, 64

63. Yu, Y. P., Landsittel, D., Jing, L., Nelson, J., Ren, B., Liu, L., McDonald, C., Thomas, R., Dhir, R., Finkelstein, S., Michalopoulos, G., Becich, M., and Luo, J. H. (2004) Gene expression alterations in prostate cancer predicting tumor aggression and preceding development of malignancy. J Clin Oncol 22, 2790- 2799

64. Huss, W. J., Hanrahan, C. F., Barrios, R. J., Simons, J. W., and Greenberg, N. M. (2001) Angiogenesis and prostate cancer: identification of a molecular progression switch. Cancer Res 61, 2736-2743

65. Lee, S., Garner, E. I., Welch, W. R., Berkowitz, R. S., and Mok, S. C. (2007) Over-expression of hypoxia-inducible factor 1 alpha in ovarian clear cell carcinoma. Gynecologic oncology 106, 311-317

66. Wong, C., Wellman, T. L., and Lounsbury, K. M. (2003) VEGF and HIF-1alpha expression are increased in advanced stages of epithelial ovarian cancer. Gynecologic oncology 91, 513-517

67. Lilja, H., Ulmert, D., and Vickers, A. J. (2008) Prostate-specific antigen and prostate cancer: prediction, detection and monitoring. Nat Rev Cancer 8, 268-278

68. Novoa, E. M., Camacho, N., Tor, A., Wilkinson, B., Moss, S., Marin-Garcia, P., Azcarate, I. G., Bautista, J. M., Mirando, A. C., Francklyn, C. S., Varon, S., Royo, M., Cortes, A., and Ribas de Pouplana, L. (2014) Analogs of natural aminoacyl- tRNA synthetase inhibitors clear malaria in vivo. Proc Natl Acad Sci U S A 111, E5508-5517

217 69. Goldberg, D. E., Slater, A. F., Cerami, A., and Henderson, G. B. (1990) Hemoglobin degradation in the malaria parasite Plasmodium falciparum: an ordered process in a unique organelle. Proc Natl Acad Sci U S A 87, 2931-2935

70. Babbitt, S. E., Altenhofen, L., Cobbold, S. A., Istvan, E. S., Fennell, C., Doerig, C., Llinas, M., and Goldberg, D. E. (2012) Plasmodium falciparum responds to amino acid starvation by entering into a hibernatory state. Proc Natl Acad Sci U S A 109, E3278-3287

71. Sherman, I. W., and Mudd, J. B. (1966) Malaria infection (Plasmodium iophurae): changes in free amino acids. Science 154, 287-289

218 APPENDIX I: IDENTIFICATION AND CHARACTERIZATION OF

THREONYL-tRNA SYNTHETASE INTERACTION PARTNERS

A1.1 Introduction

The Q566W TARS and A538W TARS mutations demonstrate a clear connection between borrelidin-induced conformational changes and the anti-angiogenic effects of the compound (see Chapter 4, section 4.2.6). Structural prediction software suggest that both mutations would shift a helix consisting of residues 492-547 into a conformation that mimics the structure of the TARS-borrelidin complex (1). Since neither of the TARS mutants exhibited the pro-angiogenic character of the wildtype protein, this conformational change alone seems responsible for borrelidin’s anti- angiogenic mechanism. Furthermore, Q566W TARS is still able to catalyze the canonical aminoacylation reaction, suggesting that this conformational change is not important for the normal function of TARS in translation. Therefore, the most likely explanation is that borrelidin inhibits TARS-mediated angiogenesis by disrupting an interaction between

TARS and another molecule.

The modular structure of TARS consists of four major domains. Three of these,

N2, catalytic, and anticodon-binding domains, possess clear functions within the canonical role of the protein (See Chapter 1, section 1.3.1). Furthermore, the ability of

ΔN1N2-TARS to rescue TARS inhibition (Min Guo, unpublished data) demonstrate that

TARS’s aminoacylation function requires only the catalytic and anticodon-binding domains. In contrast, the fourth domain, known as the N1 or TGS domain, possesses an ubiquitin-like structure and is dispensable for canonical TARS function. Likewise, the

219 eukaryotic TARS N-terminal extension appears to be unstructured making it difficult to predict its purpose. However, the association of ubiquitin-like proteins and intrinsically disordered sequences as mediators of protein-protein interactions suggests that possibility that these structures are important for the non-canonical functions of TARS.

Analysis of the eukaryotic TARS N-terminal 82 amino acid extension using the

Prediction of Order and Disorder by machine LEarning (POODLE) system indicates a high probability of disorder (>50%) in amino acids 1-52 (2). The conformational plasticity of disordered proteins provide many benefits in the formation of protein-protein interactions. The less compact structure of these disordered regions can facilitate interactions through an extended capture radius and increased interaction surface area

(3,4). Additionally, these regions can adopt new conformations upon complex formation allowing for induced fit interactions with several possible binding partners (5-7).

Interestingly, around 70% of proteins involved in signaling functions are predicted to contain unstructured regions, implicating these regions in important biological functions

(8).

Ubiquitin-like proteins (UBLs) represent a large and diverse superfamily of proteins recognizable by their characteristic β-grasp superfold (9). Several family members, such as the eponymous ubiquitin, SUMO, and Nedd, consist exclusively of the ubiquitin-like fold and modify target proteins through the formation of covalent adducts.

Alternatively, other UBLs contain ubiquitin-like domains as part of their modular structure. Despite their common fold, however, the sequence and function of UBLs can vary greatly. These variations between UBLs alter their surface properties and contributes to the specificity of their interactions with proteins containing ubiquitin-

220 binding domains (UBDs). The binding affinities between UBDs and UBLs are relatively weak with Kd values ranging from 5-500 μM for single ubiquitin like domains but can be augmented through the attachment of additional UBL monomers or contributions from neighboring protein domain, allowing for both tight and transient binding events.

Furthermore, like their corresponding UBLs, UBDs consist of a broad range of sequences and structures (10,11). Together, this great variance between ligand and binding partner allows for the highly specific interactions critical for their wide range of cellular functions (12). Therefore, protein-protein interactions appears to be a common characteristic of UBLs.

The combination of a disordered region, ubiquitin-like domain, and conformational dependent pro-angiogenic function emphasizes the importance of protein- protein interactions in TARS’s secondary function. Unfortunately, the complexity of angiogenesis provides a list of potential candidate interactions too numerous to investigate individually within a reasonable amount of time. Using existing protein- interaction databases and our own affinity purification followed by mass spectrometry experiments this list can be reduced to a manageable quantity. Based on literature searches three candidates were elected for further investigation: the von Hippel-Lindau tumor suppressor (VHL), eukaryotic elongation factor 1α (eEF1α), and poly (ADP- ribose) polymerase 1 (PARP1).

A1.2 Results

A1.2.1 Characterizing the TAR-VHL interaction

221 One of the most important regulators of angiogenesis is the von Hippel Lindau

(VHL) tumor suppressor protein, a master regulator of the potent angiogenesis-inducer hypoxia inducible factor 1α (HIF1α). HIF1α is constitutively expressed under most conditions; however, its cellular levels are strictly controlled by the Von Hippel Lindau

E3-ligase system. During normoxia, a family of prolyl hydroxylases (PHDs) hydroxylate prolines P402 and P564 of HIF1α (13). VHL recognizes these hydroxyl groups and initiates the ubiquitination and eventual degradation of HIF1α. During hypoxia the activities of the PHDs are significantly reduced, rescuing HIF1α from VHL-mediated ubiquitination. This stabilization increases levels of HIF1α, which is transported into the nucleus to form an active transcription factor complex with HIF1β. This complex is then able to bind to hypoxia response elements (HRE), consequently up-regulating hypoxic response genes related to angiogenesis, such as vascular endothelial growth factor

(VEGF), as well as genes involved in glycolysis and fermentation (14-16). This shift to fermentative processes under aerobic conditions, classically referred to as the Warburg effect, provides critical metabolic intermediates for macromolecule biosynthesis and emphasizes the significance of the VHL/HIF1α pathway dysregulation in cancer progression (17-19). Of particular interest to this work, a large scale protein-protein interaction study has identified interactions between the VHL tumor suppressor and

TARS, suggesting that TARS’s non-canonical, angiogenic function could be related to the regulation of VHL.

As an E3 ubiquitin ligase, VHL serves as the target recognition component for protein complexes that catalyze the attachment of poly-ubiquitin chains to other proteins.

In the regulation of HIF1α this requires forming a complex with Rbx-1, elongin C,

222

Figure 40. TARS may regulate VHL E3-ligase function. (A) Structure of VHL (yellow) in complex with elongin C (purple) and elongin B (cyan) (PDB: 1LQB). Inset – Superposition of TARS TGS domain (green; PDB: 1WWT) and elongin B (cyan; PDB: 1LQB). (B) Western blot for HIF1α and TARS from cell lysates under normoxic and hypoxic conditions at various concentrations of borrelidin or BC194. The prolyl hydroxylase inhibitor, CoCl2, was used as a positive control for HIF1α stabilization. Data were kindly provided by Karen Lounsbury. elongin B, and Cullin-2 (20). Structural comparisons have revealed that one of these partners, elongin B, shares a similar ubiquitin-like fold with the N1-domain of TARS. As

TARS has been shown to interact with VHL (21), it may be that elongin B is displaced by

223 the N1 domain of TARS and the assembly of the complex is mediated through elongin C

(Figure 40A). Disruption of this interaction could then modify VHL’s ability to regulate

HIF1α and, through it angiogenesis. In support of this data from the lab of Karen

Lounsbury have shown that treatment of human embryonic kidney 293 cells (HEK293) with 10 nM of the TARS inhibitors borrelidin or BC194 reduces the protein levels of

HIF1α under hypoxic conditions (Figure 40B), thereby establishing a connection between

TARS and VHL function. Therefore, the TARS-VHL interaction was characterized in further detail.

Initial validation of the TARS-VHL interaction was achieved by investigating the ability of various TARS constructs to “pull-down” VHL. Each TARS construct was cloned into an overexpression vector containing a BirA biotinylation site and an HA epitope and co-transfected into HEK293 cells with a plasmids encoding the biotin transferase BirA and c-Myc-tagged VHL. The presence of BirA would catalyze the biotinylation of the TARS construct, allowing for the protein to be affinity-purified by streptavidin-coated beads from the cell lysates along with anything currently interacting with it. Interactions are then identified by Western blot. To account for possible non- specific binding to the streptavidin beads, the ability of VHL to recapitulate the TARS pull-down experiments was also performed pulling VHL out of solution with anti-c-Myc sepharose-conjugated beads and observing interaction partners by Western blot. Both full length TARS (Figure 41A) and the isolated N1-domain (Figure 41B) were found to co-immunoprecipitate with VHL, affirming the TARS-VHL interaction and supporting the N1-domain as the specific action site. Interestingly, the bands for N1-TARS were noticeably more intense than those for the full-length protein, suggesting that the isolated

224

Figure 41. TARS interacts with VHL through its N1-domain. (A, B) Western blots of TARS and VHL following affinity purification (AP: Biotin) of Full length TARS (A) or N1-TARS (B) and VHL immunoprecipitation (IP: Myc). Immunoprecipition with IgG-coated beads was used as a control for VHL immunoprecipitation to demonstrate the specificity of the binding. (C) Western blots of TARS affinity-purified lysates treated with various concentrations of BC194 (0.01 to 1000 nM) for full-length TARS (IB: HA) and VHL (IB: Myc). Data in (C) were kindly provided by Joann Russo and Jason Botten.

N1-domain forms stronger interactions with VHL. However, it may also be that the smaller size of the N1-domain allows for its increased expression or reduces steric hindrance with other molecules during the pull-down process and warrants further investigation. While these data would support the hypothesis that TARS binds to VHL

225 through elongin C by displacing elongin B, additional information is required to confirm this interaction.

Borrelidin and its derivative BC194 have been previously shown to inhibit

TARS-mediated angiogenesis. Therefore, if the TARS-VHL interaction was important for TARS’s non-canonical role in angiogenesis, borrelidin’s anti-angiogenic properties may occur by disrupting this interaction. Since the borrelidin-binding site is located within the TARS catalytic domain, only the full-length TARS construct was investigated

(Figure 41C). The addition of 0.01 nM, 10 nM, 100 nM, and 1000 nM of BC194 had no effect on the ability of TARS to pull-down VHL, suggesting that the interaction is not susceptible to inhibition by borrelidin and related compounds.

Owing to its association with tumor angiogenesis and anaerobic metabolism,

VHL is considered a tumor suppressor protein. The inhibition of alteration of normal

VHL function through point mutations in the protein have been associated with the onset of several tumor types. Several of these mutations are located along the VHL-elongin C interface. Therefore, it may be possible that these mutations could disrupt normal VHL function by destabilizing its interaction with elongin C. Since elongin B binds through elongin C it seems likely that these mutations would also disrupt the TARS-VHL interaction if it truly binds by displacing elongin B. To test this hypothesis, three mutations were introduced into VHL, R167A, R167W, and C162A, corresponding to the two most-common pathogenic mutations at the VHL-elongin C interface, and subsequently investigated for their ability to disrupt interactions between full-length

TARS (Figure 42A) and N1-TARS (Figure 42B) (22). The C162A mutation appeared to disrupt the interactions between VHL and both the full-length TARS and N1-domain.

226 Additionally, R167W also was found to inhibit the N1-domain interaction with VHL

(Likely R167W would have inhibited the VHL interaction with full-length TARS as well, but due to a sample processing error these data were not obtained). Together, these data support the hypothesis that TARS interacts with VHL through elongin C by displacing elongin B with its N1-domain. However, it should be noted that these experiments were only performed once and should be repeated for confirmation. Furthermore, a stronger band was anticipated for the N1-TARS-WT VHL interaction than was received (see

Figure 41B), suggesting that an error may have occurred in the isolation of this sample.

Figure 42. TARS interacts with VHL by displacing elongin B. (A, B) Western blots of TARS (IB:HA) and VHL (IB: Myc) following affinity purification (AP: Biotin) of Full length TARS (A) or N1-TARS (B). Where indicated the wildtype myc- tagged VHL plasmid was substituted with plasmids encoding myc-tagged R167A, R167W, or C162A VHL constructs.

227

As an alternative method to demonstrate that TARS binds to VHL by displacing elongin B, previous reports indicate that peptides mimicking the elongin C component of the elongin C-VHL interface could disrupt both elongin B and C interactions with VHL.

Single amino acid alanine substitutions in the peptide, notably L158, were able to block the effect of this compound, which indicates the importance of these residues for this interaction (23). Unfortunately, this study was not discovered until after the above experiments were performed and suggests that future repetitions of these experiments could be improved using a L158A peptide.

A1.2.2 Identifying the TARS interactome through mass-spectrometry

Since the interaction between VHL and TARS appeared to be unaffected by borrelidin, the ability of TARS to interact with other proteins partners was investigated using mass spectrometry. Specific emphasis was placed on any interaction with ties to angiogenesis or VHL function and were subsequently investigated by Western blot to determine the effects of borrelidin. For these experiments HEK293 cell lysates overexpressing HA- and biotin-tagged TARS alone (condition 1) or with myc-tagged VHL

(condition 2) were affinity purified using streptavidin coated beads. Unbound proteins were removed through successive washing and the prey proteins identified through mass spec analysis of SDS-PAGE resolved bands. Non-specific interactions were determined using an empty vector control and removed from subsequent analyses. High confidence was given to any interaction with a cross-correlation value (XC) greater than 20.00.

Observed interactions are shown in Table 13.

228 Table 13. Interactions for affinity purified TARS identified by mass spectrometry Translation Related Name XC Peptides Condition Function TARS 218.32 36 1,2 Threonyl-tRNA Synthetase (cytoplasmic) TARSL2 38.28 4 1,2 Likely TARS isoform EPRS 30.19 4 2 Bifunctional aminoacyl-tRNA synthetase RPS15A 10.22 1 2 40S Ribosomal protein RPS20; SNORD54 10.18 1 1,2 40S Ribosomal protein RPS3A; SNORD73A 10.18 1 1 40S ribosomal protein S3a eEF1A1 10.16 2,1 1,2 Elongation factor 1-alpha 1 MARS 10.15 1 2 Methionyl-tRNA Synthetase (cytoplasmic) LARS 10.14 1 1 Leucyl-tRNA Synthetase (cytoplasmic) eEF1G 10.11 1 1 Elongation factor 1-gamma TARS2 8.13 1 1,2 Mito TARS

Protein Degradation # of Name XC Condition Function Ions Sim. E2D2-2 10.13 1 2 Similar to Ubq-conjugating enzyme E2D 2 Med8 10.10 1 1 E3-ligase- Elongin B,C, cullin, rbx based VHL-1 10.10 1 1 24kDa VHL-3 10.10 1 1 18kDa

Mitochondrial/metabolism # of Name XC Condition Function Ions SLC25A5 20.22 3 2 ADP/ATP Translocase 2 Glyceraldehyde-3-phosphate GAPDH 10.16 1 1 dehydrogenase ENO1 10.15 1 1 Alpha-enolase ENO2 10.15 1 1 Gamma-enolase ENO3 10.15 1 1 Beta-enolase HSPD1 10.15 1 2 Mito Heat shock protein

Nuclear-cytoplasmic trafficking # of Name XC Condition Function Ions eEF1A1 10.16 1 1,2 Elongation factor 1-alpha 1 NUP107 10.15 1 2 Nuclear pore protein

Name XC Peptides Condition Function PARP1 30.16 4 1 Poly [ADP-ribose] polymerase 1 Isoform 3 Myosin 1C 20.18 2 1 Myosin-Ic HNRNPC 20.18 2 1 Heterogeneous nuclear RNPs C1/C2 POTEE 20.14 2 1 POTE ankyrin domain family member E CALD1 10.18 1 1 Caldesmon Putative ATP-dependent RNA helicase DHx30-1 10.17 1 1 DHX30

229 Putative ATP-dependent RNA helicase DHx30-2 10.17 1 1 DHX30 Putative ATP-dependent RNA helicase DHx30-3 10.17 1 1 DHX30 HNRNPU 10.14 1 1 Heterogeneous nuclear ribonucleoprotein U RUVBL1 10.14 1 1 RuvB Like NEFM 10.13 1 1 Neurofilament medium polypeptide PRPH 10.13 1 1 Peripherin ILF2 10.11 1 1 Interleukin enhancer-binding factor 2 HSP90AB3P 8.16 1 1 Putative heat shock protein HSP 90-beta-3 CLTB 8.15 1 1 Clathrin light chain B TDRD6 8.13 1 1 Tudor domain-containing protein 6 ALMS1 8.11 1 1 Alstrom syndrome protein 1 1Condition 1 - Over-expressed TARS; Condition 2 - Over-expressed TARS and VHL

Several notable interactions were identified in these experiments including many related to metabolism, protein translation, and nuclear-cytoplasmic trafficking. Although

XC values indicated low confidence, the results once again confirmed the previously identified interactions with VHL. Furthermore, several members of the multi-synthetase complex were observed even though TARS itself is not recognized as a component of the complex (24). Specifically, high confidence was observed for its interaction with the glutamyl/prolyl-tRNA synthetase (EPRS) and less confidently with leucyl-tRNA synthetase (LARS) and methionyl-tRNA synthetase (MARS). Interestingly, both EPRS and LARS are reported to be key components of various survival responses through

GAIT and mTOR mediated responses respectively (25,26). The significance of these interactions is not yet apparent, but may indicate a higher degree of cross-talk between the translational machinery than previously known. Finally, two other interactions were also of interest: poly(ADP-ribose)polymerase1 (PARP1) and eukaryotic elongation factor

1α1 (eEF1A1).

A1.2.3 Characterizing TARS’s interaction with PARP1

230 Canonically, PARP1 is viewed as a member of the DNA-damage response in eukaryotic cells. In response to genotoxic stress PARP1 will homodimerize and catalyze the attachment of ADP-ribose polymers to itself (27,28). The presence of this modification recruits other factors to the damaged site to initiate the necessary repair events (29,30).

However, over the last decade the functions of PARP1 have expanded beyond DNA repair and stability including inflammation and angiogenesis. Over ADP-ribosylation by

PARP1 can lead death associated pathways due to the depletion of ATP from the synthesis of NAD+ (31,32). Alternatively, direct modification of inflammatory genes can lead to their activation, such as NF-κB (33,34). While the connection between inflammation and angiogenesis has been well established PARP1 also appears to have a more direct connection as well. Treatment with several PARP1 inhibitors was found to decrease angiogenesis as determined by in vitro HUVEC tube-formation and aortic ring branching assays as well as in vivo rat matrigel plug assays (35,36). Specifically, these inhibitors disrupted endothelial cell migration, similar to the effects observed with TARS inhibition. Furthermore, PARP1 has been demonstrated to form protein-protein interactions with the potent angiogenic factor HIF1α and synergistically up-regulates genes under its control (37). The importance of this connection to TARS’s secondary function is only made greater by the interaction with the master regulator of HIF1α,

VHL, established above. Finally, PARP1 has been associated with the non-canonical functions of another ARS, WARS, with known angiostatic properties. PARP1 and

WARS have been shown to form a nuclear protein complex with DNA-dependent protein

231

Figure 43. TARS interacts with poly(ADP-ribose) polymerase 1 (PARP1). (A,B) Plasmids encoding HA-tagged TARS with BirA biotinylation site and BirA were transfected into HEK cells. After 48 hours, TARS was “pulled-down” with streptavidin- coated beads. (A) Western blots of TARS affinity-purified lysates analyzed using antibodies against PARP1 (IB: PARP1) and β-tubulin (IB: Tubulin) at various concentrations of BC194 (1-1000 nM). (B) Quantification of PARP1 relative to β-tubulin by densitometric analysis; n=1. Without repetitions, no meaningful analysis of statistics can be performed. kinase following stimulation by interferon γ (IFNγ) (38). Not only does this observation reveal a role of PARP1 in ARS secondary function but IFNγ itself is also known to regulate the expression of WARS and other angiostatic signals, implicating a role of this pathway in the regulation of angiogenesis as well (39). Together, these findings emphasize PARP1 as a strong candidate to mediate the angiogenic functions of TARS.

232 The TARS-PARP1 interaction identified by mass-spectrometry was subsequently confirmed by Western blot of affinity-purified TARS lysate samples (Figure 43A).

Treatment with BC194 appeared to disrupt this interaction at concentrations of 10 nM and 100 nM, maintaining only 73% and 3% of the interaction exhibited uninhibited condition respectively. Interestingly, the disruption of this interaction was not observed at higher concentrations of BC194 since the interaction returned to 61% relative to the uninhibited conditions in lysates treated with 1000 nM. It should be noted, however, that these data have not been repeated (Figure 43B). Additionally, the input samples, representing the original PARP1 present in the lysate before TARS affinity-purification, appear to vary with BC194-treatment concentrations as well. While these signals are clearly saturated, as demonstrated by the empty space in the middle of each band, and may not be completely accurate, it could be that the apparent reduction in TARS-PARP1 interaction is due to a reduction in PARP1 levels rather than a disruption of protein- protein interactions. Therefore, repetitions are required to validate this result.

A1.2.4 Characterizing TARS’s interaction with eEF1A1

The protein eEF1A became a target of interest for several reasons: 1) a truncated version of eEF1A is referred to as prostate tumor inducer (PTI) for its relation to cancer (40); 2) eEF1A is known to interact with ubiquitin (41); 3) eEF1A1 is known to assist in the nuclear export of VHL (42); and 4) a population of eEF1A has been found to function as an extracellular membrane receptor (a component that is currently lacking from the extracellular TARS-mediated angiogenesis model) (43).

All of these functions can be connected in some way to angiogenesis and, therefore,

233

Figure 44. TARS interacts with eukaryotic elongation factor 1α1 (eEF1A1). (A) Plasmids encoding HA-tagged TARS with BirA biotinylation site and BirA were transfected into HEK cells. After 48 hours, TARS was “pulled-down” with streptavidin-coated beads. The presence of eEF1A1 was determined by western blot of TARS affinity-purified lysates using antibodies against eEF1A1 (IB: eEF1A1) and β- tubulin (IB: Tubulin) at various concentrations of BC194 (1-1000 nM). (B) GTPase activity of eEF1A1 (blue), TARS (red), or combined eEF1A1 and TARS (green) as determined by formation of GDP per pmol of enzyme per minute. Human TARS or eEF1A1 (2 μM) was incubated with [α-32P]-GTP and quenched at various time points. Reaction products were resolved via thin-layer chromatography, and the formation of GDP over time was quantified by phosphorimaging; mean ± SD, n≥3. could be putatively modified through interactions with TARS. Before investigating any of these individual effects the interaction between eEF1A1 and TARS was verified by TARS affinity purification and Western blot methods similar to VHL

234 except that a plasmid encoding for a C-terminally myc-tagged eEF1A1 constructs was used in place of VHL. The samples appeared to confirm the TARS-eEF1A1 interaction indicated by the mass spectrometry data (Figure 44A). Exposure of these lysates to BC194, however, had no effect over all at the concentrations investigated.

app As the highest concentration, 1 μM, is well above the 4.09 nM predicted Ki there should be plenty of compound available to inhibit this interaction if it were capable.

Like with the PARP1 observations, more repetitions are necessary to confirm these results. However, effects of BC194 have been investigated at least one other time at a concentration of 1 μM (data not shown) without any effect, suggesting BC194 does not inhibit the eEF1A1’s interaction with TARS.

While BC194 treatment did not seem to affect the TARS-eEF1A1 interaction, the possibility that TARS could influence the GTPase activity of the enzyme was also investigated. Canonically, eEF1A functions to transport charged-tRNA to the ribosome and regulates this transport using guanine nucleotide cycling (44). This GTPase activity has also been proposed to be involved in eEF1A non-canonical activities (40,42). In support of this mechanism, a recent report demonstrated that leucyl-tRNA synthetase

(LARS) stimulates the GTPase activity of the mTOR regulator, RagD, in response to changes in amino acid concentrations (45). Furthermore, as established in Chapter 5,

TARS itself possesses a cryptic GTPase activity of unknown function. Therefore, we investigated if TARS influences the guanine nucleotide cycling of eEF1A (Figure 44B).

Individually, both TARS and eEF1A1 were found to exhibit slow GTPase activity.

However, when the proteins were combined the rates were only additive, indicating that no enhancement to eEF1A1’s GTPase activity was observed.

235 Table 14. eEF1A1 and TARS GTPase activity Condition Activity (min-1) eEF1A1 0.1132 ± 0.02033 TARS 0.04745 ± 0.002536 Both 0.1771 ± 0.01632

A1.3 Discussion

A1.3.1 Possible contributions of the TARS-VHL interaction to the regulation of angiogenesis.

The connection between conformational changes in TARS and the inhibition of angiogenesis strongly supports the role of protein-protein interactions in TARS- mediated angiogenesis. Using affinity-purification and mass spectrometry several cellular binding partners for TARS have been identified. Based on literature searches indicating potential connections to angiogenesis three of these interactions have been chosen for further investigation, including the E3 ligase VHL, eEF1A1, and PARP1.

The interaction between TARS and VHL opens up several opportunities through which

TARS could regulate angiogenesis. The regulation of the VHL/HIF1α pathway is notably complicated and includes post-translational modification and transportation based mechanisms, some of which could possibly be influenced through interactions with

TARS (15,46,47). The presence of cellular localization signals in TARS has not been fully studied, however, some immunofluorescence and Western blot data provide some evidence of TARS peri-nuclear and nuclear localization (See Chapter 6, section 6.2.2).

As such, complex formation with TARS may provide VHL with a sub-cellular localization signal that leads to its sequestration from HIF1α and allows for angiogenesis to occur. Alternatively, TARS’s interaction with VHL may alter the formation of the 236 canonical VHL E3-ligase complex by disrupting the binding of elongin B, a hypothesis supported by the disruption of TARS-VHL interaction through the introduction of mutations along the VHL-elongin C interface shown above. The resulting loss of VHL

E3-ligase function would then, in turn, stimulate angiogenesis through HIF-mediated upregulation of VEGFA expression. Therefore, future experiments investigating the effects of TARS on HIFα stabilization and VHL E3-ligase activity would be beneficial.

The inability of BC194 to disrupt the interaction between TARS and VHL is not surprising. The interaction appears to be mediated by the N1-domain, which is located two domains away from the borrelidin binding site in the TARS catalytic domain.

Therefore, it seems unlikely that any conformational change induced by borrelidin would be propagated to the N1-domain interaction interface. Alternatively, if TARS was essential for bridging VHL with another protein essential in the regulation of angiogenesis, borrelidin might be able to elicit an anti-angiogenic response through disruption of this second TARS interaction. The loss of angiogenic potential by the

Q566W TARS, borrelidin-binding mimetic (Chapter 4, Section 4.2.6) already supports the role of borrelidin-induced conformational changes as the mechanism of the compound’s anti-angiogenic properties. However, no interaction has been identified that is disrupted by this structural change.

A1.3.2 Possible contributions of the TARS-eEF1α and TARS-PARP1 interactions.

The identification of interactions with eEF1A1 and PARP1 appeared interesting owing to a variety of different reports tying these proteins to angiogenesis. Although the experiments need to be repeated, preliminary investigations suggest that the PARP1-

237 TARS interaction might be susceptible to disruption by borrelidin treatment. However, the dose-responsive nature of this effect is complicated by prominent re-establishment of an interaction from 100 nM to 1000 nM BC194. One possible explanation for this observation is that 1000 nM is the lowest concentration observed to induce amino acid starvation with BC194 (Chapter 3, section 3.2.1). Therefore, it may be possible that amino acid starvation initiates an effect that overrides BC194’s inhibition of this interaction. Conversely, while Western blot analyses appear to confirm the TARS- eEF1A1 interaction identified in mass-spectrometry experiments, there is no clear connection between TARS-angiogenesis and complex formation with eEF1A1.

Additionally, BC194 treatment showed no effect on eEF1A1 GTPase activity as well.

However, the recent development of a negative-control database for experiments that couple affinity purification with mass spectrometry, referred to as the contaminant repository for affinity purification-mass spectrometry data (CRAPome), suggests that the promiscuity and abundance of both eEF1A1 and PARP1 leads to their frequent appearance as non-specific interactions in whole interactome studies (48). Therefore, subsequent investigations of these interactions may be valuable, but should be undertaken with caution.

A1.4 Experimental Procedures

A1.4.1 Affinity and Immunoprecipitation Procedures

Culture dishes were seeded with 2x106 human embryonic kidney cells (HEK293) and maintained in DMEM (Mediatech) supplemented with 10% fetal bovine serum (Gibco,

Carlsbad, CA), penicillin/streptomycin (Gibco), and L-glutamine (Gibco) at 37°C and 5%

238 CO2 in a humidified incubator. Cells were transfected by polyethylenimine (0.05 mg/ml per μg DNA) with plasmids (3 μg each) encoding a TARS construct with C-terminal HA tag and BirA biotinylation site an C-terminally myc-tagged VHL or eEF1A1 as required.

Additionally, a BirA plasmid was included in all experiments using the TARS construct to ensure biotinylation of the protein. Control experiments substituted an empty vector plasmid for the TARS construct. Following a 48 hour incubation, cells were lysed with

1% Triton X, 0.5% NP-40, 140 mM NaCl, 25 mM Tris-HCl pH 7.6, and 1 Complete

Mini protease inhibitor tablet (Roche) per 10 ml and the entire lysate mixed with 100 μl

(1 mg) of streptavidin-coated magnetic beads (Invitrogen, Dynabeads MyOne

Streptavidin) or 60-70 μl of EZviewTM Red Anti-c-Myc Affinity Gel (Sigma-Aldrich) to pull down TARS or VHL and eEF1A1 respectively. As a non-specific control for anti-c-

Myc experiments, separate lysates were treated with 60-70 μl of rabbit IgG-coated agarose beads (Sigma-Aldrich). Unbound proteins were washed away with three exchanges of lysis buffer. The bound proteins were eluted from the beads through boiling and resolved on a reducing, SDS-PAGE gel for use in Western blotting or mass spectrometry experiments.

A1.4.2 Mass spectrometry analysis of affinity purified TARS from cell lysate

Major bands from SDS-resolved TARS affinity purified samples and the empty vector controls were detected using SilverSNAP Stain Kit II (Pierce) and excised. Fragments were trypically digested using the in gel procedure for ProteaseMAX Surfactant

(Promega) according to the manufacturers specifications. Briefly, free cysteines were alkylated by incubation with 55 mM iodoacetamide:50 mM NH4HCO3 followed by

239 trypsin digestion (2 ng/µl) in 0.01% ProteaseMAX surfactant:50 mM NH4HCO3.

Peptides were analyzed by electospray ionization (ESI) liquid chromatography mass spectrometry (LC-MS). Samples were resolved over a fused-silica microcapillary

MagicC18 LC column (12cm x100 µm i.d.) using a 5-50% acetonitrile gradient in 0.1% formic acid. Spectra were obtained using collision-induced dissociation with an LTQ linear quadrupole ion trap-Orbitrap mass spectrometer (Thermo Electro, San Jose, CA) and analyzed using SEQUEST (Bioworks software package, version 3.3.1; Thermo

Electron, San Jose, CA). Acquired TARS data were compared to empty vector equivalents in order to identify non-specific interactions.

240 A1.5 References

1. Fang, P., Yu, X., Jae Jeong, S., Mirando, A. C., Chen, K., Chen, X., Sunghoon, K., Francklyn, C. S., and Guo, M. Structural basis for full-spectrum inhibition of translational functions on a tRNA synthetase. Nature communications In press, 1- 11

2. Hirose, S., Shimizu, K., Kanai, S., Kuroda, Y., and Noguchi, T. (2007) POODLE- L: a two-level SVM prediction system for reliably predicting long disordered regions. Bioinformatics 23, 2046-2053

3. Marsh, J. A., and Teichmann, S. A. (2011) Relative solvent accessible surface area predicts protein conformational changes upon binding. Structure 19, 859-867

4. Shoemaker, B. A., Portman, J. J., and Wolynes, P. G. (2000) Speeding molecular recognition by using the folding funnel: the fly-casting mechanism. Proc Natl Acad Sci U S A 97, 8868-8873

5. Demchenko, A. P. (2001) Recognition between flexible protein molecules: induced and assisted folding. Journal of molecular recognition : JMR 14, 42-61

6. Dyson, H. J., and Wright, P. E. (2002) Coupling of folding and binding for unstructured proteins. Current opinion in structural biology 12, 54-60

7. Dyson, H. J., and Wright, P. E. (2005) Intrinsically unstructured proteins and their functions. Nat Rev Mol Cell Biol 6, 197-208

8. Iakoucheva, L. M., Brown, C. J., Lawson, J. D., Obradovic, Z., and Dunker, A. K. (2002) Intrinsic disorder in cell-signaling and cancer-associated proteins. J Mol Biol 323, 573-584

9. Kiel, C., and Serrano, L. (2006) The ubiquitin domain superfold: structure-based sequence alignments and characterization of binding epitopes. J Mol Biol 355, 821-844

10. Grabbe, C., and Dikic, I. (2009) Functional roles of ubiquitin-like domain (ULD) and ubiquitin-binding domain (UBD) containing proteins. Chem Rev 109, 1481- 1494

11. Hicke, L., Schubert, H. L., and Hill, C. P. (2005) Ubiquitin-binding domains. Nat Rev Mol Cell Biol 6, 610-621

12. Welchman, R. L., Gordon, C., and Mayer, R. J. (2005) Ubiquitin and ubiquitin- like proteins as multifunctional signals. Nat Rev Mol Cell Biol 6, 599-609

241 13. Cummins, E. P., Berra, E., Comerford, K. M., Ginouves, A., Fitzgerald, K. T., Seeballuck, F., Godson, C., Nielsen, J. E., Moynagh, P., Pouyssegur, J., and Taylor, C. T. (2006) Prolyl hydroxylase-1 negatively regulates IkappaB kinase- beta, giving insight into hypoxia-induced NFkappaB activity. Proc Natl Acad Sci U S A 103, 18154-18159

14. Koukourakis, M. I., Giatromanolaki, A., Simopoulos, C., Polychronidis, A., and Sivridis, E. (2005) Lactate dehydrogenase 5 (LDH5) relates to up-regulated hypoxia inducible factor pathway and metastasis in colorectal cancer. Clinical & experimental metastasis 22, 25-30

15. Maxwell, P. H., Wiesener, M. S., Chang, G. W., Clifford, S. C., Vaux, E. C., Cockman, M. E., Wykoff, C. C., Pugh, C. W., Maher, E. R., and Ratcliffe, P. J. (1999) The tumour suppressor protein VHL targets hypoxia-inducible factors for oxygen-dependent proteolysis. Nature 399, 271-275

16. Semenza, G. L., Roth, P. H., Fang, H. M., and Wang, G. L. (1994) Transcriptional regulation of genes encoding glycolytic enzymes by hypoxia-inducible factor 1. J Biol Chem 269, 23757-23763

17. Levine, A. J., and Puzio-Kuter, A. M. (2010) The control of the metabolic switch in cancers by oncogenes and tumor suppressor genes. Science 330, 1340-1344

18. Semenza, G. L. (2007) HIF-1 mediates the Warburg effect in clear cell renal carcinoma. Journal of bioenergetics and biomembranes 39, 231-234

19. Warburg, O., Wind, F., and Negelein, E. (1927) The Metabolism of Tumors in the Body. J Gen Physiol 8, 519-530

20. Pugh, C. W., and Ratcliffe, P. J. (2003) The von Hippel-Lindau tumor suppressor, hypoxia-inducible factor-1 (HIF-1) degradation, and cancer pathogenesis. Semin Cancer Biol 13, 83-89

21. Ewing, R. M., Chu, P., Elisma, F., Li, H., Taylor, P., Climie, S., McBroom- Cerajewski, L., Robinson, M. D., O'Connor, L., Li, M., Taylor, R., Dharsee, M., Ho, Y., Heilbut, A., Moore, L., Zhang, S., Ornatsky, O., Bukhman, Y. V., Ethier, M., Sheng, Y., Vasilescu, J., Abu-Farha, M., Lambert, J. P., Duewel, H. S., Stewart, II, Kuehl, B., Hogue, K., Colwill, K., Gladwish, K., Muskat, B., Kinach, R., Adams, S. L., Moran, M. F., Morin, G. B., Topaloglou, T., and Figeys, D. (2007) Large-scale mapping of human protein-protein interactions by mass spectrometry. Mol Syst Biol 3, 89

22. Stebbins, C. E., Kaelin, W. G., Jr., and Pavletich, N. P. (1999) Structure of the VHL-ElonginC-ElonginB complex: implications for VHL tumor suppressor function. Science 284, 455-461

242 23. Ohh, M., Takagi, Y., Aso, T., Stebbins, C. E., Pavletich, N. P., Zbar, B., Conaway, R. C., Conaway, J. W., and Kaelin, W. G., Jr. (1999) Synthetic peptides define critical contacts between elongin C, elongin B, and the von Hippel-Lindau protein. J Clin Invest 104, 1583-1591

24. Van Dang, C., and Yang, D. C. (1979) Disassembly and gross structure of particulate aminoacyl-tRNA synthetases from rat liver. Isolation and the structural relationship of synthetase complexes. J Biol Chem 254, 5350-5356

25. Han, J. M., Jeong, S. J., Park, M. C., Kim, G., Kwon, N. H., Kim, H. K., Ha, S. H., Ryu, S. H., and Kim, S. (2012) Leucyl-tRNA synthetase is an intracellular leucine sensor for the mTORC1-signaling pathway. Cell 149, 410-424

26. Sampath, P., Mazumder, B., Seshadri, V., Gerber, C. A., Chavatte, L., Kinter, M., Ting, S. M., Dignam, J. D., Kim, S., Driscoll, D. M., and Fox, P. L. (2004) Noncanonical function of glutamyl-prolyl-tRNA synthetase: gene-specific silencing of translation. Cell 119, 195-208

27. Ogata, N., Ueda, K., Kawaichi, M., and Hayaishi, O. (1981) Poly(ADP-ribose) synthetase, a main acceptor of poly(ADP-ribose) in isolated nuclei. J Biol Chem 256, 4135-4137

28. Pion, E., Ullmann, G. M., Ame, J. C., Gerard, D., de Murcia, G., and Bombarda, E. (2005) DNA-induced dimerization of poly(ADP-ribose) polymerase-1 triggers its activation. Biochemistry 44, 14670-14681

29. Benjamin, R. C., and Gill, D. M. (1980) ADP-ribosylation in mammalian cell ghosts. Dependence of poly(ADP-ribose) synthesis on strand breakage in DNA. J Biol Chem 255, 10493-10501

30. Durkacz, B. W., Omidiji, O., Gray, D. A., and Shall, S. (1980) (ADP-ribose)n participates in DNA excision repair. Nature 283, 593-596

31. Alano, C. C., Garnier, P., Ying, W., Higashi, Y., Kauppinen, T. M., and Swanson, R. A. (2010) NAD+ depletion is necessary and sufficient for poly(ADP-ribose) polymerase-1-mediated neuronal death. The Journal of neuroscience : the official journal of the Society for Neuroscience 30, 2967-2978

32. Ha, H. C., and Snyder, S. H. (1999) Poly(ADP-ribose) polymerase is a mediator of necrotic cell death by ATP depletion. Proc Natl Acad Sci U S A 96, 13978- 13982

33. Hassa, P. O., Covic, M., Hasan, S., Imhof, R., and Hottiger, M. O. (2001) The enzymatic and DNA binding activity of PARP-1 are not required for NF-kappa B coactivator function. J Biol Chem 276, 45588-45597

243 34. Kauppinen, T. M., Gan, L., and Swanson, R. A. (2013) Poly(ADP-ribose) polymerase-1-induced NAD(+) depletion promotes nuclear factor-kappaB transcriptional activity by preventing p65 de-acetylation. Biochim Biophys Acta 1833, 1985-1991

35. Rajesh, M., Mukhopadhyay, P., Godlewski, G., Batkai, S., Hasko, G., Liaudet, L., and Pacher, P. (2006) Poly(ADP-ribose)polymerase inhibition decreases angiogenesis. Biochem Biophys Res Commun 350, 1056-1062

36. Tentori, L., Lacal, P. M., Muzi, A., Dorio, A. S., Leonetti, C., Scarsella, M., Ruffini, F., Xu, W., Min, W., Stoppacciaro, A., Colarossi, C., Wang, Z. Q., Zhang, J., and Graziani, G. (2007) Poly(ADP-ribose) polymerase (PARP) inhibition or PARP-1 gene deletion reduces angiogenesis. European journal of cancer 43, 2124-2133

37. Elser, M., Borsig, L., Hassa, P. O., Erener, S., Messner, S., Valovka, T., Keller, S., Gassmann, M., and Hottiger, M. O. (2008) Poly(ADP-ribose) polymerase 1 promotes tumor cell survival by coactivating hypoxia-inducible factor-1- dependent gene expression. Mol Cancer Res 6, 282-290

38. Sajish, M., Zhou, Q., Kishi, S., Valdez, D. M., Jr., Kapoor, M., Guo, M., Lee, S., Kim, S., Yang, X. L., and Schimmel, P. (2012) Trp-tRNA synthetase bridges DNA-PKcs to PARP-1 to link IFN-gamma and p53 signaling. Nat Chem Biol 8, 547-554

39. Wakasugi, K., Slike, B. M., Hood, J., Otani, A., Ewalt, K. L., Friedlander, M., Cheresh, D. A., and Schimmel, P. (2002) A human aminoacyl-tRNA synthetase as a regulator of angiogenesis. Proc Natl Acad Sci U S A 99, 173-177

40. Leclercq, T. M., Moretti, P. A., and Pitson, S. M. Guanine nucleotides regulate sphingosine kinase 1 activation by eukaryotic elongation factor 1A and provide a mechanism for eEF1A-associated oncogenesis. Oncogene 30, 372-378

41. Chuang, S. M., Chen, L., Lambertson, D., Anand, M., Kinzy, T. G., and Madura, K. (2005) Proteasome-mediated degradation of cotranslationally damaged proteins involves translation elongation factor 1A. Mol Cell Biol 25, 403-413

42. Khacho, M., Mekhail, K., Pilon-Larose, K., Pause, A., Cote, J., and Lee, S. (2008) eEF1A is a novel component of the mammalian nuclear protein export machinery. Mol Biol Cell 19, 5296-5308

43. Itagaki, K., Naito, T., Iwakiri, R., Haga, M., Miura, S., Saito, Y., Owaki, T., Kamiya, S., Iyoda, T., Yajima, H., Iwashita, S., Ejiri, S., and Fukai, F. Eukaryotic translation elongation factor 1A induces anoikis by triggering cell detachment. J Biol Chem 287, 16037-16046

244 44. Moon, H. M., Redfield, B., and Weissbach, H. (1972) Interaction of eukaryote elongation factor EF 1 with guanosine nucleotides and aminoacyl-tRNA. Proc Natl Acad Sci U S A 69, 1249-1252

45. Han, J. M., Jeong, S. J., Park, M. C., Kim, G., Kwon, N. H., Kim, H. K., Ha, S. H., Ryu, S. H., and Kim, S. (2012) Leucyl-tRNA synthetase is an intracellular leucine sensor for the mTORC1-signaling pathway. Cell 149, 410-424

46. Kwon, S. J., Song, J. J., and Lee, Y. J. (2005) Signal pathway of hypoxia- inducible factor-1alpha phosphorylation and its interaction with von Hippel- Lindau tumor suppressor protein during ischemia in MiaPaCa-2 pancreatic cancer cells. Clinical cancer research : an official journal of the American Association for Cancer Research 11, 7607-7613

47. Mekhail, K., Gunaratnam, L., Bonicalzi, M. E., and Lee, S. (2004) HIF activation by pH-dependent nucleolar sequestration of VHL. Nat Cell Biol 6, 642-647

48. Mellacheruvu, D., Wright, Z., Couzens, A. L., Lambert, J. P., St-Denis, N. A., Li, T., Miteva, Y. V., Hauri, S., Sardiu, M. E., Low, T. Y., Halim, V. A., Bagshaw, R. D., Hubner, N. C., Al-Hakim, A., Bouchard, A., Faubert, D., Fermin, D., Dunham, W. H., Goudreault, M., Lin, Z. Y., Badillo, B. G., Pawson, T., Durocher, D., Coulombe, B., Aebersold, R., Superti-Furga, G., Colinge, J., Heck, A. J., Choi, H., Gstaiger, M., Mohammed, S., Cristea, I. M., Bennett, K. L., Washburn, M. P., Raught, B., Ewing, R. M., Gingras, A. C., and Nesvizhskii, A. I. (2013) The CRAPome: a contaminant repository for affinity purification-mass spectrometry data. Nat Methods 10, 730-736

245 COMPREHENSIVE BIBLIOGRAPHY

Abramsson, A., Lindblom, P., & Betsholtz, C. (2003). Endothelial and nonendothelial sources of PDGF-B regulate pericyte recruitment and influence vascular pattern formation in tumors. J Clin Invest, 112(8), 1142-1151.

Adams, P. D., Afonine, P. V., Bunkoczi, G., Chen, V. B., Davis, I. W., Echols, N., . . . Zwart, P. H. (2010). PHENIX: a comprehensive Python-based system for macromolecular structure solution. Acta Crystallogr D Biol Crystallogr, 66(Pt 2), 213-221.

Adams, R. H., & Alitalo, K. (2007). Molecular regulation of angiogenesis and lymphangiogenesis. Nat Rev Mol Cell Biol, 8(6), 464-478.

Aggarwal, B. B. (2003). Signalling pathways of the TNF superfamily: a double-edged sword. Nat Rev Immunol, 3(9), 745-756.

Ahmed, S. A., Gogal, R. M., Jr., & Walsh, J. E. (1994). A new rapid and simple non- radioactive assay to monitor and determine the proliferation of lymphocytes: an alternative to [3H]thymidine incorporation assay. J Immunol Methods, 170(2), 211-224.

Alano, C. C., Garnier, P., Ying, W., Higashi, Y., Kauppinen, T. M., & Swanson, R. A. (2010). NAD+ depletion is necessary and sufficient for poly(ADP-ribose) polymerase-1-mediated neuronal death. J Neurosci, 30(8), 2967-2978.

Antonellis, A., Ellsworth, R. E., Sambuughin, N., Puls, I., Abel, A., Lee-Lin, S. Q., . . . Green, E. D. (2003). Glycyl tRNA Synthetase Mutations in Charcot-Marie-Tooth Disease Type 2D and Distal Spinal Muscular Atrophy Type V. Am J Hum Genet, 72(5), 1293-1299.

Arias-Salgado, E. G., Lizano, S., Sarkar, S., Brugge, J. S., Ginsberg, M. H., & Shattil, S. J. (2003). Src kinase activation by direct interaction with the integrin beta cytoplasmic domain. Proc Natl Acad Sci U S A, 100(23), 13298-13302.

Arnaoutova, I., & Kleinman, H. K. (2010). In vitro angiogenesis: endothelial cell tube formation on gelled basement membrane extract. Nat Protoc, 5(4), 628-635.

B'Chir, W., Chaveroux, C., Carraro, V., Averous, J., Maurin, A. C., Jousse, C., . . . Bruhat, A. (2014). Dual role for CHOP in the crosstalk between autophagy and apoptosis to determine cell fate in response to amino acid deprivation. Cell Signal, 26(7), 1385-1391.

B'Chir, W., Maurin, A. C., Carraro, V., Averous, J., Jousse, C., Muranishi, Y., . . . Bruhat, A. (2013). The eIF2alpha/ATF4 pathway is essential for stress-induced autophagy gene expression. Nucleic Acids Res, 41(16), 7683-7699. 246 Babbitt, S. E., Altenhofen, L., Cobbold, S. A., Istvan, E. S., Fennell, C., Doerig, C., . . . Goldberg, D. E. (2012). Plasmodium falciparum responds to amino acid starvation by entering into a hibernatory state. Proc Natl Acad Sci U S A, 109(47), E3278-3287.

Bambace, N. M., Levis, J. E., & Holmes, C. E. (2010). The effect of P2Y-mediated platelet activation on the release of VEGF and endostatin from platelets. Platelets, 21(2), 85-93.

Bashkin, P., Doctrow, S., Klagsbrun, M., Svahn, C. M., Folkman, J., & Vlodavsky, I. (1989). Basic fibroblast growth factor binds to subendothelial extracellular matrix and is released by heparitinase and heparin-like molecules. Biochemistry, 28(4), 1737-1743.

Belrhali, H., Yaremchuk, A., Tukalo, M., Berthet-Colominas, C., Rasmussen, B., Bosecke, P., . . . Cusack, S. (1995). The structural basis for seryl-adenylate and Ap4A synthesis by seryl-tRNA synthetase. Structure, 3(4), 341-352.

Benjamin, R. C., & Gill, D. M. (1980). ADP-ribosylation in mammalian cell ghosts. Dependence of poly(ADP-ribose) synthesis on strand breakage in DNA. J Biol Chem, 255(21), 10493-10501.

Beresten, S. F., Zargarova, T. A., Favorova, O. O., Rubikaite, B. I., Ryazanov, A. G., & Kisselev, L. L. (1989). Molecular and cellular studies of tryptophanyl-tRNA synthetase using monoclonal antibodies. Evaluation of a common antigenic determinant in eukaryotic, prokaryotic and archaebacterial enzymes which maps outside the catalytic domain. Eur J Biochem, 184(3), 575-581.

Berger, J., Jampolsky, L. M., & Goldberg, M. W. (1949). Borrelidin, a new antibiotic with antiborrelia activity and penicillin enhancement properties. Arch Biochem, 22(3), 476-478.

Bergers, G., & Benjamin, L. E. (2003). Tumorigenesis and the angiogenic switch. Nat Rev Cancer, 3(6), 401-410.

Bergers, G., & Hanahan, D. (2008). Modes of resistance to anti-angiogenic therapy. Nat Rev Cancer, 8(8), 592-603.

Bergstrom, J., Furst, P., Noree, L. O., & Vinnars, E. (1974). Intracellular free amino acid concentration in human muscle tissue. J Appl Physiol, 36(6), 693-697.

Bobbert, P., Schluter, H., Schultheiss, H. P., & Reusch, H. P. (2008). Diadenosine polyphosphates Ap3A and Ap4A, but not Ap5A or Ap6A, induce proliferation of vascular smooth muscle cells. Biochem Pharmacol, 75(10), 1966-1973.

Bourne, H. R., Sanders, D. A., & McCormick, F. (1991). The GTPase superfamily: conserved structure and molecular mechanism. Nature, 349(6305), 117-127. 247 Bovee, M. L., Pierce, M. A., & Francklyn, C. S. (2003). Induced fit and kinetic mechanism of adenylation catalyzed by Escherichia coli threonyl-tRNA synthetase. Biochemistry, 42(51), 15102-15113.

Bromann, P. A., Korkaya, H., & Courtneidge, S. A. (2004). The interplay between Src family kinases and receptor tyrosine kinases. Oncogene, 23(48), 7957-7968.

Brooks, R. F. (1977). Continuous protein synthesis is required to maintain the probability of entry into S phase. Cell, 12(1), 311-317.

Brown, C. J., Takayama, S., Campen, A. M., Vise, P., Marshall, T. W., Oldfield, C. J., . . . Dunker, A. K. (2002). Evolutionary rate heterogeneity in proteins with long disordered regions. J Mol Evol, 55(1), 104-110.

Burnstock, G. (2002). Purinergic signaling and vascular cell proliferation and death. Arterioscler Thromb Vasc Biol, 22(3), 364-373.

Buschow, S. I., van Balkom, B. W., Aalberts, M., Heck, A. J., Wauben, M., & Stoorvogel, W. (2010). MHC class II-associated proteins in B-cell exosomes and potential functional implications for exosome biogenesis. Immunol Cell Biol, 88(8), 851-856.

Cai, Q., & Robertson, E. S. Ubiquitin/SUMO modification regulates VHL protein stability and nucleocytoplasmic localization. PLoS One, 5(9).

Cao, X., Tay, A., Guy, G. R., & Tan, Y. H. (1996). Activation and association of Stat3 with Src in v-Src-transformed cell lines. Mol Cell Biol, 16(4), 1595-1603.

Carmeliet, P., & Jain, R. K. (2000). Angiogenesis in cancer and other diseases. Nature, 407(6801), 249-257.

Carmeliet, P., & Jain, R. K. (2011). Molecular mechanisms and clinical applications of angiogenesis. Nature, 473(7347), 298-307.

Chandran, U. R., Ma, C., Dhir, R., Bisceglia, M., Lyons-Weiler, M., Liang, W., . . . Monzon, F. A. (2007). Gene expression profiles of prostate cancer reveal involvement of multiple molecular pathways in the metastatic process. BMC Cancer, 7, 64.

Charo, I. F., Nannizzi, L., Smith, J. W., & Cheresh, D. A. (1990). The vitronectin receptor alpha v beta 3 binds fibronectin and acts in concert with alpha 5 beta 1 in promoting cellular attachment and spreading on fibronectin. J Cell Biol, 111(6 Pt 1), 2795-2800.

Chen, T. T., Luque, A., Lee, S., Anderson, S. M., Segura, T., & Iruela-Arispe, M. L. (2010). Anchorage of VEGF to the extracellular matrix conveys differential signaling responses to endothelial cells. J Cell Biol, 188(4), 595-609.

248 Chen, Z., & Han, Z. C. (2008). STAT3: a critical transcription activator in angiogenesis. Med Res Rev, 28(2), 185-200.

Chuang, S. M., Chen, L., Lambertson, D., Anand, M., Kinzy, T. G., & Madura, K. (2005). Proteasome-mediated degradation of cotranslationally damaged proteins involves translation elongation factor 1A. Mol Cell Biol, 25(1), 403-413.

Clark, A. J. (1937). General pharmacology, handbook of experimental pharmacology. Berlin: Springer Verlag.

Copeland, Robert Allen. (2005). Evaluation of enzyme inhibitors in drug discovery : a guide for medicinal chemists and pharmacologists. Hoboken, N.J.: Wiley- Interscience.

Copeland, Robert Allen. (2013). Evaluation of enzyme inhibitors in drug discovery : a guide for medicinal chemists and pharmacologists (2nd ed.). Hoboken, N.J.: Wiley.

Corriden, R., & Insel, P. A. (2012). New insights regarding the regulation of chemotaxis by nucleotides, adenosine, and their receptors. Purinergic Signal, 8(3), 587-598.

Cummins, E. P., Berra, E., Comerford, K. M., Ginouves, A., Fitzgerald, K. T., Seeballuck, F., . . . Taylor, C. T. (2006). Prolyl hydroxylase-1 negatively regulates IkappaB kinase-beta, giving insight into hypoxia-induced NFkappaB activity. Proc Natl Acad Sci U S A, 103(48), 18154-18159.

Dang, C. V., & Traugh, J. A. (1989). Phosphorylation of threonyl- and seryl-tRNA synthetase by cAMP-dependent protein kinase. A possible role in the regulation of P1, P4-bis(5'-adenosyl)-tetraphosphate (Ap4A) synthesis. J Biol Chem, 264(10), 5861-5865.

Demchenko, A. P. (2001). Recognition between flexible protein molecules: induced and assisted folding. J Mol Recognit, 14(1), 42-61.

Dey, S., Savant, S., Teske, B. F., Hatzoglou, M., Calkhoven, C. F., & Wek, R. C. (2012). Transcriptional repression of ATF4 gene by CCAAT/enhancer-binding protein beta (C/EBPbeta) differentially regulates integrated stress response. J Biol Chem, 287(26), 21936-21949.

Dickinson, L., Griffiths, A. J., Mason, C. G., & Mills, R. F. (1965). Anti-viral activity of two antibiotics isolated from a species of Streptomyces. Nature, 206(981), 265- 268.

Dock-Bregeon, A., Sankaranarayanan, R., Romby, P., Caillet, J., Springer, M., Rees, B., . . . Moras, D. (2000). Transfer RNA-mediated editing in threonyl-tRNA synthetase. The class II solution to the double discrimination problem. Cell, 103(6), 877-884. 249 Durkacz, B. W., Omidiji, O., Gray, D. A., & Shall, S. (1980). (ADP-ribose)n participates in DNA excision repair. Nature, 283(5747), 593-596.

Dyson, H. J., & Wright, P. E. (2002). Coupling of folding and binding for unstructured proteins. Curr Opin Struct Biol, 12(1), 54-60.

Dyson, H. J., & Wright, P. E. (2005). Intrinsically unstructured proteins and their functions. Nat Rev Mol Cell Biol, 6(3), 197-208.

Easson, L.H., & Stedman, E. (1936). The Absolute Activity of Choline-Esterase. Proc R Soc London Series B, 121, 142-164.

Eastwood, E. L., & Schaus, S. E. (2003). Borrelidin induces the transcription of amino acid biosynthetic enzymes via a GCN4-dependent pathway. Bioorganic & Medicinal Chemistry Letters, 13(13), 2235-2237.

Elser, M., Borsig, L., Hassa, P. O., Erener, S., Messner, S., Valovka, T., . . . Hottiger, M. O. (2008). Poly(ADP-ribose) polymerase 1 promotes tumor cell survival by coactivating hypoxia-inducible factor-1-dependent gene expression. Mol Cancer Res, 6(2), 282-290.

Emsley, P., Lohkamp, B., Scott, W. G., & Cowtan, K. (2010). Features and development of Coot. Acta Crystallogr D Biol Crystallogr, 66(Pt 4), 486-501.

Esteban, M. A., Tran, M. G., Harten, S. K., Hill, P., Castellanos, M. C., Chandra, A., . . . Maxwell, P. H. (2006). Regulation of E-cadherin expression by VHL and hypoxia-inducible factor. Cancer Res, 66(7), 3567-3575.

Ewing, R. M., Chu, P., Elisma, F., Li, H., Taylor, P., Climie, S., . . . Figeys, D. (2007). Large-scale mapping of human protein-protein interactions by mass spectrometry. Mol Syst Biol, 3, 89.

Fang, P., Yu, X., Jae Jeong, S., Mirando, A. C., Chen, K., Chen, X., . . . Guo, M. Structural basis for full-spectrum inhibition of translational functions on a tRNA synthetase. Nat Commun, In press, 1-11.

Fernandez, I., Harlow, L., Zang, Y., Liu-Bryan, R., Ridgway, W. M., Clemens, P. R., & Ascherman, D. P. (2013). Functional redundancy of MyD88-dependent signaling pathways in a murine model of histidyl-transfer RNA synthetase-induced myositis. J Immunol, 191(4), 1865-1872.

Francklyn, C. S., First, E. A., Perona, J. J., & Hou, Y. M. (2008). Methods for kinetic and thermodynamic analysis of aminoacyl-tRNA synthetases. Methods, 44(2), 100- 118.

Fukui, H., Hanaoka, R., & Kawahara, A. (2009). Noncanonical activity of seryl-tRNA synthetase is involved in vascular development. Circ Res, 104(11), 1253-1259.

250 Funahashi, Y., Wakabayashi, T., Semba, T., Sonoda, J., Kitoh, K., & Yoshimatsu, K. (1999). Establishment of a quantitative mouse dorsal air sac model and its application to evaluate a new angiogenesis inhibitor. Oncol Res, 11(7), 319-329.

Gafiuc, D., Weiss, M., Mylonas, I., & Bruning, A. (2013). Borrelidin has limited anti- cancer effects in bcl-2 overexpressing breast cancer and leukemia cells and reveals toxicity in non-malignant breast epithelial cells. J Appl Toxicol.

Gantt, J. S., Bennett, C. A., & Arfin, S. M. (1981). Increased levels of threonyl-tRNA synthetase in a borrelidin-resistant Chinese hamster ovary cell line. Proc Natl Acad Sci U S A, 78(9), 5367-5370.

Gao, Y. M., Wang, X. J., Zhang, J., Li, M., Liu, C. X., An, J., . . . Xiang, W. S. (2012). Borrelidin, a potent antifungal agent: insight into the antifungal mechanism against Phytophthora sojae. J Agric Food Chem, 60(39), 9874-9881.

Garcia-Roman, J., & Zentella-Dehesa, A. (2013). Vascular permeability changes involved in tumor metastasis. Cancer Lett, 335(2), 259-269.

Ghosh, R., Lipson, K. L., Sargent, K. E., Mercurio, A. M., Hunt, J. S., Ron, D., & Urano, F. (2010). Transcriptional regulation of VEGF-A by the unfolded protein response pathway. PLoS One, 5(3), e9575.

Gianni, T., Leoni, V., Chesnokova, L. S., Hutt-Fletcher, L. M., & Campadelli-Fiume, G. (2012). alphavbeta3-integrin is a major sensor and activator of innate immunity to herpes simplex virus-1. Proc Natl Acad Sci U S A, 109(48), 19792-19797.

Goerlich, O., Foeckler, R., & Holler, E. (1982). Mechanism of synthesis of adenosine(5')tetraphospho(5')adenosine (AppppA) by aminoacyl-tRNA synthetases. Eur J Biochem, 126(1), 135-142.

Goldberg, D. E., Slater, A. F., Cerami, A., & Henderson, G. B. (1990). Hemoglobin degradation in the malaria parasite Plasmodium falciparum: an ordered process in a unique organelle. Proc Natl Acad Sci U S A, 87(8), 2931-2935.

Grabbe, C., & Dikic, I. (2009). Functional roles of ubiquitin-like domain (ULD) and ubiquitin-binding domain (UBD) containing proteins. Chem Rev, 109(4), 1481- 1494.

Greenberg, J. I., Shields, D. J., Barillas, S. G., Acevedo, L. M., Murphy, E., Huang, J., . . . Cheresh, D. A. (2008). A role for VEGF as a negative regulator of pericyte function and vessel maturation. Nature, 456(7223), 809-813.

Greenberg, Y., King, M., Kiosses, W. B., Ewalt, K., Yang, X., Schimmel, P., . . . Tzima, E. (2008). The novel fragment of tyrosyl tRNA synthetase, mini-TyrRS, is secreted to induce an angiogenic response in endothelial cells. FASEB J, 22(5), 1597-1605. 251 Guo, M., & Schimmel, P. (2013). Essential nontranslational functions of tRNA synthetases. Nat Chem Biol, 9(3), 145-153.

Guo, R. T., Chong, Y. E., Guo, M., & Yang, X. L. (2009). Crystal structures and biochemical analyses suggest a unique mechanism and role for human glycyl- tRNA synthetase in Ap4A homeostasis. J Biol Chem, 284(42), 28968-28976.

Guth, E. C., & Francklyn, C. S. (2007). Kinetic discrimination of tRNA identity by the conserved motif 2 loop of a class II aminoacyl-tRNA synthetase. Mol Cell, 25(4), 531-542.

Guth, E., Farris, M., Bovee, M., & Francklyn, C. S. (2009). Asymmetric amino acid activation by class II histidyl-tRNA synthetase from Escherichia coli. J Biol Chem, 284(31), 20753-20762.

Ha, H. C., & Snyder, S. H. (1999). Poly(ADP-ribose) polymerase is a mediator of necrotic cell death by ATP depletion. Proc Natl Acad Sci U S A, 96(24), 13978- 13982.

Habibi, D., Ogloff, N., Jalili, R. B., Yost, A., Weng, A. P., Ghahary, A., & Ong, C. J. (2011). Borrelidin, a small molecule nitrile-containing macrolide inhibitor of threonyl-tRNA synthetase, is a potent inducer of apoptosis in acute lymphoblastic leukemia. Invest New Drugs, 30(4), 1361-1370.

Hamanaka, R. B., Bennett, B. S., Cullinan, S. B., & Diehl, J. A. (2005). PERK and GCN2 contribute to eIF2alpha phosphorylation and cell cycle arrest after activation of the unfolded protein response pathway. Mol Biol Cell, 16(12), 5493-5501.

Han, J. M., Jeong, S. J., Park, M. C., Kim, G., Kwon, N. H., Kim, H. K., . . . Kim, S. (2012). Leucyl-tRNA synthetase is an intracellular leucine sensor for the mTORC1-signaling pathway. Cell, 149(2), 410-424.

Harisi, R., Kenessey, I., Olah, J. N., Timar, F., Babo, I., Pogany, G., . . . Jeney, A. (2009). Differential inhibition of single and cluster type tumor cell migration. Anticancer Res, 29(8), 2981-2985.

Harten, S. K., Shukla, D., Barod, R., Hergovich, A., Balda, M. S., Matter, K., . . . Maxwell, P. H. (2009). Regulation of renal epithelial tight junctions by the von Hippel-Lindau tumor suppressor gene involves occludin and claudin 1 and is independent of E-cadherin. Mol Biol Cell, 20(3), 1089-1101.

Hassa, P. O., Covic, M., Hasan, S., Imhof, R., & Hottiger, M. O. (2001). The enzymatic and DNA binding activity of PARP-1 are not required for NF-kappa B coactivator function. J Biol Chem, 276(49), 45588-45597.

252 Herzog, W., Muller, K., Huisken, J., & Stainier, D. Y. (2009). Genetic evidence for a noncanonical function of seryl-tRNA synthetase in vascular development. Circ Res, 104(11), 1260-1266.

Hicke, L., Schubert, H. L., & Hill, C. P. (2005). Ubiquitin-binding domains. Nat Rev Mol Cell Biol, 6(8), 610-621.

Hirose, S., Shimizu, K., Kanai, S., Kuroda, Y., & Noguchi, T. (2007). POODLE-L: a two-level SVM prediction system for reliably predicting long disordered regions. Bioinformatics, 23(16), 2046-2053.

Hocking, D. C., & Kowalski, K. (2002). A cryptic fragment from fibronectin's III1 module localizes to lipid rafts and stimulates cell growth and contractility. J Cell Biol, 158(1), 175-184.

Hu, X., Chen, J., Wang, L., & Ivashkiv, L. B. (2007). Crosstalk among Jak-STAT, Toll- like receptor, and ITAM-dependent pathways in macrophage activation. J Leukoc Biol, 82(2), 237-243.

Huber, M. A., Kraut, N., & Beug, H. (2005). Molecular requirements for epithelial- mesenchymal transition during tumor progression. Curr Opin Cell Biol, 17(5), 548-558.

Humphries, J. D., Byron, A., Bass, M. D., Craig, S. E., Pinney, J. W., Knight, D., & Humphries, M. J. (2009). Proteomic analysis of integrin-associated complexes identifies RCC2 as a dual regulator of Rac1 and Arf6. Sci Signal, 2(87), ra51.

Huss, W. J., Hanrahan, C. F., Barrios, R. J., Simons, J. W., & Greenberg, N. M. (2001). Angiogenesis and prostate cancer: identification of a molecular progression switch. Cancer Res, 61(6), 2736-2743.

Hutter, R., Poralla, K., Zachau, H. G., & Zahner, H. (1966). [Metabolic products of microorganisms. 5l. On the mechanism of action of borrelidin-inhibition of the threonine incorporation in sRNA]. Biochem Z, 344(2), 190-196.

Iakoucheva, L. M., Brown, C. J., Lawson, J. D., Obradovic, Z., & Dunker, A. K. (2002). Intrinsic disorder in cell-signaling and cancer-associated proteins. J Mol Biol, 323(3), 573-584.

Ibba, M., & Soll, D. (2000). Aminoacyl-tRNA synthesis. Ann. Rev. Biochem., 69, 617- 650.

Itagaki, K., Naito, T., Iwakiri, R., Haga, M., Miura, S., Saito, Y., . . . Fukai, F. Eukaryotic translation elongation factor 1A induces anoikis by triggering cell detachment. J Biol Chem, 287(19), 16037-16046.

253 Jordanova, A., Irobi, J., Thomas, F. P., Van Dijck, P., Meerschaert, K., Dewil, M., . . . Timmerman, V. (2006). Disrupted function and axonal distribution of mutant tyrosyl-tRNA synthetase in dominant intermediate Charcot-Marie-Tooth neuropathy. Nat Genet, 38(2), 197-202.

Kapoor, M., Zhou, Q., Otero, F., Myers, C. A., Bates, A., Belani, R., . . . Schimmel, P. (2008). Evidence for annexin II-S100A10 complex and plasmin in mobilization of cytokine activity of human TrpRS. J Biol Chem, 283(4), 2070-2077.

Kauppinen, T. M., Gan, L., & Swanson, R. A. (2013). Poly(ADP-ribose) polymerase-1- induced NAD(+) depletion promotes nuclear factor-kappaB transcriptional activity by preventing p65 de-acetylation. Biochim Biophys Acta, 1833(8), 1985- 1991.

Kawai, T., & Akira, S. (2006). TLR signaling. Cell Death Differ, 13(5), 816-825.

Kawamura, T., Liu, D., Towle, M. J., Kageyama, R., Tsukahara, N., Wakabayashi, T., & Littlefield, B. A. (2003). Anti-angiogenesis effects of borrelidin are mediated through distinct pathways: threonyl-tRNA synthetase and caspases are independently involved in suppression of proliferation and induction of apoptosis in endothelial cells. J Antibiot (Tokyo), 56(8), 709-715.

Khacho, M., Mekhail, K., Pilon-Larose, K., Pause, A., Cote, J., & Lee, S. (2008). eEF1A is a novel component of the mammalian nuclear protein export machinery. Mol Biol Cell, 19(12), 5296-5308.

Kiel, C., & Serrano, L. (2006). The ubiquitin domain superfold: structure-based sequence alignments and characterization of binding epitopes. J Mol Biol, 355(4), 821-844.

Kim, D. G., Lee, J. Y., Kwon, N. H., Fang, P., Zhang, Q., Wang, J., . . . Kim, S. (2013). Chemical inhibition of prometastatic lysyl-tRNA synthetase-laminin receptor interaction. Nat Chem Biol, 10(1), 29-34.

Kimmel, C. B., Ballard, W. W., Kimmel, S. R., Ullmann, B., & Schilling, T. F. (1995). Stages of embryonic development of the zebrafish. Dev Dyn, 203(3), 253-310.

Kodera, H., Osaka, H., Iai, M., Aida, N., Yamashita, A., Tsurusaki, Y., . . . Matsumoto, N. (2014). Mutations in the glutaminyl-tRNA synthetase gene cause early-onset epileptic encephalopathy. J Hum Genet.

Kosugi, S., Hasebe, M., Matsumura, N., Takashima, H., Miyamoto-Sato, E., Tomita, M., & Yanagawa, H. (2009). Six classes of nuclear localization signals specific to different binding grooves of importin alpha. J Biol Chem, 284(1), 478-485.

Koukourakis, M. I., Giatromanolaki, A., Simopoulos, C., Polychronidis, A., & Sivridis, E. (2005). Lactate dehydrogenase 5 (LDH5) relates to up-regulated hypoxia

254 inducible factor pathway and metastasis in colorectal cancer. Clin Exp Metastasis, 22(1), 25-30.

Kwon, S. J., Song, J. J., & Lee, Y. J. (2005). Signal pathway of hypoxia-inducible factor- 1alpha phosphorylation and its interaction with von Hippel-Lindau tumor suppressor protein during ischemia in MiaPaCa-2 pancreatic cancer cells. Clin Cancer Res, 11(21), 7607-7613.

Langmuir, I. (1916). The constituion and fundamental properties of solids and liquids. J Am Chem Soc, 38, 2221-2295.

Lechertier, T., & Hodivala-Dilke, K. (2012). Focal adhesion kinase and tumour angiogenesis. J Pathol, 226(2), 404-412.

Leclercq, T. M., Moretti, P. A., & Pitson, S. M. Guanine nucleotides regulate sphingosine kinase 1 activation by eukaryotic elongation factor 1A and provide a mechanism for eEF1A-associated oncogenesis. Oncogene, 30(3), 372-378.

Lee, S., Garner, E. I., Welch, W. R., Berkowitz, R. S., & Mok, S. C. (2007). Over- expression of hypoxia-inducible factor 1 alpha in ovarian clear cell carcinoma. Gynecol Oncol, 106(2), 311-317.

Lee, S., Jilani, S. M., Nikolova, G. V., Carpizo, D., & Iruela-Arispe, M. L. (2005). Processing of VEGF-A by matrix metalloproteinases regulates bioavailability and vascular patterning in tumors. J Cell Biol, 169(4), 681-691.

Lee, Y. N., Nechushtan, H., Figov, N., & Razin, E. (2004). The function of lysyl-tRNA synthetase and Ap4A as signaling regulators of MITF activity in FcepsilonRI- activated mast cells. Immunity, 20(2), 145-151.

Levine, A. J., & Puzio-Kuter, A. M. (2010). The control of the metabolic switch in cancers by oncogenes and tumor suppressor genes. Science, 330(6009), 1340- 1344.

Lewis, M. D., & Roberts, B. J. (2003). Role of nuclear and cytoplasmic localization in the tumour-suppressor activity of the von Hippel-Lindau protein. Oncogene, 22(26), 3992-3997.

Lien, S., & Lowman, H. B. (2008). Therapeutic anti-VEGF antibodies. Handb Exp Pharmacol(181), 131-150.

Lilja, H., Ulmert, D., & Vickers, A. J. (2008). Prostate-specific antigen and prostate cancer: prediction, detection and monitoring. Nat Rev Cancer, 8(4), 268-278.

Lin, M. T., Yen, M. L., Lin, C. Y., & Kuo, M. L. (2003). Inhibition of vascular endothelial growth factor-induced angiogenesis by resveratrol through

255 interruption of Src-dependent vascular endothelial cadherin tyrosine phosphorylation. Mol Pharmacol, 64(5), 1029-1036.

Lumb, M., Macey, P. E., Spyvee, J., Whitmarsh, J. M., & Wright, R. D. (1965). Isolation of vivomycin and borrelidin, two antibiotics with anti-viral activity, from a species of Streptomyces (C2989). Nature, 206(981), 263-265.

Luthje, J., & Ogilvie, A. (1988). Catabolism of Ap4A and Ap3A in whole blood. The dinucleotides are long-lived signal molecules in the blood ending up as intracellular ATP in the erythrocytes. Eur J Biochem, 173(1), 241-245.

Maehr, H., & Evans, R. H. (1987). Identity of borrelidin with treponemycin. J Antibiot (Tokyo), 40(10), 1455-1456.

Magnusson, L. U., Farewell, A., & Nystrom, T. (2005). ppGpp: a global regulator in Escherichia coli. Trends Microbiol, 13(5), 236-242.

Marsh, J. A., & Teichmann, S. A. (2011). Relative solvent accessible surface area predicts protein conformational changes upon binding. Structure, 19(6), 859-867.

Maxwell, P. H., Wiesener, M. S., Chang, G. W., Clifford, S. C., Vaux, E. C., Cockman, M. E., . . . Ratcliffe, P. J. (1999). The tumour suppressor protein VHL targets hypoxia-inducible factors for oxygen-dependent proteolysis. Nature, 399(6733), 271-275.

McAuslan, B. R., Reilly, W. G., Hannan, G. N., & Gole, G. A. (1983). Angiogenic factors and their assay: activity of formyl methionyl leucyl phenylalanine, adenosine diphosphate, heparin, copper, and bovine endothelium stimulating factor. Microvasc Res, 26(3), 323-338.

McKenzie, A. J., Campbell, S. L., & Howe, A. K. Protein kinase A activity and anchoring are required for ovarian cancer cell migration and invasion. PLoS One, 6(10), e26552.

Mekhail, K., Gunaratnam, L., Bonicalzi, M. E., & Lee, S. (2004). HIF activation by pH- dependent nucleolar sequestration of VHL. Nat Cell Biol, 6(7), 642-647.

Mellacheruvu, D., Wright, Z., Couzens, A. L., Lambert, J. P., St-Denis, N. A., Li, T., . . . Nesvizhskii, A. I. (2013). The CRAPome: a contaminant repository for affinity purification-mass spectrometry data. Nat Methods, 10(8), 730-736.

Minajigi, A., & Francklyn, C. S. (2008). RNA-assisted catalysis in a protein enzyme: The 2'-hydroxyl of tRNA(Thr) A76 promotes aminoacylation by threonyl-tRNA synthetase. Proc Natl Acad Sci U S A, 105(46), 17748-17753.

256 Miranti, C. K., Ohno, S., & Brugge, J. S. (1999). Protein kinase C regulates integrin- induced activation of the extracellular regulated kinase pathway upstream of Shc. J Biol Chem, 274(15), 10571-10581.

Moon, H. M., Redfield, B., & Weissbach, H. (1972). Interaction of eukaryote elongation factor EF 1 with guanosine nucleotides and aminoacyl-tRNA. Proc Natl Acad Sci U S A, 69(5), 1249-1252.

Morabito, A., De Maio, E., Di Maio, M., Normanno, N., & Perrone, F. (2006). Tyrosine kinase inhibitors of vascular endothelial growth factor receptors in clinical trials: current status and future directions. Oncologist, 11(7), 753-764.

Morrison, John F., & Walsh, Christopher T. (2006). The Behavior and Significance of Slow-Binding Enzyme Inhibitors Advances in Enzymology and Related Areas of Molecular Biology (pp. 201-301): John Wiley & Sons, Inc.

Moss, S. J., Carletti, I., Olano, C., Sheridan, R. M., Ward, M., Math, V., . . . Wilkinson, B. (2006). Biosynthesis of the angiogenesis inhibitor borrelidin: directed biosynthesis of novel analogues. Chem Commun (Camb)(22), 2341-2343.

Mukhopadhyay, R., Jia, J., Arif, A., Ray, P. S., & Fox, P. L. (2009). The GAIT system: a gatekeeper of inflammatory gene expression. Trends Biochem Sci, 34(7), 324- 331.

Newton, A. C. (1995). Protein kinase C: structure, function, and regulation. J Biol Chem, 270(48), 28495-28498.

Novoa, E. M., Camacho, N., Tor, A., Wilkinson, B., Moss, S., Marin-Garcia, P., . . . Ribas de Pouplana, L. (2014). Analogs of natural aminoacyl-tRNA synthetase inhibitors clear malaria in vivo. Proc Natl Acad Sci U S A, 111(51), E5508-5517.

Ofir-Birin, Y., Fang, P., Bennett, S. P., Zhang, H. M., Wang, J., Rachmin, I., . . . Guo, M. (2013). Structural switch of lysyl-tRNA synthetase between translation and transcription. Mol Cell, 49(1), 30-42.

Ogata, N., Ueda, K., Kawaichi, M., & Hayaishi, O. (1981). Poly(ADP-ribose) synthetase, a main acceptor of poly(ADP-ribose) in isolated nuclei. J Biol Chem, 256(9), 4135-4137.

Ohh, M., Takagi, Y., Aso, T., Stebbins, C. E., Pavletich, N. P., Zbar, B., . . . Kaelin, W. G., Jr. (1999). Synthetic peptides define critical contacts between elongin C, elongin B, and the von Hippel-Lindau protein. J Clin Invest, 104(11), 1583-1591.

Otoguro, K., Ui, H., Ishiyama, A., Kobayashi, M., Togashi, H., Takahashi, Y., . . . Omura, S. (2003). In vitro and in vivo antimalarial activities of a non-glycosidic 18-membered macrolide antibiotic, borrelidin, against drug-resistant strains of Plasmodia. J Antibiot (Tokyo), 56(8), 727-729. 257 Otwinowski, Z., & Minor, W. (1997). Processing of X-ray Diffraction Data Collected in Oscillation Mode. Methods in Enzymology, 276, 307-326.

Paetz, W., & Nass, G. (1973). Biochemical and immunological characterization of threonyl-tRNA synthetase of two borrelidin-resistant mutants of Escherichia coli K12. Eur J Biochem, 35(2), 331-337.

Pang, Y. L., & Martinis, S. A. (2009). A paradigm shift for the amino acid editing mechanism of human cytoplasmic leucyl-tRNA synthetase. Biochemistry, 48(38), 8958-8964.

Park, S. G., Kim, H. J., Min, Y. H., Choi, E. C., Shin, Y. K., Park, B. J., . . . Kim, S. (2005). Human lysyl-tRNA synthetase is secreted to trigger proinflammatory response. Proc Natl Acad Sci U S A, 102(18), 6356-6361.

Peinado, H., Aleckovic, M., Lavotshkin, S., Matei, I., Costa-Silva, B., Moreno-Bueno, G., . . . Lyden, D. (2012). Melanoma exosomes educate bone marrow progenitor cells toward a pro-metastatic phenotype through MET. Nat Med, 18(6), 883-891.

Pierschbacher, M. D., & Ruoslahti, E. (1984). Cell attachment activity of fibronectin can be duplicated by small synthetic fragments of the molecule. Nature, 309(5963), 30-33.

Pion, E., Ullmann, G. M., Ame, J. C., Gerard, D., de Murcia, G., & Bombarda, E. (2005). DNA-induced dimerization of poly(ADP-ribose) polymerase-1 triggers its activation. Biochemistry, 44(44), 14670-14681.

Pizzirani, C., Ferrari, D., Chiozzi, P., Adinolfi, E., Sandona, D., Savaglio, E., & Di Virgilio, F. (2007). Stimulation of P2 receptors causes release of IL-1beta-loaded microvesicles from human dendritic cells. Blood, 109(9), 3856-3864.

Porter, J. C., & Hogg, N. (1998). Integrins take partners: cross-talk between integrins and other membrane receptors. Trends Cell Biol, 8(10), 390-396.

Puffenberger, E. G., Jinks, R. N., Sougnez, C., Cibulskis, K., Willert, R. A., Achilly, N. P., . . . Strauss, K. A. (2012). Genetic mapping and exome sequencing identify variants associated with five novel diseases. PLoS One, 7(1), e28936.

Pugh, C. W., & Ratcliffe, P. J. (2003). The von Hippel-Lindau tumor suppressor, hypoxia-inducible factor-1 (HIF-1) degradation, and cancer pathogenesis. Semin Cancer Biol, 13(1), 83-89.

Rajesh, M., Mukhopadhyay, P., Godlewski, G., Batkai, S., Hasko, G., Liaudet, L., & Pacher, P. (2006). Poly(ADP-ribose)polymerase inhibition decreases angiogenesis. Biochem Biophys Res Commun, 350(4), 1056-1062.

258 Ribatti, D., Nico, B., Vacca, A., & Presta, M. (2006). The gelatin sponge-chorioallantoic membrane assay. Nat Protoc, 1(1), 85-91.

Rock, F. L., Mao, W., Yaremchuk, A., Tukalo, M., Crepin, T., Zhou, H., . . . Alley, M. R. (2007). An antifungal agent inhibits an aminoacyl-tRNA synthetase by trapping tRNA in the editing site. Science, 316(5832), 1759-1761.

Ruan, B., Bovee, M. L., Sacher, M., Stathopoulos, C., Poralla, K., Francklyn, C. S., & Soll, D. (2005). A unique hydrophobic cluster near the active site contributes to differences in borrelidin inhibition among threonyl-tRNA synthetases. J Biol Chem, 280(1), 571-577.

Ruhrberg, C., Gerhardt, H., Golding, M., Watson, R., Ioannidou, S., Fujisawa, H., . . . Shima, D. T. (2002). Spatially restricted patterning cues provided by heparin- binding VEGF-A control blood vessel branching morphogenesis. Genes Dev, 16(20), 2684-2698.

Rumjahn, S. M., Yokdang, N., Baldwin, K. A., Thai, J., & Buxton, I. L. (2009). Purinergic regulation of vascular endothelial growth factor signaling in angiogenesis. Br J Cancer, 100(9), 1465-1470.

Sajish, M., Zhou, Q., Kishi, S., Valdez, D. M., Jr., Kapoor, M., Guo, M., . . . Schimmel, P. (2012). Trp-tRNA synthetase bridges DNA-PKcs to PARP-1 to link IFN- gamma and p53 signaling. Nat Chem Biol, 8(6), 547-554.

Sampath, P., Mazumder, B., Seshadri, V., Gerber, C. A., Chavatte, L., Kinter, M., . . . Fox, P. L. (2004). Noncanonical function of glutamyl-prolyl-tRNA synthetase: gene-specific silencing of translation. Cell, 119(2), 195-208.

Sankaranarayanan, R., Dock-Bregeon, A. C., Rees, B., Bovee, M., Caillet, J., Romby, P., . . . Moras, D. (2000). Zinc ion mediated amino acid discrimination by threonyl- tRNA synthetase.[see comment]. Nature Structural Biology, 7(6), 461-465.

Scheper, G. C., van der Klok, T., van Andel, R. J., van Berkel, C. G., Sissler, M., Smet, J., . . . van der Knaap, M. S. (2007). Mitochondrial aspartyl-tRNA synthetase deficiency causes leukoencephalopathy with brain stem and spinal cord involvement and lactate elevation. Nature Genetics, 39(4), 534-539.

Schreiber, G., Ron, E. Z., & Glaser, G. (1995). ppGpp-mediated regulation of DNA replication and cell division in Escherichia coli. Curr Microbiol, 30(1), 27-32.

Schwarz, M. A., Zheng, H., Liu, J., Corbett, S., & Schwarz, R. E. (2005). Endothelial- monocyte activating polypeptide II alters fibronectin based endothelial cell adhesion and matrix assembly via alpha5 beta1 integrin. Exp Cell Res, 311(2), 229-239.

259 Seiradake, E., Mao, W., Hernandez, V., Baker, S. J., Plattner, J. J., Alley, M. R., & Cusack, S. (2009). Crystal structures of the human and fungal cytosolic Leucyl- tRNA synthetase editing domains: A structural basis for the rational design of antifungal benzoxaboroles. J Mol Biol, 390(2), 196-207.

Semenza, G. L. (2007). HIF-1 mediates the Warburg effect in clear cell renal carcinoma. J Bioenerg Biomembr, 39(3), 231-234.

Semenza, G. L., Roth, P. H., Fang, H. M., & Wang, G. L. (1994). Transcriptional regulation of genes encoding glycolytic enzymes by hypoxia-inducible factor 1. J Biol Chem, 269(38), 23757-23763.

Shen, J., & DiCorleto, P. E. (2008). ADP stimulates human endothelial cell migration via P2Y1 nucleotide receptor-mediated mitogen-activated protein kinase pathways. Circ Res, 102(4), 448-456.

Sherman, I. W., & Mudd, J. B. (1966). Malaria infection (Plasmodium iophurae): changes in free amino acids. Science, 154(3746), 287-289.

Shi, Y., Xu, X., Zhang, Q., Fu, G., Mo, Z., Wang, G. S., . . . Yang, X. L. (2014). tRNA synthetase counteracts c-Myc to develop functional vasculature. Elife (Cambridge), 3, e02349.

Shoemaker, B. A., Portman, J. J., & Wolynes, P. G. (2000). Speeding molecular recognition by using the folding funnel: the fly-casting mechanism. Proc Natl Acad Sci U S A, 97(16), 8868-8873.

Singh, S. K., Gurusiddaiah, S., & Whalen, J. W. (1985). Treponemycin, a nitrile antibiotic active against Treponema hyodysenteriae. Antimicrob Agents Chemother, 27(2), 239-245.

Somanath, P. R., Malinin, N. L., & Byzova, T. V. (2009). Cooperation between integrin alphavbeta3 and VEGFR2 in angiogenesis. Angiogenesis, 12(2), 177-185.

Sottile, J., & Hocking, D. C. (2002). Fibronectin polymerization regulates the composition and stability of extracellular matrix fibrils and cell-matrix adhesions. Mol Biol Cell, 13(10), 3546-3559.

Stebbins, C. E., Kaelin, W. G., Jr., & Pavletich, N. P. (1999). Structure of the VHL- ElonginC-ElonginB complex: implications for VHL tumor suppressor function. Science, 284(5413), 455-461.

Strilic, B., Kucera, T., Eglinger, J., Hughes, M. R., McNagny, K. M., Tsukita, S., . . . Lammert, E. (2009). The molecular basis of vascular lumen formation in the developing mouse aorta. Dev Cell, 17(4), 505-515.

260 Stupack, D. G., & Cheresh, D. A. (2002). ECM remodeling regulates angiogenesis: endothelial integrins look for new ligands. Sci STKE, 2002(119), pe7.

Sulzmaier, F. J., Jean, C., & Schlaepfer, D. D. (2014). FAK in cancer: mechanistic findings and clinical applications. Nat Rev Cancer, 14(9), 598-610.

Sundrud, M. S., Koralov, S. B., Feuerer, M., Calado, D. P., Kozhaya, A. E., Rhule-Smith, A., . . . Rao, A. (2009). Halofuginone inhibits TH17 cell differentiation by activating the amino acid starvation response. Science, 324(5932), 1334-1338.

Takahashi, A., Camacho, P., Lechleiter, J. D., & Herman, B. (1999). Measurement of intracellular calcium. Physiol Rev, 79(4), 1089-1125.

Tentori, L., Lacal, P. M., Muzi, A., Dorio, A. S., Leonetti, C., Scarsella, M., . . . Graziani, G. (2007). Poly(ADP-ribose) polymerase (PARP) inhibition or PARP-1 gene deletion reduces angiogenesis. Eur J Cancer, 43(14), 2124-2133.

Teplyakov, A., Obmolova, G., Chu, S. Y., Toedt, J., Eisenstein, E., Howard, A. J., & Gilliland, G. L. (2003). Crystal structure of the YchF protein reveals binding sites for GTP and nucleic acid. J Bacteriol, 185(14), 4031-4037.

Teske, B. F., Wek, S. A., Bunpo, P., Cundiff, J. K., McClintick, J. N., Anthony, T. G., & Wek, R. C. (2011). The eIF2 kinase PERK and the integrated stress response facilitate activation of ATF6 during endoplasmic reticulum stress. Mol Biol Cell, 22(22), 4390-4405.

Thomas, C. M., Hothersall, J., Willis, C. L., & Simpson, T. J. (2010). Resistance to and synthesis of the antibiotic mupirocin. Nat Rev Microbiol, 8(4), 281-289.

Torres-Larios, A., Sankaranarayanan, R., Rees, B., Dock-Bregeon, A. C., & Moras, D. (2003). Conformational movements and cooperativity upon amino acid, ATP and tRNA binding in threonyl-tRNA synthetase. J Mol Biol, 331(1), 201-211.

Traut, T. W. (1987). Synthesis of hybrid bisnucleoside 5',5"'-P1,P4-tetraphosphates by aminoacyl-tRNA synthetases. Mol Cell Biochem, 75(1), 15-21.

Tsai, Y. C., & Weissman, A. M. (2010). The Unfolded Protein Response, Degradation from Endoplasmic Reticulum and Cancer. Genes Cancer, 1(7), 764-778.

Tsuchiya, E., Yukawa, M., Miyakawa, T., Kimura, K. I., & Takahashi, H. (2001). Borrelidin inhibits a cyclin-dependent kinase (CDK), Cdc28/Cln2, of Saccharomyces cerevisiae. J Antibiot (Tokyo), 54(1), 84-90.

Turkson, J., Bowman, T., Garcia, R., Caldenhoven, E., De Groot, R. P., & Jove, R. (1998). Stat3 activation by Src induces specific gene regulation and is required for cell transformation. Mol Cell Biol, 18(5), 2545-2552.

261 Vagin, A., & Teplyakov, A. (2010). Molecular replacement with MOLREP. Acta Crystallogr D Biol Crystallogr, 66(Pt 1), 22-25.

Van Dang, C., & Yang, D. C. (1979). Disassembly and gross structure of particulate aminoacyl-tRNA synthetases from rat liver. Isolation and the structural relationship of synthetase complexes. J Biol Chem, 254(12), 5350-5356.

Van de Vijver, P., Ostrowski, T., Sproat, B., Goebels, J., Rutgeerts, O., Van Aerschot, A., . . . Herdewijn, P. (2008). Aminoacyl-tRNA synthetase inhibitors as potent and synergistic immunosuppressants. J Med Chem, 51(10), 3020-3029. van der Giet, M., Jankowski, J., Schluter, H., Zidek, W., & Tepel, M. (1998). Mediation of the vasoactive properties of diadenosine tetraphosphate via various purinoceptors. J Hypertens, 16(12 Pt 2), 1939-1943.

Vestweber, D. (2008). VE-cadherin: the major endothelial adhesion molecule controlling cellular junctions and blood vessel formation. Arterioscler Thromb Vasc Biol, 28(2), 223-232.

Wakabayashi, T., Kageyama, R., Naruse, N., Tsukahara, N., Funahashi, Y., Kitoh, K., & Watanabe, Y. (1997). Borrelidin is an angiogenesis inhibitor; disruption of angiogenic capillary vessels in a rat aorta matrix culture model. J Antibiot (Tokyo), 50(8), 671-676.

Wakasugi, K., & Schimmel, P. (1999). Two distinct cytokines released from a human aminoacyl-tRNA synthetase. Science, 284(5411), 147-151.

Wakasugi, K., Slike, B. M., Hood, J., Ewalt, K. L., Cheresh, D. A., & Schimmel, P. (2002). Induction of angiogenesis by a fragment of human tyrosyl-tRNA synthetase. J Biol Chem, 277(23), 20124-20126.

Wakasugi, K., Slike, B. M., Hood, J., Otani, A., Ewalt, K. L., Friedlander, M., . . . Schimmel, P. (2002). A human aminoacyl-tRNA synthetase as a regulator of angiogenesis. Proc Natl Acad Sci U S A, 99(1), 173-177.

Wang, L., Karlsson, L., Moses, S., Hultgardh-Nilsson, A., Andersson, M., Borna, C., . . . Erlinge, D. (2002). P2 receptor expression profiles in human vascular smooth muscle and endothelial cells. J Cardiovasc Pharmacol, 40(6), 841-853.

Warburg, O., Wind, F., & Negelein, E. (1927). The Metabolism of Tumors in the Body. J Gen Physiol, 8(6), 519-530.

Weis, S. M., & Cheresh, D. A. (2011). alphaV integrins in angiogenesis and cancer. Cold Spring Harb Perspect Med, 1(1), a006478.

Wek, R. C., Jiang, H. Y., & Anthony, T. G. (2006b). Coping with stress: eIF2 kinases and translational control. Biochem Soc Trans, 34(Pt 1), 7-11.

262 Welchman, R. L., Gordon, C., & Mayer, R. J. (2005). Ubiquitin and ubiquitin-like proteins as multifunctional signals. Nat Rev Mol Cell Biol, 6(8), 599-609.

Wellman, T. L., Eckenstein, M., Wong, C., Rincon, M., Ashikaga, T., Mount, S. L., . . . Lounsbury, K. M. (2014). Threonyl-tRNA synthetase overexpression correlates with angiogenic markers and progression of human ovarian cancer. BMC Cancer, 14, 620.

Wellman, T. L., Jenkins, J., Penar, P. L., Tranmer, B., Zahr, R., & Lounsbury, K. M. (2004). Nitric oxide and reactive oxygen species exert opposing effects on the stability of hypoxia-inducible factor-1alpha (HIF-1alpha) in explants of human pial arteries. FASEB J, 18(2), 379-381.

Welton, J. L., Khanna, S., Giles, P. J., Brennan, P., Brewis, I. A., Staffurth, J., . . . Clayton, A. (2010). Proteomics analysis of bladder cancer exosomes. Mol Cell Proteomics, 9(6), 1324-1338.

Wilkinson, B., Gregory, M. A., Moss, S. J., Carletti, I., Sheridan, R. M., Kaja, A., . . . Zhang, M. Q. (2006). Separation of anti-angiogenic and cytotoxic activities of borrelidin by modification at the C17 side chain. Bioorg Med Chem Lett, 16(22), 5814-5817.

Williams, J. W., & Morrison, J. F. (1979). The kinetics of reversible tight-binding inhibition. Methods Enzymol, 63, 437-467.

Williams, T. F., Mirando, A. C., Wilkinson, B., Francklyn, C. S., & Lounsbury, K. M. (2013). Secreted Threonyl-tRNA synthetase stimulates endothelial cell migration and angiogenesis. Sci Rep, 3, 1317.

Wolf, Y. I., Aravind, L., Grishin, N. V., & Koonin, E. V. (1999). Evolution of aminoacyl- tRNA synthetases--analysis of unique domain architectures and phylogenetic trees reveals a complex history of horizontal gene transfer events. Genome Res, 9(8), 689-710.

Wong, C., Wellman, T. L., & Lounsbury, K. M. (2003). VEGF and HIF-1alpha expression are increased in advanced stages of epithelial ovarian cancer. Gynecol Oncol, 91(3), 513-517.

Woolard, J., Vousden, W., Moss, S. J., Krishnakumar, A., Gammons, M. V., Nowak, D. G., . . . Wilkinson, B. (2011). Borrelidin modulates the alternative splicing of VEGF in favour of anti-angiogenic isoforms. Chem Sci, 2011(2), 273-278.

Yamada, K. M., & Even-Ram, S. (2002). Integrin regulation of growth factor receptors. Nat Cell Biol, 4(4), E75-76.

Yao, P., & Fox, P. L. (2013). Aminoacyl-tRNA synthetases in medicine and disease. EMBO Mol Med, 5(3), 332-343. 263 Yu, Y. P., Landsittel, D., Jing, L., Nelson, J., Ren, B., Liu, L., . . . Luo, J. H. (2004). Gene expression alterations in prostate cancer predicting tumor aggression and preceding development of malignancy. J Clin Oncol, 22(14), 2790-2799.

Zamecnik, P. C., Stephenson, M. L., Janeway, C. M., & Randerath, K. (1966). Enzymatic synthesis of diadenosine tetraphosphate and diadenosine triphosphate with a purified lysyl-sRNA synthetase. Biochem Biophys Res Commun, 24(1), 91-97.

Zetter, B. R. (1998). Angiogenesis and tumor metastasis. Annu Rev Med, 49, 407-424.

Zhou, Q., Kapoor, M., Guo, M., Belani, R., Xu, X., Kiosses, W. B., . . . Yang, X. L. (2010). Orthogonal use of a human tRNA synthetase active site to achieve multifunctionality. Nat Struct Mol Biol, 17(1), 57-61.

264