Genomic Integration of Wnt/Β-Catenin and BMP/Smad1 Signaling Coordinates Digestive System Development

Home , HOXA3, HOXB1, HOXC6, HOXC9, HOXD10, MSX1

Genomic integration of Wnt/β-catenin and BMP/Smad1 signaling coordinates digestive system development

A dissertation submitted to the Division of Graduate Studies and Research of the University of Cincinnati in partial fulfillment of the requirements for the degree of

Doctor of Philosophy in the Graduate Program of Molecular and Developmental Biology in the College of Medicine

2017

Mariana L. Stevens B.S., Universidade Federal do Rio de Janeiro 2009

M.S., Universidade Federal do Rio de Janeiro 2010

Dissertation Committee:

Aaron Zorn, Ph.D. (Chair) James Wells, Ph.D. Katherine Yutzey, Ph.D. Brian Gebelein, Ph.D. Dan Buchholz, Ph.D. Joo-Seop Park, Ph.D. Abstract

The digestive system forms during early fetal development, through intricate events coordinated by growth factors. The embryonic endoderm and mesoderm layers coordinate the formation of the gut progenitors that will give rise to foregut-derived organs such as the liver, pancreas and stomach, and hindgut derived intestine.

Signaling pathways like Bone Morphogenetic Protein (BMP) and Wnt are reiteratively used during development. Their downstream targets are variable upon time and target tissue, and the specificity of the outcome is achieved through crosstalk with other factors. Although we know what growth factors act when and where during development, it is still largely unknown how multiple signaling pathways integrate their signals at the genomic level for correct embryonic patterning. In this study we aimed to unveil the transcriptional program of foregut and hindgut progenitors and how

BMP/Smad1 and Wnt/β-catenin coordinate their signals to pattern these progenitors.

We used genome wide analysis to determine the foregut and hindgut programs in early Xenopus laevis embryos. We found that these transcriptional programs are highly conserved with mammals. Upon signaling manipulation, we found that a subset of foregut endoderm and mesoderm genes depend on BMP signaling. Moreover, we found that BMP/Smad1 directly activates ventral genes while repressing dorsal genes.

Meanwhile we showed that the Wnt signaling pathway acts on anterior-posterior patterning, directly activating the hindgut while repressing the foregut programs.

Detailed analysis identified several factors, other than Tcf, to be involved with β-catenin direct repression of foregut genes. We further reported that hundreds of key foregut and

ii hindgut genes presented Smad1 and β-catenin integration on the same CRMs.

Functional analysis showed some of these cis-regulatory modules (CRMs) to be functional and responsive to BMP and Wnt manipulations. Overall, this part of our study revealed a new layer of complexity to BMP/Smad1 and Wnt/β-catenin pathways showing their transcriptional integration at the genomic level to coordinate hundreds of target genes during foregut and hindgut patterning.

Once we defined the mechanism of BMP/Smad1 and Wnt/β-catenin regulation of foregut and hindgut, we investigated the temporal role dynamic Wnt/β-catenin plays during these progenitors’ specification. We used genome wide transcriptome and transcription factor binding analysis to explore the temporal binding dynamics of β- catenin associated with Wnt-regulated genes. We first found that ~88% of all β-catenin associated genes are not expressed or Wnt-regulated at the same developmental stage. Moreover, we showed that among Wnt-regulated genes, β-catenin is recruited by two different mechanisms depending on whether it is activating or repressing the target.

Hindgut-enriched Wnt-activated genes are primed by β-catenin since early stages of development, mostly at the same CRMs. Meanwhile foregut-enriched Wnt-repressed genes rarely presented β-catenin at early stages and most of the CRMs seem to be stage-specific. Taken together our findings deepen our understanding of combinatorial signaling crosstalk in developmental processes and provide new insights in how multiple signaling pathways can utilize the same machinery for different biological outputs.

iii

iv Acknowledgments

Graduate School is a challenging journey, but being surrounded by so many great people definitely made it easier. First, I would like to thank my mentor Aaron Zorn for being so patient and dedicated to my training. I have grown so much in these last 6 years and I could not have done it without him. From past and current lab mates, there are a lot of people that have contributed to this work. However, my favorite bioinformatician Praneet Chaturvedi, our beloved lab manager Scott Rankin, and my amazing undergraduate Melissa Macdonald were hands down the most essential pieces in this puzzle.

I would also like to thank the Xenbase folks (especially Malcolm Fisher), the

Wells lab, the Kenny lab, and the Kofron/Cha labs for the great discussions during our weekly meetings. You have all made an impact in my career and for that I will be forever thankful. Since no great science is done inside closed doors, I am very grateful for our collaborations. I would like to thank Jin Cho and Professor Ken Cho for opening their lab doors for me at the University of California, Irvine for a delightful week of training. Also, I would like to thank the whole Barski’s lab for their endless advice and partnership.

Lastly, I have to thank my support system that has kept me going during these intense grad school years. From my family members abroad; Mom, Dad, my sister

Carol, and my lovely grandparents. To the closest ones at home, my amazing husband

Bobby and his awesome kids Emma and Parker. You guys are the definition of the word

“family” to me.

v Table of contents

Abstract ...... ii

Acknowledgments ...... v

Table of contents ...... vi

List of tables and illustrations ...... viii

List of abbreviations ...... xi

Chapter 1: Introduction ...... 1

Overview ...... 2 Bone Morphogenetic Protein Signaling Pathway ...... 4 The Canonical Wnt Signaling Pathway ...... 7 Transcriptional Regulation and Epigenetics ...... 12 Embryonic Axial Patterning ...... 17 Progenitors Patterning ...... 18 Crosstalk between BMP and Wnt signaling pathways ...... 22 Central Hypothesis ...... 28 References ...... 29

Chapter 2: Genomic integration of Wnt/β-catenin and BMP/Smad1 signaling coordinates foregut and hindgut transcriptional program ...... 38

Abstract ...... 39 Introduction ...... 40 Results ...... 42 Discussion ...... 68 Materials and methods ...... 72 References ...... 75 Supplementary figures ...... 81 List of supplemental tables ...... 91 Supplementary materials and methods ...... 94 References ...... 101

Chapter 3: Genome-wide analysis of β-catenin occupancy at early stages of development ...... 103

Abstract ...... 104 Introduction ...... 105 Materials and Methods ...... 109 Results ...... 113 Discussion ...... 134

vi References ...... 143

Chapter 4: Discussion ...... 148

Major Findings ...... 149 Potential Implications ...... 150 Experimental Limitations ...... 155 Future Directions ...... 158 References ...... 162

vii List of tables and illustrations

Chapter 1: Introduction

Figure 1: The BMP transcriptional mechanism of activation...... 5

Figure 2: The canonical Wnt signaling pathway...... 11

Figure 3: Epigenetic characteristics of active and poised enhancers...... 16

Figure 4: Endoderm specification and axial patterning during frog development...... 18

Figure 5: Endoderm patterning in Xenopus laevis...... 22

Figure 6: BMP and Wnt signaling interaction...... 27

Chapter 2: Genomic integration of Wnt/β-catenin and BMP/Smad1 signaling coordinates foregut and hindgut transcriptional program

Fig. 1: Transcriptional program of FG and HG progenitors correlates with differential BMP and Wnt signaling...... 47

Fig. 2: BMP signaling coordinates D-V patterning...... 50

Fig. 3: Smad1 chromatin binding to BMP-regulated genes...... 53

Fig. 4: Wnt signaling promotes HG transcriptional program and represses FG transcriptional program...... 56

Fig. 5: β-catenin chromatin binding to Wnt-regulated genes...... 61

Fig. 6: Smad1 and β-catenin converge on common CRMs...... 66

Fig. 7: BMP/Smad1 and Wnt/β-catenin converge on the same FG and HG CRMs...... 67

Fig. S1: Transcriptional program of FG and HG progenitors in vivo...... 81

Fig. S2: RNA-seq of DMSO and DMH1 treated embryos identified BMP-regulated genes...... 83

Fig. S3: Smad1 ChIP-seq of embryos stage NF20...... 85

Fig. S4: RNA-seq of control, Tg(hsp70:dkk1) and BIO treated embryos...... 87

viii Fig. S5: b-catenin ChIP-seq of embryos stage NF20...... 88

Fig. S6: BMP/Smad1 and Wnt/β-catenin crosstalk...... 89

Fig.S7: Smad1 and β-catenin syntenic peaks in Xenopus laevis and Homo sapiens. ... 90

Table S1: Transcriptional program of FG and HG progenitors...... 91

Table S2: Tables show the number of transcripts overlapping in the following pairwise differential expression analyses...... 91

Table S3: FG and HG transcriptome conservation among vertebrates...... 91

Table S4: BMP regulated genes from RNA-seq analysis...... 91

Table S5: Smad1 and p300 peaks of whole embryos stage NF20 ...... 91

Table S6: Wnt regulated genes from RNA-seq analysis...... 92

Table S7: β-catenin and p300 peaks of whole embryos and FG+HG tissues stage NF20 ...... 92

Table S8: Genes associated with Smad1 and β-catenin peaks...... 92

Table S9: Syntenic Smad1 and β-catenin peaks between Xenopus and human ChIP- seq data...... 92

Table S10: Summary of FG- and HG-enriched genes indicating BMP and Wnt and association of Smad1 or β-catenin peaks ...... 92

Chapter 3: Genome-wide analysis of β-catenin occupancy at early stages of development

Fig. 1: Summary diagram of β-catenin activity levels of early developing Xenopus laevis embryos...... 107

Fig. 2: During gastrulation, foregut progenitors transition to a low Wnt/β-catenin environment...... 115

Fig. 3: Wnt/β-catenin has dynamic roles in FG and HG progenitors patterning...... 118

Fig. 4: Wnt-regulated transcriptome and β-catenin binding on early stage embryos. .. 121

Fig. 5: Wnt-regulated genes and β-catenin occupancy at different stages of development...... 123

ix Fig. 6: Non-productive β-catenin binding at stage NF20 binding is not residual from stage NF11.5 regulation...... 126

Fig. 7: Stage NF20 direct Wnt-target genes are associated with β-catenin at earlier stages of development...... 130

Fig. 8: Hindgut marker CRMs are primed by β-catenin prior to regulation...... 133

x List of abbreviations

A-P - Anterior-posterior

BMP - Bone Morhogenetic Protein

CBP - CREB-binding protein

CRMs - Cis-Regulatory Modules

D-V - Dorso-ventral

Endo - Endoderm

FG - Foregut

FGF - Fibroblast growth factor

HAT - Histone acetyltransferase

HDAC - histone deacetylase

HG - Hindgut

HMG - High mobility group

HMT - histone methyltransferase

MBT - Mid-blastula-transition

Meso - Mesoderm

PSC - Pluripotent stem cell

xi Chapter 1: Introduction

Mariana L. Stevens

Division of Developmental Biology, Perinatal Institute, Cincinnati Children's Research

Foundation and Department of Pediatrics College of Medicine, University of Cincinnati,

Cincinnati, Ohio 45229, USA.

1 Overview

Digestive system formation starts during the first trimester of gestation, and is coordinated by a limited number of growth factors. The same growth factors involved in digestive organ formation are reiteratively used in other compartments of the embryo.

Therefore, one of the major questions in biology that we will address in this study is how a restricted number of signaling factors can have so many different roles during embryogenesis, and more specifically how they coordinate their signals on the genome.

Development of the digestive system relies on growth factor crosstalk between endodermal and mesodermal cell populations. Animal model studies have shown that growth factors like Wnt, Bone Morphogenetic Protein (BMP), Fibroblast Growth Factor

(FGF) and Retinoic acid pattern the naïve endoderm and lateral plate mesoderm along the anterior and posterior axis into foregut and hindgut progenitors. Hours later, foregut progenitors are subdivided into organs like esophagus, lungs, stomach, liver and pancreas, while hindgut progenitors subdivide into the different regions of the intestine.

The same growth factors are reused at multiple steps of digestive system formation with a wide variety of roles. Recently, stem cell differentiation protocols have applied insights from basic developmental biology to mimic the stepwise process of cell lineage specification in their cultures. Although advances have been recently made to the extent of what factors regulate progenitor patterning, the precise molecular mechanisms behind cell fate specification are largely unknown. Moreover, it is yet to be resolved how these same factors can play multiple roles during embryonic development. Therefore, a comprehensive understanding of how these patterning growth factors regulate

2 transcriptional programs seems necessary for the advancement of stem cell derived tissue protocols.

Here we will introduce the BMP and Wnt signaling pathways and what is known about how they transcriptionally regulate target genes, focusing on how modifications on the genome can impact transcription. We will then review the embryonic steps prior to foregut and hindgut patterning, focusing on the many different roles BMP and Wnt signaling pathways play during digestive system formation. We will finish the chapter with the possible ways BMP and Wnt can coordinate their signals to regulate their targets. In chapter 2 we will discuss the mechanisms by which BMP/Smad1 and Wnt/β- catenin coordinate gut tube patterning during early somitogenesis and we will report our findings of how Smad1 and β-catenin integrate on the genome to regulate hundreds of target genes. Moreover, in chapter 3 we will discuss preliminary data on how Wnt/β- catenin might prime foregut and hindgut genes for later transcriptional activation or repression, which provides new insights on the different regulatory outcomes Wnt can produce. In chapter 4, we will thoroughly discuss how the presented findings provide new and exciting data that impacts the way we see how signaling pathways are mechanistically integrated and how this can benefit disease and stem cell differentiation studies.

3 Bone Morphogenetic Protein Signaling Pathway

BMP signaling

The TGF-β superfamily plays diverse roles in development, homeostasis and disease and is generally activated by ligand (TGF-β, Activin, Nodal or BMPs) binding to transmembrane serine/threonine kinase receptors, type I and type II. In the case of the

BMP signaling pathway, a BMP ligand binding to the receptor type II induces the recruitment of receptor type I. Receptor type II phosphorylates serine and threonine residues within the intracellular glycine-serine-rich domain of the receptor type I subunit

(Wrana et al., 1994). Activated type I receptor phosphorylates Smad1/5/8, which binds to its partner, Smad4, and the Smad complex travels to the nucleus, binding to defined elements on the DNA (Fig. 1). Smad proteins contain an MH1 domain that possesses the DNA-binding and nuclear import functions. Phosphorylated Smad1/5/8 recognizes the motif CAGAC, SMAD-binding-element (SBE), as well as GC-rich sequences

(Morikawa et al., 2011). The MH2 domain contains sites for receptor phosphorylation as well as for interaction with a variety of nuclear co-factors, including co-activators and co- repressors. The linker region between the MH1 and MH2 domains contains binding sites for kinases and a nuclear export signal (Massague, 2012). Smads are constantly shuttling between the cytoplasm and the nuclei. The receptor-mediated phosphorylation reduces cytoplasm- and increases nuclei-affinity for Smad; once Smad is dephosphorylated, it returns to the cytoplasm, conferring a prompt sensing mechanism to the pathway activation (Massague et al., 2005).

Figure 1: The BMP transcriptional mechanism of activation. Upon ligand binding to receptor type 2, a downstream cascade starts with phosphorylation of the receptor type 1. Activated receptor type 1 phosphorylates Smad1/5/8, promoting the recruitment of co-factor Smad4 and forming a complex. This complex enters the nucleus and binds to BMP responsive DNA elements, triggering transcription of target genes. Known inhibitors and co-factors are shown. Modified from (Gamez et al., 2013).

5 BMP inhibitors

Intracellular inhibitors can tightly regulate BMP activity; with some examples including BMP and activin membrane bound inhibitor (BAMBI), inhibitory Smads (I-

Smads: Smad6 and Smad7), Smad ubiquitination regulatory factor-1 (Smurf), Ski and

Tob. BAMBI, for example, is a dominant-negative receptor sequestering functional BMP receptors and inhibiting the downstream pathway (Onichtchouk et al., 1999). In the cytoplasm, however, I-Smads can bind to the phosphorylated and activated receptor type I, inhibiting its interaction with Smad1/5/8 (Imamura et al., 1997). At another level of regulation, Smurfs are able to interact with Smad1/5 through the linker region, triggering ubiquitination followed by proteosomal degradation (Zhu et al., 1999).

However, inside the nuclei, the proteins Ski and Tob can act as co-repressors, inhibiting

BMP/Smad1 response and negatively regulating the pathway (Fig. 1) (von Bubnoff and

Cho, 2001). Moreover, nuclear co-factors perform a pivotal role in the BMP/Smad pathway, not only increasing DNA-binding affinity but also being crucial for transcriptional regulation, either activation or repression.

BMP/Smad1 regulated transcription

The Smad1-Smad4 complex doesn’t have high DNA-binding affinity (Kd≈ 1 ×

10−7 M), and it is believed that once inside the nucleus, Smads must interact with other

DNA-binding cofactors. The interaction with other co-factors promotes not only better affinity to DNA, but also selectivity for target genes (Massague et al., 2005).

BMP/Smad1 signaling’s effect – activation or repression - and target specificity comes mostly from other binding partners. Smad1-Smad4 complex can directly interact with

6 co-activators like histone acetyltransferases (HATs) p300 and CBP (Pouponnot et al.,

1998) to loosen chromatin and promote gene transcription, or with DNA-binding co- factors like the zinc finger protein OAZ, in the case of Xenopus gene ventx2.1 (Hata et al., 2000; Karaulanov et al., 2004). Although about two thirds of the binding events initiate transcriptional activation, the rest initiates transcriptional repression (Massague et al., 2005). BMP-induced repression originates from interaction with repressive co- factors, as has been described with Nkx3-2. Smad1-Smad4 and Nkx3-2 were demonstrated to form a complex in vitro, recruiting histone deacetylase 1 (HDAC1) and repressing luciferase activity of a BMP responsive reporter (Kim and Lassar, 2003).

Another binding partner of Smad1 is the zinc-finger protein Schnurri, considered to be a pivotal player in the BMP activation/repression switch. Schnurri and Smad1 regulate brinker (brk) expression and gradient formation through binding to a silencer element

(Blitz and Cho, 2009; Marty et al., 2000; Muller et al., 2003).

The Canonical Wnt Signaling Pathway

Canonical Wnt signaling and pathway components

The Wnt/β-catenin signaling pathway has many roles in embryonic development and tissue homeostasis, being involved in mechanisms ranging from cell proliferation to lineage determination. The Wnt/β-catenin pathway is activated by a well-conserved mechanism, starting with the binding of Wnt ligands, secreted lipid-modified proteins, to seven-pass transmembrane receptors of the Frizzled (Fzd) family and single-pass transmembrane co-receptor lipoprotein receptor-related protein 5/6 (LRP5/6). In total,

19 Wnt ligands have been discovered, which allows for the complexity of outcomes 7 executed by this pathway (MacDonald et al., 2009). Wnt proteins undergo two types of lipid modifications prior to their secretion, which is a crucial step, not just for secretion, but also for binding to the transmembrane receptors Fzd and LRP5/6 (Fig. 2)

(Komekado et al., 2007). Wnt binding to receptors Fzd and LRP5/6 can be antagonized by the Wnt inhibitors Secreted Frizzled-related proteins (Sfrps) and Dickkopf (Dkk)

(Bovolenta et al., 2008; Glinka et al., 1998). Sfrps are known to inhibit the Wnt pathway by binding to both Wnt ligands and to Fzd proteins, impairing the β-catenin/Fzd interaction (Bovolenta et al., 2008). Dkk1, on the other hand, binds to LRP5/6 inhibiting its interaction with Fzd and promoting its internalization and further degradation (Fig. 2)

(Mao et al., 2002).

The β-catenin destruction complex

The Wnt pathway can be divided between canonical and non-canonical, the former involving the downstream factor β-catenin while the latter involves either the

Calcium or the JNK downstream signals. In the canonical Wnt pathway, upon Wnt ligand binding, a complex of Fzd and the intracellular protein Dishevelled (Dvl) is formed, recruiting Axin and the β-catenin degradation complex, which initiates phosphorylation of LRP5/6. LRP5/6 phosphorylation itself recruits even more degradation complexes to the membrane, creating a positive feedback loop (Tamai et al., 2004; Zeng et al., 2008a). This complex is formed by previously mentioned, Axin, as well as Adenomatous Polyposis Coli (APC), glycogen synthase kinase 3β (GSK3β) and

Casein Kinase 1 alpha (CKIα). Axin is the core of this complex interacting with all its components and also with β-catenin. CKIα and GSK3β phosphorylate and

8 polyubiquinate a well-conserved consensus site in the N terminus of β-catenin, which is recognized by the E3 ubiquitin ligase subunit, β-Trcp, and targeted for proteasomal degradation (Aberle et al., 1997; Liu et al., 2002). However, upon pathway activation and ligand binding, the degradation complex is inactived, allowing the non- phosphorylated cytoplasmic β-catenin to enter the nucleus and activate target genes

(Fig. 2).

Wnt/β-catenin transcriptional mechanism

Although β-catenin does not have a DNA-binding domain, its central Arm-repeats interact with DNA binding factors lymphoid enhancer binding factor 1 or transcription factor 3 (Tcf3/Lef1) (Clevers and Nusse, 2012). Vertebrates normally encode four Tcf genes (Tcf1, Lef1, Tcf3 and Tcf4), which bind to the well-conserved Tcf consensus motif

AGATCAAAGG through their high-mobility-group (HMG) DNA-binding domain (van de

Wetering et al., 1997). Upon Wnt activation, free cytosolic β-catenin translocates to the nucleus, binding to Tcf and triggering Tcf phosphorylation (Hikasa et al., 2010). It appears that this phosphorylation switches Tcf from a repressor to an activator, since

Wnt induction displaces repressive proteins Groucho/TLE and CBP, and recruits transcriptional co-activators like the histone acetyltransferases (HATs) CBP/p300 to Wnt target genes (Fig. 2) (Hoppler and Kavanagh, 2007). Like Smads, β-catenin is constantly shuttling between the cytoplasm and nuclei, conferring another level of regulation to the Wnt pathway (Clevers and Nusse, 2012). Interestingly, in the cytoplasm, the same β-catenin is part of the adherens junction, together with E-cadherin and α-catenin, conferring even more complexity to the Wnt pathway, since the 9 cytoplasmic β-catenin is a target for both E-cadherin and APC protein-protein interactions (Hulsken et al., 1994). The toggle mechanism between being a transcriptional activator or a cytoskeletal component remains to be investigated. Once in the nuclei, β-catenin can bind to co-activators like BCLP9, Pygo and CBP/p300, through its amino-terminal or carboxyl terminal portions (Fig. 2) (Mosimann et al., 2009). β- catenin binding and transcriptional activation is accompanied by chromatin remodeling, such as histone acetylation and histone H3K4 tri-methylation, both indicative of active transcription, as well as removal of histone H3K27 tri-methylation, a marker for transcriptional repression (Neijts et al., 2016; Parker et al., 2008; Sierra et al., 2006).

Although most of β-catenin’s effect is through transcriptional activation, there are a few examples of β-catenin being associated with recruitment of repressors (Jamora et al.,

2003; Olson et al., 2006; Theisen et al., 2007).

Figure 2: The canonical Wnt signaling pathway. In the absence of Wnt ligand binding, the β-catenin destruction complex, composed of Axin, CK1α, GSK3β and APC, phosphorylate and polyubiquinate β-catenin for degradation. Upon Wnt binding to receptors LRP5/6 and Fzd, the destruction complex is inhibited and non-phosphorylated β-catenin enters the nucleus binding to DNA-binding factors Tcf/Lef, recruiting co-activators and inducing downstream target genes. Wnt repressors Sfrp and Dkk1 are shown in the figure. Modified from (Barker and Clevers, 2006).

11 Transcriptional Regulation and Epigenetics

Chromatin accessibility

Most of our understanding of how Wnt and BMP signaling control cell identity in vivo has come from an analysis of a few key genes, but how these impact genome scale regulation is unknown. Moreover, it is yet to be elucidated how signaling pathways modify the chromatin landscape to achieve transcriptional regulation. Since the ‘70s, researchers have shown that DNA is wrapped in nucleosomal structures formed by octamers of histones, which keep chromatin in a very compact and almost inaccessible form (Kornberg and Thomas, 1974). Modifications of the histone tails are directly associated with chromatin remodeling, for either activation or repression.

Acetylation of the N-terminal region of histones H3 and H4, for example, are one of the marks of transcriptional activation, because addition of these acetyl groups promotes chromatin opening, making it accessible to transcription factors. This modification is regulated by histone acetyltransferases (HATs) and counteracted by histone deacetylases (HDACs) (Bogdanovic et al., 2012). Smad1 and β-catenin co- factors CBP/p300, the most well studied HATs, are implicated with deposition of acetyl groups on histone 3 lysine 27 around enhancer sites. Therefore, this factor is currently used as a marker for active enhancers genome-wide (Holmqvist and Mannervik, 2013).

On the other hand, Tcf co-factors Groucho and TLE are co-repressors that recruit

HDACs, removing acetyl groups from nearby histone tails and inducing chromatin compaction (Chen and Courey, 2000). Meanwhile, histone methylation is generally correlated with transcriptional repression, with the exception of tri-methylation of histone

3 lysine 4, which is a marker of active promoters (Fig. 3). The placement and

12 displacement of such modifications is performed by histone methyltransfereases

(HMTs) and demethyltransferases, respectively. The polycomb repressive group is responsible for addition of a methyl group, done by the H3K27 HMT subunit, which with the help of the chromatin compaction subunit, induces chromatin closing and inaccessibility to transcription activators (Fig. 3) (Bogdanovic et al., 2012). Histone modifications can either create a permissive environment for transcription factors like

Smad1 and TCF/β-catenin to access DNA or a nucleosome dense domain obstructing their binding.

Several papers have demonstrated that the chromatin state is correlated with cell lineage choices. Chromatin modifications can make certain cis-regulatory-modules more or less available for transcription factor binding and transcriptional regulation, or in other words, can poise these regions for rapid activation (Fig. 3) (Kraus and Grapin-

Botton, 2012). Our most current knowledge of foregut and hindgut transcriptional regulation has been made through advances in the in vitro differentiation of human stem cells. Recent genomic analyses in human stem cell cultures have reported the epigenetic profile of digestive progenitors. These studies revealed that many key progenitor genes have poised bivalent enhancers, which are marked by the presence of

H3K4me1 and H3K27ac. As these cells commit to more defined lineages, like liver, pancreas or intestine, lineage-specific transcription factors are recruited and the bivalency is resolved. Genes that are necessary for that developmental step are activated, while genes relevant for alternative lineages remain silent (Loh et al., 2014;

Tsankov et al., 2015; Wang et al., 2015).

13 Gene priming

The first factors to bind and promote chromatin opening are known as pioneer factors. They precede lineage specific factors, like Smad1 and β-catenin, which do not possess the ability of direct histone remodeling. Pioneer factors, like Foxa1/2, are able to read the chromatin and bind to unopened regions providing a permissive state for future regulation (Sekiya et al., 2009; Xu et al., 2007). They do so by opening up the chromatin with nucleosome repositioning and providing more accessibility to upcoming factors that will regulate nearby genes (Zaret and Mango, 2016). Once chromatin is made accessible the lineage-specific factors, like Smad1 and β-catenin, are responsible for initiation of transcriptional regulation. Upon binding, lineage-specific factors recruit the general transcription factors, which are part of the basic transcription factor machinery, followed by the RNA polymerase II. RNA pol II is activated by phosphorylation, starting the production of messenger RNA and initiating transcription

(Spitz and Furlong, 2012). Interestingly, transcription factors that bind to distal enhancers can promote chromatin looping to a nearby promoter and initiate the transcription machinery. Sometimes lineage specific factors occupy poised bi-valent enhancers, which contain both active and repressive marks hindering initiation of transcription until further factors are recruited (Fig. 3) (Calo and Wysocka, 2013).

Although much is known about how Smad1 and β-catenin in general can activate transcription of downstream target genes, it is still not clear what are all the targets in different contexts of development, especially during digestive system patterning when both BMP and Wnt are reiteratively used throughout development (discussed next).

Moreover, the detailed mechanism of regulation is also unknown; it is not clear if other

14 factors cooperate with Smad1 and β-catenin and whether chromatin-remodeling proteins are recruited. In chapter 2, we will address these target genes and how BMP and Wnt coordinate their signals in the genome to transcriptionally regulate hundreds of key lineage specific genes. Moreover, in chapter 3 we will provide data to support the hypothesis that β-catenin acts to prime downstream Wnt-targets for future transcription.

Figure 3: Epigenetic characteristics of active and poised enhancers. (A) Active enhancers contain both H3K27ac and H3K4me1 marks. Lineage-specific transcription factors recruit HAT p300, RNA pol II and Mediator complex to interact with the gene promoter through DNA looping. The promoter is marked by active marker H3K4me3. (B) Poised enhancers contain H3K27me3 repressive mark, deposited by the polycomb repressive group (PRC2). Transcription factors are bound but transcription is inhibited by repressive marks. The promoter is marked by both active marker H3K4me3 and repressive mark H3K27me3. Modified from (Calo and Wysocka, 2013).

16 Embryonic Axial Patterning

The future digestive and respiratory system’s specification start at early embryogenesis, when distinct cell movements regulated by growth factors like TGFβ,

Wnt/β-catenin, Sox and Fox gene families give rise to the three germ layers: endoderm, mesoderm and ectoderm. The molecular pathways underlying germ layer formation are well-conserved between species (Zorn and Wells, 2009). In the frog marginal dorsal side of the embryo, maternally encoded Wnt/β-catenin and the T-box transcription factor

VegT induce the expression of Nodal-related ligands (nodal1/5/6). This induction creates a gradient of Nodal activity, high in the dorsal and low in the ventral domains

(Fig. 4) (Shen, 2007). Initially, mesendoderm common precursors are formed, but due to the Nodal activity gradient, these bipotent progenitors separate between high nodal- endoderm and low nodal-mesoderm cell populations (Clements et al., 1999; D'Amour et al., 2005; Green and Smith, 1990; Kubo et al., 2004). High nodal levels promote the expression of Mix-like proteins Foxa2, Sox17, Eomesodermin and Gata4-6, all components of the endoderm regulatory network (Sinner et al., 2006). Conversely, low

Nodal in the prospective mesoderm induces the expression of FGF ligands and

Brachyury, which activate the mesoderm gene regulatory network while inhibiting the endoderm identity (Mizoguchi et al., 2006). Nodal and maternal Wnt signals induce the expression of organizer transcription factors gsc, otx2 and lim1/lhx1. These dorsal signals not only instruct the cells to start the process of gastrulation and convergent extension, but also initiate the foregut progenitor’s identity.

Figure 4: Endoderm specification and axial patterning during frog development. VegT and Vg1 initiate a Nodal gradient that specify endoderm and mesoderm layers at early stages of development. High Nodal induces endoderm while low Nodal induces mesoderm. At the beginning of gastrulation, the dorsal domain expresses factors like Chordin, Noggin, Sfrp2 and Dkk1 inhibiting the ventral ligands Bmp4 and zWnt8. The anterior foregut cells migrate from the dorsal signaling center, while the posterior endoderm derives from the ventral cells of the developing embryo. Modified from (De Robertis and Kuroda, 2004).

Progenitors Patterning

Upon germ layer formation, both endoderm and mesoderm are patterned in antero-posterior (A-P) and dorso-ventral (D-V) axes. The initial patterning steps take place during early gastrulation, and are directed by surrounding growth factors. The high nodal and maternal Wnt/β-catenin environment of the organizer signaling center, through a conserved mechanism, induces the expression of the anterior endoderm gene hhex in both frogs and mammals (Fig. 4) (Rankin et al., 2011; Thomas et al.,

1998). Nodal, through transcription factors FoxH1/Smad2, directly activates hhex, while

Wnt/β-catenin induces the homeodomain factors Siamois and Twin, which stimulate hhex transcription (Rankin et al., 2011). As gastrulation proceeds, the first cells to migrate from the dorsal signaling center are the hhex positive (hhex+) anterior 18 endoderm population, which will reach the most anterior domain (presumptive foregut) at early neurula.

Apart from instructive cues, the dorsal signaling center also has antagonistic signals to protect from ventral factors and maintain the dorsal fate (De Robertis and

Kuroda, 2004). Some of these factors include Chordin, Noggin, Sfrp2 and Dkk1, which inhibit the ventral ligands Bmp4 and zygotic Wnt8 (zWnt8) (Fig. 4). This ventral marginal zone, low in Nodal signals and high in Bmp4 and zWnt8, will give rise to the ventral posterior embryonic region, the future hindgut, and through homeobox factors ventx1 and ventx2, inhibit the dorso-anterior development, as has been shown in both fish and frogs (Dosch et al., 1997; Onichtchouk et al., 1996; Ramel et al., 2005; Ramel and Lekven, 2004; Rankin et al., 2011). In general, low levels of BMP are necessary for specification and patterning of the most anterior gut cell population (Tiso et al., 2002;

Wills et al., 2008). Human pluripotent stem cell researchers have applied these findings to pattern definitive endoderm in foregut progenitors, doing so with BMP inhibition

(Green et al., 2011; Li et al., 2011; Willems and Leyns, 2008).

Shortly after gastrulation, the roles of BMP and Wnt in the foregut drastically change (Fig. 5). There is evidence that BMP from the septum transversum mesenchyme promotes the development of several ventral foregut lineages including the liver, lung and thyroid within the endoderm (Kenny et al., 2012; Rossi et al., 2001;

Shin et al., 2007; Wandzioch and Zaret, 2009). Moreover, in the adjacent mesoderm,

BMP is necessary for cardiac fate as well as for erythroid and myeloid progenitors

(Klaus and Birchmeier, 2009; Walmsley et al., 2002). Meanwhile, at post-gastrulation, high levels of Wnt are inhibitory for foregut progenitors (McLin et al., 2007; Zhang et al.,

19 2013). We know from animal models that Wnt signaling must be repressed in the cardiac mesenchyme to initiate cardiogenesis and specification of first and second heart fields (Klaus and Birchmeier, 2009). In fact, the endoderm progenitor cells adjacent to this cardiac mesenchyme express Wnt antagonists, sfrp1 and sfrp5, protecting them from Wnt signals and allowing cardiac and foregut identity maintenance in both frog and fish embryos (Gibb et al., 2013; Li et al., 2008; Stuckenholz et al., 2013). Although BMP and Wnt’s roles in foregut lineages are very tissue- and time-dependent, both are necessary overall for posterior endoderm and mesoderm even after gastrulation

(Willems and Leyns, 2008) (Fig. 5). Mouse embryo foregut explants, when induced for

Wnt, activated the expression of the hindgut markers cdx2 and hoxd9, while repressed foregut markers hhex and sox2 (Sherwood et al., 2011). Moreover, Wnt induction is a key step for the generation of hindgut lineages from stem cells (McLin et al., 2007;

Spence et al., 2011).

Once these progenitor domains are established, foregut cells express hhex, sox2, foxa2, hoxb1, hoxa3 and hoxb4 factors, while hindgut cells express cdx1/2/3, hoxc5, hoxc6, hoxb6, hoxb8, hoxc9, hoxb9, hox/d13 and ventx genes (Zorn and Wells,

2009). Although these cells have an established transcriptional identity, they are still plastic and several experiments have addressed their potential to become another region of the gut upon instructive cues (Grapin-Botton, 2005; Horb and Slack, 2001;

Wells and Melton, 2000). Moreover, gene expression domains do not have a defined boundary and the different combinations of genes will give identity to the tissue along its

A-P and D-V axes. Therefore, it is necessary to have the correct set of genes in each

20 domain in order to correctly pattern the endoderm (De Robertis, 2009; Rankin et al.,

2011).

The gastrulation morphogenetic movements create the future gut cavity, or archenteron, which is formed by endoderm epithelial cells surrounded by mesoderm inducing tissue. After specification, these endodermal progenitors continue receiving

BMP and Wnt signals from the adjacent mesoderm – which play several different roles during endoderm patterning (Zorn and Wells, 2009). For example, hhex+ foregut progenitor’s cells bifurcate between liver and pancreas lineages, and BMP is thought to favor the hepatic fate (Chung et al., 2008; Deutsch et al., 2001). However, later in development, BMP plays a different role in the separation of the common foregut tube between trachea and esophagus, by inhibition of the digestive lineage marker Sox2 and promoting the respiratory fate (Domyan et al., 2011). Dynamic Wnt signaling has also been demonstrated to be necessary for respiratory and hepatic lineages (McLin et al.,

2007; Rankin et al., 2012; So et al., 2013). However, different Wnt levels during development are necessary for correct heart development, being inhibitory shortly after gastrulation, but later necessary to maintain and expand cardiac progenitor cell lineages

(Klaus and Birchmeier, 2009). Overall, these same signaling pathways are reiteratively used during foregut and hindgut progenitor patterning and although the extracellular components required for proper foregut and hindgut patterning seem to be widely studied, the molecular mechanism by which BMP and Wnt downstream effectors,

Smad1 and β-catenin, coordinate the genomic profile of foregut and hindgut genes still remains to be investigated. In chapter 2 we will address this question and show how

21 signaling integration between BMP and Wnt is necessary to correctly establish foregut and hindgut progenitors impacting later organogenesis.

Figure 5: Endoderm patterning in Xenopus laevis. Schematic illustrating endoderm development and fate mapping of foregut (orange) and hindgut (green) progenitor cells. During gastrulation, foregut cells require low levels of BMP and high levels of Wnt, while hindgut cells require high levels of BMP and low levels of Wnt. About 10 hours later, the requirements change such that foregut requires high levels of bmp and low levels of Wnt, while hindgut requires high levels of both BMP and Wnt. Foregut cells will derive several organs, like esophagus, lungs, liver, stomach and pancreas, while hindgut cells will form the different regions of the intestine.

Crosstalk between BMP and Wnt signaling pathways

Synergy during axial patterning stages

Wnt and BMP pathway components are co-expressed in several tissues during embryonic development. There are many events where the pathways are not only co- expressed but also crosstalking. It is believed that one way pathways fine-tune

22 transcriptional regulation is through interaction with other signaling cascades, which amplifies the variety of signals they can generate. These signaling integration events can be of different natures: additive (where they achieve the same goal) or antagonistic

(where they behave in opposing manners); it can happen in different compartments: extracellularly, in the cytoplasm or inside the nucleus, all depending largely on the biological context (Fig. 6) (Itasaki and Hoppler, 2010). Many examples have been described in the context of embryonic development, where tight regulation of both BMP and Wnt are necessary for correct patterning. During axial patterning for example bmp4/7 and zWnt8, expressed in the ventral marginal zone, cooperate to regulate

Xenopus mesodermal fate. In this scenario bmp4/7 acts upstream of wnt8 to induce ventrolateral mesoderm (Hoppler and Moon, 1998). During blood formation, an example of a mesoderm derivative, BMP was shown to be a necessary signal acting through activation of the Wnt ligand Wnt3a (Lengerke et al., 2008; Walmsley et al., 2002).

However, it seems that in most cases BMP and Wnt antagonize each other and this behavior seems to be fundamental to establish tissue boundaries during important cell fate decisions.

BMP and Wnt antagonizing activities

Most of the interactions between these two signaling pathways seem to have contrary outcomes. In the intestinal crypt, stem cells make the decision between differentiation and renewal, and BMP and Wnt have been implicated in this choice. Wnt signals are necessary in the intestinal crypt to maintain the stem cell niche. This can only be achieved with the help of BMP antagonists, also present in the crypts, that keep

23 BMP levels low and allow for Wnt activity (Kosinski et al., 2007). Our group has also shown that during lung development, Wnt activity requires low levels of BMP. BMP repression allows for proper Wnt2/2b expression, which is necessary for lung progenitor specification in Xenopus (Rankin et al., 2012). The mechanisms of crosstalk between these signaling pathways are so dynamic that later in development, in the same cell population, BMP signaling becomes necessary for lung branching and morphogenesis.

In these cells, BMP induces the expression of Wnt inhibitory factor 1 (Wif1), which is a

Wnt antagonist that keeps canonical Wnt levels low, allowing for correct fetal lung development (Xu et al., 2011). On the other hand, Wnt has also been shown to inhibit the BMP pathway on multiple occasions. The embryonic ectoderm can be differentiated into neural or epidermal fates, and it appears that the balance between BMP and Wnt dictates this decision. The Harland group has shown that Wnt activates neural development in ectoderm explants through inhibition of bmp4 expression (Baker et al.,

1999). Some would argue that the specific mechanism of inhibition would involve a role for the Wnt component GSK3β, acting on Smad1 degradation. BMP/Smad1 activity duration was shown to be tightly regulated by GSK3β through the phosphorylation of the

Smad1 linker region, which causes ubiquitination, degradation and subsequent suppression of BMP activity (Eivers et al., 2009; Fuentealba et al., 2007). Conversely,

Smad1 can also inhibit canonical Wnt signaling through association with Dvl in the cytoplasm, resulting in β-catenin degradation by the GSK3β complex (Liu et al., 2006).

Sometimes, in order to maintain developmental territories BMP and Wnt mutually inhibit each other, as has been shown in fly’s leg and eye/antennal discs. Dpp and wg, or BMP

24 and Wnt in flies, repress each other’s expression separating the dorso-ventral territories

(Theisen et al., 1996).

BMP and Wnt integration at cis-regulatory-modules

In some cases, BMP and Wnt pathways can converge to the same cis-regulatory modules (CRMs) to regulate downstream gene transcription. The neural factors emx2 and msx2 were shown to be regulated by mutual binding of Smad1 and Tcf, in a synergistic manner (Hussein et al., 2003; Theil et al., 2002). Interestingly, Smad1 and

Tcf mutual binding was observed in more than a handful of genes. It was recently shown with genome-wide analysis of hematopoietic progenitors that Smad1 and Tcf4 converge to the same CRMs to interact with master regulators of erythroid and myeloid lineages to participate in these cell fate decisions (Trompouki et al., 2011). Moreover, a possible mechanism of recruitment to enhancers could be through the formation of protein complexes between Smad1, β-catenin and TCF. This ternary complex was shown to regulate expression of the genes xtwn, in the frog Spemann’s organizer, as well as myc, in polycystic kidneys (Hu and Rosenblum, 2005; Nishita et al., 2000).

However, there are cases where β-catenin and Smad1 converge to the same cis- regulatory modules to antagonize each other’s signals. The fly visceral mesoderm gene ultrabithorax contains binding motifs for both Mad (Smad1) and pangolin (Tcf), however, it also contains a binding site for the repressor brinker overlapping with Smad1 site.

Brinker is activated by high levels of Wnt and competes with Smad1 for binding to the ultrabithorax promoter, regulating it’s transcription (Itasaki and Hoppler, 2010). Also in flies, Mad (Smad1) can bind to pangolin (Tcf) and compete for Armadillo (β-catenin)

25 binding, interfering with activation of Wnt target genes like distal-less (Zeng et al.,

2008b). It has become more evident that integration of signaling pathways governs multiple developmental processes in very different manners. We will investigate in chapter 2 how these pathways interact in the context of foregut and hindgut progenitor regulation and show that one mechanism of regulation is through binding to the same cis-regulatory modules of key lineage specific genes.

Figure 6: BMP and Wnt signaling interaction. BMP and Wnt signaling pathways integrate their signals through different mechanisms depending on the cellular context. (A) Regulation of each other’s components. (B) Communication outside of the cell. (C) Intracellular crosstalk (either cytoplasmic or nuclear). (D) Integration on the same CRMs. Adopted from (Itasaki and Hoppler, 2010).

27 Central Hypothesis

Although pioneer factor binding, chromatin poising and accessibility account for activation and silencing of a given set of genes, it does not explain how signaling pathways read chromatin instructions and initiate gene regulation. It is still unclear how transcription factors read the chromatin landscape and initiate binding followed by transcriptional regulation. Moreover, we still lack an understanding of the mechanism by which BMP/Smad1 and Wnt/β-catenin regulate foregut and hindgut transcriptional programs and confer identity to the endoderm and adjacent mesenchyme. Knowing the role of BMP and Wnt during development and how much is still unknown on the specific mechanisms these pathways utilize to pattern foregut and hindgut progenitors, we proposed to tackle these unanswered questions. We hypothesized that differential spatial and temporal dynamic BMP/Smad1 and Wnt/β-catenin integrate on common CRMs to regulate the transcriptional program of foregut and hindgut progenitors.

In this study we proposed to use genome-wide RNA-seq and ChIP-seq analysis to identify the different roles BMP and Wnt have on these progenitors. In order to do so, we took advantage of Xenopus laevis embryos due to their large size (allowing for meticulous dissections of foregut and hindgut tissues), external development (for precise temporal analysis), and ease of accessibility (for the demanding genomic assays). In this study we have reported several CRMs that integrate Smad1 and β- catenin signals and are associated with important endoderm and mesenchyme genes involved in patterning (Chapter 2). Moreover, looking at different developmental stages we also identified a pattern for Wnt/β-catenin transcriptional regulation that involves 28 early priming of target genes (Chapter 3). We finalize this study by highlighting the implications of our findings in the context of stem cell differentiation and tissue regeneration, as well as in the general understanding of how BMP/Smad1 and Wnt/β- catenin coordinate transcriptional regulation.

References

Aberle, H., Bauer, A., Stappert, J., Kispert, A. and Kemler, R. (1997). beta-catenin is a target for the ubiquitin-proteasome pathway. EMBO J 16, 3797-3804. Baker, J. C., Beddington, R. S. and Harland, R. M. (1999). Wnt signaling in Xenopus embryos inhibits bmp4 expression and activates neural development. Genes Dev 13, 3149-3159. Barker, N. and Clevers, H. (2006). Mining the Wnt pathway for cancer therapeutics. Nat Rev Drug Discov 5, 997-1014. Blitz, I. L. and Cho, K. W. (2009). Finding partners: how BMPs select their targets. Dev Dyn 238, 1321-1331. Bogdanovic, O., van Heeringen, S. J. and Veenstra, G. J. (2012). The epigenome in early vertebrate development. Genesis 50, 192-206. Bovolenta, P., Esteve, P., Ruiz, J. M., Cisneros, E. and Lopez-Rios, J. (2008). Beyond Wnt inhibition: new functions of secreted Frizzled-related proteins in development and disease. J Cell Sci 121, 737-746. Calo, E. and Wysocka, J. (2013). Modification of enhancer chromatin: what, how, and why? Mol Cell 49, 825-837. Chen, G. and Courey, A. J. (2000). Groucho/TLE family proteins and transcriptional repression. Gene 249, 1-16. Chung, W. S., Shin, C. H. and Stainier, D. Y. (2008). Bmp2 signaling regulates the hepatic versus pancreatic fate decision. Dev Cell 15, 738-748. Clements, D., Friday, R. V. and Woodland, H. R. (1999). Mode of action of VegT in mesoderm and endoderm formation. Development (Cambridge, England) 126, 4903-4911. Clevers, H. and Nusse, R. (2012). Wnt/beta-catenin signaling and disease. Cell 149, 1192-1205.

29 D'Amour, K. A., Agulnick, A. D., Eliazer, S., Kelly, O. G., Kroon, E. and Baetge, E. E. (2005). Efficient differentiation of human embryonic stem cells to definitive endoderm. Nat Biotechnol 23, 1534-1541. De Robertis, E. M. (2009). Spemann's organizer and the self-regulation of embryonic fields. Mech Dev 126, 925-941. De Robertis, E. M. and Kuroda, H. (2004). Dorsal-ventral patterning and neural induction in Xenopus embryos. Annu Rev Cell Dev Biol 20, 285-308. Deutsch, G., Jung, J., Zheng, M., Lora, J. and Zaret, K. S. (2001). A bipotential precursor population for pancreas and liver within the embryonic endoderm. Development (Cambridge, England) 128, 871-881. Domyan, E. T., Ferretti, E., Throckmorton, K., Mishina, Y., Nicolis, S. K. and Sun, X. (2011). Signaling through BMP receptors promotes respiratory identity in the foregut via repression of Sox2. Development (Cambridge, England) 138, 971- 981. Dosch, R., Gawantka, V., Delius, H., Blumenstock, C. and Niehrs, C. (1997). Bmp-4 acts as a morphogen in dorsoventral mesoderm patterning in Xenopus. Development (Cambridge, England) 124, 2325-2334. Eivers, E., Demagny, H. and De Robertis, E. M. (2009). Integration of BMP and Wnt signaling via vertebrate Smad1/5/8 and Drosophila Mad. Cytokine Growth Factor Rev 20, 357-365. Fuentealba, L. C., Eivers, E., Ikeda, A., Hurtado, C., Kuroda, H., Pera, E. M. and De Robertis, E. M. (2007). Integrating patterning signals: Wnt/GSK3 regulates the duration of the BMP/Smad1 signal. Cell 131, 980-993. Gamez, B., Rodriguez-Carballo, E. and Ventura, F. (2013). BMP signaling in telencephalic neural cell specification and maturation. Front Cell Neurosci 7, 87. Gibb, N., Lavery, D. L. and Hoppler, S. (2013). sfrp1 promotes cardiomyocyte differentiation in Xenopus via negative-feedback regulation of Wnt signalling. Development (Cambridge, England) 140, 1537-1549. Glinka, A., Wu, W., Delius, H., Monaghan, A. P., Blumenstock, C. and Niehrs, C. (1998). Dickkopf-1 is a member of a new family of secreted proteins and functions in head induction. Nature 391, 357-362. Grapin-Botton, A. (2005). Antero-posterior patterning of the vertebrate digestive tract: 40 years after Nicole Le Douarin's PhD thesis. Int J Dev Biol 49, 335-347. Green, J. B. and Smith, J. C. (1990). Graded changes in dose of a Xenopus activin A homologue elicit stepwise transitions in embryonic cell fate. Nature 347, 391-394. Green, M. D., Chen, A., Nostro, M. C., d'Souza, S. L., Schaniel, C., Lemischka, I. R., Gouon-Evans, V., Keller, G. and Snoeck, H. W. (2011). Generation of anterior foregut endoderm from human embryonic and induced pluripotent stem cells. Nat Biotechnol 29, 267-272.

30 Hata, A., Seoane, J., Lagna, G., Montalvo, E., Hemmati-Brivanlou, A. and Massague, J. (2000). OAZ uses distinct DNA- and protein-binding zinc fingers in separate BMP-Smad and Olf signaling pathways. Cell 100, 229-240. Hikasa, H., Ezan, J., Itoh, K., Li, X., Klymkowsky, M. W. and Sokol, S. Y. (2010). Regulation of TCF3 by Wnt-dependent phosphorylation during vertebrate axis specification. Dev Cell 19, 521-532. Holmqvist, P. H. and Mannervik, M. (2013). Genomic occupancy of the transcriptional co-activators p300 and CBP. Transcription 4, 18-23. Hoppler, S. and Kavanagh, C. L. (2007). Wnt signalling: variety at the core. J Cell Sci 120, 385-393. Hoppler, S. and Moon, R. T. (1998). BMP-2/-4 and Wnt-8 cooperatively pattern the Xenopus mesoderm. Mech Dev 71, 119-129. Horb, M. E. and Slack, J. M. (2001). Endoderm specification and differentiation in Xenopus embryos. Dev Biol 236, 330-343. Hu, M. C. and Rosenblum, N. D. (2005). Smad1, beta-catenin and Tcf4 associate in a molecular complex with the Myc promoter in dysplastic renal tissue and cooperate to control Myc transcription. Development (Cambridge, England) 132, 215-225. Hulsken, J., Birchmeier, W. and Behrens, J. (1994). E-cadherin and APC compete for the interaction with beta-catenin and the cytoskeleton. J Cell Biol 127, 2061- 2069. Hussein, S. M., Duff, E. K. and Sirard, C. (2003). Smad4 and beta-catenin co- activators functionally interact with lymphoid-enhancing factor to regulate graded expression of Msx2. J Biol Chem 278, 48805-48814. Imamura, T., Takase, M., Nishihara, A., Oeda, E., Hanai, J., Kawabata, M. and Miyazono, K. (1997). Smad6 inhibits signalling by the TGF-beta superfamily. Nature 389, 622-626. Itasaki, N. and Hoppler, S. (2010). Crosstalk between Wnt and bone morphogenic protein signaling: a turbulent relationship. Dev Dyn 239, 16-33. Jamora, C., DasGupta, R., Kocieniewski, P. and Fuchs, E. (2003). Links between signal transduction, transcription and adhesion in epithelial bud development. Nature 422, 317-322. Karaulanov, E., Knochel, W. and Niehrs, C. (2004). Transcriptional regulation of BMP4 synexpression in transgenic Xenopus. EMBO J 23, 844-856. Kenny, A. P., Rankin, S. A., Allbee, A. W., Prewitt, A. R., Zhang, Z., Tabangin, M. E., Shifley, E. T., Louza, M. P. and Zorn, A. M. (2012). Sizzled-tolloid interactions maintain foregut progenitors by regulating fibronectin-dependent BMP signaling. Dev Cell 23, 292-304. Kim, D. W. and Lassar, A. B. (2003). Smad-dependent recruitment of a histone deacetylase/Sin3A complex modulates the bone morphogenetic protein-

31 dependent transcriptional repressor activity of Nkx3.2. Mol Cell Biol 23, 8704- 8717. Klaus, A. and Birchmeier, W. (2009). Developmental signaling in myocardial progenitor cells: a comprehensive view of Bmp- and Wnt/beta-catenin signaling. Pediatr Cardiol 30, 609-616. Komekado, H., Yamamoto, H., Chiba, T. and Kikuchi, A. (2007). Glycosylation and palmitoylation of Wnt-3a are coupled to produce an active form of Wnt-3a. Genes Cells 12, 521-534. Kornberg, R. D. and Thomas, J. O. (1974). Chromatin structure; oligomers of the histones. Science 184, 865-868. Kosinski, C., Li, V. S., Chan, A. S., Zhang, J., Ho, C., Tsui, W. Y., Chan, T. L., Mifflin, R. C., Powell, D. W., Yuen, S. T., et al. (2007). Gene expression patterns of human colon tops and basal crypts and BMP antagonists as intestinal stem cell niche factors. Proc Natl Acad Sci U S A 104, 15418-15423. Kraus, M. R. and Grapin-Botton, A. (2012). Patterning and shaping the endoderm in vivo and in culture. Curr Opin Genet Dev 22, 347-353. Kubo, A., Shinozaki, K., Shannon, J. M., Kouskoff, V., Kennedy, M., Woo, S., Fehling, H. J. and Keller, G. (2004). Development of definitive endoderm from embryonic stem cells in culture. Development (Cambridge, England) 131, 1651- 1662. Lengerke, C., Schmitt, S., Bowman, T. V., Jang, I. H., Maouche-Chretien, L., McKinney-Freeman, S., Davidson, A. J., Hammerschmidt, M., Rentzsch, F., Green, J. B., et al. (2008). BMP and Wnt specify hematopoietic fate by activation of the Cdx-Hox pathway. Cell Stem Cell 2, 72-82. Li, F., He, Z., Li, Y., Liu, P., Chen, F., Wang, M., Zhu, H., Ding, X., Wangensteen, K. J., Hu, Y., et al. (2011). Combined activin A/LiCl/Noggin treatment improves production of mouse embryonic stem cell-derived definitive endoderm cells. J Cell Biochem 112, 1022-1034. Li, Y., Rankin, S. A., Sinner, D., Kenny, A. P., Krieg, P. A. and Zorn, A. M. (2008). Sfrp5 coordinates foregut specification and morphogenesis by antagonizing both canonical and noncanonical Wnt11 signaling. Genes Dev 22, 3050-3063. Liu, C., Li, Y., Semenov, M., Han, C., Baeg, G. H., Tan, Y., Zhang, Z., Lin, X. and He, X. (2002). Control of beta-catenin phosphorylation/degradation by a dual-kinase mechanism. Cell 108, 837-847. Liu, Z., Tang, Y., Qiu, T., Cao, X. and Clemens, T. L. (2006). A dishevelled-1/Smad1 interaction couples WNT and bone morphogenetic protein signaling pathways in uncommitted bone marrow stromal cells. J Biol Chem 281, 17156-17163. Loh, K. M., Ang, L. T., Zhang, J., Kumar, V., Ang, J., Auyeong, J. Q., Lee, K. L., Choo, S. H., Lim, C. Y., Nichane, M., et al. (2014). Efficient endoderm induction from human pluripotent stem cells by logically directing signals controlling lineage bifurcations. Cell Stem Cell 14, 237-252.

32 MacDonald, B. T., Tamai, K. and He, X. (2009). Wnt/beta-catenin signaling: components, mechanisms, and diseases. Dev Cell 17, 9-26. Mao, B., Wu, W., Davidson, G., Marhold, J., Li, M., Mechler, B. M., Delius, H., Hoppe, D., Stannek, P., Walter, C., et al. (2002). Kremen proteins are Dickkopf receptors that regulate Wnt/beta-catenin signalling. Nature 417, 664-667. Marty, T., Muller, B., Basler, K. and Affolter, M. (2000). Schnurri mediates Dpp- dependent repression of brinker transcription. Nat Cell Biol 2, 745-749. Massague, J. (2012). TGF-beta signaling in development and disease. FEBS Lett 586, 1833. Massague, J., Seoane, J. and Wotton, D. (2005). Smad transcription factors. Genes Dev 19, 2783-2810. McLin, V. A., Rankin, S. A. and Zorn, A. M. (2007). Repression of Wnt/beta-catenin signaling in the anterior endoderm is essential for liver and pancreas development. Development (Cambridge, England) 134, 2207-2217. Mizoguchi, T., Izawa, T., Kuroiwa, A. and Kikuchi, Y. (2006). Fgf signaling negatively regulates Nodal-dependent endoderm induction in zebrafish. Dev Biol 300, 612- 622. Morikawa, M., Koinuma, D., Tsutsumi, S., Vasilaki, E., Kanki, Y., Heldin, C. H., Aburatani, H. and Miyazono, K. (2011). ChIP-seq reveals cell type-specific binding patterns of BMP-specific Smads and a novel binding motif. Nucleic Acids Res 39, 8712-8727. Mosimann, C., Hausmann, G. and Basler, K. (2009). Beta-catenin hits chromatin: regulation of Wnt target gene activation. Nat Rev Mol Cell Biol 10, 276-286. Muller, B., Hartmann, B., Pyrowolakis, G., Affolter, M. and Basler, K. (2003). Conversion of an extracellular Dpp/BMP morphogen gradient into an inverse transcriptional gradient. Cell 113, 221-233. Neijts, R., Amin, S., van Rooijen, C., Tan, S., Creyghton, M. P., de Laat, W. and Deschamps, J. (2016). Polarized regulatory landscape and Wnt responsiveness underlie Hox activation in embryos. Genes Dev 30, 1937-1942. Nishita, M., Hashimoto, M. K., Ogata, S., Laurent, M. N., Ueno, N., Shibuya, H. and Cho, K. W. (2000). Interaction between Wnt and TGF-beta signalling pathways during formation of Spemann's organizer. Nature 403, 781-785. Olson, L. E., Tollkuhn, J., Scafoglio, C., Krones, A., Zhang, J., Ohgi, K. A., Wu, W., Taketo, M. M., Kemler, R., Grosschedl, R., et al. (2006). Homeodomain- mediated beta-catenin-dependent switching events dictate cell-lineage determination. Cell 125, 593-605. Onichtchouk, D., Chen, Y. G., Dosch, R., Gawantka, V., Delius, H., Massague, J. and Niehrs, C. (1999). Silencing of TGF-beta signalling by the pseudoreceptor BAMBI. Nature 401, 480-485.

33 Onichtchouk, D., Gawantka, V., Dosch, R., Delius, H., Hirschfeld, K., Blumenstock, C. and Niehrs, C. (1996). The Xvent-2 homeobox gene is part of the BMP-4 signalling pathway controlling [correction of controling] dorsoventral patterning of Xenopus mesoderm. Development (Cambridge, England) 122, 3045-3053. Parker, D. S., Ni, Y. Y., Chang, J. L., Li, J. and Cadigan, K. M. (2008). Wingless signaling induces widespread chromatin remodeling of target loci. Mol Cell Biol 28, 1815-1828. Pouponnot, C., Jayaraman, L. and Massague, J. (1998). Physical and functional interaction of SMADs and p300/CBP. J Biol Chem 273, 22865-22868. Ramel, M. C., Buckles, G. R., Baker, K. D. and Lekven, A. C. (2005). WNT8 and BMP2B co-regulate non-axial mesoderm patterning during zebrafish gastrulation. Dev Biol 287, 237-248. Ramel, M. C. and Lekven, A. C. (2004). Repression of the vertebrate organizer by Wnt8 is mediated by Vent and Vox. Development (Cambridge, England) 131, 3991-4000. Rankin, S. A., Gallas, A. L., Neto, A., Gomez-Skarmeta, J. L. and Zorn, A. M. (2012). Suppression of Bmp4 signaling by the zinc-finger repressors Osr1 and Osr2 is required for Wnt/beta-catenin-mediated lung specification in Xenopus. Development (Cambridge, England) 139, 3010-3020. Rankin, S. A., Kormish, J., Kofron, M., Jegga, A. and Zorn, A. M. (2011). A gene regulatory network controlling hhex transcription in the anterior endoderm of the organizer. Dev Biol 351, 297-310. Rossi, J. M., Dunn, N. R., Hogan, B. L. and Zaret, K. S. (2001). Distinct mesodermal signals, including BMPs from the septum transversum mesenchyme, are required in combination for hepatogenesis from the endoderm. Genes Dev 15, 1998-2009. Sekiya, T., Muthurajan, U. M., Luger, K., Tulin, A. V. and Zaret, K. S. (2009). Nucleosome-binding affinity as a primary determinant of the nuclear mobility of the pioneer transcription factor FoxA. Genes Dev 23, 804-809. Shen, M. M. (2007). Nodal signaling: developmental roles and regulation. Development (Cambridge, England) 134, 1023-1034. Sherwood, R. I., Maehr, R., Mazzoni, E. O. and Melton, D. A. (2011). Wnt signaling specifies and patterns intestinal endoderm. Mech Dev 128, 387-400. Shin, D., Shin, C. H., Tucker, J., Ober, E. A., Rentzsch, F., Poss, K. D., Hammerschmidt, M., Mullins, M. C. and Stainier, D. Y. (2007). Bmp and Fgf signaling are essential for liver specification in zebrafish. Development (Cambridge, England) 134, 2041-2050. Sierra, J., Yoshida, T., Joazeiro, C. A. and Jones, K. A. (2006). The APC tumor suppressor counteracts beta-catenin activation and H3K4 methylation at Wnt target genes. Genes Dev 20, 586-600.

34 Sinner, D., Kirilenko, P., Rankin, S., Wei, E., Howard, L., Kofron, M., Heasman, J., Woodland, H. R. and Zorn, A. M. (2006). Global analysis of the transcriptional network controlling Xenopus endoderm formation. Development (Cambridge, England) 133, 1955-1966. So, J., Martin, B. L., Kimelman, D. and Shin, D. (2013). Wnt/beta-catenin signaling cell-autonomously converts non-hepatic endodermal cells to a liver fate. Biol Open 2, 30-36. Spence, J. R., Mayhew, C. N., Rankin, S. A., Kuhar, M. F., Vallance, J. E., Tolle, K., Hoskins, E. E., Kalinichenko, V. V., Wells, S. I., Zorn, A. M., et al. (2011). Directed differentiation of human pluripotent stem cells into intestinal tissue in vitro. Nature 470, 105-109. Spitz, F. and Furlong, E. E. (2012). Transcription factors: from enhancer binding to developmental control. Nat Rev Genet 13, 613-626. Stuckenholz, C., Lu, L., Thakur, P. C., Choi, T. Y., Shin, D. and Bahary, N. (2013). Sfrp5 modulates both Wnt and BMP signaling and regulates gastrointestinal organogenesis [corrected] in the zebrafish, Danio rerio. PLoS One 8, e62470. Tamai, K., Zeng, X., Liu, C., Zhang, X., Harada, Y., Chang, Z. and He, X. (2004). A mechanism for Wnt coreceptor activation. Mol Cell 13, 149-156. Theil, T., Aydin, S., Koch, S., Grotewold, L. and Ruther, U. (2002). Wnt and Bmp signalling cooperatively regulate graded Emx2 expression in the dorsal telencephalon. Development (Cambridge, England) 129, 3045-3054. Theisen, H., Haerry, T. E., O'Connor, M. B. and Marsh, J. L. (1996). Developmental territories created by mutual antagonism between Wingless and Decapentaplegic. Development (Cambridge, England) 122, 3939-3948. Theisen, H., Syed, A., Nguyen, B. T., Lukacsovich, T., Purcell, J., Srivastava, G. P., Iron, D., Gaudenz, K., Nie, Q., Wan, F. Y., et al. (2007). Wingless directly represses DPP morphogen expression via an armadillo/TCF/Brinker complex. PLoS One 2, e142. Thomas, P. Q., Brown, A. and Beddington, R. S. (1998). Hex: a homeobox gene revealing peri-implantation asymmetry in the mouse embryo and an early transient marker of endothelial cell precursors. Development (Cambridge, England) 125, 85-94. Tiso, N., Filippi, A., Pauls, S., Bortolussi, M. and Argenton, F. (2002). BMP signalling regulates anteroposterior endoderm patterning in zebrafish. Mech Dev 118, 29-37. Trompouki, E., Bowman, T. V., Lawton, L. N., Fan, Z. P., Wu, D. C., DiBiase, A., Martin, C. S., Cech, J. N., Sessa, A. K., Leblanc, J. L., et al. (2011). Lineage regulators direct BMP and Wnt pathways to cell-specific programs during differentiation and regeneration. Cell 147, 577-589.

35 Tsankov, A. M., Gu, H., Akopian, V., Ziller, M. J., Donaghey, J., Amit, I., Gnirke, A. and Meissner, A. (2015). Transcription factor binding dynamics during human ES cell differentiation. Nature 518, 344-349. van de Wetering, M., Cavallo, R., Dooijes, D., van Beest, M., van Es, J., Loureiro, J., Ypma, A., Hursh, D., Jones, T., Bejsovec, A., et al. (1997). Armadillo coactivates transcription driven by the product of the Drosophila segment polarity gene dTCF. Cell 88, 789-799. von Bubnoff, A. and Cho, K. W. (2001). Intracellular BMP signaling regulation in vertebrates: pathway or network? Dev Biol 239, 1-14. Walmsley, M., Ciau-Uitz, A. and Patient, R. (2002). Adult and embryonic blood and endothelium derive from distinct precursor populations which are differentially programmed by BMP in Xenopus. Development (Cambridge, England) 129, 5683-5695. Wandzioch, E. and Zaret, K. S. (2009). Dynamic signaling network for the specification of embryonic pancreas and liver progenitors. Science 324, 1707-1710. Wang, A., Yue, F., Li, Y., Xie, R., Harper, T., Patel, N. A., Muth, K., Palmer, J., Qiu, Y., Wang, J., et al. (2015). Epigenetic priming of enhancers predicts developmental competence of hESC-derived endodermal lineage intermediates. Cell Stem Cell 16, 386-399. Wells, J. M. and Melton, D. A. (2000). Early mouse endoderm is patterned by soluble factors from adjacent germ layers. Development (Cambridge, England) 127, 1563-1572. Willems, E. and Leyns, L. (2008). Patterning of mouse embryonic stem cell-derived pan-mesoderm by Activin A/Nodal and Bmp4 signaling requires Fibroblast Growth Factor activity. Differentiation 76, 745-759. Wills, A., Dickinson, K., Khokha, M. and Baker, J. C. (2008). Bmp signaling is necessary and sufficient for ventrolateral endoderm specification in Xenopus. Dev Dyn 237, 2177-2186. Wrana, J. L., Attisano, L., Wieser, R., Ventura, F. and Massague, J. (1994). Mechanism of activation of the TGF-beta receptor. Nature 370, 341-347. Xu, B., Chen, C., Chen, H., Zheng, S. G., Bringas, P., Jr., Xu, M., Zhou, X., Chen, D., Umans, L., Zwijsen, A., et al. (2011). Smad1 and its target gene Wif1 coordinate BMP and Wnt signaling activities to regulate fetal lung development. Development (Cambridge, England) 138, 925-935. Xu, J., Pope, S. D., Jazirehi, A. R., Attema, J. L., Papathanasiou, P., Watts, J. A., Zaret, K. S., Weissman, I. L. and Smale, S. T. (2007). Pioneer factor interactions and unmethylated CpG dinucleotides mark silent tissue-specific enhancers in embryonic stem cells. Proc Natl Acad Sci U S A 104, 12377-12382. Zaret, K. S. and Mango, S. E. (2016). Pioneer transcription factors, chromatin dynamics, and cell fate control. Curr Opin Genet Dev 37, 76-81.

36 Zeng, X., Huang, H., Tamai, K., Zhang, X., Harada, Y., Yokota, C., Almeida, K., Wang, J., Doble, B., Woodgett, J., et al. (2008a). Initiation of Wnt signaling: control of Wnt coreceptor Lrp6 phosphorylation/activation via frizzled, dishevelled and axin functions. Development (Cambridge, England) 135, 367-375. Zeng, Y. A., Rahnama, M., Wang, S., Lee, W. and Verheyen, E. M. (2008b). Inhibition of Drosophila Wg signaling involves competition between Mad and Armadillo/beta-catenin for dTcf binding. PLoS One 3, e3893. Zhang, Z., Rankin, S. A. and Zorn, A. M. (2013). Different thresholds of Wnt-Frizzled 7 signaling coordinate proliferation, morphogenesis and fate of endoderm progenitor cells. Dev Biol 378, 1-12. Zhu, H., Kavsak, P., Abdollah, S., Wrana, J. L. and Thomsen, G. H. (1999). A SMAD ubiquitin ligase targets the BMP pathway and affects embryonic pattern formation. Nature 400, 687-693. Zorn, A. M. and Wells, J. M. (2009). Vertebrate endoderm development and organ formation. Annu Rev Cell Dev Biol 25, 221-251.

Chapter 2: Genomic integration of Wnt/β-catenin and BMP/Smad1 signaling coordinates foregut and hindgut transcriptional program

AUTHORS

Mariana L. Stevens1, Praneet Chaturvedi1, Scott A. Rankin1, Melissa Macdonald1,

Sajjeev Jagannathan2, Masashi Yukawa2, Artem Barski2 and Aaron M. Zorn1*

AFFILIATIONS

1Division of Developmental Biology, Perinatal Institute 2 Division of Allergy &

Immunology and Human Genetics, Cincinnati Children's Research Foundation and

Department of Pediatrics College of Medicine, University of Cincinnati, Cincinnati, Ohio

45229, USA.

* Correspondence: [email protected]

KEY WORDS

BMP, Smad1, Wnt, beta-catenin, foregut, hindgut, ChIP-seq, RNA-seq, Xenopus, endoderm, mesoderm

SUMMARY STATEMENT

Foregut and hindgut transcriptional programs are regulated by BMP and Wnt signaling through integration of Smad1 and β-catenin on cis-regulatory elements.

Accepted to Development on February 3, 2017

Abstract

Digestive system development is orchestrated by combinatorial signaling interactions between endoderm and mesoderm, but how these signals are integrated in the genome is poorly understood. Here we identified the transcriptomes of Xenopus foregut and hindgut progenitors, which are conserved with mammals. Using RNA-seq and ChIP-seq we show that BMP/Smad1 regulates dorsal-ventral gene expression in both the endoderm and mesoderm, whereas Wnt/β-catenin acts as a genome-wide toggle between foregut and hindgut programs. Unexpectedly β-catenin and Smad1 binding were associated with both transcriptional activation and repression, with Wnt-repressed genes often lacking canonical Tcf DNA-binding motifs, suggesting a novel mode of direct repression. Combinatorial Wnt-BMP signaling was mediated by Smad1 and β- catenin co-occupying hundreds of DNA cis-regulatory elements, and by a crosstalk where Wnt negatively regulated BMP ligand expression in the foregut. These results extend our understanding of gastrointestinal organogenesis and how Wnt and BMP may coordinate genomic responses in other contexts.

Introduction

Embryonic development of the digestive and respiratory systems is controlled by a reiterative series of growth factor interactions between the epithelium and mesenchyme

(reviewed in Zorn and Wells, 2009). Our understanding of how combinatorial signals orchestrate organogenesis in animal models has been the foundation for strategies to direct the differentiation of human pluripotent stem cells (hPSCs) into organoids for disease modeling and regenerative medicine (reviewed in Lancaster and Knoblich,

2014). Despite our growing knowledge of which growth factors act where and when during organogenesis, how combinatorial signals are integrated at the genomic level to coordinate gene expression through DNA cis-regulatory modules (CRMs) is still poorly understood. Here we investigated how spatially restricted BMP and Wnt signals coordinate the genomic transcriptional programs of foregut (FG) and hindgut (HG) progenitors in Xenopus embryos.

In post-gastrula vertebrate embryos and during hPSC differentiation, Wnt and

BMP pattern the naïve endoderm (endo) and mesoderm (meso) germ layers along the anterior-posterior (A-P) axis into FG and HG progenitors (Loh et al., 2014; Zorn and

Wells, 2009). Bmp4/7 and Wnt8 ligands expressed in the ventral-posterior mesendoderm promote HG fate and inhibit FG lineages (McLin et al., 2007; Rankin et al., 2011; Sherwood et al., 2011; Spence et al., 2011), whereas the anterior mesendoderm secretes Wnt- and BMP-antagonists (e.g., Dkk1, Sfrp5, Noggin and

Chordin) that protect the FG from these posteriorizing signals (De Robertis, 2009;

Green et al., 2011; Li et al., 2008). Within the meso, these same Wnt-antagonists also promote anterior lateral plate and cardiac fates (reviewed in Gibb et al., 2013; Klaus and

Birchmeier, 2009), thus coordinating the development of the FG endo and meso lineages. The effects of these pathways on patterning are temporally restricted such that several hours later, spatially distinct Wnt and/or BMP signals no longer suppress

FG identity, but promote lung, thyroid, liver, pancreas and heart organogenesis (Kenny et al., 2012; Klaus and Birchmeier, 2009; Zorn and Wells, 2009).

In both the BMP and Wnt pathways, ligand-receptor binding stimulates the translocation of transcriptional effectors to the nucleus. Activated BMP receptors phosphorylate cytosolic Smad1,5,8 (Smad1), which forms a complex with Smad4 and enter the nucleus to interact with DNA-binding transcription factors, such as Schnurri,

Gata or Runx (reviewed in Gaarenstroom and Hill, 2014). In the canonical Wnt pathway, receptor binding results in stabilization and nuclear translocation of β-catenin, which interacts with DNA-binding Tcf/Lef transcription factors, displacing a co-repressor complex containing Groucho/TLE and recruiting transcriptional co-activators (reviewed in Cadigan and Waterman, 2012). Tcf/β-catenin and Smad1/Smad4 both recruit co- activator complexes containing the p300 or CBP histone acetyltransferases (HATs), which acetylate H3 histones, to promote chromatin opening, RNA polymerase binding and transcription (Cadigan and Waterman, 2012; Gaarenstroom and Hill, 2014).

Combinatorial Wnt and BMP signaling governs cellular responses in a variety of development and disease contexts, and in a few well-characterized target genes β- catenin and Smad1 converge on the same DNA cis-regulatory modules (CRMs) to stimulate transcription (reviewed in Itasaki and Hoppler, 2010). ChIP-seq of in vitro differentiated myeloid cells indicates that β-catenin and Smad1 can co-occupy many genomic loci suggesting this may be a widespread mechanism of signal integration

(Trompouki et al., 2011). Recent genome-scale studies in Xenopus embryos and hPSCs have begun to reveal how β-catenin and Smad2, which transduces

Activin/Nodal signals, regulate transcription during germ layer formation and gastrulation (Estaras et al., 2015; Gupta et al., 2014; Kim et al., 2011; Kjolby and

Harland, 2016; Nakamura et al., 2016; Tsankov et al., 2015). Despite these advances it is still unknown how Wnt/β-catenin and BMP/Smad1 signals are integrated in the genome to regulate cell fate choices during digestive system organogenesis.

Here we identify the transcriptional program of FG and HG progenitors in

Xenopus laevis embryos, which are largely conserved with mammals. RNA-seq and

ChIP-seq revealed how BMP and Wnt signals coordinate spatially restricted FG and HG gene expression with β-catenin and Smad1 co-binding CRMs of hundreds of key cell identity regulators. We identify a Wnt-BMP crosstalk in the FG and unexpectedly find that many genes inhibited by BMP or Wnt are associated with Smad1 or β-catenin binding, suggesting direct repression. These findings advance our understanding of how combinatorial signaling is integrated in the genome during gastrointestinal development and serve as a paradigm for other development and disease contexts.

Results

Transcriptional program of foregut and hindgut progenitors

In order to examine how BMP and Wnt regulate early gut development at the genomic level, we first defined the FG and HG transcriptomes. Taking advantage of large and abundant Xenopus laevis embryos we microdissected the ventral FG-endo, FG-meso,

HG-endo and HG-meso tissues from 50 sibling embryos at stage NF20 (22 hours post fertilization; ~ 6-7 somites), and performed RNA-seq (Fig. 1A,B). This time point is similar to E8.5 in mice, when the FG is being patterned and is still plastic. Differential expression analysis of FG (endo + meso) versus the HG (endo + meso), identified 906

FG-enriched genes and 987 HG-enriched genes, whereas endoderm (FG + HG) compared to the mesoderm (FG + HG) identified 3439 endo-enriched and 4829 meso- enriched transcripts (log2FC ≥1, FDR ≤5%) (Fig. 1A-C, Fig. S1A, Tables S1, S2). These gene lists contained over 98% of well-known FG and HG transcripts (n=74) manually curated from mouse and differentiated hPSCs, confirming the extensive conservation

(Fig. S1B, Table S3). FG-endo included the transcription factors hhex and pdx1, while the FG-meso included key regulators of cardiac (hand2), myeloid (spib and cebpa) and endothelial (fli1) lineages (Fig. 1D). Gene Ontology (GO) analysis showed epithelial, vasculature and circulatory system development among the top FG-enriched terms (Fig.

S1C). The HG transcriptome was enriched for GO terms related to A-P patterning and included the key intestinal regulators cdx1, 2 and 4 in the HG-endo, and homeobox genes hox5-11 in the HG-meso (Fig. 1D, Fig. S1B,C, Table S1).

BMP and Wnt regulate gut tube patterning

Consistent with a role in patterning, GO terms related to Wnt and BMP were enriched in the FG and HG datasets (Fig. S1C). Examination of BMP and Wnt pathway components in the total RNA-seq data indicated that BMP ligands (bmp2, 4 and 7), receptors (bmpr1a, bmpr2) and BMP-target genes (id3 and szl) were variably expressed in both FG and HG tissues, with no obvious A-P difference (Fig. S1D). In contrast, Wnt

ligands were enriched in the HG while Wnt-antagonists (dkk1, sfrp2 and sfpr5) were restricted to the FG-endo (Fig. S1D).

We next compared the levels of nuclear (n)β-catenin and pSmad1 immunostaining in FG and HG domain, which are marked by hhex and ventx2 respectively (Fig. 1E) (McLin et al., 2007). Consistent with previous reports, the gastrula

NF11 anterior mesendoderm (presumptive FG) had low levels of both pSmad1 and

(n)β-catenin, while the posterior mesendoderm (presumptive HG) was the opposite

(Schohl and Fagotto, 2002). After gastrulation, at stage NF20, (n)β-catenin was still low in the FG and high in the HG, however we observed an up-regulation of pSmad1 in the ventral (but not dorsal) FG (Fig. 1F), consistent with the known de novo expression of bmp2/4/7 in the pre-cardiac FG-meso (Kenny et al., 2012). Thus during FG-HG patterning, BMP and Wnt are differentially active along orthologous axes; pSmad1 is high in the ventral and low in the dorsal FG and HG, whereas (n)β-catenin is low in the

FG and high in the HG (Fig. 1G).

We next examined the impact of BMP or Wnt inhibition on progenitor patterning by treating embryos with the BMP-receptor inhibitor DMH1 (40 µM), or inhibiting Wnt with a heatshock inducible Dkk1 transgenic line Tg(hsp70:dkk1) (Lin and Slack, 2008).

We added DMH1 or heatshock between stages NF12-20, during FG-HG patterning but after gastrulation to avoid disruption of axial patterning. BMP-inhibition reduced both hhex and ventx2.1, whereas Wnt-inhibition expanded hhex and reduced ventx2.1 expression (Fig. S1E). To test whether these changes in patterning impacted subsequent organogenesis, we analyzed organ lineages at NF35. BMP inhibition between NF12-20 resulted in an expansion of the dorsal esophageal marker sox2, and

loss of liver (nr1h5), lung (nkx2-1) and heart (nkx2-5). In contrast, Wnt-inhibition expanded the liver (nr1h5) and reduced the intestinal marker (darmin) (Fig. S1E). Thus spatially restricted Wnt and BMP pattern the FG and HG progenitors.

Fig. 1: Transcriptional program of FG and HG progenitors correlates with differential BMP and Wnt signaling. (A) Fate map showing that FG progenitors (yellow) give rise to lungs, liver, pancreas and stomach, whereas HG progenitors (green) give rise to intestine (Chalmers and Slack, 2000). (B) Experimental design. RNA-seq was performed on FG-endo, FG-meso, HG-endo and HG-meso explants dissected from stage NF20 embryos. Differentially expressed transcripts were identified by pairwise comparisons of FG (endo and meso) versus HG (endo and meso), as well as endo (FG and HG) versus meso (FG and HG) tissue (log2 fold change (FC) ≤-1 or ≥1, false discovery rate (FDR) ≤5%). (C) Venn diagram showing the intersection of two separate differential expression analysis: FG versus HG and endo versus meso, showing mutually exclusive lists of transcripts with enriched expression. (D) Heatmap clustering of the 906 FG-enriched and 987 HG- enriched genes showing expression in the indicated tissues with representative FG (orange) and HG (green) genes listed on the right. (E) In situ hybridization of sagittal bisected stage NF20 embryos with hhex and ventx2.1 marking FG and HG domains, respectively. (F) BMP and Wnt activity shown by pSmad1 (red) and nuclear (n)β-catenin (red) immunostaining in NF20 embryos. Nuclei staining with DAPI (green) (G) pSmad1 is high in the ventral and low in the dorsal FG and HG, whereas (n)β-catenin is low in the FG and high in the HG.

BMP-regulated foregut and hindgut transcriptome

To identify BMP-regulated transcripts, we isolated FG and HG explants (containing both endo and meso) from DMH1 or vehicle treated NF20 embryos and analyzed these by

RNA-seq (Fig. 2A). Transcripts with reduced expression in DMH1 compared to controls were classified as normally activated by BMP (n=697), while increased expression in

DMH1 indicated that they are repressed by BMP (n=1063) (log2FC ≥1, FDR ≤5%) (Fig.

2B, Fig. S2A,B, Tables S2, S4). Eight genes showed variable regulation by DMH1 and were not included in further analysis (Fig. S2B), resulting in 1760 unique BMP-regulated genes. Approximately 17% (155/906) of FG-enriched genes were activated by BMP, whereas 21% (185/906) were repressed. Of the HG-enriched transcripts, 10% (97/987) were activated by BMP and 9% (89/987) were repressed (Fig. 2B). BMP-activated genes were enriched for GO terms related to cardiovascular, blood vessel and digestive system development, while BMP-repressed genes were enriched for skeletal and renal system, indicative of dorsal gene expression, normally low in FG and HG explants (Fig.

S2C).

BMP signaling was required to maintain expression of key posterior homeobox genes cdx2, cdx4, ventx2 and msx1 in both the HG-endo and HG-meso (Fig. 2C-E).

The role of BMP was more complex in the FG. Approximately 85% of the BMP- regulated FG-endo genes appear to be repressed by BMP (Fig. 2C,D), in contrast to the

FG-meso where BMP was required for ~60% of the BMP-regulated genes including known regulators of the heart (tbx20 and hand2) and myeloid (spib and cebpa) lineages

(Fig. 2D).

BMP regulates dorsal-ventral patterning of the early foregut

Although most BMP-regulated FG-endo genes were repressed, a few FG-endo genes required BMP for their expression, including hhex and sfrp5 (Fig. 2D,E), which are implicated in liver and ventral pancreas development (Li et al., 2008; McLin et al., 2007).

Together with the observation that the ventral FG has higher nuclear pSmad1 levels than the dorsal FG (Fig. 1F) this prompted us to examine dorso-ventral (D-V) patterning in more detail.

Since some digestive organs, including the esophagus, stomach, intestine and pancreas originate from both ventral and dorsal endoderm cells (Chalmers and Slack,

2000), we hypothesized that ventral and dorsal transcripts might be differentially regulated by BMP. Unsupervised clustering of BMP-regulated transcripts in control and

DMH1-treated FG explants along with isolated dorsal explants (which contain a thin layer of dorsal endoderm), revealed that the DMH1-treated FG had an expression profile similar to dorsal tissue, suggesting that BMP represses dorsal lineages (Fig.

S2D). In situ hybridization confirmed that DMH1 treatment reduced the expression of ventral FG genes sfrp5 (endo), cebpa (myeloid) and nkx2.5 (cardiac), whilst causing the expansion of dorsal endo transcripts mnx1 and hrg and the paraxial meso gene foxc2.

Injection of BMP2 protein into the FG had the opposite effect causing an expansion of ventral genes, and repression of dorsal markers (Fig. 2E, Fig. S2E). Thus in the FG between stages NF12-20, BMP promotes ventral (presumptive liver, lung, cardiac) and represses dorsal fates (presumptive esophagus, kidney and paraxial mesoderm).

Fig. 2: BMP signaling coordinates D-V patterning. (A) Experimental design. FG and HG (endo + meso) explants were dissected from DMSO and DMH1 treated NF20 embryos and submitted for RNA-seq (in triplicate). (B) Venn diagram showing overlap of FG-enriched genes, HG-enriched genes and BMP- activated and repressed transcripts (log2FC ≥1, FDR ≤5%); see Fig. S2B for details. (C) Expression heatmap clustering of FG-endo, FG-meso, HG-endo and HG-meso enriched transcripts in vehicle and DMH1-treated explants. (D) Different categories of BMP activated (ACT) or repressed (REP) transcripts based on whether the gene is normally enriched in the endo, meso or expressed in both endo and meso (en=me). (E) In situ hybridization of vehicle control, DMH1-treated or BMP2 injected stage NF20 embryos in mid-sagittal section (cdx2, sfrp5, cebpa and nkx2-5; anterior left and dorsal up) or cross section (mnx1 and hrg; dorsal up, white lines indicate expression domain), n>20 for each probe.

Smad1 chromatin binding to BMP-regulated hindgut and foregut genes

In efforts to identify direct BMP target genes we performed Smad1 ChIP-seq on NF20 embryos and identified 7976 Smad1 peaks, located within +/- 20kb of 5252 genes (Fig.

3A, Table S5). This represents ~18% of the 45,099 predicted genes in the allotetraploid

Xenopus laevis genome (Session et al., 2016). Smad consensus DNA-binding sites, as well as Gata and Tbx motifs were enriched in Smad1 peaks (Fig. 3B). Specificity of the

Smad1 antibody was confirmed by reduction of Smad1 binding to known BMP-target genes msx1, id3 and ventx2.1 (Karaulanov et al., 2004) upon DMH1 treatment (Fig.

S3A,B). Of the Smad1 peaks, 27% overlapped with p300 ChIP-seq peaks indicative of active transcription, and ~56% of the Smad1 associated genes had detectable expression in NF20 embryos (Fig. S3C,D).

Of the total 1760 BMP-regulated genes in the FG or HG (from Fig. 2B), ~35%

(615/1760) were associated with Smad1-binding, a statistically significant enrichment based on a hypergeometric test (HGT) (2.7 fold enrichment (FE), p<0.05) (Fig. 3A).

These included 48% of the BMP-activated HG genes (47/97, 3.7 FE, HGT *p<0.05), such as cdx2, ventx2.1 and msx1, consistent with direct activation by Smad1 (Fig.

3C,D). In contrast, Smad1-binding was only associated with 17% (26/155) of BMP- activated FG genes such as hand2 (meso) and hhex (endo) (Fig. 3C,D), suggesting that most BMP-activated FG genes are indirect targets. Unexpectedly, Smad1-binding was more associated with BMP-repressed genes (n=410) than with BMP-activated genes

(n=205). Indeed, ~55% of BMP-repressed FG genes (103/185, 4.3 FE, HGT *p<0.05) and ~49% of BMP-repressed HG genes (44/89, 3.8 FE, HGT *p<0.05) were associated with Smad1 peaks, suggesting direct BMP repression, including dorsal genes such as foxc2 and mnx1 (Fig. 3D). Analysis of p300 co-occupancy on different classes of 51

Smad1 peaks did not however show a statistically significant enrichment on activated versus repressed genes, presumably because the ChIP was done on whole embryos containing both expressing and non expressing cells. In sum, about a third of the BMP regulated FG and HG transcriptome was associated with Smad1-binding suggesting direct transcriptional activation and repression.

Fig. 3: Smad1 chromatin binding to BMP-regulated genes. (A) Smad1 ChIP-seq analysis of NF20 whole embryos identified 7976 peaks within +/- 20 Kb of 5252 genes. The Venn diagram intersecting 5252 Smad1-bound genes with the 1760 BMP-regulated genes (from Fig. 2B) grouped as FG-enriched (n=340), HG- enriched (n=186) or expressed at similar levels in the FG and HG (FG=HG; n=1234), identifies 615 BMP-regulated genes associated with Smad1-binding, BMP-activated genes in red and BMP-repressed in blue. *Statistically enriched based on hypergeometric test (p<0.05). (B) DNA-binding protein motif enrichment analysis of 7976 Smad1 peaks. (C) Chart illustrates Smad1-bound genes within the different categories of BMP-activated (ACT) and repressed (REP) genes. (D) Genome browser view of Smad1 peaks on BMP-activated HG gene cdx2, and ventral mesoderm hand2, as well as BMP-repressed dorsal genes foxc2 and mnx1. Red boxes indicate statistically significant Smad1 peaks.

Canonical Wnt promotes hindgut and represses foregut transcriptomes

To determine the genomic basis of Wnt-mediated A-P patterning, we performed RNA- seq on FG and HG explants (endo + meso) from NF20 embryos where Wnt/β-catenin activity was either activated with BIO treatment or inhibited with secreted Wnt- antagonist Dkk1 expressed from Tg(hsp70:dkk1) from stages NF12-20 stages (Fig. 4A).

Transcripts repressed by Dkk1 or induced by BIO, relative to controls, were considered

Wnt-activated, while transcripts up regulated by Dkk1 or repressed by BIO were considered Wnt-repressed (Fig. S4A). Differential expression analysis identified 959

Wnt-activated transcripts and 2032 Wnt-repressed transcripts (log2FC ≥1, FDR ≤5%).

41 transcripts showed variable regulation by BIO and Dkk1 and were not included in further analysis (Fig. 4B, Fig. S4A, Tables S2, S6). Thus we classified a total of 2991

Wnt-regulated genes. Interestingly, over half of the FG-enriched genes (496/906) were repressed by Wnt, whereas only 3% (28/906) were activated. HG-enriched genes exhibited the opposite behavior, with 25% (247/987) being Wnt-activated and only 6%

(57/987) repressed (Fig. 4B). GO term enrichment showed A-P patterning among the top terms for Wnt-activated genes, while Wnt-repressed genes were enriched for terms related to the circulatory system, consistent with the known role of Wnt in repressing early cardiac fate (Fig. S4B). Unsupervised clustering of FG and HG transcript levels in the various experimental conditions revealed that BIO-treated FG samples cluster with

HG samples, indicating a genome-scale switch from FG to HG upon Wnt stimulation

(Fig. 4C), which is also clearly shown in the scatter plots (Fig. 4D, Fig. S4C,D).

Wnt had the same impact on endo and meso transcripts - repressing in the FG and activating in the HG - regardless of germ layer. Wnt-repressed FG transcripts

included hhex and sfrp2 in the FG-endo, as well as spib and hand2 in the FG-meso. HG transcripts activated by Wnt included ventx2.1 and cdx2 in the HG-endo, as well as hoxa9 and hoxd10 in the HG-meso (Fig. 4E). In situ hybridization confirmed that Dkk1 expanded expression of the FG genes gata4, cebpa and tbx1, whilst reducing the HG transcripts cdx2 and msx1, with BIO treatment having the opposite effect (Fig. 4F). Thus

Wnt/β-catenin acts as a genome wide switch, promoting the HG transcriptional program and repressing the FG program.

Fig. 4: Wnt signaling promotes HG transcriptional program and represses FG transcriptional program. (A) Experimental design. FG and HG (endo + meso) explants were dissected from DMSO, BIO and Tg(hsp70:dkk1) NF20 embryos and submitted for RNA-seq (in triplicates). (B) Venn diagram illustrates overlap of FG-enriched, HG-enriched, Wnt- activated and Wnt-repressed genes (log2FC ≤-1 or ≥1, FDR ≤5%); see Fig. S4A for details. (C) Unsupervised clustering of FG- and HG-enriched genes, showing that BIO- treated FG has an expression profile similar to HG control. (D) Scatter plot showing 56

log2FC in expression between control, BIO and Tg(hsp70:dkk1) FG and HG explants. Transcripts are colored based on the normal control expression; HG-enriched in green, FG-enriched in orange, transcripts expressed similarly FG and HG (FG=HG) in grey and normally not expressed in FG or HG in black. (E) Different categories of Wnt activated (ACT) or repressed (REP) transcripts group based on whether the gene is normally enriched in the endo, meso or expressed in both endo and meso (en=me). (F) In situ hybridization of Control, Tg(hsp70:dkk1) or BIO treated embryos in mid-sagittal section, anterior left and dorsal up. Anterior genes are gata4, cebpa and tbx1 and posterior genes are cdx2, and msx1, n>20 for each probe.

Chromatin binding of β-catenin is associated with activation of HG and repression of FG genes

To determine how Wnt regulates A-P patterning at the genomic level, we performed β- catenin ChIP-seq on stage NF20 control and BIO-treated FG and HG explants (Fig.

S5A). We merged sequence files from the four different ChIPs to call β-catenin peaks.

This identified 16303 β-catenin peaks associated with 11007 genes (+/-20kb) (Fig. 5A,

Fig. S5B, Table S7), which represents ~24% of the genes in the genome. As expected,

β-catenin peaks were enriched for Tcf/Lef DNA-binding motifs and ChIP-PCR of known

Wnt-target genes in Tg(hsp70:dkk1) embryos confirmed the expected reduction in β- catenin binding (Fig. S5C,D).

Of the 11,007 genes associated with β-catenin peaks only 1,243 (~11%) were

Wnt regulated based on our RNA-seq data (Fig. 5A), suggesting that ~90% of all genomic β-catenin binding events in the FG or HG tissue were not associated with Wnt- regulated transcription, similar to recent findings in the Xenopus gastrula (Nakamura et al., 2016). Of all Wnt-regulated genes, ~41% (1243/2991, 1.5 FE, HGT p<0.05) were associated with β-catenin peaks, including 73% (180/247, 2.7 FE, HGT *p<0.05) of the

HG-enriched Wnt-activated genes (Fig. 5Aa), suggesting direct regulation.

Unexpectedly, 42% (208/496, 1.6 FE, HGT *p<0.05) of all FG-enriched Wnt-repressed genes were also associated with β-catenin peaks (Fig. 5Ab), suggesting direct repression. This was surprising since β-catenin is usually thought to stimulate transcription and thus we expected that Wnt would indirectly repress FG genes. We next assessed whether activation or repression correlated with changes in recruitment of β-catenin to chromatin through differential peak enrichment analysis of control versus

BIO explants (Fig. 5B). As expected, 88% (159/180) of the Wnt-activated HG genes had 58

increased β-catenin binding in BIO-treated FG tissue, as exemplified by cdx2 (Fig.

5Ba’,C). Surprisingly, among the 208 Wnt-repressed FG genes (Fig. 5Ab), we observed different behaviors; 59 genes were associated with increased β-catenin binding upon

BIO treatment (Fig. 5Bb’), 43 genes experienced reduced β-catenin binding (Fig. 5Bb”), whereas 91 Wnt-repressed genes had no change in β-catenin (Fig. 5B). For example,

β-catenin binding near the sfrp2 promoter was increased upon BIO treatment, whereas the nkx2-3 peak was reduced (Fig. 5C). Plotting the average tag density of these different classes of peaks confirmed the significant changes in β-catenin recruitment upon BIO treatment (Fig. 5D, Fig. S5E,F). These data suggest that, in the context of

Wnt-repressed genes, elevated nuclear β-catenin levels (from BIO) do not necessarily correlate with increased chromatin recruitment.

We next performed p300 ChIP-seq in control and BIO treated FG and HG tissue to determine how recruitment of the HAT co-activator complex correlated with changes in β-catenin binding (Fig. 5E, Table S7). Examination of individual genes such as cdx2

(Fig. 5C), as well as average tag density analysis confirmed that p300 was recruited to

β-catenin peaks of HG Wnt-activated genes upon BIO treatment (Fig. 5E, Fig. S5E). For both classes of Wnt-repressed FG genes, with either increased β-catenin (e.g. sfrp2) or reduced β-catenin (e.g. nkx2-3) recruitment upon BIO, we observed a trend of reduced p300 recruitment to β-catenin peaks, consistent with repression (Fig. 5C), although this was not statistically significant (Fig. 5E, Fig. S5F). To further investigate how β-catenin recruitment might activate some genes and repress others, we performed motif enrichment analysis on different classes of peaks. Tcf was the most enriched motif in

HG Wnt-activated peaks. Surprisingly, Tcf DNA-binding sites were not enriched in β-

catenin peaks from FG Wnt-repressed genes; rather these were enriched for Gata, Sox and TEAD motifs (Fig. 5F), suggesting that other DNA-binding proteins might form co- repressor complexes with β-catenin.

Fig. 5: β-catenin chromatin binding to Wnt-regulated genes. (A) β-catenin ChIP-seq of FG and HG explants identified 16303 peaks within +/- 20kb of 11007 genes. The Venn diagram shows the overlap between 11007 β-catenin-bound genes and 2991 Wnt-regulated genes (from Fig. 4B) grouped as FG-enriched (n=524), HG-enriched (n=304) or expressed at similar levels in the FG and HG (FG=HG; n=2163) based on RNA-seq data. BMP-activated genes in red and BMP-repressed in blue. *Statistically enriched based on hypergeometric test (p<0.05). (B) Overlap

between 180 HG Wnt-activated genes (a), 208 FG Wnt-repressed genes (b), and genes with gain or loss of β-catenin ChIP enrichment upon BIO treatment. (C) Genome browser view of β-catenin peaks on HG-activated gene cdx2 and FG-repressed genes sfrp2 and nkx2-3. Red boxes indicate β-catenin significant peaks. (D-E) Average tag density of β-catenin (D) and p300 (E) peaks on HG Wnt-activated genes (a’), and two classes of FG Wnt-repressed genes (b’ and b”) comparing control to BIO treatment. *p<0.05, Wilcoxon test. (F) Motif enrichment analysis of β-catenin peaks.

Smad1 and β-catenin converge on common CRMs to coordinate transcription

We next investigated the extent to which BMP and Wnt cooperate to regulate FG and

HG transcription. A comparison of BMP-regulated and Smad1-bound genes with Wnt- regulated and β-catenin-bound genes revealed 229 coordinately regulated genes (Fig.

6A, Fig. S6A). Of these, 33 genes had a total 111 non-overlapping Smad1 and β- catenin peaks suggesting regulation through separate CRMs, while the other 196 genes had a total of 369 overlapping β-catenin and Smad1 peaks suggesting co-occupied

CRMs (Fig. 6B). This included 21 genes activated by both pathways, primarily regulatory factors known to confer HG fate like cdx2, ventx2 and msx1 (Fig. 6C,D). 62 genes with overlapping β-catenin/Smad1 peaks were Wnt-activated but BMP- repressed, many of which had a paraxial and intermediate mesoderm signature, whereas 55 genes were repressed by both Wnt and BMP including FG genes irx1-3, sfrp2 and tbx1. Finally, a group of 56 genes were BMP-activated and Wnt-repressed, including the ventral FG-endo gene hhex, cardiac genes such as nkx2-3 as well as bmp4, bmp7 and the BMP-target gene szl (Fig. 6C,D, Table S8). Although comparable

FG and HG datasets are not available in mammals, examination of public ChIP-seq data from human ES and transformed cell lines (Estaras et al., 2015; Tsankov et al.,

2015; Watanabe et al., 2014) revealed that ~47% (92/194) of the human orthologs have

β-CATENIN and/or SMAD1 peaks in regions of the genome syntenic to the co-occupied peaks in Xenopus (Fig. S7, Table S9) suggesting considerable conservation.

Motif analysis of these distinct classes of co-regulated genes revealed differential enrichment for Tcf, Smad, Gata, Lhx and/or Foxa sites (Fig. 6E) suggesting that different transcription factors probably influence whether β-catenin and Smad1

recruitment to chromatin results in transcriptional activation or repression. As a proof-of- principle that co-bound CRMs are coordinately regulated by Wnt and BMP, we tested the hhex distal region (lacking a Tcf-motif) and a cdx2 intronic region (that contains a

Tcf-motif) in embryo injection luciferase reporter assays (Fig. 6F). The cdx2:luc reporter was more active in the HG than the FG as expected, whereas the hhex:luc was active in both the FG and the HG, indicating that this CRM alone was not sufficient to confer FG restricted expression. As predicted, both hhex:luc and cdx2:luc were repressed upon

DMH1 treatment (Fig. 6F), consistent with both genes being BMP-activated. On the other hand, BIO suppressed the hhex:luc construct but stimulated cdx2:luc activity consistent with Wnt repressing the FG gene hhex and activating the HG gene cdx2 (Fig.

6F).These data indicate that binding of Smad1 and β-catenin to the same CRMs coordinate Wnt and BMP responsive gene expression to pattern the FG and HG progenitors.

The observation that the bmp4 and bmp7 loci are repressed by Wnt and activated by BMP, and that they both had CRMs with overlapping β-catenin and Smad1 peaks suggested an additional layer of Wnt-BMP signaling crosstalk (Fig. S6B). In situ hybridization confirmed that Wnt negatively regulates expression of bmp4/7 as well as bmp2 in the FG, and to a lesser extent in the HG. In contrast BMP signaling was required to maintain robust bmp2/4/7 expression in both the FG and HG (Fig. S6C,D).

Interestingly, in the gastrula (when the BIO and DMH1 treatment is started) bmp2 is expressed in the anterior mesendoderm (future FG), a low Wnt environment, whereas bmp4/7 are expressed in the ventral/posterior mesendoderm (future HG) a high Wnt environment (Hoppler and Moon, 1998). Taken together this suggests that low Wnt is

required to maintain bmp2 in the presumptive FG, which then initiates a known positive

BMP feedback loop to promote bmp4/7 expression in the FG (Karaulanov et al., 2004;

Kenny et al., 2012) (Fig. 7A).

Overall these data reveal a complex BMP-Wnt gene regulatory network that coordinates A-P and D-V patterning of the FG and HG progenitors (Fig. 7A), with β- catenin and Smad1 converging on CRMS to regulate both transcriptional activation and repression (Fig. 7B, Table S10).

Fig. 6: Smad1 and β-catenin converge on common CRMs. (A) Overlap of BMP-regulated/Smad1-bound genes with Wnt-regulated/β-catenin bound genes. (B) Schematic of 33 genes with distinct β-catenin and Smad1 peaks and 196 genes with overlapping β-catenin and Smad1 peaks. Right panel shows Smad1 and β- catenin read density in the corresponding peaks. FG genes (yellow) and HG genes (green). (WE = Whole Embryo). (C) Venn shows 196 genes with overlapping Smad1 and β-catenin peaks categorized based on activation (act) or repression (rep) by BMP and Wnt signaling. (D) Browser view of β-catenin and Smad1 peaks on hhex and cdx2 with illustration of luciferase constructs. Red boxes indicate overlapping Smad1 and β- catenin peaks. (E) Motif enrichment analysis of gene sets based on BMP/Wnt regulation. (F) Luciferase assays of reporter constructs with CRMs depicted in panel D. FG cells were injected in the C1 blastomere and HG cells in the C4 blastomere at 32- cell stage embryos. Error bars represent standard deviation of three biological replicates, * p<0.05 in student T-test. 66

Fig. 7: BMP/Smad1 and Wnt/β-catenin converge on the same FG and HG CRMs. (A) A model of how spatially restricted Wnt and BMP activity coordinate A-P and D-V patterning of FG and HG progenitors. A signaling crosstalk in the FG with low Wnt promoting BMP ligand expression in the pre-cardiac mesoderm. (B) Schematic of how overlapping Smad1 and β-catenin peaks regulate FG and HG transcription.

Discussion

In embryonic development and hPSC differentiation, combinatorial growth factor signaling controls cell fate decisions, but how these signals coordinate gene expression at the genomic level is poorly understood. We investigated how BMP and Wnt signaling, differentially active along orthologous embryonic axes, are integrated in the genome to coordinate the FG and HG transcriptomes (Fig. 7A). BMP/Smad1 activity, high in the ventral and low in the dorsal gut tube, regulates D-V identity in the FG, and promotes expression of key HG genes. On the other hand, low Wnt/β-catenin activity is required for FG fate, while high Wnt induces HG and represses FG transcription. We defined

Smad1- and β-catenin-bound CRMs of many Wnt- and BMP-target genes, including lineage specifying transcription factors. Our data suggest a much more complicated regulatory landscape than previously appreciated with Smad1 or β-catenin binding correlated with either activation or repression, and combinatorial Wnt and BMP signaling converging on hundreds of common CRMs to regulate FG and HG specific responses (Fig. 7, Table S10). While there are many aspects of the model to be resolved with detailed cis-regulatory analysis, this provides a framework for understanding how β-catenin and Smad1 are integrated in the genome to control early gut patterning.

We found that combinatorial Wnt and BMP signaling coordinates A-P and D-V patterning in the Xenopus embryo in a manner, almost identically to recent hPSC differentiation protocols where BMPlow/Wntlow specify general FG-endo, but

BMPhigh/Wntlow promote ventral FG lineages (Green et al., 2011; Loh et al., 2014).

BMPhigh/Wntlow also promotes cardiac, myeloid and endothelial fates, whilst repressing 68

intermediate and paraxial mesoderm fates, which require BMPlow/Wnthigh (Loh et al.,

2016). Finally, BMPhigh/Wnthigh promotes intestinal fate (Spence et al., 2011).

Our ChIP-seq results suggest a model for how this differential Wnt and BMP signaling regulates FG and HG transcription (Fig. 7B), and identified hundreds of putative β-catenin and Smad1-bound CRMs associated with known and novel target genes. In a comparison to recent β-catenin ChIP-seq of Xenopus gastrula, we found that 81% (695/849) of X. laevis (Kjolby and Harland, 2016) and 43% (1970/4529) of X. tropicalis (Nakamura et al., 2016) genes associated with β-catenin occupancy at the gastrula were also in our dataset. These included HG homeobox genes such as cdx2, ventx1, hoxd1 and msx1, consistent with Wnt promoting the expression of these posterior genes starting at gastrulation. We also identified several hundred genes co- regulated by Wnt and BMP that exhibit overlapping β-catenin and Smad1 peaks, many of which appear to be conserved in humans based on public ChIP-seq data from hPSCs and transformed cell lines (Benahmed et al., 2008; Gaunt et al., 2003; Tsankov et al.,

2015; Watanabe et al., 2014), suggesting a conserved paradigm.

Most BMP-activated genes in the FG were not associated with Smad1-binding, suggesting indirect regulation. Despite this, a small cohort of genes encoding lineage- specifying transcription factors such as, nkx2-3, nkx2-5, hand2, spib, gata5, and hhex were associated with Smad1-bound CRMs. In the case of Nkx2-5 this Smad1-bound

CRM is conserved in mammals (Lien et al., 2002), and public ChIP-seq data from hPCS indicates that SMAD1-binding can also occur at the human HHEX loci (Tsankov et al.,

2015) (Fig. S7). This suggests that in the FG, BMP/Smad1 initiates a cascade of transcription factors that promote cardiac, myeloid and hepatic fates. Gata motifs were

enriched in our Smad1 peaks, suggesting they might be key components of this regulatory cascade. Gata factors are known to be required for cardiac and foregut development and can physically interact with Smad1 to regulate target gene transcription (Benchabane and Wrana, 2003; Brown et al., 2004; Haworth et al., 2008;

Rossi et al., 2001; Trompouki et al., 2011; Xuan et al., 2012).

For most BMP- and Wnt-activated genes, our data support the canonical model of Smad1 and β-catenin associating with DNA-binding proteins (Tcf in the case of β- catenin) to recruit HAT co-activator complex and stimulate transcription. Unexpectedly about 55% of BMP-repressed and 73% of Wnt-repressed FG genes associated with

Smad1 and β-catenin peaks, respectively, suggesting direct repression. Motif analysis indicated that these peaks were not enriched in consensus Tcf or Smad DNA-binding sites; Wnt-repressed genes were enriched for Gata, Sox and TEAD motifs, whilst BMP- repressed genes were enriched for Gata, Lhx and Nkx. Transcriptional repression by

Smad or β-catenin has only been documented in a handful of cases. For example,

Smad1 can bind to Nkx2-3 on DNA and recruit Sin3/HDAC1 co-repressors to inhibit reporter construct expression in response to BMP signaling (reviewed in Blitz and Cho,

2009; Kim and Lassar, 2003; Marty et al., 2000), consistent with Nkx motifs in many of our Smad1 peaks (Fig. S3E). In mammalian and Drosophila cells, β-catenin can recruit co-repressor complexes to repress e-cadherin and dpp (bmp) transcription respectively

(Jamora et al., 2003; Olson et al., 2006; Theisen et al., 2007), consistent with Wnt repression of bmp4/7 that we observed in the FG. In recent years, a number of transcription factors have been shown to interact with β-catenin in different cellular contexts, including Sox, homeobox and TEAD (Estaras et al., 2015; reviewed in

Kormish et al., 2010), suggesting that different DNA-binding proteins determine whether

Smad1 or β-catenin recruit co-activator or co-repressor complexes.

Interestingly most Wnt-repressed FG genes had β-catenin peaks in both FG and

HG tissues. One possibility is that the FG explants contained some HG cells, consistent with low levels of cdx2 mRNA and expression of the cdx2:luc reporter in the FG.

Alternatively, β-catenin activity levels might impact a switch between activation and repression, since we have previously shown that low levels of Wnt/Fzd7 are required for hhex expression, while high levels are inhibitory (Zhang et al., 2013). However, unlike

Wnt-activated genes, BIO treatment (which inhibits GSK3β and stabilizes β-catenin) did not strictly correlate with increased β-catenin recruitment to Wnt-repressed FG genes.

Since GSK3β, has other substrates besides β-catenin (Ding et al., 2000; Wu and Pan,

2010), is it possible that some of the BIO regulated gene expression is Wnt- independent.

Another striking observation was that the majority of Smad1 and β-catenin peaks were not associated with BMP or Wnt regulated transcription. This is similar to recent findings in Xenopus gastrula and hPSCs (Nakamura et al., 2016; Tsankov et al., 2015), and is consistent with an emerging concept that transcription factor binding is pervasive throughout the genome even when they are not engaged in productive transcription

(Nakamura et al., 2016; reviewed in Skalska et al., 2015). One possibility is that these

β-catenin/Smad1 binding events are due to earlier Wnt/BMP signaling. Indeed ~40% of our stage NF20 β-catenin-bound genes were also reported in β-catenin ChIP-seq from

Xenopus tropicalis gastrula (NF10.5) (Nakamura et al., 2016). This is consistent with the idea that β-catenin and/or Smad binding can prime genes for future activation, perhaps

by modulating epigenetic poising and/or interacting with pioneering factors, which are important for gut tube lineages (Blythe et al., 2010; Loh et al., 2014; Tsankov et al.,

2015; Wang et al., 2015)

In summary this study has advanced our understanding of how BMP/Smad1 and

Wnt/β-catenin signaling are integrated in the genome to regulate FG and HG transcriptional programs, which should inform hPSC differentiation mechanisms and how Wnt and BMP interact in other development and disease contexts.

Materials and methods

Embryo experiments and manipulations

Animal experiments were performed according to CCHMC IACUC approved protocols.

Xenopus laevis embryos were staged according to Nieuwkoop and Faber (Nieuwkoop and Faber, 1967). Injections, small molecule treatments, luciferase assays (see supplementary Materials and Methods), in situ hybridizations and immunostaining were performed as previously described (McLin et al., 2007). Protein injections into the closing blastocoel of the foregut were performed at stage NF12, with either 40 nl of recombinant human BMP2 (5.8 µM; R&D Systems) in PBS + 0.1%BSA or PBS +

0.1%BSA as control. Embryos were cultured from stages NF12-20 with either DMSO vehicle in 0.1XMBS or DMH1 (40 µM; TOCRIS) or BIO (60 µM; TOCRIS) dissolved in

DMSO. Stage NF12 transgenic Xenopus laevis Xla.Tg(hgem:Xtr.dkk1)Jmws, referred to as Tg(hsp70:dkk1) (Lin and Slack, 2008), were heatshocked at 370C for 30 minutes followed by incubation at 130C to NF20. For immunofluorescence we used anti-

phospho-Smad1/5/8 (1:300, Cell Signaling Technology, 13820S) and anti-β-catenin

(1:300, Santa Cruz Biotechnologies, sc-7199).

Genomic analysis

RNA/DNA-seq analyses were performed using the Xenopus laevis genome v9.1,

(Session et al., 2016). Since Xenopus laevis is allotetraploid most genes contain two copies, designated as .L or .S (eg. cdx2.L and cdx2.S). For simplicity we dropped the .L or .S in most figures, however this is reported in all the supplementary gene lists.

For each RNA-seq sample, 50 explants were microdissected and when necessary cultured in 10 µg/ml dispase for 15-20 minutes to separate endo and meso.

Total RNA was extracted from two or three independent biological replicates and libraries were sequenced with ~7-10 million reads/library with 75 bp length. Quality trimmed reads were mapped to the X. laevis genome 9.1, quantified using RSEM and mapped with bowtie2 using default thresholds (Li and Dewey, 2011). Differential gene expression analysis was carried out with RUVSeq (Risso et al., 2014) with log2FC ≥1 or

≤-1, p <0.05 and FDR ≤5%.

ChIP was carried out as previously described (Akkers et al., 2012; Blythe et al.,

2009) with 25-50 whole embryos or 100 FG or HG explants using the following antibodies: anti-Smad1 (Invitrogen, 38-5400), anti-β-catenin (Life technologies, 712700) and anti-p300 (Santa Cruz sc-585 X). Libraries were sequenced with ~30 million reads/library. Reads were mapped to the X. laevis genome assembly v9.1 using

Bowtie2 at default thresholds (Langmead and Salzberg, 2012). ChIP-seq peaks were called using MACS2 at default thresholds (Zhang et al., 2008). Irreproducibility

Discovery Rate (IDR) was performed with standard thresholds (Li et al., 2011). Publicly available data for SMAD1 (GSM1505734), and β-CATENIN (GSM1579346 and

GSM1303695) (Estaras et al., 2015; Tsankov et al., 2015; Watanabe et al., 2014) were processed with bowtie and MACS2 using BioWardrobe Toolkit at default thresholds

(Kartashov and Barski, 2015). For further details of RNA-seq analysis, ChIP and ChIP- seq analysis, see the supplementary Materials and Methods.

Acknowledgements

We specially thank Dr. Ken Cho and Jin Cho for the ChIP-seq training, Dr. Gert

Veenstra’s lab for ChIP-seq advice and Dr. Matt Weirauch for valuable discussions. We are grateful to the Zorn and Wells labs and Endoderm Club for helpful discussions. We thank Brian Gebelein and Malcolm Fisher for comments on the manuscript.

Authors contribution

MLS and AMZ designed the study, interpreted the data, and wrote the manuscript. MLS,

SAR, MM and AMZ performed Xenopus experiments. SJ, MY and AB prepared the

ChIP-seq libraries. PC and MLS performed the bioinformatics analysis.

Funding

This work was funded by NIH DK070858 to AMZ. MLS was supported by NIH

T32HL007752-2, and a University of Cincinnati Research Council Summer fellowship. 74

NGS library construction and sequencing was supported by NIGMS DP2 GM119134 and NCATS 5UL1TR001425-02 to AB. This project was supported in part by NIH P30

DK078392.

Data availability

All RNA-seq and ChIP-seq datasets from this study are available in NCBI Gene

Expression Omnibus (GEO) under accession number GSE87654.

Competing interests

A.B. is co-founders of Datirium, LLC. Which provides software development and bioinformatics support services, including installation of BioWardrobe.

References

Akkers, R. C., Jacobi, U. G. and Veenstra, G. J. (2012). Chromatin immunoprecipitation analysis of Xenopus embryos. Methods Mol Biol 917, 279- 292. Benahmed, F., Gross, I., Gaunt, S. J., Beck, F., Jehan, F., Domon-Dell, C., Martin, E., Kedinger, M., Freund, J. N. and Duluc, I. (2008). Multiple regulatory regions control the complex expression pattern of the mouse Cdx2 homeobox gene. Gastroenterology 135, 1238-1247. Benchabane, H. and Wrana, J. L. (2003). GATA- and Smad1-dependent enhancers in the Smad7 gene differentially interpret bone morphogenetic protein concentrations. Mol Cell Biol 23, 6646-6661. Blitz, I. L. and Cho, K. W. (2009). Finding partners: how BMPs select their targets. Dev Dyn 238, 1321-1331.

Blythe, S. A., Cha, S. W., Tadjuidje, E., Heasman, J. and Klein, P. S. (2010). beta- Catenin primes organizer gene expression by recruiting a histone H3 arginine 8 methyltransferase, Prmt2. Dev Cell 19, 220-231. Blythe, S. A., Reid, C. D., Kessler, D. S. and Klein, P. S. (2009). Chromatin immunoprecipitation in early Xenopus laevis embryos. Dev Dyn 238, 1422-1432. Brown, C. O., 3rd, Chi, X., Garcia-Gras, E., Shirai, M., Feng, X. H. and Schwartz, R. J. (2004). The cardiac determination factor, Nkx2-5, is activated by mutual cofactors GATA-4 and Smad1/4 via a novel upstream enhancer. J Biol Chem 279, 10659-10669. Cadigan, K. M. and Waterman, M. L. (2012). TCF/LEFs and Wnt signaling in the nucleus. Cold Spring Harb Perspect Biol 4, 1-22. Chalmers, A. D. and Slack, J. M. (2000). The Xenopus tadpole gut: fate maps and morphogenetic movements. Development 127, 381-392. De Robertis, E. M. (2009). Spemann's organizer and the self-regulation of embryonic fields. Mech Dev 126, 925-941. Ding, V. W., Chen, R. H. and McCormick, F. (2000). Differential regulation of glycogen synthase kinase 3beta by insulin and Wnt signaling. J Biol Chem 275, 32475- 32481. Estaras, C., Benner, C. and Jones, K. A. (2015). SMADs and YAP compete to control elongation of beta-catenin:LEF-1-recruited RNAPII during hESC differentiation. Mol Cell 58, 780-793. Gaarenstroom, T. and Hill, C. S. (2014). TGF-beta signaling to chromatin: how Smads regulate transcription during self-renewal and differentiation. Semin Cell Dev Biol 32, 107-118. Gaunt, S. J., Drage, D. and Cockley, A. (2003). Vertebrate caudal gene expression gradients investigated by use of chick cdx-A/lacZ and mouse cdx-1/lacZ reporters in transgenic mouse embryos: evidence for an intron enhancer. Mech Dev 120, 573-586. Gibb, N., Lavery, D. L. and Hoppler, S. (2013). sfrp1 promotes cardiomyocyte differentiation in Xenopus via negative-feedback regulation of Wnt signalling. Development 140, 1537-1549. Green, M. D., Chen, A., Nostro, M. C., d'Souza, S. L., Schaniel, C., Lemischka, I. R., Gouon-Evans, V., Keller, G. and Snoeck, H. W. (2011). Generation of anterior foregut endoderm from human embryonic and induced pluripotent stem cells. Nat Biotechnol 29, 267-272. Gupta, R., Wills, A., Ucar, D. and Baker, J. (2014). Developmental enhancers are marked independently of zygotic Nodal signals in Xenopus. Dev Biol 395, 38-49. Haworth, K. E., Kotecha, S., Mohun, T. J. and Latinkic, B. V. (2008). GATA4 and GATA5 are essential for heart and liver development in Xenopus embryos. BMC Dev Biol 8, 74.

Hoppler, S. and Moon, R. T. (1998). BMP-2/-4 and Wnt-8 cooperatively pattern the Xenopus mesoderm. Mech Dev 71, 119-129. Itasaki, N. and Hoppler, S. (2010). Crosstalk between Wnt and bone morphogenic protein signaling: a turbulent relationship. Dev Dyn 239, 16-33. Jamora, C., DasGupta, R., Kocieniewski, P. and Fuchs, E. (2003). Links between signal transduction, transcription and adhesion in epithelial bud development. Nature 422, 317-322. Karaulanov, E., Knochel, W. and Niehrs, C. (2004). Transcriptional regulation of BMP4 synexpression in transgenic Xenopus. EMBO J 23, 844-856. Kartashov, A. V. and Barski, A. (2015). BioWardrobe: an integrated platform for analysis of epigenomics and transcriptomics data. Genome Biol 16, 158. Kenny, A. P., Rankin, S. A., Allbee, A. W., Prewitt, A. R., Zhang, Z., Tabangin, M. E., Shifley, E. T., Louza, M. P. and Zorn, A. M. (2012). Sizzled-tolloid interactions maintain foregut progenitors by regulating fibronectin-dependent BMP signaling. Dev Cell 23, 292-304. Kim, D. W. and Lassar, A. B. (2003). Smad-dependent recruitment of a histone deacetylase/Sin3A complex modulates the bone morphogenetic protein- dependent transcriptional repressor activity of Nkx3.2. Mol Cell Biol 23, 8704- 8717. Kim, S. W., Yoon, S. J., Chuong, E., Oyolu, C., Wills, A. E., Gupta, R. and Baker, J. (2011). Chromatin and transcriptional signatures for Nodal signaling during endoderm formation in hESCs. Dev Biol 357, 492-504. Kjolby, R. A. and Harland, R. M. (2016). Genome-wide identification of Wnt/beta- catenin transcriptional targets during Xenopus gastrulation. Dev Biol. Klaus, A. and Birchmeier, W. (2009). Developmental signaling in myocardial progenitor cells: a comprehensive view of Bmp- and Wnt/beta-catenin signaling. Pediatr Cardiol 30, 609-616. Kormish, J. D., Sinner, D. and Zorn, A. M. (2010). Interactions between SOX factors and Wnt/beta-catenin signaling in development and disease. Dev Dyn 239, 56- 68. Lancaster, M. A. and Knoblich, J. A. (2014). Organogenesis in a dish: modeling development and disease using organoid technologies. Science 345, 1247125. Langmead, B. and Salzberg, S. L. (2012). Fast gapped-read alignment with Bowtie 2. Nat Methods 9, 357-359. Li, B. and Dewey, C. N. (2011). RSEM: accurate transcript quantification from RNA- Seq data with or without a reference genome. BMC Bioinformatics 12, 323. Li, Q., Brown, J. B., Huang, H. and Bickel, P. J. (2011). Measuring reproducibility of high-throughput experiments. 1752-1779.

Li, Y., Rankin, S. A., Sinner, D., Kenny, A. P., Krieg, P. A. and Zorn, A. M. (2008). Sfrp5 coordinates foregut specification and morphogenesis by antagonizing both canonical and noncanonical Wnt11 signaling. Genes Dev 22, 3050-3063. Lien, C. L., McAnally, J., Richardson, J. A. and Olson, E. N. (2002). Cardiac-specific activity of an Nkx2-5 enhancer requires an evolutionarily conserved Smad binding site. Dev Biol 244, 257-266. Lin, G. and Slack, J. M. (2008). Requirement for Wnt and FGF signaling in Xenopus tadpole tail regeneration. Dev Biol 316, 323-335. Loh, K. M., Ang, L. T., Zhang, J., Kumar, V., Ang, J., Auyeong, J. Q., Lee, K. L., Choo, S. H., Lim, C. Y., Nichane, M., et al. (2014). Efficient endoderm induction from human pluripotent stem cells by logically directing signals controlling lineage bifurcations. Cell Stem Cell 14, 237-252. Loh, K. M., Chen, A., Koh, P. W., Deng, T. Z., Sinha, R., Tsai, J. M., Barkal, A. A., Shen, K. Y., Jain, R., Morganti, R. M., et al. (2016). Mapping the Pairwise Choices Leading from Pluripotency to Human Bone, Heart, and Other Mesoderm Cell Types. Cell 166, 451-467. Marty, T., Muller, B., Basler, K. and Affolter, M. (2000). Schnurri mediates Dpp- dependent repression of brinker transcription. Nat Cell Biol 2, 745-749. McLin, V. A., Rankin, S. A. and Zorn, A. M. (2007). Repression of Wnt/beta-catenin signaling in the anterior endoderm is essential for liver and pancreas development. Development 134, 2207-2217. Nakamura, Y., de Paiva Alves, E., Veenstra, G. J. and Hoppler, S. (2016). Tissue- and stage-specific Wnt target gene expression is controlled subsequent to beta- catenin recruitment to cis-regulatory modules. Development 143, 1914-1925. Nieuwkoop, P. D. and Faber, J. (1967). Normal Table of Xenopus laevis (Daudin): A Systematical and Chronological Survey of the Development from the Fertilized Egg to the End of Metamorphosis.: New York: Garland Publishing. Olson, L. E., Tollkuhn, J., Scafoglio, C., Krones, A., Zhang, J., Ohgi, K. A., Wu, W., Taketo, M. M., Kemler, R., Grosschedl, R., et al. (2006). Homeodomain- mediated beta-catenin-dependent switching events dictate cell-lineage determination. Cell 125, 593-605. Rankin, S. A., Kormish, J., Kofron, M., Jegga, A. and Zorn, A. M. (2011). A gene regulatory network controlling hhex transcription in the anterior endoderm of the organizer. Dev Biol 351, 297-310. Risso, D., Ngai, J., Speed, T. P. and Dudoit, S. (2014). Normalization of RNA-seq data using factor analysis of control genes or samples. Nat Biotechnol 32, 896- 902. Rossi, J. M., Dunn, N. R., Hogan, B. L. and Zaret, K. S. (2001). Distinct mesodermal signals, including BMPs from the septum transversum mesenchyme, are required in combination for hepatogenesis from the endoderm. Genes Dev 15, 1998-2009.

Schohl, A. and Fagotto, F. (2002). Beta-catenin, MAPK and Smad signaling during early Xenopus development. Development 129, 37-52. Session, A. M., Uno, Y., Kwon, T., Chapman, J. A., Toyoda, A., Takahashi, S., Fukui, A., Hikosaka, A., Suzuki, A., Kondo, M., et al. (2016). Genome evolution in the allotetraploid frog Xenopus laevis. Nature 538, 336-343. Sherwood, R. I., Maehr, R., Mazzoni, E. O. and Melton, D. A. (2011). Wnt signaling specifies and patterns intestinal endoderm. Mech Dev 128, 387-400. Skalska, L., Stojnic, R., Li, J., Fischer, B., Cerda-Moya, G., Sakai, H., Tajbakhsh, S., Russell, S., Adryan, B. and Bray, S. J. (2015). Chromatin signatures at Notch-regulated enhancers reveal large-scale changes in H3K56ac upon activation. EMBO J 34, 1889-1904. Spence, J. R., Mayhew, C. N., Rankin, S. A., Kuhar, M. F., Vallance, J. E., Tolle, K., Hoskins, E. E., Kalinichenko, V. V., Wells, S. I., Zorn, A. M., et al. (2011). Directed differentiation of human pluripotent stem cells into intestinal tissue in vitro. Nature 470, 105-109. Theisen, H., Syed, A., Nguyen, B. T., Lukacsovich, T., Purcell, J., Srivastava, G. P., Iron, D., Gaudenz, K., Nie, Q., Wan, F. Y., et al. (2007). Wingless directly represses DPP morphogen expression via an armadillo/TCF/Brinker complex. PLoS One 2, e142. Trompouki, E., Bowman, T. V., Lawton, L. N., Fan, Z. P., Wu, D. C., DiBiase, A., Martin, C. S., Cech, J. N., Sessa, A. K., Leblanc, J. L., et al. (2011). Lineage regulators direct BMP and Wnt pathways to cell-specific programs during differentiation and regeneration. Cell 147, 577-589. Tsankov, A. M., Gu, H., Akopian, V., Ziller, M. J., Donaghey, J., Amit, I., Gnirke, A. and Meissner, A. (2015). Transcription factor binding dynamics during human ES cell differentiation. Nature 518, 344-349. Wang, A., Yue, F., Li, Y., Xie, R., Harper, T., Patel, N. A., Muth, K., Palmer, J., Qiu, Y., Wang, J., et al. (2015). Epigenetic priming of enhancers predicts developmental competence of hESC-derived endodermal lineage intermediates. Cell Stem Cell 16, 386-399. Watanabe, K., Biesinger, J., Salmans, M. L., Roberts, B. S., Arthur, W. T., Cleary, M., Andersen, B., Xie, X. and Dai, X. (2014). Integrative ChIP-seq/microarray analysis identifies a CTNNB1 target signature enriched in intestinal stem cells and colon cancer. PLoS One 9, e92317. Wu, D. and Pan, W. (2010). GSK3: a multifaceted kinase in Wnt signaling. Trends Biochem Sci 35, 161-168. Xuan, S., Borok, M. J., Decker, K. J., Battle, M. A., Duncan, S. A., Hale, M. A., Macdonald, R. J. and Sussel, L. (2012). Pancreas-specific deletion of mouse Gata4 and Gata6 causes pancreatic agenesis. J Clin Invest 122, 3516-3528.

Zhang, Y., Liu, T., Meyer, C. A., Eeckhoute, J., Johnson, D. S., Bernstein, B. E., Nusbaum, C., Myers, R. M., Brown, M., Li, W., et al. (2008). Model-based analysis of ChIP-Seq (MACS). Genome Biol 9, R137. Zhang, Z., Rankin, S. A. and Zorn, A. M. (2013). Different thresholds of Wnt-Frizzled 7 signaling coordinate proliferation, morphogenesis and fate of endoderm progenitor cells. Dev Biol 378, 1-12. Zorn, A. M. and Wells, J. M. (2009). Vertebrate endoderm development and organ formation. Annu Rev Cell Dev Biol 25, 221-251.

Supplementary figures

Fig. S1: Transcriptional program of FG and HG progenitors in vivo. (A) Scatter plot of log fold change (log2FC) in expression between HG versus FG samples and endo versus meso samples. HG-enriched transcripts (green), FG-enriched transcripts (orange) meso-enriched transcripts (red) and endo-enriched transcripts (yellow) based on log2 fold change (FC) ≤-1 or ≥1 and false discovery rate (FDR) ≤5%. (B) Xenopus orthologs of genes known to be involved in human and mouse GI development (manually curated list from the literature) are present in our FG-enriched and HG-enriched gene lists. The heatmap shows that the Xenopus transcripts have restricted expression in manner predicted from the mouse and human literature, illustrating high conservation across species. (C) GO term enrichment analysis of 172 FG-endo, 294 FG-meso, 518 HG-endo and 202 HG-meso genes from Fig. 1C. (D) BMP and Wnt pathway components that are expressed in any sample (FG-endo, FG- meso, HG-endo or HG-meso) above one transcripts per-million reads (TPM >1; lower than this is considered not expressed). The heatmap shows that BMP pathway genes are expressed in both the FG and HG, whereas Wnt ligands are generally restricted to the HG and Wnt-antagonists enriched in the FG. (E) In situ hybridization of mid-sagittal section stage NF20 (hhex and ventx2.1) or NF35 embryos (nr1h5, nkx2-1, nkx2-5, sox2 and darmin) in DMH1 or Tg(hsp70:dkk1) embryos; anterior left and dorsal up.

Fig. S2: RNA-seq of DMSO and DMH1 treated embryos identified BMP-regulated genes. (A) Scatter plot showing log2FC in expression between DMSO and DMH1 treated FG (left) and between DMSO and DMH1 treated HG (right) samples. Transcripts are colored based on the normal control expression; HG-enriched in green, FG-enriched in orange, genes expressed similarly in FG and HG (FG=HG) in grey and normally not expressed in control FG or HG in black. (B) Venn diagram illustrates overlap between transcripts up regulated (é) or down regulated (ê) upon DMH1 treatment in FG and HG tissues. FG or HG transcripts repressed upon DMH1 treatment (log2FC ≤-1, FDR ≤5% 83

relative to controls) are considered to be BMP-activated genes (n=697), whereas FG or HG transcripts that are increased upon DMH1 (log2FC ≥1, FDR ≤5%) are classified as BMP-repressed genes (n=1063). Eight transcripts had ambiguous regulation being both activated and repressed by DMH1 in FG or HG tissues, and were excluded from further analysis. Overall we categorized a total of 1760 (697+1063) BMP-regulated genes in the FG and HG tissue. (C) GO term analysis of BMP-activated and -repressed genes. (D) Unsupervised clustering of BMP-regulated genes in control (CONT) and DMH1 treated FG and HG samples compared to dorsal explants (DORS), which contain a thin layer of dorsal endoderm (yellow) along with neural and somite tissue. The DMH1- treated FG showed similarities to the dorsal tissue suggesting that BMP induces ventral mesendoderm fate and represses dorsal fate. (E) In situ hybridization of control, DMH1 treated or BMP2 injected embryos, in a cross section confirms that expression of the dorsal mesoderm gene foxc2 is expanded ventrally with DMH1 and restricted dorsally upon BMP injection. White line indicates expression domain.

Fig. S3: Smad1 ChIP-seq of embryos stage NF20. (A) Genomic distribution of Smad1 ChIP-seq peaks in stage NF20 embryos categorized as upstream (-20kb), downstream (+20kb), intragenic and promoter (-1kb to +1kb) regions. (B) Smad1 ChIP-PCR of known BMP-target genes showing reduced Smad1-binding to CRMs of msx1, id3 and ventx2.1 promoters in DMH1 treated embryos compare to DMSO controls. (C) Peak overlap between Smad1 and p300 ChIP-seq of stage NF20 whole embryos. (D) Venn showing the overlap between Smad1-bound genes, p300-bound genes and genes expressed in NF20 embryo at levels higher than 5 transcripts per million (TPM>5) based on RNA-seq. (E) Motif enrichment analysis of Smad1 ChIP-seq peaks associated with activated (act) or repressed (rep) genes. 85

Fig. S4: RNA-seq of control, Tg(hsp70:dkk1) and BIO treated embryos. (A) Venn diagram illustrates transcripts with expression increased (é) or decreased (ê) upon Tg(hsp70:dkk1) or BIO treatment in FG or HG tissues. Wnt-activated genes were log2FC ≤-1 upon heat-shock or log2FC ≥1 upon BIO treatment FDR ≤5% (n=959). Wnt- repressed genes were log2FC ≥1 upon heat-shock or log2FC ≤-1 upon BIO treatment FDR ≤5% (n=2032). Forty-one transcripts had ambiguous regulation with evidence of being both Wnt-activated and Wnt-repressed, and were excluded from further analysis. Overall we categorized a total of 2991 (959+2032) Wnt-regulated genes in the FG and HG tissue. (B) GO term enrichment analysis of Wnt-activated and Wnt–repressed genes. (C-D) Scatter plot showing log2FC in expression between FG non-heatshock and FG Tg(hsp70:dkk1) (C) and HG DMSO and HG BIO (D) explants. Transcripts are colored based on the normal control expression; HG-enriched in green, FG-enriched in orange, expressed similarly FG and HG (FG=HG) in grey and normally not expressed in FG or HG in black.

Fig. S5: b-catenin ChIP-seq of embryos stage NF20. (A) Experimental design showing β-catenin ChIP-seq of 100 FG or 100 HG explants dissected from DMSO vehicle control or BIO treated NF20 embryos. Reads were merged and MACS2 peak calling was performed followed by irreproducibility discovery rate (IDR) filtering with standard thresholds (Li et al., 2011) identified 16303 statistically significant peaks associated with 11007 genes (+/- 20 Kb from transcription start site) in the FG and HG samples. (B) Genomic distribution of β-catenin ChIP-seq peaks classified as upstream (-20kb), downstream (+20kb), intragenic and promoter (-1kb to +1kb) regions. (C) DNA-binding protein motif enrichment analysis of all β-catenin ChIP- seq peaks in the genome. (D) β-catenin ChIP-PCR of known CRMs in Wnt-target genes ventx2.1, cdx2 and sp5 from Tg(hsp70:dkk1) embryos with and without heat shock (HS). (E) Read density of different classes of β-catenin and p300 peaks in DMSO or BIO treated FG and HG explants from Fig. 5B. +/-2kb centered on the β-catenin peak summit. (F) Box plots of average tag density of β-catenin and p300 peaks on HG Wnt- activated genes (a’), and FG Wnt-repressed genes (b’ and b”) upon BIO treatment. *p<0.05, Wilcoxon test. 88

Fig. S6: BMP/Smad1 and Wnt/β-catenin crosstalk. (A) Overlap between BMP-regulated/Smad1-bound and Wnt-regulated/β-catenin-bound genes. (B) Browser view of β-catenin and Smad1 peaks in Xenopus and human genes (from GSM1505734 and GSM1579346). Red boxes indicate overlapping Smad1 and β- catenin peaks. Black boxes indicate syntenic peaks. (C) In situ hybridization of DMH1, Tg(hsp70:dkk1) and BIO treated embryos in mid-sagittal section, anterior left and dorsal up. Embryos are either stage NF12 wild-type or treated at stage NF11 and fixed at stage NF20. * indicates hhex-expressing FG cells. (D) Expression heatmap of BMP and Wnt ligands and targets present in FG and HG samples of controls (CO), DMH1, Tg(hsp70:dkk1) and BIO-treated embryos. 89

Fig.S7: Smad1 and β-catenin syntenic peaks in Xenopus laevis and Homo sapiens. Browser view of β-catenin and Smad1 peaks in Xenopus and human genes (from the following public data: GSM1505734 and GSM1579346). Red boxes indicate overlapping Smad1 and β-catenin peaks. Black boxes indicate syntenic peaks. 90

List of supplemental tables

Table S1: Transcriptional program of FG and HG progenitors. Differential expression analysis between FG and HG samples identified 906 FG- enriched and 987 HG-enriched genes. Comparison between FG+HG endo and FG+HG meso identified 3439 endo-enriched and 4829 meso-enriched genes. log2FC ≤-1 or ≥1, FDR ≤5%.

Table S2: Tables show the number of transcripts overlapping in the following pairwise differential expression analyses. Enriched transcripts have log2 fold change ≤-1 or ≥1 difference in expression and false discovery rate ≤5%. (A) Intersection of FG versus HG and endo versus meso transcripts. This table is supplementary to Fig. 1B-C. (B) Intersection of FG versus HG and BMP-activated versus BMP-repressed transcripts. This table is supplementary to Fig. 2B. (C) Intersection of FG versus HG and Wnt-activated versus Wnt-repressed. This table is supplementary to Fig. 4B.

Table S3: FG and HG transcriptome conservation among vertebrates. Manually curated list of genes expressed in FG and HG tissue from mouse embryos and direct differentiation of human stem cells. Genes expressed with tpm≥1 in at least one of the frog samples are considered present.

Table S4: BMP regulated genes from RNA-seq analysis. BMP differentially expressed genes in DMH1 treated FG (Sheet 1) or HG (Sheet 2) samples compared to DMSO control. Differentially expressed genes are identified with their gene name, log2FC, p value and FDR. Log fold change and FDR indicate those values of DMH1 experiments compared to DMSO control. Experiments were done in triplicate with log2FC ≤-1 or ≥1, FDR ≤5%.

Table S5: Smad1 and p300 peaks of whole embryos stage NF20 Smad1 (Sheet1) and p300 (Sheet2) ChIP-seq identified 7976 and 4727 peaks, respectively. The position of each Smad1 peak is indicated by “Chromosome/Scaffold”, “Peak_Start” and “Peak_Stop”, with “Summit” of the peak. The nearest genes are indicated by “gene” with the “Gene_Start” and “Gene_Stop” positions and “Distance_to_Gene_TSS_in_bp_from_Summit”. Peaks were categorized depending on where they fall related to each gene. Peaks inside the gene were categorized as ”intragenic”, peaks +1kb/-1kb of the TSS are “promoter”, peaks downstream of the gene are “proximal downstream” (+10kb) or “distal downstream” (+20kb), and peaks upstream of the gene are “proximal upstream” (-10kb) or “distal upstream” (-20kb).

Table S6: Wnt regulated genes from RNA-seq analysis. Wnt differentially expressed genes in Tg(hsp70:dkk1) or BIO treated FG (Sheet 1 and 3) or HG (Sheet 2 and 4) samples compared to non-heatshock or DMSO control. Differentially expressed genes are identified with their gene name, log2FC, p value and FDR. Log fold change and FDR indicate those values of experimental manipulation compared to control. Experiments were done in triplicate with log2FC ≤-1 or ≥1, FDR ≤5%.

Table S7: β-catenin and p300 peaks of whole embryos and FG+HG tissues stage NF20 β-catenin (sheet1) and p300 (sheet2) ChIP-seq identified 16303 and 15146 peaks MACS2 IDR, respectively. The position of each β-catenin peak is indicated by “Chromosome/Scaffold”, “Peak_Start” and “Peak_Stop”, with “Summit” of the peak. The nearest genes are indicated by “gene” with the “Gene_Start” and “Gene_Stop” positions and “Distance_to_Gene_TSS_in_bp_from_Summit”. Peaks were categorized depending on where they fall related to each gene. Peaks inside the gene were categorized as ”intragenic”, peaks +1kb/-1kb of the TSS are “promoter”, peaks downstream of the gene are “proximal downstream” (+10kb) or “distal downstream” (+20kb), and peaks upstream of the gene are “proximal upstream” (-10kb) or “distal upstream” (-20kb).

Table S8: Genes associated with Smad1 and β-catenin peaks. Overlapping peaks are considered when the overlap is of at least 1 nucleotide. Genes were separated in different lists according to BMP and Wnt regulation, based on the RNA-seq data. Activated (act), repressed (rep), FG-enriched (orange) and HG-enriched (green).

Table S9: Syntenic Smad1 and β-catenin peaks between Xenopus and human ChIP-seq data. Peaks considered syntenic have similar positions in relation to both Xenopus and human genes. Human publicly available data for SMAD1 GSM1505734 (Tsankov et al., 2015) and β-CATENIN GSM1579346 (Estaras et al., 2015) and GSM1303695 (Watanabe et al., 2014).

Table S10: Summary of FG- and HG-enriched genes indicating BMP and Wnt and association of Smad1 or β-catenin peaks Table with 906 FG-enriched and 987 HG-enriched genes and how they were affected by the different BMP and Wnt manipulations, as well as whether they were associated with Smad1 or β-catenin peak within +/-20kb. FG-enriched genes in orange with log2FC ≤1 and HG-enriched genes in green with log2FC ≥1. Endoderm-enriched genes in yellow with log2FC ≤1 and mesoderm-enriched genes in red with logFC ≥1. BMP inhibition with DMH1 in FG and HG tissues with activated genes in pink with logFC ≥1 and repressed genes in blue with logFC ≥1. Wnt activation with BIO in FG and HG 92

tissues with activated genes in pink with logFC ≥1 and repressed genes in blue with log2FC ≤1. Wnt inhibition with Tg(hsp70:dkk1) in FG and HG tissues with activated genes in pink with log2FC ≤1 and repressed genes in blue with log2FC ≥1. For simplicity tpm values are represented by average of the replicates. Smad1 and β- catenin peaks association within +/-20kb of each transcript is shown with genome coordinates. NS = non-significant log2FC (-1< log2FC >1).

Supplementary materials and methods

Luciferase reporter assay

The following CRMs were cloned into the pGL4.23 luc2 miniP vector (Promega) to generate hhex:luc and cdx2:luc luciferase constructs: hhex CRM:

TTGTCTCTGCTCCCCTTGCTCATTACCTGCCCAGTCCCTATACACACCTTGCTGCT

CACACTGAGAGGGTAGAGACAAGGAATCTTCTCCCATCTGAGCGGCGCCGA cdx2 CRM with Tcf motif in bold (based on Cis-BP TF binding tool (PWMs – LogOdds

>10) (Weirauch et al., 2014):

CGGCGGCGTTTGTTCAGTAGTGGTAATTCCAAATATCTATAGGCCTGATAACATTTT

GCCTTGTAGCTCATTGTTAGCCCCTGTGTTCTCCATTCATTGACACTGCCCAATTCT

CTCTGATCTGCCTTGTCCCCTCTCCA

One hundred picograms of hhex:luc and cdx2:luc luciferase constructs were co-injected with PRL-SV40 control Renilla vector (Promega) (25pg) into C1 (presumptive FG) or

C4 (presumptive HG) cells of 32-cell stage embryos. At stage NF12 embryos were treated with DMSO, DMH1 or BIO (as described) and 3 embryos were frozen in triplicate at stage NF20. Embryos were lysed by pipetting in 75uL of 100mM TRIS-HCl pH7.4 + 0.2% NP-40 and then 25 uL of embryo Lysate was assayed using a dual luciferase assay kit (Biotium,Inc). Luciferase activity was normalized to co-injected

TK:renilla and the mean relative activity of the triplicate samples was shown ±S.D. with pairwise student T-tests to determine significant differences in expression. Each experiment was repeated a minimum of two times and a representative result is shown.

RNA-seq analysis

For each RNA-seq sample, 50 explants were microdissected and when necessary cultured in 10 µg/ml dispase for 15-20 minutes to separate endo and meso.

Total RNA was extracted from two or three independent biological replicates with the

Nucleo-spin RNA kit (Machery-Nagel). Libraries were constructed with TruSeq Stranded mRNA Library Prep Kit and sequenced ~7-10 million reads/library with 75 bp length using Illumina HiSeq2500. FastQC reports identified adapters, over-represented sequences, low quality bases and overall low quality reads. Trimmomatic was used to clip off adapters, over-represented sequences and low quality bases. Reads were trimmed keeping minimum length as 50, thereby after trimming, read lengths ranged from 50 to 75 base pairs. Quality trimmed reads were mapped to the X. laevis genome

9.1, quantified using RSEM and mapped with bowtie2 using default thresholds (Li and

Dewey, 2011). Differential gene expression analysis was carried out using CSBB’s

[https://github.com/csbbcompbio/CSBB-v1.0] Differential Expression Module, which uses RUVSeq (Risso et al., 2014). With RUVSeq we performed two-way normalization on the count’s matrix 1) Upper Quantile and 2) Empirical gene normalization with default settings, and differential expression analysis. Pairwise comparisons create mutually exclusive lists of enriched genes with log2FC ≤-1 or ≥1, p <0.05 and FDR ≤5% differences in expression for each of the following pairwise comparison analysis:

• To define FG- and HG-enriched transcripts we merged fastq files from FG-endo and

FG-meso from the same biological replicates, as well as HG-endo and HG-meso

samples. This resulted in ~14-20 million reads for each biological replicate of FG

(endo+meso) and HG (endo+meso). Correlation analysis indicated that merging 95

fastq was very similar to sequencing intact FG and HG (with meso and endo not

separated) r2=0.93, compared to different biological replicates of intact FG or intact

HG r2=0.94-0.90, validating this approach. We then performed a differential

expression analysis comparing FG and HG. Transcripts with log2FC ≤-1 are

classified as FG-enriched (n=906) and log2FC ≥1 as HG-enriched (n=987).

• To define endo- and meso-enriched genes we merged fastq files from FG endo to

HG endo from the same biological replicates to generate an “endo’” transcriptome

as well as FG meso to HG meso samples to generate a “meso” sample, similar to

our approach described above. We then performed a differential expression

analysis comparing endo and meso. Transcripts with log2FC ≤-1 are classified as

endo-enriched (n=3439) and log2FC ≥1 as meso-enriched (n=4829).

• To identify BMP-regulated genes we compared FG DMH1 with FG DMSO as well as

HG DMH1 with HG DMSO samples. Transcripts with log2FC ≤-1 are classified as

BMP-activated genes (n=697) and log2FC ≥1 as BMP-repressed genes (n=1063).

Eight transcripts had ambiguous regulation being both activated and repressed by

DMH1 in FG or HG tissues, and were excluded from further analysis. Overall we

categorized a total of 1760 (697+1063) BMP-regulated genes in the FG and HG

tissue.

• To identify Wnt-regulated genes we compared Tg(hsp70:dkk1) heatshocked with

non-heatshocked FG, as well as heatshocked with non-heatshocked HG. We also

compared FG BIO with FG DMSO as well as HG BIO with HG DMSO samples. Wnt-

activated genes were log2FC ≤-1 upon heat-shock or log2FC ≥1 upon BIO treatment

FDR ≤5% (n=959). Wnt-repressed genes were log2FC ≥1 upon heat-shock or

log2FC ≤-1 upon BIO treatment FDR ≤5% (n=2032). Forty-one transcripts had

ambiguous regulation with evidence of being both Wnt-activated and Wnt-repressed,

and were excluded from further analysis. Overall we categorized a total of 2991

(959+2032) Wnt-regulated genes in the FG and HG tissue.

GO term enrichment analyses were performed using ToppGene Suite (Chen et al.,

2009). Heatmaps were generated using GeneE from Broad Institute

(https://software.broadinstitute.org/GENE-E/index.html). Scatter Plots were generated using CSBB’s InteractiveScatterPlot module.

Chromatin immunoprecipitation

Embryos (25-50 whole embryos or 100 FG or HG explants) at stage NF20 were harvested and fixed at room temperature with 1% formaldehyde in 0.1XMBS for 45 minutes. Immediately after fixation, the embryos were incubated with 125 mM glycine/MBS for 10 minutes and washed three times with ice-cold RIPA buffer (50 mM

Tris pH 7.6,150 mM NaCl, 1 mM EDTA, 1% IGEPAL CA-630, 0.25% Sodium deoxycholate, 0.1% SDS, 0.5 mM DTT, and supplemented with Protease Inhibitor

Cocktail (Sigma,P8340)) for 5 minutes. Batches of 50 embryos were snap-frozen in

liquid nitrogen and stored at -80°C for future use. Embryos were thawed on ice, 1 ml of

RIPA buffer was added, homogenized, and then kept on ice for 10 minutes. The lysate was centrifuged at 14,000 rpm for 10 minutes at 4°C, and the pellet was ressuspended in 1ml of RIPA buffer and transferred to a Bioruptor tube (Diagenode) for sonication.

The lysate was sonicated for 15 cycles of 20 seconds ON and 60 seconds OFF on the

Bioruptor Pico Instrument (Diagenode). The sonicated samples were centrifuged at

14,000 rpm for 10 minutes at 4°C, and the supernatant were transferred to a 1.5ml tube.

The supernatant was blocked for 2 hours at 4°C with Dynabeads Protein G (Life technologies). In a separate tube, 20 µl of Dynabeads Protein G was blocked with 1 ml

5% BSA/PBS for 1 hour at 4°C. Followed by another 1 hour incubation with the following antibodies on ice: 20 µl of anti-Smad1 per IP (Invitrogen, 38-5400), 20 µl of anti-β-catenin per IP (Life technologies, 712700) and 3 µl of anti-p300 per IP (Santa

Cruz sc-585 X). A small chromatin aliquot was saved for input (50 µl) and the rest was transferred to the tube with beads and antibody, and incubated overnight at 4°C. The input material was stored at -20°C for later usage. The beads were successively washed with ChIP buffer 1 (50 mM HEPES-KOH pH 7.5, 150 mM NaCl, 2 mM EDTA,

1% Triton X-100, 0.1% sodium deoxycholate), ChIP buffer 2 (50 mM HEPES-KOH pH

7.5, 500 mM NaCl, 2 mM EDTA, 1% Triton X-100, 0.1% sodium deoxycholate), ChIP buffer 3 (10 mM Tris pH 8.0, 250 mM LiCl, 1 mM EDTA, 0.5% IGEPAL CA-630, 0.5%

Sodium deoxycholate), ChIP buffer 4 (10 mM Tris pH 8.0, 1 mM EDTA) for 20 minutes each. Chromatin was eluted from the beads with 105 µl of elution buffer (50 mM Tris pH

8.0, 10 mM EDTA, 1% SDS) for 2 washes of 30 minutes at 65°C. At this stage, the frozen input samples were supplemented with elution buffer and incubated overnight at

65°C for reverse crosslinking. ChIP and input samples were incubated with RNase A at

37°C for 1 hour and treated with proteinase K for 1 hours at 55°C. The de-crosslinked

DNA fragments were purified with phenol:chloroform:isoamylalcohol and precipitated in ethanol for qPCR. qPCR was performed using iQ SYBR Green Supermix (BIORAD) on a QuantStudio 3 Real-time PCR System (ThermoFisher). qPCR primers used were: sp5 (F, 5’- TGT CCC GCC TTT TGT CAC CTC-3’ and R, 5’- GCC GCC CAA TCA TCA

AAG AAG-3’); ventx2.1 (F, 5’- CAT AGC CAG CTG AGC ATA ATA AA-3’ and R, 5’- TCA AAG GCA

GAG ATC ACT ACC A-3’); msx1 (F, 5’- CAT ATG TTT GGG TTT GGA GAG-3’ and R, 5’-GTG CAG AAC ATG

GGA GAT TAG-3’); id3 (F, 5’- TTC GGC GCC GTT GGT TAC TTT ACT -3’ and R, 5’- GTC TCC ACG GGC

AAC CAC TCC TT -3’); cdx2 (F, 5’- AGG TTT CGG CGG CGT TTG TT-3’ and R, 5’- TTG GGC AGT GTT AGT

GAA TGG AGA -3’); sp5 -15kb (F, 5’- GTG ATA AAG TAG TCC CAG CAG TGA-3’ and R, 5’- AAG GGG

GAA ATT TAA ACC AGA TA-3’); ventx2.1 -15kb (F, 5’- GTA GGA ACC CAC AGC CAA TAA TC-3’ and R, 5’- GTC AGT

AAG AAA ATC GCC CAT AAG-3’); id3 -8.5kb (F, 5’- TTC CCT GTG CCT GTG TTG AT-3’ and R, 5’- TTG GGG GCA TTT

ATT TAG TTA TT-3’).

ChIP-seq analysis

ThruPLEX® DNA-seq libraries were constructed from ChIP and input control DNA and sequenced (~30 million reads/library) using Illumina HiSeq2500. Raw reads quality check and quality trimming was performed using FastQC and Trimmomatic. Duplicate mapped and multi-mapped reads were removed using picard and samtools respectively. Peaks were called with MACS2 at default thresholds [--qvalue 0.01, -- mfold 5:50, --call-summits] (Zhang et al., 2008). IDR (Irreproducibility Discovery Rate) was performed with standard thresholds (Li et al., 2011) to identify high-confidence:

Smad1, β-catenin and p300 reproducible peaks as follows:

• Smad1 and p300 whole embryo ChIP-seq were individually mapped to the X.

laevis genome assembly v9.1 (Session et al., 2016) using Bowtie2 at default

thresholds (Langmead and Salzberg, 2012).

• β-catenin and p300 in either FG or HG explants with or without BIO. Fastq files

from individual ChIP-seq experiments were merged for FG/HG and DMSO/BIO

explants from β-catenin or p300 ChIP-seq datasets. Pooled fastq’s reads were

mapped to the X. laevis genome assembly v9.1 (Session et al., 2016) using

Bowtie2 at default thresholds (Langmead and Salzberg, 2012).

Described bam files (from merged and not merged fastq) were converted to tagAlign format with only mapped reads with mapping quality ≥30 using samtools and bedtools.

With this tagAlign file we created three replicates of equal sizes by shuffling and

100

randomly placing tags in each replicate. We performed the IDR analysis on the 3 replicates and input using Rscript as described for ENCODE with a threshold of 0.01

[https://sites.google.com/site/anshulkundaje/projects/idr]. The merged input tagAlign file was used as the input for MACS2 peak calling with [-p 1e-3, --to-large] thresholds. Our

IDR pipeline resulted in higher confidence peaks than with MACS2 peak calling alone.

HOMER findMotifsGenome.pl script was used for motif analysis (Heinz et al., 2010). For the Hypergeometric test we used dhyper function in R (https://stat.ethz.ch/R-manual/R- devel/library/stats/html/Hypergeometric.html). Genome browser views were visualized with Integrative Genomics Viewer (IGV) (Robinson et al., 2011).

The sum of all of our genomic analysis is provided in the Table S10.

References

Chen, J., Bardes, E. E., Aronow, B. J. and Jegga, A. G. (2009). ToppGene Suite for gene list enrichment analysis and candidate gene prioritization. Nucleic Acids Res 37, W305-311. Estaras, C., Benner, C. and Jones, K. A. (2015). SMADs and YAP compete to control elongation of beta-catenin:LEF-1-recruited RNAPII during hESC differentiation. Mol Cell 58, 780-793. Heinz, S., Benner, C., Spann, N., Bertolino, E., Lin, Y. C., Laslo, P., Cheng, J. X., Murre, C., Singh, H. and Glass, C. K. (2010). Simple combinations of lineage- determining transcription factors prime cis-regulatory elements required for macrophage and B cell identities. Mol Cell 38, 576-589. Langmead, B. and Salzberg, S. L. (2012). Fast gapped-read alignment with Bowtie 2. Nat Methods 9, 357-359. Li, B. and Dewey, C. N. (2011). RSEM: accurate transcript quantification from RNA- Seq data with or without a reference genome. BMC Bioinformatics 12, 323. Li, Q., Brown, J. B., Huang, H. and Bickel, P. J. (2011). Measuring reproducibility of high-throughput experiments. 1752-1779.

101

Risso, D., Ngai, J., Speed, T. P. and Dudoit, S. (2014). Normalization of RNA-seq data using factor analysis of control genes or samples. Nat Biotechnol 32, 896- 902. Session, A. M., Uno, Y., Kwon, T., Chapman, J. A., Toyoda, A., Takahashi, S., Fukui, A., Hikosaka, A., Suzuki, A., Kondo, M., et al. (2016). Genome evolution in the allotetraploid frog Xenopus laevis. Nature 538, 336-343. Tsankov, A. M., Gu, H., Akopian, V., Ziller, M. J., Donaghey, J., Amit, I., Gnirke, A. and Meissner, A. (2015). Transcription factor binding dynamics during human ES cell differentiation. Nature 518, 344-349. Watanabe, K., Biesinger, J., Salmans, M. L., Roberts, B. S., Arthur, W. T., Cleary, M., Andersen, B., Xie, X. and Dai, X. (2014). Integrative ChIP-seq/microarray analysis identifies a CTNNB1 target signature enriched in intestinal stem cells and colon cancer. PLoS One 9, e92317. Weirauch, M. T., Yang, A., Albu, M., Cote, A. G., Montenegro-Montero, A., Drewe, P., Najafabadi, H. S., Lambert, S. A., Mann, I., Cook, K., et al. (2014). Determination and inference of eukaryotic transcription factor sequence specificity. Cell 158, 1431-1443. Zhang, Y., Liu, T., Meyer, C. A., Eeckhoute, J., Johnson, D. S., Bernstein, B. E., Nusbaum, C., Myers, R. M., Brown, M., Li, W., et al. (2008). Model-based analysis of ChIP-Seq (MACS). Genome Biol 9, R137.

102

Chapter 3: Genome-wide analysis of β-catenin occupancy at early stages of development

AUTHORS

Mariana L. Stevens1, Praneet Chaturvedi1, Scott A. Rankin1, Melissa Macdonald1,

Sajjeev Jagannathan2, Masashi Yukawa2, Artem Barski2 and Aaron M. Zorn1

AFFILIATIONS

1Division of Developmental Biology, Perinatal Institute 2 Division of Allergy &

Immunology and Human Genetics, Cincinnati Children's Research Foundation and

Department of Pediatrics College of Medicine, University of Cincinnati, Cincinnati, Ohio

45229, USA.

103

Abstract

The Wnt/β-catenin pathway is reiteratively used during digestive and respiratory system development for endoderm layer formation, progenitor patterning and proliferation and differentiation of organ buds. The detailed mechanism by which this pathway regulates different steps of development is still unknown. In chapter 2, I demonstrated that during early somitogenesis, Wnt/β-catenin directly induces hindgut while it inhibits foregut transcriptional programs. However, it is well known that 12 hours earlier, Wnt has the opposite effect: inducing presumptive foregut and inhibiting presumptive hindgut progenitors. Although in the previous chapter we started to gain insights on the role Wnt pathway plays in anterior-posterior patterning at the genomic level, we still do not know how opposing temporal Wnt/β-catenin activities account for this regulation. In this chapter, I present preliminary experiments and analyze recently published data that begin to address whether Wnt signaling, active at different times in different regions of the embryo, results in β-catenin occupying different cis-regulatory modules at distinct developmental stages. We use genome-wide analysis to investigate

β-catenin binding at stage NF9 (blastula) and NF11.5 (gastrula) to understand the temporal mechanism by which Wnt/β-catenin pattern foregut and hindgut progenitors at stage NF20 (early neurula). Preliminary data show that at stage NF20 ~45% of the β- catenin peaks were not associated with either expressed or Wnt-regulated genes, implying they could be regulated at later stages. Moreover, we found that hindgut genes, which are activated by Wnt at stage NF20, have the same CRMs already bound at blastula and gastrula stages. Meanwhile, foregut genes, which are repressed by Wnt

104

at stage NF20, acquire β-catenin only prior to regulation at CRMs not occupied at the earlier stages analyzed. Since Wnt/β-catenin pathway has opposite roles in foregut and hindgut progenitors at different times of development, our findings of differential binding dynamics deepen our understanding of how Wnt target genes can be marked for future transcriptional regulation at different times of development, and use either the same or distinct CRMs depending on whether they are activated or repressed.

Introduction

The formation of the digestive and respiratory systems are governed by dynamic signaling factors that very often play different roles throughout development. The canonical Wnt signaling pathway is known to be involved in many stages from endoderm formation, patterning and tissue specification. Foregut (FG) and hindgut (HG) progenitors, in particular, are good example of cells that are differentially affected by

Wnt signals pre- and post-gastrulation. In the previous chapter I showed how post- gastrula Wnt/β-catenin signaling is critical to induce HG and inhibit FG progenitors.

However, it is already established that in the blastula, maternal β-catenin is active in the dorsal anterior cells (the future FG), and is low in the ventral-posterior cells (the future

HG). In this chapter, I have begun to address at the genomic level how β-catenin can have a different role at different stages of development. We postulated that Wnt differentially regulates the same genes by β-catenin binding to different cis-regulatory modules (CRMs) at distinct developmental stages.

105

The maternal Wnt pathway activity is high in the dorsal-marginal zone of early stage vertebrate embryos and low in the ventral-marginal zone. Together with other factors like VegT, this dorsal Wnt/β-catenin induces Nodal-related ligands (nodal1/5/6), promoting an environment for initiation of the definitive endoderm gene regulatory network (Clements et al., 1999; D'Amour et al., 2005; Green and Smith, 1990; Kubo et al., 2004). As the gastrulation process starts, the maternal to zygotic transition takes place, and the presumptive FG cells, which required high levels of Wnt, transition to a low Wnt state. Our group and others have shown that after gastrulation, high levels of

Wnt have an inhibitory role on FG development (McLin et al., 2007; Sherwood et al.,

2011; Zhang et al., 2013). Wnt antagonists like secreted Frizzled-related proteins

(sFRPs) and Dickkopf (DKK) become expressed in this domain during gastrulation protecting presumptive FG cells from high Wnt levels, and allow for maintenance of the

FG identity (Li et al., 2008b; Rankin et al., 2011; Stuckenholz et al., 2013). Meanwhile, at the ventral marginal zone, zygotic Wnt8 and Bmp4 start the ventral posterior embryonic program, the presumptive HG (De Robertis and Kuroda, 2004). HG progenitors continue to require high levels of Wnt post-gastrula, being widely used as a step for stem cell differentiation in posterior gut lineages (McLin et al., 2007; Sherwood et al., 2011; Spence et al., 2011). Overall, FG and HG progenitors are exposed to varying levels of Wnt signals, such as before gastrulation Wnt/β-catenin activity is high in the dorso-anterior and low in the ventro-posterior domain, while after gastrulation the anterior hhex+ domain is low in Wnt/β-catenin and the posterior ventx2.1+ domain is high in Wnt/β-catenin (Fig. 1).

106

Fig. 1: Summary diagram of β-catenin activity levels of early developing Xenopus laevis embryos. The drawings are based on β-catenin immunostaining patterns at the specified stages of development. At stage NF9 Wnt/β-catenin activity is high in the dorso-anterior and low in the ventro-posterior domain. During gastrulation, a transition occurs such as by stage NF20 the anterior hhex+ domain is low in Wnt/β-catenin and the posterior ventx2.1+ domain is high in Wnt/β-catenin. Red is strongest activity and pale yellow is weak activity. h – hhex expression domain, v – ventx2.1 expression domain, D - dorsal, V - ventral, A - anterior, P – posterior. Modified from (Schohl and Fagotto, 2002).

The Wnt signaling pathway is activated by the binding of Wnt ligands, which initiates a phosphorylation cascade that sequesters the β-catenin destruction complex to the membrane. The β-catenin destruction complex is formed by Axin, Adenomatous

Polyposis Coli (APC), glycogen synthase kinase 3β (GSK3β) and Casein Kinase 1 alpha (CKIα). In the absence of ligand binding, the pathway is inactive, and CKIα and

GSK3β phosphorylate and polyubiquinate β-catenin targeting for proteasomal degradation (Aberle et al., 1997; Liu et al., 2002). However, upon ligand binding and inactivation of the destruction complex, β-catenin is free to travel to the nuclei and through DNA-binding factor T-cell transcription factor or lymphocyte enhancer factor

(Tcf3/Lef1), activate downstream target genes (Clevers and Nusse, 2012). Although the signaling cascade downstream of Wnt is known, the mechanism by which it can have different impacts on the same cell populations is still largely unresolved.

107

The concept that β-catenin enters the nucleus for immediate activation of target genes upon Wnt ligand binding has been a dogma in the signaling field. However, recent genome-wide analysis studies done by others and us have shown that there are several instances where β-catenin is associated with genes not regulated by Wnt

(Kjolby and Harland, 2016; Nakamura et al., 2016) (Chapter 2). It is becoming prominent that, in general, transcription factors can bind to genes prior to transcriptional regulation; moreover, it has already been shown that maternal β-catenin primes genes that will only be transcribed after the maternal to zygotic transition (Blythe et al., 2010;

Skalska et al., 2015). However, thus far, it is still unknown whether the β-catenin priming mechanism is used in a genome-wide scale to mark genes that will be Wnt-regulated only at further developmental stages.

In this chapter we tested two mechanistic models: 1) β-catenin primes all Wnt- responsive CRMs prior to transcriptional regulation; 2) β-catenin binds to different

CRMs at distinct developmental stages, reflecting the differential Wnt signaling activities on FG and HG progenitors. Using β-catenin ChIP-seq at different developmental stages and RNA-seq of Wnt regulated genes at gastrula and neurula stages, we found that the mechanism is a mix of both proposed models. We report that ~45% of stage NF20 β- catenin binding events are not associated with transcribed genes, suggesting non- productive binding. The majority of these binding events are not residual from gastrula

Wnt-regulation, indicating priming for future regulation. More importantly, we show that

HG-activated genes are marked for regulation much earlier in development with the same CRMs being occupied at multiple stages. Meanwhile, FG-repressed genes do not appear to be marked for regulation, and most of the β-catenin binding seems to be at

108

stage NF20 specific CRMs. Our preliminary findings begin to address early dynamics of the canonical Wnt pathway and how Wnt/β-catenin distinctly marks activated genes, through common CRMs at distinct developmental stages, and repressed genes, through stage-specific CRMs.

Materials and Methods

Embryo experiments and manipulations

Animal experiments were performed according to CCHMC IACUC approved protocols.

Xenopus laevis embryos were staged according to Nieuwkoop and Faber (Nieuwkoop and Faber, 1967). β-catenin morpholino injections were performed as previously described (Rankin et al., 2011). Small molecule inhibitor BIO was dissolved in DMSO and embryos were cultured from stages NF12-20 with either DMSO vehicle in 0.1XMBS or BIO (60 µM; TOCRIS). Stage NF12 transgenic Xenopus laevis

Xla.Tg(hgem:Xtr.dkk1)Jmws, (Lin and Slack, 2008), were heatshocked at 370C for 30 minutes followed by incubation at 130C to NF20.

In situ hybridizations and immunostaining were performed as previously described (McLin et al., 2007). Clones and details on mRNA probe construction available upon request. For immunofluorescence anti-β-catenin (1:300, Santa Cruz

Biotechnologies, sc-7199) was used.

109

RNA-seq

RNA/DNA-seq analyses were performed using the Xenopus laevis genome v9.1,

(Session et al., 2016). Since Xenopus laevis is allotetraploid most genes contain two copies, which are designated as .L or .S (eg. cdx2.L and cdx2.S). For simplicity we dropped the .L or .S in the figures.

For stage NF20 RNA-seq, 50 explants were microdissected and total RNA was extracted from three independent biological replicates with the Nucleo-spin RNA kit

(Machery-Nagel). Libraries were constructed with TruSeq Stranded mRNA Library Prep

Kit and sequenced ~10 million reads/library with 75 bp length using Illumina HiSeq2500.

Publicly available RNA-seq data (GSE77364) was obtained from GEO (Kjolby and

Harland, 2016). Both public and our own data were analyzed as follows. FastQC reports identified adapters, over-represented sequences, low quality bases and overall low quality reads. Trimmomatic was used to clip off adapters, over-represented sequences and low quality bases. Reads were trimmed keeping minimum length as 50, thereby after trimming, read lengths ranged from 50 to 75 base pairs. Quality trimmed reads were mapped to the X. laevis genome 9.1, quantified using RSEM and mapped with bowtie2 using default thresholds (Li and Dewey, 2011). Differential gene expression analysis was carried out using CSBB’s [https://github.com/csbbcompbio/CSBB-v1.0]

Differential Expression Module, which uses RUVSeq (Risso et al., 2014). With RUVSeq, we performed two-way normalization on the count’s matrix 1) Upper Quantile and 2)

Empirical gene normalization with default settings, and differential expression analysis.

Log2FC ≤-1 or ≥1, p <0.05 and FDR ≤5% was used for all of the analysis. For detailed information on stage NF20 differential expression comparison please refer to chapter 2.

110

Chromatin immunoprecipitation

Twenty-five whole embryos at stages NF9 and NF20 were harvested and fixed at room temperature with 1% formaldehyde in 0.1XMBS for 45 minutes. Immediately after fixation, the embryos were incubated with 125 mM glycine/MBS for 10 minutes and washed three times with ice-cold RIPA buffer (50 mM Tris pH 7.6,150 mM NaCl, 1 mM

EDTA, 1% IGEPAL CA-630, 0.25% Sodium deoxycholate, 0.1% SDS, 0.5 mM DTT, and supplemented with Protease Inhibitor Cocktail (Sigma,P8340)) for 5 minutes. Batches of

50 embryos were snap-frozen in liquid nitrogen and stored at -80°C for future use.

Embryos were thawed on ice, 1 ml of RIPA buffer was added, homogenized, and then kept on ice for 10 minutes. The lysate was centrifuged at 14,000 rpm for 10 minutes at

4°C, and the pellet was resuspended in 1ml of RIPA buffer and transferred to a

Bioruptor tube (Diagenode) for sonication. The lysate was sonicated for 15 cycles of 20 seconds ON and 60 seconds OFF on the Bioruptor Pico Instrument (Diagenode). The sonicated samples were centrifuged at 14,000 rpm for 10 minutes at 4°C, and the supernatant was transferred to a 1.5ml tube. The supernatant was blocked for 2 hours at 4°C with Dynabeads Protein G (Life technologies). In a separate tube, 20 µl of

Dynabeads Protein G was blocked with 1 ml 5% BSA/PBS for 1 hour at 4°C. Followed by another 1 hour incubation with 20 µl of anti-β-catenin (Life technologies, 712700) on ice. A small chromatin aliquot was saved for input (50 µl) and the rest was transferred to the tube with beads and antibody, and incubated overnight at 4°C. The input material was stored at -20°C for later usage. The beads were successively washed with ChIP buffer 1 (50 mM HEPES-KOH pH 7.5, 150 mM NaCl, 2 mM EDTA, 1% Triton X-100,

0.1% sodium deoxycholate), ChIP buffer 2 (50 mM HEPES-KOH pH 7.5, 500 mM NaCl,

2 mM EDTA, 1% Triton X-100, 0.1% sodium deoxycholate), ChIP buffer 3 (10 mM Tris 111

pH 8.0, 250 mM LiCl, 1 mM EDTA, 0.5% IGEPAL CA-630, 0.5% Sodium deoxycholate),

ChIP buffer 4 (10 mM Tris pH 8.0, 1 mM EDTA) for 20 minutes each. Chromatin was eluted from the beads with 105 µl of elution buffer (50 mM Tris pH 8.0, 10 mM EDTA,

1% SDS) for 2 washes of 30 minutes at 65°C. At this stage, the frozen input samples were supplemented with elution buffer and incubated overnight at 65°C for reverse crosslinking. ChIP and input samples were incubated with RNase A at 37°C for 1 hour and treated with proteinase K for 1 hour at 55°C. The de-crosslinked DNA fragments were purified with phenol:chloroform:isoamylalcohol and precipitated in ethanol for qPCR. qPCR was performed using iQ SYBR Green Supermix (BIORAD) on a

QuantStudio 3 Real-time PCR System (ThermoFisher). qPCR primers used were: sp5 (F, 5’- TGT CCC GCC TTT TGT CAC CTC-3’ and R, 5’- GCC GCC CAA TCA TCA

AAG AAG-3’); ventx2.1 (F, 5’- CAT AGC CAG CTG AGC ATA ATA AA-3’ and R, 5’- TCA AAG GCA

GAG ATC ACT ACC A-3’); cdx2 (F, 5’- AGG TTT CGG CGG CGT TTG TT-3’ and R, 5’- TTG GGC AGT GTT AGT

GAA TGG AGA -3’); sp5 -15kb (F, 5’- GTG ATA AAG TAG TCC CAG CAG TGA-3’ and R, 5’- AAG GGG

GAA ATT TAA ACC AGA TA-3’); ventx2.1 -15kb (F, 5’- GTA GGA ACC CAC AGC CAA TAA TC-3’ and R, 5’- GTC AGT

AAG AAA ATC GCC CAT AAG-3’);

112

ChIP-seq

Five nanograms of ChIP material were used for library preparation. ThruPLEX® DNA- seq libraries were constructed from ChIP and input control DNA and sequenced (~10 million reads/library) using Illumina HiSeq2500. Raw reads quality check and quality trimming was performed using FastQC and Trimmomatic. Public stage NF11.5 flag

ChIP-seq (GSE77363) (Kjolby and Harland, 2016) and our own stage NF9 and NF20 β- catenin ChIP-seq datasets were analyzed as follows. Fastq’s reads were mapped to the

X. laevis genome assembly v9.1 (Session et al., 2016) using Bowtie2 at default thresholds (Langmead and Salzberg, 2012). Duplicate mapped and multi-mapped reads were removed using picard and samtools respectively. Peaks were called with MACS2 at default thresholds [--qvalue 0.01, --mfold 5:50, --call-summits] (Zhang et al., 2008).

HOMER findMotifsGenome.pl script was used for motif analysis (Heinz et al., 2010).

Genome browser views were visualized with Integrative Genomics Viewer (IGV)

(Robinson et al., 2011).

Results

During gastrulation, foregut progenitors transition to a low Wnt/β-catenin environment

Many previous studies have established that maternal nuclear β-catenin is high in the dorsal anterior mesendoderm of the organizer and low in the ventral posterior region, and that this switches by the late gastrula (Blythe et al., 2010; McLin et al., 2007;

Rankin et al., 2011; Schohl and Fagotto, 2002; Zorn et al., 1999b). In order to determine precisely when β-catenin becomes down regulated in the future FG and up regulated in

113

the future HG, we compared the levels of nuclear (n)β-catenin in hhex+ future FG and ventx2.1+ future HG cells by immunostaining from stages NF10.5 to NF20. As gastrulation starts and until mid-gastrula stages (NF10.5-12), hhex+ FG transition from a high (n)β-catenin state to a low (n)β-catenin state. Meanwhile ventx2.1+ HG cells transition from a low (n)β-catenin state to a high (n)β-catenin state. At the end of gastrulation and before stage NF15, hhex+ FG cells completely exit the active Wnt/β- catenin domain, while ventx2.1+ remains positive for (n)β-catenin (Fig. 2A). This pattern persists until early neurula stage (NF20), the main focus in the previous chapter

(Chapter 2 – Fig. 1E-G). Overall these data confirm previous studies and show that FG- progenitors experience a transition during gastrulation from (n)β-cateninhigh to (n)β- cateninlow, whereas the HG-progenitors transition from (n)β-cateninlow to a (n)β- cateninhigh state (Fig. 2B).

114

Fig. 2: During gastrulation, foregut progenitors transition to a low Wnt/β-catenin environment. (A) In situ hybridization of sagittal bisected stages NF10.5, NF12, NF15 and NF20 embryos with hhex and ventx2.1 marking FG and HG domains, respectively. Wnt activity shown by nuclear (n)β-catenin immunostaining. (B) During early stages of development hhex+ FG cells transition from (n)β-cateninhigh to (n)β-cateninlow, whereas ventx2.1+ HG cells continue to experience (n)β-cateninhigh throughout the analyzed stages.

115

Wnt/β-catenin has distinct roles in FG and HG progenitors pre- and post- gastrulation

Given the dynamic subcellular localization of Wnt/β-catenin, we decided to compare the impact of maternal (pre-MBT) Wnt manipulations on FG and HG progenitors to stage NF12 (post-MBT) manipulations (reported in the chapter 2). To disrupt maternal Wnt pathway, embryos were injected with β-catenin morpholino to inhibit the protein translation and impair proper downstream signals. Oligo morpholino was injected at 2-cell stage and embryos were fixed at stages NF11, NF20 and NF35 for further analysis. Meanwhile, to inhibit the canonical Wnt pathway at later stages of development we used the heatshock inducible Dkk1 transgenic line Tg(hsp70:dkk1) (Lin and Slack, 2008). Embryos were heatshocked between stages NF12-20, during FG-HG patterning but after gastrulation to avoid disruption of axial patterning, and fixed at stages NF20 and NF35. For in situ hybridization, hhex and ventx2.1 were used as FG and HG progenitors, respectively, as well as the organ markers nrh1 (liver), ptf1a

(pancreas), sox2 (stomach) and darmin (intestine).

As expected, depletion of β-catenin in the early blastula stage resulted in loss of

FG marker hhex and expansion of HG marker ventx2.1 by stage NF11, as previously shown (Rankin et al., 2011). The effects persisted during early neurula stage (NF20) and affected subsequent organogenesis, since loss of hhex+ FG progenitors and expansion of ventx2.1+ HG domain coincided with absent liver (nr1h5) and stomach

(sox2), with a concomitant expansion of the pancreas (ptf1a) and intestine (darmin)

(Fig. 3A). Compared to previously described data (Chapter 2), inhibition of canonical

Wnt signaling with ectopic expression of Dkk1 from the heatshock transgene resulted in an expansion of hhex+ FG at NF20 with an enlarged liver (nr1h5) at NF35 at the

116

expense of reduced pancreas (ptf1a) and stomach (sox2), while ventx2.1+ HG progenitors are reduced, resulting in a small intestine expression domain (darmin). This data confirms a developmental switch in the canonical Wnt requirements for FG and HG progenitors. FG progenitors require high levels of Wnt/β-catenin prior to gastrulation and low levels of Wnt/β-catenin post-gastrulation, while HG progenitors require low levels of

Wnt/β-catenin prior to gastrulation and high levels of Wnt/β-catenin post-gastrulation

(Fig. 3B).

117

Fig. 3: Wnt/β-catenin has dynamic roles in FG and HG progenitors patterning. (A) In situ hybridization of mid-sagittal section stages NF11 (hhex and ventx2.1; anterior right and dorsal up) and NF20 (hhex and ventx2.1; anterior left and dorsal up) or NF35 embryos (nr1h5, ptf1a, sox2 and darmin; anterior left and dorsal up) in β-catenin morpholino injected or Tg(hsp70:dkk1) embryos. (B) FG progenitors require high levels of maternal Wnt/β-catenin and low levels of zygotic Wnt/β-catenin, while HG progenitors require low levels of maternal Wnt/β-catenin and high levels of zygotic Wnt/β-catenin. (Partially extracted from chapter 2 Fig. S1E).

118

Wnt-regulated transcriptome and β-catenin binding at stages NF9, NF11.5 and NF20

These data and previous work show that Wnt/β-catenin has opposing roles on

FG and HG genes pre- and post-gastrulation, and we hypothesized that this difference comes from β-catenin utilizing different CRMs at distinct stages of development. In order to address the temporal binding differences, we compared RNA-seq of Wnt- regulated genes and ChIP-seq of genomic β-catenin binding from publicly available stage NF11.5 (Kjolby and Harland, 2016) data to our stage NF20 data. Kjolby and

Harland identified Wnt-regulated genes at NF11 by injection of Wnt inhibitor dkk1

(100pg) at 4-cell stage embryos, followed by RNA-seq at stage NF11.5 (Kjolby and

Harland, 2016). The post-gastrulation manipulations with heatshock transgenic dkk1 line and BIO treatment were done as previously described (Chapter 2) (Fig. 4A). The potential limitation of stages NF11.5 and NF20 Wnt manipulations not being completely equivalent will be addressed in the discussion. We also investigated β-catenin binding at stages NF9, NF11.5 and NF20. The stage NF9 ChIP-seq was the earliest stage we were able to obtain good quality ChIP’ed material due to technical limitations since these embryos contain an average of only 104 cells (limitation addressed in the discussion). The stages NF11.5 and NF20 were analyzed for straight comparison with the RNA-seq data. For stage NF11.5, we obtained the data from the mentioned study where the authors injected both blastomeres of 2-cell stage embryos with a β-catenin-

3X-flag (500pg) construct that was used for the chromatin-immunoprecipitation (Kjolby and Harland, 2016). Overexpression of a flag tagged construct is commonly used when no good antibody is available for ChIP. The authors confirmed no off-target effects with their overexpression approach. For stages NF9 and NF20 ChIP-seq, we performed β- 119

catenin ChIP in whole embryos and submitted for high throughput sequencing (Fig. 4B).

The RNA-seq and ChIP-seq datasets were analyzed side-by-side for straight comparisons (refer to methods for detailed information).

Genome-wide motif analysis of β-catenin ChIP-seq datasets revealed Tcf among the top motifs in all stages analyzed. Sox, Smad, Gata and Cdx2 motifs were also present in the dataset, however, at different frequencies depending on the stage (Fig.

4C). This indicates that during development, β-catenin might interact with different transcription factors in addition to Tcf at distinct developmental stages, which could partially explain how Wnt can activate different sets of genes at different times (Fig. 3A).

Our peak distribution analysis revealed occupancy mainly in intergenic regions and to a lesser extent in promoters and intragenic sites (Fig. 4D). The results did not change during the different stages analyzed, inferring that although other co-factors are involved depending on the stage of development, β-catenin binding preferences follow the same pattern.

120

Fig. 4: Wnt-regulated transcriptome and β-catenin binding on early stage embryos. (A) RNA-seq experimental design. RNA was isolated from whole embryos stage NF11.5 from dkk1 injected or non-injected controls and submitted for RNA-seq (Kjolby and Harland, 2016). FG and HG (endo + meso) explants were dissected from DMSO, BIO and Tg(hsp70:dkk1) NF20 embryos and assayed for RNA-seq. Differentially expressed transcripts were identified by pairwise comparisons of control and treated samples with (log2 fold change (FC) ≤-1 or ≥1, false discovery rate (FDR) ≤5%). (B) Experimental design showing β-catenin or flag ChIP-seq of whole embryos at stages NF9, NF11.5 and NF20 embryos. Reads were merged and MACS2 peak calling was performed. (C) DNA-binding protein motif enrichment analysis of all β-catenin or flag ChIP-seq peaks in the genome. (D) Genomic distribution of β-catenin or flag ChIP-seq peaks classified as upstream (-20kb), downstream (+20kb), intragenic and promoter (-1kb to +1kb) regions.

121

By pair-wise comparison between dkk1 injected and control embryos we have identified 101 differentially expressed genes at stage NF11.5. We compared the list of genes regulated by Wnt at NF11.5 with our previously published Wnt-regulated genes at stage NF20 (refer to chapter 2 Fig. 4 for more details). We identified 46 genes regulated by Wnt/β-catenin at both stages of development, including several known

Wnt-target genes like sp5, axin2, hoxd1 and cdx2 (Fig. 5A). The β-catenin and flag

ChIP-seq identified 1509, 10577, 11818 peaks at stages NF9, NF11.5 and NF20, respectively. We performed a peak overlap analysis between all three stages and 20% of the stage NF20 CRMs were also occupied at stage NF11.5, while 2.4% of them were occupied at all stages analyzed (Fig. 5C). These peaks are associated with 1678, 8720 and 7909 genes at stages NF9, NF11.5 and NF20, respectively. We have found several overlapping genes among the three stages analyzed. About 47% of all NF20 β-catenin- bound genes were also present on the NF11.5 dataset, and 10% of all of them were present on all datasets (Fig. 5B). These analyses revealed that ~53% of NF20 β-catenin associated genes are also associated with β-catenin earlier, but only ~25% of NF20 peaks overlap at earlier stages suggests that many genes have multiple peaks and that different peaks are engaged at different times of development. Overall, these results indicate that several genes are bound by β-catenin at multiple stages of development, however, only half of them are occupied at the same CRMs in distinct stages.

122

Fig. 5: Wnt-regulated genes and β-catenin occupancy at different stages of development. (A) Overlap of Wnt-regulated genes at stages NF11.5 and NF20. log2FC ≤-1 or ≥1, FDR ≤5%. (B) Overlap between β-catenin associated genes (+/- 20kb) at different stages of development. (C) Overlapping peaks have at least 1 nucleotide in common at both stages analyzed. Venn shows overlapping β-catenin peaks at stages NF9, NF11.5 and NF20. 123

Non-productive β-catenin binding at stage NF20 binding is not residual from previous Wnt regulation

As we started the investigation of β-catenin binding to stage NF20 Wnt-regulated genes, we noticed that only 12% of all β-catenin binding events in the genome were associated with Wnt-regulated genes at stage NF20. Indeed, examination of RNA-seq expression levels showed that ~55% of these β-catenin peaks were associated with genes that were not even expressed with tpm≥5 − lower than this level is generally considered not expressed (Fig. 6A). For the purpose of discussion, we classified β- catenin associated genes that were not Wnt-regulated and expressed with tpm below 5 as being “non-productive”. These findings raised the question of what role these non- productive peaks have in the Wnt signaling cascade. Therefore, we next asked whether some of these non-productive CRMs were associated with Wnt-regulated genes at prior stages, and β-catenin remained bound until stage NF20, even though these genes were not expressed anymore. Using the published stage NF11.5 Wnt-regulated RNA-seq we compared the non-productive bound genes (blue diagram) with transcripts expressed

(tmp≥5) and regulated by Wnt at stage NF11.5 (Fig. 6B). The results strikingly suggest that a very small percentage (0.1%) of the non-productive binding events at stage NF20 were Wnt-regulated at prior stages and furthermore, less than 6% were even expressed prior to stage NF20. Unexpectedly, some of these peaks associated with genes not expressed at stage NF20 were also present at stages NF9 and NF11.5 ChIP-seq datasets. These data indicate that non-productive β-catenin peaks identified at stage

NF20 are overall not associated with earlier expression or regulation (stage NF11.5). It

124

still remains to be investigated whether these genes are future targets of Wnt/β-catenin and further studies remain necessary.

We next asked whether expressed genes (tmp≥5) bound by β-catenin at stage

NF20 but not Wnt-regulated at NF20 were regulated and/or bound at earlier stages

(NF9 and NF11.5) (green diagram). By comparison of stage NF20 β-catenin bound and expressed (tmp≥5) with stage NF11.5 expressed and Wnt-regulated genes we found that only 0.6% (17+4/3388) are previously regulated by Wnt and almost all of them directly bound by β-catenin at stages NF9 and NF11.5 (Fig. 6C). Meanwhile, over 70%

(2427/3388) of the genes, although expressed and bound at stage NF11.5, are not associated with regulation at this stage (Fig. 6C).

Overall, these results show that only 24 of the genes directly regulated by Wnt at stage NF11.5 remained bound by β-catenin at stage NF20, which does not explain the

~7k binding events not associated with Wnt-regulation at stage NF20. Since the majority of the binding events are not associated with prior regulation, these results also suggest that β-catenin associates with these genes for several developmental stages (NF9,

NF11.5 and NF20) without signs of transcriptional regulation, perhaps as mechanism of transcriptional priming.

125

Fig. 6: Non-productive β-catenin binding at stage NF20 binding is not residual from stage NF11.5 regulation. (A) Overlap of expressed (tmp≥5), Wnt-regulated (log2FC ≤-1 or ≥1, FDR ≤5%) and β- catenin bound (+/- 20kb) genes at stage NF20. (B) Overlap between stage NF11.5 expressed genes, stage NF11.5 Wnt-regulated genes and stage NF20 β-catenin bound genes not expressed or regulated at stage NF20 (blue diagram). (C) Overlap between stage NF11.5 expressed genes, Wnt-regulated genes and stage NF20 β-catenin bound genes not regulated at stage NF20 (green diagram).

126

Stage NF20 direct Wnt-target genes are associated with β-catenin at earlier stages of development

After investigation of non-productive β-catenin binding at stage NF20, we investigated Wnt-regulated genes at stage NF20 and whether they are already regulated at stage NF11.5 and β-catenin bound at blastula (NF9) and/or gastrula

(NF11.5) stages (Fig. 7A). By comparing these 935 Wnt-regulated and β-catenin-bound

NF20 genes (601+334 genes, red diagram) with genes that were expressed with tpm above 5 and/or Wnt-regulated from stage NF11.5 we found that only 4% (29+7/935) of them were regulated prior to stage NF20 (Fig. 7B). This list includes key hindgut progenitor markers and Wnt-target genes like axin2, cdx2, sp5 and ventx1.1. Being that only 4% of 935 Wnt-regulated stage NF20 genes are also regulated at stage NF11.5 we can conclude, at a genome-wide level, that Wnt signaling is temporally dynamic, such as only hours apart can impact largely distinct gene pools. Another interesting finding was that 33.4% (312/935) of Wnt-regulated genes at stage NF20 were already expressed at stage NF11.5 but they were not regulated by experimental dkk1 overexpression. Although these genes had tmp values above 5 at stage NF11.5, they were not yet regulated by the canonical Wnt pathway, which happened only at stage

NF20 (Fig. 7B). These results indicate three different mechanisms by which Wnt/β- catenin regulates these 935 genes during development: 1) 36 genes are Wnt-regulated and β-catenin bound in both stages NF11.5 and NF20, 2) 312 genes are expressed at

NF11.5 but not regulated by Wnt prior to stage NF20, and 3) 587 genes are neither expressed nor regulated prior to stage NF20.

Given these distinct categories of regulation, we hypothesized that β-catenin would bind preferentially to Wnt-regulated genes at stage NF11.5 compared to the other 127

two gene groups. By comparison with β-catenin associated genes at stages NF9 and

NF11.5, we identified that 75% (27/36) of the genes that were regulated by Wnt at both stages NF11.5 and NF20, were already associated with β-catenin since stage NF9 (Fig.

7C). This list includes known Wnt-target gene sp5 and HG markers cdx2 and ventx1.1.

A detailed peak overlap analysis even revealed that about 92% of these genes had β- catenin occupying the same CRMs in at least two out of the three stages analyzed, indicating that the same regulatory modules are responsible for their expression across development (Fig. 7C). However, we observed that β-catenin is bound since stage NF9 to only 25% (78/312) and since stage NF11.5 to 50% (156/312) of the genes that are expressed but not regulated by Wnt until stage NF20. Moreover ~55% of the CRMs identified are occupied in at least two out of the three stages analyzed (Fig. 7C). We hypothesized that β-catenin associates with these genes, priming them for future regulation, which happens only at stage NF20. Interestingly, FG-enriched genes like hhex and gata4, although expressed at stage NF11.5, did not seem to be occupied prior to stage NF20. Moreover, we also found that β-catenin occupancy near genes that were not even expressed at stage NF11.5 happened mostly at stage NF20, with only 8%

(51/587) and 36% (211/587) of the CRMs occupied since stages NF9 and NF11.5, respectively. With only 25% of overlapping peaks across at least two out of the three analyzed stages, we conclude that several identified CRMs are stage NF20-specific and not observed at earlier stages (Fig. 7C). Taken together, these data suggest that Wnt- regulated genes undergo a very dynamic mechanism of β-catenin binding. Some genes are Wnt-regulated prior to stage NF20 and maintain β-catenin on the same CRMs,

128

whereas other genes are only Wnt-regulated at stage NF20 but acquire β-catenin either at much earlier stages or right prior to regulation.

129

Fig. 7: Stage NF20 direct Wnt-target genes are associated with β-catenin at earlier stages of development. (A) Overlap of expressed (tmp≥5), Wnt-regulated (log2FC ≤-1 or ≥1, FDR ≤5%) and β- catenin bound (+/- 20kb) genes at stage NF20. (B) Overlap between stages NF11.5 expressed genes, stage NF11.5 Wnt-regulated genes and stage NF20 directly regulated genes (red diagram). (C) Mutually bound and peak overlap of genes classified by whether or not they were Wnt-regulated or expressed at stage NF11.5.Top panels show mutually bound genes across different stages of development. Bottom panels show stage NF20 β-catenin peak read density at stages NF9 and NF11.5. 0 is the position of the stage NF20 β-catenin peak summit.

130

Hindgut marker CRMs are primed by β-catenin prior to regulation

We have identified different β-catenin binding dynamics for the different classes of genes analyzed. When the gene is Wnt-regulated at both stages NF11.5 and NF20,

β-catenin seems to use the same CRMs at distinct developmental stages, however, the genes not regulated prior to stage NF11.5 are occupied by β-catenin at distinct CRMs throughout development. Therefore, we asked whether Wnt activation or repression could be related to β-catenin use of the same or different CRMs during development.

We analyzed two groups of Wnt-regulated genes described in the previous chapter: 496

FG-enriched Wnt-repressed genes and 247 HG-enriched Wnt-activated genes.

Although we gained insights on how the spatial differences coordinate the regulation of these genes, these recent findings suggest that differential temporal Wnt/β-catenin might also play a role. Therefore, we used stages NF9, NF11.5 and NF20 ChIP-seq datasets to compare β-catenin binding to these two groups of genes at different developmental stages. We found that among FG-repressed genes, only 30% are β- catenin bound since stage NF9 while twice as many (60%) HG-activated genes were previously associated with β-catenin (Fig. 8A). Interestingly, only 8% of the peaks associated with FG-repressed genes are the same at different developmental stages

(Fig. 8B). For example, the FG expressed gene sfrp2, has two β-catenin peaks, one upstream and one downstream of the transcription start site, the former bound at both stages NF9 and NF11.5, while the latter only at stage NF20. While the gene hhex, contains an upstream β-catenin peak only at stage NF20, and no other peaks identified at earlier developmental stages (Fig. 8C). Interestingly, our peak overlap analysis shows ~40% of the β-catenin peaks were present in at least two of the three analyzed

131

stages, suggesting that HG-activated genes are bound by β-catenin since early development (Fig. 8B). Cdx2 and msx1, known intestinal lineage determinants, have the same β-catenin peaks at stages NF9, NF11.5 and NF20 (Fig. 8C). These results suggest that Wnt/β-catenin regulation mechanism involves early β-catenin binding of

HG-activated genes through the same CRMs starting at least at stage NF9, and β- catenin binding to FG-repressed genes only prior to regulation.

132

Fig. 8: Hindgut marker CRMs are primed by β-catenin prior to regulation. (A) Overlap of β-catenin bound (+/- 20kb) genes at stages NF9 and NF11.5 and Wnt- regulated FG-repressed and HG-activated genes (log2FC ≤-1 or ≥1, FDR ≤5%). (B) β- catenin read density analysis of ChIP-seq from embryos at stages NF9, NF11.5 and NF20 associated with FG-repressed and HG-activated genes. (C) Genome browser view of β-catenin peaks on FG-repressed genes sfrp2 and hhex and HG-activated genes cdx2 and msx1. Red boxes indicate β-catenin significant peaks.

133

Discussion

During developmental progression, the same growth factors are reiteratively used in different tissues and embryonic stages. How these factors coordinate temporal and spatial outputs is still largely unknown. We raised the question whether Wnt signaling, active at different times in different regions of the embryo, would result in β- catenin occupying different CRMs at distinct developmental stages. We tested two mechanistic models for Wnt specificity: 1) β-catenin primes all Wnt-responsive CRMs and specificity comes from further combinatorial signals, 2) specificity is controlled by β- catenin binding to different CRMs at distinct developmental stages. Our described results suggest a combination of both events. We found that among Wnt-regulated genes, HG-enriched Wnt-activated genes are bound by β-catenin at the same CRMs as early as blastula stage, while FG-enriched Wnt-repressed genes are not bound until before transcriptional regulation. These findings suggest two different mechanisms by which Wnt activates and represses downstream target genes, and we started to understand the extent to which β-catenin binds to distinct or overlapping regions in the genome as a result of temporal and spatially distinct Wnt signaling events.

Limitations of the study

In our preliminary analysis, we compared Wnt-regulated genes at two distinct developmental stages, NF11.5 and NF20. Our experimental limitation in this analysis was that the transcriptome for stage NF11.5 was originated from whole embryos, while the stage NF20 was microdissected FG and HG explants available from the chapter 2

134

analysis. The ideal comparison would be to either dissect stage NF11.5 dorso-anterior

(presumptive FG) and ventral-posterior (presumptive HG) tissues or to use whole stage

NF20 embryos transcriptome. Moreover, being a pilot study we used both gain- and loss-of function results from stage NF20 RNA-seq data, while the available stage

NF11.5 was generated with only loss-of-function. We understand that some differences on Wnt-regulated genes at the two stages could be due to the described limitations.

In this chapter we begin to address β-catenin binding profile at different stages of development. One of our major findings was that at stage NF20 almost half of β-catenin peaks were not associated with genes expressed or Wnt-regulated at that stage. We further asked whether these genes were previously expressed and/or regulated by Wnt.

Our findings suggested that most of these binding events were not associated with expression or Wnt-regulation at stage NF11.5 according to publicly available data, however, we are lacking the analysis at stages prior to NF11.5, like stage NF9 for example. Ideally, we would like to perform Wnt manipulations with the use of our heatshock transgenic dkk1 line to inhibit Wnt during early embryogenesis and address at stage NF9, by RNA-seq, which genes are affected by such manipulation. Afterward we can make the conclusion whether β-catenin binding at stage NF20 is residual from regulation events prior to stage NF11.5. This Wnt-regulated transcriptome will also allow us to analyze the maternal Wnt role in FG and HG progenitors, since we know the requirements at this stage are distinct from the requirements starting post-gastrulation until stage NF20.

In this chapter we have also started to analyze early stage β-catenin ChIP-seq data. The frog embryo contains approximately 104 cells at stage NF9 and 105 cells at

135

stage NF20, therefore early stages chromatin immunoprecipitation become very challenging due to less chromatin content. We have found Tcf as the top enriched motif in our stage NF9 β-catenin ChIP-seq, indicating specificity of immunoprecipitation, however, we did not find binding on some known target genes at this stage like some of the nodals, xtwin and siamois. Moreover, the publicly available data for stage NF11.5 is a flag ChIP-seq, which means that a β-catenin flagged construct was injected in the embryo and a flag antibody was used for the pool-down. This is not the ideal approach to identify β-catenin genome-wide binding, since overexpression of flagged constructs can result in several non-specific binding sites. Although both ChIP-seq datasets contain Tcf as top enriched motif and contain β-catenin binding at many target genes, it would still be necessary to repeat them, with a direct β-catenin antibody, to further confirm the results.

We have reported that prior to stage NF10.5 Wnt/β-catenin activity is high in the presumptive FG and low in the presumptive HG cells. Therefore, our results of HG- enriched Wnt-activated genes like cdx2 and msx1 being already bound by β-catenin since stage NF9 suggest that the β-catenin peaks identified might be from the dorso- anterior cells. Cdx2 and msx1 transcriptional activation only happens after maternal-to- zygotic transition, which supports the idea that β-catenin binds to several Wnt-target genes prior to transcriptional regulation. Nevertheless, the fact that the reported β- catenin ChIP-seqs were done in whole embryo does not allow for such distinction between different cell populations. With the data available, we cannot conclude whether these CRMs are occupied by β-catenin peaks in the future FG or HG cells. The ideal

136

experiment would be to perform dorso-anterior and ventral-posterior dissections at stages NF9 and NF11.5 followed by β-catenin ChIP-seq.

β-catenin widespread genome binding

One of our major findings was that of all genes associated with β-catenin peaks only ~12% were Wnt-regulated based on our previously published RNA-seq data

(Chapter 2). Moreover, we have also shown that most of these binding events are associated with non-expressed genes (tpm<5), suggesting several non-productive peaks. Several studies of transcription factor widespread genomic binding found that one way to distinguish productive from non-productive binding could be through different peak heights or intensity of binding. In fact, in fruit flies, it has been shown that over 20 transcription factors present a genome-wide spread binding, just like we have reported for β-catenin, but they also determined that the binding affinity is the determinant factor for a functional transcription factor (Li et al., 2008a; MacArthur et al.,

2009). It still remains to be elucidated if peak height correlation with productive binding is present in vertebrates. Myod, for example, a transcription factor involved in myogenesis did not follow the same behavior. Researchers have shown that Myod also presents widespread binding, however, they could not correlate peak height (as a measure of peak affinity) with Myod regulated transcripts (Cao et al., 2010). Although we have reported several instances where β-catenin binding was associated with genes that are not expressed, further work needs to be done to determine the role widespread binding plays in the Wnt pathway.

137

Lineage determinant’s factors are expected to bind to specific CRMs during developmental progression, however, more recent evidences are challenging this concept. Our findings of β-catenin’s widespread genome binding are confirmed by recent genome-wide β-catenin binding profiles in gastrula Xenopus laevis and Xenopus tropicalis embryos, where both studies have found many β-catenin non-productive binding events (Kjolby and Harland, 2016; Nakamura et al., 2016). Beyond Wnt signaling, other factors were also reported to associate with non-regulated genes. Gli3 for example, presents genome-spread binding in developing limbs, where only 12.4% of binding was associated with Shh transcriptional targets (Vokes et al., 2008). Moreover, in embryonic stem cells Sox2 was also shown to bind multiple regions unrelated to active enhancers. However, it appears that although the sites are non-functional, Sox2 might act as a place-holder to maintain the nucleosome opening for subsequent tissue- specific enhancers like Sox3, 4 and 11 (Spitz and Furlong, 2012). Although we have no evidence thus far for β-catenin being a place-holder, it would be very interesting to know if other lineage specific factors are recruited to these sites to allow further transcriptional regulation. One approach would be to further investigate what role the factors that were identified during our β-catenin ChIP-seq motif analysis could play in this context. For example, whether they are necessary for digestive progenitor key genes and even if they bind to the same CRMs occupied by β-catenin.

β-catenin priming genes for future regulation

The most current model for gene poising and transcriptional activation involves deposition of pioneer factors in closed chromatin regions, where they are able to move 138

nucleosomes and open up chromatin promoting a permissive domain for further factor’s binding. Lineage specific transcription factors recognize these open regions and bind in proximity to downstream target genes, recruiting histone acetylases and promoting a transition from poised to active state (Li et al., 2012; Wang et al., 2015; Zaret and

Mango, 2016). Based on our findings, we could hypothesize that β-catenin recognizes these open chromatin sites and binds to them, mostly through Tcf, prior to transcriptional regulation, using a mechanism of priming. Supporting this hypothesis, β- catenin has been implicated with priming of enhancer elements of dorsal organizer genes prior to their transcription. Genes like siamois and xnr3 are poised by β-catenin in a mechanism involving initiating RNA pol II, H3K4me3 and H3K9/14ac, as well as chromatin remodeling of histone 3 arginine 8 (H3R8) by the Prmt2 methyltransferase

(Blythe et al., 2010).

The idea that transcription factors mark enhancers for future regulation is gaining strength with the advent of deep-sequencing techniques. Moreover, the ability of temporal investigation in frog embryos makes it possible to determine whether these factors are recruited to enhancer domains prior to events of regulation. An interesting study in frogs has addressed this question with several stages during gastrulation and has shown that Smad2/3, downstream of Nodal signaling, are deposited nearby genes many stages prior to their transcriptional regulation. The authors also correlated

Smad2/3 binding with active chromatin epigenetic markers H3K4me1 and H3K27ac

(Gupta et al., 2014). Also in frogs, it has been shown that epigenetic modifications descendant from maternal factors persist during development and influence transcriptional regulation later in development (Hontelez et al., 2015). Chromatin marks

139

like H3K4me3 and H3K27me3 create bivalent domains during the first step of development poising genes for future transcription (Bernstein et al., 2006; Hontelez et al., 2015; Vastenhouw et al., 2010). These findings suggest that perhaps transcription factor binding itself is not enough to dictate the transcriptional state of a nearby gene, but it requires the right combination of epigenetic marks.

Interestingly, recent studies in human stem cells confirmed these findings, showing that many key foregut and hindgut progenitor genes are epigenetically marked with H3K4me1 and H3K27me3 prior to commitment to organ lineages. The studies showed that lineage specific genes are transcriptionally activated only with the correct combination of epigenetic marks and instructional signals from growth factors (Loh et al., 2014; Tsankov et al., 2015; Wang et al., 2015). Therefore, it remains to be elucidated whether our reported β-catenin peaks are within poised bivalent chromatin domains, marked by H3K4me1 and H3K27me3. These peaks could be waiting for the correct downstream signals to recruit chromatin remodelers like Prmt2, deposit H3R8 methylation and regulate nearby genes. Having our current knowledge in chromatin immunoprecipitation, this hypothesis could be tested initially with ChIP-PCR for the chromatin marks mentioned. It would be important to test at different stages if β-catenin occupancy is accompanied by poised chromatin marks prior to transcriptional regulation of key HG genes like cdx2, msx2 and ventx2.1.

Wnt/β-catenin promoting epigenetic remodeling

The addition of active marks, like H3K27ac, and removal of repressive marks, like H3K27me3, could be directly downstream of β-catenin binding itself. In chapter 2,

140

we demonstrated that Wnt/β-catenin regulate foregut and hindgut program partially through chromatin modifications. We have shown that histone acetyltransferase p300 was involved on the mechanism of regulation and that this factor is recruited upon induction of Wnt (Chapter 2). The same mechanism has been reported in erythroid cells lines, when Wnt induction by BIO treatment promoted the recruitment of p300 to

TCF7L2 sites in enhancers of blood genes (Trompouki et al., 2011), as well as at a temporal restricted window of heart development where β-catenin recruits histone acetyltransferase CBP with subsequent acetylation of H3K9 to promote Isl1 transcription (Lu et al., 2014). Wnt/β-catenin is not only involved in recruitment of activators but also in removal of repression, as shown in human induced pluripotent stem cells. The pluripotency factor OCT4 was demonstrated to require Wnt/β-catenin not only for activation of endoderm genes, but also for removal of repressive marks

(Ying et al., 2015). It is a possibility that in the context of foregut and hindgut patterning, these poised enhancers contain H3K4me1 mark and β-catenin binding until the embryo reaches the correct developmental stage when H3K27ac is added, H3K27me3 is removed, and transcriptional regulation can be initiated.

Despite the changes in chromatin marks, β-catenin has also been involved in changes of chromatin landscape. A recent study has shown that β-catenin binding to a distal enhancer, through TCF, can promote DNA looping to a nearby promoter. More importantly, β-catenin binding and looping itself was not enough for transcription initiation, which was directly dependent on Smad3 binding to the promoter (Estaras et al., 2015). In fact, our β-catenin peak analysis indicated that among all 3 stages, only

~35% of the peaks were located in promoters or intragenic regions, indicating an overall

141

distal enhancer preference. Our results could also indicate that several non-productive

β-catenin binding events could have initiated chromatin looping, but without further activation signals, transcription did not move forward, highlighting the importance of co- factors to achieve optimal transcription. In chapter 2 we have reported a combinatorial signal between β-catenin and Smad1, which could perfectly fit this hypothesis that β- catenin might bind to genes prior to regulation, promote chromatin looping, and only in the presence of Smad1, it could transcriptionally regulate the nearby gene. We have reported several putative β-catenin co-factors by motif analysis of the different ChIP-seq datasets in this study. Besides Tcf, the only motif present in all stages analyzed was

Sox. Physical interaction between Sox proteins and β-catenin has already been reported, in both agonistic and antagonistic manners. Moreover, the complex Sox-β- catenin-Tcf was shown to coordinate endoderm regulatory network (Sinner et al., 2007;

Sinner et al., 2004; Zorn et al., 1999a). Although evidence points to a cooperation mechanism between β-catenin and other co-factors, it remains to be investigated the different roles these co-factors can have during the different stages of endoderm patterning.

In this study, we have started to understand how Wnt/β-catenin primes genes during development, and how it uses different mechanisms for activation and repression. Widespread β-catenin binding could be involved with poised epigenetic marks and chromatin landscape changes, adding extra layers of transcriptional regulation that goes beyond the simple way we thought to see Wnt pathway. Ligand binding, signaling cascade initiation and β-catenin travelling to the nuclei is not, by itself,

142

able to explain widespread and non-productive binding, and there are still many unanswered questions to be resolved.

References

Aberle, H., Bauer, A., Stappert, J., Kispert, A. and Kemler, R. (1997). beta-catenin is a target for the ubiquitin-proteasome pathway. EMBO J 16, 3797-3804. Bernstein, B. E., Mikkelsen, T. S., Xie, X., Kamal, M., Huebert, D. J., Cuff, J., Fry, B., Meissner, A., Wernig, M., Plath, K., et al. (2006). A bivalent chromatin structure marks key developmental genes in embryonic stem cells. Cell 125, 315- 326. Blythe, S. A., Cha, S. W., Tadjuidje, E., Heasman, J. and Klein, P. S. (2010). beta- Catenin primes organizer gene expression by recruiting a histone H3 arginine 8 methyltransferase, Prmt2. Dev Cell 19, 220-231. Cao, Y., Yao, Z., Sarkar, D., Lawrence, M., Sanchez, G. J., Parker, M. H., MacQuarrie, K. L., Davison, J., Morgan, M. T., Ruzzo, W. L., et al. (2010). Genome-wide MyoD binding in skeletal muscle cells: a potential for broad cellular reprogramming. Dev Cell 18, 662-674. Clements, D., Friday, R. V. and Woodland, H. R. (1999). Mode of action of VegT in mesoderm and endoderm formation. Development (Cambridge, England) 126, 4903-4911. Clevers, H. and Nusse, R. (2012). Wnt/beta-catenin signaling and disease. Cell 149, 1192-1205. D'Amour, K. A., Agulnick, A. D., Eliazer, S., Kelly, O. G., Kroon, E. and Baetge, E. E. (2005). Efficient differentiation of human embryonic stem cells to definitive endoderm. Nat Biotechnol 23, 1534-1541. De Robertis, E. M. and Kuroda, H. (2004). Dorsal-ventral patterning and neural induction in Xenopus embryos. Annu Rev Cell Dev Biol 20, 285-308. Estaras, C., Benner, C. and Jones, K. A. (2015). SMADs and YAP compete to control elongation of beta-catenin:LEF-1-recruited RNAPII during hESC differentiation. Mol Cell 58, 780-793. Green, J. B. and Smith, J. C. (1990). Graded changes in dose of a Xenopus activin A homologue elicit stepwise transitions in embryonic cell fate. Nature 347, 391-394. Gupta, R., Wills, A., Ucar, D. and Baker, J. (2014). Developmental enhancers are marked independently of zygotic Nodal signals in Xenopus. Dev Biol 395, 38-49.

143

Heinz, S., Benner, C., Spann, N., Bertolino, E., Lin, Y. C., Laslo, P., Cheng, J. X., Murre, C., Singh, H. and Glass, C. K. (2010). Simple combinations of lineage- determining transcription factors prime cis-regulatory elements required for macrophage and B cell identities. Mol Cell 38, 576-589. Hontelez, S., van Kruijsbergen, I., Georgiou, G., van Heeringen, S. J., Bogdanovic, O., Lister, R. and Veenstra, G. J. (2015). Embryonic transcription is controlled by maternally defined chromatin state. Nat Commun 6, 10148. Kjolby, R. A. and Harland, R. M. (2016). Genome-wide identification of Wnt/beta- catenin transcriptional targets during Xenopus gastrulation. Dev Biol. Kubo, A., Shinozaki, K., Shannon, J. M., Kouskoff, V., Kennedy, M., Woo, S., Fehling, H. J. and Keller, G. (2004). Development of definitive endoderm from embryonic stem cells in culture. Development (Cambridge, England) 131, 1651- 1662. Langmead, B. and Salzberg, S. L. (2012). Fast gapped-read alignment with Bowtie 2. Nat Methods 9, 357-359. Li, B. and Dewey, C. N. (2011). RSEM: accurate transcript quantification from RNA- Seq data with or without a reference genome. BMC Bioinformatics 12, 323. Li, X. Y., MacArthur, S., Bourgon, R., Nix, D., Pollard, D. A., Iyer, V. N., Hechmer, A., Simirenko, L., Stapleton, M., Luengo Hendriks, C. L., et al. (2008a). Transcription factors bind thousands of active and inactive regions in the Drosophila blastoderm. PLoS Biol 6, e27. Li, Y., Rankin, S. A., Sinner, D., Kenny, A. P., Krieg, P. A. and Zorn, A. M. (2008b). Sfrp5 coordinates foregut specification and morphogenesis by antagonizing both canonical and noncanonical Wnt11 signaling. Genes Dev 22, 3050-3063. Li, Z., Gadue, P., Chen, K., Jiao, Y., Tuteja, G., Schug, J., Li, W. and Kaestner, K. H. (2012). Foxa2 and H2A.Z mediate nucleosome depletion during embryonic stem cell differentiation. Cell 151, 1608-1616. Lin, G. and Slack, J. M. (2008). Requirement for Wnt and FGF signaling in Xenopus tadpole tail regeneration. Dev Biol 316, 323-335. Liu, C., Li, Y., Semenov, M., Han, C., Baeg, G. H., Tan, Y., Zhang, Z., Lin, X. and He, X. (2002). Control of beta-catenin phosphorylation/degradation by a dual-kinase mechanism. Cell 108, 837-847. Loh, K. M., Ang, L. T., Zhang, J., Kumar, V., Ang, J., Auyeong, J. Q., Lee, K. L., Choo, S. H., Lim, C. Y., Nichane, M., et al. (2014). Efficient endoderm induction from human pluripotent stem cells by logically directing signals controlling lineage bifurcations. Cell Stem Cell 14, 237-252. Lu, H., Li, Y., Wang, Y., Liu, Y., Wang, W., Jia, Z., Chen, P., Ma, K. and Zhou, C. (2014). Wnt-promoted Isl1 expression through a novel TCF/LEF1 binding site and H3K9 acetylation in early stages of cardiomyocyte differentiation of P19CL6 cells. Mol Cell Biochem 391, 183-192.

144

MacArthur, S., Li, X. Y., Li, J., Brown, J. B., Chu, H. C., Zeng, L., Grondona, B. P., Hechmer, A., Simirenko, L., Keranen, S. V., et al. (2009). Developmental roles of 21 Drosophila transcription factors are determined by quantitative differences in binding to an overlapping set of thousands of genomic regions. Genome Biol 10, R80. McLin, V. A., Rankin, S. A. and Zorn, A. M. (2007). Repression of Wnt/beta-catenin signaling in the anterior endoderm is essential for liver and pancreas development. Development (Cambridge, England) 134, 2207-2217. Nakamura, Y., de Paiva Alves, E., Veenstra, G. J. and Hoppler, S. (2016). Tissue- and stage-specific Wnt target gene expression is controlled subsequent to beta- catenin recruitment to cis-regulatory modules. Development (Cambridge, England) 143, 1914-1925. Nieuwkoop, P. D. and Faber, J. (1967). Normal Table of Xenopus laevis (Daudin): A Systematical and Chronological Survey of the Development from the Fertilized Egg to the End of Metamorphosis.: New York: Garland Publishing. Rankin, S. A., Kormish, J., Kofron, M., Jegga, A. and Zorn, A. M. (2011). A gene regulatory network controlling hhex transcription in the anterior endoderm of the organizer. Dev Biol 351, 297-310. Risso, D., Ngai, J., Speed, T. P. and Dudoit, S. (2014). Normalization of RNA-seq data using factor analysis of control genes or samples. Nat Biotechnol 32, 896- 902. Robinson, J. T., Thorvaldsdottir, H., Winckler, W., Guttman, M., Lander, E. S., Getz, G. and Mesirov, J. P. (2011). Integrative genomics viewer. Nat Biotechnol 29, 24-26. Schohl, A. and Fagotto, F. (2002). Beta-catenin, MAPK and Smad signaling during early Xenopus development. Development (Cambridge, England) 129, 37-52. Session, A. M., Uno, Y., Kwon, T., Chapman, J. A., Toyoda, A., Takahashi, S., Fukui, A., Hikosaka, A., Suzuki, A., Kondo, M., et al. (2016). Genome evolution in the allotetraploid frog Xenopus laevis. Nature 538, 336-343. Sherwood, R. I., Maehr, R., Mazzoni, E. O. and Melton, D. A. (2011). Wnt signaling specifies and patterns intestinal endoderm. Mech Dev 128, 387-400. Sinner, D., Kordich, J. J., Spence, J. R., Opoka, R., Rankin, S., Lin, S. C., Jonatan, D., Zorn, A. M. and Wells, J. M. (2007). Sox17 and Sox4 differentially regulate beta-catenin/T-cell factor activity and proliferation of colon carcinoma cells. Mol Cell Biol 27, 7802-7815. Sinner, D., Rankin, S., Lee, M. and Zorn, A. M. (2004). Sox17 and beta-catenin cooperate to regulate the transcription of endodermal genes. Development (Cambridge, England) 131, 3069-3080. Skalska, L., Stojnic, R., Li, J., Fischer, B., Cerda-Moya, G., Sakai, H., Tajbakhsh, S., Russell, S., Adryan, B. and Bray, S. J. (2015). Chromatin signatures at

145

Notch-regulated enhancers reveal large-scale changes in H3K56ac upon activation. EMBO J 34, 1889-1904. Spence, J. R., Mayhew, C. N., Rankin, S. A., Kuhar, M. F., Vallance, J. E., Tolle, K., Hoskins, E. E., Kalinichenko, V. V., Wells, S. I., Zorn, A. M., et al. (2011). Directed differentiation of human pluripotent stem cells into intestinal tissue in vitro. Nature 470, 105-109. Spitz, F. and Furlong, E. E. (2012). Transcription factors: from enhancer binding to developmental control. Nat Rev Genet 13, 613-626. Stuckenholz, C., Lu, L., Thakur, P. C., Choi, T. Y., Shin, D. and Bahary, N. (2013). Sfrp5 modulates both Wnt and BMP signaling and regulates gastrointestinal organogenesis [corrected] in the zebrafish, Danio rerio. PLoS One 8, e62470. Trompouki, E., Bowman, T. V., Lawton, L. N., Fan, Z. P., Wu, D. C., DiBiase, A., Martin, C. S., Cech, J. N., Sessa, A. K., Leblanc, J. L., et al. (2011). Lineage regulators direct BMP and Wnt pathways to cell-specific programs during differentiation and regeneration. Cell 147, 577-589. Tsankov, A. M., Gu, H., Akopian, V., Ziller, M. J., Donaghey, J., Amit, I., Gnirke, A. and Meissner, A. (2015). Transcription factor binding dynamics during human ES cell differentiation. Nature 518, 344-349. Vastenhouw, N. L., Zhang, Y., Woods, I. G., Imam, F., Regev, A., Liu, X. S., Rinn, J. and Schier, A. F. (2010). Chromatin signature of embryonic pluripotency is established during genome activation. Nature 464, 922-926. Vokes, S. A., Ji, H., Wong, W. H. and McMahon, A. P. (2008). A genome-scale analysis of the cis-regulatory circuitry underlying sonic hedgehog-mediated patterning of the mammalian limb. Genes Dev 22, 2651-2663. Wang, A., Yue, F., Li, Y., Xie, R., Harper, T., Patel, N. A., Muth, K., Palmer, J., Qiu, Y., Wang, J., et al. (2015). Epigenetic priming of enhancers predicts developmental competence of hESC-derived endodermal lineage intermediates. Cell Stem Cell 16, 386-399. Ying, L., Mills, J. A., French, D. L. and Gadue, P. (2015). OCT4 Coordinates with WNT Signaling to Pre-pattern Chromatin at the SOX17 Locus during Human ES Cell Differentiation into Definitive Endoderm. Stem Cell Reports 5, 490-498. Zaret, K. S. and Mango, S. E. (2016). Pioneer transcription factors, chromatin dynamics, and cell fate control. Curr Opin Genet Dev 37, 76-81. Zhang, Y., Liu, T., Meyer, C. A., Eeckhoute, J., Johnson, D. S., Bernstein, B. E., Nusbaum, C., Myers, R. M., Brown, M., Li, W., et al. (2008). Model-based analysis of ChIP-Seq (MACS). Genome Biol 9, R137. Zhang, Z., Rankin, S. A. and Zorn, A. M. (2013). Different thresholds of Wnt-Frizzled 7 signaling coordinate proliferation, morphogenesis and fate of endoderm progenitor cells. Dev Biol 378, 1-12.

146

Zorn, A. M., Barish, G. D., Williams, B. O., Lavender, P., Klymkowsky, M. W. and Varmus, H. E. (1999a). Regulation of Wnt signaling by Sox proteins: XSox17 alpha/beta and XSox3 physically interact with beta-catenin. Mol Cell 4, 487-498. Zorn, A. M., Butler, K. and Gurdon, J. B. (1999b). Anterior endomesoderm specification in Xenopus by Wnt/beta-catenin and TGF-beta signalling pathways. Dev Biol 209, 282-297.

147

Chapter 4: Discussion

Mariana L. Stevens

Division of Developmental Biology, Perinatal Institute, Cincinnati Children's Research

Foundation and Department of Pediatrics College of Medicine, University of Cincinnati,

Cincinnati, Ohio 45229, USA.

148

Major Findings

Growth factor signaling regulates cell proliferation, homeostasis, cell fate decisions, diseases and many other biological processes. How combinatorial extracellular signals are integrated in the genome to regulate these different outcomes is still not well understood. In this study, we addressed how BMP/Smad1 and Wnt/β- catenin coordinate their intracellular signals in time and space during development of foregut and hindgut progenitors to properly pattern the complex digestive system. In chapter 2 we identified the transcriptional program of foregut (FG) and hindgut (HG) progenitors in Xenopus embryos, which are largely conserved with mammals. Using

RNA-seq and ChIP-seq we determined how BMP/Smad1 and Wnt/β-catenin signals integrate their signals to refine FG and HG patterning. We show that BMP/Smad1 coordinates D-V patterning in both the endoderm (endo) and mesoderm (meso) layers.

BMP, through Smad1, directly induces ventral mesoderm genes while repressing dorsal endoderm genes. Wnt/β-catenin on the other hand, acts as a genome-wide toggle between direct repression of FG and induction of HG programs. Surprisingly, we found that FG-enriched genes appear to be repressed by β-catenin binding DNA through factors other than Tcfs. Finally, we found that Wnt regulates BMP signaling and that

Smad1 and β-catenin co-occupy several cis-regulatory modules (CRMs) associated with key FG and HG transcriptomes to control tissue restricted gene expression. Since these pathways are reiteratively used during development, we have begun in chapter 3 to examine how Wnt signaling can temporally regulate different sets of target genes. By comparing temporal transcriptome and β-catenin binding profiles, we investigated the mechanism by which β-catenin primes future Wnt target genes at early stages of 149

development. We show that β-catenin has many non-productive binding events, in which associated genes are neither expressed nor Wnt-regulated. However, among

Wnt-regulated genes, several HG-enriched activated transcripts were primed by β- catenin several stages prior to regulation. Meanwhile, FG-enriched Wnt-repressed genes acquired β-catenin binding just a few hours before transcriptional regulation.

Taken together, our data indicate that BMP and Wnt signaling pattern FG and HG progenitors largely by direct Smad1 and β-catenin binding, respectively. BMP coordinates dorso-ventral domains by acting on both the endoderm and mesoderm layers, while Wnt acts on antero-posterior patterning directly repressing FG-enriched genes. The mechanism utilized by Wnt entails activation of the HG program, with early

β-catenin priming events alongside FG program repression by immediate binding of β- catenin through factors other than Tcf/Lef.

Potential Implications

The Cartesian model of BMP and Wnt patterning

We have shown how BMP and Wnt pathways synergize both agonistically and antagonistically to offer a fine-tuned mechanism to regulate FG and HG genes. We have shown that BMP and Wnt crosstalk provide positional information for these progenitor cells: low levels of Wnt and high levels of BMP induces ventral FG, low levels of both Wnt and BMP induces dorsal anterior tissue, high levels of Wnt and BMP induces HG, and high levels of Wnt and low levels of BMP induces paraxial/intermediate mesoderm (Chapter 2 Fig.7). A Cartesian model has been

150

previously proposed in which BMP and Wnt levels provide the positional information for general D-V and A-P patterning in not only amphibians, but also all bilaterians (Eivers et al., 2009; Niehrs, 2010). This model was originally proposed for early gastrulation events, in which conserved BMP and Wnt signals would pattern the Central Nervous

System (CNS) and the mesodermal axis through two perpendicular morphogen gradients. We have found that endoderm D-V and A-P axis patterning meticulously fit the Cartesian model, extending it beyond CNS and early mesoderm patterning to more differentiated progenitors like FG and HG lineages. Furthermore, our CRMs analysis tested the Cartesian model and proved that BMP and Wnt directly synergize on the genome with Smad1 and β-catenin co-occupying several hundreds of genes expressed in different D-V and A-P coordinates during embryonic development.

A new paradigm for BMP and Wnt induced transcriptional repression

β-catenin and Smad1 binding to DNA have been widely used as markers for

BMP and Wnt transcriptional activation of target genes. However, our results showed that about 42% of FG-enriched Wnt-repressed genes have an associated β-catenin peak, while 55% of FG-enriched BMP-repressed genes have a Smad1 peak.

BMP/Smad1 transcriptional repression has been previously documented and it seems to be involved with recruitment of co-repressors, like Schnurri, which compete with

HATs like p300, or with Smad4 binding (Blitz and Cho, 2009; Gaarenstroom and Hill,

2014; Marty et al., 2000). Surprisingly, there have been few reports of β-catenin association with transcriptional repression, especially independent of Tcf as we described (Chapter 2 Fig. 5F). There are two viable possibilities: Tcf/Lef could be

151

binding to non-traditional sites (not recognized by our database) or β-catenin could be using other DNA-binding factors. Both have been demonstrated in a gene-by-gene case, where β-catenin either used non-traditional TCF motifs or partnered with the DNA- binding factor Prop1 in different contexts to transcriptionally repress genes (Blauwkamp et al., 2008; Olson et al., 2006). Interestingly, we were able to show that β-catenin associated repression is more frequent than initially appreciated, and our findings challenge the belief that active Wnt/β-catenin is always accompanied by transcriptional activation.

Smad1 and β-catenin signal integration genome-wide

Our findings of BMP and Wnt integration on the genome add another layer of complexity to their transcriptional regulation. Although BMP and Wnt signaling crosstalk is a topic of interest of many (Itasaki and Hoppler, 2010), very few reports were able to tackle this question with a genome-wide perspective. A comprehensive transcription factor-binding atlas was published in 2015 with over 35 different factors in human ESC, including Smad1 and Tcf (Tsankov et al., 2015). The study correlated transcription factors binding in the different cell types analyzed, and showed that several of these factors overlap genome-wide at different frequencies, depending on the cell type.

Smad1 and Tcf showed some overlap in ESC and mesendoderm progenitor cells, however, the study did not further investigate what genes are associated with these

CRMs and whether they are relevant to that cell population (Tsankov et al., 2015).

Moreover, the authors restricted their analysis to early progenitor cells differentiated in general endoderm, mesoderm and ectoderm tissues without further patterning, which

152

we have been able to tackle in our study. The most relevant work in differentiated cell types has shown Smad1 and Tcf binding profiles in hematopoietic lineages and how they partner with erythroid and myeloid key factors to regulate cell differentiation

(Trompouki et al., 2011). Nonetheless the authors did not further investigate to what extent BMP/Smad1 and Wnt/Tcf signals co-occupy the same enhancers in this context.

It appears that many CRMs are co-occupied by these factors; however, since the study lacks data on which genes were associated with these bindings sites, further analysis would be necessary for a direct comparison with our findings. It is also important to note that β-catenin can use several other DNA-binding domains besides Tcf (Cadigan and

Waterman, 2012); therefore, it might not reflect all Wnt/β-catenin target genes. While the various studies discussed in chapter 1 addressed how these pathways crosstalk on a gene-by-gene basis, until now it was still largely unknown whether and how this could be extrapolated to a genome-wide scale. We show that several key FG and HG genes,

Smad1 and β-catenin integrate on the same CRMs to coordinately pattern the endoderm. Our findings are very relevant for a growing field of stem cell differentiation protocols in endoderm derivatives.

Development informing stem cells studies

The endoderm layer gives rise to several organs, therefore, successful differentiation protocols have many potential therapeutic targets. These protocols commonly recapitulate developmental steps; they rely heavily on available information regarding normal organ development, like the data we have described in our study.

Even though Xenopus embryos are not the closest model organisms to humans, the

153

level of tissue dissections performed in this study would have been very difficult in any mammalian species (Chapter 2 Fig.1A-B). Therefore, while we took advantage of our model system we also highlighted the relevance of our findings to mammalian development. We have shown a high degree of conservation between FG and HG transcription profiles in Xenopus compared to human and mice studies (Chapter 2 Fig.

S1B). Moreover, we have also shown that several key endoderm and mesoderm genes, regulated by BMP and Wnt, contained conserved CRMs occupied by Smad1 and β- catenin factors. Strikingly 42% of mutually bound peaks were syntenic to the human

Smad1 and β-catenin/Tcf peaks (Chapter 2 Fig. S7). We have demonstrated, for the first time in vivo, how two pathways widely used with embryonic stem cell differentiation protocols, BMP and Wnt, interact during patterning stages and how their signals widely impact FG and HG programs. Moreover, we were able to not only study the digestive system progenitors but our data also expanded the understanding on hematopoietic, cardiac and lateral plate mesoderm progenitors, providing valuable knowledge to a vast range of stem cell studies.

Development and disease

Signaling pathways have been implicated in several disease contexts, and understanding the transcriptional mechanism by which these pathways function provides not only insights on the cellular basis of the disease but also on potential therapeutic targets. Abnormal BMP and/or canonical Wnt levels are highly correlated with colon cancer (Bertrand et al., 2012; Schepers and Clevers, 2012), cardiovascular

(Morrell et al., 2016) and bone diseases (Monroe et al., 2012). Mutations have been

154

described in many components of the pathways, but the mechanism adds up to either up- or down-regulation of target genes. Knowing BMP and Wnt direct target genes in the different biological contexts becomes necessary to further understand the biology of each disease. Therefore, we have further contributed to the issue with our genome-wide analysis that comprehends thousands of target transcripts in both endoderm and mesoderm derivative lineages.

Overall, we believe that this work advances our understanding of how combinatorial signaling is integrated in the genome during GI organogenesis and serves as a paradigm for other development and disease contexts where Wnt and BMP interact.

Experimental Limitations

ChIP-seq in whole embryos

Some of the novelty of our findings came from the FG and HG dissections performed for our β-catenin ChIP-seq. In these experiments we were able to see how their transcriptional profiles responded to the signaling manipulations, by RNA-seq, and how β-catenin occupancy changed, using ChIP-seq. However, due to a lesser quality antibody, we were not able to achieve the same level of detail with our Smad1 analysis, which was done in whole embryos. We have described Smad1 peaks associated with

BMP-repressed genes in FG and HG tissues, however, due to the mixed cell population used in the ChIP-seq we could not pinpoint the cell lineages engaged in Smad1 binding.

One alternative to circumvent this issue could be with the use of human pluripotent stem

155

cell (PSC) cultures differentiated in FG or HG progenitors. There have been several studies in which pure FG and HG cell populations were obtained in vitro, where we could test our findings (Green et al., 2011; Loh et al., 2014; McCracken et al., 2017;

Spence et al., 2011; Wang et al., 2015). Upon further differentiation of definitive endoderm into FG and HG progenitors, we could also manipulate BMP and Wnt pathways and perform Smad1 and β-catenin ChIP-seq analysis. Although the results would reflect a pure cell population, they would not fully represent the complex in vivo tissue, where both endoderm and mesoderm layers are present and exchanging crucial signals for embryonic development.

Small molecules and off target effects

BMP and Wnt pathways have distinct roles on FG and HG patterning during development (Chapter 1), therefore our experimental design had to account for such differences. In order to disrupt the pathways without affecting gastrulation and axial patterning we relied on small molecule treatments to inhibit BMP (DMH1) and to induce

Wnt (BIO). DMH1, like other BMP inhibitors, interacts with the transmembrane receptor blocking phosphorylation of Smad1/5/8, hindering the subsequent signaling cascade.

Although with a lower affinity, DMH1 not only interacts with BMP receptor ALK2, but it also has the potential to bind the TGFβ receptor ALK5, having the potential to cause non-specific responses (Alsamarah et al., 2015). The use of a heatshock inducible transgenic line for temporal control - like the Tg(hsp70:dkk1) (Lin and Slack, 2008) - inducing a BMP antagonist, like noggin for example, or treatment of FG and HG

156

dissected explants with Noggin protein, would minimize potential non-specific phenotypes and could be a target of future studies.

BIO has many more potentially confounding off target effects than DMH1 because it inhibits the multifunctional kinase GSK3β. GSK3β has several targets, being implicated not only with the β-catenin destruction complex of the Wnt signaling pathway, but also with glycogen synthesis, apoptosis, microtubules dynamics and the Hedgehog pathway (Ding et al., 2000; Forde and Dale, 2007; Wu and Pan, 2010). In fact, when we analyzed β-catenin recruitment to FG-enriched Wnt-repressed genes, surprisingly we observed several instances where β-catenin was removed from these CRMs, which could indicate off-target effects from BIO treatment. An alternative approach to induce the pathway could be with a heatshock transgenic line inducing the genes Wnt5/8/11, or treatment of FG and HG dissected explants with Wnt proteins, all highly expressed in the HG tissue (Chapter 2 Fig. S1D). Overall, although both small molecules have the potential for off target effects, we cautiously controlled for responses on known target gene expression with both RNA-seq and in situ hybridization, as well as responses of

Smad1 and β-catenin binding to target genes through the use of ChIP-PCR.

Testing in other animal models

The aim of this study was to investigate how BMP/Smad1 and Wnt/β-catenin coordinate FG and HG development. We hoped to inform stem cell differentiation protocols as well as studies on diseases like cancer and bone malformation that involve these pathways. Although we found great conservation with mouse and human cells, our study revealed that in several instances the transcripts were not expressed as seen

157

in mammalian cells or the CRMs were not syntenic to human embryonic stem cells data. Given species-specific differences, it is possible that on a gene-to-gene basis, distinct outcomes can come out from mammalian cells experiments. Nevertheless, our reported findings should be easily testable in any other study models.

Future Directions

Testing our mechanistic model

Genome-wide analyses, like we have performed, does not allow for detailed investigation on individual genes. Although we have reported thousands of genes of interest for endoderm studies, we were not able to test a larger number of CRMs. It would be interesting to see if the genes that fall in the same categories of regulation contain CRMs that respond to BMP and Wnt in similar ways. In chapter 2 we have shown hundreds of BMP and Wnt target genes that contain co-occupied CRMs (Fig. 6).

With the use of luciferase assays we demonstrated, as predicted, that one hhex’s associated CRM is positively regulated by BMP while it is negatively regulated by Wnt.

Moreover both BMP and Wnt positively regulate the tested cdx2 CRM, as predicted.

Our hypothesis is that other genes of the same category would behave in the same manner, however, only individual testing would be able to confirm this. Therefore, my first goal would be to test a larger number of CRMs, including the other 2 categories

(Wnt-activated/BMP-repressed and Wnt-repressed/BMP-repressed) with BMP and Wnt manipulations followed by luciferase assays. Once these responsive CRMs are tested we would like to know which co-factors are orchestrating BMP and Wnt signals. We 158

have shown that β-catenin and Smad are likely to use different co-factors for activation and repression, therefore to fully understand their regulation mechanism it would be necessary to know what factors are playing a role. One way this could be tested is with the same luciferase constructs containing mutations on putative co-factor binding motifs. Moreover, once CRMs are validated for their response to BMP and Wnt signaling manipulations, it would be crucial to determine which ones are responsible for the FG or HG expression domains. Tissue expression validations in Xenopus can be done with either gfp labeled transgenesis (McLin et al., 2007) or CRISPR-Cas9 (Tandon et al., 2016). This directed approach of testing other CRMs, identifying co-factors and determining function/expression would further test our hypothesis and potentially show that BMP and Wnt function in a bigger gene regulatory network to coordinate endoderm patterning.

Smad1 and β-catenin as part of a protein complex

Understanding how Smad1 and β-catenin integration at these CRMs affects transcription and expression could be further explored as far as how these transcription factors recognize and bind to the DNA. There are at least two main mechanisms by which these factors could co-occupy the same DNA domains: through independent binding or through complex formation. It has been shown that Smad1 and Tcf/β-catenin can form protein complexes to regulate some downstream targets like xtwn, (Nishita et al., 2000), myc (Hu and Rosenblum, 2005), msx2 (Hussein et al., 2003) and emx2

(Theil et al., 2002). It would be important to know whether these complexes are formed in FG and HG progenitor cells prior to binding to the CRMs described in our study. Due

159

to the easy access of Xenopus embryo explants, one could test Smad1/β-catenin interaction in FG and HG tissues using co-immunoprecipitations (co-IPs). With easy tissue dissections, we can test if Smad1/β-catenin form a protein-protein complex in FG or HG tissues, or even in both domains. The putative co-factors identified in the previously proposed analysis can also be tested at this point. Moreover, to address if one factor is upstream of the other, we could assay Smad1 or β-catenin recruitment to target genes after BMP or Wnt inhibition with ChIP-PCR. For example, if Smad1 binding is inhibited upon Wnt/β-catenin inhibition, we would be able to infer that BMP/Smad1 transcriptional regulation is dependent on normal Wnt levels. All of these different approaches would further test our model, providing a detailed mechanism for how

BMP/Smad1 and Wnt/β-catenin can effect transcriptional activation or repression in the context of endoderm patterning.

Smad1 and β-catenin impacting chromatin remodeling

Once Smad1 and β-catenin bind to target CRMs, the subsequent steps prior to transcriptional regulation have yet to be fully elucidated. We have shown that upon

Wnt/β-catenin induction, p300 is recruited to HG-activated genes, while it is removed during repression of FG-enriched genes (Chapter 2 Fig. 5C-D). HATs like p300 have the important role of creating a permissive environment for transcriptional activation through histone remodeling and increased DNA accessibility (Holmqvist and Mannervik,

2013). Once p300 is recruited or removed from these CRMs upon Wnt induction, it would be important to know the impact on chromatin remodeling. One would expect that

H3K27ac marks (the histone modification promoted by p300) are increased, while 160

H3K27me3 marks (the repressive modification) are removed from these sites.

Moreover, in a broader way, it is also unknown whether in the context of FG and HG patterning, Wnt/β-catenin is involved with DNA-looping, a process just recently studied

(Estaras et al., 2015; Yochum et al., 2010). It is believed that β-catenin binds to distal enhancer elements and with the help of Cohesin, promotes a loop in the DNA to nearby promoters of mesendoderm genes. Interestingly, this study also showed that β-catenin, at the enhancer, and Smad2 (the TGFβ transcription factor), at the promoter, interact to regulate these genes (Estaras et al., 2015). The idea of multiple enhancers being involved in the regulation of the same gene is gaining strength recently and the chromatin looping could be the conformational change providing the physical proximity between these CRMs (Spitz and Furlong, 2012). Therefore, it is reasonable to believe that, in the context of endoderm patterning, β-catenin and Smad1 could be promoting epigenetic modifications and DNA looping to achieve optimal transcriptional regulation, and depending on the co-factors involved, this regulation could exert activation or repression.

Overall, in this study we have used powerful genome-wide tools to understand how BMP and Wnt coordinate FG and HG patterning. We were able to look at transcriptional changes as well as dynamics of Smad1 and β-catenin binding over time and in different tissues. We are providing future studies in stem cells or any BMP and

Wnt related diseases with a comprehensive database that has just started to unfold the mechanisms of BMP/Smad1 and Wnt/β-catenin regulation during digestive and respiratory system formation.

161

References

Alsamarah, A., LaCuran, A. E., Oelschlaeger, P., Hao, J. and Luo, Y. (2015). Uncovering Molecular Bases Underlying Bone Morphogenetic Protein Receptor Inhibitor Selectivity. PLoS One 10, e0132221. Bertrand, F. E., Angus, C. W., Partis, W. J. and Sigounas, G. (2012). Developmental pathways in colon cancer: crosstalk between WNT, BMP, Hedgehog and Notch. Cell Cycle 11, 4344-4351. Blauwkamp, T. A., Chang, M. V. and Cadigan, K. M. (2008). Novel TCF-binding sites specify transcriptional repression by Wnt signalling. EMBO J 27, 1436-1446. Blitz, I. L. and Cho, K. W. (2009). Finding partners: how BMPs select their targets. Dev Dyn 238, 1321-1331. Cadigan, K. M. and Waterman, M. L. (2012). TCF/LEFs and Wnt signaling in the nucleus. Cold Spring Harb Perspect Biol 4. Ding, V. W., Chen, R. H. and McCormick, F. (2000). Differential regulation of glycogen synthase kinase 3beta by insulin and Wnt signaling. J Biol Chem 275, 32475- 32481. Eivers, E., Demagny, H. and De Robertis, E. M. (2009). Integration of BMP and Wnt signaling via vertebrate Smad1/5/8 and Drosophila Mad. Cytokine Growth Factor Rev 20, 357-365. Estaras, C., Benner, C. and Jones, K. A. (2015). SMADs and YAP compete to control elongation of beta-catenin:LEF-1-recruited RNAPII during hESC differentiation. Mol Cell 58, 780-793. Forde, J. E. and Dale, T. C. (2007). Glycogen synthase kinase 3: a key regulator of cellular fate. Cell Mol Life Sci 64, 1930-1944. Gaarenstroom, T. and Hill, C. S. (2014). TGF-beta signaling to chromatin: how Smads regulate transcription during self-renewal and differentiation. Semin Cell Dev Biol 32, 107-118. Green, M. D., Chen, A., Nostro, M. C., d'Souza, S. L., Schaniel, C., Lemischka, I. R., Gouon-Evans, V., Keller, G. and Snoeck, H. W. (2011). Generation of anterior foregut endoderm from human embryonic and induced pluripotent stem cells. Nat Biotechnol 29, 267-272. Holmqvist, P. H. and Mannervik, M. (2013). Genomic occupancy of the transcriptional co-activators p300 and CBP. Transcription 4, 18-23. Hu, M. C. and Rosenblum, N. D. (2005). Smad1, beta-catenin and Tcf4 associate in a molecular complex with the Myc promoter in dysplastic renal tissue and cooperate to control Myc transcription. Development (Cambridge, England) 132, 215-225.

162

Hussein, S. M., Duff, E. K. and Sirard, C. (2003). Smad4 and beta-catenin co- activators functionally interact with lymphoid-enhancing factor to regulate graded expression of Msx2. J Biol Chem 278, 48805-48814. Itasaki, N. and Hoppler, S. (2010). Crosstalk between Wnt and bone morphogenic protein signaling: a turbulent relationship. Dev Dyn 239, 16-33. Lin, G. and Slack, J. M. (2008). Requirement for Wnt and FGF signaling in Xenopus tadpole tail regeneration. Dev Biol 316, 323-335. Loh, K. M., Ang, L. T., Zhang, J., Kumar, V., Ang, J., Auyeong, J. Q., Lee, K. L., Choo, S. H., Lim, C. Y., Nichane, M., et al. (2014). Efficient endoderm induction from human pluripotent stem cells by logically directing signals controlling lineage bifurcations. Cell Stem Cell 14, 237-252. Marty, T., Muller, B., Basler, K. and Affolter, M. (2000). Schnurri mediates Dpp- dependent repression of brinker transcription. Nat Cell Biol 2, 745-749. McCracken, K. W., Aihara, E., Martin, B., Crawford, C. M., Broda, T., Treguier, J., Zhang, X., Shannon, J. M., Montrose, M. H. and Wells, J. M. (2017). Wnt/beta- catenin promotes gastric fundus specification in mice and humans. Nature 541, 182-187. McLin, V. A., Rankin, S. A. and Zorn, A. M. (2007). Repression of Wnt/beta-catenin signaling in the anterior endoderm is essential for liver and pancreas development. Development (Cambridge, England) 134, 2207-2217. Monroe, D. G., McGee-Lawrence, M. E., Oursler, M. J. and Westendorf, J. J. (2012). Update on Wnt signaling in bone cell biology and bone disease. Gene 492, 1-18. Morrell, N. W., Bloch, D. B., ten Dijke, P., Goumans, M. J., Hata, A., Smith, J., Yu, P. B. and Bloch, K. D. (2016). Targeting BMP signalling in cardiovascular disease and anaemia. Nat Rev Cardiol 13, 106-120. Niehrs, C. (2010). On growth and form: a Cartesian coordinate system of Wnt and BMP signaling specifies bilaterian body axes. Development (Cambridge, England) 137, 845-857. Nishita, M., Hashimoto, M. K., Ogata, S., Laurent, M. N., Ueno, N., Shibuya, H. and Cho, K. W. (2000). Interaction between Wnt and TGF-beta signalling pathways during formation of Spemann's organizer. Nature 403, 781-785. Olson, L. E., Tollkuhn, J., Scafoglio, C., Krones, A., Zhang, J., Ohgi, K. A., Wu, W., Taketo, M. M., Kemler, R., Grosschedl, R., et al. (2006). Homeodomain- mediated beta-catenin-dependent switching events dictate cell-lineage determination. Cell 125, 593-605. Schepers, A. and Clevers, H. (2012). Wnt signaling, stem cells, and cancer of the gastrointestinal tract. Cold Spring Harb Perspect Biol 4, a007989. Spence, J. R., Mayhew, C. N., Rankin, S. A., Kuhar, M. F., Vallance, J. E., Tolle, K., Hoskins, E. E., Kalinichenko, V. V., Wells, S. I., Zorn, A. M., et al. (2011). Directed differentiation of human pluripotent stem cells into intestinal tissue in vitro. Nature 470, 105-109. 163

Spitz, F. and Furlong, E. E. (2012). Transcription factors: from enhancer binding to developmental control. Nat Rev Genet 13, 613-626. Tandon, P., Conlon, F., Furlow, J. D. and Horb, M. E. (2016). Expanding the genetic toolkit in Xenopus: Approaches and opportunities for human disease modeling. Dev Biol. Theil, T., Aydin, S., Koch, S., Grotewold, L. and Ruther, U. (2002). Wnt and Bmp signalling cooperatively regulate graded Emx2 expression in the dorsal telencephalon. Development (Cambridge, England) 129, 3045-3054. Trompouki, E., Bowman, T. V., Lawton, L. N., Fan, Z. P., Wu, D. C., DiBiase, A., Martin, C. S., Cech, J. N., Sessa, A. K., Leblanc, J. L., et al. (2011). Lineage regulators direct BMP and Wnt pathways to cell-specific programs during differentiation and regeneration. Cell 147, 577-589. Tsankov, A. M., Gu, H., Akopian, V., Ziller, M. J., Donaghey, J., Amit, I., Gnirke, A. and Meissner, A. (2015). Transcription factor binding dynamics during human ES cell differentiation. Nature 518, 344-349. Wang, A., Yue, F., Li, Y., Xie, R., Harper, T., Patel, N. A., Muth, K., Palmer, J., Qiu, Y., Wang, J., et al. (2015). Epigenetic priming of enhancers predicts developmental competence of hESC-derived endodermal lineage intermediates. Cell Stem Cell 16, 386-399. Wu, D. and Pan, W. (2010). GSK3: a multifaceted kinase in Wnt signaling. Trends Biochem Sci 35, 161-168. Yochum, G. S., Sherrick, C. M., Macpartlin, M. and Goodman, R. H. (2010). A beta- catenin/TCF-coordinated chromatin loop at MYC integrates 5' and 3' Wnt responsive enhancers. Proc Natl Acad Sci U S A 107, 145-150.

164