Novel Mechanisms of Transcriptional Repression by the Paired- Like Homeodomain Transcription Factor Goosecoid

Novel mechanisms of transcriptional repression by the paired- like homeodomain transcription factor Goosecoid.

Luisa Izzi

A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy Department of Medical Biophysics University of Toronto

Novel mechanisms of transcriptional repression by the paired-like homeodomain transcription factor Goosecoid

Doctor of Philosophy, 2008

Luisa Izzi

Department of Medical Biophysics, University of Toronto

Abstract

Gastrulation is the process by which the three germ layers are generated during vertebrate

development. Nodal ligands, which form a subgroup of the Transforming Growth Factor β

(TGFβ) superfamily, regulate the expression several transcription factors implicated in gastrulation. Among these are the paired-like homeodomain transcription factors Goosecoid

(Gsc) and Mixl1. At the molecular level, Gsc has been described to function as a transcriptional repressor by directly binding to paired homedomain binding sites on target promoters. Here, I describe a novel mechanism of transcriptional repression by Gsc. Using a molecular and embryological approach, I demonstrate that the forkhead transcription factor Foxh1, a major transducer of Nodal signaling, associates with Gsc which in turn recruits histone deacetylases to negatively regulate Mixl1 expression during early mouse development. Post-translational modification of transcription factors by SUMO proteins represents an important mechanism through which their activity is controlled. Here, I also demonstrate that Gsc is sumoylated in mammalian cells by members of the PIAS family of proteins and this modification potentiates the repressive activity of Gsc on direct targets such as the Xbra and Gsc promoters, but not on indirect targets such as Mixl1. Taken together, work presented in this thesis describes two novel mechanisms of transcriptional repression by Gsc.

Dedication

I would like to dedicate this thesis to my beloved and late grandmother Onorina Fedele (July 27 1923-February 10, 1999), a woman who faced life’s hardships with strength, courage and grace. She will forever be a source of inspiration!

“Nella vita, si deve essere sempre guerriere!”

Elisabetta Fedele (Nettuno, Italy, February 25, 2008)

iii

Acknowledgments

Completing a PhD thesis is a Herculean task that can be exasperating at times. However, the exhilaration that is felt when a scientific discovery is made compensates for all the sacrifices that are made along the way. During the last eight years, so many people have contributed on many different levels to help me achieve this goal.

I would first like to thank my supervisor Dr. Liliana Attisano for welcoming me in her lab and giving me the opportunity to pursue my degree under her guidance. Lil, thank you for all you have done over the years to help me find it within myself to achieve my goals. I would also like to thank Dr. Jeff Wrana, a close scientific collaborator and a member of my graduate supervisory committee, for all his contributions. Jeff, I greatly appreciate the time and ressources you dedicated to my PhD research project over the years. As well, I would like to extend my gratitude to Drs. Jane McGlade and Corinne Lobe, who as members of my graduate committee members have followed my progress and provided insight into my doctoral research.

As a member of the Attisano lab, I had the opportunity to work with a select group of very talented young scientists. Over the years, so many of you have become very close friends. I am very appreciative for the kindness and friendship you have shown me over the years. I am especially thankful to Cristoforo Silvestri and Etienne Labbé. I began this journey with you so many years ago, and I have learned so much from you (both scientific and non-scientific). A very special thank you to Garnet Lau for the compassion you showed me when I was working long hours to complete the revisions for my paper. I was so lucky to have you as a little sister in the lab, always making sure that I had something to eat for dinner and a little snack for my walk home. Thank you to Stephen Perusini and Roberta De Oliviera for their camaraderie. I would also like to extend my gratitude to members of the Wrana lab, especially to Masahiro Narimatsu and Abi Ogunjimi. Over the years, you have all been so generous with your time and your expertise.

I made so many friends while I was in Toronto, and all had such a great impact on my life. I am so thankful to Andrea, Arun, Hassina, Miriam, Tuba, and Francoise for being such great friends over the years. A special thank you to Lesley MacNeil, my thesis writing buddy, for listening to my complaints while we were writing our dissertations at the library. Also, I would like to thank Christine for being my big sister in Toronto, for your sound advice and your continued support. To all my friends and family back in Montreal, thank you for keeping in touch and encouraging me along the way. To my surrogate family in Toronto, the Santillis, I will never forget the kindness and generosity you have shown me over the years.

Finally, I am grateful to my parents Elena and Beniamino and my brother Lorenzo, without your love and support none of this would have been possible. Mom and Dad, I’ll never forget all the sacrifices you have made to allow me to pursue my dreams. I hope I can always make you proud!

Table of Contents

Dedication...... iii

Acknowledgments...... iv

Table of Contents...... v

List of Tables...... viii

List of Figures ...... ix

List of Abbreviations...... x

CHAPTER 1

Introduction ...... 1

1.1 Overview of early development in model systems ...... 2 1.1.1 Early amphibian development ...... 3 1.1.2 Early fish development...... 6 1.1.3 Early mammalian development ...... 10

1.2 Nodal signaling during early development...... 14 1.2.1 Ligands ...... 14 1.2.2 Receptors and Co-receptors...... 15 1.2.3 Signal transducers: Smads ...... 20 1.2.3.1 Smad Activation...... 20 1.2.3.2 Role of Smads in early mouse development ...... 22 1.2.4 Smad DNA binding partners ...... 24 1.2.4.1 Foxh1 ...... 25 1.2.4.1.1 Foxh1 structure-function ...... 25 1.2.4.1.2 Foxh1 expression in early development ...... 26 1.2.4.1.3 Foxh1 function during early development...... 28 1.2.4.1.4 Foxh1 target genes...... 30 1.2.4.1.5 Modulation of Foxh1-dependent transcription ...... 32 1.2.4.2 Mixer and p53 ...... 33 1.2.5 Smad transcriptional cofactors...... 34 1.2.5.1 Transcriptional coactivators ...... 34 1.2.5.2 Transcriptional co-repressors ...... 35

1.3 Regulation of Nodal/Smad signaling pathway ...... 36 v

1.3.1 Ligand antagonists...... 37 1.3.2 Receptor antagonists...... 38 1.3.3 Transmembrane-bound antagonist...... 38 1.3.4 Smad-associating proteins ...... 39 1.3.5 Ubiquitin-dependent regulation of Nodal/Smad pathway components ...... 39

1.4 Regulation of Transcription by Sumoylation...... 41 1.4.1 Small ubiquitin-like modifier ...... 41 1.4.2 Sumoylation Pathway...... 42 1.4.3 Targets and transcriptional function of sumoylation ...... 43 1.4.4 Regulation of Smad activity by sumoylation...... 45

1.5 Regulation of the paired-like homeodomain transcription factors Goosecoid and Mix-like proteins by Nodal signaling...... 46 1.5.1 Goosecoid in embryogenesis and cancer ...... 46 1.5.2 Mix-like transcription factors ...... 50

1.6 Thesis Perspectives...... 54

CHAPTER 2

Foxh1 recruits Gsc to negatively regulate Mixl1 expression during early mouse development ...... 56

2.1 Abstract ...... 57

2.2 Introduction...... 58 2.2.1 Materials and Methods ...... 61 2.2.2 Reporter constructs and transcriptional reporter assays...... 61 2.2.3 Construction of mammalian and bacterial expression vectors...... 61 2.2.4 Gel-shifts, immunoprecipitation, GST pull-downs and immunoblotting ...... 64 2.2.5 ES cells and embryoid body differentiation...... 65 2.2.6 Quantitave RT-PCR and in-situ hybridization analysis of Mixl1 in mouse embryos...... 66

2.3 Results ...... 67 2.3.1 Foxh1 and Smads mediate TGFβ/Activin-dependent transcription of Mixl1...... 67 2.3.2 Nodal induces transcriptional activation of Mixl1...... 72 2.3.3 Foxh1 negatively regulates Mixl1 expression during early mouse embryogenesis...... 76 2.3.4 Goosecoid negatively regulates Foxh1-dependent activation of Mixl1...... 79 2.3.5 Physical interaction of Foxh1 with Goosecoid ...... 81 2.3.6 Gooseceoid is recruited to the Mixl1 promoter through Foxh1...... 84 2.3.7 The repressive activity of Goosecoid is mediated by histone deacetylases ...... 87 2.3.8 Goosecoid-mediated repression of endogenous Mixl1 requires Foxh1...... 92

2.4 Discussion ...... 97 vi

2.4.1 Foxh1 negatively regulates Mixl1 expression in mouse embryos ...... 97 2.4.2 Gsc is recruited to Foxh1 and represses Mixl1 expression via histone deacetylases...... 98 2.4.3 Repression of Mixl1 expression by Goosecoid requires Foxh1 in embryoid bodies...... 99

CHAPTER 3

Sumoylation differentially regulates Goosecoid-mediated transcriptional repression...... 102

3.1 Abstract ...... 103

3.2 Introduction...... 104

3.3 Material and Methods ...... 107 3.3.1 Reporter constructs and transcriptional reporter assays...... 107 3.3.2 Construction of mammalian expression vectors and generation of stable cell lines...... 107 3.3.3 Immunoprecipitation and immunoblotting ...... 108 3.3.4 Immunofluorescence staining...... 109

3.4 Results ...... 110 3.4.1 Goosecoid is modified by Sumoylation...... 110 3.4.2 PIAS proteins regulate Gsc Sumoylation ...... 111 3.4.3 Gsc comprises multiple sumoylation sites...... 115 3.4.4 Overexpression of Gsc 6Km leads to a change in MDA-MB-231 cell size ...... 118 3.4.5 Role of Sumoylation on Gsc transcriptional activity...... 121

3.5 Discussion ...... 128 3.5.1 Goosecoid is modified by SUMO proteins...... 128 3.5.2 Sumoylation potentiates Gsc repressive activity on direct targets...... 129 3.5.3 Stable expression of Gsc sumoylaton mutant alters MDA-MB-231 cell morphology ...... 130

CHAPTER 4

General Discussion and Future Directions...... 131

4.1 General Discussion ...... 132 4.1.1 Foxh1 as a transcriptional repressor ...... 132 4.1.2 Foxh1 recruits co-factors to regulate transcription ...... 135 4.1.3 Co-regulation of target genes by Foxh1 and Gsc...... 135 4.1.4 Regulation of Gsc activity ...... 137

4.2 Conclusions ...... 139

CHAPTER 5

References ...... 140 vii

List of Tables

Table 1.1 Molecules involved in Nodal signaling across vertebrate species...... 19 Table 2.1 List of primer used...... 62

viii

List of Figures

Figure 1.1: Amphibian Gastrulation ...... 4 Figure 1.2: Zebrafish Gastrulation ...... 8 Figure 1.3: Mouse Gastrulation ...... 12 Figure 1.4: General mechanism of the Nodal signaling pathway...... 17 Figure 1.5: The Foxh1 family of forkhead/winged helix transcription factors...... 27 Figure 1.6: The SUMOylation pathway...... 44 Figure 1.7: The Gsc family of proteins...... 49 Figure 1.8: The Mix-like family of proteins ...... 53 Figure 2.1: Alignment of Mixl1 promoter sequences ...... 68 Figure 2.2 Foxh1 and Smads bind the Mixl1 promoter and mediate the TGFβ-dependent induction of Mixl1...... 70 Figure 2.3 Foxh1 and Smad binding requirements for the TGFβ-dependent induction of Mixl1...... 73 Figure 2.4: Nodal signaling induces Mixl1 promoter activity...... 75 Figure 2.5: Mixl1 expression is upregulated in gastrulating Foxh1-null embryos...... 77 Figure 2.6: Goosecoid (Gsc) represses Foxh1-mediated induction of the Mixl1 promoter...... 80 Figure 2.7:Gsc interacts with Foxh1 and represses Foxh1-mediated activation of the Mixl1 promoter...... 82 Figure 2.8: Gsc interacts with Foxh1 and Smads on the Mixl1 promoter...... 85 Figure 2.9: Gsc repressive activity is inhibited by Trichostatin A (TSA) treatment...... 88 Figure 2.10: Goosecoid recruits histone deacetylases (HDACs)...... 89 Figure 2.11: A model for Foxh1-dependent regulation of Mixl1...... 91 Figure 2.12: Mixl1 expression is downregulated in wild-type but not in Foxh1-null embryoid bodies overexpressing Gsc...... 94 Figure 2.13: Recruitment of Gsc to the endogenous Mixl1 promoter is Foxh1-dependent...... 96 Figure 2.14: A model representing positive and negative feedback loops regulating Mixl1 and Gsc expression through Foxh1...... 101 Figure 3.1:Goosecoid is post-translationally modified by SUMO proteins ...... 112 Figure 3.2: PIAS proteins regulate Goosecoid sumoylation ...... 114 Figure 3.3: Mutation of six lysine residues decreases Goosecoid sumoylation levels...... 116 Figure 3.4: Stable expression of Gsc 6Km alters cell size in MDA-MB-231 cells ...... 119 Figure 3.5: Sumoylation does not appear to affect the repressive activity of Gsc on the Mixl1 promoter ...... 123 Figure 3.6: Sumoylation potentiates Gsc repressive activity on direct targets...... 125 Figure 3.7: Model illustrating the effect of sumoylation on Gsc-mediated repression...... 127

List of Abbreviations

ActRI: Activin Receptor I E: Embryonic day ActRII: Activin Receptor II E1A: Early region 1A AES: Amino Enhancer of Split EB: embryoid body Aldh: Alcohol dehydrogenase EDTA: ethylene diamine tetraacetic acid. ALK: Activin Receptor Like Kinase EEA1: early endosome associated protein 1 Aos1: Activator of SUMO-1 EGF: Epidermal GrowthFactor AR: androgen receptor EGO: Early Gastrula Organizer ARC: activator recruited cofactor Elk-1: Ets-like protein 1 ARE: Activin Response Element EMSA: Electromobility Shift Assay ASE: asymmetric enhancer EMT: Epithelial-to-mesenchymal transition ATP: Adenosine tri-phosphate Eh-1: engrailed homology 1 AVE: Anterior visceral endoderm Eomes: Eomesodermin ER: Estrogen Receptor BF-1: Brain-factor 1 ES: Embryonic Stem βgal: Beta-galactosidase Evi-1: Ectopic viral integration 1 Bix:Brachyury inducer in Xenopus EVL: enveloping layer BMPs: Bone morphogenetic protein Bon: Bonnie and Clyde FGF: Fibroblast Growth Factor bp: base pair FKBP12: FK506-binding protein FKHD: forkhead C/EBP: CCAAT/enhancer binding protein Flh: floating head CAD: C-terminal acid domain flk1: fetal liver kinase 1 CBP: CREB binding protein FM: Foxh1 motif cDNA: complementary DNA FYVE: Fab1, YOTB, Vac1, EEA1 Cer: Cerberus GCMa: Glial Cell Missing Drosophila Cerl : Cerberus-like homolog a CFC: Cryptic-FRL1-Cripto GCN5: general control of amino acid co-Smad: common Smad synthesis protein 5 cpm: counts per minute GDF: Growth and Differentiation Factor c-PML:cytoplasmic-promyelocytic leukemia GEH: Goosecoid Engrailed Homology CREB: cAMP responsive element binding GPI: glycosyl phosphoatidylinositol protein Ct: Cycle threshold GR: Glucocortoid receptor CtBP: C-terminal binding protein Grg: Groucho-related gene cyc: cyclops Gro: Groucho Gsc: Goosecoid DAN: differential screening-selected gene GST: Glutathione-S-transferase aberrative in neuroblastoma DAPI: 4,6-diamidino-2-phenylindole HA: Haemagglutinin DMZ: dorsal marginal zone hAIP4: human Atrophin1 interacting protein DNA: Deoxyribonucleic acid 4 dpc: days post-coitum HAT: histone acetyltransferase Drap1: DR1-associated protein 1 HD: Homeodomain HDAC: Histone deacetyase

HECT: homologous to the E6-AP carboxyl terminus NCoR/SMRT: Nuclear receptor co- Hesx1: homeo box gene expressed in ES repressor/silencing mediator of retinoid and cells thyroid hormone receptors HMEC: human mammary epithelial cell Nedd4:Neuronal precursor cell-expressed HNF6: Hepatocyte nuclear factor-6 developmentally downregulated 4 hpf: hours post-fertilization NF-E2: nuclear factor erythroid-derived 2 HPRT:Hypoxanthine NF-κB: nuclear factor-kappa B phosphoribosyltransferase Nkx2.5: NK2 transcription factor related HSF1: Heat Shock Factor 1 locus 5 hZIMP10: human zinc finger-containing, NOMO: Nicalin-Nodal Modulator Miz-1, PIAS-like protein on chromosome 10 oep one-eyed pinhead IB: Immunoblotting ICN1: intracellular domain of Notch 1 PACE: Paired basic amino acid cleaving IRF-1: Insuling Response Factor-1 enzyme ISH: In situ hybridization PBS: phosphate buffered saline I-Smad: Inhibitory-Smad PCAF: p300/CBP associated factor PCR: polymerase chain reaction Jab1: Jun activation domain-binding protein- PFA: paraformaldehyde 1 PIAS: protein inhibitor of STAT PML-NB: promyelocytic leukemia- nuclear LEF1: lymphoid enhancer binding factor 1 bodies Lgr4: leucine-rich repeat-containing G PPM1A: protein phosphatase 1A protein-coupled receptor 4 PR: progesterone receptor LIF: leukemia inhibitor factor Lmo: LIM domain only 1 RING: Really INteresting Gene L-R: Left-right RNA: ribonucleic acid LSLD: low salt low detergent R-Smad: receptor-regulated Smad LSE: left-side specific enhancer RT-PCR: Reverse transcriptase-polymerase LUMIER: Luminescence-Based Mammalian chain reaction Interactome Mapping SARA: Smad Anchor for Receptor MAP: mitogen-activated protein Activation MeCP2: methyl CpG binding protein 2 SBE: Smad binding element MEF: mouse embryonic fibrobast SCF: Skp1/Cdc53/F-box Mefc2: Myocyte Enhancer Factor c 2 SDS: Sodium dodecylsulfate MDCK: Madin-Darby canine kidney SENP: SUMO/Sentrin protease MGO: mid-gastrula organizer SFE: Smad/Foxh1 element MH1: MAD-homology 1 SIM: Smad interaction motif MH2: MAD-homology 2 Smad: contraction of Sma and MAD Milk: Mix-like gene Smif: Smad4 interaction cofactor 1 Mix: Mesoderm inducer in Xenopus Smurf: Smad ubiquitination regulatory Mixer: Mix-like endodermal regulator factor MO: morpholino oligonucleotide SNIP1: Smad nuclear interacting protein-1 mRNA: messenger ribonucleic acid Sno: ski-related novel MSG-1: Melanocyte Specific Gene-1 spaw: southpaw Mtx: mix-type Spc: subtilisin-like proprotein convertase MZ: Maternal-Zygotic sqt: squint xi

SRC-1 steroid receptor coactivator-1 SRF: Serum Response Factor-1 SUMO: Small ubiquitin modifier sur: schmalspur SWI/SNF: SWItch/Sucrose NonFermentable

TALE: three-amino-acid loop extension TBP: TATA binding protein TGFβ: Transforming Growth Factor Beta TGIF: TG-interacting factor Tiul1: TGIF interacting ubiquitin ligase 1 TLE: Transducin-like Enhancer of split TMEFF: transmembrane protein with EGF- like and two follistatin-like domains 1 TRF: TGFβ Response Factor TSA: Trichostatin A

Uba2: ubiquitin-like modifier activating enzyme 2 Ubc: Ubiquitin conjugating enzyme VEGF: Vascular Epidermal Growth Factor Xbra: Xenopus Brachyury XCR: Xenopus Cripto-related Xnr: Xenopus nodal-related XTwn: Xenopus homeobox gene Twin XWBSCR11: Xenopus homolog of the human gene Williams-Beuren syndrome critical region 11

YB-1: Y box binding protein-1 YSL: Yolk syncitial layer YY1: Ying Yang 1

ZEB-1/δEF1: zinc finger E-box binding homeobox 1/delat-crystallin/E2-box factor 1 ZNF76: zinc finger protein 76

xii 1

CHAPTER 1

1 Introduction

Portions of this chapter were published as:

Izzi L. and Attisano L. (2006) Ubiquitin-dependent regulation of TGFbeta signaling in cancer. Neoplasia 8(8), 677-88. Review and

Izzi L. and Attisano L. (2004) Regulation of the TGFbeta signalling pathway by ubiquitin- mediated degradation. Oncogene 23 (11), 2071-8. Review

1.1 Overview of early development in model systems

Most multi-cellular organisms arise from a single fertilized egg, a zygote, which undergoes multiple rounds of mitotic divisions and morphological changes to produce a multi-layered, patterned embryo. Through continued growth and development, the embryo ultimately becomes a new living entity capable of functioning independently. In animal species, gastrulation is the most important event occurring during early embryogenesis, leading towards the development of a complex body form. It consists of massive cellular movements which result in the rearrangement of a simple ball of cells into a complex embryo containing three primary germ layers and distinguishable primary body axes. The primary germ layers include the ectoderm, mesoderm and endoderm. While the outer ectodermal layer gives rise to the epidermis and the nervous system, the inner endodermal layer will develop into the digestive tube and its associated organs. The mesoderm, which forms the intermediate layer, produces several organs including the heart, kidney, and gonads, as well as connective tissues such as bone, muscles tendons and blood cells. Although gastrulation movements may vary greatly between species, the dramatic structural reorganization of the embryo allows different groups of cells to come together and establish new intercellular signaling programs which are essential for subsequent morphogenesis and organogenesis.

Several model systems, each offering a unique set of advantages, have been developed to study embryogenesis. The African clawed-toed frog Xenopus laevis, the zebrafish Danio Rerio, and the common house mouse Mus musculus are among the most studied organisms in developmental biology.

1.1.1 Early amphibian development

The main advantage of using Xenopus as a model system is that its eggs and embryos, which develop outside the mother, can be obtained in large numbers and are easily microinjected, grafted and labeled, thus, making the Xenopus a model of choice to study early developmental events such as embryonic axis formation and morphogenesis.

At the blastula stage, the embryo is a simple sphere of cells enclosing a large fluid-filled cavity, known as the blastocoel, which can be demarcated into three regions: the animal and vegetal regions and the marginal zone (Figure 1.1A). While it is generally believed that the ectodermal and endodermal germ layers are specified by maternal factors distributed in the animal (ectoderm) and vegetal (endoderm) regions of the Xenopus oocyte, the mesodermal layer is specified by inductive signals from the vegetal region of the embryo on prospective ectodermal cells present in the marginal zone of the embryo or blastula (reviewed in [1]). A series of explants experiments showed that cells isolated from animal pole, which typically differentiate into ectoderm, can be induced to form mesoderm when incubated with vegetal pole cells (reviewed in [1]). Induction of zygotic gene transcription during mid-blastula transition (MBT) promotes a series of molecular events that lead to the initiation of gastrulation (reviewed in [2]).

In Xenopus, gastrulation begins with the invagination of prospective endodermal cells on the dorsal side of the embryo, forming a slit-like structure or blastopore (Figure 1.1B) [3]. While maintaining contact with the outside surface of the embryo, these cells, which are also known as bottle cells, move towards the inside the embryo and initiate the formation of the archenteron (Figure 1.1B-C). Gastrulation proceeds as marginal zone cells move towards the dorsal lip of the blastopore, involute and continue to migrate along the wall of the blastocoel. As the tip of the archenteron moves animally and then laterally across the gastrula, the bastocoel is reduced in size and is shifted opposite the dorsal blastopore lip (Figure 1.1C). Meanwhile, ectodermal cells from the animal region undergo epiboly, spreading as a sheet, to enclose the inner layers of the embryo. With the formation of new bottle cells in mid-gastrulation stages, the blastopore

Figure 1.1: Amphibian Gastrulation

(A) At blastula stage, the embryo encloses a fluid-filed cavity known as the blastocoel and is demarcated in three regions: the animal region, the vegetal region and the marginal zone. (B) At the onset of gastrulation, bottle cells move inward and form the dorsal blastopore lip. Mesodermal precursors within the dorsal marginal zone involute and migrate under the roof of the blastocoel. (C and D) At mid-gastrula stages, the archenteron forms and displaces the blastocoel opposite the dorsal blastopore lip. Ectodermal cells from the animal region undergo epiboly (cell movements depicted by arrows). Mesodermal cells continue to involute, first moving animally and then laterally within the embryo. Bottle cells from in the ventral region of the embryo and begin to move inward, forming the ventral blastopore lip. (E and F) In late gastrula stages, the blastocoel is obliterated, mesoderm and endoderm have completely internalized and the embryo is surrounded by ectoderm.

* Figure adapted from Gilbert S.F., 2000 [4]

Figure 1.1

expands ventrally and laterally forming a blastopore crescent, where additional mesodermal and endodermal cells migrate inward to pattern the Xenopus embryo (Figure 1.1D). As the blastopore crescent closes into a ring, the remaining endodermal cells are internalized, marking the end of gastrulation (Figure 1.1E-F).

Fate map studies have demonstrated that the first cells to involute inside the gastrula become head mesoderm, while chondramesodermal cells, the next group of cells to migrate within the embryo, are fated to become the notochord, a transient mesodermal structure which is essential for the induction of the nervous system (reviewed in [3]). The dorsal lip of the blastopore, also known as Spemann Organizer, identified by Hans Spemann and Hilde Mangold, was found to be an important dorsal signaling centre with axis-inducing activities [5]. Transplantation experiments demonstrated that grafting of the Spemman Organizer to the ventral marginal zone of another gastrula resulted in the formation of a complete secondary body axis (reviewed in [6]). Additional experiments showed that while transplantation of the organizer of an early gastrula generated a complete ectopic axis with a recognizable head structure, grafting of an organizer from a late gastrula induced a headless secondary axis, thereby, suggesting that the axis-inducing activities of the Spemann Organizer changes during the course of gastrulation (reviewed in [6]).

1.1.2 Early fish development

In the last twenty-five years, Danio rerio, also known as zebrafish, has emerged as important model organism to study vertebrate development. Several features including high fecundity, rapid development, accessibility and optical transparency of the embryos make the zebrafish amenable to genetic and embryological analyses. Because of its experimental tractability, the zebrafish system has provided important insight in both embryonic patterning and organogenesis.

Unlike the Xenopus oocyte, the mature zebrafish egg is radially symmetric prior to fertilization. Approximately 2 hours post-fertilization (hpf), following a rapid cleavage period, the embryo contains three separable regions: the yolk cell, the enveloping layer (EVL) and the deep cells of the blastoderm (reviewed in [7]). The yolk cell is an extra-embryonic tissue which serves mainly as a source of nutrition for the developing embryo. The EVL is a thin layer of epithelium that protects the deep cells from extra-embryonic medium. Later on, the EVL becomes the periderm which is ultimately replaced by definitive skin. The deep cells of the blastoderm give rise to proper embryonic tissues (reviewed in [8]). Approximately 2.5 hpf, the blastoderm undergoes MBT and zygotic transcription begins (reviewed in [7]). As well, cells lying at the blastoderm margin collapse into the yolk and release their nuclei, forming the yolk syncytial layer (YSL). This unique organ functions as a signaling center for mesendoderm induction during later development (reviewed in [7]).

The first cellular movements occur approximately 4 hpf when epiboly begins (reviewed in [7, 8]). During this process, deep cells of the blastoderm move outwardly and intercalate radially, leading to a thinning of the blastoderm which ultimately completely envelops the yolk (Figure 1.2A-B). Gastrulation movements begin approximately 5 hpf when 50% epiboly is achieved. At this time epiboly is arrested and deep cells at the blastoderm margin involute and migrate towards the animal pole of the embryo (Figure 1.2C).

Involution at the margin occurs simultaneously around the embryo and leads to the formation of a germ ring, which is composed of a superficial layer, the epiblast, and an inner layer, the hypoblast [7, 8] (Figure 1.2D). By 6 hpf, convergent extension movements resulting from the migration of cells from lateral and ventral regions to the dorsal region give rise to the formation of the embryonic shield (Figure 1.2D), a localized thickening at the dorsal side of the embryo which is a functional equivalent of the Xenopus dorsal blastopore lip. Similar to Xenopus, the first cells to involute at the embryonic shield give rise to the prechordal plate mesoderm, whereas cells immediately following give rise to chondromesoderm which is a notochord precursor.

Figure 1.2: Zebrafish Gastrulation

(A) At the dome stage (4.3 hpf), epiboly begins as cells of the blastodisc intercalate and start to spread over the yolk. As a result of the radial intercalation of deep cells, the internal yolk syncytial layer (YSL) bulges towards the animal pole (small black arrows). (B) Approximately, 4.7 hpf, when 30% epiboly is achieved, the blastoderm is nearly uniform in thickness. (C and D) Approximately 5 hpf, when epiboly has progressed half-way (50% epiboly), mesoderm inducing factors (indicated by red arrowhead) emanating from the YSL induce mesendodermal cell fates. Involution of these mesodermal precursors at the blastoderm/YSL margin leads to the formation of the hypoblast (D). Convergence and extension movements result in the accumulation of cells at the dorsal margin and the formation of the embryonic shield. (E) At 90% epiboly (9 hpf), the endoderm, mesoderm and ectoderm have formed and surround the yolk. The dorsal side of the blastoderm, which is thicker than the ventral side, will give rise the embryo proper. (F) At the bud stage (10 hpf), the blastoderm completely encloses the yolk cells and epiboly comes to an end, also marking the completion of gastrulation. A tailbud if formed at the caudal end of the embryonic axis.

* Figure adapted from Gilbert S.F., 2000 [4]

Figure 1.2

As epiboly resumes at approximately 7.5 hpf, the shield moves towards the animal pole and gradually becomes less distinctive as convergence extension movements continue promote the lengthening of the AP axis (reviewed in [7, 8]). At 90% epiboly (9 hpf), the anterior dorsal epiblast thickens, forming the neural plate (Figure 1.2E). As epiboly is completed and gastrulation comes to an end, the blastoderm completely encloses the yolk, the neural plate runs along the embryonic axis, and a tail bud forms at the most caudal end of the AP axis (Figure 1.2F). Cells originating from the tail bud contribute to the formation of tail structures, while cells located proximally to the tail bud give rise to the trunk spinal cord (reviewed in [7, 8]).

1.1.3 Early mammalian development

Because of its relatively short life cycle and its amenability to classical genetic experiments, the mouse is an important model organism used to study mammalian development (reviewed in [9, 10]). As with other mammals, the mouse embryo develops within the mother’s uterus and development begins with the fertilization of the mouse egg within the fallopian tube. Approximately 4.5 days post-coitum (dpc), the conceptus or blastocyst is a fluid-filled ball of cells, consisting of three tissue lineages, the trophoectoderm, the primitive endoderm and the epiblast which is of ectodermal origin (reviewed in [9, 10]) (Figure 1.3A). While the

trophoectoderm gives rise to structures important for the implantation and nutrition of the embryo, such as trophoblast giant cells, the ecoplacental cone and extra-embryonic ectoderm, cells from the primitive endoderm form parietal and visceral endoderm. For its part, the epiblast contributes to the embryo proper and additional extra-embryonic tissues. Between 4.5 and 6.0 dpc, the epiblast elongates and develops an internal cavity, forming a U-shaped structure (Figure 1.3B-C) (reviewed in [9, 10]). Approximately 6.5 dpc gastrulation begins with the formation of the primitive streak in the posterior region of the embryo (Figure 1.3D) (reviewed in [11]). There, epiblast cells undergo epithelial-to-mesenchymal transition and migrate through the primitive streak, emerging as a mesodermal layer between the epiblast and visceral endoderm.

Some epiblast cells also intercalate into the visceral endoderm, gradually displacing it and ultimately forming definitive endoderm (reviewed in [11]).

During the course of gastrulation, the primitive streak elongates from the posterior embryonic/extra-embryonic junction to the distal tip of the epiblast (Figure 1.3E), becoming a forerunner of the anterior-posterior axis of the embryo. Located at the anterior end of the elongating primitive streak, a specialized group of cells with organizer activity have been shown to induce the formation of a secondary axis in orthotropic transplantation experiments [12]. This heterogenous population of cells is referred to as early gastrula organizer (EGO), mid-gastrula organizer (MGO) or node depending on the stage of gastrulation (Figure 1.3E). Similar to that of the amphibian dorsal blastopore lip, the patterning activities of the mouse organizer are dynamic throughout gastrulation (reviewed in [6]). Grafting experiments have demonstrated that EGO cells induce a partial axis where anterior neural features are absent. Likewise, transplantation of the node results in the formation of a secondary axis composed mainly of trunk-related structures such as the notochord and somites [12, 13]. In contrast, transplantation experiments with the MGO have shown that cells from the mid-streak organizer are capable of inducing a complete secondary axis with both anterior and posterior features [14].

The observation that grafted nodes failed to induce the formation of head-related structures led to the notion that a second signaling center was required for anterior specification. The anterior visceral endoderm (AVE) is thought to provide inductive signals to underlying epiblast cells specifying them for anterior development (Figure 1.3C-D). While tissue ablation studies have demonstrated that removal of the AVE results in anterior defects [15], transplantation studies have shown that the AVE alone cannot induce anterior neural specification but also requires co- transplantation of underlying epiblast tissues [16]. Evidence also supports the notion that anterior neural induction can occur in the absence of AVE formation [17], suggesting that the AVE may play a secondary role in anterior specification and may function in conjunction with the gastrula organizer to specify anterior neural development.

Figure 1.3: Mouse Gastrulation

(A-C) At 4.5 days post-coitum, the blastocyst is a simple ball of cells consisting mainly of trophoectoderm, primitive endoderm epiblast. Between 4.5 and 6.0 days post-coitum, the epiblast elongates and develops a proamniotic cavity. Around 5.5 days post-coitum, the distal visceral endoderm begins to migrate in the anterior region giving rise to the anterior visceral endoderm. (D) At 6.5 days post-coitum, the primitive streak forms in the posterior end of the epiblast opposite the anterior visceral endoderm. Newly formed mesoderm emerges from the primitive streak and begins to populate the embryo. (E) By 7.5 days post-coitum, the primitive streak reaches its maximal length spanning from the posterior end to the distal tip of the embryo. The organizer or node is visible at the anterior end of the primitive streak. Axial mesendoderm consisting of definitive endoderm and mesoderm is also formed.

*Figure adapted from Hogan B., 1994 [18]

Figure 1.3

1.2 Nodal signaling during early development

In order for embryonic development to proceed normally, a number of crucial events including specification of primary germ layers, determination of embryonic axes and gastrulation movements have to take place in a correct spatial and temporal manner. The proper occurrence of these key developmental processes is dependent on the activity of several signaling pathways, including those of several members of the Transforming Growth Factor β (TGFβ) superfamily, such as bone morphogenetic proteins (BMPs), activins, Vg1, nodals and growth differentiation factors (GDFs) (reviewed in [19-23]). This dissertation seeks to elucidate novel molecular mechanisms underlying the transcriptional regulation of nodal target genes, therefore, the following sections will focus on the role of nodal signaling during early development.

1.2.1 Ligands

Nodal and nodal-related ligands constitute a subgroup of the TGFβ superfamily, and homologues have been identified in various species (Table 1.1; reviewed in [23]). Nodals as well as downstream components of its signaling pathway have been recognized to play important roles in axis specification, mesendoderm induction, left-right asymmetry determination as well as neural induction (reviewed in [23]). While humans, mice and chicks have a single nodal gene,

six nodal-related genes (Xnr1-6) have been described in Xenopus and three nodal-like genes, including squint (sqt); cyclops (cyc), and southpaw (spaw) have been identified in zebrafish so far [24-30].

Nodals, like most TGFβ ligands, are synthesized as pro-proteins homodimers and are subsequently proteolytically processed by substilin-like convertases Spc1/Furin and Spc4/PACE4 to generate a mature and functional form of the ligand [31-33] (Figure 1.4A). Genetic studies in mice indicate that Spc1/Furin and Spc4/PACE4 are essential for the maturation of nodal and the full induction of the its downstream signaling pathway [31, 32].

While Spc1 mutants exhibit defects in heart morphogenesis and reduced definitive endoderm formation, Spc4 mutants exhibit left-right asymmetry defects and cyclopia [32]. Interestingly, these phenotypes are consistent with milder nodal loss-of-function phenotypes (reviewed in [23]), whereas double Spc1/Spc4 mutant mice which exhibit severe reduction of mesendodermal cell fates and loss of anterior visceral endoderm are reminiscent of strong nodal loss-of-function phenotypes [31]. Similarly, Xenopus PACE4 has been found to be required for mesoderm induction as depletion of maternal XPACE4 mRNA results in a reduction of mesendodermal markers as well as a delay in blastopore lip formation [34].

1.2.2 Receptors and Co-receptors

TGFβ ligands induce signaling by bringing together a heteromeric complex of type I and type II serine/threonine kinase receptors [35] (Figure 1.4B). Type I and type II receptors are glycoproteins of approximately 50 and 70 KDa, respectively, comprising a cysteine-rich extracellular domain, a single-span transmembrane region, and a serine/threonine kinase domain (reviewed in [36]). Type II receptors also typically contain a short C-terminal extension following the kinase domain. Upon ligand binding, type II receptors, which are capable of autophosphorylation and are constitutively active, transphosphorylate the type I receptors [37].

This phosphorylation event occurs on multiple serine and threonine residues present within the GS domain which is unique to the type I receptors [38] (Figure 1.4B). Once phosphorylated, the activated type I receptors can signal to downstream effectors.

Genetic studies suggest that activin type II receptors ActRIIA and ActRIIB and activin type I receptor ActRIB (ALK4) mediate nodal signaling in vivo [39-41]. Although early mouse development is not disrupted with the individual mutation of ActRIIA and ActRIIB [42, 43], mice carrying mutations of both activin type II receptors arrested at the egg cylinder stage and failed to undergo gastrulation [40]. Interestingly, ActRIIA-/-;ActRIIB+/- embryos displayed a range of

gastrulation defects, including lack of embryonic axes formation, impairment of primitive streak elongation and mesoderm formation, constriction of the embryonic/extra-embryonic junction as well as anterior patterning defects [40], all of which are also observed in mice carrying mutations in the nodal gene. Likewise, disruption of the ActRIB gene in mice resulted in arrest of egg cylinder development prior to gastrulation [41]. Chimeric analysis showed that although of ActRIB-/- embryonic stem (ES) cells were able to contribute to mesodermal tissues, they did not participate in primitive streak formation [41], further evidence of a function for ActRIB in nodal signaling. Although gain- and loss-of-function studies demonstrated that the orphan receptor ALK7 was implicated in mesendoderm induction and nodal signal transduction during Xenopus embryonic development [39], genetic studies in mouse have demonstrated that ALK7 function is dispensable for nodal signaling during early mouse development [44]. In zebrafish, the type I receptor TARAM-A has been suggested to mediate certain functions of nodal signaling during mesendoderm formation [45].

Unlike activins, TGFβ and BMPs, nodal ligands require the presence of EGF-CFC co-receptors to elicit signaling from the receptor complex (Figure 1.4B) (reviewed in [23]). EGF-CFC co- receptors are glycosyl phosphatidylinositol (GPI) membrane-anchored proteins containing two cystein-rich domains, one with the predicted structure of epidermal growth factor (EGF) and the other known as the CFC domain (for Cryptic-FRL1-Cripto) [46]. Members of this family of co- receptors include zebrafish one-eyed pinhead (oep), frog XCR1/FRL1, XCR2, XCR3, chick CFC, and mouse and human Cripto and Cryptic (Table 1.1) [47-54].

Genetic studies in zebrafish have demonstrated that loss of maternal and zygotic oep phenocopies cyc;sqt double mutants, lacking head and trunk mesoderm and endoderm [55].

Figure 1.4: General mechanism of the Nodal signaling pathway

(A) Nodal ligands are expressed as homodimeric proproteins which are processed by convertases into mature and active forms. The ligand is then able to interact with a receptor complex containing type I (ALK4) and type II (ActRIIA and ActRIIB) receptors as well as EGF-CFC co- receptors. (B) Ligand binding leads to the phosphorylation of type I receptors by type II receptors resulting in type I receptor activation which is followed by the phosphorylation of receptor-regulated Smads (R-Smad, Smad2 or Smad3). (C) Activated R-Smads interact with Smad4 (co-Smad) and translocate to the nucleus where they interact with (D) transcription factors (Foxh1 and Mixer) as well as cofactors to regulate gene transcription. (E) Nodal signaling is inhibited by soluble antagonists such as Cerberus and Lefty proteins which bind the Nodal ligand. Lefty can also block signaling through interaction with the EGF-CFC co- receptors. Tomoregulin (TMEFF) inhibits pathway activity by preventing the recruitment of EGF-CFC coreceptors to the receptor complex. (F) Receptor complexes are targeted for lysosomal degradation by Dapper2 which promotes their internalization into endosomes. (H) PPM1A inhibits signaling by promoting the dephosphorylation and nuclear export of Smad2.

Figure 1.4

Table 1.1 Molecules involved in Nodal signaling across vertebrate species.

In frogs, although XCR1, XCR2 and XCR3 are all involved in mediating nodal signaling, each protein appears to potentiate the activity of specific ligands and elicit diverse biological effects [47, 49]. Loss-of-function studies showed that while XCR3 is essential for mesendoderm specification and XCR1 is required for the formation anterior and neural tissues [49], XCR2 is implicated in L-R patterning [47]. Mice lacking Cripto function, similar to those lacking nodal activity, display defects in anterior-posterior axis patterning [56]. Disruption of the Cryptic gene, however, results in mice with L-R laterality defects, including randomization of embryo turning, cardiac looping and abdominal situs as well as pulmonary right isomerism [57, 58]. Interestingly, human L-R laterality defects are also associated with loss-of-function mutations in the CRYPTIC/CFC-1 gene [59]. Consistent with genetic studies, biochemical data indicate that Cripto facilitates the association of nodal or Xnr-1 with ALK4 either alone or complexed with ActRIIB [39, 60, 61]. Cripto has also been found to promote Xnr-1 signaling downstream of the ALK7/ActRIIB receptor complex [39].

1.2.3 Signal transducers: Smads

1.2.3.1 Smad Activation

Downstream of the activated the type I receptor, TGFβ superfamily signal transduction is mediated by the Smad family of proteins which comprises eight members categorized in three functional subgroups: the receptor-regulated Smads (R-Smads), common Smad (co-Smad) and inhibitory Smads (I-Smad) (Table 1.1) [62-64]. The type I receptor kinase directly phosphorylates two serine residues at the carboxy-terminus of R-Smads, Smad1, 2, 3, 5, and 8 (Figure 1.4B) [62-64]. In general, the specificity of the responses elicited by TGFβ superfamily members is dictated by the ability of BMP type I receptors to phosphorylate and activate the R- Smads Smad1, 5 and 8, and TGFβ or activin type I receptors to phosphorylate the R-Smads, Smad2 and Smad3. Consistent with this, nodal ligands signal through the activin type I receptor

ActRIB and induction of the nodal signaling pathway results in the phosphorylation of R-Smads, Smad2 and Smad3 [60, 65, 66]. Once phosphorylated, R-Smads can then associate with the co- Smad, Smad4 and translocate to the nucleus where the complex is able to interact with DNA binding proteins to regulate transcriptional responses [62-64] (Figure 1.4C-D). In addition to associating with a variety of transcription factors, Smads recruit both transcriptional co-activator and co-repressor to modulate levels of transcription (reviewed in [67]).

A third class of Smads, the inhibitory Smads or I-Smads, which includes Smad6 and Smad7, have been identified as negative regulators of TGFβ and BMP signaling. By interacting with type I receptors, I-Smads block the access of R-Smads to their specific receptors and inhibit signaling. In addition, I-Smads can downregulate signaling via the cell surface receptors by targeting them for ubiquitin-dependent proteosomal degradation [62, 63, 68-71].

Smads comprise two conserved globular domains, the N-terminal MH1 domain and C-terminal MH2 domain, which are separated by a proline-rich linker region [62-64, 72]. While the MH1 domain of R-Smads and Smad4 are highly conserved, the amino-terminal region of the I-Smads, only shows a weak sequence similarity with the MH1 domain of other Smads. The MH2 domain is highly conserved among all Smads, whereas the linker regions are divergent. Functionally, the MH1 domain is implicated in nuclear import and DNA binding, whereas the MH2 domain is involved in receptor interaction and Smad oligomerization [62-64, 72]. The MH2 domain has also been shown to mediate the interaction between Smad2 and Smad3 with the Smad anchor for receptor activation (SARA). This FYVE domain-containing protein, which mainly localizes to early endosomes, enhances the recruitment of R-Smads to TGFβ receptors and facilitates TGFβ signaling [73]. Both the MH1 and MH2 domains mediate interactions with transcription factors, and transcriptional co-activators or co-repressors [62, 74, 75]. The linker region contains a number of phosphorylation sites crucial for crosstalk with other signalling pathways and a PY motif, which specifically interact with E3 ubiquitin ligases implicated in the regulation of the TGFβ and BMP pathways [63].

Biochemical studies as well as functional studies in frog, fish and tissue culture cells have demonstrated a role for Smad2 and Smad3 signaling downstream of the nodal signaling pathway [65, 66, 76-80]. Although these two molecules share 92% amino acid identity, Smad2 and Smad3 exhibit differential DNA binding capabilities because of structural differences in their MH1 domain [81]. While Smad3 can bind directly to DNA through a β-hairpin DNA binding motif located within its MH1 domain, a thirty amino acid insert within the MH1 domain of Smad2 is thought to impose steric constraints which prevents Smad2 DNA binding [81]. Therefore, Smad2 must form complexes with Smad4 or other transcription factors to be recruited to target promoters. Consistent with their functional differences, Smad2 and Smad3 display very different transcriptional activities [62-64]. For example, in human HepG2 hepatoma cells, while Smad2/4 expression positively regulates the induction of mouse Goosecoid promoter mediated by the forkhead/winged helix transcription factor Foxh1, Smad3/4 expression downregulates promoter activity [82]. Similarly, whereas activation of the Xenopus Mix2 ARE-luciferase reporter is observed in Smad3-deficient mouse embryonic fibroblasts (MEFs), reporter activity is suppressed in Smad2-null MEFs [83]. Taken together, these biochemical data suggest unique roles for Smad2 and Smad3 in the regulation of TGFβ-dependent target gene expression.

1.2.3.2 Role of Smads in early mouse development

Consistent with these findings, genetic studies in mice have revealed very different phenotypes for Smad2 and Smad3 mutant animals. While Smad3 mutant mice are viable and fertile and show only subtle developmental abnormalities [84-86], Smad2 mutant mice fail to establish anteroposterior polarity and undergo gastrulation resulting in arrested development between E7.5 and E12.5 [87-89]. Analysis of chimeras in which wild-type ES cells were injected into Smad2-/- blastocysts revealed that early patterning defects observed in Smad2 mutant mice are largely due to aberrant specification of the AVE [89]. Reciprocal chimeras demonstrated that while Smad2-deficient ES cells contributed extensively to embryonic mesoderm and ectoderm,

they failed to colonize definitive endodermal lineages [90]. In agreement with this, without affecting primitive streak formation and gastrulation, conditional inactivation of Smad2 expression in the epiblast disrupts the specification of anterior mesendoderm [91]. Furthermore, genetic manipulation of Smad2 and Smad3 expression ratios showed that loss of Smad3 in the context of one wild-type copy of Smad2 (Smad2+/-;Smad3-/-) results in defects in anterior mesendoderm specification [92]. In embryos with Smad2-deficient epiblasts, loss of one allelic copy of Smad3 results in embryos lacking all anterior primitive streak derivatives, including prechordal plate mesoderm, anterior definitive endoderm, notochord and node [92]. Complete loss of Smad2 and Smad3 in the epiblast results in embryos in which middle-streak mesodermal lineages, such as paraxial and lateral plate mesoderm, are missing [92]. Taken together, these genetic studies indicate that while Smad2 functions in the AVE to specify AP polarity of the embryo, it cooperates with Smad3 in the epiblast to specify mesendodermal lineages. Furthermore, the increasingly severe phenotypes generated by graded loss of Smad2/3 function are reminiscent of phenotypes obtained from graded loss of nodal activity [91, 93].

Embryos lacking the Smad4 gene exhibit a phenotype that is similar but slightly more severe than that of Smad2 mutant mice [94, 95]. Smad4 mutant mice result in embryonic lethality prior to gastrulation, displaying severe embryonic disorganization, lack of extra-embryonic ectoderm and mesoderm formation as well as reduced ectodermal cell proliferation [94, 95]. Complementation experiments where Smad4-/- ES cells were aggregated with tetraploid wild- type embryonic cells demonstrated that the gastrulation defect exhibited by Smad4 mutant embryos was a result of a defect in visceral endoderm function [94]. Furthermore, conditional inactivation of Smad4 activity demonstrated that although Smad4-deficient epiblasts are capable of forming mesoderm, they fail to develop structural derivatives of the anterior primitive streak, including the notochord, node and definitive endoderm [96]. Because abundant biochemical data demonstrates a central role for Smad4 in the signal transduction of TGFβ superfamily members, the ability of Smad4 mutants to form mesoderm suggests that Smad4-independent pathways may be elicited to mediate mesodermal induction by nodal signal. Although it has not been

implicated in mesoderm specification, the nuclear protein TIFγ (also known as ectodermin and Trim33) has recently been shown to associate with Smad2/3 and mediate TGFβ signaling in hematopoietic progenitors [97], therefore, R-Smads can potentially transduce Nodal signals independently of Smad4 activity.

1.2.4 Smad DNA binding partners

The recruitment of the R-Smad/Smad4 complex to target promoters is a crucial step determining the transcriptional response to TGFβ signaling. Although Smads can directly bind DNA, their affinity is relatively low and would not support efficient transcriptional activation in vivo [81]. Numerous studies have also demonstrated that Smads can directly interact with target promoters through either a minimal Smad Binding Element (SBE), defined as 5’-AGAC-3’, or GC-rich motifs (reviewed in [98]). Because the SBE is found approximately once every 1 Kbp in the genome and GC-rich motifs are prominent in promoter regions, induction of gene expression solely through Smad binding would result in a non-specific response to TGFβ superfamily signaling. Thus, R-Smad/Smad4 complexes associate with other DNA-binding transcription factors to achieve a level of affinity and selectivity conducive to specific transcriptional activation. Thus far, over forty Smad DNA-binding partners have been identified [68]. These transcription factors belong to diverse families of transcription factors, including the bHLH, bZIP, forkhead, homeodomain, nuclear receptor, runx, and zinc finger families [68]. The extensive range of transcription factors with which Smads are able to cooperate reflects the versatile transcriptional responses elicited by TGFβ superfamily signaling. Of the numerous transcription factors able to interact with Smads, only Foxh1, Mixer and p53 have been shown to mediate transcriptional responses downstream of the Nodal signaling pathway (Figure 1.4D).

1.2.4.1 Foxh1

1.2.4.1.1 Foxh1 structure-function

Foxh1 (initially known as Forkhead Activin Signal Transducer or FAST1) is the first and most extensively studied transcription factor associated with the regulation of target genes downstream of Nodal-like signaling pathways. Foxh1 is a member of the forkhead/winged helix family of transcription factors that was originally identified based on its ability to mediate activin-dependent induction of the Xenopus Mix2 gene [99]. Foxh1 family members, which include Xenopus FoxH1a (FAST1) and FoxH1b (FAST3), mouse Foxh1 (FAST2), zebrafish foxh1/schmalspur (sur) and human FOXH1 (Table 1.1, Figure 1.5) are now known to regulate the transcription of a myriad of genes expressed during the embryonic development of frog, fish and mouse [82, 100-109].

In vitro binding site selection analysis originally identified the sequence 5’- AAT(C/A)(A/C)ACA-3’ as an optimal Foxh1 binding site [110]. However, a functional analysis of the Smad/Foxh1 enhancer (SFE) of the mouse Gsc promoter demonstrated that nucleotides flanking the core 8-bp Foxh1 binding site as well as the relative orientation and distance between Smad and Foxh1 binding sites were critical for maximal SFE activity [109]. Optimal SFE are defined as a 5’-(C/G)AAT(C/A)CACA(A/T)-3’ Foxh1 binding site with Smad binding elements located within 50 bp upstream [109]. Although Foxh1 family members can directly bind DNA through their forkhead/winged helix domain (Figure 1.5), they lack transactivating activity [82, 110, 111]. In response to activin and Nodal signaling, Smad2/4 or Smad3/4 complexes associate to constitutively DNA-bound Foxh1 to mediate the transcriptional regulation of target genes [82, 110, 111]. While Foxh1 interacts with the MH2 domain of Smad2 and Smad3 [82, 110, 111], two highly conserved motifs present within the C-terminal region of Foxh1 family members, the Smad interaction (SIM) and FAST/Foxh1 motif (FM), are required for the recruitment of Smads (Figure 1.5) [112, 113]. The SIM is a proline-rich sequence that interacts with the MH2 domains of both Smad2 and Smad3 [112, 113]. In contrast, the FM was found to only interact with the

MH2 domain of activated Smad2 but not Smad3 [113]. Since the MH2 domain of Smad2 and Smad3 are nearly identical, it is not yet clear how the FM is able to discriminate between these two proteins. Regardless, the FM may be providing a mechanism through which genes are regulated by one Smad versus the other in specific contexts. It was demonstrated that while Smad2 promoted the Foxh1-dependent induction of the mouse Goosecoid gene, Smad3 negatively regulated the expression of the gene [82]. Although it was demonstrated that the differential effect of Smad2 and Smad3 on the regulation of Gsc expression resided in the ability of their MH1 domain to bind DNA [82], the FM may required in certain contexts to recruit Smad2 instead of Smad3 to allow for Gsc expression in vivo. In contexts where Gsc expression requires to be turned off, additional transcription factors may be present to aid in the recruitment of Smad3 to Foxh1.

1.2.4.1.2 Foxh1 expression in early development

Foxh1 family members are expressed primarily during early development [99, 114-117]. In mouse, before and during gastrulation, Foxh1 is expressed predominantly throughout the epiblast although low levels of expression are also detected in extra-embryonic endoderm [115]. During early somite stages (~E8.0), Foxh1 is expressed bilaterally in the lateral plate mesoderm and

weakly in the neural plate [118]. In later stages, Foxh1 expression becomes restricted to the embryonic heart, allantois and forebrain [109, 115]. Cross-sections of E8.25 embryos show a strong expression of Foxh1 in the surface ectoderm and neuroectoderm of the forebrain as well as the head mesenchyme [109].

In frog and fish, Foxh1 family members are expressed both maternally and zygotically [99, 114, 116]. While XFoxh1a mRNA is detected in oocytes and expression is maintained throughout gastrulation, XFoxh1b expression is detected only in gastrulating embryos [99, 114].

Figure 1.5: The Foxh1 family of forkhead/winged helix transcription factors.

A schematic representation of the Foxh1 family members is shown (frog, xFoxh1a and xFoxh1b; mouse, mFoxh1; zebrafish, zFoxh1; human, hFOXH1). The forkhead/winged helix domain (FKHD) is depicted in blue, the Smad interaction motif (SIM) is shown in green) and the FoxH1 motif (FM) is represented in yellow.

At early gastrula stages (Stage 10.25), XFoxh1a and XFoxh1b are both expressed in the animal cap and prospective mesoderm [114]. Similarly, in zebrafish, foxh1 transcripts are detected in oocytes prior to fertilization, where they are prominently localized in the prospective animal pole [116]. Although foxh1 expression is widely detected through the embryo during blastula and early gastrula stages, a ventral to dorsal gradient of foxh1 expression is progressively established during gastrulation, with highest levels being observed in the embryonic shield [116, 117]. During somitogenesis, foxh1 expression becomes restricted to the midline, lateral plate mesoderm and dorsal anterior neuroectoderm [116, 117]. After the segmentation period, foxh1 expression is no longer detected [116, 117].

1.2.4.1.3 Foxh1 function during early development

As suggested by the gene expression studies, Foxh1 family members play important roles during early development. Loss of function studies in mouse demonstrated that Foxh1-/- embryos display a variable range of phenotypes which can be classified into three general groups. Type I embryos, showing the least severe phenotype, exhibit axial defects which include loss or reduction of midline structures such as the node, notochord, and prechordal plate, while the intermediate type II embryos, in addition to midline defects, completely lack anterior structures

[119, 120]. Consistent with these findings, zebrafish embryos lacking both maternal and zygotic foxh1 activity (MZschmalspur;MZsur) exhibit similar morphological defects, including loss of the embryonic shield, notochord disruptions and reduction of prechordal plate [116, 117]. Axial deficiencies ranging from anterior truncations, reduced notochords to fused somites are also observed in FoxH1-depleted Xenopus embryos [104].

Foxh1-/- embryos with the most severe phenotype, type III mutants, fail to establish the anteroposterior axis and lack all development of the epiblast [119]. Chimeric studies as well as specific deletion of Foxh1 in the epiblast but not the extra-embryonic lineages revealed that the

AP patterning defects resulted from the improper specification of the AVE in Foxh1-/- animals [119]. While MZsur mutants display relatively mild gastrulation defects [116, 117], recent studies in which foxh1 expression in zebrafish embryos is abrogated using anti-sense morpholino oligonucleotides (MO) revealed severe defects in epiboly and convergence extension movements [108]. Although both sur alleles contain mutations that result in loss of Foxh1 DNA binding activity, the mutant protein may function as an hypomorph that still retains activity associated with the SIM and FM domains, thus, explaining the discrepancy in the phenotypes observed between MZsur mutants and foxh1-depleted zebrafish embryos [108]. Sequencing of the zebrafish genome is not yet complete, thus, the distinct phenotypes may also be due to an additional foxh1 gene that remains to be identified that is targeted by the MO but remains intact in the MZsur mutants [108].

In addition to gastrulation defects, Foxh1-/- mutant mice display defects in anterior heart field derivatives, including a disorganized myocardium, reduced ventricular trabeculation and outflow tract malformations [105]. Foxh1 has recently been shown to be required in forebrain patterning and development [109]. As well, Foxh1-/- mutant mice fail to establish definitive endoderm [120]. Progenitor cells from the anterior primitive streak contribute to the formation of definitive endoderm. Chimeric analyses demonstrated that Foxh1-/- cells fail to colonize the embryonic foregut, indicating that the lack of definitive endoderm in Foxh1-null mice is not related to APS defects but rather to the requirement of Foxh1 activity in specification of definitive endoderm [120]. A reduction or loss of endodermal markers foxa2/axial and sox17 is also observed in zebrafish Foxh1 morphants [108]. As disruption of other Nodal signaling components results in defects in definitive endoderm formation [23], similar phenotypes observed with the loss of Foxh1 activity support a conserved role for Foxh1 in nodal signaling and endoderm specification.

1.2.4.1.4 Foxh1 target genes

In agreement with the expression pattern of Foxh1 family members, Foxh1 target genes primarily function during early development [23, 121, 122]. Ectopic expression of FoxH1 fusion constructs in frog have identified several putative FoxH1 target genes, including the pan- mesodermal marker Xbra, the organizer genes Gsc and Xlim1, and the mesendodermal marker Mix2 [123]. While the Mix2 gene was already known to be a FoxH1 target gene [99], subsequent analysis confirmed Xlim1 as a bona fide Foxh1 target gene [103]. Foxh1 binding sites have also been identified and confirmed to be functional in the promoter of Eomesodermin (Eomes) [101]. Although direct transcriptional regulation by Foxh1 remains to be confirmed, gain-of-function studies in zebrafish have identified mesodermal markers gsc and floating-head (flh) as well as endodermal markers gata5, axial and sox17 as potential Foxh1 targets [116]. Analysis of the U3 region of the bhikhari LTR-retroelement revealed a Foxh1 binding site which was found to be required for nodal-dependent expression of bhikhari in zebrafish mesendoderm [124]. Consistent with findings in Xenopus and zebrafish, Foxh1 consensus sites mediating activin/nodal responsiveness have been described in the promoters of the mouse Gsc and Mixl1 genes, both primitive streak markers of gastrulating embryos ([82, 106, 120] and this thesis). In mouse, expression of Foxa2, a homologue of zebrafish axial, is lost in Foxh1-/- mutant embryos, and although Foxh1 binding sites have not been identified in Foxa2 regulatory elements, it has

been suggested that Foxh1 may be part of a genetic pathway regulating Foxa2 expression in vivo [120]. Furthermore, this genetic pathway is likely to be conserved as foxh1 also appears to regulate axial in zebrafish [116]. While most known Foxh1 targets are involved vertebrate gastrulation, an in silico-based genome-wide analysis identified a novel set of Foxh1 target genes involved in forebrain development, and these include Lgr4, Lmo1, Hesx1, Fgf8, Aldh1a1, Aldh1a2, Aldh1a3 [109].

Though the initial expression of nodal-related ligands does not require Foxh1 activity, maintenance and propagation of nodal signals is dependent on Foxh1 in various chordate species.

In mouse, nodal is positively autoregulated by an asymmetric enhancer (ASE) located within the first intronic region and a left-side specific enhancer (LSE) present within the promoter region, both of which contain Foxh1 binding sites required for their induction [93, 118, 125]. Conserved Foxh1 binding sites are also essential for the expression in Xenopus Xnr-1 [100]. Although the direct transcriptional regulation of cyc and sqt through foxh1 has not yet been reported, maintenance of cyc and sqt expression is lost in MZschmalspur mutants [116].

Several studies have also demonstrated that Foxh1 functions in negative feedback loops to modulate expression of nodal-related genes by regulating the expression of nodal antagonists, mouse lefty2 and zebrafish antivin/lefty2 [100, 116]. Analysis of left-side specific enhancers (ASE) within regulatory regions of the mouse lefty1 [107] and lefty2 [100] and human LEFTY1 and LEFTY2 genes [126] confirmed the direct regulation of these genes by Foxh1. Known antagonists of nodal signaling also include members of the Cerberus/Dan family of proteins (reviewed [23]). While several studies have demonstrated the nodal-dependent regulation of Cerberus-like proteins [127-129], chicken Cerberus is likely to be the only family member to be directly regulated by Foxh1 as comparison of regulatory sequences of other family members failed to identify conserved Foxh1 binding sequences [130].

The positive and negative feedback loops through which Foxh1 regulates the expression of nodal and lefty genes are critical to the establishment of left-right asymmetry in the vertebrate embryo. Further supporting a role for Foxh1 in left-right determination, asymmetric expression of the mouse Pitx2 gene is controlled by the induction of an intronic ASE that contains multiple Foxh1 binding sites [102]. Interestingly, newborn transgenic mice lacking the Pitx2 ASE manifest numerous L-R laterality defects, including right isomerism in the lungs and heart atrium and reversed positioning of the great arteries and heart apex [131].

Although Foxh1 is commonly known as a positive regulator of transcription, several recent studies described a new role for Foxh1 as a negative regulator of gene expression ([104, 132, 133] and this thesis). Depletion of maternal FoxH1 results in the ectopic expression of Xenopus

Xnr-5 and Xnr-6 in ventral vegetal of the late blastula [104]. The mechanism through which Foxh1 negatively regulates these genes remains to be elucidated, however, regulation is likely to be direct as Foxh1 binding sites have been identified in the promoter region of Xnr-5 [132]. In zebrafish, Foxh1 negatively regulates the expression of the VEGF receptor flk1 during vasculature formation through conserved Foxh1 consensus sequences identified in its upstream regulatory region [133].

1.2.4.1.5 Modulation of Foxh1-dependent transcription

Foxh1 binding to activin/nodal responsive regulatory elements is typically sufficient to induce the optimal expression of target genes. However, several studies have described mechanisms through which Foxh1 interacts with other transcription factors to synergistically activate gene expression. The Xenopus homolog of the human Williams-Beuren syndrome critical region 11 (XWBSCR11) factor cooperates with Foxh1 to synergistically activate the activin-dependent response of the gsc promoter [134]. Similarly, the maximal TGFβ-dependent induction of the mouse Mefc2 gene during cardiac development requires the combined activity of Foxh1 and the cardiac specific homeobox transcription factor Nkx-2.5 [105].

Several mechanisms have also been described where activin/nodal-dependent transcriptional

activation is down-regulated through the inhibition of the Foxh1/Smad2 complex formation. The forkhead/winged helix protein FoxG1 (BF-1) and the MADS box-containing transcription factor Serum Responsive Factor (SRF) block transcriptional activation through their association with Foxh1 and inhibition of the Foxh1/Smad2 complex[135, 136]. In contrast, Drap1, a repressor of basal transcription by RNA polymerase II, antagonizes nodal signaling by binding to Foxh1 and inhibiting DNA binding to target promoters [137].

1.2.4.2 Mixer and p53

The gastrulation, axial patterning and endoderm specification defects observed in Foxh1 mutants in mouse, frogs and fish are consistent with a role for Foxh1 downstream of the Nodal signaling pathway (reviewed in [23]). However, phenotypes displayed by nodal loss-of-function mutants are not fully recapitulated by FoxH1 mutants. Mesoderm specification is largely unaffected in Foxh1 loss-of-function mutants as evidenced by the expression of mesodermal markers [104, 116, 117, 119, 120], while zebrafish cyc-/-;sqt-/- and mouse nodal mutants exhibit a severe disruption in mesoderm formation (reviewed in [23]). These observations indicate that additional transcription factors are also mediating important functions downstream of the nodal signaling pathway.

So far, two transcription factors have been identified as alternative factors mediating nodal- dependent transcription during early development (Figure 1.4D). Biochemical studies have demonstrated that Mixer, a paired-like homeobox transcription factor, recruits active Smad2/Smad4 to mediate the activin/TGFβ-dependent induction of the Xenopus gsc promoter [112]. Further supporting a role for Mixer in regulating gsc expression in vivo, Mixer and gsc are co-expressed in prospective mesendodermal cells of the gastrulating Xenopus embryo [112] and the zebrafish mixer/bonnie & clyde (bon) mutant displays a decreased gsc expression in mesendodermal cells [138]. Gain- and loss- of function studies in Xenopus demonstrated that the tumor suppressor p53 specifically activates the expression of a subset of activin target genes to promote mesoderm as well as endoderm formation [139]. Mechanistically, activated Smad2 physically interacts with p53 to regulate activin-responsive regulatory elements [139].

1.2.5 Smad transcriptional cofactors

1.2.5.1 Transcriptional coactivators

In addition to binding a range of transcription factors, Smads interact with transcriptional cofactors to control the level of expression of target genes. Smads mediate transcriptional activation, at least in part, through the recruitment of general co-activators with intrinsic histone acetyltransferase (HAT) activity, including p300/CREB Binding Protein (CBP), p300/CBP associated factor (PCAF) and GCN5 [140-146]. HATs mediate the acetylation of lysine residues within the N-terminal tail of core histones. This modification results in the weakening of the interaction between histones and chromatin and ensues in the loosening of chromatin, which in turn results in the recruitment of the basal transcriptional machinery and upregulation of transcription. In addition to transcriptional co-activators with intrinsic HAT activity, Smads recruit proteins such as SRC-1, MSG-1, ZEB-1/δEF1, PIAS3, Smif, which themselves recruit p300/CBP to regulate transcription [147-153].

Recent evidence indicates that, in P19 cells, activated Smad2/Smad4 complexes recruit p300 to target regulatory sequences to specifically acetylate nucleosomal histone H3 to activate the expression of the lefty1 promoter [154]. Studies also suggest that p300/CBP and P/CAF promote TGFβ-dependent transcription through the direct acetylation of Smad2 and Smad3, however, it is not known how acetylation of Smad2 and Smad3 regulates their transcriptional activities [155- 158]. Smad7 acetylation antagonizes its ubiquitin-dependent proteosomal degradation [159], thus, Smad2 and Smad3 might be regulated in a similar fashion. Alternatively, p300/CBP- dependent acetylation of Smad2 and Smad3 might serve to recruit other transcriptional co-factors that function to modulate Smad-dependent gene expression.

In addition to HATs, phosphorylated Smad2 interacts with Brg1, a component of the multi- subunit SWI/SNF complex, an ATP-dependent chromatin remodeling complex [154]. Furthermore, in response to TGFβ signaling, Brg1 is recruited to the lefty1 and nodal promoters,

thus, activated Smad2 likely facilitates this recruitment to enhance the transcription of these two genes [154].

The activator recruited cofactor (ARC) complex (also known as the mediator complex) facilitates transcription by bridging transcription factors to the basal transcriptional machinery, including RNA polymerase II [160]. ARC105, a component of the mediator complex, interacts with Smad2/3 and Smad4 and is recruited to activin/nodal responsive regulatory elements [161]. As well, depletion of ARC105 mimics loss of Smad-dependent Nodal signaling in Xenopus embryos, thus suggesting that Smads are able to promote transcription through their interaction with the ARC/mediator complex [161].

1.2.5.2 Transcriptional co-repressors

Transcriptional repression of Smad target genes occurs through several mechanisms. These include (1) recruitment of histone deacetylases (HDAC) to Smad complexes directly or indirectly through general repressor complexes, (2) disruption of Smad/co-activator complexes and (3) disruption of Smad binding to DNA.

The proto-oncogene products c-Ski and SnoN negatively regulate Smad-dependent transcription

induced by activin/TGFβ signaling primarly by recruiting histone deacetylase (HDAC) activity through their interaction with repressor complexes NCoR/SMRT, Sin3 and MeCP2 complexes [162-166]. Similarly, Dachshund-1, a protein structurally related to c-Ski and SnoN, attenuates TGF-β induced transcription by recruiting NCoR to target genes [167]. The transcriptional repressors Evi-1 and SIP/ZEB2 as well as Smad6 negatively regulate Smad-mediated gene transcription by recruiting the CtBP/histone deacetylase (HDAC) complex to target promoters [151, 168, 169].

TG-interacting factor (TGIF), an atypical three-amino-acid loop extension (TALE) homeodomain protein, inhibits Foxh1-mediated transcription through its association with Smad2/4 complexes [170]. Mechanistically, TGIF mediates repression by interefering with the binding of p300 to Smad2 and promoting the recruitment of the mSin3A/HDAC1 complex to the Foxh1/Smad complex [171, 172]. As well, TGIF may also regulate Smad2-dependent transcription by functioning as an adaptor for the E3 ubiquitin ligase Tiul1 which targets Smad2 for ubiquitin-dependent proteosomal degradation [173].

c-Ski has also been shown to antagonize Smad-dependent transcription by disrupting R- Smad/Smad4 complex formation and preventing Smad DNA binding to target genes as well as competing with Smads for p300/CBP binding [162, 163, 174, 175]. Similarly, E1A, SNIP1, YB- 1, Tax, p53 and Notch1 (ICN1) repress Smad-mediated gene expression by hindering the recruitment of p300/CBP to Smad complexes [176-181]. Like c-Ski, the proteins Tax, YY1, p53, androgen receptors and the human papilloma virus E7 protein as well as the Drosophila brinker protein block Smad-dependent transcription by inhibiting Smad binding to DNA [174, 178, 180, 182-185].

1.3 Regulation of Nodal/Smad signaling pathway

At first glance, Nodal, like other TGFβ family members, appears to induce a linear signaling cascade. However, this apparent simplicity is inconsistent with the broad action of Nodal signaling during embryonic development. To ensure the specificity of Nodal responses in various cell types and at various developmental stages, regulatory mechanisms targeting components at each step of the Nodal signaling pathway function to modulate the duration and intensity of the signaling responses.

1.3.1 Ligand antagonists

Among the numerous extracellular inhibitors that have been described to regulate the action of TGFβ family members, Lefty and Cerberus proteins are the best characterized antagonists of Nodal signal transduction (Figure 1.4E). Lefty proteins are divergent members of the TGFβ superfamily expressed in all chordates. They lack a conserved cysteine residue which is required for the formation of covalently linked dimers and, unlike other family members, Lefty proteins are expressed as monomers [186]. Overexpression of lefty/antivin in zebrafish induces phenotypes that mimic cyc;sqt double mutants and MZoep mutants [187-190]. In contrast, lefty loss-of-function in frog, fish and mouse leads to an expansion Nodal signaling, mesoderm formation and dorsal gene expression [187, 190-199]. At the molecular level, lefty proteins antagonize Nodal signaling by inhibiting receptor complex formation through their interaction with Nodal itself and the EGF-CFC co-receptors [197-199]. Although controversial, there is also evidence that lefty proteins block Nodal signaling through their interaction with ActRIIA and ActRII B receptors [186].

Various members of the Cerberus/DAN family of proteins, which are secreted cysteine knot proteins, also function to limit Nodal signaling during early vertebrate development (Figure 1.4E). Biochemical studies have demonstrated that Xenopus Cerberus (Cer) and Coco, as well as mouse Cerberus-like 2 (Cerl-2) and zebrafish Charon block Nodal signaling through direct binding of Nodal ligands [127, 128, 200, 201]. During gastrulation, Nodal activity is blocked in the anterior region of the mouse epiblast by Cerberus-like 1 (Cerl-1) to promote anterior development [202], whereas Cerl-2 restricts Nodal function to the left side of the embryo at later stages of development for proper establishment of left-right asymmetry [200]. Similarly in zebrafish, knockdown of charon which binds the nodal ligand spaw results in bilateral expression of left-side specific genes [128]. In Xenopus, overexpression of Cer promotes anterior neural development at the expense of mesoderm and endoderm formation in part by inhibiting Nodal signaling [127, 129, 203].

1.3.2 Receptor antagonists

Tomoregulin-1 (TMEFF-1) is a transmembrane protein that inhibits Nodal signaling by physically interacting with the Nodal co-receptor Cripto and preventing its interaction with the type I receptor ALK-4. TMEFF1 is expressed in the animal region of Xenopus gastrulae from midgastrula stages onwards, where it is thought to modulate Nodal activity during neural development [204].

DAPPER2 was originally identified in silico as a human homolog of Xenopus Dapper an inhibitor of Wnt signaling [205]. In zebrafish, however, dapper2 appears to antagonize Nodal signaling as loss of axial mesendodermal tissue and cyclopia induced by dapper2 overexpression is rescued by lefty1 knockdown [206, 207]. Furthermore, lefty1 overexpression reverses the induction of mesodermal tissue formation promoted by dapper2 knockdown [206]. Biochemical studies demonstrated that Dapper2, which is localized to late endosomes, downregulates Nodal signaling by promoting the internalization and lysosomal-dependent degradation of Nodal receptor ALK4 (Figure 1.4F) [206, 207].

1.3.3 Transmembrane-bound antagonist

Nicalin, a transmembrane protein belonging to the aminopeptidase/transferrin receptor superfamily, and its binding protein NOMO inhibit both activin and Nodal signaling in mammalian cell cultures. Accordingly, knockdown of NOMO in zebrafish promotes the expansion of anterior mesoderm and endoderm derivatives at the expense of posterior mesoderm, which is consistent with a loss of Nodal activity [208]. As these proteins are expressed in the membrane of the endoplasmic reticulum (ER), they are thought to inhibit Nodal signaling either by modifying or trapping components of the Nodal signaling pathway in the ER [208].

1.3.4 Smad-associating proteins

In addition to functioning as a transcriptional co-repressor in the nucleus, cytoplasmic SnoN can sequester Smads in the cytoplasm, preventing them from participating in TGFβ-dependent transcriptional activation [209]. In resting cells, Smad2, 3 and 4 are retained in the cytoplasm through their interaction with microtubules and upon receptor stimulation, they are released from the microtubule network to accumulate in the nucleus and activate transcription [210]. The integral nuclear protein Man1 inhibits TGFβ signaling by sequestering R-Smads in the inner nuclear membrane [211, 212]. The nuclear serine/threonine phosphatase PPM1A terminates Nodal signaling by directly dephosphorylating Smad2 and Smad3 and promoting their nuclear export [213] (Figure 1.4H). Supporting a role for PPM1A in limiting Nodal signaling, Ppm1A overexpression in zebrafish downregulates the expression of Nodal target genes gsc and lefty2 [213].

1.3.5 Ubiquitin-dependent regulation of Nodal/Smad pathway components

Ubiquitination is an important mechanism through which the strength and duration of TGFβ superfamily signal transduction are regulated (reviewed [71, 214, 215]). Covalent modification by ubiquitin occurs on lysine residues of target proteins through a three-step process involving ubiquitin-activating (E1), ubiquitin-conjugating (E2) and ubiquitin ligase (E3) enzymes [216]. E3 ubiquitin ligases are generally divided in 3 classes: the HECT (homologous to the E6- associated protein C-terminus)-type class, the RING (really interesting gene)-type class and the finally the U-box-type class [216]. While HECT domain-containing E3 ligases directly catalyze the transfer of ubiquitin to the substrate, RING-domain and U-box type E3 ligases act as molecular scaffolds which facilitate the ubiquitination of target proteins [216].

Smad ubiquitination-related factor 1 (Smurf1) and Smurf2 were the first HECT-type class E3 ubiquiting ligases to be identified as regulators of TGFβ superfamily signaling, and like other

40 members of this class of enzymes, they contain WW domains which bind PY motifs of target proteins. Smurfs control spurious signaling by targeting cytoplasmic pools of R-Smads for proteosomal degradation [217-219]. In addition, Smurf1 and Smurf2 as well as two other E3 HECT-domain ubiquitin ligases, Nedd4-2 and WWP1/Tiul1, terminate signaling by targeting activated Smad2 for proteosomal destruction [173, 218, 220, 221]. Roc-1, a component of the SCFFbwa1/β-TrcP1, interacts with Smad3 and promotes the SCFFbwa1/β-TrcP1 -dependent ubiquitination and degradation of phosphorylated Smad3 in the cytoplasm [222].

Like R-Smads, Smad4 levels are also regulated by HECT-domain E3 ubiquitin ligases such as Smurf1, Smurf2, Nedd4-2 and WWP1/Tiul1 [223]. However, because Smad4 lacks a PY motif, it cannot directly associate with HECT-containing E3 ligases, but rather recruits the enzymes through adaptors such as I-Smads and R-Smads [223]. Overexpression of the Jun-activating domain binding protein 1 (Jab1), a subunit of the COP9 signalosome, promotes the interaction between Smad4 and the Roc1/SCFβTrCP1 complex, resulting in the ubiquitination and proteosomal degradation of Smad4 [224, 225]. Ectodermin, a single subunit RING type E3 ligase, was shown to prevent excessive BMP signaling in the animal pole of Xenopus blastula, allowing for the proper development of ectodermal and neuronal tissues as well as restricting TGFβ/nodal-mediated mesodermal induction to the vegetal hemisphere of the embryo [226]. Interestingly, while polyubiquitination negatively regulates Smad4 activity, monoubiquitination of Smad4 has been shown to promote its transcriptional activity [227].

Smads can also function as adaptors for ubiquitin ligases, recruiting them to various substrates. While the anaphase-promoting complex engages Smad3, Smurf2 associates with Smad2 to mediate the proteosomal degradation of the transcriptional co-repressor SnoN [228-230]. Likewise, I-Smads, Smad6 and Smad7, recruit Smurfs and other HECT-E3 ligases such as Nedd4-2, WWP1/Tiul1, and hAIP4/Itch to TGFβ superfamily receptors complexes to promote their ubiquitin-dependent destruction [173, 220, 221, 231-233]. Interestingly, recruitment of Smad7/Smurf1 to the activin and nodal receptor, ALK4 (ActRIB), is enhanced by FKBP12, an

intracellular inhibitor of TGFβ signaling [234].

The ubiquitin-proteasome pathway also plays an important role in controlling TGFβ receptor trafficking [235]. TGFβ receptors can be internalized via two distinct routes, a clathrin- and a raft/caveolar-dependent pathway each of which has distinct outcomes. Receptors internalized via the clathrin pathway enter the EEA1-positive early endosome, where SARA and c-PML are also enriched, are competent to transduce TGFβ signals [235-239]. In contrast, receptor complexes present in lipid-rich raft domains of the plasma membrane associate with the Smad7/Smurf2 complex and are internalized into a caveolin-positive compartment that leads to the degradation of the receptor complex [235].

Although polyubiquitination of R-Smads by a variety of E3 ligases functions to restrict TGFβ superfamily signaling, emerging evidence also suggests a role for this posttranslational modification in the amplification of TGFβ signaling. Mouse Itch E3 ubiqutin ligase activity promotes TGFβ-dependent Smad2 phosphorylation [240]. Arkadia is a RING-type E3 ubiquitin ligase that enhances Nodal signaling and is essential for organizer formation as well anterior primitive streak development in frog and mouse [241, 242]. At the molecular level, Arkadia enhances signaling by targeting I-Smad, Smad7, as well as transcriptional co-repressors SnoN and c-Ski for proteosomal degradation [241-244]. Arkadia also promotes Nodal/TGFβ signaling by coupling phospho-Smad2/3 activity and degradation [245]. This mechanism may be necessary to clear gene regulatory elements, allowing them to respond to new incoming signals [245].

1.4 Regulation of Transcription by Sumoylation

1.4.1 Small ubiquitin-like modifier

The small ubiquitin-like modifier (SUMO) proteins are a family of small polypeptides conserved

from yeast to human, consisting of four members (SUMO1, SUMO2, SUMO3 and SUMO4) that are distantly related to ubiquitin. While SUMO1, SUMO2 and SUMO3 are expressed in most tissues, SUMO4 is mainly in expressed in kidney [246, 247]. SUMO proteins, like ubiquitin, are covalently linked to their targets through an isopeptide bond between the C-terminal carboxyl group of the SUMO moiety and the ε-amino group of lysine residues on substrate proteins. Although the majority of SUMO substrates are modified on lysine residues present within a ψKxE/D motif (where ψ is a large hydrophobic residue, K is the lysine, x is any amino acid and E/D is either glutamic acid or aspartic acid), several proteins, including Mdm2, Daxx, CREB and CtBP-2, are modified at sites which do not conform to the consensus sequence ψKxE/D [246].

1.4.2 Sumoylation Pathway

SUMO conjugation, like ubiquitination, occurs through an ATP-dependent three-tier enzymatic cascade, involving a heterodimeric E1 SUMO activating enzyme (Aos1/Uba2), an E2 SUMO conjugating enzyme (Ubc9), and an E3 SUMO ligase (Figure 1.6) [246]. So far, four types of E3 SUMO ligases have been described. The largest group of SUMO E3 ligases consists of the PIAS (protein inhibitor of activated STAT) family of proteins, which include the yeast Siz1, Siz2, and MMS21 proteins, the mammalian PIAS 1, 3 Xα, Xβ, γ, and the more distantly related hZIMP7

and hZIMP10 proteins [248]. The three other SUMO E3 ligases include the nucleoporin RanBP2, the polycomb group protein Pc2 and the topoisomerase I binding protein TOPORS [249-256]. While structurally unrelated to each other, SUMO E3 ligases appear to facilitate the transfer of SUMO moieties by stabilizing the interaction between SUMO-bound Ubc9 and target substrates [246]. As with other post-translational modifications, sumoylation is a reversible process that is mediated by desumoylases known as SUMO/sentrin-specific proteases (SENP) [246]. In mammals, seven SENP proteins have been described and several of these have been shown to also be required for the processing of SUMO precursor proteins to mature active forms [246].

1.4.3 Targets and transcriptional function of sumoylation

Although sumoylation has been implicated in the regulation of a wide range of cellular processes, including signal transduction, nucleocytoplasmic shuttling, chromosome assembly and segregation, and maintenance of genomic integrity, the vast majority of SUMO substrates are proteins involved in transcriptional regulation [246, 257, 258]. In most cases, sumoylation promotes transcriptional repression, and several molecular mechanisms through which this occurs have been described so far. Sumoylation of the NF-κB inhibitor IκBα promotes its proteolytic stability by blocking ubiquitin-dependent proteosomal degradation, which consequently results in the inhibition of NF-κB-dependent transcription [259]. Conjugation of SUMO moieties to lysine residues within the synergy control motifs of c-Myb, Sp3, C/EBP and several nuclear receptors (AR, GR, and PR) has been shown to block their transcriptional activity [260]. Similarly, transcriptional repression by Elk-1 and CBP/p300 is mediated by the sumoylation of lysine residues within their negative regulatory domains which promotes the recruitment of HDAC2 and HDAC6, respectively [261, 262]. Interestingly, sumoylation also enhances transcriptional repression by potentiating the enzymatic activity of HDAC1 and HDAC4 [263, 264]. As well, SUMO modification inhibits transcriptional activity by interfering with the recruitment of transcription factors to target promoters. The inhibitory effect exerted by sumoylation on the transcriptional activity of the Glial Cell Missing Drosophila homolog a (GCMa), a transcription factor critical for placental development, is attributed to decreased DNA binding by GCMa [265].

Several sumoylated transcription factors, including p53, LEF1, SRF, and IRF-1 are targeted to promyelocytic leukemia nuclear bodies (PML-NB) [246, 258]. These subnuclear structures are thought to serve as nuclear storage sites or specific active sites for transcription factor modification and assembly. Although the exact function of PML-NB is ambiguous,

Figure 1.6: The SUMOylation pathway

(A) SUMO proteins are synthesized as precursor proteins which are processed by SUMO proteases (SENPs) to make the double glycine (GG) motif available for conjugation. (B) Mature SUMO proteins are conjugated to the heteromeric SUMO E1 activating enzyme via a thioester bond in an ATP-dependent reaction. (C) The SUMO moiety is transferred to an active cysteine of the SUMO E2 conjugating enzyme (Ubc9) forming a thioester bond. (D) A SUMO E3 ligase (PIAS, RanBP2, PcG or Topors) facilitates the transfer of the SUMO peptide from Ubc9 onto its target substrate via an isopeptide bond with ε-amino groups of target lysines (K). (E) Desumoylation of substrates occurs via SENPs desumoylases.

recruitment of sumoylated transcription factors into these structures is thought to regulate their transcriptional activity [246, 258]

Sumoylation is most commonly associated with transcriptional repression, emerging, however, evidence also suggests that SUMO modification can also enhance transcriptional activation [257, 258]. Sumoylation of HSF1, p45/NF-E2, Oct-4 promotes their transcriptional activity by enhancing their DNA binding activity [266-268]. In hypoxic conditions, HIF1α is modified by SUMO-1, promoting its stabilization and transcriptional activity [269]. Sumoylation of the transcriptional repressor ZNF76 inhibits its interaction with the TATA-binding protein (TBP), which results in derepression of transcription [270]. Thus, as with transcriptional repression, sumoylation appears to modulate transcriptional activation through various molecular mechanisms.

1.4.4 Regulation of Smad activity by sumoylation

Sumoylation of Smad4, which can be mediated by several PIAS family members, has been shown to enhance TGFβ-dependent transcriptional activation by promoting Smad4 protein stability and nuclear accumulation [271-274]. In contrast, there is also evidence to show that sumoylation decreases the ability of Smad4 to transactivate the artificial GAL4 promoter, suggesting that regulation of Smad4 activity by SUMO modification may be promoter specific [275]. Interestingly, Smad4 sumoylation appears to be regulated by TGFβ-induced activation of the p38/MAP kinase pathway [276]. PIASγ has also been shown to modify Smad3 [277], therefore, the contradictory effects resulting from Smad4 SUMOylation may be in part explained by simultaneous SUMOylation of Smad3.

1.5 Regulation of the paired-like homeodomain transcription factors Goosecoid and Mix-like proteins by Nodal signaling.

Nodal signaling is essential for gastrulation to proceed normally in all vertebrate species studied. Accordingly, numerous genes activated downstream of the Nodal signaling pathway regulate embryonic axis formation as well as mesoderm and endoderm development in all vertebrate species studied [23]. Several known targets of Nodal signaling are transcription factors belonging to the paired-like homeodomain family, including Goosecoid and several members of Mix-like family of transcription factors [23].

1.5.1 Goosecoid in embryogenesis and cancer

The goosecoid (gsc) gene was originally isolated from a Xenopus dorsal blastopore lip cDNA library screen aimed at identifying genes involved in the induction of the Spemann Organizer [278]. The goosecoid protein is a member of the paired-like homeodomain family and was named as such to reflect the homology of its homeodomain to that of Drosophila proteins Gooseberry and Bicoid [279]. Goosecoid homologues have since been cloned in hydra, fly, mouse, zebrafish, chick, and human [280-286] (Figure 1.7). A Gsc-related gene, Gsc-2 has also been identified in human and mouse[287, 288].

In Xenopus, Gsc expression is first detected at the onset of gastrulation in the dorsal marginal zone (DMZ), marking the site of dorsal blastopore lip formation [289]. As gastrulation proceeds, gsc-expressing cells involute and move away from the dorsal lip and ultimately populate the prechordal plate and anterior mesendoderm in late gastrula stages [289]. Similarly, in zebrafish, Gsc transcripts are maximally expressed in the embryonic shield at 50% epiboly, at which time gastrulation begins in zebrafish [283, 284]. Gsc-expressing cells subsequently populate the anterior end of the axial hypoblast, which gives rise to the zebrafish prechordal plate and anterior mesendoderm [283, 284]. Miroinjection of gsc mRNA in the ventral blastomeres of 4-cell stage

Xenopus embryos induces the formation of secondary axis, thus, supporting a role for Gsc in axis specification [289, 290]. Overexpression studies have indicated that Gsc dorsalizes the ventral marginal zone in a dose-dependent manner [291]. Loss-of-function studies indicate that Gsc and the ventralizing genes Vent1 and Vent2 reciprocally repress each other to achieve proper DV mesodermal patterning [292]. Antagonism of Gsc activity through the expression of antimorphic gsc proteins as well as anti-sense Gsc mRNA have also shown a requirement for Gsc function in head formation [290, 293, 294]. In addition to its patterning function, Gsc has been shown to promote the dorso-anterior migration in Xenopus embryos, suggesting a role for Gsc in gastrulation movements [295].

In mouse, Gsc expression marks the onset of gastrulation and is first detected in the posterior epiblast at the site of primitive streak formation [282]. At mid-streak stages, when Gsc expression reaches its peak, transcripts are detected in both the anterior primitive streak as well as the anterior visceral endoderm [282, 296]. In late gastrulation stages, gsc expression is restricted to anterior mesendoderm, more specifically in the prechordal plate and foregut, and the ventral diencephalon [297, 298]. Given that Gsc expression is conserved in the organizer as well as anterior mesendoderm in all vertebrates examined, Gsc function was thought to be essential for axis specification as well axial mesoderm patterning across species [279]. However, functional studies demonstrated that Gsc-null mice are born alive and show no overt gastrulation or axial midline defects [299, 300]. The lack of gastrulation abnormalities in mice suggests that functionally redundant genes may be compensating for the loss of Gsc activity [279]. Foxa2/HNF3β, a forkhead transcription factor co-expressed with Gsc in the anterior primitive streak, has been to shown to partly compensate for loss of Gsc activity during neural patterning [298]. While Gsc-null mice do not exhibit abnormalities related to gastrulation, they die shortly after birth and display various craniofacial as well as skeletal defects, consistent with a second phase of Gsc expression in neural crest cells, limb buds and branchial arches [299, 300].

Structurally, Gsc contains a paired-like homeodomain at its C-terminus and a conserved Goosecoid-Engrailed Homology (GEH) domain at its N-terminus (Figure 1.7) [278, 301, 302]. Subclasses of the paired-like homeodomain family are defined according to the residue present at position 50 of the homeodomain, which defines the DNA binding specificity [303]. Gsc contains a lysine residue at position 50 of its homeodomain, and, thus, exhibit the DNA binding specificity of the Bicoid subclass of the paired-like homeodomain family [279]. Oligonucleotide-binding selection studies have described the near-palindromic DNA sequence TAATCCGATTA as the optimal Gsc binding site [304]. At the molecular level, Gsc functions to repress transcription of its own promoter as well as those of the Xenopus Brachyury and Wnt8 promoters through direct DNA binding to paired homeodomain binding sites [290, 305, 306].

In Drosophila, transcriptional repression is mediated through the recruitment of the Groucho co- repressor by the Gsc GEH domain [301]. The Drosophila Groucho (Gro) protein is the prototypic member of a large family of corepressors widely expressed among vertebrate and invertebrate species [307]. Gro homologues include transducin-like Enhancer of split (TLE) proteins, groucho-related gene (Grg) proteins, and the amino Enhancer of split (AES) proteins [307]. Gro/TLE family members can function as corepressor by several different mechanisms, including direct interaction with the basal transcriptional machinery, recruitment of HDACs and impairment of co-activator access [308-310].

Although Gsc expression is primarily induced by Nodal/Foxh1 signaling in mouse [82, 119, 120], in Xenopus, Gsc expression is regulated by both the Wnt/β-catenin and Activin/Nodal signaling pathways through distinct promoter elements [112, 311]. A proximal element mediates the β-catenin-dependent induction through binding of the homeodomain proteins Siamois or XTwn, whereas a distal element mediates the Activin/Nodal induction through binding of DNA binding complex containing Mixer and Smad2/4 [112, 311].

Figure 1.7: The Gsc family of proteins.

(top) Schematic representation of the general structure of Gsc proteins. The paired-like homeodomain which is responsible for DNA binding is depicted in blue, the Goosecoid Engrailed Homology domain is depicted in red. (bottom) A midpoint rooted phylogenetic tree constructed with the amino acid sequences of described Gsc proteins using ClustalW alignment and a best tree neighbor-joining method. Branch lengths are indicated and represent uncorrected distance between tree nodes. Dm, Drosophila melanogaster; Mm, Mus musculus; Hs, Homo sapiens; Dr, Danio rerio; Gg, Gallus gallus; Xl, Xenopus laevis; Hv, Hydra vulgaris

In addition to playing a key role during early embryonic development, Gsc has recently been implicated in the development of cancer. GSC mRNA expression was found to be upregulated in the vast majority of human ductal-type breast tumors when compared to patient-matched normal tissues [312]. Stable Gsc expression was also found to elicit the epithelial-to-mesenchymal transition of immortalized human mammary cells and promote the migration of human MDA- MB-231 breast cancer cells [312]. Furthermore, Gsc expression enhanced the ability of MDA- MB-231 cells to form pulmonary metastasis in mice [312]. Interestingly, TGFβ signaling which is known to promote metastasis was also found to induce Gsc expression in human breast cells [312]. As members of the TGFβ superfamily are known to regulate Gsc expression during early development, it is likely that TGFβ signaling also reactivates other embryonic genes to promote the development of neoplastic disease

1.5.2 Mix-like transcription factors

The founding member of the Mix-like family of transcription factors, Mix1, was originally identified in Xenopus as an activin early response gene capable of inducing ventral mesoderm and endoderm [313]. So far, fourteen Mix-related genes have been described, including seven genes in Xenopus (Mix1, Mix2, Mix3/Mixer, Mix4/Bix1, Bix2/Milk, Bix3, Bix4), four in zebrafish

(bonnie and clyde (bon)/mixer, mezzo, mtx1 and mtx2) and one gene each for mouse (Mixl1/mml), chick (CMIX) and human (MIXL1) [313-327]. Mix-like proteins share a conserved paired-like homeodomain and a conserved carboxy-terminal acidic region (CAD) with predicted transcriptional activity (Figure 1.8) [324, 328].

Xenopus Mix-like genes are transiently expressed during early development from mid-blastula transition to the end of gastrulation. In late blastula stages, Mix-like transcripts are initially detected in the marginal zone where presumptive mesendoderm is formed, and by late gastrula stages, they become restricted to the endoderm [314-317, 319, 329]. In zebrafish, mixer, mezzo

and mtx2 transcripts are first detected at the sphere stage in a patch of dorsal blastomeres in the marginal region, and as epiboly begins, expression of Mix-like genes extends over the whole blastoderm marginal zone in both presumptive mesoderm and endoderm [321-323, 330]. Transcript levels begin to decline soon after the shield stage (50% epiboly) and are undetectable by late stages of gastrulation (70% epiboly). In addition, mixer, mtx1 and mtx2 expression is also detected in the YSL, an extraembryonic tissue known to play an important role in the induction of endoderm [321, 323, 330]. In mouse, Mixl1 mRNA expression is first detected throughout the visceral endoderm in late blastula stages (5.5 dpc) [324, 327]. As the blastula develops, Mixl1 becomes asymmetrically expressed in the posterior epiblast at the embryonic/extraembryonic margin. As gastrulation begins, Mixl1 expression marks the nascent primitive streak and the emerging mesoderm. While this expression pattern is maintained throughout streak elongation, by early bud stages (~E8.0), Mixl1 expression is gradually restricted to posterior primitive streak where it is maintained until late head fold stages (E8.5) [324, 327]. Mixl1 transcripts are noticeably absent from the anterior primitive streak during streak elongation as well as the definitive node [324, 327]. Similar expression patterns are observed in the chick embryo, with CMIX expression marking primitive streak and nascent mesoderm during gastrulation but not Hensen’s node [325, 326].

In all species studied, the spatial and temporal expression patterns of Mix-like genes imply a pivotal function for these genes in mesoderm and endoderm development. Functional studies in frog, fish and mouse have in fact demonstrated an essential role for these genes in the specification and patterning of mesoderm and endoderm. Overexpression of Mix3/Mixer, Bix2/Milk, and Bix4 induces the formation of endoderm in explanted animal caps which typically gives rise to ectoderm [315, 317-319]. Interestingly, high levels of Bix1 mis-expression induces endoderm development, while low levels cause ventral mesodermal differentiation [316]. Endoderm development is severely impaired during gastrulation in zebrafish bon mutants as revealed by the decrease of sox17- and foxa2/axial-expressing endodermal cells [320, 321]. Although depletion of mezzo alone by morpholino knockdown does not cause overt endoderm

defects, loss of mezzo expression in bon mutants completely abolishes sox17 expression [322]. Moreover, overexpression of mezzo mRNA partially rescues the loss of sox17-expressing cells in bon mutants, thus, suggesting that mezzo and mixer/bon share partially redundant functions in endoderm formation [322].

In addition to its role in endoderm formation, mixer/bon cooperates with the Nodal signaling pathway components sqt and sur/foxh1 to regulate the patterning of axial mesendoderm patterning. Genetic studies have demonstrated that double bon-/-;sqt-/- and bon-/-;sur-/- mutants display axial mesendoderm defects which are much more severe than in single mutants, including lack of prechordal plate and cardiac mesoderm [138]. Moreover, bon controls anterior neural patterning through the transcriptional regulation of mesendodermal gene expression [331]. Mixl1-/- mutant mice embryos display a thickened primitive streak and lack of morphologically distinct node. While molecular marker analysis demonstrated that germ layer derivatives are properly specified in Mixl1-/- mutant mice, abnormalities in notochord, heart, gut and neural tube formation suggest that Mixl1 activity is required for proper tissue and organ morphogenesis [332].

Consistent with phenotypes induced in functional studies, ectopic expression of Mix3/Mixer, Bix2/Milk, and Bix3 in the marginal zone has been shown to induce endodermal gene expression while repressing mesodermal gene transcription [315, 317, 333]. Moreover,

Mix3/Mixer loss-of-function results in the decreased expression of endodermal genes Sox17α, Gata-5 and Endodermin and enhanced expression of several mesodermal genes, including Fgf3, Fgf8 and Eomesodermin [334]. Mix1 has been shown to synergize with siamois to induce endodermal gene expression, and cooperate with goosecoid to repress the expression of the pan- mesodermal marker Xbra [335, 336]. Expression of several Mix-like genes including Mix1, Mix2, Mix3, Mixer, Bix4, bon, mezzo and Mixl1 are regulated by Nodal-related signaling [104, 106, 138, 314, 322, 330, 331, 337]. Biochemical studies have in fact confirmed that Foxh1 and

Figure 1.8: The Mix-like family of proteins

(top) Schematic representation of the general structure of Mix-like proteins. The paired-like homeodomain which is responsible for DNA binding is depicted in green, the Smad interaction motif (SIM) is depicted in red and the C-terminal acid domain (CAD) is depicted in orange. (bottom) A midpoint rooted phylogenetic tree constructed with the amino acid sequences of described Mix-like proteins using ClustalW alignment and a best tree neighbor-joining method. Branch lengths are indicated and represent uncorrected distance between tree nodes. Dm, Drosophila melanogaster; Mm, Mus musculus; Hs, Homo sapiens; Dr, Danio rerio; Gg, Gallus gallus; Xl, Xenopus laevis.

Smads not only directly regulate Mix2 expression in Xenopus but also Mixl1 in mouse ([99, 106, 338] and this thesis). In addition, like Mixer, Bix2/Milk, Bix3 and bon/mixer can interact with Smad2 through a conserved PPNK motif to mediate TGFβ/activin signaling [112, 339]. Structure-function studies demonstrated that the Smad2 interaction motif, the CAD as well as the homeodomain present in Mixer/Mix3 are all critical for endoderm formation [340]. As well, while DNA binding was required for both endoderm induction and mesoderm repression, the CAD was only required for endoderm induction [340]. Although the molecular mechanisms are yet to be defined, Mixer and Mixl1 loss-of-function results in the enhancement of Nodal signaling in vivo [332, 334]. Thus, Mix-like genes and Nodal-related genes appear to function in an intricate transcriptional regulatory network to regulate mesendodermal specification [341].

1.6 Thesis Perspectives

The goal of my doctoral dissertation has been to gather a better understanding of the transcriptional mechanisms downstream of Nodal signaling pathway which serve to regulate the expression of genes essential for early embryonic development. Previous studies have demonstrated that the forkhead/winged helix transcription factor Foxh1 is a major effector of the Nodal signaling pathway and typically functions as a transcriptional activator. In contrast to

these studies, our analysis of Foxh1-mutant mice revealed a role for Foxh1 in the transcriptional repression of the Mixl1 gene during early embryonic development. To delineate the molecular mechanism underlying this repressive activity, I undertook a series of biochemical studies and demonstrated that Foxh1 recruits the paired-like homeodomain protein Gsc to negatively regulation Mixl1 expression. While Gsc has previously been shown to repress transcription through direct DNA binding and recruitment of the Groucho co-repressor [301, 305, 336], this work revealed a distinct mechanism whereby Gsc associates with DNA-bound Foxh1 and mediates the recruitment of HDACs to repress Mixl gene expression.

During the course of these studies, we observed that Gsc was post-translationally modified in mammalian cells, I was thus interested in identifying the nature of this modification and determining whether it was required for Gsc repressive activity. Using a molecular approach, I demonstrated that Gsc was covalently modified by the small-ubiquitin modifiers SUMO-1, SUMO-2 and SUMO-3. Moreover, our studies suggested that regulation of Gsc repressive activity by sumoylation may serve to differentially regulate gene expression during embryonic development and neoplastic disease.

Taken together, the work presented in this dissertation provides insight into mechanisms regulating transcriptional events occurring during mouse gastrulation, and further supports a role for embryonic genes in the development of cancer.

CHAPTER 2

2 Foxh1 recruits Gsc to negatively regulate Mixl1 expression during early mouse development

Note: The work included in this chapter was published in the following manuscript:

Izzi L., Silvestri C., von Both I., Labbé E., Zakin L., Wrana J.L., and Attisano, L. (2007) Foxh1 recruits Gsc to negatively regulate Mixl1 expression during early mouse development. EMBO J. 26(13):3132-43.

L. Izzi compiled Fig. 2.1 and generated Figs. 2.2, 2.3, 2.6, 2.7, 2.8, 2.9, 2.10, 2.11, 2.12, 2.13 and 2.14

C. Silvestri generated Fig. 2.5A, contributed to Fig.2.3D and 2.6C and generated the Flag-Foxh1 cDNA construct used in Fig. 2.7B

I. von Both generated Fig. 2.5B

E. Labbé generated the Gsc cDNA constructs used in Fig. 2.7C,D, Fig. 2.9A and Fig. 2.10B,C

L. Zakin set up breddings and dissected mouse embryos used in Fig. 2.6C.

L. Attisano generated Fig. 2.4

2.1 Abstract

Mixl1 is a member of the Mix/Bix family of paired-like homeodomain proteins and is required for proper axial mesendoderm morphogenesis and endoderm formation during mouse development. Mix/Bix proteins are transcription factors that function in Nodal-like signaling pathways and are themselves regulated by Nodal. Here, we show that Foxh1 forms a DNA binding complex with Smads to regulate TGFβ/Nodal-dependent Mixl1 gene expression. While Foxh1 is commonly described as a transcriptional activator, we observed that Foxh1-null embryos exhibit expanded and enhanced Mixl1 expression during gastrulation, indicating that Foxh1 negatively regulates expression of Mixl1 during early mouse embryogenesis. We demonstrate that Foxh1 associates with the homeodomain-containing protein Goosecoid (Gsc), which in turn recruits histone deacetylases to repress Mixl1 gene expression. Ectopic expression of Gsc in embryoid bodies represses endogenous Mixl1 expression and this effect is dependent on Foxh1. Since Gsc is itself induced in a Foxh1-dependent manner, we propose that Foxh1 initiates positive and negative transcriptional circuits to refine cell fate decisions during gastrulation.

2.2 Introduction

During gastrulation the three germ layers, ectoderm, mesoderm and endoderm, are generated and in mice, this is first apparent with the formation of the primitive streak at the posterior end of the epiblast near the embryonic/extraembryonic junction [342]. The primitive streak elongates towards the distal end of the epiblast and cells move through the streak to emerge as mesoderm and endoderm. The organizer or node, which is located at the anterior end of the primitive streak, is a specialized population of cells with axis inducing abilities that also contributes to the formation of axial mesendoderm, prechordal mesoderm and notochord [6].

Numerous growth factors regulate gastrulation, including Transforming Growth Factor β (TGFβ) superfamily members such as Nodal and Activin [23]. Nodal, like other TGFβ ligands, initiates signaling through its interaction with a heteromeric complex of serine/threonine kinase receptors, but also requires the presence of the EGF-CFC co-receptors, Tdgf1 (Cripto) or Cfc1 (Cryptic) [23, 63, 68]. Following activation of the receptor complex, the signal is transmitted to Smad proteins which translocate to the nucleus and interact with specific transcription factors to regulate the expression of target genes [63, 68]. Loss- and gain-of-functions studies in frog, fish and mice have demonstrated that several components of the Nodal signaling pathway are implicated in the establishment of mesoderm and endoderm as well as axial mesendoderm patterning [23].

Several transcription factors are targeted by Nodal signaling and function to regulate gastrulation. For example, Mix/Bix paired-like homeobox genes including Xenopus Mix2, zebrafish og9x/mezzo and mouse Mixl1 act downstream of Nodal-like signaling pathways to regulate both mesoderm and endoderm formation [111, 322, 332]. In mice, Mixl1 is expressed in the primitive streak and emerging mesoderm at the onset of gastrulation and is then restricted to the posterior primative streak at the early bud stage [324, 327]. The important role played by Mixl1 during mouse development is revealed by the numerous defects displayed by Mixl1-/-

59 mutant embryos, which include enlarged primitive streak, abnormal anterior midline structures, absence of heart tube formation and defective gut morphogenesis.

Goosecoid (Gsc), another paired-like homeodomain-containing transcription factor whose expression is regulated by Nodal-like signaling pathways, functions in gastrulation as well as in axial mesendoderm formation [14, 282]. In mice, Gsc expression marks the onset of gastrulation and is first detected in the primitive streak. As gastrulation proceeds, Gsc expression is restricted to the anterior primitive streak and the anterior visceral endoderm [282, 296]. Although its role in early development remains unclear as Gsc-null mice present no overt gastrulation or axial midline defects [299, 300], Gsc can function as a transcriptional repressor by directly binding to paired homeodomain binding sites to repress transcription of its own promoter as well as those of other genes including Xenopus brachyury (Xbra) and wnt8 (Xwnt8) [290, 305, 306]. The N- terminus of Gsc comprises a conserved Goosecoid Engrailed Homology (GEH) domain, and in Drosophila, the domain has been shown to interact with Groucho and mediate the repressive activity of Gsc [301].

A key mediator of the Nodal signaling pathway is the forkhead/winged-helix transcription factor, Foxh1 [121]. Genetic ablation of Foxh1 in mice results in embryonic lethality and a range of defects, including a total lack of embryonic development, aberrant anterior primitive streak patterning, loss of anterior and midline structures and failure to form definitive endoderm [119, 120]. In Xenopus and zebrafish, loss of Foxh1 activity results in anterior and axial defects as well as aberrant mesoderm development [23]. Thus, loss of Foxh1 activity mimics numerous phenotypes observed in mutants where components of the Nodal signaling pathway have been disrupted. In general, Foxh1 binds directly to DNA and cooperates with Smad2/4 complexes to activate Nodal-dependent expression of target genes such as the TGFβ family members, nodal and Lefty2 and the homeobox factors, Mix2, Gsc and Pitx2 [82, 102, 111, 118]. However, Foxh1-dependent induction of the Mef2c gene in the anterior heart field also requires cooperation with Nkx2-5, a heart-specific homeodomain transcription factor [105]. Here, we demonstrate

that recruitment of Gsc to the Mixl1 promoter by Foxh1 represses Mixl1 expression. Thus, our work reveals that Foxh1 can function either positively or negatively to control target gene expression and we propose that this precise control of gene expression contributes to cell fate determination during gastrulation.

2.2.1 Materials and Methods

2.2.2 Reporter constructs and transcriptional reporter assays

The Mixl1 promoter (position -390 to -143 from the translational start site) was amplified by PCR from a mouse Mixl1 genomic DNA clone (kindly provided by Jonathan J. Pearce) and subcloned into the SacI/BglII sites of a modified pGL2-promoter vector (Promega), as described [82]. All mutant Mixl1-luc reporters were generated by overlap PCR mutagenesis using primers indicated in Table 2.1. For luciferase assays, HepG2 cells were transiently-transfected using the calcium phosphate DNA precipitation method, as described previously [82]. Unless otherwise indicated, transfections contained 0.083 µg of Mixl1-luc reporter, 0.0035 µg of Foxh1, 0.035 µg of each Smad, 0.25 to 250 ng of Gsc or its derived mutants, 0.1 µg of pCMV-βgal and pCMV5 empty vector to a total of 1 µg per well in a 24-well dish. Trichostatin A (Sigma) in 100% ethanol was added as indicated 18 hours prior to lysis.

2.2.3 Construction of mammalian and bacterial expression vectors

The mature ligand region of Nodal was amplified by PCR and subcloned into a pCMV5 vector containing the Activin pro-region. Carboxy-terminal Flag-tagged mouse Cripto was generated

by PCR using NIA clone H3029H05 as template. Epitope-tagged Gsc full-length and deletion mutants in pCMV5B were generated by PCR using mouse Gsc cDNA and subcloned in pGEX4T1 for bacterial expression. Flag-Foxh1 ∆N where the first 183 amino acids were deleted was generated by PCR. GST-HDAC1 was generated by PCR using HDAC1 cDNA and subcloned pGEX4T1 for bacterial expression. All primers are listed in Table 2.1.

Table 2.1 List of primer used

Primer Name Sequence

Primers for reporter constructs

Mixl1For 5’-CCG GAG CTC ACG AAC CAA GCC CCC AAG-3’

Mixl1 Rev 5’-CCG AGA TCT GCC AGA CGC CGC GGG AAT-3’

Mixl1 mF For 5’-AGG TAT TTA GAT TGG TGG TTG GAC-3’

Mixl1 mF Rev 5’-CCA CCA ATC TAA ATA CCT CGA ACC GC-3’

Mixl1 SBEm1 For 5’-GGT GGT TGA ATT AGA TAT CGG ATG GGC GGG-3’

Mixl1 SBEm1 Rev 5’-CCC ATC CGA TAT CTA ATT CAA CCA CCA ATC C-3’

Mixl1 SBEm2 For 5’-TCG GAT GAG TGG AGA AGG GAC GGA-3’

Mixl1 SBEm2 Rev 5’-CTT CTC CAC TCA TCC GAT ATC TA-3’

Mixl1 ∆SBE Rev 5’-CCG AGA TCT ACC AAT CCA CAT ACC TCG-3’

Primers for expression constructions

Gsc fl For 5’-CCG GTC GAC CCC GCC AGC ATG TTC AGC-3’

Gsc fl Rev 5’-CCG GGA TCC TCT AGA TCA GCT GTCCGA GTC CAA ATC-3’

Gsc 1-220 Rev 5’-CCG GGA TCC TTA CTT CTG TCG TCT CCA CTT GGC TC-3’

Gsc 1-143 Rev 5’-CCG GGA TCC GCT CAC GTG CCC ACC GTT CAT GTA-3’

Gsc 1-116 Rev 5’-CCG GGA TCC GCG GCC GCT CAC GTC GGG ACG CAG GAG CA-3’

Gsc 1-96 Rev 5’-CCG CTG CAG GCC CAC GGG CGC CGC CTG CAC-3’

Gsc 1-63 Rev 5’-CCG CTG CAG AGC CGC GGC GCT GGG CGC CA-3’

Gsc ∆14 For 5’-CCG GTC GAC GC CGG CCG CGC TGC AAA GAC-3’

Gsc ∆81 For 5’-CCG GTC GAC AGC TAC TTC TAC GGG CAG-3’

Gsc ∆96 For 5’ CCG GTC GAC GC CCG GCT TGC TGC GGG GCT-3’

Gsc ∆111 For 5’-CCG GTC GAC GC TCC TGC GTC CCG ACG C-3’

Gsc ∆143 For 5’- CCG GTC GAC GCC TGT CGC GCA CTG AGC TG-3’

Gsc HD For 5’-CCG GTC GAC GGA AGC GGC GGC ACC GCA CCA TC-3’

Primer Name Sequence

Primers for expression constructions

Gsc HD Rev 5’-CCG GGA TCC TTA CCG CCG ACA GTG CAG CTG GTT-3’

Gsc NG (N210G) For 5’-CGC CGA GCC AAG TGG AGA CGA C-3’

Gsc NG (N210G) Rev 5’-TCT CCA CTT GGC TCG GCG GCC CTT AAA CCA GAC CTC-3’

Foxh1 ∆N For 5’-CC GGT CGA CGC CAG CCT TAT CAG CCA CCC AG-3’

Foxh1 ∆N Rev 5’-CCG TCT AGA GGA TCC TTA CAT GCT GTA CCA GGA AAG G-3’

HDAC1 For 5’-CCG GTC GAC TGG CGC AGA CGC AGG GCA CC-3’

HDAC1 Rev 5’-CCG GCG GCC GCG GAT CCT TAG GCC AAC TTG ACC TCC TCC TTG A-3’

DNA and Chromatin IP QPCR primers

Mixl1 ChIP/DNA IP For 5’- GAG GTA TGT GGA TTG GTG GTT GGA-3’

Mixl1 DNAIP Rev 5’-CCG AGA TCT GCC AGA CGC CGC GGG AAT-3’

Mixl1 ChIP Rev 5’-AAT GAG GGA GGC GCG AAC TTG A-3’

Mixl1 ctl For 5’-TGG CAG AGG ACA GTG ATG GAC AAA-3’

Mixl1 ctl Rev 5’-AAA CCA GCC TGA GAA GAC ACC AGA-3’

QPCR primers for mRNA expression

Mixl1 cDNA For 5’-AAC CGA CGG GCC AAG TC-3

Mixl1 cDNA Rev 5’-TCC CCG CCT TGA GGA TAA G-3’

Hprt For 5’-AAA CAA TGC AAA CTT TGC TTT CC-3’

Hprt Rev 5’-GGT CCT TTT CAC CAG CAA GCT-3’

Genotyping primers

NeoF2 5’-GAG GAT CTC GTC GTG ACC CAT GG-3’

GF1 5’-CAG ATG CTG CCC TAC ATG AAC GTG G-3’

GR1 5’-GGC GTT TTC TGA CTC CTC CGA GG-3’

2.2.4 Gel-shifts, immunoprecipitation, GST pull-downs and immunoblotting

Electrophoretic mobility shift assays were performed as described [82] except that cell extracts or bacterial fusion proteins were incubated with 30 000 cpm [32P]-labeled DNA prior to non- denaturing electrophoresis.

For immunoprecipitations and GST pull-downs, COS-1 cells were transfected using the polyethylenimine (PEI; Sigma-Aldrich catalog # 408727) method. Briefly, for a 100 mm dish, 25 µl of 2 mg/ml of PEI stock was added to 10 µg of DNA made up to 750 µl in serum-free medium. The DNA/PEI solution was vortexed, incubated at room temperature for 5 min and added to cells containing 10 mL of growth medium. Immunoprecipitations, GST pull-downs and immunoblotting were carried out as reported [82] using M2 anti-Flag (Sigma) or anti-T7 (Novagen) monoclonal antibodies.

For DNA immunoprecipitations, COS-1 cells were transfected with the appropriate cDNA and Mixl1-luc reporter using LipofectAMINE (Invitrogen), and immunoprecipitations were performed as described [105] with the following modifications. For double DNA immunoprecipitation, protein-DNA complexes were immunoprecipitated with 2 µg of Flag antibody or T7 antibody, followed by incubation with protein G Sepharose for 1h. Protein-DNA complexes were eluted by incubation with 1% SDS, rediluted to a final concentration of 0.1% SDS, and subjected to immunoprecipitation with 2 µg of T7 or Flag antibody overnight. For triple DNA immunoprecipitations, lysates were sequentially immunoprecipitated with 2 µg of T7 antibody for 3 hours, followed by an overnight immunoprecipitation with 2 µg of Flag antibody, and a 3-hour immunoprecipitation with 2 µg of Y11 anti-HA antibody (Santa Cruz). Protein- DNA complexes were eluted each time with 1% SDS. For endogenous ChIP, EB cultures were established as previously described [105] with the following modifications. ES cells were suspended in 0.05% trypsin and differentiated in bacterial grade Petri-dishes. After 5 days of

differentiation, EBs were collected and ChIPs were carried out as described [105] with the following modification. EBs were fixed in 1% formaldehyde for 20 minutes at room temperature. Protein G sepharose beads were blocked with 0.5 mg/ml BSA and 0.2 mg/ml salmon sperm DNA. Levels of Mixl1 promoter precipitated DNA was analyzed by QPCR with the ABI Prism 7000 or 7900 sequence detection system (Applied Biosystems) using SYBR Green PCR Master Mix (Applied Biosystems) and primers indicated in Table 2.1. For DNA IP, Ct values for ChIP samples were normalized with corresponding DNA inputs Ct values. Net Ct values were plotted using Comparative Ct method (docs.appliedbiosystems.com/pebiodocs/ 04303859.pdf). For ChIPs, Net Ct values for the anti-Flag IP was corrected with Net Ct values of control IP and plotted using the Comparative Ct method.

2.2.5 ES cells and embryoid body differentiation

Flag-tagged Gsc cDNA was subcloned in the episomal expression vector pCAGIP, and this vector was linearized with PvuI and electroporated into the previously described Foxh1+/+(#9b) and Foxh1-/-(#3) ES cell lines [120]. Individual puromycin-resistant colonies were selected, and Gsc mRNA expression was confirmed by quantitative RT-PCR. For in vitro differentiation assays, ES cells were grown in ES medium [343] were passaged twice on 0.1%

gelatin-coated plates and grown to about 70% confluency. ES cells were trypsinized in 0.05% trypsin, harvested and resuspended in EB medium (ES medium minus LIF). Hanging drop cultures were established as previously described [344]. Embryoid bodies from two-day hanging drop cultures were transferred to 24-well ultra-low attachement plates at a density of 50 EBs/well. Three days later, EBs were transferred to gelatin-coated 6-wells dishes and maintained in EB media until harvested for RNA extraction.

2.2.6 Quantitave RT-PCR and in-situ hybridization analysis of Mixl1 in mouse embryos

Embryonic RNA was isolated using TRIzol reagent (Invitrogen) according to standard procedures from pools of E7.5 embryos genotyped for Foxh1 as described previously [120]. Gsc mutant embryos were genotyped by PCR as described previously [297] with DNA extracted from the extraembryonic region of E7.5 embryos. Embryonic RNA from wild-type and Gsc-null embryos were isolated using the Nanoprep kit (Stratagene) according to manufacturer’s instruction. ES cell and EB RNA was isolated using the RNeasy Mini Kit (Qiagen) according to manufacturer’s instructions. RNA samples were treated with DNaseI (Fermentas), primed with oligo p(dT)20 (ACGT Corp.) and reverse transcribed using RevertAid™ H Minus M-MuLV Reverse Transcriptase (Fermentas). Quantitave PCR on embryonic samples was performed using SYBR Green PCR Master Mix (Applied Biosystems) and Mixl1 and Hprt primers indicated in Supplementary Table 1. Quantitative PCR on ES and EB samples and wild-type and Gsc-null embryos was performed with TaqMan Universal PCR Master Mix (Applied Biosystems) and pre-developed Taqman Gene expression assays (Applied Biosystems) for mouse Mixl1, Gsc and GAPDH using the ABI Prism 7000 or 7900 sequence detection system (Applied Biosystems). Analysis was performed using the Comparative Ct method. Whole- mount in-situ hybridization analysis of Mixl1 in mouse embryos was performed as described

[105] using a Mixl1 probe [324].

2.3 Results

2.3.1 Foxh1 and Smads mediate TGFβ/Activin-dependent transcription of Mixl1

Activin-dependent induction of the Xenopus Mix2 gene, a Mix/Bix family member, occurs via Foxh1, a DNA-binding forkhead protein, in complex with activated Smads [111]. As expression of Foxh1 and Mixl1 appears to overlap during early mouse embryogenesis [115, 332], we sought to determine whether Foxh1 might mediate the TGFβ/Activin-dependent transcriptional regulation of the murine Mixl1 gene. Examination of human, mouse, rat and rhesus monkey Mixl1 promoters revealed the presence of a putative Foxh1 binding site (Fig. 2.1), suggesting a conserved role for Foxh1 in Mixl1 transcription. Thus, to examine whether Foxh1 regulates Mixl1 expression, a 248-bp fragment from the murine Mixl1 promoter, encompassing the putative Foxh1 site, was subcloned upstream of a luciferase reporter gene (Mixl1-luc) and the response to TGFβ was determined in the human hepatocarcinoma cell line, HepG2, which lacks Foxh1 activity [82]. Co-expression of Foxh1 with the Mixl1-luc reporter yielded strong TGFβ- dependent activation of the Mixl1-luc reporter (Fig. 2.2A). Direct binding of Foxh1 to the Mixl1 promoter fragment was confirmed by electrophoretic mobility shift assays (EMSA) using bacterially-expressed Foxh1 (Fig. 2.2B). Co-expression of Smad2 and Smad4 enhanced both the basal and TGFβ-induced Foxh1-dependent activation of the Mixl1-luc reporter (Fig. 2.2C). Although expression of Smad3 and Smad4 also increased the basal Foxh1-mediated activation of the Mixl1 promoter, TGFβ-dependent responsiveness was lost (Fig. 2.2C).

We next examined whether Foxh1 and Smads cooperate to form a higher-order DNA binding complex by EMSA. Comparison of DNA-binding complexes from mock transfected COS-1 cells versus myc-Foxh1 expressing cells revealed the appearance of a slower migrating band in both the presence and absence of the activated Activin type I receptor, ActRIB(TD) (Fig. 2.2D). Co-expression of Smad4 and either Smad2 or Smad3 with Foxh1 resulted in a further decrease in

Figure 2.1: Alignment of Mixl1 promoter sequences

ClustalW alignment the promoter sequences of the mouse (Mus musculus), rat (Rattus norvegicus), rhesus monkey (Macaca mulatta), and human (Homo sapiens) Mixl1 gene shows extensive conservation in the region encompassing the putative Foxh1 binding site. The Foxh1 binding site is indicated above the sequence alignment, conserved nucleotides are indicated with * and numbering relative to the start site are shown on the left.

Figure 2.2 Foxh1 and Smads bind the Mixl1 promoter and mediate the TGFβ-dependent induction of Mixl1.

(A and C) HepG2 cells were transiently-transfected with the Mixl1-luc reporter, Foxh1 and Smads (S2, S3, S4) as indicated. Relative luciferase (Luc) reporter activity is plotted and error bars represent standard deviation of the mean. (B and D) Electrophoretic Mobility Shift Assays. A 248 bp Mixl1 promoter fragment containing a wild-type Foxh1 binding site was incubated with bacterially-expressed proteins (B) or crude extracts from COS-1 transiently- transfected with the indicated DNA (D). Protein-DNA complexes were visualized by autoradiography. For supershift assays (D), anti-myc (M), anti-Flag (F), and anti-Smad4 (S4) antibodies were added to the reactions.

Figure 2.2

DNA migration, which was most evident in the presence of ActRIB(TD) (Fig. 2.2D). Incubation with antibodies resulted in supershift or loss of DNA-binding complexes, demonstrating the presence Foxh1 and Smads in the TGFβ Responsive Factor (TRF) (Fig. 2.2D). These observations indicate that Foxh1 can bind the Mixl1 promoter and, upon activation of the signaling pathway, forms a DNA-binding complex with Smad2 or Smad3 and Smad4.

To confirm a requirement for Foxh1 binding, two point mutations that prevent Foxh1 binding [82] were introduced in the putative Foxh1 site. These mutations abrogated Foxh1 binding to the Mixl1 promoter (Fig. 2.3A) and abolished TGFβ-dependent signaling (Fig. 2.3B). Smads bind to GC-rich sequences, thus we generated Mixl1 promoter constructs harboring either 8 GC to AT point mutations (SBEmut) or a complete deletion of a GC-rich region located downstream of the Foxh1 site (∆SBE) (Fig. 2.3C). The point mutations reduced, while complete deletion abolished, both TGFβ responsiveness and TRF formation on the promoter (Fig. 2.3C,D). Thus, our results, in agreement with previous studies [106], show that Foxh1 and Smad DNA binding is required for maximal TGFβ-dependent activation of Mixl1.

2.3.2 Nodal induces transcriptional activation of Mixl1

Foxh1 expression is confined to early embryogenesis and is thought to be downstream of the Nodal signaling pathway [23]. Nodal activates a TGFβ-like pathway by binding Activin receptors and the GPI-linked co-receptors, Cripto or Cryptic [23]. To investigate whether Nodal signaling could induce Foxh1-dependent activation of the Mixl1 promoter, HepG2 cells were transiently-transfected with the Mixl1-luc reporter and Foxh1 with various combinations of Activin type I and type II receptors and the co-receptor, Cripto. While co-transfection of Activin receptor type IB (ActRIB) alone or in the presence of Activin receptor type II (ActRII) mediated induction of the Mixl1 promoter upon Activin treatment,

Figure 2.3 Foxh1 and Smad binding requirements for the TGFβ-dependent induction of Mixl1.

Foxh1 and Smad binding requirements for the TGFβ-dependent induction of Mixl1. (A, D) Electrophoretic Mobility Shift Assays. Indicated Mixl1 promoter fragment was incubated with bacterially-expressed proteins (A) or crude extracts from COS-1 transiently-transfected with the indicated DNA (D). Protein-DNA complexes were visualized by autoradiography. (E) For supershift assays, anti-Smad4 (S4) antibodies were added to the reactions. (B and C) HepG2 cells were transiently-transfected with the indicated Mixl1-luc reporter (wt, mF, SBEmut or ∆SBE), Foxh1 and Smads (S2, S3, S4). Relative luciferase (Luc) reporter activity is plotted and error bars represent standard deviation of the mean.

Figure 2.3

Figure 2.4: Nodal signaling induces Mixl1 promoter activity

HepG2 cells were transiently-transfected with the wt Mixl1-luc reporter, Cripto, ActRIB and ActRII as indicated. Relative luciferase (Luc) reporter activity is plotted and error bars represent standard deviation of the mean.

Nodal-dependent activation of Mixl1 required co-expression of Cripto (Fig. 2.4) in agreement with a recent study [106]. Taking into account that Nodal plays an essential role in the induction of mesendodermal cell fates [23] and that Mixl1 is implicated in axial mesendoderm morphogenesis and patterning [332], Foxh1-dependent induction of Mixl1 is most likely driven by Nodal signaling during embryogenesis.

2.3.3 Foxh1 negatively regulates Mixl1 expression during early mouse embryogenesis

To evaluate the contribution of Foxh1 to the regulation of Mixl1 transcription in vivo, we examined Mixl1 expression levels in wild-type and Foxh1 mutant (Foxh1-/-) embryos. RNA was extracted from genotyped pools of wild-type and Foxh1-/- embryos at embryonic day 7.5 (E7.5), and transcript levels were determined by quantitative PCR (QPCR) using primers specific for Mixl1 transcripts. In RNA extracted from a pool of Foxh1-/- embryos, Mixl1 expression was increased by almost 2 fold over levels expressed in wild-type embryos (Fig. 2.5A). The level of Foxh1 expression and the absence of Foxh1 mRNA in our pool of wild-type and Foxh1-/- embryos, respectively, were confirmed by QPCR (Fig. 2.5A).

We next used whole-mount in-situ hybridization to examine Mixl1 expression patterns in Foxh1- /- mutant embryos. At E7.0, Mixl1 expression is confined to the nascent primitive streak and emerging mesodermal wings in both wild-type (Fig. 2.5B) [324, 327] and Foxh1-/-embryos (Fig. 2.5B). However, in all Foxh1-/- mutant embryos examined, we observed an increase in the intensity of the staining relative to the wild-type embryos (Fig. 2.5B). At E7.5, Mixl1 staining becomes limited to the primitive streak in wild-type embryos (Fig. 2.5B) [324, 327], while in Foxh1-/- mutant embryos, Mixl1 expression was expanded anteriorly in the embryonic mesoderm and proximally in the embryonic/extra-embryonic border (Fig. 2.5B). At E8, Mixl1, which becomes restricted to the posterior primitive streak and the base of the allantois in wild-type

Figure 2.5: Mixl1 expression is upregulated in gastrulating Foxh1-null embryos.

(A) RNA was extracted from E7.5 Foxh1+/+ and Foxh1-/- embryos and Mixl1 and Foxh1 levels (mean ± standard deviation for each gene) quantified by QPCR, normalized to Hprt expression. (B) At E7.0 and E7.5, Mixl1 expression is expanded and upregulated in Foxh1 mutant (-/-) embryos within the posterior primitive streak and the wings of the nascent mesoderm. At about E8.0, Mixl1 is ectopically expressed in the endoderm overlying the head process. Images taken from the left. Background staining (*) is noted.

embryos (Fig. 2.5B) [324, 327], exhibited an expanded expression domain in the caudal end of the primitive streak of Foxh1-/- embryos (Fig. 2.5B). Ectopic Mixl1 staining was also observed anteriorly in mutant embryos (Fig. 2.5B). Thus, increased Mixl1 transcript levels are due to both enhanced gene activity and spatial expansion of the Mixl1 expression domain. Although Foxh1 is best known as transcriptional activator downstream of the Activin/Nodal signaling pathway [121], our data suggests that Foxh1 can also act as a negative regulator of target gene expression.

2.3.4 Goosecoid negatively regulates Foxh1-dependent activation of Mixl1

The repressive activity of Foxh1 on endogenous Mixl1 expression observed in mouse embryos was not recapitulated in HepG2 cells, where Foxh1 promotes activation of the Mixl1-luc reporter (Fig. 2.2A). This suggested the possibility that a co-factor present in embryos, but absent in HepG2 cells, was required for the repressive effect. Forkhead and homeodomain-containing proteins have been shown to physically interact to negatively regulate target genes [345-347]. Since, the expression of Foxh1, Mixl1 and the paired-like homeodomain gene Goosecoid (Gsc) partially overlap during gastrulation [115, 282, 297, 324, 327] and Gsc expression is lost in Foxh1 mutant embryos [119, 120] we next determined whether Gsc might modulate Foxh1- dependent expression of Mixl1. For this, HepG2 cells were transfected with Foxh1 and

increasing amounts of Gsc and TGFβ-dependent activation of the Mixl1-luc reporter was examined. Ectopic expression of Gsc strongly repressed TGFβ-dependent induction of luciferase activity mediated by Foxh1 alone (Fig. 2.6A) or when Smad2 and Smad4 were co-expressed with Foxh1 (Fig. 2.6B). Thus, Gsc functionally interacts with Foxh1 to repress Mixl1 transcription. To determine whether Gsc modulates the expression of Mixl1 in vivo, we next examined the expression of Mixl1 in E7.5 wild-type and Gsc-null embryos by QPCR. In RNA extracted from Gsc-null embryos, average Mixl1 expression was increased by almost 2 fold over wild-type levels, (Fig. 2.6C). Interestingly, the range of Mixl1 mRNA expression in Gsc-null

Figure 2.6: Goosecoid (Gsc) represses Foxh1-mediated induction of the Mixl1 promoter.

(A, B and D) HepG2 cells were transiently-transfected with the Mixl1-luc reporter with combinations of Foxh1, increasing amounts (0.25 to 250 ng) of Gsc or Gsc N210G. Relative luciferase (Luc) reporter activity is plotted and error bars represent standard deviation of the mean. (C) Quantitative RT-PCR analysis of Mixl1 expression in wild-type and Gsc-null embryos. Expression was normalized to GAPDH. Black line represents average and grey box represents range of Mixl1 expression. Data represents the average of two QPCR experiments. Number of embryos: 5 wild-type, 3 Gsc-null.

embryos was wider as compared to wild-type embryos. As Gsc null embryos do not display overt defects in early embryogenesis [299, 300], we speculate that compensatory mechanisms overcome the increased expression of Mixl1 to allow gastrulation to proceed. As loss of Gsc enhances expression of Mixl1, these results suggest that Gsc can negatively regulate Mixl expression in vivo.

2.3.5 Physical interaction of Foxh1 with Goosecoid

Goosecoid can repress transcription of target genes such as Xbra and its own promoter by binding directly to DNA at paired homeodomain sites [305, 336]. The Mixl1 promoter does not contain consensus paired homeodomain sites, and the Gsc DNA-binding mutant (Gsc N210G) repressed Foxh1-mediated induction of the Mixl1 promoter (Fig. 2.6D), indicating that Mixl1 repression by Gsc occurs independent of its DNA binding activity.

We next examined whether Gsc might repress transcription by associating with Foxh1. Analysis of immunoprecipitates of cell lysates from COS-1 cells transfected with Flag-Foxh1 and T7-Gsc revealed an interaction between Foxh1 and Gsc (Fig. 2.7A). To determine the region in Foxh1 that mediates the binding to Gsc, we evaluated the ability of two deletion constructs of Flag-

Foxh1 to interact with Gsc expressed as a GST fusion. We observed that GST-Gsc bound full- length Flag-Foxh1 and a mutant (∆C) containing only the forkhead domain but not Foxh1 lacking the forkhead domain (∆N, Fig. 2.7B). A similar approach was used to identify the region of Gsc mediating the association with Foxh1. In this case, full-length GST-Gsc and GST-Gsc 1- 220 bound Flag-Foxh1, while larger deletions lacking the homeodomain did not (Fig. 2.7C). The Gsc homeodomain alone was not sufficient to interact with Flag-Foxh1 (Fig. 2.7C), suggesting that in the context of full length Gsc, other regions are also necessary for efficient binding to Foxh1. Gsc lacking the homeodomain (Gsc 1-143) did not block Foxh1-mediated induction of the Mixl1 promoter in luciferase reporter assays (Fig. 2.7D). Thus, our data shows that the

Figure 2.7:Gsc interacts with Foxh1 and represses Foxh1-mediated activation of the Mixl1 promoter.

(A) COS-1 cells were cotransfected with T7-Gsc and Flag-Foxh1 or vector control. Cell lysates were immunoprecipitated (IP) with anti-T7 antibody, and immunoblotted (IB) with anti-Flag antibody. (B and C) GST pulldown assay (GST PD) was performed by incubating Gsc or Gsc deletion constructs expressed as GST-fusion proteins with lysates from COS-1 cells expressing full-length or deletion mutants of Flag-Foxh1. Bound proteins were detected by immunoblotting with anti-Flag antibody. Equal protein expression was confirmed by immunoblotting or by Ponceau red or Coomassie blue staining as indicated. (D) HepG2 cells were transiently- transfected with the Mixl1-luc reporter, with Foxh1 and increasing amounts (0.25 to 250 ng) of Gsc or Gsc 1-143 (closed triangles). Relative luciferase (Luc) reporter activity is plotted and error bars represent standard deviation of the mean.

interaction between Gsc and Foxh1 is mediated by their conserved DNA binding regions, and that the Gsc homeodomain is required for the repressive activity of Gsc.

2.3.6 Gooseceoid is recruited to the Mixl1 promoter through Foxh1.

As Gsc can bind Foxh1, we next examined whether Gsc could be recruited to the Mixl1 promoter through its interaction with Foxh1 by performing sequential protein-DNA immunoprecipitations. COS-1 cells were transiently-transfected with the Mixl1-luc reporter together with cDNAs encoding Flag-Foxh1 and T7-Gsc. Protein-DNA complexes were cross-linked by formaldehyde treatment, and cell lysates were sequentially immunoprecipitated with anti-Flag and anti-T7 antibodies (Fig. 2.8A) and the amount of immunoprecipitated Mixl1-luc reporter was then assessed by QPCR. When Flag-Foxh1 and T7-Gsc were co-expressed, the Mixl1 promoter fragment was detected (Fig. 2.8A), whereas only basal levels were observed when either protein was expressed alone. In contrast, Gsc 1-143, which does not interact with Foxh1 nor blocks Foxh1-dependent induction of Mixl1, did not bind to the Mixl1 promoter (Fig. 2.8B). Thus, Gsc can be recruited to the Mixl1 promoter through its interaction with Foxh1.

We next evaluated whether Gsc could interact with Foxh1 within the context of a TRF complex by performing a three-step sequential protein-DNA immunoprecipitation. COS-1 cells were transiently-transfected with the Mixl1-luc reporter along with cDNAs encoding Flag-Foxh1, Smad4-HA, T7-Gsc and ActRIB(TD). Cell lysates were subjected to immunoprecipitation with anti-T7 antibody, followed by elution with 1% SDS, immunoprecipitation of the eluate with anti- Flag antibody, a second elution with 1% SDS and a final immunoprecipitation with anti-HA antibody (Fig. 2.8C). QPCR analysis demonstrated that T7-Gsc bound Flag-Foxh1 within a TRF complex (Fig. 2.8C, right panel).

Figure 2.8: Gsc interacts with Foxh1 and Smads on the Mixl1 promoter.

(A-C) COS-1 cells were transiently-transfected with the Mixl1-luc reporter and the indicated cDNAs. Protein-DNA complexes were collected by sequential immunoprecipitations, followed by elution in 1% SDS, using anti-Flag, anti-T7 and anti-HA antibodies as indicated (left panels). Recovered DNA was analyzed by QPCR, and corrected for DNA inputs (middle panels). Protein expression was confirmed by immunoblotting (left panels).

In contrast, negligible amounts of Mixl1 were immunoprecipitated when T7-Gsc was co- expressed with a Foxh1 DNA binding mutant (Foxh1 dm; R61H, K64N) or with a control plasmid (Fig. 2.8C). Our data show that Foxh1, Smads and Gsc can coexist on the Mixl1 promoter and suggest that the transcriptional repression mediated by Gsc is unlikely to occur through disruption of the TRF.

2.3.7 The repressive activity of Goosecoid is mediated by histone deacetylases

We next examined whether the conserved GEH repressor domain present at the N-terminus of Gsc is required for the repressive activity. However, Gsc ∆14, which lacks the GEH domain, was able to repress Mixl1-luc reporter activity as well as wild-type Gsc (Fig. 2.9A). To investigate whether Gsc might recruit transcriptional co-repressors such as HDACs to abrogate transcriptional activation of Mixl1, we examined the effect of Trichostatin A (TSA), an inhibitor of HDAC activity on Mixl1-luc reporter activity. HepG2 cells, transfected with the Mixl1-luc reporter, Foxh1, in the presence or absence of Gsc were incubated with TGFβ and TSA. Treatment with increasing doses of TSA resulted in the enhancement of Mixl1-luc activity by up to 3 fold (Fig. 2.9B), suggesting that recruitment of HDACs by Gsc contributes to the repression of Mixl1 promoter activity. The association of Gsc with HDACs was then examined using COS- 1 cells transfected with T7-Gsc and various Flag-HDACs. Gsc interacted with Class I HDACs, HDAC1 and HDAC3 and the Class IV HDAC, HDAC11 but not with the others (Fig. 2.10A). To determine the region in Gsc that mediates the interaction with HDACs, we tested the ability of Gsc N-terminal deletion mutants to interact with HDAC1 expressed as a GST-fusion protein (Fig. 2.10B). HDAC1 bound full-length Gsc as well as deletion mutant ∆81 but not deletion mutants ∆111 and ∆143 (Fig. 2.10B).

Figure 2.9: Gsc repressive activity is inhibited by Trichostatin A (TSA) treatment.

(A) HepG2 cells were transiently-transfected with the Mixl1-luc reporter, with Foxh1 and

increasing amounts (0.25 to 250 ng) of full-length or ∆14 Gsc (closed triangles). Relative luciferase activity is plotted and error bars represent standard deviation of the mean. (B) HepG2 cells were transiently-transfected with the Mixl1-luc reporter, Foxh1 and either a control or Gsc expression vector. Cells were incubated overnight with 100 pM TGFβ and the indicated amount of TSA. Luciferase activity from cell lysates was determined and presented as luciferase activity induced by TSA in cells transfected with Gsc relative to luciferase activity induced by TSA in cells transfected with the control plasmid. Error bars represent propagated error on the quotient calculated from standard errors of the mean of numerators and denominators.

Figure 2.10: Goosecoid recruits histone deacetylases (HDACs).

(A) Lysates of COS-1 cells co-transfected with T7-Gsc and the indicated Flag-HDAC constructs or vector control were immunoprecipitated with anti-T7 antibody and immunoblotted with anti- Flag antibody. Protein expression was verified by immunoblotting. (B) GST pulldown assay (GST PD) was performed by incubating GST-HDAC1 with lysates from COS-1 cells expressing full-length or deletion mutants of Flag-Gsc. Bound proteins were detected by immunoblotting with anti-Flag antibody. Equal protein expression was confirmed by immunoblotting. (C) HepG2 cells were transiently-transfected with the Mixl1-luc reporter, with Foxh1 and increasing amounts (0.25 to 250 ng) of full-length Gsc, or indicated deletion mutants (closed triangles). Relative luciferase (Luc) reporter activity is plotted and error bars represent standard deviation of the mean.

Figure 2.11: A model for Foxh1-dependent regulation of Mixl1.

Foxh1 constitutively binds the Mixl1 promoter and upon TGFβ signaling, Smads are recruited to Foxh1. Gsc binds Foxh1 and recruits HDACs to repress Mixl1 promoter activity.

The effect of these deletion mutants on Mixl1 promoter activation demonstrated that Gsc ∆81 repressed transcription, while the larger deletion mutants did not (Fig. 2.10C). Thus, our data shows that the interaction between Gsc and HDAC1 is mediated by a region located between residues 81 and 111 of Gsc, and that this region is required for Gsc-mediated repression of the Mixl1 promoter. Taken together, our biochemical analysis suggests a model (Fig. 2.11) in which Gsc associates with Foxh1 that is bound to the Mixl1 promoter and then functions to recruit histone deacetylases and thereby repress Mixl1 expression.

2.3.8 Goosecoid-mediated repression of endogenous Mixl1 requires Foxh1

Analyses of Foxh1-/- mutant embryos suggested that Foxh1 negatively regulates Mixl1 expression in vivo and our biochemical studies showed that Gsc associates with Foxh1 to mediate the transcriptional repression of Mixl1. Thus, we sought to determine whether increased expression of Gsc could repress endogenous Mixl1 expression and whether this repressive effect of Gsc requires Foxh1. Since Gsc is itself a Foxh1-regulated gene, the interpretation of in vivo data is particularly complex. Therefore, we utilized an in vitro assay in which mouse embryonic stem cells are induced to differentiate into embryoid bodies (EBs), a model system that recapitulates gastrulation events, including the generation of all three germ layers [348]. For

this, wild-type or Foxh1-/- mutant ES cell lines stably expressing murine Gsc or an empty vector control were generated. Individual lines were isolated and clones overexpressing Gsc were identified by QPCR using RNA extracted from ES cell colonies (Fig. 2.12A). Using the hanging drop method [344], ES cell lines were induced to differentiate by the removal of leukemia inhibiting factor (LIF) and embryoid bodies were collected for up to 7 days post-differentiation. Increased Gsc expression in the stable transfectants was maintained throughout the differentiation time course (Fig. 2.12A). Consistent with previous reports, Mixl1 was transiently induced [349-351], with maximal expression detected at day 5 in both wild-type (Fig. 2.12B, top

panel) and Foxh1-/- control lines (Fig. 2.12B, bottom panel). This concurs with our in vivo expression data, demonstrating that activation of Mixl1 expression occurs independent of Foxh1. However, in EBs derived from Foxh1 wild-type ES cells stably overexpressing Gsc, the enhanced expression of Mixl1 was abrogated (Fig. 2.12B, top panel). In marked contrast, overexpression of Gsc in all four Foxh1-/- EBs, resulted in enhanced Mixl1 expression at either day 5 (for clones 3, 6) or with a short delay on day 6 (for clones 9, 21) (Fig. 2.12B, bottom panel). Thus, in cells lacking Foxh1, Gsc is unable to repress Mixl1 expression.

To determine whether Gsc is recruited to the endogenous Mixl1 promoter, we next performed chromatin immunoprecipitation assays (ChIP) using cell lysates from differentiated embryoid bodies. Wild-type or Foxh1-/- ES cells lines overexpressing Flag-Gsc were induced to differentiate into EBs. After 5 days of differentiation, EBs were collected and protein-DNA complexes were fixed by formaldehyde treatment and immunoprecipitated with either control or anti-Flag antibody. Binding of Flag-Gsc to the region of the Mixl1 promoter encompassing the Foxh1 site was then assessed by QPCR. Efficient amplification of the target Mixl1 promoter fragment was detected in Foxh1 wild-type EBs as compared to EBs derived from Foxh1 null cells (Fig 2.13). No amplification of a control DNA fragment, located 2.5 Kb away from the Foxh1 binding site, was detected, though efficient amplification of the DNA fragment was observed in total cell lysates (Fig. 2.13 and data not shown). Thus, our results demonstrate that

Gsc is recruited to the endogenous Mixl1 promoter and that this is dependent on Foxh1. Taken together, our results indicate that Gsc binds to the Mixl promoter to downregulate the expression of Mixl1, and that this binding and repressive activity requires Foxh1.

Figure 2.12: Mixl1 expression is downregulated in wild-type but not in Foxh1-null embryoid bodies overexpressing Gsc.

Quantitative real-time RT-PCR analysis of Gsc (A) and Mixl1 (B) expression in control and Gsc- overexpressing undifferentiated ES cells (EB0) and embryoid bodies cultured for the indicated number of days. Expression levels were normalized to GAPDH. Gsc and Mixl1 levels are presented as mean expression ± standard deviation.

Figure 2.13: Recruitment of Gsc to the endogenous Mixl1 promoter is Foxh1-dependent.

EBs were differentiated for 5 days. Protein-DNA complexes were immunoprecipitated with control or anti-Flag antibodies. Recovered DNA was analyzed by QPCR with control or target primer pairs. Relative occupancy (mean ± standard deviation) for each set of primers represents levels of recovered DNA corrected by levels of DNA input. * No amplification detected.

2.4 Discussion

2.4.1 Foxh1 negatively regulates Mixl1 expression in mouse embryos

The mouse Mix-like gene, Mixl1, is essential for normal gastrulation and for proper development of the node, notochord, axial mesendoderm, heart and gut [332]. In Xenopus and zebrafish, the related Mix/Bix family members play pivotal roles in early development, functioning in Nodal-like signaling pathways to induce and specify mesoderm and endoderm [111, 322, 332]. While numerous studies have investigated the transcription factor requirements for induction of Mix/Bix genes in frogs and fish, little is known of the molecular mechanism controlling the expression of the mouse Mixl1 gene. Our phylogenetic analysis of the Mixl1 promoter revealed the presence of a conserved DNA binding site for the forkhead protein Foxh1, a positive transcriptional mediator of Nodal-like signals in early embryos [121]. However, in contrast to a role in promoting gene expression, our data demonstrates that Foxh1 functions as a negative regulator of Mixl1 transcription. Furthermore, we show that association of Foxh1 with the homeodomain protein Gsc is required for this repressive effect. Thus, we propose that positive and negative regulation of target gene transcription by Foxh1, a key mediator of Nodal signaling contributes to refining cell fate decisions in the primitive streak.

Our initial studies (this paper) and those of others [106] using mammalian tissue culture cells, have shown that Foxh1 binds to the Mixl1 promoter and can co-operate with Smads to mediate Nodal-dependent signaling. These results are consistent with previous reports defining a requirement for Nodal-like signaling for the induction of Mix/Bix family members such as Xenopus Mix2, and Mix3/Mixer [111, 315] as well as zebrafish ogx9/mezzo and bonnie and clyde (bon) [322] and more specifically via Foxh1 for Mix2 [111]. However, in contrast to these studies, our in vivo analysis of embryos lacking Foxh1 revealed prominent expression of Mixl1, clearly indicating that in mice Foxh1 is dispensable for induction of Mixl1 expression. Interestingly, use of morpholinos to knock-down maternally-expressed FAST1 (i.e. Foxh1) in

Xenopus oocytes prior to the onset of zygotic gene expression did not significantly affect Mix2 or Mix3/Mixer expression [104]. Furthermore, reduction of both FAST1 and the Xenopus-specific Foxh1-related gene, FAST3, in animal caps dissected from morpholino-injected early gastrulae, similarly revealed minimal effects on Mix1/Mix2 expression [114]. Thus, elimination of Foxh1 in either frogs or mice reveals that Foxh1 is not required for the initial induction of Mix-like gene expression. Instead, our data indicates that Foxh1 is a negative regulator of Mixl1 expression. Intriguingly, in the late Xenopus blastula, maternal Foxh1 is required to prevent ectopic expression of Xnr5 and Xnr6 but not Xnr3 in the ventral vegetal area [104]. The mechanism for the inhibitory Foxh1 activity was not determined, but the authors speculate that perhaps these promoters might harbor Foxh1 binding sites that function to negatively regulate their expression. Recently, a role for Foxh1 in negatively-regulating flk1 expression in zebrafish was reported [133]. This observations together with our data, suggest that Foxh1 may function as a negative regulator of transcription for diverse genes.

2.4.2 Gsc is recruited to Foxh1 and represses Mixl1 expression via histone deacetylases

Negative regulation of certain Foxh1 target genes can occur when the TRF is comprised of

Smad3/Smad4 heteromers rather than the positively-acting Smad2/Smad4 complex [82]. In addition, disruption of Foxh1 DNA binding via DRAP1, a co-repressor of basal RNA polymerase II transcription, has been reported to prevent Foxh1 target gene expression [137]. Gsc has been shown to negatively regulate gene expression by binding directly to DNA at specific sites [305, 336] and by recruiting the co-repressor protein Groucho via the N-terminal GEH domain [301]. In contrast, our analysis has revealed a distinct mechanism whereby, the homeobox transcription factor Gsc associates with DNA-bound Foxh1 and mediates the recruitment of HDACs to repress Mixl gene expression. Foxh1 can also recruit the homeodomain-containing protein Nkx2-5 [105] but, unlike Gsc, recruitment of Nkx2-5 mediates

heart-specific induction of Mef2c during cardiogenesis [105]. The ability of homeodomain proteins to act as either co-activators or co-repressors of forkhead transcription factors has been previously reported [346, 352] and may, thus, be a general property of this class of transcription factors.

2.4.3 Repression of Mixl1 expression by Goosecoid requires Foxh1 in embryoid bodies

Our analysis demonstrated that Mixl1 mRNA levels are enhanced in the nascent primitive streak and emerging mesoderm of Foxh1-/- mutant embryos. Gsc expression also marks the newly formed primitive streak at the onset of gastrulation [282], and this expression is lost in Foxh1-/- mutant embryos [119, 120]. Although we show that loss of Foxh1 relieves inhibition of Mixl1 mRNA expression in vivo, since Gsc is itself under the control of Foxh1, we cannot conclude from these data that loss of Gsc enhances this effect. To circumvent this difficulty, we examined Mixl1 expression in EBs ectopically expressing Gsc driven by a heterologous promoter. We observed that Mixl1 expression was downregulated by Gsc in wild-type but not in Foxh1-null EBs. Furthermore, downregulation was most pronounced on day 5 of EB differentiation, a time point which approximately corresponds to primitive streak development in gastrulating mouse embryos [348]. This demonstrates that Foxh1 is required for the repressive effect of Gsc on Mixl1 expression.

During early mouse development, Nodal expression is highly dynamic and its activity is tightly regulated. It is first expressed throughout the epiblast and in the visceral endoderm and later is confined to the posterior proximal epiblast where it marks the site of primitive streak induction [23]. Our data suggest that at the onset of gastrulation Nodal induces expression of Gsc in the primitive streak via Foxh1 [82, 119, 120], while Mixl1 expression is induced via a Foxh1- independent mechanism (Fig. 2.14). As gastrulation proceeds, Gsc accumulates in cells fated for

100

the anterior region of the streak, where it is recruited through Foxh1 to the Mixl1 promoter to repress transcription, thus restricting Mixl1 expression to the posterior region. Accordingly, in Foxh1 mutant embryos, we observed anterior expansion of Mixl1 expression (Fig. 2.5B). Thus, Foxh1 functions in a transcriptional regulatory loop that acts through Gsc to negatively regulate Mixl1. This mechanism likely reflects part of a much larger transcriptional network that acts to refine gene expression patterning and cell fate decisions in the primitive streak. For instance, the expression of Gsc itself requires Foxh1 [82], and thus downregulation of Gsc expression could occur via an autoregulatory loop either by direct interaction with Foxh1 or through direct DNA binding of Gsc to its own promoter as has been previously reported [305]. Furthermore, Foxh1- dependent enhancer elements present in the first intron of the nodal gene are required for the maintenance and amplification of nodal expression [118]. Thus, it will be interesting to investigate whether Foxh1 can also function to negatively regulate nodal expression through the recruitment of Gsc. In Xenopus, Nodal-dependent induction has also been reported to occur either through Mixer [112], Xenopus homolog of the human gene Williams-Beuren syndrome critical region 11 (XWBSCR11) [134] or p53 [139], though the role of these modulators has not been examined during early mouse development. The exact mechanisms controlling transcriptional events occurring during mouse gastrulation are likely to have specific spatial, temporal and promoter context requirements and will require further investigation. However, our results indicate that Foxh1, functioning within positive and negative feedback loops may serve to spatially and temporally refine Gsc and Mixl1 activities to ensure appropriate cell fate choices during gastrulation.

101

Figure 2.14: A model representing positive and negative feedback loops regulating Mixl1 and Gsc expression through Foxh1.

102

CHAPTER 3

3 Sumoylation differentially regulates Goosecoid- mediated transcriptional repression

Note: The work included in this chapter was published in the following manuscript:

Izzi L., Narimatsu M. and Attisano L. (2008) Sumoylation differentially regulates Goosecoid- mediated transcriptional repression. Exp Cell Res. 314(7):1585-94

L. Izzi generated all figures.

M. Narimatsu performed confocal microscopy on immunofluorescence stainings presented as data not shown.

103

3.1 Abstract

Goosecoid (Gsc), a paired-like homeobox gene expressed in the vertebrate organizer, functions as a transcriptional repressor either by direct DNA binding to paired TAAT homeodomain sites or through recruitment by the forkhead/winged-helix transcription factor Foxh1. Here, we report that Gsc is post-translationally modified by small ubiquitin-like modifier proteins (SUMO). Members of the PIAS family of proteins enhance Gsc sumoylation and this modification occurs on at least six lysine residues. Stable expression of a SUMO-defective Gsc mutant (Gsc 6Km) in MDA-MB-231 breast cancer cells results in morphological changes giving rise to cells with increased cell area. We demonstrate that Gsc 6Km can effectively repress Foxh1-mediated induction of the Mixl1 promoter, indicating that sumoylation is not required for Gsc-mediated repression of promoters where recruitment occurs through Foxh1. In contrast, Gsc 6Km exhibits a decreased ability to repress the induction of promoters to which it is directly recruited through paired homeodomain binding sites, including its own promoter and that of the Xenopus Brachyury gene. Taken together, our data suggests that regulation of Gsc repressive activity by SUMO modification is promoter specific and may serve to differentially regulate genes that function to control cell morphology during early development and cancer.

104

3.2 Introduction

Prior to gastrulation, the vertebrate embryo is a simple hollow ball of cells. As gastrulation proceeds, coordinated cell movements results in a massive reorganization of the embryo, transforming it from a simple epithelial structure to a complex multilayered organism, comprising ectoderm, mesoderm and endoderm. In addition to creating the three germ layers, gastrulation movements also lead to the establishment of the vertebrate body plan [342]. In amphibians, the dorsal lip of the blastopore, also known as the Spemann organizer, regulates the fate of neighboring cells and thus early embryonic patterning [5]. Specialized sub-populations of cells with organizer activity have also been identified in other species including, fish, chicken and mouse (reviewed in [6, 7, 353]).

Although gastrulation and early embryonic patterning differs widely between species, a number of conserved genes with organizer activity have been described so far [6]. The paired-like homeobox gene goosecoid (gsc) was the first gene with organizer properties to be identified from a Xenopus dorsal blastopore lip cDNA library [289]. In addition, numerous overexpression and loss-of-function studies in Xenopus have firmly established the requirement for gsc in dorso- ventral patterning of mesodermal tissues [289, 291-293, 295]. In mouse, Gsc is first expressed at the onset of gastrulation at the site of primitive streak formation, and although expression is maintained throughout gastrulation in organizer cells [282], knockout studies have revealed that loss of Gsc activity does not significantly affect gastrulation [299, 300]. This has led to the suggestion that complementary mechanisms involving other organizer specific genes may compensate for the lack of Gsc expression and allow gastrulation to proceed [279]. Interestingly, nodes from Gsc-/- mice grafted into primitive streak stage chick embryos have decreased neural inducing activities [354]. Gsc-/- embryos also show increased expression of the primitive streak marker Mixl1 when compared to wild-type embryos [355]. In addition to its role during gastrulation, Gsc has recently been implicated in cancer. Gsc mRNA expression is found to be elevated in human breast tumors when compared to normal tissues [312]. Moreover, stable

105

expression of Gsc in immortalized mammary cells promotes epithelial-to-mesenchymal transition (EMT) and cell motility in vitro and enhances the metastatic ability of MDA-MB-231 in mice [312]. This newly identified role for Gsc in enhancing metastasis is consistent with Gsc activities during gastrulation, which include promoting the dorso-anterior migration of blastomeres in Xenopus embryos [295].

Molecularly, Gsc functions as a transcriptional repressor. Gsc can repress its own promoter and that of the Xbrachyury (Xbra) and Xwnt-8 genes through direct DNA binding to paired homeodomain binding sites [290, 305, 306]. The N-terminus of Gsc contains a short conserved sequence known as the Goosecoid Engrailed Homology (GEH) motif, which has been shown, in Drosophila, to interact with the co-repressor Groucho and mediate the repressive activity of Gsc [301]. In addition, Gsc can be recruited to the Mixl1 promoter through the forkhead/winged helix transcription factor Foxh1, and through the recruitment of histone deacetylases represses the Foxh1/Smad2/4-dependent induction of the Mixl1 gene [355].

Post-translational modification with the Small Ubiquitin-related Modifier (SUMO) family of proteins has been shown to play a crucial role in the regulation of the activity of numerous transcription factors (reviewed in [257]). SUMO proteins (SUMO1, SUMO2 and SUMO3) are small polypeptides distantly related to ubiquitin that are covalently attached by an isopeptide bond to lysine residues of substrate proteins. SUMO conjugation, which often requires a consensus sequence ψKxE/D (where ψ is a large hydrophobic residue and x is any amino acid), occurs in a stepwise enzymatic cascade involving a SUMO-activating E1 enzyme (Aos1/Uba2), a SUMO-conjugating E2 enzyme (Ubc9), and a SUMO E3 ligase [257]. Although the biological significance of sumoylation remains largely unknown, a recent study demonstrated that XSUMO- 1 is required for mesoderm induction and axis elongation in Xenopus [356]. At the molecular level, sumoylation regulates the functional properties of target proteins. For example, sumoylation of transcription factors has been shown to mediate their ability to directly interact with DNA and recruit co-repressors such as histone deacetylases [257].

106

Here, we demonstrate that Gsc is covalently modified by SUMO proteins and identify six lysines, five of which are present within the homeodomain, that are targeted for modification. Sumoylation of Gsc is enhanced by members of the protein inhibitor of activated STATs (PIAS) family of proteins and expression of a version of Gsc that cannot be sumoylated leads to increased cell spreading in breast cancer cells. We also demonstrate that the ability of a SUMO mutant of Gsc to repress transcription is promoter specific. Therefore, we speculate that sumoylation may serve to differentially regulate Gsc-dependent repression of target genes and thereby affects processes occurring during normal gastrulation and cancer development.

107

3.3 Material and Methods

3.3.1 Reporter constructs and transcriptional reporter assays

The mouse Gsc promoter (position -336 to -456 from the transcription initiation start site) and Xenopus brachyury (Xbra) promoter (position -150 to -1000 from the transcription initiation start site) were amplified by PCR from a mouse Gsc genomic DNA clone and Xenopus genomic DNA, respectively and subcloned into the SacI/BglII sites of the pGL2-promoter vector (Promega). The Mixl1-luciferase reporter has been described previously [355]. For luciferase assays, HepG2 cells were transiently-transfected using the calcium phosphate DNA precipitation method, as described previously [82]. Transfections contained 0.083 µg of Mixl1-luc reporter, 0.2 µg of Gsc-luc reporter, 0.2 µg of Xbra-luc reporter, 0.0035 µg of Flag-Foxh1, 0.25 to 250 ng of T7-Gsc or its derived mutants, 0.1 µg of pCMV-βgal and pCMV5 empty vector to a total of 1 µg per well in a 24-well dish.

3.3.2 Construction of mammalian expression vectors and generation of stable cell lines

Epitope-tagged Gsc full length, deletion and point mutants were generated by PCR using a mouse Gsc cDNA and subcloned into SalI/BamHI sites in pCMV5B. HA-SUMO-1, HA- SUMO-2 and HA-SUMO-3 (obtained from Ronald T. Hay) were subcloned from a pCDNA3 vector into pCMV5B in KpnI/XbaI sites. T7-PIASγ was provided by Rudolf Grosschedl and Flag-TLE1 by Stefano Stifani. Flag-PIASγ, Flag-PIAS1 and Flag-PIAS Xα were subcloned in pCMV5C as described in [357]. Flag-Foxh1 and Flag-HDAC1 have been described previously [82, 355].

For stables, Flag-Gsc WT and 6Km cDNAs were cloned in a pCAGIP vector as described in [355]. MDA-MB-231 cells, maintained in 5% FBS RPMI, were transfected with

108

LipofectAMINE 2000 (Invitrogen) according to the manufacturer’s protocol. Puromycin selection (1.5 µg/mL) was applied to cell cultures 48 hrs post-transfection. Following selection, pooled clones were maintained in 1 µg/mL puromycin.

3.3.3 Immunoprecipitation and immunoblotting

For immunoprecipitations, COS-1 cells were transfected using the polyethylenimine (PEI; Sigma-Aldrich catalog # 408727) method as described previously [355]. For immunoprecipitations performed to detect Gsc sumoylation, COS-1 cell lysates were prepared in LSLD buffer (50 mM Hepes pH 7.4, 50 mM NaCl, 0.1% Tween-20, 10% glycerol) and immunoprecipitated with anti-T7 antibodies (Novagen) for 1 hr at 4oC. For sequential immunoprecipitations, COS-1 cell lysates were prepared in LSLD buffer as described above and subjected to immunoprecipitation with anti-Flag antibodies, followed by incubation with protein G Sepharose for 1 hr. Protein complexes were eluted by boiling in 1% SDS, rediluted 10 times in TNTE lysis buffer (0.5% Triton X-100, 2 mM EDTA, 150 mM NaCl, 20 mM Tris-HCl pH 8.1), and subjected to a second immunoprecipitation with anti-Flag antibody. Immunoblotting was carried out as reported [82] using polyclonal rabbit anti-HA (Y11) antibodies (Santa Cruz), M2 anti-Flag (Sigma), anti-T7 (Novagen) monoclonal antibodies. Quantitation of the levels of

sumoylated Gsc normalized to total Gsc in immunoblots was determined using QuantityOne software (BioRad).

For DNA immunoprecipitations, COS-1 cells were transfected with the appropriate cDNA and pGL2p-Gsc-luc reporter using LipofectAMINE (Invitrogen), and immunoprecipitations were performed as described [355]. Levels of Gsc promoter precipitated DNA was analyzed by QPCR with the ABI Prism 7900 Sequence Detection System (Applied Biosystems) using SYBR Green PCR Master Mix (Applied Biosystems) and the following primers, forward 5’-AGA TTA ACC TGG GCA ATT AGG CCG-3’ and reverse 5’-GGC CTC GGC CTC TGC ATA AAT AAA G -

109

3’. Ct values for immunoprecipitated samples were normalized to corresponding DNA input Ct values. Net Ct values were plotted using the Comparative Ct method (docs.appliedbiosystems.com/pebiodocs/ 04303859.pdf).

3.3.4 Immunofluorescence staining

For immunofluorescence staining, MDA-MB-231 cell lines were plated onto fibronectin in 2.5% FBS RPMI, 36 hrs after plating, cells were washed once with PBS (prewarmed at 37oC), fixed with 4% paraformaldehyde (prewarmed at 37oC) for 10 minutes, permeabilized with 0.5% Triton X-100/PBS for 10 minutes, and blocked with 10% heat-inactivated goat serum for 30 minutes. Cells were sequentially incubated with M2 anti-Flag antibody (Sigma), and Alexa 488- conjugated goat anti-mouse antibody (Molecular Probes) for 1 hr at room temperature. Cells were stained for F-actin using rhodamine-conjugated phalloidin (Molecular Probes) for 20 min at room temperature. All antibodies and fluorescent stains were prepared in 10% heat-inactivated goat serum/PBS. Cells were washed extensively with 0.1% Triton X-100/PBS in between each antibody or fluorescent stain incubation. Bright field and epifluorescence images were captured with a Zeiss Axiovert 200 inverted microscope and analyzed with Volocity software (Improvision). Quantifications were done using the ImageJ software (http://rsb.info.nih.gov/ij/).

110

3.4 Results

3.4.1 Goosecoid is modified by Sumoylation

Goosecoid is a paired-like homeodomain transcription factor that functions to negatively regulate the expression of target genes [290, 305, 336, 355]. In general, proteins are regulated by various post-translational modifications such as phosphorylation, ubiquitination, sumoylation and acetylation. Because sumoylation is closely associated with the regulation of transcription factor activity, we examined whether Gsc was modified by SUMO in mammalian cells. For this, COS- 1 cells transfected with T7-tagged Gsc in presence or absence of HA-tagged SUMO isoforms were immunoprecipitated with anti-T7 antibodies and immunoblotted with anti-HA antibodies. We observed that co-expression of SUMO1, SUMO2 or SUMO3 with Gsc resulted in the appearance of four major higher molecular weight protein species which were absent when either Gsc or SUMO isoforms were expressed alone (Fig. 3.1A, compare lanes 6-8 with lanes 2-4). Since the SUMO-conjugated species have apparent molecular masses that exceed unmodified Gsc by approximately 20 kDa or greater, this analysis suggested that the slower mobility bands represented sumoylated forms of Gsc.

To confirm that the SUMO-conjugated species were indeed modified forms of Gsc rather than sumoylated proteins associating with Gsc, lysates of COS-1 cell transiently-transfected with Flag-Gsc and either HA-SUMO1 or HA-SUMO3 were subjected to two sequential anti-Flag immunoprecipitations under native and denaturing conditions, respectively. Immunoblotting with anti-HA antibodies revealed the appearance of either SUMO1- or SUMO3-conjugated species in Flag-Gsc immunoprecipitates isolated under denaturing conditions (Fig. 3.1B, lanes 4 and 9), indicating that the SUMO-conjugated proteins detected in our immunoblots were covalently modified forms of Gsc and not sumoylated binding partners.

We next examined whether Gsc could also be modified by endogenous SUMO1. Lysates of COS-1 cells transiently-transfected with Flag-Gsc were immunoprecipitated with anti-Flag

111

antibodies and immunoblotted with anti-SUMO1 antibodies. We observed the presence of SUMO-conjugated species in Flag-Gsc immunoprecipitates (Fig. 3.1C). These SUMO- immunoreactive bands migrated at the same level as bands detected in samples which expressed HA-SUMO1 (Fig. 3.1C, compare lanes 3 and 4). Taken together, our data indicate that Gsc is covalently modified by SUMO isoforms and this modification can occur at physiological levels of SUMO1.

3.4.2 PIAS proteins regulate Gsc Sumoylation

SUMO moieties are attached to target proteins in a stepwise process involving three enzymes [358, 359]. A SUMO-activating enzyme (E1) first activates SUMO isoforms in an ATP- dependent manner. Activated SUMO isoforms are then passed from the active-site cysteine of the E1 to the active-site cysteine of a SUMO-conjugating enzyme (E2). Finally, a SUMO E3 ligase catalyzes the transfer of SUMO isoforms from the E2 enzyme to the ε-amino group of lysines on target proteins [358, 359].

Members of the PIAS family of proteins, including PIAS1, PIAS3, PIASXα, PIASXβ and PIASγ, function as SUMO E3 ligases and have been described to promote the sumoylation of numerous transcription factors [257, 360]. Thus, we examined whether PIAS proteins regulate

the sumoylation of Gsc in mammalian cells. For this, COS-1 cells transiently-transfected with T7-Gsc, Flag-PIAS proteins and HA-SUMO3 were subjected to immunoprecipitation with anti- T7 antibodies and immunoblotted with anti-HA antibodies. We observed that all three PIAS proteins tested, that is, PIAS1, PIASγ and PIASXα promoted the sumoylation of Gsc (Fig. 3.2A, compare lanes 4-6 with lane 3). Co-immunoprecipitation studies also showed that Gsc can associate with PIASγ in COS-1 cells (Fig. 3.2B). Taken together, our data suggests that PIAS proteins function as SUMO E3 ligases that catalyze the attachment of SUMO onto Gsc.

112

Figure 3.1:Goosecoid is post-translationally modified by SUMO proteins

(A) COS-1 cells were transfected with T7-Gsc and either HA-SUMO1, HA-SUMO2, HA- SUMO3 or vector control. Cell lysates were immunoprecipitated (IP) with anti-T7 antibody, followed by immunoblotting (IB) with anti-HA antibody. (B) COS-1 lysates from cells expressing Flag-Gsc and either HA-SUMO1 or HA-SUMO3 were subjected to a double immunoprecipitation with anti-Flag antibodies under native (1X) and denaturing (2X) conditions to precipitate native and covalently modified forms of Gsc, respectively. Sumoylated Gsc was detected by immunoblotting with anti-HA antibodies. (C) COS-1 lysate from cells expressing Flag-Gsc, HA-SUMO1 or vector control, were immunoprecipitated with anti-Flag antibodies and immunoblotted with anti-SUMO1 antibodies to detect ectopic (black arrowheads) and endogenous (*) SUMO1 modification. (A, B, C). Protein expression was confirmed by immunoblotting of total cell lysates with the indicated antibodies (bottom panels). Sumoylated Gsc (black arrowheads), antibody heavy chains (arrows) and non-specific bands (open arrowheads) are indicated.

113

114

Figure 3.2: PIAS proteins regulate Goosecoid sumoylation

(A) Lysates from COS-1 cells transfected with T7-Gsc, HA-SUMO3 and Flag-PIASγ, Flag-

PIAS1 or Flag-PIASXα were immunoprecipitated with anti-T7 antibody and immunoblotted with anti-HA antibodies. (B) COS-1 lysates expressing Flag-Gsc and T7-PIASγ or vector control were immunoprecipitated with anti-Flag antibodies and immunoblotted with anti-T7 antibodies. (A and B) Protein expression was confirmed by immunoblotting of total cell lysates with the indicated antibodies (bottom panels). Sumoylated Gsc (black arrowheads), non-specific bands (open arrowhead) and antibody heavy chains (arrows) are indicated.

115

3.4.3 Gsc comprises multiple sumoylation sites

We next focused on identifying the lysine residues on Gsc that were targeted by this post- translational modification. Previous studies have shown that SUMO moieties are typically added to lysine residues present within a ψKx(E/D) consensus motif, where ψ is a large hydrophobic residue and x is any amino acid, though exceptions have been reported [361-363]. Because none of the thirteen lysine residues in Gsc are located within a ψKx(E/D), we used a candidate approach to identify the lysine residues in Gsc targeted for SUMO-modification. Our initial mutation analysis revealed that none of the single or double lysine mutations as well as deletion of the C-terminal lysines (Gsc 1-219) had a significant effect on the SUMO-modification of Gsc (Fig. 3.3A and data not shown), thus, we next examined whether point mutation of multiple lysine residues affected the overall sumoylation status of Gsc. We generated two multiple lysine point mutants of Gsc (Gsc 4Km and 6Km) where four (K183R, K203A, K209R, K214R) or six lysines were mutated (K183R, K196R, K203A, K209R, K214R, K238R). These mutants as well as wild-type Gsc were transiently-transfected in COS-1 in the presence or absence of HA- SUMO3. Lysates were subjected to immunoprecipitation with anti-T7 antibodies and immunoblotting with anti-HA antibodies. We found that Gsc 4Km was less efficiently sumoylated than wild-type Gsc (Fig. 3.3B, compare lanes 5 and 6). Moreover, sumoylation was further decreased in Gsc 6Km, which consists of Gsc 4Km with two additional mutations (K196R and K238R) (Fig. 3.3B, compare lanes 5 ,6 and 7). These observations were confirmed by quantitative analysis of three independent experiments (Fig 3C). Taken together, these data suggest that multiple lysine residues are targets for sumoylation on Gsc.

116

Figure 3.3: Mutation of six lysine residues decreases Goosecoid sumoylation levels.

(A) Single or double lysine mutations do not affect sumoylation of Goosecoid. A schematic of Gsc with the Gsc Engrailed Homology (GEH) and the homeodomain (HD) indicated. The amino acid sequence of Goosecoid downstream of residue 159 is shown. Homeodomain (HD) residues are underlined (dotted line) and lysine residues mutated in Gsc 6Km are indicated by asterisks (*). Single and multiple point mutants of Gsc are listed and their sumoylation status is noted. Degree of sumoylation is compared (+). (B) Lysates of COS-1 cells expressing HA-SUMO3, T7-tagged WT, 4Km and 6Km Gsc were immunoprecipitated (IP) with anti-T7 antibody and immunoblotted (IB) with anti-HA antibody. Protein expression was confirmed by immunoblotting of total cell lysates with the indicated antibodies (bottom panels). (C) Quantitation of immunoprecipitated Gsc-SUMO species normalized to total Gsc is shown. The average of three experiments and the standard error of the mean are plotted. p values obtained from one-tailed Student’s T-Test comparing Gsc 4Km or Gsc 6Km sumoylation levels to wild- type Gsc sumoylation levels are indicated. Sumoylated Gsc (black arrowheads), antibody heavy chains (arrows) and non-specific bands (open arrowhead) are indicated.

117

118

3.4.4 Overexpression of Gsc 6Km leads to a change in MDA-MB-231 cell size

In addition to the well-studied role for Gsc during early gastrulation, a recent study reported enhanced expression of GSC in human breast tumors and when overexpressed in MDA-MB-231 cells can promote metastasis [312]. Thus, to determine whether the SUMO mutant of Gsc might differentially affect breast cancer cell motility or morphology we generated pools of MDA-MB- 231 cells lines stably expressing either wild-type or the SUMO-defective mutant of Flag-Gsc (Flag-Gsc 6Km). Ectopic expression of Gsc was verified by immunoblotting (Fig. 3.4A) and immunofluorescence microscopy (Fig. 3.4B). Analysis of cell motility by scratch assays combined with live cell time-lapse phase contrast microscopy revealed no change in the rate of cell motility between controls, wild-type or 6Km Gsc-expressing cells (data not shown). However, a marked difference in cell morphology was observed in cells expressing Gsc 6Km. To further investigate this, we fixed our stable cell lines and examined cell morphology by immunofluorescence microscopy. Quantitation of individual cell area showed that control cells displayed an average area of 27.1±9.5 x103 µm2 and cells overexpressing wild-type Flag-Gsc exhibited a modest decrease averaging 18.8±6.5 x103 µm2 (p<0.0001) (Fig. 3.4B and C). In contrast, cells overexpressing 6Km Gsc exhibited a cell area that averaged 57.4 ±21.5 x103 µm2 (p<0.0001), corresponding approximately to a two-fold increase in cell area when compared to the control cell line (Fig. 3.4B and C). Similar changes in cell area were observed by immunofluoresence confocal microscopy using the membrane-specific dye, DiI (data not shown). Taken together, our data suggests stable expression of the sumoylation mutant of Gsc promotes a cell phenotype that results in an increased cellular area, indicating that sumoylation of Gsc may be involved in modulating the expression of genes regulating cell morphology.

119

Figure 3.4: Stable expression of Gsc 6Km alters cell size in MDA-MB-231 cells

(A) Immunoblotting of MDA-MB-231 cells stably expressing vector control, WT or 6Km Flag- tagged Gsc. (B) Immunofluorescence staining of MDA-MB-231 cells stably expressing vector control, WT or 6Km Flag-Gsc. Anti-Flag antibody staining (green), phalloidin staining (red) and DAPI staining (blue) is shown. Cell outlines are illustrated (far-right). (C) The cell area of MDA-MB-231 cells stably expressing vector control, WT or 6Km Flag-Gsc was quantitated in three independent experiments and results from a representative experiment are shown. Error bars represent standard deviation. p values were obtained from a two-tailed Mann-Whitney U test.

120

121

3.4.5 Role of Sumoylation on Gsc transcriptional activity

We next conducted a molecular analysis to determine the basis of the differential effects of the Gsc SUMO mutant (Gsc 6Km) on cell morphology. We previously demonstrated that Gsc associates with Foxh1 and mediates the recruitment of histone deacetylases (HDACs) to repress Mixl1 promoter activity [355]. Because sumoylation regulates the functional properties of transcription factors [257], we first examined whether sumoylation modulated the repression of promoters to which Gsc was recruited by associating with Foxh1. For this, HepG2 cells where transiently-transfected with the Mixl1-luc reporter, Flag-Foxh1 and increasing amounts of either wild-type or 6Km T7-Gsc. Gsc 6Km repressed the TGFβ-dependent induction of the Mixl1 promoter activity as well as wild-type Gsc (Fig. 3.5A). Consistent with this data, Gsc 6Km retained the ability to interact with Foxh1 (Fig. 3.5B) and HDAC1 (Fig. 3.5C) when immunoprecipitated from transiently-transfected COS-1 cell lysates.

Although Gsc can repress the transcription of Foxh1 target genes by binding to Foxh1 and recruiting HDACs [355], Gsc can also repress transcription by direct binding to paired- homeodomain sites such as those found in its own promoter as well as in the promoters of the Xbra and Wnt8 genes [290, 305, 306]. Sumoylation has been reported to regulate protein-DNA interactions [265, 268, 364], thus, we next investigated whether mutation of the sumoylation sites affected Gsc DNA binding activity. We performed protein-DNA immunoprecipitations lysates of COS-1 cells transiently-transfected with the Gsc-luc reporter together with wild-type or 6Km T7-Gsc. Protein-DNA complexes were cross-linked by formaldehyde treatment, cell lysates were immunoprecipitated with anti-T7 antibodies, and the amount of immunoprecipitated Gsc-luc reporter was assessed by quantitative PCR (QPCR). We found that Gsc 6Km bound the Gsc promoter as well as the wild-type protein (Fig. 3.6A), indicating that mutation of the sumoylation sites does not significantly affect Gsc DNA binding.

In Drosophila, the GEH motif present within the N-terminal region of Gsc has been shown interact with the Groucho co-repressor and mediate Gsc-dependent repression [301]. Thus, we

122 next tested whether mutation of Gsc sumoylation sites interfered with the ability of Gsc to recruit mammalian Groucho family member, transducin-like Enhancer of split 1 (TLE1). For this, lysates of COS-1 cells transiently-transfected with Flag-TLE1 and either wild-type or 6Km T7- Gsc where immunoprecipitated with anti-T7 antibodies and immunoblotted with anti-Flag antibodies. We found that both wild-type and Gsc 6Km interacted with Flag-TLE1 (Fig. 3.6B), indicating that recruitment of TLE1 is not significantly affected by the sumoylation status of Gsc. We next investigated whether mutation of sumoylation sites altered Gsc-mediated repression of its own promoter or that of the Xbra gene. For this, HepG2 cells were transiently- transfected with Gsc-luc or Xbra-luc reporters in the presence of increasing concentrations of either wild-type or 6Km Gsc. We found that Gsc 6Km has a decreased ability to repress transcription when compared to wild-type Gsc on both the Gsc and Xbra promoters that was statistically significant (Fig. 3.6C). Taken together, our data suggests a model where sumoylation does not appear to alter the repressive activity of Gsc in cases where recruitment to the promoter occurs through the association with the transcription factor Foxh1, present within a complex also containing Smad2 and Smad4 (Fig.3.7A), but is important when Gsc represses transcription through direct DNA binding to paired TAAT homeodomain consensus sequences (Fig. 3.7B).

123

Figure 3.5: Sumoylation does not appear to affect the repressive activity of Gsc on the Mixl1 promoter

(A) HepG2 cells were transiently-transfected with the Mixl1-luc reporter, Foxh1 and increasing amounts (0.25 to 250 ng) of WT or 6Km Gsc and incubated with or without 100 pM TGFβ. Luciferase activity from cells lysates is plotted as mean ± standard deviation. p values obtained from a two-tailed Student’s T-Test comparing the repressive activity of wild-type and Gsc 6Km at increasing doses are p=0.4, p=0.07, p=0.34, p=0.13, indicating no statistical difference in their activity. (B and C) Lysates from COS-1 cells transfected with the indicated constructs were immunoprecipitated with either anti-Flag (B) or anti-T7 (C) antibodies. Interactions were detected by immunoblotting with either anti-T7 (B) or anti-Flag (C) antibodies. (A-C) Protein expression was confirmed by immunoblotting of total cell lysates with the indicated antibodies (bottom panels)

124

125

Figure 3.6: Sumoylation potentiates Gsc repressive activity on direct targets.

(A) COS-1 cells were transiently-transfected with the Gsc-luc reporter and the indicated cDNAs. Protein-DNA complexes were immunoprecipitated with anti-T7 antibodies. Recovered DNA was analyzed by QPCR, and normalized with DNA inputs. (B) Lysates from COS-1 cells transfected with the indicated constructs were immunoprecipitated with anti-T7 and immunoblotted with anti-Flag antibodies. (C) HepG2 cells were transiently-transfected with either the Gsc-luc or Xbra-luc reporters and increasing amounts (0.25 to 250 ng) of WT or 6Km Gsc. Luciferase activity from cells lysates is plotted as mean ± standard deviation. p values obtained from a two-tailed Student’s T-Test comparing the repressive activity of wild-type and Gsc 6Km are indicated. (A-C) Protein expression was confirmed by immunoblotting of total cell lysates with the indicated antibodies (bottom panels).

126

127

Figure 3.7: Model illustrating the effect of sumoylation on Gsc-mediated repression.

(A) Sumoylation does not alter Gsc repressive activity on promoters where Gsc is recruited through Foxh1 present within a complex containing Smad2 and Smad4. (B) Sumoylation enhances Gsc-mediated repression on promoters where Gsc is directly recruited through paired TAAT homeodomain binding sites.

128

3.5 Discussion

3.5.1 Goosecoid is modified by SUMO proteins

Goosecoid is a paired-like homeodomain-containing transcription factor expressed in the organizer region of gastrulating embryos [282, 289]. Gsc regulates the transcription of its target genes through direct DNA-binding to paired TAAT binding sites [290, 305, 306, 336] as well as indirect recruitment through the forkhead/winged helix transcription factor Foxh1 [355]. Although several Gsc target genes have been identified, little is known about the regulatory mechanisms that govern Gsc-mediated repression. Sumoylation is a post-translational modification that can regulate the activity of many transcription factors (reviewed in [365, 366]). While there are some examples where sumoylation enhances the activity of transcriptional regulators, it is typically found to mediate the repressive activity of target transcription factors. In this study, we demonstrate that PIAS proteins promote the sumoylation of Gsc in mammalian cells and that this post-translational modification regulates Gsc repressive activity in a promoter specific manner.

While none of the thirteen lysine residues present within Gsc conform to the canonical sumoylation consensus [365, 366], we found that five lysine residues present within the Gsc homeodomain and one lysine residue located in the Gsc C-terminal region were targeted for sumoylation as their mutation resulted in a significant decrease in overall Gsc sumoylation levels. Because Gsc with mutations in six lysines (Gsc 6Km) still retained some degree of sumoylation, additional residues may be targeted for covalent modification by SUMO. Alternatively, similar to ubiquitination, the residual sumoylation we observed on Gsc 6Km may reflect sumoylation at secondary sites as primary sites were no longer available.

129

3.5.2 Sumoylation potentiates Gsc repressive activity on direct targets

A number of studies have demonstrated that sumoylation regulates transcriptional activity by modulating protein-protein interactions [257]. For example, sumoylation of Elk-1 promotes association with HDAC2 [261], whereas SUMO modification of p300/CBP enhances recruitment of HDAC6 [262]. We previously demonstrated that Gsc recruits HDACs to mediate the repression of the Mixl1 gene, however, in this study we observed that sumoylation did not regulate Gsc repressive activity by modulating its interaction with HDACs. In agreement with the lack of effect of sumoylation on the Gsc-HDAC interaction, SUMO modification of Gsc did not appear to be critical for the regulation of Mixl1 promoter activity.

Although sumoylation of Gsc does not appear to modulate the transcriptional activity of the Mixl1 promoter, we found that the Gsc SUMO mutant (Gsc 6Km) had a decreased ability to repress the activity of the Gsc and Xbra promoters, both of which contain paired-homeodomain DNA binding sites [305, 306]. Previous studies have demonstrated that sumoylation of transcripton factors modulates their ability to bind their target DNA elements [265, 268, 364]. However, we found that although Gsc 6Km displayed lower levels of SUMO modification when compared to wild-type, it was still able to bind the Gsc promoter in protein-DNA immunoprecipitations, suggesting that sumoylation of Gsc does not affect its ability to bind DNA. In addition, sumoylation does not appear to regulate the recruitment of the Groucho family member TLE1 to Gsc. The precise mechanism for the differential effects on specific promoters remains unclear. However, paired-homeodomain proteins have been reported to homo- and hetero-dimerize to repress transcription [304, 367], so perhaps sumoylation can differentially regulate the nature of such interactions.

130

3.5.3 Stable expression of Gsc sumoylaton mutant alters MDA-MB-231 cell morphology

During gastrulation, mouse epiblast cells present within the primitive streak undergo epithelial- to-mesenchymal transition (EMT) and ingress within the embryo to become incorporated as either mesoderm or endoderm [368]. In Xenopus, gsc can cooperate with Mix1 to promote adhesion of blastocoel roof cells to fibronectin in in vitro cell adhesion assays [369]. EMT is also an important process involved in cancer metastasis, and it is thought that genes which regulate these processes during gastrulation may be reactivated during cancer progression [368]. Recently, Gsc, which was known to promote cell migration in Xenopus gastrulae [295], was also found to enhance the invasiveness of MDA-MB-231 breast carcinoma cells [312]. The mechanism through which this occurs has not been studied, however, Gsc likely regulates the expression of either direct or indirect target genes involved in promoting EMT. Expression of the Gsc sumoylation mutant (6Km) in MDA-MB-231 cells resulted in a change in cell morphology that led to an increase in cellular area, suggesting a role for Gsc in the transcriptional regulation of genes that are involved in controlling cell morphology. Previous studies have demonstrated that Gsc overexpression in MDA-MB-231 cells increases the number of lung metastasis in an orthotopic mouse model [312]. Alterations in cell adhesion properties often precede morphological changes [370], thus, Gsc may also be involved in the transcriptional regulation of adhesion molecules such as integrins, cadherins, catenins and other extracellular matrix molecules which have been shown to be important in the homing of tumor cells in new environments [312, 368]. Future work will be required to determine which genes are regulated by sumoylated forms of Gsc and how this regulation contributes to the promotion of EMT and the the development of cancer. It will also be interesting to examine whether sumoylated forms of Gsc regulate genes important for early gastrulation events in Xenopus.

131

CHAPTER 4

4 General Discussion and Future Directions

132

4.1 General Discussion

In the initial stages of embryogenesis, the zygote which is a totipotent cell divides and gives rise to equally totipotent blastomeres. After several rounds of cleavage, the blastomeres progressively loose their potency and begin to adopt different cell fates. As embryogenesis proceeds, cells continue to differentiate to give rise to various tissues and organs. Nodal-related signals have emerged as key regulators of early vertebrate development. Genetic and functional studies in Xenopus, zebrafish and mouse have demonstrated that Nodal signals are involved in mesoderm and endoderm induction, as well as neural patterning and establishment of left-right asymmetry [23]. Delineation of the molecular mechanisms underlying the transcriptional regulation of Nodal-responsive genes, thus, provides insights into how cells differentiate and contribute to the formation of specific tissues.

4.1.1 Foxh1 as a transcriptional repressor

Transcriptional regulation of Nodal-responsive genes occurs largely through the forkhead/winged helix transcription Foxh1 and Smads. While Foxh1 is commonly viewed as a transcriptional activator, the data presented here (Chapter 2) demonstrates that Foxh1 functions to negatively regulate Mixl1 expression during early mouse development. Our results corroborate emerging evidence which indicates that Foxh1 can act as a transcriptional repressor during Xenopus as well as zebrafish embryonic development. In Xenopus, Foxh1-depletion using anti-sense MO technology revealed that expression of Xnr-5 and Xnr-6, two potent inducers of dorsal mesoderm, was negatively regulated by Foxh1 activity [104]. Moreover, Foxh1 was shown to directly repress flk1 gene expression in zebrafish through the binding of a Foxh1 consensus site located within the promoter region of the gene [133]. While all three studies support a role for Foxh1 in transcriptional repression, my work is the first to define a molecular mechanism through which Foxh1 negatively regulates transcription. I demonstrated

133

that Foxh1 mediates the transcriptional repression of Mixl1 through binding of the paired-like homeodomain protein Gsc. Furthermore, I identified a novel region within Gsc capable of mediating repression of the Mixl1 promoter and determined that this region was required for the recruitment of HDACs. As well, I showed that Trichostatin A, a potent inhibitor of histone deacetylase activity, blocked Gsc-mediated repression of the Mixl1 promoter. We thus proposed a model where Gsc is recruited to the Mixl1 promoter through its association with DNA-bound Foxh1 and mediates transcriptional repression through the recruitment of HDACs (Chapter 2).

Although Foxh1 is implicated in numerous developmental processes including anterior-posterior axis formation, anterior primitive streak patterning, and anterior heart field development [105, 119, 120], thus far, only a few bona fide Foxh1 target genes have been identified in mouse [82, 102, 105, 107, 118]. Of these known targets, Mixl1 is the only gene to be negatively regulated by Foxh1 (this thesis). Given the broad expression pattern of Foxh1 during embryogenesis [107, 115, 118] and the complex phenotypes exhibited by Foxh1-mutant mice [119, 120], additional genes are likely to be regulated by Foxh1. Thus, an interesting area of study is the identification of novel Foxh1 target genes during early mouse development. To this end, an integrated approach consisting of a functional mapping of Smad/Foxh1 elements (SFE) coupled to an in silico-based search has been undertaken in the lab. This approach revealed 169 evolutionary conserved genes with putative SFEs in their regulatory regions [109]. Ranking of these genes for conserved positioning of the SFEs identified 21 genes that were more carefully examined. Five of the 21 genes are known Foxh1 targets, including Nodal, Lefty1, Pitx2, Gsc and Mixl1. Furthermore, experimental verification identified 5 novel targets that are positively regulated by Foxh1, including Lgr4, Lmo1, Hesx1, Aldh1a1, Aldh1a2 and Aldh1a3 [109].

Interestingly, of the top 21 candidate genes, Mix11 is the only gene found to be negatively regulated by Foxh1. Even if genes negatively regulated by Foxh1 are fewer in number than positively regulated genes, their identification is crucial to understanding the full regulatory function of Foxh1. In recent years, several groups have established protocols to perform high-

134

throughput whole-mount in situ hybridization (ISH) screens on mouse embryos of various developmental stages, which have proven to be successful in identifying novel regulators of mouse embryonic development [371-373]. Automated ISH to sections of E15.5 wild-type and Pax6-mutant mouse embryos identified 16 novel genes downstream of the Pax6 regulatory pathway [374]. Fourteen of these genes where confirmed by EMSA to be direct regulatory targets of Pax6 in the developing brain cortex [374]. Thus, future work to expand our list of genes negatively regulated by Foxh1 could examine the expression pattern of the remaining 148 putative Foxh1 target genes in both wild-type and Foxh1-null embryos through large-scale whole-mount in situ hybridization (WISH) analysis. Subsequently, the regulatory regions of genes that are repressed by Foxh1 would be further analyzed by EMSA and luciferase-reporter assays to determine whether the transcriptional regulation is direct.

Although the in silico method used by our group to identify novel Foxh1 targets has a 75% success rate in identifying putative Foxh1 targets, it did not recognize the confirmed Lefty 2 gene as a target due to divergent Foxh1 binding sites in the Lefty 2 regulatory elements [109]. Moreover, the full TGFβ-dependent activation of the Mef2c gene requires the binding of both Foxh1 and the homeodomain transcription factor Nkx2.5 [105]. Interestingly, the verified Foxh1 binding site present the TGFβ-responsive element of the Mef2c promoter does not conform with the Foxh1 consensus site, thus, like Lefty2, the Mef2c gene would not have been detected by the in silico search program. High throughput WISH screens could thus be an interesting avenue to pursue to not only describe whether specific genes are positively or negatively regulated by Foxh1 but also to identify novel Foxh1 targets that would not be detected by computational analysis. Together these approaches would provide global perspective on the role played by Foxh1 during early development.

135

4.1.2 Foxh1 recruits co-factors to regulate transcription

Forkhead transcription factors have been described to function as transcriptional activators as well as repressors, and in some instances, they can function as both depending on the cellular context and their binding partners [345, 346, 352, 375-380]. Emerging evidence suggests that Fox proteins cooperate with homeodomain transcription factors to regulate transcription of target genes [345-347, 352, 379, 380]. Consistently, our work and that of others indicates that Foxh1 can regulate transcription either positively or negatively through combinatorial interactions with homeodomain proteins such as Gsc (Chapter 2) and Nkx2.5 [105]. Interestingly, Foxa2 has been shown to play distinct regulatory functions through its interactions with diverse homeodomain transcription factors such as HNF6, Engrailed and Pax6 [347, 379, 380]. Combinatorial interactions of Fox transcription factors with multiple homeodomain proteins may be an important mechanism through which Fox proteins can diversify their transcriptional targets in a spatial and temporal manner. It would thus be of interest to determine whether Foxh1 can interact with additional homeodomain transcription factors proteins to mediate novel transcriptional programs critical for the development different tissues and organs. This could be accomplished using the LUMIER method [357], a mammalian cell-based protein-protein interaction assay. Subsequently, ChIP-on-chip assays could be performed to identify the target genes co-regulated by Foxh1 and its newly identified homeobox partners. Mapping interactions between Foxh1 and homeodomain proteins as well as identifying the transcriptional targets of such combinatorial interactions could provide insight on the multiple roles played by Foxh1 during early development.

4.1.3 Co-regulation of target genes by Foxh1 and Gsc

My work demonstrated that Foxh1 recruited Gsc to negatively regulate the activity of the Mixl1 promoter. Supporting my biochemical data, in situ hybridization and QPCR analysis

136 demonstrated that Mixl1 expression was upregulated in both Foxh1- and Gsc-mutant embryos. Interestingly, we observed that the range of Mixl1 mRNA expression in Gsc-null embryos was wider than in wild-type embryos. Given that the early development of Gsc-null embryos is normal, compensatory mechanisms are likely coming into play to overcome ectopic Mixl1 expression and allow gastrulation to proceed normally. In addition to Gsc, a second closely related gene designated Gsc-2 has been described in mouse [287, 381]. It would therefore be of interest to determine whether Gsc-2 could interact with Foxh1 and mediate Mixl1 repression in vivo. At present, Gsc-2 expression as only been examined in mid-gestation stages (E8.5 to E11.5) of mouse development and was found to be detected in primordial germ cells and the developing brain [382]. It is thus imperative to determine whether Gsc-2 is expressed during gastrulation stages and if its expression pattern is consistent with a role in repressing Mixl1 epxression.

Forkhead transcription factors have also been shown to mediate transcriptional repression through the recruitment of transcriptional co-repressors of the TLE/Groucho family via engrailed homology 1 (eh1) -like motifs [375, 383, 384]. A systemic analysis of Fox proteins revealed the presence of eh1-like motifs in over 50% of Fox proteins [385]. The family of forkhead transcription factors is organized in 19 subclasses, 10 of which were described to contain eh1- like motifs, including the A, B, C, D, E, G, H, I, L, and Q subclasses. The prevalence of eh1-like motifs in Fox proteins suggests that recruitment of TLE/Groucho protein may be a common mechanism through which Fox proteins repress gene activity. Recent work from Daniel Kessler’s laboratory showed that Groucho cooperates with Foxh1 to repress Nodal-responsive genes as well as mesoderm induction during early Xenopus development [386]. Similar to its Xenopus homologue, mouse Foxh1 contains an eh1-like domain in its C-terminus, and likely interacts with murine Groucho-related gene (grg) products. An interesting avenue to pursue would be to determine whether Foxh1 recruits Grg proteins to negatively regulate Mixl1 as well as other genes during mouse development. As levels of Nodal signaling activity are critical in determining cell fate during early development, notably during the allocation of mesendodermal

137

lineages, it is conceivable that Foxh1, a major regulator of Nodal signaling, utilizes various cofactors to maintain appropriate gene responses during embryogenesis.

4.1.4 Regulation of Gsc activity

An important aspect of the work presented in this thesis was to understand how post-translational modifications regulate the transcriptional activity of Gsc. In the last ten years, sumoylation has emerged as an important mechanism through which the activity of numerous transcription factors is controlled. In this thesis (Chapter 3), I demonstrated that Gsc was covalently modified by SUMO moieties in mammalian cells. Moreover, I showed Gsc sumoylation occurred on multiple lysines and was promoted by various members of the PIAS family of proteins, which are known to function as SUMO E3 ligases for numerous transcription factors.

Sumoylation regulates the activity of transcription factor through several distinct mechanisms, including changes in subcellular localization, DNA binding activity and protein-protein interactions. My work demonstrated that sumoylation did not appear to regulate the ability of Gsc to interact with either Foxh1 or HDACs, which is consistent with the observation that sumoylation did not affect the repressive activity of Gsc on indirect targets such as Mixl1. Since five of the six lysines targeted for sumoylation are located within the homeodomain, I reasoned that sumoylation might be regulating the repression of direct targets such as the Xbra and Gsc promoters through the modulation of Gsc DNA binding activity. Though Gsc 6Km, a mutant in which sumoylation is significantly reduced, had a decrease ability to repress transcription of direct target targets such as the Xbra and Gsc promoters, it retained the ability to bind DNA. While Gsc 6Km may still interact with DNA, it might be associating with its target sequence with a lower affinity. It would thus be of interest to perform competitive EMSA analysis to calculate the DNA binding affinity of native and sumoylated Gsc to determine whether the reduced ability of Gsc 6Km to inhibit transcription is related changes in its affinity for DNA.

138

Gsc contains an eh1-like motif (also known as Goosecoid Engrailed Homology (GEH) motif) at its N-terminus [301]. Given that this motif has been shown to be required for direct transcriptional repression through the recruitment of the Groucho co-repressor in Drosophila [301], I tested whether sumoylation regulated the interaction between Gsc and TLE1, a mammalian homologue of Groucho. Gsc 6Km interacted with TLE1 as well as the wild-type protein, thus, suggesting that sumoylation does not appear to regulate the repressive activity of Gsc on direct targets through the recruitment of Groucho-homologues. Future studies aimed at identifying novel Gsc binding proteins would provide insight onto the mechanism by which sumoylation regulates Gsc activity on direct targets.

Gsc which was known for its role in promoting cell migration during Xenopus gastrulation was recently shown to enhance the migratory abilities of human MDA-MB-231 breast carcinoma cells in transwell migration assays [295, 312]. In my hands, however, stable expression of wild- type Gsc as well as a Gsc sumoylation mutant (Gsc 6Km) did not significantly affect cell migration in in vitro wound-healing assays (data not shown). Although the transwell migration and in vitro wound-healing assays are commonly used to assess the invasiveness potential of cells, these two assays examine different aspects of cell invasion. While the transwell migration assay evaluates the intravasation potential of cells, the in vitro wound-healing assay assesses the motility of cells. Thus, the discrepancy may be due to the fact that the assays reflect distinct aspects of cell invasion. Alternatively, the incongruous results may be due differences in Gsc expression in the stable lines I generated and those previously described by Hartwell et al. [312]. Studies in Xenopus have demonstrated that ectopic expression of Gsc dorsalizes the ventral marginal zone in a dose-dependent manner [291]. Therefore, it is conceivable that different Gsc expression levels in MDA-MB-231 cells may result in distinct phenotypes. Future studies should be directed at addressing this issue. Nevertheless, I did observe differences in the morphology of the cells expressing Gsc 6Km when compared to cells expressing either vector control or wild-type Gsc. Studies by Hartwell et al. demonstrated that stable expression of Gsc in immortalized human mammary epithelial cells (HMEC) and Madin-Darby canine kidney

139

(MDCK) epithelial cells elicited EMT which was evidenced by the loss of epithelial markers such as E-cadherin, α- and γ-catenin and induction of mesenchymal markers such as vimentin and N-cadherin [312]. As Gsc functions as a transcriptional repressor, it may be directly or indirectly regulating the expression of markers mentioned above as well as other genes involved in cell adhesion and cytoskeletal rearrangements. Sumoylation may be fine-tuning the transcriptional activity of Gsc towards these genes or direct Gsc to regulate a completely distinct set of genes leading to different cell morphology. It will be important to characterize the transcriptional programs of native and sumoylated forms of Gsc to determine the biological relevance of Gsc sumoylation in neoplastic disease. Alternatively, the function of Gsc sumoylation in vivo could be explored in Xenopus. For instance, the activity of Gsc 6Km could be assessed by microinjecting in Xenopus embryos and examining the response of endogenous Gsc target genes. The effect of Gsc 6Km on axis duplication, cell migration and mesodermal patterning could also be characterized.

4.2 Conclusions

In summary, the work presented in this thesis delineates novel mechanisms regulating the activity of two transcription factors which are known to be critical for early vertebrate development. In the first study, I present the finding that Foxh1, which was commonly known to function as a transcriptional activator, could also function as a repressor through the recruitment

of the homeodomain protein Gsc. In the second study, I demonstrated that sumoylation regulates the repressive activity of Gsc in promoter-specific manner and may be involved in controlling cell morphogenesis. Together, these studies contribute to our understanding of transcriptional regulation during early embryogenesis and possibly cancer development.

140

CHAPTER 5

5 References

141

1. Kimelman, D., Mesoderm induction: from caps to chips. Nat Rev Genet, 2006. 7(5): p. 360-72.

2. Heasman, J., Patterning the early Xenopus embryo. Development, 2006. 133(7): p. 1205- 17.

3. Keller, R. and D. Shook, Gastrulation in Amphibians, in Gastrulation: From Cells to Embryos, C.D. Stern, Editor. 2004, Cold Spring Harbor Laboratory Press: Cold Spring Harbor, N.Y. p. 171-203.

4. Gilbert, S.F., Developmental biology. 6th ed. 2000, Sunderland, Mass.: Sinauer Associates. xviii, 749 p.

5. Spemann, H. and H. Mangold, Uber induction von embryonanlagen durch implantation artfremder organis atoren. Roux's Arch. Dev. Biol., 1924. 100: p. 599-638.

6. Robb, L. and P.P. Tam, Gastrula organiser and embryonic patterning in the mouse. Semin. Cell. Dev. Biol., 2004. 15(5): p. 543-54.

7. Schier, A.F. and W.S. Talbot, Molecular genetics of axis formation in zebrafish. Annu. Rev. Genet., 2005. 39: p. 561-613.

8. Kane, D.A. and R.M. Warga, Teleost Gastrulation, in Gastrulation: From Cells to Embryo, C.D. Stern, Editor. 2004, Cold Spring Harbor Laboratory Press: Cold Spring Harbor, N.Y. p. 157-169.

9. Beddington, R.S.P. and E.J. Robertson, Axis development and early asymmetry in mammals. Cell, 1999. 96: p. 195-209.

10. Lu, C.C., J. Brennan, and E.J. Robertson, From fertilization to gastrulation: axis formation in the mouse embryo. Curr Opin Genet Dev, 2001. 11(4): p. 384-92.

11. Tam, P.P.L. and J.M. Gad, Gastrulation in the Mouse Embryo, in Gastrualtion: From Cells to Embryo, C.D. Stern, Editor. 2004, Cold Spring Harbor Laboratory Press: Cold Spring Harbor, N.Y. p. 233-262.

12. Beddington, R.S.P., Induction of a second neural axis by the mouse node. Dev., 1994. 120: p. 613-620.

13. Tam, P.P., et al., Lineage and functional analyses of the mouse organizer. Cold Spring Harb Symp Quant Biol, 1997. 62: p. 135-44.

14. Kinder, S.J., et al., The organizer of the mouse gastrula is composed of a dynamic population of progenitor cells for the axial mesoderm. Development, 2001. 128(18): p. 3623-34.

15. Thomas, P. and R. Beddington, Anterior primitive endoderm may be responsible for patterning the anterior neural plate in the mouse embryo. Curr Biol, 1996. 6(11): p. 1487-96.

142

16. Tam, P.P. and K.A. Steiner, Anterior patterning by synergistic activity of the early gastrula organizer and the anterior germ layer tissues of the mouse embryo. Development, 1999. 126(22): p. 5171-9.

17. Camus, A., et al., Absence of Nodal signaling promotes precocious neural differentiation in the mouse embryo. Dev Biol, 2006. 295(2): p. 743-55.

18. Hogan, B., Manipulating the mouse embryo : a laboratory manual. 2nd ed. 1994, Plainview, N.Y.: Cold Spring Harbor Laboratory Press. xvii, 497 p.

19. Houston, D.W. and C. Wylie, The Role of Wnts in Gastrulation, in Gastrulation: From Cells to Embryo, C.D. Stern, Editor. 2004, Cold Spring Harbor Laboratory Press: Cold Spring Harbor, N.Y. p. 521-537.

20. Sivak, J. and E. Amaya, FGF signaling during gastrulation, in Gastrulation: From Cells to Embryo, C.D. Stern, Editor. 2004, Cold Spring Harbor Laboratory Press: Cold Spring Harbor, N.Y. p. 463-473.

21. Sanjuan-Munoz, I. and A.H. Brivanlou, Modulation of BMP signaling during vertebrate gastrulation, in Gastrulation: From Cells to Embryo, C.D. Stern, Editor. 2004, Cold Spring Harbor Laboratory Press: Cold Spring Harbor, N.Y. p. 475-489.

22. Kessler, D.S., Activin and Vg1 and the search for Embryonic Inducers, in Gastrulation: From Cells to Embryo, C.D. Stern, Editor. 2004, Cold Spring Harbor Laboratory Press: Cold Spring Harbor, N.Y. p. 505-520.

23. Schier, A.F., Nodal signaling in vertebrate development. Annu Rev Cell Dev Biol, 2003. 19: p. 589-621.

24. Conlon, F.L., et al., A primary requirement for nodal in the formation and maintenance of the primitive streak in the mouse. Development, 1994. 120: p. 1919-1928.

25. Feldman, B., et al., Zebrafish organizer development and germ-layer formation require nodal-related signals. Nature, 1998. 395(6698): p. 181-5.

26. Rebagliati, M.R., et al., cyclops encodes a nodal-related factor involved in midline signalling. Proc. Natl. Acad. Sci. USA, 1998. 95: p. 9932-9937.

27. Rebagliati, M.R., et al., Zebrafish nodal related genes are implicated in axial patterning and establishing left-right asymmetry. Dev. Biol., 1998. 199: p. 261-272.

28. Sampath, K., et al., Induction of the zebrafish ventral brain and floorplate required cyclops/nodal signalling. Nature News & Views, 1998. 395: p. 185-189.

29. Zhou, X., et al., Nodal is a novel TGF-ß-like gene expressed in the mouse node during gastrulation. Nature, 1993. 361: p. 543-547.

30. Jones, C.M., et al., Nodal-related signals induce axial mesodem and dorsalize mesoderm during gastrulation. Dev., 1995. 121: p. 3651-3662.

143

31. Beck, S., et al., Extraembryonic proteases regulate Nodal signalling during gastrulation. Nat Cell Biol, 2002. 4(12): p. 981-5.

32. Constam, D.B. and E.J. Robertson, SPC4/PACE4 regulates a TGFbeta signaling network during axis formation. Genes Dev, 2000. 14(9): p. 1146-55.

33. Constam, D.B. and E.J. Robertson, Tissue-specific requirements for the proprotein convertase furin/SPC1 during embryonic turning and heart looping. Development, 2000. 127(2): p. 245-54.

34. Birsoy, B., et al., XPACE4 is a localized pro-protein convertase required for mesoderm induction and the cleavage of specific TGFbeta proteins in Xenopus development. Development, 2005. 132(3): p. 591-602.

35. Wrana, J.L., et al., TGFβ signals through a heteromeric protein kinase receptor complex. Cell, 1992. 71: p. 1003-1014.

36. Massagué, J., TGF-β Signal Transduction. Annu. Rev. Biochem., 1998. 67: p. 753-91.

37. Wrana, J.L., et al., Mechanism of activation of the TGF–ß receptor. Nature, 1994. 370: p. 341-347.

38. Wieser, R., J.L. Wrana, and J. Massagué, GS domain mutations that constitutively activate TßR-I, the downstream signalling component in the TGF-ß receptor complex. EMBO J., 1995. 14: p. 2199-2208.

39. Reissmann, E., et al., The orphan receptor ALK7 and the Activin receptor ALK4 mediate signaling by Nodal proteins during vertebrate development. Genes Dev, 2001. 15(15): p. 2010-22.

40. Song, J., et al., The type II activin receptors are essential for egg cylinder growth, gastrulation, and rostral head development in mice. Dev Biol, 1999. 213(1): p. 157-69.

41. Gu, Z., et al., The type I activin receptor ActRIB is required for egg cylinder organization and gastrulation in the mouse. Genes Dev., 1998. 12: p. 844-857.

42. Matzuk, M.M., T.R. Kumar, and A. Bradley, Different phenotypes for mice deficient in either activins or activin receptor type II. Nature, 1995. 374: p. 356-359.

43. Oh, S.P. and E. Li, The signaling pathway mediated by the type IIB activin receptor controls axial patterning and lateral assymetry in the mouse. Genes Dev., 1997. 11: p. 1812-1826.

44. Jornvall, H., et al., ALK7, a receptor for nodal, is dispensable for embryogenesis and left- right patterning in the mouse. Mol Cell Biol, 2004. 24(21): p. 9383-9.

45. Aoki, T.O., et al., Regulation of nodal signalling and mesendoderm formation by TARAM-A, a TGFbeta-related type I receptor. Dev Biol, 2002. 241(2): p. 273-88.

144

46. Minchiotti, G., et al., Membrane-anchorage of Cripto protein by glycosylphosphatidylinositol and its distribution during early mouse development. Mech Dev, 2000. 90(2): p. 133-42.

47. Onuma, Y., C.Y. Yeo, and M. Whitman, XCR2, one of three Xenopus EGF-CFC genes, has a distinct role in the regulation of left-right patterning. Development, 2006. 133(2): p. 237-50.

48. Shen, M.M., H. Wang, and P. Leder, A differential display strategy identifies Cryptic, a novel EGF-related gene expressed in the axial and lateral mesoderm during mouse gastrulation. Development, 1997. 124(2): p. 429-42.

49. Dorey, K. and C.S. Hill, A novel Cripto-related protein reveals an essential role for EGF-CFCs in Nodal signalling in Xenopus embryos. Dev Biol, 2006. 292(2): p. 303-16.

50. Colas, J.F. and G.C. Schoenwolf, Subtractive hybridization identifies chick-cripto, a novel EGF-CFC ortholog expressed during gastrulation, neurulation and early cardiogenesis. Gene, 2000. 255(2): p. 205-17.

51. Dono, R., et al., The murine cripto gene: expression during mesoderm induction and early heart morphogenesis. Development, 1993. 118(4): p. 1157-68.

52. Zhang, J., W.S. Talbot, and A.F. Schier, Positional cloning identifies zebrafish one-eyed pinhead as a permissive EGF-related ligand required during gastrulation. Cell, 1998. 92(2): p. 241-51.

53. Kinoshita, N., J. Minshull, and M.W. Kirschner, The identification of two novel ligands of the FGF receptor by a yeast screening method and their activity in Xenopus development. Cell, 1995. 83(4): p. 621-30.

54. Ciccodicola, A., et al., Molecular characterization of a gene of the 'EGF family' expressed in undifferentiated human NTERA2 teratocarcinoma cells. Embo J, 1989. 8(7): p. 1987-91.

55. Gritsman, K., et al., The EGF-CFC protein one-eyed pinhead is essential for nodal signaling. Cell, 1999. 97: p. 121-132.

56. Ding, J., et al., Cripto is required for correct orientation of the anterior-posterior axis in the mouse embryo. Nature, 1998. 395(6703): p. 702-7.

57. Gaio, U., et al., A role of the cryptic gene in the correct establishment of the left-right axis. Curr Biol, 1999. 9(22): p. 1339-42.

58. Yan, Y.T., et al., Conserved requirement for EGF-CFC genes in vertebrate left-right axis formation. Genes Dev, 1999. 13(19): p. 2527-37.

59. Bamford, R.N., et al., Loss-of-function mutations in the EGF-CFC gene CFC1 are associated with human left-right laterality defects. Nat Genet, 2000. 26(3): p. 365-9.

145

60. Yeo, C. and M. Whitman, Nodal signals to Smads through Cripto-dependent and Cripto- independent mechanisms. Mol Cell, 2001. 7(5): p. 949-57.

61. Yan, Y.T., et al., Dual roles of Cripto as a ligand and coreceptor in the nodal signaling pathway. Mol Cell Biol, 2002. 22(13): p. 4439-49.

62. Moustakas, A., S. Souchelnytskyi, and C.-H. Heldin, Smad regulation in TGF-β signal transduction. J. Cell Sci., 2001. 114: p. 4359-4369.

63. Attisano, L. and J.L. Wrana, Signal transduction by the TGF-β superfamily. Science, 2002. 296: p. 1646-1647.

64. Shi, Y. and J. Massague, Mechanisms of TGF-beta signaling from cell membrane to the nucleus. Cell, 2003. 113(6): p. 685-700.

65. Lee, M.A., J. Heasman, and M. Whitman, Timing of endogenous activin-like signals and regional specification of the Xenopus embryo. Development, 2001. 128(15): p. 2939-52.

66. Kumar, A., et al., Nodal signaling uses activin and transforming growth factor-beta receptor-regulated Smads. J Biol Chem, 2001. 276(1): p. 656-61.

67. Miyazono, K., S. Maeda, and T. Imamura, Smad transcriptional co-activators and co- repressors, in Smad signal transduction: Smads in proliferation, differentiation and disease, P. ten Dijke and C.H. Heldin, Editors. 2006, Springer: Dordrecht. p. 277-293.

68. Feng, X.H. and R. Derynck, Specificity and versatility in tgf-beta signaling through Smads. Annu Rev Cell Dev Biol, 2005. 21: p. 659-93.

69. ten Dijke, P., K. Miyazono, and C.-H. Heldin, Signaling inputs converge on nuclear effectors in TGF-β signaling. Trends Biochem. Sci., 2000. 25: p. 64-70.

70. Siegel, P.M. and J. Massague, Cytostatic and apoptotic actions of TGF-beta in homeostasis and cancer. Nat Rev Cancer, 2003. 3(11): p. 807-21.

71. Izzi, L. and L. Attisano, Regulation of the TGFbeta signalling pathway by ubiquitin- mediated degradation. Oncogene, 2004. 23(11): p. 2071-8.

72. de Caestecker, M.P., E. Piek, and A.B. Roberts, Role of Transforming Growth Factor-β signaling in cancer. J. Natl. Cancer Inst., 2000. 92: p. 1388-1402.

73. Tsukazaki, T., et al., SARA, a FYVE domain protein that recruits Smad2 to the TGF-β receptor. Cell, 1998. 95: p. 779-791.

74. Attisano, L. and J.L. Wrana, Smads as transcriptional co-modulators. Curr. Op. Cell Biol., 2000. 12: p. 235-243.

75. Wotton, D. and J. Massagué, Transcriptional control by the TGFβ/Smad signaling system. EMBO J., 2000. 19: p. 745-54.

146

76. Baker, J.C. and R. Harland, A novel mesoderm inducer, Madr2, functions in the activin signal transduction pathway. Genes Dev., 1996. 10: p. 1880-1889.

77. Graff, J.M., A. Bansal, and D.A. Melton, Xenopus Mad proteins transduce distinct subsets of signals for the TGFß superfamily. Cell, 1996. 85: p. 479-487.

78. Dick, A., et al., Cloning and characterization of zebrafish smad2, smad3 and smad4. Gene, 2000. 246(1-2): p. 69-80.

79. Yeo, C.-Y. and M. Whitman, Nodal signals to Smads through cripto-dependent and cripto-independent mechanisms. Mol. Cell, 2001. 7: p. 949-957.

80. Jia, S., et al., Smad2 and Smad3 are required for mesendoderm induction by TGF-beta /Nodal signals in zebrafish. J Biol Chem, 2007.

81. Shi, Y., et al., Crystal structure of a Smad MH1 domain bound to DNA: insights on DNA binding in TGF-β signalling. Cell, 1998. 94(5): p. 585-594.

82. Labbé, E., et al., Smad2 and Smad3 positively and negatively regulate TGFβ-dependent transcription through the forkhead DNA binding protein, FAST2. Mol. Cell, 1998. 2: p. 109-120.

83. Piek, E., et al., Functional characterization of transforming growth factor β signaling in Smad2- and Smad3-deficient fibroblasts. J. Biol. Chem., 2001. 276: p. 19945-19953.

84. Datto, M.B., et al., Targeted disruption of Smad3 reveals an essential role in Transforming Growth Factor beta-mediated signal transduction. Mol. Cell. Biol., 1999. 19: p. 2495-2504.

85. Yang, X., et al., Targeted disruption of Smad3 results in impaired mucosal immunity and diminished T cell responsiveness to TGF beta. EMBO J., 1999. 18(5): p. 1280-1291.

86. Zhu, Y., et al., Smad3 Mutant Mice Develop Metastatic Colorectal Cancer. Cell, 1998. 94: p. 703-714.

87. Nomura, M. and E. Li, Smad2 role in mesoderm formation, left-right patterning and craniofacial development. Nature, 1998. 393: p. 786.

88. Brennan, J., et al., Nodal signalling in the epiblast patterns the early mouse embryo. Nature, 2001. 411(6840): p. 965-9.

89. Waldrip, W.R., et al., Smad2 Signalling in Extraembryonic Tissues Determines Anterior- Posterior Polarity of the Early Mouse Embryo. Cell, 1998. 92: p. 797-808.

90. Tremblay, K.D., et al., Formation of the definitive endoderm in mouse is a Smad2- dependent process. Develop., 2000. 127: p. 3079-3090.

91. Vincent, S.D., et al., Cell fate decisions within the mouse organizer are governed by graded Nodal signals. Genes Dev, 2003. 17(13): p. 1646-62.

147

92. Dunn, N.R., et al., Combinatorial activities of Smad2 and Smad3 regulate mesoderm formation and patterning in the mouse embryo. Development, 2004. 131(8): p. 1717-28.

93. Norris, D.P., et al., The Foxh1-dependent autoregulatory enhancer controls the level of nodal signals in the mouse embryo. Develop., 2002. 129: p. 3455-3468.

94. Sirard, C., et al., The tumor suppressor gene Smad4/Dpc4 is required for gastrulation and later for anterior development of the mouse embryo. Genes & Dev., 1998. 12: p. 107- 119.

95. Yang, X., et al., The tumor suppressor SMAD4/ DPC4 is essential for epiblast proliferation and mesoderm induction in mice. Proc. Natl. Acad. Sci. USA, 1998. 95: p. 3667-3672.

96. Chu, G.C., et al., Differential requirements for Smad4 in TGFbeta-dependent patterning of the early mouse embryo. Development, 2004. 131(15): p. 3501-12.

97. He, W., et al., Hematopoiesis controlled by distinct TIF1gamma and Smad4 branches of the TGFbeta pathway. Cell, 2006. 125(5): p. 929-41.

98. Massague, J., J. Seoane, and D. Wotton, Smad transcription factors. Genes Dev, 2005. 19(23): p. 2783-810.

99. Chen, X., M.J. Rubock, and M. Whitman, A transcriptional partner for MAD proteins in TGF-β signalling. Nature, 1996. 383: p. 691-696.

100. Osada, S.I., et al., Activin/nodal responsiveness and asymmetric expression of a Xenopus nodal-related gene converge on a FAST-regulated module in intron 1. Development, 2000. 127(11): p. 2503-14.

101. Ryan, K., et al., The Xenopus eomesodermin promoter and its concentration-dependent response to activin. Mech. Dev., 2000. 94: p. 133-146.

102. Shiratori, H., et al., Two-step regulation of left-right asymmetric expression of Pitx2: initiation by nodal signaling and maintenance by Nkx2. Mol. Cell, 2001. 7: p. 137-149.

103. Watanabe, M., et al., Regulation of the Lim-1 gene is mediated through conserved FAST- 1/FoxH1 sites in the first intron. Dev Dyn, 2002. 225(4): p. 448-56.

104. Kofron, M., et al., New roles for FoxH1 in patterning the early embryo. Development, 2004. 131(20): p. 5065-78.

105. von Both, I., et al., Foxh1 is essential for development of the anterior heart field. Dev Cell, 2004. 7(3): p. 331-45.

106. Hart, A.H., et al., Transcriptional regulation of the homeobox gene Mixl1 by TGF-beta and FoxH1. Biochem Biophys Res Commun, 2005. 333(4): p. 1361-9.

107. Takaoka, K., et al., The mouse embryo autonomously acquires anterior-posterior polarity

148

at implantation. Dev Cell, 2006. 10(4): p. 451-9.

108. Pei, W., et al., An early requirement for maternal FoxH1 during zebrafish gastrulation. Dev Biol, 2007. 310(1): p. 10-22.

109. Silvestri, C., et al., Genome-wide Identification of Smad/Foxh1 Targets Reveals a Role for Foxh1 in Retinoic Acid Regulation and Forebrain Development analysis of Foxh1 target genes. Dev Cell, 2008.

110. Zhou, S., et al., Characterization of human FAST-1, a TGFβ and activin signal transducer. Mol. Cell, 1998. 2: p. 121-127.

111. Chen, X., et al., Smad4 and FAST-1 in the assembly of activin-responsive factor. Nature, 1997. 389: p. 85-89.

112. Germain, S., et al., Homeodomain and winged-helix transcription factors recruit activated Smads to distinct promoter elements via a common Smad interaction motif. Genes Dev, 2000. 14(4): p. 435-51.

113. Randall, R.A., et al., Recognition of phosphorylated-Smad2-containing complexes by a novel Smad interaction motif. Mol Cell Biol, 2004. 24(3): p. 1106-21.

114. Howell, M., G.J. Inman, and C.S. Hill, A novel Xenopus Smad-interacting forkhead transcription factor (XFast-3) cooperates with XFast-1 in regulating gastrulation movements. Development, 2002. 129(12): p. 2823-34.

115. Weisberg, E., et al., A mouse homologue of FAST-1 transduces TGFβ superfamily signals and is expressed during early embryogenesis. Mech. Dev., 1998. 79: p. 17-27.

116. Pogoda, H.-M., et al., The zebrafish forkhead transcription factor FoxH1/Fast1 is a modulator of nodal signaling required for organizer formation. Curr. Biol., 2000. 10: p. 1041-1049.

117. Sirotkin, H.I., et al., fast1 is required for the development of dorsal axial structures in zebrafish. Curr. Biol., 2000. 10: p. 1051-1054.

118. Saijoh, Y., et al., Left-right asymmetric expression of lefty2 and nodal is induced by a signaling pathway that includes the transcription factor FAST2. Mol. Cell, 2000. 5: p. 35-47.

119. Yamamoto, M., et al., The transcription factor FoxH1 (FAST) mediates nodal signaling during anterior-posterior patterning and node formation in the mouse. Genes Dev., 2001. 15: p. 1242-1256.

120. Hoodless, P.A., et al., FoxH1 (Fast) functions to specify the anterior primitive streak in the mouse. Genes Dev., 2001. 15: p. 1257-12371.

121. Attisano, L., et al., The transcriptional role of Smads and FAST (FoxH1) in TGFβ and activin signalling. Mol. Cell. Endo., 2001. 180: p. 3-11.

149

122. Whitman, M., Nodal signaling in early vertebrate embryos: Themes and variations. Dev. Cell, 2001. 1: p. 605-617.

123. Watanabe, M. and M. Whitman, FAST-1 is a key maternal effector of mesoderm inducers in the early Xenopus embryo. Develop., 1999. 126: p. 5621-5634.

124. Vogel, A.M. and T. Gerster, Promoter activity of the zebrafish bhikhari retroelement requires an intact activin signaling pathway. Mech. Dev., 1999. 85: p. 133-146.

125. Saijoh, Y., et al., Two nodal-responsive enhancers control left-right asymmetric expression of Nodal. Dev Dyn, 2005. 232(4): p. 1031-6.

126. Yashiro, K., et al., Distinct transcriptional regulation and phylogenetic divergence of human LEFTY genes. Genes Cells, 2000. 5(5): p. 343-57.

127. Piccolo, S., et al., The head inducer Cerberus is a multifunctional antagonist of Nodal, BMP and Wnt signals. Nature, 1999. 397(6721): p. 707-10.

128. Hashimoto, H., et al., The Cerberus/Dan-family protein Charon is a negative regulator of Nodal signaling during left-right patterning in zebrafish. Development, 2004. 131(8): p. 1741-53.

129. Agius, E., et al., Endodermal Nodal-related signals and mesoderm induction in Xenopus. Development, 2000. 127(6): p. 1173-83.

130. Tavares, A.T., et al., Cerberus is a feedback inhibitor of Nodal asymmetric signaling in the chick embryo. Development, 2007. 134(11): p. 2051-60.

131. Shiratori, H., et al., Conserved regulation and role of Pitx2 in situs-specific morphogenesis of visceral organs. Development, 2006. 133(15): p. 3015-25.

132. Hilton, E., M. Rex, and R. Old, VegT activation of the early zygotic gene Xnr5 requires lifting of Tcf-mediated repression in the Xenopus blastula. Mech Dev, 2003. 120(10): p. 1127-38.

133. Choi, J., et al., FoxH1 negatively modulates flk1 gene expression and vascular formation in zebrafish. Dev Biol, 2007. 304(2): p. 735-44.

134. Ring, C., et al., The role of a Williams-Beuren syndrome-associated helix-loop-helix domain-containing transcription factor in activin/nodal signaling. Genes Dev, 2002. 16(7): p. 820-35.

135. Dou, C., et al., BF-1 interferes with Transforming Growth Factor β signaling by associating with Smad partners. Mol. Cell. Biol., 2000. 20: p. 6201-6211.

136. Yun, C.H., et al., Negative regulation of Activin/Nodal signaling by SRF during Xenopus gastrulation. Development, 2007. 134(4): p. 769-77.

137. Iratni, R., et al., Inhibition of excess nodal signaling during mouse gastrulation by the

150

transcriptional corepressor DRAP1. Science, 2002. 298(5600): p. 1996-9.

138. Kunwar, P.S., et al., Mixer/Bon and FoxH1/Sur have overlapping and divergent roles in Nodal signaling and mesendoderm induction. Development, 2003. 130(23): p. 5589-99.

139. Cordenonsi, M., et al., Links between tumor suppressors: p53 is required for TGF-beta gene responses by cooperating with Smads. Cell, 2003. 113(3): p. 301-14.

140. Janknecht, R., N.J. Wells, and T. Hunter, TGF-β-stimulated cooperation of Smad proteins with the coactivators CBP/p300. Genes Dev., 1998. 12: p. 2114-2119.

141. Feng, X.-H., et al., The tumor suppressor Smad4/DPC4 and transcriptional adaptor CBP/p300 are coactivators for Smad3 in TGF-β-induced transcriptional activation. Genes Dev., 1998. 12: p. 2153-2163.

142. Pouponnot, C., L. Jayaraman, and J. Massagué, Physical and Functional Interaction of SMADs and p300/CBP. J. Biol. Chem., 1998. 273(36): p. 22865-22868.

143. Shen, X., et al., TGF-beta-induced Phosphorylation of Smad3 Regulates Its Interaction with Coactivator p300/CREB-binding protein. Mol Biol Cell, 1998. 9(12): p. 3309-3319.

144. de Caestecker, M.P., et al., The Smad4 activation domain (SAD) is a proline-rich, p300- dependent transcriptional activation domain. J Biol Chem, 2000. 275(3): p. 2115-22.

145. Kahata, K., et al., Regulation of transforming growth factor-beta and bone morphogenetic protein signalling by transcriptional coactivator GCN5. Genes Cells, 2004. 9(2): p. 143-51.

146. Itoh, S., et al., The transcriptional co-activator P/CAF potentiates TGF-beta/Smad signaling. Nucleic Acids Res, 2000. 28(21): p. 4291-8.

147. Dennler, S., et al., The steroid receptor co-activator-1 (SRC-1) potentiates TGF- beta/Smad signaling: role of p300/CBP. Oncogene, 2005. 24(11): p. 1936-45.

148. Shioda, T., et al., Transcriptional activating activity of Smad4: roles of SMAD hetero- oligomerization and enhancement by an associating transactivator. Proc Natl Acad Sci U S A, 1998. 95(17): p. 9785-90.

149. Yahata, T., et al., The MSG1 non-DNA-binding transactivator binds to the p300/CBP coactivators, enhancing their functional link to the Smad transcription factors. J Biol Chem, 2000. 275(12): p. 8825-34.

150. Postigo, A.A., Opposing functions of ZEB proteins in the regulation of the TGFbeta/BMP signaling pathway. Embo J, 2003. 22(10): p. 2443-52.

151. Postigo, A.A., et al., Regulation of Smad signaling through a differential recruitment of coactivators and corepressors by ZEB proteins. Embo J, 2003. 22(10): p. 2453-62.

152. Long, J., et al., Activation of Smad transcriptional activity by protein inhibitor of

151

activated STAT3 (PIAS3). Proc Natl Acad Sci U S A, 2004. 101(1): p. 99-104.

153. Bai, R.Y., et al., SMIF, a Smad4-interacting protein that functions as a co-activator in TGFbeta signalling. Nat Cell Biol, 2002. 4(3): p. 181-90.

154. Ross, S., et al., Smads orchestrate specific histone modifications and chromatin remodeling to activate transcription. Embo J, 2006. 25(19): p. 4490-502.

155. Simonsson, M., et al., The DNA binding activities of Smad2 and Smad3 are regulated by coactivator-mediated acetylation. J Biol Chem, 2006.

156. Inoue, Y., et al., Smad3 is acetylated by p300/CBP to regulate its transactivation activity. Oncogene, 2006.

157. Tu, A.W. and K. Luo, Acetylation of Smad2 by the co-activator p300 regulates activin and transforming growth factor beta response. J Biol Chem, 2007. 282(29): p. 21187-96.

158. Inoue, Y., et al., Smad3 is acetylated by p300/CBP to regulate its transactivation activity. Oncogene, 2007. 26(4): p. 500-8.

159. Gronroos, E., et al., Control of Smad7 stability by competition between acetylation and ubiquitination. Mol Cell, 2002. 10(3): p. 483-93.

160. Naar, A.M., B.D. Lemon, and R. Tjian, Transcriptional coactivator complexes. Annu Rev Biochem, 2001. 70: p. 475-501.

161. Kato, Y., et al., A component of the ARC/Mediator complex required for TGF beta/Nodal signalling. Nature, 2002. 418(6898): p. 641-6.

162. Akiyoshi, S., et al., c-Ski acts as a transcriptional co-repressor in transforming growth factor-β signaling through interaction with Smads. J. Biol. Chem., 1999. 274: p. 35269- 35277.

163. Luo, K., et al., The Ski oncoprotein interacts with the Smad proteins to repress TGFß signaling. Genes Dev., 1999. 13: p. 2196-2206.

164. Sun, Y., et al., Interaction of the Ski oncoprotein with Smad3 regulates TGF-β signaling. Mol. Cell, 1999. 4: p. 499-509.

165. Wang, W., et al., Ski represses bone morphogenic protein signaling in Xenopus and mammalian cells. Proc. Natl. Acad. Sci. USA, 2000. 97: p. 14394-14399.

166. Ueki, N. and M.J. Hayman, Signal-dependent N-CoR requirement for repression by the Ski oncoprotein. J Biol Chem, 2003. 278(27): p. 24858-64.

167. Wu, K., et al., DACH1 inhibits transforming growth factor-beta signaling through binding Smad4. J Biol Chem, 2003. 278(51): p. 51673-84.

168. Izutsu, K., et al., The corepressor CtBP interacts with Evi-1 to repress transforming

152

growth factor beta signaling. Blood, 2001. 97(9): p. 2815-22.

169. Lin, X., et al., Smad6 recruits transcription corepressor CtBP to repress bone morphogenetic protein-induced transcription. Mol Cell Biol, 2003. 23(24): p. 9081-93.

170. Wotton, D., et al., A Smad transcriptional corepressor. Cell, 1999. 97(1): p. 29-39.

171. Wotton, D., et al., Multiple modes of repression by the Smad transcriptional corepressor TGIF. J Biol Chem, 1999. 274(52): p. 37105-10.

172. Wotton, D., et al., The Smad transcriptional corepressor TGIF recruits mSin3. Cell Growth Differ, 2001. 12(9): p. 457-63.

173. Seo, S.R., et al., The novel E3 ubiquitin ligase Tiul1 associates with TGIF to target Smad2 for degradation. Embo J, 2004. 23(19): p. 3780-92.

174. Denissova, N.G. and F. Liu, Repression of endogenous Smad7 by Ski. J Biol Chem, 2004. 279(27): p. 28143-8.

175. Wu, J.W., et al., Structural mechanism of Smad4 recognition by the nuclear oncoprotein Ski: insights on Ski-mediated repression of TGF-beta signaling. Cell, 2002. 111(3): p. 357-67.

176. Nishihara, A., et al., E1A inhibits Transforming Growth Factor-β signaling through binding to Smad proteins. J. Biol. Chem., 1999. 274: p. 28716-28723.

177. Kim, R.H., et al., A novel smad nuclear interacting protein, SNIP1, suppresses p300- dependent TGF-beta signal transduction. Genes Dev., 2000. 14: p. 1605-1616.

178. Lee, D.K., et al., Human T-cell lymphotropic virus type 1 tax inhibits transforming growth factor-beta signaling by blocking the association of Smad proteins with Smad- binding element. J Biol Chem, 2002. 277(37): p. 33766-75.

179. Higashi, K., et al., Interferon-gamma interferes with transforming growth factor-beta signaling through direct interaction of YB-1 with Smad3. J Biol Chem, 2003. 278(44): p. 43470-9.

180. Ghosh, A.K., S. Bhattacharyya, and J. Varga, The tumor suppressor p53 abrogates Smad- dependent collagen gene induction in mesenchymal cells. J Biol Chem, 2004. 279(46): p. 47455-63.

181. Masuda, S., et al., Notch1 oncoprotein antagonizes TGF-beta/Smad-mediated cell growth suppression via sequestration of coactivator p300. Cancer Sci, 2005. 96(5): p. 274-82.

182. Kurisaki, K., et al., Nuclear factor YY1 inhibits transforming growth factor beta- and bone morphogenetic protein-induced cell differentiation. Mol Cell Biol, 2003. 23(13): p. 4494-510.

183. Chipuk, J.E., et al., The androgen receptor represses transforming growth factor-beta

153

signaling through interaction with Smad3. J Biol Chem, 2002. 277(2): p. 1240-8.

184. Lee, D.K., et al., The human papilloma virus E7 oncoprotein inhibits transforming growth factor-beta signaling by blocking binding of the Smad complex to its target sequence. J Biol Chem, 2002. 277(41): p. 38557-64.

185. Saller, E. and M. Bienz, Direct competition between Brinker and Drosophila Mad in Dpp target gene transcription. EMBO Rep, 2001. 2(4): p. 298-305.

186. Sakuma, R., et al., Inhibition of Nodal signalling by Lefty mediated through interaction with common receptors and efficient diffusion. Genes Cells, 2002. 7(4): p. 401-12.

187. Bisgrove, B.W., J.J. Essner, and H.J. Yost, Regulation of midline development by antagonism of lefty and nodal signaling. Development, 1999. 126(14): p. 3253-62.

188. Thisse, C. and B. Thisse, Antivin, a novel and divergent member of the TGFbeta superfamily, negatively regulates mesoderm induction. Development, 1999. 126(2): p. 229-40.

189. Thisse, B., C.V. Wright, and C. Thisse, Activin- and Nodal-related factors control antero-posterior patterning of the zebrafish embryo. Nature, 2000. 403(6768): p. 425-8.

190. Meno, C., et al., Mouse Lefty2 and zebrafish antivin are feedback inhibitors of nodal signaling during vertebrate gastrulation. Mol Cell, 1999. 4(3): p. 287-98.

191. Agathon, A., B. Thisse, and C. Thisse, Morpholino knock-down of antivin1 and antivin2 upregulates nodal signaling. Genesis, 2001. 30(3): p. 178-82.

192. Branford, W.W. and H.J. Yost, Lefty-dependent inhibition of Nodal- and Wnt-responsive organizer gene expression is essential for normal gastrulation. Curr Biol, 2002. 12(24): p. 2136-41.

193. Feldman, B., et al., Lefty antagonism of Squint is essential for normal gastrulation. Curr Biol, 2002. 12(24): p. 2129-35.

194. Chen, Y. and A.F. Schier, Lefty proteins are long-range inhibitors of squint-mediated nodal signaling. Curr Biol, 2002. 12(24): p. 2124-8.

195. Cha, Y.R., S. Takahashi, and C.V. Wright, Cooperative non-cell and cell autonomous regulation of Nodal gene expression and signaling by Lefty/Antivin and Brachyury in Xenopus. Dev Biol, 2006. 290(2): p. 246-64.

196. Meno, C., et al., Diffusion of nodal signaling activity in the absence of the feedback inhibitor Lefty2. Dev Cell, 2001. 1(1): p. 127-38.

197. Chen, C. and M.M. Shen, Two modes by which Lefty proteins inhibit nodal signaling. Curr Biol, 2004. 14(7): p. 618-24.

198. Cheng, S.K., et al., Lefty blocks a subset of TGFbeta signals by antagonizing EGF-CFC

154

coreceptors. PLoS Biol, 2004. 2(2): p. E30.

199. Tanegashima, K., et al., Xantivin suppresses the activity of EGF-CFC genes to regulate nodal signaling. Int J Dev Biol, 2004. 48(4): p. 275-83.

200. Marques, S., et al., The activity of the Nodal antagonist Cerl-2 in the mouse node is required for correct L/R body axis. Genes Dev, 2004. 18(19): p. 2342-7.

201. Bell, E., et al., Cell fate specification and competence by Coco, a maternal BMP, TGFbeta and Wnt inhibitor. Development, 2003. 130(7): p. 1381-9.

202. Perea-Gomez, A., et al., Nodal antagonists in the anterior visceral endoderm prevent the formation of multiple primitive streaks. Dev Cell, 2002. 3(5): p. 745-56.

203. Bouwmeester, T., et al., Cerberus is a head-inducing secreted factor expressed in the anterior endoderm of Spemann's organizer. Nature, 1996. 382: p. 595-601.

204. Harms, P.W. and C. Chang, Tomoregulin-1 (TMEFF1) inhibits nodal signaling through direct binding to the nodal coreceptor Cripto. Genes Dev, 2003. 17(21): p. 2624-9.

205. Katoh, M., Identification and characterization of human DAPPER1 and DAPPER2 genes in silico. Int J Oncol, 2003. 22(4): p. 907-13.

206. Zhang, L., et al., Zebrafish Dpr2 inhibits mesoderm induction by promoting degradation of nodal receptors. Science, 2004. 306(5693): p. 114-7.

207. Su, Y., et al., The evolutionally conserved activity of Dapper2 in antagonizing TGF-beta signaling. Faseb J, 2007. 21(3): p. 682-90.

208. Haffner, C., et al., Nicalin and its binding partner Nomo are novel Nodal signaling antagonists. Embo J, 2004. 23(15): p. 3041-50.

209. Krakowski, A.R., et al., Cytoplasmic SnoN in normal tissues and nonmalignant cells antagonizes TGF-beta signaling by sequestration of the Smad proteins. Proc Natl Acad Sci U S A, 2005. 102(35): p. 12437-42.

210. Dong, C., et al., Microtubule binding to Smads may regulate TGF beta activity. Mol Cell, 2000. 5(1): p. 27-34.

211. Lin, F., et al., MAN1, an integral protein of the inner nuclear membrane, binds Smad2 and Smad3 and antagonizes transforming growth factor-beta signaling. Hum Mol Genet, 2005. 14(3): p. 437-45.

212. Pan, D., et al., The integral inner nuclear membrane protein MAN1 physically interacts with the R-Smad proteins to repress signaling by the transforming growth factor-{beta} superfamily of cytokines. J Biol Chem, 2005. 280(16): p. 15992-6001.

213. Lin, X., et al., PPM1A functions as a Smad phosphatase to terminate TGFbeta signaling. Cell, 2006. 125(5): p. 915-28.

155

214. Itoh, S. and P. ten Dijke, Negative regulation of TGF-beta receptor/Smad signal transduction. Curr Opin Cell Biol, 2007. 19(2): p. 176-84.

215. Izzi, L. and L. Attisano, Ubiquitin-dependent regulation of TGFbeta signaling in cancer. Neoplasia, 2006. 8(8): p. 677-88.

216. Pickart, C.M., Mechanisms underlying ubiquitination. Annu Rev Biochem, 2001. 70: p. 503-33.

217. Zhu, H., et al., A SMAD ubiquitin ligase targets the BMP pathway and affects embryonic pattern formation. Nature, 1999. 400: p. 687-693.

218. Lin, X., M. Liang, and X.-H. Feng, Smurf2 is a ubiquitin E3 ligase mediating proteasome-dependent degradation of Smad2 in Transforming Growth Factor-β signaling. J. Biol. Chem., 2000. 275: p. 36818-36822.

219. Zhang, Y., et al., Regulation of Smad degradation and activity by Smurf2, an E3 ubiquitin ligase. Proc. Natl. Acad. Sci. USA, 2001. 98: p. 974-979.

220. Kuratomi, G., et al., NEDD4-2 (neural precursor cell expressed, developmentally downregulated 4-2) negatively regulates TGF-beta (transforming growth factor-beta) signalling by inducing ubiquitin-mediated degradation of Smad2 and TGF-beta type I receptor. Biochem J, 2005. 386(Pt 3): p. 461-70.

221. Komuro, A., et al., Negative regulation of transforming growth factor-beta (TGF-beta) signaling by WW domain-containing protein 1 (WWP1). Oncogene, 2004. 23(41): p. 6914-23.

222. Fukuchi, M., et al., Ligand-dependent degradation of Smad3 by a ubiquitin ligase complex of Roc1 and associated proteins. Mol. Biol. Cell, 2001. 12: p. 1430-1443.

223. Moren, A., et al., Degradation of the tumor suppressor Smad4 by WW and HECT domain ubiquitin ligases. J Biol Chem, 2005. 280(23): p. 22115-23.

224. Wan, M., et al., Jab1 antagonizes TGF-beta signaling by inducing Smad4 degradation. EMBO Rep., 2002. 3: p. 171-176.

225. Wan, M., et al., Smad4 protein stability is regulated by ubiquitin ligase SCF beta-TrCP1. J Biol Chem, 2004. 279(15): p. 14484-7.

226. Dupont, S., et al., Germ-layer specification and control of cell growth by Ectodermin, a Smad4 ubiquitin ligase. Cell, 2005. 121(1): p. 87-99.

227. Moren, A., et al., Differential ubiquitination defines the functional status of the tumor suppressor smad4. J Biol Chem, 2003. 278(35): p. 33571-82.

228. Bonni, S., et al., TGF-β induces assembly of a Smad2–Smurf2 ubiquitin ligase complex that targets SnoN for degradation. Nat. Cell Biol., 2001. 3: p. 587-595.

156

229. Stroschein, S.L., et al., Smad3 recruits the anaphase promoting complex for ubiquitination and degradation of SnoN. Genes Dev., 2001. 15: p. 2822-2836.

230. Wan, Y., X. Liu, and M.W. Kirschner, The anaphase-promoting complex mediates TGF- β signaling by targeting SnoN for destruction. Mol. Cell, 2001. 8: p. 1027-1039.

231. Kavsak, P., et al., Smad7 binds to Smurf2 to form an E3 ubiquitin ligase that targets TGFß receptor for degradation. Mol. Cell, 2000. 6: p. 1365-1375.

232. Ebisawa, T., et al., Smurf1 interacts with transforming growth factor-beta type I receptor through Smad7 and induces receptor degradation. J. Biol. Chem., 2001. 276: p. 12477- 12480.

233. Lallemand, F., et al., AIP4 restricts transforming growth factor-beta signaling through a ubiquitination-independent mechanism. J Biol Chem, 2005. 280(30): p. 27645-53.

234. Yamaguchi, T., et al., FKBP12 functions as an adaptor of the Smad7-Smurf1 complex on activin type I receptor. J Mol Endocrinol, 2006. 36(3): p. 569-79.

235. Di Guglielmo, G.M., et al., Distinct endocytic pathways regulate TGF-beta receptor signalling and turnover. Nat Cell Biol, 2003. 5(5): p. 410-21.

236. Ehrlich, M., A. Shmuely, and Y.I. Henis, A single internalization signal from the di- leucine family is critical for constitutive endocytosis of the type II TGF-beta receptor. J Cell Sci, 2001. 114(Pt 9): p. 1777-86.

237. Hayes, S., A. Chawla, and S. Corvera, TGF beta receptor internalization into EEA1- enriched early endosomes: role in signaling to Smad2. J Cell Biol, 2002. 158(7): p. 1239- 49.

238. Penheiter, S.G., et al., Internalization-dependent and -independent requirements for transforming growth factor beta receptor signaling via the Smad pathway. Mol Cell Biol, 2002. 22(13): p. 4750-9.

239. Lin, H.K., S. Bergmann, and P.P. Pandolfi, Cytoplasmic PML function in TGF-beta signalling. Nature, 2004. 431(7005): p. 205-11.

240. Bai, Y., et al., Itch E3 ligase-mediated regulation of TGF-beta signaling by modulating smad2 phosphorylation. Mol Cell, 2004. 15(5): p. 825-31.

241. Niederlander, C., et al., Arkadia enhances nodal-related signalling to induce mesendoderm. Nature, 2001. 410(6830): p. 830-4.

242. Episkopou, V., et al., Induction of the mammalian node requires Arkadia function in the extraembryonic lineages. Nature, 2001. 410(6830): p. 825-30.

243. Koinuma, D., et al., Arkadia amplifies TGF-beta superfamily signalling through degradation of Smad7. Embo J, 2003. 22(24): p. 6458-70.

157

244. Liu, W., et al., Axin is a scaffold protein in TGF-beta signaling that promotes degradation of Smad7 by Arkadia. Embo J, 2006. 25(8): p. 1646-58.

245. Mavrakis, K.J., et al., Arkadia enhances Nodal/TGF-beta signaling by coupling phospho- Smad2/3 activity and turnover. PLoS Biol, 2007. 5(3): p. e67.

246. Johnson, E.S., Protein modification by SUMO. Annu Rev Biochem, 2004. 73: p. 355-82.

247. Bohren, K.M., et al., A M55V polymorphism in a novel SUMO gene (SUMO-4) differentially activates heat shock transcription factors and is associated with susceptibility to type I diabetes mellitus. J Biol Chem, 2004. 279(26): p. 27233-8.

248. Sharrocks, A.D., PIAS proteins and transcriptional regulation--more than just SUMO E3 ligases? Genes Dev, 2006. 20(7): p. 754-8.

249. Pichler, A., et al., The nucleoporin RanBP2 has SUMO1 E3 ligase activity. Cell, 2002. 108(1): p. 109-20.

250. Pichler, A., et al., The RanBP2 SUMO E3 ligase is neither HECT- nor RING-type. Nat Struct Mol Biol, 2004. 11(10): p. 984-91.

251. Kagey, M.H., T.A. Melhuish, and D. Wotton, The polycomb protein Pc2 is a SUMO E3. Cell, 2003. 113(1): p. 127-37.

252. Li, B., et al., Polycomb protein Cbx4 promotes SUMO modification of de novo DNA methyltransferase Dnmt3a. Biochem J, 2007. 405(2): p. 369-78.

253. Hammer, E., R. Heilbronn, and S. Weger, The E3 ligase Topors induces the accumulation of polysumoylated forms of DNA topoisomerase I in vitro and in vivo. FEBS Lett, 2007.

254. Pungaliya, P., et al., TOPORS functions as a SUMO-1 E3 ligase for chromatin-modifying proteins. J Proteome Res, 2007. 6(10): p. 3918-23.

255. Weger, S., E. Hammer, and R. Heilbronn, Topors acts as a SUMO-1 E3 ligase for p53 in vitro and in vivo. FEBS Lett, 2005. 579(22): p. 5007-12.

256. Weger, S., E. Hammer, and M. Engstler, The DNA topoisomerase I binding protein topors as a novel cellular target for SUMO-1 modification: characterization of domains necessary for subcellular localization and sumolation. Exp Cell Res, 2003. 290(1): p. 13- 27.

257. Gill, G., Something about SUMO inhibits transcription. Curr. Opin. Genet. Dev., 2005. 15(5): p. 536-41.

258. Zhao, J., Sumoylation regulates diverse biological processes. Cell Mol Life Sci, 2007.

259. Desterro, J.M., M.S. Rodriguez, and R.T. Hay, SUMO-1 modification of IkappaBalpha inhibits NF-kappaB activation. Mol Cell, 1998. 2(2): p. 233-9.

158

260. Verger, A., J. Perdomo, and M. Crossley, Modification with SUMO. A role in transcriptional regulation. EMBO Rep, 2003. 4(2): p. 137-42.

261. Yang, S.H. and A.D. Sharrocks, SUMO promotes HDAC-mediated transcriptional repression. Mol. Cell, 2004. 13(4): p. 611-7.

262. Girdwood, D., et al., P300 transcriptional repression is mediated by SUMO modification. Mol. Cell, 2003. 11(4): p. 1043-54.

263. David, G., M.A. Neptune, and R.A. DePinho, SUMO-1 modification of histone deacetylase 1 (HDAC1) modulates its biological activities. J Biol Chem, 2002. 277(26): p. 23658-63.

264. Kirsh, O., et al., The SUMO E3 ligase RanBP2 promotes modification of the HDAC4 deacetylase. Embo J, 2002. 21(11): p. 2682-91.

265. Chou, C.C., et al., SUMO modification regulates the DNA-binding activity of GCMa. J. Biol. Chem., 2007.

266. Hong, Y., et al., Regulation of heat shock transcription factor 1 by stress-induced SUMO- 1 modification. J Biol Chem, 2001. 276(43): p. 40263-7.

267. Shyu, Y.C., et al., Sumoylation of p45/NF-E2: nuclear positioning and transcriptional activation of the mammalian beta-like globin gene locus. Mol Cell Biol, 2005. 25(23): p. 10365-78.

268. Wei, F., H.R. Scholer, and M.L. Atchison, Sumoylation of Oct4 Enhances Its Stability, DNA Binding, and Transactivation. J. Biol. Chem., 2007. 282(29): p. 21551-21560.

269. Carbia-Nagashima, A., et al., RSUME, a small RWD-containing protein, enhances SUMO conjugation and stabilizes HIF-1alpha during hypoxia. Cell, 2007. 131(2): p. 309-23.

270. Zheng, G. and Y.C. Yang, ZNF76, a novel transcriptional repressor targeting TATA- binding protein, is modulated by sumoylation. J Biol Chem, 2004. 279(41): p. 42410-21.

271. Lin, X., et al., SUMO-1/Ubc9 promotes nuclear accumulation and metabolic stability of tumor suppressor Smad4. J Biol Chem, 2003. 278(33): p. 31043-8.

272. Lee, P.S., et al., Sumoylation of Smad4, the common Smad mediator of transforming growth factor-beta family signaling. J Biol Chem, 2003. 278(30): p. 27853-63.

273. Lin, X., et al., Activation of transforming growth factor-beta signaling by SUMO-1 modification of tumor suppressor Smad4/DPC4. J Biol Chem, 2003. 278(21): p. 18714-9.

274. Liang, M., et al., Regulation of Smad4 sumoylation and transforming growth factor-beta signaling by protein inhibitor of activated STAT1. J Biol Chem, 2004. 279(22): p. 22857- 65.

159

275. Long, J., et al., Repression of Smad4 transcriptional activity by SUMO modification. Biochem J, 2004. 379(Pt 1): p. 23-9.

276. Ohshima, T. and K. Shimotohno, Transforming growth factor-beta-mediated signaling via the p38 MAP kinase pathway activates Smad-dependent transcription through SUMO-1 modification of Smad4. J Biol Chem, 2003. 278(51): p. 50833-42.

277. Imoto, S., et al., Regulation of Transforming Growth Factor-{beta} Signaling by Protein Inhibitor of Activated STAT, PIASy through Smad3. J Biol Chem, 2003. 278(36): p. 34253-34258.

278. Blumberg, B., et al., Organizer-specific homeobox genes in Xenopus laevis embryos. Science, 1991. 253(5016): p. 194-6.

279. De Robertis, E.M., Goosecoid and Gastrulation, in Gastrulation: From Cells to Embryo, C.D. Stern, Editor. 2004, Cold Spring Harbor Laboratory Press: Cold Spring Harbor, N.Y. p. 581-589.

280. Broun, M., S. Sokol, and H.R. Bode, Cngsc, a homologue of goosecoid, participates in the patterning of the head, and is expressed in the organizer region of Hydra. Development, 1999. 126(23): p. 5245-54.

281. Goriely, A., et al., A functional homologue of goosecoid in Drosophila. Development, 1996. 122(5): p. 1641-50.

282. Blum, M., et al., Gastrulation in the mouse: the role of the homeobox gene goosecoid. Cell, 1992. 69(7): p. 1097-106.

283. Schulte-Merker, S., et al., Expression of zebrafish goosecoid and no tail gene products in wild-type and mutant no tail embryos. Development, 1994. 120(4): p. 843-52.

284. Stachel, S.E., D.J. Grunwald, and P.Z. Myers, Lithium perturbation and goosecoid expression identify a dorsal specification pathway in the pregastrula zebrafish. Development, 1993. 117(4): p. 1261-74.

285. Izpisua-Belmonte, J.C., et al., The homeobox gene goosecoid and the origin of organizer cells in the early chick blastoderm. Cell, 1993. 74(4): p. 645-59.

286. Blum, M., et al., Molecular cloning of the human homeobox gene goosecoid (GSC) and mapping of the gene to human chromosome 14q32.1. Genomics, 1994. 21(2): p. 388-93.

287. Funke, B., et al., Characterization and mutation analysis of goosecoid-like (GSCL), a homeodomain-containing gene that maps to the critical region for VCFS/DGS on 22q11. Genomics, 1997. 46(3): p. 364-72.

288. Gottlieb, S., et al., The DiGeorge syndrome minimal critical region contains a goosecoid- like (GSCL) homeobox gene that is expressed early in human development. Am J Hum Genet, 1997. 60(5): p. 1194-201.

160

289. Cho, K.W.Y., et al., Molecular nature of Spemann's organizer: the role of the Xenopus homeobox gene goosecoid. Cell, 1991. 67: p. 1111-1120.

290. Yao, J. and D.S. Kessler, Goosecoid promotes head organizer activity by direct repression of Xwnt8 in Spemann's organizer. Development, 2001. 128(15): p. 2975-87.

291. Niehrs, C., H. Steinbeisser, and E.M. De Robertis, Mesodermal patterning by a gradient of the vertebrate homeobox gene goosecoid. Science, 1994. 263(5148): p. 817-20.

292. Sander, V., B. Reversade, and E.M. De Robertis, The opposing homeobox genes Goosecoid and Vent1/2 self-regulate Xenopus patterning. EMBO J., 2007. 26(12): p. 2955-65.

293. Steinbeisser, H., et al., The role of gsc and BMP-4 in dorsal-ventral patterning of the marginal zone in Xenopus: a loss-of-function study using antisense RNA. EMBO J., 1995. 14(21): p. 5230-43.

294. Ferreiro, B., et al., Antimorphic goosecoids. Development, 1998. 125(8): p. 1347-59.

295. Niehrs, C., et al., The homeobox gene goosecoid controls cell migration in Xenopus embryos. Cell, 1993. 72(4): p. 491-503.

296. Belo, J.A., et al., Cerberus-like is a secreted factor with neutralizing activity expressed in the anterior primitive endoderm of the mouse gastrula. Mech. Dev., 1997. 68: p. 45-57.

297. Belo, J.A., et al., The prechordal midline of the chondrocranium is defective in Goosecoid-1 mouse mutants. Mech Dev, 1998. 72(1-2): p. 15-25.

298. Filosa, S., et al., Goosecoid and HNF-3beta genetically interact to regulate neural tube patterning during mouse embryogenesis. Development, 1997. 124(14): p. 2843-54.

299. Yamada, G., et al., Targeted mutation of the murine goosecoid gene results in craniofacial defects and neonatal death. Development, 1995. 121(9): p. 2917-22.

300. Rivera-Perez, J.A., et al., Goosecoid is not an essential component of the mouse gastrula organizer but is required for craniofacial and rib development. Development, 1995. 121(9): p. 3005-12.

301. Jimenez, G., C.P. Verrijzer, and D. Ish-Horowicz, A conserved motif in goosecoid mediates groucho-dependent repression in Drosophila embryos. Mol Cell Biol, 1999. 19(3): p. 2080-7.

302. Mailhos, C., et al., Drosophila Goosecoid requires a conserved heptapeptide for repression of paired-class homeoprotein activators. Development, 1998. 125(5): p. 937- 47.

303. Treisman, J., et al., A single amino acid can determine the DNA binding specificity of homeodomain proteins. Cell, 1989. 59(3): p. 553-62.

161

304. Wilson, D., et al., Cooperative dimerization of paired class homeo domains on DNA. Genes Dev., 1993. 7(11): p. 2120-34.

305. Danilov, V., et al., Negative autoregulation of the organizer-specific homeobox gene goosecoid. J. Biol. Chem., 1998. 273: p. 627-635.

306. Latinkic, B.V., et al., The Xenopus Brachyury promoter is activated by FGF and low concentrations of activin and suppressed by high concentrations of activin and by paired- type homeodomain proteins. Genes Dev, 1997. 11(23): p. 3265-76.

307. Chen, G. and A.J. Courey, Groucho/TLE family proteins and transcriptional repression. Gene, 2000. 249(1-2): p. 1-16.

308. Yu, X., et al., Inhibition of androgen receptor-mediated transcription by amino-terminal enhancer of split. Mol Cell Biol, 2001. 21(14): p. 4614-25.

309. Chen, G., et al., A functional interaction between the histone deacetylase Rpd3 and the corepressor groucho in Drosophila development. Genes Dev, 1999. 13(17): p. 2218-30.

310. Sekiya, T. and K.S. Zaret, Repression by Groucho/TLE/Grg proteins: genomic site recruitment generates compacted chromatin in vitro and impairs activator binding in vivo. Mol Cell, 2007. 28(2): p. 291-303.

311. Laurent, M.N., et al., The Xenopus homeobox gene Twin mediates Wnt induction of Goosecoid in establishment of Spemann's organizer. Development, 1997. 124: p. 4905- 4916.

312. Hartwell, K.A., et al., The Spemann organizer gene, Goosecoid, promotes tumor metastasis. Proc. Natl. Acad. Sci. U.S.A., 2006. 103(50): p. 18969-74.

313. Rosa, F.M., Mix.1, a homeobox mRNA inducible by mesoderm inducers, is expressed mostly in the presumptive endodermal cells of Xenopus embryos. Cell, 1989. 57(6): p. 965-74.

314. Vize, P.D., DNA sequences mediating the transcriptional response of the Mix.2 homeobox gene to mesoderm induction. Dev Biol, 1996. 177(1): p. 226-31.

315. Henry, G.L. and D.A. Melton, Mixer, a homeobox gene required for endoderm development. Science, 1998. 281(5373): p. 91-6.

316. Tada, M., et al., Bix1, a direct target of Xenopus T-box genes, causes formation of ventral mesoderm and endoderm. Development, 1998. 125(20): p. 3997-4006.

317. Ecochard, V., et al., A novel Xenopus mix-like gene milk involved in the control of the endomesodermal fates. Development, 1998. 125(14): p. 2577-85.

318. Mead, P.E., et al., Cloning of Mix-related homeodomain proteins using fast retrieval of gel shift activities, (FROGS), a technique for the isolation of DNA-binding proteins. Proc Natl Acad Sci U S A, 1998. 95(19): p. 11251-6.

162

319. Casey, E.S., et al., Bix4 is activated directly by VegT and mediates endoderm formation in Xenopus development. Development, 1999. 126(19): p. 4193-200.

320. Alexander, J. and D.Y. Stainier, A molecular pathway leading to endoderm formation in zebrafish. Curr Biol, 1999. 9(20): p. 1147-57.

321. Kikuchi, Y., et al., The zebrafish bonnie and clyde gene encodes a Mix family homeodomain protein that regulates the generation of endodermal precursors. Genes Dev, 2000. 14(10): p. 1279-89.

322. Poulain, M. and T. Lepage, Mezzo, a paired-like homeobox protein is an immediate target of Nodal signalling and regulates endoderm specification in zebrafish. Development, 2002. 129(21): p. 4901-14.

323. Hirata, T., et al., Novel mix-family homeobox genes in zebrafish and their differential regulation. Biochem Biophys Res Commun, 2000. 271(3): p. 603-9.

324. Pearce, J.J.H. and M.J. Evans, Mml, a mouse Mix-like gene expressed in the primitive streak. Mech. Dev., 1999. 87: p. 189-192.

325. Peale, F.V., Jr., L. Sugden, and M. Bothwell, Characterization of CMIX, a chicken homeobox gene related to the Xenopus gene mix.1. Mech Dev, 1998. 75(1-2): p. 167-70.

326. Stein, S., T. Roeser, and M. Kessel, CMIX, a paired-type homeobox gene expressed before and during formation of the avian primitive streak. Mech Dev, 1998. 75(1-2): p. 163-5.

327. Robb, L., et al., Cloning, expression analysis, and chromosomal localization of murine and human homologues of a Xenopus mix gene. Dev Dyn, 2000. 219(4): p. 497-504.

328. Guo, W., et al., A human Mix-like homeobox gene MIXL shows functional similarity to Xenopus Mix.1. Blood, 2002. 100(1): p. 89-95.

329. Mead, P.E., et al., BMP-4-responsive regulation of dorsal-ventral patterning by the homeobox protein Mix.1. Nature, 1996. 382(6589): p. 357-60.

330. Alexander, J., et al., casanova plays an early and essential role in endoderm formation in zebrafish. Dev Biol, 1999. 215(2): p. 343-57.

331. Trinh, L.A., D. Meyer, and D.Y. Stainier, The Mix family homeodomain gene bonnie and clyde functions with other components of the Nodal signaling pathway to regulate neural patterning in zebrafish. Development, 2003. 130(20): p. 4989-98.

332. Hart, A.H., et al., Mixl1 is required for axial mesendoderm morphogenesis and patterning in the murine embryo. Development, 2002. 129(15): p. 3597-608.

333. Trindade, M., et al., Regulation of apoptosis in theXenopus embryo by Bix3. Development, 2003. 130(19): p. 4611-22.

163

334. Kofron, M., C. Wylie, and J. Heasman, The role of Mixer in patterning the early Xenopus embryo. Development, 2004. 131(10): p. 2431-41.

335. Lemaire, P., et al., A role for the vegetally expressed Xenopus gene Mix.1 in endoderm formation and in the restriction of mesoderm to the marginal zone. Development, 1998. 125(13): p. 2371-80.

336. Latinkic, B.V. and J.C. Smith, Goosecoid and mix.1 repress Brachyury expression and are required for head formation in Xenopus. Development, 1999. 126(8): p. 1769-79.

337. Xanthos, J.B., et al., Maternal VegT is the initiator of a molecular network specifying endoderm in Xenopus laevis. Development, 2001. 128(2): p. 167-80.

338. Yeo, C.-Y., X. Chen, and M. Whitman, The role of FAST-1 and Smads in transcriptional regulation by activin during early Xenopus embryogenesis. J. Biol. Chem., 1999. 274: p. 26584-26590.

339. Randall, R.A., et al., Different Smad2 partners bind a common hydrophobic pocket in Smad2 via a defined proline-rich motif. Embo J, 2002. 21(1-2): p. 145-56.

340. Doherty, J.R., et al., Determination of the minimal domains of Mix.3/Mixer required for endoderm development. Mech Dev, 2006. 123(1): p. 56-66.

341. Loose, M. and R. Patient, A genetic regulatory network for Xenopus mesendoderm formation. Dev Biol, 2004. 271(2): p. 467-78.

342. Tam, P.P. and R.R. Behringer, Mouse gastrulation: the formation of a mammalian body plan. Mech. Dev., 1997. 68(1-2): p. 3-25.

343. Nagy, A., Manipulating the mouse embryo : a laboratory manual. 3rd ed. 2003, Cold Spring Harbor, N.Y.: Cold Spring Harbor Laboratory Press. x, 764 p.

344. Spector, D.L., R.D. Goldman, and L.A. Leinwand, Culture and in vitro differentiation of mouse embryonic stem cells, in Cells A Laboratory Manual, D.L. Spector, R.D. Goldman, and L.A. Leinwand, Editors. 1998, Cold Spring Harbor Laboratory Press: New York. p. 8.1-8.22.

345. Guo, Y., et al., The embryonic stem cell transcription factors Oct-4 and FoxD3 interact to regulate endodermal-specific promoter expression. Proc Natl Acad Sci U S A, 2002. 99(6): p. 3663-7.

346. Foucher, I., et al., Hoxa5 overexpression correlates with IGFBP1 upregulation and postnatal dwarfism: evidence for an interaction between Hoxa5 and Forkhead box transcription factors. Development, 2002. 129(17): p. 4065-74.

347. Rausa, F.M., Y. Tan, and R.H. Costa, Association between hepatocyte nuclear factor 6 (HNF-6) and FoxA2 DNA binding domains stimulates FoxA2 transcriptional activity but inhibits HNF-6 DNA binding. Mol Cell Biol, 2003. 23(2): p. 437-49.

164

348. Keller, G., Embryonic stem cell differentiation: emergence of a new era in biology and medicine. Genes Dev, 2005. 19(10): p. 1129-55.

349. Willey, S., et al., Acceleration of mesoderm development and expansion of hematopoietic progenitors in differentiating ES cells by the mouse Mix-like homeodomain transcription factor. Blood, 2006. 107(8): p. 3122-30.

350. Mossman, A.K., et al., Mixl1 and oct4 proteins are transiently co-expressed in differentiating mouse and human embryonic stem cells. Stem Cells Dev, 2005. 14(6): p. 656-63.

351. Ng, E.S., et al., The primitive streak gene Mixl1 is required for efficient haematopoiesis and BMP4-induced ventral mesoderm patterning in differentiating ES cells. Development, 2005. 132(5): p. 873-84.

352. Kim, J.J., et al., Regulation of insulin-like growth factor binding protein-1 promoter activity by FKHR and HOXA10 in primate endometrial cells. Biol Reprod, 2003. 68(1): p. 24-30.

353. Boettger, T., et al., The avian organizer. Int. J. Dev. Biol., 2001. 45(1): p. 281-7.

354. Zhu, L., et al., Goosecoid regulates the neural inducing strength of the mouse node. Dev. Biol., 1999. 216(1): p. 276-81.

355. Izzi, L., et al., Foxh1 recruits Gsc to negatively regulate Mixl1 expression during early mouse development. EMBO J., 2007. 26(13): p. 3132-43.

356. Yukita, A., et al., XSUMO-1 is required for normal mesoderm induction and axis elongation during early Xenopus development. Dev Dyn, 2007. 236(10): p. 2757-66.

357. Barrios-Rodiles, M., et al., High-throughput mapping of a dynamic signaling network in mammalian cells. Science, 2005. 307(5715): p. 1621-5.

358. Muller, S., et al., SUMO, ubiquitin's mysterious cousin. Nat. Rev. Mol. Cell Biol., 2001. 2(3): p. 202-10.

359. Gill, G., SUMO and ubiquitin in the nucleus: different functions, similar mechanisms? Genes Dev, 2004. 18(17): p. 2046-59.

360. Schmidt, D. and S. Muller, PIAS/SUMO: new partners in transcriptional regulation. Cell Mol. Life Sci., 2003. 60(12): p. 2561-74.

361. Hoege, C., et al., RAD6-dependent DNA repair is linked to modification of PCNA by ubiquitin and SUMO. Nature, 2002. 419(6903): p. 135-41.

362. Rui, H.L., et al., SUMO-1 modification of the C-terminal KVEKVD of Axin is required for JNK activation but has no effect on Wnt signaling. J. Biol. Chem., 2002. 277(45): p. 42981-6.

165

363. Yamamoto, H., et al., Sumoylation is involved in beta-catenin-dependent activation of Tcf-4. EMBO J., 2003. 22(9): p. 2047-59.

364. Anckar, J., et al., Inhibition of DNA binding by differential sumoylation of heat shock factors. Mol. Cell. Biol., 2006. 26(3): p. 955-64.

365. Hay, R.T., SUMO: a history of modification. Mol. Cell, 2005. 18(1): p. 1-12.

366. Gill, G., Post-translational modification by the small ubiquitin-related modifier SUMO has big effects on transcription factor activity. Curr. Opin. Genet. Dev., 2003. 13(2): p. 108-13.

367. Tucker, S.C. and R. Wisdom, Site-specific heterodimerization by paired class homeodomain proteins mediates selective transcriptional responses. J. Biol. Chem., 1999. 274(45): p. 32325-32.

368. Thiery, J.P. and J.P. Sleeman, Complex networks orchestrate epithelial-mesenchymal transitions. Nat. Rev. Mol. Cell Biol., 2006. 7(2): p. 131-42.

369. Wacker, S., et al., Patterns and control of cell motility in the Xenopus gastrula. Development, 1998. 125(10): p. 1931-42.

370. Li, S., J.L. Guan, and S. Chien, Biochemistry and biomechanics of cell motility. Annu. Rev. Biomed. Eng., 2005. 7: p. 105-50.

371. Sousa-Nunes, R., et al., Characterizing embryonic gene expression patterns in the mouse using nonredundant sequence-based selection. Genome Res, 2003. 13(12): p. 2609-20.

372. Yoshikawa, T., et al., High-throughput screen for genes predominantly expressed in the ICM of mouse blastocysts by whole mount in situ hybridization. Gene Expr Patterns, 2006. 6(2): p. 213-24.

373. Neidhardt, L., et al., Large-scale screen for genes controlling mammalian embryogenesis, using high-throughput gene expression analysis in mouse embryos. Mech Dev, 2000. 98(1-2): p. 77-94.

374. Visel, A., et al., Regulatory pathway analysis by high-throughput in situ hybridization. PLoS Genet, 2007. 3(10): p. 1867-83.

375. Yaklichkin, S., et al., FoxD3 and Grg4 physically interact to repress transcription and induce mesoderm in Xenopus. J Biol Chem, 2007. 282(4): p. 2548-57.

376. Schuff, M., et al., FoxN3 is required for craniofacial and eye development of Xenopus laevis. Dev Dyn, 2007. 236(1): p. 226-39.

377. Li, B. and M.I. Greene, FOXP3 actively represses transcription by recruiting the HAT/HDAC complex. Cell Cycle, 2007. 6(12): p. 1432-6.

378. Freddie, C.T., et al., Functional interactions between the Forkhead transcription factor

166

FOXK1 and the MADS-box protein SRF. Nucleic Acids Res, 2007. 35(15): p. 5203-12.

379. Foucher, I., et al., Joint regulation of the MAP1B promoter by HNF3beta/Foxa2 and Engrailed is the result of a highly conserved mechanism for direct interaction of homeoproteins and Fox transcription factors. Development, 2003. 130(9): p. 1867-76.

380. Gauthier, B.R., et al., Hepatic nuclear factor-3 (HNF-3 or Foxa2) regulates glucagon gene transcription by binding to the G1 and G2 promoter elements. Mol Endocrinol, 2002. 16(1): p. 170-83.

381. Galili, N., et al., A region of mouse chromosome 16 is syntenic to the DiGeorge, velocardiofacial syndrome minimal critical region. Genome Res, 1997. 7(1): p. 17-26.

382. Galili, N., et al., Gscl, a gene within the minimal DiGeorge critical region, is expressed in primordial germ cells and the developing pons. Dev Dyn, 1998. 212(1): p. 86-93.

383. Marcal, N., et al., Antagonistic effects of Grg6 and Groucho/TLE on the transcription repression activity of brain factor 1/FoxG1 and cortical neuron differentiation. Mol Cell Biol, 2005. 25(24): p. 10916-29.

384. Yao, J., E. Lai, and S. Stifani, The winged-helix protein brain factor 1 interacts with groucho and hes proteins to repress transcription. Mol Cell Biol, 2001. 21(6): p. 1962- 72.

385. Yaklichkin, S., et al., Prevalence of the EH1 Groucho interaction motif in the metazoan Fox family of transcriptional regulators. BMC Genomics, 2007. 8: p. 201.

386. Steiner, A.B., Fox family transcription factors in the regulation of Nodal signaling and germ layer formation, in Cell and Molecular Biology Graduate Group. 2006, University of Pennsylvania: Philadelphia. p. 216.