NOTCH4 is an inhibitor of canonical Notch signalling
Alexander James B.Sc. (hons)
A thesis in fulfilment of the requirements for the degree of Doctor of Philosophy.
St Vincent’s Clinical School
Faculty of Medicine
Victor Chang Cardiac Research Institute
December 2012 i | Page
THE UNIVERSITY OF NEW SOUTH WALES Thesis/Dissertation Sheet
Surname or Family name: James
First name: Alexander Other name/s: Campbell
Abbreviation for degree as given in the University calendar: PhD
School: St Vincent’s Clinical School Faculty: Medicine
Title: Notch4 is an inhibitor of canonical Notch signalling.
Abstract 350 words maximum: (PLEASE TYPE) Notch signal transduction is evolutionarily conserved with essential functions identified in numerous species from C.elegans to H.Sapiens. Notch is critical during embryonic development, in the adult, and in cancer. In mammals there are four Notch receptors; much of our understanding of Notch signalling comes from the study of NOTCH1. During embryogenesis, Notch1 is widely expressed and is required in numerous processes including blood vessel formation and remodelling (angiogenesis). The view of Notch signal transduction, achieved primarily through studying NOTCH1, is that Notch receptors undergo processing and are transported to the cell surface as a heterodimer. Ligand binding triggers a series of proteolytic cleavages leading to the release of the Notch intracellular domain, which translocates to the nucleus to activate transcription. In contrast to NOTCH1, little is known about NOTCH4; Notch4 is expressed in endothelial cells of the circulatory system, trans-activation by ligand is difficult to detect, and a phenotype in mice, reported to be null for Notch4 (Notch4d1 allele), has not been identified. Here, further analysis of NOTCH4 has revealed many features that are in contrast to NOTCH1: (i) unprocessed NOTCH4 was present on the cell surface, (ii) ligand did not trans-activate NOTCH4 signalling, (iii) NOTCH4 inhibited NOTCH1 signal transduction in a dose dependent manner, (iv) NOTCH4 induced the differentiation of myoblasts, and (v) subcellular localisation of NOTCH4 was different to NOTCH1, and coexpression resulted in NOTCH1 adopting the NOTCH4 pattern of localisation. In addition, postnatal retinal angiogenesis was examined in mice carrying the Notch4d1 allele. A delay in the expansion of the retinal vasculature was equally observed in both Notch4d1+/- and Notch4d1-/- mice. This suggested that Notch4d1 was not a null allele. Fittingly, the Notch4d1 allele produced transcripts that encode for much of the extracellular domain of NOTCH4. The expression pattern of this transcript was equivalent to Notch4 in mouse embryos. Like full length NOTCH4, expression of a cDNA based on this transcript also inhibited NOTCH1 signal transduction. These studies reveal for the first time a function for NOTCH4 that is, as an inhibitor of NOTCH1. Accordingly it is postulated that Notch4 will have a crucial function in angiogenesis, since Notch1 is required for angiogenesis and Notch4 is expressed in endothelial cells of the circulatory system.
Declaration relating to disposition of project thesis/dissertation
I hereby grant to the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or in part in the University libraries in all forms of media, now or here after known, subject to the provisions of the Copyright Act 1968. I retain all property rights, such as patent rights. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertation.
I also authorise University Microfilms to use the 350 word abstract of my thesis in Dissertation Abstracts International (this is applicable to doctoral theses only).
………………………………………………… ……………………………………..……………… ……….……………………...…….… ………… Witness Date Signature
The University recognises that there may be exceptional circumstances requiring restrictions on copying or conditions on use. Requests for restriction for a period of up to 2 years must be made in writing. Requests for a longer period of restriction may be considered in exceptional circumstances and require the approval of the Dean of Graduate Research.
FOR OFFICE USE ONLY Date of completion of requirements for Award:
THIS SHEET IS TO BE GLUED TO THE INSIDE FRONT COVER OF THE THESIS
ii | Page
Abstract Notch signal transduction is evolutionarily conserved with essential functions identified in numerous species from C.elegans to H.Sapiens. Notch is critical during embryonic development, in the adult, and in cancer. In mammals there are four Notch receptors; much of our understanding of Notch signalling comes from the study of NOTCH1. During embryogenesis, Notch1 is widely expressed and is required in numerous processes including blood vessel formation and remodelling (angiogenesis). The view of Notch signal transduction, achieved primarily through studying NOTCH1, is that Notch receptors undergo processing and are transported to the cell surface as a heterodimer. Ligand binding triggers a series of proteolytic cleavages leading to the release of the Notch intracellular domain, which translocates to the nucleus to activate transcription. In contrast to NOTCH1, little is known about NOTCH4; Notch4 is expressed in endothelial cells of the circulatory system, trans-activation by ligand is difficult to detect, and a phenotype in mice, reported to be null for Notch4 (Notch4d1 allele), has not been identified. Here, further analysis of NOTCH4 has revealed many features that are in contrast to NOTCH1: (i) unprocessed NOTCH4 was present on the cell surface, (ii) ligand did not trans-activate NOTCH4 signalling, (iii) NOTCH4 inhibited NOTCH1 signal transduction in a dose dependent manner, (iv) NOTCH4 induced the differentiation of myoblasts, and (v) subcellular localisation of NOTCH4 was different to NOTCH1, and coexpression resulted in NOTCH1 adopting the NOTCH4 pattern of localisation. In addition, postnatal retinal angiogenesis was examined in mice carrying the Notch4d1 allele. A delay in the expansion of the retinal vasculature was equally observed in both Notch4d1+/- and Notch4d1-/- mice. This suggested that Notch4d1 was not a null allele. Fittingly, the Notch4d1 allele produced transcripts that encode for much of the extracellular domain of NOTCH4. The expression pattern of this transcript was equivalent to Notch4 in mouse embryos. Like full length NOTCH4, expression of a cDNA based on this transcript also inhibited NOTCH1 signal transduction. These studies reveal for the first time a function for NOTCH4 that is, as an inhibitor of NOTCH1. Accordingly it is postulated that Notch4 will have a crucial function in angiogenesis, since Notch1 is required for angiogenesis and Notch4 is expressed in endothelial cells of the circulatory system.
iii | Page
Originality statement ‘I hereby declare that this submission is my own work and to the best of my knowledge it contains no materials previously published or written by another person, or substantial proportions of material which have been accepted for the award of any other degree or diploma at UNSW or any other educational institution, except where due acknowledgement is made in the thesis. Any contribution made to the research by others, with whom I have worked at UNSW or elsewhere, is explicitly acknowledged in the thesis. I also declare that the intellectual content of this thesis is the product of my own work, except to the extent that assistance from others in the project's design and conception or in style, presentation and linguistic expression is acknowledged.’
Signed ……………………………………………......
Date ……………………………………………......
iv | Page
Copyright statement ‘I hereby grant the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or part in the University libraries in all forms of media, now or here after known, subject to the provisions of the Copyright Act 1968. I retain all proprietary rights, such as patent rights. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertation. I also authorise University Microfilms to use the 350 word abstract of my thesis in Dissertation Abstract International (this is applicable to doctoral theses only). I have either used no substantial portions of copyright material in my thesis or I have obtained permission to use copyright material; where permission has not been granted I have applied/will apply for a partial restriction of the digital copy of my thesis or dissertation.' Authenticity statement ‘I certify that the Library deposit digital copy is a direct equivalent of the final officially approved version of my thesis. No emendation of content has occurred and if there are any minor variations in formatting, they are the result of the conversion to digital format.’
Signed ……………………………………………......
Date ……………………………………………......
v | Page
Acknowledgments
I would like to thank both my supervisor Professor Sally Dunwoodie and co-supervisor Dr Gavin Chapman for all their help and support throughout my PhD. Their approachability, advice and knowledge were invaluable during the course of these studies. I would also like to thank all the present and past members of the Dunwoodie laboratory whose help made this work possible. There have been too many people to thank individually among the staff and students of the Victor Chang Cardiac Research Institute who have provided advice and help throughout these studies. I would like to acknowledge the NHMRC and Australian Heart Foundation for generously supporting my Scholarship which made this work possible.
vi | Page
Abbreviations
A 260 Absorbance at 280nm A 280 Absorbance at 260nm AmpR Ampicillin resistance gene ANK Ankyrin repeat ANOVA Analysis of variance AP Alkaline phosphatase ATP Adenosine 5'-triphosphate BA Branchial arteries BCA Bicinchoninic acid BCIP 5-bromo-4-chloro-3’-indolyphosphate BSA Bovine serum albumin BTD Beta-trefoil domain C-terminus Carboxy terminus CADASIL Cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy CAT Chloramphenicol transferase cDNA Coding DNA CHAPS 3-[(3-cholamidopropyl) dimethylammonio]-1- propanesulfonate ChIP Chromatin immunoprecipitation CMV Cytomegalovirus CR Cysteine rich CTP Cytidine 5'-triphosphate DAPT N-[N-(3,5-Difluorophenacetyl)-L-alanyl]-S-phenylglycine t-butyl ester DEPC Diethylpyrocarbonate DIG Digoxigenin DMEM Dulbecco’s modified eagle medium DMSO Dimethylsulphoxide DNA Deoxyribose nucleic acid dNTP Deoxyribonucleotide DSL Delta/Serrate/LAG-1 DTT Dithiothreitol E Embryonic day EC Endothelial cell ECT Extracellular truncation EDTA Ethylene diamide tetra-acetic acid EGF Epidermal growth factor EGTA Ethylene glycol tetraacetic acid EJC Exon junction complex ER Endoplasmic reticulum ES Embryonic stem cell EtBr Ethidium bromide FAB Fragment antigen binding FCS Foetal calf serum FL Full length GlcNAc N-Acetylglucosamine GFP Green fluorescent protein GTP Guanosine 5'-triphosphate HA Hemagglutinin HDD Heterodimerisation domain HEPES 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid
vii | Page
HRP Horseradish peroxidase HUVEC Human umbilical venous endothelial cells IC Immunocytochemistry ICD Intracellular domains ICR Imprinting Control Region IHC Immunohistochemistry IP Immunoprecipitation ISV Intersomitic vessels KanR Kanamycin resistance gene KCM KCl, CaCl2 , MgCl2 KOMP Knockout out mouse project LAR Luciferase assay reagent LB Luria broth LEC Lymphatic endothelial cell Levamisole Tetramisole hydrochloride LNR Lin-12/Notch repeats LP Lens pit LTR Long terminal repeat MAEC Murine arterial endothelial cells MGI Mouse genome informatics MMTV Mouse mammary tumour virus MOPS 3-(N-morpholino) propanesulfonic acid Mya Million years ago N-linked Amino-linked N-terminus Amino-terminus NBT Nitro-blue tetrazolium NEXT Notch extracellular truncation NLS Nuclear localisation sequence NP-40 NonidetP-40 or Octylphenoxy polyethoxy ethanol (Igepal CA-630) NRR Negative regulatory region NTD N-terminal domain NTMT NaCl, Tris, MgCl2, Tween-20 OD600 Optical density at 600nm P Day post birth PBS Phosphate buffered saline ++ PBS PBS. MgCl2 , CaCl2 PBST PBS, Triton X-100 PBT PBS, Tween-20 PCR Polymerase chain reaction PDZ Post synaptic density protein, Drosophila disc large tumor suppressor and zonula occludens-1 protein domain PEG Polyethylene glycol PEST Proline, glutamine, serine and threonine rich PFA Paraformaldehyde PMSF Phenylmethanesulfonylfluoride PolyA Poly adenosine PVDF Polyvinylidene difluoride RACE Rapid amplification of cDNA ends RAM Rbpj-associated molecule RIPA Radioimmunoprecipitation assay buffer RNA Ribonucleic acid RNAase Ribonuclease
viii | Page
rRNA Ribosomal RNA RT-PCR Reverse transcription PCR SA Streptavidin SCD Spondylocostal dysostosis SDS Sodium dodecyl sulphate SSC NaCl, Trisodium citrate SV40 Simian vacuolating virus 40 T-ALL T-cell acute lymphoblastic leukaemia TA Tris acetate TAD Transactivation domain TAE Tris, sodium acetate, EDTA TB Tail bud TBST TBS, 0.1% Tween-20 TE Tris, EDTA TGN Trans-Golgi network TMIC Transmembrane and intracellular domains Tris Tris (hydroxymethyl) aminomethane TTBS TBS, 0.05% Tween-20 TM Transmembrane domain Tween-20 Polyoxyethylene-sorbitan monolaurate UTP Uridine 5'-triphosphate UV Ultraviolet light VWB Von Willebrand factor type C domain WAP Whey acidic promoter WB Western blot
Gene and protein abbreviations are defined within the text. Unless otherwise indicated gene and protein names refer to the mouse nomenclature in the MGI database.
ix | Page
x | Page
Table of Contents
Abstract ...... iii Originality statement ...... iv Copyright statement ...... v Authenticity statement ...... v Acknowledgments ...... vi Abbreviations ...... vii 1 Introduction ...... 1 1.1 Vascular development ...... 1 1.1.1 Vasculogenesis ...... 2 1.1.2 Angiogenesis ...... 6 1.1.2.1 Notch and tip cell formation ...... 7 1.1.2.2 VEGFA gradient ...... 10 1.1.2.3 Tubulogenesis ...... 10 1.1.2.4 Arterial/venous specification ...... 11 1.1.2.5 Vessel stabilisation ...... 14 1.1.3 Lymphangiogenesis ...... 16 1.2 The Notch pathway ...... 17 1.2.1 Mammalian Notch Receptors ...... 18 1.2.1.1 Evolution ...... 18 1.2.1.2 Structure ...... 19 1.2.1.3 Vascular expression of the Notch receptors ...... 21 1.2.2 Notch ligands ...... 22 1.2.2.1 DSL ligands ...... 22 1.2.2.2 Non-DSL ligands ...... 24 1.2.3 The canonical Notch pathway ...... 25 1.2.3.1 Receptor maturation ...... 28 1.2.3.2 FURIN and the regulation of surface expression ..... 32 1.2.4 Receptor recycling and degradation ...... 33 1.2.5 Ligand binding and receptor activation ...... 34 1.2.6 Receptor activation, S2 cleavage ...... 35 1.2.7 Receptor activation, S3 and S4 cleavage ...... 37 1.2.8 Transcription complex ...... 38 1.2.9 cis-Inhibition ...... 39 xi | Page
1.2.10 Non-canonical Notch signalling ...... 42 1.3 Aims ...... 43 2 Materials and Methods ...... 45 2.1 Materials ...... 45 2.1.1 Chemicals and reagents ...... 45 2.1.2 Enzymes ...... 45 2.1.3 Kits ...... 46 2.1.4 Miscellaneous materials ...... 46 2.1.5 Buffers and solutions ...... 47 2.1.6 Plasmids ...... 49 2.1.6.1 Basic cloning vectors ...... 49 2.1.6.2 pCAGiPuro based vectors ...... 49 2.1.6.3 pCMX based vectors ...... 50 2.1.6.4 pCS2 based vectors...... 51 2.1.6.1 Miscellaneous ...... 51 2.1.7 Mammalian Cell lines ...... 52 2.1.8 Oligonucleotides ...... 53 2.1.9 Bacterial strains and reagents ...... 55 2.1.1 Antibodies ...... 55 2.2 Methods ...... 56 2.2.1 Basic Molecular Biology Techniques ...... 56 2.2.1.1 Restriction digests ...... 56 2.2.1.2 Alkaline phosphatase ...... 57 2.2.1.3 Agarose gel electrophoresis ...... 57 2.2.1.4 Gel purification ...... 57 2.2.1.5 DNA ligations ...... 57 2.2.1.6 Gateway Cloning ...... 57 2.2.1.7 Bacterial Transformation ...... 57 2.2.1.8 Miniprep purification of plasmids ...... 58 2.2.1.9 Maxiprep purification of plasmids ...... 58 2.2.1.10 Sequencing ...... 58 2.2.1.11 Site Directed Mutagenesis ...... 58 2.2.2 Vector construction ...... 59 2.2.2.1 pCAGiPuro based vector construction ...... 59 2.2.2.2 pCMX based vector construction ...... 60
xii | Page
2.2.2.3 pCS2 based vector construction ...... 61 2.2.2.4 pGEM-T Easy based vectors ...... 61 2.2.2.5 Vectors for in situ hybridisation and northern probes ...... 62 2.2.3 Mammalian cell culture and transfection ...... 62 2.2.3.1 C2C12 and NIH3T3 lines ...... 62 2.2.3.2 MAEC ...... 62 2.2.3.3 HUVEC ...... 63 2.2.4 Mouse lines ...... 63 2.2.4.1 Genotyping ...... 63 2.2.5 Reverse Transcription-PCR (RT-PCR) ...... 64 2.2.6 3’RACE ...... 64 2.2.7 Real time PCR ...... 65 2.2.8 Northern Blotting ...... 65 2.2.9 N41437 Antibody purification ...... 66 2.2.10 Western Blotting ...... 67 2.2.11 Immunoprecipitation ...... 67 2.2.12 Biotinylation ...... 67 2.2.13 Luciferase Assays ...... 68 2.2.13.1 Luciferase Activity ...... 68 2.2.13.2 Co-culture Assays ...... 68 2.2.13.3 EDTA Assays ...... 69 2.2.13.4 NotchECT assays ...... 69 2.2.14 Immunocytochemistry (IC) ...... 71 2.2.14.1 Fixed cells ...... 71 2.2.14.2 Live imaging ...... 72 2.2.15 Immunohistochemistry (IHC)-Retina staining ...... 72 2.2.16 Wholemount RNA in situ hybridisation ...... 73 3 NOTCH4 is not activated by ligand ...... 75 3.1 Introduction ...... 75 3.1.1 Notch4 and the canonical Notch pathway ...... 75 3.1.2 Extracellular domain structure and ligand binding ...... 75 3.1.3 Ligand binding and activation of NOTCH4 ...... 77 3.1.4 The NOTCH4 negative regulatory region (NRR) and surface presentation ...... 78 3.1.5 NOTCH4ICD ...... 79 xiii | Page
3.1.5.1 Structural Divergence of NOTCH4ICD ...... 79 3.1.5.2 Transactivation potential of NotchICD family ...... 80 3.1.5.3 Degradation of NotchICD ...... 80 3.1.6 Aims ...... 81 3.2 Results ...... 82 3.2.1 Ligand dependent Notch activation - co-culture assays . 82 3.2.2 NOTCH4 does not signal in response to ligand in co- culture assays ...... 86 3.2.3 NOTCH4 does not signal in response to ligand in co- culture assays using a variety of cell types...... 88 3.2.4 Gamma-secretase cleavage of Notch4ICD constructs. .. 92 3.2.5 Gamma-secretase dependent entry of NOTCH4ICD to the nucleus...... 95 3.2.6 Notch4ECT is a weak activator of the Notch responsive reporter relative to Notch1ECT...... 96 3.2.7 Stably expressed NOTCH4 is not activated by ligand .. 100 3.2.8 NOTCH4 does not signal upon artificial activation with EDTA ...... 102 3.2.9 NOTCH4 is inefficiently S1 processed ...... 104 3.2.10 Cell surface presentation of NOTCH4 differs from that of NOTCH1...... 108 3.2.11 NOTCH4 is not activated by ligand despite introduction of a consensus FURIN cleavage site...... 110 3.2.12 NOTCH1 and NOTCH4 exhibit distinct subcellular localisations ...... 114 3.2.13 Chimaeric Notch receptors reveal the importance of the extracellular domain and protease cleavage region...... 116 3.3 Discussion ...... 122 3.3.1 NOTCH4 is inefficiently S1 processed and is presented on the cell surface as a full length receptor ...... 122 3.3.2 NOTCH4 is not activated in response to EDTA ...... 125 3.3.3 3.3.3 The extracellular and intracellular domains contribute to the inability of NOTCH4 to signal ...... 125 3.3.4 Conclusions ...... 127 4 NOTCH4 inhibits NOTCH1 signalling ...... 128 4.1 Introduction ...... 128 4.1.1 Paralogue specific functions of mammalian Notch receptors ...... 128
xiv | Page
4.1.2 Aims ...... 130 4.2 Results ...... 132 4.2.1 NOTCH4 dose dependently inhibits NOTCH1 signalling...... 132 4.2.2 The intracellular domains of NOTCH1 and NOTCH4 do not compete...... 136 4.2.3 NOTCH4 acts as a dose dependent inhibitor of NOTCH1 in a stable cell line ...... 141 4.2.4 NOTCH4 inhibits myogenic differentiation of C2C12 myoblasts...... 144 4.2.5 NOTCH1 expression in the presence of NOTCH4 ...... 152 4.2.6 NOTCH4 co-expression alters the subcellular distribution of NOTCH1 ...... 154 4.3 Discussion ...... 156 4.3.1 NotchICD competition ...... 156 4.3.2 Plasmid competition ...... 156 4.3.3 Full length NOTCH1 accumulates in NOTCH4 co- expressing cells ...... 157 4.3.3.1 NOTCH1 alters its subcellular localisation and co- localises with NOTCH4 in response to co-expression ...... 158 4.3.4 Conclusions ...... 158 5 Notch4d1 is not a true null allele ...... 159 5.1 Introduction ...... 159 5.1.1 Gain-of-function Notch4 models ...... 159 5.1.1.1 NOTCH4ICD models-breast cancer ...... 159 5.1.1.2 Notch4ICD models – vascular specific expression 160 5.1.2 Loss-of-function – the Notch4d1 mouse ...... 160 5.1.3 The mouse retina as a model of angiogenesis ...... 162 5.1.3.1 Overview of retinal angiogenesis ...... 162 5.1.3.2 Initiation of the retinal vasculature ...... 164 5.1.3.3 Vessel guidance ...... 164 5.1.3.4 Vascular remodelling and maturation ...... 165 5.1.4 Aims ...... 166 5.2 Results ...... 168 5.2.1 Establishment, genotyping and breeding of Notch4d1 mice ...... 168 5.2.2 Confirmation of the insertion site of the neo cassette 169
xv | Page
5.2.3 Notch4d1/d1 mice display no overt phenotype compared to Notch4+/+ littermates...... 169 5.2.4 Retinal angiogenesis in Notch4d1 mice ...... 173 5.2.5 Transcription from the Notch4d1 Allele...... 188 5.2.6 Isolation and analysis of the Notch4d1 transcript ...... 191 5.2.7 Northern analysis of Notch4d1 mice ...... 194 5.2.8 Analysis by in situ hybridisation of Notch4d1 mice. .... 196 5.2.9 The putative NOTCH4d1 protein ...... 197 5.2.10 NOTCH4d1 induces differentiation in C2C12 cells .... 201 5.2.11 NOTCH4d1 and NOTCH4 interact with NOTCH1 ...... 204 5.3 Discussion ...... 206 5.3.1 Notch4d1 mice have a transient delay in retinal angiogenesis ...... 206 5.3.2 Notch4d1 transcript ...... 208 5.3.2.1 Splicing artefacts in transgenic mice...... 209 5.3.2.2 Structure of the Notch4d1 transcript ...... 209 5.3.2.3 Possible functions of the NOTCH4d1 protein ...... 210 5.3.2.4 NOTCH4d1 as a secreted protein ...... 211 5.3.2.5 Notch4 functions left intact in Notch4d1 ...... 211 5.3.3 A definitive Notch4 knockout ...... 213 5.3.4 Conclusions ...... 214 6 Discussion ...... 216 6.1 Evolutionary evidence for Notch4 function ...... 216 6.2 NOTCH4ICD Functions ...... 217 6.3 Notch4 extracellular domain functions...... 219 6.4 A model of Notch4 function ...... 220 7 Bibliography ...... 227
List of Figures
Figure 1.1 Major vessels and vascular beds of the mouse embryo. 5 F i g u r e 1 . 2 T i p C e l l F o r m a t i o n ...... 8 Figure 1.3 Artery and vein structure ...... 12 F i g u r e 1 . 4 S t r u c t u r e o f N o t c h R e c e p t o r s ...... 20 Figure 1.5 Structure of DSL ligands ...... 23 Figure 1.6 Canonical Notch Signalling ...... 26 F i g u r e 1 . 7 G l y c o s y l a t i o n o f N o t c h r e c e p t o r s ...... 30
xvi | Page
Figure 3.1 A co-culture assay to measure ligand dependent Notch a c t i v a t i o n ...... 84 Figure 3.2 NOTCH4 did not signal in response to ligand in a co- c u l t u r e a s s a y ...... 87 Figure 3.3 NOTCH4 signalling is not induced by ligand co-culture a s s a y s u s i n g a v a r i e t y o f c e l l t y p e s ...... 90 Figure 3.4 The N41463 antibody reacts specifically with the neo- epitope produced by gamma-secretase cleavage of NOTCH4...... 94 Figure 3.5 The N41463 antibody specifically detects nuclear NOTCH4ICD...... 97 Figure 3.6 NOTCH4ICD can activate Notch dependent transcription albeit at a low level compared to NOTCH1ICD...... 98 Figure 3.7 NOTCH4 does not signal in response to ligand when e x p r e s s e d a t s t e a d y s t a t e l e v e l s ...... 101 Figure 3.8 EDTA treatment does not activate NOTCH4...... 103 Figure 3.9 Unlike NOTCH1, NOTCH4 is not S1-processed e f f i c i e n t l y i n a v a r i e t y o f c e l l t y p e s ...... 106 Figure 3.10 Unprocessed full length NOTCH4 is presented on the c e l l s u r f a c e ...... 109 Figure 3.11 NOTCH4 is not activated by ligand despite introduction of a consensus FURIN cleavage site...... 112 Figure 3.12 NOTCH4 is localised differently to NOTCH1 ...... 114 F i g u r e 3 . 1 3 C h i m a e r i c N o t c h r e c e p t o r c o n s t r u c t s ...... 117 Figure 3.14 Activation potential of Notch chimaeric constructs. .. 120 Figure 4.1 NOTCH4 dose dependently inhibits NOTCH1 signalling i n a c o - c u l t u r e a s s a y ...... 134 Figure 4.2 NOTCH4ICD does not inhibit NOTCH1ICD signalling. 137 Figure 4.3 Notch4ECT has a synergistic effect on NOTCH1 a c t i v a t i o n ...... 140 Figure 4.4 NOTCH4 inhibits signalling by stably expressed N O T C H 1 ...... 142 Figure 4.5 NOTCH4 over expression induces myotube formation in C 2 C 1 2 c e l l s ...... 146 Figure 4.6 NOTCH4 induces MHC expression in C2C12 cell lines...... 148 Figure 4.7 NOTCH4 induces a 30-fold increase in MHC expression i n C 2 C 1 2 c e l l s ...... 150 Figure 4.8 NOTCH4 co-expression with NOTCH1 increases unprocessed NOTCH1...... 153 F i g u r e 4 . 9 N O T C H 4 c o - e x p r e s s i o n a l t e r s N O T C H 1 l o c a l i s a t i o n ... 155 Figure 5.1 Retinal Angiogenesis ...... 163 Figure 5.2 The Notch4d1 allele ...... 170 Figure 5.3 Notch4d1/d1 mice displayed normal postnatal growth and were produced in Mendelian ratios...... 172 Figure 5.4 Retinal angiogenesis in Notch4d1 mice at day 3 post b i r t h ...... 176
xvii | Page
Figure 5.5 Measurements of retinal angiogenesis at day 3 post birth ...... 178 Figure 5.6 Retinal angiogenesis of Notch4d1 mice at 5 days post birth ...... 180 Figure 5.7 Measurements of retinal angiogenesis at 5 days post birth ...... 182 Figure 5.8 Retinal angiogenesis in Notch4d1 mice at 7 days post birth ...... 184 Figure 5.9 Measurements of retinal angiogenesis at 7 days post birth ...... 186 Figure 5.10 RT-PCR of Notch4d1 mice...... 190 Figure 5.11 The transcript produced from the Notch4d1 allele contains two novel exons...... 192 Figure 5.12 Northern analysis of Notch4d1 m i c e ...... 195 Figure 5.13 Expression domain of the Notch4d1 t r a n s c r i p t ...... 198 Figure 5.14 The NOTCH4d1 protein inhibits NOTCH1 activation .. 200 Figure 5.15 NOTCH4d1 induces differentiation in C2C12 cells ..... 202 Figure 5.16 NOTCH1 interacts with NOTCH4d1 and NOTCH4 ...... 205 Figure 6.1 Model of NOTCH4 function...... 222
List of Tables
Table 2.1 NIH3T3 and C2C12 cell lines...... 52 Table 2.2 Oligonucleotides ...... 53 Table 2.3 Antibodies ...... 55 Table 2.4 Transfection mixes for luciferase assays...... 69
xviii | Page
1 Introduction The Notch pathway is a highly conserved signal transduction pathway that is used reiteratively throughout development. Notch is involved in cell fate decisions in all three germ layers and almost every tissue as well as having roles in tissue homeostasis (Hofmann and Iruela-Arispe, 2007). This wide range of functions is reflected in the diverse array of disease states and developmental processes in which the Notch pathway has been implicated (Andersson et al., 2011; Hofmann and Iruela-Arispe, 2007; Penton et al., 2012). Homologues of both Notch receptors and ligands have been found in members of all metazoan lineages including basal metazoans such as sponges and placazoa but are absent from other lineages (Gazave et al., 2009). This ubiquity of Notch signalling in development and homeostasis in the metazoan lineage has led to the popularisation of the quote “there are 2 types of biologists: those who work on Notch and those who do not yet realise they work on Notch” (Hofmann and Iruela-Arispe, 2007). The pathway derives its name from the “Notched” phenotype described in 1917 in a paper entitled “The Theory of the Gene” (Morgan, 1917) showing that the study of Notch genetics is almost as old as the concept of the gene itself. The phenotype described was due to haploinsufficiency of Notch which resulted in some strains of Drosophila with “Notches” at the margin of the wing.
The Notch pathway is absolutely required for the formation of the circulatory system. Our interest in both the circulatory system and the Notch pathway led us to examine the role of Notch4 in vascular development.
1.1 Vascular development It is essential for all aerobic cells to have access to sufficient nutrients and oxygen to maintain metabolism. Equally, it is essential for these cells to be able to efficiently remove cellular waste. Among metazoans, a variety of systems have evolved to deal with this problem. In smaller animals, most notably in the phyla Platyhelminthes, Nematoda and Porifera, the small body size allows every cell to have direct access to nutrients and oxygen (Jin and Patterson, 2009). An example of this is in the well known model organism Caenorhabditis elegans (C.elegans) which contains an internal cavity called the pseudocoelem allowing all cells access to nutrients and oxygen. Waste products are dealt with by two cells, which extend along the body axis. These cells regulate osmotic pressure and excrete waste products through a pore located on the ventral side of the head (Riddle et al., 1997). In larger animals not all cells in the body can have direct access 1 | Page
to nutrients and oxygen. Thus a circulatory system is required. There are two major types of circulatory system in metazoans. These are referred to as open and closed systems. In open systems, e.g. in Arthropods, the blood and interstitial fluid mix and are referred to as haemolymph. The internal organs are submerged in this fluid within a cavity called the haemocoel and can thus receive nutrients and oxygen and excrete waste (Wasserthal, 2007). The second type of system is the closed system. In a closed system the blood circulates only within the lumen of blood vessels and does not mix with the interstitial fluid. This type of circulatory system is found in Chordata, Annelida and some Mollusca (Cephalopods) (Jin and Patterson, 2009). In vertebrates the circulatory system ensures that every cell in the body is within 100-200 m of a vessel which is the diffusion limit for oxygen (Folkman, 1971). The absolute requirement for access to this system by every cell of the body requires strictly controlled development of the circulatory system. In addition the circulatory system must be able to adjust to growth and repair after injury. Thus fine tuning of the circulatory system requires precise developmental and physiological control. Below is a review of the vertebrate vascular system with special emphasis on the mouse as a model organism and the role Notch signalling plays in the formation of the vascular system.
1.1.1 Vasculogenesis The development of the mammalian vasculature can be divided into three distinct processes, vasculogenesis, angiogenesis and lymphangiogenesis. Vasculogenesis is the de novo formation of vessels by the differentiation and association of endothelial progenitor cells (Risau and Flamme, 1995). Angiogenesis is the formation of new vessels from existing ones (Section 1.1.2). Lymphangiogenesis is the development of the lymphatic system (Section 1.1.3). Thus the very early stages of development are vasculogenic in nature.
Endothelial progenitors originate from the mesoderm. Prior to their incorporation into vessels these primitive endothelial cells are called angioblasts. One of the earliest markers of this lineage is the vascular endothelial growth factor A (Vegfa) receptor 2 ( Vegfr2) also known as Flk1 and renamed kinase insert domain protein receptor; (Kdr), which is co-expressed with brachyury, a mesodermal marker, in angioblast precursors in the primitive streak (Huber et al., 2004). The initial in situ differentiation occurs in mesoderm that is in contact with endoderm suggesting that the endoderm plays a role in endothelial specification. However, the endoderm is not absolutely required for endothelial induction since several organs including the allantois, kidney and coronary mesenchyme give rise to angioblasts in the absence of endoderm 2 | Page
(Cumano et al., 2000; Perez-Pomares et al., 1998; Robert et al., 1998). In addition angioblasts can still develop in mesoderm that has been experimentally separated from the endoderm (Vokes and Krieg, 2002). Although the endoderm does not appear to specify endothelial cell fate on its own, endoderm signals do influence vascular development.
Early endothelial progenitor cells migrate in response to endoderm derived signals including fibroblast growth factor 2 (FGF2) and bone morphogenetic protein 4 (BMP4), first at embryonic day 6-6.5 (E6-6.5) to the extra-embryonic mesoderm tissue of the yolk sac (Coultas et al., 2005; Ferkowicz and Yoder, 2005; Risau and Flamme, 1995) followed by the embryo proper (Eichmann et al., 1993; Quinn et al., 1993), the allantois (Caprioli et al., 1998; Caprioli et al., 2001) and later the placenta (Demir et al., 2007; Yamaguchi et al., 1993) (Figure 1a). In the yolk sac angioblasts then proliferate, aggregate and differentiate to form blood islands (Haar and Ackerman, 1971; Heinke et al., 2012). Once formed the blood islands fuse and form a primitive vascular plexus which is later remodelled by angiogenesis (Figure 1b). Blood islands contain haematopoietic progenitors surrounded by a peripheral layer of angioblasts; this suggests that blood cells arise intra- vascularly and thus share a common progenitor with endothelial cells, the haemangioblast. In vitro experiments using embryoid bodies show that both lineages can arise from a single clone (Choi, 2002). However, it is not clear if this clone is really a bi- potent progenitor or a more primitive multi-potent progenitor. These cells, when detected, are also exceedingly rare. One study found only 1-5 haemangioblast cells per embryo (Huber et al., 2004). In vivo lineage tracing studies found that the cells within blood islands are polyclonal and thus do not arise in situ by differentiation (Ueno and Weissman, 2006). The rarity of the haemangioblast and the demonstration that blood islands are polyclonal suggests that the haemangioblast plays a small role in vascular development. The situation is further complicated by the ability of some endothelial cells to trans-differentiate and produce haematopoietic cells. These cells are referred to as haemogenic endothelium (Goldie et al., 2008). Cells able to differentiate into endothelial cells can be isolated from the circulation both in the embryo (LaRue et al., 2003) and the adult (Shi et al., 1994). The extent of the contributions these cells make to blood vessels remains under dispute. A role in the adult was supported by observations that the endothelialisation of prosthetic devices does not necessarily proceed from the margins of the device, which a purely angiogenic process would predict, and was described as “fallout endothelialisation” (Drake, 2003; Shi et al., 1994). Fallout endothelialisation was later shown to involve circulating CD34+ cells. Circulating CD34+ cells also contribute to 3 | Page
neovascularisation after ischemic injury (Asahara et al., 1997) and circulating cells have been shown to contribute to vasculogenesis in the embryo (Drake, 2003). The above mechanisms for endothelial differentiation remain under dispute and most endothelial cells in the embryo derive from progressive restriction of the mesoderm under the influence of Indian hedgehog (IHH), FGF2, BMP4 and VEGFA (Heinke et al., 2012).
After the blood islands form in the yolk sac vasculogenesis is initiated in the embryo proper forming the dorsal aorta and cardinal veins (Coultas et al., 2005) (Figure 1a). In response to VEGFA from the endoderm angioblasts migrate towards, but do not cross, the midline. Crossing of the midline is prevented by the notochord that produces the BMP antagonists Noggin (NOG) and Chordin (CHRD) (Coultas et al., 2005; Nimmagadda et al., 2005; Reese et al., 2004). Thus the aorta forms as two parallel cords along the midline which later fuse to form the single midline dorsal aorta and connect to the heart and yolk sac.
It was thought that the subsequent vascularisation of organs was angiogenic in nature i.e. vessels sprouting from existing vessels which does occur in the intersomitic vessels and the vessels of the brain. However, it is now clear that vasculogenesis occurs in a number of organs including the liver (Matsumoto et al., 2001), kidney (Robert et al., 1998), pancreas (Lammert et al., 2001), lung (Gebb and Shannon, 2000) and heart (Perez-Pomares et al., 1998). The developing organs do not just instruct the vasculature to develop. The vasculature itself plays an inductive role in organogenesis. For example, the liver bud first appears as a multilayered epithelium which is surrounded by endothelial cells. The endothelial cells subsequently migrate in and aggregate into sinusoids. Hepatocytes then migrate from the endoderm to the septum transversum and proliferate. In the absence of endothelial cell derived signals the hepatocytes fail to migrate and differentiate (Coultas et al., 2005; Matsumoto et al., 2001). A similar process occurs in the pancreas where endothelial cells are required for the maturation of islet endocrine cells and insulin secretion (Lammert et al., 2001).
The initial vessels formed by vasculogenesis undergo further growth and remodelling in the process of angiogenesis.
4 | Page
Figure 1.1 Major vessels and vascular beds of the mouse embryo.
Reproduced and adapted with permission from (Wang et al., 1998) a. Arterial (a. red) and venous (v. blue) vessels in the developing mouse embryo. Arteries and veins are identified by their expression of Efnb2 and Ephb4 respectively. b. Remodelling of the yolk sac vascular plexus. Stage names are according to (Risau, 1997).
5 | Page
1.1.2 Angiogenesis In contrast to vasculogenesis, the de novo formation of vessels via differentiation and aggregation of endothelial precursors, angiogenesis is the formation of new vessels from existing vessels. Developmental angiogenesis is of great medical interest as neoangiogenesis in the adult utilises largely similar processes and a number of disease states are characterised by aberrant angiogenesis (Carmeliet, 2003). Angiogenesis can be divided into two distinct processes; intussusception, the splitting of existing vessels, and sprouting angiogenesis, the sprouting of new vessels from existing ones.
Intussusception can occur more quickly than sprouting angiogenesis and can proceed under conditions of blood flow (Heinke et al., 2012). Intussusception occurs in the development of the lung, heart and yolk sac. Endothelial cells invaginate until the opposing walls of the vessel are in contact thus forming a trans-capillary pillar. The pillar is then stabilised by collagen fibres produced from infiltrating pericytes (Heinke et al., 2012).
Sprouting angiogenesis has been studied intensively in a few well described model systems; the mouse retina and hindbrain (Fruttiger, 2007; Hellström et al., 2007; Lobov et al., 2007; Ridgway et al., 2006; Suchting et al., 2007), xenograft tumour angiogenesis (Noguera-Troise et al., 2006; Ridgway et al., 2006; Thurston et al., 2007) and the intersomitic vessels in the zebrafish (Leslie et al., 2007; Siekmann and Lawson, 2007). Sprouting angiogenesis is initiated in response to endothelial growth factors, most notably VEGFA (Figure 1.2a). The first response made by an existing vessel to sprouting cues is the release of extracellular matrix degrading enzymes that degrade the surrounding matrix. Matrix components and blood plasma provide a scaffold along which cells of the developing sprout can migrate (Davis and Senger, 2005; Heinke et al., 2012). VEGFA increases proliferation and migration in endothelial cells and the sprout develops by both migration and proliferation. The developing sprout must be able to control both of these events separately. If all cells that are initially equivalent in the existing vessel proliferated, the vessel would just enlarge. If all cells migrated, the vessel would disintegrate. Thus a cell must be selected to migrate while others proliferate (Figure 1.2c). The migratory cell is called the tip cell and those following, the stalk cells. The mechanisms underlying tip cell selection has been most extensively studied in the mouse retina (see Chapter 5).
6 | Page
1.1.2.1 Notch and tip cell formation In response to VEGFA the tip cell extends long filopodia (Gerhardt et al., 2003) (Figure 1.2c). The length and stability of the filopodia are increased by VEGFA. The filopodia contain the VEGFA receptor KDR. Thus the increase in length increases the surface area covered and the amount of VEGFA encountered, this further reinforces the tip cell state (Bentley et al., 2008) (Figure 1.2c). In addition to increasing filopodia length, VEGFA drives the expression of the Notch ligand Delta-like 4 (Dll4) in vascular endothelial cells (Figure 1.2b). Of the many genes involved in vasculogenesis and angiogenesis lethality caused by haploinsufficiency has only been reported for VegfA and Dll4 (Carmeliet et al., 1996; Duarte et al., 2004; Ferrara et al., 1996; Gale et al., 2004; Krebs et al., 2004). The importance of Dll4 in establishing tip cells is demonstrated by the phenotype of Dll4-/- mice. Vasculogenesis in Dll4-/- mice proceeds in an apparently normal manner as the dorsal aorta forms (Duarte et al., 2004). However, during angiogenesis Dll4-/- mice display a hyperbranched capillary network. The haploinsufficient phenotype of Dll4+/- mice is strain dependent. On an outbred ICR background Dll4 heterozygous pups survive, which has allowed analysis of post-natal angiogenesis in the retina (Lobov et al., 2007) (see Chapter 5 for a discussion of this model). DLL4, via activation of Notch in neighbouring vascular endothelial cells, suppresses tip cell formation and the neighbouring cells become stalk cells. In mice where Notch1 was incompletely removed (via cre/lox recombination) from vascular endothelial cells further support this model (Hellström et al., 2007). The majority, although not all, Notch1 deficient cells were found to be tip cells.
However, stalk cells are also exposed to VEGFA. The suppression of the tip cell state can in part be explained by Notch regulation of the VEGFA receptors VEGFR1 (renamed FMS-like tyrosine kinase 1; FLT1) and KDR (Figure 1.2b). KDR is activated by VEGFA while FLT1 acts as a decoy receptor to suppress VEGFA induced signalling through KDR. Notch activation causes the down regulation of Kdr and the upregulation of Flt1 (Suchting et al., 2007). Thus a Notch signalling cell has less VEGF signalling and therefore produces less DLL4. This model is too simplistic implying that tip and stalk cell formation is static; however gene expression and live imaging suggest otherwise. For example, expression of Dll4 is heterogenous and there is no clear cut distinction between tip and stalk cells (Claxton and Fruttiger, 2004; Hellström et al., 2007; Hofmann and Luisairuelaarispe, 2007; Lobov et al., 2007). Moreover, the growth of intersomitic vessels in the zebrafish proved its value as a model in the analysis of tip and stalk cell switching. Live video microscopy demonstrated that tip cells can rapidly switch from a migratory phenotype and integrate into a 7 | Page
vessel (Leslie et al., 2007; Torres-Vazquez et al., 2004). Dll4 is regulated by other pathways in addition to VEGF including Notch. The dynamic switching of tip versus stalk cell fate has been described as a tug of war for the tip position (Phng and Gerhardt, 2009). Further complexity and fine tuning was observed in mice in which Jagged1 (Jag1) another Notch ligand, was knocked out of the vasculature. These mice exhibited reduced tip cell number and increased Notch signalling (Benedito et al., 2009). This action of a ligand as a suppressor of Notch signalling is discussed in Section 1.1.2.4.
Figure 1.2 Tip Cell Formation
Reproduced and adapted with permission from (Potente et al., 2011). a. In response to VEGFA endothelial cells initiate sprouting angiogenesis by releasing enzymes to degrade the extracellular matrix (ECM). b. Tip cell selection is driven by Delta/Notch signalling in a feedback loop involving the VEGFA receptors KDR and FLT1. c. Tip cells migrate along the VEGFA gradient while trailing stalk cells proliferate. VEGFA induces filopodia formation and expression of KDR on filopodia further reinforces the tip cell state.
8 | Page
9 | Page
1.1.2.2 VEGFA gradient Once a tip cell is selected the growing vasculature needs to establish the direction of growth. Heparin sulphate proteoglycans in the extracellular matrix can bind signalling molecules and provide guidance information to new vessels. A number of VegfA isoforms are expressed due to alternative splicing. These isoforms, and their differential ability to bind the extracellular matrix, drive the formation of a VEGFA signalling gradient (Carmeliet and Tessier-Lavigne, 2005; Ruhrberg et al., 2002). VEGFA120 is the smallest isoform and is the most readily diffusible. VEGFA188 contains two heparin binding domains and is found exclusively in association with the surface of the secreting cell and surrounding extracellular matrix. VEGFA164 is intermediate between VEGFA120 and VEGFA188 in its diffusion due to the lack of one of the two heparin binding domains found in VEGFA188 (Park et al., 1993). In VegfA120/120 mice exon 6 and 7 and the intervening sequences have been replaced by a cDNA construct. These mice only express the VEGFA120 isoform. They survive until birth but display defects in postnatal angiogenesis in a number of organs (Carmeliet et al., 1999; Ng et al., 2001). These defects result in vessels with fewer branch points and an increased luminal diameter. The defect observed in VegfA120/120 mice is due to an abnormal distribution of endothelial cells rather than impaired proliferation.
In mice that express only VEGFA188 the opposite phenotype was observed. There was an increase in vessel sprouts with excessive filopodia that were poorly directed. Vessel patterning appeared normal in VegfA120/188 double heterozygotes that do not express Vegfa164, ruling out a specific role for VEGFA164 signalling. These observations provide a model for VEGFA guidance of vessels in which VEGFA120 provides a long range signal to drive proliferation while the less diffusible heparin binding isoform, VEGFA188, provides high local concentrations driving guidance (Ruhrberg et al., 2002).
1.1.2.3 Tubulogenesis The vascular lumen is formed in two distinct ways. The first mechanism involves a rearrangement of endothelial cells. The cells initially form a cord and adhere to each other. The cell/cell junctions either dissolve or migrate around the plasma membrane to the centre of the cord. A lumen is then formed in the centre (Jin et al., 2005). These endothelial movements form the lumen of the dorsal aorta (Strilic et al., 2009). The other mechanism used to form lumens involves intra- and inter-cellular fusion of vacuoles (Blum et al., 2008; Kamei et al., 2006). Fusion of intracellular vacuoles can create a single cell with a central lumen or can be
10 | Page
achieved by endothelial cell movements and subsequent inter- cellular fusions.
1.1.2.4 Arterial/venous specification The vasculature forms a closed loop with a delivery (arterial) and return (venous) system (Figure 1.3). In addition the lymphatic vasculature has important roles in fluid balance and immune function (Section 1.1.3). The specification of arterial/venous character shows a high degree of plasticity. This is best demonstrated by vein grafts adopting arterial character after transplant under the influence of blood pressure and flow (Bush et al., 1986; Henderson et al., 1986). However, it has become clear that arterial/venous identity is genetically predetermined and occurs before the onset of blood flow (Aitsebaomo et al., 2008; Hong et al., 2008; Swift and Weinstein, 2009; Wang et al., 1998).
Arteries and veins are defined by their calibre and position within the vascular tree (Figure 1.3). The largest artery in the body, the aorta, connects to the left ventricle of the heart and successively branches into smaller and smaller vessels that reach their target tissues and organs. These smallest arteries are called arterioles. Branching continues, with calibre reduction, until the smallest vessels, the diameter of a red blood cell, are called capillaries. The capillaries then return the deoxygenated blood via the venous system. The smallest veins are referred to as venules. The venules progressively increase in size to form veins until they converge on the largest veins, the venae cavae. The superior vena cava, from the upper trunk of the body and the inferior vena cava from the lower trunk both empty into the right atrium of the heart (dela Paz and D'Amore, 2009). With the exception of the smallest vessels, the capillaries, both veins and arteries have the same basic structure (Figure 1.3). The outer layer, tunica externa, is composed of connective tissue, predominantly collagen and elastic fibres and lends structural support to the vessel. In veins this outer layer constitutes the majority of the vessel wall whereas in arteries it is only about half the total thickness. Underneath the tunica externa is the tunica media. This layer consists of smooth muscle cells and elastic fibres which is the thickest layer of the vessel in arteries and less pronounced in veins. The internal layer, tunica intima, lines the lumen and is composed of endothelial cells and an extracellular matrix of collagen and elastic fibres. The venous circulation also contains venous valves that extend into the lumen and prevent the backflow of blood.
11 | Page
Figure 1.3 Artery and vein structure
Oxygenated blood is pumped by the heart through arteries. Arteries progressively branch to form arterioles and finally capillaries to supply tissues. The deoxygenated blood is returned via the capillaries which fuse to form venules and finally veins. Arteries and veins have the same basic structure of three “tunics”; the tunica intima composed of endothelial cells and associated collagen and elastic fibres, the tunica media composed of smooth muscle and elastic fibres and the tunica externa composed of connective tissue. The degree of oxygen saturation is symbolised by blue (low oxygen) to red (high oxygen).
12 | Page
13 | Page
Arterial or venous fate was originally thought to be established by haemodynamic forces. However, arterial expression of Ephrin B2 ( Efnb2) and venous expression of Eph receptor B4 (Ephb4) was found to occur before the onset of blood flow (Wang et al., 1998) (Figure 1a). Arterial fate is acquired by the combined effect of forkhead box C1 (FOXC1) and C2 (FOXC2) transcription factors and VEGFA. Loss of both Foxc1 and Foxc2 results in arterial/venous shunts and a loss of arterial identity (Seo et al., 2006), a phenotype also observed in Notch pathway mutants (Carlson et al., 2005; Krebs et al., 2010; Miniati et al., 2009; Murphy et al., 2008). FOXC1 and FOXC2 directly bind and activate the promoter of the Notch ligand Dll4 (Seo et al., 2006) which is enhanced by VEGFA activation (Kume, 2010). DLL4 then activates Notch signalling which drives the expression of the arterial markers Efnb2 and Neuropilin 1 (Nrp1) (Grego-Bessa et al., 2007). Venous endothelial cells express COUP-TFII, renamed nuclear receptor subfamily 2, group F, member 2 (NR2F2), which drives the expression of Efnb4 and suppresses the expression of Notch pathway genes and Nrp1. Bidirectional signalling through EFNB2 expressed in arterial cells, and EPHB4 expressed in venous cells, establishes and maintains arterial-venous interactions and identity (Kullander and Klein, 2002).
1.1.2.5 Vessel stabilisation Vessel stabilisation requires the recruitment and maturation of smooth muscle cells. Endothelial cells secrete platelet derived growth factor B (PDGFB) which binds to its receptor, platelet derived growth factor receptor B (PDGFRB) expressed on undifferentiated mesenchymal cells and recruits them to the developing vessel (Hellstrom et al., 1999; Patel-Hett and D'Amore, 2011). The interaction between the recruited pericyte and endothelial cell is further strengthened via signalling mediated by angiopoietin and endothelial-specific receptor tyrosine kinase ( Tek). Signalling via angiopoietin and TEK are not required for the initial specification or assembly of the vascular plexus. Deletion of angiopoietin 1 (Angpt1) results in embryonic death between E9.5 and E12.5 due to vascular haemorrhage (Suri et al., 1996). This is due to a failure of the developing vessels to recruit smooth muscle mural cells. Deletion of Tek phenocopies deletion of Angpt1 (Dumont et al., 1994; Sato et al., 1995; Suri et al., 1996). In addition, over-expression of Angpt2 resembles this phenotype (Heinke et al., 2012; Maisonpierre et al., 1997). Secretion of ANGPT1 strengthens the pericyte/endothelial interaction while conversely the secretion of ANGPT2 inhibits this interaction. The opposing effects of ANGPT1 and ANGPT2 are due to differing abilities to activate the TEK receptor. ANGPT1 can induce signalling from TEK whilst ANGPT2 fails to activate the receptor and acts as a competitive inhibitor of ANGPT1 (Fiedler et al., 14 | Page
2003; Maisonpierre et al., 1997; Valenzuela et al., 1999). The balance between ANGPT1 and ANGPT2 expressed by endothelial cells prevents premature investment of the vessel by smooth muscle.
Maturation of the investing cells to smooth muscle cells is controlled by the Notch pathway. The Notch ligand, Jag1, is expressed in endothelial cells and is induced by Notch signalling (Liu et al., 2009; Ross and Kadesch, 2004). Selective deletion of Jag1 in the vasculature leads to embryonic lethality due to a failure of smooth muscle gene expression (High et al., 2008). Notch3 expression is largely confined to vascular smooth muscle cells (Joutel et al., 2000; Villa et al., 2001). Mutations of the Notch3 receptor cause the inherited neurovascular genetic disorder cerebral autosomal dominant arteriopathy with subcortical i nfarcts and leukoencephalopathy (CADASIL) (Joutel et al., 1996). CADASIL is characterised by a degeneration of the smooth muscle lining of vessels. In addition, genetic deletion of Notch3, while not embryonic lethal, does result in defects in arterial smooth muscle cell investment (Domenga et al., 2004; Liu et al., 2010b). Constitutive activation of the Notch pathway induces the expression of smooth muscle genes including smooth muscle alpha actin (Doi et al., 2006; Noseda et al., 2004a; Noseda et al., 2004b; Tang et al., 2008). In addition activation of NOTCH3 results in the induction of expression of Notch3 itself and Jag1 (Liu et al., 2009).
These observations have led to the following model of Notch induced vascular smooth muscle cell maturation (Liu et al., 2009). Endothelial cells expressing JAG1 activate NOTCH3 in the premature smooth muscle cells investing the vessel. Activation of NOTCH3 leads to an upregulation of smooth muscle genes. In addition both Notch3 and Jag1 are also upregulated. The feedback loop of NOTCH3 activation leading to upregulation of both receptor and ligand reinforces the maturation of neighbouring smooth muscle.
The role of Jag1 in the endothelium appears to be twofold. JAG1 inhibits Notch signalling in endothelial cells whilst activating Notch signalling in maturing smooth muscle. The separation of JAG1 activation of smooth muscle versus endothelial activation is mediated by O -fucosylpeptide 3-beta-N- acetylglucosaminyltransferase (Fringe) activity (Section 1.2.3.1). Fringe modification of Notch receptors in endothelial cells inhibits activation by JAG1 and enhances activation by DLL4 (Chen et al., 2001; Evrard et al., 1998; Hicks et al., 2000; Moloney et al., 2000; Shimizu et al., 2001; Yang et al., 2005; Zhang and Gridley, 1998). These observations explain the surprising result that while
15 | Page
deletion of Dll4 causes decreased Notch signalling and a hyper- branched network, Jag1 deletion increases Notch signalling and reduces tip cell formation and branching (Benedito et al., 2009). The inhibition is likely to be due to JAG1 competing with DLL4 for binding to Fringe modified Notch (Benedito et al., 2009; Yang et al., 2005). Thus the Notch ligand JAG1 acts as a suppressor of Notch signalling in the developing endothelial sprouts whilst retaining its ability to activate NOTCH3 in recruited pericytes. This dual activity of JAG1 highlights the complexity of Notch signalling in mammals where 4 receptors and 5 ligands are present.
1.1.3 Lymphangiogenesis In addition to the blood vessels mammals also have a lymphatic circulation. Although 90% of the fluid that is released to tissues returns via the venous system, the remaining 10% of the interstitial fluid returns via the lymphatic vasculature. In addition to regulating fluid balance the lymphatic vasculature also plays important roles in the generation of immune cells, the initiation of the immune response and transport of dietary fat from the intestine to the liver. The lymphatic vasculature is found in all vascularised organs except the brain and retina where drainage is either venous or perivascular in nature (Albrecht and Christofori, 2011).
The lymphatic system begins in the periphery with blind ending capillaries approximately 30-80m in diameter. These lymphatic capillaries are composed of a single layer of loosely connected lymphatic endothelial cells (LECs) (Baluk et al., 2007). Under low interstitial pressure these vessels are collapsed. As the interstitial pressure rises, swelling of the surrounding tissue leads to gaps forming between the LECs to allow entry of fluid and cells. The fluid drains into lymphatic collection vessels that have a continuous LEC lining. Lymphatic collection vessels are also sparsely covered with smooth muscle cells. In contrast to the blood supply the lymphatic system does not have a central pump i.e. the heart. Flow in the lymphatic system is driven by contraction of the vessels and skeletal muscle, and respiratory movements. Backflow of lymph is prevented by intraluminal valves. The lymph drains to two major collection vessels, the right lymphatic duct and the thoracic duct. These two vessels then return the fluid to the blood supply via connections with the right and left subclavian veins, respectively. The lymphatic system also contains a number of lymphoid organs. Primary lymphoid organs include the bone marrow and the thymus where differentiation and selection of lymphocytes occurs. Secondary lymphoid organs, including the lymph nodes, spleen and Peyer’s patches, provide an environment to initiate immune responses. Tertiary lymphoid organs form as accumulations of immune cells during chronic inflammation (Albrecht and Christofori, 2011). 16 | Page
The lymphatic system develops shortly after the separation of veins and arteries at E9.0. A subset of endothelial cells, in the interior cardinal vein, express Vegf receptor 3 (renamed FMS-like tyrosine kinase 4 (Flt4)) and the LEC marker lymphatic vessel endothelial hyaluronan receptor 1 (Lyve1) (Banerji et al., 1999). SRY-box containing gene 18 (SOX18), in conjunction with NR2F2, then induce prospero-related homeobox 1 (Prox1) expression (Francois et al., 2008). Prox1 is essential for lymphatic development (Wigle et al., 2002; Wigle and Oliver, 1999), driving venous differentiation to LECs. Conditional deletion of Prox1 can cause LECs to revert to a venous phenotype (Johnson et al., 2008). Thus venous/lymphatic endothelial fate is reversible and dependent on Prox1.
PROX1 and NR2F2 directly interact and induce genes required to generate and maintain LECs (Lee et al., 2009; Lin et al., 2010; Yamazaki et al., 2009). The specified LECs then bud off the cardinal vein at E10.5. The cells migrate and proliferate to form the primary lymphatic sacs. The primary lymphatic sacs then expand by proliferation and sprouting. The process of sprouting in the lymphatic system is not as well defined as the blood supply but does involve leading, migratory tip cells (Albrecht and Christofori, 2011). The major growth factor receptor for LEC guidance and proliferation is FLT4 which responds to VEGFC (Karkkainen et al., 2004). Both Notch1 and Notch4 are expressed in LECs and during angiogenesis in the blood supply (Shawber et al., 2007). Notch signalling can upregulate Efnb2 and Flt4 in the blood endothelial cells and a similar role in the lymphatic system seems likely (Grego-Bessa et al., 2007; Kume, 2010; Shawber et al., 2007).
1.2 The Notch pathway The core Notch pathway, called the canonical pathway, is deceptively simple. Notch receptors bind their ligands presented on neighbouring cells and through a series of proteolytic cleavages the Notch intracellular domain (NotchICD) is released and enters the nucleus. Here the NotchICD interacts with the DNA binding factor recombination signal binding protein for immunoglobulin kappa J region (RBPJ), also known as CSL (Cbf1/Su(H)/Lag1). The interaction between NotchICD and RBPJ displaces co-repressors and recruits transcriptional activators (Bray, 2006). Thus the Notch pathway is unusual compared to other signal transduction pathways. Both the ligand and receptor are type I transmembrane proteins, thus signalling occurs between cells in contact i.e. in trans. In addition there are no second messengers or kinase cascades between the receptor and nuclear transcription. One receptor on the surface can at most lead to one NotchICD molecule in the nucleus. In addition the receptor is consumed in the signalling process so the entry of one NotchICD 17 | Page
molecule into the nucleus means that there is at least one less receptor on the plasma membrane.
Although the effects of Notch signalling are extremely broad they can be broadly categorised into three groups (Bray, 2006);
1. Lateral inhibition: In a group of roughly equivalent cells initially small differences in Notch ligand and receptor levels are amplified through feedback loops resulting in the formation of two distinct cell populations. One cell adopts a specific cell fate and inhibits others from adopting the same fate.
2. Lineage decisions: Two daughter cells asymmetrically inherit Notch pathway components. One cell becomes a Notch receiving cell (activated Notch signalling) while the other becomes a ligand expressing signal sending cell.
3. Boundary formation/inductive signalling: Notch signalling occurs between two populations of cells and segregates them into two separate groups and keeps the two fields of cells separate.
1.2.1 Mammalian Notch Receptors
1.2.1.1 Evolution In the vertebrate lineage Notch receptors and ligands have undergone amplification. Mammals have four Notch receptors; Notch1 (Ellisen et al., 1991), Notch2 (Weinmaster et al., 1992), Notch3 (Lardelli et al., 1994) and Notch4 (Uyttendaele et al., 1996) and five Delta/Serrate/LAG-1 (DSL) ligands; Dll1 (Bettenhausen and Gossler, 1995), Dll3 (Dunwoodie et al., 1997), Dll4 (Shutter et al., 2000), Jag1 (Lindsell et al., 1995) and Jag2 (Shawber et al., 1996). There have been reports of additional ligands although these remain poorly described (Section 1.2.2.2). The ancestry of the mammalian Notch receptors has been largely deduced from sequence comparisons and the analysis of syntenic regions surrounding Notch genes. Syntenic regions surrounding the notch genes support a descent of Notch2, flanked by Brd, Pbx, Lhx and Camsap, from Notch1 flanked by Brd, Pbx, Lhx, Nrarp and Camsap. Notch3 is phylogenetically closer to Notch2 and is flanked by Brd, Pbx and Camsap supporting a model of Notch3 deriving from a duplication of Notch2 following the divergence of Notch1 and Notch2. Notch4 is by far the most divergent of the family and is only flanked only by Brd and Pbx family members (Theodosiou et al., 2009). Although there are four Notch homologues in fish, Notch1a and Notch1b clearly cluster with other Notch1 receptors and probably arose in a whole genome duplication near the teleost/tetrapod divergence (Kortschak et al., 2001). Indeed the expression domains of Notch1a and Notch1b in zebrafish additively correspond to the expression domain of
18 | Page
Notch1 in the mouse, supporting a process of complementation and degeneration (Westin and Lardelli, 1997). Notch4 homologues have not been found outside the mammalian group. The divergence of Notch4 from other family members is highlighted by the fact that in phylogenetic trees Notch4 forms a branch separate from the other paralogues even before the branching of Drosophila Notch. This suggests that either Notch4 was ancestral to Drosophila Notch and has been independently lost in every lineage except mammals or, more parsimoniously, that Notch4 has undergone a period of rapid divergence. Although Kortschak et al., 2001 found some phylogenetic relationship supporting descent of Notch4 from Notch3 using isolated domains, Theodosiou et al., 2009 found none. The highly divergent nature of Notch4 makes definitive conclusions difficult to justify.
1.2.1.2 Structure Notch receptors are single pass transmembrane receptors whose extracellular and intracellular domains perform distinct functions (Figure 1.4). The extracellular domain contains of 29-36 epidermal growth factor-like (EGF-like) repeats that are required for ligand binding (Rebay et al., 1991). The EGF-like repeats are followed by three LNR (Lin-12/Notch repeats) repeats and a membrane proximal segment that together form the negative regulatory region (NRR). The NRR inhibits ligand independent activation (Kopan et al., 1996). This region also includes a number of protease cleavage sites important for the regulation of Notch activity and formation of a heterodimer (Sections 1.2.3.2, 1.2.6 and 1.2.7). Following the membrane spanning peptide and stop translocation sequence is the Rbpj-associated molecule (RAM) region and Ankyrin repeat structure. These domains are involved in the formation of nuclear transcription complexes although they have been implicated in other interactions that are less well characterised (Section 1.2.10). Following the Ankyrin repeats is a poorly conserved C-terminal domain which in NOTCH1 and NOTCH2 was found to act as a transactivation domain when linked to the GAL4 DNA binding domain (Kurooka et al., 1998). In the same assay NOTCH3 and NOTCH4 demonstrated no transactivation activity. NOTCH3 was shown to have transactivation activity in this region but only on promoters with a proximal putative zinc finger transcription factor binding site (Ong et al., 2006). NOTCH4 is highly divergent in this region and no transactivation domain has been identified. At the C-terminal is a PEST motif (proline, serine, glutamate and threonine-rich) involved in protein degradation.
19 | Page
Figure 1.4 Structure of Notch Receptors
Domain structure of mammalian Notch receptors. NOTCH1 and NOTCH2 contain 36 EGF-Like repeats, NOTCH3 34 EGF-Like repeats and NOTCH4 29 EGF-like repeats. These are followed by three LNR domains and the transmembrane region. On the intracellular side there is a RAM domain which contains a nuclear localisation signal (NLS). This is followed by seven Ankyrin repeats (ANK). In NOTCH1, NOTCH2 and NOTCH3 this is followed by a second NLS. NOTCH1 and NOTCH2 have a transactivation domain C-terminal to the Ankyrin repeats and all four receptors have a C-terminal PEST sequence.
EGF-like repeat-Epidermal growth factor-like repeat, LNR- Lin- 12/Notch repeats, RAM- Rbpj-associated molecule, NLS- Nuclear localisation sequence, ANK- Ankyrin repeat, TAD- Transactivation domain, PEST- Proline, glutamine, serine and threonine rich.
20 | Page
1.2.1.3 Vascular expression of the Notch receptors Notch1 is widely expressed and plays vital roles in a large number of developmental and physiological processes. Notch1 is expressed broadly in most tissues including brain, liver, heart, lung, kidney, intestine, bone marrow, skeletal muscle, spinal cord, eye and thymus (Swiatek et al., 1994). Targeted deletion of Notch1 is embryonic lethal with wide spread cell death and impaired somitogenesis, neurogenesis and angiogenesis (Conlon et al., 1995; Krebs et al., 2000; Swiatek et al., 1994). Importantly Notch1-/- mice die mid-gestation (E9.5-10.5) from massive defects in the vasculature. In these mice the major vessels form indicating that Notch1 deletion causes developmental arrest during angiogenesis rather than vasculogenesis. NOTCH1 has been identified as the major Notch receptor driving endothelial tip cell selection and arterial fate (Krebs et al., 2000).
Notch2 is expressed predominantly in the brain, liver, kidney and stomach and partially overlaps with Notch1 expression (McCright et al., 2001). Targeted deletion of Notch2, like Notch1, is embryonic lethal with cell death in neural tissues as well as cardiovascular and kidney defects (Hamada et al., 1999; McCright et al., 2001) NOTCH2 is not thought to play a role in vascular development although it has been implicated in vascular regression (McCright et al., 2001).
In contrast to Notch1 and Notch2, Notch3 expression is more restricted. Notch3 is expressed predominantly in vascular smooth muscle (Joutel et al., 2000) but is also in the central nervous system and in certain hematopoietic cells (Lardelli et al., 1994). Notch3 null mice exhibit marked defects in vascular smooth muscle cell maturation with enlarged arteries and abnormal distribution of elastic laminae (Domenga et al., 2004). NOTCH3 signalling controls smooth muscle cell recruitment, maturation and maintenance (Section 1.1.2.4).
Mice homozygous for Notch4tm1Grid, referred to as Notch4d1, (considered to be a null allele) are viable, fertile and display no overt phenotype. However a genetic interaction was observed when both the Notch4 and Notch1 targeted alleles were homozygous; 50% of the embryos displaying a more severe vascular phenotype than Notch1 deletion alone (Krebs et al., 2000). Mice with either Notch4 gain- or loss-of-function are discussed in Chapter 5. Notch4 expression was initially reported as endothelial specific (Shirayoshi et al., 1997; Uyttendaele et al., 1996). Notch4 expression was detected as early as E7.5 using PCR based methods. However, these transcripts could not be reliably detected by in situ methods. At E8.5 Notch4 can be detected in endothelial cells as well as the cephalic mesenchyme
21 | Page
which may indicate angioblast expression (Shirayoshi et al., 1997). In the developing vasculature Notch4 expression is restricted to the arterial endothelium and developing capillaries although it is also expressed in the lymphatic system (Shawber et al., 2007; Villa et al., 2001). The endothelial expression and timing of Notch4 expression supports a role for Notch4 in angiogenesis and maintenance of arterial identity and a potential role in lymphangiogenesis.
Although Notch4 is often referred to as an endothelial specific receptor it is expressed outside the endothelium. Notch4 has been detected in the breast (Raafat et al., 2011), the haematopoietic system (Sekine et al., 2009; Singh et al., 2000; Yuan et al., 2010), kidney (Cummins et al., 2011) and the neural tube between E9- 11.5 although expression was absent in the nervous system by E13.5 (Uyttendaele et al., 1996). However, what role it may play in these tissues has not been well established. Notch4 has been linked to schizophrenia in humans although this association is disputed (Ivo et al., 2006; McGinnis et al., 2001; Skol et al., 2003).
1.2.2 Notch ligands
1.2.2.1 DSL ligands The best characterised Notch ligands belong to the DSL family (named for Delta and Serrate in Drosophila and Lag-2 from C.elegans). DSL ligands are single pass membrane proteins with an N-terminal domain with limited conservation. This is followed by the DSL domain and a number of EGF-like repeats (Figure 1.5). The N-terminal domain, DSL and the first two EGF-like repeats form the receptor binding region (Parks et al., 2006; Shimizu et al., 1999). The Serrate group of ligands, JAG1 and JAG2 in vertebrates, also contain a cysteine rich region between the EGF- like repeats and the membrane which has homology to the von Willebrand factor type C domain (Vitt et al., 2001). The intracellular domain of DSL ligands are poorly conserved although they do contain, with the exception of DLL3, multiple lysine residues that are required for signalling activity through interactions with the endocytic machinery (D'Souza et al., 2008).
The major DSL ligands in vascular development are DLL4 and JAG1. The vital role of Dll4 in angiogenesis and in particular tip cell selection is well established (Section 1.1.2.1). Jag1 plays important roles in smooth muscle recruitment, maturation and maintenance (Section 1.1.2.4). Although Dll1 deletion does not lead to defects in embryonic vascular development a role in adult angiogenesis after ischemia has been identified (Limbourg et al., 2007). In mature tissues recovering from injury Dll1 may play a similar role to Dll4 in development. The expression pattern of both 22 | Page
Dll4 and Jag1 coincides with that of Notch4 supporting a role for these ligands in Notch4 activation. In addition Notch4 and Dll4 are the only receptor and ligand detected in developing capillaries (Villa et al., 2001).
Figure 1.5 Structure of DSL ligands
Domain structure of DSL ligands. All DSL ligands, with the exception of DLL3 have a N-terminal DSL domain followed by a variable number of EGF-like repeats, 8 in DLL1 and DLL4, 6 in DLL3 and 16 in JAG1 and JAG2. In JAG1 and JAG2 this is followed by VWB domain. Following the transmembrane region, all DSL ligands, again with the exception of DLL3 have a cytoplasmic PDZ domain.
EGF-like repeat-Epidermal growth factor-like repeat, DSL- Delta/Serrate/LAG-1, VWB- von Willebrand factor type C domain, PDZ-post synaptic density protein, Drosophila disc large tumour suppressor and zonula occludens1 protein domain.
23 | Page
1.2.2.2 Non-DSL ligands There have been a number of Notch ligands identified in addition to DSL family members. These ligands are referred to as non- canonical ligands and are relatively poorly characterised in comparison to DSL ligands. None of the non-canonical ligands are essential for embryonic development and their role appears to be as modulators of signal output. One of the first non-canonical ligands reported was Delta-like 1 homolog (Dlk1) (Laborda et al., 1993; Smas and Sul, 1993). DLK1 was reported to be a cis- inhibitor of Notch signalling (Baladron et al., 2005; Nueda et al., 2007) (Section 1.2.9). However, alternative spliced isoforms are soluble and may bind without activating Notch and thus prevent productive interaction with cell surface DSL ligands. Another complication of Dlk1 is the closely related gene Dlk2, which is reciprocally regulated (Sanchez-Solana et al., 2011). DLK1 and DLK2 can directly interact with each other. In addition both can interact with NOTCH1. Co-expression of DLK1 and DLK2 can lead to reciprocal inhibition and thus Notch activation. A role for Dlk1 in regulating tip cell selection in the murine retina has been identified (Rodriguez et al., 2012).
Delta/Notch-like EGF-related receptor (DNER) is another potential Notch ligand that may play a role in glial cell development (Eiraku et al., 2005). Its role is poorly understood but DNER appears to bind surface Notch on adjacent cells to activate signalling (Eiraku et al., 2002; Eiraku et al., 2005). Contactin1 (CNTN1) and Contactin6 (CNTN6) have also been reported as activators of Notch signalling and may play a role in late oligodendrocyte differentiation (Cui et al., 2004; Hu et al., 2004). There are a number of proteins identified that can act to enhance Notch DSL dependent signalling. CCN3 (renamed nephroblastoma over expressed gene (NOV)) potentiates Notch signalling when expressed in the same cell i.e. in cis (Minamizato et al., 2007; Sakamoto et al., 2002b). Microfibrillar-associated protein 1 and 2 (Mfap1 and Mfap2) have also been shown to enhance activation of Notch when expressed in cis (Miyamoto et al., 2006). The mechanism of action is thought to involve destabilisation of the Notch heterodimer. However, an inhibitory action for Magp2 has been identified in some cell types (Albig et al., 2008). Two additional enhancers of Notch activation have been identified that interact with NOTCH3. Tumour suppressor region 2 (Tsp2) enhances activation when expressed in cis or as a soluble protein (Meng et al., 2009). The interaction is thought to occur at the cell surface. A closely related protein, TSP1, can also interact with NOTCH3 but does not enhance signalling. Y box protein 1 (YBX1 can enhance NOTCH3 activation through a direct interaction but does not interact with NOTCH1 (Rauen et al., 2009).
24 | Page
A secreted potential ligand, EGF-like domain 7 (EGFL7), directly interacts with all four mammalian Notch receptors and is expressed in a similar pattern to Notch4 i.e. in the arterial vascular endothelium (Schmidt et al., 2009). EGFL7 has been reported to contain a DSL domain. However, the domain is highly divergent and lacks the characteristic N-terminal cysteines (Fitch et al., 2004). EGFL7 acts as both a Notch agonist and antagonist (Nichol et al., 2010; Schmidt et al., 2009). EGFL7 appears to compete for binding with JAG1 but may enhance DLL4 signalling by interacting directly with DLL4. The differential effects may depend on a number of factors. EGFL7 may prevent binding of Fringe modified Notch receptors to JAG1 thus allowing increased access to DLL4 (Section 1.2.3.1)(Nichol and Stuhlmann, 2012). In addition, EGFL7 can bind the extracellular matrix and its activity may be dependent on whether it is soluble or matrix bound. The mechanism by which EGFL7 modulates Notch signalling remains poorly characterised.
1.2.3 The canonical Notch pathway The best established mode of Notch function is described as the canonical pathway. The canonical pathway describes Notch receptor activation via trans interactions with DSL ligands. Ligand activation of Notch leads to a series of proteolytic cleavages of the Notch receptor. The NotchICD is released into the cytoplasm and is translocated to the nucleus. There NotchICD binds RBPJ and acts as a transcriptional activator (Kopan and Ilagan, 2009) (Figure 1.6).
25 | Page
Figure 1.6 Canonical Notch Signalling
The Notch receptor is translated as a full length peptide which undergoes extensive glycosylation in the ER (a.) (Section 1.2.3.1). After passage through the ER the glycosylation is further modified by Fringe in the Golgi (b.) (Section 1.2.3.1) and the receptor is cleaved by FURIN, referred to as S1 in the TGN (b.) (Section 1.2.3.2). The N- and C-terminals remain non-covalently bound, referred to as the Notch heterodimer and are presented on the cell surface (c.) (Section 1.2.3.2). In the absence of ligand the receptor is endocytosed and either recycled to the plasma membrane (d.) or degraded (e.) (Section 1.2.4). Ligand binding (f.) and associated endocytosis (g.) disassociate the heterodimer (Section 1.2.5 and 1.2.6). The transmembrane and intracellular domain (TMIC) is then cleaved a second time, referred to as S2 (h.) (1.2.6) to form Notch extracellular truncated (NEXT). NEXT is then cleaved twice, S3 and S4, within the membrane by gamma- secretase (i.) (Section 1.2.7) to generate the Notch intracellular d omain (NotchICD). The NotchICD translocates to the nucleus and activates gene expression via RBPJ bound promoters (j.) (Section 1.2.8). The NotchICD is then rapidly degraded (k.).
ER- endoplasmic reticulum, TMIC - Transmembrane and intracellular domains, NEXT- Notch extracellular truncated, NotchICD - Notch intracellular domains, TGN – trans Golgi network
26 | Page
27 | Page
1.2.3.1 Receptor maturation Notch receptors undergo a large number of post translational modifications in the secretory pathway which include both N - and O -linked glycosylation (Figure 1.6a). Notch is initially produced as a single peptide, which undergoes an unusual type of glycosylation involving the addition of o-fucose to serine and threonine residues that occur before the 3rd cysteine of the EGF-like repeats (Okajima and Irvine, 2002) (Figure 1.7a). This relatively loose 2 3 2 3 consensus site (C X 3-5S/TC ) where C and C are the 2nd and 3rd cysteines of the EGF-like repeat, is present in 23, 20, 15 and 17 of the EGF-like repeats of NOTCH1, NOTCH2, NOTCH3 and NOTCH4, respectively (Harris and Spellman, 1993; Panin et al., 1997; Wang and Spellman, 1998) (Figure 1.7b). These have been best characterised in Drosophila NOTCH and mammalian NOTCH1. Not all of the repeats are O -fucosylated but multiple repeats are (Moloney et al., 2000; Panin et al., 1997). This activity is due the action of protein O -fucosyltransferase 1 (POFUT1), which is localised in the endoplasmic reticulum (ER) (Moloney et al., 2000). Mice null for Pofut1 show profound developmental defects that are reminiscent of those seen in Presenilin 1 and 2 (Psen1 and Psen2- components of the gamma-secretase complex) (Section 1.2.7) combined knockout and the Rbpj knockout, both of which display severely disrupted or absent Notch signalling (Okamura and Saga, 2008; Shi and Stanley, 2003). In Pofut1 null mice NOTCH1 fails to traffic to the cell surface and accumulates in the ER. Notch trafficking was normal in Drosophila carrying Pofut1 mutants lacking fucosyltransferase activity, which has led to the hypothesis that POFUT1 also acts as a chaperone to ensure the correct folding of the Notch extracellular domain prior to export to the Golgi (Okajima et al., 2005).
In addition to O -fucosylation EGF-like repeats of Notch are also O - glucosylated (Figure 1.7a). O -glucosylation occurs at the consensus site C1 XSXPC2 (Jafar-Nejad et al., 2010). The O - glucosylation consensus sequence is present in 16, 17, 14 and 11 EGF-like repeats in NOTCH1, NOTCH2, NOTCH3 and NOTCH4, respectively (Figure 1.7b). The enzymatic activity required for O - glucosylation has been identified in Drosophila as Rumi (Acar et al., 2008). Drosophila Rumi mutants display a temperature sensitive loss of Notch signalling. The temperature sensitive phenotype is present even in mutants where ~95% of the coding region of Rumi has been deleted (Jafar-Nejad et al., 2010). These results suggest that it is the Notch receptor that is temperature sensitive in Rumi mutants rather than Rumi itself. The surface presentation of NOTCH is unaffected by the temperature which suggests that O -glucosylation serves to stabilise a ligand receptive conformation of the NOTCH receptor. Although
28 | Page
homologues of Rumi have been identified in mammals biochemical data is lacking.
After passage through the ER, Notch is further modified in the Golgi. The O -linked fucose sites are substrates for β 1,3-N - acetylglucosaminyltransferases encoded by Fringe genes (Moloney et al., 2000) (Figure 1.6b). Modification by Fringe has been shown to enhance Delta and inhibit Serrate induced Notch signalling both in vivo and in cell based assays (Chen et al., 2001; Evrard et al., 1998; Hicks et al., 2000; Moloney et al., 2000; Shimizu et al., 2001; Yang et al., 2005; Zhang and Gridley, 1998). Fringe glycosylation increases Delta-Notch binding thereby potentiating Delta dependent Notch signalling. However, Jagged-Notch binding is not affected by Fringe activity even though Jagged dependent Notch signalling is reduced by Fringe. This suggests that Jagged binds but this interaction is non-productive (Yang et al., 2005). This result may explain the observation that JAG1 can inhibit DLL4 by competing for ligand but not producing a signal (Benedito et al., 2009). The Fringe modification is present on multiple EGF-like repeats. In an extensive mutational analysis it was demonstrated that no single modification mediates Fringe’s affect on Notch signalling (Xu et al., 2005).
In mammals there are three Fringe genes, Manic Fringe (Mfng), Lunatic Fringe (Lfng) and Radical Fringe (Rfng). Although all three recognise the same sites there have been reports that have noted differences (Rampal et al., 2005). Galactose addition to O-fucose glycans was found necessary to augment DLL1 mediated signalling to LFNG but not MFNG modified NOTCH1. LFNG and MFNG suppressed Jagged dependent signalling but RFNG was found to increase both Delta and Jagged dependent signalling (Benedito et al., 2009). However, in mice null for either Rfng (Moran et al., 1999) or Mfng no overt phenotype was observed and double mutants of Rfng and Lfng (Zhang et al., 2002) and triple mutants of Rfng, Mfng and Lfng showed no defects other than those associated with loss of Lfng (Moran et al., 2009). Mice null for Lfng have skeletal defects that result in defects in somite patterning (Zhang and Gridley, 1998) and skeletal defects have also been observed in humans carrying mutations in LFNG (Dunwoodie, 2009; Sparrow et al., 2006). The complex glycosylation patterns of Notch receptors and their affects remain an active area of study.
29 | Page
Figure 1.7 Glycosylation of Notch receptors
Reproduced and adapted with permission from (Jafar-Nejad et al., 2010). a. Diagram of an EGF-like repeat based on the crystal structure of one of the EGF-like repeats of human coagulation factor IX (Huang et al., 1989). The sites of O-glucosylation (blue) and O- fucosylation (red) are indicated by arrows. Green lines indicate the pattern of disulfide bonding between the six cysteine residues of the EGF-like repeat. b. Diagram of the extracellular regions of human Notch receptors. EGF-like repeats with consensus O -fucosylation (yellow boxes), O - glucosylation (blue boxes) and calcium binding (red lines) motifs are shown. LNR-Lin-12/Notch repeat, HDD-heterodimerisation domain, TM-transmembrane domain.
30 | Page
31 | Page
1.2.3.2 FURIN and the regulation of surface expression A number of proteolytically processed forms of Notch receptors can be detected in both cell lines and in vivo (Blaumueller et al., 1997). The cleavages are referred to as S1, S2, and S3 and 4 and are performed as part of the activation process. The first cleavage, S1, is performed by a Furin-like convertase and occurs prior to ligand binding in the trans-Golgi network (Blaumueller et al., 1997; Jarriault et al., 1998; Logeat et al., 1998) (Figure 1.6b). The role of S1 cleavage and its significance for signalling (Section 1.2.6) remains controversial and results differ between Notch family members. The best characterised Notch family member in terms of S1 cleavage is mammalian NOTCH1. FURIN cleaves proteins that contain the consensus sequence RXR/KR although a broader consensus of an R at P1 and at least two other basic residues at P2, P4 or P6 has been identified (Henrich et al., 2005; Nakayama, 1997). In the case of mammalian NOTCH1, S1 cleavage is a requirement for surface expression (Logeat et al., 1998) (Figure 1.6c). Following S1 cleavage the N- and C-terminal parts of the protein remain non-covalently associated and is referred to as the Notch heterodimer. The association of the N- and C-termini is calcium dependent and can be disrupted with EDTA (Rand et al., 2000). NOTCH1 is predominantly cleaved at RQRR at position 1654. Mutation of this site leads to the use of two additional dibasic sites nearby. Only when all three sites are mutated is S1 cleavage abolished (Logeat et al., 1998). Mutation of all three sites causes a complete loss of signalling and intracellular accumulation of NOTCH1. There have been conflicting results suggesting full length NOTCH1 can be found on the cell surface (Bush et al., 2001). However, over 95% of the surface NOTCH1 was found in the heterodimeric form. In addition inhibition of FURIN activity led to an increase in unprocessed NOTCH1 found on the cell surface but was unable to be activated. This result suggested that the formation of the heterodimer is not only part of surface presentation but also has important consequences for ligand dependent activation (Section 1.2.6). S1 cleavage of NOTCH1 was thought to be a constitutive process and part of normal receptor maturation. However, recent evidence has shown that S1 cleavage is a regulated process. FURIN processing of NOTCH1 is induced by growth factor stimulation and a direct interaction with Rous sarcoma oncogene (SRC) (Ma et al., 2012). Additionally, a direct interaction between cryptic family 1 (CFC1) and the last two EGF-like repeats of NOTCH1 has been shown to recruit FURIN and augment NOTCH1 processing (Watanabe et al., 2009). FURIN processing is also under negative control by cation transport regulator 1 (Chac1 also known as blocks Notch (Botch)) (Chi et al., 2012). CHAC1 binds to the S1 site of NOTCH1 and prevents S1 cleavage, surface presentation and signalling.
32 | Page
The situation in Drosophila NOTCH is more controversial. One group has reported that FURIN does not cleave Drosophila NOTCH and full length NOTCH is readily detected on the cell surface (Kidd and Lieber, 2002). However, the authors identified FURIN cleavage sites and products in deuterostomes and C.elegans leading them to conclude that the lack of FURIN cleavage in Drosophila is a derived rather than ancestral characteristic. In complete contrast other groups have identified S1 processing in Drosophila NOTCH occurring in RLKK at position 1637 (Lake et al., 2009). However, they were unable to inhibit processing with FURIN inhibitors and could not rule out alternative proteases. When the S1 site of Drosophila NOTCH was mutated it led to intracellular accumulation and lack of surface expression as has been noted in some NOTCH1 mutants (Logeat et al., 1998). However the residues mutated, based on the structure of mammalian NOTCH1 and NOTCH2 (Gordon et al., 2007; Gordon et al., 2009), may play important structural roles and a structural defect in these mutants may explain the intracellular accumulation.
Variation in S1 cleavage is also observed amongst mammalian Notch paralogues. In a crystal structure of the NRR of NOTCH1 and NOTCH2 FURIN cleavage was identified to occur in an unstructured loop (Gordon et al., 2007; Gordon et al., 2009). Consensus FURIN cleavage sites can be identified in NOTCH1, NOTCH2 and NOTCH3. In contrast NOTCH4 contains no identifiable FURIN cleavage site (see Chapter 3). The majority of cell surface NOTCH2 is S1 cleaved. In contrast to NOTCH1, mutation of the S1 cleavage site in NOTCH2 does not lead to reduced surface expression or an inability to signal (Gordon et al., 2009). Thus S1 cleavage is regulated in a paralogue specific manner.
1.2.4 Receptor recycling and degradation The canonical Notch pathway is sensitive to dosage and timing of signalling and requires fine tuning of the response (Bray, 2006). The level of surface expression of the receptor, and thus its ability to bind ligand, provides a mechanism for controlling signalling output. In the absence of ligand surface Notch is constitutively endocytosed and either recycled to the plasma membrane (Figure 1.6d) or trafficked to the late endosome for subsequent lysosomal degradation (Chastagner et al., 2008; Jehn et al., 2002; Mcgill et al., 2009) (Figure 1.6e). Endocytosed proteins are transported to the early endosome. Proteins are sorted in the early endosome to the recycling endosome for return to the plasma membrane, transport to the trans Golgi network or retained as the endosome matures. The endosome matures to form the late endosome and through a series of fission and fusion events forms acidified lysosomes where the cargo of proteins is degraded. 33 | Page
Regulation of endosomal sorting of Drosophila NOTCH is dependent on Nedd4 family members such as suppressor of deltex ( Su(dx)) (Sakata et al., 2004; Wilkin et al., 2004). Su(dx) is an E3 ubiquitin ligase that ubiquitylates Notch and targets it for lysosomal degradation. Su(dx) interacts with the PPSY motif in Drosophila NOTCH (Sakata et al., 2004). In mammals the crucial tyrosine residue is absent and the mouse homologue of Su(dx), E3 ubiquitin protein ligase (Itch), may interact and form different complexes with Notch (Chastagner et al., 2008). Thus extrapolations of findings in Drosophila need to be made with caution with respect to mammalian Notch receptors. Itch was initially described in the mutant Itchy mice that develop a progressive immune disorder (Perry et al., 1998). This disorder can be phenocopied in transgenic mice with increased Notch signalling (Matesic et al., 2006). Ligand mediated receptor activation is normal in Itch-/- fibroblasts but ITCH facilitates degradation of the receptor in the absence of ligand (Chastagner et al., 2008). ITCH poly ubiquitylates NOTCH1 via K29-linked chains, which target NOTCH1 for lysosomal degradation. Poly ubiquitylation and thus degradation of Notch is enhanced by NUMB (McGill and McGlade, 2003). NUMB is associated with membrane bound Notch and can recruit ITCH to enhance degradation (Mcgill et al., 2009). Regulation of surface expression of Notch receptors is undoubtedly more complex. Other players include additional E3 ubiquitin ligases such as Casitas B-lineage lymphoma (CBL) which has also been implicated in lysosomal degradation of Notch (Jehn et al., 2002). The endocytic sorting of Notch remains an active area of research and has multiple levels of regulation.
1.2.5 Ligand binding and receptor activation Cell aggregation assays have demonstrated that EGF-like repeats 11 and 12 of Notch constitute the minimal region required for binding to Delta and Serrate family members (Rebay et al., 1991). The importance of repeats 11 and 12 has been confirmed by many groups (de Celis et al., 1993; Lawrence et al., 2000; Lei et al., 2003; Shimizu et al., 1999) but only a few studies have investigated how other repeats modulate ligand binding (Pei and Baker, 2008; Xu et al., 2005). The importance of the other repeats is highlighted by the observation that mutations in EGF-like repeat 12 that cannot be O -fucosylated, still differentially respond to Delta and Serrate following Fringe modification (Lei et al., 2003). Although disparate parts of Notch are involved in the Fringe modification of signalling, swapping the N-terminus of Serrate for Delta is sufficient to stop the inhibition by Fringe (Fleming et al., 1997), which suggests that elements of Notch that are widely separated by primary sequence contribute to ligand binding.
34 | Page
A possible explanation for this comes from a proposed model for the quaternary structure of Notch (Xu et al., 2005). One piece of evidence that supports this model is from a structural analysis of EGF-like repeats 11-13. EGF-like repeat 13 was included to correctly fold EGF-like repeats 11 and 12 and this construct could bind ligand when complexed with calcium. The calcium bound form adopts an elongated rod-like structure (Hambleton et al., 2004). The spacing of calcium binding domains in Notch receptors is well conserved and led to the suggestion that calcium containing EGF- like repeats 6-9, 10-21 and 22-26 could form rigid rods separated by more flexible hinge regions (Xu et al., 2005). In addition, in a solid phase binding assay, an association was found between Notch EGF-like repeats 11-20 and 21-30 (Pei and Baker, 2008). In this speculative model EGF-like repeats 6-9 and 22-26 fold back into close contact with the ligand binding region within EGF-like repeats 10-21 and glycosylation could affect the flexibility of the hinge regions (Gordon et al., 2008a). Ligand binding would require disruption of these structures. This raises the intriguing possibility that extracellular calcium concentration may modulate Notch activity by affecting this quaternary structure. In a co-culture assay extracellular calcium was found to enhance Notch responsiveness to ligand and may play a role in modulating Notch signalling in left/right axis determination where a calcium gradient is established (Raya et al., 2004).
1.2.6 Receptor activation, S2 cleavage The Notch receptor is maintained in an inactive state in the absence of ligand. Notch receptors that contain relatively short extracellular regions are constitutively active and signal in the absence of ligand. However, a construct that has the EGF-like repeats removed but contains 277 amino acids of the extracellular domain of Notch is inactive and unable to respond to ligand (Kopan et al., 1994; Kopan et al., 1996). This region constitutes the negative regulatory region (NRR) of Notch. The structure of this region is calcium dependent and Notch signalling can be induced by treatment with EDTA which disrupts the structure (Rand et al., 2000). Additional support for the importance of this region is the identification of hyperactive Notch mutants which cluster in this region in Drosophila and C.elegans (Greenwald, 1998; Kimble et al., 1998) and in Notch1 mutations in T-cell acute lymphoblastic leukaemia (T-ALL) (Gordon et al., 2008b). S2 cleavage is required to produce the Notch extracellular truncated (NEXT) C-terminal peptide before further processing can lead to receptor activation (Section 1.2.6) (Figure 1.5h). S2 cleavage occurs at V1711 and takes place at the plasma membrane (Lieber et al., 2002; Mumm et al., 2000). In Drosophila this activity was attributed to Kuz, the Drosophila homologue of mammalian a d isintegrin and metallopeptidase domain 10 (Adam10) (Jarriault et 35 | Page
al., 1998; Pan and Rubin, 1997; Rooke et al., 1996; Sotillos et al., 1997). However, in mammals the protease activity responsible for S2 cleavage was purified from the plasma membrane but did not co-purify with ADAM10 (Brou et al., 2000). The activity was found to be due to a closely related family member ADAM17. However, genetic deletion of Adam17 results in embryonic death at E17.5 or post birth in contrast to Notch1 mutants which die mid-gestation at E9.5-10 (Krebs et al., 2000; Peschon et al., 1998). These results indicate that there may be a partial functional redundancy between Adam family members. However, the role of ADAM10 or ADAM17 in Notch signalling remains controversial and cell culture studies have suggested that ADAM17 cannot compensate for ADAM10 activity associated with Notch activation (Bozkulak and Weinmaster, 2009).
The inability of S2 cleavage to take place in the absence of ligand is explained by the crystal structure of the NRR region (Gordon et al., 2007). The NRR region is composed of three Lin-12/Notch repeats (LNR) and the heterodimerisation domain (HDD). The LNR domains wrap around the HDD burying the S2 cleavage site in addition to a total solvent accessible surface area of approximately 3000Å2 . This extensive interaction needs to be disrupted to uncover the S2 site. First the LNR-A domain, followed by the LNRA-B linker, then the LNR-B domain and finally LNR-C need to be removed before S2 cleavage can take place. Exposure of the S2 site thus requires substantial force and a strong ligand/receptor interaction (Gordon et al., 2007). The strength of the Notch ligand/receptor interaction is supported by measurements of cell adhesion strength using atomic force microscopy (Ahimou et al., 2004). Ligand induced signalling requires endocytosis in both the ligand and receptor presenting cell (Seugnet et al., 1997). Endocytosis after ligand binding is thought to provide the force necessary to disrupt the NRR and thus expose the S2 site (Parks et al., 2000) (Figure 1.6g). Binding of Notch produces a resistance to ligand endocytosis which leads to ubiquitylation of the ligand and recruitment of epsins. Epsins act as endocytic adaptors involved in clathrin-mediated endocytosis in the ligand cell to provide the pulling force to disrupt the Notch NRR (Meloty-Kapella et al., 2012).
NOTCH1 receptors that are not FURIN processed can be found on the cell surface but are unable to signal (Section 1.2.3.2). This led to the suggestion that disassociation of the heterodimer was required prior to S2 cleavage. The N-terminal portion of the NOTCH1 heterodimer can be detected within the ligand cell and this translocation was not inhibited by inhibiting S2 cleavage (Nichols et al., 2007) (Figure 1.6g). These results suggest that dissociation does take place before S2 cleavage. However, 36 | Page
NOTCH2 receptors that have not been S1 cleaved can still signal (Gordon et al., 2009). The requirement for disassociation of the heterodimer prior to S2 cleavage remains controversial and may differ between homologues and/or cell type.
1.2.7 Receptor activation, S3 and S4 cleavage Following S2 cleavage the NEXT peptide is a constitutive substrate for the gamma-secretase complex (Saxena et al., 2001; Schroeter et al., 1998). The exact cellular location of gamma- secretase cleavage remains controversial (Figure 1.6i). Gamma- secretase is found at the cell surface as well as in endosomal compartments. Indirect evidence that gamma-secretase cleaves in intracellular compartments comes from the observation that factors that promote internalisation and entry into the early endosome are required for productive signalling and endocytosis enhances gamma-secretase cleavage (Gupta-Rossi et al., 2004; Lu and Bilder, 2005; Vaccari et al., 2008). However, other groups disagree and have shown productive gamma-secretase cleavage to occur at the plasma membrane (Sorensen and Conner, 2010). The conflicting results may reflect differences in experimental protocols, in particular the Notch constructs analysed. In addition, gamma-secretase may productively cleave Notch at multiple subcellular locations. The gamma secretase complex predominantly cleaves the NEXT peptide at a valine residue just inside the membrane, termed S3, and an additional N-terminal intra-membrane cleavage S4 (Saxena et al., 2001; Schroeter et al., 1998). Although the cleavage of NEXT is constitutive there is evidence that regulation of cleavage can occur. Initially the gamma-secretase cleavage site of NOTCH1 was identified to occur at V1744 and mutation of this site severely reduced NotchICD production and signalling (Saxena et al., 2001; Schroeter et al., 1998; Tagami et al., 2007). However, using more sensitive techniques, gamma-secretase cleavage was shown to occur at multiple sites. These alternative forms of the NOTCH1ICD are unstable and have a reduced capacity to signal. The alternative forms were generated preferentially by endosomal compared to plasma membrane preparations and the production of these forms was proportional to the rate of endocytosis (Tagami et al., 2007). The cleavage site preference of gamma-secretase may be affected by factors such as membrane curvature and pH that is lowered in late endosomal compartments (Fukumori et al., 2006). Control of alternative gamma-secretase cleavage sites is an ongoing and active area of research due their importance in Alzheimers disease (Haapasalo and Kovacs, 2011). These results suggest that NOTCH1ICD produced at the plasma membrane preferentially generates active signalling molecules while NEXT cleaved intracellularly is preferentially targeted for degradation. Cleavage by gamma-secretase releases the NotchICD into the cytoplasm. 37 | Page
Nuclear localisation sequences in the NotchICD then target the NotchICD to the nucleus.
1.2.8 Transcription complex Canonical Notch signalling acts via the DNA binding factor RBPJ (Figure 1.6j). RBPJ binds to the consensus sequence GTGGGAA, although variants of this sequence such as CTGAGAA are also bound with high affinity (Lam and Bresnick, 1998; Wu and Bresnick, 2007a). Potential binding sites for RBPJ are plentiful in the mammalian genome. The best characterised direct targets of Notch signalling are members of the hairy and enhancer of split (Hes) and hairy/enhancer-of-split related with YRPW motif (Hey) families of basic helix loop helix transcription factors. Hes and Hey transcription factors are involved in a wide variety of developmental and physiological processes (Fischer and Gessler, 2007; Kageyama et al., 2007).Their main functions have been identified as regulating binary cell fate decisions and in particular the prevention of differentiation of progenitor cells. In addition, by self regulating their own expression, their expression can oscillate and form biological clocks important in processes such as somitogenesis (Kageyama et al., 2007; Niwa et al., 2007).
In the absence of Notch signalling RBPJ acts as a transcriptional repressor. RBPJ interacts with a variety proteins and recruits multiprotein repressor complexes such as histone deacetylase 1 and 2 (HDAC1 and HDAC2), SIN3A, transducin-like enhancer of split (TLE) and C-terminal binding protein (CTBP) (Lai, 2002). RBPJ contains a three domain conserved core which consists of the N-terminal domain (NTD), the beta-trefoil domain (BTD) and a C-terminal domain (Kovall and Hendrickson, 2004). NotchICD interacts with RBPJ through the RAM domain and Ankyrin repeats. Analysis of the RAM domain using circular dichroism showed that in the absence of bound RBPJ, the RAM domain has very little or no secondary structure (Bertagna et al., 2008; Nam et al., 2003). Upon interaction with RBPJ the RAM domain adopts a rigid ordered structure. Only the first 20 residues of the RAM domain interact with RBPJ. The following 70 amino acids do not form a structure in the complex and are poorly conserved amongst Notch receptors. However, although the primary sequence is not well conserved the number of residues is. In a polymer chain model of these residues the distance spanned is approximately 50Å (Bertagna et al., 2008). This distance accurately corresponds to the distance between the RAM domain and the Ankyrin repeats in the bound conformation. Using this modelling approach an effective concentration of the Ankyrin repeats can be calculated after RAM domain binding. The effective concentration of approximately of 0.5mM allows even low affinity interactions to take place once RBPJ is bound by the RAM domain. After binding of the NotchICD to RBPJ, a beta hairpin loop 38 | Page
in the NTD of RBPJ adopts an open confirmation (Friedmann et al., 2008). The open confirmation creates a binding site for the protein mastermind (MAML). MAML binds both RBPJ and the NotchICD to form the ternary complex. The ternary complex displaces repressor complexes and recruits histone acetyltransferases, such as K(lysine) acetyltransferase 2B (KAT2B) (also known as p300/CBP associated factor), leading to acetylation of chromatin and gene transcription (Fryer et al., 2002; Kurooka and Honjo, 2000; Wallberg et al., 2002). However, it has been suggested that the interaction between RBPJ and the DNA binding site may be more dynamic than the above model implies. Notch activation transiently leads to increased promoter occupancy by RBPJ (Krejci and Bray, 2007). This may be due to the formation of a more stable complex and longer residency time when NotchICD is present.
In addition to recruiting transcriptional activators MAML also recruits proteins involved in the degradation of the NotchICD (Figure 1.6k). MAML promotes the hyper-phosphorylation and subsequent degradation of the NotchICD (Fryer et al., 2002). Notch signalling is used reiteratively and a rapid turnover of the signalling complex is essential for this to occur (Fior and Henrique, 2009). The cyclin C (CCNC):cyclin dependent kinase 8 (CDK8) complex binds to MAML and is co-recruited to the transcription complex where it phosphorylates the NotchICD in the PEST domain (Fryer et al., 2004). Phosphorylation of the PEST domain makes it a substrate F-box and WD-40 domain protein 7 (FBXW7), an E3 ubiquitin ligase, which targets the NotchICD for proteosomal degradation (Fryer et al., 2004; Fryer et al., 2002; Wu et al., 2001).
1.2.9 cis-Inhibition In the canonical pathway of Notch activation ligand and receptor are presented in trans i.e. Notch and its receptor are expressed in separate cells. Ligands expressed in the same cell as Notch act as inhibitors of Notch signalling (del Alamo et al., 2011). This process is referred to as cis-inhibition. cis-Inhibition was first described in Drosophila mutants where gene dosage phenotypes supported an inhibitory function for ligands (de-la-Concha et al., 1988; de Celis and Bray, 2000; Ramos et al., 1989). These phenotypes included loss-of-function mutants where both receptor and ligand loss suppressed the phenotype of single loss of either receptor or ligand. In addition, ligand over expression enhanced receptor loss-of-function phenotypes and vice versa. The cis- inhibitory nature of these interactions was investigated in dorsoventral boundary formation in the wing imaginal disc in Drosophila (de Celis and Bray, 1997; del Alamo et al., 2011; Irvine, 1999; Micchelli et al., 1997). In the developing Drosophila wing, Notch is expressed throughout the wing disc while Delta (Dl) 39 | Page
and Serrate ( Ser) are expressed predominantly in the ventral and dorsal compartments, respectively. Cells in the dorsal compartment also express Fringe, which enhances DL and suppresses SER signalling (Section 1.2.3.1). Thus Fringe modified NOTCH in the dorsal compartment responds to DL expressed in the ventral compartment while SER in the dorsal compartment can activate NOTCH in the ventral compartment. This process leads to a strong activation of NOTCH along the dorsoventral boundary. Evidence that this process was in part regulated by cis-inhibition came from over expression studies. Dorsal over expression of DL resulted in Notch activation. However, cells over expressing DL did not display any Notch activation. Notch activation was only seen in neighbouring cells. Over expression of SER was found to elicit similar properties when expressed in the ventral compartment. Over expression of either DL or SER in cells at the dorsoventral boundary, where Notch signalling is endogenously high also displayed this pattern. Thus high levels of ligands expressed in cis inhibit NOTCH activation in that cell. These observations are supported by ligand deletion experiments (del Alamo et al., 2011; Fiuza et al., 2010; Micchelli et al., 1997). Notch activation is not detected in cells flanking the dorsoventral border even though after boundary formation both Dl and Ser are expressed in these compartments. Loss of both Dl and Ser led to Notch activation in these clones by ligands expressed in their wild type neighbours.
In vertebrates evidence for cis-inhibition has largely been confined to over expression studies in cell culture. Cell culture experiments have demonstrated that co-expression of both receptor and ligand leads to a reduction in Notch activation (Ladi et al., 2005; Sakamoto et al., 2002a). Ligands that have been C-terminally truncated and are unable to signal in trans display dominant negative phenotypes in terms of Notch activation when over expressed (Franklin et al., 1999). Additionally, ligand expression in cis affects Notch dependent biological processes such as somite formation (Chapman et al., 2011), T-cell development (Hoyne et al., 2011), neurite outgrowth (Franklin et al., 1999), retinal neurogenesis (Henrique et al., 1997) and keratinocyte differentiation (Lowell et al., 2000; Lowell and Watt, 2001) consistent with cis-inhibition.
The mechanism of cis-inhibition remains poorly defined and highly controversial. The inability of ligands to activate in cis could be due to an inability to induce the conformational changes necessary to expose the S2 site (Section 1.2.6). The binding sites for both the cis and trans interactions overlap and map to the EGF-like repeats 11 and 12 (Becam et al., 2010; Cordle et al., 2008; Fiuza et al., 2010). However, it has been suggested that although the 40 | Page
binding sites for cis and trans overlap, binding is in fact in a different orientation. However, this in silico model lacks experimental support (Cordle et al., 2008). While Notch signalling in trans is linear over a range of concentrations, cis-inhibition occurs as a sharp switch between signalling and non-signalling as ligand levels increase (Sprinzak et al., 2010). The sharp switch could be utilised during development to establish boundaries that have previously been explained by transcriptional feedback loops (Sprinzak et al., 2011; Sprinzak et al., 2010). The major controversy surrounding cis-inhibition is where the interaction takes place, at the cell surface (Glittenberg et al., 2006; Ladi et al., 2005) or in the secretory pathway (Chapman et al., 2011; Sakamoto et al., 2002a). cis-Binding at the cell surface would suggest that ligands in cis cannot activate Notch and compete for binding with ligands in trans. A pure competition model is at odds with results using EDTA to artificially activate Notch (Fiuza et al., 2010; Rand et al., 2000). Ligands were still able to cis-inhibit EDTA activation of Notch, suggesting that cis-inhibition is not due to blocking trans interactions. Intracellular binding may suggest that surface levels of receptor and ligand would be diminished. Cell surface labelling has not detected drops in surface expression under conditions of cis-inhibition (Ladi et al., 2005; Sakamoto et al., 2002a). However, surface expression of Notch is dynamically regulated during ligand binding and changes in surface expression under signalling conditions have not been thoroughly examined. Support for an intracellular interaction comes from work in our laboratory examining the role of Dll3 in somite formation. DLL3 is the most divergent of the DSL ligands with a short intracellular domain. This domain lacks lysine residues that have been shown to be essential for trafficking and signalling of other DSL ligands. The divergent nature of DLL3 is exemplified by its inability to activate Notch in trans (Geffers et al., 2007; Hoyne et al., 2011; Ladi et al., 2005). However, DLL3 can cis-inhibit Notch signalling. These properties make analysis of Dll3 an important tool for analysing cis-inhibition. Although one report where DLL3 was over expressed in L-cells found DLL3 on the cell surface others have found DLL3 to be purely intracellular (Geffers et al., 2007; Ladi et al., 2005). Importantly DLL3 is purely intracellular in vivo where its phenotype is displayed (Chapman et al., 2011) (see below). In addition, DLL3 only interacts with NOTCH1 prior to S1 cleavage and the predominant interaction of NOTCH1 with DLL1 when expressed in cis is with full length NOTCH1 (Chapman et al., 2011). Since the majority, if not all, NOTCH1 at the cell surface is S1 cleaved (Section 1.2.3.2) this supports cis interactions occurring intracellularly. Dll3-/- mice display malformations of the spine and mutations have been found in DLL3 in humans with the congenital spinal abnormality spondylocostal dysostosis (SCD)
41 | Page
(Bulman et al., 2000; Dunwoodie et al., 2002). The defects observed are caused by abnormalities in somite development. During somite development, among other functions, Notch signalling establishes somite boundaries and anterior-posterior polarity (Barrantes et al., 1999; Conlon et al., 1995; Feller et al., 2008; Jiang et al., 2000; Morimoto et al., 2005). The anterior- posterior polarity is established by the restriction of Notch signalling to a narrow band of cells. In the Dll3-/- mouse Notch signalling fails to restrict and a broad area of Notch signalling cells remains. These findings are consistent with cis-inhibition of Notch by DLL3 (Chapman et al., 2011).
1.2.10 Non-canonical Notch signalling In addition to activation and/or inhibition by non-canonical ligands several other non-canonical roles for Notch have been identified. Non-canonical signalling is independent of RBPJ and can be divided into two groups, NotchICD dependent and NotchICD independent (Sanalkumar et al., 2010). Non-canonical signalling by Notch remains poorly defined and controversial. In addition, the signalling attributed to Notch may be due to modification of the signalling of other pathways rather than through Notch itself. Many of the pathways affected by non-canonical Notch signalling also interact with the canonical pathway at multiple levels with transcriptional control of each pathway intertwined (Bash et al., 1999; Cheng et al., 2001; Corada et al., 2010; Estrach et al., 2006).
The best explored example of a role for NotchICD independent of Notch signalling is via interactions with the Wnt signalling pathway (Andersen et al., 2012). Wnt signalling controls a variety of developmental and physiological processes. Briefly, Wnt signalling is activated when the secreted ligand Wnt binds to Frizzled and activates the cytoplasmic protein Dishevelled. Dishevelled inhibits phosphorylation of cadherin associated protein, beta 1 (CTNNB1 also known as beta-catenin) which translocates to the nucleus to activate Wnt target genes. In the absence of Wnt signalling, CTNNB1 is phosphorylated by glycogen synthase kinase 3 beta (GSK3B) which targets CTNNB1 for proteosomal degradation (Andersen et al., 2012). Cell surface Notch can bind CTNNB1 and thus prevent nuclear entry (Hayward et al., 2005). In addition, the normal turnover of Notch receptors (Section 1.2.4) can carry this bound active CTNNB1 and target it to lysosomal mediated degradation. The Notch and Wnt signalling pathways interact at a variety of levels and NotchICD interacts with multiple components of the pathway including GSK3B (Espinosa et al., 2003).
Non-canonical, i.e. RBPJ independent, Notch signalling has also been reported that is dependent on NotchICD production and thus
42 | Page
ligand activation. NotchICD can directly interact with a wide variety of other proteins that are not part of the canonical pathway. Many of the interactions are only detected in over expression systems and remain poorly described and controversial.
The best described examples include interactions with hypoxia inducible factor 1, alpha subunit (HIF1A) and nuclear factor of kappa light polypeptide gene enhancer in B cells (NF B). NF B proteins are a family of transcription factors that regulate immune functions. NF B proteins are retained in the cytoplasm by inhibitor of B (I B) family members. Phosphorylation of I B by I B kinases (IKK) targets it for degradation and allows nuclear entry of NF B (Osipo et al., 2008). When over expressed, NOTCH1ICD can directly interact with NF B family members through a poorly conserved region near the RAM domain (Wang et al., 2001). Interactions with IKK and NF B lead to increased nuclear entry of NF B and transcription (Shin et al., 2006; Song et al., 2008; Vilimas et al., 2007). These interactions occur under conditions of over expression but may have relevance in some types of cancer (Song et al., 2008; Vilimas et al., 2007).
HIF1A plays an essential role in oxygen sensing and interacts with the Notch pathway at multiple levels. NotchICD can bind to HIF1A and recruit it to RBPJ bound promoters enhancing Notch dependent activation (Gustafsson et al., 2005). A non-canonical role for Notch is suggested by an interaction with hypoxia- inducible factor 1, alpha subunit inhibitor (HIF1AN), also known as factor inhibiting Hif1a (FIH). HIF1AN hydroxylates HIF1A which blocks its transcriptional activity and targets it for degradation (Hirota and Semenza, 2005; Wilkins et al., 2009). Notch is also hydroxylated by HIF1AN but functional consequences for Notch of hydroxylation are unknown. However, Notch may compete with HIF1A for access to HIF1AN thus regulating HIF1A activity (Coleman et al., 2007). The ability of Notch receptors to interact with HIF1AN varies between family members. NOTCH4 lacks the critical residues required for hydroxylation but can still bind HIF1AN. Although this interaction is an order of magnitude lower compared to NOTCH1-3, NOTCH4 binds with higher affinity than HIF1A. Thus it is possible NOTCH4 could compete for HIF1AN although evidence that this takes place is lacking (Wilkins et al., 2009).
1.3 Aims The Notch pathway has been extensively studied and a large body of information is available. However, there are still many areas that remain poorly described. Of particular interest to us was the role of the mammalian Notch family member Notch4. The canonical 43 | Page
Notch pathway has largely been described in terms of Drosophila Notch and mammalian Notch1. In results described in Chapter 3 we investigated the ability of NOTCH4 to act as a canonical Notch receptor. We found that not only was NOTCH4 not activated by ligand but potentially inhibited Notch signalling. This observation was pursued in Chapter 4 where we confirmed that NOTCH4 inhibited signalling by NOTCH1. We also wished to evaluate the in vivo role of Notch4. To achieve this we utilised the Notch4d1 mouse reported to be a null allele. Although we observed a previously undescribed angiogenesis phenotype, we also demonstrated that the Notch4d1 mouse is not a true null allele. cDNA constructs of the transcript produced from the Notch4d1 allele were shown to retain the inhibitory capacity of Notch4 identified in Chapter 4.
44 | Page
2 Materials and Methods
2.1 Materials
2.1.1 Chemicals and reagents Ajax Finechem Ethanol, Isopropanol and methanol Amyl Media Bacto tryptone, Bacto yeast extract Astral PBS tablets BDH Chemicals Formamide, Triton X-100 BioRad All Blue Precision Plus markers (10- 250kDa) Diploma Skim milk Geneworks 100bp DNA ladder (low), 500bp DNA ladder Gibco/BRL DMEM, FCS, L-glutamine, OptiMEM-1, Sodium Pyruvate, TripleE express, 1M HEPES buffer ICN Glycerol, NP-40, Tween-20 Invitrogen Lipofectamine 2000 Reagent, Lipofectamine Ltx Reagent, PLUS Reagent, Protein G sepharose 4B, Hoechst 33342, TO-PRO-3 Molecular Probes ProLong Antifade Mounting Medium PerkinElmer EasyTides Uridine 5’-triphosphate - 32P Progen Ampicillin (sodium salt) Promega RNasin Roche Anti-DIG-AP FAB fragments, Complete protease inhibitors, DIG-11-dUTP, dNTPs, NBT/BCIP, PMSF Sigma Agar, Agarose, BSA, BSA (Fraction V), CaCl2 , CHAPS, Citric Acid, DAPT, Disodium hydrogen orthophosphate, Dextran sulfate, DMSO, Donkey serum, DTT, EtBr, EDTA, EGTA, Ficoll 400, 37% Formaldehyde, Glucose, Gluteraldehyde, Glycine, Guanidine hydrochloride, HCl, Heparin sodium, HEPES, Kanamycin sulphate, KCl, Levamisole, MgCl2 , MOPS, Magnesium sulfate, 2-mercaptoethanol, NaCl, Orange G, paraformaldehyde, phenol:chloroform:isoamyl alcohol , Phenol Red, Polyvinylpyrrolidone, Potassium chloride, Potassium phosphate, Mouse IgG, Sodium acetate, Sodium bicarbonate, Sodium deoxycholate, Sodium dihydrogen orthophosphate, SDS, Sheep serum, Sodium lactate, Sodium pyruvate, Spermidine, Sucrose, Torula yeast RNA, TRI Reagent, Tris-HCl, Trisodium citrate, VEGF. Thermo Scientific EZ-Link Sulfo-NHS-SS-Biotin, Streptavidin Agarose Resin 2.1.2 Enzymes Agilent PfuUltra HF DNA Polymerase, 10xPfuUltra HF reaction buffer
45 | Page
Ambion T3 and T7 RNA polymerase Applied Biosystems Big Dye 3.1 Epicentre Monsterscript Reverse Transcriptase Invitrogen Platinum Taq Polymerase, SuperScript III, RNAaseA Kapa Biosystems Kapa-Hi-Fi DNA polymerase New England BioLabs Restriction Endonucleases, T4 DNA ligase Roche Proteinase K RNase DNase free, Taq DNA polymerase, Alkaline phosphatase, calf intestine 1U/L
2.1.3 Kits Applied Biosystems Taqman Gene Expression Assay Mm00440525_m1 FAM labelled, Taqman Gene Expression Assay Mm01205647_gl VIC labelled Clontech Chromaspin 100 DEPC columns Qiagen QIAGEN Plasmid Maxi Kit, QIAquick Gel Extraction Kit, QIAquick PCR Purification Kit Thermo Scientific Pierce BCA Protein Assay Kit, Supersignal West Pico Chemiluminescent Substrate Invitrogen Purelink micro-midi RNA extraction kit, SuperScript ViLo cDNA Synthesis Kit, NuPAGE Tris-acetate 3-8% Pre-Cast gradient gels, NuPAGE Bis-Tris 4-12% Pre- Cast gradient gels, NuPAGE 4xLDS loading buffer, NuPAGE MOPS Running buffer, NuPAGE Transfer buffer, NuPAGE TA Running buffer, LR Clonase, BP Clonase Roche LightCycler 480 Probes Master Promega Dual-Luciferase Reporter 1000 Assay System, pGEM-T Easy Vector System I 2.1.4 Miscellaneous materials BD Biosciences Tissue culture grade plates, wells and tubes Corning CellBind plates 60mm GE Healthcare Hybond N+ , illustra MicroSpin G-50 Columns, Amersham Hyperfilm ECL Menzel-Glaser Coverslips and glass slides Millipore Polyvinylidene Fluoride (PVDF) Thermo Scientific Superfrost Plus slides
46 | Page
2.1.5 Buffers and solutions 5xABI dilution buffer 400mM Tris-Hl (pH 9.0) and 10mM MgCl2
Bacterial lysis buffer 0.1M NaOH, 0.5% SDS, 10mM Tris-HCl pH 8.0, 1mM EDTA
Competant cell storage buffer 10% PEG 8000, 5% DMSO, 10mM MgCl2 , 10mM MgSO4 , 10% glycerol in LB media
50XDenhardt’s solution 1% Ficoll 400, 1% polyvinylpyrolidone, 1% BSA (fraction V)
DNA loading dye 50% glycerol and OrangeG to colour
In situ Hybridisation Solution 50% Formamide, 5X SSC (pH 7.0), 0.1% Tween-20, Heparin (50 μ g/mL), Torula yeast RNA (100 μ g/mL), herring sperm DNA (100 μ g/mL), antisense RNA probe (1/200)
In situ Wash Solution I 50% Formamide, 5X SSC (pH 4.5), 1% SDS
In situ Wash Solution II 0.5M NaCl, 0.01M Tris pH 7.5, 0.1% Tween-20
In situ Wash Solution III 50% Formamide, 2X SSC (pH 4.5)
In situ Prehybridisation Solution 50% Formamide, 5X SSC (pH 7.0), 0.1% Tween-20, Heparin (50 μ g/mL)
KCM 100mM KCl, 30mM CaCl2 , 50mM MgCl2
M2 0.251g/L CaCl2 .H2 O, 10% FCS, 1g/L glucose, 4.969g/L HEPES, 0.356g/L KCl, 0.162g/L KH2 PO4 , 0.143g/L MgSO4 , 5.532g/L NaCl, 0.35g/L NaHCO3 ,4.349g/L Na lactate, 0.33μ M Na pyruvate, 0.01g/L Phenol Red (pH7.4)
10xMOPS 0.4M MOPS (3-(N-morpholino)-propanesulfonic acid) (pH 7.0), 0.5M sodium acetate and 10mM EDTA
NBT/BCIP stain 337.5 μ g NBT and 175μ g BCIP (dissolved in DMF) per mL of NTMT
47 | Page
Northern hybridisation buffer 5xSSC, 5XDenhardt’s solution, 50% formamide, 1%SDS, 100g/mL Torula yeast RNA, 100g/mL herring sperm DNA.
NTMT 0.1 M NaCl, 0.1 M Tris pH 9.5, 0.05 M MgCl2 , 0.1% Tween-20
PBS Four phosphate buffered saline tablets (Austral) dissolved in 200mL of milli-Q water.
PBS++ PBS plus 1mM MgCl2 and 0.1mM CaCl2
PBST PBS with 0.3% Triton X-100
PBT Phosphate buffered saline with 0.1% Tween-20
4% PFA 4% paraformaldehyde in PBS
RIPA 20mM Tris-Cl (pH 7.5), 150mM NaCl, 2mM EDTA, 1% NP-40, 1% sodium deoxycholate and 0.1% SDS
RNA loading dye 1xMOPS buffer, 7.5% formaldehyde, 50% formamide, 1XDNA loading dye and 10g/mL EtBr
20xSSC 3M NaCl, 0.3M Trisodium citrate pH adjusted to 4, 5 or 7 using citric acid
TAE 40mM Tris-HCl (pH8.2), 20mM sodium acetate and 10mM EDTA (pH8.2)
Tail lysis buffer 100mM Tris-HCl (pH 8.8), 1M Tris-HCl (pH 8.8), 200mM NaCl, 5mM EDTA, 0.2% SDS and 0.5mg/mL Proteinase K
10x TBS 1.37M NaCl, 26.83mM Potassium chloride, 250mM Tris (pH 7.5)
TBST 1x TBS, 0.1% Tween-20
TE 10mM Tris (pH8.0) and 1mM EDTA
TTBS 50mM Tris-HCl (pH7.6), 150mM NaCl and 0.05% Tween-20
48 | Page
WCE 20mM HEPES (pH 7.8), 420mM NaCl, 0.5% NP-40, 25% glycerol 0.2mM EDTA and 1.5mM MgCl2
Yolk sac lysis buffer 50mM Tris-HCl (pH 8.0), 1mM EDTA, 0.5% Tween-20 and 0.5mg/mL Proteinase K. 2.1.6 Plasmids
2.1.6.1 Basic cloning vectors pBluescript II KS(-) The plasmid pBluescript II KS(-) (pBS) (Agilent) is a basic bacterial cloning vector that contains a multiple cloning site and the ampicillin resistance gene (AmpR). The multiple cloning site is flanked by T7 and T3 promoters and transcription start sites to allow for in vitro transcription. pGEM-T Easy The vector pGEM-T Easy was supplied linearised with a single 3’ terminal thymidine at both ends as part of the pGEM-T Easy Vector System I (Promega). The vector contains the AmpR gene for bacterial selection and the cloning site is flanked by T7 and SP6 Transcription start sites to allow sequencing with T7 and SP6 primers. pENTR2B Entry vector for use with the Gateway LR recombination system (Gibco BRL). The pENTR2B plasmid contains a multiple cloning site flanked by attL1 and attL2 recombination sites and the kanamycin resistance gene for bacterial selection (KanR). Cloning into the multiple cloning sites removes the ccdB gene. The ccdB protein interferes with E.coli DNA gyrase and prevents growth of most E.coli strains including DH10B. The pENTR2B plasmid was prepared in the E.coli strain DB3.1 that contains a mutation in the DNA gyrase mutation gyrA462 which provides resistance to ccdB. pDONR201 Entry vector for use with the Gateway BP recombination system (Gibco BRL). The pDONR201 plasmid contains attP1 and attP2 recombination sites and the kanamycin resistance gene for bacterial selection (KanR). Recombination of the attP1 and attP2 sites removes the ccdB gene. The ccdB protein interferes with E.coli DNA gyrase and prevents growth of most E.coli strains including DH10B. The pENTR2B plasmid was prepared in the E.coli strain DB3.1 that contains a mutation in the DNA gyrase mutation gyrA462 which provides resistance to ccdB.
2.1.6.2 pCAGiPuro based vectors pCAGiPuro The pCAGiPuro vector is a mammalian expression vector was used to establish stable cell lines (Miyahara et al., 2000). The vector contains an ampiciilin resistance gene (AmpR) for bacterial selection, a puromycin resistance gene (PuroR) for mammalian selection and attR2 and attR1 sequences added by Gavin Chapman 49 | Page
in our laboratory for recombination reactions using the Gateway cloning system (GibcoBRL). Expression of cloned cDNA is driven by the CAG promoter, a fusion of the CMV early enhancer element and the chicken beta actin promoter. The CAG promoter drives the expression of a bicistronic message of the cloned cDNA, an internal ribosome entry site and PuroR. pCAGiPuroHA The pCAGiPuroHA plasmid contains all of the features of pCAGiPuro plasmid in addition to an HA tag at the 3’ end of the cloned cDNA. pCAGiPuroDll4HA The open reading frame of mouse Dll4 cloned into the pCAGiPuroHA vector that drives the expression of DLL4 with a C- terminal HA tag and PuroR. pCAGiPuroJag1 The open reading frame of mouse Jag1 cloned into the pCAGiPuro vector that drives the expression of JAG1 and PuroR. pCAGiPuroNotch1HA The open reading frame of mouse Notch1 cloned into the pCAGiPuroHA vector that drives the expression of NOTCH1 with a C-terminal HA tag and PuroR. Made by Sharon Pursglove in our laboratory. pCAGiPuroNotch1myc The open reading frame of mouse Notch1 with a myc tag cloned into the HindIII site of the cDNA sequence cloned into pCAGiPuro that drives expression of myc tagged NOTCH1 and PuroR (Nye et al., 1994). pCAGiPuroNotch1-GFP The open reading frame of mouse Notch1 with GFP fused to the C- terminal. Constructed by Sharon Pursglove in our laboratory.
2.1.6.3 pCMX based vectors pCMX The pCMX mammalian expression vector was used for transient expression in mammalian cell lines (Umesono et al., 1991). The vector contains an ampicillin resistance gene (AmpR) for bacterial selection and attR2 and attR1 sequences were added by Gavin Chapman in our laboratory for recombination reactions using the Gateway cloning system (GibcoBRL). Expression of cloned cDNA is driven by the CMV promoter of cytomegalovirus. pCMXHA The pCMXHA plasmid contains all of the features of the pCMX plasmid in addition to an HA tag at the 3’ end of the cloned cDNA. pCMXCAT Control plasmid expressing the bacterial gene chloramphenicol transferase (CAT) (Hoyne et al., 2011).
50 | Page
pCMXN1HA Constructed by Sharon Pursglove in our laboratory. The complete open reading frame of Notch1 with a C-terminal HA tag. pCMXN1myc The open reading frame of mouse Notch1 with a myc tag cloned into the HindIII site of the cDNA sequence cloned into pCMX that drives expression of myc tagged NOTCH1 and PuroR (Nye et al., 1994). pCMXren The Renilla luciferase gene from pRL-tk (Promega) cloned into pCMX by Gavin Chapman in our laboratory.
2.1.6.4 pCS2 based vectors pCS2+MT The pCS2+MT expression vector was used for transient expression in mammalian cell lines. The vector contains an ampicillin resistance gene (AmpR) for bacterial selection and 6 myc tags for detection of expressed proteins. Expression of cloned cDNA is driven by the simian CMV IE94 promoter of cytomegalovirus (Roth et al., 1991).
N1 E, N2 E, N3 E and N4 E Notch E plasmids contain cDNA that codes for the NOTCH1 signal peptide plus amino acids 1704-2813 of NOTCH1 (N1 E), amino acids 1660-2145 of NOTCH2 (N2 E), amino acids 1623-2099 of NOTCH3 (N3 E) and amino acids 1421-1864 of NOTCH4 (N4 E) followed by 6 myc tags. In addition, methionine residues in NOTCH1 to 3 were mutated to prevent alternative translation start sites (M1727V, M1697L and M1663L for NOTCH1, 2 and 3 respectively) (Saxena et al., 2001; Schroeter et al., 1998).
2.1.6.1 Miscellaneous pGL46xTP1 The plasmid pGL4 (Promega) containing the firefly luciferase gene under the control of 6 copies of the Notch responsive promoter (TP1) of Epstein Barr virus (Kato et al., 1996).
pYXNotch4 I.M.A.G.E. clone number 6855960 containing a 5’ truncated Notch4 transcript (Lennon et al., 1996).
pSport6Dll4 I.M.A.G.E. clone number 4017786 containing the complete open reading frame of Dll4 (Lennon et al., 1996).
Notch E-GVP N1 E based plasmid with the GAL4/VP16 sequences cloned into an introduced AscI site C-terminal of the transmembrane region (Karlstrom et al., 2002).
51 | Page
pEFBosNotch4mRuby The open reading frame of Notch4 fused to mRuby cloned into the expression vector pEFbos (Mizushima et al., 1990). Constructed by Gavin Chapman in our laboratory.
Additional plasmids constructed for this work are described in Section 2.2.2.
2.1.7 Mammalian Cell lines NIH3T3 and C2C12 based cell lines
NIH3T3 is an adherent mouse embryonic fibroblast line. C2C12 cells are an adherent mouse muscle myoblast line isolated from dystrophic mouse muscle (Yaffe and Saxel, 1977). Cell lines based on NIH3T3 and C2C12 contained a pCAGiPuro based vector incorporated in the genome. Cell lines, the associated vector and the cDNA expressed are shown in Table 2.1.
Table 2.1 NIH3T3 and C2C12 cell lines.
Cell Line Parent pCAGiPuro Vector cDNA line NIH3T3Dll4HA NIH3T3 pCAGiPuroDll4HA Dll4HA NIH3T3Jag1 NIH3T3 pCAGiPuroJag1 Jag1 NIH3T3iPuro NIH3T3 pCAGiPuro n/a NIH3T3Notch1myc NIH3T3 pCAGiPuroNotch1myc Notch1myc NIH3T3Notch1HA NIH3T3 pCAGiPuroNotch1HA Notch1HA C2C12iPuro C2C12 pCAGiPuro n/a C2C12Notch1HA C2C12 pCAGiPuroNotch1HA Notch1HA C2C12Notch1GFP C2C12 pCAGiPuroNotch1GFP Notch1GFP
MAEC An endothelial cell line derived from the aorta of p53 deficient mice (Nishiyama et al., 2007).
HUVEC HUVEC umbilical venous endothelial cells were purchased from Cambrex.
Additional cell lines developed for this work are described in Section 2.2.3.1.
52 | Page
2.1.8 Oligonucleotides All oligonucleotides were ordered from Sigma at a scale of 0.05 mol, desalted, and resuspended at 100 M in TE for storage at -20⁰ C. Working aliquots for short term use were diluted to 10 M in TE and stored at -20⁰ C.
Notch4 primers are named N4ABC where A is the exon number or I followed by the intron number, B is the orientation (forward (F) or reverse (R)) and C a unique identifier if multiple primers are in a single exon. If a restriction enzyme site is introduced the enzyme name is appended to the end. Mutagenic primers are followed by the amino acids introduced. Primers within the neo cassette are neoForR and a unique identifier. Notch1 oligonucleotides begin with N1.
Table 2.2 Oligonucleotides
Primer Sequence
3’RACE RT CCAGTGAGCAGAGTGACGAGGACTCGAGCTCAAGCTTTTTTTTT TTTTTTTTVN N1FBamHI CTCGGGCCCACGTAGTCCCACCTG
N1RAscI AGGGAACCAGAGCTGGCCATGGGC
N413F CTGCATCTCCACACCCTGT
N415F ATCGAGCAGTGTGTGGACAG
N415F GCTGCACTGTGAGGAGAAGA
N416F CCTCGTTCCAGTGCCTGT
N416R ATCGAGCAGTGTGTGGACAG
N419F TAGCCAACGCCTTCTACTGC
N41F GCTCTTGCCACTCAATTTCCC
N41FEcoRI GAGGGGGAATTCCTGAAGAGGGAGAGGAGA
N421F GAGGAGTGTCTCTTTGATGGC
N421F2 CGCAGTGTGACTCTGAGGAG
N421F2 CGCAGTGTGACTCTGAGGAG
N421F3 AGGAGGAGACTGGGATGGAG
N421RXhoI ATAGCTCGAGATGCAGGTTAGAGGGATTTC
53 | Page
N422F CCTATGACCAGTACTGCCGAG
N422F1 CTGGGCAAGGAGACAGAGTC
N422R GAGGAATCCCTAGCTCCACTC
N424FAscI ACGGCGGCGCGCCCATGGGGCCCT
N424RAscI AGGGCCCCATGGGCGCGCCGCCGT
N425R TCCATCCTCATCCACTTCGGCCTC
N426F GGCTGAAGAAACAGCCTCAG
N428F GGACGGGACTACACCTTTGA
N42R GAGATAGCCTCAGGCAGGTG
N430F CACTTGGTCGGTGGACTTG
N430F2 CCACGCGCGGCCGCAGGTTC
N430REagI CGCAGCCGGCCGGTTCAGATTTCTTACAACCG
N43R GCCACCATTCTTGCAGAGTTG
N44F GCGACATCAACGAGTGCTT
N45FPmlI GACACCTACACGTGCCCCTGCCCCAAG
N45RPmlI CTTGGGGCAGGGGCACGTGTAGGTGTC
N46F GGATCCACCTGCATCGAC
N49F CTGCCATGACCTGCTCAAC
N4I21F CATAGAATGCCTCCCTGGAA
N4RQRRF GGAGCTAGGGATCGCCAGCGACGGGAAAGACAAGCC
N4RQRRR GGCTTGTCTTTCCCGTCGCTGGCGATCCCTAGCTCC
NeoF2 CTGGGCACAACAGACAATCGGC
NeoR3 TATTCGGCAAGCAGGCATCGCC
NeoF4 ATTGCATCGCATTGTCTGAG
NeoR4 TGTCTGTTGTGCCCAGTCAT
NeoR1 AAGCGCATGCTCCAGACTGCC
OligodT TTTTTTTTTTTTTTTTT
54 | Page
Q 0 CCAGTGAGCAGAGTGACG
Q 1 GAGGACTCGAGCTCAAGC
SP6 ATTTAGGTGACACTATAGAA
T7 TAATACGACTCACTATAGGG
2.1.9 Bacterial strains and reagents All liquid bacterial cultures were performed in Luria broth (g/L Bacto tryptone peptone digest, 5 g/L Bacto yeast extract and 10 g/L Sodium chloride) in a orbital shaker at 37⁰ C. Antibiotic was added where appropriate (ampicillin 100g/mL and kanamycin 50g/mL). Bacterial agar plates were made with Luria broth containing 15g/L agar and the appropriate antibiotic (ampicillin 100g/mL and kanamycin 50g/mL) and incubated in a 37⁰ C oven.
DH5B (F– mcrA ∆ ( mrr- hsdRMS-mcrBC) Φ 80lacZ ∆ M15 ∆ lacX74 recA1 endA1 araD139 ∆ ( ara leu) 7697 galU galK rpsL nupG λ ). An E.coli strain used for routine cloning.
DB3.1 (F– gyrA462 endA1 ∆ (sr1-recA) mcrB mrr hsdS20(rB–, mB–) supE44 ara-14 galK2 lacY1 proA2 rpsL20(SmR) xyl-5 λ – leu mtl1). An E.coli strain used to amplify plasmids that were part of the Gateway system containing the ccdB gene.
2.1.1 Antibodies
Table 2.3 Antibodies
Antibody Species Use:Dilution Source
Beta-Actin Mouse WB:1/10000 Sigma (AC-15)
Beta-Tubulin Mouse WB:1/10000 Sigma (TUB 2.1)
Guinea pig Donkey IC:1/1000 Molecular IgG Alexa Probes Fluor 488
Guinea pig Donkey WB:1/1000 Jackson IgG HRP- Immunoresearch conjugated
55 | Page
HA Mouse WB:1/1000 Covance (HA.11) IC:1/500 IP:1/50 MHC Mouse WB:1/50 Developmental (MF20) IC:1/20 Studies Hybridoma bank Mouse Donkey IHC:1/1000 Molecular Alexa Fluor Probes 488
Mouse IgG Donkey IC:1/1000 Jackson Cy3- Immunoresearch conjugated
Mouse IgG Donkey WB:1/10000 Molecular Alexa Fluor Probes 680nm
Myc Mouse WB:1/1000 Developmental (9E10) Studies Hybridoma bank Notch1 Rabbit IP:1/100 Abcam (ab27526)
Notch4 Guinea WB:1/1000 See Section (N41437) pig IC:1/500 2.2.9
PECAM Rat IHC:1/200 BD Pharmingen (MEC 13.3)
Rat IgG Donkey IHC:1/1000 Jackson Cy3- Immunoresearch conjugated
SMA Mouse IHC:1/100 Dako (1A4) CytoMation
WB-Western blot, IP-Immunoprecipitation, IC- Immunocytochemistry, IHC-Immunocytochemistry
2.2 Methods
2.2.1 Basic Molecular Biology Techniques
2.2.1.1 Restriction digests Restriction endonuclease digestions were performed in 20µl volumes containing either 1µg of maxiprep purified plasmid or 2µl of miniprep plasmid, 2µl of the appropriate buffer (NEB), 0.5µl of the appropriate restriction enzyme (NEB) and made up to 20µl with MilliQ water. The digestion was performed at 37°C from one hour to overnight.
56 | Page
2.2.1.2 Alkaline phosphatase Restriction digests were incubated in the presence of 0.5U of calf intestinal phosphatase (Roche) for 30 minutes at 37°C. Fragments were then immediately agarose gel purified (Sections 2.2.1.4).
2.2.1.3 Agarose gel electrophoresis 0.7-2% Agarose gels were used to analyse plasmid DNA and PCR products. Agarose was dissolved by boiling in TAE buffer before adding 0.5mg/mL of EtBr. Gels were then poured into horizontal gel boxes. One tenth volume of DNA loading buffer was added to the samples and size makers (Geneworks) before loading. The gels were electrophoresed in TAE at 5V/cm for approximately 40 minutes. The bands were visualised by medium wavelength UV light and photographed using a Gel Doc System (BioRad).
2.2.1.4 Gel purification Bands of DNA separated by agarose gel electrophoresis were visualised under UV light and the desired band excised using a clean scalpel blade. The gel fragment was weighed and purified according to manufacturer’s instructions using the QIAquick Gel Extraction Kit (Qiagen).
2.2.1.5 DNA ligations Ligations were performed in 20µl volumes consisting of plasmid:insert ratios varying from 1:1 to 1:6. The components of the reaction were 2µl T4 ligase buffer (NEB), 0.5µl T4 DNA ligase
(NEB), 100ng DNA and H2 O to 20µl. The reactions were incubated at room temperature from 1 hour to overnight.
2.2.1.6 Gateway Cloning Gateway recombination cloning was performed according to the manufacturer’s instructions using either LR Clonase or BP Clonase as indicated (Invitrogen).
2.2.1.7 Bacterial Transformation Competant cells were prepared by seeding 100µl of an overnight culture of DH10B into 100ml of LB broth and grown with vigorous shaking for approximately 3 hours at 37°C. The optical density of the culture was measured at 600nm. At an OD600 of 0.6-0.8 the bacteria were collected by centrifugation at 4000g for 20 minutes at 4⁰ C and resuspended in cold (4°C) competent cell storage buffer. The bacteria were immediately frozen in a dry ice/ethanol bath and stored until use at -80°C. Competant cells (50µL) were incubated with plasmid DNA diluted in 50µL KCM buffer for 30 minutes on ice. Following a 2 minute heat shock at 42°C the cells were allowed to recover for 30 minutes in LB media at 37°C before being selected on LB agar plates containing the either 100g/mL ampicillin or 50g/mL kanamycin.
57 | Page
2.2.1.8 Miniprep purification of plasmids Single colonies were picked from agar plates and used to inoculate 1mL of LB media plus the appropriate antibiotic. Bacterial cultures were grown for at least 5 hours to overnight at 37°C in an orbital shaking incubator. The bacteria were pelleted by centrifugation at 20000g for 30 seconds in a benchtop centrifuge. The pellets were resuspended in 300µL of bacterial lysis buffer followed by the addition of 150µL of 3M sodium acetate (pH 5.2). After mixing by inversion the lysate was centrifuged for 4 minutes at 20000g and the supernatant removed and added to 1mL of 100% ethanol. The precipitated DNA was collected by centrifugation at 20000g for 4 minutes followed by a 400µL wash in 70% ethanol. The pellet was air dried and resuspended in 20µL of 20µg/mL RNAaseA. Restriction digests used 2 L of miniprep DNA in a total volume of 20L.
2.2.1.9 Maxiprep purification of plasmids 100mL of LB media containing either 100 µg/mL ampicillin or 50g/mL kanamycin was inoculated and incubated overnight at 37°C in an orbital shaker. The cells were harvested by centrifugation at 4000g for 15 minutes at 4°C, and the bacterial pellets drained. Plasmid DNA was extracted using the Purelink Invitrogen Maxiprep kit, according to the manufacturer’s instructions. Yield and quality of plasmid DNA was determined, at wavelengths of 260nm and 280nm and resuspended at 1mg/mL. Maxiprep DNA was stored at -20⁰ C.
2.2.1.10 Sequencing One microgram of maxiprep purified plasmid DNA was subjected to cycle sequencing in the presence of 25pmol of primer, 1l Big Dye terminator mix (PE Biosystems), 3.5 l of 5xABI dilution buffer in a total volume of 20l. The reaction was cycled 25 times through the following steps; 96°C for 30 seconds, 50°C for 15 seconds and 60°C for 4 minutes. Reactions were analysed at the DNA sequencing facility, University of New South Wales, Sydney, Australia and viewed using SeqScanner software (Applied Biosystems).
2.2.1.11 Site Directed Mutagenesis Plasmid DNA was amplified using the Kapa HiFi PCR kit (Kapa Biosystems) according to the manufacturer’s instructions. The reaction consisted of 10ng plasmid, 0.375 M of both forward and reverse primers, 0.375mM of each dNTP, 1x fidelity buffer (Kapa Biosystems) and 0.5U KapaHiFi. The reaction was incubated at 95⁰ C for 3 minutes followed by 18 cycles of 98⁰ C for 20 seconds, 55⁰ C for 15 seconds and 72⁰ C for 30seconds/kilobase of plasmid then 10 minutes at 72⁰ C. The PCR reaction was purified using the QIAquick PCR Purification Kit and digested with DpnI prior to
58 | Page
transformation into DH5B. See individual plasmid constructions for plasmids and primers used. All mutant plasmids were confirmed to be correct by sequencing.
2.2.2 Vector construction
2.2.2.1 pCAGiPuro based vector construction pCAGiPuroNotch4HA The 3’ end of Notch4 was PCR amplified from pYXNotch4 using the primers N430F and N430FEagI to create an in frame fusion with the pENTR2B vector. The reaction consisted of 10ng pYXNotch4, 1 M N430F, 1 M N430REagI, 2.5 M of each dNTP, 1x Taq polymerase buffer and 0.5U Taq polymerase (Roche). The reaction was incubated at 95⁰ C for 30 seconds followed by 30 cycles of 95⁰ C for 30 seconds, 55⁰ C for 30 seconds and 72⁰ C for 30 second then 72⁰ C for 10 minutes. The PCR product was gel purified and digested with EagI. The digested PCR product was cloned into pENTR2B, that had been digested with NotI and treated with alkaline phosphatise, to create plasmid pENTR2BN4NotI. RNA was extracted from the lung tissue of C57BL/6J mice using the Purelink micro-midi RNA extraction kit according to manufacturer’s instructions (Thermo Fisher Scientific). The RNA was reversed transcribed using Monsterscript reverse transcriptase (Epicentre). The primer N416R and 1g RNA were heat denatured at 65⁰ C for 5 minutes prior to addition of 2mM of each dNTP, 1x Monsterscript buffer and 5U of Monsterscript reverse transcriptase. The reaction was incubated at 42⁰ C for 5 minutes, 60⁰ C 20 minutes, 65⁰ C 20 minutes then 90⁰ C for 5 minutes. The cDNA was PCR amplified using PfuUltra polymerase in a touchdown procedure. The reaction consisted of 1L of cDNA, 1 M N41FEcoRI, 1 M N416R, 1x PfuUltra HF buffer, 0.25 M of each dNTP and 2.5U PfuUltra HF DNA polymerase in a total volume of 20L. The reaction was incubated at 95⁰ C for 3 minutes followed by 12 touchdown cycles of 95⁰ C for 30 seconds, 72⁰ C, at cycle one and dropping 1 ⁰ C/cycle, for 30 seconds and 72⁰ C for 5 minutes. The touchdown cycles were followed by 28 cycles of 95⁰ C for 30 seconds, 60⁰ C for 30 seconds and 72⁰ C for 5 minutes followed by 72⁰ C for 10 minutes. The PCR product was gel purified, digested with EcoRI and XhoI, and cloned into EcoRI and XhoI digested pYXNotch4 to create pYXNotch4FL. An EcoRI to NotI fragment was subcloned from pYXNotch4FL into EcoRI and NotI digested pENTR2BN4NotI to create pENTR2BNotch4tag. A Gateway LR reaction was used to transfer the cDNA to pCAGiPuroHA.
59 | Page
pCAGiPuroDll4 The plasmid pSport6dll4 was used in a BP Gateway reaction to make pDONRDll4. This plasmid was used in a subsequent LR Gateway reaction to create pCAGDll4. pCAGNotch4d1 A fragment of Notch4 was PCR amplified from pCMXNotch4 (see below) with the primers N421RXhoI and N432F. The reaction consisted of 10ng plasmid, 0.375 M N432F and N421RXhoI, 0.375 M of each dNTPs, 1x fidelity buffer (Kapa Biosystems) and 0.5U KapaHiFi. The reaction was incubated at 95⁰ C for 3 minutes followed by 18 cycles of 98⁰ C for 20 seconds, 55⁰ C for 15 seconds and 72⁰ C for 90 seconds then 10 minutes at 72⁰ C. The PCR product was gel purified, digested with XhoI and subcloned into XhoI cut and alkaline phosphatase treated pENTR2BNotch4tag to create pENTR2BNotch4d1. The cDNA was transferred to pCAGiPuroHA in a LR Gateway reaction.
2.2.2.2 pCMX based vector construction pCMXNotch4HA A gateway LR reaction was used to transfer the Notch4 sequences from pENTR2BNotch4tag into pCMXHA. pCMXNotch4 A NotI fragment from pYXNotch4 was subcloned into NotI digested and alkaline phosphatase treated pENTR2BNotch4tag. The resulting cDNA was transferred in a gateway reaction to the pCMX vector.
Notch1:1 A NotI site in the 3’ untranslated region in the plasmid pENTR2BNotch1myc was removed by digesting with SpeI and XhoI, gel purifying the vector and re-ligating. The GAL4/VP16 sequences of Notch EGAL4/VP16 (Karlstrom et al., 2002) were removed by cutting with AscI, gel purifying the plasmid and religating. A NotI fragment was subcloned into pBS to make pBSN1NA. A fragment of Notch1 from pCMXNotch1myc was amplified using the N1FBamHI and N1RAscI primers. The reaction consisted of 10ng of pCMXNotch1myc, 1M N1FBamHI, 1 M N1RAscI, 1x PfuUltra HF buffer, 0.25M of each dNTP and 2.5U PfuUltra HF DNA polymerase in a total volume of 20L. The reaction was incubated at 95⁰ C for 3 minutes followed by 30 cycles of 95⁰ C for 30 seconds, 55⁰ C for 30 seconds and 72⁰ C for minutes. The PCR product was gel purified, digested with BamHI and AscI and cloned into BamHI and AscI cut pBSN1NA to make pBS1:1. A BamHI to HindIII fragment was subcloned into pCMXNotch1myc to make pCMXNotch1:1.
Notch4:1 A fragment of Notch4 was amplified from the plasmid pCMXNotch4HA using the primers N413F and N424RAscI. The reaction consisted of 10ng of pCMXNotch4HA, 1M N413F, 1 M N424RAscI, 1x PfuUltra HF buffer, 0.25M of each dNTP and 2.5U
60 | Page
PfuUltra HF DNA polymerase in a total volume of 20L. The reaction was incubated at 95⁰ C for 3 minutes followed by 30 cycles of 95⁰ C for 30 seconds, 55⁰ C for 30 seconds and 72⁰ C for 2 minutes followed by a 10 minute incubation at 72⁰ C. The PCR product was gel purified, digested with XhoI and AscI and cloned into AscI and XhoI cut pBS1:1 resulting in pBS4:1. A BstEII to NotI fragment from pBS4:1 was cloned into BstEII and NotI cut pCMXNotch4HA to make pCMXNotch4:1.
Notch4:4 A fragment of Notch4 was amplified from the plasmid pCMXNotch4HA using the primers N424FAscI and N425R. The reaction consisted of 10ng of pCMXNotch4HA, 1M N424FAscI, 1 M N425R, 1x PfuUltra HF buffer, 0.25 M of each dNTP and 2.5U PfuUltra HF DNA polymerase in a total volume of 20L. The reaction was incubated at 95⁰ C for 3 minutes followed by 30 cycles of 95⁰ C for 30 seconds, 55⁰ C for 30 seconds and 72⁰ C for 30 seconds, followed by a 10 minute incubation at 72⁰ C. An AflII to AscI fragment was cloned into AflII and AscI cut pBS4:1 to make pBS4:4. A BstEII to AflII fragment was subcloned from pBS4:4 into BstEII and AflII cut pENTR2BNotch4tag and an LR Gateway reaction used to transfer the cDNA to make pCMXNotch4:4.
Notch1:4 An EcoRI to AscI fragment of pCMXNotch1:1was cloned into EcoRI and AscI cut pCMXNotch4:4 to make Notch1:4.
pCMXRQRR Site directed mutagenesis was performed on the plasmid pBSN4XN using the primers N4RQRRF and N4RQRRR. A BstEII to AflII was subcloned into BstEII and AflII cut pENTR2BNotch4tag and an LR Gateway reaction used to transfer the cDNA to pCMXHA. pCMXNotch4d1HA The cDNA from the plasmid pENTR2BNotch4d1 was transferred in a Gateway reaction to pCMXHA.
2.2.2.3 pCS2 based vector construction N1ECT The plasmid pCMXN1HA was cut with BglII and XhoI, gel purified and ligated into the BglII and XhoI sites of N1 E.
N4ECT An AflII to NheI fragment of pCMXNotch4HA was gel purified and cloned into the AflII and XbaI sites of N4 E.
2.2.2.4 pGEM-T Easy based vectors PCR products to be sequenced were cloned using the pGEM-T Easy Vector System I kit. PCR products were gel purified and cloned into pGEM-T Easy according to manufacturer’s instructions. Maxiprep purified plasmid was then used for sequencing using the primers T7 and SP6.
61 | Page
2.2.2.5 Vectors for in situ hybridisation and northern probes pBSN4EP The plasmid pBSEP was made to generate the 5’ probe for in situ hybridisation. An EcoRI to XhoI fragment was subcloned from pCMXNotch4HA into pBS. A PmlI site was introduced by site directed mutagenesis using the forward and reverse primers N45FPmlI and N45RPmlI respectively. The EcoRI to PmlI fragment was subcloned into the EcoRI and SmaI sites in pBS. pBSN4XN The plasmid pBSN4XN was used to generate the 3’ probe for in situ hybridisation. A XhoI to NotI fragment of pCMXNotch4HA was subcloned into the XhoI and NotI sites of pBS. 2.2.3 Mammalian cell culture and transfection
2.2.3.1 C2C12 and NIH3T3 lines C2C12 and NIH3T3 lines were grown in DMEM containing 2mM L- glutamine and 10% FCS in a humidified incubator at 37⁰ C and 5%
CO2 . Sub-confluent cultures were passaged by detaching the cells with TripleE express (Gibco/BRL) and diluting 1/10 every 2-3 days. For transfections cells were seeded at 2.5x104 cells per cm2 and grown overnight at 37°C in a humidified chamber containing 5%
CO2 . Cells were transfected with Lipofectamine™ Ltx Reagent (Invitrogen) under standard conditions following the manufacturer’s instructions. The stable cell lines NIH3T3Notch4HA, NIH3T3Dll4, C2C12Notch4HA, C2C12iPuro were established using pCAGiPuro based vectors. Cells were transfected using standard techniques and cultured for 48 hours. The cultures were detached with TripleE express and passaged at dilutions of 1/20, 1/40 and 1/80 in standard media plus 1g/mL puromycin. The cells were cultured until individual colonies were established. Isolated single colonies were collected and passaged into both standard cultures and onto coverslips for immunocytochemistry. Lines that displayed even expression with similar morphology and growth to the parent line were used. Lines were maintained in standard media plus 1g/mL puromycin.
2.2.3.2 MAEC MAEC cells were grown in M199 media containing 2mM L- glutamine, 10mM HEPES, 0.1mg/mL Heparin sodium, 5ng/mL VEGF and 5% FCS in a humidified incubator at 37⁰ C and 5%CO2 . Sub- confluent cultures were passaged by detaching the cells with TripleE express and diluting 1/5 every 3-4 days. For transfections cells were seeded at 2.5x104 cells per cm2 and grown overnight. DNA complexes were prepared by diluting DNA, with the addition of 1L/ g Plus reagent, and Lipofectamine 2000 (Invitrogen) in a ratio of 1:5 in OptiMEM and incubating at room temperature for 30 minutes. The media on the cells was aspirated and replaced with serum free OptiMEM. The cells were incubated with the DNA 62 | Page
complexes for 4 hours. The complexes were removed and the media replaced with standard media.
2.2.3.3 HUVEC HUVEC cells (Cambrex) were grown in EGM-2 with the addition of EGM-2 SingleQuots (Cambrex) which contained hydrocortisone, hEGF, FBS, VEGF, hFGF-B, R3-IGF-1, ascorbic acid, heparin and gentamicin/amphotericin-B) in a humidified incubator at 37⁰ C and
5% CO2. Media was exchanged every 2-3 days and the subconfluent cultures passaged every 5-6 days by detaching in TripleE express and diluting 1/5. For transfection cells were seeded at 2.5x104 cells per cm2 and grown overnight. DNA complexes were prepared by diluting DNA, with the addition of 1 L/g Plus reagent, and Lipofectamine 2000 (Invitrogen) in a ratio 1:2.5 in OptiMEM and incubating at room temperature for 30 minutes. The media on the cells was aspirated and replaced with serum free OptiMEM. The cells were incubated with the DNA complexes for 4 hours. The complexes were removed and the media replaced with standard media.
2.2.4 Mouse lines Animal ethics permission was obtained under the animal ethics number 09/33. All mice were housed initially at the Biological Testing Facility in the Garvan Institute of Medical Research, Sydney and subsequently at the Biocore, Victor Chang Cardiac Research Institute, Sydney. The mice were kept in a perpetual 12 hour light/dark cycle and fed ad libidum. Male and female mice were separately housed unless needed for specific breeding purposes.
Notch4tm1Grid, referred to as Notch4d1, was a kind gift from Thomas Gridley. Notch4d1 carry a neomycin insertion cassette in the Notch4 gene that spans exons 22 and 23 (Krebs et al., 2000).They were received and housed in the quarantine room of the Biological Testing Facility. Female Notch4+/d1 mice were induced to super ovulate and ova removed at day E2.5. Natalie Wise isolated blastocysts and implanted them in psuedopregnant recipients.The breeding colony was maintained by crossing heterozygous Notch4+/d1 with purebred C57BL/6J mice. C57BL/6J mice are referred to as wild type.
2.2.4.1 Genotyping DNA was extracted from mice by cutting a small piece of tail (approximately 3 mm) or ear clip from each mouse. This tissue was lysed in 500µL murine tail DNA lysis solution containing Proteinase K (0.5 mg/mL) at 55 C overnight. Any undigested tissue was removed by centrifugation for 5 minutes at 20000g. Following centrifugation the sample was precipitated with an equal volume of isopropanol and centrifuged for 5 minutes. The sample was then 63 | Page
washed in 70% ethanol and air-dried at room temperature for 15 minutes before resuspending in 200 l of milli-Q H2 O. For embryo genotyping a small section of yolk sac was rinsed in PBS and resuspended in yolk sac buffer with 5mg/mL Proteinase K overnight at 55⁰ C. Undigested tissue was removed by centrifugation at 20000g for 5 minutes followed by heat inactivation at 95⁰ C for 5 minutes. The primers to detect wild type Notch4 allele were N422F and N422R, the Notch4d1 allele were N421F and Neo1R and the insertion cassette Neo2F and Neo3F. The PCR reactions contained 2L DNA, 0.5U Taq polymerase (Roche), 1xPCR buffer, 1 M of each primer and 0.25 M of each dNTP. The PCR conditions were 95⁰ C for 1 minute followed by 35 cycles of 95⁰ C for 30 seconds, 58⁰ C for 30 seconds, 72⁰ C for 30 seconds followed by 10 minutes at 72⁰ C.
2.2.5 Reverse Transcription-PCR (RT-PCR) Neonatal mice at day 5 post birth (P5) were sacrificed and the lungs removed. RNA was extracted using the Purelink micro-midi RNA extraction kit according to the manufacturer’s instructions. One microgram of RNA was reverse transcribed using SuperScript III (Invitrogen). One microgram of RNA was heat denatured at 65°C for 5 minutes in the presence of 50pmol of oligodT primer and 0.25 M of each dNTP in a final volume of 17µL. Two microlitres of 10xPCR buffer (Roche) and 1µL SuperScript III (Invitrogen) was added and incubated at 50°C for 1 hour followed by heat inactivation at 70°C for 20 minutes. The cDNA was treated with 1µL RNaseH for 20 minutes at 37°C. The primer pairs used to amplify the cDNA were N422F and N422R both within exon 22, neoF2 and NeoR3 within the neo gene, N41F and N43R to span exons 1 to 3, N415F and N416R to span exons 15 and 16 and N422F and neoR4 to amplify the Notch4d1 transcript. All PCR reactions contained 1L cDNA, 0.5U Taq polymerase (Roche), 1xPCR buffer, 1 M of each primer and 0.25 M of each dNTP. The PCR conditions were 95⁰ C for 1 minute followed by 35 cycles of 95⁰ C for 30 seconds, 58⁰ C for 30 second, 72⁰ C for 30 seconds followed by 10 minutes at 72⁰ C.
2.2.6 3’RACE Neonatal mice at day 5 post birth (P5) were sacrificed and the lungs removed. RNA was extracted using the Purelink micro-midi RNA extraction kit according to the manufacturer’s instructions. The 3’RACE protocol was adapted from (Borson et al., 1992; Scotto-Lavino et al., 2006). One microgram of RNA was heat denatured at 65°C for 5 minutes in the presence of 50pmol of primer 3’RACE RT and 0.25 M of each dNTP in a final volume of 17µL. Two microlitres of 10xPCR buffer (Roche) and 1µL SuperScript III (Invitrogen) was added and incubated at 50°C for 1 64 | Page
hour followed by heat inactivation at 70°C for 20 minutes. The cDNA was treated with 1µL RNaseH for 20 minutes at 37°C. The cDNA was purified with the QIAquick PCR Purification Kit (Qiagen) with an additional 35% guanidine hydrochloride wash. One microlitre of cDNA was PCR amplified in a 20µl reaction containing
0.5U Platinum Taq, 1x PCR buffer (Roche), 1µM Qo primer, 1µM N419F primer and 0.25mM of each dNTP. The reaction was denatured at 94°C for 5 minutes followed by 35 cycles of 94°C for 10 seconds, 64°C for 10 seconds and 72°C for 4 minutes and 30 seconds followed by a final incubation at 72°C for 15 minutes. The PCR reaction was purified with the QIAquick PCR Purification Kit (Qiagen) including a wash with 35% guanidine hydrochloride according to the manufacturer’s instructions. One microlitre of the purified PCR reaction was amplified through a second PCR reaction consisting of 0.5U Platinum Taq, 1x PCR buffer, 1 M primer Q1 , 1 M N421F3 primer and 0.25mM of each dNTP. The reaction was heat denatured at 94°C for 5 minutes followed by 35 cycles of 94°C for 10 seconds, 60°C for 10 seconds and 72°C for 1 minute followed by a final incubation at 72°C for 15 minutes. The resulting PCR product was cloned into pGEMeasyT according to the manufacturer’s instructions. The cloned product was sequenced using the primers T7 and SP6.
2.2.7 Real time PCR Neonatal mice at 5 days post birth (P) were sacrificed and their lungs removed. The lungs were washed in cold PBS and snap frozen in a dry ice/ethanol bath. The tissue was stored at -80°C until used. Concurrently the tips of the tail (~3mm) were collected for genotyping. RNA was extracted using TRI reagent (Sigma) according to the manufacturer’s instructions. Purity and concentration of RNA was assessed by measuring absorbance at
A 260 and A280. One microgram of RNA was reverse transcribed using the SuperScript ViLo cDNA synthesis kit (Invitrogen) in a 20l reaction according to the manufacturer’s instructions. Real time PCR was performed using the LightCycler 480 Probes Master kit and the Taqman Gene Expression Assay Mm00440525_m1 (Notch4, FAM labelled) and Taqman Gene Expression Assay Mm01205647_g1 ( -actin, VIC labelled) according to the manufacturer’s instructions in a single tube 20 L reaction using 2 L of cDNA.
2.2.8 Northern Blotting Probe Preparation 25 μ g of pBSEP was linearised with EcoRI in a 200 μ L reaction volume. The linearised DNA was purified by phenol extraction. An equal volume of phenol:chloroform:isoamyl alcohol (25:24:1, equilibrated with TE) was added and vortexed vigorously for 30 seconds followed by centrifugation at 20000g for 2 minutes. The
65 | Page
aqueous phase was transferred to a fresh tube and extracted twice with chloroform by adding an equal volume of chloroform, vortexing for 30 seconds and centrifugation at 20000g. The DNA was precipitated by adding 1/10 volume of 3M sodium acetate (pH 5.2) and 2 volumes of ethanol followed by centrifugation at 20000g for 15 minutes. The resulting pellet was washed in 70% ethanol, air dried and resuspended at 0.5mg/mL in RNAase free water. In vitro transcription was subsequently performed in a 50μ L reaction containing 2.5 μ g template DNA, 40 U RNAsin (Promega), 1x T7 (Ambion), 0.25 U of T7 polymerase as appropriate, 0.5mM each of GTP, ATP and CTP and 62.5pM EasyTides Uridine 5’-triphosphate - 32P (PerkinElmer). The reaction was incubated at 37⁰ C for 1 hour followed by the addition of 25mM EDTA. Free nucleotides were removed using illustra MicroSpin G-50 columns according to the manufacturer’s instructions.
Northern Blotting Neonatal mice at 5 days post birth (P) were sacrificed and their lungs removed. The lungs were washed in cold PBS and snap frozen in a dry ice/ethanol bath. The tissue was stored at -80 until used. Concurrently the tips of the tail (~3mm) were collected for genetyping. RNA was extracted from the tissue using TRI reagent according to the manufacturer’s instructions. 10ug of RNA was heated (55⁰ C for 15 minutes) in RNA loading dye then separated on a 0.7% agarose gel containing 1xMOPS buffer and 3.7% formaldehyde overnight at 1.5V/cm2 . The 18S and 28S rRNA bands were visualised on a UV tansilluminator and photographed. The RNA was transferred by capilliary transfer in 10xSSC to Hybond N+ membrane overnight at room temperature. The membrane was rinsed in 2xSSC, air dried and the RNA crosslinked using a UV crosslinker. The membrane was prehybridised in northern hybridisation buffer for 3 hours at 60°C. The labelled probed was added (1/1000) after heating at 60⁰ C for 10 minutes in Northern hybridisation buffer. The membrane was incubated over night at 60⁰ C. The membrane was washed sequentially 3 times in 2xSSC, 0.1%SDS for 15 minutes at room temperature, 2 times in 0.2xSSC, 0.1%SDS at 60°C for 15 minutes and 2 times in 0.2XSSC, 0.1%SDS at 68°C. The washed membrane was dried and exposed to X-ray film overnight at -80°C. 2.2.9 N41437 Antibody purification A peptide of sequence VLQLIRGGC conjugated to ovalbumin was used to immunise guinea pigs. The resulting anti-sera and peptide linked via its cysteine residue to column matrix were used to purify antibody N41437 (Peptide Specialty Laboratories GmbH). Serum was diluted 1:2 in PBS and incubated with peptide column matrix overnight at 4°C. The matrix was loaded into a column and washed with 3x10mL of PBS, 2x10mL sodium phosphate (pH 6.8) and eluted with 8x500µl 0.1M glycine (pH 2.4) into 8x 35µl of 2M
K 2 HPO4 . The absorbance of each fraction was measured at 280nm and the two peak fractions combined. The purified antibody was diluted 1:2 in glycerol and stored at -20°C.
66 | Page
2.2.10 Western Blotting The NuPAGE Pre-Cast polyacrylamide gel system and reagents from Invitrogen were used for protein gel electrophoresis. The protein concentration in samples was measured using the BCA kit with the supplied BSA standard. Samples were prepared by heating at 70⁰ C for 2 minutes in a one tenth volume of LDS buffer containing 10% 2-mercaptoethanol and 50g/lane were loaded on either NuPAGE Bis-Tris 4-12% Pre-Cast gradient gels or NuPAGE Tris-Acetate Pre-Cast 4-8% gradient gels. The gels were run in optimised buffers (NuPAGE MOPS and NuPAGE TA Running Buffers) according to the manufacturer’s instructions. Lane one contained All Blue precision plus protein molecular weight markers. The gels were transferred to PVDF membranes in NuPAGE transfer buffer containing 10% methanol at 30V for 60 minutes. The membranes were blocked in 5% skim milk in TTBS for at least 1 hour at room temperature. Primary antibody was added in TTBS containing 1% skim milk and incubated with agitation overnight at 4°C. The membranes were washed on an orbital shaker with TTBS for four times 15 minutes at room temperature. The secondary antibody was added diluted in TTBS/1% skim milk and incubated with agitation at room temperature for 1 hour followed by four 15 minute washes in TTBS. For HRP detection the membranes were incubated for 5 minutes in Supersignal West Pico Chemiluminescent Substrate then exposed to X-ray film. For fluorescent detection membranes were rinsed in water and air dried before detection on an Odessey infrared scanner. 2.2.11 Immunoprecipitation C2C12 cells and C2C12 cells stably expressing Notch1myc were transfected with Notch4HA, Notch4d1HA or pCMXCAT expression vectors. After overnight incubation the culture media was aspirated and the cells washed twice in PBS prior to scraping into microfuge tubes. Cells were lysed with ice cold RIPA buffer containing 1x complete protease inhibitors (Roche) and 1mM PMSF on ice for 30 minutes with intermittent physical dissociation of cells using a pipette. The lysate was cleared of cellular debris by centrifugation at 20000g for 15 minutes at 4°C. A fraction of the supernatant was kept as an input control, with the rest incubated with either anti- HA or IgG control gently rotating overnight at 4°C. Subsequently, 50μ L of protein G sepharose beads (Invitrogen) was added per 1 mL of the lysate/antibody mixture and allowed to rotate gently for 2 hours at 4°C. The slurry was then allowed to settle by gravity in ice. The pellet containing the antibody-mediated pull down of protein complexes was washed four times with RIPA buffer at 4°C. Following the last wash lysates were centrifuged at 20000g for 5 minutes at 4°C. The pellet was resuspended in LDS loading buffer containing 10% 2-mercaptoethanol and heat denatured at 70⁰ C for 10 minutes. Western blotting was performed as described above. 2.2.12 Biotinylation NIH3T3Notch4HA and NIH3T3Notch1HA cells were seeded into 60mm CellBind tissue culture dishes, grown overnight, then washed three times with PBS containing 0.1mM CaCl2 and 1mM
67 | Page
++ MgCl2 (PBS ). The cells were labelled with 0.5mg/mL EZ-Link Sulfo-NHS-SS-Biotin (Thermo Scientific) in PBS++ for 2 hours at 4°C. The reaction was quenched by incubating the cells in DMEM for 10 minutes at 4°C followed by three washes in ice cold PBS++. The labelled cells were harvested and lysed for 30 minutes on ice in RIPA buffer containing 1mM PMSF and 1x complete protease inhibitors (Roche). The lysate was cleared by centrifugation at 4°C at 20000g for 30 minutes. The protein concentration was determined using the Pierce BCA Protein Assay Kit and 50g stored at -20°C as a load control. The remaining lysate was incubated overnight at 4°C with 20µl streptavidin (SA) agarose beads (Thermo Scientific) or control Protein G agarose beads (Invitrogen). Following incubation the beads were allowed to settle under gravity on ice and the supernatant removed. The beads were washed three times in RIPA buffer followed by one wash in 150mM NaCl. Following the last wash lysates were centrifuged at 20000g for 5 minutes at 4°C, subsequently the pellet was resuspended in LDS loading buffer and heat denatured at 70⁰ C for 10 minutes. Western blotting was performed as described above.
2.2.13 Luciferase Assays
2.2.13.1 Luciferase Activity Luciferase assays were performed using the Dual Luciferase Kit according to the manufacturer’s instructions. Cell cultures in 24 well plates were washed once with PBS and lysed in 100µl of passive lysis buffer (Dual luciferase kit) for 20 minutes at room temperature on a rocking platform. 20µl samples of the lysate were added to 96 well luminescence plates. The luminescence was read under the following conditions. LAR reagent (Dual luciferase kit), was added at 50µl per well and the luminescence averaged over 4 seconds, 1 second lag time followed by addition of 50µl of Stop and Glow reagent (Dual luciferase kit) and the luminescence averaged over 4 seconds with a 1 second lag time. The units are expressed as the average luminescence measured during the LAR reaction (due to firefly luciferase activity) divided by the average luminescence measured during the stop and glow reaction (due to the Renilla luciferase).
2.2.13.2 Co-culture Assays Cells were seeded at 5x104 cells per well into 24 well tissue culture plates and grown overnight. The cells were transfected with the plasmid combinations indicated in Table 2.4. After 6 hours the transfection mixes were aspirated and 5x104 of the ligand presenting cells indicated in the figures were added. The co- cultures were cultured overnight and then assayed for luciferase activity.
68 | Page
2.2.13.3 EDTA Assays NIH3T3Notch1Myc and NIH3T3Notch4HA cells were co-transfected with pGL46xTP1 and pCMXren in a 10:1 ratio and cultured overnight. The transfected cells were incubated in 0.5mM EDTA in PBS for 5 minutes at 37°C, washed in PBS++ then returned to complete media and incubated at 37°C overnight. The treated cells were then assayed for luciferase activity.
2.2.13.4 NotchECT assays NIH3T3 cells were co-transfected with the plasmid combinations indicated in Table 2.4 and cultured overnight. The cells were then assayed for luciferase activity.
Table 2.4 Transfection mixes for luciferase assays.
Figure 3.2 vector Notch1HA Notch4HA Notch4 pCMXCAT 250ng pCMXNotch1HA 250ng pCMXNotch4HA 250ng pCMXNotch4 250ng pCMXren 25ng 25ng 25ng 25ng pGL46xTP1 225ng 225ng 225ng 225ng Figure 3.3a vector Notch1myc Notch4HA pCMXCAT 400ng pCMXNotch1myc 400ng pCMXNotch4HA 400ng pCMXren 40ng 40ng 40ng pGL46xTP1 360ng 360ng 360ng Figure 3.3b vector Notch1myc Notch4HA pCMXCAT 250ng pCMXNotch1myc 250ng pCMXNotch4HA 250ng pCMXren 25ng 25ng 25ng pGL46xTP1 225ng 225ng 225ng Figure 3.6 vector Notch1ECT (25-200ng) pCS2+MT 250ng 225ng 200ng 150ng 50ng Notch1ECT 25ng 50ng 100ng 200ng pCMXren 25ng 25ng 25ng 25ng 25ng pGL46xTP1 225ng 225ng 225ng 225ng 225ng Notch4ECT (25-200ng) pCS2+MT 225ng 200ng 150ng 50ng Notch4ECT 25ng 50ng 100ng 200ng pCMXren 25ng 25ng 25ng 25ng pGL46xTP1 225ng 225ng 225ng 225ng Figure 3.7 vector cells Notch1HA Notch4HA cells cells pCMXren 50ng 50ng 50ng pGL46xTP1 450ng 450ng 450ng Figure 3.12c vector Notch4HA Notch4 RQRR pCMXCAT 250ng pCMXNotch4HA 250ng pCMXRQRR 250ng 69 | Page
pCMXren 25ng 25ng 25ng pGL46xTP1 225ng 225ng 225ng Figure 3.14a vector Notch1myc Notch4HA Notch1:1 Notch4:4 pCMXCAT 250ng pCMXNotch1myc 250ng pCMXNotch4HA 250ng pCMXNotch1:1 250ng pCMXNotch4:4 250ng pCMXren 25ng 25ng 25ng 25ng 25ng pGL46xTP1 225ng 225ng 225ng 225ng 225ng Figure 3.14b Notch1:1 Notch1:4 Notch4:1 Notch4:4 Notch1:1 250ng Notch1:4 250ng Notch4:1 250ng Notch4:4 250ng pCMXren 25ng 25ng 25ng 25ng pGL46xTP1 225ng 225ng 225ng 225ng Figure 4.1a vector Notch1HA (25ng) Notch4HA (25-200ng) pCMXCAT 225ng 200ng 175ng 125ng 25ng pCMXNotch1HA 25ng 25ng 25ng 25ng 25ng pCMXNotch4HA 25ng 50ng 100ng 200ng pCMXren 25ng 25ng 25ng 25ng 25ng pGL463xTP1 225ng 225ng 225ng 225ng 225ng Figure 4.1b vector Notch1HA (25ng) Notch4HA (25-200ng) pCMXCAT 225ng 200ng 175ng 125ng 25ng pCMXNotch1HA 25ng 25ng 25ng 25ng 25ng pCMXNotch4HA 25ng 50ng 100ng 200ng pCMXren 25ng 25ng 25ng 25ng 25ng pGL463xTP1 225ng 225ng 225ng 225ng 225ng Figure 4.1c vector Notch1myc (125ng) Notch4HA (125-375ng) pCMXCAT 400ng 275ng 150ng 25ng 275ng pCMXNotch1myc 125ng 125ng 125ng 125ng pCMXNotch4HA 125ng 250ng 375ng pCMXren 40ng 40ng 40ng 40ng 40ng pGL46xTP1 360ng 360ng 360ng 360ng 360ng Notch4HA (125-375ng) pCMXCAT 375ng 150ng 25ng pCMXNotch4HA 125ng 250ng 375ng pCMXren 40ng 40ng 40ng pGL46xTP1 360ng 360ng 360ng Figure 4.2 vector Notch1ECT (25-200ng) pCMXCAT 250ng 225ng 200ng 150ng 50ng pCMXNotch1ECT 25ng 50ng 100ng 200ng pCMXNotch4ECT pCMXren 25ng 25ng 25ng 25ng 25ng pGL46xTP3 225ng 225ng 225ng 225ng 225ng Notch4ECT (25-200ng) pCMXCAT 250ng 225ng 200ng 150ng pCMXNotch1ECT pCMXNotch4ECT 25ng 50ng 100ng 200ng 70 | Page
pCMXren 25ng 25ng 25ng 25ng pGL46xTP3 225ng 225ng 225ng 225ng Notch1ECT (25ng) Notch4ECT (25-200ng) pCMXCAT 200ng 175ng 125ng 25ng pCMXNotch1ECT 25ng 25ng 25ng 25ng pCMXNotch4ECT 25ng 50ng 100ng 200ng pCMXren 25ng 25ng 25ng 25ng pGL46xTP3 225ng 225ng 225ng 225ng Figure 4.3 Notch4ECT (25-200ng) pCMXCAT 225ng 175ng 150ng 100ng 25ng pCMXNotch4ECT 25ng 50ng 100ng 200ng pCMXNotch1HA 25ng 25ng 25ng 25ng 25ng pCMXren 25ng 25ng 25ng 25ng 25ng pGL46xTP1 225ng 225ng 225ng 225ng 225ng Figure 4.4a vector Notch4HA (25-200ng) pCMXCAT 250ng 225ng 200ng 150ng 50ng pCMXNotch4HA 25ng 50ng 100ng 200ng pCMXren 25ng 25ng 25ng 25ng 25ng pGL46xTP1 225ng 225ng 225ng 225ng 225ng Figure 4.4b vector Notch4HA (25-200ng) pCMXCAT 250ng 225ng 200ng 150ng 50ng pCMXNotch4HA 25ng 50ng 100ng 200ng pCMXren 25ng 25ng 25ng 25ng 25ng pGL46xTP1 225ng 225ng 225ng 225ng 225ng Figure 5.13 Notch1HA Notch4HA Notch4d1 Notch1HA Notch1HA Notch4HA Notch4d1HA pCMXCAT 225ng 50ng 50ng 25ng 25ng pCMXNotch1HA 25ng 25ng 25ng pCMXNotch4HA 200ng 200ng pCMXNotch4d1HA 200ng 200ng pCMXren 25ng 25ng 25ng 25ng 25ng pGL46xTP1 225ng 225ng 225ng 225ng 225ng
2.2.14 Immunocytochemistry (IC)
2.2.14.1 Fixed cells Cells were grown on coverslips and transfected where appropriate using standard conditions. Cells were rinsed twice with 1x PBS++ and fixed for 15 minutes in 4% PFA. The fixative was removed and excess aldehyde groups were quenched by incubating the cells in 150mM glycine in PBS for 15 minutes at room temperature. Glycine was aspirated, the cells rinsed twice in PBS++ and then blocked for one hour in blocking solution (5% BSA, 0.3% Triton X-100 and 10% goat serum in PBS). Subsequently, cells were incubated with primary antibody in blocking solution overnight at 4⁰ C. The cells were washed three times in PBS and incubated in the dark with fluorochrome-conjugated secondary antibody (and 50 M TO-PRO-3 where appropriate) in blocking solution for 60 minutes. Cells were washed three times in PBS. In addition, where appropriate, cells were stained in 1 g/mL Hoechst 33342 (Invitrogen)/PBS++ for 10 71 | Page
minutes at room temperature in the dark, and then washed once in PBS++. Coverslips were mounted in ProLong antifade mounting medium (Molecular Probes) and allowed to dry in the dark overnight prior to sealing with clear nail polish. Confocal images were taken on a LSM 700 upright confocal microscope (Zeiss) using Zen software. Brightness and contrast were adjusted in Adobe photoshop for printing purposes. All images were adjusted using identical conditions.
2.2.14.2 Live imaging Performed by Gavin Chapman in our laboratory. C2C12Notch1GFP cells were seeded onto coverslips and the following day transfected with pEFBosNotch4mRuby. GFP (green) and mRuby (red) fluorescence was excited in live cells using 488 and 561 nm lasers, respectively. The image was captured with a 63x 1.4 NA objective on an Axio-Observer equipped with a 710 scan head (Zeiss).
2.2.15 Immunohistochemistry (IHC)-Retina staining Mouse pups were collected at 3, 5 and 7 days post birth (P3, P5 and P7 respectively). Mice were sacrificed by decapitation and tail clips taken for genotyping. Eyes were enucleated and fixed in 4% PFA at 4°C for 10 minutes on ice. An incision was made using dissecting forceps just below the lens and the pigmented epithelium and outer layers removed. The vitreous and associated hyaloid vasculature was removed and the dissected retina fixed again in 4% PFA for 2 hours at 4°C followed by three 5 minute washes in PBS/0.3% Triton X-100 (PBST).
Retinas were then either processed immediately or dehydrated by sequential washes in 25% methanol in PBS, 50% methanol in PBS, 75% methanol in PBS followed by three washes in 100% methanol and storage at -20°C. To rehydrate stored retina prior to staining they were passed through methanol washes of 75%, 50% and 25% methanol in PBS followed by three washes in PBST.
Non-specific binding was blocked by incubation in PBST containing 5% donkey serum for at least 2 hours at room temperature. Anti- PECAM (1/200) and anti-SMA (1/100) were added and incubated with gentle agitation overnight at 4°C. The retina were washed six times for 15 minutes in PBST then incubated in the anti-mouse488 and anti-RatCy3 diluted 1/500 in PBST containing 5% donkey serum for 2 hours at room temperature in the dark. Following six 15 minute washes in PBST at room temperature, four radial cuts were made in the retina to allow it to be laid flat on a microscope slide. The retina were then mounted in ProLong antifade mounting medium (Molecular Probes) and allowed to dry in the dark overnight prior to sealing with clear nail polish.
72 | Page
A 7x7 (P7 mice) or 6x6 (P3 and P5 mice) tiled z-stack was captured for each retina using a 7Duo confocal microscope (20x objective) and Zen software. The samples were concurrently excited with a 405nm and 555nm laser. The tiled stacks were flattened in the Z dimension using the average intensity function in ImageJ. After setting a threshhold, kept constant for each litter, the binary image of the Cy3 channel was traced using the image tracing tool and the area calculated in ImageJ. The vascular density was calculated as the number of positive pixels/area. The retinal radius was measured on the alexa fluor 488 channel using the line tool. Representative images were adjusted for brightness and contrast in Adobe Photoshop for printing purposes. All samples were adjusted using identical conditions.
2.2.16 Wholemount RNA in situ hybridisation Probe synthesis 25 μ g of pYXNotch4, pBSEP (5’ probe) and pBSXN (3’ probe) were linearised with EcoRI, EcoRI and AflII respectively in a 200 μ L reaction volume. The linearised DNA was purified by phenol extraction. An equal volume of phenol:chloroform:isoamyl alcohol (25:24:1, equilibrated with TE) was added and vortexed vigorously for 30 seconds followed by centrifugation at 20000g for 2 minutes. The aqueous phase was transferred to a fresh tube and extracted twice with chloroform by adding an equal volume of chloroform, vortexing for 30 seconds and centrifugation at 20000g. The DNA was precipitated by adding 1/10 volume of 3M sodium acetate (pH 5.2) and 2 volumes of ethanol followed by centrifugation at 20000g for 15 minutes. The resulting pellet was washed in 70% ethanol, air dried and resuspended at 0.5mg/mL in RNAase free water. In vitro transcription was subsequently performed in a 50μ L reaction containing 2.5 μ g template DNA, 1x Transcription buffer (T3 for pYXNotch4, T7 for pBSEP, T7 for pBSXN (Ambion)), 0.25U of T7 or T3 polymerase as appropriate, 0.5mM each of GTP, ATP and CTP, 0.32mM UTP and 0.18mM DIG-11-rUTP (Roche). The reaction was incubated at 37⁰ C for 2 hours. Chromaspin 100 DEPC columns (Clontech) were centrifuged at 500 g for 3 minutes. The samples were loaded onto the columns and the products were collected in RNase-free tubes by centrifugation at 500g for 5 minutes. Probes were stored in aliquots at -80⁰ C.
Wholemount RNA in situ hybridisation Embryos were dissected as described in (Hogan et al., 1994). In brief, after the detection of the vaginal plug, pregnant females were sacrificed on day E10.5. Females were dissected to reveal the uterine horns, which were then placed in M2 media. The deciduas were removed from the uterus by making longitudinal tears, adjacent to each decidua. The embryo was freed by scoring the adjacent tissue around it. The embryo was removed from the yolk sac and a small piece kept for genotyping. Embryos were fixed in 4% PFA overnight at 4°C then dehydrated by washing in PBS, 25%, 50%, 75% and 100% methanol in PBS for 10 minutes 73 | Page
each at room temperature then stored at -20°C. When ready to proceed, embryos were rehydrated by washing in 75, 50 and 25% methanol in PBS, and subsequently washed twice in PBT. Embryos were bleached for 1 hour in 6% H2O2 in PBS after which they were washed three times in PBT and then incubated in 10g/mL of Proteinase K in PBT for 20 minutes. The digestion was stopped by washing the embryos in freshly prepared 2 mg/mL glycine in PBS. After washing the embryos twice in PBT the embryos were refixed in 0.2% gluteraldehyde/4% PFA in PBS for 20 minutes, followed by an additional two washes in PBT. The embryos were blocked in in situ prehybridisation solution at 70⁰ C for one hour, after which they were incubated overnight at 70⁰ C in in situ hybridisation solution containing denatured probe. The following day embryos were washed, twice in in situ wash solution I for 30 minutes at 70⁰ C, once in a one-to-one ratio of in situ wash solution I and II for 10 minutes, and three times in in situ wash solution II. Embryos were incubated in in situ wash solution II containing 100 μ g/mL of RNaseA for 30 minutes at 37⁰ C. Following this, the embryos were washed once in both in situ wash solution II and in situ wash solution III. Subsequently they were washed in in situ wash solution III for 30 minutes at 65⁰ C, and then three times in TBST plus 0.5mg/mL levamisole before blocking in 10% sheep serum in TBST for 2 hours. The embryos were then incubated overnight in a 1 in 2000 dilution of anti-DIG-AP FAB fragments in 1% sheep serum in TBST. The following day the embryos were washed 12 times in TBST for approximately 30 minutes/wash with a final wash overnight at 4⁰ C. On the following morning the embryos were washed three times in NTMT for 10 minutes and stained by incubating at room temperature in NBT/BCIP diluted 1/50 in NTMT in the dark for 6 hours. The reaction was stopped by washing the embryos twice in NTMT and PBT, and subsequently fixed overnight in 4% PFA/0.1% gluteraldehyde at 4⁰ C before being stored in 0.1% PFA/PBT. All washes were for 5 minutes at room temperature unless otherwise indicated.
74 | Page
3 NOTCH4 is not activated by ligand
3.1 Introduction
3.1.1 Notch4 and the canonical Notch pathway Canonical Notch signalling (Section 1.2.3) is the major Notch dependent signalling route and is important in most situations where Notch is known to be involved. The canonical pathway describes the activation of Notch receptors via Delta/Serrate/LAG- 2 (DSL) ligands, which culminates in the release of the intracellular domain and RBPJ (also known as CBF1/Su(H)/LAG-1 (CSL)) dependent activation of Notch target genes. Ligand binding results in a sequence of protease cleavages of the receptor which releases the Notch intracellular domain (NotchICD) from the plasma membrane. NotchICD translocates to the nucleus where it binds RBPJ and activates transcription. The canonical Notch pathway has largely been elucidated through studies of Drosophila Notch and mammalian Notch1. The biochemistry of the remaining mammalian Notch receptors has for the most part been assumed to be identical to NOTCH1.
The major site of Notch4 expression is in the vascular endothelium (Uyttendaele et al., 1996). This, and the fact that vascular expression of the activated form of NOTCH4, NOTCH4ICD, in transgenic mice results in vascular abnormalities (Carlson et al., 2005; Miniati et al., 2009; Murphy et al., 2008; Uyttendaele et al., 2001) points to an important role for Notch4 in angiogenesis. Our interest in angiogenesis has led us to investigate the functions of NOTCH4. Notch4 is the most divergent and least well characterised of the Notch family of receptors. The aim of this project was to investigate signal transduction through the NOTCH4 receptor. We wished to distinguish between what was known of the constitutively active oncogenic form of NOTCH4 (initially referred to as Int-3) (Gallahan and Callahan, 1997; Gallahan et al., 1987) and the native receptor. This is a necessary comparison as over expressed activated fragments of NOTCH4 do not recapitulate all aspects of physiological signalling. In order to do this we needed to activate NOTCH4 in a more physiologically relevant manner. Ligand induced activation is the pivotal step that distinguishes the full length receptor from constitutively active constructs. Below I will summarise what is known specifically about NOTCH4 ligand induced activation i.e. the canonical Notch signal transduction pathway.
3.1.2 Extracellular domain structure and ligand binding The extracellular regions of Notch receptors contain a large number of epidermal growth factor (EGF)-like repeats. EGF-like repeats are evolutionary conserved domains commonly found in 75 | Page
the extracellular region of membrane proteins. An EGF-like repeat is 30-40 amino acids long with six cysteine residues involved in three disulfide bonds (Jafar-Nejad et al., 2010). NOTCH4 is the most highly divergent Notch receptor containing only 29 EGF-like repeats as opposed to 36 in NOTCH1 and NOTCH2 and 34 in NOTCH3. The EGF-like repeats in analogous regions of Notch paralogues show greater similarity to each other than EGF-like repeats within a single receptor. Thus a number of EGF-like repeats of NOTCH4 have been found to correspond to NOTCH1. The EGF-like repeats 1 to 13, 22-24 and 26-29 of NOTCH4 correspond to EGF-like repeats 1-13, 28-30 and 33-36 of NOTCH1 respectively (Uyttendaele et al., 1996). Although the remaining repeats show a more limited similarity to other Notch family members there is some similarity when the N- and C-terminal portions of these repeats are analysed separately. These novel domains retain the critical spacing of the cysteine residues within the EGF-like repeat and support a model where five of the repeats are derived from fusions of repeat 14 and 15, 16 and 17, 20 and 23, 26 and 27 and 31 and 32 (Gallahan and Callahan, 1997). A similar scenario has been proposed for Notch3 which contains a novel EGF-like repeat derived from a fusion of repeats 2 and 3 (Lardelli et al., 1994).
It has been shown that the evolutionarily conserved EGF-like repeats 11 and 12 are alone sufficient for ligand binding (de Celis et al., 1993; Lawrence et al., 2000; Rebay et al., 1991; Shimizu et al., 1999). EGF-like repeats 11 and 12 of Notch4 have not been tested for ligand binding capacity but are more highly conserved than the protein as a whole, consistent with a functional role (58% identical to Notch1 compared to 43% identical to the protein as a whole). However, EGF-like repeats 11 and 12 of NOTCH4 are still the most divergent in the Notch family, with NOTCH2 being 70% identical to NOTCH1 and NOTCH3 being 73% identical to NOTCH1 over repeats 11 and 12. Although repeats 11 and 12 alone are sufficient for ligand binding, other regions of the EGF-like repeat domains influence binding (Lawrence et al., 2000; Perez et al., 2005; Powell et al., 2001; Ramain et al., 2001) (Section 1.2.5). The evidence for co-operation by other EGF-like repeats in binding comes from analysis of the fucose-specific β 1,3-N- acetylglucosaminyltransferase, Fringe, modification of Notch (Section 1.2.3.1). Modification by Fringe enhances Delta induced signalling and inhibits Serrate induced signalling (Chen et al., 2001; Evrard et al., 1998; Hicks et al., 2000; Moloney et al., 2000; Shimizu et al., 2001; Yang et al., 2005; Zhang and Gridley, 1998). This differential response to Delta and Serrate family members is retained when the modification site in EGF-like repeat 12 is mutated indicating that modification of other repeats affect ligand binding (Lei et al., 2003). Extensive mutagenesis of Fringe 76 | Page
modifiable sites showed that no individual site was required to modify binding (Xu et al., 2005). However, a small N-terminal region of the ligands was found to mediate the differential response to Fringe modification (Fleming et al., 1997). Thus parts of the Notch receptor widely separated in primary sequence contribute to ligand binding. This has led to a speculative model of the quaternary structure of Notch receptors (Xu et al., 2005) (Section 1.2.5). Of particular interest to this observation is the abruptex mutants found in Drosophila. Abruptex mutations cluster in EGF-like repeats 24-29 and cause a gain-of-function phenotype suggesting these repeats are involved in an inter-domain structure that affects interaction with ligand (de Celis and Garcia-Bellido, 1994; Kelley et al., 1987; Xu et al., 2005). NOTCH4 is highly divergent in this region having lost EGF-like repeats and generated new fusion repeats (Gallahan and Callahan, 1997; Uyttendaele et al., 1996). The divergence of these regions suggests that the inter-domain structure of NOTCH4 may differ from other Notch receptors. This divergence may have implications for its ability to bind and be activated by ligand.
3.1.3 Ligand binding and activation of NOTCH4 The divergence of NOTCH4 from other family members led to the speculation that it may respond to a separate set of ligands (Gallahan and Callahan, 1997). However, no ligands other than Delta and Serrate family members have been reported for NOTCH4. In order for NOTCH4 to be activated by ligand the first obvious requirement is that the ligand and receptor be expressed by cells in contact. The major expression domain of Notch4, the vascular endothelium, overlaps with the expression domain of Dll4 and Jag1 (Section 1.2.2.1). Notch4 and Dll4 were found to be the only Notch receptor and ligand expressed in the capillaries (Villa et al., 2001). This pattern of tissue expression suggests that DLL4 and JAG1 are the major ligands for NOTCH4.
There are relatively few studies that have looked directly at ligand binding and/or activation of NOTCH4. Based on similar expression domains, DLL4 and JAG1 have been widely identified as the major NOTCH4 ligands. DLL4 has been shown to bind NOTCH4 in two reports. A soluble NOTCH4 protein containing all the EGF-like repeats could competitively inhibit antibody binding to DLL4 (Yamanda et al., 2009). Antibody binding was also blocked by soluble NOTCH1 protein indicating that the two receptors bound ligand in a similar way. The second independent confirmation that NOTCH4 can interact with DLL4 was in a yeast two-hybrid assay (Nichol et al., 2010). Although reports of direct binding by JAG1 with NOTCH4 are lacking, there are reports of JAG1 induced activation (see below). While these reports demonstrate that NOTCH4 can bind DLL4 they are not necessarily indicative of 77 | Page
signal activation. For example, soluble ligands that bind receptor can act as inhibitors of Notch signalling (Hicks et al., 2002; Nichols et al., 2007; Varnum-Finney et al., 2000).
The most compelling reports of direct canonical ligand mediated activation of NOTCH4 were generated in the Kitajewski laboratory (Funahashi et al., 2008; Shawber et al., 2003; Shawber et al., 2007). In one report, NOTCH4 was able to activate RBPJ dependent reporter transcription by 3.8-fold upon addition of DLL4 expressing cells (Shawber et al., 2003). NOTCH4 could also upregulate the Notch target genes Flt4, Hey1 and Hey2 in response to both DLL4 and JAG1 (Shawber et al., 2007). In contrast, NOTCH4 was not activated by canonical Notch ligands in a study which measured signalling from a NOTCH4-GAL4 activation domain fusion (Aste-Amezaga et al., 2010). The Notch4 construct was unable to respond to ligand in co-culture assays although similar constructs based on Notch1, 2 and 3 could (Aste- Amezaga et al., 2010). In addition, Notch4 is conspicuous by its absence when comparisons of differences in mammalian Notch receptors are investigated (Aste-Amezaga et al., 2010; Shimizu et al., 2002).
3.1.4 The NOTCH4 negative regulatory region (NRR) and surface presentation The NRR region contains three Lin-12/Notch repeat (LNR) domains, A, B and C, and the heterodimerisation domain (HDD). The NRR is essential in preventing the ligand independent activation of Notch. The LNR domains protect the S2 site from cleavage in the absence of ligand. Exposure of the S2 site requires the sequential removal of first the LNR-A then the linker between LNR-A and LNR-B and finally the LNR-B to expose the S2 site (Section 1.2.6).
The NRR also contains the site of FURIN cleavage (S1). The unstructured loop in which FURIN cleaves was identified in a structure of the NOTCH2 NRR (Gordon et al., 2007). S1 cleavage of Notch results in the N- and C-terminals remaining non- covalently bound and this form of Notch is referred to as the Notch heterodimer. The C-terminal fragment that contains the t ransmembrane and intracellular domains is referred to as the TMIC. The Notch heterodimer, in the case of NOTCH1 is generally considered the only form found on the cell surface (Blaumueller et al., 1997; Logeat et al., 1998). There have been some conflicting reports using over expression systems (Bush et al., 2001). However, for NOTCH1, 2 and 3 the vast majority (90-95%) of surface protein is in the form of the heterodimer. In contrast to other family members no identifiable FURIN cleavage site exists in NOTCH4. However, proteolytic processed peptides of NOTCH4 that
78 | Page
are approximately the expected size of NOTCH4TMIC are seen (Shawber et al., 2003). In contrast to the other family members the full length peptide and not the NOTCH4TMIC is the major species found on the cell surface (Shawber et al., 2003; Wu and Bresnick, 2007b). These results show that NOTCH4 is divergent to other Notch family members in the regulation of surface expression levels. This is particularly important in the canonical Notch pathway. In contrast to many other signalling pathways there are no second messengers or enzymatic amplification of the signal received by the receptor. One ligand engaged receptor can, at a maximum, deliver one molecule to the nucleus. This makes regulation of surface expression of Notch receptors of critical importance and the divergence of NOTCH4 in this respect of particular interest.
3.1.5 NOTCH4ICD The divergence of the extracellular domain points to differences in ligand induced activation of the NOTCH4 receptor. The final steps in ligand induced activation require the release, nuclear translocation and the formation of an active transcription complex by the intracellular domain. NOTCH4 is also highly divergent in this respect. The divergence of the NOTCH4ICD with respect to canonical Notch signalling will be discussed below.
3.1.5.1 Structural Divergence of NOTCH4ICD The NotchICD contains a number of structural elements that are conserved across the family. The most important for RBPJ dependent gene activation have been identified as the Rbpj- a ssociated molecule (RAM) domain and the Ankyrin repeats. These are the most highly conserved regions of the NOTCH4ICD. Isolated RAM domains of each mammalian Notch receptor, including that of NOTCH4, has been demonstrated to bind to the beta trefoil domain of RBPJ with very similar affinities (Lubman et al., 2007). In addition, the region between the RAM domain and Ankyrin repeats of NOTCH4 maintains the critical 70 residue spacing between the domains required for a productive interaction with RBPJ (Section 1.2.8). It has been demonstrated, using Notch4ICD constructs in tissue culture and transgenic animals that the NOTCH4ICD does retain the ability to activate RBPJ bound promoters (Carlson et al., 2005; Ong et al., 2006). These studies use the constitutively active fragment which is expressed at a much higher level than would be potentially produced by ligand dependent activation of NOTCH4. Although NOTCH4 exhibits the basic interactions with the transcription complex, there are other areas of divergence.
Gamma-secretase cleavage occurs at valine 1463 in NOTCH4 and valine 1744 in NOTCH1 (Mumm et al., 2000; Saxena et al., 2001; Schroeter et al., 1998). Mutation of this residue severely reduced
79 | Page
NOTCH1ICD production and signalling. However, later studies found that the gamma-secretase cleavage site was variable in NOTCH1 although the NOTCH1ICDs with alternative N-termini were rapidly degraded (Tagami et al., 2007). The variability in gamma-secretase processing was proportional to the rate of endocytosis and was enhanced in endosomal as opposed to plasma membrane fractions. This indicated that these alternate short lived forms were cleaved in an alternative sub-cellular compartment to the signalling active valine 1744 NOTCH1ICD. Although these alternative forms may represent degradative intermediates, it highlights the control of NotchICD signalling by internalisation. Internalisation and gamma-secretase cleavage is controlled, in part, by the monoubiquitylation of lysine 1749 in the stop translocation sequence just inside the membrane. Monoubiquitylation can enhance internalisation and signalling (Gupta-Rossi et al., 2004). Mutation of this residue caused aberrant processing and rapid degradation of NOTCH1. Interestingly, the mutation in NOTCH1 (lysine-arginine) occurs naturally in NOTCH4 at this site. The presence of alternative gamma-secretase cleavage sites and alternative entry into the early endosome, have not been investigated in NOTCH4. However, the absence of this key residue, which is conserved in other family members, indicates that activation of NOTCH4 may be divergent.
3.1.5.2 Transactivation potential of NotchICD family C-terminal of the Ankyrin repeats is a poorly conserved terminal domain. In NOTCH1 and NOTCH2 this region acts as a transactivation domain when linked to the GAL4 DNA binding domain (Kurooka et al., 1998). In the same assay NOTCH3 and NOTCH4 demonstrated no transactivation activity. In a later study NOTCH3 was shown to have transactivation activity in this region but only on promoters with a proximal putative zinc finger transcription factor binding site (Ong et al., 2006). No transactivation activity has been identified for NOTCH4, which is highly divergent in this region. Size exclusion profiling of nuclear extracts from cells transfected with Notch1, 2, 3 and 4, NOTCH4 displayed a unique elution profile despite similarities between NOTCH1, 2 and 3 (Han et al., 2011). These results suggest that NOTCH4 may form unique complexes with similarly unique transcriptional results.
3.1.5.3 Degradation of NotchICD At the extreme C-terminal of the NotchICD is a PEST sequence, named because it is rich in proline, glutamine, serine and threonine residues (Rechsteiner and Rogers, 1996). PEST sequences enhance and control protein degradation. The PEST domain of NOTCH4 differs from that of the other receptors. The PEST sequence is targeted by FBXW7 (also known as SEL10), part
80 | Page
of an E3 ubiquitin ligase complex that mediates ubiquitylation and proteosomal degradation of the NOTCH1ICD (Fryer et al., 2004; Wu et al., 2001). Although NOTCH4ICD was efficiently bound and ubiquitylated by FBXW7 it was not as efficiently degraded as NOTCH1. FBXW7 binds preferentially to phosphorylated forms of the NotchICD and this phosphorylation is mediated by the CCNC:CDK8 complex. The CCNC:CDK8 complex is recruited to the NotchICD via an interaction with MAML indicating that this process is controlled by the formation of an active transcription complex. Additional phosphorylation sites have been identified in Notch1ICD which control degradation. A sequence at the N-terminal end of the NOTCH1 PEST sequence, WSSSSP, is thought to prime the NOTCHICD for subsequent phosphorylation in the transcription complex (Chiang, 2006). This sequence is highly conserved in the Notch family of receptors (100% identical in Drosophila, Xenopus and mammalian NOTCH1 and NOTCH2 and slightly divergent in NOTCH3 (WASPSP)). The complete absence of any sequence resembling this motif in the NOTCH4 again highlights a divergence in the regulation of the NOTCH4ICD.
3.1.6 Aims Notch4 is the most divergent and least well characterised of the Notch family of receptors. Some of the key interactions involved in canonical Notch signalling have been maintained in the NOTCH4 receptor. The ability to bind and be activated by Delta and Serrate family members is at least partly conserved though poorly described. In addition the final steps of the canonical Notch signal transduction pathway are also at least partly conserved, i.e. RBPJ dependent gene transcription. However, there are key areas where the NOTCH4 receptor is clearly divergent to the other family members. We wished to examine the canonical Notch pathway with respect to NOTCH4 to compare and contrast what is known about Notch receptors in general with NOTCH4 specifically. In order to do this we used co-culture experiments to induce ligand dependent activation and investigated the processing, cell surface presentation and transcriptional output of NOTCH4.
81 | Page
3.2 Results
3.2.1 Ligand dependent Notch activation - co-culture assays The canonical Notch signalling pathway is activated by DSL ligands. DSL ligands are transmembrane proteins that are presented on the cell surface; they bind and activate Notch receptors on adjacent cells. Therefore, signal transduction is restricted to cells in contact. Ligand: receptor interaction results in a series of proteolytic cleavages which release the Notch i ntracellular domain (NotchICD) from the membrane enabling its translocation to the nucleus where it activates transcription from RBPJ dependent promoters (Section 1.2.3). The first proteolytic cleavage after ligand binding (termed S2), occurs N-terminal to the plasma membrane. Artificial constructs that mimic this cleaved form of Notch (termed Notch extracellular truncation-NEXT) have been very successfully employed to characterise the Notch pathway post ligand binding (Mumm et al., 2000; Schroeter et al., 1998). However, this approach fails to model the physiological activation of Notch, which requires interaction with ligand presented on the surface of adjacent cells. In order to recreate the complete and physiologically relevant Notch pathway in vitro, a co- culture based luciferase assay was employed to measure Notch signalling induced by transmembrane ligand (Figure 3.1). Cells stably expressing DSL ligands are added to plates of Notch receptor expressing cells that have been transfected with a synthetic Notch responsive firefly luciferase reporter construct termed pGL46xTP1 (Figure 3.1). It is important to note that the receptor and ligand are expressed in different cell populations so that signalling occurs in trans. The reporter construct pGL46xTP1 contains six copies of the TP1 promoter sequence from Epstein Barr virus cloned into pGL4 and driving the expression of firefly luciferase. The TP1 sequence has previously been shown to specifically induce robust Notch dependent gene expression with very low constitutive levels of expression (Kato et al., 1996). To ensure that differences in reporter activity were not due to variable transfection or loading, pCMXren, a plasmid expressing Renilla luciferase from the CMV promoter was also transfected and used to normalise results.
NIH3T3 cells stably expressing DSL ligands were generated by transfection of pCAGiPuro plasmids containing ligand cDNAs. In these cell lines, ligand cDNAs are expressed with the puromycin resistance gene under the control of the CAG promoter, a fusion between the CMV early enhancer element and the chicken beta actin promoter that generates strong ubiquitous expression. The CAG promoter drives expression of a bicistronic message of the
82 | Page
ligand of interest, an internal ribosome entry site, and the puromycin resistance gene, so that resistance to puromycin is directly coupled to ligand cDNA expression (Miyahara et al., 2000). After transfection, cells were selected for puromycin resistance and single clones isolated. The lines used included a control line that had integrated the parent plasmid pCAGiPuro (referred to as control cells) and lines that stably over expressed either DLL4 (Dll4 cells), DLL4 with a C-terminal HA tag (Dll4HA cells) or JAG1 (Jag1 cells). These lines were maintained under selection with 1µg/ml puromycin to ensure continued expression of the ligand.
Notch1 with a C-terminal HA tag (Notch1HA) (constructed in our laboratory by Sharon Pursglove) or an internal myc tag (Notch1myc) (Nye et al., 1994) and Notch4 with (Notch4HA) and without a C-terminal HA tag (Notch4) were cloned into the transient expression vector pCMX (Umesono et al., 1991). cDNA expression was driven from the constitutive CMV promoter of cytomegalovirus. The entire open reading frame was sequenced to confirm that the Notch4 constructs matched the mouse reference sequence. The bacterial gene chloramphenicol transferase (CAT), which has no effect on Notch signalling, cloned into pCMX (Hoyne et al., 2011) was used as a vector only negative control (referred to as vector). Some experiments used lines generated by transfection with pCAGiPuro based constructs analogous to the ligand cells described above containing HA tagged versions of Notch1 and Notch4 and a myc tagged version of Notch1.
83 | Page
Figure 3.1 A co-culture assay to measure ligand dependent Notch activation. a. The co-culture assay depends on two distinct cell populations: NIH3T3 cells that stably over express ligand (ligand cells) and cells co-transfected with either vector or a Notch expressing plasmid in addition to the Notch responsive reporter pGL46xTP1 and the transfection control pCMXren (signalling cells). b. A constant number of ligand presenting or control cells are added to the receptor expressing cells and cultured overnight. c. Following overnight culture, cells are lysed and luciferase activity measured. The Notch responsive reporter expresses firefly luciferase (luc.) and the transfection control expresses Renilla luciferase (luc.). The luciferase activity is normalised by dividing the firefly luciferase activity by the Renilla luciferase activity.
84 | Page
85 | Page
3.2.2 NOTCH4 does not signal in response to ligand in co- culture assays The Notch family of receptors has four paralogues in mammals. All four receptors retain the same basic domain structure and are thought to transduce signal in a similar manner. NOTCH4 is by far the most divergent of the Notch receptors and the least well characterised. The field largely assumes, based on homology and the lack of an overt phenotype seen in the Notch4d1 mouse (Krebs et al., 2000), that NOTCH4 functions largely analogously to, and redundantly with NOTCH1. We performed co-culture assays to characterise the function of NOTCH4 and how it may diverge from NOTCH1.
NIH3T3 cells were co-transfected with Notch1HA, Notch4HA or the vector control, plus the Notch responsive reporter pGL46xTP1 and the Renilla luciferase transfection control, pCMXren. The transfected cells were co-cultured overnight with equal numbers of either Dll4 cells, Jag1 cells, or control cells (Figure 3.2). Co- cultures were lysed and the relative luciferase activity was measured and normalised for transfection efficiency by dividing the firefly luciferase, generated from pGL46xTP1, by the Renilla luciferase activity generated from pCMXren.
Notch1HA transfected NIH3T3 cells exhibited a small, though not statistically significant activation of the reporter upon co-culture with control cells (Figure 3.2, columns 1 and 2). This trend can be explained by the presence of low level expression of endogenous ligands in NIH3T3 cells. Co-culture of Notch1HA transfected cells with Dll4 or Jag1 cells induced robust activation of the reporter compared to either the vector transfected cells (P<0.001 for both Dll4 and Jag1, Figure 3.2 columns 5, 6, 9 and 10) or the Notch1HA transfected cells co-cultured with the control cells (P<0.001 for both Dll4 and Jag1, Figure 3.2 columns 2, 6 and 10). Although markedly different levels of activation (P<0.001) are seen in response to Dll4 cells compared to Jag1 cells (Figure 3.2 columns 6 and 10), no conclusions regarding the relative activation potential of either delta or serrate families can be drawn because Dll4 cells and Jag1 cells are likely to express different amounts of their respective ligands. In contrast to the results obtained for Notch1HA, transfection with Notch4HA did not activate the reporter when co-cultured with Dll4 cells, Jag1 cells or control cells (Figure 3.2 columns 3, 7 and 11). To ensure that the lack of activation observed was not due to the presence of the C-terminal HA tag, an untagged version of Notch4 was also assayed. No significant difference was observed between the tagged and untagged versions of Notch4 co-cultured with Dll4 cells, Jag1 cells or control cells (Figure 3.2 columns 3, 4, 7, 8, 11 and 12).
86 | Page
Figure 3.2 NOTCH4 did not signal in response to ligand in a co-culture assay.
NIH3T3 cells were co-transfected with Notch1 with a C-terminal HA tag (Notch1HA), Notch4, Notch4 with a C-terminal HA tag (Notch4HA) or a vector control (pCMXCAT), plus a Notch responsive reporter (pGL46xTP1) and a Renilla luciferase transfection control (pCMXren). The transfected cells were co- cultured overnight with either NIH3T3 cell lines over expressing DLL4 (Dll4 cells), JAG1 (Jag1 cells) or control cells. These data points represent relative luciferase units normalised for transfection efficiency by dividing by the Renilla luciferase activity. This figure represents the mean of three independent experiments with bars indicating the standard deviation. Using ANOVA and Tukeys’ post test experimentally relevant and statistically significant differences were found between columns: 2 and 6 (P<0.001), 2 and 10 (P<0.001), 5 and 6 (P<0.001), 9 and 10 (P<0.001) and 6 and 10 (P<0.001).
87 | Page
3.2.3 NOTCH4 does not signal in response to ligand in co- culture assays using a variety of cell types. Ligand activated NOTCH1 in NIH3T3 cells, while NOTCH4 was unresponsive in the same assay (Figure 3.2). This may indicate that in NIH3T3 cells NOTCH4 requires, or is inhibited by, factors that do not prevent NOTCH1 activation. To investigate if the lack of NOTCH4 signalling is cell line specific, two additional lines were analysed. Since both Dll4 and Jag1 ligand-expressing cells gave similar results, and Dll4 is co-expressed with Notch4 in vascular endothelial cells in vivo (Villa et al., 2001), Dll4HA cells were used as the ligand presenting cells in the majority of the following assays.
The murine arterial endothelial cell line (MAEC) was derived from mouse aortic endothelial cells, which are part of the endogenous expression domain Notch4 (Nishiyama et al., 2007). MAECs were transfected with Notch1myc, Notch4HA or the vector control in addition to the reporter (pGL46xTP1) and transfection control (pCMXren). After transfection the cells were co-cultured overnight with either Dll4HA cells or control cells. Similar to NIH3T3 cells, co-culture of Dll4HA cells with Notch1myc transfected MAECs led to robust Notch reporter induction (Figure 3.3a, columns 2 and 5). In agreement with the results obtained with NIH3T3 cells (Section 3.2.2) co-culture of Notch4HA transfected MAECs with Dll4HA cells did not activate the Notch reporter above the level of vector transfected MAECs (Figure 3.3a, columns 4 and 6).
A third cell line, C2C12, was also tested. C2C12 cells were derived from mouse myoblasts and are commonly used in assays assessing Notch activation (Kopan et al., 1994). C2C12 cells support active Notch signalling which prevents their myogenic differentiation indicating that they possess endogenous ligands and receptors (Chapman et al., 2006; Dahlqvist et al., 2003). Vector transfected C2C12 cells were able to activate the reporter by approximately 4-fold in response to Dll4HA stimulation (Figure 3.3b, columns 1 and 4). Again, expression of Notch1myc robustly activated reporter expression in response to Dll4HA ligand cells (Figure 3.3b, columns 2 and 5). In agreement with both NIH3T3 and MAECs, Notch4HA did not activate the Notch reporter in C2C12 cells upon co-culture with Dll4HA cells (Figure 3.3b, columns 3 and 6).
88 | Page
In the above assays, NOTCH4 was unable to activate the Notch reporter above background. In fact, expression of NOTCH4 resulted in lower reporter activity than the vector control in both the MAEC and C2C12 lines. This observation is further characterised in Chapter 4. The co-culture assay has the advantage of modelling the entire Notch signal transduction pathway. As such the lack of activation of NOTCH4 in response to ligand could be due to a divergence of function of NOTCH4 from NOTCH1 at any point within the pathway. NOTCH4, in comparison to NOTCH1, may not be well expressed and presented on the cell surface. Since activation of Notch receptors occurs in trans, a lack of surface expression would prevent interaction with ligand and hence signal transduction. NOTCH4 may also interact with ligand less efficiently once on the cell surface. Differences further downstream in the pathway may also explain the lack of NOTCH4 activation in response to ligand. The following experiments were conducted to identify where in the Notch signal transduction pathway NOTCH4 is not being activated.
89 | Page
Figure 3.3 NOTCH4 signalling is not induced by ligand co- culture assays using a variety of cell types. a. The murine arterial endothelial cell line (MAEC) was transfected with Notch1myc, Notch4HA or a vector control, plus a Notch responsive reporter (pGL46xTP1) and a Renilla luciferase transfection control (pCMXren). Transfected MAECs were co- cultured overnight with either a NIH3T3 cell line stably expressing DLL4 (Dll4HA cells) or control cells. These data points represent the relative luciferase units normalised for transfection efficiency by dividing by the Renilla luciferase activity. This figure represents the mean of four replicate transfections with bars as standard deviations. b. The cell line C2C12 was transfected with Notch1myc, Notch4HA or a vector control, plus a Notch responsive reporter (pGL46xTP1) and a Renilla luciferase transfection control (pCMXren). Transfected C2C12 cells were co-cultured overnight with either a NIH3T3 cell line expressing DLL4 (Dll4HA cells) or control cells. These data points represent the relative luciferase units normalised for transfection efficiency by dividing by the Renilla luciferase activity. This figure represents the mean of four replicate transfections with bars as standard deviations.
90 | Page
91 | Page
3.2.4 Gamma-secretase cleavage of Notch4ICD constructs. Notch receptors are kept in an inactive state prior to ligand binding by the presence of a negative regulator region (NRR). The folding of this region buries the S2 site where ADAM10/17 cleaves. Ligand binding induces a conformational change, which exposes the S2 site and allows cleavage (Gordon et al., 2007). Notch constructs that do not contain the NRR and sequences N-terminal to it are constitutively active because gamma-secretase cleavage of such truncated Notch molecules occurs independently of ligand (Jarriault et al., 1995; Kopan et al., 1996). A construct of this type, termed N1∆ E, was used to identify the gamma-secretase cleavage site of NOTCH1 (Schroeter et al., 1998). In order to isolate enough protein for N-terminal sequencing, N1∆ E was also truncated at the C-terminal to remove the PEST sequence responsible for rapid NotchICD degradation (Schroeter et al., 1998). The gamma-secretase site was identified as occurring at valine 1744 in mouse NOTCH1. This information was exploited in the generation of antibodies that recognise the neo-epitope produced following gamma-secretase cleavage. Such antibodies specifically recognise the S3-cleaved and hence activated form of NOTCH1ICD and have proven a valuable experimental tool to detect endogenous NOTCH1 signalling in tissues and the embryo (Chapman et al., 2011; Morimoto et al., 2005).
Using equivalent constructs to N1∆ E, the gamma-secretase cleavage site of NOTCH4 has been previously mapped to valine 1463 by N-terminal sequencing (Saxena et al., 2001). We sought to generate an equivalent anti-cleaved NOTCH4 antibody in order to detect NOTCH4 activation independently of other Notch family members. A peptide corresponding to the first six amino acid residues following position 1463 of NOTCH4ICD was fused to two glycines as a spacer and a cysteine for coupling (VLQLIRGGC, designed in our laboratory by Gavin Chapman). This peptide was used to immunise guinea pigs and antisera collected (Peptide Specialty GmbH). The antiserum was affinity purified and the resulting antibody was called N41463.
Although there is little to no sequence similarity at the gamma- secretase cleavage site of mouse Notch receptors (Figure 3.4a) all four mouse Notch receptors were analysed for cross-reaction with N41463. C2C12 cells were transfected with N∆ E constructs of each mouse Notch receptor (a gift from Raphael Kopan; (Saxena et al., 2001; Schroeter et al., 1998)). Lysates from cells transfected with the parent vector (Figure 3.4b, lane 1), N1∆ E (lane 2), N2∆ E (lane3), N3∆ E (lane4) and N4∆ E (lane5) were analysed by western blot using the N41463 antibody. In addition cells transfected with the N4∆ E were incubated with the gamma-secretase inhibitor, N-
92 | Page
[(3,5-Difluorophenyl)acetyl]-L-alanyl-2-phenylglycine-1,1- dimethylethyl ester (DAPT) to test if the antibody was specific for the neo-epitope generated by gamma-secretase cleavage. The antibody specifically recognised only a band of the correct size (~57kDa) in cells transfected with N4∆ E (Figure 3.4b, lane 5). No cross reaction was observed in cells transfected with the parental vector (Figure 3.4b, lane 1) or cells transfected with N1∆ E, N2∆ E or N3∆ E (Figure 3.4b, lanes 2, 3 and 4). DAPT treatment of cells transfected with N4∆ E abrogated antibody recognition (Figure 3.4b, lane 6) confirming that the antibody was indeed specific for the gamma-secretase cleaved form of NOTCH4.
93 | Page
Figure 3.4 The N41463 antibody reacts specifically with the neo-epitope produced by gamma-secretase cleavage of NOTCH4.
a. The N-terminal sequence of the four mouse Notch receptors after gamma-secretase cleavage. The sequence was confirmed for NOTCH1 and NOTCH4 by N-terminal sequencing and inferred by homology for NOTCH2 and NOTCH3 (Saxena et al., 2001). b. C2C12 cells were transfected with constitutively active Notch constructs based on all four mouse Notch receptors (N1∆ E, N2∆ E, N3∆ E and N4∆ E) and the parental empty vector (control). In addition, cells that had been transfected with N4∆ E were treated with DAPT (N4∆ EDAPT) overnight while the others were treated with an equivalent volume of vehicle (DMSO). The cell lysates were western blotted with the N41463 antibody followed by an HRP conjugated anti-guinea pig secondary antibody. 94 | Page
3.2.5 Gamma-secretase dependent entry of NOTCH4ICD to the nucleus. We have been unable to detect ligand dependent NOTCH4 activation (Section 3.2.1). The inability of NOTCH4 to activate our reporter could be due to a large number of factors. We wished to investigate and confirm where possible that NOTCH4 can in fact be activated in a manner similar to NOTCH1. The end point of our assay was transcription from the Notch responsive reporter pGL46xTP1. The ability of Notch receptors to induce transcription is dependent on nuclear entry of the NotchICD. The inability of NOTCH4 to transduce signal in response to ligand may be due to a lack of nuclear accumulation of NOTCH4ICD. The ability of the NOTCH4ICD to activate transcription, if produced, was investigated in the following experiments.
Immunocytochemistry was performed to determine if the N41463 antibody could specifically detect nuclear accumulation of NOTCH4ICD (Figure 3.5). In this and following experiments Notch constructs were made based on modification of those used in Figure 3.4. The original constructs were not only truncated at the extracellular domain but were also truncated at the C-terminal to aid detection and N-terminal protein sequencing (Saxena et al., 2001; Schroeter et al., 1998). To more closely mimic the peptide fragment produced after ligand binding, constructs for both Notch1 and Notch4 were made that contained the extracellular truncation (ECT) of the ∆ E constructs, in addition to an intact C-terminus and a C-terminal HA tag to aid detection (N1ECTand N4ECT, respectively). These constructs, along with the parental vector (Figure 3.5, panels 10-12) were transfected into NIH3T3 cells and co-stained with N41463 and mouse anti-HA. An anti-mouse Cy3 conjugated secondary antibody was used to detect the anti-HA antibody (shown in red) and a secondary anti-guinea pig Alexa Fluor488 was used to detect N41463 (green). Cells transfected with N4ECT displayed cytoplasmic and weak nuclear staining for HA (Figure 3.5, panel 1). In contrast, N41463 antibody staining was restricted to the nucleus, consistent with nuclear translocation of NOTCH4ICD (Figure 3.5, panel 2). The merged image (Figure 3.5, panel 3) indicated that the two channels were independent. Strong staining in the Cy3 channel (anti-HA) did not correlate with strong staining in the 488 channel (anti-N41463, Figure 3.5, panels 1, 2 and 3). Cells transfected with Notch1ECT (Figure 3.5, panels 4-6) stained with anti-HA (Figure 3.5, panel 4) indicating that Notch1ECT was expressed but no signal was observed from the N41463 antibody (Figure 3.5, panel 5). Thus, the N41463 antibody did not recognise gamma-secretase cleaved Notch1 by immunofluorescence, consistent with the western blot (Figure 3.4). HA staining was unchanged by treatment of N4ECT transfected cells with DAPT prior to staining (Figure 3.5, panels 1 and 7). 95 | Page
However, inhibition of gamma-secretase activity by DAPT ablated N41463 antibody staining, demonstrating the specificity of this antibody for only the gamma-secretase cleaved form of NOTCH4 (Figure 3.5, panel 8). No signal was observed in cells transfected with the parental vector alone (Figure 3.5, panels 10-12).
Our work with the Notch4ECT construct demonstrated that the NOTCH4ICD, once produced, can enter the nucleus. Attempts to detect the NOTCH4ICD in co-cultures with the N41463 antibody proved unsuccessful. Attempts to detect the NOTCH4ICD in tissue samples from sites of active angiogenesis, the murine retina model described in Chapter 5, were also unsuccessful. This indicated that NOTCH4 was not activated or activated at a level below the detection limit of the antibody. The Notch responsive reporter, pGL46xTP1, amplifies the signal generated from the NotchICD and is a more sensitive measure of Notch activation. However, NOTCH4 did not activate the pGL46xTP1 reporter upon ligand co- culture (Section 3.1). All NotchICD family members bind common factors and thus our reporter should be responsive to all family members including NOTCH4 (Lubman et al., 2007; Mizutani et al., 2001). Although there have been no reports of NOTCH4ICD being inhibited by or requiring additional factors to NOTCH1 to activate transcription, the ability of the NOTCH4ICD to activate transcription from the pGL46xTP1 reporter was investigated.
3.2.6 Notch4ECT is a weak activator of the Notch responsive reporter relative to Notch1ECT. Notch4 did not signal in response to ligand in co-culture assays (Figure 3.2). Moreover, NOTCH4ICD was not detected using the N41463 antibody despite its demonstrated activity and specificity (Figure 3.4 and Figure 3.5). To confirm that once produced and translocated to the nucleus NOTCH4ICD can activate transcription from the pGL46xTP1 reporter, a luciferase assay was performed using Notch4ECT. NIH3T3 cells were transfected with increasing quantities of Notch1ECT, Notch4ECT or the vector control, plus the pGL46xTP1 reporter and the pCMXren transfection control. The absolute quantity of DNA transfected was kept constant by adjusting with the control parent vector. After 24 hours the cells were lysed and the luciferase activity measured.
The mean of three independent experiments is represented in Figure 3.6. The relative luciferase activity was normalised to cells transfected with the parental vector alone and expressed as a fold induction. NIH3T3 cells transfected with Notch1ECT displayed a large increase in luciferase activity. This effect was dependent on the amount of Notch1ECT plasmid transfected (note the increase in the first four bars). Notch1ECT enhanced the expression of luciferase more than 5000-fold over the vector control.
96 | Page
Figure 3.5 The N41463 antibody specifically detects nuclear NOTCH4ICD.
NIH3T3 cells were transfected with Notch4ECT (N4ECT) (panels 1- 3 and 7-9), Notch1ECT (N1ECT) (panels 4-6) or vector (panels 10- 12). The cells were grown overnight on coverslips in DAPT (panels 7-9) or DMSO (vehicle) (panels 1-6 and 10-12). The cells were fixed and co-stained with an anti-HA antibody followed by a secondary labelled with Cy3 (panels 1, 4, 7 and 10) and the N41463 antibody followed by a secondary labelled with Alexa Fluor488 (panels 2, 5, 8 and 11). The two channels are overlaid in panels 3, 6, 9 and 12.
97 | Page
Notch4ECT in comparison is a relatively weak activator, inducing the reporter up to 480-fold over the control. Thus the NOTCH4ICD is a much weaker activator of Notch target genes than NOTCH1ICD but is capable of robust activation in our hands.
Figure 3.6 NOTCH4ICD can activate Notch dependent transcription albeit at a low level compared to NOTCH1ICD.
NIH3T3 cells were transfected with Notch1ECT, Notch4ECT, or the vector control in increasing amounts (25ng, 50ng, 100ng and 200ng), plus a Notch responsive reporter (pGL46xTP1) and a Renilla luciferase transfection control (pCMXren). These data were normalised to the vector control and are expressed as fold luciferase induction. This figure represents the mean of three independent experiments with bars indicating the standard deviation. 98 | Page
99 | Page
3.2.7 Stably expressed NOTCH4 is not activated by ligand NOTCH4 does not signal in response to ligand (Figure 3.2) even though its intracellular domain can transduce a signal if released by gamma-secretase (Figure 3.6) . A possible limitation of the co- culture assay described in Sections 3.2.2 and 3.2.3 was the use of transient transfections to express Notch. In these assays the cells were transfected for six hours and then co-cultured for a further eighteen hours. The readout of this assay was luciferase activity which requires Notch expression, signal transduction and the transcription and translation of the firefly luciferase gene. Although this was sufficient time for robust NOTCH1 activation, the lack of NOTCH4 activation may reflect a slower receptor maturation rate. To overcome this limitation cell lines were established that stably expressed Notch1HA, Notch4HA or the vector control. The stable lines expressed the Notch receptors at a steady state level, overcoming possible maturation issues associated with transient transfection. The stable lines were co- transfected with the pGL46xTP1 reporter and the pCMXren transfection control, followed by overnight co-culture with either Dll4 cells or control cells. Results similar to the transient transfection assay were obtained. No significant difference was observed between Notch1HA and Notch4HA stable lines that were co-cultured with control cells (Figure 3.7, columns 2 and 3). Co- culture of vector control cells with Dll4 cells induced a small but significant increase in luciferase activity indicating that these cells possess endogenous Notch receptors (Figure 3.7, column 4). There was a robust activation when Notch1HA expressing cells were co-cultured with Dll4 expressing cells. The reporter was induced in Notch1HA cells 35-fold (P<0.0001) over vector control cells co-cultured with Dll4 cells. In contrast to these results, and in agreement with transient transfection experiments, the Notch4HA line did not stimulate expression of luciferase above the level of control. In fact, as noted for MAEC and C2C12 cells (Section 3.2.3) the luciferase activity measured, although not statistically significant was in fact lower in Notch4HA cells than vector control cells exposed to ligand expressing cells (Figure 3.7, columns 4 and 6).
100 | Page
Figure 3.7 NOTCH4 does not signal in response to ligand when expressed at steady state levels.
NIH3T3 cells lines stably expressing Notch1HA, Notch4HA or the vector only control were transfected with the pGL46xTP1 Notch responsive reporter and a pCMXren transfection control plasmid. Following transfection the cells were co-cultured overnight with Dll4 cells or control cells. Data points represent the relative luciferase units normalised to the Renilla luciferase activity. This figure represents the mean of three independent experiments with bars as standard deviations. Statistics were calculated using ANOVA and Tukey’s post test. 101 | Page
3.2.8 NOTCH4 does not signal upon artificial activation with EDTA NOTCH4ICD has the ability to activate Notch responsive promoters (Section 3.2.4 and 3.2.5) but the full length native receptor was refractory to activation by ligand (Section 3.2.2-3 and 3.2.6). Although the DLL4 ligand has been identified as a ligand responsible for NOTCH4 activation (Shawber et al., 2003), we have been unable to detect ligand-mediated activation in our co- culture assays by either delta or serrate family members. Prior to ligand binding the negative regulatory region (NRR) of Notch receptors maintains the receptors in an “off” state. This region maintains a structure which protects the S2 cleavage site thus preventing activation (Gordon et al., 2007). The structure of the NRR is dependent on calcium ions and can be disrupted using the divalent cation chelator EDTA (Rand et al., 2000). Upon chelation of the calcium ions the NRR region unfolds, exposes the S2 cleavage site and allows ligand independent Notch activation.
To investigate the ability of full length NOTCH4 to respond to this form of activation, NIH3T3 cell lines stably expressing Notch1myc and Notch4HA were transfected with the pGL46xTP1 reporter and the pCMXren transfection control in a similar manner to the co- culture experiments described previously. The cells were then treated with EDTA, washed and allowed to recover in standard media overnight. The Notch1myc and Notch4HA cells displayed low signalling in the absence of EDTA (Figure 3.8, control). Upon EDTA treatment, strong activation of the Notch1myc cells was observed (Figure 3.8, column 3). The Notch4HA cells in contrast failed to activate the reporter upon EDTA treatment (Figure 3.8, column 4). Thus the NOTCH4 receptor is incapable of being activated both in response to ligand (Sections 3.2.2-3.2.3) and by the disruption of the NRR region by calcium chelation.
102 | Page
Figure 3.8 EDTA treatment does not activate NOTCH4.
The NIH3T3 cell lines expressing Notch1myc, Notch4HA were co- transfected with the Notch responsive reporter (pGL46xTP1) and a Renilla luciferase transfection control (pCMXren). The cells were then incubated for 5 minutes in either 0.5mM EDTA or PBS (control). After washing the cells were allowed to recover overnight in standard media. The cells were lysed and the relative luciferase calculated by dividing the firefly luciferase by the
103 | Page
Renilla luciferase. These data represent the mean of three independent experiments with bars as standard deviation.
3.2.9 NOTCH4 is inefficiently S1 processed In the case of NOTCH1, the receptor must be presented on the cell surface for signalling to occur. Surface presentation of NOTCH1 is preceded by S1-processing of the full length peptide by FURIN. After this proteolytic cleavage the C- and N-terminal portions of NOTCH1 remain non-covalently bound to each other and are referred to as the Notch heterodimer (Blaumueller et al., 1997; Logeat et al., 1998). S1 processing is an absolute requirement for signalling from NOTCH1 (Bush et al., 2001; Logeat et al., 1998).
To investigate the ability of NOTCH4 to undergo S1 processing, the stable NIH3T3 cell lines Notch1HA, Notch4HA and the vector control were grown and cell lysates analysed by western blot using the anti-HA antibody and a secondary antibody labelled with a 680nm fluorophore. The resulting blots were analysed with an infrared scanner for accurate quantitation (Figure 3.9a). As expected, NOTCH1 was detected as a full length unprocessed peptide at 300kDa and a S1 processed band at approximately 100kDa previously identified to encompass the transmembrane and i ntracellular domains (TMIC) (Logeat et al., 1998) (Figure 3.9a, lane 2). In the Notch4HA cells a prominent band at 200kDa, which is the expected size of the full length peptide, was observed. A second faint band was also observed at 55kDa which is approximately the expected size of NOTCH4TMIC (Figure 3.9a, lane 3). Faint non-specific bands were visible in all lanes at approximately 180kDa due to the overexposure needed to visualise the NOTCH4TMIC. Specific bands were not observed in the control cells at the expected sizes of NOTCH1 and NOTCH4 (Figure 3.9a, lane 1).
Cell lysates were made in triplicate and the bands were quantified and normalised to beta-actin (Figure 3.9b). The total HA tagged protein produced by the two cell lines was not significantly different. The proportion of NOTCH1TMIC was approximately 60% in this cell line (Figure 3.9b). In contrast, less than 8% of NOTCH4 was processed.
This result was confirmed in two additional cell lines (Figure 3.9c). In the C2C12 cell line no processing of NOTCH4 was observed and only a full length peptide was detected. In transiently transfected HUVEC cells, a very faint but detectable amount of NOTCH4TMIC was observed. Due to the low transfection efficiency and expression achieved in these cells, the proportion of NOTCH4TMIC could not be accurately quantified. When stable lines expressing Notch1HA and Notch4HA were made in C2C12 cells and western blotted (Figure 3.9d), an even more dramatic difference was 104 | Page
observed than in the NIH3T3 cells. In the Notch1HA C2C12 cells, almost 90% of NOTCH1 was S1 processed (Figure 3.9e). In agreement with the transient transfection results (Figure 3.9c), no processing of NOTCH4 was detected (Figure 3.9d).
105 | Page
Figure 3.9 Unlike NOTCH1, NOTCH4 is not S1-processed efficiently in a variety of cell types. a. Cell lysates from NIH3T3 cell lines expressing Notch1HA (lane 2), Notch4HA (lane3) or the vector control (lane 1) were western blotted with anti-HA and anti-beta-actin antibodies and detected with an anti-mouse 680 secondary antibody. The full length peptides (NOTCH1Fl and NOTCH4Fl) and the TMIC are indicated (NOTCH1TMIC and NOTCH4TMIC). The position of the molecular weight markers is indicated on the left. b. Three replicate lysates were quantified on an Odyssey infrared scanner and analysed using ImageJ software. These data points represent the mean of the Notch TMIC divided by the total Notch with bars as standard deviation. c. Cell lysates from NIH3T3N4HA cells (lane 1) and C2C12 (lane 2) and HUVEC cells (lane 3) transiently transfected with Notch4HA were western blotted with anti-HA followed by anti-mouse HRP. The position of the full length, NOTCH4Fl, and NOTCH4TMIC are indicated on the right. d. Cell lysates from C2C12 cells stably expressing Notch1HA (lane 1) or Notch4HA (lane2) were western blotted with an anti-HA antibody and detected with an anti-mouse 680 secondary antibody. The expected positions of the full length and TMIC of NOTCH1 (NOTCH1Fl and NOTCH1TMIC) and Notch4 (NOTCH4Fl and NOTCH4TMIC) are indicated on the right. e. Three replicate lysates were quantified on an Odyssey infrared scanner and analysed using ImageJ software. These data points represent the mean of the Notch TMIC divided by the total Notch with bars as standard deviation.
106 | Page
107 | Page
3.2.10 Cell surface presentation of NOTCH4 differs from that of NOTCH1. NOTCH1 is exclusively presented on the cell surface as an S1 processed heterodimer, which indicates an intimate link between S1 processing and cell surface expression. The lack of NOTCH4 signalling observed may therefore be caused by poor surface presentation, preventing interaction with ligand presented by adjacent cells. To evaluate the surface expression of NOTCH4 a surface biotinylation assay was performed (Bush et al., 2001). A cell impermeable biotin reagent (Sulfo-NHS-SS-Biotin; Thermo Scientific) that reacts to form covalent linkages between biotin and amide residues of proteins on the cell surface was used to label surface proteins. This allows the capture of proteins that have been exposed on the cell surface.
To compare the surface expression of NOTCH1 and NOTCH4, stable NIH3T3 cell lines expressing Notch1HA and Notch4HA were labelled with Sulfo-NHS-SS-Biotin and the surface proteins captured using streptavidin coupled beads (SA).The captured proteins were analysed by western blotting using an antibody to the HA tag. Notch1HA and Notch4HA were expressed at equal levels in cell lysates prior to capture (Figure 3.10, lanes 1 and 2). Biotin labelled Notch4HA lysates that were captured with control beads (no streptavidin) failed to capture Notch4HA indicating that the capture was specific to biotin labelled proteins. Notch4HA was not detected in SA captures of unlabelled Notch4HA cell lysates, also indicating that the streptavidin capture was specific to labelled proteins (Figure 3.10, lane 5). Lastly, beta-actin was not detected in any of the capture lanes, indicating that intracellular proteins were not labelled by biotin (Figure 3.10, lanes 3-6). As expected, only the S1 processed NOTCH1TMIC was found in SA captures and not unprocessed NOTCH1 receptor (Figure 3.10, lane 3). In contrast, only full length NOTCH4 was found in SA captures. The NOTCH4TMIC was not detectable in SA captures although a faint 55kDa band can be seen in cell lysates (Figure 3.10, lanes 2 and 6). A lack of surface S1 processed NOTCH4 in this assay may be due to low levels in the cell lysate prior to SA capture (Figure 3.10, lane 2). The bands detected on the western blot were quantified. Although very similar amounts of NOTCH1 and NOTCH4 were detected in the cell lysates used for the pull down approximately 5-fold more NOTCH1 was detected on the cell surface compared to NOTCH4. An equal quantity of protein was used for each capture after calculating the protein concentration using a BCA assay. Therefore, not only was NOTCH4 presented on the surface without being S1 processed it was also inefficiently presented compared to NOTCH1.
108 | Page
Figure 3.10 Unprocessed full length NOTCH4 is presented on the cell surface
NIH3T3 cell lines stably expressing Notch1HA or Notch4HA were surface biotinylated. The surface proteins were captured with streptavidin beads (SA) or control beads and analysed by western blot (WB) with anti-HA and anti-beta-actin antibodies. Lanes correspond to lysates from Notch1HA (lane 1) and Notch4HA (lane 2) prior to streptavidin capture, Notch1HA cells after streptavidin capture (lane 3), Notch4HA cells captured with control beads (lane 4), Notch4HA cells mock biotinylated and captured with streptavidin (lane 5) and labelled Notch4HA cells after streptavidin capture (lane 6). The position of full length Notch (NOTCH1Fl and NOTCH4Fl) and TMIC peptides (NOTCH1TMIC and NOTCH4TMIC) are indicated on the right. The positions of the molecular weight markers are shown on left.
109 | Page
3.2.11 NOTCH4 is not activated by ligand despite introduction of a consensus FURIN cleavage site. The lack of NOTCH4 signalling in response to ligand appears to be due to a combination of factors. The relatively low ability of the NOTCH4ICD to activate transcription may not be detectable above background in our assay. In conjunction with this the NOTCH4 receptor was poorly S1 processed and presented on the cell surface at an approximately 5-fold less efficient rate than NOTCH1. A lack of S1-processing may be a direct cause of the failure of NOTCH4 to be efficiently presented on the cell surface, which is well described for S1 site mutations in NOTCH1 (Blaumueller et al., 1997; Logeat et al., 1998). The FURIN processing region of the Notch receptors was elucidated in a crystal structure of the NOTCH2 NRR (Gordon et al., 2007). An unstructured loop, termed the S1 loop, was removed in the construct used to crystallise the NOTCH2 NRR region because its disordered nature interfered with crystallisation.
The sequence of this region in mouse Notch receptors is highly divergent (Figure 3.11a). However, consensus FURIN cleavage sites can be readily identified in mouse NOTCH1, 2 and 3 (shown in bold Figure 3.11a). A FURIN consensus site is absent in NOTCH4 orthologues (Figure 3.11b). However, we cannot exclude the possibility of a non-consensus site.
We reasoned that NOTCH4 could potentially be activated in our assay if FURIN cleavage could be enhanced by the introduction of a strong FURIN consensus site. To achieve this, the full length NOTCH4 construct was engineered to contain the sequence RQRR within the S1 loop (Figure 3.11b Notch4RQRR). This construct replaces amino acids SSSW with RQRR.
This construct, Notch4HA and the control vector were co- transfected into C2C12 cells with the pGL46xTP1 reporter pGL46xTP1 and the pCMXren control plasmid. After overnight incubation with either Dll4HA cells or control cells, the signalling capacity of Notch4RQRR was assessed by luciferase assay. In similar results to those presented previously, the control transfected cells displayed mild Notch activation in response to ligand (Figure 3.11c, columns 1 and 4) indicating the presence of endogenous Notch receptors. C2C12 cells express NOTCH1 and activation of this receptor prevents differentiation in this cell line. In contrast to this and in agreement with previous results (Section 3.2.1), NOTCH4 failed to activate the reporter above control in response to ligand (Figure 3.11c, columns 4 and 5). In fact as previously noted the NOTCH4 receptor actually produced less signal than control (see Chapter 4). The NOTCH4 receptor
110 | Page
engineered to have a consensus FURIN site within the S1 loop, Notch4RQRR, was indistinguishable from the native receptor (Figure 3.11c, column 2 vs. 3 and column 5 vs. 6). In addition there was no increase in the production of the processed form of NOTCH4 when NIH3T3 cells were transiently transfected with Notch4RQRR (Figure 3.11d). Therefore, addition of a strong FURIN consensus cleavage site to NOTCH4 does not enhance its capacity to signal.
111 | Page
Figure 3.11 NOTCH4 is not activated by ligand despite introduction of a consensus FURIN cleavage site. a. Sequence of the S1 loops of mouse Notch receptors. The FURIN cleavage sites are shown in red. b. Alignment of the S1 loops of NOTCH4 orthologues and the construct Notch4RQRR. c. C2C12 cells were co-transfected with Notch4HA, Notch4RQRR or the vector control in addition to the pGL46xTP1 reporter and pCMXren transfection control. This was followed by an overnight co-culture with cells expressing Dll4HA or control cells. The relative luciferase activity was measured. The figure represents the mean of four replicate transfections with bars as standard deviations. d. Cell lysates from NIH3T3 cells transiently transfected with vector, Notch4HA or Notch4RQRR were western blotted with anti- HA followed by anti-mouse HRP. The molecular weight markers are indicated on the left.
112 | Page
113 | Page
3.2.12 NOTCH1 and NOTCH4 exhibit distinct subcellular localisations Experiments thus far have highlighted functional differences between the NOTCH1 and NOTCH4 receptors. The intracellular domain of NOTCH4 has a reduced transactivation capacity compared with that of NOTCH1. NOTCH4 also exhibits reduced S1- processing and surface presentation compared with NOTCH1. There is an intimate connection between FURIN processing and subsequent intracellular trafficking and surface presentation of NOTCH1. Addition of a consensus FURIN recognition site did not improve the signalling capacity of NOTCH4. To investigate the intracellular trafficking of NOTCH4 the subcellular localisation of NOTCH1 and NOTCH4 was examined.
NIH3T3 cells were transfected with Notch1HA (Figure 3.12a) or Notch4HA (Figure 3.12b). The transfected cells were fixed then stained with an anti-HA antibody followed by an anti-mouse antibody labelled with Cy3 and counterstained with the nuclear stain Hoechst 33342. The cells were examined by confocal microscopy. Three representative fields are shown in Figure 3.12a (panels 1-3, 4-6 and 7-9). Notch1HA was located in distinct puncta throughout the cell in addition to a diffuse staining in the peri- nuclear region corresponding to the known location of NOTCH1 within the secretory pathway i.e. within the Golgi apparatus and intracellular vesicles (Jarriault et al., 1995). Notch4HA was diffusely expressed in contrast to the distinct puncta observed in Notch1HA (Figure 3.12b, panels1-3, 4-6, and 7-9). Although Notch4HA is diffusely spread throughout the cell, the major region of expression of Notch4HA is in close proximity to the nuclear envelope. This different expression pattern may have consequences for NOTCH4 trafficking to the cell surface and signalling.
Figure 3.12 NOTCH4 is localised differently to NOTCH1
NIH3T3 cells were transfected with HA tagged Notch1 (a) or Notch4 (b). Cells were stained with anti-mouse HA followed by a Cy3 labelled secondary antibody (red). The nucleus was counter stained with Hoechst33342 nuclear stain (blue). Three representative images for Notch4HA and Notch1HA are shown. Confocal images were taken on a LSM 700 upright confocal microscope (Zeiss) at 63X. Images were taken sequentially with a 555 laser (column 1) and a 405 laser (column 2). Merged images are shown in column 3.
114 | Page
115 | Page
3.2.13 Chimaeric Notch receptors reveal the importance of the extracellular domain and protease cleavage region. The lack of signalling observed for NOTCH4 can be explained by a number of factors. NOTCH4 displays a distinct subcellular localisation pattern to NOTCH1 and is poorly presented on the cell surface in the unprocessed rather than heterodimeric form. In addition to this, NOTCH4ICD is a relatively weak activator of transcription. We have demonstrated that the differences observed between NOTCH1 and NOTCH4 are due to both upstream events (receptor maturation) and downstream events (transcriptional activation). In order to separate these activities a series of constructs were generated in which extracellular and intracellular domains of NOTCH1 and NOTCH4 were exchanged (Figure 3.13). In experiments to look at NOTCH1 activation separate from endogenous Notch activity, a Notch1 construct was used where the yeast transcriptional activator Gal4/VP16 was cloned into an AscI site introduced into a Notch1 E construct (Notch E-GVP) (Karlstrom et al., 2002). The introduction of the site caused a point mutation changing a histidine to an alanine. The Gal4/VP16 sequences were removed and the region containing the AscI site cloned into the full length receptor resulting in the construct Notch1:1 (Figure 3.13). A similar mutation was made in Notch4 to generate an AscI site at the same relative position in the gene. This introduced a mutation converting a histidine to an alanine (Figure 3.13, Notch4:4). Two further constructs were made. Notch1:4 was generated by combining the sequences 5' of the AscI site in Notch1:1 with those 3' of the AscI site in Notch4:4 (Figure 3.13, Notch1:4). The reciprocal construct, Notch4:1 was made using the 5’ end of Notch4:4 and the 3’ end of Notch1:1 (Figure 3.13, Notch4:1). The protease cleavages involved in NOTCH1 activation (S1/FURIN, S2/ADAM10/17 and S3 and S4/gamma- secretase) are all N-terminal to the introduced AscI site. The region involved in protease activation of NOTCH1 is present in Notch1:4 with the corresponding region of NOTCH4 present in Notch4:1. Therefore Notch1:4 has NOTCH1's ability to be activated by ligand with the NOTCH4ICD's ability to stimulate transcription. Notch4:1 has NOTCH4's ability to be activated by ligand with the transcriptional activity of the NOTCH1ICD.
116 | Page
Figure 3.13 Chimaeric Notch receptor constructs.
A diagrammatical representation of the chimaeric Notch receptor constructs. EGF-epidermal growth factor-like domain, LNR-Lin- 12/Notch repeat, TM-transmembrane domain, ANK-Ankyrin repeats, HA-HA tag. a. Notch1:1 contains an AscI site that converts an alanine to a histidine C-terminal to the gamma-secretase cleavage site. b. Notch1:4 comprises Notch1 sequences N-terminal to the introduced AscI site and Notch4 ICD sequence C-terminal of the AscI site. c. Notch4:1 comprises Notch4 sequences N-terminal to the introduced AscI site into Notch1 ICD sequence C-terminal of the AscI site. d. Notch4:4 contains an AscI site that converts an alanine to a histidine C-terminal to the gamma-secretase cleavage site equivalent to that of Notch1:1. 117 | Page
To ensure that the introduction of the point mutation in both NOTCH1 and NOTCH4 did not change the behaviour of the protein, Notch1:1 and Notch4:4 were compared to their unmodified counterparts. NIH3T3 cells were co-transfected with the Notch constructs Notch1myc, Notch4HA, Notch1:1, Notch4:4 or the vector control in addition to the Notch responsive reporter (pGL46xTP1) and a Renilla luciferase transfection control (pCMXren). This was followed by co-culture with Dll4HA cells or control cells and measurement of the relative luciferase activity. As observed previously Notch4HA was unable to activate luciferase expression above the level of control in cells either co-cultured with control cells or Dll4HA cells. The Notch4:4 construct behaved in a manner indistinguishable from wild type Notch4HA (Figure 3.14a, columns 3 and 5 and 8 and 10). Compared to control (Figure 3.14a, column 1), cells transfected with either Notch1myc or Notch1:1 were activated upon co-culture either with control cells (Figure 3.14a, columns 2 and 4) or Dll4HA cells (Figure 3.14a, columns 7 and 9).
The ability of the chimaeric receptors to activate the Notch responsive reporter was analysed using a co-culture assay. NIH3T3 cells were transfected with the four receptor constructs described above or the vector control plus the Notch responsive reporter (pGL46xTP1) and the Renilla transfection control (pCMXren). These cells were co-cultured overnight with Jag1 cells or control cells. The results were normalised to the relative luciferase activity of the control transfected cells.
As expected the Notch1:1 transfected cells induced a luciferase activity approximately 15-fold higher than the control plasmid (Figure 3.14b, column 1). Luciferase activity was significantly reduced (P<0.05) when the Notch4 intracellular domain was present in the context of Notch1 extracellular domain and protease cleavage region (Figure 3.14b, Notch1:4). Nonetheless, the Notch1:4 construct was able to enhance luciferase activity approximately 5-fold over control transfected cells, a significantly greater activity than Notch4:4 (P<0.01). The observation that the NOTCH4ICD can signal, albeit at a reduced level, when incorporated into a functional receptor such as NOTCH1 is consistent with the signalling capacity of Notch4ECT (Figure 3.6). Thus NOTCH4 fails to signal predominately because of its extracellular domains and protease cleavage region.
Replacement of the NOTCH1 extracellular domain with those of NOTCH4 resulted in a significant reduction in luciferase activity compared to Notch1:1 (Figure 3.14b, Notch4:1; P<0.01). This construct was only capable of inducing a 2-fold increase in luciferase activity over control but was significantly (P<0.05) better than the Notch4:4 construct. Thus, loss of Notch signalling
118 | Page
capacity to only twice that of the control is again due to the NOTCH4 extracellular domain and protease cleavage region. However, the capacity to generate significantly more signal than vector only shows that the NOTCH4 extracellular domain has some capacity to respond to ligand.
119 | Page
Figure 3.14 Activation potential of Notch chimaeric constructs. a. NIH3T3 cells were transfected with Notch1myc, Notch4HA, Notch1:1, Notch4:4 or vector control plus a Notch responsive reporter (pGL46xTP1) and a Renilla luciferase transfection control (pCMXren). The cells were co-cultured with Dll4HA ligand cells or control cells and the relative luciferase activity measured. These data represent the mean of four independent transfections with bars as standard deviations. b. NIH3T3 cells were transfected with Notch1:1, Notch1:4, Notch4:1, Notch4:4 or the vector control plus a Notch responsive reporter (pGL46xTP1) and a Renilla luciferase transfection control (pCMXren).The cells were co-cultured with Jag1 ligand cells and the luciferase activity measured. These data represent the mean (n=3) of the relative luciferase activity normalised to control transfected cells of three independent experiments. A t-test was performed to compare Notch1:1 to Notch1:4 (P<0.05) and Notch4:1 (P<0.01) and Notch4:4 to Notch1:4 (P<0.01) and Notch4:1 (P<0.05).
120 | Page
121 | Page
3.3 Discussion A co-culture system that recapitulates all the stages of Notch signalling was used to investigate the ligand-induced signalling capacity of the NOTCH4 receptor. NOTCH4, unlike NOTCH1, did not signal in response to ligand in each of three cell types tested (NIH3T3, MAEC, C2C12). Ligand-induced signalling of Notch relies on a number of key events occurring: receptor processing and cell surface presentation, receptor cleavage following ligand binding, nuclear entry of NotchICD, and transcriptional activation of RBPJ- bound promoters. A comparison of NOTCH4 with NOTCH1 was made with respect to these key signalling events. These studies revealed that NOTCH4 differed from NOTCH1 in the following ways: (i) NOTCH4 is inefficiently S1-processed, (ii) unprocessed NOTCH4 is present on the cell surface, (iii) NOTCH4 is present on the cell surface in reduced amounts, (iv) NOTCH4 cannot be activated by ligand co-culture or by EDTA activation, (v) NOTCH4 was expressed in a different subcellular localisation to NOTCH1 and (vi) the NOTCH4ICD has a reduced potential to activate transcription in comparison to the NOTCH1ICD. These differences will be discussed in the following sections.
3.3.1 NOTCH4 is inefficiently S1 processed and is presented on the cell surface as a full length receptor In our co-culture experiments NOTCH4 was not activated by either DLL4 or JAG1 in three different cell types. In all cases NOTCH1 activation was clearly evident. Our aim was to characterise where and when in the canonical pathway NOTCH4 diverged from a canonical Notch receptor. A key point of divergence we found was the ability of NOTCH4 to undergo S1 processing.
The most compelling evidence for ligand dependent activation of NOTCH4 was generated by the Kitajewski laboratory (Funahashi et al., 2008; Shawber et al., 2003; Shawber et al., 2007). In their initial report they found that NOTCH4 could be activated in primary neonatal dermal microvascular cells by DLL4 in a luciferase reporter assay similar to ours (Shawber et al., 2003). The activation was mild (3.8-fold) and no comparison was made with other Notch family members. A simple explanation for the distinct findings is that human dermal microvascular cells and human umbilical venous endothelial cells (HUVEC) used in their studies are capable of activating NOTCH4 in response to ligand while NIH3T3, C2C12 and MAEC that we have analysed are not. It is of note that S1-processed NOTCH4 was detected on the surface of dermal microvascular cells by this group (Shawber et al., 2003). In contrast we and others (Wu and Bresnick, 2007b) found S1 processed NOTCH4 was undetectable on the cell surface of
122 | Page
NIH3T3 cells. Mutations to the S1 cleavage site in NOTCH1 are known to inhibit receptor signalling (Bush et al., 2001). Similar to NOTCH1, NOTCH4 may also need to be presented on the cell surface as a heterodimer in order to be activated by ligand. Nonetheless, NOTCH4 is processed to some degree in NIH3T3 cells and yet NOTCH4 expressing NIH3T3 cells did not signal in response to ligand. Even EDTA treatment, which also activates intracellular Notch heterodimers (Gavin Chapman, Pers. Comm.), did not elicit a signal from NOTCH4 above background in our hands. Thus, it is likely that other properties, in addition to its S1 processing, also impede the ability of NOTCH4 to signal.
The cell type dependence of S1 processing we noted also points to differences in the regulation of S1 processing between NOTCH1 and NOTCH4. In NIH3T3 cell lines we found approximately 10% of NOTCH4 was S1 cleaved compared to 60% for NOTCH1. This result contrasted with the result obtained for C2C12 cells where NOTCH1 was more efficiently processed (90%) and NOTCH4 processing was absent. The opposite effects on S1 processing, seen between NOTCH1 and NOTCH4 in NIH3T3 and C2C12 cells, suggest that S1 processing is not dependent on FURIN expression alone. The NOTCH4 receptor does not contain any identifiable FURIN consensus site in the S1 loop, suggesting NOTCH4 family members are not FURIN substrates. We confirmed this by showing that the addition of a FURIN consensus site to the S1 loop did not enhance NOTCH4 S1-processing. Nonetheless, the identification of a NOTCH4TMIC by us and others (Shawber et al., 2003; Wu and Bresnick, 2007b) indicates that NOTCH4 is cleaved in or near the S1 loop.
One possible explanation for a lack of NOTCH4 S1 processing is that NOTCH4 is processed by an unknown protease, which is not a member of the Furin-like convertase family. Expression of such a protease in some cell types but not others would explain observed cell type specific differences in the S1 processing of NOTCH4. The potential importance of S1 processing to signalling makes the identification of both the S1 processing site and factor/s responsible important for understanding NOTCH4 function. We attempted to purify the NOTCH4TMIC for N-terminal sequencing in order to identify the site of S1 processing of NOTCH4. We were unsuccessful due to the low abundance of the NOTCH4TMIC. There are a number of strategies that could be employed to successfully purify the NOTCH4TMIC. A variety of cell lines could be screened for increased S1 processing of NOTCH4 and then used, after scaling up, to generate sufficient quantities of protein for N-terminal sequencing. The identification of cell types in which NOTCH4 was efficiently S1 cleaved would also have the added benefit of enabling the analysis of the relationship between S1 123 | Page
processing of NOTCH4 and its ability to be activated by ligand. In addition, identification of where and when NOTCH4 is S1 processed could point to events regulated by canonical NOTCH4 signalling. Perhaps more importantly identification of endogenous cell types where NOTCH4 is not S1 processed could point to areas where NOTCH4 has a non-overlapping function with NOTCH1.
Another possibility for a lack of NOTCH4 S1 processing is that it may not come in contact with FURIN during its normal intracellular trafficking. The localisation of NOTCH1 at least in terms of surface presentation is regulated by FURIN cleavage. We have observed that the NOTCH4 receptor has a distinct subcellular localisation pattern to NOTCH1. NOTCH1 is S1 processed in the trans-Golgi network (TGN) and FURIN localises to the TGN and endosomal vesicles (Schafer et al., 1995). NOTCH1 expression is detected as discrete puncta (Section 3.2.13). These vesicles are late- and post-secretory pathway (Chapman et al., 2011). The lack of clear vesicle staining for NOTCH4 indicates that it may not efficiently reach the later stages of the secretory pathway. Consistent with this idea is our observation that approximately 5-fold less NOTCH4 is presented on the cell surface compared with NOTCH1. Therefore significant amounts of NOTCH4 may not be available for S1 cleavage in the TGN. A more thorough characterisation of the subcellular localisation of NOTCH4 may indicate if trafficking of NOTCH4 prevents S1 cleavage or whether S1 cleavage regulates trafficking. Co-staining with FURIN and/or other subcellular markers to look at co-localisation of NOTCH4 would shed light on whether NOTCH4 maturation is divergent from NOTCH1 before or at S1 processing. Nonetheless, the classical secretory pathway dictates that NOTCH4 should pass through the TGN on its way to the plasma membrane, implying that full length NOTCH4 observed on the cell surface had the opportunity to come in contact with FURIN in the TGN but was not S1 processed at the time. Alternate secretory pathways exist that bypass the Golgi apparatus, making it possible that NOTCH4 could, by taking such a route, avoid coming in contact with FURIN (Nickel and Rabouille, 2009; Schotman et al., 2008; Schotman et al., 2009).
We demonstrated that the major species of NOTCH4 that is presented on the cell surface is the full length peptide i.e. not S1 processed (Section 3.2.9). This is in stark contrast to NOTCH1, which is exclusively presented on the surface as a S1 cleaved heterodimer. There have been reports of other Notch receptors expressed on the cell surface as full length peptides in over expression systems and the ability of full length receptors to respond to ligand remains controversial (Bush et al., 2001; Gordon et al., 2009). NOTCH2 and Drosophila NOTCH, in contrast to NOTCH1, can signal without S1 cleavage. These results suggest 124 | Page
that S1 cleavage may be a marker of correct surface presentation rather than an absolute prerequisite. However, for NOTCH1, 2 and 3 the exclusive or majority (90-95%) species on the cell surface in these over expression systems is in the heterodimeric form. We and others (Shawber et al., 2003; Wu and Bresnick, 2007b), find that the majority of NOTCH4 on the cell surface is unprocessed. This includes detection of endogenously expressed NOTCH4 (Wu and Bresnick, 2007b) indicating that this is not an artefact of over expression. These results may be linked to the lack of S1 processing and have important implications for ligand activation of NOTCH4. Mutating the FURIN cleavage sites of NOTCH1 can force unprocessed NOTCH1 onto the surface. However, the full length receptor on the surface is incapable of ligand mediated activation.
3.3.2 NOTCH4 is not activated in response to EDTA The NOTCH4 receptor was not activated in response to co-culture with Dll4 cells or Jag1 cells. In addition we were unable to artificially activate NOTCH4 using EDTA (Section 3.2.7). EDTA is a powerful activator of NOTCH1. The NOTCH1 NRR is maintained in a S2 protease resistant state in a calcium dependent structure. By chelating calcium, EDTA disrupts this structure and allows for ligand independent activation. The NOTCH4 NRR is highly divergent to that of other family members which may have structural consequences. Recent attempts to generate antibodies to the NRR of mouse NOTCH4 were unsuccessful because the authors were unable to express the NOTCH4 NRR although they could generate protein from all three other mammalian homologues (Falk et al., 2012). In fact the authors tried to generate an expressible protein by mutating the surface exposed residues of the NOTCH2 NRR to match NOTCH4. The NOTCH4/NOTCH2 hybrid NRR also failed to express pointing to deeper structural differences between the two receptors. Additionally there are important differences particularly in the LNR-A domain. The LNR-A domain of NOTCH4 is missing a cysteine pair. LNR-A is also missing key residues involved in calcium binding and it has been shown that the folding of the NOTCH4 LNR-A is independent of calcium in contrast to the NOTCH1 LNR-A (Hao, 2009). The unfolding of the LNR-A from the NRR is a key first step in the activation of the receptor. The divergence of NOTCH4 in this region points to key differences in ligand dependent activation.
3.3.3 The extracellular and intracellular domains contribute to the inability of NOTCH4 to signal In addition to receptor maturation and ligand mediated activation we also analysed the ability of the NOTCH4ICD to enter the nucleus and activate transcription. The NOTCH4ICD is highly 125 | Page
divergent from other Notch family members but does retain the ability to activate Notch responsive reporters, despite the absence of a transactivation domain found in NOTCH1 and NOTCH2 (Section 3.4.1.2). We found that although robust induction of luciferase was detected, this was approximately 15-fold lower than that seen for NOTCH1ICD (Section 3.2.5). The low level of signalling compared to NOTCH1ICD is in agreement with previous reports (Ong et al., 2006). There are a number of features that diverge in the NOTCH4ICD from other family members, which could account for these differences.
The lack of activation seen in our reporter may suggest that the NOTCH4ICD has alternative targets and/or functions. There is some evidence that the context of the RBPJ binding site may affect transcription. Both NOTCH1 and NOTCH2 contain sequences in the C-terminal region that act as a transcriptional activation d omain (TAD). In the case of the NOTCH3ICD it was first reported that NOTCH3ICD did not contain a TAD (Beatus et al., 1999; Beatus et al., 2001; Kurooka et al., 1998). However, a later study reported that NOTCH3 required a zinc finger binding site in close proximity which could then display transactivation activity (Ong et al., 2006). No such requirement has been reported for NOTCH4 and no TAD has been identified. NOTCH4, like NOTCH3, may also require additional factors to efficiently activate transcription. Our reporter may represent a suboptimal promoter structure for NOTCH4ICD interaction. To establish if the NOTCH4ICD does recognise a different set of promoters a ChIP-on-Chip assay could be performed, comparing NotchICD family members. Such an experiment may reveal different binding preferences for NOTCH4ICD and may shed light on whether NOTCH4 performs better in specific contexts.
NotchICD family members also display cell type specific relative abilities to drive transcription (see Section 4.4.1). This suggests that the ICD of individual receptors may form unique transcriptional complexes. In the case of NOTCH4 there is incomplete but good evidence that this is the case. Size exclusion fractionation of NOTCH4ICD complexes displayed a unique elution profile to those of NOTCH1, 2 and 3 (Han et al., 2011). Further characterisation of these complexes could shed light on what additional regulations and/or co-factors are required for NOTCH4ICD activation of transcription.
We have shown that NOTCH4ICD is a poor activator of transcription compared to NOTCH1ICD. In addition the cell surface presentation of NOTCH4 is low compared to NOTCH1. These two factors may account for the lack of NOTCH4 dependent transcription in response to ligand. This idea was addressed by 126 | Page
the analysis of chimaeric Notch receptors. The Notch4:1 chimaera measured the ability of NOTCH4 to respond to ligand, detected by the strong transactivation activity of NOTCH1ICD. All of the protease cleavage sites were derived from NOTCH4 in this chimaera: (i) the S1 site responsible for TMIC production, (ii) the S2 cleavage site uncovered after ligand dependent activation and (iii) the gamma secretase cleavage sites. Addition of the NOTCH1ICD to NOTCH4 generated low-level reporter activation. Thus, despite low surface presentation and S1-processing, NOTCH4 can generate some NOTCH4ICD in response to ligand. In the reciprocal chimaera, we exploited the ability of NOTCH1 to be efficiently presented on the cell surface to measure ligand induced signalling potential of NOTCH4ICD (Notch1:4 Section 3.2.14). Addition of NOTCH4ICD to NOTCH1 produced a somewhat improved transcriptional response compared with Notch4:1 and Notch4:4. The observation that Notch1:4 signals better than Notch4:1 indicates that differences in the extracellular domain of NOTCH4 are the primary cause of a lack of NOTCH4 signal transduction. The weak activation potential of NOTCH4ICD contributes to a lesser degree to NOTCH4’s apparent lack of signalling. In combination, the extracellular and intracellular domains of Notch4 reduce its ligand dependent signalling capacity to below a detectable level. These results support a model where NOTCH4 is capable of canonical Notch signalling albeit to a low degree. The small amount of NOTCH4 that makes it to the surface either S1 cleaved or not can be activated by ligand to release a weak but canonically functional NotchICD.
3.3.4 Conclusions We were unable to detect canonical NOTCH4 signalling in a variety of cell types. The inability of NOTCH4 to detectably act as a canonical Notch receptor may be due to a number of factors: (i) reduced and divergent S1 cleavage (ii) reduced surface expression, (ii) surface expression of full length receptor, (iii) divergent subcellular distribution and (v) relatively poor activation potential of the NOTCH4ICD. However, we showed that by uncoupling these processes using chimaeric constructs that the ability of NOTCH4 to participate in canonical Notch signalling was preserved. However, not only did we not detect NOTCH4 activation, we noted that in NIH3T3 (Section 3.2.6), C2C12 and MAEC (Section 3.2.2) cells the reporter was in fact inhibited below the level of the endogenous Notch response. The inhibition of Notch signalling seen in these experiments is pursued in Chapter 4.
127 | Page
4 NOTCH4 inhibits NOTCH1 signalling 4.1 Introduction Our analysis of the canonical signalling potential of NOTCH4 revealed that NOTCH4 lacked the ability to activate transcription in response to ligand. The reasons for this were due to a combination of factors. The NOTCH4ICD has a reduced ability to transactivate RBPJ dependent promoters compared to NOTCH1ICD. NOTCH4 was poorly S1 processed and displayed on the cell surface at reduced levels compared to NOTCH1. In addition the NOTCH4 receptor was presented on the cell surface as an intact polypeptide rather than as a S1 processed heterodimer. In the course of our experiments we found that NOTCH4, rather than simply lacking transactivation potential, also inhibited endogenous Notch activity in C2C12 cells, MAEC cells (Section 3.2.2) and NIH3T3 cells (Section 3.2.3 and 3.2.6). The ability of NOTCH4 to suppress canonical Notch signalling was a novel finding which we investigated further in this chapter.
4.1.1 Paralogue specific functions of mammalian Notch receptors In the mammalian lineage there has been an amplification of both Notch receptor and ligand homologues. The family consists of four Notch receptors (Notch1-4) and five ligands (Dll1, 3 and 4 and Jag1 and 2 ). The diversity of Notch receptors suggests that paralogue specific functions may have evolved even though it is generally accepted that the biochemical functions of Notch paralogues are at least in part interchangeable (Bray, 2006; Guruharsha et al., 2012). The tissue specific expression of each receptor does not necessarily imply differences in function. For example, in zebrafish there are two Notch1 paralogues. The two paralogues have diverged in their tissue distribution such that the combined expression domain of the two paralogues reflects the expression domain of the single Notch1 paralogue in mammals (Westin and Lardelli, 1997). Specific timing and expression in various tissues could evolve to control the Notch pathway while preserving the basic function of the protein. However, in many cells multiple receptors are co-expressed. The co-expression of multiple Notch receptors may indicate non-overlapping functions.
There have been various reports that Notch1 and Notch2 expression in various cancers can predict opposite prognostic outcomes (Chu et al., 2011; Parr et al., 2004). These effects are correlative only but some investigation of biochemical divergence has been conducted. These studies have concentrated on the effects of NotchICD constructs which, in the case of NOTCH1 and
128 | Page
NOTCH2, have been found to have opposing effects in some circumstances. Malignant mesothelioma cells respond to NOTCH1ICD by down regulating PTEN which leads to increased proliferation. In contrast NOTCH2ICD increased PTEN levels and caused a G1 arrest (Graziani et al., 2008). Medulloblastoma cells expressing NOTCH2ICD had increased proliferation, growth in soft agar and xenograft growth. Expression of NOTCH1ICD in medulloblastoma cells had the opposite effects (Fan et al., 2004). There has also been some evidence that while Notch1 deletion can inhibit normal haematopoietic differentiation, Notch2 deletion cannot, even though both are expressed (Kumano et al., 2003).
The NOTCH4ICD has also been shown to functionally diverge from other NotchICD paralogues. Haematopoietic stem cell emergence and their functional development was blocked by the expression of NOTCH1ICD in endothelial/haematopoietic progenitors whereas NOTCH4ICD expression had no effect (Tang et al., 2012). All four mammalian NotchICDs were able to drive T-cell development in haematopoietic progenitors. However, in contrast to NOTCH1- 3ICD, NOTCH4ICD was unable to induce T-cell acute lymphoblastic leukaemia, possibly because myelocytomatosis oncogene (Myc) expression was not induced by NOTCH4ICD (Aster et al., 2011). These divergent functions of NotchICDs have not been well characterised and may reflect different strengths and duration of signalling rather than divergent function as Notch signalling is very sensitive to dosage and duration (Bray, 2006).
There have been reports of the NotchICD regions forming dimers (Liu et al., 2010a; Nam et al., 2007). RBPJ binding sites on promoters can be found as conserved pairs separated, in a “head to head” orientation, by 15-19 nucleotides. In a crystal structure containing the NOTCH1ICD Ankyrin repeats, RBPJ and fragments of MAML bound to a paired promoter, residues in the second and third Ankyrin repeats were found to make contacts (Nam et al., 2007). The formation of this complex was dependent on the addition of MAML which may explain why NotchICDs have not been reported to form dimers in isolation. The critical residues in contact are conserved across species from sea urchin, Drosophila and vertebrates although they are divergent in C.elegans. The loss of these contacts in C.elegans coincides with the loss of paired promoter sequences. In addition these residues are conserved in all four mammalian Notch paralogues raising the possibility of NotchICD heterodimers forming. Dimer formation has important consequences for transcriptional activation. Mutants that have a charge reversal in one of the salt bridges formed by dimerisation no longer efficiently activate paired promoters. Importantly mutants that have a complementary charge reversal in the salt
129 | Page
bridge complement each other (Nam et al., 2007). Thus the defect in these mutants is due to an inability to form dimers rather that a defect in transactivation potential in general. The formation of dimers drives leukaemiagenesis (Liu et al., 2010a). Expression of wild type NOTCH1ICD in haematopoietic stem cells induces leukaemia with 100% penetrance. Dimerisation deficient mutants are unable to induce leukaemia, dependent on dimerisation, but can induce T-cell specification, a Notch process dependent on unpaired promoters. The significance of these findings in respect to the potential for individual paralogues to modulate each other has not been investigated.
There have been some reports of competitive inhibitory interactions between NotchICD paralogues. There has been in vitro evidence presented that the NOTCH3ICD can competitively inhibit NOTCH1ICD at Hes1 and Hes5 promoters (Beatus et al., 1999). These in vitro over expression studies were supported by in vivo observations of mice over expressing Notch3ICD. Notch3ICD expression in the developing central nervous system led to a reduction in Hes-5 expression. In addition, expression of Notch3ICD in progenitors of the developing pancreas leads to reduced Hes-1 expression which induces aberrant endocrine cell differentiation (Apelqvist et al., 1999). Similar effects in the pancreas were also observed in mice deficient for Rbpj or the ligand Dll1 suggesting that NOTCH3ICD suppresses canonical Notch signalling. NOTCH2ICD has also been reported as a competitive inhibitor. Over expression of NOTCH2ICD led to an inhibitory effect on NOTCH1ICD and NOTCH3ICD mediated gene activation (Shimizu et al., 2002). It is important to note that all of these inhibitory affects were seen in over expression systems. The inhibitory affects may be due to sequestration of factors, such as RBPJ, which under normal physiological levels of expression are in large excess compared to NotchICDs. However, the possibility remains that Notch receptor paralogues, by differing in transactivation potential at various promoters and competing for common factors, can act as inhibitors of Notch signalling.
4.1.2 Aims In Chapter 3 we were unable to generate ligand dependent activation of the NOTCH4 receptor. In fact not only did we see a lack of activation we saw an inhibition of Notch signalling. NOTCH4 reduced signalling via endogenous receptors when transiently transfected into C2C12 and MAEC cells (Section 3.2.2). Stable over expression of NOTCH4 in NIH3T3 cells reduced signalling compared to control co-cultures (Section 3.2.3 and 3.2.6). Inhibition of Notch signalling by a full length Notch receptor was a novel finding. Our aim in this chapter was to
130 | Page
further characterise this inhibitory effect with particular reference to Notch1 which is co-expressed with Notch4 in the endothelium.
131 | Page
4.2 Results
4.2.1 NOTCH4 dose dependently inhibits NOTCH1 signalling. The lack of NOTCH4 signal transduction in response to ligand observed in Chapter 3 could be due to a number of differences between NOTCH4 and the other Notch receptors. Notch4 is by far the most divergent of the Notch family of receptors. Although the NOTCH4 receptor contains many of the functional domains of the Notch family of receptors including the ligand binding EGF-like repeats, the NRR region, the RAM and Ankyrin repeats and the C- terminal PEST domain, there are important differences. The NOTCH4ICD has a markedly reduced capacity to activate gene expression compared to NOTCH1ICD (Section 3.2.5) and lacks an identifiable transactivation domain (TAD). Full length NOTCH4 is found on the cell surface in lower amounts than NOTCH1, reducing its capacity to interact with ligand expressed in trans (Section 3.2.9). The reduced surface expression of NOTCH4 may be due to minimal S1 processing (Section 3.2.8), which in the case of NOTCH1 is an absolute requirement for receptor surface expression and activation (Bush et al., 2001; Logeat et al., 1998). However, in NIH3T3 cells where some S1 processing was evident, there was no enrichment on the cell surface for the S1 processed over the full length NOTCH4 receptor (Section 3.2.9). These results indicate NOTCH4 may differ from NOTCH1 in the requirement of S1 processing for surface presentation. Weak S1 processing, surface expression of full length receptor and the subcellular localisation of NOTCH4 suggest a different maturation pathway for this receptor compared to NOTCH1. These differences, in addition to the relatively low transactivation potential of NOTCH4ICD (Section 3.2.5), can explain a lowered activation potential, possibly below the sensitivity of our assay system leading to an undetectable level of NOTCH4 receptor activation. Importantly NOTCH4 not only lacks detectable signal transduction in response to ligand but also reduced signalling via endogenous receptors when transiently transfected into C2C12 and MAEC cells (Section 3.2.2). Stable over expression of NOTCH4 in NIH3T3 cells reduced signalling compared to control co-cultures (Section 3.2.3 and 3.2.6). An inhibitory activity for a full length Notch receptor has not been reported for any member of the Notch family and has important implications for the possible function of NOTCH4. To further explore possible inhibition of NOTCH1 by NOTCH4, a number of competition experiments based on co- culture assays were conducted in cells co-expressing NOTCH1 and NOTCH4.
132 | Page
To investigate the effect of NOTCH4 expression on Notch signalling, NIH3T3 cells were co-transfected with a constant amount of Notch1HA, increasing amounts of Notch4HA or control vector plus the Notch reporter pGL46xTP1 and the transfection control pCMXren. Transfected cells were co-cultured with either Dll4 cells (Figure 4.1a) or Jag1 cells (Figure 4.1b) or control cells. As previously seen, Notch1HA transfected cells were robustly activated upon co-culture with Dll4 cells (Figure 4.1a, columns 1 and 6) and Jag1 cells (Figure 4.1b, columns 1 and 6) compared to co-culture with control cells. Addition of increasing amounts of Notch4HA to the transfection resulted in a dose dependent decrease in the level of NOTCH1 signalling observed despite the fact that all transfections included the same amount of Notch1HA plasmid. NOTCH4 inhibited NOTCH1 signalling both in response to DLL4 (Figure 4.1a, columns 6 and 7-10) and JAG1 (Figure 4.1b, columns 6 and 7-10). The inhibition of Notch activation was statistically significant (P<0.05) in response to Dll4 cells at 100ng of Notch4HA (Figure 4.1a, column 9) and highly significant (P<0.01) at 200ng of Notch4HA (Figure 4.1b, column 10). The inhibition of JAG1 dependent NOTCH1 signalling was statistically significant (P<0.05) at 50ng of Notch4HA (Figure 4.1b, column 8), highly significant (P<0.01) at 100ng (Figure 4.1b, column 9) and very highly significant (P<0.001) at 200ng (Figure 4.1b, column 10). Although not reaching statistical significance at lower levels of Notch4HA, the overall trend of NOTCH4 inhibition of NOTCH1 signalling was consistent at all levels of transfection.
NOTCH4 expression inhibited endogenous Notch receptor signalling in C2C12 and MAEC cells (Section 3.2.3). MAEC cells are derived from murine arterial endothelial cells (Nishiyama et al., 2007) and form part of the endogenous expression domain of both Notch1 and Notch4. To determine if NOTCH4 could inhibit NOTCH1 signalling in an endothelial cell line, MAEC cells were co- transfected with a constant amount of Notch1myc (Figure 4.1c, columns 2-5 and 10-13), increasing amounts of Notch4HA (Figure 4.1c, columns 3-8 and 11-16) or control vector (Figure 4.1c, columns 1 and 9) plus the Notch reporter pGL46xTP1 and the transfection control pCMXren. Following transfection, the MAEC cells were co-cultured with Dll4HA or control cells and the relative luciferase activity measured. In agreement with the results presented in Chapter 3, introduction of Notch1myc on its own enhanced the relative luciferase activity compared to vector control both when co-cultured with control cells (Figure 4.1c, columns 1 and 2), and Dll4HA cells (Figure 4.1c, columns 9 and 10). Co-culture with Dll4HA cells induced a robust activation of luciferase activity in Notch1myc transfected cells (Figure 4.1c, columns 2 and 10). Transfection of increasing amounts of Notch4HA dose dependently inhibited NOTCH1 signalling upon co- 133 | Page
culture with either control cells (Figure 4.1c, columns 2-5) or Dll4HA cells (Figure 4.1c, columns 10-13). A similar trend was observed in MAEC cells co-transfected with Notch4HA alone. A modest induction of signalling occurred in response to Dll4HA cell co-culture (Figure 4.1c, columns 1 and 9), which was progressively inhibited by transfection of increasing amounts of Notch4HA plasmid (Figure 4.1c, columns 14-16).
Figure 4.1 NOTCH4 dose dependently inhibits NOTCH1 signalling in a co-culture assay.
a. NIH3T3 cells were co-transfected with Notch1HA or Notch1HA and increasing amounts (25ng, 50ng, 100ng and 200ng) of Notch4HA plus the Notch responsive reporter (pGL46xTP1) and a Renilla luciferase transfection efficiency control (pCMXren). The transfected cells were co-cultured overnight with either a NIH3T3 control cell line or a DLL4 over expressing line. The data points represent fold over vector control of the relative luciferase units normalised for transfection efficiency by dividing by the Renilla luciferase activity. This figure represents the mean of three independent experiments with bars as standard deviation. Notch1HA alone co-cultured with ligand expressing cells was compared to Notch1HA and Notch4HA co-expression using ANOVA in Prism software (*P<0.05 and **P<0.01). b. NIH3T3 cells were co-transfected with Notch1HA or Notch1HA and increasing amounts (25ng, 50ng, 100ng and 200ng) of Notch4HA plus the Notch responsive reporter (pGL46xTP1) and a Renilla luciferase transfection efficiency control (pCMXren). The transfected cells were co-cultured overnight with either a NIH3T3 control cell line or a JAG1 over expressing cell line. The data points represent fold over vector control of the relative luciferase units normalised for transfection efficiency by dividing by the Renilla luciferase activity. This figure represents the mean of three independent experiments with bars as standard deviation. Notch1HA alone co-cultured with ligand expressing cells was compared to Notch1HA and Notch4HA co-expression using ANOVA in Prism software (*P<0.05, **P<0.01 and ***P<0.001). c. MAEC cells were co-transfected with Notch1myc, increasing amounts (125ng, 250ng and 375ng) of Notch4, Notch1myc and increasing amounts (125ng, 250ng and 375ng) of Notch4 or vector control plus the Notch responsive reporter (pGL46xTP1) and a Renilla luciferase transfection efficiency control (pCMXren). The transfected cells were co-cultured overnight with either a NIH3T3 control cell line or a DLL4HA over expressing line. The data points represent the relative luciferase units normalised for transfection efficiency by dividing by the Renilla luciferase activity. The figure represents the mean of four replicate transfections with bars as standard deviations.
134 | Page
135 | Page
4.2.2 The intracellular domains of NOTCH1 and NOTCH4 do not compete. The activation of the Notch pathway requires a number of protease cleavages that result in the release of the NotchICD which translocates to the nucleus to activate target gene expression. In Chapter 3, Notch constructs lacking the extracellular domain (NotchECT) were used to demonstrate that the NOTCH4ICD was capable of activating transcription. It was observed that Notch4ECT, although capable of activating the Notch responsive reporter 480-fold over control, was relatively weak in comparison to Notch1ECT which was able to induce more than a 5,000-fold increase of the reporter. Although there have been no reports of a full length Notch receptor acting as an inhibitor of transcription there have been some reports that NOTCH3ICD can inhibit the activation of a reporter by NOTCH1ICD (Beatus et al., 1999; Beatus et al., 2001). The competition observed in Section 4.2.1 may have been due to competition between the NOTCH1ICD and NOTCH4ICD.
To investigate the potential for an inhibitory interaction between the NOTCH1ICD and NOTCH4ICD a competition experiment was performed. NIH3T3 cells were co-transfected with increasing amounts of Notch1ECT, increasing amounts of Notch4ECT, a constant amount of Notch1ECT co-transfected with increasing quantities of Notch4ECT or vector control plus the Notch responsive reporter (pGL46xTP1) and the transfection control plasmid (pCMXren). The cells were cultured overnight and the relative luciferase activity measured. Transfection of increasing amounts of the Notch1ECT (Figure 4.2, columns 2-5) resulted in a corresponding increase in reporter activation. Notch4ECT transfected cells also displayed increased activation as more Notch4ECT plasmid was introduced (Figure 4.2, columns 6-9). The signal produced from Notch1ECT was not saturated at the lowest quantity used in this experiment (Figure 4.2, column 2) since greater activation was observed with higher quantities of plasmid transfected. The amount of Notch1ECT was kept constant and co- transfected at this non-saturating level with increasing quantities of Notch4ECT (Figure 4.2, columns 10-13). Co-transfection of increasing amounts of Notch4ECT had no affect on the reporter activity induced by Notch1ECT (Figure 4.2, columns 2 and 10-13). Thus in this over-expression assay, competition was not observed between the NOTCH4ICD and NOTCH1ICD.
136 | Page
Figure 4.2 NOTCH4ICD does not inhibit NOTCH1ICD signalling.
NIH3T3 cells were co-transfected with increasing quantities of Notch1ECT (25ng, 50ng, 100ng and 200ng), increasing quantities of Notch4ECT (25ng, 50ng, 100ng and 200ng), a constant amount of Notch1ECT (25ng) and increasing quantities of Notch4ECT (25ng, 50ng, 100ng and 200ng) or vector control plus the Notch responsive reporter (pGL46xTP1) and the Renilla transfection control (pCMXren). The data points represent the relative luciferase units normalised for transfection efficiency by dividing by the Renilla luciferase activity. This figure represents the mean of four replicate transfections experiments with bars indicating the standard deviation.
137 | Page
NOTCH4ICD did not appear to compete with NOTCH1ICD for activation of the Notch responsive reporter. However, additive effects of Notch4ECT were not observed even though the Notch1ECT was transfected at non-saturating levels. The Notch1ECT construct is an extremely potent activator of the Notch reporter which could possibly be covering a potential interaction. The maximal activation potential of the Notch4ECT was within the measurement error of the results obtained with the Notch1ECT. In addition the levels of gene activation seen when using the NotchECT constructs are far in excess of those seen in ligand dependent activation and therefore may not accurately represent the physiological scenario. For example, when ligand is used to activate NOTCH1 signalling by co-culture, the Notch reporter is typically activated by approximately 30-fold over the vector control. To further investigate if Notch4ECT inhibits NOTCH1 signalling, a co-culture assay was performed as it better represents endogenous Notch signalling than that achieved using ECT constructs of Notch.
NIH3T3 cells were co-transfected with a constant amount of Notch1HA and increasing quantities of Notch4ECT or control vector plus the Notch responsive reporter pGL46xTP1 and the transfection control plasmid (pCMXren). The transfected cells were then co-cultured with Dll4HA cells and the relative luciferase activity measured. In the absence of Notch4ECT, luciferase activity was induced 27-fold upon co-culture with Dll4HA cells compared to control cells (Figure 4.3, columns 1 and 6) due to signalling from full length NOTCH1 receptor. Co-transfection of increasing amounts of Notch4ECT induced a dose dependent increase in Notch reporter activity in cells co-cultured with control cells (Figure 4.3, columns 2-5) or Dll4HA cells (Figure 4.3, columns 7-10). The strong activation of the reporter by the Notch4ECT in the control co-culture compared to the activation seen with full length NOTCH1 co-cultured with Dll4HA cells (Figure 4.3, columns 6 and 2-5) highlights the difference in activity between the constitutively active construct (Notch4ECT) and the activity produced by ligand activation. Co-culture of cells expressing Notch1HA and Notch4ECT with Dll4HA cells induced higher reporter activity than control cell co-culture (Figure 4.3, columns 2 and 7, 3 and 8, 4 and 9 and 5 and 10), an effect attributable to signalling from intact NOTCH receptor in response to Dll4HA cells. The increase in activity indicates that even with the very large excess of NOTCH4ICD introduced by the use of the Notch4ECT there is no inhibition of ligand dependent signalling from the full length NOTCH1. In fact co-transfection of Notch4ECT not only had an additive but synergistic effect on signalling from NOTCH1 in response to ligand. The co-operative effect of NOTCH1ICD and NOTCH4ICD may be due to the presence of six 138 | Page
RBPJ binding sites in the Notch reporter requiring multiple NotchICDs binding to initiate transcription. These results indicate that the competition seen between NOTCH1 and NOTCH4 in the co-culture assays was not due to competition between the NotchICDs for access to the promoter.
139 | Page
Figure 4.3 Notch4ECT has a synergistic effect on NOTCH1 activation.
NIH3T3 cells were co-transfected with a constant amount of Notch1HA (25ng) and increasing quantities of Notch4ECT (25ng, 50ng, 100ng and 200ng) or vector control plus the Notch responsive reporter (pGL46xTP1) and a Renilla transfection efficiency control (pCMXren). The transfected cells were co- cultured overnight with either NIH3T3 cells over expressing DLL4HA (Dll4HA cells) or control cells. Data points represent the relative luciferase units normalised for transfection efficiency by dividing by the Renilla luciferase activity. This figure represents the mean of three independent experiments with bars as standard deviation.
140 | Page
4.2.3 NOTCH4 acts as a dose dependent inhibitor of NOTCH1 in a stable cell line The ability of full length NOTCH4 to inhibit NOTCH1 signalling was a novel observation that required careful evaluation of the experimental protocol used. The competition experiments described above demanded a complex transfection strategy combining a variety of plasmids. In all cases discussed in Sections 4.2.1 and 2, the transfection mixes used contained an equal quantity of DNA. Control plasmid (pCMXCAT), which has an identical parent vector to both the Notch1 and the Notch4 constructs was used to keep the amount of DNA transfected into cells constant and to control for potential competition among promoters of transfected expression constructs. One argument against promoter competition occurring in these assays is the observation that NOTCH4 also suppressed endogenous Notch activity (Figure 4.1c). To confirm that NOTCH4 inhibits NOTCH1 signalling in an assay not reliant on co-transfection, a number of experiments were performed using stable Notch1myc and Notch1HA cell lines. The use of stable lines keeps the amount of Notch1 constant independently of transfection.
The NIH3T3 cell line Notch1myc stably over expresses the NOTCH1 protein with an internal myc tag. These cells were co- transfected with increasing quantities of Notch4HA or vector control plus the Notch responsive reporter (pGL46xTP1) and the transfection control (pCMXren). Transfected cells were co-cultured with either Dll4HA cells or control cells. In the absence of Notch4HA, Notch1myc cells were robustly activated (~40-fold) upon co-culture with Dll4HA cells (Figure 4.4a, columns 1 and 6). Transfection of increasing amounts of Notch4HA resulted in a dose dependent decrease in reporter activity, consistent with the co- transfection experiments. Notch4HA decreased reporter activity in response to co-culture with control cells (Figure 4.4a, columns 1- 5) and ligand cells (Figure 4.4a, columns 6-10).
In order to show that this effect was independent of the cell line used, a similar experiment was conducted using a C2C12 line stably expressing Notch1HA. Again, this line displayed robust activation of the Notch reporter in response to ligand (Figure 4.4b, columns 1 and 6). Consistent with the above experiments, introduction of Notch4HA dose dependently inhibited the activation of NOTCH1 signalling in response to ligand (Figure 4.4b, columns 7-10). Taken together, these findings indicate that the ability of NOTCH4 to inhibit NOTCH1 is a general phenomenon, not restricted to one type of cell line or experimental design.
141 | Page
Figure 4.4 NOTCH4 inhibits signalling by stably expressed NOTCH1. a. A NIH3T3 cell line stably expressing NOTCH1 (NIH3T3Notch1myc) was co-transfected with increasing quantities of Notch4HA (25ng, 50ng, 100ng or 200ng) or vector control plus the Notch responsive reporter (pGL46xTP1) and the Renilla transfection control (pCMXren). The transfected cells were co- cultured with either Dll4HA cells. The data points represent the relative luciferase units normalised for transfection efficiency by dividing by the Renilla luciferase activity. This figure represents the mean of four replicate transfections with bars as standard deviation. b. A C2C12 cell line stably expressing NOTCH1 (C2C12Notch1HA) was co-transfected with increasing quantities of Notch4HA (25ng, 50ng, 100ng or 200ng) or vector control plus the Notch responsive reporter (pGL46xTP1) and the Renilla transfection control (pCMXren). The transfected cells were co-cultured with either Dll4HA cells. The data points represent the relative luciferase units normalised for transfection efficiency by dividing by the Renilla luciferase activity. This figure represents the mean of four replicate transfections with bars as standard deviation.
142 | Page
143 | Page
4.2.4 NOTCH4 inhibits myogenic differentiation of C2C12 myoblasts. Data presented thus far indicate that NOTCH4 does not signal (Chapter 3). Instead, NOTCH4 inhibits NOTCH1 signal transduction (this chapter). These experiments have mostly relied on luciferase assays and a Notch responsive luciferase reporter to generate quantitative data of signal output. However, Notch signalling in vivo has many distinct developmental consequences including maintenance of the undifferentiated state, induction of differentiation, border formation and cell fate specification (Bray, 2006; Lai, 2004; Penton et al., 2012). Therefore it is important to determine if NOTCH4 can alter Notch dependent biological processes via its ability to inhibit signalling from NOTCH1.
C2C12 cells are a myoblast line that resembles cultured satellite cells in their ability to form multinucleated myotubes at confluency, following serum withdrawal (Kopan et al., 1994). C2C12 cells remain in an undifferentiated state at low cell density and the induction of differentiation in confluent cultures can be inhibited by suppression of Notch signalling. This process is characterised by a dramatic change in cell morphology and cell fusion to form myotubes and induction of myosin heavy chain (MHC) expression. The differentiation of C2C12 cells was used to determine if the inhibitory capacity of NOTCH4 alters the Notch dependent differentiation of C2C12 cells into myotubes. In order to achieve this C2C12 cell lines were established that over expressed Notch4HA or the control vector stably integrated into the genome. The use of these cells overcame complications inherent in the preceding experiments. Firstly, since stable lines were established all cells in the population expressed Notch4. Secondly, the readout of the assay was cell differentiation and so no artificial reporter constructs were required. In addition, Notch1 was expressed from its endogenous promoter at physiological levels.
C2C12 cells were transfected with pCAGNotch4HAiPuro or pCAGiPuro and selected for puromycin resistance. Individual clones were isolated and immunostained to confirm NOTCH4 expression. Corresponding clones that had incorporated the parental vector pCAGiPuro were used as controls. Most C2C12 cell differentiation protocols use a low percentage of horse serum at confluence to achieve the optimal myotube formation. We found that serum deprivation was not necessary to induce differentiation in cells over expressing NOTCH4 (Figure 4.5b). In these experiments the cell lines were grown in standard media used to propagate the C2C12 line in its undifferentiated state (DMEM supplemented with 10% foetal calf serum). Initially a single clone of C2C12Notch4HA (C2C12Notch4HA6) and a single clone of
144 | Page
C2C12iPuro (C2C12iPuro1) were analysed. C2C12Notch4HA and C2C12iPuro cells were seeded at similar densities and grown to confluence over three days.
The C2C12iPuro cells maintained an undifferentiated morphology with single nuclei in each cell and a compact cell area (Figure 4.5a). In contrast, many C2C12Notch4HA cells had undergone a striking change in cell morphology. The C2C12Notch4HA cultures contained cells that had differentiated into myotubes (Figure 4.5b; marked with an asterisk). These cells displayed characteristics of myotube differentiation. Firstly, there was a dramatic increase in cell volume among selected cells. Secondly, these cells displayed syncytium formation with a large number of nuclei sharing a common cytoplasm. No cells with these morphological characteristics were observed in the control cells.
The extent of myoblast differentiation varies between individual C2C12 clones. The plasmid introduced into these cells randomly integrates into the genome and in the case of the C2C12Notch4HA lines examined may have affected the differentiation rate irrespective of NOTCH4 expression. To establish that Notch4 does indeed drive differentiation in C2C12 cells, Kavitha Iyer in our laboratory examined additional clones of C2C12Notch4HA and C2C12iPuro. Examination of the cells by morphological characteristics alone requires a subjective assessment. To objectively assess differentiation, expression of the differentiation marker myosin heavy chain (MHC) was used.
145 | Page
Figure 4.5 NOTCH4 over expression induces myotube formation in C2C12 cells.
C2C12 were transfected with pCAGiPuro (a) and pCAGNotch4HA (b). The cells were selected in 1µg/mL puromycin and single clones isolated. The cells were grown to confluence over three days. Photographs were taken of ten random fields in each culture (10x objective). Representative images are shown in the figure. Myotubes are indicated with an asterisk (*).
146 | Page
Three clones of C2C12iPuro (C2C12iPuro2, C2C12iPuro3 and C2C12iPuro4) and three clones of C2C12Notch4HA (C2C12Notch4HA4, C2C12Notch4HA6 and C2C12Notch4HA7) were chosen for analysis. The C2C12Notch4HA clones were chosen after immunostaining with anti-HA to confirm NOTCH4 expression. Clones were seeded at the same density and grown to confluence over three days. The confluent cultures were stained for MHC (Figure 4.6, green) and counterstained with the nuclear stain TO- PRO-3 (Figure 4.6, red). There was a marked increase in the number of cells positive for MHC in the C2C12Notch4HA lines examined (Figure 4.6b) compared to the C2C12iPuro lines (Figure 4.6a). The Notch4HA lines displayed some variation in the extent of differentiation. C2C12Notch4HA6 displayed the most robust response. C2C12Notch4HA6 had high MHC expression and large multi-nucleated syncytia aligned as myotubes (Figure 4.6b). The two additional clones differentiated to a lesser extent. However, in comparison to the C2C12iPuro lines (Figure 4.6a) they had higher MHC expression and the cultures contained multinucleated cells. The C2C12ipuro lines contained a small number of cells that were positive for MHC (Figure 4.6a). No multinucleated cells were observed in these cultures.
The extent of differentiation was measured by quantifying MHC expression. The three clones of C2C12iPuro and C2C12Notch4HA were seeded at the same density and grown to confluence over three days. The cultures were analysed by western blot for MHC expression (Figure 4.7a). Expression of MHC could be clearly identified in the three C2C12Notch4HA clones examined as a band running at 220kDa (Figure 4.7a, lanes 2, 4 and 6). MHC expression was faint to undetectable in the C2C12iPuro lines (Figure 4.7a, lanes 1, 3 and 5). The bands were quantified and the expression levels of MHC, normalised to beta-tubulin, were measured (Figure 4.7b). There was a significant (P<0.05) 30-fold mean induction of MHC expression in the C2C12Notch4HA lines compared to the C2C12iPuro lines.
These results establish that NOTCH4 can act as an inhibitor of Notch dependent biological processes.
147 | Page
Figure 4.6 NOTCH4 induces MHC expression in C2C12 cell lines. a. C2C12iPuro cell lines (C2C12iPuro2, C2C12iPuro3, C2C12iPuro4) and b. C2C12Notch4HA cell lines (C2C12Notch4HA4, C2C12Notch4HA6, C2C12Notch4HA7). Cell lines were grown to confluence over three days and immunostained for myosin heavy chain (MHC) expression (green) and the nuclei counterstained with TO-PRO-3 (red).
148 | Page
149 | Page
Figure 4.7 NOTCH4 induces a 30-fold increase in MHC expression in C2C12 cells. a. Western blot of three control C2C12 lines (C2C12iPuro4, C2C12iPuro3, and C2C12iPuro2) and three Notch4HA lines (C2C12Notch4HA7, C2C12Notch4HA6 and C2C12Notch4HA4). The positions of myosin heavy chain (MHC) and the loading control beta-tubulin are indicated on the right. Molecular weight markers are indicated on the left. b. The expression levels of MHC normalised to beta-tubulin expression. These data are the mean of the three control lines (C2C12iPuro) and the three Notch4 lines (C2C12Notch4HA) with bars as standard deviation. Statistics were calculated by t-test (P<0.05).
150 | Page
151 | Page
4.2.5 NOTCH1 expression in the presence of NOTCH4 Data presented thus far indicate that NOTCH4 acts as an inhibitor of NOTCH1 signalling. The inhibitory effect of NOTCH4 has been confirmed in co-cultures using an artificial reporter in a variety of cell types. In addition, NOTCH4 inhibits a Notch dependent biological process, namely the maintenance of the undifferentiated myoblast cell state in vitro. NOTCH4 specifically alters signalling through the NOTCH1 receptor. NOTCH4 may inhibit NOTCH1 signalling by altering NOTCH1 receptor maturation or stability. To address these possibilities, the affect of NOTCH4 co-expression on NOTCH1 protein levels was analysed by western blot.
NIH3T3 cells expressing NOTCH1 (Notch1myc) cells were transfected with increasing quantities of Notch4HA (Figure 4.8a, lanes 2-5) or control plasmid (Figure 4.8a, lane 1). The protein was extracted and analysed by western blot using the myc antibody to detect Notch1myc, the HA antibody to detect Notch4HA and beta-actin as a loading control. As expected the stable cell line Notch1myc expressed bands corresponding to both full length NOTCH1 (NOTCH1Fl) at 300kDa and S1 processed NOTCH1TMIC at 100kDa in all samples. A band corresponding to full length NOTCH4 (NOTCH4Fl) was observed at 200kDa which increased with increasing levels of Notch4HA transfection (Figure 4.8a, lanes 2-5). NOTCH4TMIC was undetectable at this level of expression. The Notch1myc bands were quantified. The luciferase experiments, transfected with similar quantities of Notch4HA, showed inhibition of NOTCH1 activation (Section 4.2.3). When the protein levels were evaluated, the amount of S1 processed NOTCH1 remained at a relatively steady state when increasing quantities of Notch4HA were transfected into the cells. However, the percentage of intact NOTCH1 (NOTCH1Fl) detected at 300kDa as a proportion of total NOTCH1 (NOTCH1Fl plus NOTCH1TMIC) increased as more Notch4HA was transfected into the cells (Figure 4.8b). In the case of NOTCH1 only the S1 processed heterodimer is expressed on the cell surface and is capable of signalling (Bush et al., 2001; Logeat et al., 1998). The accumulation of full length NOTCH1 suggests an alteration in S1 processing.
152 | Page
Figure 4.8 NOTCH4 co-expression with NOTCH1 increases unprocessed NOTCH1.
A NIH3T3 cell line expressing myc tagged Notch1 was transfected with increasing quantities (100ng, 200ng, 400ng and 800ng) of HA tagged Notch4 or control vector. Cell lysates were analysed sequentially by western blot using antibodies for myc, HA and beta-actin (a). The position of the unprocessed NOTCH1 (NOTCH1Fl) and NOTCH4 (NOTCH4Fl) and the S1 processed NOTCH1 (NOTCH1TMIC) are shown on the right. The position of the molecular weight markers is shown on the left. The bands were quantified using ImageJ software (b). The data points represent the percentage of NOTCH1Fl with respect to total NOTCH1.
153 | Page
4.2.6 NOTCH4 co-expression alters the subcellular distribution of NOTCH1 NOTCH4 did not reduce NOTCH1 protein levels. However, NOTCH4 increased levels of full length NOTCH1, suggesting that NOTCH4 may inhibit S1 processing of NOTCH1. S1 processing of NOTCH1 is intimately associated with trafficking and surface presentation. To investigate the effect NOTCH4 has on NOTCH1 trafficking, the subcellular distribution of both receptors was assessed when co- expressed. The dynamics of subcellular trafficking of NOTCH1 is currently being investigated in our laboratory by Gavin Chapman using live imaging techniques. The data presented in Figure 4.9 are still images from these experiments.
Fluorescent derivatives of NOTCH1 and NOTCH4 were made in order to detect these proteins in live cells. A C2C12 line was established that stably expresses NOTCH1 fused to GFP (Notch1GFP) (Figure 4.9, green). A construct of Notch4 fused to mRuby (Notch4mRuby) was transiently transfected into these cells (Figure 4.9, red). Two representative fields are shown. Transient transfection results in only a subset of Notch1GFP cells co- expressing Notch4mRuby. Untransfected Notch1GFP cells (Figure 4.9, panels 1 and 2 white arrows) express Notch1GFP in distinct puncta as previously noted (Section 3.2.13). At least some of these puncta represent vesicles of the late endocytic compartment (Chapman et al 2011). Notch4mRuby in contrast, was found at high intensity in large intracellular accumulations as well as at lower levels in a more diffuse cytoplasmic pattern (Figure 4.9, panel 3 and 4, red arrows) in agreement with previous results (Section 3.2.13). In Notch4mRuby expressing cells, the expression of Notch1GFP colocalised with that Notch4mRuby (Figure 4.9, panels 5 and 6). There was a reduction in the vesicle expression of Notch1GFP upon co-expression with Notch4mRuby (compare cells highlighted by arrows in Figure 4.9, panels 1 and 2 vs. 3 and 4). Thus, in the presence of Notch4Ruby, Notch1GFP adopts the NOTCH4 subcellular distribution.
The inability of NOTCH4 to be activated by ligand was in part due to its subcellular distribution as the vast majority of NOTCH4 is localised inside the cell rather than at the cell surface (Section 3.2.13). The adoption of the NOTCH4 expression pattern by NOTCH1 in cells co-expressing both receptors has important consequences for signalling. NOTCH1 must enter vesicles to traffic from the Golgi to the cell surface. The association between vesicle trafficking and signal transduction suggests that the inhibition of NOTCH1 signalling by NOTCH4 is due to alterations in NOTCH1 subcellular distribution.
154 | Page
Figure 4.9 NOTCH4 co-expression alters NOTCH1 localisation C2C12 cells stably expressing a Notch1GFP fusion protein (green in panel 1, 3, 4, 6, 7 and 9) were transfected with a Notch4mRuby fusion construct (red in panel 2, 3, 5, 6, 7 and 9). White arrows in panel 1 and 4 indicate Notch1GFP cells not expressing Notch4mRuby. Red arrows in panel 2 and 5 indicate Notch1GFP cells transfected with Notch4mRuby. Two fields are shown (1, 2, 3 and 4, 5, 6) at high magnification and one field (7, 8 and 9) at low magnification. 155 | Page
4.3 Discussion In our initial experiments investigating canonical signalling by NOTCH4 we found that transfection with Notch4 inhibited endogenous Notch signalling. We further confirmed this effect in co-transfection experiments. NOTCH4 dose dependently inhibited signalling from NOTCH1 in the three cell lines tested (NIH3T3, C2C12 and MAEC). Stable NOTCH4 expression induced myotube formation in C2C12, a process that is inhibited by canonical Notch signalling. The observation that an intact Notch receptor can act as an inhibitor of canonical signalling is a novel finding that requires careful evaluation of the experimental methods used. Below is a discussion of the various mechanisms that could be responsible for this novel effect.
4.3.1 NotchICD competition NotchICD paralogues display a range of activation strengths on a variety of promoters both in vitro (Maier and Gessler, 2000; Ong et al., 2006) and in vivo (Aster et al., 2011). Previous in vitro studies have identified inhibitory functions for Notch3ICD when co- transfected with the Notch1ICD, and for Notch2ICD when co- transfected with either Notch1ICD or Notch3ICD (Beatus et al., 1999; Shimizu et al., 2002). The authors concluded that the NotchICD paralogues competed for a limited common activator. Alternatively, the NotchICDs may compete for a limited pool of reporter constructs. We found that the inhibitory effect of NOTCH4 was not due to the NOTCH4ICD. Not only did transfection with Notch4ECT not inhibit Notch activation, there was a synergistic effect evident. The synergism seen was probably due to the large number of RBPJ binding sites in the Notch reporter. NOTCH4ICD was able to activate the reporter in the presence of NOTCH1ICD. Thus the inhibitory effect of the NOTCH4 receptor could not be explained by NotchICD competition for a limited pool of either co- activators or reporters.
4.3.2 Plasmid competition Co-transfection of Notch1 and Notch4 resulted in a dose dependent inhibition of signalling from NOTCH1. The expression constructs used for both Notch1 and Notch4 were based on the same parental vector and hence the same constitutively active promoter. It was possible that the two vectors could be transfected at different efficiencies. Transfection efficiency can be influenced by the degree of supercoiling which can be variable between plasmid preparations (Sousa et al., 2009). Multiple independent plasmid preparations produced identical results making it unlikely that this effect could explain inhibition. In addition transfection with Notch4 could inhibit signalling from lines that stably 156 | Page
expressed NOTCH1 making it unlikely those transfection artefacts could be responsible for inhibition. In all co-transfection experiments performed, an equal quantity of total vector was transfected. The balance of the DNA was normalised with the addition of our control plasmid pCMXCAT. This plasmid is based on an identical parent plasmid to both the Notch1 and Notch4 constructs. In all cases an equal quantity of parent vector driving protein expression was transfected making competition for transcription between vectors an unlikely explanation.
We also observed inhibition of endogenous Notch signalling in co- culture assays. In these cases expression of endogenous Notch receptors was entirely independent of the expression plasmids. Finally NOTCH4 was able to induce C2C12 cell differentiation into myotubes, an event inhibited by Notch signalling, consistent with NOTCH4 inhibiting signalling through other Notch receptors. Expression of NotchICD constructs including NOTCH4ICD inhibits C2C12 cell differentiation (Ye et al., 2004). However, stable expression of full length NOTCH4 receptor induced C2C12 differentiation and hence inhibited Notch signalling.
4.3.3 Full length NOTCH1 accumulates in NOTCH4 co- expressing cells The co-expression of NOTCH4 in NOTCH1 expressing cells led to the accumulation of full length NOTCH1 receptor. S1 processing of NOTCH1 is essential for its function in cell surface presentation and its ability to be activated by ligand (Bush et al., 2001; Logeat et al., 1998). The accumulation of intact NOTCH1 receptor suggests that the inhibitory effect of NOTCH4 may occur before or at S1 processing and the divergence of NOTCH4 in terms of S1 processing is noteworthy here. Recent results have shown that S1 processing is a regulated, not a constitutive, event in NOTCH1 receptor maturation. FURIN processing of NOTCH1 is induced by growth factor stimulation and a direct interaction with Rous sarcoma oncogene (SRC) (Ma et al., 2012). Additionally a direct interaction between cryptic family 1 (CFC1) and the last two EGF- like repeats of NOTCH1 has been shown to recruit FURIN and augment NOTCH1 processing (Watanabe et al., 2009). FURIN processing is also under negative control by Chac1 (also known as blocks Notch (Botch) (Chi et al., 2012). CHAC1 binds to the S1 site of NOTCH1 and prevents S1 cleavage, surface presentation and signalling. Although NOTCH4 is highly divergent in this region, CHAC1 also binds to NOTCH4, albeit at lower affinity (3.3nM vs. 6.3nM). The regulated S1 cleavage and surface expression of NOTCH1 may therefore be the step where NOTCH4 inhibits NOTCH1 signalling.
157 | Page
The lack of any down regulation of the NOTCH1TMIC, essential for signalling, may be due to a number of factors. Firstly, the NOTCH1 heterodimer is expressed in all cells in the population while Notch4 was transfected into a subset of cells. A decrease in NOTCH1TMIC may therefore have been obscured by the presence of untransfected cells. In addition in our experiment there was no ligand activation of NOTCH1 and thus no consumption of NOTCH1TMIC via signalling. The pool of NOTCH1TMIC on the cell surface and in recycling endosomes may not have encountered NOTCH4. The potential for NOTCH4 to down regulate NOTCH1 S1 processing could be addressed by depleting the cells of mature NOTCH1TMIC. Depletion could be achieved through co-culture with ligand or the use of EDTA activation to remove active receptor and measure the rate of return of cell surface NOTCH1. In addition the timing of co-expression could be better controlled through the use of stable lines expressing Notch receptors from inducible promoters.
4.3.3.1 NOTCH1 alters its subcellular localisation and co- localises with NOTCH4 in response to co-expression Notch activation by ligand causes the clustering of receptor at the cell surface. Upon encountering ligand presenting cells, receptor containing intracellular vesicles are depleted in the vicinity of the point of contact and receptor/ligand clusters form on the cell surface (Bardot et al., 2005). This is followed by internalisation of receptor containing vesicles. When NOTCH4 was co-expressed in NOTCH1 cells there was a loss of the punctate staining of NOTCH1 (Figure 4.9). The loss of receptor positive vesicles available to participate in signalling could explain the inhibition of NOTCH1 signalling by NOTCH4.
4.3.4 Conclusions NOTCH4 can inhibit signalling from NOTCH1 in a variety of cell types (Section 4.2.1). The inhibition is not due to the intracellular domains (Section 4.2.2). Additionally, expression of NOTCH4 inhibits a Notch dependent biological process; C2C12 myoblast differentiation (Section 4.2.4). These results indicate that the inhibitory function of NOTCH4 is a general one and not due to the specific experimental protocols applied to each case. The co- expression of both NOTCH1 and NOTCH4 leads to NOTCH1 adopting the intracellular localisation of NOTCH4 (Section 4.2.5). This was associated with the accumulation of unprocessed full length NOTCH1 and provides a potential mechanism for the inhibitory capacity of NOTCH4. To our knowledge this is the first report of a full length Notch receptor acting as an inhibitor of canonical Notch signalling.
158 | Page
5 Notch4d1 is not a true null allele
5.1 Introduction Our investigations thus far have involved in vitro experiments. There are a number of transgenic mouse models available to study the role of Notch4 in vivo. These models can be broadly divided into gain- and loss-of-function groups. The gain-of-function models rely on the expression of truncated NOTCH4 proteins containing the intracellular domain (ICD). There is currently only one loss-of- function model available, Notch4tm1Grid, in which exons 22 and 23 of Notch4 have been deleted; herein referred to as the Notch4d1 mouse (Krebs et al., 2000).
5.1.1 Gain-of-function Notch4 models
5.1.1.1 NOTCH4ICD models-breast cancer Notch4 was first identified in the mouse as a common insertion site of a mouse mammary tumour virus (MMTV). A colony of mice derived from a single pair of trapped feral mice (M. Musculus subsp. Musculus strain Czech II), were found to have a 12% incidence of pregnancy independent type A mammary adenocarcinoma. The breast cancers were triggered by an infectious retrovirus transmitted via milk (Gallahan et al., 1987). The insertion of the virus drove oncogene expression. One site which was frequently (20% of cases) targeted was the Notch4 locus (Gallahan et al., 1987). The long terminal repeats (LTR) of the virus integrated within the Notch4 gene and drove expression of a truncated NOTCH4 protein utilising a methionine 30 amino acids N-terminal to the transmembrane domain as a start codon. The resulting allele was referred to as Int-3 for integration site 3. The Int-3 protein product therefore lacks the extracellular domain of NOTCH4 and is constitutively active like other extracellular domain truncated Notch receptors. The integration caused the over expression of Notch4ICD. A mouse model containing a construct based on the LTR/Notch4ICD developed mammary adenocarcinomas, severe ductal hyperplasia in the salivary glands and various other hyperproliferative lesions. In addition all male mice were sterile (Jhappan et al., 1992). Notch4ICD expression driven by mammary specific whey acidic protein (WAP) promoter also induced adenocarcinomas with 100% penetrance (Gallahan et al., 1996). It is important to note that the expression of the Notch4ICD was not driven by its endogenous promoter and does not reflect the normal tissue distribution/expression level of Notch4. However, NOTCH4 is expressed endogenously in breast tissue (Raafat et al., 2011) and has been reported to be over expressed in some breast cancers and cell lines (Imatani and Callahan, 2000). In the Notch4d1 mouse, where the Notch4ICD is deleted (see below), no abnormalities in breast development were 159 | Page
observed (Krebs et al., 2000). A role for intact endogenous NOTCH4 in normal mammary gland development and homeostasis has not been described.
5.1.1.2 Notch4ICD models – vascular specific expression Notch4 is expressed in a variety of cells and tissue types including breast (Raafat et al., 2011), kidney (Bonegio, 2009) and the haematopoietic system (Singh et al., 2000). However, the major expression domain of Notch4 is the vascular endothelium (Uyttendaele et al., 1996). Several mouse models, based on the expression of Notch4ICD, have been developed to investigate vascular specific roles for NOTCH4. The Notch4 transcript was found expressed throughout embryogenesis, in a pattern similar to Kdr (also known as Flk1) (Shirayoshi et al., 1997; Uyttendaele et al., 2001). Therefore mice were made where Notch4ICD was knocked-in to the Kdr locus (Uyttendaele et al., 2001). The mice died mid gestation (E9.5-10.5) due to vascular defects. At E8.5 there were no apparent differences in the vascular plexus of the embryo and yolk sac. Major vessels such as the dorsal aorta, the cardinal veins, intersomitic arteries and arteries of the brain formed in the mutant embryos suggesting the initial stages of vasculogenesis were unaffected. However, major defects became apparent in the vasculature by E9.5. The vasculature failed to remodel into fine branched structures throughout the embryo, including the brain and yolk sac. The major arterial vessels were dilated and vessel wall integrity was lost. Additional mouse models expressing Notch4ICD under the control of another endothelial specific inducible promoter, endothelial-specific receptor tyrosine kinase (Tek also known as Tie2), generated similar results (Carlson et al., 2005). There was an increase in the expression of arterial markers indicating defects in arterial/venous identity. These included shunts between arteries and veins that bypass capillary beds, compromising tissue perfusion. The arterial shunts observed in both the brain (Murphy et al., 2008) and lungs (Miniati et al., 2009) were found to be reversible. Removal of repression on the construct allowed for adult expression of NOTCH4ICD and re-administration suppressed expression. The reversible nature of these arterial/venous defects suggests that Notch signalling has a role in not just establishing but maintaining arterial identity.
5.1.2 Loss-of-function – the Notch4d1 mouse The extracellular domain of Notch receptors controls the rate, timing and ligand dependent production of the NotchICD while the NotchICD, upon its release, generates the Notch transcriptional response. A Notch4 “knockout” allele was created in which the Notch4ICD was no longer expressed. This allele, Notch4d1 (Krebs et al., 2000), contains a neomycin resistance cassette that removes exons 22 and 23 and prevents transcription of the
160 | Page
Notch4ICD coding sequence (Krebs et al., 2000). Deletion of NOTCH4ICD was thought to ablate the receptor's main function, canonical Notch signalling (Section 1.2.3). Notch4d1 homozygous mice are viable, fertile and display no overt phenotype. The author is aware of two reports of mild phenotypes for these mice on wild type backgrounds. Notch4d1 homozygous mice display a very mild elevation of systolic blood pressure compared to wild type littermates (Takeshita et al., 2007). Notch4d1 homozygous mice also display a persistently elevated level of urinary albumin excretion after nine months of age although no differences were observed in younger mice. The phenotype never progressed to renal failure (Bonegio, 2009).
Although no phenotype was reported for Notch4d1 homozygous mice on a wild type background, the original description of the Notch4d1 mice found a genetic interaction with Notch1 (Krebs et al., 2000). Notch4d1 homozygous mice on a haploinsufficient Notch1+/- background displayed mild growth retardation. This became apparent at weaning where Notch4d1/d1 Notch1+/- mice were on average 80% of the weight of Notch4d1/d1 Notch1+/+ littermates. Notch4d1 homozygous mice on a Notch1-/- background displayed more severe phenotypes than Notch4+/+ Notch1-/- mice in 50% of cases. These two results suggest that Notch4 has a largely redundant role to Notch1. The phenotypes in the compound mutants were apparent at E9.5. These included fewer somites, incomplete embryonic turning and open neural tubes. In the vasculature there was a more severe disruption of the anterior cardinal vein and dorsal aorta. The interpretation of these phenotypes is complicated by the already severe disruption of normal development in the Notch1-/- mice. These mice are growth arrested at or slightly before E9.5 (Conlon et al., 1995; Swiatek et al., 1994) and display massive defects in angiogenesis. The 50% penetrance of the more severe phenotypes in compound mutants could be due to a modifying locus present in the mixed genetic background (C57BL/6J × 129/SvImJ) on which these animals were maintained.
The description of the Notch4d1 homozygous mice only included a gross description of the vasculature at E9.5. The consensus from all the mouse models of the potential function of Notch4 in the vasculature is that it affects angiogenesis and maintenance of arterial identity not the earlier events of vasculogenesis. We reasoned that the action of Notch4 may be as a subtle modifier of vascular development. We turned to the mouse retina as a sensitive model of vascular development, which had the potential to uncover subtle defects in Notch4d1 homozygous mice.
161 | Page
5.1.3 The mouse retina as a model of angiogenesis The mouse retina has proven a valuable tool for investigating angiogenesis and has been particularly useful for elucidating the role of Notch signalling (Hellström et al., 2007; Hofmann and Iruela-Arispe, 2007; Suchting et al., 2007). Among the advantages of this model is the ease of microscopic analysis, because the vasculature initially develops in a single plane. Secondly, the vasculature develops in a well-described spatiotemporal pattern allowing analysis of various stages of angiogenesis. Below is a description of the development of the mouse retinal vasculature.
5.1.3.1 Overview of retinal angiogenesis The vasculature that supplies the inner portion of the retina develops postnatally in the mouse. Initially, the inner part of the eye is supplied by the arterial hyaloid vasculature. Blood enters through the central hyaloid artery in the optic nerve, runs through the vessels in the vitrous and exits through collection vessels around the circumference of the front of the eye. At birth the hyaloid vessels regress and a vascular plexus emerges from the optic nerve. The network extends radially in a strict spatiotemporal pattern to give rise to the retinal vasculature (Fruttiger, 2007). The growing network extends towards the periphery and reaches the edge by approximately day 8 post birth. This planar bed then extends deeper into the inner plexiform layer and extends a second plane followed by extension into the outer plexiform layer to form the final 3 tiered structure by 21 days post birth (Dorrell and Friedlander, 2006; Fruttiger, 2007) (Figure 5.1).
The network is remodelled into distinct arteries and veins and matures via vascular pruning and mural cell recruitment. Since these processes are occurring simultaneously in the retina with extension at the periphery (distal to the optic nerve) and maturation at the established centre (proximal to the optic nerve) this allows the simultaneous observation of different stages of angiogenesis making it an attractive model for studying angiogenic development.
162 | Page
Figure 5.1 Retinal Angiogenesis
Reproduced with permission from (Gerhardt et al., 2003).
Representation of angiogenesis in the mouse retina. On the left are top views of retinas of P1, P5 and P8 mice. Boxed areas indicate the side views shown in the centre panels. Panels on the right are micrographs of retinas stained with the endothelial specific dye isolectin (green). Arrows indicate where sprouting occurs at the periphery (P1 and P5) and into the deeper layers of the retina (P8).
163 | Page
5.1.3.2 Initiation of the retinal vasculature Prior to the development of the retinal vasculature (embryonic day 17, two days before birth) a population of cells derived from an astrocyte precursor lineage in the optic nerve responds to secretion of platelet derived growth factor alpha (PDGFA) by retinal ganglion cells (Fruttiger et al., 2000). The astrocytes proliferate rapidly and form a network across the inner surface of the retina (West et al., 2005). Prior to the development of blood vessels this network of cells experiences hypoxia and express the angiogenic factor Vegfa. Endothelial cells respond to VEGFA and use this network and the gradient of VEGFA generated as a template to lay down a vascular bed (Section 1.1.2.2). There is a strong developmental and evolutionary link between the astrocyte network and the retinal vasculature. In mammals only areas of the retina covered by blood vessels contain astrocytes. Astrocytes are absent from the avascular retina of the possum (Stone and Dreher, 1987), present only in the small vascularised region surrounding the optic nerve in the horse (Schnitzer, 1987), present only in a broad band associated with myelinated nerve fibres in the rabbit (Stone and Dreher, 1987) and absent from the avascular fovea of primates. In mice, the entire retina is covered by a network of both astrocytes and blood vessels (Engerman, 1976; Huxlin et al., 1992). As the developing blood vessels supply these cells with oxygen the astrocytes differentiate into a more mature phenotype. The astrocytes cease to proliferate and Vegfa is downregulated, providing a feedback loop to limit angiogenesis (Chu et al., 2001; Fruttiger, 2002; West et al., 2005). This process is thought to be angiogenic i.e. vessels sprouting from existing vessels. However, there have been suggestions that this process also incorporates circulating endothelial progenitors (Friedlander et al., 2007). The extent to which vasculogenesis may contribute remains controversial.
5.1.3.3 Vessel guidance The gradient of VEGFA guides the endothelial cells through interactions with a specialised “tip” cell. The tip cell phenotype is induced in endothelial cells in response to VEGFA (Gerhardt et al., 2003; Suchting et al., 2007). The tip cell extends long filopodia displaying the VEGFA receptor, KDR. These filopodia are of uniform thickness (~100nm) with the longest extending over 100µm (Gerhardt et al., 2003). Filopodia not associated with the underlying network of VEGFA producing astrocytes appear shorter and undulating. In a model proposed by (Bentley et al., 2008), tip cell fate is self reinforcing through purely geometrical considerations. Cells that bind VEGFA extend more filopodia which increases the surface area covered by the cell. This increases the amount of VEGFA bound and thus sets up a positive feedback loop. 164 | Page
The cells following the migratory tip cell, stalk cells, produce few filopodia and instead proliferate when exposed to VEGFA (Gerhardt et al., 2003). Stalk cells also form the vascular lumen (Iruela-Arispe and Davis, 2009). Stalk cells establish adherens junctions and tight junctions to maintain the integrity of the new sprout and establish luminal/abluminal polarity which leads to basal lamina deposition and mural cell recruitment/attachment (Dejana et al., 2009).
The induction of a tip cell phenotype cannot be solely explained in terms of a VEGFA gradient as the local concentration difference encountered by a tip cell and neighbouring stalk cell is very small. The tip/stalk cell phenotype is further enforced by Notch signalling. In response to VEGFA the tip cells express the Notch ligand Dll4 (Lobov et al., 2007). The tip cells activate Notch signalling in their neighbours which suppresses the tip cell phenotype (Duarte et al., 2004; Gale et al., 2004; Hellström et al., 2007; Krebs et al., 2004). Anastomosis, the fusion of tip cells, forms a new cell/cell boundary and an opportunity for further Notch driven signalling. Initial small differences in expression leads to one tip cell being inhibited and adopting a stalk cell fate. Potentially this signal can be passed on to neighbouring cells. Thus tip cell selection is a dynamic process where individual cells can alternate between tip and stalk cell fate (Bentley et al., 2009).
5.1.3.4 Vascular remodelling and maturation In the region following the leading tip cells, a dense capillary network is laid down. As more vessels are added at the growing edge this plexus is remodelled and matures into a hierarchical vascular tree. Differences in vessel diameter form and arteries and veins can be distinguished. Arterial identity is driven by further Notch signalling. Efbn2 is a direct transcriptional target of Notch and is a marker of arterial identity (Grego-Bessa et al., 2007). These initial studies suggested that venous identity is a default state. However, the venous transcription factor nuclear receptor subfamily 2, group F, member 2 (NR2F2, also known as COUP- TFII) actively represses the arterial marker neuropilin 1 (Nrp1) and Notch. Loss of Nr2f2 leads to an expansion of arterial cell fate which challenges the idea that venous is purely a default state (You et al., 2005). Bidirectional signalling through EFBN2, expressed in arterial cells, and EPHB4, expressed in venous cells, establishes and maintains arterial venous interactions and identity (Kullander and Klein, 2002) (Section 1.1.2.4).
Some of the capillaries formed are pruned while others are strengthened. This can occur both by migration of endothelial cells (Hughes and Chang-Ling, 2000) and apoptosis (Ishida et al., 2003). Apoptotic cell death is selective and driven by leukocytes
165 | Page
(Ishida et al., 2003). Interestingly, injection of liposomes containing clodronate, which ablates macrophages, has the opposite effect (Checchin et al., 2006). This indicates that there are complex interactions between the immune cells and the developing vasculature. Pruning is most pronounced in areas of highest oxygen and lowest VEGFA concentration i.e. at the centre near the optic nerve and around the arteries (Claxton and Fruttiger, 2003; Riva et al., 1986). If mouse pups are exposed to increased atmospheric oxygen within the first two weeks this effect becomes more pronounced. After the first two weeks the vasculature becomes resistant to hyperoxia induced ablation due to the recruitment of pericytes and smooth muscle cells which stabilise the vessels (Hellstrom et al., 1999; Ishida et al., 2003). Smooth muscle cell maturation is also driven by Notch signalling. JAG1 expression on endothelial cells (High et al., 2008) activates NOTCH3, which drives expression of both Jag1 and Notch3 and in turn promotes and maintains a differentiated smooth muscle cell phenotype (Liu et al., 2003). There is a cross talk between the endothelial and smooth muscle cells. Platelet derived growth factor B (PDGFB), secreted by endothelial cells, and angiopoietin 1 (ANGPT1), secreted by mural cells (Nishishita and Lin, 2004; Satchell et al., 2001; Suri et al., 1996) oppose differentiation (Campos et al., 2002). Thus the timing of vessel maturation is influenced by contributions from both the endothelium and smooth muscle.
After the initial capillary plexus reaches the periphery of the retina (post birth day 8) the outer plexus of the retinal vasculature develops (Dorrell and Friedlander, 2006). This is driven by transient expression of Vegfa from cells within the inner nuclear layer (Stone et al., 1995). Vascular sprouts grow along Mϋ ller cell processes perpendicular to the plane of the primary plexus. When the processes reach the inner and outer boundary of the inner nuclear layer they extend a second and third plane parallel to the first. The regulation of the development of the deeper plexus remains poorly understood although it is believed to depend on ANGPT1/TEK (Hackett et al., 2000; Hackett et al., 2002; Maisonpierre et al., 1997) and Wnt signalling (Luhmann et al., 2005; Ohlmann et al., 2005; Xu et al., 2004). By day 21 post birth the final three tiered structure is complete.
5.1.4 Aims
The retina provides a sensitive model system for evaluating angiogenesis. The evidence from in vivo models of Notch4 function suggests that the role of Notch4 in vessel development involves angiogenic growth and/or establishment or maintenance of arterial identity. This model system was used to evaluate angiogenesis in 166 | Page
mice homozygous for the Notch4d1 allele. During the course of this work the Notch4d1 allele was found to produce a transcript. The Notch4d1 transcript was characterised and the potential intact functions of Notch4 in the Notch4d1 mouse evaluated.
167 | Page
5.2 Results
5.2.1 Establishment, genotyping and breeding of Notch4d1 mice The Notch4d1 allele was constructed as a null allele by Krebs et al., 2000 (Figure 5.2a). The insertion of the neomycin resistance cassette (neo) replaces 1009bp of genomic sequence in the Notch4 locus (Figure 5.2a). This sequence contains exons 22 and 23 which encode amino acids 1249 to 1434 of the NOTCH4 protein. The neo cassette replaced the transmembrane domain and the methionine residue at amino acid 1411 which was identified as a start codon in the Int-3 locus (Robbins et al., 1992). The insertion of the cassette in the correct location was confirmed by Southern blot and PCR analysis (Krebs et al., 2000). Transcripts 3’ to the insertion site were shown to be absent by RNA in situ hybridisation and RT-PCR.
Notch4d1 mice were a kind gift from Thomas Gridley. Notch4+/d1 mice were rederived by embryo transfer into C57BL/6J mice and used to establish a breeding colony. Genotypes of mice within the Notch4d1 colony were determined by performing PCR on DNA isolated from either tail clips or ear clips. The positions of the primers used are shown in Figure 5.2a. Primers within exon 22 amplify a 313bp product from the unmodified Notch4 allele in Notch4+/+ and Notch4+/d1 mice (Figure 5.2b, lanes one and two of row one). Exon 22 was replaced by the neo cassette in Notch4 d1/d1 mice and no PCR product was observed (Figure 5.2b lane three of row one). Primers hybridising to sequences within exon 21 and the neo cassette amplify a 469bp product from the Notch4d1 allele (Figure 5.2b lanes two and three of row two). No product was observed in Notch4+/+ mice using this primer pair (Figure 5.2b lane one of row two). Mice were additionally genotyped with primers hybridising within the neo cassette that amplify a 518bp product (Figure 5.2b, lanes two and three of row three).
It was not known to what extent the Notch4d1 mice had been backcrossed into the C57Bl/6J line but were likely to be on a mixed CJ7/C57BL/6J background. In the original study 50% of Notch1-/- Notch4d1/d1 double mutant embryos displayed a more severe phenotype than Notch1-/- embryos. The 50% penetrance of this phenotype indicated the possible involvement of a second unlinked locus in the observed phenotype. To minimise this confounding factor the breeding colony was maintained by crossing heterozygous Notch4+/d1 with purebred C57BL/6J mice.
168 | Page
5.2.2 Confirmation of the insertion site of the neo cassette The insertion site of the neo cassette was confirmed by PCR amplifying the region surrounding the 5’ and 3’ insertion sites (Figure 5.2b rows two and four, respectively). The PCR product was cloned and sequenced. The 5’ integration site, GCCATGAGCTT (Figure 5.2c) was formed by blunt end ligation of the end-filled NcoI site of Notch4 and the end-filled HindIII site of neo cassette (Krebs et al., 2000). The 3’ site, AGCGGCCAAC was formed by the blunt end ligation of the end filled NotI site of the neo cassette and the end-filled HpaI site of Notch4 (Krebs et al., 2000) (Figure 5.2c). The detection of the PCR products appropriate to genotype and the sequencing of the integration site confirmed the identity of the Notch4d1 line.
5.2.3 Notch4d1/d1 mice display no overt phenotype compared to Notch4+/+ littermates. Previous investigations of the Notch4d1 allele reported that the homozygous mice were viable and displayed no overt phenotype. In order to confirm our colony of Notch4d1 mice displayed no overt phenotype early in life Notch4d1/d1, Notch4+/d1 and Notch4+/+ mice were generated by crossing pairs of Notch4+/d1 heterozygotes. Mouse pups were collected at day 3 post birth (P3), day 5 post birth (P5) and day 7 post birth (P7). The mice were weighed and genotyped. These time points were chosen due to the timing of retinal angiogenesis (see Section 5.2.4).
The weight of each mouse was normalised to the average weight of the heterozygotes in each litter to correct for interlitter variation. There was no significant difference in the weight of either the Notch4d1/d1 or the Notch4+/d1 mice compared to wild type littermates at any of the time points examined (Figure 5.3a, b and c). The expected progeny of a heterozygote cross with Mendelian inheritance would be 1:2:1 (Notch4+/+:Notch4+/d1:Notch4d1/d1). No significant deviation from Mendelian inheritance was observed at P3 (chi test=0.73, Figure 5.3a), P5 (chi test= 0.74 Figure 5.3b) or P7 (chi test=0.98 Figure 5.3c). Our results confirmed previous findings that the Notch4d1/d1 mice displayed no overt phenotype.
169 | Page
Figure 5.2 The Notch4d1 allele a. The Notch4 locus was targeted by Thomas Gridley via homologous recombination of a 5’ 3.3kb EcoRI to NcoI fragment and a 3’ 1.1kb HpaI to XbaI fragment of Notch4 flanking the neo cassette (Krebs et al 2000). The neo cassette replaced 1009bp of Notch4 including exons 22 and 23 (coding for amino acids 1249- 1434). Exons are shown as boxes and PCR primer annealing sites shown with arrows. b. The structure of the Notch4d1 locus and the genotypes of mice were confirmed by PCR with the primers indicated in (a). Primers within exon 22, flanking exon 21 and neo, within neo and flanking neo and exon 24 yielded specific products of 313bp, 469bp, 518bp and 735bp respectively. Using the above primers PCR products were size fractionated in agarose and representative genotypes shown. c. The PCR products containing the 5’ and 3’ insertion sites were cloned and sequenced. The sequence chromatograms of the integration sites are shown.
170 | Page
171 | Page
Figure 5.3 Notch4d1/d1 mice displayed normal postnatal growth and were produced in Mendelian ratios.
Neonatal mice were generated by crossing Notch4+/d1 mice. Mice were genotyped and weighed at day 3 post birth (a), day 5 post birth (b) and day 7 post birth (c). Bodyweight was normalised to the mean of the Notch4+/d1 pups in each litter. No significant difference in bodyweight was observed between genotypes at P3, P5 or P7 (P=0.58, 0.86, 0.62 respectively). Results are displayed as the mean with error bars representing standard deviation. Statistical significance was calculated using a one way ANOVA and Tukey’s post test using Graphpad Prism software. The genotypes displayed no significant difference to the expected Mendelian distribution at each time point (Chi test, P=0.73, 0.74 and 0.98 respectively, calculated using Microsoft Excel software).
172 | Page
5.2.4 Retinal angiogenesis in Notch4d1 mice We conjectured that the Notch4d1/d1 mice may in fact have a vascular phenotype that was not detected in previous studies due to the sensitivity of the methods used. The development of the vasculature in the mouse retina has proven to be a powerful model of angiogenesis and has been particularly useful for elucidating the role of Notch signalling (Hellström et al., 2007; Hofmann and Iruela-Arispe, 2007; Suchting et al., 2007). The mouse retina is avascular at birth. At birth endothelial cells migrate and proliferate from the region of the optic nerve. The vasculature develops in a strict spatiotemporal manner and reaches the periphery of the retina at P7-8 (Fruttiger, 2007) (Figure 5.1). The vasculature begins as a capillary plexus which then develops into distinct arterial, venous and capillary beds. These processes were analysed in the Notch4d1 mice at three distinct stages, P3, P5 and P7.
Notch4+/d1 mice were crossed and litters collected at P3, P5 and P7. The eyes were enucleated and the retina dissected out. The endothelial cells were identified by staining with an anti-PECAM1 antibody and a secondary antibody labelled with the fluorescent dye Cy3 (red). Retinas were co-stained with smooth muscle alpha- a ctin (SMA), a smooth muscle marker, and a secondary antibody labelled with Alexa-488 (green). Samples of tail tissue were also collected and used to genotype the mouse pups as described in Section 5.2.1.
Representative retinas from Notch4+/+ (Figure 5.4a), Notch4+/d1 (Figure 5.4b) and Notch4d1/d1 (Figure 5.4c) mice at P3 from a single litter are shown as examples. At P3 PECAM1 positive endothelial cells have developed into a capillary plexus that extends approximately one third of the distance to the edge of the retina in all three genotypes (Figure 5.4a, b and c). The vascular bed does not contain identifiable venous and arterial vessels at this stage. SMA expression was limited to areas surrounding the optic nerve at the centre. Prior to birth the retina is supplied with oxygen and nutrients by the hyaloid artery which we attempted to remove in the dissection of the retina. Additional staining can be seen in the Notch4+/+ example due to the presence of remnants of the hyaloid artery (Figure 5.4a). Dissections and staining were performed blind to genotype and there was no association between genotype and the amount of hyaloid tissue. Visual inspection of the retina revealed no obvious difference in retinas from Notch4+/+, Notch4+/d1 or Notch4 d1/d1 mice. We attempted to identify differences that were not immediately obvious by quantifying the area covered by the vasculature (Figure 5.5a), the retinal radius (Figure 5.5b) and the vascular density (Figure 5.5c).
173 | Page
The area covered by the vasculature at P3 was measured and no significant difference was found between genotypes (Figure 5.5a). Between birth and P3 there is a rapid expansion of the vasculature. Expansion was variable within genotypes with wild type Notch4+/+ mice spanning a range of 56 to 152% of the mean of the wild type range, Notch4+/d1 74 to 126%, and Notch d1/d1 47 to 129%. The large variation may potentially cover a phenotype at this time point. No significant difference between genotypes was detected in the radius of the retina (Figure 5.5b). The retina itself is not undergoing the rapid expansion seen in the endothelial cells at this stage. This was reflected in the relatively low variation in retinal radius measured within genotypes, Notch4+/+ 89-106%, Notch4+/d1 97-103% and Notch4d1/d1 91-106%. The density of the vascular plexus was variable at this time point, Notch4+/+ 61-151%, Notch4+/d1 69-140% and Notch4d1/d1 49- 144% (Figure 5.5c). Although variable within genotypes, there was no significant difference between genotypes.
Retinas from P5 mice generated by Notch4+/d1 crosses were analysed as above. At P5 the PECAM1 positive vascular plexus laid down at P3 approximately doubled in size in all genotypes (Figure 5.6a, b and c, red). In addition, the vasculature had remodelled into alternating arterial and venous vessels separated by capillary beds. The arterial vessels were identified as the major PECAM1 positive vessels radiating from the optic nerve towards the periphery of the retina. Initiating at the optic nerve, these vessels were also partially lined with smooth muscle cells expressing SMA (Figure 5.6a, b and c, green). The venous vessels were identified as the radial major vessels intervening between arteries. In all genotypes the arterial and venous vessels were correctly separated by an intervening capillary bed. No differences were observed by gross inspection.
Measurements of the vascular area at P5 revealed a significant difference between genotypes (Figure 5.7a). Notch4+/d1 and Notch4d1/d1 mice had a reduced vascular area compared to wild type (121 vs. 100%, P<0.05 and 121 vs. 96%, P<0.01 respectively). There was no significant difference between the Notch4+/d1 and Notch4d1/d1 mice (100 vs. 96%). At this time point the variation within genotypes was reduced compared to the P3 results, Notch4+/+ 107-141%, Notch4+/d1 81-118% and Notch4d1/d1 84-118%.
The difference in vascular area may have been due to a change in the growth rate of the retina itself. Measurements of the retinal radius revealed no significant difference between genotypes (Figure 5.7b). The differences observed in vascular area were therefore due to endothelial coverage rather than a change in the
174 | Page
size of the retina. The vascular density was also not significantly different between genotypes (Figure 5.7c), supporting the gross morphological observation that the vasculature in Notch4+/d1 and Notch4d1/d1 mice, although reduced in surface area, was otherwise normal.
The analysis of retinas was continued into P7 (Figure 5.8a, b and c). At P7 the PECAM1 positive vascular plexus reached the periphery of the retina in all three genotypes (Figure 5.8 in red). The gross morphology of the vasculature was similar in all three genotypes. Clearly identifiable arterial vessels characterised by high PECAM1 expression and SMA expression radiate and branch from the optic nerve. Venous vessels alternate between the arteries distinguishable by their reduced PECAM1 and SMA staining and higher calibre than the arterial vessels. In all genotypes the arterial and venous supply was correctly separated by capillary beds. Thus the gross morphology of the vasculature at P7 continues to appear normal in all three genotypes.
Measurements of the vascular area at P7 revealed no significant difference between genotypes (Figure 5.9a). At P7 the vasculature had reached the edge of the retina in all genotypes. There was also no significant difference in the retinal radius (Figure 5.9b). The vascular density was also not significantly different between genotypes (Figure 5.9c), supporting the gross morphological observation that the vasculature in Notch4+/d1 and Notch4d1/d1 mice was normal.
We have identified a growth delay in the retinal vasculature of both the Notch4+/d1 and Notch4d1/d1 mice at P5. The vasculature produced appeared otherwise normal and had recovered to be indistinguishable from wild type littermates by P7. It is of note that the growth delay was equally apparent in mice homozygous and heterozygous for the Notch4d1 allele. The presence of a phenotype in the heterozygous and homozygous Notch4d1 pups raised the possibility that the Notch4d1 allele was not a true null allele.
175 | Page
Figure 5.4 Retinal angiogenesis in Notch4d1 mice at day 3 post birth
Notch4+/d1 heterozygous mice were crossed and the retinas collected and stained with PECAM1 (red) and SMA (green) at day P3. Tail clippings (~2mm) were collected at the same time and used for genotyping. A 6x6 tile z-stack was captured for each retina using a 7Duo confocal microscope (Zeiss). The tiled stacks were flattened in the Z dimension using the average intensity function in ImageJ. Representative images of Notch4+/+ (a), Notch4+/d1 (b) and Notch4d1/d1 (c) mice from a single litter are shown.
176 | Page
177 | Page
Figure 5.5 Measurements of retinal angiogenesis at day 3 post birth
Retinas from mouse pups at P3 were collected and stained for PECAM1 and SMA expression. Tail samples were collected concurrently and used to genotype the pups. a. The vascular area was measured in ImageJ using the trace tool. The results were normalised for each litter by dividing by the mean of the heterozygotes. The data points represent the mean of each genotype with bars as standard deviation. Representative pictures shown in Figure 5.4 are indicated by open circles (Notch4+/+), open squares (Notch4+/d1) and open triangles (Notch4d1/d1). No significant difference (P>0.05) was observed between genotypes using ANOVA and Tukey’s post test. b. The retinal radius was measured using ImageJ software. The results were normalised for each litter by dividing by the mean of the heterozygotes. Representative pictures shown in Figure 5.4 are indicated by open circles (Notch4+/+), open squares ( Notch4+/d1) and open triangles (Notch4d1/d1). No significant difference (P>0.05) was observed between genotypes using ANOVA and Tukey’s post test. c. The number of pixels positive for PECAM1 was divided by the vascular area to calculate vascular density. The results were normalised for each litter by dividing by the mean of the heterozygotes. Representative pictures shown in Figure 5.4 are indicated by open circles (Notch4+/+), open squares (Notch4+/d1) and open triangles (Notch4d1/d1). No significant difference (P>0.05) was observed between genotypes using ANOVA and Tukey’s post test.
178 | Page
179 | Page
Figure 5.6 Retinal angiogenesis of Notch4d1 mice at 5 days post birth
Notch4+/d1 heterozygous mice were crossed and the retinas collected and stained with PECAM1 (red) and SMA (green) at day P5. Tail clippings (~2mm) were collected at the same time and used for genotyping. A 6x6 tile z-stack was captured for each retina using a 7Duo confocal microscope. The tiled stacks were flattened in the Z dimension using the average intensity function in ImageJ. Representative images of Notch4+/+ (a), Notch4+/d1 (b) and Notch4d1/d1 (c) mice from a single litter are shown.
180 | Page
181 | Page
Figure 5.7 Measurements of retinal angiogenesis at 5 days post birth
Retinas from mouse pups at P5 were collected and stained for PECAM1 and SMA expression. Tail samples were collected concurrently and used to genotype the pups. a. The vascular area was measured in ImageJ using the trace tool. The results were normalised for each litter by dividing by the mean of the heterozygotes. The data points represent the mean of each genotype with bars as standard deviation. Representative pictures shown in Figure 5.6 are indicated by open circles (Notch4+/+), open squares (Notch4+/d1) and open triangles (Notch4d1/d1). A significant (P>0.05) and highly significant (P<0.01) difference was observed in Notch4+/d1 and Notchd1/d1 mice compared to Notch+/+. No significant difference (P>0.05) was observed between Notch4 +/d1 and Notch4 d1/d1. Statistics were calculated using ANOVA and Tukey’s post test. b. The retinal radius was measured using ImageJ software. The results were normalised for each litter by dividing by the mean of the heterozygotes. Representative pictures shown in Figure 5.6 are indicated by open circles (Notch4+/+), open squares ( Notch4+/d1) and open triangles (Notch4d1/d1). No significant difference (P>0.05) was observed between genotypes using ANOVA and Tukey’s post test. c. The number of pixels positive for PECAM1 was divided by the vascular area to calculate vascular density. The results were normalised for each litter by dividing by the mean of the heterozygotes. Representative pictures shown in Figure 5.6 are indicated by open circles (Notch4+/+), open squares (Notch4+/d1) and open triangles (Notch4d1/d1). No significant difference (P>0.05) was observed between genotypes using ANOVA and Tukey’s post test.
182 | Page
183 | Page
Figure 5.8 Retinal angiogenesis in Notch4d1 mice at 7 days post birth
Notch4+/d1 heterozygous mice were crossed and the retinas collected and stained with PECAM1 (red) and SMA (green) at day P7. Tail clippings (~2mm) were collected at the same time and used for genotyping. A 7x7 tile z-stack was captured for each retina using a 7Duo confocal microscope. The tiled stacks were flattened in the Z dimension using the average intensity function in ImageJ. Representative images of Notch4+/+ (a), Notch4+/d1 (b) and Notch4d1/d1 (c) mice from a single litter are shown.
184 | Page
185 | Page
Figure 5.9 Measurements of retinal angiogenesis at 7 days post birth
Retinas from mouse pups at P7 were collected and stained for PECAM1 and SMA expression. Tail samples were collected concurrently and used to genotype the pups. a. The vascular area was measured in ImageJ using the trace tool. The results were normalised for each litter by dividing by the mean of the heterozygotes. The data points represent the mean of each genotype with bars as standard deviation. Representative pictures shown in Figure 5.8 are indicated by open circles (Notch4+/+), open squares (Notch4+/d1) and open triangles (Notch4d1/d1). No significant difference (P>0.05) was observed between genotypes using ANOVA and Tukey’s post test. b. The retinal radius was measured using ImageJ software. The results were normalised for each litter by dividing by the mean of the heterozygotes. Representative pictures shown in Figure 5.8 are indicated by open circles (Notch4+/+), open squares ( Notch4+/d1) and open triangles (Notch4d1/d1). No significant difference (P>0.05) was observed between genotypes using ANOVA and Tukey’s post test. c. The number of pixels positive for PECAM1 was divided by the vascular area to calculate vascular density. The results were normalised for each litter by dividing by the mean of the heterozygotes. Representative pictures shown in Figure 5.8 are indicated by open circles (Notch4+/+), open squares (Notch4+/d1) and open triangles (Notch4d1/d1). No significant difference (P>0.05) was observed between genotypes using ANOVA and Tukey’s post test.
186 | Page
187 | Page
5.2.5 Transcription from the Notch4d1 Allele. The results of the analysis of the Notch4d1 mice uncovered a previously undescribed phenotype in the vasculature i.e. a transient reduction in the vascular area at P5. The phenotype was significantly different in both Notch4+/d1 and Notch4d1/d1 mice compared to wild type with no significant difference observed between the heterozygous and homozygous Notch4d1 mice. A reduced vascular area phenotype in heterozygous and homozygous Notch4d1 pups suggests that the phenotype is caused by the presence of the Notch4d1 allele rather than the absence of Notch4.
Although we thoroughly characterised the genotype of the Notch4d1 mice (Section 5.2.1) both our analysis and that of the original report only characterised the Notch4d1 allele in terms of the genome and any transcripts covering the gene 3’ to the insertion of the neo cassette. To determine if the Notch4d1 allele was a true null allele a series of reverse transcription/PCR reactions were performed. Lung tissue from Notch4+/+, Notch4+/d1 and Notch4d1/d1 mice was collected from litters at P5 generated by heterozygous crosses. Tail tissue was collected concurrently and genotyped. Each genotype was represented by three individual mice. RNA was isolated and reverse transcribed using an oligodT primer. The cDNA was then analysed by PCR. Intron spanning primers were used to amplify across exons 1-3, exons 15-16, within exon 22 and the neo cassette (Figure 5.10a)
Primers within exon 22, which were routinely used for genotyping confirmed that a 313bp product was amplified from exon 22 and expressed in all Notch4+/+ (Figure 5.10a, lanes 1,4 and 7) and all Notch4+/d1 mice (Figure 5.10a, lanes 2, 5 and 8). This band was absent in all three Notch4d1/d1 mice (Figure 5.10a, lanes 3, 6 and 9). The lack of transcript was unsurprising given that exon 22 has been removed in the Notch4d1 allele. The control containing no DNA (Figure 5.10a, lane 10) gave no band. A set of primers designed to amplify the introduced neo gene amplified a 518bp product from all Notch4+/d1 mice (Figure 5.10a, lanes 2, 5 and 8) and all three Notch4d1/d1 mice (Figure 5.10a, lanes 3, 6 and 9). Again the lack of any product in the water control indicated there was no cross contamination. These results also confirmed the genotyping as the neo cassette is absent from the wild type mice (Figure 5.10a, lanes 1, 4 and 7). Two additional PCR reactions were performed that used primers targeted to transcripts upstream of the neo insertion site. A primer pair that amplified a 199bp product spanning exons 1-3 was used to confirm the presence of the 5’ end of the Notch4 transcript in all Notch4+/+ mice (Figure 5.10a, lanes 1, 4 and 7) and all Notch4+/d1 mice (Figure 5.10a, lanes 2, 5 and 8). However, this primer also amplified a specific 188 | Page
product of the correct size in all three Notch4d1/d1 mice (Figure 5.10a, lanes 3, 6 and 9). The primers flanked exon boundaries, so the amplification product observed could not be derived from contaminating genomic DNA. Moreover, the RNA was reverse transcribed using an oligodT primer indicating the presence of a polyadenylated 3’ tail. The water control was also negative indicating a lack of cross contamination. Another set of primers upstream of the neo insertion site that amplified a 160bp product derived from sequences spanning exons 15 and 16 yielded identical results (Figure 5.10a, lanes 3, 6 and 9). In all mice, regardless of genotype, a specific product was amplified. Again the primers were designed to span exons to rule out genomic DNA contamination.
Reverse transcription/PCR (RT-PCR) is an extremely sensitive technique that can potentially pick up even single transcripts. The transcripts detected in the null mice may have been extremely rare and/or unstable transcripts arising from the modified allele. In order to quantify this transcript the cDNA described above was analysed by quantitative PCR. Primers spanning exons 18 and 19 were used to quantify expression. The relative expression normalised to beta-actin (Actb) detected for each mouse is shown in Figure 5.10b. These results confirmed our previous analysis that the Notch4d1 mice were producing polyadenylated transcripts derived from the modified allele. There was no significant difference in expression in the mice with each genotype. The transcript produced in the Notch4d1 mice was being made at equivalent levels to the wild type allele and is thus potentially of physiological significance. The following experiments were aimed at characterising this transcript further.
189 | Page
Figure 5.10 RT-PCR of Notch4d1 mice.
Lung tissue from three mice of each genotype, Notch4+/+, Notch4+/d1 and Notch4d1/d1, was collected and reverse transcribed using an oligodT primer. a. PCR primers within exon 22, within neo, spanning exon 1 and 3 and spanning exon 15 and 16 were used to amplify cDNA from Notch4+/+ (lane 1, 4 and 7), Notch4+/d1 (lane 2,5 and 8) and Notch4d1/d1 (lane 3, 6 and 9). A water control with no cDNA is shown in lane 10. b. Real time PCR of cDNA from three Notch4+/+, three Notch4+/d1 and three Notch4d1/d1 mice amplified with primers spanning exon 18 and 19 normalised to beta-actin (Actb). There is no significant difference (P=0.92) between groups (ANOVA).
190 | Page
5.2.6 Isolation and analysis of the Notch4d1 transcript Experiments thus far detected a transcript from the Notch4d1 allele that was as abundant as the native Notch4 transcript. The Notch4d1 transcript was spliced and polyadenylated and could be detected in all mice tested. The transcript we detected contained sequences 5’ of the insertion of the neo cassette. These included exons 1-3, 15-16 and 18-19 but the nature of the 3’ end and polyadenylation site was unknown. A 3’RACE protocol was used to identify and characterise the 3’ end of the Notch4d1 transcript (Scotto-Lavino et al., 2006). A PCR product was amplified using a 5’ primer annealing within exon 21 of Notch4 and an anchored oligodT primer. A second set of primers annealing 125 nucleotides downstream of the 5’ primer and within the first round 3’ primer was used to amplify a product which was cloned and sequenced. The product was aligned to the genomic sequence and was found to be a processed mRNA transcript. The 5’ end was derived from exon 21 of the Notch4 locus with the correct 3’ exon end spliced to a cryptic 3’ intron acceptor site within the neo cassette (Figure 5.11a, red and blue). A second undescribed intron within the neo cassette was also removed. The 3’ end of the transcript was polyadenylated 18 nucleotides 3’ of the SV40 polyadenylation signal of the neo gene. The translation frame of the Notch4 locus, extended into the novel exon, codes for an additional 10 amino acids before encountering a stop codon. Transcripts that contain an intron greater than 50-55 bases downstream of the stop codon can be substrates for nonsense mediated decay (Maquat, 2004). This stop codon is 51 bases upstream of the 3’ exon/exon boundary formed by the second novel exon. The Notch4d1 transcript was detected in equal abundance to the Notch4 transcript (Section 5.2.5) confirming that the transcript was stable.
The 3’ RACE protocol used a nested PCR technique which was extremely sensitive and could result in the cloning of an extremely rare transcript specific to the individual tested. To confirm that this transcript was generally present in mice carrying the Notch4d1 allele the region corresponding to the novel splice products was amplified using primers flanking this region. A specific product was isolated from Notch4+/d1 (Figure 5.11b, lanes 2, 5 and 8) and Notch4d1/d1 (Figure 5.10b, lanes 3, 6 and 9) mice. No band was observed in Notch4+/+ mice (Figure 5.11b, lanes 1, 4 and 7). The PCR product produced was cloned and sequenced. The PCR products derived from all mice carrying the Notch4d1 allele (n=6) were found to be identical to the 3’ RACE transcript we isolated.
191 | Page
Figure 5.11 The transcript produced from the Notch4d1 allele contains two novel exons.
Lung tissue from Notch4d1/d1 mice was isolated and reverse transcribed. a. A 3’RACE protocol was employed to amplify the 3’ end of the Notch4d1 transcript. The first round of amplification primers are shown underlined and the second round primers are double underlined. The 3’ end of exon 22 of Notch4 is shown in red. The 1 st novel exon (blue) codes for an additional 10 amino acids before a stop codon (bold). The polyA consensus sequence derived from the SV40 PolyA signal of the neo cassette is shown as bold underlined. Primers used in (b) are overlined in black. b. Primers hybridising to exon 22 of Notch4 and within the neo cassette (overlined in black in a.) were used to amplify cDNA from Notch4+/+ (lanes 1, 4 and 7), Notch4+/d1 (lanes 2, 5 and 8) and Notch4d1/d1 (lanes 3, 6 and 9). A water control is shown in lane 10.
192 | Page
193 | Page
5.2.7 Northern analysis of Notch4d1 mice A transcript was detected from the Notch4d1 allele and was identical at the sequence level in all Notch4d1 mice tested (Section 5.2.6). The PCR-based RACE technique employed to detect this transcript is extremely sensitive, and as such has the potential to detect very rare transcripts. Real time PCR detected transcripts containing exons 18 and 19 present in equal quantities in Notch4+/+ and Notch4d1/d1 mice (Section 5.2.5). To establish that a single transcript was produced from the Notch4d1 allele a northern blot was performed.
Heterozygous crosses of Notch4+/d1 mice were performed and the lung tissue collected from pups at P5. Tail clips were collected concurrently and used to genotype the mice. RNA extracted from lung tissue from Notch4+/+ (Figure 5.12, lanes 1, 4 and 7), Notch4+/d1 (Figure 5.12, lanes 2, 5, 8) and Notch4d1/d1 (Figure 5.12, lanes 3, 6 and 9) mice were analysed by northern blot. The probe hybridised to sequences upstream of the neo cassette. Prior to hybridisation the blot was stained with ethidium bromide and the position of the 18S (1.9kb) and 28S (4.7kb) are indicated on the left (Figure 5.12). A single band was detected in Notch4d1/d1 RNA indicating that the Notch4d1 allele produced one major transcript (Figure 5.12). The Notch4d1 transcript migrated at the expected size of approximately 4.9kb (Section 5.2.6), slightly faster than the wild type transcript detected in Notch4+/+ mice. Bands corresponding to the wild type and Notch4d1 transcripts were difficult to resolve in Notch4+/d1 mice and smeared across the expected size of both transcripts (Figure 5.12, lanes 2, 5 and 8). Higher molecular weight background staining can be observed in all lanes.
194 | Page
Figure 5.12 Northern analysis of Notch4d1 mice
RNA isolated from the lungs of three mice of each genotype, Notch4+/+ (lanes 1, 4 and 7), Notch4+/d1 (lanes 2, 5 and 8) and Notch4d1/d1 (lanes 3, 6 and 9) was northern blotted using a probe hybridising to a region of Notch4 upstream of the insertion site of neo. The position of the Notch4, Notch4d1, and the 28s and 18s rRNA transcripts identified on the ethidium bromide stained gel are indicated on the left.
195 | Page
5.2.8 Analysis by in situ hybridisation of Notch4d1 mice. Northern blot analysis revealed that the Notch4d1 allele produces a single major transcript (Section 5.2.7). The Notch4d1 transcript is a unique polyadenylated and processed transcript and present in equal quantities to the native transcript. All mice carrying the Notch4d1 allele produce an identical transcript. The preceding results used lung tissue from neonatal mice as a source of RNA. The lung is a rich source of Notch4 mRNA due to its high vascular content. To exclude the possibility that the Notch4d1 transcript is tissue specific an RNA in situ hybridisation assay was employed.
Initial experiments determined the peak expression level detected for Notch4 in wild type embryos by whole mount RNA in situ hybridisation was at embryonic day 10.5 (E10.5). Earlier embryos (E9.5) had low expression levels and later embryos (E11.5) had reduced staining due to reduced penetrance of the probe into the embryo. Wild type E10.5 C57BL/6J embryos were hybridised with a Notch4 probe, pYXNotch4 probe (Figure 5.13a). The pYXNotch4 probe used to determine the expression domain of Notch4 overlapped both the Notch4 and Notch4d1 transcripts. The Notch4 transcript was detected in blood vessels throughout the embryo. The most prominent staining in the circulatory system was in the intersomitic vessels (ISV) and the branchial arteries (BA). Strong staining was also observed in the vessels supplying the lens pit (LP). In addition to the vasculature, there was strong expression in the tail bud (TB) extending into the presomitic mesoderm.
Next the expression of Notch4d1 transcript was examined using two probes designed to hybridise to sequences downstream of the neo cassette (3’ probe) or upstream of the neo cassette (5’ probe). The 5’ probe hybridises to sequences common to both Notch4+ and Notch4d1 transcripts. The 3’ probe hybridises to sequences in the wild type transcript that are absent in the Notch4d1 transcript and acts as a negative control. To determine the expression domain of the Notch4d1 transcript, embryos from heterozygous crosses of Notch4+/d1 mice were collected at E10.5. Samples of the yolk sac were collected concurrently and used to genotype the embryos. Three Notch4+/+ and three Notch4d1/d1 embryos were stained with each probe. Representative examples of littermates are shown in Figure 5.13b.
The Notch4d1 transcript was expressed in an identical expression domain to wild type Notch4 and at a similar intensity (Figure 5.13b, Notch4+/+ 5’ probe and Notch4d1/d1 5’ probe). In common with wild type Notch4, Notch4d1 was detected in the intersomitic vessels (ISV), branchial arteries (BA) and the vessels of the lens pit (LP). The Notch4d1 transcript was also expressed in the tail bud (TB) at similar levels to wild type Notch4. No transcripts were
196 | Page
detected in Notch4d1/d1 embryos using the 3’ probe (Figure 5.13b, Notch4d1/d1 3’ probe). The 3’ probe detected Notch4 transcripts only in the wild type littermate controls, demonstrating the specificity of this probe (Figure 5.13b, Notch4+/+ 3’ probe).
5.2.9 The putative NOTCH4d1 protein The Notch4d1 allele produces a stable polyadenylated transcript in quantities equal to and in the same expression domain as Notch4. The putative protein produced would contain all of the EGF-like repeats and the LNR-A and B domains of NOTCH4 plus 10 additional novel amino acids. We were unable to detect endogenous NOTCH4 using commercially available antibodies to the extracellular domain. Two peptides corresponding to extracellular region of NOTCH4 were designed in our laboratory by Gavin Chapman and used to produce antibodies. The antibodies produced lacked the necessary specificity and sensitivity to detect endogenous levels of NOTCH4. In order to analyse the function of the NOTCH4d1 protein, a construct was designed that contained all of the Notch4 sequences contained in Notch4d1 with a C- terminal HA tag. We have shown in Chapter 4 that NOTCH4 acts as an inhibitor of canonical NOTCH1 signalling. This inhibition appears to be due to an interaction with NOTCH1 that alters its subcellular localisation. We hypothesised that this function may be retained in the NOTCH4d1 protein.
To test the ability of NOTCH4d1 to act as an inhibitor of NOTCH1 activation a co-culture assay was performed. NIH3T3 cells were transfected with Notch4d1HA, Notch4HA, Notch1HA, or were co- transfected with Notch1HA and Notch4d1HA, or Notch1HA and Notch4HA plus the Notch reporter pGL46xTP1 and the transfection control pCMXren. Transfected cells were co-cultured with either Dll4 expressing cells or control cells and the relative luciferase activity measured. As previously seen, Notch1HA transfected cells robustly induced reporter activity 30-fold over control cell co- cultures (Figure 5.14, column 8). Notch4HA was not activated under the same conditions (Figure 5.14, column 6). The Notch4d1HA construct lacks the NOTCH4ICD and unsurprisingly was unable to activate the reporter above control (Figure 5.14, column 7). Addition of Notch4HA to Notch1HA transfected cells significantly (P<0.01) inhibited Notch1 activation down from 30- fold to 10-fold over control (Figure 5.14, column 9). Co- transfection of Notch4d1HA with Notch1HA also significantly (P<0.01) inhibited NOTCH1 activation in response to ligand, down from 30 fold to 6 fold over control (Figure 5.14, column 10). There was no significant difference in the ability of NOTCH4 and NOTCH4d1 to inhibit NOTCH1 activation in response to ligand (Figure 5.14, columns 9 and 10).
197 | Page
Figure 5.13 Expression domain of the Notch4d1 transcript a. In situ hybridisation using a probe to Notch4 (pYXNotch4) in E10.5 wild type C57BL/6J embryos. Regions of strong expression are indicated with arrows. BA, branchial arteries; ISV intersomitic vessels; LP, lens pit; TB, tail bud. b. Representative examples of a single litter of embryos derived from Notch4+/d1 crosses at E10.5. The embryos were stained with two probes. For Notch4 expression a probe hybridising to sequences downstream of the neo insertion site (Notch4+/+ 3’ probe and Notch4d1/d1 3’ probe) was used. Both Notch4 and Notch4d1 transcripts were detected using a probe to sequences upstream of the neo insertion site (Notch4+/+ 5’ probe and Notch4+/d1 5’ probe).
198 | Page
199 | Page
Figure 5.14 The NOTCH4d1 protein inhibits NOTCH1 activation
NIH3T3 cells were transfected with Notch4HA, Notch4d1HA, Notch1HA, or co-transfected with Notch1 and Notch4HA, or Notch1 and Notch4d1HA, or vector control plus the Notch responsive reporter (pGL46xTP1) and a Renilla luciferase transfection efficiency control (pCMXren). The transfected cells were co- cultured overnight with either a NIH3T3 control cell line or a DLL4 expressing line. The data points represent fold over vector control of the relative luciferase units normalised for transfection efficiency by dividing by the Renilla luciferase activity. This figure represents the mean of three independent experiments with bars as standard deviation. Notch1HA alone co-cultured with ligand expressing cells was compared to Notch1HA and Notch4HA co- expression and Notch1HA and Notch4d1HA co-expression using ANOVA in Prism software (**P<0.01).
200 | Page
5.2.10 NOTCH4d1 induces differentiation in C2C12 cells NOTCH4d1 can inhibit Notch signalling in our co-culture model in a similar manner to NOTCH4. To test that NOTCH4d1 retains the ability of wild type NOTCH4 to inhibit Notch signalling in a biological setting, the C2C12 myogenesis model (Section 4.2.4) was used. The following experiment was performed by Kavitha Iyer in our laboratory.
C2C12 cells were transfected with pCAGNotch4d1HA and selected for puromycin resistance. Selected clones were isolated and grown and NOTCH4d1 expression was confirmed by immunostaining with an HA antibody. Corresponding C2C12 lines expressing Notch4HA (C2C12Notch4HA) or the parental vector (C2C12iPuro) were used as controls (see Section 4.2.4). C2C12Notch4d1HA, C2C12Notch4HA and C2C12iPuro lines were seeded at the same density and grown to confluence over three days. As described previously the C2C12iPuro cells maintained their undifferentiated morphology with single nuclei and compact cell area (Figure 5.15a). The cells were stained for myosin heavy chain (MHC) expression and the nuclei counterstained with TO-PRO-3. The C2C12iPuro cells had a low level of MHC expression. In contrast to these control cells, C2C12Notch4d1HA cells displayed a marked change in morphology. Large polynucleated myotubes expressing MHC were easily identifiable (Figure 5.15a). A similar differentiated morphology was observed in the C2C12Notch4HA cells.
C2C12Notch4d1HA, C2C12Notch4HA and C2C12iPuro cells were seeded at the same density and western blots for MHC expression were performed at 24, 48 and 72 hours. There was a clear induction of MHC expression beginning at 48 hours (Figure 5.15b, lanes 4-6) in both the C2C12Notch4d1HA and C2C12Notch4HA cultures compared to control where MHC expression was below the level of detection. This induction increased in both C2C12Notch4d1HA and C2C12Notch4HA cells at 72 hours while still below the level of detection in control (Figure 5.15b, lanes 7- 9). Thus NOTCH4d1, like NOTCH4, is able to inhibit a Notch dependent biological process.
201 | Page
Figure 5.15 NOTCH4d1 induces differentiation in C2C12 cells a. C2C12 cell lines were established by transfecting with Notch4d1HA, Notch4HA or control (iPuro). The cell lines were seeded at the same density and allowed to come to confluence over three days. C2C12Notch4d1HA, C2C12Notch4HA and control (C2C12iPuro) cells were stained for MHC expression (green) and the nuclei counterstained with TO-PRO-3 (red). b. C2C12Notch4d1HA (lanes 3, 6 and 9), C2C12Notch4HA (lanes 2, 5 and 8) or C2C12iPuro cell lines (lanes 1, 4 and 7) were seeded at the same density and grown for 24 (lanes 1-3), 48 (lanes 4-6) or 72 hours (lanes 7-9). The cultures were western blotted for MHC and beta-tubulin expression.
202 | Page
203 | Page
5.2.11 NOTCH4d1 and NOTCH4 interact with NOTCH1 Co-expression of either NOTCH4d1 or NOTCH4 with NOTCH1, inhibited NOTCH1 activation (Section 5.2.9). This inhibition was confirmed in a C2C12 myogenesis assay (Section 5.2.10). Thus the protein encoded by the Notch4d1 allele retains the novel inhibitory function of NOTCH4 identified in Chapter 4. Co- expression of NOTCH4 and NOTCH1 led to a change in the subcellular distribution of NOTCH1 (Section 4.2.6). We conjectured that this may be due to an interaction between NOTCH4 and NOTCH1 and that the mechanism of NOTCH4d1 inhibition of Notch activation was shared with NOTCH4. Co- immunoprecipitation experiments were performed by Joelene Major in our laboratory to investigate if NOTCH4 and NOTCH4d1 interacted with NOTCH1 in cultured cells.
C2C12 cells and C2C12 cells stably expressing Notch1 were transfected with either Notch4HA, Notch4d1HA or vector control plasmid (Figure 5.16). Protein complexes were immunoprecipitated with either mouse anti-HA antibody (Figure 5.16, lanes 1, 2, 5, 6 and 7) or non-specific mouse IgG antibody (Figure 5.16, lanes 3 and 4). The immunoprecipitated complexes were western blotted and detected with an anti-NOTCH1 antibody. Immunoprecipitation of HA proteins from cells co-expressing NOTCH4d1 and NOTCH1 captured complexes containing full length NOTCH1 (NOTCH1Fl) (Figure 5.16, lane 7). Co-expression of NOTCH4HA and NOTCH1 also resulted in the capture of NOTCH1Fl when immunoprecipitated with HA (Figure 5.16, lane 6). NOTCH1Fl was not detected in control IgG immunoprecipitates when either NOTCH4d1HA and NOTCH1 (Figure 5.16, lane 4) or NOTCH4HA and NOTCH1 (Figure 5.16, lane 3) were co-expressed. Immunoprecipitation with HA in NOTCH1 expressing cells (Figure 5.16, lane 1), NOTCH4d1HA (Figure 5.16, lane 2) or NOTCH4HA (Figure 5.16, lane 5) alone did not capture any NOTCH1Fl. In all NOTCH1 expressing cells (Figure 5.16, lanes 1, 3, 4, 6 and 7) a non-specific band was detected at approximately 100kDa. The non- specific band was present in equal quantities in the negative controls. NOTCH4 and NOTCH4d1 specifically interacted with the unprocessed form of NOTCH1.
204 | Page
Figure 5.16 NOTCH1 interacts with NOTCH4d1 and NOTCH4
C2C12 cells (lanes 2 and 5) and C2C12 cells stably expressing NOTCH1 (lanes 1, 3, 4, 6 and 7) were transfected with Notch4HA (lanes 3, 5 and 6) or Notch4d1HA (lanes 2, 4 and 7). Protein complexes were immunoprecipitated with either mouse anti-HA (lanes 1, 2, 5, 6 and 7) or non-specific mouse IgG antibody (lanes 3 and 4). The immunoprecipitated complexes were western blotted with an anti-NOTCH1 antibody directed against the intracellular domain (ab27526, Abcam). The positions of full length NOTCH1 (NOTCH1Fl) are indicated on the right. The positions of the molecular makers are indicated on the left.
205 | Page
5.3 Discussion The study of NOTCH4 in vivo has heavily relied on the use of activated forms of the protein which are independent of ligand. These studies have revealed that the NOTCH4ICD can have dramatic influences on the development of the vasculature (Carlson et al., 2005; Uyttendaele et al., 2001). Over expression of the Notch4ICD produces phenotypes that are similar to Notch1 knockout phenotypes (Krebs et al., 2000; Uyttendaele et al., 2001). The similar phenotypes seen in both loss and gain-of- function mutants, demonstrates that the correct balance of Notch signalling is critical to the growth and patterning of the vasculature. The phenotypes displayed by Notch pathway mutants indicate that Notch signalling has roles in angiogenesis and the maintenance of arterial/venous identity. However, the Notch4d1 mouse, a putative loss-of-function mutant, displays no overt vascular phenotype.
In addition to supplying a large variety of organs and tissues, each with unique requirements, the vasculature must be able to repair itself after a variety of injuries. The correct patterning of the vasculature thus has many in built redundancies, needed to withstand a variety of insults. Notch4 is a mammalian specific gene. There are no Notch4 homologues in the chicken, crocodile or fish (Theodosiou et al., 2009). These species all build and maintain a functioning vasculature without the help of Notch4. The lack of phenotype in the Notch4d1 mouse suggests that the role of Notch4 in the vascular endothelium is likely to be an auxiliary one. However, the vasculature is extremely sensitive to the dosage of Notch signalling in development as evidenced by the haploinsufficient lethal phenotype of both Dll4 and Vegfa (Carmeliet et al., 1996; Duarte et al., 2004; Ferrara et al., 1996; Gale et al., 2004; Krebs et al., 2004). We reasoned that although loss of canonical NOTCH4 signalling can be compensated for by other mechanisms, e.g. NOTCH1, a more thorough analysis of angiogenesis would have the potential to uncover a non-redundant role of NOTCH4. We therefore utilised the mouse retina as a sensitive and experimentally manipulatable model of angiogenesis.
5.3.1 Notch4d1 mice have a transient delay in retinal angiogenesis The analysis of the retinal angiogenesis in the Notch4d1 mouse revealed a transient delay in the area of vascular coverage at P5 (Section 5.2.4). There was no significant difference between Notch4 and Notch4d1 mice at either P3 or P7. The increase in vascular area was not due to differences in the size or growth of the retina itself (Section 5.2.4) nor a global growth delay (Section 5.2.3). There are a number of reasons to believe that P5 is the
206 | Page
most sensitive time point for an analysis of the affects of Notch signalling.
In the initial stages of angiogenic growth of the retinal vasculature there is a rapid migration and expansion of endothelial cells arising from the vicinity of the optic nerve. The rapidity of the migration/expansion leads to a large degree of variation in the extent of vasculature between individual mice. The wild type and Notch4d1 mice both displayed over 2.5-fold difference in endothelial coverage and vascular density (Figure 5.5) between individuals at this time point. In contrast there was approximately a 30 and 40% difference within genotypes in the Notch4 and Notch4d1 mice at P5 (Figure 5.7). Any differences at P3 between genotypes may have been obscured by the variation within genotypes. Thus P3 was not an informative stage. The variation encountered at P7 may be partly artifactual. There is a distinct outlier in the wild type group greater than 2 standard deviations below the mean. The outlier could not be excluded on experimental grounds and so was included in all analyses. If this outlier was not representative we may have found a similar result to that obtained at P5. However, at P7 the vasculature is at or close to the edge of the retina. The data is thus bounded on one side which further complicates analysis. The retinas at P7 were otherwise normal and no phenotype additional to that observed at P5 could be inferred even if the outlier was removed. Thus P5 was the most informative time point and a significant reduction in the vascular coverage was observed at this time point.
There are a number of explanations for the phenotype being most apparent at P5. At this stage not only is there the tip cell selection process occurring but there are also other active processes occurring. The arterial and venous vessels have diverged and the arterial vessels are being actively stabilised by smooth muscle cells. The processes of pruning around areas of high oxygen are also occurring and the potential for a delay at P5 has been well demonstrated by deletion of Notch-regulated Ankyrin repeat protein (Nrarp). Nrarp is a direct target of RBPJ dependent Notch signalling and is expressed in stalk cells in the developing retinal vasculature (Krebs et al., 2001; Pirot et al., 2004; Phng et al., 2009). NRARP functions, in part, to promote the degradation of Notch and thus attenuate signalling. Nrarp-/- mice have a transiently reduced vascular coverage of the retina at P5. The retinas of Nrarp-/- mice have a decreased vessel density characteristic of excessive Notch signalling (Phng et al., 2009). NRARP also plays a Notch independent role in vessel stabilisation. Nrarp-/- mice have an increase in vessel regression contributing to the growth delay seen at P5 which is independent of Notch (Phng et al., 2009). 207 | Page
In tip cell selection the process has been described as a tug of war between Notch and Dll4 (Phng and Gerhardt, 2009). The tip and stalk cells are initially functionally equivalent. Small stochastic differences are then reinforced. The adoption of tip and stalk cell fate is not dependent on the absolute levels of either Notch receptor or ligand but rather the ratio of the two. NOTCH4 could perhaps fine tune this process and a lack of NOTCH4ICD may prolong the struggle leading to the growth delay observed at P5. We did not observe any increase in vascular density so although the decision between stalk and tip, although potentially delayed, is reaching the correct stalk/tip ratio.
Vessels are stabilised in part by Notch signalling. The expression of Jag1 in endothelial cells is driven by Notch signalling (High et al., 2008). There is also the potential for the lack of canonical NOTCH4 signalling acting at this point as well. However, the vessels do form correctly and recruit smooth muscle. Vessel maturation is regulated at many levels and compensatory mechanisms are available in addition to Notch signalling. The arterialisation of vessels is influenced by the induction of blood flow which may compensate for mild defects in Notch signalling.
There are a number of ways in which the Notch4d1 phenotype could be explored further. Although the retina provides an excellent model for studying angiogenesis it is not representative of all vascular beds. Endothelial cells from different tissues and organs display different phenotypes. An investigation of angiogenesis in a variety of tissues may highlight more pronounced differences. However, the lack of any overt phenotype indicates that, despite the slight growth delay we observed, a functional circulatory system does develop in all organs and tissues.
The phenotype observed in the Notch4d1 mice was mild and the mice recovered. Further work using this model is complicated by our observation that the Notch4d1 allele is not a true null allele. Any interpretations of angiogenesis in this mouse are complicated by this allele which has the potential to not only preserve some Notch4 function but may also introduce novel functions.
5.3.2 Notch4d1 transcript The existence of a transcript arising from the Notch4d1 allele was an unexpected result. The transcript was present in all mice carrying the Notch4d1 allele, was produced and as stable as the wild type transcript and was expressed in the same expression pattern as the wild type Notch4 transcript. The Notch4d1 allele was designed to be a null allele. The canonical function of the Notch receptors is mediated by the NotchICD. The extracellular domain regulates the production of the NotchICD. cis-Inhibition 208 | Page
represents an additional role for the extracellular domain and will be discussed in Section 5.3.2.5. Thus by removing the ICD of Notch4 from the genome, the intention of Krebs et al., 2000 was to create a Notch4 null allele. In agreement with previous reports we confirmed that the allele is indeed null for the NOTCH4ICD and thus canonical NOTCH4 signalling. However, we have evidence that the novel inhibitory role of NOTCH4 revealed in Chapter 4 may be preserved in these mice. Below is a discussion on the identification of this transcript and its possible functional consequences.
5.3.2.1 Splicing artefacts in transgenic mice There are many examples of alleles originally targeted to generate a null that actually produce a product from the disrupted locus. There are varied mechanisms by which this can occur. Knockouts that remove the start codon can initiate translation from a previously internal methionine, generating an N-terminal truncated protein (e.g. the signal transducer and activator of transcription 1 ( Stat1) mouse (Meraz et al., 1996)). The construction of the Notch4d1 mouse took this into account by removing the methionine residue utilised as the start codon in the Int-3 transcript (Uyttendaele et al., 1996). Targeting strategies that generate C- terminal truncations (e.g. cannabinoid receptor 2 (macrophage) ( Cnr2 also known asCB2 (Buckley et al., 2000)) or insertion events resembling recombinations can leave an allele intact (e.g. v-myc myelocytomatosis viral related oncogene, neuroblastoma derived (avian) Nmyc (Moens et al., 1992)). Aberrant splicing can occur around the insertion cassette (e.g. TGFalpha (Luetteke et al., 1993) and Vegfa (Carmeliet et al., 1996)). The potential for splicing around the insertion cassette was taken into account for the Notch4d1 allele. Exons 21 and 24, either side of the insertion cassette are out of frame with respect to each other. In agreement with the initial report of the Notch4d1 mouse (Krebs et al., 2000) we could not detect any transcripts containing sequences 3’ of the insertions cassette.
There are also examples that are similar to the splicing event we observed i.e. splicing into the insertion cassette e.g. Vegfa (Carmeliet et al., 1996) and estrogen receptor 1 (alpha) (Esr1) (Kos et al., 2002). The structure of the Notch4d1 transcript, produced by this form of splicing, is discussed below.
5.3.2.2 Structure of the Notch4d1 transcript The transcript from the Notch4d1 allele is a single species that comprises the first 21 exons of the Notch4 gene with a novel 3’ end generated from the insertion cassette (Figure 5.10). The novel 3’ end is the result of two unexpected splicing events. In mammalian cells the signals regulating splicing are not completely understood. The strict consensus sequences required in yeast are 209 | Page
divergent in mammals which makes their prediction on purely primary sequence difficult (Gao et al., 2008). The first intron removed makes use of the native 5’ end of intron 22. In the current model of splicing the 5’ end of the intron is the first structure recognised and bound by the spliceosome and this occurs co- transcriptionally. The polyA site is also recognised by the splicing machinery (Will and Luhrmann, 2011). Thus the transcript produced from the Notch4d1 allele could be generated if the splicing machinery recognises, binds and splices legitimate sites both within the 5’ end of Notch4 and the 3’ polyA end of the insertion cassette. Thus the transcript has actively recruited the splicing machinery which could “search” for less favourable aberrant sites to generate a functional mature transcript.
The Notch4d1 transcript contains an intron 3’ of a novel stop codon introduced by the first novel exon. The presence of an intron 3’ of a stop codon can mark transcripts for nonsense mediated decay. Nonsense mediated decay relies on the presence of the exon junction complex (EJC) (Maquat, 2004). The EJC binds 20-24 nucleotides 5’ of the removed intron and remains bound to the transcript during export from the nucleus and until the initiation of transcription (Maquat, 2004). As the transcript is translated the ribosomal proteins displace the EJC until they reach the stop codon. The ribosome is a large multiprotein complex and will displace EJCs that are less than 50-55 nucleotides 3’ of the stop codon. EJCs that are not removed from the transcript recruit factors involved in nonsense mediated decay. The stop codon of the Notch4d1 transcript falls with this critical zone at 51 nucleotides 3’ of the stop codon. The detection of comparable quantities of Notch4d1 transcript, compared to Notch4, by quantitative PCR (Figure 5.10b) and northern blot analysis (Figure 5.12) demonstrates that the transcript is not unstable and thus not a substrate for nonsense mediated decay.
5.3.2.3 Possible functions of the NOTCH4d1 protein We have not been able to detect endogenous NOTCH4 with antibodies to the extracellular domain. We have therefore been unable to directly demonstrate the existence of a protein produced from the Notch4d1 allele. However, analysis of the Notch4d1 transcript reveals no indication that the transcript would not be translated.
The putative protein would contain the EGF-like repeats and the LNR-A and B domains of Notch4 and an additional, novel, 10 amino acids. The LNR-B domain C-terminal boundary is two amino acids N-terminal of the novel sequences. The NRR of Notch4 is divergent from that of other Notch receptors and other groups have had difficulty with its expression (Falk et al., 2012). The LNR-B is
210 | Page
less divergent that the LNR-A domain and may have a more conserved structure. It contains a complete set of cysteine residues in a conserved spacing. LNR-B domains of NOTCH1 and NOTCH2 are true domains and can independently fold (Gordon et al., 2008b; Gordon et al., 2007). Thus if the LNR-B of NOTCH4 does fold in a conserved way compared to NOTCH1 and NOTCH2, the C-terminal of the NOTCH4d1 could form a correctly folded structure thus lending stability to the protein. The potential structure of the novel additional 10 amino acids is unknown and was not investigated.
5.3.2.4 NOTCH4d1 as a secreted protein The NOTCH4d1 protein does not contain a transmembrane domain. This would suggest that it could be secreted. However, although Notch receptors are generally referred to as plasma membrane receptors, the majority of Notch is intracellular (Aster et al., 1994; Fehon et al., 1991; Zagouras et al., 1995). Surface expression is strictly regulated. Thus the question of whether and to what extent the NOTCH4d1 protein would be secreted remains open. If the NOTCH4d1 protein is secreted this raises the additional possibility that not only does the NOTCH4d1 protein potentially retain some NOTCH4 function (see below) it may also have novel functions. The extracellular domain of Notch proteins have been employed as soluble inhibitors of Notch signalling (Funahashi et al., 2008). The soluble receptor binds ligand and thus acts as a decoy receptor. The NOTCH4d1 protein could act as an extracellular inhibitor of canonical Notch signalling. We have not yet attempted to detect secreted NOTCH4d1 in serum. If NOTCH4d1 is secreted into the serum, we could determine the ability of soluble exogenous NOTCH4d1 protein to inhibit Notch signalling. Isolation of cells from Notch4d1 mice that could be expanded in vitro that are known to express NOTCH4, e.g. endothelial cells, could be used to condition media. Such conditioned media could be used to assay for Notch signalling inhibition as well as provide a more manipulatable system with which to detect NOTCH4d1.
5.3.2.5 Notch4 functions left intact in Notch4d1 The ability of the NOTCH4d1 protein to inhibit Notch signalling was assessed in co-culture assays, the myotube differentiation assay and by co-immunoprecipitation. NOTCH4d1 inhibited NOTCH1 dependent signalling in a co-culture assay. NOTCH4d1 expression induced myotube formation when expressed in C2C12 cells. The NOTCH4d1 protein interacts with unprocessed NOTCH1 as does NOTCH4. These functions are distinct from the actions of the NOTCH4ICD. We demonstrated that in our assays the NOTCH4ICD was not the source of the inhibition. NOTCH4ICD does not inhibit signalling in a co-culture assay. Other groups have demonstrated that the NOTCH4ICD, in contrast to NOTCH4
211 | Page
and NOTCH4d1, inhibited myotube differentiation in C2C12 cells. Thus although the Notch4d1 mouse is null for the Notch4ICD, the inhibitory role we identified is intact.
Both NOTCH4d1 and NOTCH4 interact with the unprocessed full length NOTCH1 protein. An early model proposed for Notch receptor activation postulated that receptor dimerisation may play a role (Greenwald and Seydoux, 1990; Kidd et al., 1989). It was thought that gamma-secretase cleavage could be inhibited by dimers of trans-membrane domains (Struhl and Adachi, 2000). The model proposed that ligand binding disassociated the Notch receptor dimer allowing gamma-secretase cleavage. This model identified cysteine residues in the heterodimerisation domain (HDD) as forming bonds in the dimer. It is now absolutely clear that this is not the case (Gordon et al., 2007). However, both the NOTCH1 and NOTCH2 receptors can be co-immunoprecipitated with NOTCH1 (Sakamoto et al., 2002a; Vooijs et al., 2004). Dimerisation of Drosophila NOTCH or mammalian NOTCH1 EGF- like repeats has been detected by electron microscopy (Kelly et al., 2010). Importantly the receptor that is captured is not S1 processed, the form of NOTCH1 found on the cell surface. Measurement of the amount of dimer present on the cell surface was less than 2%. Thus the vast majority, if not all, the dimers formed by Notch receptors are intracellular. The dimerisation of Notch receptors is dependent on the EGF-like repeats and independent of calcium. Since full length, not S1 cleaved, receptor is largely localised to the ER, the point at which dimerisation of the receptors is thought to occur is early in the secretory pathway (Vooijs et al., 2004) which raises the speculative theory that dimerisation and/or heterodimerisation may play a role in endoplasmic reticulum (ER) export as has been found for other proteins (Ma and Jan, 2002; Salahpour et al., 2004). The role of receptor dimerisation remains poorly described but it seems that dimerisation may be involved in receptor maturation and/or ER export and not in ligand activation. These results provide a model for the inhibitory effect of NOTCH4 on NOTCH1 activation. A direct interaction between NOTCH4 and NOTCH1 results in NOTCH1 adopting a NOTCH4 like subcellular distribution (Section 4.2.9). The interruption to normal NOTCH1 receptor maturation prevents ligand dependent activation. Importantly, NOTCH4d1 retains the ability to inhibit Notch signalling and to interact with the full length NOTCH1 protein. These results provide evidence that the Notch4d1 mouse is not a true null and may retain the inhibitory function of NOTCH4. An additional function that may be conserved in the Notch4d1 mouse is cis-inhibition. Notch and ligand co-expressed in a single cell i.e. in cis can lead to their mutual inhibition (Section 1.2.9). Although the precise biochemical mechanisms of cis-inhibition 212 | Page
have yet to be unambiguously identified the cis-interaction between receptor and ligand is mediated by the EGF-like repeats (Section 1.2.9). The cis-interaction would thus be available for the NOTCH4d1 protein. It has yet to be established if NOTCH4 can interact with ligands in cis. An investigation of the cis interaction, not only of NOTCH4 but NOTCH4d1 as well, may point to functions of NOTCH4 that have been preserved in the Notch4d1 mouse.
The potential for intact functions and novel functions makes interpretation of the Notch4d1 phenotype in terms of Notch4 function problematic. Below is a discussion of an alternative approach to generate a Notch4 loss-of-function mutant.
5.3.3 A definitive Notch4 knockout In order to generate a true null Notch4 mouse we investigated a number of options. There was insufficient time to generate and characterise a new mouse for this PhD. Below is a discussion of the approach we took. The production of loss-of-function mutants for every mouse gene is being conducted by an international consortium, KOMP (knockout out mouse project, http://www.knockoutmouse.org). Many of the embryonic stem (ES) cell lines available for Notch4 loss-of-function involve gene trap methods. Gene trap methods are useful for generating large numbers of clones and thus maximal genome coverage. However, these mutants have the potential for incomplete knockout due to alternative splicing events. A guaranteed null Notch4 ES line was generated as part of the definitive knockout mouse project, Velocigene’s KOMP definitive null allele (http://www.velocigene.com/komp/detail/10800). The aim of this project is to replace all coding sequence from start to stop codon. The advantage of this approach is that there is no possibility of even partial protein expression, thus the alleles generated are a definitive null. However, there are drawbacks to this approach. The total deletion of the Notch4 open reading frame removes 23,547 base pairs of genomic sequence. The deletion of intronic sequences may have unintended consequences aside from deleting Notch4. The mouse genome is predicted to contain more than 1000 microRNAs. Over 50% of the known microRNA species are located within introns, making it possible that the Notch4 definitive knockout allele also deletes microRNA species (Osokine et al., 2008). This caveat not only applies to “definitive” knockouts. Osokine et al., 2008 identified almost 200 cases where disruption of a protein coding allele also disrupted a microRNA, pointing to a potentially widespread issue with knockouts. This is particularly well demonstrated in the Egfl7 gene knockout. Initial descriptions attributed 50% embryonic lethality with vascular defects as being due to the loss of Egfl7 (Schmidt et al., 2009). However, Egfl7 intron 7 also contains a miRNA, miRNA126. Selective deletion of 213 | Page
miRNA126 or Egfl7 found that these phenotypes were due the deletion of miRNA126 and not Egfl7 (Kuhnert et al., 2008). Although there are no miRNA genes identified in the introns of the Notch4 gene (miRBase database, http://mirbase.org), we cannot exclude the possibility of unidentified functional sequences that are deleted in addition to the Notch4 coding region. A disruption of the genome on this scale may have additional unexpected consequences. Although enhancers usually act within tens of kilobases of the genes they influence, some have been described to act as much as a megabase away (West and Fraser, 2005). Any phenotype observed in these mice will need to take the potential for functional sequences in addition to the Notch4 coding region into account.
The need to produce this mouse was driven by the incomplete knockout of Notch4 in the Notch4d1 mouse. We decided the advantage of the definitive null design outweighed the disadvantages of such a large deletion. Mice have been produced from ES cells by the Australian Phenomics Network (http:/www.australianphenomics.org.au). The correct integration of the insertion cassette has been confirmed by PCR analysis using primers that spanned both the 3’ and 5’ insertion sites. The establishment of this line and analysis of these mice will form future work.
5.3.4 Conclusions Analysis of the Notch4d1 mouse identified a previously unidentified phenotype. We observed a transient growth delay in the retinal vasculature at P5. The interpretation of this mild phenotype in terms of Notch4 function is complicated by the existence of a transcript from the Notch4d1 locus. The transcript is likely to produce a protein that not only retains some NOTCH4 function but could also have novel functions. In order to analyse a true null mouse a new model is required. We identified the definitive Notch4 knockout mouse as being an appropriate model system with the caveat that the large deletion could potentially have unintended consequences. The potential phenotypes of this mouse in terms of a model of NOTCH4 function based on the results of Chapters 3 and 4 will be discussed in the final chapter.
214 | Page
215 | Page
6 Discussion Notch4 is the most divergent and least well characterised of the mammalian Notch family. The targeted deletion of Notch1 and Notch2 are embryonic lethal and Notch3-/- mice display defects in smooth muscle cell investment (Conlon et al., 1995; Swiatek et al., 1994; Hamada et al., 1999; McCright et al., 2001; Domenga et al., 2004). In contrast the function of Notch4 remains elusive. The expression domain and the vascular phenotypes in gain-of-function mutants suggest a role in angiogenesis (Uyttendaele et al., 2001; Carlson et al., 2005). The loss-of-function model, the Notch4d1 mouse, does not display overt developmental defects. However, on a Notch1-/- background the Notch4d1 allele exacerbates the angiogenic defects in Notch1-/- mice (Krebs et al., 2000). We detected a mild angiogenic phenotype in Notch4d1 mice but in addition found that the NOTCH4 receptor can act as an inhibitor of canonical NOTCH1 signalling and that NOTCH4d1 retains this function. Below is a discussion of the possible functions of Notch4.
6.1 Evolutionary evidence for Notch4 function Notch4 is the most divergent member of the mammalian Notch receptor family. After the initial duplication that created the Notch4 gene in the mammalian lineage Notch4 underwent a period of rapid divergence (Theodosiou et al., 2009). The relaxation of selection pressure at this time makes the relationship between Notch4 and other mammalian family members ambiguous (Kortschak et al., 2001; Theodosiou et al., 2009). However, there are a number of lines of reasoning that support selection for maintenance of the NOTCH4 protein and thus a functional consequence to NOTCH4 expression. Notch4 homologues were present in at least as early as the placental/marsupial divergence 160 million years ago (mya) (Luo et al., 2011) as there is a Notch4 homologue present in the marsupial Monodelphis domestica (opossum) (NCBI Reference Sequence: XP_001376558.2). The average rate of divergence of a gene without selection pressure equals the mutation rate which has been estimated in mammals as 2.2x10-9/base/year (Kumar and Subramanian, 2002). In a comparison of the mouse and human genomes, which diverged 75 mya, the estimate of the substitution rate was as high as 1 substitution for every 2 nucleotides (Mouse Genome Sequencing et al., 2002). Thus there has been ample time for Notch4, in the absence of selection pressure, to have drifted to pseudogene status. Notch4 encodes a functional open reading frame suggestive of selection pressure favouring protein production. Although the EGF-like repeat region of NOTCH4 is divergent compared to other Notch receptors, the domain structure has been conserved. This is exemplified in the EGF-like repeats derived 216 | Page
from fusions of repeat 14 and 15, 16 and 17, 20 and 23, 26 and 27 and 31 and 32 of the Notch1 and Notch2 receptors (Gallahan and Callahan, 1997). The deletion events that produced these fusion repeats in Notch4 maintained the EGF-like repeat structure including the critical spacing and conservation of the cysteine residues that stabilise the EGF-like repeat fold. The maintenance of structure in these domains suggests that they were under selection pressure to maintain a folded protein structure and thus function.
6.2 NOTCH4ICD Functions The Notch receptors can be divided into two regions that control their function, the extracellular domain and the intracellular domain. Our analysis of the Notch4d1 mouse revealed only a very mild phenotype, namely a small delay in the growth of the retinal blood vessels (Section 5.2.4). Although we detected transcripts for the extracellular domains of Notch4 in these mice they are null for the Notch4ICD (Sections 5.2.5 - 5.2.8). These results may suggest that the NOTCH4ICD is largely dispensable. However, the NOTCH4ICD remains able to transactivate RBPJ dependent promoters (Section 3.2.5) and its potential for function is highlighted in the strong phenotypes of mice over expressing Notch4ICD constructs (Carlson et al., 2005; Miniati et al., 2009; Murphy et al., 2008; Uyttendaele et al., 2001). There is evidence that Notch4ICD may be under negative selection. Notch4 was initially identified as the insertion site of a murine mammary t umour virus (MMTV), the Int-3 locus (Gallahan et al., 1987). This insertion detected in a colony of wild type mice results in NOTCH4ICD expression driven by the viral promoter. Int-3 mice exhibit infertility in males and the development of breast cancer in females with 100% penetrance in addition to other hyperproliferative lesions (Jhappan et al., 1992). Thus in populations exposed to MMTV, such as the mice described, the Notch4 locus becomes a potentially lethal locus and is likely to be under selection to deactivate its transcriptional capacity.
NOTCH4ICD may also play as yet undescribed roles. Transgenic mice are routinely housed and maintained in disease free and stress free conditions. Under such conditions many of the selection pressures that act on wild mice are removed. Notch4 could have functions under specific conditions faced by wild rather than laboratory mice. For example, Notch4 is expressed in cells of the immune system but its function has not been investigated. The Notch4 promoter contains binding sites for GATA transcription factors, known to be important in lineage commitment within the haematopoietic system (Orkin, 1992; Tsai et al., 1994; Vercauteren and Sutherland, 2004). Expression of Notch4ICD in haematopoietic progenitors results in enhanced stem cell activity, 217 | Page
impaired differentiation and altered lymphoid development (Vercauteren and Sutherland, 2004). Notch4 is also expressed in maturing macrophages (Singh et al., 2000), dendritic cells (Sekine et al., 2009) and megakaryocyte-erythrocyte precursors (Yuan et al., 2010). A thorough analysis of haematopoietic development and the immune response in both Notch4d1 and the Notch4 definitive null (Notch4-/-) mice may uncover a role for NOTCH4.
Another limitation of the use of transgenic mice involves the use of inbred strains. Genes evolve in the context of other genes in the gene pool of the species. In the extremely limited gene pool of the inbred strains used, Notch4 is not exposed to all the genetic interactions possible in an outbred population. Potentially, the function of Notch4 could become apparent on other genetic backgrounds. This is exemplified in the case of the Dll4-/- mouse where haploinsufficiency is embryonic lethal in some strains while others survive post birth (Duarte et al., 2004; Gale et al., 2004; Hellström et al., 2007; Krebs et al., 2004).
Although others have established that NOTCH4 at least has the capacity for ligand dependent activation (Funahashi et al., 2008; Shawber et al., 2003; Shawber et al., 2007) we could not detect ligand-dependent signalling by NOTCH4 (Chapter 3). Our chimaeric Notch1:4 constructs (Section 3.2.14) indicate that the NOTCH4 extracellular domain responds to canonical Notch activation by ligand if combined with the more powerful transcriptional activation domains of NOTCH1. Thus the relatively low transactivation potential of the NOTCH4ICD obscured this weak activation of the receptor in response to ligand. The lack of transactivation by NOTCH4 with our reporter may indicate that NOTCH4 may be more active at a subset of promoters as has been found for NOTCH3 (Ong et al., 2006) or potentially has a non- canonical role (Section 1.2.10). The NOTCH4ICD displays a unique elution profile in comparison to other Notch family members which was not due to interactions with RBPJ and MAML and was detected as a free monomer over a range of concentrations in contrast to NOTCH1-3 (Han et al., 2011). Although peptide fragments of the RAM domain of NOTCH4 bound to RBPJ with similar affinities to that of other family members (Lubman et al., 2007), the elution results support a lower affinity of the NOTCH4ICD for RBPJ/MAML complexes. Many of the non-canonical interactions of NotchICDs are only detected in over expression systems. Under normal physiological conditions RBPJ is in excess while free Notch1ICD is not detectable (Lubman et al., 2007). The low interaction of the NOTCH4ICD with RBPJ/MAML (Han et al., 2011) may allow more NOTCH4ICD to be available for non-canonical roles. Non-canonical roles for NOTCH4 have been suggested for HIF1A and transforming growth factor beta (TGFB) regulation. The ability of 218 | Page
NOTCH4ICD to interact with HIF1AN (also known as FIH), but not be hydroxylated, has been suggested as a means by which NOTCH4 could regulate HIF1A (Section 1.2.10) (Wilkins et al., 2009). Notch receptors have been shown to interact with the TGFB pathway. Briefly, the TGFB pathway is activated by ligand leading to the phosphorylation of the receptor-regulated small body size, mothers against decapentaplegic (SMAD) proteins, SMAD2 and SMAD3. SMAD2/3 interact with the co-SMAD, SMAD4 and this complex is translocated to the nucleus to activate transcription (Attisano and Wrana, 2002). The interaction between the Notch and TGFB pathways is highly complex and occurs at multiple levels. Notch1ICD can interact with SMAD proteins and recruit them to RBPJ complexes to further enhance Notch mediated transcription (Dahlqvist et al., 2003). In addition, SMADs can recruit Notch1ICD to SMAD-dependent promoters and enhance their transcription (Tang et al., 2010). These interactions can have complex outcomes because recruitment to one class of promoters can inhibit recruitment to the other (Itoh et al., 2004). Key components of each pathway are also regulated via signalling from the other pathway (Fu et al., 2009; Kennard et al., 2007). However, there have been reports that the interaction with NOTCH4ICD has distinct outcomes from interactions with NOTCH1ICD (Sun et al., 2005; Tang et al., 2010). NOTCH4ICD, but not NOTCH1ICD or NOTCH2ICD, co-immunoprecipitates with phospho-SMAD2/3 in smooth muscle cells. NOTCH4ICD binds to and inhibits SMAD3-dependent gene expression in a variety of cell lines while Notch1ICD has more variable cell line specific effects (Sun et al., 2005).
However, our inability to detect NOTCH4ICD production with either antibodies or reporters suggests little, if any, NOTCH4ICD is produced upon ligand stimulation. The lack of phenotype observed by others (Krebs et al., 2000) and the very mild phenotype we observed in the Notch4d1 mice (Section 5.2.4) suggest that the NOTCH4ICD plays a minor role or that its functions are not evident in laboratory-housed inbred mice.
6.3 Notch4 extracellular domain functions. The extracellular regions of Notch receptors control ligand dependent activation and also bind ligand in cis (Sections 1.2.5 and 1.2.9). In the Notch4d1 mouse these regions are intact and thus phenotypic consequences of their deletion are unknown.
cis-Inhibition is used in development to establish signal sending and signal receiving populations. The separation of these two populations of cells is then reinforced through transcriptional feedback loops (Sprinzak et al., 2011; Sprinzak et al., 2010). Although not as well characterised as ligand cis-inhibition of Notch
219 | Page
receptors, Notch receptors appear to also inhibit ligand function in cis (del Alamo et al 2011). We have not investigated cis-inhibition of ligand by NOTCH4 although this may be an attractive area of study. We noted in our RNA in situ hybridisation analysis of Notch4 expression that there was strong expression in the tail bud and presomitic mesoderm (Section 5.2.8). During somite development, among other functions, Notch signalling establishes somite boundaries and anterior-posterior polarity (Barrantes et al., 1999; Conlon et al., 1995; Feller et al., 2008; Jiang et al., 2000; Morimoto et al., 2005). The anterior-posterior polarity is established by the restriction of Notch signalling to a narrow band of cells. In Dll3-/- embryos, NOTCH1 signalling fails to restrict and a broad area of signalling cells remains. These findings are consistent with cis-inhibition of NOTCH1 by DLL3 (Chapman et al., 2011). The interplay between multiple receptors and ligands and both trans and cis effects in developing somites makes this an attractive model system in which to study potential modulators of canonical Notch signalling such as Notch4. In addition our laboratory has developed techniques to stress this process (Sparrow et al., 2012). Mouse embryos exposed to low oxygen at mid-gestation (E9.5) have more severe somitogenesis defects in association with Notch pathway mutations. This system would allow stresses to be placed on the Notch4-/- embryos which may uncover a role for Notch4.
6.4 A model of Notch4 function The major and novel function of the NOTCH4 extracellular domains that we detected was as an inhibitor of NOTCH1. The NOTCH4 extracellular domains interact with NOTCH1 (Section 5.2.11). This interaction was accompanied by an altered subcellular distribution of NOTCH1 (Section 4.2.6). The inhibition of NOTCH1 was traced to an intracellular event as the NOTCH4 extracellular domains interacted with NOTCH1 prior to S1 processing (Section 5.2.11). The interaction between NOTCH4 and NOTCH1 led to a decrease in canonical NOTCH1 signalling (Section 4.2.1). We cannot exclude the possibility that NOTCH4 also inhibited NOTCH1 at the cell surface by competing unproductively for ligand. In addition, NOTCH4 may also share components of the maturation pathway with NOTCH1. The level of surface expression of NOTCH1 is tightly regulated. If surface expression of NOTCH1 is co-regulated with NOTCH4, NOTCH4 may serve to decrease NOTCH1 by competing for surface presentation. However, we found NOTCH4 was poorly surface expressed indicating a predominantly intracellular role.
Both weak canonical NOTCH4 signalling and inhibition of NOTCH1 canonical signalling could be occurring concurrently with the net result of inhibition. This would predict circumstances under which 220 | Page
canonical NOTCH4 signalling could produce a net positive effect. In the absence of NOTCH4 expression canonical NOTCH1 signalling proceeds (Figure 6.1a). In cells where Notch signalling from other receptors is very low or absent the inhibitory effect of NOTCH4 would be relaxed and may reveal canonical NOTCH4 signalling (Figure 6.1b). In cells expressing both NOTCH1 and NOTCH4 the inhibitory role of NOTCH4 would become apparent (Figure 6.1c). It is of interest to note in this respect that Notch4 expression is under the control of canonical Notch signalling. The Notch4 promoter contains RBPJ binding sites (Li et al., 1998) and Notch4 expression is upregulated by transfection with NotchICD constructs (Carlson et al., 2005; Uyttendaele et al., 2000). Thus NOTCH4 may serve as an attenuator of NOTCH1 signalling, limiting its duration. A similar mechanism has been described for Nrarp (Phng et al., 2009) (Section 5.3.1). Notch signalling induces NRARP expression which in turn promotes NotchICD degradation thus attenuating Notch signalling.
221 | Page
Figure 6.1 Model of NOTCH4 function a. Canonical NOTCH1 signalling. NOTCH1 is cleaved by FURIN (S1) to form the Notch heterodimer (i.). NOTCH1 is exclusively presented on the cell surface post S1 cleavage (II.) and only the NOTCH1 heterodimer is activated by DSL ligands (iii.). Following ligand binding the S2 site is exposed and cleaved to form the Notch1 TMIC. The NOTCH1 TMIC is further processed by gamma- secretase (S3 and S4) (iv.) and the Notch1ICD translocates to the nucleus to activate RBPJ promoters (v.). b. In contrast to NOTCH1, NOTCH4 undergoes minimal S1 processing (i.). NOTCH4 is poorly presented on the cell surface in comparison to NOTCH1 and the majority of cell surface NOTCH4 is not S1 processed (ii.). We were unable to detect signalling via the full length NOTCH4 receptor in response to ligand. NOTCH4 TMIC constructs are processed by gamma-secretase (S3 and S4) (iv.), translocate to the nucleus and are relatively poor activators of RBPJ promoters compared to NOTCH1 (v.). c. NOTCH4 co-expression with NOTCH1 leads to inhibition of canonical Notch signalling. NOTCH4 directly interacts with NOTCH1 prior to S1 cleavage which leads to NOTCH1 altering its subcellular localisation (i.) and suppression of the canonical pathway. The NOTCH4 TMIC does not inhibit the NOTCH1 TMIC and does not affect the downstream events; gamma-secretase processing (S3 and S4) or activation of RBPJ promoters.
222 | Page
223 | Page
The Notch4d1 mouse, initially designed as a null allele, expresses a transcript containing most of the Notch4 extracellular domain (Section 5.2.6). Expression of a cDNA construct based on this transcript inhibited NOTCH1 signalling in co-culture assays (Section 5.2.9), promoted myoblast differentiation (Section 5.2.10) and co-immunprecipitates NOTCH1 (Section 5.2.11) in a similar way to NOTCH4. Thus NOTCH4d1 retains the inhibitory ability of NOTCH4. Therefore we developed the definitive Notch4 knockout mouse (Notch4-/-) to investigate the role of the NOTCH4 extracellular domains. The major expression domain of Notch4 in the vasculature and the effect NOTCH4ICD has on vascular development and maintenance of arterial identity suggest a role in the endothelium. Notch4 is a mammalian specific Notch homologue which is absent in other vertebrate lineages. However, vertebrate lineages which lack Notch4 are still able to generate and maintain a functional vasculature. This suggests that the role of Notch4 is a subsidiary one that may only be required under specific circumstances and contexts. The retinal angiogenesis model remains an attractive system for studying Notch4 function. Similar to the role of Notch in somitogenesis, multiple receptors and ligands interact to fine tune this process. Additionally there is the potential, although not established, for interactions in cis to affect this process. The dosage dependence of Notch signalling in retinal angiogenesis has been established in a number of ways. Mice deficient for a single allele of Dll4 display dramatic defects in angiogenesis (Duarte et al., 2004; Gale et al., 2004; Hellström et al., 2007; Krebs et al., 2004). Additionally, inducible over expression of JAG1 in endothelial cells suppresses DLL4 mediated activation of Notch in the retina by competing for binding but not activating signal due to Fringe modification (Benedito et al., 2009). The importance of the Fringe modification is highlighted by the retinal defects seen in Lfng-/- mice. These mice still expressed other Fringe members, Rfng and Mfng, but these failed to entirely compensate for loss of Lfng (Benedito et al., 2009). Thus the retina is an attractive model for investigating what may be subtle defects associated with the overall strength of Notch activation.
If we can detect even a mild phenotype in the Notch4-/- mouse, the Notch4d1 mouse would become an extremely valuable tool. The Notch4d1 mouse is null for the NOTCH4ICD and therefore could be used to separate functions attributable to the ICD and the extracellular domains. If Notch4-/d1 heterozygotes do not display any phenotype compared to a Notch4-/- then we could attribute the phenotype to a role of the extracellular domains independent of NOTCH4ICD functions.
224 | Page
7 Bibliography Acar, M., H. Jafar-Nejad, H. Takeuchi, A. Rajan, D. Ibrani, N.A. Rana, H. Pan, R.S. Haltiwanger, and H.J. Bellen. 2008. Rumi is a CAP10 domain glycosyltransferase that modifies Notch and is required for Notch signaling. Cell. 132:247-258. Ahimou, F., L.P. Mok, B. Bardot, and C. Wesley. 2004. The adhesion force of Notch with Delta and the rate of Notch signaling. J Cell Biol. 167:1217-1229. Aitsebaomo, J., A.L. Portbury, J.C. Schisler, and C. Patterson. 2008. Brothers and sisters: molecular insights into arterial- venous heterogeneity. Circ Res. 103:929-939. Albig, A., D. Becenti, T. Roy, and W. Schiemann. 2008. Microfibril- associate glycoprotein-2 (MAGP-2) promotes angiogenic cell sprouting by blocking notch signaling in endothelial cells. Microvasc Res. 76:7-14. Albrecht, I., and G. Christofori. 2011. Molecular mechanisms of lymphangiogenesis in development and cancer. Int J Dev Biol. 55:483-494. Andersen, P., H. Uosaki, L.T. Shenje, and C. Kwon. 2012. Non- canonical Notch signaling: emerging role and mechanism. Trends Cell Biol. 22:257-265. Andersson, E.R., R. Sandberg, and U. Lendahl. 2011. Notch signaling: simplicity in design, versatility in function. Development. 138:3593-3612. Asahara, T., T. Murohara, A. Sullivan, M. Silver, R. van der Zee, T. Li, B. Witzenbichler, G. Schatteman, and J.M. Isner. 1997. Isolation of putative progenitor endothelial cells for angiogenesis. Science. 275:964-967. Aste-Amezaga, M., N. Zhang, J.E. Lineberger, B.A. Arnold, T.J. Toner, M. Gu, L. Huang, S. Vitelli, K.T. Vo, P. Haytko, J.Z. Zhao, F. Baleydier, S. L'Heureux, H. Wang, W.R. Gordon, E. Thoryk, M.B. Andrawes, K. Tiyanont, K. Stegmaier, G. Roti, K.N. Ross, L.L. Franlin, H. Wang, F. Wang, M. Chastain, A.J. Bett, L.P. Audoly, J.C. Aster, S.C. Blacklow, and H.E. Huber. 2010. Characterization of Notch1 antibodies that inhibit signaling of both normal and mutated Notch1 receptors. PLoS One. 5:e9094. Aster, J., W. Pear, R. Hasserjian, H. Erba, F. Davi, B. Luo, M. Scott, D. Baltimore, and J. Sklar. 1994. Functional analysis of the TAN-1 gene, a human homolog of Drosophila notch. Cold Spring Harbor symposia on quantitative biology. 59:125-136. Aster, J.C., N. Bodnar, L. Xu, F. Karnell, J.M. Milholland, I. Maillard, G. Histen, Y. Nam, S.C. Blacklow, and W.S. Pear. 2011. Notch ankyrin repeat domain variation influences leukemogenesis and Myc transactivation. PLoS One. 6:e25645. Attisano, L., and J.L. Wrana. 2002. Signal transduction by the TGF-beta superfamily. Science. 296:1646-1647. Baladron, V., M.J. Ruiz-Hidalgo, M.L. Nueda, M.J. Diaz-Guerra, J.J. Garcia-Ramirez, E. Bonvini, E. Gubina, and J. Laborda. 2005. dlk acts as a negative regulator of Notch1 activation through interactions with specific EGF-like repeats. Exp Cell Res. 303:343-359. 225 | Page
Baluk, P., J. Fuxe, H. Hashizume, T. Romano, E. Lashnits, S. Butz, D. Vestweber, M. Corada, C. Molendini, E. Dejana, and D.M. McDonald. 2007. Functionally specialized junctions between endothelial cells of lymphatic vessels. J Exp Med. 204:2349-2362. Banerji, S., J. Ni, S.X. Wang, S. Clasper, J. Su, R. Tammi, M. Jones, and D.G. Jackson. 1999. LYVE-1, a new homologue of the CD44 glycoprotein, is a lymph-specific receptor for hyaluronan. J Cell Biol. 144:789-801. Barrantes, I.B., A.J. Elia, K. Wunsch, M.H. Hrabe de Angelis, T.W. Mak, J. Rossant, R.A. Conlon, A. Gossler, and J.L. de la Pompa. 1999. Interaction between Notch signalling and Lunatic fringe during somite boundary formation in the mouse. Curr Biol. 9:470-480. Bash, J., W.X. Zong, S. Banga, A. Rivera, D.W. Ballard, Y. Ron, and C. Gelinas. 1999. Rel/NF-kappaB can trigger the Notch signaling pathway by inducing the expression of Jagged1, a ligand for Notch receptors. EMBO J. 18:2803-2811. Beatus, P., J. Lundkvist, C. Oberg, and U. Lendahl. 1999. The notch 3 intracellular domain represses notch 1-mediated activation through Hairy/Enhancer of split (HES) promoters. Development. 126:3925-3935. Beatus, P., J. Lundkvist, C. Oberg, K. Pedersen, and U. Lendahl. 2001. The origin of the ankyrin repeat region in Notch intracellular domains is critical for regulation of HES promoter activity. Mech Dev. 104:3-20. Becam, I., U.M. Fiuza, A.M. Arias, and M. Milan. 2010. A role of receptor Notch in ligand cis-inhibition in Drosophila. Curr Biol. 20:554-560. Benedito, R., C. Roca, I. Sörensen, S. Adams, A. Gossler, M. Fruttiger, and R.H. Adams. 2009. The Notch Ligands Dll4 and Jagged1 Have Opposing Effects on Angiogenesis. Cell. 137:1124-1135. Bentley, K., H. Gerhardt, and P.A. Bates. 2008. Agent-based simulation of notch-mediated tip cell selection in angiogenic sprout initialisation. J Theor Biol. 250:25-36. Bentley, K., G. Mariggi, H. Gerhardt, and P.A. Bates. 2009. Tipping the balance: robustness of tip cell selection, migration and fusion in angiogenesis. PLoS Comput Biol. 5:e1000549. Bertagna, A., D. Toptygin, L. Brand, and D. Barrick. 2008. The effects of conformational heterogeneity on the binding of the Notch intracellular domain to effector proteins: a case of biologically tuned disorder. Biochem. Soc. Trans. 36:157. Bettenhausen, B., and A. Gossler. 1995. Efficient isolation of novel mouse genes differentially expressed in early postimplantation embryos. Genomics. 28:436-441. Blaumueller, C.M., H. Qi, P. Zagouras, and S. Artavanis-Tsakonas. 1997. Intracellular cleavage of Notch leads to a heterodimeric receptor on the plasma membrane. Cell. 90:281-291. Blum, Y., H.G. Belting, E. Ellertsdottir, L. Herwig, F. Luders, and M. Affolter. 2008. Complex cell rearrangements during
226 | Page
intersegmental vessel sprouting and vessel fusion in the zebrafish embryo. Dev Biol. 316:312-322. Bonegio, R. 2009. Notch signaling during kidney development and injury responses. PhD Thesis, Boston University. Borson, N.D., W.L. Salo, and L.R. Drewes. 1992. A lock-docking oligo(dT) primer for 5' and 3' RACE PCR. PCR methods and applications. 2:144-148. Bozkulak, E.C., and G. Weinmaster. 2009. Selective use of ADAM10 and ADAM17 in activation of Notch1 signaling. Mol Cell Biol. 29:5679-5695. Bray, S.J. 2006. Notch signalling: a simple pathway becomes complex. Nat Rev Mol Cell Biol. 7:678-689. Brou, C., F. Logeat, N. Gupta, C. Bessia, O. LeBail, J.R. Doedens, A. Cumano, P. Roux, R.A. Black, and A. Israël. 2000. A novel proteolytic cleavage involved in Notch signaling: the role of the disintegrin-metalloprotease TACE. Mol Cell. 5:207-216. Buckley, N.E., K.L. McCoy, E. Mezey, T. Bonner, A. Zimmer, C.C. Felder, M. Glass, and A. Zimmer. 2000. Immunomodulation by cannabinoids is absent in mice deficient for the cannabinoid CB(2) receptor. Eur J Pharmacol. 396:141-149. Bulman, M.P., K. Kusumi, T.M. Frayling, C. McKeown, C. Garrett, E.S. Lander, R. Krumlauf, A.T. Hattersley, S. Ellard, and P.D. Turnpenny. 2000. Mutations in the human delta homologue, DLL3, cause axial skeletal defects in spondylocostal dysostosis. Nat Genet. 24:438-441. Bush, G., G. diSibio, A. Miyamoto, J.B. Denault, R. Leduc, and G. Weinmaster. 2001. Ligand-induced signaling in the absence of furin processing of Notch1. Dev Biol. 229:494-502. Bush, H.L., Jr., J.A. Jakubowski, G.R. Curl, D. Deykin, and D.C. Nabseth. 1986. The natural history of endothelial structure and function in arterialized vein grafts. J Vasc Surg. 3:204- 215. Campos, A.H., W. Wang, M.J. Pollman, and G.H. Gibbons. 2002. Determinants of Notch-3 receptor expression and signaling in vascular smooth muscle cells: implications in cell-cycle regulation. Circ Res. 91:999-1006. Caprioli, A., T. Jaffredo, R. Gautier, C. Dubourg, and F. Dieterlen- Lievre. 1998. Blood-borne seeding by hematopoietic and endothelial precursors from the allantois. Proc Natl Acad Sci U S A. 95:1641-1646. Caprioli, A., K. Minko, C. Drevon, A. Eichmann, F. Dieterlen- Lievre, and T. Jaffredo. 2001. Hemangioblast commitment in the avian allantois: cellular and molecular aspects. Dev Biol. 238:64-78. Carlson, T.R., Y. Yan, X. Wu, M.T. Lam, G.L. Tang, L.J. Beverly, L.M. Messina, A.J. Capobianco, Z. Werb, and R. Wang. 2005. Endothelial expression of constitutively active Notch4 elicits reversible arteriovenous malformations in adult mice. Proc Natl Acad Sci USA. 102:9884-9889. Carmeliet, P. 2003. Angiogenesis in health and disease. Nat Med. 9:653-660. Carmeliet, P., V. Ferreira, G. Breier, S. Pollefeyt, L. Kieckens, M. Gertsenstein, M. Fahrig, A. Vandenhoeck, K. Harpal, C. Eberhardt, C. Declercq, J. Pawling, L. Moons, D. Collen, W.
227 | Page
Risau, and A. Nagy. 1996. Abnormal blood vessel development and lethality in embryos lacking a single VEGF allele. Nature. 380:435-439. Carmeliet, P., Y.S. Ng, D. Nuyens, G. Theilmeier, K. Brusselmans, I. Cornelissen, E. Ehler, V.V. Kakkar, I. Stalmans, V. Mattot, J.C. Perriard, M. Dewerchin, W. Flameng, A. Nagy, F. Lupu, L. Moons, D. Collen, P.A. D'Amore, and D.T. Shima. 1999. Impaired myocardial angiogenesis and ischemic cardiomyopathy in mice lacking the vascular endothelial growth factor isoforms VEGF164 and VEGF188. Nat Med. 5:495-502. Carmeliet, P., and M. Tessier-Lavigne. 2005. Common mechanisms of nerve and blood vessel wiring. Nature. 436:193-200. Chapman, G., L. Liu, C. Sahlgren, C. Dahlqvist, and U. Lendahl. 2006. High levels of Notch signaling down-regulate Numb and Numblike. J Cell Biol. 175:535-540. Chapman, G., D.B. Sparrow, E. Kremmer, and S.L. Dunwoodie. 2011. Notch inhibition by the ligand DELTA-LIKE 3 defines the mechanism of abnormal vertebral segmentation in spondylocostal dysostosis. Hum Mol Genet. 20:905-916. Chastagner, P., A. Israel, and C. Brou. 2008. AIP4/Itch regulates Notch receptor degradation in the absence of ligand. PLoS One. 3:e2735. Checchin, D., F. Sennlaub, E. Levavasseur, M. Leduc, and S. Chemtob. 2006. Potential role of microglia in retinal blood vessel formation. Invest Ophthalmol Vis Sci. 47:3595-3602. Chen, J., D.J. Moloney, and P. Stanley. 2001. Fringe modulation of Jagged1-induced Notch signaling requires the action of beta 4galactosyltransferase-1. Proc Natl Acad Sci U S A. 98:13716-13721. Cheng, P., A. Zlobin, V. Volgina, S. Gottipati, B. Osborne, E.J. Simel, L. Miele, and D.I. Gabrilovich. 2001. Notch-1 regulates NF-kappaB activity in hemopoietic progenitor cells. J Immunol. 167:4458-4467. Chi, Z., J. Zhang, A. Tokunaga, M.M. Harraz, S.T. Byrne, A. Dolinko, J. Xu, S. Blackshaw, N. Gaiano, T.M. Dawson, and V.L. Dawson. 2012. Botch promotes neurogenesis by antagonizing Notch. Dev Cell. 22:707-720. Chiang, M.Y. 2006. Identification of a Conserved Negative Regulatory Sequence That Influences the Leukemogenic Activity of NOTCH1. Mol Cell Biol. 26:6261-6271. Choi, K. 2002. The hemangioblast: a common progenitor of hematopoietic and endothelial cells. Journal of hematotherapy & stem cell research. 11:91-101. Chu, D., Z. Zhang, Y. Zhou, W. Wang, Y. Li, H. Zhang, G. Dong, Q. Zhao, and G. Ji. 2011. Notch1 and Notch2 have opposite prognostic effects on patients with colorectal cancer. Annals of oncology : official journal of the European Society for Medical Oncology / ESMO. 22:2440-2447. Chu, Y., S. Hughes, and T. Chan-Ling. 2001. Differentiation and migration of astrocyte precursor cells and astrocytes in human fetal retina: relevance to optic nerve coloboma. FASEB J. 15:2013-2015.
228 | Page
Claxton, S., and M. Fruttiger. 2003. Role of arteries in oxygen induced vaso-obliteration. Exp Eye Res. 77:305-311. Claxton, S., and M. Fruttiger. 2004. Periodic Delta-like 4 expression in developing retinal arteries. Gene Expr Patterns. 5:123-127. Coleman, M.L., M.A. Mcdonough, K.S. Hewitson, C. Coles, J. Mecinovic, M. Edelmann, K.M. Cook, M.E. Cockman, D.E. Lancaster, B.M. Kessler, N.J. Oldham, P.J. Ratcliffe, and C.J. Schofield. 2007. Asparaginyl Hydroxylation of the Notch Ankyrin Repeat Domain by Factor Inhibiting Hypoxia- inducible Factor. Journal of Biological Chemistry. 282:24027- 24038. Conlon, R.A., A.G. Reaume, and J. Rossant. 1995. Notch1 is required for the coordinate segmentation of somites. Development. 121:1533-1545. Corada, M., D. Nyqvist, F. Orsenigo, A. Caprini, C. Giampietro, M.M. Taketo, M.L. Iruela-Arispe, R.H. Adams, and E. Dejana. 2010. The Wnt/beta-catenin pathway modulates vascular remodeling and specification by upregulating Dll4/Notch signaling. Dev Cell. 18:938-949. Cordle, J., S. Johnson, J.Z. Tay, P. Roversi, M.B. Wilkin, B.H. de Madrid, H. Shimizu, S. Jensen, P. Whiteman, B. Jin, C. Redfield, M. Baron, S.M. Lea, and P.A. Handford. 2008. A conserved face of the Jagged/Serrate DSL domain is involved in Notch trans-activation and cis-inhibition. Nat Struct Mol Biol. 15:849-857. Coultas, L., K. Chawengsaksophak, and J. Rossant. 2005. Endothelial cells and VEGF in vascular development. Nature. 438:937-945. Cui, X.Y., Q.D. Hu, M. Tekaya, Y. Shimoda, B.T. Ang, D.Y. Nie, L. Sun, W.P. Hu, M. Karsak, T. Duka, Y. Takeda, L.Y. Ou, G.S. Dawe, F.G. Yu, S. Ahmed, L.H. Jin, M. Schachner, K. Watanabe, Y. Arsenijevic, and Z.C. Xiao. 2004. NB-3/Notch1 pathway via Deltex1 promotes neural progenitor cell differentiation into oligodendrocytes. J Biol Chem. 279:25858-25865. Cumano, A., F. Dieterlen-Lievre, and I. Godin. 2000. The splanchnopleura/AGM region is the prime site for the generation of multipotent hemopoietic precursors, in the mouse embryo. Vaccine. 18:1621-1623. Cummins, T.D., M.D. Mendenhall, M.N. Lowry, A.K. E, M.T. Barati, S.J. Khundmiri, S.A. Salyer, J.B. Klein, and D.W. Powell. 2011. Elongin C is a mediator of Notch4 activity in human renal tubule cells. Biochim Biophys Acta. 1814:1748-1757. D'Souza, B., A. Miyamoto, and G. Weinmaster. 2008. The many facets of Notch ligands. Oncogene. 27:5148-5167. Dahlqvist, C., A. Blokzijl, G. Chapman, A. Falk, K. Dannaeus, C.F. Ibanez, and U. Lendahl. 2003. Functional Notch signaling is required for BMP4-induced inhibition of myogenic differentiation. Development. 130:6089-6099. Davis, G.E., and D.R. Senger. 2005. Endothelial extracellular matrix: biosynthesis, remodeling, and functions during vascular morphogenesis and neovessel stabilization. Circ Res. 97:1093-1107.
229 | Page
de-la-Concha, A., U. Dietrich, D. Weigel, and J.A. Campos-Ortega. 1988. Functional interactions of neurogenic genes of Drosophila melanogaster. Genetics. 118:499-508. de Celis, J.F., R. Barrio, A. del Arco, and A. Garcia-Bellido. 1993. Genetic and molecular characterization of a Notch mutation in its Delta- and Serrate-binding domain in Drosophila. Proc Natl Acad Sci U S A. 90:4037-4041. de Celis, J.F., and S. Bray. 1997. Feed-back mechanisms affecting Notch activation at the dorsoventral boundary in the Drosophila wing. Development. 124:3241-3251. de Celis, J.F., and S.J. Bray. 2000. The Abruptex domain of Notch regulates negative interactions between Notch, its ligands and Fringe. Development. 127:1291-1302. de Celis, J.F., and A. Garcia-Bellido. 1994. Modifications of the notch function by Abruptex mutations in Drosophila melanogaster. Genetics. 136:183-194. Dejana, E., E. Tournier-Lasserve, and B.M. Weinstein. 2009. The control of vascular integrity by endothelial cell junctions: molecular basis and pathological implications. Dev Cell. 16:209-221. del Alamo, D., H. Rouault, and F. Schweisguth. 2011. Mechanism and significance of cis-inhibition in Notch signalling. Curr Biol. 21:R40-47. dela Paz, N.G., and P.A. D'Amore. 2009. Arterial versus venous endothelial cells. Cell Tissue Res. 335:5-16. Demir, R., Y. Seval, and B. Huppertz. 2007. Vasculogenesis and angiogenesis in the early human placenta. Acta histochemica. 109:257-265. Doi, H., T. Iso, H. Sato, M. Yamazaki, H. Matsui, T. Tanaka, I. Manabe, M. Arai, R. Nagai, and M. Kurabayashi. 2006. Jagged1-selective notch signaling induces smooth muscle differentiation via a RBP-Jkappa-dependent pathway. J Biol Chem. 281:28555-28564. Domenga, V., P. Fardoux, P. Lacombe, M. Monet, J. Maciazek, L.T. Krebs, B. Klonjkowski, E. Berrou, M. Mericskay, Z. Li, E. Tournier-Lasserve, T. Gridley, and A. Joutel. 2004. Notch3 is required for arterial identity and maturation of vascular smooth muscle cells. Genes Dev. 18:2730-2735. Dorrell, M., and M. Friedlander. 2006. Mechanisms of endothelial cell guidance and vascular patterning in the developing mouse retina. Progress in Retinal and Eye Research. 25:277- 295. Drake, C.J. 2003. Embryonic and adult vasculogenesis. Birth defects research. Part C, Embryo today : reviews. 69:73-82. Duarte, A., M. Hirashima, R. Benedito, A. Trindade, P. Diniz, E. Bekman, L. Costa, D. Henrique, and J. Rossant. 2004. Dosage-sensitive requirement for mouse Dll4 in artery development. Genes Dev. 18:2474-2478. Dumont, D.J., G. Gradwohl, G.H. Fong, M.C. Puri, M. Gertsenstein, A. Auerbach, and M.L. Breitman. 1994. Dominant-negative and targeted null mutations in the endothelial receptor tyrosine kinase, tek, reveal a critical role in vasculogenesis of the embryo. Genes Dev. 8:1897-1909.
230 | Page
Dunwoodie, S.L. 2009. Mutation of the fucose-specific beta1,3 N- acetylglucosaminyltransferase LFNG results in abnormal formation of the spine. Biochim Biophys Acta. 1792:100-111. Dunwoodie, S.L., M. Clements, D.B. Sparrow, X. Sa, R.A. Conlon, and R.S. Beddington. 2002. Axial skeletal defects caused by mutation in the spondylocostal dysplasia/pudgy gene Dll3 are associated with disruption of the segmentation clock within the presomitic mesoderm. Development. 129:1795-1806. Dunwoodie, S.L., D. Henrique, S.M. Harrison, and R.S. Beddington. 1997. Mouse Dll3: a novel divergent Delta gene which may complement the function of other Delta homologues during early pattern formation in the mouse embryo. Development. 124:3065-3076. Eichmann, A., C. Marcelle, C. Breant, and N.M. Le Douarin. 1993. Two molecules related to the VEGF receptor are expressed in early endothelial cells during avian embryonic development. Mech Dev. 42:33-48. Eiraku, M., Y. Hirata, H. Takeshima, T. Hirano, and M. Kengaku. 2002. Delta/notch-like epidermal growth factor (EGF)-related receptor, a novel EGF-like repeat-containing protein targeted to dendrites of developing and adult central nervous system neurons. J Biol Chem. 277:25400-25407. Eiraku, M., A. Tohgo, K. Ono, M. Kaneko, K. Fujishima, T. Hirano, and M. Kengaku. 2005. DNER acts as a neuron-specific Notch ligand during Bergmann glial development. Nat Neurosci. 8:873-880. Ellisen, L.W., J. Bird, D.C. West, A.L. Soreng, T.C. Reynolds, S.D. Smith, and J. Sklar. 1991. TAN-1, the human homolog of the Drosophila notch gene, is broken by chromosomal translocations in T lymphoblastic neoplasms. Cell. 66:649- 661. Engerman, R.L. 1976. Development of the macular circulation. Invest Ophthalmol. 15:835-840. Espinosa, L., J. Ingles-Esteve, C. Aguilera, and A. Bigas. 2003. Phosphorylation by glycogen synthase kinase-3 beta down- regulates Notch activity, a link for Notch and Wnt pathways. J Biol Chem. 278:32227-32235. Estrach, S., C.A. Ambler, C. Lo Celso, K. Hozumi, and F.M. Watt. 2006. Jagged 1 is a beta-catenin target gene required for ectopic hair follicle formation in adult epidermis. Development. 133:4427-4438. Evrard, Y.A., Y. Lun, A. Aulehla, L. Gan, and R.L. Johnson. 1998. lunatic fringe is an essential mediator of somite segmentation and patterning. Nature. 394:377-381. Falk, R., A. Falk, M.R. Dyson, A. Melidoni, K. Parthiban, J. Young, W. Roake, and J. McCafferty. 2012. Generation of anti-Notch antibodies and their application in blocking Notch signalling in neural stem cells. Methods. Fan, X., I. Mikolaenko, I. Elhassan, X. Ni, Y. Wang, D. Ball, D.J. Brat, A. Perry, and C.G. Eberhart. 2004. Notch1 and notch2 have opposite effects on embryonal brain tumor growth. Cancer Res. 64:7787-7793. Fehon, R.G., K. Johansen, I. Rebay, and S. Artavanis-Tsakonas. 1991. Complex cellular and subcellular regulation of notch
231 | Page
expression during embryonic and imaginal development of Drosophila: implications for notch function. J Cell Biol. 113:657-669. Feller, J., A. Schneider, K. Schuster-Gossler, and A. Gossler. 2008. Noncyclic Notch activity in the presomitic mesoderm demonstrates uncoupling of somite compartmentalization and boundary formation. Genes Dev. 22:2166-2171. Ferkowicz, M.J., and M.C. Yoder. 2005. Blood island formation: longstanding observations and modern interpretations. Exp Hematol. 33:1041-1047. Ferrara, N., K. Carver-Moore, H. Chen, M. Dowd, L. Lu, K.S. O'Shea, L. Powell-Braxton, K.J. Hillan, and M.W. Moore. 1996. Heterozygous embryonic lethality induced by targeted inactivation of the VEGF gene. Nature. 380:439-442. Fiedler, U., T. Krissl, S. Koidl, C. Weiss, T. Koblizek, U. Deutsch, G. Martiny-Baron, D. Marme, and H.G. Augustin. 2003. Angiopoietin-1 and angiopoietin-2 share the same binding domains in the Tie-2 receptor involving the first Ig-like loop and the epidermal growth factor-like repeats. J Biol Chem. 278:1721-1727. Fior, R., and D. Henrique. 2009. "Notch-Off": a perspective on the termination of Notch signalling. Int J Dev Biol. 53:1-7. Fischer, A., and M. Gessler. 2007. Delta-Notch--and then? Protein interactions and proposed modes of repression by Hes and Hey bHLH factors. Nucleic Acids Res. 35:4583-4596. Fitch, M.J., L. Campagnolo, F. Kuhnert, and H. Stuhlmann. 2004. Egfl7, a novel epidermal growth factor-domain gene expressed in endothelial cells. Dev Dyn. 230:316-324. Fiuza, U.M., T. Klein, A. Martinez Arias, and P. Hayward. 2010. Mechanisms of ligand-mediated inhibition in Notch signaling activity in Drosophila. Dev Dyn. 239:798-805. Fleming, R.J., Y. Gu, and N.A. Hukriede. 1997. Serrate-mediated activation of Notch is specifically blocked by the product of the gene fringe in the dorsal compartment of the Drosophila wing imaginal disc. Development. 124:2973-2981. Folkman, J. 1971. Tumor angiogenesis: therapeutic implications. N Engl J Med. 285:1182-1186. Francois, M., A. Caprini, B. Hosking, F. Orsenigo, D. Wilhelm, C. Browne, K. Paavonen, T. Karnezis, R. Shayan, M. Downes, T. Davidson, D. Tutt, K.S. Cheah, S.A. Stacker, G.E. Muscat, M.G. Achen, E. Dejana, and P. Koopman. 2008. Sox18 induces development of the lymphatic vasculature in mice. Nature. 456:643-647. Franklin, J.L., B.E. Berechid, F.B. Cutting, A. Presente, C.B. Chambers, D.R. Foltz, A. Ferreira, and J.S. Nye. 1999. Autonomous and non-autonomous regulation of mammalian neurite development by Notch1 and Delta1. Curr Biol. 9:1448-1457. Friedlander, M., M.I. Dorrell, M.R. Ritter, V. Marchetti, S.K. Moreno, M. El-Kalay, A.C. Bird, E. Banin, and E. Aguilar. 2007. Progenitor cells and retinal angiogenesis. Angiogenesis. 10:89-101.
232 | Page
Friedmann, D.R., J.J. Wilson, and R.A. Kovall. 2008. RAM-induced allostery facilitates assembly of a notch pathway active transcription complex. J Biol Chem. 283:14781-14791. Fruttiger, M. 2002. Development of the mouse retinal vasculature: angiogenesis versus vasculogenesis. Invest Ophthalmol Vis Sci. 43:522-527. Fruttiger, M. 2007. Development of the retinal vasculature. Angiogenesis. 10:77-88. Fruttiger, M., A.R. Calver, and W.D. Richardson. 2000. Platelet- derived growth factor is constitutively secreted from neuronal cell bodies but not from axons. Curr Biol. 10:1283-1286. Fryer, C., J. White, and K. Jones. 2004. Mastermind Recruits CycC:CDK8 to Phosphorylate the Notch ICD and Coordinate Activation with Turnover. Mol Cell. 16:509-520. Fryer, C.J., E. Lamar, I. Turbachova, C. Kintner, and K.A. Jones. 2002. Mastermind mediates chromatin-specific transcription and turnover of the Notch enhancer complex. Genes Dev. 16:1397-1411. Fu, Y., A. Chang, L. Chang, K. Niessen, S. Eapen, A. Setiadi, and A. Karsan. 2009. Differential regulation of transforming growth factor beta signaling pathways by Notch in human endothelial cells. J Biol Chem. 284:19452-19462. Fukumori, A., M. Okochi, S. Tagami, J. Jiang, N. Itoh, T. Nakayama, K. Yanagida, Y. Ishizuka-Katsura, T. Morihara, K. Kamino, T. Tanaka, T. Kudo, H. Tanii, A. Ikuta, C. Haass, and M. Takeda. 2006. Presenilin-dependent gamma- secretase on plasma membrane and endosomes is functionally distinct. Biochemistry. 45:4907-4914. Funahashi, Y., S.L. Hernandez, I. Das, A. Ahn, J. Huang, M. Vorontchikhina, A. Sharma, E. Kanamaru, V. Borisenko, D.M. Desilva, A. Suzuki, X. Wang, C.J. Shawber, J.J. Kandel, D.J. Yamashiro, and J. Kitajewski. 2008. A Notch1 Ectodomain Construct Inhibits Endothelial Notch Signaling, Tumor Growth, and Angiogenesis. Cancer Res. 68:4727-4735. Gale, N.W., M.G. Dominguez, I. Noguera, L. Pan, V. Hughes, D.M. Valenzuela, A.J. Murphy, N.C. Adams, H.C. Lin, J. Holash, G. Thurston, and G.D. Yancopoulos. 2004. Haploinsufficiency of delta-like 4 ligand results in embryonic lethality due to major defects in arterial and vascular development. Proc Natl Acad Sci USA. 101:15949-15954. Gallahan, D., and R. Callahan. 1997. The mouse mammary tumor associated gene INT3 is a unique member of the NOTCH gene family (NOTCH4). Oncogene. 14:1883-1890. Gallahan, D., C. Jhappan, G. Robinson, L. Hennighausen, R. Sharp, E. Kordon, R. Callahan, G. Merlino, and G.H. Smith. 1996. Expression of a truncated Int3 gene in developing secretory mammary epithelium specifically retards lobular differentiation resulting in tumorigenesis. Cancer Res. 56:1775-1785. Gallahan, D., C. Kozak, and R. Callahan. 1987. A new common integration region (int-3) for mouse mammary tumor virus on mouse chromosome 17. J Virol. 61:218-220.
233 | Page
Gao, K., A. Masuda, T. Matsuura, and K. Ohno. 2008. Human branch point consensus sequence is yUnAy. Nucleic Acids Res. 36:2257-2267. Gazave, E., P. Lapebie, G.S. Richards, F. Brunet, A.V. Ereskovsky, B.M. Degnan, C. Borchiellini, M. Vervoort, and E. Renard. 2009. Origin and evolution of the Notch signalling pathway: an overview from eukaryotic genomes. BMC Evol Biol. 9:249. Gebb, S.A., and J.M. Shannon. 2000. Tissue interactions mediate early events in pulmonary vasculogenesis. Dev Dyn. 217:159-169. Geffers, I., K. Serth, G. Chapman, R. Jaekel, K. Schuster-Gossler, R. Cordes, D.B. Sparrow, E. Kremmer, S.L. Dunwoodie, T. Klein, and A. Gossler. 2007. Divergent functions and distinct localization of the Notch ligands DLL1 and DLL3 in vivo. J Cell Biol. 178:465-476. Gerhardt, H., M. Golding, M. Fruttiger, C. Ruhrberg, A. Lundkvist, A. Abramsson, M. Jeltsch, C. Mitchell, K. Alitalo, D. Shima, and C. Betsholtz. 2003. VEGF guides angiogenic sprouting utilizing endothelial tip cell filopodia. J Cell Biol. 161:1163- 1177. Glittenberg, M., C. Pitsouli, C. Garvey, C. Delidakis, and S. Bray. 2006. Role of conserved intracellular motifs in Serrate signalling, cis-inhibition and endocytosis. EMBO J. 25:4697- 4706. Goldie, L.C., J.L. Lucitti, M.E. Dickinson, and K.K. Hirschi. 2008. Cell signaling directing the formation and function of hemogenic endothelium during murine embryogenesis. Blood. 112:3194-3204. Gordon, W.R., K.L. Arnett, and S.C. Blacklow. 2008a. The molecular logic of Notch signaling - a structural and biochemical perspective. J Cell Sci. 121:3109-3119. Gordon, W.R., M. Roy, D. Vardar-Ulu, M. Garfinkel, M.R. Mansour, J.C. Aster, and S.C. Blacklow. 2008b. Structure of the Notch1 negative regulatory region: Implications for normal activation and pathogenic signaling in T-ALL. Blood:37. Gordon, W.R., D. Vardar-Ulu, G. Histen, C. Sanchez-Irizarry, J.C. Aster, and S.C. Blacklow. 2007. Structural basis for autoinhibition of Notch. Nat Struct Mol Biol. 14:295-300. Gordon, W.R., D. Vardar-Ulu, S. L'Heureux, T. Ashworth, M.J. Malecki, C. Sanchez-Irizarry, D.G. McArthur, G. Histen, J.L. Mitchell, J.C. Aster, and S.C. Blacklow. 2009. Effects of S1 cleavage on the structure, surface export, and signaling activity of human Notch1 and Notch2. PLoS ONE. 4:e6613. Graziani, I., S. Eliasz, M.A. De Marco, Y. Chen, H.I. Pass, R.M. De May, P.R. Strack, L. Miele, and M. Bocchetta. 2008. Opposite effects of Notch-1 and Notch-2 on mesothelioma cell survival under hypoxia are exerted through the Akt pathway. Cancer Res. 68:9678-9685. Greenwald, I. 1998. LIN-12/Notch signaling: lessons from worms and flies. Genes Dev. 12:1751-1762. Greenwald, I., and G. Seydoux. 1990. Analysis of gain-of-function mutations of the lin-12 gene of Caenorhabditis elegans. Nature. 346:197-199.
234 | Page
Grego-Bessa, J., L. Luna-Zurita, G. del Monte, V. Bolos, P. Melgar, A. Arandilla, A.N. Garratt, H. Zang, Y.S. Mukouyama, H. Chen, W. Shou, E. Ballestar, M. Esteller, A. Rojas, J.M. Perez-Pomares, and J.L. de la Pompa. 2007. Notch signaling is essential for ventricular chamber development. Dev Cell. 12:415-429. Gupta-Rossi, N., E. Six, O. LeBail, F. Logeat, P. Chastagner, A. Olry, A. Israël, and C. Brou. 2004. Monoubiquitination and endocytosis direct gamma-secretase cleavage of activated Notch receptor. J Cell Biol. 166:73-83. Guruharsha, K.G., M.W. Kankel, and S. Artavanis-Tsakonas. 2012. The Notch signalling system: recent insights into the complexity of a conserved pathway. Nat Rev Genet. Gustafsson, M.V., X. Zheng, T. Pereira, K. Gradin, S. Jin, J. Lundkvist, J.L. Ruas, L. Poellinger, U. Lendahl, and M. Bondesson. 2005. Hypoxia requires notch signaling to maintain the undifferentiated cell state. Dev Cell. 9:617-628. Haapasalo, A., and D.M. Kovacs. 2011. The many substrates of presenilin/gamma-secretase. Journal of Alzheimer's disease : JAD. 25:3-28. Haar, J.L., and G.A. Ackerman. 1971. A phase and electron microscopic study of vasculogenesis and erythropoiesis in the yolk sac of the mouse. The Anatomical record. 170:199- 223. Hackett, S.F., H. Ozaki, R.W. Strauss, K. Wahlin, C. Suri, P. Maisonpierre, G. Yancopoulos, and P.A. Campochiaro. 2000. Angiopoietin 2 expression in the retina: upregulation during physiologic and pathologic neovascularization. J Cell Physiol. 184:275-284. Hackett, S.F., S. Wiegand, G. Yancopoulos, and P.A. Campochiaro. 2002. Angiopoietin-2 plays an important role in retinal angiogenesis. J Cell Physiol. 192:182-187. Hamada, Y., Y. Kadokawa, M. Okabe, M. Ikawa, J.R. Coleman, and Y. Tsujimoto. 1999. Mutation in ankyrin repeats of the mouse Notch2 gene induces early embryonic lethality. Development. 126:3415-3424. Hambleton, S., N.V. Valeyev, A. Muranyi, V. Knott, J.M. Werner, A.J. McMichael, P.A. Handford, and A.K. Downing. 2004. Structural and functional properties of the human notch-1 ligand binding region. Structure. 12:2173-2183. Han, J., J. Allalunis-Turner, and M.J. Hendzel. 2011. Characterization and comparison of protein complexes initiated by the intracellular domain of individual Notch paralogs. Biochem Biophys Res Commun. 407:479-485. Hao, P. 2009. Characterization of the Ca2+ Binding Affinity and Coordination Site of the LIN-12/Notch-Repeat. PhD Thesis, Wellesley College. Harris, R.J., and M.W. Spellman. 1993. O-linked fucose and other post-translational modifications unique to EGF modules. Glycobiology. 3:219-224. Hayward, P., K. Brennan, P. Sanders, T. Balayo, R. DasGupta, N. Perrimon, and A. Martinez Arias. 2005. Notch modulates Wnt signalling by associating with Armadillo/beta-catenin and
235 | Page
regulating its transcriptional activity. Development. 132:1819-1830. Heinke, J., C. Patterson, and M. Moser. 2012. Life is a pattern: vascular assembly within the embryo. Frontiers in bioscience. 4:2269-2288. Hellstrom, M., M. Kalen, P. Lindahl, A. Abramsson, and C. Betsholtz. 1999. Role of PDGF-B and PDGFR-beta in recruitment of vascular smooth muscle cells and pericytes during embryonic blood vessel formation in the mouse. Development. 126:3047-3055. Hellström, M., L.-K. Phng, J.J. Hofmann, E. Wallgard, L. Coultas, P. Lindblom, J. Alva, A.-K. Nilsson, L. Karlsson, N. Gaiano, K. Yoon, J. Rossant, M.L. Iruela-Arispe, M. Kalén, H. Gerhardt, and C. Betsholtz. 2007. Dll4 signalling through Notch1 regulates formation of tip cells during angiogenesis. Nature. 445:776-780. Henderson, V.J., R.G. Cohen, R.S. Mitchell, J.C. Kosek, and D.C. Miller. 1986. Biochemical (functional) adaptation of "arterialized" vein grafts. Annals of surgery. 203:339-345. Henrich, S., I. Lindberg, W. Bode, and M.E. Than. 2005. Proprotein convertase models based on the crystal structures of furin and kexin: explanation of their specificity. J Mol Biol. 345:211-227. Henrique, D., E. Hirsinger, J. Adam, I. Le Roux, O. Pourquie, D. Ish-Horowicz, and J. Lewis. 1997. Maintenance of neuroepithelial progenitor cells by Delta-Notch signalling in the embryonic chick retina. Curr Biol. 7:661-670. Hicks, C., S.H. Johnston, G. diSibio, A. Collazo, T.F. Vogt, and G. Weinmaster. 2000. Fringe differentially modulates Jagged1 and Delta1 signalling through Notch1 and Notch2. Nat Cell Biol. 2:515-520. Hicks, C., E. Ladi, C. Lindsell, J.J. Hsieh, S.D. Hayward, A. Collazo, and G. Weinmaster. 2002. A secreted Delta1-Fc fusion protein functions both as an activator and inhibitor of Notch1 signaling. J Neurosci Res. 68:655-667. High, F.A., M.M. Lu, W.S. Pear, K.M. Loomes, K.H. Kaestner, and J.A. Epstein. 2008. Endothelial expression of the Notch ligand Jagged1 is required for vascular smooth muscle development. Proc Natl Acad Sci U S A. 105:1955-1959. Hirota, K., and G.L. Semenza. 2005. Regulation of hypoxia- inducible factor 1 by prolyl and asparaginyl hydroxylases. Biochem Biophys Res Commun. 338:610-616. Hofmann, J., and M. Luisairuelaarispe. 2007. Notch expression patterns in the retina: An eye on receptor–ligand distribution during angiogenesis. Gene Expression Patterns. 7:461-470. Hofmann, J.J., and M.L. Iruela-Arispe. 2007. Notch signaling in blood vessels: who is talking to whom about what? Circ Res. 100:1556-1568. Hogan, B., R. Beddington, F. Costantini, and E. Lacey. 1994. Manipulating the mouse embryo; A laboratory manual. Cold Spring Harbour laboratory Press. Hong, C.C., T. Kume, and R.T. Peterson. 2008. Role of crosstalk between phosphatidylinositol 3-kinase and extracellular
236 | Page
signal-regulated kinase/mitogen-activated protein kinase pathways in artery-vein specification. Circ Res. 103:573-579. Hoyne, G.F., G. Chapman, Y. Sontani, S.E. Pursglove, and S.L. Dunwoodie. 2011. A cell autonomous role for the Notch ligand Delta-like 3 in alphabeta T-cell development. Immunology and cell biology. 89:696-705. Hu, Q.D., X.Y. Cui, Y.K. Ng, and Z.C. Xiao. 2004. Axoglial interaction via the notch receptor in oligodendrocyte differentiation. Annals of the Academy of Medicine, Singapore. 33:581-588. Huang, L.H., X.H. Ke, W. Sweeney, and J.P. Tam. 1989. Calcium binding and putative activity of the epidermal growth factor domain of blood coagulation factor IX. Biochem Biophys Res Commun. 160:133-139. Huber, T.L., V. Kouskoff, H.J. Fehling, J. Palis, and G. Keller. 2004. Haemangioblast commitment is initiated in the primitive streak of the mouse embryo. Nature. 432:625-630. Hughes, S., and T. Chang-Ling. 2000. Roles of endothelial cell migration and apoptosis in vascular remodeling during development of the central nervous system. Microcirculation. 7:317-333. Huxlin, K.R., A.J. Sefton, and J.H. Furby. 1992. The origin and development of retinal astrocytes in the mouse. J Neurocytol. 21:530-544. Imatani, A., and R. Callahan. 2000. Identification of a novel NOTCH-4/INT-3 RNA species encoding an activated gene product in certain human tumor cell lines. Oncogene. 19:223- 231. Iruela-Arispe, M.L., and G.E. Davis. 2009. Cellular and molecular mechanisms of vascular lumen formation. Dev Cell. 16:222- 231. Irvine, K.D. 1999. Fringe, Notch, and making developmental boundaries. Curr Opin Genet Dev. 9:434-441. Ishida, S., K. Yamashiro, T. Usui, Y. Kaji, Y. Ogura, T. Hida, Y. Honda, Y. Oguchi, and A.P. Adamis. 2003. Leukocytes mediate retinal vascular remodeling during development and vaso-obliteration in disease. Nat Med. 9:781-788. Itoh, F., S. Itoh, M.J. Goumans, G. Valdimarsdottir, T. Iso, G.P. Dotto, Y. Hamamori, L. Kedes, M. Kato, and P. ten Dijke Pt. 2004. Synergy and antagonism between Notch and BMP receptor signaling pathways in endothelial cells. EMBO J. 23:541-551. Ivo, R., T.G. Schulze, J. Schumacher, K. Kesper, D.J. Muller, I. Kremer, M. Dobrusin, M. Mujaheed, I. Murad, M. Blanaru, I. Bannoura, A. Reshef, R. Bachner-Melman, R.P. Ebstein, P. Propping, R.H. Belmaker, W. Maier, M. Rietschel, M.M. Nothen, and S. Cichon. 2006. No evidence for association between NOTCH4 and schizophrenia in a large family-based and case-control association analysis. Psychiatric genetics. 16:197-203. Jafar-Nejad, H., J. Leonardi, and R. Fernandez-Valdivia. 2010. Role of glycans and glycosyltransferases in the regulation of Notch signaling. Glycobiology. 20:931-949.
237 | Page
Jarriault, S., C. Brou, F. Logeat, E.H. Schroeter, R. Kopan, and A. Israel. 1995. Signalling downstream of activated mammalian Notch. Nature. 377:355-358. Jarriault, S., O. Le Bail, E. Hirsinger, O. Pourquie, F. Logeat, C.F. Strong, C. Brou, N.G. Seidah, and A. Isra l. 1998. Delta-1 activation of notch-1 signaling results in HES-1 transactivation. Mol Cell Biol. 18:7423-7431. Jehn, B.M., I. Dittert, S. Beyer, K. von der Mark, and W. Bielke. 2002. c-Cbl binding and ubiquitin-dependent lysosomal degradation of membrane-associated Notch1. J Biol Chem. 277:8033-8040. Jhappan, C., D. Gallahan, C. Stahle, E. Chu, G.H. Smith, G. Merlino, and R. Callahan. 1992. Expression of an activated Notch-related int-3 transgene interferes with cell differentiation and induces neoplastic transformation in mammary and salivary glands. Genes Dev. 6:345-355. Jiang, Y.J., B.L. Aerne, L. Smithers, C. Haddon, D. Ish-Horowicz, and J. Lewis. 2000. Notch signalling and the synchronization of the somite segmentation clock. Nature. 408:475-479. Jin, S.W., D. Beis, T. Mitchell, J.N. Chen, and D.Y. Stainier. 2005. Cellular and molecular analyses of vascular tube and lumen formation in zebrafish. Development. 132:5199-5209. Jin, S.W., and C. Patterson. 2009. The opening act: vasculogenesis and the origins of circulation. Arterioscler Thromb Vasc Biol. 29:623-629. Johnson, N.C., M.E. Dillard, P. Baluk, D.M. McDonald, N.L. Harvey, S.L. Frase, and G. Oliver. 2008. Lymphatic endothelial cell identity is reversible and its maintenance requires Prox1 activity. Genes Dev. 22:3282-3291. Joutel, A., F. Andreux, S. Gaulis, V. Domenga, M. Cecillon, N. Battail, N. Piga, F. Chapon, C. Godfrain, and E. Tournier- Lasserve. 2000. The ectodomain of the Notch3 receptor accumulates within the cerebrovasculature of CADASIL patients. J Clin Invest. 105:597-605. Joutel, A., C. Corpechot, A. Ducros, K. Vahedi, H. Chabriat, P. Mouton, S. Alamowitch, V. Domenga, M. Cecillion, E. Marechal, J. Maciazek, C. Vayssiere, C. Cruaud, E.A. Cabanis, M.M. Ruchoux, J. Weissenbach, J.F. Bach, M.G. Bousser, and E. Tournier-Lasserve. 1996. Notch3 mutations in CADASIL, a hereditary adult-onset condition causing stroke and dementia. Nature. 383:707-710. Kageyama, R., T. Ohtsuka, and T. Kobayashi. 2007. The Hes gene family: repressors and oscillators that orchestrate embryogenesis. Development. 134:1243-1251. Kamei, M., W.B. Saunders, K.J. Bayless, L. Dye, G.E. Davis, and B.M. Weinstein. 2006. Endothelial tubes assemble from intracellular vacuoles in vivo. Nature. 442:453-456. Karkkainen, M.J., P. Haiko, K. Sainio, J. Partanen, J. Taipale, T.V. Petrova, M. Jeltsch, D.G. Jackson, M. Talikka, H. Rauvala, C. Betsholtz, and K. Alitalo. 2004. Vascular endothelial growth factor C is required for sprouting of the first lymphatic vessels from embryonic veins. Nature immunology. 5:74-80.
238 | Page
Karlstrom, H., A. Bergman, U. Lendahl, J. Naslund, and J. Lundkvist. 2002. A sensitive and quantitative assay for measuring cleavage of presenilin substrates. J Biol Chem. 277:6763-6766. Kato, H., T. Sakai, K. Tamura, S. Minoguchi, Y. Shirayoshi, Y. Hamada, Y. Tsujimoto, and T. Honjo. 1996. Functional conservation of mouse Notch receptor family members. FEBS Lett. 395:221-224. Kelley, M.R., S. Kidd, W.A. Deutsch, and M.W. Young. 1987. Mutations altering the structure of epidermal growth factor- like coding sequences at the Drosophila Notch locus. Cell. 51:539-548. Kennard, Liu, and Lilly. 2007. TGF-beta 1 downregulates Notch3 in fibroblasts to promote smooth muscle gene expression. J Biol Chem. Kidd, S., M.K. Baylies, G.P. Gasic, and M.W. Young. 1989. Structure and distribution of the Notch protein in developing Drosophila. Genes Dev. 3:1113-1129. Kidd, S., and T. Lieber. 2002. Furin cleavage is not a requirement for Drosophila Notch function. Mech Dev. 115:41-51. Kimble, J., S. Henderson, and S. Crittenden. 1998. Notch/LIN-12 signaling: transduction by regulated protein slicing. Trends Biochem Sci. 23:353-357. Kopan, R., and M.X.G. Ilagan. 2009. The canonical Notch signaling pathway: unfolding the activation mechanism. Cell. 137:216- 233. Kopan, R., J.S. Nye, and H. Weintraub. 1994. The intracellular domain of mouse Notch: a constitutively activated repressor of myogenesis directed at the basic helix-loop-helix region of MyoD. Development. 120:2385-2396. Kopan, R., E.H. Schroeter, H. Weintraub, and J.S. Nye. 1996. Signal transduction by activated mNotch: importance of proteolytic processing and its regulation by the extracellular domain. Proc Natl Acad Sci U S A. 93:1683-1688. Kortschak, R.D., R. Tamme, and M. Lardelli. 2001. Evolutionary analysis of vertebrate Notch genes. Dev Genes Evol. 211:350-354. Kos, M., S. Denger, G. Reid, K.S. Korach, and F. Gannon. 2002. Down but not out? A novel protein isoform of the estrogen receptor alpha is expressed in the estrogen receptor alpha knockout mouse. Journal of molecular endocrinology. 29:281- 286. Kovall, R.A., and W.A. Hendrickson. 2004. Crystal structure of the nuclear effector of Notch signaling, CSL, bound to DNA. EMBO J. 23:3441-3451. Krebs, L.T., J.R. Shutter, K. Tanigaki, T. Honjo, K.L. Stark, and T. Gridley. 2004. Haploinsufficient lethality and formation of arteriovenous malformations in Notch pathway mutants. Genes Dev. 18:2469-2473. Krebs, L.T., C. Starling, A.V. Chervonsky, and T. Gridley. 2010. Notch1 activation in mice causes arteriovenous malformations phenocopied by ephrinB2 and EphB4 mutants. Genesis. 48:146-150.
239 | Page
Krebs, L.T., Y. Xue, C.R. Norton, J.R. Shutter, M. Maguire, J.P. Sundberg, D. Gallahan, V. Closson, J. Kitajewski, R. Callahan, G.H. Smith, K.L. Stark, and T. Gridley. 2000. Notch signaling is essential for vascular morphogenesis in mice. Genes Dev. 14:1343-1352. Krejci, A., and S. Bray. 2007. Notch activation stimulates transient and selective binding of Su(H)/CSL to target enhancers. Genes Dev. 21:1322-1327. Kuhnert, F., M.R. Mancuso, J. Hampton, K. Stankunas, T. Asano, C.Z. Chen, and C.J. Kuo. 2008. Attribution of vascular phenotypes of the murine Egfl7 locus to the microRNA miR- 126. Development. 135:3989-3993. Kullander, K., and R. Klein. 2002. Mechanisms and functions of Eph and ephrin signalling. Nat Rev Mol Cell Biol. 3:475-486. Kumano, K., S. Chiba, A. Kunisato, M. Sata, T. Saito, E. Nakagami-Yamaguchi, T. Yamaguchi, S. Masuda, K. Shimizu, T. Takahashi, S. Ogawa, Y. Hamada, and H. Hirai. 2003. Notch1 but not Notch2 is essential for generating hematopoietic stem cells from endothelial cells. Immunity. 18:699-711. Kumar, S., and S. Subramanian. 2002. Mutation rates in mammalian genomes. Proc Natl Acad Sci U S A. 99:803-808. Kume, T. 2010. Specification of arterial, venous, and lymphatic endothelial cells during embryonic development. Histology and histopathology. 25:637-646. Kurooka, H., and T. Honjo. 2000. Functional interaction between the mouse notch1 intracellular region and histone acetyltransferases PCAF and GCN5. J Biol Chem. 275:17211-17220. Kurooka, H., K. Kuroda, and T. Honjo. 1998. Roles of the ankyrin repeats and C-terminal region of the mouse notch1 intracellular region. Nucleic Acids Res. 26:5448-5455. Laborda, J., E.A. Sausville, T. Hoffman, and V. Notario. 1993. dlk, a putative mammalian homeotic gene differentially expressed in small cell lung carcinoma and neuroendocrine tumor cell line. J Biol Chem. 268:3817-3820. Ladi, E., J.T. Nichols, W. Ge, A. Miyamoto, C. Yao, L.T. Yang, J. Boulter, Y.E. Sun, C. Kintner, and G. Weinmaster. 2005. The divergent DSL ligand Dll3 does not activate Notch signaling but cell autonomously attenuates signaling induced by other DSL ligands. J Cell Biol. 170:983-992. Lai, E.C. 2002. Keeping a good pathway down: transcriptional repression of Notch pathway target genes by CSL proteins. EMBO Rep. 3:840-845. Lai, E.C. 2004. Notch signaling: control of cell communication and cell fate. Development. 131:965-973. Lake, R.J., L.M. Grimm, A. Veraksa, A. Banos, and S. Artavanis- Tsakonas. 2009. In vivo analysis of the Notch receptor S1 cleavage. PLoS ONE. 4:e6728. Lam, L.T., and E.H. Bresnick. 1998. Identity of the beta-globin locus control region binding protein HS2NF5 as the mammalian homolog of the notch-regulated transcription factor suppressor of hairless. J Biol Chem. 273:24223-24231.
240 | Page
Lammert, E., O. Cleaver, and D. Melton. 2001. Induction of pancreatic differentiation by signals from blood vessels. Science. 294:564-567. Lardelli, M., J. Dahlstrand, and U. Lendahl. 1994. The novel Notch homologue mouse Notch 3 lacks specific epidermal growth factor-repeats and is expressed in proliferating neuroepithelium. Mech Dev. 46:123-136. LaRue, A.C., R. Lansford, and C.J. Drake. 2003. Circulating blood island-derived cells contribute to vasculogenesis in the embryo proper. Dev Biol. 262:162-172. Lawrence, N., T. Klein, K. Brennan, and A. Martinez Arias. 2000. Structural requirements for notch signalling with delta and serrate during the development and patterning of the wing disc of Drosophila. Development. 127:3185-3195. Lee, S., J. Kang, J. Yoo, S.K. Ganesan, S.C. Cook, B. Aguilar, S. Ramu, J. Lee, and Y.K. Hong. 2009. Prox1 physically and functionally interacts with COUP-TFII to specify lymphatic endothelial cell fate. Blood. 113:1856-1859. Lei, L., A. Xu, V.M. Panin, and K.D. Irvine. 2003. An O-fucose site in the ligand binding domain inhibits Notch activation. Development. 130:6411-6421. Lennon, G., C. Auffray, M. Polymeropoulos, and M.B. Soares. 1996. The I.M.A.G.E. Consortium: an integrated molecular analysis of genomes and their expression. Genomics. 33:151-152. Leslie, J.D., L. Ariza-McNaughton, A.L. Bermange, R. McAdow, S.L. Johnson, and J. Lewis. 2007. Endothelial signalling by the Notch ligand Delta-like 4 restricts angiogenesis. Development. 134:839-844. Li, L., G.M. Huang, A.B. Banta, Y. Deng, T. Smith, P. Dong, C. Friedman, L. Chen, B.J. Trask, T. Spies, L. Rowen, and L. Hood. 1998. Cloning, characterization, and the complete 56.8-kilobase DNA sequence of the human NOTCH4 gene. Genomics. 51:45-58. Lieber, T., S. Kidd, and M.W. Young. 2002. kuzbanian-mediated cleavage of Drosophila Notch. Genes Dev. 16:209-221. Limbourg, A., M. Ploom, D. Elligsen, I. Sorensen, T. Ziegelhoeffer, A. Gossler, H. Drexler, and F.P. Limbourg. 2007. Notch ligand Delta-like 1 is essential for postnatal arteriogenesis. Circ Res. 100:363-371. Lin, F.J., X. Chen, J. Qin, Y.K. Hong, M.J. Tsai, and S.Y. Tsai. 2010. Direct transcriptional regulation of neuropilin-2 by COUP-TFII modulates multiple steps in murine lymphatic vessel development. J Clin Invest. 120:1694-1707. Lindsell, C.E., C.J. Shawber, J. Boulter, and G. Weinmaster. 1995. Jagged: a mammalian ligand that activates Notch1. Cell. 80:909-917. Liu, H., A.W. Chi, K.L. Arnett, M.Y. Chiang, L. Xu, O. Shestova, H. Wang, Y.M. Li, A. Bhandoola, J.C. Aster, S.C. Blacklow, and W.S. Pear. 2010a. Notch dimerization is required for leukemogenesis and T-cell development. Genes Dev. 24:2395-2407. Liu, H., S. Kennard, and B. Lilly. 2009. NOTCH3 Expression Is Induced in Mural Cells Through an Autoregulatory Loop That
241 | Page
Requires Endothelial-Expressed JAGGED1. Circ Res. 104:466-475. Liu, H., W. Zhang, S. Kennard, R.B. Caldwell, and B. Lilly. 2010b. Notch3 is critical for proper angiogenesis and mural cell investment. Circ Res. 107:860-870. Liu, Z.J., T. Shirakawa, Y. Li, A. Soma, M. Oka, G.P. Dotto, R.M. Fairman, O.C. Velazquez, and M. Herlyn. 2003. Regulation of Notch1 and Dll4 by vascular endothelial growth factor in arterial endothelial cells: implications for modulating arteriogenesis and angiogenesis. Mol Cell Biol. 23:14-25. Lobov, I.B., R.A. Renard, N. Papadopoulos, N.W. Gale, G. Thurston, G.D. Yancopoulos, and S.J. Wiegand. 2007. Delta- like ligand 4 (Dll4) is induced by VEGF as a negative regulator of angiogenic sprouting. Proc Natl Acad Sci USA. 104:3219-3224. Logeat, F., C. Bessia, C. Brou, O. LeBail, S. Jarriault, N.G. Seidah, and A. Israel. 1998. The Notch1 receptor is cleaved constitutively by a furin-like convertase. Proc Natl Acad Sci U S A. 95:8108-8112. Lowell, S., P. Jones, I. Le Roux, J. Dunne, and F.M. Watt. 2000. Stimulation of human epidermal differentiation by delta-notch signalling at the boundaries of stem-cell clusters. Curr Biol. 10:491-500. Lowell, S., and F.M. Watt. 2001. Delta regulates keratinocyte spreading and motility independently of differentiation. Mech Dev. 107:133-140. Lu, H., and D. Bilder. 2005. Endocytic control of epithelial polarity and proliferation in Drosophila. Nat Cell Biol. 7:1232-1239. Lubman, O.Y., M.X.G. Ilagan, R. Kopan, and D. Barrick. 2007. Quantitative dissection of the Notch:CSL interaction: insights into the Notch-mediated transcriptional switch. J Mol Biol. 365:577-589. Luetteke, N.C., T.H. Qiu, R.L. Peiffer, P. Oliver, O. Smithies, and D.C. Lee. 1993. TGF alpha deficiency results in hair follicle and eye abnormalities in targeted and waved-1 mice. Cell. 73:263-278. Luhmann, U.F., J. Lin, N. Acar, S. Lammel, S. Feil, C. Grimm, M.W. Seeliger, H.P. Hammes, and W. Berger. 2005. Role of the Norrie disease pseudoglioma gene in sprouting angiogenesis during development of the retinal vasculature. Invest Ophthalmol Vis Sci. 46:3372-3382. Luo, Z.X., C.X. Yuan, Q.J. Meng, and Q. Ji. 2011. A Jurassic eutherian mammal and divergence of marsupials and placentals. Nature. 476:442-445. Ma, D., and L.Y. Jan. 2002. ER transport signals and trafficking of potassium channels and receptors. Curr Opin Neurobiol. 12:287-292. Ma, Y.C., C. Shi, Y.N. Zhang, L.G. Wang, H. Liu, H.T. Jia, Y.X. Zhang, F.H. Sarkar, and Z.S. Wang. 2012. The tyrosine kinase c-Src directly mediates growth factor-induced Notch-1 and Furin interaction and Notch-1 activation in pancreatic cancer cells. PLoS One. 7:e33414.
242 | Page
Maier, M.M., and M. Gessler. 2000. Comparative analysis of the human and mouse Hey1 promoter: Hey genes are new Notch target genes. Biochem Biophys Res Commun. 275:652-660. Maisonpierre, P.C., C. Suri, P.F. Jones, S. Bartunkova, S.J. Wiegand, C. Radziejewski, D. Compton, J. McClain, T.H. Aldrich, N. Papadopoulos, T.J. Daly, S. Davis, T.N. Sato, and G.D. Yancopoulos. 1997. Angiopoietin-2, a natural antagonist for Tie2 that disrupts in vivo angiogenesis. Science. 277:55-60. Maquat, L.E. 2004. Nonsense-mediated mRNA decay: splicing, translation and mRNP dynamics. Nat Rev Mol Cell Biol. 5:89- 99. Matesic, L.E., D.C. Haines, N.G. Copeland, and N.A. Jenkins. 2006. Itch genetically interacts with Notch1 in a mouse autoimmune disease model. Hum Mol Genet. 15:3485-3497. Matsumoto, K., H. Yoshitomi, J. Rossant, and K.S. Zaret. 2001. Liver organogenesis promoted by endothelial cells prior to vascular function. Science. 294:559-563. McCright, B., X. Gao, L. Shen, J. Lozier, Y. Lan, M. Maguire, D. Herzlinger, G. Weinmaster, R. Jiang, and T. Gridley. 2001. Defects in development of the kidney, heart and eye vasculature in mice homozygous for a hypomorphic Notch2 mutation. Development. 128:491-502. Mcgill, M.A., S.E. Dho, G. Weinmaster, and C.J. Mcglade. 2009. Numb regulates post-endocytic trafficking and degradation of notch1. Journal of Biological Chemistry:1-25. McGill, M.A., and C.J. McGlade. 2003. Mammalian numb proteins promote Notch1 receptor ubiquitination and degradation of the Notch1 intracellular domain. J Biol Chem. 278:23196- 23203. McGinnis, R.E., H. Fox, P. Yates, L.A. Cameron, M.R. Barnes, I.C. Gray, N.K. Spurr, O. Hurko, and D. St Clair. 2001. Failure to confirm NOTCH4 association with schizophrenia in a large population-based sample from Scotland. Nat Genet. 28:128- 129. Meloty-Kapella, L., B. Shergill, J. Kuon, E. Botvinick, and G. Weinmaster. 2012. Notch ligand endocytosis generates mechanical pulling force dependent on dynamin, epsins, and actin. Dev Cell. 22:1299-1312. Meng, H., X. Zhang, K.D. Hankenson, and M.M. Wang. 2009. Thrombospondin 2 potentiates notch3/jagged1 signaling. J Biol Chem. 284:7866-7874. Meraz, M.A., J.M. White, K.C. Sheehan, E.A. Bach, S.J. Rodig, A.S. Dighe, D.H. Kaplan, J.K. Riley, A.C. Greenlund, D. Campbell, K. Carver-Moore, R.N. DuBois, R. Clark, M. Aguet, and R.D. Schreiber. 1996. Targeted disruption of the Stat1 gene in mice reveals unexpected physiologic specificity in the JAK-STAT signaling pathway. Cell. 84:431-442. Micchelli, C.A., E.J. Rulifson, and S.S. Blair. 1997. The function and regulation of cut expression on the wing margin of Drosophila: Notch, Wingless and a dominant negative role for Delta and Serrate. Development. 124:1485-1495. Minamizato, T., K. Sakamoto, T. Liu, H. Kokubo, K. Katsube, B. Perbal, S. Nakamura, and A. Yamaguchi. 2007. CCN3/NOV
243 | Page
inhibits BMP-2-induced osteoblast differentiation by interacting with BMP and Notch signaling pathways. Biochem Biophys Res Commun. 354:567-573. Miniati, D., E.B. Jelin, J. Ng, J. Wu, T.R. Carlson, X. Wu, M.R. Looney, and R.A. Wang. 2009. Constitutively active endothelial Notch4 causes lung arteriovenous shunts in mice. Am J Physiol Lung Cell Mol Physiol. Miyahara, M., H. Nakanishi, K. Takahashi, K. Satoh-Horikawa, K. Tachibana, and Y. Takai. 2000. Interaction of nectin with afadin is necessary for its clustering at cell-cell contact sites but not for its cis dimerization or trans interaction. J Biol Chem. 275:613-618. Miyamoto, A., R. Lau, P.W. Hein, J.M. Shipley, and G. Weinmaster. 2006. Microfibrillar proteins MAGP-1 and MAGP-2 induce Notch1 extracellular domain dissociation and receptor activation. J Biol Chem. 281:10089-10097. Mizutani, T., Y. Taniguchi, T. Aoki, N. Hashimoto, and T. Honjo. 2001. Conservation of the biochemical mechanisms of signal transduction among mammalian Notch family members. Proc Natl Acad Sci U S A. 98:9026-9031. Moens, C.B., A.B. Auerbach, R.A. Conlon, A.L. Joyner, and J. Rossant. 1992. A targeted mutation reveals a role for N-myc in branching morphogenesis in the embryonic mouse lung. Genes Dev. 6:691-704. Moloney, D.J., V.M. Panin, S.H. Johnston, J. Chen, L. Shao, R. Wilson, Y. Wang, P. Stanley, K.D. Irvine, R.S. Haltiwanger, and T.F. Vogt. 2000. Fringe is a glycosyltransferase that modifies Notch. Nature. 406:369-375. Moran, J.L., J.M. Levorse, and T.F. Vogt. 1999. Limbs move beyond the radical fringe. Nature. 399:742-743. Moran, J.L., E.T. Shifley, J.M. Levorse, S. Mani, K. Ostmann, A. Perez-Balaguer, D.M. Walker, T.F. Vogt, and S.E. Cole. 2009. Manic fringe is not required for embryonic development, and fringe family members do not exhibit redundant functions in the axial skeleton, limb, or hindbrain. Dev Dyn. 238:1803-1812. Morgan, T.H. 1917. The Theory Of The Gene. The American Naturlist. 51:513-544. Morimoto, M., Y. Takahashi, M. Endo, and Y. Saga. 2005. The Mesp2 transcription factor establishes segmental borders by suppressing Notch activity. Nature. 435:354-359. Mouse Genome Sequencing, C., R.H. Waterston, K. Lindblad-Toh, E. Birney, J. Rogers, J.F. Abril, P. Agarwal, R. Agarwala, R. Ainscough, M. Alexandersson, P. An, S.E. Antonarakis, J. Attwood, R. Baertsch, J. Bailey, K. Barlow, S. Beck, E. Berry, B. Birren, T. Bloom, P. Bork, M. Botcherby, N. Bray, M.R. Brent, D.G. Brown, S.D. Brown, C. Bult, J. Burton, J. Butler, R.D. Campbell, P. Carninci, S. Cawley, F. Chiaromonte, A.T. Chinwalla, D.M. Church, M. Clamp, C. Clee, F.S. Collins, L.L. Cook, R.R. Copley, A. Coulson, O. Couronne, J. Cuff, V. Curwen, T. Cutts, M. Daly, R. David, J. Davies, K.D. Delehaunty, J. Deri, E.T. Dermitzakis, C. Dewey, N.J. Dickens, M. Diekhans, S. Dodge, I. Dubchak, D.M. Dunn, S.R. Eddy, L. Elnitski, R.D. Emes, P. Eswara, E.
244 | Page
Eyras, A. Felsenfeld, G.A. Fewell, P. Flicek, K. Foley, W.N. Frankel, L.A. Fulton, R.S. Fulton, T.S. Furey, D. Gage, R.A. Gibbs, G. Glusman, S. Gnerre, N. Goldman, L. Goodstadt, D. Grafham, T.A. Graves, E.D. Green, S. Gregory, R. Guigo, M. Guyer, R.C. Hardison, D. Haussler, Y. Hayashizaki, L.W. Hillier, A. Hinrichs, W. Hlavina, T. Holzer, F. Hsu, A. Hua, T. Hubbard, A. Hunt, I. Jackson, D.B. Jaffe, L.S. Johnson, M. Jones, T.A. Jones, A. Joy, M. Kamal, et al. 2002. Initial sequencing and comparative analysis of the mouse genome. Nature. 420:520-562. Mumm, J.S., E.H. Schroeter, M.T. Saxena, A. Griesemer, X. Tian, D.J. Pan, W.J. Ray, and R. Kopan. 2000. A ligand-induced extracellular cleavage regulates gamma-secretase-like proteolytic activation of Notch1. Mol Cell. 5:197-206. Murphy, P.A., M.T.Y. Lam, X. Wu, T.N. Kim, S.M. Vartanian, A.W. Bollen, T.R. Carlson, and R.A. Wang. 2008. Endothelial Notch4 signaling induces hallmarks of brain arteriovenous malformations in mice. Proc Natl Acad Sci USA. 105:10901- 10906. Nakayama, K. 1997. Furin: a mammalian subtilisin/Kex2p-like endoprotease involved in processing of a wide variety of precursor proteins. Biochem J. 327 ( Pt 3):625-635. Nam, Y., P. Sliz, W.S. Pear, J.C. Aster, and S.C. Blacklow. 2007. Cooperative assembly of higher-order Notch complexes functions as a switch to induce transcription. Proc Natl Acad Sci USA. 104:2103-2108. Nam, Y., A.P. Weng, J.C. Aster, and S.C. Blacklow. 2003. Structural requirements for assembly of the CSL.intracellular Notch1.Mastermind-like 1 transcriptional activation complex. J Biol Chem. 278:21232-21239. Ng, Y.S., R. Rohan, M.E. Sunday, D.E. Demello, and P.A. D'Amore. 2001. Differential expression of VEGF isoforms in mouse during development and in the adult. Dev Dyn. 220:112-121. Nichol, D., C. Shawber, M.J. Fitch, K. Bambino, A. Sharma, J. Kitajewski, and H. Stuhlmann. 2010. Impaired angiogenesis and altered Notch signaling in mice overexpressing endothelial Egfl7. Blood. 116:6133-6143. Nichol, D., and H. Stuhlmann. 2012. EGFL7: a unique angiogenic signaling factor in vascular development and disease. Blood. 119:1345-1352. Nichols, J.T., A. Miyamoto, S.L. Olsen, B. D'Souza, C. Yao, and G. Weinmaster. 2007. DSL ligand endocytosis physically dissociates Notch1 heterodimers before activating proteolysis can occur. J Cell Biol. 176:445-458. Nickel, W., and C. Rabouille. 2009. Mechanisms of regulated unconventional protein secretion. Nat Rev Mol Cell Biol. 10:148-155. Nimmagadda, S., P. Geetha Loganathan, R. Huang, M. Scaal, C. Schmidt, and B. Christ. 2005. BMP4 and noggin control embryonic blood vessel formation by antagonistic regulation of VEGFR-2 (Quek1) expression. Dev Biol. 280:100-110.
245 | Page
Nishishita, T., and P.C. Lin. 2004. Angiopoietin 1, PDGF-B, and TGF-beta gene regulation in endothelial cell and smooth muscle cell interaction. J Cell Biochem. 91:584-593. Nishiyama, T., K. Mishima, F. Ide, K. Yamada, K. Obara, A. Sato, N. Hitosugi, H. Inoue, K. Tsubota, and I. Saito. 2007. Functional Analysis of an Established Mouse Vascular Endothelial Cell Line. J Vasc Res. 44:138-148. Niwa, Y., Y. Masamizu, T. Liu, R. Nakayama, C.-X. Deng, and R. Kageyama. 2007. The initiation and propagation of Hes7 oscillation are cooperatively regulated by Fgf and notch signaling in the somite segmentation clock. Dev Cell. 13:298- 304. Noguera-Troise, I., C. Daly, N.J. Papadopoulos, S. Coetzee, P. Boland, N.W. Gale, H.C. Lin, G.D. Yancopoulos, and G. Thurston. 2006. Blockade of Dll4 inhibits tumour growth by promoting non-productive angiogenesis. Nature. 444:1032- 1037. Noseda, M., L. Chang, G. McLean, J.E. Grim, B.E. Clurman, L.L. Smith, and A. Karsan. 2004a. Notch activation induces endothelial cell cycle arrest and participates in contact inhibition: role of p21Cip1 repression. Mol Cell Biol. 24:8813-8822. Noseda, M., G. McLean, K. Niessen, L. Chang, I. Pollet, R. Montpetit, R. Shahidi, K. Dorovini-Zis, L. Li, B. Beckstead, R.E. Durand, P.A. Hoodless, and A. Karsan. 2004b. Notch activation results in phenotypic and functional changes consistent with endothelial-to-mesenchymal transformation. Circ Res. 94:910-917. Nueda, M.L., V. Baladron, B. Sanchez-Solana, M.A. Ballesteros, and J. Laborda. 2007. The EGF-like protein dlk1 inhibits notch signaling and potentiates adipogenesis of mesenchymal cells. J Mol Biol. 367:1281-1293. Nye, J.S., R. Kopan, and R. Axel. 1994. An activated Notch suppresses neurogenesis and myogenesis but not gliogenesis in mammalian cells. Development. 120:2421- 2430. Ohlmann, A., M. Scholz, A. Goldwich, B.K. Chauhan, K. Hudl, A.V. Ohlmann, E. Zrenner, W. Berger, A. Cvekl, M.W. Seeliger, and E.R. Tamm. 2005. Ectopic norrin induces growth of ocular capillaries and restores normal retinal angiogenesis in Norrie disease mutant mice. J Neurosci. 25:1701-1710. Okajima, T., and K.D. Irvine. 2002. Regulation of notch signaling by o-linked fucose. Cell. 111:893-904. Okajima, T., A. Xu, L. Lei, and K.D. Irvine. 2005. Chaperone activity of protein O-fucosyltransferase 1 promotes notch receptor folding. Science. 307:1599-1603. Okamura, Y., and Y. Saga. 2008. Pofut1 is required for the proper localization of the Notch receptor during mouse development. Mech Dev. 125:663-673. Ong, C.T., H.T. Cheng, L.W. Chang, T. Ohtsuka, R. Kageyama, G.D. Stormo, and R. Kopan. 2006. Target selectivity of vertebrate notch proteins. Collaboration between discrete domains and CSL-binding site architecture determines activation probability. J Biol Chem. 281:5106-5119.
246 | Page
Orkin, S.H. 1992. GATA-binding transcription factors in hematopoietic cells. Blood. 80:575-581. Osipo, C., T.E. Golde, B.A. Osborne, and L.A. Miele. 2008. Off the beaten pathway: the complex cross talk between Notch and NF-κ B. Lab Invest. 88:11-17. Osokine, I., R. Hsu, G.B. Loeb, and M.T. McManus. 2008. Unintentional miRNA ablation is a risk factor in gene knockout studies: a short report. PLoS genetics. 4:e34. Pan, D., and G.M. Rubin. 1997. Kuzbanian controls proteolytic processing of Notch and mediates lateral inhibition during Drosophila and vertebrate neurogenesis. Cell. 90:271-280. Panin, V.M., V. Papayannopoulos, R. Wilson, and K.D. Irvine. 1997. Fringe modulates Notch-ligand interactions. Nature. 387:908-912. Park, J.E., G.A. Keller, and N. Ferrara. 1993. The vascular endothelial growth factor (VEGF) isoforms: differential deposition into the subepithelial extracellular matrix and bioactivity of extracellular matrix-bound VEGF. Mol Biol Cell. 4:1317-1326. Parks, A.L., K.M. Klueg, J.R. Stout, and M.A. Muskavitch. 2000. Ligand endocytosis drives receptor dissociation and activation in the Notch pathway. Development. 127:1373- 1385. Parks, A.L., J.R. Stout, S.B. Shepard, K.M. Klueg, A.A. Dos Santos, T.R. Parody, M. Vaskova, and M.A. Muskavitch. 2006. Structure-function analysis of delta trafficking, receptor binding and signaling in Drosophila. Genetics. 174:1947-1961. Parr, C., G. Watkins, and W.G. Jiang. 2004. The possible correlation of Notch-1 and Notch-2 with clinical outcome and tumour clinicopathological parameters in human breast cancer. Int J Mol Med. 14:779-786. Patel-Hett, S., and P.A. D'Amore. 2011. Signal transduction in vasculogenesis and developmental angiogenesis. Int J Dev Biol. 55:353-363. Pei, Z., and N.E. Baker. 2008. Competition between Delta and the Abruptex domain of Notch. BMC developmental biology. 8:4. Penton, A.L., L.D. Leonard, and N.B. Spinner. 2012. Notch signaling in human development and disease. Semin Cell Dev Biol. 23:450-457. Perez-Pomares, J.M., D. Macias, L. Garcia-Garrido, and R. Munoz- Chapuli. 1998. The origin of the subepicardial mesenchyme in the avian embryo: an immunohistochemical and quail-chick chimera study. Dev Biol. 200:57-68. Perez, L., M. Milan, S. Bray, and S.M. Cohen. 2005. Ligand- binding and signaling properties of the Ax[M1] form of Notch. Mech Dev. 122:479-486. Perry, W.L., C.M. Hustad, D.A. Swing, T.N. O'Sullivan, N.A. Jenkins, and N.G. Copeland. 1998. The itchy locus encodes a novel ubiquitin protein ligase that is disrupted in a18H mice. Nat Genet. 18:143-146. Peschon, J.J., J.L. Slack, P. Reddy, K.L. Stocking, S.W. Sunnarborg, D.C. Lee, W.E. Russell, B.J. Castner, R.S. Johnson, J.N. Fitzner, R.W. Boyce, N. Nelson, C.J. Kozlosky,
247 | Page
M.F. Wolfson, C.T. Rauch, D.P. Cerretti, R.J. Paxton, C.J. March, and R.A. Black. 1998. An essential role for ectodomain shedding in mammalian development. Science. 282:1281-1284. Phng, L.K., and H. Gerhardt. 2009. Angiogenesis: a team effort coordinated by notch. Dev Cell. 16:196-208. Potente, M., H. Gerhardt, and P. Carmeliet. 2011. Basic and therapeutic aspects of angiogenesis. Cell. 146:873-887. Powell, P.A., C. Wesley, S. Spencer, and R.L. Cagan. 2001. Scabrous complexes with Notch to mediate boundary formation. Nature. 409:626-630. Quinn, T.P., K.G. Peters, C. De Vries, N. Ferrara, and L.T. Williams. 1993. Fetal liver kinase 1 is a receptor for vascular endothelial growth factor and is selectively expressed in vascular endothelium. Proc Natl Acad Sci U S A. 90:7533- 7537. Raafat, A., A.S. Goldhar, M. Klauzinska, K. Xu, I. Amirjazil, D. McCurdy, K. Lashin, D. Salomon, B.K. Vonderhaar, S. Egan, and R. Callahan. 2011. Expression of Notch receptors, ligands, and target genes during development of the mouse mammary gland. J Cell Physiol. 226:1940-1952. Ramain, P., K. Khechumian, L. Seugnet, N. Arbogast, C. Ackermann, and P. Heitzler. 2001. Novel Notch alleles reveal a Deltex-dependent pathway repressing neural fate. Curr Biol. 11:1729-1738. Ramos, R.G., B.G. Grimwade, K.A. Wharton, T.N. Scottgale, and S. Artavanis-Tsakonas. 1989. Physical and functional definition of the Drosophila Notch locus by P element transformation. Genetics. 123:337-348. Rampal, R., A.S. Li, D.J. Moloney, S.A. Georgiou, K.B. Luther, A. Nita-Lazar, and R.S. Haltiwanger. 2005. Lunatic fringe, manic fringe, and radical fringe recognize similar specificity determinants in O-fucosylated epidermal growth factor-like repeats. J Biol Chem. 280:42454-42463. Rand, M.D., L.M. Grimm, S. Artavanis-Tsakonas, V. Patriub, S.C. Blacklow, J. Sklar, and J.C. Aster. 2000. Calcium depletion dissociates and activates heterodimeric notch receptors. Mol Cell Biol. 20:1825-1835. Rauen, T., U. Raffetseder, B.C. Frye, S. Djudjaj, P.J. Muhlenberg, F. Eitner, U. Lendahl, J. Bernhagen, S. Dooley, and P.R. Mertens. 2009. YB-1 acts as a ligand for Notch-3 receptors and modulates receptor activation. J Biol Chem. 284:26928- 26940. Raya, A., Y. Kawakami, C. Rodriguez-Esteban, M. Ibanes, D. Rasskin-Gutman, J. Rodriguez-Leon, D. Buscher, J.A. Feijo, and J.C. Izpisua Belmonte. 2004. Notch activity acts as a sensor for extracellular calcium during vertebrate left-right determination. Nature. 427:121-128. Rebay, I., R.J. Fleming, R.G. Fehon, L. Cherbas, P. Cherbas, and S. Artavanis-Tsakonas. 1991. Specific EGF repeats of Notch mediate interactions with Delta and Serrate: implications for Notch as a multifunctional receptor. Cell. 67:687-699. Rechsteiner, M., and S.W. Rogers. 1996. PEST sequences and regulation by proteolysis. Trends Biochem Sci. 21:267-271.
248 | Page
Reese, D.E., C.E. Hall, and T. Mikawa. 2004. Negative regulation of midline vascular development by the notochord. Dev Cell. 6:699-708. Riddle, D.L., T. Blumenthal, B.J. Meyer, and J.R. Priess. 1997. Introduction to C. elegans. In C. elegans II. D.L. Riddle, T. Blumenthal, B.J. Meyer, and J.R. Priess, editors, Cold Spring Harbor (NY). Ridgway, J., G. Zhang, Y. Wu, S. Stawicki, W.C. Liang, Y. Chanthery, J. Kowalski, R.J. Watts, C. Callahan, I. Kasman, M. Singh, M. Chien, C. Tan, J.A. Hongo, F. de Sauvage, G. Plowman, and M. Yan. 2006. Inhibition of Dll4 signalling inhibits tumour growth by deregulating angiogenesis. Nature. 444:1083-1087. Risau, W. 1997. Mechanisms of angiogenesis. Nature. 386:671- 674. Risau, W., and I. Flamme. 1995. Vasculogenesis. Annu Rev Cell Dev Biol. 11:73-91. Riva, C.E., C.J. Pournaras, and M. Tsacopoulos. 1986. Regulation of local oxygen tension and blood flow in the inner retina during hyperoxia. J Appl Physiol. 61:592-598. Robbins, J., B.J. Blondel, D. Gallahan, and R. Callahan. 1992. Mouse mammary tumor gene int-3: a member of the notch gene family transforms mammary epithelial cells. J Virol. 66:2594-2599. Robert, B., P.L. St John, and D.R. Abrahamson. 1998. Direct visualization of renal vascular morphogenesis in Flk1 heterozygous mutant mice. Am J Physiol. 275:F164-172. Rodriguez, P., M.A. Higueras, A. Gonzalez-Rajal, A. Alfranca, M. Fierro-Fernandez, R.A. Garcia-Fernandez, M.J. Ruiz-Hidalgo, M. Monsalve, F. Rodriguez-Pascual, J.M. Redondo, J.L. de la Pompa, J. Laborda, and S. Lamas. 2012. The non-canonical NOTCH ligand DLK1 exhibits a novel vascular role as a strong inhibitor of angiogenesis. Cardiovasc Res. 93:232- 241. Rooke, J., D. Pan, T. Xu, and G.M. Rubin. 1996. KUZ, a conserved metalloprotease-disintegrin protein with two roles in Drosophila neurogenesis. Science. 273:1227-1231. Ross, D.A., and T. Kadesch. 2004. Consequences of Notch- mediated induction of Jagged1. Exp Cell Res. 296:173-182. Roth, M.B., A.M. Zahler, and J.A. Stolk. 1991. A conserved family of nuclear phosphoproteins localized to sites of polymerase II transcription. J Cell Biol. 115:587-596. Ruhrberg, C., H. Gerhardt, M. Golding, R. Watson, S. Ioannidou, H. Fujisawa, C. Betsholtz, and D.T. Shima. 2002. Spatially restricted patterning cues provided by heparin-binding VEGF- A control blood vessel branching morphogenesis. Genes Dev. 16:2684-2698. Sakamoto, K., O. Ohara, M. Takagi, S. Takeda, and K. Katsube. 2002a. Intracellular cell-autonomous association of Notch and its ligands: a novel mechanism of Notch signal modification. Dev Biol. 241:313-326. Sakamoto, K., S. Yamaguchi, R. Ando, A. Miyawaki, Y. Kabasawa, M. Takagi, C.L. Li, B. Perbal, and K. Katsube. 2002b. The nephroblastoma overexpressed gene (NOV/ccn3) protein
249 | Page
associates with Notch1 extracellular domain and inhibits myoblast differentiation via Notch signaling pathway. J Biol Chem. 277:29399-29405. Sakata, T., H. Sakaguchi, L. Tsuda, A. Higashitani, T. Aigaki, K. Matsuno, and S. Hayashi. 2004. Drosophila Nedd4 regulates endocytosis of notch and suppresses its ligand-independent activation. Curr Biol. 14:2228-2236. Salahpour, A., S. Angers, J.F. Mercier, M. Lagace, S. Marullo, and M. Bouvier. 2004. Homodimerization of the beta2-adrenergic receptor as a prerequisite for cell surface targeting. J Biol Chem. 279:33390-33397. Sanalkumar, R., S.B. Dhanesh, and J. James. 2010. Non-canonical activation of Notch signaling/target genes in vertebrates. Cell Mol Life Sci. 67:2957-2968. Sanchez-Solana, B., M.L. Nueda, M.D. Ruvira, M.J. Ruiz-Hidalgo, E.M. Monsalve, S. Rivero, J.J. Garcia-Ramirez, M.J. Diaz- Guerra, V. Baladron, and J. Laborda. 2011. The EGF-like proteins DLK1 and DLK2 function as inhibitory non-canonical ligands of NOTCH1 receptor that modulate each other's activities. Biochim Biophys Acta. 1813:1153-1164. Satchell, S.C., S.J. Harper, and P.W. Mathieson. 2001. Angiopoietin-1 is normally expressed by periendothelial cells. Thromb Haemost. 86:1597-1598. Sato, T.N., Y. Tozawa, U. Deutsch, K. Wolburg-Buchholz, Y. Fujiwara, M. Gendron-Maguire, T. Gridley, H. Wolburg, W. Risau, and Y. Qin. 1995. Distinct roles of the receptor tyrosine kinases Tie-1 and Tie-2 in blood vessel formation. Nature. 376:70-74. Saxena, M.T., E.H. Schroeter, J.S. Mumm, and R. Kopan. 2001. Murine notch homologs (N1-4) undergo presenilin-dependent proteolysis. J Biol Chem. 276:40268-40273. Schafer, W., A. Stroh, S. Berghofer, J. Seiler, M. Vey, M.L. Kruse, H.F. Kern, H.D. Klenk, and W. Garten. 1995. Two independent targeting signals in the cytoplasmic domain determine trans-Golgi network localization and endosomal trafficking of the proprotein convertase furin. EMBO J. 14:2424-2435. Schmidt, M.H., F. Bicker, I. Nikolic, J. Meister, T. Babuke, S. Picuric, W. Muller-Esterl, K.H. Plate, and I. Dikic. 2009. Epidermal growth factor-like domain 7 (EGFL7) modulates Notch signalling and affects neural stem cell renewal. Nat Cell Biol. 11:873-880. Schnitzer, J. 1987. Retinal astrocytes: their restriction to vascularized parts of the mammalian retina. Neurosci Lett. 78:29-34. Schotman, H., L. Karhinen, and C. Rabouille. 2008. dGRASP- mediated noncanonical integrin secretion is required for Drosophila epithelial remodeling. Dev Cell. 14:171-182. Schotman, H., L. Karhinen, and C. Rabouille. 2009. Integrins mediate their unconventional, mechanical-stress-induced secretion via RhoA and PINCH in Drosophila. J Cell Sci. 122:2662-2672.
250 | Page
Schroeter, E.H., J.A. Kisslinger, and R. Kopan. 1998. Notch-1 signalling requires ligand-induced proteolytic release of intracellular domain. Nature. 393:382-386. Scotto-Lavino, E., G. Du, and M.A. Frohman. 2006. 3' end cDNA amplification using classic RACE. Nat Protoc. 1:2742-2745. Sekine, C., Y. Moriyama, A. Koyanagi, N. Koyama, H. Ogata, K. Okumura, and H. Yagita. 2009. Differential regulation of splenic CD8- dendritic cells and marginal zone B cells by Notch ligands. Int Immunol. 21:295-301. Seo, S., H. Fujita, A. Nakano, M. Kang, A. Duarte, and T. Kume. 2006. The forkhead transcription factors, Foxc1 and Foxc2, are required for arterial specification and lymphatic sprouting during vascular development. Dev Biol. 294:458- 470. Seugnet, L., P. Simpson, and M. Haenlin. 1997. Requirement for dynamin during Notch signaling in Drosophila neurogenesis. Dev Biol. 192:585-598. Shawber, C., J. Boulter, C.E. Lindsell, and G. Weinmaster. 1996. Jagged2: a serrate-like gene expressed during rat embryogenesis. Dev Biol. 180:370-376. Shawber, C.J., I. Das, E. Francisco, and J. Kitajewski. 2003. Notch signaling in primary endothelial cells. Ann N Y Acad Sci. 995:162-170. Shawber, C.J., Y. Funahashi, E. Francisco, M. Vorontchikhina, Y. Kitamura, S.A. Stowell, V. Borisenko, N. Feirt, S. Podgrabinska, K. Shiraishi, K. Chawengsaksophak, J. Rossant, D. Accili, M. Skobe, and J. Kitajewski. 2007. Notch alters VEGF responsiveness in human and murine endothelial cells by direct regulation of VEGFR-3 expression. J. Clin. Invest. 117:3369-3382. Shi, Q., M.H. Wu, N. Hayashida, A.R. Wechezak, A.W. Clowes, and L.R. Sauvage. 1994. Proof of fallout endothelialization of impervious Dacron grafts in the aorta and inferior vena cava of the dog. J Vasc Surg. 20:546-556; discussion 556-547. Shi, S., and P. Stanley. 2003. Protein O-fucosyltransferase 1 is an essential component of Notch signaling pathways. Proc Natl Acad Sci U S A. 100:5234-5239. Shimizu, K., S. Chiba, K. Kumano, N. Hosoya, T. Takahashi, Y. Kanda, Y. Hamada, Y. Yazaki, and H. Hirai. 1999. Mouse jagged1 physically interacts with notch2 and other notch receptors. Assessment by quantitative methods. J Biol Chem. 274:32961-32969. Shimizu, K., S. Chiba, T. Saito, K. Kumano, Y. Hamada, and H. Hirai. 2002. Functional diversity among Notch1, Notch2, and Notch3 receptors. Biochem Biophys Res Commun. 291:775- 779. Shimizu, K., S. Chiba, T. Saito, K. Kumano, T. Takahashi, and H. Hirai. 2001. Manic fringe and lunatic fringe modify different sites of the Notch2 extracellular region, resulting in different signaling modulation. J Biol Chem. 276:25753-25758. Shin, H.M., L.M. Minter, O.H. Cho, S. Gottipati, A.H. Fauq, T.E. Golde, G.E. Sonenshein, and B.A. Osborne. 2006. Notch1 augments NF-kappaB activity by facilitating its nuclear retention. EMBO J. 25:129-138.
251 | Page
Shirayoshi, Y., Y. Yuasa, T. Suzuki, K. Sugaya, E. Kawase, T. Ikemura, and N. Nakatsuji. 1997. Proto-oncogene of int-3, a mouse Notch homologue, is expressed in endothelial cells during early embryogenesis. Genes Cells. 2:213-224. Shutter, J.R., S. Scully, W. Fan, W.G. Richards, J. Kitajewski, G.A. Deblandre, C.R. Kintner, and K.L. Stark. 2000. Dll4, a novel Notch ligand expressed in arterial endothelium. Genes Dev. 14:1313-1318. Siekmann, A.F., and N.D. Lawson. 2007. Notch signalling limits angiogenic cell behaviour in developing zebrafish arteries. Nature. 445:781-784. Singh, N., R.A. Phillips, N.N. Iscove, and S.E. Egan. 2000. Expression of notch receptors, notch ligands, and fringe genes in hematopoiesis. Exp Hematol. 28:527-534. Skol, A.D., K.A. Young, D.W. Tsuang, S.V. Faraone, S.L. Haverstock, S. Bingham, S. Prabhudesai, F. Mena, A.S. Menon, C.E. Yu, P. Rundell, J. Pepple, F. Sauter, C. Baldwin, D. Weiss, J. Collins, T. Keith, M. Boehnke, G.D. Schellenberg, and M.T. Tsuang. 2003. Modest evidence for linkage and possible confirmation of association between NOTCH4 and schizophrenia in a large Veterans Affairs Cooperative Study sample. American journal of medical genetics. Part B, Neuropsychiatric genetics : the official publication of the International Society of Psychiatric Genetics. 118B:8-15. Smas, C.M., and H.S. Sul. 1993. Pref-1, a protein containing EGF- like repeats, inhibits adipocyte differentiation. Cell. 73:725- 734. Song, L.L., Y. Peng, J. Yun, P. Rizzo, V. Chaturvedi, S. Weijzen, W.M. Kast, P.J. Stone, L. Santos, A. Loredo, U. Lendahl, G. Sonenshein, B. Osborne, J.Z. Qin, A. Pannuti, B.J. Nickoloff, and L. Miele. 2008. Notch-1 associates with IKKalpha and regulates IKK activity in cervical cancer cells. Oncogene. 27:5833-5844. Sorensen, E.B., and S.D. Conner. 2010. gamma-secretase- dependent cleavage initiates notch signaling from the plasma membrane. Traffic. 11:1234-1245. Sotillos, S., F. Roch, and S. Campuzano. 1997. The metalloprotease-disintegrin Kuzbanian participates in Notch activation during growth and patterning of Drosophila imaginal discs. Development. 124:4769-4779. Sousa, F., D.M. Prazeres, and J.A. Queiroz. 2009. Improvement of transfection efficiency by using supercoiled plasmid DNA purified with arginine affinity chromatography. The journal of gene medicine. 11:79-88. Sparrow, D.B., G. Chapman, A.J. Smith, M.Z. Mattar, J.A. Major, V.C. O'Reilly, Y. Saga, E.H. Zackai, J.P. Dormans, B.A. Alman, L. McGregor, R. Kageyama, K. Kusumi, and S.L. Dunwoodie. 2012. A mechanism for gene-environment interaction in the etiology of congenital scoliosis. Cell. 149:295-306. Sparrow, D.B., G. Chapman, M.A. Wouters, N.V. Whittock, S. Ellard, D. Fatkin, P.D. Turnpenny, K. Kusumi, D. Sillence, and S.L. Dunwoodie. 2006. Mutation of the LUNATIC FRINGE
252 | Page
gene in humans causes spondylocostal dysostosis with a severe vertebral phenotype. Am J Hum Genet. 78:28-37. Sprinzak, D., A. Lakhanpal, L. LeBon, J. Garcia-Ojalvo, and M.B. Elowitz. 2011. Mutual inactivation of Notch receptors and ligands facilitates developmental patterning. PLoS Comput Biol. 7:e1002069. Sprinzak, D., A. Lakhanpal, L. Lebon, L.A. Santat, M.E. Fontes, G.A. Anderson, J. Garcia-Ojalvo, and M.B. Elowitz. 2010. Cis-interactions between Notch and Delta generate mutually exclusive signalling states. Nature. 465:86-90. Stone, J., and Z. Dreher. 1987. Relationship between astrocytes, ganglion cells and vasculature of the retina. J Comp Neurol. 255:35-49. Stone, J., A. Itin, T. Alon, J. Pe'er, H. Gnessin, T. Chan-Ling, and E. Keshet. 1995. Development of retinal vasculature is mediated by hypoxia-induced vascular endothelial growth factor (VEGF) expression by neuroglia. J Neurosci. 15:4738- 4747. Strilic, B., T. Kucera, J. Eglinger, M.R. Hughes, K.M. McNagny, S. Tsukita, E. Dejana, N. Ferrara, and E. Lammert. 2009. The molecular basis of vascular lumen formation in the developing mouse aorta. Dev Cell. 17:505-515. Struhl, G., and A. Adachi. 2000. Requirements for presenilin- dependent cleavage of notch and other transmembrane proteins. Mol Cell. 6:625-636. Suchting, S., C. Freitas, F. le Noble, R. Benedito, C. Bréant, A. Duarte, and A. Eichmann. 2007. The Notch ligand Delta-like 4 negatively regulates endothelial tip cell formation and vessel branching. Proc Natl Acad Sci USA. 104:3225-3230. Sun, Y., W. Lowther, K. Kato, C. Bianco, N. Kenney, L. Strizzi, D. Raafat, M. Hirota, N.I. Khan, S. Bargo, B. Jones, D. Salomon, and R. Callahan. 2005. Notch4 intracellular domain binding to Smad3 and inhibition of the TGF-beta signaling. Oncogene. 24:5365-5374. Suri, C., P.F. Jones, S. Patan, S. Bartunkova, P.C. Maisonpierre, S. Davis, T.N. Sato, and G.D. Yancopoulos. 1996. Requisite role of angiopoietin-1, a ligand for the TIE2 receptor, during embryonic angiogenesis. Cell. 87:1171-1180. Swiatek, P.J., C.E. Lindsell, F.F. del Amo, G. Weinmaster, and T. Gridley. 1994. Notch1 is essential for postimplantation development in mice. Genes Dev. 8:707-719. Swift, M.R., and B.M. Weinstein. 2009. Arterial-Venous Specification During Development. Circ Res. 104:576-588. Tagami, Okochi, Yanagida, Ikuta, Fukumori, Matsumoto, Ishizuka- Katsura, Nakayama, Itoh, Jiang, Nishitomi, Kamino, Morihara, Hashimoto, Tanaka, Kudo, Chiba, and Takeda. 2007. REGULATION OF NOTCH SIGNALING BY DYNAMIC CHANGES IN THE PRECISION IN S3 CLEAVAGE OF NOTCH- 1. Mol Cell Biol. Takeshita, K., M. Satoh, M. Ii, M. Silver, F.P. Limbourg, Y. Mukai, Y. Rikitake, F. Radtke, T. Gridley, D.W. Losordo, and J.K. Liao. 2007. Critical role of endothelial Notch1 signaling in postnatal angiogenesis. Circ Res. 100:70-78.
253 | Page
Tang, Y., H. Bai, S. Urs, Z. Wang, and L. Liaw. 2012. Notch1 activation in embryonic VE-cadherin populations selectively blocks hematopoietic stem cell generation and fetal liver hematopoiesis. Transgenic research. Tang, Y., S. Urs, J. Boucher, T. Bernaiche, D. Venkatesh, D.B. Spicer, C.P. Vary, and L. Liaw. 2010. Notch and transforming growth factor-beta (TGFbeta) signaling pathways cooperatively regulate vascular smooth muscle cell differentiation. J Biol Chem. 285:17556-17563. Tang, Y., S. Urs, and L. Liaw. 2008. Hairy-Related Transcription Factors Inhibit Notch-Induced Smooth Muscle -Actin Expression by Interfering With Notch Intracellular Domain/CBF-1 Complex Interaction With the CBF-1-Binding Site. Circ Res. 102:661-668. Theodosiou, A., S. Arhondakis, M. Baumann, and S. Kossida. 2009. Evolutionary scenarios of Notch proteins. Mol Biol Evol. Thurston, G., I. Noguera-Troise, and G.D. Yancopoulos. 2007. The Delta paradox: DLL4 blockade leads to more tumour vessels but less tumour growth. Nat Rev Cancer. 7:327-331. Torres-Vazquez, J., A.D. Gitler, S.D. Fraser, J.D. Berk, N.P. Van, M.C. Fishman, S. Childs, J.A. Epstein, and B.M. Weinstein. 2004. Semaphorin-plexin signaling guides patterning of the developing vasculature. Dev Cell. 7:117-123. Tsai, F.Y., G. Keller, F.C. Kuo, M. Weiss, J. Chen, M. Rosenblatt, F.W. Alt, and S.H. Orkin. 1994. An early haematopoietic defect in mice lacking the transcription factor GATA-2. Nature. 371:221-226. Ueno, H., and I.L. Weissman. 2006. Clonal analysis of mouse development reveals a polyclonal origin for yolk sac blood islands. Dev Cell. 11:519-533. Umesono, K., K.K. Murakami, C.C. Thompson, and R.M. Evans. 1991. Direct repeats as selective response elements for the thyroid hormone, retinoic acid, and vitamin D3 receptors. Cell. 65:1255-1266. Uyttendaele, H., V. Closson, G. Wu, F. Roux, G. Weinmaster, and J. Kitajewski. 2000. Notch4 and Jagged-1 induce microvessel differentiation of rat brain endothelial cells. Microvasc Res. 60:91-103. Uyttendaele, H., J. Ho, J. Rossant, and J. Kitajewski. 2001. Vascular patterning defects associated with expression of activated Notch4 in embryonic endothelium. Proc Natl Acad Sci U S A. 98:5643-5648. Uyttendaele, H., G. Marazzi, G. Wu, Q. Yan, D. Sassoon, and J. Kitajewski. 1996. Notch4/int-3, a mammary proto-oncogene, is an endothelial cell-specific mammalian Notch gene. Development. 122:2251-2259. Vaccari, T., H. Lu, R. Kanwar, M.E. Fortini, and D. Bilder. 2008. Endosomal entry regulates Notch receptor activation in Drosophila melanogaster. J Cell Biol. 180:755-762. Valenzuela, D.M., J.A. Griffiths, J. Rojas, T.H. Aldrich, P.F. Jones, H. Zhou, J. McClain, N.G. Copeland, D.J. Gilbert, N.A. Jenkins, T. Huang, N. Papadopoulos, P.C. Maisonpierre, S. Davis, and G.D. Yancopoulos. 1999. Angiopoietins 3 and 4:
254 | Page
diverging gene counterparts in mice and humans. Proc Natl Acad Sci U S A. 96:1904-1909. Varnum-Finney, B., L. Wu, M. Yu, C. Brashem-Stein, S. Staats, D. Flowers, J.D. Griffin, and I.D. Bernstein. 2000. Immobilization of Notch ligand, Delta-1, is required for induction of notch signaling. J Cell Sci. 113 Pt 23:4313- 4318. Vercauteren, S.M., and H.J. Sutherland. 2004. Constitutively active Notch4 promotes early human hematopoietic progenitor cell maintenance while inhibiting differentiation and causes lymphoid abnormalities in vivo. Blood. 104:2315- 2322. Vilimas, T., J. Mascarenhas, T. Palomero, M. Mandal, S. Buonamici, F. Meng, B. Thompson, C. Spaulding, S. Macaroun, M.L. Alegre, B.L. Kee, A. Ferrando, L. Miele, and I. Aifantis. 2007. Targeting the NF-kappaB signaling pathway in Notch1-induced T-cell leukemia. Nat Med. 13:70-77. Villa, N., L. Walker, C.E. Lindsell, J. Gasson, M.L. Iruela-Arispe, and G. Weinmaster. 2001. Vascular expression of Notch pathway receptors and ligands is restricted to arterial vessels. Mech Dev. 108:161-164. Vitt, U.A., S.Y. Hsu, and A.J. Hsueh. 2001. Evolution and classification of cystine knot-containing hormones and related extracellular signaling molecules. Mol Endocrinol. 15:681-694. Vokes, S.A., and P.A. Krieg. 2002. Endoderm is required for vascular endothelial tube formation, but not for angioblast specification. Development. 129:775-785. Vooijs, M., E.H. Schroeter, Y. Pan, M. Blandford, and R. Kopan. 2004. Ectodomain shedding and intramembrane cleavage of mammalian Notch proteins is not regulated through oligomerization. J Biol Chem. 279:50864-50873. Wallberg, A.E., K. Pedersen, U. Lendahl, and R.G. Roeder. 2002. p300 and PCAF act cooperatively to mediate transcriptional activation from chromatin templates by notch intracellular domains in vitro. Mol Cell Biol. 22:7812-7819. Wang, H.U., Z.F. Chen, and D.J. Anderson. 1998. Molecular distinction and angiogenic interaction between embryonic arteries and veins revealed by ephrin-B2 and its receptor Eph-B4. Cell. 93:741-753. Wang, J., L. Shelly, L. Miele, R. Boykins, M.A. Norcross, and E. Guan. 2001. Human Notch-1 inhibits NF-kappa B activity in the nucleus through a direct interaction involving a novel domain. J Immunol. 167:289-295. Wang, Y., and M.W. Spellman. 1998. Purification and characterization of a GDP-fucose:polypeptide fucosyltransferase from Chinese hamster ovary cells. J Biol Chem. 273:8112-8118. Wasserthal, L.T. 2007. Drosophila flies combine periodic heartbeat reversal with a circulation in the anterior body mediated by a newly discovered anterior pair of ostial valves and 'venous' channels. The Journal of experimental biology. 210:3707- 3719.
255 | Page
Watanabe, K., T. Nagaoka, J.M. Lee, C. Bianco, M. Gonzales, N.P. Castro, M.C. Rangel, K. Sakamoto, Y. Sun, R. Callahan, and D.S. Salomon. 2009. Enhancement of Notch receptor maturation and signaling sensitivity by Cripto-1. J Cell Biol. 187:343-353. Weinmaster, G., V.J. Roberts, and G. Lemke. 1992. Notch2: a second mammalian Notch gene. Development. 116:931-941. West, A.G., and P. Fraser. 2005. Remote control of gene transcription. Hum Mol Genet. 14 Spec No 1:R101-111. West, H., W.D. Richardson, and M. Fruttiger. 2005. Stabilization of the retinal vascular network by reciprocal feedback between blood vessels and astrocytes. Development. 132:1855-1862. Westin, J., and M. Lardelli. 1997. Three Novel Notch genes in zebrafish: implications for vertebrate Notch gene evolution and function. Dev Genes Evol. 207:51-63. Wigle, J.T., N. Harvey, M. Detmar, I. Lagutina, G. Grosveld, M.D. Gunn, D.G. Jackson, and G. Oliver. 2002. An essential role for Prox1 in the induction of the lymphatic endothelial cell phenotype. EMBO J. 21:1505-1513. Wigle, J.T., and G. Oliver. 1999. Prox1 function is required for the development of the murine lymphatic system. Cell. 98:769- 778. Wilkin, M.B., A.M. Carbery, M. Fostier, H. Aslam, S.L. Mazaleyrat, J. Higgs, A. Myat, D.A. Evans, M. Cornell, and M. Baron. 2004. Regulation of notch endosomal sorting and signaling by Drosophila Nedd4 family proteins. Curr Biol. 14:2237- 2244. Wilkins, S.E., J. Hyvärinen, J. Chicher, J.J. Gorman, D.J. Peet, R.L. Bilton, and P. Koivunen. 2009. Differences in hydroxylation and binding of Notch and HIF-1α demonstrate substrate selectivity for factor inhibiting HIF-1 (FIH-1). Int J Biochem Cell Biol. 41:1563-1571. Will, C.L., and R. Luhrmann. 2011. Spliceosome structure and function. Cold Spring Harb Perspect Biol. 3. Wu, G., S. Lyapina, I. Das, J. Li, M. Gurney, A. Pauley, I. Chui, R.J. Deshaies, and J. Kitajewski. 2001. SEL-10 is an inhibitor of notch signaling that targets notch for ubiquitin- mediated protein degradation. Mol Cell Biol. 21:7403-7415. Wu, J., and E.H. Bresnick. 2007a. Bare rudiments of notch signaling: how receptor levels are regulated. Trends Biochem Sci. 32:477-485. Wu, J., and E.H. Bresnick. 2007b. Glucocorticoid and growth factor synergism requirement for Notch4 chromatin domain activation. Mol Cell Biol. 27:2411-2422. Xu, A., L. Lei, and K.D. Irvine. 2005. Regions of Drosophila Notch that contribute to ligand binding and the modulatory influence of Fringe. J Biol Chem. 280:30158-30165. Xu, Q., Y. Wang, A. Dabdoub, P.M. Smallwood, J. Williams, C. Woods, M.W. Kelley, L. Jiang, W. Tasman, K. Zhang, and J. Nathans. 2004. Vascular development in the retina and inner ear: control by Norrin and Frizzled-4, a high-affinity ligand- receptor pair. Cell. 116:883-895.
256 | Page
Yaffe, D., and O. Saxel. 1977. Serial passaging and differentiation of myogenic cells isolated from dystrophic mouse muscle. Nature. 270:725-727. Yamaguchi, T.P., D.J. Dumont, R.A. Conlon, M.L. Breitman, and J. Rossant. 1993. flk-1, an flt-related receptor tyrosine kinase is an early marker for endothelial cell precursors. Development. 118:489-498. Yamanda, S., S. Ebihara, M. Asada, T. Okazaki, K. Niu, T. Ebihara, A. Koyanagi, N. Yamaguchi, H. Yagita, and H. Arai. 2009. Role of ephrinB2 in non-productive angiogenesis induced by Delta-like 4 blockade. Blood:1-38. Yamazaki, T., Y. Yoshimatsu, Y. Morishita, K. Miyazono, and T. Watabe. 2009. COUP-TFII regulates the functions of Prox1 in lymphatic endothelial cells through direct interaction. Genes Cells. 14:425-434. Yang, L.T., J.T. Nichols, C. Yao, J.O. Manilay, E.A. Robey, and G. Weinmaster. 2005. Fringe glycosyltransferases differentially modulate Notch1 proteolysis induced by Delta1 and Jagged1. Mol Biol Cell. 16:927-942. Ye, Q., J.H. Shieh, G. Morrone, and M.A. Moore. 2004. Expression of constitutively active Notch4 (Int-3) modulates myeloid proliferation and differentiation and promotes expansion of hematopoietic progenitors. Leukemia. 18:777-787. You, L.R., F.J. Lin, C.T. Lee, F.J. DeMayo, M.J. Tsai, and S.Y. Tsai. 2005. Suppression of Notch signalling by the COUP- TFII transcription factor regulates vein identity. Nature. 435:98-104. Yuan, J.S., P.C. Kousis, S. Suliman, I. Visan, and C.J. Guidos. 2010. Functions of notch signaling in the immune system: consensus and controversies. Annual review of immunology. 28:343-365. Zagouras, P., S. Stifani, C.M. Blaumueller, M.L. Carcangiu, and S. Artavanis-Tsakonas. 1995. Alterations in Notch signaling in neoplastic lesions of the human cervix. Proc Natl Acad Sci U S A. 92:6414-6418. Zhang, N., and T. Gridley. 1998. Defects in somite formation in lunatic fringe-deficient mice. Nature. 394:374-377. Zhang, N., C.R. Norton, and T. Gridley. 2002. Segmentation defects of Notch pathway mutants and absence of a synergistic phenotype in lunatic fringe/radical fringe double mutant mice. Genesis. 33:21-28.
257 | Page