UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Identifying gene regulatory networks controlled by bone morphogenetic protein-signaling in Drosophila… Othonos, Katerina Maria 2020

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata


24-ubc_2020_may_othonos_katerina.pdf [ 10.45MB ]
JSON: 24-1.0388817.json
JSON-LD: 24-1.0388817-ld.json
RDF/XML (Pretty): 24-1.0388817-rdf.xml
RDF/JSON: 24-1.0388817-rdf.json
Turtle: 24-1.0388817-turtle.txt
N-Triples: 24-1.0388817-rdf-ntriples.txt
Original Record: 24-1.0388817-source.json
Full Text

Full Text

  IDENTIFYING GENE REGULATORY NETWORKS CONTROLLED BY BONE MORPHOGENETIC PROTEIN-SIGNALING IN DROSOPHILA AND MURINE GENOMES by Katerina Maria Othonos B.Sc., University of Cyprus, 2013    A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY  in  THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Neuroscience)   THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  February 2020  © Katerina Maria Othonos, 2020 ii  The following individuals certify that they have read, and recommend to the Faculty of Graduate and Postdoctoral Studies for acceptance, the dissertation entitled:  Identifying Gene Regulatory Networks Controlled by Bone Morphogenetic Protein-Signaling in Drosophila and Murine Genomes  submitted by Katerina Maria Othonos  in partial fulfillment of the requirements for the degree of Doctor of Philosophy  in Neuroscience  Examining Committee: Dr. Douglas Allan Supervisor  Dr. Timothy O`Connor Supervisory Committee Member  Dr. Brad Hoffman University Examiner Dr. Daniel Goldowitz University Examiner  Additional Supervisory Committee Members: Dr. Michael Underhill Supervisory Committee Member Dr. Vanessa Auld Supervisory Committee Member   iii  Abstract  Bone morphogenetic proteins (BMPs) are a group of phylogenetically conserved signaling proteins, first identified to play important roles in bone formation. Since their discovery, they have been recognized to contribute to embryonic development and adult homeostasis in a multitude of tissues, by regulating cellular lineage commitment, morphogenesis, differentiation, proliferation, and apoptosis. BMPs transduce their signals through intracellular downstream effectors, primarily the Smad transcription factors, many of which bind to genomic BMP-responsive cis-regulatory elements (BMP-CREs) to direct gene expression. Despite their importance in cellular processes and maintenance, BMP-CREs remain largely unidentified at a genomic level for most BMP-dependent cellular processes. The overall objectives of this thesis were to experimentally characterize the widespread function of a novel low-affinity BMP-CRE motif in the Drosophila nervous system and to identify the BMP-driven regulatory network underlying mammalian chondrogenesis.  To address the first goal, we used computational methods to identify this novel BMP-CRE through the Drosophila genome and used in vivo transgenic reporters to determine their function in the Drosophila nervous system. Our results show that this BMP-CRE is used within multiple enhancers to mediate their BMP-dependent activity. For our second goal, we used poly-A transcriptome sequencing (RNA-seq) to characterize differentially expressed genes (DEGs) during chondrogenesis in primary murine cells. Amongst these DEGs, we identified transcription factors/cofactors with previously unknown roles in chondrogenesis that are of interest for further study. Further, we used histone modification ChIP-seq to identify more than 2000 candidate iv  regulatory regions in the vicinity of BMP-responsive DEGs. Using computational tools, we examined these candidate regulatory regions for Smad-binding sites using BMP-CRE motifs identified in Drosophila. We then applied multiple selection criteria to prioritize likely BMP-responsive regulatory regions and assessed four novel regions for BMP-responsive reporter expression, using mouse primary limb mesenchymal (PLM) cells. Among these, we identified two BMP-responsive regulatory regions, including one within 50kb of the transcription factor Jdp2, a gene with previously unknown roles in chondrogenesis.   The genomic mapping of BMP-CREs remains incomplete. Mutations in these cis-regulatory sites and BMP-regulated genes could potentially result in disease, and therefore their identification is of critical importance to help further our understanding of disorders in various human tissues.         v  Lay Summary  From fruit flies to humans, bone morphogenetic protein (BMP) signaling plays critical roles in many tissues. Defects in this pathway are implicated in many human diseases; however, the exact roles of BMPs in these diseases are mostly unknown. BMPs control proteins called transcription factors that can direct gene expression through DNA sequence sites called cis-regulatory elements (CREs). Mutations occurring in CREs have been linked to several human disorders. Despite being extremely important, BMP-dependent CREs are not well characterized. This thesis identifies and studies CREs in fruit fly neurons, providing insight into BMP-dependent gene regulation. Further, it identifies important new genes and BMP-dependent CREs that regulate the process of cartilage formation during mouse embryonic development. This work will help us understand the process of cartilage formation in humans and lays the foundations for identifying cartilage-related BMP-dependent CREs that are associated with human disease.  vi  Preface The contents of this dissertation are my original work. All experiments were designed and conducted by me in conjunction with my supervisor, Dr. Douglas Allan, and the guidance of my supervisory committee.  Chapter 1 is a general introduction to my thesis, beginning with a review of the relevant literature and ending with the main objectives and overarching hypothesis of my research. Chapter 2: “Genomic identification and validation of a novel BMP cis-regulatory element in the Drosophila CNS” The work entailed in this chapter is part of a manuscript currently under review, following the first round of review. The publication is entitled: A low affinity cis-regulatory BMP response element confers subtype-specific gene activation in Drosophila neurons, Anthony J.E. Berndt*, Katerina M. Othonos*, Tianshun Lian, Mo Miao, Shamsuddin Buiyan, Raymond Y. Cho, Justin S. Fong, Seo Am Hur, Paul Pavlidis, Douglas W. Allan   *= co-first authorship Dr. Douglas Allan, Dr. Anthony Berndt and I conceptualized the study and the required experiments that are detailed in Chapter 2. The initial experiments were conducted by Dr. Anthony Berndt, as part of his Ph.D. research. This included cloning of the initial transgenic flies which was done with the help of Raymond Y. Cho, Justin S. Fong, and Seo Am Hur. I generated all further transgenic lines used in the study, performed all dissections and immunocytochemistry, collected data and prepared figures/graphs. Mo Miao assisted with transgenic fly maintenance and dissections. Computational motif discovery was performed by Shamsuddin Bhuiyan, a Ph.D. candidate in the Pavlidis lab, UBC and Dr. Stephane Flibotte. Tianshun performed all EMSAs. Along with Dr. Douglas Allan and Dr. Anthony Berndt, I helped write and revise the manuscript. vii  Chapter 3: “Determining the BMP-dependent regulatory network underlying the early stages of chondrogenesis in the murine limb bud.” is a manuscript in preparation, the title of which is currently tentative. I am the first author on this manuscript, co-authored by Mo Miao, Stephane Flibotte, Shamsuddin Bhuiyan, Ryan Vander Werff, Paul Pavlidis, Michael T. Underhill, and Douglas W. Allan.  Dr. Douglas Allan, Dr. Michael Underhill and I conceptualized the study. I prepared samples for all experiments, conducted the GO term analyses, analyzed the data and prepared figures. RNA-seq sample Bioanalyzer Quality Control and PolyA-RNA-seq library preparation and sequencing were performed by Ryan Vander Werff, at the BRC-Seq Next Generation Sequencing Core, UBC. For the histone ChIP-seq experiments I prepared all samples that were used for chromatin immunoprecipitation and library preparation. These were conducted in the Laboratory of Epigenomics and Chromatin Biology, UBC, under the supervision of Dr. Martin Hirst. Sequencing was performed at the BC Cancer Genome Sciences Centre. RNA-seq and ChIP-seq raw data analysis, as well as the generation of all corresponding quality control figures, were performed by Dr. Stephane Flibotte in the Allan lab. Computational motif discovery was performed by Shamsuddin Bhuiyan, a Ph.D. candidate in the Pavlidis lab, UBC and Dr. Stephane Flibotte. Mo Miao, a graduate student in the Allan lab, assisted with the construction of reporter plasmids and cloning of genomic fragments.  Chapter 4 is a summary of the overall conclusions of my thesis, including contributions and significance to the field, and future directions. viii  Table of Contents Abstract ......................................................................................................................................... iii Lay Summary .................................................................................................................................v Preface ........................................................................................................................................... vi Table of Contents ....................................................................................................................... viii List of Tables ..................................................................................................................................x List of Figures ............................................................................................................................... xi List of Abbreviations ................................................................................................................. xiii Acknowledgements .................................................................................................................... xvi Dedication .................................................................................................................................. xvii Chapter 1: Introduction ................................................................................................................1 1.1 cis-Regulatory Modules and cis-Regulatory Elements .................................................... 1 1.1.1 cis-Regulatory Element Tuning and Specificity ....................................................... 2 1.1.2 Chromatin Modifications and their use in CRM Discovery ..................................... 5 1.2 Bone Morphogenetic Protein Signaling ........................................................................... 9 1.2.1 BMP-CREs, Smads, and their Co-factors ............................................................... 12 1.3 Conservation of Bone Morphogenetic Protein cis-Regulatory Elements ...................... 13 1.3.1 Bone Morphogenetic Protein cis-Regulatory Elements in Drosophila .................. 13 1.3.2 Bone Morphogenetic Protein cis-Regulatory Elements in Vertebrates .................. 16 1.4 Challenges in Studying cis-Regulatory Modules and cis-Regulatory Elements ............ 18 1.4.1 Combining Computational and Experimental Methods to Detect CREs ................ 18 1.5 Thesis Research Objectives and Aims ........................................................................... 21 Chapter 2: Genomic Identification and Validation of a Novel BMP cis-Regulatory Rlement in the Drosophila CNS. ................................................................................................................22 2.1 Introduction .................................................................................................................... 22 2.2 Preliminary Work ........................................................................................................... 24 2.3 Materials and Methods ................................................................................................... 25 2.3.1 Fly Genetics ............................................................................................................ 25 2.3.2 EGFP Reporter Transgene and Transgenic Fly Construction................................. 25 2.3.3 Immunocytochemistry ............................................................................................ 26 2.3.4 Bioinformatic Detection of BMP-AE2 ................................................................... 26 2.3.5 Imaging Analysis and Quantification of Reporter Expression ............................... 28 2.3.6 Statistical Analysis .................................................................................................. 28 2.4 Results ............................................................................................................................ 29 2.4.1 Identification of 128 Highly Conserved AE2 Genomic Fragments in Drosophila 29 2.4.2 Identification of 7 BMP-Responsive Genomic Fragments in Neurons .................. 31 2.4.3 Identification of other BMP-CRE Binding Sites within the BMP-AE2 Fragments 49 2.5 Discussion ...................................................................................................................... 51 2.6 Strengths and Limitations............................................................................................... 53 2.7 Conclusions .................................................................................................................... 55 Chapter 3: Determining the BMP-Dependent Regulatory Network Underlying the Early Stages of Chondrogenesis in the Murine Limb Bud. ................................................................57 3.1 Introduction .................................................................................................................... 57 3.1.1 Chondrogenesis ....................................................................................................... 58 ix  3.1.2 Current Knowledge of Chondrogenic Regulatory Mechanisms ............................. 59 SOX9, the “Master” Regulator of Chondrogenesis ......................................... 59 BMP Signaling in Chondrogenesis.................................................................. 60 Signaling Pathway Crosstalk During Chondrogenesis .................................... 62 cis-Regulatory Elements in Chondrogenesis ................................................... 64 BMP cis-Regulatory Elements in Chondrogenesis ....................................... 65 3.1.3 Project Rationale ..................................................................................................... 68 3.2 Materials and Methods ................................................................................................... 69 3.2.1 Animals and Timed Mating .................................................................................... 69 3.2.2 Primary Cell Culture and Treatments ..................................................................... 69 3.2.3 Alcian Blue Staining ............................................................................................... 70 3.2.4 Plasmids .................................................................................................................. 70 3.2.5 Reporter Construction ............................................................................................. 71 3.2.6 PLM Cell Transfections .......................................................................................... 73 3.2.7 Poly-A RNA-seq ..................................................................................................... 73 3.2.8 qPCR Validation ..................................................................................................... 74 3.2.9 RNA-seq Data Analysis .......................................................................................... 75 3.2.10 Gene Ontology Term Analysis ............................................................................... 76 3.2.11 Histone Modification ChIP-seq .............................................................................. 77 3.2.12 ChIP-seq Data Analysis .......................................................................................... 77 3.2.13 Graphs ..................................................................................................................... 78 3.2.14 Bioinformatic Detection of BMP-CREs ................................................................. 78 3.3 Results ............................................................................................................................ 79 3.3.1 Differential Gene Expression During BMP4-Driven Chondrogenesis ................... 79 Identification of Differentially Expressed Genes Between BMP4, NOG and Control Treatments ........................................................................................................... 81 Gene Ontology Terms used to Identify Genes Regulated via BMP-Signaling 83 Gene Ontology Terms used to Identify Chondrogenic Regulatory Proteins ... 92 3.3.2 Identification of Active Regulatory Regions with Histone Modifications ........... 101 3.3.3 Bioinformatic Motif Discovery and Integration with Genomic Data ................... 107 Prioritization of Regulatory Regions ............................................................. 113 3.3.4 Preliminary Reporter Assay Experiments to Functionalize Identified CREs ....... 121 3.4 Discussion .................................................................................................................... 125 3.5 Strengths and Limitations............................................................................................. 128 3.6 Conclusions .................................................................................................................. 130 Chapter 4: Conclusions .............................................................................................................131 4.1 Significance and Translation of the Study ................................................................... 132 4.2 Future Directions .......................................................................................................... 133 References ...................................................................................................................................136 Appendix .....................................................................................................................................166  x  List of Tables  Table 2.1: Q5 mutagenesis primers .............................................................................................. 26 Table 2.2: Cloned AE2 fragment genomic coordinates and they primer sequences used for their amplification ................................................................................................................................. 27 Table 2.3: Cloned AE2 fragment conservation, genomic location and corresponding nearby genes....................................................................................................................................................... 30 Table 2.4: AE2-enhancer genomic fragment-driven GFP reporter expression patterns. .............. 36 Table 3.1: Reporter Plasmid PCR and Q5 mutagenesis primers .................................................. 73 Table 3.2: Transcription Factors and Cofactors with no chondrogenic GO annotations, upregulated with BMP4 treatment. ................................................................................................................... 88 Table 3.3: Transcription Factors and Cofactors with no chondrogenic GO annotations, downregulated with BMP4 treatment. .......................................................................................... 90 Table 3.4: Transcription Factors and Cofactors with no chondrogenic GO annotations, upregulated or downregulated with NOG treatment......................................................................................... 91 Table 3.5: Transcription Factors and Cofactors with no chondrogenic GO annotations, differentially expressed only at 12hrs ........................................................................................... 96 Table 3.6: Transcription Factors and Cofactors with no chondrogenic GO annotations, differentially expressed only at 24hrs and with RNA-seq |log2foldChange| ≥ 0.55 .................... 97 Table 3.7: Transcription Factors and Cofactors with no chondrogenic GO annotations, differentially expressed at both 12hrs and 24hrs and with RNA-seq |log2foldChange| ≥ 1 ......... 99 Table 3.8: Murine BMP-CREs with Average PhastCons Score > 0.7 ....................................... 107 Table 3.9: BMP-CRE distance from the nearest BMP-responsive gene TSS ............................ 108 Table 3.10: Revisiting and Examining Published Vertebrate BMP Regulatory Elements ......... 110  xi  List of Figures Figure 1.1: Representation of genome browser tracks for histone modifications corelated to gene and enhancer activity. ..................................................................................................................... 8 Figure 1.2: Summary of the BMP Signaling Pathway in Drosophila and mouse......................... 11 Figure 2.1 AE2 motifs with no detectable expression in the CNS of third Instar (L3) larvae.               I.. ................................................................................................................................................... 37 Figure 2.2: Non-BMP responsive AE2 motifs with expression in the VNC of third Instar larvae..38 Figure 2.3: Three of the fragments tested are non wit-responsive in the L3 VNC………………...40 Figure 2.4: wit-mutants exhibit a significant loss in reporter-expressing cells compared to controls in the VNC of third Instar larvae. ................................................................................................. 42 Figure 2.5: wit-mutants and AE2 Mad-binding site mutants exhibit significant loss of reporter-expressing cell numbers compared to controls in the VNC of third Instar larvae. ....................... 44 Figure 2.6: Near complete loss of pMad stain and reporter-expression overlap in pMad-binding site mutant Atyp15mad. .................................................................................................................. 45 Figure 2.7: Significant change in relative reporter intensity was observed in 4 mutant genotypes compared to their respective controls. .......................................................................................... 47 Figure 2.8: Medea is necessary for BMP-dependent activity of 4 wit-responsive BMP-AE2s. ... 48 Figure 2.9: Comparison of two reporter lines with a common BMP-CRE binding site reveals strikingly different expression patterns. ........................................................................................ 50 Figure 2.10: Schematic representation of gene regulation model in the Atyp15 reporter line. .... 56 Figure 3.1: Chondrocyte differentiation ....................................................................................... 63 Figure 3.2: RNA-seq Analysis Pipeline ........................................................................................ 75 Figure 3.3: Key chondrogenic genes respond to our treatments as expected, based on previously published work.. ............................................................................................................................ 80 Figure 3.4: Venn diagrams representing the number and overlap of differentially expressed genes with different treatments and at different time points. .................................................................. 82 Figure 3.5: Workflow for generating the “strict” gene lists from the RNA-seq results. .............. 84 Figure 3.6: Gene Ontology Term analysis of genes from the BMP4-treated list of differentially regulated genes identifies several potentially uncharacterized transcription factors and cofactors that may be involved in chondrogenesis. ...................................................................................... 87 xii  Figure 3.7: Gene Ontology Term analysis of genes from the NOG-treated list of differentially regulated genes identifies several potentially uncharacterized transcription factors and cofactors that may be involved in chondrogenesis. ...................................................................................... 91 Figure 3.8: Gene Ontology Term analysis of genes separated into 3 lists based on their expression time (only at 12hrs, only at 24hrs or at both 12hrs and 24hrs) identifies several potentially uncharacterized transcription factors and cofactors that may be involved in chondrogenesis. .... 95 Figure 3.9: Histone mark enrichment distribution near the TSS of  differentially regulated genes identified via RNA-seq. .............................................................................................................. 104 Figure 3.10: H3K27ac peaks near DEGs at 12hrs and 24hrs post-treatment. ............................ 105 Figure 3.11: RNA-seq, histone modification and Sox9 ChIP-seq for Sox5. .............................. 106 Figure 3.12: Putative BMP-CREs are enriched up to 50kb away from BMP-regulated genes. . 108 Figure 3.13: H3K27ac peaks with at least 1 BMP-CRE motif within 1Mb of a DEG. .............. 113 Figure 3.14: Histone modification marks, SOX9 ChIP-seq and BMP-CREs in the Klf2 locus. 114 Figure 3.15: Histone modification marks in the Grhl1 locus...................................................... 115 Figure 3.16: Histone modification marks and SOX9 ChIP-seq data in the Wwp2 locus. .......... 115 Figure 3.17: Histone modification marks, SOX9 ChIP-seq and BMP-CREs in the Dlx2 locus. 118 Figure 3.18: Histone modification marks, SOX9 ChIP-seq and BMP-CREs in the Msx2 locus.119 Figure 3.19: Histone modification marks, SOX9 ChIP-seq and BMP-CREs in the Jdp2 locus. 120 Figure 3.20: PLM cell transfection with positive control enhancers and minimal promoter Id3 (mpId3) driving eGFP plasmids confirms BMP-driven expression of previously used BMP-regulated enhancers. .................................................................................................................... 123 Figure 3.21: The Jdp2 BMP-AE drives GFP expression in differentiating chondrogenic PLM cells in a BMP4-dependent manner. ................................................................................................... 124     xiii  List of Abbreviations  AE   Activating Element  ACAN  Aggregan  bHLH  basic helix-loop-helix BMP  Bone Morphogenetic Protein  BMPR1A BMP Receptor 1A BMPR1B BMP Receptor 1B BMPRII  BMP Receptor II CDC  Centers for Disease Control and Prevention ChIP   Chromatin Immunoprecipitation  CHRD  Chordin Ci  Cubitus interruptus CNS   Central Nervous System Col2a1 Collagen type II alpha 1 CRE  cis-Regulatory Element CRISPR Clustered Regularly Interspaced Short Palindromic Repeats CRM  cis-Regulatory Module Dad  Daughters against dpp  DEG   Differentially expressed gene dfd  deformed (dfd) enhancer DLX2  Distal-Less Homeobox 2 DNA  Deoxyribonucleic Acid Dpp   Decapentaplegic  ECM  Extracellular matrix  EGFP   Enhanced Green Fluorescent Protein  ENCODE Encyclopedia of DNA Elements EVI1  Ecotropic Viral Integration Site 1 Protein Homolog FAIRE-Seq Formaldehyde-Assisted Isolation of Regulatory Elements Sequencing  FDR  False Discovery Rate xiv  FMRFa FMRFamide (Drosophila neuropeptide) Gbb   Glass bottom boat  GDF  Growth Differentiation Factors  GFP   Green Fluorescent Protein  GMR  eye-specific glass multimer reporter (GMR) enhancer GO  Gene Ontology GWAS Genome-Wide Association Studies HAT  histone acyteltransferase  HDAC  histone deacetylase  HD-RE Homeodomain-Response Element Hh  Hedgehog HMT   Histone methyltransferases  IHH  Indian Hedgehog JDP2  Jun Dimerization Protein 2 Mad   Mothers against dpp  Matn1  Matrilin-1 MEF2  Myocyte Enhancer actor-2 family transcription factors  MH1  Mad Homolog 1  MH2  Mad Homolog 2 MSC  Mesenchymal Stem Cells MSX2  Msh Homeobox 2 NKX  NKX-homeodomain factors NMJ   Neuromuscular junction  NOG  Noggin OA  Osteoarthritis  OAZ  Ornithine Decarboxylase Antizyme 1 PANTHER Protein ANalysis THrough Evolutionary Relationships PLM  Primary Limb Mesenchymal  xv  PRC2  Polycomb Repressive Complex 2 PRMT  Protein arginine methyltransferases PTHrP  Parathyroid Hormone-related Peptide RNA   Ribonucleic Acid  RT  Room Temperature  RUNX  Runt-related transcription factors Sax  Saxophone SBE  Smad binding element  SE   Silencer Element  seq   Sequencing (usually referring to high throughput methods)  Shn  Schnurri  Smad  acronym from the fusion of the C.elegans Sma and Drosophila Mad genes Sog  Short gastrulation  SOX  Sex-Determining Region Y-Box family transcription factors STARR-Seq Self-Transcribing Active Regulatory Region Sequencing STRING Search Tool for the Retrieval of Interacting Genes/Proteins svb  shavenbaby TCF4  Transcription Factor 4 TF  Transcription factor TFBS  Transcription factor binding site TGFβ  Transforming Growth Factor β Tkv   Thickveins  TSS  Transcription start site UTR   Untranslated region  VNC   Ventral Nerve Cord  wit   wishful thinking  WT   Wildtype ZEB2  Zinc Finger E-Box Binding Homeobox 2 Zen  zerknüllt  xvi  Acknowledgements First and foremost, I would like to express my gratitude to my supervisor, Dr. Douglas Allan, for his continuous support during my Ph.D. work. I am grateful for his patience, guidance, motivation and for consistently inspiring me with his passion for science. Thank you for taking a chance on me and for investing all this time and effort to ensure my development as a student and scientist. I could not have imagined having a better advisor for my Ph.D. I would like to extend my appreciation to my supervisory committee members Drs. Tim O`Connor, Michael Underhill and Vanessa Auld, for their time, encouragement and feedback that helped shape this thesis and my progress as a researcher. To our collaborators, the Underhill lab, the Hirst lab, Dr. Stephane Flibotte, Shamsuddin Bhuiyan and Margot Gunning, I am fortunate to have had the opportunity to work with you and learn from you. This research would not have been possible without the passion and knowledge of all of the aforementioned.  I would also like to thank all members of the Allan lab, past and present, for their friendship and support during my time here. Thank you from the bottom of my heart for the countless hours you have spent training, teaching and troubleshooting with me. Special thanks go to past lab members: Anthony, Monica, Kathleen, Justin, Katiana and Ray. To all of my present lab members: Payel, Mo, Robin, Helen, Tian, Mriga and Aarya, I have enjoyed working with all of you and will always cherish our friendships. It is rare to find a lab where you not only enjoy the science, but also appreciate all the people you work with. Last but not least, I would like to thank my friends and family for their continuous support, love and guidance. To my old lab-mates, Sarah, Tissa, Adam and Negar, thank you for helping me when I first moved to Canada and for supporting me through some tough times. To my closest friends and partners-in-crime, Payel, Lakshana and Aarya, thank you for being there for all the ups and downs over the years. I am fortunate to have found such wonderful friends that I can work with every day. To my amazing partner and best friend, Abhinav, thank you for coming into my life so unexpectedly and for keeping me sane during this degree. Words cannot express how much your love, patience, devotion and never-ending support has meant to me over the years. You are the best thing that has ever happened to me and I am excited to see what the future will bring. To my family, thank you for your unconditional love and support through the years. To my baby brothers, Χρήστο, Αλέξη and Μάρκο, thank you for your encouragement and love, even from so far away. To γιαγιά Αναστασία, παππού Σπύρο and γιαγιά Παναγιώτα thank you for believing in me. To θείο Γιώργο, θεία Ελεάνα, Χριστίνα, Σπύρο and Μαρία, thank you for making the move to Canada worth it. Finally, to my parents Αντρέα and Δήμητρα, thank you for being the most encouraging, inspirational and supportive people in my life. Thank you for always believing in me and telling me that I can make all my dreams come true. Your generosity is truly amazing, and I would not have been able to survive this degree without your help. I am forever indebted to you and can only hope that I make you proud every day. xvii  Dedication    to my parents,  for always encouraging me to follow my dreams and to my partner, for keeping me sane and pushing me to always do my best 1  Chapter 1:    Introduction 1.1 cis-Regulatory Modules and cis-Regulatory Elements The focus of my thesis is identifying and experimentally verifying cis-regulatory regions that regulate gene expression in bone morphogenetic protein (BMP)-dependent cellular differentiation. To this end, the next section provides a basic background on the nature of cis-regulatory elements in transcription regulation.  Differential gene expression is a fundamental process that underlies cellular development and differentiation. An important step in this process is the transcription of a DNA sequence into mRNA by the RNA polymerase II (Pol II). Transcription requires the recruitment of Pol II and the formation of the pre-initiation complex (i.e. general TFs, Mediator complex) at the core promoter, which is comprised of genomic sequences around the transcription start site (TSS). However, control of the recruitment of this basal transcriptional machinery and/or clearance of this complex from the promoter to initiate transcription is regulated by transcription factor (TF) complexes bound to non cis-regulatory regions, commonly referred to as enhancers or silencers (Shlyueva et al., 2014). Cis-regulatory elements (CREs) are short 4-12bp DNA binding sites for TFs (Shlyueva et al., 2014; Suryamohan and Halfon, 2015; Guo and Huo, 2017). TFs that bind CREs act to enhance or repress transcription, depending on the factors recruited and/or their post-translational modifications. One of the defining hallmarks of a CRE is that they can influence transcription regardless of their location or orientation relative to the promoter of the gene they regulate (Banerji et al., 1981; Ong and Corces, 2011; Shlyueva et al., 2014). CREs are typically organized in 50bp-2kb clusters (Levine 2010; Cho 2012; Long et al., 2016; Hojo et al., 2017), termed cis-regulatory 2  modules (CRMs, i.e. enhancers, silencers, insulators, promoters; Hardison and Taylor, 2012). CRMs may be up to 1Mb away from the TSS they regulate, in intergenic non-coding sequence or 5′ UTR, 3′ UTR or introns relative to their target genes (Lettice et al., 2008; Spitz and Furlong, 2012; Evans et al., 2012; Hardison and Taylor, 2012; Suryamohan and Halfon, 2015; Long et al., 2016). Thus, there are CREs that are located within a few kb of the promoter, referred to as promoter-proximal CREs, and those that interact with distally located promoters, called distal CREs. DNA looping allows distal CRM complexes to come into proximity with their target promoters, where they help recruit or stabilize interactions with the core transcriptional machinery (Weake and Workman, 2010; Bulger and Groudine, 2011; Beagrie and Pombo, 2016). The modular nature of CRMs allows the integration of signaling and tissue-specific inputs via the assembly of several TFs, which then recruit co-activators or co-repressors and chromatin modifiers, including post-translational modifiers or nucleosome-remodeling complexes (Levine, 2010; Weake and Workman, 2010; Mathelier et al., 2015). The combined regulatory input of all bound factors results in the spatiotemporal-specific regulation of target gene transcription (Levine, 2010; Yanez-Cuna et al., 2013).   1.1.1 cis-Regulatory Element Tuning and Specificity As Chapter 2 deals largely with a novel pSmad CRE binding site that relies upon a low-affinity interaction for its appropriate function, here I will be discussing some of our current knowledge on CRE specificity, how it arises and the use of low-affinity binding sites in transcriptional regulation.  The binding of TFs to CREs is dependent upon the specific DNA sequence of the CRE. Historically, CREs have been defined and searched as high affinity, stringent binding sites. 3  However, most TFs can tolerate a degree of sequence diversity, or degeneracy, at the CRE (Mathelier et al., 2015). Beyond the linear organization, spacing and number of transcription factor binding sites, the relative strength/quality of regulatory binding sites is key for tuning enhancer activity for appropriate expression. For each TF, there is an optimal motif with highly constrained nucleotide composition conferring the lowest possible binding free energy, and thus, making the binding between TF and DNA more stable, favorable and high affinity. Since each nucleotide within a sequence contributes additively to the free energy of the binding site, it stands to reason that there are multiple nucleotide substitutions that can convert a sequence to a lower affinity binding site (Badis et al., 2009; Slattery et al., 2012; Ramos and Barolo, 2013; Farley et al., 2015; Crocker et al., 2016). Thus, CREs that exhibit high degeneracy have been called low-affinity motifs.  A growing body of literature has identified low-to-moderate affinity binding sites that are required for regulating spatiotemporal gene expression. Importantly, in these cases, conversion to a high affinity site often disrupts gene expression, or results in ectopic expression, indicating that low-affinity sites are necessary to specify precise gene expression patterns (Swanson et al., 2011; Farley et al., 2015; Crocker et al., 2015; Lorberbaum et al., 2016; Crocker et al., 2016). One such example is from the Drosophila Cubitus interruptus (Ci) TF that mediates gene expression in response to a Hedgehog (Hh) morphogen signaling gradient (Von Ohlen et al., 1997). Depending on the intensity of Hh signaling, Ci activates or actively represses the transcription of Hh responsive genes. Multiple low-affinity binding sites for Ci in Hh-responsive enhancers were found to be required for normal activation in regions with relatively low Hh signaling. Changing these low-affinity sites into optimal Ci binding motifs resulted in a switch from Ci-mediated activation to Ci-mediated repression (Parker et al. 2011). Therefore, the low-affinity nature of the 4  binding sites contributed to limiting the response to a pleiotropic and broadly active pathway, allowing for adjustment of that response in specific tissues and cells (Ramos and Barolo 2013). Another example is the regulation of the shavenbaby (svb) gene by (homotypic) clusters of low-affinity binding sites for the Ultrabithorax TF in Drosophila. Clustering of many low-affinity sites was shown to confer robustness of svb regulation and replacing these sites with high affinity ones not only resulted in ectopic expression, but it also decreased normal expression patterns (Crocker et al. 2015). Even with sequence tolerance, we cannot explain how a family of TFs, with overlapping or similar sequence recognition, can carry out distinct functions in vivo. In other words, how can each of the nearly 60 basic helix-loop-helix (bHLH) transcription factors that bind to a variation of the sequence “CACGTG” carry out their discrete functions (Murre et al., 1989; Jones, 2004)? Spatiotemporal differences in expression, as well as the small variances in binding sequence preference between TF family members alone, cannot explain how TFs can regulate their unique set of target genes. A mechanism called “latent specificity” could account for the added TF specificity observed in vivo. Based on work with the Drosophila Hox homeodomain TFs, it was shown that individual members of the same TF family acquire novel DNA binding sequence preferences by forming complexes with cofactors (Joshi et al., 2007; Slattery et al., 2011). Further work with a larger set of TFs verified that interactions between cofactors can significantly alter a TF’s sequence preference (Jolma et al. 2015). This added complexity further hinders our ability to identify CREs, as TFs form different complexes under different contexts and bind to degenerate sequences, making it challenging to define universal and computationally identifiable consensus sequences.   5  1.1.2 Chromatin Modifications and their use in CRM Discovery Another layer of transcriptional regulation is epigenetic regulation by modification of chromatin structure. DNA is compacted into repeated structures termed nucleosomes; these are the basic subunits of chromatin and consist of 147bp of packaged DNA wrapped around a histone octamer. Two copies of each H2A, H2B, H3, and H4 core histone proteins form this octamer. The core histones are mainly globular with unstructured N-termini (histone tails). The linker histone H1 binds on the outside and stabilizes the folding on the nucleosome (Luger et al., 1997).  Chromatin structure can be altered to facilitate or suppress transcription; for example, actively transcribed regions tend to be in looser chromatin structures so that transcription factors and RNA polymerases can access the genes. Post-translational modifications on histone tail residues have a large impact, by disrupting chromatin structure or by recruiting non-histone proteins, such as a class of co-factors called histone-modifying enzymes, described below. For instance, acetylation of a lysine residue can neutralize its positive charge and weaken the binding between the histone and the negatively charged DNA, exposing the DNA to regulatory proteins (Kouzarides 2007), and facilitate the recruitment of transcriptional machinery.  Histone modifying enzymes promote covalent modifications of histone protein tails. Out of the many distinct types of modifications (e.g. phosphorylation, ubiquitylation, sumoylation; Butler et al., 2012), here I will only focus on acetylation and methylation since these are commonly used to identify active/repressed genes and enhancers. Histone acetylation was first identified in the 60s (Allfrey et al., 1964) and is characterized by the addition of an acetyl moiety to the ε-amino group of a conserved lysine residue (Postberg et al., 2010). Acetylation is controlled by histone acetyltransferases (HATs) and histone deacetylases (HDACs). As both HATs and HDACs are 6  unable to directly bind to DNA, they are recruited to the promoter or regulatory regions along with other multiprotein complexes by DNA-bound activators. Histone methylation can occur at both lysine and arginine residues. Lysines may be mono-, di-, or tri-methylated and arginines may be mono- or di-methylated. Methylation of lysine on histone subunits H3 and H4 is catalyzed by histone methyltransferases (HMTs), and methylation of arginines is catalyzed by the protein arginine methyltransferases (PRMT) family. (Whetstine 2010; Burke and Grant 2010) Examining histone modification patterns during various stages of cell development in many tissues and organisms revealed that certain marks correlate with specific gene expression or transcriptional states (Bernstein et al., 2006; Calo and Wysocka, 2013) (Figure1.2). While a significant number of histone post-translational modifications have been identified to date (Huang et al., 2014; Zhao and Garcia, 2015), the most extensively studied H3 methylation sites are lysines 4, 9, 27, 36, 79 and acetylation sites at lysines 9, 14, 18, 23, 27, 56 (Lennartsson and Ekwall, 2009; Greer and Shi, 2012). The following section will outline what is currently known about H3 post-translational modifications at K4me1, K4me3, K27ac, and K27me3, as they are widely used to identify active/repressive enhancer function or gene expression.  Methylation of histone H3K4 is predominantly found at CREs and promoters during transcription. Monomethylated H3K4 (H3K4me1) marks active, primed and poised enhancers, while H3K4me3 is found at the promoters/TSS of actively transcribed genes (Santos-Rosa et al., 2002; Bernstein et al., 2006). Histone H3 lysine 27 acetylation (H3K27ac) identifies both actively transcribed gene promoters/TSS and active enhancers (Wang et al., 2008; Tie et al., 2009). In fact, active enhancers bear both the H3K4me1 and H3K27ac marks and are typically bound by the histone acetyltransferase P300 (Heintzman et al., 2007; Heintzman et al., 2009; Creyghton et al., 7  2010; Zentner et al., 2011). Primed enhancers drive low levels of gene expression and are characterized by a lack of H3K27ac and retention of H3K4me1 and P300 (Creyghton et al., 2010; Zentner et al., 2011). Finally, at repressed genes, regulatory elements are enriched for H3K27me3, a mark which is also associated with Polycomb Repressive Complex 2 (PRC2) silencing, and are depleted of other activating modifications (Bernstein et al., 2006; Boros et al., 2014). However, a bivalent modification pattern containing the “opposing” H3K4me3 and H3K27me3 marks was identified in ES cells and found to keep certain developmental genes in a silent but poised-for-activation state (Bernstein et al. 2006). This type of chromatin modification pattern tends to be resolved during ES cell differentiation, with differentiated cells only maintaining one of the two marks on each given gene. Poised enhancers are inactive, and are marked by H3K4me1 and P300 while also bearing the repressive H3K27me3 mark (Rada-Iglesias et al., 2011; Shlyueva et al., 2014). During differentiation, these poised enhancers transition to their active state and H3K27me3 residues replace their methylation with acetylation marks (Rada-Iglesias et al. 2011).  In short, combinations of H3K4me1, H3K27ac, and H3K27me3 are widely used to classify active enhancers (containing both H3K4me1 and H3K27ac), inactive enhancers (containing H3K4me1 but lacking H3K27ac) or poised enhancers (containing both H3K4me1 and H3K27me3) (Heintzman et al. 2007; Creyghton et al. 2010; Rada-Iglesias et al. 2011). 8      Figure 1.1: Representation of genome browser tracks for histone modifications corelated to gene and enhancer activity. Each column represents the transcriptional state of a gene/enhancer region and the histone modification marks we expect to observe in each case. Figure generated using (Pundhir et al., 2015) as source material. 9  1.2 Bone Morphogenetic Protein Signaling BMPs were first identified in 1965 (Urist, 1965), but were only characterized as critical components of bone extract that direct cartilage and de novo bone formation in the late 1980s (Wozney et al. 1988). Since then, multiple studies have confirmed the role of BMPs in cartilage and bone formation, and expanded our understanding of their important functions in the formation, maintenance and regeneration of multiple tissues and organs (Mehler et al. 1997; Whitman 1998; Massague et al., 2000; Liu and Niswander 2005; Wang et al. 2014). As would be expected by these diverse functions, altered BMP signaling is associated with several human diseases including amyotrophic lateral sclerosis, arthritis, osteoporosis, kidney diseases, fibroses, multiple cancers and pulmonary hypertension (Wang et al. 2014).  BMPs belong to the highly conserved transforming growth factor (TGF) β superfamily of signaling molecules, which also includes activins, inhibins and nodal, growth differentiation factors (GDFs) and the TGFβ subfamily (Massagué et al., 2005; Mueller, 2015). Although BMPs can signal via a variety of pathways, they do so primarily through the canonical Smad transduction pathway. In this pathway, BMP ligand dimers bind to a heterotetrameric complex of transmembrane receptors comprised of two dimers each of type I and type II receptor kinases. Ligand binding allows type II receptor kinases to transphosphorylate the glycine/serine rich domain of type I receptors (Miyazono et al., 2010). Receptor engagement results in type I receptor-driven phosphorylation of the conserved C-terminal SXS (Ser-X-Ser) motif of Receptor-Smads (R-Smad), which are Smad1/5/9 in vertebrates and Mad in Drosophila. Phosphorylated R-Smads form a 2:1 complex with their cofactor, the Co-Smad (Smad4 in vertebrates and Medea in Drosophila), that translocates to the nucleus to act as the sequence specific transcription factor pSmad complex (Fig.1.1) (Zawel et al., 1998; Massagué et al., 2005; Mueller, 2015). While in the 10  nucleus, the activated pSmad complex binds to a CRE (BMP-CREs are described in Section 1.3.1) and interacts with various transcriptional co-activators and co-repressors to fine-tune cell-type specific transcription of target genes (Miyazono et al., 2006; Ross and Hill, 2008; Hill, 2016).  R-Smads and co-Smads both have two conserved domains that are connected by a proline-rich linker. These domains are the Mad Homolog 1 (MH1) domain, which contains the sequence-specific DNA binding domain, and the MH2 domain, which contains transactivation and protein-protein interaction domains (Heldin et al., 1997; Kretzschmar and Massagué, 1998). Canonical BMP signaling can be regulated at all steps along the pathway (Zhu et al., 1999; Massagué et al., 2005; Xu, 2006); however, here I will only briefly discuss those directly relevant to this thesis. Extracellular molecules, such as Drosophila Short gastrulation (Sog) and vertebrate Noggin (NOG) act as extracellular antagonists of BMP signaling via sequestering of ligand and suppression of receptor activation. In addition, intracellular regulators of BMP signaling such as inhibitory Smads (I-Smads; Smad6, Smad7 in vertebrates and Dad in Drosophila) can compete with R-Smads for binding to the activated type I receptors and thereby reduce R-Smad phosphorylation (Mulloy and Rider, 2015). Smad complex co-regulators, such as Schnurri (Shn) in Drosophila, will be discussed further in Sections 1.2.2 and 1.3.1. 11                Signaling Component Drosophila Mouse BMP Ligand Decapentaplegic (Dpp), Glass bottom boat (Gbb), Screw (Scw) BMP2/4/5/8 BMP Antagonist Short gastrulation (Sog), Twisted gastrulation (Tsg) Chordin (CHRD), Noggin (NOG) Type I Receptor Thickveins (Tkv), Saxophone (Sax) BMPR1A,BMPR1B,ACVRL1 Type II Receptor Punt (Pnt), Wishful thinking (Wit) BMPRII, ACTRIIA,ACTRIIB R-Smad Mothers against decapentaplegic (Mad) SMAD1/5/9 Co-Smad Medea SMAD4 Inhibitory Smad Daughters against dpp (Dad) SMAD6/7 Figure 1.2: Summary of the BMP Signaling Pathway in Drosophila and mouse.  Schematic representation of the canonical BMP signaling pathway. Table containing the orthologues of relevant BMP signaling components in Drosophila and mouse.  12  1.2.1 BMP-CREs, Smads, and their Co-factors  One of the long-term goals of this project is to identify TFs that act together with Smads to define cell-type specific gene expression. Smads have weak affinity for DNA relative to many other TFs (Shi et al., 1998; Shi et al., 1999), and have often been found to require cooperative or collaborative interactions with other TFs or cofactors for high-affinity and specificity in recruitment to CRE binding sites (Ross and Hill, 2008; Blitz and Cho, 2009). This is seen by many as a mechanistic basis for diversifying Smad transcriptional activities in a wide variety of cellular processes. Homeodomain TFs, such as HOXC8 (Shi et al., 1999) and Zen (Xu et al., 2005), and BMP-activated Smads have been commonly observed to maintain cooperative or collaborative interactions (Henningfeld et al., 2002; Li et al., 2006; Walsh and Carroll, 2007; Zhou et al., 2008; Liang et al., 2012). There are several other well-documented examples of TFs that interact with Smads. Some of these include: transcriptional coactivators like EP300/CREB binding protein (CBP) (Pouponnot et al., 1998) and OAZ (Hata et al. 2000), transcriptional corepressors like ZEB2 (Verschueren et al. 1999) and EVI1 (Alliston et al. 2005), as well as other transcription factors like RUNX2 (Hanai et al. 1999), NKX3-2 (Kim and Lassar, 2003), YY1 (Lee et al., 2004), TCF4 (Hu and Rosenblum, 2005), GATA4 (Brown et al. 2004), GATA5 and GATA6 (Benchabane and Wrana, 2003). More recently, SOX5 was shown to physically interact with R-Smads and that it is essential for the recruitment of Smad1/4 to BMP response elements during Xenopus laevis ectodermal patterning (Nordin and LaBonne, 2014). While we still do not have a comprehensive list of all potential cofactors that cooperate with Smads, according to the BioGRID database (version 3.5.170; Oughtred et al., 2019) it is estimated that human SMAD1, SMAD5, and SMAD9 interact with 126, 62 and 114 proteins, respectively. While these estimates may not be completely 13  reliable, due to potential false-positives, it opens up the possibility of identifying many more Smad coregulators.   1.3 Conservation of Bone Morphogenetic Protein cis-Regulatory Elements BMP signaling is highly conserved from invertebrates to vertebrates and used to regulate a diverse array of processes across development (Schmierer and Hill, 2007; Blitz and Cho, 2009). Likewise, evidence suggests that BMP-CREs are also highly conserved. Studies have shown that BMP-CREs from one organism can recapitulate the function of autologous enhancers in another organism. For instance, Weiss et al. showed that a multimerized BMP-CRE from the Drosophila Dad gene enhancer induced reporter expression in zebrafish embryos (Weiss et al., 2010). Additionally, the Allan lab demonstrated that a BMP-CRE taken from the Xenopus bambi BMP-responsive enhancer functionally replaced a fly BMP-CRE in vivo and a multimerized fly Dad BMP-AE could drive reporter expression in chick spinal cord (Vuilleumier et al. 2018).  The following subsections discuss current knowledge on BMP-CREs in Drosophila and vertebrate systems.  1.3.1 Bone Morphogenetic Protein cis-Regulatory Elements in Drosophila The pMad-binding consensus motif GCCGnCGC was originally identified in Drosophila in the promoters of Vestigial (Kim et al., 1997) and Tinman (Xu et al., 1998). Since then, work primarily in the wing imaginal disc has identified two bipartite BMP-CREs sequences for Mad and Medea binding found in the regulatory regions of Drosophila BMP-regulated genes. These are termed the BMP-Silencer Element (BMP-SE) (Pyrowolakis et al., 2004) and the BMP-Activation Element (BMP-AE). These motifs have since been found to be operational in other 14  tissues of developing Drosophila; however, it is notable that less than 10 of these motifs had been identified and functionally analyzed prior to recent work by the Allan lab in the fly nervous system (Vuilleumier et al. 2018). Thus, we still do not know how broadly they are used to regulate gene expression in other tissues. The BMP-SE and BMP-AE motifs operate to pattern BMP-regulated gene expression across a gradient of BMP ligand availability across the developing wing imaginal disc. Phosphorylated Mad (pMad; Drosophila R-Smad) binds to GRCGNC or GGCGCC motifs, while Medea (Drosophila co-Smad) binds to GTCT or GNCV. Two additional regulators are also recruited to these motifs, the conserved co-regulator Schnurri and the Dipteran specific repressor Brinker, that play critical roles in BMP-dependent developing wing imaginal disc patterning. BMP-Silencing Element (BMP-SE)                 GRCGNC(N5)GTCT                            (R= G or A) BMP-Activating Element (BMP-AE)              GGCGCCA(N4)GNCV                     (V= A, C or G) Recruitment of the conserved co-repressor Schnurri requires the N5 spacer and the terminal thymidine of GTCT of the BMP-SE. This recruitment makes the BMP-SE a repressor motif. The lack of Schnurri recruitment to the BMP-AE, due to the lack of the terminal thymidine in the GNCV motif, allows the pMad and Medea complex to activate gene expression (Pyrowolakis et al., 2004; Gao et al., 2005; Weiss et al., 2010). In many tissues of the fly, a major gene silenced by the BMP-SE is brinker. Brinker is itself a sequence-specific transcriptional repressor. In the wing disc and other tissues, brinker repression by BMP signaling results in de-repression of genes that would be otherwise repressed by Brinker. Notably, many BMP-activated genes in the developing wing imaginal disc are in fact de-repressed via BMP-dependent brinker repression (Sivasankaran et al., 2000; Zhang et al., 2001; Weiss et al., 2010). However, Brinker not only acts 15  to promote BMP-dependent gene activation through de-repression but also competes with pMad binding to BMP-AE motifs to counteract pMad/Medea-driven gene activation from this motif. Brinker binds sequences with the consensus sequence GGCGYY (Y=C, T). This overlaps with the pMad binding site of the BMP-AE (GGCGCC), serving as a template for competitive binding to this motif, placing gene activation from the BMP-AE under the control of the relative expression of pMad and Brinker. This serves to finely tune gene expression across the BMP ligand gradient of the developing wing imaginal disc (Rushlow et al., 2001). In addition to the work in the imaginal wing disc, the Allan lab has recently verified the widespread use of the BMP-AE motif in the Drosophila CNS, by identifying and experimentally verifying 34 BMP-AE containing genomic fragments near BMP-activated genes (Vuilleumier et al. 2018). This is the first study to show that the BMP-AE motif is utilized by a large number of BMP-responsive genes in any tissue.  Despite mounting evidence that the BMP-SE and BMP-AE are widespread motifs governing BMP-regulated genes, several diverse pMad/Medea-responsive motifs have been identified in Drosophila (Xu et al., 1998; Rushlow et al., 2001; Lin et al., 2006; Walsh and Carroll, 2007; Deignan et al., 2016). The deviation of some of these motifs from the AE/SE consensus sequences may be required in order to enable pMad/Medea interactions with other TFs and cofactors (Ross and Hill, 2008; Blitz and Cho, 2009). Adding to this body of work, we have identified and characterized a novel BMP-CREs motif that we have termed the BMP-AE2 in the Drosophila CNS (see Chapter 2). This motif has a low affinity, compared to the BMP-AE and BMP-SE, and our evidence indicates that this lowered affinity is required for cell-specific activation of the nearby gene. Additionally, biochemical studies have shown Smad complex binding to the bipartite BMP-CRE sequence with varying linker spacing, demonstrating that linker ranges of 5±1bp and even 5+ multimers of 10bp may be tolerated (Gao and Laughon, 2007). This 16  further indicates that novel BMP-CREs with varying sequences or linker site lengths may exist and should be further investigated. Studying the deployment and roles of these varied motifs will deepen our understanding of BMP-dependent gene regulation. 1.3.2 Bone Morphogenetic Protein cis-Regulatory Elements in Vertebrates Early insight into the DNA-binding specificity of human Smads came from oligonucleotide binding screens, where the palindromic 5`-GTCTAGAC-3` termed Smad binding element (SBE), was identified as a binding site for Smad4 (vertebrate co-Smad) and Smad3 (TGFβ pathway R-Smad) (Zawel et al. 1998). This was further corroborated by the identification of SBE sites that bind Smads and regulate the expression of genes like Nkx2.5, the mammalian homologue of the BMP-responsive Drosophila transcription factor Tinman (Lien et al., 2002; Brown et al., 2004). Furthermore, the X-ray crystal structure of the N-terminal MH1 domain of Smads 1, 3, 4 and 5 indicated that they all recognize and bind to the GTCT motif (Shi et al., 1998; Baburajendran et al., 2010; Baburajendran et al., 2011; Chai et al., 2015).  Though able to bind to the SBE (GTCT) motif as cited above, Smad1 binds to GC-rich sequences with higher affinity (Kim et al., 1997; Korchynskyi and Ten Dijke, 2002). X-ray crystallography of the Smad5 MH1 domain confirms that it binds to GC-rich sequences (in addition to the SBE sequence) (Chai et al. 2015). Additionally, canonical BMP pathway Smads1/5/9 were shown to bind GC-rich sequences (such as GCCG, GGCGCC), termed a BMP Response Element (BRE), near well-known BMP-regulated genes such as the Id1, Id2, Id3 and Id4 genes (Ishida et al., 2000; López-Rovira et al., 2002; Korchynskyi and Ten Dijke, 2002; Karaulanov et al., 2004). There are many variations of this GC-rich BRE sequence, for example, 17  a ChIP-seq study mapping SMAD1/5 occupancy in primary human cells revealed enrichment for both GGCGCC and GGAGCC sequences (Morikawa et al. 2011).  Akin to Drosophila BMP-CREs, bipartite arrangements of SBE and GC-rich sequences (i.e. TGGCGGC(N5)GTCT) have been observed in the regulatory regions of numerous vertebrate BMP-regulated genes such as Ihh, Bambi, Smad7 (I-Smad) and Id genes (Katagiri et al., 2002; Korchynskyi and Ten Dijke, 2002; Karaulanov et al., 2004; Seki and Hata, 2004; Nakahiro et al., 2010). A notable difference between the fly and vertebrates is that these BMP-CREs serve as activators (Yao et al., 2006), despite the fact that these are BMP-SE silencers in the Drosophila wing imaginal disc and early embryonic blastoderm (Gao et al., 2005; Pyrowolakis et al., 2004; Weiss et al., 2010). For a more detailed description of these vertebrate BMP-CREs see Table 3.10 in Chapter 3, Section 3.3.3. Recently, published work by the Allan lab demonstrated that a multimerized fly BMP-AE-type motif functions as a functional activator element in the vertebrate CNS (Vuilleumier et al. 2018). This opens the possibility that BMP-AE-type BMP-CREs may regulate BMP-responsive genes in vertebrates.   If these more ‘complex’ bipartite motifs are widely utilized in vertebrates, as opposed to a degenerate “GC-rich” sequence, this would improve our ability to use computational methods to detect putative BMP-CREs, as was previously done in Drosophila. However, the question remains as to the fraction of BMP-CREs that have this bipartite structure and conform to the BMP-SE consensus. This has not been examined directly.  Experimentally validating the existence of diverse bipartite BMP-CRE sequence types, perhaps with different functions, will be an important step in understanding BMP-gene regulation in vertebrates. Identifying discrete sequences and the extent to which each of these is utilized in 18  vertebrates will help us further elucidate BMP regulatory mechanisms, as well as improve our ability to accurately detect such motifs by purely computational methods. My project partially aims to identify such motifs, and ultimately test their contribution to the BMP-regulatory landscape.   1.4 Challenges in Studying cis-Regulatory Modules and cis-Regulatory Elements  Mutations in non-coding CREs are known to cause or contribute to human disease, resulting in a growing class of diseases classified as “enhanceropathies” (Smith and Shilatifard, 2014; Mathelier et al., 2015). Mapping BMP-CREs on a genomic scale and defining their sequence variation tolerances would help interpret disease susceptibility arising from sequence variants in non-coding regions (Mathelier, Shi, and Wasserman 2015). CRE discovery has been historically challenging, as CREs are located at varying distances to the genes they regulate and can have high sequence degeneracy. However, their crucial role in gene regulation makes their identification very important, particularly in light of human genome sequencing efforts to identify disease susceptibility loci in the genome (Wasserman and Sandelin, 2004; Petersen et al., 2017).  1.4.1 Combining Computational and Experimental Methods to Detect CREs Several computational and experimental methods have been developed to predict and identify CREs; however, none of these strategies alone can confidently identify functional CREs. In this section, I will be discussing the advantages and disadvantages of using some of these methods.  Computational methods include enrichment and clustering of characterized transcription factor binding sites (TFBS), as well as comparative genomics based on sequence conservation 19  between species. TFBS are typically short (4-12bp), and such short motifs can be frequently found in the genome, with only a few functioning as in vivo binding sites. Searching for clusters of short TFBS motifs can improve predictions but these can also be widespread in large genomes, resulting in many false-positive predictions. Chromatin context, cell type, developmental stage, as well as the presence or absence of regulatory co-factors, all contribute to whether a predicted binding site is occupied in vivo (Gaulton et al., 2010; Kaplan et al., 2011), making it almost impossible to experimentally verify a false-positive prediction as false. The high variability in TFBS sequences, especially when TFs act cooperatively (Jolma et al. 2015) can also result in a high degree of false negatives. Incorporation of sequence conservation information can improve prediction accuracy, assuming that important non-coding functional DNA sequences are more likely to be conserved across multiple species during evolution. However, CRE evolution is a driving force of evolutionary change, hence phylogenetic conservation can also be misleading (Long et al., 2016; Douglas and Hill, 2014; Rebeiz and Williams, 2011). Additionally, as many computational methods of enhancer prediction rely on the strongest biochemical TF binding signals (high-affinity binding sites), they typically fail to detect low to moderate affinity binding sites (Jaeger et al., 2010; Ramos and Barolo, 2013).  Experimentally, one of the ways occupies CREs can be identified is by using ChIP-seq either against transcription factors or against histone post-translational modifications characteristic of active transcription (Hardison and Taylor, 2012; Shlyueva et al., 2014). This provides experimentally verified, unbiased, genome-wide detection of CREs in a systematic manner. Using ChIP-seq, we can recover binding sites of a transcription factor with a resolution of 150-300bp (Park 2009). However, this method only allows the study of single factors in a snapshot of time and space. This hinders the detection of cell-specific CREs unless they are directly examined 20  (Schmidt et al. 2010). An additional caveat of this method is the recovery of a substantial number of candidate CREs, many of which are false positives and do not contribute to gene regulation. Finally, it is possible that more “rare” binding sites can be dismissed as false-positives or “background”. In order to better discriminate functional TF binding sites, one can also assess chromatin accessibility and histone modifications that mark genomic regions areas by their functional status, such as the enrichment of H3K4me1 and H3K27ac. H3K4me1 enrichment at enhancers can be quite broad, extending 1kb or more on either side of the CREs, thereby providing a large sequence area to identify the actual TF binding site (Heintzman et al., 2007; Creyghton et al., 2010).  In isolation, any single strategy is limited in its ability to confidently predict functional CREs. However, predictions can be improved by the combined use of ChIP-seq data for TF occupancy and chromatin modifications, coupled with TF binding motif models and multi-species sequence comparisons. In addition to data generated in-house, public datasets such as those available through the ENCODE projects can be added to enhance CRE prediction. These available datasets include, but are not limited to, DNA-seq and FAIRE-seq data (from various cell types and conditions) to identify nucleosome-depleted and open chromatin regulatory regions (Dailey 2015), as well as STARR-seq data which are massively parallel functional assays for the discovery of active enhancers in vertebrates and invertebrates (Arnold et al., 2013). Combining all these complimentary methods to identify and experimentally verify TF binding sites will markedly enhance the predictive power of any CRE characterization project.   21  1.5 Thesis Research Objectives and Aims Despite the importance of BMP-CREs to BMP-driven genomic responses and the potential for BMP-CRE sequence variants in disease susceptibility, they remain largely unidentified on a genomic scale. The overall objective of my thesis is two-fold; to characterize the widespread deployment of a novel low-affinity BMP-CRE motif our lab found to be active in the Drosophila central nervous system and to start the process of identifying the BMP-driven gene regulatory network underlying mammalian chondrogenesis.    My specific experimental aims were to: 1)  Experimentally validate a novel set of BMP-CRE motifs in Drosophila. 2)  Perform genomic identification of BMP-regulated genes and active enhancer regions during mouse chondrogenesis. 3) Experimentally verify candidate BMP-CREs that are active during mouse chondrogenesis.             22  Chapter 2:                                                                                                         Genomic Identification and Validation of a Novel BMP cis-Regulatory Element in the Drosophila CNS.  2.1 Introduction Neurogenesis is a complex and tightly regulated process driven by local extrinsic cues and genetically encoded intrinsic programs that control cell fate (Hobert et al., 2010; Allan and Thor, 2015). Late stages of neural circuitry differentiation and maturation, such as the formation and growth of synapses, arborization of axons and dendrites and acquisition neurotransmitter phenotype, also require “retrograde” signaling from dendritic and axonal targets (Hippenmeyer et al., 2004; Marqués, 2005; da Silva and Wang, 2011). A well-studied example of retrograde signaling in Drosophila is retrograde BMP signaling in efferent neurons. Retrograde BMP-signaling is required for the induction and maintenance of neuronal subtype-specific gene expression including neuropeptide genes, such as FMRFa, in developing and mature neurons in the VNC (Allan et al., 2003; Marqués et al., 2003; Eade and Allan, 2009; Veverytsa and Allan, 2011). Additionally, retrograde BMP signaling is required in motor neurons for increased neuromuscular junction (NMJ) growth and neurotransmission to homeostatically match muscle growth (Aberle et al., 2002; Marqués, 2005; Goold and Davis, 2007; Berke et al., 2013).  In Drosophila efferent neurons, retrograde BMP-signaling is triggered by Glass bottom boat (Gbb) ligand engagement at a presynaptic BMP-Receptor complex of Wishful thinking (Wit), Thickveins (Tkv) and Saxophone (Sax) that phosphorylates Mad (Aberle et al., 2002; Rawson et al., 2003; Allan et al., 2003; Mccabe et al., 2004). Phospho-Mad binds Medea to form the Smad complex (Gao, Steffen, and Laughon 2005) Gao and Laughon, 2006) that translocates to the 23  nucleus and regulates transcription by binding DNA in a sequence-specific manner at BMP-response elements (Kim et al., 1997; Xu et al., 1998; Shi and Massague, 2003; Berndt et al., 2015; Vuilleumier et al., 2018). The role of retrograde BMP-signaling in neuronal gene regulation is best defined for the neuropeptide gene FMRFa. This gene is selectively initiated and maintained in Tv4 neurons through the integration of BMP-activated Smad transcription factors with a Tv4-specific transcription factor code at two closely spaced cis-elements, a homeodomain-response element (HD-RE) and a BMP-response element (BMP-CRE, within a Tv4-neuron specific enhancer of the FMRFa gene (Allan et al., 2003; Miguel-Aliaga et al., 2004; Allan et al., 2005; Berndt et al., 2015).  These results lead to a model wherein retrograde signaling contributes a BMP input to complete a combinatorial transcription factor code that determines selective gene expression in differentiating neurons (Allan et al., 2003; Berndt et al., 2015).  Studies primarily in the Drosophila wing imaginal disc have established core principles regarding BMP-CRE sequence and function. Drosophila BMP-dependent gene activation is mediated directly by Smads acting as activators, or indirectly through de-repression whereby BMP/Smad-dependent repression of the transcriptional repressor brinker reduces its activity, and de-represses numerous genes (Affolter and Basler, 2007; Hamaratoglu et al., 2014). The existence of activating/repressing BMP cis-Response Element motifs (see Section 1.3.1) in Drosophila suggests that the nature of the BMP-mediated gene regulation can be textured by the specific BMP-CRE sequence contained within a gene’s cis-regulatory region. Therefore, in addition to a model where the BMP-CRE is a simple docking site for BMP input to a combinatorial transcription factor code, a second model presents itself in which the BMP-CRE sequence plays 24  an additional functional role to diversify the regulatory outcomes of BMP input. These models have not been tested for their relevance in diversifying BMP-dependent gene expression in neurons. 2.2 Preliminary Work This section summarizes work performed by a former graduate student, Anthony Berndt, upon which my work for Chapter 2 builds on, for manuscript submission. In studying the highly restricted expression of the FMRFa neuropeptide gene, our lab identified a minimal BMP-responsive cis-regulatory element (BMP-CRE) that mediates BMP-dependent FMRFa gene transcription in the 6 Tv4 neurons of the Drosophila VNC (Berndt et al., 2015). Follow-up work revealed that this BMP-CRE is a novel Activating-like Element (BMP-AE2; GGCGCCA(N4)GTAT) that differs from the previously identified BMP-SE and BMP-AE sequences (see Section 1.3.1), mainly by virtue of a C > A transversion of a functionally critical C nucleotide in the Medea binding site (GTAT vs. GTCT/GNCV), that confers reduced Smad recruitment compared to BMP-AE (GNCV) or BMP-SE-type (GTCT) motifs (see attached manuscript for details). Regardless of reduced Smad recruitment, functional testing of BMP-AE2 mutants demonstrated that this atypical motif directly mediates Smad recruitment and FMRFa expression, without the contribution of brinker or schnurri co-regulators. Moreover, conversion of GTAT to an optimal BMP-AE sequence GNCV-type motif (GACG) enhanced Smad recruitment in vitro and led to substantial ectopic BMP-dependent reporter expression in other neuronal populations. This indicates that the low affinity of the BMP-AE2 motif directly contributes to the high neuronal subtype-specific expression of BMP-dependent FMRFa. Mutational analysis further revealed that the T in position two of the Medea binding site (GTAT) is the only nucleotide that 25  can be replaced without eliminating reporter activity. Taken together, this suggests that a BMP-AE2 consensus sequence (GGCGCCA(N4) GNAT) may be a novel BMP-CRE motif.  2.3 Materials and Methods 2.3.1 Fly Genetics Strains used: med C246 (McCabe et al., 2004), med13 (Hudson et al., 1998), med1 (Das et al. 1998), witA12 and witB11 (Aberle et al. 2002) were provided by the Bloomington Drosophila Stock Center. Mutants were kept over TM3, Sb, dfd-GMR-nvYFP (Le et al. 2006) TM3,Sb,Ser,twiGAL4,UAS-2xEGFP (Halfon et al. 2002) or TM6B (Craymer 1984). w1118 was used as the control genotype. Flies were reared on standard medium 25°C, 70% humidity.  2.3.2 EGFP Reporter Transgene and Transgenic Fly Construction  A list of all primers used for the construction of these transgenic flies can be found in Table 2.1. To generate reporter constructs, approximately 2kb genomic DNA fragments (see Table 2.1) were amplified by PCR and cloned into an empty pThunderbird EGFP vector (Berndt et al. 2015). These fragment sizes correspond with those tested in other studies performing large scale enhancer identification in Drosophila (Weiss et al., 2010; Kvon et al., 2014; Pfeiffer et al., 2008; Vuilleumier et al., 2018) and are sufficiently large fragments to reduce the chance of excluding important enhancer elements on either side of the putative BMP-CREs. Coordinates of the fragments amplified, as well as the primers used for fragment amplification from genomic DNA,  are summarized in Table 2.2.  Mutagenesis was performed by Q5® Site-Directed Mutagenesis Kit (New England Biolabs), using primers designed to introduce specific base pair substitutions to the Mad binding site (GCCGGC 26  > tgatga), according to manufacturer`s protocols. Primers were designed using the NEBase Changer v1.2.8 tool and are summarized in Table2.1. All constructs were verified by sequencing before the generation of transgenic fly lines.  For fly transgenesis, constructs were inserted via PhiC31 mediated site-specific integration (Bischof et al., 2007) into the attP2 site (Groth et al, 2004) on chromosome 3 by Rainbow Transgenics Flies Inc. (CA, USA).  Table 2.1: Q5 mutagenesis primers  2.3.3 Immunocytochemistry Standard protocols were used throughout (Eade and Allan, 2009; Vuilleumier et al., 2018). Primary antibodies used: rabbit anti-pSmad1/5 (1:100; 41D10; Cell Signaling Technology). Secondary antibodies used: donkey anti-rabbit conjugated to Cy3 or Cy5 (1:250; Jackson ImmunoResearch).  2.3.4 Bioinformatic Detection of BMP-AE2 HOMER v4.10 software suite (Heinz et al. 2011) was used to scan the reference dm6 Drosophila genome for the BMP-AE2 (GGCGCCN5GTAT) motif. Base-specific PhastCons scores (Siepel et al., 2005; 27-insect conservation) were obtained from the UCSC genome browser (https://genome.ucsc.edu/)  to annotate each motif instance with evolutionary conservation scores.  DNA fragment Primer Sequence  (5`→ 3`)  Mutated BMP-AE2 site  Atyp3 F: GGCGTATACGTGATAAGTGGCGCTA tgatgaATGGCGTAT R: ATtcatcaGTCACCAGATTCGGTGGTT Atyp8 F: ACTGTATCTGTATCTGTGTTTCTTTTTTTTG tgatgaAGACTGTAT R: CTtcatcaTCTTGCAGCTCGAATGTG Atyp9 F: TTTGTATTTCGAGGGTAAGGCCGAA tgatgaATTTTGTAT R: ATtcatcaAATGGGTTTCAATGGTCCC Atyp11 F: TtcatcaTCTAATGTTTGCTCGGTTTG tgatgaAAAAAGTAT R: AAAAGTATGCAACACTACAGAAGCCC Atyp15 F:  TtcatcaAGCACTTTGTCGTGGGCG tgatgaAAAAAGTAT R: AAAAGTATGCTATAATATTTAAAGCTCACGCAAGC 27  Table 2.2: Cloned AE2 fragment genomic coordinates and they primer sequences used for their amplification 28  2.3.5 Imaging Analysis and Quantification of Reporter Expression Images were acquired as z-stacks on an Olympus FluoView FV1000 confocal microscope or a Zeiss Axio Imager VIS LSM880 confocal microscope. We examined native GFP reporter expression (without anti-GFP immunoreactivity enhancement) in late L3 larval VNCs, in the context of anti-pMad immunoreactivity (to mark nuclei with active BMP-signaling). In all cases, four or more VNC were dissected and imaged for each genotype. All tissues to be compared were processed with the same reagents, imaged and analyzed in identical ways. To quantitate reporter activity, we used Bitplane: Imaris v9.2 software (in Spots Mode) to identify reporter-positive nuclei in the VNC (excluding the brain lobes). When comparing control and pMad-binding site mutant genomic fragment reporters, we additionally assessed GFP positive nuclei that were co-marked by pMad immunoreactivity (by mean intensity thresholding). Imaris settings were established independently for each set of reporters, in order to provide optimal ‘spot’ marking of the GFP reporter and pMad co-immunoreactive nuclei, with minimal background fluorescence spot marking. Each image was further subtracted, manually, for spots that erroneously labelled background fluorescence. 2.3.6 Statistical Analysis  Statistical analysis and graphing were performed with GraphPad Prism Version 8.0.1 (GraphPad Software, San Diego, CA). The normality of sample distribution was determined with Shapiro-Wilk normality tests. All multiple comparisons were done with One-Way ANOVA and a Tukey post-hoc test or Student’s two-tailed t-test when only two groups were compared. Mann-Whitney U test was used when the samples were not normally distributed. Differences between genotypes were considered significant when p-value < 0.05. Data are presented as mean ± Standard Error of Mean (SEM). 29  2.4 Results 2.4.1 Identification of 128 Highly Conserved AE2 Genomic Fragments in Drosophila Following the identification of the novel BMP-AE2 regulating FMRFa expression in Tv4 neurons, we examined if this cis-regulatory motif was widely distributed through the genome to regulate numerous neuronal genes in a BMP-dependent manner. Therefore, I examined the distribution of conserved BMP-AE2 motifs in the genome and tested the BMP-dependent activity of a 2kb genomic fragment surrounding the motif, following established methods (Vuilleumier et al. 2018). We identified all 178 BMP-AE2 motifs in the D. melanogaster genome using the motif discovery tool HOMER (v4.10) (Heinz et al. 2011). These were filtered for high sequence conservation across 24 sequenced Drosophila species using PhastCons scores, limiting the list to 128 BMP-AE2 with an average PhastCons score over 0.55 (Supplemental Table S1). Out of these 128 motifs, we found that 68% (87/128) are located within an intron or UTR, while the remaining 32% (41/128) are intergenic relative to the nearest gene. Towards selecting a subset of these for functional testing in vivo, we prioritized 24 of the BMP-AE2 motifs based on their proximity to genes expressed in the larval nervous system (Chintapalli et al., 2007; Weiszmann et al., 2009) (see Table 2.3). Out of the 24 selected motifs, we selected 20 highly conserved with an average PhastCons score above 0.9 in order to optimize the chance of characterizing functional BMP-CRE's, but also added 4 motifs with scores 0.75-0.55 in order to test of lesser conserved motifs were also functional (Table2.3, Supplemental Table S1). We additionally identified the nearest wit-responsive gene TSS for each motif (unpublished data, Allan lab). These data are included in Table 2.3. 30  Table 2.3: Cloned AE2 fragment conservation, genomic location, and corresponding nearby genes 31  2.4.2 Identification of 7 BMP-Responsive Genomic Fragments in Neurons To test the in vivo activity of these BMP-AE2s, 24 genomic DNA fragments (of∼2kb) each containing one of these motifs were cloned in front of a minimal promoter and a nuclear-localizing GFP reporter.  We examined reporter activity driven from these genomic fragments in wandering third instar larvae. Out of the 24 reporters, 5 showed no expression in the VNC, but only 1 of these exhibited no reporter expression in any of the other tissues examined (Table 2.4, Figure 2.1). The remaining 19 reporters exhibited low to robust reporter activity in the VNC (Table 2.4). Additionally, a subset of the 19 reporters exhibited reporter expression in the optic lobes and/or other larval tissues, indicating that these reporters may not only have a restricted function in the CNS but may have roles outside the CNS during larval development. Of these active reporters, 10 exhibited expression in subsets of pMad-positive cells in the VNC (which at this developmental stage comprises motor and neuropeptidergic neurons) (Allan et al., 2003), as well as pMad-negative glia and neurons (Figures 2.3, 2.4, 2.5). The other 9 active reporters only exhibited reporter expression in pMad-negative glia and neurons (Figures 2.2). We tested the BMP-responsiveness of the 10 reporters expressed in subsets of pMad-positive cells reporters, by placing all 10 into a wit mutant background. The transgenic flies were crossed to a wit mutant allele line (witA12) to create recombinant animals that would contain both the enhancer reporter and wit mutant allele on the third chromosome. As third instar larvae heterozygous for witA12 are viable and have normal BMP signaling in the VNC (pMad+ stain) they were used as controls (witA12/+). The corresponding wit mutants were made by crossing with another wit mutant line (witB11), generating witA12/witB11 transheterozygotes that are pupal lethal 32  (Veverytsa and Allan, 2011) and lack normal BMP signaling and pMad stain in the VNC. Out of those 10 reporters, 7 showed partial to total loss of reporter expression, and 3 reporters showed no significant change (Figures 2.3, 2.4, 2.5). Three of these wit-responsive fragments were Atyp18, Atyp23, and Atyp26 (Figure 2.4). Atyp18 reporter was expressed in 339±15 nuclei, of which 134±25 (34%) nuclei were pMad-positive (n=16). In wit mutants, reporter activity was reduced by 47% to 181±5 nuclei (n=17) (Figure 2.4-A,D). Atyp26 reporter had very robust and widespread expression across the whole L3 brain. In the VNC it was expressed in 805±22 nuclei, of which 217±10 (27%) nuclei were pMad-positive (n=6). In wit mutants, reporter activity was reduced by 30% to 592±39 nuclei (n=6) (Figure 2.4-B,E). Finally, Atyp23 had a sparser expression pattern and was expressed in 61±6 nuclei (n=5). In wit mutants, reporter activity was reduced by 50% to 30±5 nuclei (n=5) (Figure 2.4-C,F). We next tested whether the activity of the identified wit-responsive fragments was dependent on the AE2 Mad binding site included in these fragments. We selected 4 wit-responsive fragments (Atyp3, 8, 11 and 15) and one non-wit responsive fragment (Atyp9, Figure 2.3-C,F), to introduce specific mutations into the pMad-binding site of the AE2 (GGCGCC > TGATGA). We placed these mutant genomic fragment reporters into the same attP2 site as the corresponding WT reporter. In all 4 wit-responsive fragments tested, there was a significant loss of reporter expressing cells; however, this loss was less pronounced than the loss in wit mutants (Figure 2.5-A-C, E-G). Atyp15mad was the only fragment with the same loss of cell numbers in the pMad-binding site mutant compared to the wit mutant (Figure 2.5-D,H).  33  We quantified the number of cells in which reporters were expressed for these 4 wit-responsive fragments (Atyp3, 8, 11 and 15). In the VNC, Atyp3 reporter was expressed in 224±16 nuclei, of which 56±6 (25%) nuclei are pMad-positive (n=7). In wit mutants, reporter activity was reduced by 61% to 88±10 nuclei (n=12). Once the pMad-binding site was mutated (Atyp3mad), reporter activity was reduced by 27% to 159±9 nuclei (n=15) compared to controls that had reporter expression in 218±19 nuclei (n=6) (Figure 2.5-A,E). Atyp8 reporter was expressed in 49±5 nuclei, of which 9±2 (18%) nuclei are pMad-positive (n=5). In wit mutants, reporter activity was reduced by 75% to 12±1 nuclei (n=5). In the pMad-binding site mutant (Atyp8mad), reporter activity was reduced by 50% to 23±3 nuclei (n=7) compared to controls that had reporter expression in 46±5 nuclei (n=6) (Figure 2.5-B,F). Atyp11 reporter was expressed in 228±17 nuclei, of which 118±10 (52%) nuclei are pMad-positive (n=10). In wit mutants, reporter activity was reduced by 88% to 24±8 nuclei (n=9). In the pMad-binding site mutant (Atyp11mad), reporter activity was reduced by 41% to 131±8 nuclei (n=9) compared to controls that had reporter expression in 221±15 nuclei (n=6) (Figure 2.5-C,G). Finally, Atyp15 reporter was expressed in 212±17 nuclei, of which 68±7 (32%) nuclei are pMad-positive (n=9). In wit mutants, reporter activity was reduced by 64% to 77±5 nuclei (n=9). In the pMad-binding site mutant (Atyp15mad), reporter activity was reduced by 67% to 76±7 nuclei (n=9) compared to controls that had reporter expression in 228±17 nuclei (n=9), of which 68±8 are pMad-positive (Figure 2.5-D,H). Coincidently, Atyp15 reporter is the only case where a near complete loss (90%) of pMad+ stain and reporter expression overlap were observed in the pMad-binding site mutant (Figure 2.6). Taken together, the data strongly suggest that these 4 motifs are functional BMP-AE2 where binding to the Mad-binding site plays an important role in driving reporter expression in the VNC. 34  Next, we examined reporter intensity effects in all 7 wit-dependent motifs and found that only 4 of these exhibited altered reporter intensities (Figure 2.7). Both Atyp18 and Atyp26 had a significant reduction in mean relative reporter intensity when comparing controls to their respective wit mutants. For Atyp18, control mean relative reporter activity per nucleus was measured at 120±11, while in wit mutants it dropped by 63% to 45±7 (Figure 2.7-C). The reduction was even more significant in Atyp26, where control mean relative reporter activity was measured at 260±30, while in wit mutants it dropped by 88% to 32±3 (Figure 2.7-D). In Atyp15, while there was no significant mean relative reporter intensity reduction between wit controls (64±2) and wit mutants (66±3), there was a 27% reduction in Atyp15mad mutants (49±2) compared to the controls (67±8) (Figure 2.7-A). In this case, we were also able to measure relative reporter intensities of pMad+ stained nuclei only, where the mean relative reporter intensity of controls (70±4) was reduced by 54% in the Mad-binding site mutants (32±2) (Figure 2.7-B). Finally, Atyp23 had the most surprising result, with the wit mutants exhibiting a 24% increase in mean relative reporter expression compared to the controls (Figure 2.7-E). Based on the distribution in the Atyp23 graph (Figure2.7-F), the highest reporter-expressing cells appear to be maintained in the wit mutants, suggesting that the increase in mean relative reporter intensity observed in Figure2.6-E could be a result of a loss of mainly low reporter-expressing cells.  Having established the necessity of Medea for the BMP-dependent activity of the FMRFa BMP-RE (Berndt et al., manuscript pending resubmission), we decided to test the necessity of Medea for reporter expression in a subset of the 7 BMP-dependent AE2 motifs. We placed Atyp3, Atyp8, Atyp11, and Atyp15 in a null Medea background, and in all these cases, reporter expression was lost with the same pattern observed when the reporters were placed in a wit mutant background (Figure 2.8). 35  Overall, we have identified 7 BMP-dependent AE2 motifs out of the 10 motifs with reporter expression overlap with pMad stain in the L3 VNC tested, which gives our approach a discovery rate of 70%. This discovery rate is in line with previous studies done using other BMP-motifs (Vuilleumier et al. 2018).   36  Table 2.4: AE2-enhancer genomic fragment-driven GFP reporter expression patterns.  Expression* in:     DNA Fragment  VNC  Optic lobes Reporter/pMad stain overlap wit responsive VNC Expression Details  Expression in Other Tissues         Atyp26 +++ +++ √ √ neurons and glia glial cells, cells in trachea  Atyp3 +++ ++ √ √ medial and lateral neurons fat body, muscle tissue, salivary glands  Atyp18 +++ + √ √ medial and lateral neurons salivary glands Atyp15 +++ - √ √ medial and lateral neurons cells in trachea  Atyp11 ++ + √ √ medial and lateral neurons salivary glands Atyp8 + + √ √ sparse  tip of mouth hooks Atyp23 + + √ √ sparse  none Atyp13 +++ + √ - medial neurons none Atyp9 ++ ++ √ - lateral neurons none Atyp7 + + √ - sparse  none Atyp17 ++ + - - neurons and glia cells in midgut, pharynx and glial cells Atyp22 + ++ - - sparse  brain glial cells, Malpighian tubule cells  Atyp14 + ++ - - sparse   none Atyp24 + + - - sparse none Atyp20 + + - - sparse none Atyp21 + - - - sparse none Atyp5 + + - - low intensity  none Atyp10 + + - - low intensity none Atyp12 + - - - low intensity ring of cells in proventriculus Atyp2 - +++ - - none disc cells Atyp16 - + - - none cells in midgut Atyp25 - - - - none  epidermis (specifically cells in segment borders)  Atyp6 - + - - none none Atyp4 - - - - none none        * Expression pattern was assessed in wandering third Instar larvae. Reporters were sorted based on VNC expression intensity and pattern, wit responsiveness and expression in other tissues. 37                 Figure 2.1 AE2 motifs with no detectable expression in the CNS of third Instar (L3) larvae.               I. (A) Generation of the transgenic flies required the generation of plasmids containing the AE2 motifs of interest upstream of a minimal promoter driving a GFP reporter. Using φC31 integrase, the plasmid was placed in an attP2 site on the 3rd chromosome. (B) Schematic representation of the L3 larvae Central Nervous System (CNS). II. (A-E) Transgenic reporter lines Atyp2, 4, 6, 16, 25 exhibit no GFP reporter expression in wandering L3 larvae brains. Genotypes: All lines examined here were heterozygous (w;;AtypX/+).  I. II. A. B. 38  Figure 2.2: Non-BMP responsive AE2 motifs with expression in the VNC of third Instar larvae.  Transgenic reporter lines Atyp5, 10, 12, 14, 17, 20, 21, 22, 24 exhibit GFP reporter expression in L3 larvae VNCs, with no observed overlap of GFP reporter (green) and pMad (magenta). Genotypes: All lines examined here were heterozygous (w;;AtypX/+).  39           40           Figure 2.3: Three of the fragments tested are non wit-responsive in the L3 VNC.                                 (A-B) Representative images of transgenic reporter lines Atyp7 and Atyp13, along with their respective wit mutants.  (C) Representative images of reporter line Atyp9 control, wit mutant and pMad-binding site mutant (Atyp9mad). Side panels indicate overlap of GFP reporter (green) and pMad stain (magenta). (D-F) Quantification revealed non-significant change of number of reporter-expressing cells in the mutant genotypes compared to the controls, indicating that these reporters do not have BMP-dependent expression in the L3 VNC, despite observed overlap of reporter and pMad stain in some cells. Each point in the graph represents the total number of reporter-expressing cells in the VNC of a single animal. Significance was calculated with one-way ANOVA with a Tukey post-hoc test: no was significance detected in any of the comparisons. Genotypes: All control and pMad-binding site mutant lines examined here were heterozygous (w;;AtypX/+); wit mutants (w;;AtypX,witA12/witB11). 41              42          Figure 2.4: wit-mutants exhibit a significant loss in reporter-expressing cells compared to controls in the VNC of third Instar larvae.   (A-C) Representative images of transgenic reporter lines Atyp18, Atyp26 and Atyp23 where reporter expression is significantly altered in wit mutants (witA12/witB11). Side panels indicate nuclei with overlapping GFP reporter (green) and pMad stain (magenta). (D-F) Quantification of numbers of reporter-expressing cells indicates a significant difference in the mutant genotypes compared to their controls (Atyp18 controls: 339±15 cells and wit: 181±5 cells; Atyp26 controls: 805±22 cells and wit: 592±39 cells; Atyp23 controls: 61±6 cells and wit: 30±5 cells). Data shown as mean±SEM and n indicate the number of VNCs analyzed. Significance was calculated with Student`s t-test: Atyp18 p<0.0001; Atyp26 p=0.0004; Atyp23 p=0.0028. Each point in the graphs represents the total number of reporter-expressing cells in the VNC of a single animal. Genotypes: All control lines examined here were heterozygous (w;;AtypX); wit mutants (w;;AtypX,witA12/witB11). 43      44     Figure 2.5: wit-mutants and AE2 Mad-binding site mutants exhibit significant loss of reporter-expressing cell numbers compared to controls in the VNC of third Instar larvae.  (A-D) Representative images of transgenic reporter lines Atyp3, Atyp8, Atyp11 and Atyp15, where reporter expression was significantly altered in wit mutants (witA12/witB11) and pMad-binding site mutants (AtypXmad). Side panels indicate nuclei with GFP reporter (green) and/or pMad stain (magenta). (E-H) Quantification of numbers of reporter-expressing cells indicates a significant difference in the wit and pMad-binding site mutant genotypes compared to their controls. Significance was calculated with one-way Anova with a Tukey multiple comparisons test for the following genotypes: Atyp3 control vs wit mutants p<0.0001 and control vs pMad-binding site mutants p=0.0143; Atyp8 control vs wit mutants  p<0.0001 and control vs pMad-binding site mutants p=0.0021; Atyp15 control vs wit mutant and control vs pMad-binding site mutants p<0.0001. For Atyp11, as the wit control samples were non-normally distributed, significance was calculated with Mann-Whitney U test for control vs wit mutants p<0.0001; Student`s t-test was used for control vs pMad-binding site mutants p<0.0001. In all cases, wit control and AtypX controls had non-significant differences in reporter-expressing cell numbers. Each point in the graphs represents the total number of reporter-expressing cells in the VNC of a single animal and n indicates the number of VNCs analyzed. Genotypes: All control and pMad-binding site mutant lines examined here were heterozygous (w;;AtypX/+); wit mutants (w;;AtypX,witA12/witB11).  45  Figure 2.6: Near complete loss of pMad stain and reporter-expression overlap in pMad-binding site mutant Atyp15mad. Quantification of reporter-expressing pMad+ cells reveals a 90% loss in pMad-binding site mutant Atyp15mad (average of 7±2 cells per VNC) compared to controls. Significance was calculated with Student`s t-test: p<0.0001.  Data shown as mean±SEM and n indicate the number of VNCs analyzed. Each point in the graph represents the number of pMad+ reporter-expressing cells in the VNC of a single animal and n indicates the number of VNCs analyzed. Genotypes: All control and pMad-binding site mutant lines examined here were heterozygous (w;;AtypX/+).    46  A. C. B. D. E. F. 47  Figure 2.7: Significant change in relative reporter intensity was observed in 4 mutant genotypes compared to their respective controls.   (A-B) Quantification of mean relative reporter intensity for Atyp15 indicates a significant difference of reporter expression in the controls and pMad-binding site mutant genotypes. Graph (A) plots the mean relative reporter intensity (in arbitrary units) of all reporter-expressing cells in the VNC. Significance was calculated with Student`s t-test: p=0.006. Graph (B) represents the mean relative reporter intensity of pMad+ cells only, indicating that the BMP-dependent cells exhibit a larger reporter intensity loss (54% loss) than all the reporter expressing cells combined (27% loss). Significance was calculated with Mann-Whitney U test: p<0.0001. It is important to note here that there was an average of 7±2 pMad+ with reporter overlap per VNC (FigureS1). (C) Quantification of mean relative reporter intensity for Atyp18 indicates a significant loss of reporter expression in wit mutants compared to controls (63% loss). Significance was calculated with Student`s t-test: p<0.0001. (D) Quantification of mean relative reporter intensity for Atyp26 indicates a significant loss of reporter expression in wit mutants compared to controls (88% loss). Significance was calculated with Mann-Whitney U test: p=0.0095. (E) Quantification of mean relative reporter intensity for Atyp23 indicates a significant increase of reporter expression in wit mutants compared to controls (24% increase). Significance was calculated with Student`s t-test: p=0.0057. (F) All reporter-expressing cells from the 5 VNCs of controls and wit mutants were distributed into groups based on their relative reporter intensity. Each column represents the average number of cells in each bin across the 5 VNCs that were quantified. Despite the significant loss of reporter-expressing cells in Atyp23 wit mutants (Figure2.4-F), the highest expressing cells appear to be maintained.     Each point in the graphs (A-E) represents the mean relative reporter intensity of reporter-expressing cells in the VNC of a single animal.  48                       Figure 2.8: Medea is necessary for BMP-dependent activity of 4 wit-responsive BMP-AE2s.  (A-D) Representative images of Atyp3, Atyp8, Atyp11 and Atyp15 control reporter expression compared to the same lines in a Med and wit mutant backgrounds. In all cases, reporter expression loss in both mutants exhibit the same pattern. Genotypes: All control lines examined here were heterozygous (w;;AtypX/+); wit mutants (w;;AtypX, witA12/witB11); Med mutants (w;;AtypX, Medc246/Med13).  49  2.4.3 Identification of other BMP-CRE Binding Sites within the BMP-AE2 Fragments  Evaluation of the genomic fragments cloned into our reporter lines for additional or high-affinity BMP-CREs revealed that the Atyp11 fragment contains a BMP-AE type enhancer previously cloned and characterized (Vuilleumier et al. 2018). This previously identified and characterized fragment, called Van23, includes the same BMP-AE but only partially overlaps with the Atyp11 genomic fragment (Figure 2.9-A). Van23 was shown to be wit-responsive in (Vuilleumier et al. 2018), as the reporter expression (Figure 2.9-B,C) was completely lost in a wit mutant background. Comparing Van23 reporter with Atyp11, we observe a striking difference between the two expression patterns with regards to the numbers of reporter expressing cells, as well as distribution of reporter expressing cells (Figure 2.9-B-D).                 50                              Figure 2.9: Comparison of two reporter lines with a common BMP-CRE binding site reveals strikingly different expression patterns.  (A) Represents a USCS genome browser (https://genome.ucsc.edu/, dm6 - Aug.2014-Release 6 plus ISO1 MT) snapshot with the location and size of the genomic fragments contained in the reporter lines Atyp11 and Van23 (Vuilleumier et al. 2018). The position of the BMP-AE2 and BMP-AE motifs are also indicated in red text. (B, C) Images taken from (Vuilleumier et al. 2018) indicating the expression pattern of the reporter line Van23, which only contains a BMP-AE. Permission to use data was granted via email. (D) Representative image of Atyp11 reporter line, which contains two CREs, one BMP-AE2 and one BMP-AE. chr2L:9317809-9318548 740bp chr2L:9318297-9320893 2597bp Van23 Atyp11 A. Supplemental Figure S3. (Vuilleumier et al. 2018)  Figure 2A.  (Vuilleumier et al. 2018)  control D. Atyp11 control C. B. BMP-AE BMP-AE2 51  2.5 Discussion  In this Chapter, we show that many instances of conserved BMP-AE2 motifs exist in the Drosophila genome. Functionalizing a subset of these in vivo revealed that several have a wit-dependent expression in the L3 VNC. In the case of Atyp3, Atyp8, and Atyp11, wit mutants exhibit a more severe loss of reporter-expressing cells compared to their respective Mad-binding site mutants. This suggests that additional wit-dependent co-factors could contribute to reporter expression. Another possibility is that alternative BMP-CRE binding sites compensate for the loss. In fact, for Atyp11, we have identified an additional BMP-CRE included in the genomic fragment cloned (Section 2.4.3). The difference in reporter expression between Van23 and Atyp11 could reflect the combinatorial effect of BMP-AE and BMP-AE2 binding sites. On the other hand, since the Atyp11 fragment (2597bp) is significantly larger than the Van23 fragment (740bp), it may include additional cofactor binding sites that may be missing in Van23, thereby strengthening reporter expression. The partial loss of reporter-expressing cells in the pMad-binding site mutant of the BMP-AE2 in the Atyp11 line compared to the wit-mutant reporter loss in Atyp11 could reflect compensation by the still intact BMP-AE in the pMad-binding site mutant. In the case of Atyp15, loss of reporter-expressing cells in wit mutants and the pMad-binding site mutants is comparable. This suggests that in this genotype, the BMP-AE2 motif could be the sole or most important driver of wit-dependent reporter expression. Intriguingly though, wit mutants maintain wild-type (control level) mean relative reporter intensity while reporter intensity significantly diminishes in the pMad-binding site mutants. A possible explanation for this is the existence of wit-independent cofactors that help maintain reporter intensity expression in populations of pMad- cells that maintain reporter expression in a wit background (see Figure 2.10 for a schematic model). The need for other BMP-dependent cofactors could explain the loss of 52  reporter expression in pMad- cells in wit mutants. In pMad-binding site mutants, the ability of the activated Smad-complex to bind to the BMP-AE2 motif is impaired, resulting in loss of reporter expression in almost 90% of the pMad+ cells. Remaining reporter-expressing cells have decreased mean relative reporter intensity, suggesting that the cofactors driving expression in these cells may also be impaired due to the pMad-binding site mutation. This could be due to direct interaction with the BMP-AE2 motif or, more likely due to conformational changes to the immediately flanking sequence of the AE2. There is increasing evidence that DNA structural parameters such as bendability, curvature, groove shape, of flanking sequences of a TF binding site play an important role in DNA-TF recognition and binding for some TF families (Yella et al. 2018). Therefore, it stands to reason that cofactors that bind to the immediate vicinity of the BMP-AE2 would be affected by the mutation of the pMad-binding site.  In line with what we observed with the FMRFa BMP-AE2, Medea is required for the expression of the reporter, as putting the Atyp BMP-AE2s in a med-mutant background phenocopied the reporter loss observed in a wit-mutant background. Since the exact grammar of the Medea binding site was crucial to restricting FMRFa expression to the TV4 neurons, it is possible that the Medea binding site (GTAT) plays an important role in directing the pattern of reporter expression in the Atyp BMP-AE2s as well.  It is becoming increasingly clear that the use of low affinity cis-regulatory sequences is a common mechanism for spatiotemporal restriction in gene expression (Crocker et al., 2016). While most other low affinity cis-regulatory elements are highly degenerate, we find that the BMP-AE2 motif is very well conserved. Furthermore, based on my work described in this chapter, we now know that FMRFa might not be the only gene controlled by a BMP-AE2 motif CRE. The 53  conservation of this motif, especially the C>A conversion in the Medea binding site that distinguishes the AE2 motif from the previously described AE and SE motifs, could be attributed to sequence specificity for a cofactor. However, due to the fact that many other functional BMP-AE2 motifs were identified, it is somewhat unlikely that the same cofactor is used in all these scenarios and different types of cells. The differences in expression pattern and intensity between the BMP-AE2 reporters tested in this chapter could be conferred partially by the combination of TFs present and contributing to each scenario. It is also likely that the specific sequence encoded by the BMP-AE2 motif may provide additional critical information that guides TF binding and dictates some of the effect observed in intensity and expression pattern. Detailed bioinformatic work could help us identify other degenerate BMP-CREs, as many low-affinity (redundant) binding sites have been identified as clusters, as well as cofactors and other transcription factors that directly interact with or bind sequences that flank the BMP-AE2 motif.  In the end, having a detailed understanding of the regulatory motif network utilized in different tissues, under different conditions will help build better bioinformatic prediction models for regulatory elements. With our growing understand of the contribution of non-coding sequences to disease, every effort to understand how regulatory elements are used across the genome is important.   2.6 Strengths and Limitations This is the first study to systematically search for and identify low-affinity non-canonical versions of the well-established BMP-AE and BMP-SE.  Preliminary, unpublished work in our lab (by both Dr. Berndt and Dr. Vuilleumier) indicates that Brinker is not be involved in the BMP-responsiveness of AE2 sequences in the CNS. 54  No orthologue of Brinker has been identified in any metazoan other than insects (Blitz and Cho, 2009), suggesting that BMP signaling in the Drosophila CNS may be more similar to that of mammals, compared to the better-studied Drosophila wing disc. This gives us the means to study the use of a low-affinity BMP-CRE in a more similar context to that of mammals, thereby enabling us to more easily transition this work in vertebrates.  However, our approach to identifying and characterizing these CREs has its limitations. By using a biased approach and restricting our search to the BMP-AE2 “GGCGCC(N5)GTAT” sequence we are potentially disregarding other non-canonical low-affinity BMP-CREs that could be employed in the Drosophila CNS. Additionally, the use of chromosomal reporter integration to a non-native site takes each tested BMP-CRE out of its native chromatin context. This could result in false-positive or more likely false-negative results.   Here we utilize episomal reporters assays to functionally characterize the genomic regions we have identified, and though this is the most commonly used tool in the field to test candidate CREs, episomes lack native chromatin features, which may affect our ability to identify all real CREs. Comparison between genomically-integrated versus episomal reporter assays, of the same sequence, show higher reproducibility and better correlation with ENCODE annotations based on chromatin accessibility assays and TF ChIP-seq for the integrated reporters (Inoue et al. 2017).  Despite this, episomal reporter assays remain useful since active sequence-fragments in these assays are likely to also be active in a genomically-integrated context (Inoue et al. 2017).   55  2.7 Conclusions   In conclusion, we have identified 7 conserved BMP-CREs with the same sequence as the newly identified FMRFa BMP-AE2 binding site, supporting that this CRE type is not a unique feature of FMRFa regulation.   56   Figure 2.10: Schematic representation of gene regulation model in the Atyp15 reporter line. Genotypes represented here should be considered the same as in the experimental figures described previously: controls (w;;AtypX/+); wit mutants (w;;AtypX, witA12/witB11); pMad-binding site mutants (w;;AtypXmad/+). 57  Chapter 3:                                                                                                       Determining the BMP-Dependent Regulatory Network Underlying the Early Stages of Chondrogenesis in the Murine Limb Bud. 3.1 Introduction Musculoskeletal conditions are the second leading cause of disabilities globally. Disrupted or enhanced cartilage and bone formation or regeneration underlie the majority of this global burden, including arthritis, osteoporosis, chondrodysplasias and heterotopic ossification. Deciphering the molecular steps required to regulate chondrogenesis is essential in understanding the cause of various cartilage diseases, including skeletal malformations, chondrodysplasias, as well as osteoarthritis. According to the CDC (Centers for Disease Control and Prevention; https://www.cdc.gov/arthritis/data_statistics/index.htm), osteoarthritis (OA) is the most common form of arthritis affecting not just older adults, but also athletes and others with physically demanding jobs (Allen and Golightly, 2015). Progressive and irreversible degradation of articular cartilage plays an important role in the pathology of osteoarthritis (Bentley 1975). Being an avascular tissue, articular cartilage lacks access to necessary nutrients. Thus, cartilage is unable to sufficiently heal post trauma or chronic damage (Huey et al., 2012).  Current strategies to repair or replace damaged cartilage have many limitations and disadvantages (Zuscik et al., 2008; Smith et al., 2011; Solchaga et al., 2011; Huang et al., 2018). It is widely believed that a better understanding of chondrogenesis will provide novel therapeutic strategies. Concentrated efforts in the past two decades to understand the process of chondrocyte differentiation has led to the identification of numerous important factors in this process (reviewed in Section 3.1.1; (Akiyama and Lefebvre, 2011; Liu et al., 2017)).  58  Despite this wealth of discoveries, our understating of the chondrogenic transcriptional networks is far from complete. Therefore, further elucidating the molecular and regulatory mechanisms that control and direct chondrogenesis will provide us with a better foundation to enhance current and develop novel therapeutics.   3.1.1 Chondrogenesis Chondrogenesis is the process by which cartilage is formed during the development and maintenance of the vertebrate skeleton (Kozhemyakina et al., 2015; Liu et al., 2017). Almost all mammalian skeletons are (mainly) cartilaginous during early fetal development. This cartilage matrix serves as a template for bone formation and is slowly replaced by bone through a process known as endochondral ossification until only a few cartilaginous tissues remain in skeletally-mature animals. Though endochondral ossification accounts for the development of the majority of the vertebrate skeleton, not all bones are formed via this process. The skull and facial bones develop through a process called intramembranous ossification, where mesenchymal cells directly differentiate into osteoblasts, without the cartilage matrix “intermediate” stage (Hojo et al., 2016).  The chondrogenic process is initiated by the recruitment and migration of embryonic mesenchymal cells (MSC) committed to the chondrogenic lineage. These pre-chondrogenic cells aggregate and form pre-cartilaginous condensations, where cells begin to differentiate into ovoid-shaped chondroblasts. Chondroblasts secrete molecules such as type II collagen, hyaluronic acid, glycoproteins, and proteoglycans (including aggrecan) to form the extracellular matrix (ECM) of the cartilage. Pre-hypertrophic chondrocytes proliferate within the condensations until they exit the proliferative phase to undergo hypertrophy. Terminal differentiation of chondrocytes into 59  hypertrophic chondrocytes is characterized by cartilage matrix calcification/mineralization and vascular invasion. (Goldring et al., 2006; Underhill et al., 2014) Historically, it was widely accepted that hypertrophic chondrocytes undergo programmed cell death, initiating the process of endochondral ossification where bone marrow-derived osteoblast cells migrate and replace the apoptotic chondrocyte population (Gibson, 1998; Kronenberg, 2003; Mackie et al., 2011). However, recent cell lineage tracing techniques revealed that a significant portion of bone cells are derived directly via transformation of chondrocytes into osteoblasts (Ono et al., 2014; Zhou et al., 2014; Yang et al., 2014). This suggests that chondrogenesis and osteogenesis are not separate processes, but rather sequential phases of the same process, where hypertrophic chondrocytes transdifferentiate into bone cells.   3.1.2 Current Knowledge of Chondrogenic Regulatory Mechanisms SOX9, the “Master” Regulator of Chondrogenesis SOX9 (or Sex-Determining Region Y-Box 9) is a key transcription factor involved in chondrogenesis. It is essential for mesenchymal condensation, proliferation, and differentiation of chondrocytes (Bi et al., 1999; Bi et al., 2001; Akiyama et al., 2002). At the onset of chondrogenesis, Sox9 induces expression of two other Sox family members, Sox5 and Sox6, and together as the ‘Sox trio’, these three TFs direct transcription of a multitude of chondrogenic genes (Smits et al., 2001; Smits et al., 2004; Akiyama et al., 2007). For example, during the early stages of chondrogenesis, SOX9 directly induces the expression of cartilage matrix components and markers of committed chondrocytes, like collagen type II alpha 1 (COL2A1), COL9A1 and aggregan (ACAN). In addition to SOX5 and SOX6, several other regulators have been reported to function with SOX9 in the transcriptional control of targets, including the co-regulators WWP2, 60  SIK3, ARID5A, and the histone deacetylase HDAC4 (Akiyama and Lefebvre, 2011; Lefebvre and Dvir-Ginzberg, 2017).  From the pre-hypertrophic stage onwards, further differentiation of chondrocytes is mainly directed by RUNX2 (and RUNX3 to some extent), and by the MADS-box-containing transcription factor MEF2C, with MEF2D enhancing MEF2C`s action (Takeda et al., 2001; Yoshida and Komori, 2005; Arnold et al., 2007). SOX9 delays pre-hypertrophy and prevents osteoblastic differentiation by downregulating Runt domain transcription factor RUNX2 (CBFA1) and β-catenin (Akiyama et al., 2004; Zhou et al., 2006; Topol et al., 2009; Cheng and Genever, 2010; Dy et al., 2012). During hypertrophy, SOX9 still has important functions as it cooperatively transactivates expression of Col10a1 (hypertrophic chondrocyte-specific collagen gene) along with MEF2C (Dy et al. 2012). BMP Signaling in Chondrogenesis  BMPs are necessary for chondrogenesis, skeletal development, and fracture repair. Disrupted BMP signaling causes many debilitating human disorders of bone, vasculature, as well as many cancers and fibroses (Gazzerro and Canalis, 2006; García et al., 2016; Morrell et al., 2016). Conditional deletion of BMPs, BMP receptors, Smads or even BMP antagonists leads to a failure of chondrogenesis, severe chondrodysplasia and bone malformations (Gong et al., 1999; Yoon et al., 2005; Bandyopadhyay et al., 2006; Tsuji et al., 2006), and even embryonic lethality in some cases (Retting et al., 2009). For example, the targeted deletion of BMP Receptor 1a (BMPR1A) in chondrocytes results in severe cartilage defects and halts the process of endochondral bone formation (Jing et al. 2014).  61  BMPs control nearly every aspect of chondrogenesis (Song et al., 2009). Components of the BMP signaling pathway are highly expressed in growth plates with specific temporal-spatial patterns that correlate with functions during growth plate development and homeostasis (Minina et al. 2001). In vitro, BMPs can promote primary mesenchymal cells to differentiate into chondrocytes in high-density cultures (e.g mesenchymal limb bud-derived micromass cultures; Karamboulas et al., 2010).  In vivo, conditional deletion of Bmp2 and Bmp4 in prechondrogenic limb mesenchyme revealed that they are both essential for the initiation of mesenchymal condensations (Bandyopadhyay et al. 2006). Additional in vivo experiments revealed a role for BMPs in chondrocyte differentiation, as well as a potential role in maintaining Sox (Sox9, Sox5, and Sox6) protein expression (Yoon et al. 2005). After chondrocytes have differentiated, continued TGF-β-mediated SMAD1/5/9 signaling leads to terminal differentiation hypertrophy, potentially via interaction with RUNX2 (Leboy et al., 2001; Bandyopadhyay et al., 2006; van der Kraan et al., 2009). To further support this, a recent study has demonstrated that SMAD4 controls chondrocyte hypertrophic differentiation by directly binding to regulatory elements in the Runx2 promoter and upregulating its expression during skeletal development (Yan et al. 2018). Finally, BMP7 has a pivotal role in postnatal maintenance of articular cartilage and has been shown to delay cartilage degeneration, such as when induced by excessive running in rats (Sekiya et al. 2009). Despite advancements in understanding the impact of BMP signaling, many questions remain unanswered regarding the specific regulatory contributions of the canonical BMP pathway during the initial stages of chondrogenesis. Remarkably, we do not know how BMP signaling and Smad TFs control chondrogenesis. How do Smads fit in with the other known TFs and cofactors mediating this process? The goal of this thesis is to provide a more comprehensive report on the 62  molecular players involved in embryonic chondrocyte development. Specifically, we aimed to identify transcription factors and regulatory regions, and subsequently characterize downstream gene regulatory processes involved in chondrogenesis. Signaling Pathway Crosstalk During Chondrogenesis At the molecular level, several signaling pathways such as WNT/β-catenin, Notch, Retinoid, Hedgehog, FGF, and TGF-β/BMP regulate the initiation of the chondrogenic process, chondrocyte maturation and subsequent bone formation during embryonic skeletogenesis (Goldring et al., 2006; Long and Ornitz 2013; Kozhemyakina et al., 2015).  Sonic hedgehog (SHH), along with BMPs, is able to promote chondrogenesis of the somitic mesoderm by inducing the expression of both Sox9 and Nkx3.2 (Zeng et al., 2005). NKX3.2 is a BMP-dependent transcriptional repressor that blocks BMP-dependent expression of GATA 4, 5 and 6 (Daoud et al. 2014). These same GATA TFs were shown to block SHH-dependent induction of Sox9 gene expression, thus NKX3.2 effectively de-represses Sox9 and promotes chondrogenesis (Daoud et al. 2014). WNT signaling proteins were shown to inhibit chondrocyte differentiation (Rudnicki and Brown, 1997), potentially by repressing Sox9 via methylation of its promoter by the methyltransferase DNMT3. Fibroblast growth factor (FGF) signaling was shown to block the recruitment of DNMT3 (Kumar and Lassar, 2014). In addition, FGF signals have been demonstrated to boost the expression of Sox9 in chondrocytes via a mitogen-associated protein kinase (MAPK)-dependent pathway (Murakami et al., 2000).  Retinoid and Notch signaling, on the other hand, were both shown to suppress chondrogenesis (Cash et al., 1997; Hardingham et al., 2006; Dong et al., 2010).  63  BMP signaling is also required for the condensation of chondroprogenitor cells and chondrocyte differentiation. In vitro studies suggest that TGF-β signaling operates prior to BMPs and regulates the formation of pre-cartilaginous condensations, potentially via a SMAD3-driven regulation of Sox9 (Leonard et al., 1991; Furumatsu et al., 2005; Lorda-Diez et al., 2009; Karamboulas et al., 2010). TGF-β can also signal via the MAPK proteins p38, ERK, and JNK in MSC cells and contribute to the progression from condensation to chondrocyte differentiation by blocking WNT-mediated β-catenin nuclear translocation, thereby reducing N-cadherin and cell-cell interactions (Tuli et al., 2003; Zhang, 2009; Mu et al., 2012). TGF-β also mediates the end of proliferation (by counteracting the FGF-mediated cell proliferation) and the subsequent initiation of chondrocyte differentiation (Cleary et al., 2015). The parathyroid hormone-related peptide/Indian Hedgehog (PTHRP/IHH) signaling pathway has been defined as an important negative regulator of chondrocyte maturation. IHH is mainly expressed in pre-hypertrophic and early hypertrophic chondrocytes and regulates Figure 3.1: Chondrocyte differentiation A general schematic representation of differentiating chondrocytes from mesenchymal stem cells. This is a non-exhaustive list of key TFs, extracellular matrix components and signaling molecules characteristic of the cells at each of the differentiation steps. The red box indicates the stages of chondrogenesis during which we are conducting our experiments.  64  chondrocyte maturation by maintaining expression of PTHRP (Vortkamp et al., 1996). PTHRP signals repress chondrocyte hypertrophy by inhibiting the activity of both Mef2 and Runx family members (Kozhemyakina et al., 2009; Correa et al., 2010).  Taken together, although we have a much better understanding of regulatory mechanisms that take place during the maturation process of already committed chondrocytes, the same cannot be said about molecular events that induce and regulate the initial steps of mesenchymal cell differentiation into chondrocytes. Although some of the signaling molecules necessary for the induction of this sequential process (chondrogenesis) have been identified, our understanding of the exact regulatory mechanisms at play and the extent of crosstalk between these pathways are yet to be clarified. cis-Regulatory Elements in Chondrogenesis Some of the earliest work in identifying the cis-regulatory elements that are utilized during chondrogenesis focused on extracellular matrix components, starting with Col2a1 (Horton et al., 1987; Mukhopadhyay et al., 1995; Krebsbach et al., 1996). It was later shown that SOX9, along with SOX5, SOX6, and NKX3.2, were bound to this enhancer and directed Col2a1 expression in chondrocytes (Bell et al., 1997; Lefebvre et al., 1997; Lefebvre et al., 1998; Leung et al., 1998; Kawato et al., 2012). Subsequent studies revealed that other extracellular matrix components like Col11a2, Col9a1, Col27a1, Col10a1, Acan (AGC1), MIA (Cd-rap), Hapln1 (CRTL1) and Matn1 (Matrilin-1), owed their cartilage-specific expression to SOX9, SOX5, SOX6 as well (Bridgewater et al., 1998; Xie et al., 1999; Bridgewater et al., 2003; Kou and Ikegawa, 2004; Jenkins et al., 2005; Rentsendorj et al., 2005; Han and Lefebvre, 2008; Liu and Lefebvre, 2015; Li et al., 2018).  65  In the era of genome-scale ChIP-seq experiments, hundreds to thousands of putative binding sites for chondrogenic and osteogenic transcription factors have been identified (Hojo et al., 2016). Many of these experiments were conducted in various cell-lines, whole limbs or other primary tissues, in different organisms (mice, rats, humans) and at various timepoints during development, which makes it difficult to utilize in conjunction with our experiments and analyses. Several SOX9 ChIP-seq experiments have been done in the developing mouse limb (Oh et al., 2014; Liu and Lefebvre, 2015; Liu et al., 2018), revealing many putative cis-regulatory elements relevant to our work (Garside et al. 2015; Yamashita et al. 2019). One of these studies further revealed that SOX9 binds to low-affinity motifs and that many of these binding sites are grouped together and form “super-enhancer like” clusters providing insights into specific regulatory strategies that might be employed by Sox9 (Ohba et al., 2016). While ChIP-seq experiments have been conducted in chondrogenic cells for SMAD2/SMAD3 (Wang et al., 2016) and more recently SMAD4 (Yan et al. 2018), there are no such published experiments for SMAD1/5/9. BMP cis-Regulatory Elements in Chondrogenesis Over the years there have been some attempts to identify bona fide chondrogenic BMP-CREs and Smad-binding sites; however, very few have been successful thus far.  One of the first attempts was by Kusanagi et al. who showed that SMAD1 and SMAD4 could bind to the GCCG motif (derived from the early GCCGnCGC consensus identified in flies) in a BMP-dependent manner (Kusanagi et al. 2000). They then constructed a multimerized version of this motif, placed it upstream of a minimal Col10a1 promoter and transfected two cell lines that can be differentiated into chondrocytes (the C3H10T1/2 mesenchymal progenitor cells and the murine teratocartinoma ATDC5 cells). The reporter responded to a combination of canonical BMP 66  signaling components (transfection with SMAD1/BMPRII/constitutively active BMPRIB receptor) in both cell lines, but not to a combination of TGF-β signaling components (transfection with SMAD3/TβRII/constitutively active TβRI receptor). It is important to note the authors did not describe the use of differentiation culture protocols for the C3H10T1/2 and ATDC5 cell lines during their experiments, which is what is normally done with these cells to differentiate them into chondrocytes (Shukunami et al. 1997; Andrew et al. 1999). Therefore, the multimerized GCCG motif was not studied in a chondrogenic context, but rather in a mesenchymal cell/teratocarcinogenic context.  A subsequent study identified a Smad1 binding site in the promoter of the chicken COL10A1 gene with the sequence ‘CAGAGATTATTCACCTCTCT’ that was responsive in chicken embryonic chondrocytes (Drissi et al. 2003). Unfortunately, the promoter region upstream of the chicken COL10A1 gene (Appendix, Figure S2A) is not very well conserved to the murine and human genomes and no such putative binding sites have been identified in other vertebrates.   A study utilizing an in vivo approach identified a 560bp fragment upstream of the Msx2 gene that was able to drive BMP-responsive lacZ-expression in 10T1/2 mouse embryonic fibroblasts, as well as various tissues in E11.5 mouse embryos, including the developing limb bud (Brugger et al. 2004). Deletion analysis narrowed down the BMP-responsive region to a 52bp fragment that drove weaker lacZ-reporter expression in a subset of the original 560bp fragment expression tissues. Further study of a multimerized 52bp fragment revealed that it responded to BMP4 stimulation in the anterior mouse limb bud, but it did not respond to stimulation in the proximal or posterior regions of the limb. This suggests that other sequences within the larger 560bp fragment help mediate BMP-responsiveness in a subset of tissues/developmental stages 67  (Appendix, Figure S2B). Finally, mutation of the 4 “GCCG” sequences or “AATTAA” sequence within the 52bp fragment revealed that they are necessary for BMP-responsiveness.   Deletional analysis, using a luciferase reporter in P19 embryo-derived murine teratocarcinoma cells, identified a 163bp BMP7-responsive region upstream of the Ihh gene (Seki and Hata, 2004). Mutation of the 6 GCCGNGC-like motifs and the palindromic “GGCGCC” sequence within this region revealed that each of these motifs contributes to the overall BMP-responsiveness of the 163bp fragment. Reviewing the promoter/enhancer fragments in their deletional assay revealed that the initial “2HC8” fragment had a much higher BMP7 response than any subsequent fragments tested and that deletion of the +1 to -123 region resulted in a substantial loss of reporter activity. However, this region (+1 to -123 from the Ihh gene TSS) was not studied further for BMP-responsiveness and potential Smad-binding sites. See Section 3.3.3, Table 3.10 for more details.  A 28bp GC-rich BMP-responsive sequence, with two overlapping GCCGNCGC-like sequences in the 5`end and another GCCGNCGC-like sequence in the 3`end, was identified upstream of the Smad6 gene via deletional analysis in P19 cells (Ishida et al., 2000). This sequence was multimerized (3x) and placed in a luciferase reporter plasmid with a core Col10 promoter and tested for reporter transcriptional activation in P19 and mouse mesenchymal C2C12 cell lines, as well as human HaCaT keratinocytes and Mv1Lu mink lung epithelial cells. Luciferase reporter activity was increased in all cases with the addition of Smad1/5 and constitutively active BMP type I receptors or BMP7. Mutation of the 5`end and 3` end GCCGNCGC-like sequences revealed that the overlapping sequence in the 5`end was the one important for BMP-driven reporter activity. Another region just upstream of the 28bp GC-rich BMP-responsive sequence was also shown to 68  have significant BMP-responsive activity, however, this region was not studied further for potential Smad-binding sites (Appendix, Figure S3; Section 3.3.3, Table 3.10). A subsequent study in C2C12 and 2T3 (mouse osteoblast cell line) cells was able to identify a BMP2/SMAD1/RUNX2 responsive sequence within the region upstream of Smad6, however, in this case, it is likely that SMAD1 doesn’t directly bind DNA but rather interacts with RUNX2 to mediate an effect its Smad6 expression (Wang et al. 2007).  In addition to a lack of information regarding the contribution of BMP signaling to the chondrogenic regulatory network, there is insufficient research identifying BMP-CREs and Smad-binding elements. Most of the aforementioned studies, except for the one conducted by Brugger et al., were conducted using cell lines from different animals, with only a few of them using chondrocytes and/or cell precursors with chondrogenic potential. Utilizing primary mouse chondrogenic cells, our study provided a robust, reliable, and consistent system to identify and test the function of several regulatory regions important in different stages of chondrogenesis.    3.1.3 Project Rationale Reviewing current literature on chondrogenesis, it is clear there are important unanswered questions regarding the contribution of canonical BMP signaling to chondrogenesis. Recent work by Liu et al. has revealed that while SOX9 plays important roles during early chondrogenesis, it is not absolutely required to activate transcription (Liu et al., 2018). Therefore, there must be some other activating factors that function upstream or alongside SOX9 to enable chondrogenic gene activation. This new insight further underscores the need to study the transcriptional network of early chondrogenesis. Additionally, there is very little knowledge available on the gene regulatory role of Smads in chondrogenesis and the CREs that the Smads utilize.  69  We chose to use high-density micromass cultures from embryonic mesenchymal cells as they have been proven a reliable and robust in vitro model for studying the chondrogenic program and identifying important factors and elements involved. Murine embryonic limb micromass cultures are one of the best studied, and they can provide insights into the BMP-regulated molecular cascade involved in directing chondrogenic lineage differentiation (Underhill et al., 2014).  3.2 Materials and Methods 3.2.1 Animals and Timed Mating Mice with a CD1 background were housed on a 12-hr light-dark cycle in a controlled climate according to protocols approved by the UBC Animal Care Committee and in accordance with guidelines established by the Canadian Council on Animal Care. Single male mice were paired with female mice between the ages of 7-10 weeks old and vaginal plugs were assessed the following morning. Plugged female mice were set as embryonic day E0.5 at midday of discovery and isolated in a separate cage until embryonic day E11.5. 3.2.2 Primary Cell Culture and Treatments  Primary limb mesenchymal (PLM) cell micromass cultures were established, as previously described (Hoffman et al., 2006; Karamboulas et al., 2010; Underhill et al., 2014). Briefly, whole limbs were harvested from E11.5 mouse embryos and dissociated in media containing 1U/mL Dispase (Life Technologies). Dissociated cells were pelleted and re-suspended to a density of ~2x107 cells/mL. Cells were then seeded at a density of ~2x105 cells/micromass 10uL droplet into 12-well or 24-well plates. This cell density allowed chondrogenic differentiation and the formation 70  of cartilage nodules. Following 1-2 hours incubation, media containing BMP4 or Noggin (NOG) was added to each well and replaced every 24 hours. PLM cells were treated with 20ng/mL BMP4 (Calbiochem) or 100ng/mL NOG (StemCell Technologies) for 0, 12 and 24 hours, to drive and block BMP signaling, respectively. For each experiment, additional cultures treated with BMP4 or NOG were stained with Alcian blue (protocol described in Section 3.2.3) on day 4 post-seeding to confirm treatment efficacy. Alcian blue stains proteoglycans and other extracellular matrix components produced by mature chondrocytes. We confirmed the doses of 20ng/mL BMP4 and 100ng/mL NOG were sufficient to induce or inhibit differentiation (Appendix, Figure S4A). 3.2.3 Alcian Blue Staining Cells were washed with PBS, fixed for 10 min with 4% paraformaldehyde, rinsed with PBS, followed by a wash with 0.2M HCl.  Cells were stained with a 4:1 mixture of 0.2M HCl:0.5% Alcian blue solution (Alcian Blue 8 GX, BioShop; 0.5gr Alcian Blue powder in 95% ethanol) overnight and rinsed with 70% ethanol. Stained cultures were stored dry, but for imaging purposes, stained cells were placed in 70% ethanol. Images were taken on a Motic SMZ-168 light microscope with the Moticam 2300 digital camera (3.0M Pixel USB2.0). 3.2.4 Plasmids  The pCAGGS-SmadSD-IRES-H2B-RFP (constitutively active pSmad1) and pCAGGS-IRES-H2B-RFP plasmids were kindly gifted by Dr. E.Marti (Dréau et al., 2014); 7xBRE-Id3-Luciferase was a gift from Dr. Ken Cho (Javier et al. 2012). The 8 times repeat concatemer of the fly dad13 BMP-AE plasmid, BMP-AE8x-EGFP, and the Mad-binding site mutant version BMP-AE8xΔmad were generated in the Allan lab and published (Vuilleumier et al. 2018). 71  A custom destination vector (pGL3-Gateway-mpId3-eGFP; Appendix Figure S1) was created using a pGL3-Promoter luciferase reporter vector (Promega) as the backbone. A ccdB-Gateway cassette (Invitrogen) with attR recombination sites flanking a chloramphenicol-resistance gene was cloned upstream of the minimal SV40 promoter to enable easy transfer of enhancer fragments from the MCS-TOPO entry vector. The minimal SV40 promoter was replaced with the minimal Id3 promoter from the 7xBRE-Id3-Luciferase plasmid, and the Luciferase reporter gene was replaced with eGFP to allow easy visualization of reporter expression. As a positive control plasmid, we generated the pGL3-7xBRE-mpId3-eGFP vector with the 7xBRE enhancer placed in the custom destination vector, using the Gateway LR-Clonase exchange reaction. Finally, as a negative control, the minimal SV40 promoter from the pGL3-Basic luciferase reporter vector (Promega) was replaced with the minimal Id3 promoter and the luciferase reporter with eGFP.  3.2.5 Reporter Construction A list of all primers used for the construction of the reporter lines can be found in Table 3.1. Genomic DNA extraction from CD1 mice E11.5 PLM cells (isolated as described above) was performed with a QIAamp DNA Mini kit (Qiagen) following the manufacturer`s protocols. To generate reporter constructs, approximately 1.5kb genomic DNA fragments (see Table 2.1) were PCR-amplified, verified for correct band size via agarose gel, prepped by restriction digest and ligated into the MCS-TOPO entry vector (Appendix, Figure S1). We used Acc65I and BglII restriction enzymes to digest the forward and reverse primers, respectfully, except for the Dlx2-SE reverse primer, where we used MfeI instead. Homemade electrocompetent E.coli were transformed with 1uL of the ligated entry vector with a BioRad MicroPulser Electroporator using standard protocols. Transformed cells were topped up to 500uL with SOC media and placed in a 37oC shaking incubator for 1 hour. Cells were plated at a 1:50 or 1:5 dilution and incubated 72  overnight at 37oC on spectinomycin-containing LB plates. Several colonies were picked and grown in 5mL spectinomycin-containing LB cultures at 37oC overnight for colony screening. For the colony screen, 100uL of each culture was pelleted, lysed for 10min at 37oC and run on an agarose gel. A comparison with an empty vector (~3kb) enabled us to identify colonies that contained the enhancer fragments (~4.5kb). Colonies with the correct size vector were mini-prepped, verified with a restriction digest and sent for sequencing (Genewiz; https://www.genewiz.com/). Following sequence verification, Gateway LR Clonase II reactions were performed to transfer the enhancer fragments to the pGL3-Gateway-mpId3-GFP destination vector, according to manufacturer’s protocols. Electrocompetent E.coli were transformed as described above, plated at a 1:50 or 1:5 dilution and incubated overnight at 37oC on ampicillin-containing LB plates. Colonies were picked, grown, screened for correct plasmid size, and used to inoculate 100mL cultures for midi-preps. Midi-prep samples were finally re-sequenced to verify that the correct enhancer fragment was intact in the destination vector.  Mutagenesis was performed by Q5® Site-Directed Mutagenesis Kit (New England Biolabs), using primers designed to introduce specific base pair substitutions to the Mad binding site (GCCGGC > tgatga), according to manufacturer`s protocols. Primers were designed using the NEBase Changer v1.2.8 tool and are summarized in Table3.1. Mutagenesis was performed using the MCS-TOPO plasmid, with subsequent cloning steps as described above. Verification of the mutations, however, was done only via sequencing (no size differences between WT and mutant enhancer).  73  Table 3.1: Reporter Plasmid PCR and Q5 mutagenesis primers Bold capital letters indicate the restriction cut sites added to each PCR amplified fragment. Bold lowercase letters indicate the BMP-CRE Mad-binding site mutations.   3.2.6 PLM Cell Transfections  Cells were transfected as previously described (Weston et al., 2002). Briefly, PLM cells were harvested and resuspended at a density of ~2x107 cells/mL, as described above. For transfection purposes, cells were mixed with a DNA/FuGene6 (Promega) mixture in a 2:1 ratio. FuGene6-DNA mixtures were prepared according to the manufacturer`s instructions. Cells were seeded as ~2x105 cells/micromass droplet into 24-well plates. Following a 1 hour incubation in a humidified 37oC incubator, cells were supplemented with media containing BMP4 or Noggin (same concentrations as above). The media was replaced every 24 hours. Cells were monitored for reporter expression with the IncuCyte ZOOM™ software until Day4 post-seeding when cells were fixed and stained with Alcian Blue to confirm chondrogenic differentiation (treatment efficacy). 3.2.7 Poly-A RNA-seq RNA from PLM cultures was used for poly-A (enriched for mRNA) RNA-seq to identify differentially expressed genes upon BMP4 and NOG exposure. This experiment was repeated for Name Primer Sequence (5`→ 3`)   PCR primers                                       Genomic coordinates (mm10) Dlx2-AE.SE F: ACTGATGGTACCGAACGTGTCATCATCAGCCTAAAATGGGA chr2:71,628,376-71,630,508 R: ACTGATAGATCTGTGACCAGGTAACCACATTGGACAGTACA Dlx2-SE F: ACTGATGGTACCCACCTCTGACTTTCAGCGTCTCCTCTT chr2:71,531,303-71,532,305 R: ACTGATCAATTGCTTGAACTTGGAGCGTTTGTTCTGGA Dlx2-AE F: ACTGATGGTACCGGGTTGCCCAGCATTCCAAAGCAGCCA chr2:71,542,844-71,543,410 R: ACTGATAGATCTGGAGTGTGTGCTTTTTCCTTGGTGCCA Msx2-SE F: ACTGATGGTACCCAGAGAAGGCTGTAGACGGGCC chr13:53,476,035-53,476,474 R: ACTGATAGATCTTGAGCTCAGAGAGGTGCCATC Jdp2-AE F: ACTGATGGTACCGGATTTGTCTCACCTCCAGCCCACGAGA chr12:85,548,072-85,549,296 R: ACTGATAGATCTGCAGTCTGAAAAGAATCAGGCCCAGGGCT Q5 mutagenesis primers                                         Mutant BMP-CRE sequence Jdp2mut F: TCCCAAGCAGtgatgaCGCCAGTCAGTCTGGCCAGG tgatgaCGCCAGTCA R: GCTGCGCCACACTGGAGA 74  3 biological replicates. To collect enough total RNA from PLM cells for RNA-seq, 9 x ~2 x 105 cells/micromass droplet were seeded per well in 12-well plates, resulting in ~1.8 x 106 cells/sample. Wells were rinsed twice with sterile PBS, homogenized in TRIzol (Invitrogen) by pipetting and vortex, incubated at room temperature (RT) for 5min and stored at -80oC until they could be further processed. Total RNA isolation was carried out using a TRizol/RNAeasy kit (Qiagen) hybrid protocol. Briefly, 0.2mL Chloroform (Sigma) per 1mL TRIzol was added to samples, followed by vigorous shaking and a 3min RT incubation. Samples were centrifuged at 13,000 rpm for 15min at 4oC to allow for phase separation. The upper aqueous phase (containing the RNA) was collected and mixed with an equal volume of 100% EtOH. Samples were then transferred to RNAeasy spin columns. Subsequent steps were performed as per the RNAeasy kit “RNA Isolation from Animal Cells using Spin technology” protocol by Qiagen. Quality control (QC) and sample integrity were validated via 18s and 28s ribosomal RNA fragment sizes using the Agilent 2100 Bioanalyzer (Agilent Technologies). High quality samples were reverse transcribed to cDNA with the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems) according to manufacturer’s specifications using the Eppendorf MasterCycler EP Gradient S Thermal Cycler. Real time PCR (qPCR) was performed to confirm that cultured cells responded as expected to the BMP4 or NOG treatments (description below). PolyA-RNA-seq sequencing libraries were prepared for select, high quality samples using standard Illumina protocols. Sequencing was performed on the Illumina NextSeq500 using a NextSeq 75cycle High Output v2 kit sequencing at 42x42 Paired End. 3.2.8 qPCR Validation Quantitative real time PCR was performed on a 7900HT Fast Real-Time PCR system (Applied Biosystems) using validated IDT (Integrated DNA Technologies) primer sets for Sox9 75  (Mm.PT.58.42739087), Acan (Mm.PT.58.10174685), Nog (Mm.PT.58.30936710.g), Aldh1a2 (Mm.PT.58.12196815) and Tbp1 (Mm.PT.39a.22214839), and the TaqMan Fast Universal PCR master mix kit (Applied Biosystems). PCR consisted of one cycle at 95°C for 20 seconds followed by 40 cycles at 95°C for 1 second and 60°C for 20 seconds, as per the manufacturer’s instructions. Tbp1 was used as a reference to normalize gene expression across samples. Standard curves were generated using serial dilutions of pooled control samples. Gene expression levels were quantified using the absolute quantification-standard curve method. 3.2.9 RNA-seq Data Analysis Multiple analysis RNA-seq pipelines have been applied and their results combined. Briefly, for the alignment stage, we used both STAR (Dobin et al. 2013) and HISAT2 (Pertea et al., 2016), as well as the pseudo-aligners kallisto (Bray et al., 2016) and Salmon (Patro et al., 2017). Quantification was performed directly with STAR, kallisto, and Salmon or with RSEM (Li and Dewey, 2011) and StringTie (Pertea et al. 2016) for the pipelines producing real alignments. RNA-Figure 3.2: RNA-seq Analysis Pipeline  76  seq reads were mapped/aligned against the mouse reference genome (GRCm38.p5, http://www.ensembl.org). In-house Perl scripts were used to sum the read counts at the transcript level for each gene and create matrices comprising the read counts for all the genes in each sample. Differential expression analysis was then performed using the R package DESeq2 (Love et al., 2014) and edgeR (Robinson et al., 2010). Each sample was assessed using the quality-control software RSeQC (Wang et al., 2012) and the PtR script from the trinity suite (Grabherr et al. 2011). The output for each pipeline is a list of genes ranked by the p-value for differential expression after correction for multiple testing. A combined list was obtained by ranking the genes according to their median rank from the various analysis pipelines. Genes with a differential expression not going in a consistent direction between pipelines were eliminated from that combined list. 3.2.10 Gene Ontology Term Analysis Gene Ontology (GO) Enrichment Analysis PANTHER (Protein ANalysis THrough Evolutionary Relationships) Version 14.1 (http://pantherdb.org/; Thomas et al. 2006; Mi et al. 2019) and STRING (Search Tool for the Retrieval of Interacting Genes/Proteins) Version 11 database (https://string-db.org/; Szklarczyk et al., 2019) were used to perform gene annotations and overrepresentation (enrichment) analysis for the differentially expressed genes (DEGs) from our RNA-seq experiment. In PANTHER we used the default statistical settings (Fisher`s Exact test to calculate the p-value and False Discovery Rate (FDR) p<0.05) and the whole mouse transcriptome (22296 genes) as the reference list. The Gene Ontology database browsing tool AmiGO 2 version: 2.5.12 was also used to verify GO terms associated with certain genes (http://amigo.geneontology.org/amigo; Carbon et al., 2009).  77  3.2.11 Histone Modification ChIP-seq Histone modification ChIP-seq was used to identify chromatin marks associated with the state of enhancer and promoter activity. PLM cell samples for histone modification ChIP-seq were prepared, as described (Lorzadeh et al. 2016). Each sample was split into tubes, one to be used for pellet cell number quantification and one to carry out the ChIP-seq protocol. One biological replicate of this experiment was carried out. Briefly, cells were lysed in mild non-ionic detergents (0.1% Triton X-100 and Deoxycholate) and protease inhibitor cocktail (Calbiochem) in order to preserve the integrity of histones harboring epitopes of interest during cell lysis. Cells were pelleted, flash-frozen and stored at -80oC. Frozen cell pellets were cross-linked by 1% formaldehyde and lysed, following chromatin shearing. Chromatin fragments were incubated with protein G and A Sepharose beads (GE Healthcare) to eliminate non-specific binding. Unbound chromatin fragments were incubated with beads and validated antibodies against the following histone markers: H3K4me1 (Diagenode: Catalogue#pAb-037-050, lot#A1657D), H3K4me3 (Cell Signaling Technology: Catalogue#9751S, lot#8), H3K27me3 (Diagenode: Catalogue#pAb-069-050, lot#A1811-001P), H3K27ac (CMA309 monoclonal antibody, (Kimura et al. 2008)). DNA fragments were then stripped from histones, purified using the QIAquick PCR Purification Kit (Qiagen) and subjected to Illumina library construction by end repair, 3′ A-addition, and Illumina sequencing adaptor ligation. Libraries were PCR amplified and sequenced on Illumina HiSeq 2500 sequencing platforms, with paired-end 75bp, following the manufacturer`s protocols. We conducted one biological replicate of this experiment.  3.2.12 ChIP-seq Data Analysis ChIP-seq sequence reads were aligned to the mouse reference genome (GRCm38.p5) using Burrows-Wheeler Aligner (BWA; Li and Durbin, 2009). Peak calling and differential peak 78  comparisons were performed with MACS2 (Model-based Analysis for ChIP-seq) version 2.1.2 (Zhang et al., 2008). Plots for histone mark distributions around TSS`s were made with deepTools2 (Ramirez et al. 2016). Annotation of histone modification markers at promoter sites, as well as of active and repressed regulatory regions, against the list of differentially expressed genes (RNA-seq data) was performed using custom scripts. Data were visualized with the IGV genome browser (Robinson et al., 2011).  3.2.13 Graphs Graphs were made using GraphPad Prism Version 8.0.1 (GraphPad Software, San Diego, CA) and R packages. The heatmap for Appendix Figure S7 was made with the web tool ClustVis (https://biit.cs.ut.ee/clustvis_large/; (Metsalu and Vilo 2015)). 3.2.14 Bioinformatic Detection of BMP-CREs The Perl script scanMotifGenomeWide.pl from the HOMER v4.10 software suite (Heinz et al. 2011) was used to scan the reference mm10 mouse genome for the BMP-AE (GGCGCCN5GNCV), BMP-SE (GRCGNCN5GNCT) and BMP-AE2 (GGCGCCN5GTAT) motifs. Base-specific PhastCons scores (Siepel et al., 2005; 60-vertebrate conservation) were obtained from the UCSC genome browser (https://genome.ucsc.edu/)  to annotate each motif instance with evolutionary conservation scores. Using a custom Python script, each motif was annotated with the average PhastCons score across all bases of the motif instance. Motifs with an average PhastCons conservation score above 0.7 across placental mammals were kept for the correlation analysis with the RNA-seq results.   79  3.3 Results 3.3.1 Differential Gene Expression During BMP4-Driven Chondrogenesis  We performed polyA-RNA-seq assays on primary limb mesenchymal (PLM) micromass cultures prepared from murine E11.5 forelimbs and hindlimbs, treated with BMP4, NOG or vehicle control and harvested at 0hrs, 12hrs, and 24hrs post-seeding, in order to identify changes in gene expression during early chondrogenesis. We used BMP4 to induce chondrogenesis and NOG (a potent inhibitor of BMP signaling) to block chondrogenesis. Genes with enhanced and opposite responses to these treatments are likely to be direct BMP targets, especially if there are induced as early as 12hrs post treatment. Before sequencing, we decided to test the quality and proper induction of our samples by qPCR. We looked at the RNA expression of a number of key chondrogenic markers (Sox9, Acan), Noggin (as it is upregulated during chondrogenesis to oppose BMPs), as well as genes that respond to NOG treatment (Aldh1a2 – part of the retinoid signaling pathway that opposes chondrogenesis). Results from our qPCRs are summarized in Figure 3.3. As expected, Sox9 is already expressed at relatively high levels by ~E11.5 during the development of the limb bud. BMP4 progressively increases Sox9 expression, while Noggin has the opposite effect. Furthermore, BMP4 upregulates the expression of Noggin and Acan at 12hrs and 24hrs compared to controls, while NOG downregulates them. Finally, Aldh1a2 has the opposite response, BMP4 abolishes expression, while NOG increases it over time. Alcian blue stain of Day 4 cultures from these experiments further verifies that the treatments successfully induced or blocked chondrogenesis (Appendix, Figure S4-B).   80              Figure 3.3: Key chondrogenic genes respond to our treatments as expected, based on previously published work. Each graph represents the average of 2 biological replicates per sample and treatment (each biological replicate value is the average of 2 technical replicates). All samples are normalized to the reference gene. Sox9 Acan Noggin Aldh1a2 81 Identification of Differentially Expressed Genes Between BMP4, NOG and Control Treatments Prior to downstream analysis and data interpretation, we performed quality assessment steps, including coverage uniformity plots and sample matrix correlation plots that verified our sample quality (Appendix, Figures S5 and S6).  Utilizing multiple analysis pipelines, we generated a combined gene list based on the median ranking across all the pipelines used, increasing our confidence in the ‘differentially expressed genes’ (DEGs) in our samples (Figure 3.2).  Comparing our samples at the 12-hour and 24-hour time points with the 0h Control, we find significant numbers of genes that are differentially expressed (Figure3.3-A,B). The largest change is observed in the BMP4-treated samples both at 12hours and 24hours compared to the equivalent Control and NOG-treated samples. Based on our RNA-seq DEG expression levels and previously published data (Karamboulas et al., 2010) it is likely that the cells in our culture system at 24-hours are proliferating chondrocytes.   We also identified many genes that are differentially regulated at 12-hours and 24-hours post-treatment when comparing BMP4 and NOG samples against the 0h Control (Figure 3.4-C). The most highly differentially regulated genes at 24-hours were mainly extracellular matrix related genes, transcription factors/cofactors, signaling proteins and growth factors/cytokines (heatmap plot, Appendix Figure S7).    82          24h time point NOG 12h time point A. B. C. BMP4 vs NOG treatments 304 1102 1497 Figure 3.4: Venn diagrams representing the number and overlap of differentially expressed genes with different treatments and at different time points. 83 Gene Ontology Terms used to Identify Genes Regulated via BMP-Signaling  To elucidate the regulatory network proteins involved in early chondrogenesis and determine the contribution of BMP signaling to this process, we decided to focus on genes that were differentially expressed between BMP4 or NOG treatments at 12-hours and 24-hours. Specifically, we were interested in genes that had an opposite response to BMP4 and NOG treatments (i.e. BMP4 induces, but NOG reduces gene expression), as we postulated that they were likely to be directly regulated by BMP signaling. To achieve this, we generated “strict” lists of genes for both BMP4 and NOG (Figure 3.5). For instance, the BMP4 list included DEGs at 24-hours post-treatment compared to Control at 0 hours and excluded DEGs present in the 24-hours Control or NOG treatment, unless they responded differently in the BMP4-treated samples. The list was separated into upregulated and downregulated DEGs. A similar process was followed to generate the NOG lists. We did not repeat the process for the 12-hour DEGs since most of the genes that this workflow would capture were already included in our 24-hour list analyses. The workflow generated 4 lists of genes: 1) BMP4-upregulated (total 417 genes; Appendix Table S2), 2) BMP4-downregulated (total 319 genes; Appendix Table S3), 3) NOG-upregulated (total 77 genes; Appendix Table S4), and 4) NOG-downregulated (total 93 genes; Appendix Table S5).       84   We used the PANTHER and STRING databases to conduct functional profiling and enrichment analysis of genes within each of the 4 “strict” lists, and identified genes already annotated as chondrogenic. We selected PANTHER for our analyses, as it integrated into the GO curation process (updated monthly) and is the most reliable GO term database (Mi et al. 2019). We selected STRING to confirm data gathered from PANTHER, as it is a widely used, well-maintained and regularly updated database and contains information from several sources, including experimental data, computational prediction methods and public text collections (Szklarczyk et al. 2019). Although less reliable, the text-mining feature of STRING makes it possible to obtain more up-to-date information. First, we looked for enrichment of chondrogenic and osteogenic-related GO terms with the PANTHER Annotation Data Set “Biological Process Complete”. This allowed us to identify genes Figure 3.5: Workflow for generating the “strict” gene lists from the RNA-seq results. 85  previously known to have a role in chondrogenesis or osteogenesis, enabling their exclusion from downstream analyses and helped establish novel chondrogenic genes.  Within the BMP4-upregulated list, we identified a total of 71 previously annotated chondrogenic/osteogenic genes (GO term list is summarized in Figure 3.6). These results were confirmed by searching for “Biological Process” chondrogenic and osteogenic-related GO terms in STRING, which yielded a ~99% overlap between the two tools. We then looked at the “Molecular Function Complete” GO terms of these 71 genes and found that 24/71 (34%) were enriched with TF function, transcription regulation and cofactor function terms (Figure 3.6). This could either mean that many transcription factors/cofactors are involved in the initial stages of chondrogenesis or, more likely, that most studies focus on chondrogenic TFs/cofactors, resulting in an overestimation of the number of such genes being annotated with GO terms.  We then wished to identify new TFs/cofactors important in chondrogenesis, among the remaining 316 genes without any chondrogenic/osteogenic GO annotations (Figure 3.6). Using the “Molecular Functions Complete” annotations we identified 31 potentially new TFs/cofactors that may have a role in early chondrogenesis. Literature review confirmed that 13/31 (42%) of these TFs/cofactors were previously shown to have important roles during chondrogenesis such as functioning as SOX9 transcriptional coactivators, regulating the expression of RUNX2 and SOX9 TFs, and regulating cartilage ECM genes such as Col2a1 and Mmp-13 (Table 3.2). Additionally, 6 of the remaining 18 TFs/cofactors (Rorc, Nr4a3, Grhl1, Isl2, Pouf3f3, and Cux2) displayed a ≥1.5-fold increase in RNA levels by 24hrs (BMP4 vs NOG sample comparison), and this may indicate a likely role in the chondrogenic pathway. 86  Following the same workflow for the BMP4-downregulated list, we identified a total of 20 previously annotated chondrogenic/osteogenic genes, of which 7 were TFs/cofactors (Figure 3.6). From the remaining 286 genes without any chondrogenic/osteogenic GO annotations, we identified 29 TFs/cofactors (Table 3.3). Literature review revealed only 4/29 (14%) to have been previously studied in chondrogenesis. Of the remaining 25 genes, 4 (Hic1, Ebf2, Ebf3, and Scrt1) displayed a ≥1.5-fold decrease in RNA levels by 24hrs, suggesting that their downregulation is important in the chondrogenic process.  A similar analysis was performed with the NOG-upregulated and NOG-downregulated gene lists (Figure 3.7; Table 3.4). For the NOG-upregulated list, we identified a total of 7 previously annotated chondrogenic/osteogenic genes, of which 2 were TFs/cofactors (Gsc and Cebpa). From the remaining 33 non-signaling genes without a chondrogenic/osteogenic GO annotation, we identified 7/33 TFs/cofactors, one of which (Nfkbia) has been previously implicated in chondrogenesis (Table 3.3). Likewise, for the NOG-downregulated list, we identified a total of 16 previously annotated chondrogenic/osteogenic genes, of which 6 were TFs/cofactors (Runx3, Msx2, Id4, Hey1, Gli1, and Maf). From the remaining 49 genes without a chondrogenic/osteogenic GO annotation, we identified 8 TFs/cofactors, and once again, two of these (Dlx3, Gata2) have been previously linked to chondrogenesis. The remaining newly identified genes may potentially be involved in the chondrogenic pathway and warrant further investigations.    87   Figure 3.6: Gene Ontology Term analysis of genes from the BMP4-treated list of differentially regulated genes identifies several potentially uncharacterized transcription factors and cofactors that may be involved in chondrogenesis.   88  Table 3.2: Transcription Factors and Cofactors with no chondrogenic GO annotations, upregulated with BMP4 treatment.  1  Nr4a3 may have a role in the inflammatory-related loss of cartilage/bone in human joint disease (Marzaioli et al. 2012). A conference abstract indicated that NR4A3 may be involved in the regulation of Mmp-13 and Adamts-5 expression (Angerer et al. 2012).  2 No direct association of Grhl1 with chondrogenesis, but GRHL3 was shown to act upstream of SP7 (OSX1) in a BMP-dependent manner, to regulate osteoblast formation (Salazar et al., 2016). TFs/cofactors with no chondrogenic GO annotations RNA-seq (log2FoldChange) Manual literature search associations with chondrogenesis/osteogenesis Sox10 6.26 (Chimal-Monroy et al. 2003) Rorc 3.28 - Nr4a3 3.18 1 Dlx3 2.67 (Hassan et al. 2006) Klf4 2.21 (Gurusinghe et al. 2019) Grhl1 2.07 2 Ppargc1a (aka PGC1A) 2.06 (Kawakami et al. 2005) Isl2 2 3 Pou3f3 2 4 Plagl1 (aka Zac1) 1.69 (Tsuda et al. 2004) Cux2 1.65 5 Prickle1 1.52 (Yang et al., 2013) Nupr1 1.42 (Yammani and Loeser, 2014) Nacc2 1.37 6 Ebf4 1.36 7 Vgll2 1.33 8 Klf15 1.32 (Song et al., 2017) Ets2 1.03 9 Pir 0.97 - Sox13 0.8 (Wang et al., 2006) Tox2 0.8 - Bsn 0.79 - Ssbp2 0.74 - Wipi1 0.68 - Ezh1 0.67 (Lui et al. 2016) Optn 0.66 10 Litaf 0.54 - Klf3 0.42 - Atf6 0.4 (Guo et al., 2016) Atf4 0.39 (Wang et al., 2009) Nfe2l1 0.36 (Kim et al., 2010) 89  3 Tzchori et al., noted the presence of Islet2 during limb bud patterning (Tzchori et al. 2009), but no association with chondrogenesis specifically.  4 No direct association with chondrogenesis but implicated in craniofacial development and the formation of some of the maxillary arch-derived skeleton (Jeong et al. 2008).  5 No direct association with chondrogenesis, but a recent publication showed that Cux2 is involved in the specification of the limb forming fields by regulating the Hoxb TF and retinoic acid synthesis (Ueda et al. 2019).  6 No direct association with chondrogenesis, but its paralog Nacc1 has been implicated in chondrocyte migration and differentiation (Yap et al. 2013).  7 No direct association with chondrogenesis, but its paralog (Ebf1) has been shown to have a crucial role in promoting the differentiation of MSCs into adipocytes and osteocytes (Almalki and Agrawal, 2016).   8 Vgll2a (zebrafish ortholog of Vgll2) was shown to be required for craniofacial development and pharyngeal cartilage development in zebrafish (Johnson et al. 2011). 9 Ets2 is known to be expressed in the developing cartilage, but it has been studied more in the context of osteogenesis (Raouf and Seth, 2000). We do not know if it has any other chondrogenesis specific functions.   10 No direct association with chondrogenesis; however, mutations in Optn have been associated with Paget`s disease of bone (PDB), the second most frequent metabolic bone disorder (Silva et al., 2018).               90  Table 3.3: Transcription Factors and Cofactors with no chondrogenic GO annotations, downregulated with BMP4 treatment. 1 No direct association with chondrogenesis. HIC1 attenuates Wnt signaling (Valenta et al., 2006), which is an important pathway in chondrogenesis.  2 No direct association with chondrogenesis; however, Dio2 is upregulated in osteoarthritic cartilage (compared to normal cartilage) and there is speculation on Dio3 having a role in chondrogenesis (Reynard and Loughlin, 2013).  3 No direct association with chondrogenesis or osteogenesis; however, Sox12 belongs to the SoxC family along with Sox4 and Sox11. SOX11 was shown to positively regulate osteogenesis (Gadi et al. 2013). TFs/cofactors with no chondrogenic GO annotations RNA-seq (log2FoldChange) Manual literature search associations with chondrogenesis/osteogenesis Hic1 -5.64 1 Scrt1 -2.38 - Ebf2 -2.05 - Ebf3 -1.65 - Glis3 -1.56 (Beak et al., 2007) Dio3 -1.35 2 Etv4 -1.26 - Prkcb -1.22 - Pls1 -0.94 - Hdac7 -0.92 (Bradley et al., 2015) Jazf1 -0.75 - H2afy -0.71 - Zcchc12 -0.67 - Id3 -0.66 (Asp et al., 1998) Nrip1 -0.62 - Wwc1 -0.61 - Etv6 -0.52 - Zfp90 -0.51 -  Sox12 -0.5 3 Zfp395 -0.48 - Pmf1 -0.47 - Tead3 -0.46 - Phf6 -0.45 - Ncoa3 -0.43 (Zhang et al., 2018) Ybx3 -0.43 - Pttg1 -0.39 - Kpna2 -0.35 - Stip1 -0.27 - Tcf20 -0.26 -    91                Table 3.4: Transcription Factors and Cofactors with no chondrogenic GO annotations, upregulated or downregulated with NOG treatment  *excludes genes that were captured by the BMP4 lists above 1 In mesenchymal cells, GATA2 blocks adipogenesis (Tong et al. 2000; Tong et al. 2005; Kamata et al. 2014). According to (Tolkachov et al. 2018) GATA2 also affects bone turnover and inhibits osteoblastogenesis of MSCs by blocking SMAD1/5/8 signaling. TFs/cofactors with no chondrogenic GO annotations RNA-seq (log2FoldChange) Manual literature search associations with chondrogenesis/osteogenesis Upregulated with NOG  Downregulated with BMP4  Macc1 -1.27 - Nfkbia -0.57 (Dehne et al. 2010) Nek6 -0.4 - Rrp1b -0.33 - Ruvbl1 -0.33 - Downregulated with NOG Upregulated with BMP4  Gata2 0.72 1 (Karamboulas, et al. 2010) Ywhah 0.36 - Figure 3.7: Gene Ontology Term analysis of genes from the NOG-treated list of differentially regulated genes identifies several potentially uncharacterized transcription factors and cofactors that may be involved in chondrogenesis. 92 Gene Ontology Terms used to Identify Chondrogenic Regulatory Proteins  Our “strict” lists aimed to identify genes most likely to be directly regulated by BMP signaling. However, a host of other genes that could also be involved in chondrogenesis were excluded from our previous analysis. Here, we aimed to characterize the other important chondrogenic-regulatory genes in order to create a comprehensive list of all genes expressed during the early stages of chondrogenesis.  Genes differentially regulated with BMP4 and NOG treatments were separated into 3 lists based on their expression time-points: a) “12-hours only”, b) “24-hours only” DEGs, c) “Overlap” DEGs between 12 and 24-hours. For each of these lists, we performed functional profiling and enrichment analysis with the PANTHER and STRING software, as described in Section As before, we looked for enrichment of chondrogenic and osteogenic-related GO terms using the “Biological Process Complete” and then the “Molecular Function Complete” functions to identify the TFs/regulators/cofactors (Figure 3.8). In all 3 lists, the Molecular Function GO terms “Transcription factor binding”, “Transcription regulator activity”, “Extracellular matrix structural constituents” and “Signaling receptor binding” were significantly enriched, an expected  result, since many transcription factors as well as signaling and ECM components are required to initiate most cellular differentiation programs, including chondrogenesis.   Within the “12-hours only” list we identified 24 chondrogenic/osteogenic annotated genes out of which 7 were annotated with GO terms relating to transcription factor function, transcription regulation, and cofactor functions. Genes identified in this set may be necessary only for a short period during the chondrogenic process since they were not detected as DEGs in the 24hrs post-treatment samples.  93  From the list of 254 genes without known roles in chondrogenesis/osteogenesis, we identified 25 potentially new TFs/cofactors that may be involved in the early stages of chondrogenesis (Table 3.5). An extensive literature review revealed that 8/25 of these TFs/cofactors have been previously studied in chondrogenesis. The remaining 17 TFs/cofactors could be pursued further to understand their potential contributions to chondrogenic lineage commitment and differentiation.  Within the “24-hours only” list we identified 145 chondrogenic or osteogenic annotated genes out of which 62 (43%) were annotated with TFs/cofactors GO terms. From the remaining 1234 genes without chondrogenic/osteogenic annotations, we identified 92 potentially new TFs/cofactors that may have a role during chondrogenesis. A subset of these genes is listed in Table 3.6. This includes 32 genes with a ≥0.55-fold increase in RNA levels, out of which 10 have been previously associated with chondrogenesis. From the remaining 22 TFs/cofactors, 7 display ≥|1.0-fold increase| in RNA levels (Ddn, Dmrta1, Nfe2l3, Lpin3, Myrf, Klf14, and Nr2f1) and could be pursued in the future to understand their potential contributions to the pre-hypertrophic/hypertrophic stages of chondrocyte development.  Finally, within the “Overlap” list we identified 187 chondrogenic or osteogenic annotated genes out of which 88 (47%) were annotated with TFs/cofactors GO terms. From the remaining 840 genes without chondrogenic/osteogenic annotations, we identified 97 potentially new TFs/cofactors that may have important roles during chondrogenesis since they exhibited sustained upregulation/downregulation in both tested time points. A subset of these genes was listed in Table 3.7. This includes 32 genes with a ≥1.0-fold increase in RNA levels, out of which 15 have been previously linked to chondrogenesis. The remaining 17 TFs/cofactors (Hr, Dusp26, Rbpjl, Jdp2, 94  Klf2, Ss18l2, Nr4a1, Fam129b, Isl1, Nr6a1, Smarca2, Esrrg, Mdfic, Eya2, Esrrb, Pdlim1 and Ikzf3), are identified as highly differentially expressed genes that are maintained upregulated/downregulated over at least 12-hours, which makes them likely to have important roles in chondrogenesis that need to be studied in the future.     95  Figure 3.8: Gene Ontology Term analysis of genes separated into 3 lists based on their expression time (only at 12hrs, only at 24hrs or at both 12hrs and 24hrs) identifies several potentially uncharacterized transcription factors and cofactors that may be involved in chondrogenesis. 96  Table 3.5: Transcription Factors and Cofactors with no chondrogenic GO annotations, differentially expressed only at 12hrs                                1 BCL6 was identified as a regulator of early adipose commitment (Hu et al. 2016).  2 While there is a focus on the function of Egr1 during chondrogenesis (Spaapen et al. 2013), there is some indication that Egr3 has increased expression in ATDC5 cells at the onset of chondrogenic differentiation. Additionally, EGR3 was shown to be involved in myoblast proliferation (Kurosaka et al. 2016).  3  Hdac11 ectopic expression was shown to inhibit myoblast differentiation (Byun et al. 2017).  4 Down-regulation of Hmga1 is necessary to initiate the myogenic (or osteogenic) program after induction of C2C12 differentiation (Brocher et al. 2010). 5 HMGA2 was shown to promote adipogenesis (Xi et al. 2016). 6  Klf12 was expressed in MSC but not in lineage-committed mesenchymal cells (Zhang et al. 2012).  TFs/cofactors with no chondrogenic GO annotations Manual literature search associations with chondrogenesis/osteogenesis Actn2 (James et al. 2005) Bcl6 1 Bhlhe40 (aka DEC1) (Shen et al. 1997) Cbfa2t3 - Crym - Dact2 (Sensiate et al. 2014) E2f2 (Yanagino et al. 2009) Egr3 2 Glis2 - Hdac11 3 Hmga1 4 Hmga2 5 Klf12 6 Plscr1 - Pou2f2 - Prmt5 (Norrie et al. 2016) Rxrg - Sik1 - Sim2 (Qiryaqoz et al. 2019) Smad7 (Zhao et al. 2017) Sox17 - Taf9b - Tfap4 - Trib2 (You et al. 2019) Zeb2 - Actn2 - 97  Table 3.6: Transcription Factors and Cofactors with no chondrogenic GO annotations, differentially expressed only at 24hrs and with RNA-seq |log2foldChange| ≥ 0.55 *excludes genes that were captured by the lists above  1 No direct association with chondrogenesis; however, according to (Pepe et al. 2010) NFE2L3 has an important role in smooth muscle cell (SMC) differentiation from stem cells. 2 No direct association with chondrogenesis; however, according to (Csaki et al. 2014) Lpin3 is expressed in adipose tissue and is induced during adipogenesis.  TFs/cofactors with no chondrogenic GO annotations RNA-seq (log2FoldChange) Manual literature search associations with chondrogenesis/osteogenesis Ddn 2.36 - Sfmbt2 1.73 (Hussain et al. 2018) Dmrta1 1.56 - Nfatc2 1.36 (Ranger et al. 2000; Chen et al. 2011) Nfe2l3 (aka Nrf3) 1.21 1 Lpin3 (aka Lipin-3) 1.2 2 Myrf 1.19 - Pax3 1.1 (Cairns et al. 2012) Klf14 1.04 3 Dcc 1.01 (Schubert et al. 2009) Dll1 0.94 4 Trib3 0.93 5 Ell3 0.92 - Glis1 0.87 6 Ccnd1 0.86 (Ito et al. 2014) Prdm16 0.8 (Warner et al. 2013) Pbxip1 (aka HPIP) 0.77 (Ji et al. 2019) Stat5a 0.77 7 Meis3 0.74 - Sertad2 (aka TRIP-Br2) 0.69 8 Sorbs1 0.63 - Klf13 0.59 - Anxa4 0.57 - Tsc22d1 -0.55 - Med12l -0.64 - Tle6 -0.68 - Nmi -0.71 - Rora -0.86 (Woods et al. 2009) Meis2 -0.97 9 Erbb4 -1.61 (Nawachi et al. 2002) Esr2 (aka Erβ) -1.61 (Zeng et al. 2016) Nr2f1 (aka COUP-TFI) -1.73 10 98  3 No direct association with chondrogenesis; however, KLF14 was shown to act as a master trans-regulator of adipose gene expression (Small et al. 2011). 4 Dll1 expression was observed in a chondrogenic (ATDC5) cell line (Watanabe et al. 2003), while (Zhang et al. 2018) have more recently shown that inhibition of Dll1 expression in rat chondrocytes and chondrosarcoma cells triggered cell death/senescence and suppressed proliferation.  5 No direct association with chondrogenesis; TRIB3 was shown to play an important role in proliferation and osteogenic differentiation in human bone marrow-derived mesenchymal stem cells (Zhang et al. 2017). 6 No direct association with chondrogenesis; GLIS1 was identified as a regulator of mesenchymal multipotency that represses lineage-specific genes (specifically adipocyte and osteoblast differentiation) (Gerard et al. 2019). 7 No direct association with chondrogenesis; STAT5A inhibition was shown to enhance bone formation by promoting osteogenesis of BMSCs (Lee et al. 2018). 8 No direct association with chondrogenesis; TRIP-Br2 was identified as a transcription co-regulator for adiposity and energy metabolism (Liew et al. 2013). 9 Restriction of Meis2 expression is essential for limb outgrowth/development, and disruption of its expression resulted in limb alterations, many of which developed into cartilage pattern alterations; therefore, MEIS2 may have a role during chondrogenesis (Capdevila et al. 1999). MEIS2 was shown to repress osteoblastic transdifferentiation of cells in the heart (Sun et al. 2019).   10 Nr2f1 was implicated in osteogenesis (Tsai and Tsai 1997; Manikandan et al. 2018).             99  Table 3.7: Transcription Factors and Cofactors with no chondrogenic GO annotations, differentially expressed at both 12hrs and 24hrs and with RNA-seq |log2foldChange| ≥ 1 *excludes genes that were captured by the lists above   1 WWP2 forms a complex with SOX9 and the transcription enhancer MED25 to enhance SOX9 transcriptional activity during chondrogenesis. Additional studies have recognized Wwp2 as an important target for cartilage diseases such as osteoarthritis (Rice et al. 2019; Chantry 2011). TFs/cofactors with no chondrogenic GO annotations RNA-seq (log2FoldChange) Manual literature search associations with chondrogenesis/osteogenesis Hr 3.35 - Dusp26 3.1 - Wwp2 2.74 1 (Zou et al. 2011; Nakamura et al. 2011)   Rbpjl 2.73 - Jdp2 2.26 2 Klf2 1.99 3 Pparg 1.65 (Wang et al. 2005; Ma et al. 2015) Ss18l2 1.44 - Nr4a1 1.35 4 Fam129b 1.19 5 Isl1 1 6 Nr6a1 -1.02 - Smarca2 -1.02 - Esrrg -1.09 7 Bcl11a -1.11 (Yamamoto et al. 2019) Mdfic (aka HIC) -1.16 8 Foxp2 -1.33 (Zhao et al. 2015) Epas1 (aka HIF2A) -1.35 (Preitschoph et al. 2016) Foxp1 -1.43 (Zhao et al. 2015) Eya2 -1.51 9 Esrrb -1.52 - Agtr2 -1.55 (James et al. 2005) Ebf1 -1.63 (El-Magd et al. 2015) Atf3 -1.68 (James et al. 2006) Nfia -1.7 (Pratap Singh et al. 2018) Pax9 -1.82 (Rodrigo et al. 2003) Barx1 -1.88 (Nichols et al. 2013) Zic3 -2 (Zhu et al. 2007) Pdlim1 -2.02 10 Hhex (aka Hex) -2.16 (Morimoto et al. 2011) Meox1 (aka Mox1) -2.29 (Candia et al. 1992; Candia et al. 1996) Ikzf3 -3.13 - 100  2 No direct association with chondrogenesis; however (Kawaida et al. 2003; Maruyama et al. 2012) showed JDP2 plays a role in osteoclast differentiation and bone homeostasis, while (Nakade et al. 2007) demonstrated that JDP2 represses adipogenesis.  3 Klf2 was found to be upregulated in diseased (osteoarthritis) chondrocytes (Aki et al. 2018; Teramura et al. 2015; Yuan et al. 2017). Additionally, it was demonstrated that KLF2 regulates osteoblast differentiation by targeting Runx2 (Hou et al. 2019). KLF2 was identified as a negative regulator of adipogenesis (Sen Banerjee et al. 2003; Wu et al. 2005).  4 Nr4a1 expression was significantly elevated in OA cartilage, and while it is normally located in nuclei of chondrocytes, it translocated to mitochondria in OA chondrocytes (Shi et al. 2017). 5 No direct association with chondrogenesis; FAM129B was identified as a Wnt/β-catenin regulator (Conrad et al. 2013), which offers a potential role for FAM129B to negatively regulate Wnt signaling, thereby enabling chondrogenesis.  6 Some indication of association with chondrogenesis; (Kawakami et al. 2011) showed Islet1 as a hindlimb-specific transcriptional regulator that regulates hindlimb outgrowth along with β-catenin, while (Akiyama et al. 2014) confirmed that loss of ISL1/β-catenin in the hindlimb results in failure to expand chondrogenic precursor cells and skeletal defects.  7 No direct association with chondrogenesis; however, work by (Jeong et al. 2009; Cardelli and Aubin 2014) suggests that ESRRG (aka ERRG) negatively regulates bone formation and osteoblast differentiation (potentially in a sex-dependent manner). 8 No direct association with chondrogenesis; but a paralog of MDFIC called MDFI or I-mfa has been shown to have important roles in chondrogenic differentiation (Kraut et al. 1998).  9 No direct association with chondrogenesis; but a EYA2 has been implicated in myogenesis and tenogenesis (Xu et al. 1997). 10  Pdlim1 expression was detected in human articular chondrocytes (Joos et al. 2008).           101  3.3.2 Identification of Active Regulatory Regions with Histone Modifications To complement our transcriptomic data from Section 3.3.1, we performed histone ChIP-seq assays to identify genome-wide histone modification profiles on PLM micromass cultures prepared as described previously. Cultures were treated with BMP4, NOG or vehicle control and harvested at 0hrs, 12hrs and 24hrs post-seeding, in order to identify histone modification mark changes during the time course corresponding to our RNA-seq data. To validate the biological relevance of the ChIP-seq peak calls from MACS2, we analyzed the profiles of the H3K27ac and H3K4me3 ChIP-seq peaks in the vicinity (±1.5kb) of all TSSs for all times points and treatments (Figure 3.9). We verified that the regions covered exhibited a bimodal pattern, as expected from observations in other studies of H3K27ac and H3K4me3 marks, and this confirmed the quality of our data (Wang et al. 2008; Nie et al. 2013). Additionally, the ratio of fold enrichment between upregulated and downregulated genes is reversed with BMP4 and NOG treatments, confirming that the histone peaks are enriched near the expected gene`s TSS as per our RNA-seq data.   We investigated potential regulatory regions utilizing the H3K27ac peak calls that were differentially regulated with BMP4 and NOG treatments (Figure 3.10). At 12hrs post-treatment we identified a total of 888 peaks within 50kb of DEG TSS, of which 563 were near upregulated DEGs and 325 near downregulated DEGs. Exclusion of the promoter-proximal peaks (±1.5kb from DEG TSS) helped locate potential enhancer regions within 50kb from the TSS. In doing so we identified 385 peaks near 234 upregulated DEGs and 212 near 122 downregulated DEGs. At 24hrs, a total of 2354 peaks were located within 50kb of DEG TSSs; 1035 peaks near upregulated DEGs and 1317 near downregulated DEGs. When promoter-proximal peaks were excluded, we 102  identified 856 peaks near 446 upregulated DEGs and 838 peaks near 410 downregulated DEGs within 1.5kb-50kb from the TSSs. We then combined all H3K27ac mark data within this range at both time points and removed duplicated gene entries and overlapping histone peaks. This resulted in a total of 1118 H3K27ac marks near 556 upregulated DEGs and 976 marks near 459 genes within 1.5kb-50kb from the TSS. An additional 398 peaks for the 12hr time point and 1153 peaks for the 24hr time point were located within 50-100kb. This substantial number of H3K27ac peaks in close proximity to DEGs provides a great starting point to locate candidate regulatory regions active during the initial stages of chondrogenesis when chondrocytes transition from condensation to active proliferation.  For example, the TF Sox5 is known to be activated by Sox9 during chondrogenesis. Figure 3.11 depicts the Sox5 locus, along with the RNA-seq data and pertinent histone mark data we generated, along with sequence conservation data and some previously published Sox9 ChIP-seq data from E12.5/E13 murine whole limb buds. Using our own data, we identified two potential regulatory sites for Sox5 that are used during early chondrogenesis and coincide with Sox5 differential expression upregulation. One of these sites (A`) also contains Sox9 binding sites, identified by two separate studies (Garside et al. 2015; Yamashita et al. 2019) around the same developmental stage as our PLM cultures, further validating our data.    103      104                           Figure 3.9: Histone mark enrichment distribution near the TSS of  differentially regulated genes identified via RNA-seq.  The average genome-wide histone enrichment calculated by MACS2 near TSSs (±1.5 Kb) was plotted for all genes detected from our RNA-seq analysis for each individual histone modification: H3K27ac (A,C) and H3K4me3 (B,D) and shown for individual treatments [Up (dark blue), Down (light blue) and Unchanged (yellow)]. Controls in (A,B) are from time point 0h and controls in (C,D) are from 24hrs.  The relative fold enrichment was calculated with the MACS2 algorithm, accounting for the background signal by comparing the ChIP peaks within an individual sample to its own input control. The scale of fold enrichment distribution for individual samples is on the right of each enrichment matrix. Projection analysis (C,D) graphically shows each gene in each          105       Figure 3.10: H3K27ac peaks near DEGs at 12hrs and 24hrs post-treatment.  We identified thousands of H3K27ac peaks near DEGs, many as close as 50kb from the gene`s TSS. Some of these peaks may represent active regulatory regions.    106   Figure 3.11: RNA-seq, histone modification and Sox9 ChIP-seq for Sox5. In the RNA-seq profiles, we observe a significant increase in exon reads for the BMP4 track compared to the Noggin track. Likewise, we note a significant increase in H3K27ac reads at multiple intronic regions (A’ and A” expanded to show closeups of BMP4 and Noggin H3K27ac reads). The sequence conservation track shows that these K27ac reads are localized to well-conserved sequences across placental mammals. There also appear to be Sox9 binding sites overlapping with the K27ac marks in panel A’, validating that area as a regulatory region.  Sequence conservation is represented with a heatmap of the UCSC genome browser PlyloP basewise conservation score derived from the Multiz alignment of 60 placental mammals (against the mm10 mouse genome). Exons are shown as wide bars and the 5’-3’ direction of the denoted by the > symbol. Scales are shown in the top right corner for each track. SOX9 ChIP-seq data used here were from E12.5 mouse whole limb buds ((Garside et al. 2015); GEO Database (GSE73225)) and E13 mouse whole limb buds (Yamashita et al. 2019). 107  3.3.3 Bioinformatic Motif Discovery and Integration with Genomic Data  Using the functionally distinct motif-types identified in Drosophila, we scanned the mouse genome using the motif discovery tool HOMER (v4.10) (Heinz et al. 2011). These were filtered for high sequence conservation across 60 mammalian species using PhastCons scores, limiting the list to 3944 BMP-CREs with an average PhastCons score over 0.7 (Table 3.8). In all cases, over 2/3 (2764/3944) of the motifs are highly conserved with a score above 0.9.  Table 3.8: Murine BMP-CREs with Average PhastCons Score > 0.7         Out of these 3944 motifs, more than half (2305/3944) are located nearby genes that were differentially expressed in our RNA-seq experiment. More specifically, 52% of these motifs (2068/3944) were located within 1Mb from a DEG`s TSS (Table 3.9). Notably, 533 of these motifs were located within 50kb of a DEG`s TSS. We then excluded promoter-proximal motifs (within 250bp of the TSS) and identified 401/533 motifs that are within 50kb of the DEG`s TSS. Finally, there were more BMP-CREs in the vicinity of upregulated DEGs (1143/2068) up to 1Mb away from the TSS than downregulated DEGs (925/2068).    Motif Type Average PhastCons Score BMP-SE BMP-AE BMP-AE2 1 181 115 25 0.99-0.9 1358 983 102 0.89-0.8 374 200 21 0.79-0.7 345 209 30 Total 2258 1507 178 108   Table 3.9: BMP-CRE distance from the nearest BMP-responsive gene TSS                    ___Motif Type___ Upregulated DEGs Downregulated DEGs Distance from TSS _BMP-SE_ _BMP-AE_ _BMP-AE2_  Total Up Down Up Down Up Down within 50kb 158 110 151 90 13 11 533 50kb-100kb 82 56 68 30 5 5 246 100kb-250kb 157 134 131 105 12 15 554 250kb-500kb 120 131 77 86 8 10 432 500kb-1Mb 106 72 51 64 4 6 303 over 1Mb 84 61 55 27 7 3 237 Total 707 564 533 402 49 50  1271 935 99 2305 Chi-Square Figure 3.12: Putative BMP-CREs are enriched up to 50kb away from BMP-regulated genes.  Statistical significance of the enrichment of BMP-AE and BMP-SE responsive genomic fragments near BMP up regulated genes was calculated using Pearson’s Chi-Square test for count data in R (function chisq.test) with continuity correction without correcting for multiple testing. We used the non-significantly DEGs from our RNA-seq as the control gene set.   109  We performed a Chi-square analysis to see if there is an enrichment of BMP-CREs near upregulated DEG from our RNA-seq data (Figure 3.12). This test indicated that there was a significant enrichment of BMP-AEs up to 50kb from upregulated DEGs. The same was not true for BMP-SEs, as there was an enrichment only up to 20kb away from the TSS. When we combined both BMP-AEs and BMP-SEs, we identified a significant enrichment up to 50kb away from TSSs. Upon review of previously identified Smad-binding sites or BMP-responsive regions in the vicinity of BMP-regulated genes (some of which have important roles in chondrogenesis) such as Msx1, Msx2, Smad6, Ihh and Id1, we identify canonical BMP-CRE motifs conforming to an AE- or SE- like consensus, with canonical n=5 linker site spacing (Table3.10). Many of these BMP-CREs were not identified as potential Smad-biding sites, but our bioinformatic search using the Drosophila BMP-CRE motifs has allowed us to identify these highly conserved motifs.     110  Table 3.10: Revisiting and Examining Published Vertebrate BMP Regulatory Elements Experimentally Verified Vertebrate BMP-Regulatory Elements  Gene Name Publication Citation Regulatory Element Coordinates (mm10) Smad-Binding Specific Sequence Comments and Observations Msx1 (Binato et al. 2006) chr5:37,824,529-37,824,683 GCCGGCG Potential Smad4 binding sites not identified in this study. CGGCGGACCCGGAGCCGGCGAGTGCGCCT Id1 (Korchynskyi and Ten Dijke, 2002) (López-Rovira et al. 2002) chr2:152,735,122-152,735,230 GGCGCC CGCC Described as individual binding sites identified and not a Smad-complex binding site, additional putative Smad4 binding site not identified in this study. CGCCCGCGCGGCGCCAGCCTGACA Id1 (Karaulanov, Knöchel, and Niehrs 2004) chr2:152,739,406-152,739,488 TGGCGCCTGGCTGTCT Identified a canonical BMP-CRE GGCGCCTGGCTGTCT Id2 (Karaulanov, Knöchel, and Niehrs 2004) chr12:25,098,842-25,098,933 TGGCGCCAGAGAGTCT Identified a canonical BMP-CRE GGCGCCAGAGAGTCT Within the same fragment tested for BMP-responsiveness in this study, we identify an additional putative BMP-CRE with a GNCV Smad4 binding site GGCGCCGCTCAGGCG Id3 (Karaulanov, Knöchel, and Niehrs 2004) upstream:  chr4:136,140,517-136,140,650 intron2:  chr4:136,144,759-136,144,808 upstream: TGGCGCCAGGCTGTCT intron2: TGGCGCCGGCCAGTCT Identified two canonical BMP-CRE motifs  GGCGCCAGGCTGTCT and GGCGCCGGCCAGTCT Our analysis reveals additional putative BMP-CREs both upstream and downstream Id3. Id4 (Karaulanov, Knöchel, and Niehrs 2004) chr13:48,262,829-48,262,957 TGGCGCCAGTTAGTCT Identified a canonical BMP-CRE GGCGCCAGTTAGTCT Bambi (Karaulanov, Knöchel, and Niehrs 2004) chr18:3,506,976-3,507,050 TGGCGCC GTCT Though the correct Smad1/5/9 binding site has been identified, since there was not evidence at the time that GNCV sequences excited, they were unable to identify the complete BMP-CRE.  CGTCTCGTTGGCGCC 111  Smad7 (Karaulanov, Knöchel, and Niehrs 2004) upstream:  chr18:75,367,158-75,367,207 intron1:  chr18:75,370,679-75,370,758 upstream: GGCCGGAGCC intron1: TGGCGCCAGGCGGCC The intron1 Smad-complex binding site is almost completely resolved in this study, however, we cannot detect a canonical BMP-CRE sequence in the upstream sequence identified. intron1: GGCGCCAGGCGGCCT Our analysis reveals multiple putative BMP-CREs in the 5`UTR. Smad6  (Ishida et al., 2000) chr9:64,022,908-64,022,934 28bp GC-rich BMP-responsive fragment: GCCGCTCCTGCCCTGGAGCCGGCGCGGC Mutational analysis revealed that GCCGGCGCGGC was the BMP-responsive sequence. Another region upstream of the fragment tested was also significantly BMP-responsive, however, that fragment was not tested further. Our bioinformatic analysis revealed a conserved putative BMP-CREs within that upstream fragment identified (Appendix, Figure S3): chr9:64,023,056-64,023,071; GGCGCCGATTGGCCC We identified additional BMP-CREs up to ~100kb from the TSS. chr9:64,021,146-64,021,159; GGCGCCGTGCAGACC chr9:64,021,029-64,021,042; ATACCCTTTGGCGCC chr9:63,968,880-63,968,893; AGACCCAGAGCCGCC chr9:63,919,401-63,919,414; AGCCACGGGGACGCC chr9:63,919,326-63,919,339; AGACTGGCTGTCGCC Ihh (Seki and Hata, 2004) chr1:74,951,380-74,951,542 TGCGGGGCGGGGGCCCGGGGCCCTGGCTGGTGGCGGCGTGCTGTCCCCCTCGGCGCCTCGACTCTGAGCTGCCCGGCTCGCCGGCCGCCAATAAATAGGCCGGCCCGTTTGTTTTGGCAACGCGGCGACGGCGGGGGGCGGCGGGCGGCGGGGCTGAGGGCCG Multiple GC-rich motifs have been identified in this fragment as necessary and sufficient for BMP-specific up-regulation of reporter expression. A larger fragment containing the entire 5`UTR of Ihh was initially identified as BMP responsive and was able to drive almost double the reporter expression as the GC-rich motif fragment the study focused on. Looking at the excluded part of the fragment (+1 to -123 from the Ihh TSS) we identified 2 highly conserved putative BMP-CREs. chr1:74951,300-74,951,314; GGCGTCTGGGTGGCT chr1:74,951,246-74,951,260; GGCGCCTGGTGGGCT We identified additional BMP-CREs up to ~5kb from the TSS. chr1:74,951,797-74,951,811; GGCGGCCCCAAGGCT chr1:74,949,942-74,949,956; GGCGCCAGGGCGCAT 112  Bold and underlined sequences denote motifs sections that have been identified previously. Grey highlight denotes Smad1/5/9 or Smad4 binding sites conforming to the canonical BMP-CRE consensus that we have identified. More putative BMP-CRE binding sites can be seen in the UCSC browser graphic in Appendix Figure S8.     Msx2 (Brugger et al. 2004) 560bp fragment:  chr13:53,475,962-53,476,524 52bp fragment: chr13:53,476,176-53,476,228  AGGCTGAGTGCCGGGCCGAGAGCAATTAACGCGGCTCCGGCGCGGGCAGCCGC The authors identified a 560bp BMP-responsive region with multiple GC-rich sequences. Deletion analysis revealed a 52bp BMP-responsive region with multiple “GCCG” and a homeodomain-like binding site “AATTAA” sequences that were shown to be required for BMP-responsiveness.  We identified a conserved BMP-SE within the previously identified 560bp fragment. The authors recognised the GTCT part of the motif as a Smad4 binding site but did not further investigate this motif: chr13:53,476,371-53,476,385;GGCGGCCATTTGTCT Predicted Vertebrate BMP-Regulatory Elements from (Karaulanov, Knöchel, and Niehrs 2004) Gene Name Regulatory Element Coordinates (mm10) Predicted Smad-Binding Specific Sequence Comments and Observations Msx2 chr13:53,476,048-53,476,075 ~3.3kb upstream TSS AGACGGGCCAGGGCTGCGGGTAGGCGCC We found a canonical BMP-CRE sequence very close to this other predicted Msx2 BMP response element. chr13:53,476,371-53,476,385 (~3.6kb upstream TSS);  GGCGGCCATTTGTCT Gata2 chr6:88,198,664-88,198,693 GTCTGTGCAGGAGTCGGCAGCTGGCGCCAG We identified a canonically spaced Smad4 binding site (GNCV-type) near the predicted Smad1/5/9 binding site: chr6:88,198,686-88,198,700; GGCGCCAGGGCGGCC Gata3 chr2:9,878,569-9,878,596 GGCGCCAGGCAGCTCAGTGTTCGCAGAC We haven’t identified a canonically spaced Smad4 binding sequence for this predicted motif, however, using our bioinformatic approach we have identified 3 canonical BMP-CRE upstream of Gata3:  chr2:9,881,170-9,881,184; AGTCTCGCCGACGCC chr2:9,881,396-9,881,410; GGCGCCGGTCTGACC chr2:9,882,867-9,882,881; ATCCAGGCCGGCGCC 113 Prioritization of Regulatory Regions We were interested to see if any of these BMP-CRE motifs we identified bioinformatically were located within regions marked with histone modifications associated with active enhancers. We decided to use H3K27ac as a proxy for active enhancer regions since it has narrow-type peaks, making it easier to confidently call as differentially “modified” with MACS2. We restricted our search to motifs within 1Mb away from the TSS of a DEG. This yielded a total of 332 H3K27ac peaks containing at least 1 motif up to 1Mb away from a DEG TSS. Out of these, 107 peaks contain at least 1 motif within 50kb of a DEG TSS (Figure 3.13; Appendix, Tables S6-S9).           We were then interested to identify BMP-CRE motifs within the regions marked by H3K27ac, H3K4me1 or H3K4me3 modifications, which could indicate potential regulation of Figure 3.13: H3K27ac peaks with at least 1 BMP-CRE motif within 1Mb of a DEG. Highly conserved BMP-CREs within 1Mb from the TSS of DEGs from our RNA-seq experiments were identified within H3K27ac peaks. These regions could represent active Smad-binding sites. 114  these genes by SMADs. We found that at least 4 of the TF/cofactors without any chondrogenic/osteogenic GO term annotations identified in Sections and had such histone marks and SOX9 binding sites: Klf2, Grh1l, Wwp2 and Jdp2 (Figures 3.14-3.16, 3.19).  Figure 3.14: Histone modification marks, SOX9 ChIP-seq and BMP-CREs in the Klf2 locus. In agreement with our RNA-seq data, we can see a significant decrease in H3K27me3 repression-associated marks and an increase in H3K4me3 marks with BMP4 treatment at 24hrs. We noted 3 regions upstream of the Klf2 TSS that contain an increase in H3K27ac and H3K4me1 marks, as well Sox9 binding sites. Sequence conservation is represented with a heatmap of the UCSC genome browser PlyloP basewise conservation score derived from the Multiz alignment of 60 placental mammals (against the mm10 mouse genome). Exons are shown as wide bars and the 5’-3’ direction of the denoted by the > symbol. SOX9 ChIP-seq data used here were from E12.5 mouse whole limb buds ((Garside et al. 2015); GEO Database (GSE73225)) and E13 mouse whole limb buds (Yamashita et al. 2019). 115  Figure 3.16: Histone modification marks and SOX9 ChIP-seq data in the Wwp2 locus. Multiple Wwp2 introns contain increases in H3K27ac and H3K4me1 histone marks that line up with SOX9 binding sites. This result is in agreement with previously published work on Wwp2 being a direct target of SOX9 (Nakamura et al. 2011). Figure 3.15: Histone modification marks in the Grhl1 locus. In the Grh1l locus we can see a significant reduction of the H3K27me3 repression-associated mark from time point 0h to 24hrs. The decrease in H3K27me3 occupancy is more apparent with BMP4 treatment. We can also see a significant increase in H3K27ac and H3K4me1 marks with BMP4 treatment 24hr post-induction in a region ~8kb upstream of Grh1l TSS. This region likely contains CREs that could regulate the expression of Grh1l.       116  We next sought to prioritize a set of regulatory regions to functionally characterize with a reporter assay in PLM cells. Regulatory regions were evaluated based on the existence of histone marks that would indicate an active regulatory site (e.g. H3K27ac or H3K4me1), their proximity to significantly upregulated DEGs from our RNA-seq data, and the existence of a conserved BMP-CRE within the region. Additionally, we searched for potential Sox9 binding sites within those regions using ChIP-seq data from previously published studies (Garside et al. 2015; Yamashita et al. 2019). Using these criteria, we selected regulatory regions near the following 3 genes: Dlx2, Msx2, Jpd2.  DLX2  (Distal-less homeobox 2) is a transcription factor and a member of the homeobox (Dlx) gene family, that is upregulated around E9.5-E10.5 of mouse embryonic development and plays important roles during chondrogenesis (Zhang et al. 2018; Ferguson et al. 2000). It was also demonstrated to be a downstream target of BMP signaling in early chondrogenesis (Xu et al. 2001). Several cis-regulatory elements regulating Dlx2 expression in the CNS have been identified previously, though none of them was shown to be a BMP-dependent regulatory region (Ghanem et al. 2003; 2007). According to our RNA-seq data, at 24hours post-induction Dlx2 exhibits a high upregulation (log2FoldChange=4.3) when comparing BMP4vs.NOG treatment. Additionally, we identified 3 conserved BMP-CREs (one of which was a double binding site), within areas of upregulated H3K27ac marks with BMP4 treatment compared to NOG treatment (Figure 3.17).  MSX2 (Muscle segment homeobox 2) is a homeobox transcription factor with a complex role during chondrogenic differentiation. It stimulates differentiation of mesenchymal cells into osteoblasts, inhibits adipocyte differentiation, stimulates chondrocyte maturation, but also negatively regulates chondrocyte differentiation of migratory cranial neural crest cells (Semba et 117  al. 2000; Takahashi et al. 2001; Cheng et al. 2003; Ichida et al. 2004; Amano et al. 2008). Msx2 was also shown to be a direct target of embryonic BMP4 signaling (Hollnagel et al. 1999). A BMP-responsive region with substantial lacZ-reporter expression in E11.5 mouse limbs was previously identified ~3.5kb upstream of the Msx2 gene (Brugger et al. 2004). According to our RNA-seq data, Msx2 exhibits a high upregulation (log2FoldChange=1.52) at 24hours post-induction when comparing BMP4vs.NOG treatment. Our histone ChIP-seq data also indicates that this region contains a highly upregulated H3K27ac mark with BMP4 treatment compared to NOG treatment Further, we identified a conserved BMP-SE within the 560bp region previously identified as BMP-responsive (Brugger et al. 2004), thus we decided to use this region as an additional positive control for our reporter assay in PLM cells (Figure 3.18). JDP2 (Jun dimerization protein 2) is a member of the AP-1 family of transcription factors and has many different functions during cell differentiation. It plays an important role during RANK-mediated osteoclast differentiation and bone homeostasis (Kawaida et al. 2003; Maruyama et al. 2012), while also being able to repress adipogenesis (Nakade et al. 2007) and inhibit retinoic-acid induced differentiation (Jin et al. 2002). JDP2 is also known to act as an epigenetic regulator of gene expression (Jin et al. 2006). Based on our RNA-seq data, Jdp2 exhibits high upregulation (log2FoldChange=2.26) at 24hours post-induction when comparing BMP4vs.NOG treatment. Our GO term analysis and literature review revealed no chondrogenic annotations or associations. Additionally, we identified a highly conserved BMP-CRE around 50kb from the TSS of Jdp2, within an area with a small increase in H3K27ac marks when comparing BMP4 and NOG treatments at 24hours, as well as a SOX9 binding site (Figure 3.19). 118  Figure 3.17: Histone modification marks, SOX9 ChIP-seq and BMP-CREs in the Dlx2 locus. In concordance with the RNA-seq data, there is great upregulation in H3K4me3 reads along the Dlx2 locus. We also note a significant increase in H3K27ac reads at multiple regions in the vicinity of the Dlx2 gene. The sequence conservation track shows that these K27ac reads are localized to well-conserved sequences across placental mammals. We selected 3 of these regions containing the highly conserved BMP-CREs located ~3.6kb, ~14.5kb and ~80kb from the TSS for reporter assays in PLM cell cultures.  Sequence conservation is represented with a heatmap of the UCSC genome browser PlyloP basewise conservation score derived from the Multiz alignment of 60 placental mammals (against the mm10 mouse genome). Exons are shown as wide bars and the 5’-3’ direction of the denoted by the > symbol. SOX9 ChIP-seq data used here were from E12.5 mouse whole limb buds ((Garside et al. 2015); GEO Database (GSE73225)) and E13 mouse whole limb buds (Yamashita et al. 2019).  119    Figure 3.18: Histone modification marks, SOX9 ChIP-seq and BMP-CREs in the Msx2 locus. In agreement with the RNA-seq data, there is great upregulation in H3K4me3 and H3K27ac reads along the Msx2 locus. There is also a significant increase in H3K27ac reads ~3.6kb upstream of the Msx2 TSS in a region which is highly conserved according to the sequence conservation track. We also identified a conserved BMP-SE sequence in the same region, which we selected to clone and functionalize in a reporter assay in PLM cell cultures. Around 3.5kb upstream of the Msx2 TSS is a predicted BMP-response element (Karaulanov, Knöchel, and Niehrs 2004), as well as the experimentally verified 560bp BMP-responsive region (Brugger et al. 2004). Approximately ~8.5kb downstream of the Msx2 gene 3` end we can note a SOX9 binding site.  Sequence conservation is represented with a heatmap of the UCSC genome browser PlyloP basewise conservation score derived from the Multiz alignment of 60 placental mammals (against the mm10 mouse genome). Exons are shown as wide bars and the 5’-3’ direction of the denoted by the > symbol. Sox9 ChIP-seq data used here were from E12.5 mouse whole limb buds ((Garside et al. 2015); GEO Database (GSE73225)) and E13 mouse whole limb buds (Yamashita et al. 2019). 120   Figure 3.19: Histone modification marks, SOX9 ChIP-seq and BMP-CREs in the Jdp2 locus. We have identified a region ~50kb upstream of the Jdp2 TSS that contains a highly conserved BMP-AE, a small increase in H3K27ac mark with BMP4 treatment (and a visible bimodal peak compared to NOG treatment sample), as well as a SOX9 binding site. We selected to clone and functionalize this boxed region with a reporter assay in PLM cell cultures.  Sequence conservation is represented with a heatmap of the UCSC genome browser PlyloP basewise conservation score derived from the Multiz alignment of 60 placental mammals (against the mm10 mouse genome). Exons are shown as wide bars and the 5’-3’ direction of the denoted by the > symbol. SOX9 ChIP-seq data used here were from E12.5 mouse whole limb buds ((Garside et al. 2015); GEO Database (GSE73225)) and E13 mouse whole limb buds (Yamashita et al. 2019). 121  3.3.4 Preliminary Reporter Assay Experiments to Functionalize Identified CREs  Following the identification of putative regulatory regions, we wanted to test whether they were able to drive BMP-dependent reporter expression in the PLM cell culture system. We first generated the plasmid backbone that would allow us to easily add in and screen enhancer fragments, as described in the Methods Section 3.2.4. During the initial optimization phase, we opted to use a GFP read-out for our reporter, as it would be easier to quantify the number of reporter-expressing cells than it would be with the use of a luciferase reporter system, which is the most commonly used system in these assays.   Before testing any of the prioritized putative regulatory fragments, we transfected PLM cells with a positive control enhancer (7xBRE enhancer; (Javier et al. 2012), as well as the 8-times repeat concatemer of the fly dad13 BMP-AE plasmid (BMP-AE8x-EGFP) and its Mad-binding site mutant version (BMP-AE8xΔmad) that were generated in the Allan lab (Vuilleumier et al. 2018). We monitored the cells for GFP expression for up to 36 hours post transfection.  As expected, the negative control plasmid (Empty mpId3-eGFP, only contains a minimal Id3 promoter) had no GFP expression, while the cultures transfected with the 7xBRE and 8xAE enhancer fragments had GFP expressing cells. The mutated Mad-binding site (8xAEmut enhancer fragment) had no GFP expressing cells. The GFP expression of both 7xBRE and 8xAE enhancer fragments was indeed BMP-driven, as we could see more GFP expressing cells with BMP4 treatment than in the control and NOG treatments (Figure 3.20). Therefore, we established that the assay was working, though we noted that the transfection efficiency was low. We next transfected the 5 enhancer fragments we prioritized in Section This experiment was only carried out once, and the following results are preliminary.  122  Two out of the 5 had no GFP-expression in our system (Dlx2-AE.SE and Dlx2-SE; data not shown). The enhancers in the fragments Dlx2-AE and Msx2-SE had few GFP-expressing cells, whereas the controls either had not reporter expression (Dlx2-AE) or had fewer reporter-expressing cells (Msx2-SE) than the BMP4 treated cultures (Appendix, Figure S9).  Jdp2-AE was the only enhancer fragment we tested that had robust BMP-driven GFP expression (Figure 3.21). We observed a noticeable upregulation of GFP-expressing cells with the addition of BMP4 compared to the Control and NOG treatments. While this experiment needs to be repeated, our assay has identified a BMP-responsive region ~50kb upstream of the Jdp2 gene.  Taken together, this data indicates that utilizing our experimental and in silico methods can help us identify bona fide BMP-dependent regulatory regions.   123  Figure 3.20: PLM cell transfection with positive control enhancers and minimal promoter Id3 (mpId3) driving eGFP plasmids confirms BMP-driven expression of previously used BMP-regulated enhancers.  124  Figure 3.21: The Jdp2 BMP-AE drives GFP expression in differentiating chondrogenic PLM cells in a BMP4-dependent manner.  125  3.4 Discussion  BMPs play a fundamental role in the chondrogenic program and are extensively involved in many/different steps of this process, from the initial commitment to terminal differentiation. Despite this, there remain unanswered questions regarding the exact mechanisms regulating BMP responsiveness and the underlying cis-regulatory network involved in early chondrogenesis. To address this gap, we employed whole-transcriptome (RNA-seq) and histone ChIP-seq sequencing approaches to study BMP-driven mouse embryonic limb bud chondrogenesis. Combining these two methods with computational motif discovery and PLM cell reporter assays, we uncovered and verified the BMP-dependency of regulatory regions that most likely help coordinate chondrogenic gene expression.  Through our RNA-seq experiments, we identified genes that were differentially regulated during chondrogenesis, several of which are novel transcription factors/regulators/cofactors (Tables 3.2-3.7). Literature review revealed that some of the genes we identified were previously studied in a chondrogenic context, though they lacked chondrogenic/osteogenic GO term annotations. Furthermore, many of the genes we identified without chondrogenic GO terms have been shown to have a role in the maturation and maintenance of other MSC-derived lineages, including osteocytes, myocytes, adipocytes, and tenocytes. Therefore, it is likely that some of these genes have roles in the differentiation/maintenance of both chondrocytes and other MSC-derived lineages. Another possible explanation is that such genes are only necessary for the differentiation of alternate MSC lineages and are thus downregulated in PLM cultures to allow chondrogenic differentiation. Finally, it is also possible that some of the genes we identified have no role in chondrogenesis and are only a byproduct of the heterogeneity of our culture system (Discussed 126  further in Section 3.5). These results call for further investigations regarding the specific roles and regulatory processes that the genes/transcription factors/cofactors we identified play in the chondrogenic lineage.  Our histone modification ChIP-seq data has helped identify thousands of candidate regulatory regions, including the ~2000 H3K27ac regions located within 1.5-50kb from the TSS of ~1000 DEGs (Section 3.3.2; a subset of peaks listed in Appendix Tables S6-S9). While in our case we selected to focus on the H3K4me3, H3K4me1, H3K27ac and H3K27me3 marks to distinguish active/inactive promoters and enhancers (Heintzman et al. 2007, 2009; Creyghton et al. 2010; Rada-Iglesias et al. 2011; Zentner et al., 2011; Kim and Shiekhattar 2015), several other histone modifications have been implicated in influencing gene expression. According to Barski et al., active genes are characterized by high levels of H3K4me1, H3K4me2, H3K4me3, H4K9me1, and H2A.Z surrounding the TSS and elevated levels of H2BK5me1, H3K36me3, H3K27me1, and H4K20me1 throughout the entire transcribed region downstream of the TSS (Barski et al. 2007). On the other hand, inactive genes are characterized by low or negligible levels of H3K4 methylation at promoter regions, low levels of H3K36me3, H3K27me1, K3K9me1, and H4K20me1 along the gene body and high levels of H3K27me3 and H3K79me3 in the promoter and along the gene body. In the case of regulatory regions, although as described in Chapter 1, H3K4me1, H3K27ac, H3K27me3, and p300-binding are commonly used to predict enhancers and their (active/inactive) state, it is becoming increasingly clear that these marks alone do not identify all active/inactive enhancers (Karmodiya et al. 2012; Taylor et al. 2013; Pradeepa et al. 2016). Several additional histone modification marks have been found to mark active enhancers, including H4K16ac, H3K122ac, and H3K64ac (Taylor et al. 2013; Pradeepa et al. 2016). While in many cases these marks co-localize with H3K27ac, several examples of active enhancers lacking 127  H3K27ac, but bearing H4K16ac or H3K122ac marks have identified (Taylor et al. 2013; Pradeepa et al. 2016). Further, it was recently shown that the most active enhancers bear high levels of H3K4me3 rather than H3K4me1 marks (Henriques et al. 2018). There is also evidence to suggest that certain cis-regulatory elements can be active but also devoid of H3K4me1 (Dorighi et al. 2017). Thus, following the conventional use of H3K4me1 and H3K27ac marks as the identifiers of active enhancers would exclude these very active H3K4me3 bearing enhancers. Taking all this into consideration, while the histone modification marks we used in our work can help us identify several regulatory regions with high confidence, we cannot exclude the possibility that other regions (that we overlooked) may also be active enhancers in our system. Although the use of histone modifications to identify regulatory elements is a challenging and constantly evolving field, the integration of such information to this kind of work is worthwhile, as it gives a nuanced and informative view of gene regulatory mechanisms. Using a bioinformatic motif discovery approach we identified thousands of highly conserved, candidate BMP-CRE motifs throughout the murine genome. More than 50% of these motifs were located within 1Mb of DEGs, but more importantly, 533 of these motifs were located within 50kb of DEGs TSS. In fact, BMP-AE motifs were found to be enriched up to 50kb from upregulated DEGs. Reviewing published articles investigating vertebrate BMP-responsive elements revealed that very few of these enhancers were exhaustively studied to discover the exact Smad-binding site. Many of these studies identified large BMP-responsive fragments and did not attempt to locate the exact Smad-binding site sequence with mutational analyses. Others only identified the SMAD1/5/9 binding-site and not the SMAD4 site or simply predicted a BRE without a canonical 128  linker site and did not verify its BMP-responsiveness. Our results draw attention to the necessity of further characterizing enhancers. Better characterization of enhancers will enable more accurate computational motif predictions that can be used to map regulatory regions in a wide range of tissues, developmental systems, and organisms, including humans.  Finally, we prioritized and selected BMP-CREs in a pilot test to validate with a reporter assay in PLM cell cultures. We were able to identify 2 fragments containing a conserved BMP-CREs that were functional and responsive to BMP4 treatment in the cells (robust GFP expression with Jdp2-AE and low GFP expression with Dlx2-AE). We were also able to confirm reporter expression driven off the Msx2 fragment that was studied before. It will be interesting to mutate the previously identified BMP-responsive 52bp fragment (Brugger et al. 2004) to study the BMP-responsiveness of the remaining sequence in these cells. Future studies will focus on identifying more BMP-dependent regulatory regions, as well as further dissecting the specific TF binding sites that confer BMP-dependency to those regulatory regions.    3.5 Strengths and Limitations One of the major limitations of our approach is the use of cells collected not only from whole E11.5 limb buds but also from both forelimbs and hindlimbs, ensuring a heterogeneous population (both in cell type and developmental stage) at the start of the experiment. The use of factors in vitro, such as BMP4 or NOG, helps “synchronize” the cells, though it is still likely that some heterogeneity persists in these cultures. In this case, it is unclear whether certain DEGs identified in our RNA-seq experiments are due to remaining blood or endothelial cells in the culture (as some of the GO terms may indicate; Appendix, Table S10) or whether these same genes 129  are also implicated in chondrocyte differentiation. Adopting strategies such as selecting/purifying cell types or using single-cell techniques can help provide a better picture of the regulatory and epigenetic landscape from chondrocyte lineage commitment to full differentiation. Primary limb mesenchymal cultures are notoriously difficult to transfect (Juhász et al. 2010), prompting us to re-evaluate the transfection protocol we used in this thesis for our future experiments. While the FuGENE 6 (Promega) reagent is optimized to be gentle of the limb mesenchyme, the transfection efficiency is very low (Song et al. 2004). This was one of the reasons we opted for our initial validation experiments to be done with a GFP-reporter assay system, as opposed to the more commonly used luciferase-reporter assay. Our next experiments will likely explore the use of the Effectene (Qiagen)/trehalose transfection method developed in the Underhill lab (Karamboulas et al., 2010), as well as a pulse electroporation method (Bobick et al., 2014). Another avenue we are currently exploring is to transition the reporter assays into the mouse chondroprogenitor cell line ATDC5, which is a more robust and easier to transfect system (Chen et al. 2005). Initial validation of regulatory fragments in this system will provide a faster and more efficient screening process, allowing us to prioritize transfection of PLM cells with verified BMP-dependent regulatory regions.  A major limitation of histone modification ChIP-seq assays is that they reflect marks that have been corelated with enhancer activity, rather than directly showing that a genomic region functions as an enhancer. Therefore, despite their ability to yield genome-wide potential CREs, these genomic regions need to be further validated with functional assays (i.e. reporter assays).    130  Difficulties with data interpretation is another limitation when working with histone modification ChIP-seq data. Part of these difficulties rises from sample heterogeneity, but also the limited spatiotemporal information one can extract from the “snapshot” of histone/protein-DNA interactions. Therefore, many replicates or time points may be necessary to confirm results, making this a costly experimental approach. Additionally, determining and prioritizing which of the thousands of identified protein-binding sites to validate with reporter assays as biologically relevant is a daunting task. As discussed previously (Section 3.4), there is a growing body of work regarding histone modifications and their function in transcriptional regulation. Since this is still an evolving field, with the identification of novel and more nuanced roles for these marks in different contexts (Guenther et al. 2007; Liu et al. 2017), it is difficult to reliably interpret this information in isolation.  3.6 Conclusions Using evolutionary conservation, in silico motif discovery tools, RNA-seq, and histone modification ChIP-seq assays, we were able to identify thousands of candidate active regulatory regions that may be involved in chondrogenesis. We then embarked on validating their functionality, confirming two novel BMP-responsive enhancers in the process, using reporter assays in primary limb mesenchymal cultures.   131  Chapter 4: Conclusions  The overall objectives of my thesis were to firstly characterize the use of a novel low-affinity BMP-CRE motif previously identified by our lab to be active in the Drosophila CNS and secondly to initiate the process of identifying the BMP-driven gene regulatory network of underlying mammalian chondrogenesis.  In so doing, I have:  1) Identified and studied 7 conserved Drosophila BMP-CREs with the same sequence as the non-canonical FMRFa BMP-AE2 binding site, proving that this type of CRE is not only used by FMRFa but also other enhancers regions for BMP-dependent activation. Other studies in the Allan lab have provided evidence of the extensive use of BMP-AEs (Vuilleumier et al. 2018) and BMP-SEs (unpublished) in the Drosophila nervous system. My work has added to this field by identifying the use of these low-affinity BMP-regulated binding sites, which are located near BMP-regulated genes, in the Drosophila CNS.  2) Identified transcription factors and other cofactors with no previously known roles in chondrogenesis. Whole-transcriptome sequencing of BMP4 or NOG treated primary limb mesenchymal cultures yielded a comprehensive list of the gene regulatory network involved in the early stages of chondrogenesis. Chondrogenic GO term annotations and literature review helped narrow down this list to determine previously unidentified transcription factors/cofactors. Further investigations are required to fully elucidate their contributions to chondrogenic differentiation.    3) Identified thousands of potential regulatory regions in close proximity to genes that are differentially expressed during BMP-driven chondrogenesis. Histone ChIP-seq experiments corresponding to the timepoints and treatments of our RNA-seq experiments provided a list of ~2000 candidate regulatory regions within 1.5-50kb of ~1000 DEGs. The addition of in silico 132  methods and other high-throughput datasets will help us further refine these lists to ascertain regions most likely to be bona fide enhancers.  4) Demonstrated the BMP-dependency of regulatory regions located near upregulated differentially expressed genes during chondrogenesis. We prioritized candidate BMP-dependent regulatory regions based on their proximity to highly upregulated DEGs, as well as the existence of pertinent histone modification marks (H3K4me1 and/or H3K27ac upregulation with BMP4 treatment), conserved putative BMP-CREs, and SOX9 binding sites. We confirmed the BMP-dependence of two regulatory regions that are located within 50kb of highly upregulated DEGs, and one of these (Jdp2) was identified as a novel chondrogenic transcription factor through our RNA-seq experiments/GO term analyses.                   4.1 Significance and Translation of the Study  In the era of reliable and affordable next-generation sequencing and genome-scale methods, there is renewed interest in the scientific community to decipher the non-coding regions of the human genome, discover and expand our understanding of enhancers and other regulatory regions.  One of the main arguments used to highlight the significance of identifying and studying cis-regulatory modules is the staggering number of genome-wide association studies (GWAS) loci identified in non-coding regions that have been associated with human diseases (Smallwood and Ren 2013; Kellis et al. 2014; Hojo et al., 2016). Seeking ways to reliably and robustly identify cis-regulatory modules is a very challenging goal. Concentrated efforts to map cis-regulatory regions in different tissues, developmental stages and diseases have been ongoing, with new techniques 133  added to the traditional “enhancer-bashing” methods each year. This thesis attempted to combine different high-throughput experimental techniques and in silico methods to identify cis-regulatory elements related to canonical BMP-signaling.  4.2 Future Directions  Several lines of work are currently being pursued to follow up on the work described in this thesis. Preliminary work has begun on optimizing SMAD1/5/9 ChIP-seq experiments in the PLM cultures to complement our other datasets. This will be followed by SOX9 ChIP-seq as well to confirm the data previously published in whole limb buds. These experiments will allow us to identify SMAD1/5/9 and SOX9 binding sites and potential novel binding motifs, as well as determine whether SMAD1/5/9 and SOX9 cooperatively regulated the expression of chondrogenic genes. Additional candidate enhancers could be functionally validated with reporter assays in PLM cultures following prioritization of regulatory regions based on SMAD1/5/9 and SOX9 binding.   Apart from the repetition of the Jdp2-AE, Dlx2-AE, and Msx2-SE transfection experiments, mutational analysis of the enhancer fragments will allow us to delineate the specific TF binding sites that contribute to their BMP-dependency. Specific mutation of the putative BMP-CRE binding sites would also allow us to establish the necessity of this motif for BMP-dependence and Smad-binding. Depending on the success of such experiments, we would transition these analyses from episomal plasmid reporter assays to either endogenous TF-binding site CRISPR mutations or CRISPR interference (Larson et al. 2013) in the ATDC5 chondrogenic cell line. These experiments would allow us to assess the effect of individual TF binding sites on the expression of neighboring genes via qPCR and/or western blots. The overall effect of the 134  chondrogenic differentiation progress of the culture could also be assessed with Alcian Blue staining and qPCR or key chondrogenic markers such as Sox9, Acan, Col2a1, and others. Further, the newly identified chondrogenic transcription factors/regulators/cofactors from Section 3.3.1 could be further studied in a chondrogenesis context. The use of RNAi or CRISPR techniques would enable us to remove said genes from differentiated ATDC5 cells, allowing us to study the effects of their loss to the chondrogenic program via Alcian Blue staining and qPCR/western blots for Col2a1, Acan, Sox9, Runx2, and other chondrogenic markers.   Finally, in the long run, the work this thesis has initiated will have important repercussions for efforts to identify causal genomic loci of human musculoskeletal diseases. Transferring knowledge and lessons learned from studying murine chondrogenesis will help identify genes and gene regulatory regions that underlie chondrogenesis in the human genome. It is likely that this future work will provide a vital resource for clinical geneticists and informaticians attempting to identify causal mutations for diseases that lie in the non-coding regions (which represent ~98% of the genome) of the human genome.            135                         To quote Shlyueva et al., 2014,  “These are exciting times to study transcriptional regulation”.   136  References  Aberle, Hermann, A. Pejmun Haghighi, Richard D. Fetter, Brian D. McCabe, Tiago R. Magalhães, and Corey S. Goodman. 2002. “Wishful Thinking Encodes a BMP Type II Receptor That Regulates Synaptic Growth in Drosophila.” Neuron 33 (4): 545–58. https://doi.org/10.1016/S0896-6273(02)00589-5. Affolter, Markus, and Konrad Basler. 2007. “The Decapentaplegic Morphogen Gradient: From Pattern Formation to Growth Regulation.” Nature Reviews Genetics 8: 663–74. https://doi.org/10.1038/nrg2166. Aki, Takashi, Ko Hashimoto, Masanori Ogasawara, and Eiji Itoi. 2018. “A Whole-Genome Transcriptome Analysis of Articular Chondrocytes in Secondary Osteoarthritis of the Hip.” PLoS ONE 13 (6): 1–17. https://doi.org/10.1371/journal.pone.0199734. Akiyama, Haruhiko and Lefebvre, Véronique. 2011. “Unraveling the Transcriptional Regulatory Machinery in Chondrogenesis.” Journal of Bone and Mineral Research 29 (4): 390–95. https://doi.org/10.1007/s00774-011-0273-9.Unraveling. Akiyama, Haruhiko, Marie-Christine Chaboissier, James F. Martin, Andreas Schedl, and Benoit De Crombrugghe. 2002. “The Transcription Factor Sox9 Has Essential Roles in Successive Steps of the Chondrocyte Differentiation Pathway and Is Required for Expression of Sox5 and Sox6.” Genes & Development 16: 2813–28. https://doi.org/10.1101/gad.1017802.overt. Akiyama, Haruhiko, Jon P Lyons, Yuko Mori-akiyama, Xiaohong Yang, Ren Zhang, Zhaoping Zhang, Jian Min Deng, et al. 2004. “Interactions between Sox9 and β-Catenin Control Chondrocyte Differentiation.” Genes & Development 18: 1072–87. https://doi.org/10.1101/gad.1171104.Wagner. Akiyama, Haruhiko, H Scott Stadler, James F Martin, Takahiro M Ishii, Philip A Beachy, Takashi Nakamura, and Benoit De Crombrugghe. 2007. “Misexpression of Sox9 in Mouse Limb Bud Mesenchyme Induces Polydactyly and Rescues Hypodactyly Mice.” Matrix Biology 26: 224–33. https://doi.org/10.1016/j.matbio.2006.12.002. Akiyama, Ryutaro, Hiroko Kawakami, M Mark Taketo, Sylvia M Evans, Naoyuki Wada, Anna Petryk, and Yasuhiko Kawakami. 2014. “Distinct Populations within Isl1 Lineages Contribute to Appendicular and Facial Skeletogenesis through the β -Catenin Pathway.” Developmental Biology 387: 37–48. https://doi.org/10.1016/j.ydbio.2014.01.001. Allan, Douglas W, Dongkook Park, Susan E St Pierre, Paul H Taghert, Stefan Thor, and St Louis. 2005. “Regulators Acting in Combinatorial Codes Also Act Independently in Single Differentiating Neurons.” Neuron 45: 689–700. https://doi.org/10.1016/j.neuron.2005.01.026. Allan, Douglas W, and Stefan Thor. 2015. “Transcriptional Selectors, Masters, and Combinatorial Codes: Regulatory Principles of Neural Subtype Specification.” WIREs Dev Biol 4: 505–28. https://doi.org/10.1002/wdev.191. Allen, Kelli D, and Yvonne M Golightly. 2015. “Epidemiology of Osteoarthritis: State of the Evidence Kelli.” Curr Opin Rheumatol. 27 (3): 276–83. https://doi.org/10.1097/BOR.0000000000000161.Epidemiology. Allfrey, G, R Faulkner, and A E Mirsky. 1964. “Possible Role in the Regulation of RNA Synthesis.” Proc. Natl. Acad. Sci U.S.A. 315 (1938): 786–94. Alliston, Tamara, Tien C Ko, Yanna Cao, Yao-yun Liang, Xin-hua Feng, Chenbei Chang, and Rik Derynck. 2005. “Repression of Bone Morphogenetic Protein and Activin-Inducible Transcription by Evi-1.” Journal of Biological Chemistry 280 (25): 24227–37. https://doi.org/10.1074/jbc.M414305200. Almalki, Sami G, and Devendra K Agrawal. 2016. “Key Transcription Factors in the Differentiation of Mesenchymal Stem Cells.” Differentiation 92 (1–2): 41–51. https://doi.org/10.1016/j.diff.2016.02.005.Key. Amano, Katsuhiko, Fumitaka Ichida, Atsushi Sugita, Kenji Hata, Masahiro Wada, Yoko Takigawa, Masako 137  Nakanishi, Mikihiko Kogo, Riko Nishimura, and Toshiyuki Yoneda. 2008. “Msx2 Stimulates Chondrocyte Maturation by Controlling Ihh Expression.” The Journal of Biological Chemistry 283 (43): 29513–21. https://doi.org/10.1074/jbc.M803681200. Andrew, E Denker, Andrew R. Haas, Steven B Nicoll, and Rocky S Tuan. 1999. “Chondrogenic Differentiation of Murine C3H10T1/2 Multipotential Mesenchymal Cells : I. Stimulation by Bone Morphogenetic Protein-2 in High-Density Micromass Cultures.” Differentiation 64: 67–76. https://doi.org/10.1046/j.1432-0436.1999.6420067.x. Angerer, H, D Krusch, A Deutsch, G Gruber, D Setznagl, M L Kremser, A Leithner, M H Stradner, and W B Graninger. 2012. “Potential Role of Nuclear Orphan Receptors NR4A1 and NR4A3 in Human Chondrocytes.” Osteoarthritis and Cartilage 20 (2012): S140. https://doi.org/10.1016/j.joca.2012.02.193. Arnold, CD;, D; Gerlach, C; Stelzer, ŁM; Boryń, and A. Rath, M and Stark. 2013. “Genome-Wide Quantitative Enhancer Activity Maps Identified by STARR-Seq.” Science (New York) 339 (March). http://f1000.com/717973160. Arnold, Michael A, Yuri Kim, Michael P Czubryt, Dillon Phan, John Mcanally, Xiaoxia Qi, John M Shelton, James A Richardson, Rhonda Bassel-duby, and Eric N Olson. 2007. “MEF2C Transcription Factor Controls Chondrocyte Hypertrophy and Bone Development.” Developmental Cell 12: 377–89. https://doi.org/10.1016/j.devcel.2007.02.004. Asp, Julia, Maria Thornemo, Sven Inerot, and Anders Lindahl Y. 1998. “The Helix-Loop-Helix Transcription Factors Id1 and Id3 Have a Functional Role in Control of Cell Division in Human Normal and Neoplastic Chondrocytes.” FEBS Letters 438: 85–90. Baburajendran, Nithya, Ralf Jauch, Clara Yueh, Zhen Tan, Kamesh Narasimhan, and Prasanna R Kolatkar. 2011. “Structural Basis for the Cooperative DNA Recognition by Smad4 MH1 Dimers” 39 (18): 8213–22. https://doi.org/10.1093/nar/gkr500. Baburajendran, Nithya, Paaventhan Palasingam, Kamesh Narasimhan, Wenjie Sun, Shyam Prabhakar, Ralf Jauch, and Prasanna R Kolatkar. 2010. “Structure of Smad1 MH1 / DNA Complex Reveals Distinctive Rearrangements of BMP and TGF- b Effectors.” Nucleic Acids Research 38 (10): 3477–88. https://doi.org/10.1093/nar/gkq046. Badis, Gwenael, Michael F Berger, Anthony A Philippakis, Shaheynoor Talukder, Andrew R Gehrke, Savina A Jaeger, Esther T Chan, et al. 2009. “Diversity and Complexity in DNA Recognition by Transcription Factors.” Science 324 (5935): 1720–23. https://doi.org/10.1126/science.1162327.Diversity. Bandyopadhyay, Amitabha, Kunikazu Tsuji, Karen Cox, Brian D Harfe, Vicki Rosen, and Clifford J Tabin. 2006. “Genetic Analysis of the Roles of BMP2, BMP4, and BMP7 in Limb Patterning and Skeletogenesis.” PLoS Genetics 2 (12). https://doi.org/10.1371/journal.pgen.0020216. Banerjee, Sucharita Sen, Mark W. Feinberg, Masafumi Watanabe, Susan Gray, Richard L. Haspel, Diane J. Denkinger, Rodney Kawahara, Hans Hauner, and Mukesh K. Jain. 2003. “The Krüppel-like Factor KLF2 Inhibits Peroxisome Proliferator-Activated Receptor-γ Expression and Adipogenesis.” Journal of Biological Chemistry 278 (4): 2581–84. https://doi.org/10.1074/jbc.M210859200. Banerji, J.;, and W. Rusconi, S. and Schaffner. 1981. “Expression of a B-Globin Gene Is Enhanced by Remote SV40 DNA Sequences.” Cell 27: 299–308. Barski, Artem;, Suresh; Cuddapah, Kairong; Cui, Tae-Young; Roh, Dustin E.; Schones, Zhibin; Wang, Gang; Wei, Iouri; Chepelev, and Keji Zhao. 2007. “High-Resolution Profiling of Histone Methylations in the Human Genome.” Cell 129: 823–37. https://doi.org/10.1016/j.cell.2007.05.009. Beagrie, Robert A, and Ana Pombo. 2016. “Gene Activation by Metazoan Enhancers: Diverse Mechanisms Stimulate Distinct Steps of Transcription.” BioEssays 38: 881–93. https://doi.org/10.1002/bies.201600032. 138  Beak, Ju Youn, Hong Soon Kang, Yong-sik Kim, and Anton M Jetten. 2007. “Krüppel-Like Zinc Finger Protein Glis3 Promotes Osteoblast Differentiation by Regulating FGF18 Expression.” Journal of Bone and Mineral Research 22: 1234–44. https://doi.org/10.1359/JBMR.070503. Bell, Donald M, Keith K H Leung, Susan C Wheatley, Ling Jim Ng, Sheila Zhou, Kam Wing Ling, Mai Har Sham, Peter Koopman, Patrick P L Tam, and Kathryn S E Cheah. 1997. “SOX9 Directly Regulates the Type-II Collagen Gene.” Nature Genetics 16: 7–11. Benchabane, Hassina, and Jeffrey L Wrana. 2003. “GATA- and Smad1-Dependent Enhancers in the Smad7 Gene Differentially Interpret Bone Morphogenetic Protein Concentrations.” Molecular and Cellular Biology 23 (18): 6646–61. https://doi.org/10.1128/MCB.23.18.6646. Bentley, George. 1975. “Articular Cartilage Studies Osteoarthrosis.” Annals of the Royal College of Surgeons of England 57: 86–100. Berke, Brett, Jessica Wittnam, Elizabeth Mcneill, David L Van Vactor, and Haig Keshishian. 2013. “Retrograde BMP Signaling at the Synapse : A Permissive Signal for Synapse Maturation and Activity-Dependent Plasticity.” Journal of Neuroscience 33 (45): 17937–50. https://doi.org/10.1523/JNEUROSCI.6075-11.2013. Berndt, Anthony J.E.; Tang, Jonathan C.Y.; Ridyard, Marc S.; Lian, Tianshun; Keatings, Kathleen and Allan, Douglas W. 2015. “Gene Regulatory Mechanisms Underlying the Spatial and Temporal Regulation of Target-Dependent Gene Expression in Drosophila Neurons.” PLoS Genetics, 1–27. Berndt, Anthony J.E., Jonathan C.Y. Tang, Marc S. Ridyard, Tianshun Lian, Kathleen Keatings, and Douglas W. Allan. 2015. “Gene Regulatory Mechanisms Underlying the Spatial and Temporal Regulation of Target-Dependent Gene Expression in Drosophila Neurons.” PLoS Genetics 11 (12): 1–27. https://doi.org/10.1371/journal.pgen.1005754. Bernstein, Bradley E., Tarjei S. Mikkelsen, Xiaohui Xie, Michael Kamal, Dana J. Huebert, James Cuff, Ben Fry, et al. 2006. “A Bivalent Chromatin Structure Marks Key Developmental Genes in Embryonic Stem Cells.” Cell 125 (2): 315–26. https://doi.org/10.1016/j.cell.2006.02.041. Bi, Weimin, Jian Min Deng, Zhaoping Zhang, Richard R Behringer, and Benoit De Crombrugghe. 1999. “Sox9 Is Required for Cartilage Formation.” Nature Genetics 22: 85–89. Bi, Weimin, Wendong Huang, Deanne J Whitworth, Jian Min Deng, Zhaoping Zhang, Richard R Behringer, and Benoit De Crombrugghe. 2001. “Haploinsufficiency of Sox9 Results in Defective Cartilage Primordia and Premature Skeletal Mineralization.” Proc Natl Acad Sci USA 98: 6698–6703 兩. Binato, Renata, Cristina E. Alvarez Martinez, Luciana Pizzatti, Benoit Robert, and Eliana Abdelhay. 2006. “SMAD 8 Binding to Mice Msx1 Basal Promoter Is Required for Transcriptional Activation.” Biochemical Journal 393 (1): 141–50. https://doi.org/10.1042/BJ20050327. Bischof, J., R. K. Maeda, M. Hediger, F. Karch, and K. Basler. 2007. “An Optimized Transgenesis System for Drosophila Using Germ-Line-Specific  C31 Integrases.” Proceedings of the National Academy of Sciences 104 (9): 3312–17. https://doi.org/10.1073/pnas.0611511104. Blitz, IL Ira;, and Ken KW Cho. 2009. “Finding Partners: How BMPs Select Their Targets.” Dev Dyn. 238 (6): 1321–31. https://doi.org/10.1002/dvdy.21984. Bobick, Brent E., Peter G. Alexander, and Rocky S. Tuan. 2014. “High Efficiency Transfection of Embryonic Limb Mesenchyme with Plasmid DNA Using Square Wave Pulse Electroporation and Sucrose Buffer.” BioTechniques 56 (2): 85–89. https://doi.org/10.2144/000114136. Boros, Joanna, Nausica Arnoult, Vincent Stroobant, and Anabelle Decottignies. 2014. “Polycomb Repressive Complex 2 and H3K27me3 Cooperate with H3K9 Methylation To Maintain Heterochromatin Protein 1 ␣ At.” Molecular and Cellular Biology 34 (19): 3662–74. https://doi.org/10.1128/MCB.00205-14. Bradley, Elizabeth W, Lomeli R Carpio, Eric N Olson, and Jennifer J Westendorf. 2015. “Histone Deacetylase 7 139  (Hdac7) Suppresses Chondrocyte Proliferation and β-Catenin Activity during Endochondral.” Journal of Biological Chemistry 290 (1): 118–26. https://doi.org/10.1074/jbc.M114.596247. Bray, Nicolas L.;, Harold; Pimentel, Pall; Melsted, and Lior Pachter. 2016. “Near-Optimal Probabilistic RNA-Seq Quantification.” Nature Biotechnology 34 (5): 525–28. Bridgewater, Laura C, Marlan D Walker, Gwen C Miller, Trevor A Ellison, L Daniel Holsinger, Jennifer L Potter, Todd L Jackson, et al. 2003. “Adjacent DNA Sequences Modulate Sox9 Transcriptional Activation at Paired Sox Sites in Three Chondrocyte-Specific Enhancer Elements.” Nucleic Acids Research 31 (5): 1541–53. https://doi.org/10.1093/nar/gkg230. Bridgewater, Laura Clarke;, Veronique; Lefebvre, and Benoit De Crombrugghe. 1998. “Chondrocyte-Specific Enhancer Elements in the Col11a2 Gene Resemble the Col2a1 Tissue-Specific Enhancer.” Journal of Biological Chemistry 273 (24): 14998–6. Brocher, Jan, Benjamin Vogel, and Robert Hock. 2010. “HMGA1 Down-Regulation Is Crucial for Chromatin Composition and a Gene Expression Profile Permitting Myogenic Differentiation.” BMC Cell Biology 11 (64): 1–12. Brown, Carl O, Xuan Chi, Eduardo Garcia-Gras, Manabu Shirai, Xin-Hua Feng, and Robert J Schwartz. 2004. “The Cardiac Determination Factor, Nkx2-5 , Is Activated by Mutual Cofactors GATA-4 and Smad1/4 via a Novel Upstream Enhancer.” Journal of Biological Chemistry 279 (11): 10659–69. https://doi.org/10.1074/jbc.M301648200. Brugger, Sean M., Amy E. Merrill, Jesus Torres-Vazquez, Nancy Wu, Man Chun Ting, Jane Y.M. Cho, Sonia L. Dobias, et al. 2004. “A Phylogenetically Conserved Cis-Regulatory Module in the Msx2 Promoter Is Sufficient for BMP-Dependent Transcription in Murine and Drosophila Embryos.” Development 131 (20): 5153–65. https://doi.org/10.1242/dev.01390. Bulger, Michael, and Mark Groudine. 2011. “Functional and Mechanistic Diversity of Distal Transcription Enhancers.” Cell 144 (3): 327–39. https://doi.org/10.1016/j.cell.2011.01.024. Burke, Tara L, and Patrick A Grant. 2010. “Histone Acetylation Complexes.” In Handbook of Cell Signaling, 2nd ed., 3:2369–78. Elsevier Inc. https://doi.org/10.1016/B978-0-12-374145-5.00285-0. Butler, Jill S, Evangelia Koutelou, Andria C Schibler, and Sharon YR and Dent. 2012. “Histone-Modifying Enzymes: Regulators of Developmental Decisions and Drivers of Human Disease.” Epigenomics 4 (2): 163–77. https://doi.org/10.2217/epi.12.3.Histone-modifying. Byun, Sang Kyung, Tae Hyeon An, Min Jeong Son, Da Som Lee, Hyun Sup Kang, Eun-woo Lee, Baek Soo Han, et al. 2017. “HDAC11 Inhibits Myoblast Differentiation through Repression of MyoD-Dependent Transcription.” Molecules and Cells 40 (9): 667–76. Cairns, Dana M, Renjing Liu, Manpreet Sen, James P Canner, Aaron Schindeler, David G Little, and Li Zeng. 2012. “Interplay of Nkx3.2, Sox9 and Pax3 Regulates Chondrogenic Differentiation of Muscle Progenitor Cells.” PloS One 7 (7): 1–17. https://doi.org/10.1371/journal.pone.0039642. Calo, Eliezer;, and Joanna Wysocka. 2013. “Modification of Enhancer Chromatin: What, How and Why?” Mol Cell 49 (5): 1–24. https://doi.org/10.1016/j.molcel.2013.01.038.Modification. Candia, Albert F, Jinxin Hu, Jeffrey Crosby, Peter A Lalley, Drew Noden, Joseph Nadeau, and Christopher V E Wright. 1992. “Mox-1 and Mox-2 Define a Novel Homeobox Gene Subfamily and Are Differentially Expressed during Early Mesodermal Patterning in Mouse Embryos.” Development 116: 1123–36. Candia, Albert F, and Christopher V E Wright. 1996. “Differential Localization of Mox-1 and Mox-2 Proteins Indicates Distinct Roles during Development.” Int. J. Dev. Biol. 40: 1179–84. Capdevila, Javier, Tohru Tsukui, Concepción Rodríguez Esteban, Vincenzo Zappavigna, and Juan Carlos Izpisúa Belmonte. 1999. “Control of Vertebrate Limb Outgrowth by the Proximal Factor Meis2 and Distal 140  Antagonism of BMPs by Gremlin.” Molecular Cell 4: 839–49. https://doi.org/10.1016/S1097-2765(00)80393-7. Carbon, Seth, Amelia Ireland, Christopher J Mungall, Shengqiang Shu, Brad Marshall, Suzanna Lewis, Amigo Hub, and Web Presence. 2009. “AmiGO: Online Access to Ontology and Annotation Data.” Bioinformatics 25 (2): 288–89. https://doi.org/10.1093/bioinformatics/btn615. Cardelli, Marco, and Jane E Aubin. 2014. “ERRgamma Is Not Required for Skeletal Development but Is a RUNX2-Dependent Negative Regulator of Postnatal Bone Formation in Male Mice.” PloS One 9 (10): 1–11. https://doi.org/10.1371/journal.pone.0109592. Cash, David E, Cheryl B Bock, Klaus Schughart, Elwood Linney, and T Michael Underhill. 1997. “Retinoic Acid Receptor α Function in Vertebrate Limb Skeletogenesis: A Modulator of Chondrogenesis.” Journal of Cell Biology 136 (2): 445–57. Chai, Nan, Wan-xin Li, Jue Wang, Zhi-xin Wang, Shi-ming Yang, and Jia-wei Wu. 2015. “Structural Basis for the Smad5 MH1 Domain to Recognize Different DNA Sequences.” Nucleic Acids Research 43 (18): 9051–64. https://doi.org/10.1093/nar/gkv848. Chantry, Andrew. 2011. “WWP2 Ubiquitin Ligase and Its Isoforms: New Biological Insight and Promising Disease Targets.” Cell Cycle 10 (15): 2437–39. https://doi.org/10.4161/cc.10.15.16874. Chen, Li, Trine Fink, Xiao-yan Zhang Peter, Ebbesen Vladimir Zachar, Li Chen, Á Trine Fink, Á Peter Ebbesen, and Á Vladimir Zachar. 2005. “Quantitative Transcriptional Profiling of ATDC5 Mouse Progenitor Cells during Chondrogenesis.” Differentiation 73 (7): 350–63. https://doi.org/10.1111/j.1432-0436.2005.00038.x. Chen, Weiwei, Xinli Zhang, Ronald K Siu, Feng Chen, Jia Shen, Janette N Zara, Cymbeline T Culiat, Sotirios Tetradis, Kang Ting, and Chia Soo. 2011. “Nfatc2 Is a Primary Response Gene of Nell-1 Regulating Chondrogenesis in ATDC5 Cells.” Journal of Bone and Mineral Research 26 (6): 1230–41. https://doi.org/10.1002/jbmr.314. Cheng, Aixin, and Paul Genever. 2010. “SOX9 Determines RUNX2 Transactivity by Directing Intracellular Degradation.” Journal of Bone and Mineral Research 25 (12): 2680–89. https://doi.org/10.1002/jbmr.174. Cheng, Su-li, Jian-su Shao, Nichole Charlton-Kachigian, Arleen P Loewy, and Dwight A Towler. 2003. “Msx2 Promotes Osteogenesis and Suppresses Adipogenic Differentiation of Multipotent Mesenchymal Progenitors.” The Journal of Biological Chemistry 278 (46): 45969–77. https://doi.org/10.1074/jbc.M306972200. Chimal-Monroy, J, J Rodriguez-Leon, J Montero, Y Ganan, D Macias, R Merino, and J Hurle. 2003. “Analysis of the Molecular Cascade Responsible for Mesodermal Limb Chondrogenesis: Sox Genes and BMP Signaling.” Developmental Biology 257: 292–301. https://doi.org/10.1016/S0012-1606(03)00066-6. Chintapalli, Venkateswara R, Jing Wang, and Julian A T Dow. 2007. “Using FlyAtlas to Identify Better Drosophila Melanogaster Models of Human Disease.” Nature Genetics Perspective 39 (6): 715–20. https://doi.org/10.1038/ng2049. Cho, Ken W.Y. 2012. “Enhancers: WIRE Review.” Wiley Interdiscip Rev Dev Biol. 1 (4): 469–78. https://doi.org/10.1002/wdev.53.Enhancers. Cleary, Mairéad A, Gerjo J V M Van Osch, Pieter A Brama, Catharine A Hellingman, and Roberto Narcisi. 2015. “FGF, TGF b and Wnt Crosstalk: Embryonic to in Vitro Cartilage Development from Mesenchymal Stem Cells.” Journal of Tissue Engineering and Regenerative Medicine 9: 332–42. https://doi.org/10.1002/term. Conrad, Willliam, Michael B Major, Michele A Cleary, Marc Ferrer, Shane Marine, Namjin Chung, William T Arthur, Randall T Moon, Jason D Berndt, and Andy J Chien. 2013. “FAM129B Is a Novel Regulator of Wnt / β-Catenin Signal Transduction in Melanoma Cells.” F1000Research 2 (134): 1–17. https://doi.org/10.12688/f1000research.2-134.v1. Correa, Diego, Eric Hesse, Dutmanee Seriwatanachai, Riku Kiviranta, Kei Yamana, Lynn Neff, Azeddine Atfi, et al. 141  2010. “Zfp521 Is a Target Gene and Key Effector of Parathyroid Hormone-Related Peptide Signaling in Growth Plate Chondrocytes.” Developmental Cell 19 (4): 533–46. https://doi.org/10.1016/j.devcel.2010.09.008.Zfp521. Craymer, L. 1984. “Synthesis of a Ring Third Chormosome and Its Use in Inserting Markers into Ln(3R)C.” Drosophila Inform. Serv., 3 (60): 234–36. Creyghton, M. P.;, W.; Cheng, A., G.G.; Welstead, T.; Kooistra, B.W.; Carey, Eveline J.; Steine, Jacob; Hanna, et al. 2010. “Histone H3K27ac Separates Active from Poised Enhancers and Predicts Developmental State.” PNAS 107 (50): 21931–36. https://doi.org/10.1073/pnas. Crocker, Justin, Namiko Abe, Lucrezia Rinaldi, Alistair P. Mcgregor, Nicolas Frankel, Shu Wang, Ahmad Alsawadi, et al. 2015. “Low Affinity Binding Site Clusters Confer Hox Specificity and Regulatory Robustness.” Cell 160: 191–203. https://doi.org/10.1016/j.cell.2014.11.041. Crocker, Justin, Ella Preger-ben Noon, and David L Stern. 2016. “The Soft Touch: Low-Affinity Transcription Factor Binding Sites in Development and Evolution.” In Essays on Developmental Biology Part B, 1st ed., 117:455–69. Elsevier Inc. https://doi.org/10.1016/bs.ctdb.2015.11.018. Csaki, Lauren S, Jennifer R Dwyer, Xia Li, Michael H K Nguyen, Jay Dewald, David N Brindley, Aldons Lusis, et al. 2014. “Lipin-1 and Lipin-3 Together Determine Adiposity in Vivo.” Molecular Metabolism 3: 145–54. https://doi.org/10.1016/j.molmet.2013.11.008. D. Allan, S. St. Pierre, I. Miguel-Aliaga and S. Thor. 2003. “Specification of Neuropeptide Cell Identity by the Integration of Retrograde BMP Signaling and a Combinatorial Transcription Factor Code.” Cell 113: 73–86. Dailey, Lisa. 2015. “High Throughput Technologies for the Functional Discovery of Mammalian Enhancers: New Approaches for Understanding Transcriptional Regulatory Network Dynamics.” Genomics 106 (3): 151–58. https://doi.org/10.1016/j.ygeno.2015.06.004. Daish, C;, R; Blanchard, K; Fox, P; Pivonka, and E Pirogova. 2018. “The Application of Pulsed Electromagnetic Fields (PEMFs) for Bone Fracture Repair: Past and Perspective Findings.” Annals of Biomedical Engineering 46 (4): 525–42. https://doi.org/10.1007/s10439-018-1982-1. Daoud, Georges, Herve Kempf, Deepak Kumar, Elena Kozhemyakina, Tamara Holowacz, Dae-won Kim, Andreia Ionescu, and Andrew B Lassar. 2014. “BMP-Mediated Induction of GATA4/5/6 Blocks Somitic Responsiveness to SHH.” Development 141: 3978–87. https://doi.org/10.1242/dev.111906. Das, Pradeep, Lisa L Maduzia, Huang Wang, Alyce L Finelli, Seo-hee Cho, Marsha M Smith, and Richard W Padgett. 1998. “The Drosophila Gene Medea Demonstrates the Requirement for Different Classes of Smads in Dpp Signaling.” Development 125: 1519–28. Dehne, Tilo, Rita Schenk, Carsten Perka, Lars Morawietz, Axel Pruss, Michael Sittinger, Christian Kaps, and Jochen Ringe. 2010. “Gene Expression Profiling of Primary Human Articular Chondrocytes in High-Density Micromasses Reveals Patterns of Recovery, Maintenance, Re- and Dedifferentiation.” Gene 462: 8–17. https://doi.org/10.1016/j.gene.2010.04.006. Deignan, Lisa, Marco T Pinheiro, Catherine Sutcliffe, Abbie Saunders, G Wilcockson, Leo A H Zeef, Ian J Donaldson, and Hilary L Ashe. 2016. “Regulation of the BMP Signaling-Responsive Transcriptional Network in the Drosophila Embryo.” PLoS Genetics 12 (7): 1–23. https://doi.org/10.1371/journal.pgen.1006164. Dobin, Alexander, Carrie A Davis, Felix Schlesinger, Jorg Drenkow, Chris Zaleski, Sonali Jha, Philippe Batut, Mark Chaisson, and Thomas R Gingeras. 2013. “STAR: Ultrafast Universal RNA-Seq Aligner.” Bioinformatics 29 (1): 15–21. https://doi.org/10.1093/bioinformatics/bts635. Dong, Yufeng, Alana M Jesse, Anat Kohn, Lea M Gunnell, Tasuku Honjo, Michael J Zuscik, Regis J O Keefe, and Matthew J Hilton. 2010. “RBPjk-Dependent Notch Signaling Regulates Mesenchymal Progenitor Cell Proliferation and Differentiation during Skeletal Development.” Development 1471: 1461–71. 142  https://doi.org/10.1242/dev.042911. Dorighi, Kristel M, Tomek Swigut, Telmo Henriques, Natarajan V Bhanu, Benjamin S Scruggs, Nataliya Nady, Christopher Still, Benjamin A Garcia, Karen Adelman, and Joanna Wysocka. 2017. “Mll3 and Mll4 Facilitate Enhancer RNA Synthesis and Transcription from Promoters Independently of H3K4 Monomethylation.” Molecular Cell 66: 568-576.e4. https://doi.org/10.1016/j.molcel.2017.04.018. Douglas, Adam T., and Robert E. Hill. 2014. “Variation in Vertebrate Cis-Regulatory Elements in Evolution and Disease.” Transcription 5 (MAY). https://doi.org/10.4161/trns.28848. Dréau, Gwenvael Le, Murielle Saade, Irene Gutiérrez-vallejo, and Elisa Martí. 2014. “The Strength of SMAD1/5 Activity Determines the Mode of Stem Cell Division in the Developing Spinal Cord.” Journal of Cell Biology 204 (4): 591–605. https://doi.org/10.1083/jcb.201307031. Drissi, M Hicham, Xufeng Li, Tzong J Sheu, Michael J Zuscik, Edward M Schwarz, J Edward Puzas, Randy N Rosier, and Regis J O Keefe. 2003. “Runx2 / Cbfa1 Stimulation by Retinoic Acid Is Potentiated by BMP2 Signaling Through Interaction With Smad1 on the Collagen X Promoter in Chondrocytes.” Journal of Cellular Biochemistry 90: 1287–98. https://doi.org/10.1002/jcb.10677. Dy, Peter, Weihuan Wang, Pallavi Bhattaram, Qiuqing Wang, Lai Wang, R. Tracy Ballock, and Véronique Lefebvre. 2012. “Sox9 Directs Hypertrophic Maturation and Blocks Osteoblast Differentiation of Growth Plate Chondrocytes.” Developmental Cell 22 (3): 597–609. https://doi.org/10.1016/j.devcel.2011.12.024.Sox9. Eade, K. T., and D. W. Allan. 2009. “Neuronal Phenotype in the Mature Nervous System Is Maintained by Persistent Retrograde Bone Morphogenetic Protein Signaling.” Journal of Neuroscience 29 (12): 3852–64. https://doi.org/10.1523/JNEUROSCI.0213-09.2009. El-Magd, Mohammed Abu, Steve Allen, Imelda McGonnell, Ali A Mansour, Anthony Otto, and Ketan Patel. 2015. “Shh Regulates Chick Ebf1 Gene Expression in Somite Development.” Gene 554: 87–95. https://doi.org/10.1016/j.gene.2014.10.028. Evans, Nicole C, Christina I Swanson, and Scott Barolo. 2012. Sparkling Insights into Enhancer Structure, Function, and Evolution. Current Topics in Developmental Biology: Transcriptional Switches During Development. 1st ed. Vol. 98. Elsevier Inc. https://doi.org/10.1016/B978-0-12-386499-4.00004-5. Farley, Emma K.;, Katrina M.; Olson, Wei; Zhang, Alexander J.; Brandt, Daniel S.; Rokhsar, and Michael S. and Levine. 2015. “Suboptimization of Developmental Enhancers.” Science 350 (6258): 325–28. https://doi.org/10.1126/science.aac6948.Suboptimization. Ferguson, Christine A, Abigail S Tucker, and Paul T Sharpe. 2000. “Temporospatial Cell Interactions Regulating Mandibular and Maxillary Arch Patterning.” Development 127: 403–12. Furumatsu, Takayuki, Masanao Tsuda, Noboru Taniguchi, Yoshitaka Tajima, and Hiroshi Asahara. 2005. “Smad3 Induces Chondrogenesis through the Activation of SOX9 via CREB-Binding Protein/P300 Recruitent.” Journal of Biological Chemistry 280 (9): 8343–50. https://doi.org/10.1074/jbc.M413913200. Gadi, Jogeswar, Seung-Hyun Jung, Min-Jung Lee, Ajita Jami, Kalyani Ruthala, Kyoung-Min Kim, Nam-Hoon Cho, Han-Sung Jung, Cheol-Hee Kim, and Sung-Kil Lim. 2013. “The Transcription Factor Protein Sox11 Enhances Early Osteoblast Differentiation by Facilitating Proliferation and the Survival of Mesenchymal and Osteoblast Progenitors.” Journal of Biological Chemistry 288 (35): 25400–413. https://doi.org/10.1074/jbc.M112.413377. Gao, Sheng, and Allen Laughon. 2006. “Decapentaplegic-Responsive Silencers Contain Overlapping.” Journal of Biol 281 (35): 25781–90. https://doi.org/10.1074/jbc.M603371200. ———. 2007. “Flexible Interaction of Drosophila Smad Complexes with Bipartite Binding Sites” 1769: 484–96. https://doi.org/10.1016/j.bbaexp.2007.05.006. Gao, Sheng, Janet Steffen, and Allen Laughon. 2005. “Dpp-Responsive Silencers Are Bound by a Trimeric” 280 143  (43): 36158–64. https://doi.org/10.1074/jbc.M506882200. García, Amaya, De Vinuesa, Salim Abdelilah-seyfried, Petra Knaus, An Zwijsen, and Sabine Bailly. 2016. “Cytokine & Growth Factor Reviews BMP Signaling in Vascular Biology and Dysfunction.” Cytokine and Growth Factor Reviews 27: 65–79. https://doi.org/10.1016/j.cytogfr.2015.12.005. Garside, Victoria C, Rebecca Cullum, Olivia Alder, Daphne Y Lu, Ryan Vander Werff, Mikhail Bilenky, Yongjun Zhao, et al. 2015. “SOX9 Modulates the Expression of Key Transcription Factors Required for Heart Valve Development.” Development 142: 4340–50. https://doi.org/10.1242/dev.125252. Gaulton, Kyle J, Takao Nammo, Lorenzo Pasquali, Jeremy M Simon, Paul G Giresi, Marie P Fogarty, Tami M Panhuis, et al. 2010. “A Map of Open Chromatin in Human Pancreatic Islets.” Nature Publishing Group 42 (3): 255–59. https://doi.org/10.1038/ng.530. Gazzerro, Elisabetta, and Ernesto Canalis. 2006. “Bone Morphogenetic Proteins and Their Antagonists.” Rev Endocr Metab Disord 7: 51–65. https://doi.org/10.1007/s11154-006-9000-6. Gerard, Deborah, Florian Schmidt, Martine Schmitz, Rashi Halder, Peter Ebert, Marcel H Schulz, Thomas Sauter, and Lasse Sinkkonen. 2019. “Temporal Enhancer Profiling of Parallel Lineages Identifies AHR and GLIS1 as Regulators of Mesenchymal Multipotency.” Nucleic Acids Research 47 (3): 1141–63. https://doi.org/10.1093/nar/gky1240. Ghanem, Noël, Olga Jarinova, Angel Amores, Qiaoming Long, Gary Hatch, Byung Keon Park, John L R Rubenstein, and Marc Ekker. 2003. “Regulatory Roles of Conserved Intergenic Domains in Vertebrate Dlx Bigene Clusters.” Genome Research 13: 533–43. https://doi.org/10.1101/gr.716103.1. Ghanem, Noel, Man Yu, Jason Long, Gary Hatch, John L R Rubenstein, and Marc Ekker. 2007. “Distinct Cis-Regulatory Elements from the Dlx1/Dlx2 Locus Mark Different Progenitor Cell Populations in the Ganglionic Eminences and Different Subtypes of Adult Cortical Interneurons.” The Journal of Neuroscience 27 (19): 5012–22. https://doi.org/10.1523/JNEUROSCI.4725-06.2007. Gibson, Gary. 1998. “Active Role of Chondrocyte Apoptosis in Endochondral Ossification.” MICROSCOPY RESEARCH AND TECHNIQUE 43: 191–204. Goldring, Mary B, Kaneyuki Tsuchimochi, and Kosei Ijiri. 2006. “The Control of Chondrogenesis.” Journal of Cellular Biochemistry 97: 33–44. https://doi.org/10.1002/jcb.20652. Gong, Yaoqin, Deborah Krakow, Jose Marcelino, Douglas Wilkin, David Chitayat, Riyana Babul-hirji, Louanne Hudgins, et al. 1999. “Heterozygous Mutations in the Gene Encoding Noggin Affect Human Joint Morphogenesis.” Nature Genetics 21: 302–4. Grabherr, Manfred, Brian Hass, Moran Yassour, Joshua Levin, Dawn Thompson, Ido Amit, Xian Adiconis, et al. 2011. “Trinity: Reconstructing a Full-Length Transcriptome without a Genome from RNA-Seq Data.” Nature Biotechnology 29 (7): 644–52. https://doi.org/10.1038/nbt.1883.Trinity. Greer, Eric and Shi, Yang. 2012. “Histone Methylation: A Dynamic Mark in Health, Disease and Inheritance.” Nature Reviews Genetics 13 (5): 343–57. https://doi.org/10.1038/nrg3173.Histone. Groth, Amy; Fish, Matthew; Nusse, Roel and Calos, Michele P. 2004. “Construction of Transgenic Drosophila by Using the Site-Specific Integrase From Phage PhiC31.” Genetics 166: 1775–82. https://doi.org/10.1534/genetics.166.4.1775. Guan Yang, Liang Zhu, Ning Hou, Yu Lan, Xi-Mei Wu, Bin Zhou, and Xiao Yang Yan Teng. 2014. “Osteogenic Fate of Hypertrophic Chondrocytes.” Cell Research 24 (10): 1266–69. https://doi.org/10.1038/cr.2014.111. Guenther, Matthew G, Stuart S Levine, Laurie A Boyer, Rudolf Jaenisch, and Richard A Young. 2007. “A Chromatin Landmark and Transcription Initiation at Most Promoters in Human Cells.” Cell 130 (1): 77–88. https://doi.org/10.1016/j.cell.2007.05.042.A. 144  Guo, Fengjin, Xiaofeng Han, Zhimeng Wu, Zhi Cheng, Qin Hu, Yunpeng Zhao, and Yingxiong Wang. 2016. “ATF6a, a Runx2-Activable Transcription Factor, Is a New Regulator of Chondrocyte Hypertrophy.” Journal of Cell Science 129: 717–28. https://doi.org/10.1242/jcs.169623. Guo, Haitao, and Hongwei Huo. 2017. “A New Algorithm for Identifying Cis-Regulatory Modules Based on Hidden Markov Model.” BioMed Research International. Gurusinghe, Saliya, Nadeeka Bandara, Bryan Hilbert, Gareth Trope, Lexin Wang, and Padraig Strappe. 2019. “Lentiviral Vector Expression of Klf4 Enhances Chondrogenesis and Reduces Hypertrophy in Equine Chondrocytes.” Gene 680: 9–19. https://doi.org/10.1016/j.gene.2018.09.013. Halfon, Marc S, Stephen Gisselbrecht, Jun Lu, Beatriz Estrada, Haig Keshishian, Alan M Michelson, and Uas-xeyfp Ah. 2002. “New Fluorescent Protein Reporters for Use With the Drosophila Gal4 Expression System and for Vital Detection of Balancer Chromosomes.” Genesis 138 (1–2): 135–38. https://doi.org/10.1002/gene.10136. Hamaratoglu, Fisun, Markus Affolter, and George Pyrowolakis. 2014. “Dpp/BMP Signaling in Flies: From Molecules to Biology.” Seminars in Cell and Developmental Biology 32: 128–36. https://doi.org/10.1016/j.semcdb.2014.04.036. Han, Yu, and Veronique; Lefebvre. 2008. “L-Sox5 and Sox6 Drive Expression of the Aggrecan Gene in Cartilage by Securing Binding of Sox9 to a Far-Upstream Enhancer �.” Molecular and Cellular Biology 28 (16): 4999–5013. https://doi.org/10.1128/MCB.00695-08. Hanai, Jun-ichi, Lin Feng Chen, Tomohiko Kanno, Naoko Ohtani-fujita, Woo Young Kim, Wei-hui Guo, Takeshi Imamura, et al. 1999. “Interaction and Functional Cooperation of PEBP2 / CBF with Smads.” Journal of Biological Chemistry 274 (44): 31577–82. Hardingham, Timothy E, Rachel A Oldershaw, and Simon R Tew. 2006. “Cartilage, SOX9 and Notch Signals in Chondrogenesis.” Journal of Anatomy 209: 469–80. https://doi.org/10.1111/j.1469-7580.2006.00630.x. Hardison, R.C. and Taylor, J. 2012. “Genomic Approaches towards Finding Cis-Regulatory Modules in Animals.” Nature Reviews Genetics 13: 469–83. Hardison, Ross C. and Taylor, James. 2012. “Genomic Approaches towards Finding Cis-Regulatory Modules in Animals.” Nature Reviews Genetics 13: 469–83. Hassan, Mohammad Q, Rahul S Tare, Suk Hee Lee, Matthew Mandeville, Maria I Morasso, Amjad Javed, Andre J Van Wijnen, Janet L Stein, Gary S Stein, and Jane B Lian. 2006. “BMP2 Commitment to the Osteogenic Lineage Involves Activation of Runx2 by DLX3 and a Homeodomain Transcriptional Network.” Journal of Biological Chemistry 281 (52): 40515–26. https://doi.org/10.1074/jbc.M604508200. Hata, Akiko, Joan Seoane, Giorgio Lagna, Ermelinda Montalvo, Ali Hemmati-brivanlou, and Joan Massague. 2000. “OAZ Uses Distinct DNA- and Protein-Binding Zinc Fingers in Separate BMP-Smad and Olf Signaling Pathways.” Cell 100: 229–40. Heintzman, Nathaniel D., Rhona K. Stuart, Gary Hon, Yutao Fu, Christina W. Ching, R. David Hawkins, Leah O. Barrera, et al. 2007. “Distinct and Predictive Chromatin Signatures of Transcriptional Promoters and Enhancers in the Human Genome.” Nature Genetics 39 (3): 311–18. https://doi.org/10.1038/ng1966. Heintzman, Nathaniel D, Gary C Hon, R David Hawkins, Pouya Kheradpour, Alexander Stark, Lindsey F Harp, Zhen Ye, et al. 2009. “Histone Modifications at Human Enhancers Reflect Global Cell Type-Specific Gene Expression.” Nature 459 (7243): 108–12. https://doi.org/10.1038/nature07829.Histone. Heinz, Sven, Christopher Benner, Nathanael Spann, Eric Bertolino, Yin C Lin, Peter Laslo, Jason X Cheng, Cornelis Murre, Harinder Singh, and K Christopher. 2011. “Simple Combinations of Lineage-Determining Transcription Factors Prime Cis-Regulatory Elements Required for Macrophage and B Cell Identities.” Mol Cell 38 (4): 576–89. https://doi.org/10.1016/j.molcel.2010.05.004.Simple. Heldin, Carl-Henrik, Kohei Miyazono, and Peter ten Dijke. 1997. “TGF-β Signalling from Cell Membrane to 145  Nucleus through SMAD Proteins.” Nature 390 (December): 465. https://doi.org/10.1038/37284. Henningfeld, Kristine A, Henner Friedle, Sepand Rastegar, and Walter Kno. 2002. “Autoregulation of Xvent-2B; Direct Interaction and Functional Cooperation of Xvent-2 and Smad1.” Journal of Biological Chemistry 277 (3): 2097–2103. https://doi.org/10.1074/jbc.M108524200. Henriques, Telmo, Benjamin S. Scruggs, Michiko O. Inouye, Ginger W. Muse, Lucy H. Williams, Adam B. Burkholder, Christopher A. Lavender, David C. Fargo, and Karen Adelman. 2018. “Widespread Transcriptional Pausing and Elongation Control at Enhancers.” Genes and Development 32 (1): 26–41. https://doi.org/10.1101/gad.309351.117. Hill, Caroline S. 2016. “Transcriptional Control by the SMADs.” Cold Spring Harbor Perspectives in Biology. Hippenmeyer, Simon, Ina Kramer, and Silvia Arber. 2004. “Control of Neuronal Phenotype : What Targets Tell the Cell Bodies.” Trends in Neurosciences 27 (8). https://doi.org/10.1016/j.tins.2004.05.012. Hobert, Oliver; Carrera, Ines; Stefanakis, Nikolaos. 2010. “The Molecular and Gene Regulatory Signature of a Neuron.” Trends in Neurosciences 33 (10): 435–45. https://doi.org/10.1016/j.tins.2010.05.006. Hoffman, Lisa M, Kamal Garcha, Konstantina Karamboulas, Matthew F Cowan, Linsay M Drysdale, William A Horton, and T Michael Underhill. 2006. “BMP Action in Skeletogenesis Involves Attenuation of Retinoid Signaling.” Journal of Cell Biology 174 (1): 101–13. https://doi.org/10.1083/jcb.200604150. Hojo, Hironori, Ung-il Chung, and Shinsuke Ohba. 2017. “Identification of the Gene-Regulatory Landscape in Skeletal Development and Potential Links to Skeletal Regeneration.” Regenerative Therapy 6: 100–107. https://doi.org/10.1016/j.reth.2017.04.001. Hojo, Hironori, Andrew P Mcmahon, and Shinsuke Ohba. 2016. “An Emerging Regulatory Landscape for Skeletal Development.” Trends in Genetics 32 (12): 774–87. https://doi.org/10.1016/j.tig.2016.10.001. Hollnagel, Angela, Verena Oehlmann, Joachim Heymer, Ulrich Ru, and Alfred Nordheim. 1999. “Id Genes Are Direct Targets of Bone Morphogenetic Protein Induction in Embryonic Stem Cells.” The Journal of Biological Chemistry 274 (28): 19838–45. Horton, W;, T; Miyashita, K; Kohno, J R; Hassell, and Y Yamada. 1987. “Identification of a Phenotype-Specific Enhancer in the First Intron of the Rat Collagen II Gene.” Proc Natl Acad Sci USA 84: 8864–68. Hou, Zhenyang, Zhen Wang, Yunxia Tao, Jiaxiang Bai, Binqing Yu, Jining Shen, Houyi Sun, et al. 2019. “KLF2 Regulates Osteoblast Differentiation by Targeting of Runx2.” Laboratory Investigation 99 (2): 271–80. https://doi.org/10.1038/s41374-018-0149-x. Hu, Ming Chang, and Norman D Rosenblum. 2005. “Smad1 , β-Catenin and Tcf4 Associate in a Molecular Complex with the Myc Promoter in Dysplastic Renal Tissue and Cooperate to Control Myc Transcription.” Development 132: 215–25. https://doi.org/10.1242/dev.01573. Hu, Xiaoming, Yuanfei Zhou, Yang Yang, Jie Peng, Tongxing Song, Tao Xu, Hongkui Wei, Siwen Jiang, and Jian Peng. 2016. “Identification of Zinc Finger Protein Bcl6 as a Novel Regulator of Early Adipose Commitment.” Open Biology 6: 1–14. Huang, He, Benjamin R Sabari, Benjamin A Garcia, C David Allis, and Yingming Zhao. 2014. “SnapShot: Histone Modifi Cations.” Cell 159 (2): 458-458.e1. https://doi.org/10.1016/j.cell.2014.09.037. Huang, Kui;, Qing; Li, Yong; Li, Zhihao; Yao, Daowen; Luo, Pengcheng; Rao, and Jingang Xiao. 2018. “Cartilage Tissue Regeneration: The Roles of Cells, Stimulating Factors and Scaffolds.” Current Stem Cell Research & Therapy 13: 547. https://doi.org/10.2174/1574888X12666170608080722. Hudson, John B, Steven D Podos, Kallie Keith, Susan L Simpson, and Edwin L Ferguson. 1998. “The Drosophila Medea Gene Is Required Downstream of Dpp and Encodes a Functional Homolog of Human Smad4.” Development 125 (8): 1407–20. http://www.ncbi.nlm.nih.gov/pubmed/9502722. 146  Huey, Daniel J.;, Jerry C.; Hu, and Kyriacos A. Athanasiou. 2012. “Unlike Bone, Cartilage Regeneration Remains Elusive.” Science 338 (6109): 917–21. https://doi.org/10.1126/science.1222454.Unlike. Hussain, S, M Sun, Y Guo, N Mushtaq, Y Zhao, Y Yuan, N Hussain, et al. 2018. “SFMBT2 Positively Regulates SOX9 and Chondrocyte Proliferation.” Journal of Molecular Medicine 42: 3503–12. Ichida, Fumitaka, Riko Nishimura, Kenji Hata, Takuma Matsubara, Fumiyo Ikeda, Kunihiro Hisada, Hirofumi Yatani, et al. 2004. “Reciprocal Roles of Msx2 in Regulation of Osteoblast and Adipocyte Differentiation.” The Journal of Biological Chemistry 279 (32): 34015–22. https://doi.org/10.1074/jbc.M403621200. Inoue, Fumitaka, Martin Kircher, Beth Martin, Gregory M Cooper, Daniela M Witten, Michael T Mcmanus, Nadav Ahituv, and Jay Shendure. 2017. “A Systematic Comparison Reveals Substantial Differences in Chromosomal versus Episomal Encoding of Enhancer Activity.” Genome Research 27: 38–52. https://doi.org/10.1101/gr.212092.116.7. Ishida, Wataru, Toshiaki Hamamoto, Kiyoshi Kusanagi, and Ken Yagi. 2000. “Smad6 Is a Smad1/5-Induced Smad Inhibitor.” Journal of Biological Chemistry 275 (9): 6075–79. https://doi.org/10.1074/jbc.275.9.6075. Ishida, Wataru, Toshiaki Hamamoto, Kiyoshi Kusanagi, Ken Yagi, Masahiro Kawabata, Kazuhiko Takehara, T Kuber Sampath, Mitsuyasu Kato, and Kohei Miyazono. 2000. “Smad6 Is a Smad1 / 5-Induced Smad Inhibitor.” Journal of Biological Chemistry 275 (9): 6075–79. Ito, K, Z Maruyama, A Sakai, S Izumi, T Moriishi, C A Yoshida, T Miyazaki, et al. 2014. “Overexpression of Cdk6 and Ccnd1 in Chondrocytes Inhibited Chondrocyte Maturation and Caused P53-Dependent Apoptosis without Enhancing Proliferation.” Oncogene, 1862–71. https://doi.org/10.1038/onc.2013.130. Jaeger, Savina A, Esther T Chan, Michael F Berger, Rolf Stottmann, Timothy R Hughes, and Martha L Bulyk. 2010. “Genomics Conservation and Regulatory Associations of a Wide Affinity Range of Mouse Transcription Factor Binding Sites.” Genomics 95 (4): 185–95. https://doi.org/10.1016/j.ygeno.2010.01.002. James, Claudine G, C Thomas G Appleton, Veronica Ulici, T Michael Underhill, and Frank Beier. 2005. “Microarray Analyses of Gene Expression during Chondrocyte Differentiation Identifies Novel Regulators of Hypertrophy.” Molecular Biology of the Cell 16: 5316–33. https://doi.org/10.1091/mbc.E05. James, Claudine G, Anita Woods, T Michael Underhill, and Frank Beier. 2006. “The Transcription Factor ATF3 Is Upregulated during Chondrocyte Differentiation and Represses Cyclin D1 and A Gene Transcription.” BMC Molecular Biology 11: 1–11. https://doi.org/10.1186/1471-2199-7-30. Javier, Anna L, Linda T Doan, Mui Luong, N Soledad Reyes De Mochel, Aixu Sun, Edwin S Monuki, and Ken W Y Cho. 2012. “Bmp Indicator Mice Reveal Dynamic Regulation of Transcriptional Response.” PloS One 7 (9). https://doi.org/10.1371/journal.pone.0042566. Jenkins, Elizabeth;, Jennie B.; Moss, James M.; Pace, and Laura C. Bridgewater. 2005. “The New Collagen Gene COL27A1 Contains SOX9-Responsive Enhancer Elements.” Matrix Biology 24 (3): 177–84. https://doi.org/10.1016/j.matbio.2005.02.004.The. Jeong, Byung-chul, Yong-soo Lee, Yun-yong Park, In-ho Bae, Don-kyu Kim, Seung-hoi Koo, Hong-ran Choi, et al. 2009. “The Orphan Nuclear Receptor Estrogen Receptor-Related Receptor Gamma-Negatively Regulates BMP2-Induced Osteoblast Differentiation and Bone Formation.” Journal of Biological Chemistry 284 (21): 14211–18. https://doi.org/10.1074/jbc.M808345200. Jeong, Juhee, Xue Li, Robert J Mcevilly, Michael G Rosenfeld, Thomas Lufkin, and John L R Rubenstein. 2008. “Dlx Genes Pattern Mammalian Jaw Primordium by Regulating Both Lower Jaw-Specific and Upper Jaw-Specific Genetic Programs.” Development 2916: 2905–16. https://doi.org/10.1242/dev.019778. Ji, Quanbo, Xiaojie Xu, Lei Kang, Yameng Xu, Jingbo Xiao, Stuart B Goodman, Xiang Zhu, et al. 2019. “Hematopoietic PBX-Interacting Protein Mediates Cartilage Degeneration during the Pathogenesis of Osteoarthritis.” Nature Communications 10: 1–15. https://doi.org/10.1038/s41467-018-08277-5. 147  Jin, Chunyuan, Kohsuke Kato, Takahiko Chimura, Takahito Yamasaki, Koji Nakade, Takehide Murata, Hongjie Li, et al. 2006. “Regulation of Histone Acetylation and Nucleosome Assembly by Transcription Factor JDP2.” Nature Structural and Molecular Biology 13 (4): 331–38. https://doi.org/10.1038/nsmb1063. Jin, Chunyuan, Hongjie Li, Takehide Murata, Kailai Sun, Masami Horikoshi, Robert Chiu, and Kazunari K Yokoyama. 2002. “JDP2 , a Repressor of AP-1 , Recruits a Histone Deacetylase 3 Complex To Inhibit the Retinoic Acid-Induced Differentiation of F9 Cells.” Molecular and Cellular Biology 22 (13): 4815–26. https://doi.org/10.1128/MCB.22.13.4815. Jing, Junjun, Robert J Hinton, Yuji Mishina, Ying Liu, Xuedong Zhou, Q Jian, and Materials Sciences. 2014. “Critical Role of Bmpr1a in Mandibular Condyle Growth.” Connect Tissue Res. 55 (01): 73–78. https://doi.org/10.3109/03008207.2014.923858.Critical. Johnson, Christopher W, Laura Hernandez-Lagunas, Weiguo Feng, Vida Senkus, Trevor Williams, and Kristin Bruk Artinger. 2011. “Vgll2a Is Required for Neural Crest Cell Survival during Zebrafish Craniofacial Development.” Developmental Biology 357 (1): 269–81. https://doi.org/10.1016/j.ydbio.2011.06.034.Vgll2a. Jolma, Arttu, Yimeng Yin, Kazuhiro R Nitta, Kashyap Dave, Alexander Popov, Minna Taipale, Martin Enge, Teemu Kivioja, Ekaterina Morgunova, and Jussi Taipale. 2015. “DNA-Dependent Formation of Transcription Factor Pairs Alters Their Binding Specificity.” Nature 527 (7578): 384–88. https://doi.org/10.1038/nature15518. Jones, Susan. 2004. “An Overview of the Basic Helix-Loop-Helix Proteins.” Genome Biology 5 (6). Joos, Helga, Wolfgang Albrecht, Stefan Laufer, Heiko Reichel, and Rolf E Brenner. 2008. “IL-1 β Regulates FHL2 and Other Cytoskeleton-Related Genes in Human Chondrocytes.” Mol Med 14 (3–4): 150–59. https://doi.org/10.2119/2007-00118.Joos. Joshi, Rohit, Jonathan M Passner, Remo Rohs, Rinku Jain, Alona Sosinsky, Michael A Crickmore, Vinitha Jacob, Aneel K Aggarwal, Barry Honig, and Richard S Mann. 2007. “Functional Specificity of a Hox Protein Mediated by the Recognition of Minor Groove Structure.” Cell 131: 530–43. https://doi.org/10.1016/j.cell.2007.09.024. Juhász, Tamás, Csaba Matta, Zoltán Mészár, Georgina Nagy, Zsolt Szíjgyártó, Zsanett Molnár, Bernadett Kolozsvári, Éva Bakó, and Róza Zákány. 2010. “Optimalized Transient Transfection of Chondrogenic Primary Cell Cultures.” Central European Journal of Biology 5 (5): 572–84. https://doi.org/10.2478/s11535-010-0053-x. Kamata, Mayumi, Yoko Okitsu, Tohru Fujiwara, Masahiko Kanehira, Shinji Nakajima, Taro Takahashi, Ai Inoue, et al. 2014. “GATA2 Regulates Differentiation of Bone Marrow-Derived Mesenchymal Stem Cells.” Hematologica 99 (11). https://doi.org/10.3324/haematol.2014.105692. Kaplan, Tommy, Xiao-yong Li, Peter J Sabo, Sean Thomas, John A Stamatoyannopoulos, D Mark, and Michael B Eisen. 2011. “Quantitative Models of the Mechanisms That Control Genome-Wide Patterns of Transcription Factor Binding during Early Drosophila Development.” PLoS Genetics 7 (2). https://doi.org/10.1371/journal.pgen.1001290. Karamboulas, Konstantina, Helen J Dranse, and T Michael Underhill. 2010. “Regulation of BMP-Dependent Chondrogenesis in Early Limb Mesenchyme by TGFb Signals.” Journal of Cell Science, 2068–76. https://doi.org/10.1242/jcs.062901. Karaulanov, Emil, Walter Knöchel, and Christof Niehrs. 2004. “Transcriptional Regulation of BMP4 Synexpression in Transgenic Xenopus.” EMBO Journal 23 (4): 844–56. https://doi.org/10.1038/sj.emboj.7600101. Karmodiya, Krishanpal, Arnaud R. Krebs, Mustapha Oulad-Abdelghani, Hiroshi Kimura, and Laszlo Tora. 2012. “H3K9 and H3K14 Acetylation Co-Occur at Many Gene Regulatory Elements, While H3K14ac Marks a Subset of Inactive Inducible Promoters in Mouse Embryonic Stem Cells.” BMC Genomics 13 (1): 1. https://doi.org/10.1186/1471-2164-13-424. 148  Katagiri, Takenobu, Mana Imada, Takeshi Yanai, Tatsuo Suda, Naoyuki Takahashi, and Ryutaro Kamijo. 2002. “Identification of a BMP-Responsive Element in Id1, the Gene for Inhibition of Myogenesis.” Genes to Cells 7 (9): 949–60. https://doi.org/10.1046/j.1365-2443.2002.00573.x. Kawaida, Reimi, Toshiaki Ohtsuka, Junichi Okutsu, Tohru Takahashi, Yuho Kadono, Hiromi Oda, Atsuhiko Hikita, Kozo Nakamura, Sakae Tanaka, and Hidehiko Furukawa. 2003. “Jun Dimerization Protein 2 (JDP2), a Member of the AP-1 Family of Transcription Factor, Mediates Osteoclast Differentiation Induced by RANKL.” Journal of Experimental Medicine 197 (8): 1029–35. https://doi.org/10.1084/jem.20021321. Kawakami, Yasuhiko, Merce Marti, Hiroko Kawakami, Junji Itou, Thu Quach, Austin Johnson, Setsuko Sahara, et al. 2011. “Islet1-Mediated Activation of the b-Catenin Pathway Is Necessary for Hindlimb Initiation in Mice.” Development 138: 4465–73. https://doi.org/10.1242/dev.065359. Kawakami, Yasuhiko, Masanao Tsuda, Shigeru Takahashi, Noboru Taniguchi, Concepcion Rordriguez Esteban, Michihisa Zemmyo, Takayuki Furumatsu, Martin Lotz, Juan Carlos Izpisua Belmonte, and Hiroshi Asahara. 2005. “Transcriptional Coactivator PGC-1α Regulates Chondrogenesis via Association with Sox9.” Proc Natl Acad Sci USA 102 (7): 2414–19. Kawato, Yoshitaka, Makoto Hirao, Kosuke Ebina, Kenrin Shi, Jun Hashimoto, Yui Honjo, Hideki Yoshikawa, and Akira Myoui. 2012. “Nkx3.2 Promotes Primary Chondrogenic Differentiation by Upregulating Col2a1 Transcription.” PloS ONE 7 (4): 1–8. https://doi.org/10.1371/journal.pone.0034703. Kellis, Manolis, Barbara Wold, Michael P Snyder, Bradley E Bernstein, Anshul Kundaje, Georgi Marinov, Lucas Ward, et al. 2014. “Defining Functional DNA Elements in the Human Genome.” PNAS Pers 111 (17): 6131–38. https://doi.org/10.1073/pnas.1318948111. Kim, Dae-won, and Andrew B Lassar. 2003. “Smad-Dependent Recruitment of a Histone Deacetylase/Sin3A Complex Modulates the Bone Morphogenetic Protein-Dependent Transcriptional Repressor Activity of Nkx3.2.” Molecular and Cellular Biology 23 (23): 8704–17. https://doi.org/10.1128/MCB.23.23.8704. Kim, Jaeseob;, Kirby; Johnson, Hui Ju; Chen, Sean; Carroll, and Allen Laughon. 1997. “Mad Binds to DNA and Directly Mediates Activation Of.” Nature 388 (July): 1132–36. Kim, Jonghyun, Weirong Xing, Jon Wergedal, Jefferson Y Chan, Subburaman Mohan, J Kim, W Xing, J Wergedal, Chan Jy, and Mohan S Targeted. 2010. “Targeted Disruption of Nuclear Factor Erythroid-Derived 2-like 1 in Osteoblasts Reduces Bone Size and Bone Formation in Mice.” Physiol Genomics 40: 100–110. https://doi.org/10.1152/physiolgenomics.00105.2009. Kim, Tae Kyung, and Ramin Shiekhattar. 2015. “Architectural and Functional Commonalities between Enhancers and Promoters.” Cell 162 (5): 948–59. https://doi.org/10.1016/j.cell.2015.08.008. Kimura, Hiroshi, Yoko Hayashi-Takanaka, Yuji Goto, Nanako Takizawa, and Naohito Nozaki. 2008. “The Organization of Histone H3 Modifications as Revealed by a Panel of Specific Monoclonal Antibodies.” Cell Structure and Function 33: 61–73. Korchynskyi, Olexander, and Peter Ten Dijke. 2002. “Identification and Functional Characterization of Distinct Critically Important Bone Morphogenetic Protein-Specific Response Elements in the Id1 Promoter.” Journal of Biological Chemistry 277 (7): 4883–91. https://doi.org/10.1074/jbc.M111023200. Kou, Ikuyo, and Shiro Ikegawa. 2004. “SOX9-Dependent and -Independent Transcriptional Regulation of Human Cartilage Link Protein.” Journal of Biological Chemistry 279 (49): 50942–48. https://doi.org/10.1074/jbc.M406786200. Kouzarides, Tony. 2007. “Chromatin Modifications and Their Function.” Cell 128 (4): 693–705. https://doi.org/10.1016/j.cell.2007.02.005. Kozhemyakina, Elena, Todd Cohen, Tso-Pang Yao, and Andrew B Lassar. 2009. “Parathyroid Hormone-Related Peptide Represses Chondrocyte Hypertrophy through a Protein Phosphatase 2A/Histone Deacetylase 4/MEF2 149  Pathway.” Molecular and Cellular Biology 29 (21): 5751–62. https://doi.org/10.1128/MCB.00415-09. Kozhemyakina, Elena, Andrew B Lassar, and Elazar Zelzer. 2015. “A Pathway to Bone: Signaling Molecules and Transcription Factors Involved in Chondrocyte Development and Maturation.” Development 142: 817–31. https://doi.org/10.1242/dev.105536. Kraan, P M van der, E N Blaney Davidson, A Blom, and W B van den Berg. 2009. “TGF-Beta Signaling in Chondrocyte Terminal Differentiation and Osteoarthritis Modulation and Integration of Signaling Pathways through Receptor-Smads.” Osteoarthritis and Cartilage 17 (12): 1539–45. https://doi.org/10.1016/j.joca.2009.06.008. Kraut, Norbert, Lauren Snider, C Amy Chen, Stephen J Tapscott, and Mark Groudine. 1998. “Requirement of the Mouse I-Mfa Gene for Placental Development and Skeletal Patterning.” The EMBO Journal 17 (21): 6276–88. Krebsbach, Paul H, Ken Nakata, Suzanne M Bernier, Osamu Hatano, Tomoyuki Miyashita, Craig S Rhodes, and Yoshihiko Yamada. 1996. “Identification of a Minimum Enhancer Sequence for the Type II Collagen Gene Reveals Several Core Sequence Motifs in Common with the Link Protein Gene.” Journal of Biological Chemistry 271 (8): 4298–4303. Kretzschmar, Marcus, and Joan Massagué. 1998. “SMADs: Mediators and Regulators of TGF-β Signaling.” Current Opinion in Genetics & Development 8 (1): 103–11. https://doi.org/10.1016/S0959-437X(98)80069-5. Kronenberg, Henry M. 2003. “Developmental Regulation of the Growth Plate.” Nature 423: 332–36. Kumar, Deepak, and Andrew Lassar. 2014. “FGF Maintains Chondrogenic Potential of Limb Bud Mesenchymal Cells by Modulating DNMT3A Recruitment.” Cell Research 8 (5): 1419–31. https://doi.org/10.1016/j.celrep.2014.07.038.FGF. Kurosaka, Marianna, Yuji Ogura, Toshiya Funabashi, and Tatsuo Akema. 2016. “Early Growth Response 3 (Egr3) Contributes a Maintenance of C2C12 Myoblast Proliferation.” Journal of Cellular Physiology 232: 1114–22. https://doi.org/10.1002/jcp.25574. Kusanagi, Kiyoshi, Hirofumi Inoue, Yasuhiro Ishidou, Hiromu K Mishima, Masahiro Kawabata, and Kohei Miyazono. 2000. “Characterization of a Bone Morphogenetic Protein- Responsive Smad-Binding Element.” Molecular Biology of the Cell 11: 555–65. Kvon Evgeny Z., Tomas Kazmar, J. Omar Yanez-Cuna Stampfel Gerald, Michaela Pagani, Schernhuber Katharina, and Dickson Barry and Stark Alexander J. 2014. “Genome-Scale Functional Characterization of Drosophila Developmental Enhancers in Vivo.” Nature Letters 512: 91–95. https://doi.org/10.1038/nature13395. Larson, Matthew H, Luke A Gilbert, Xiaowo Wang, Wendell A Lim, Jonathan S Weissman, and Lei S Qi. 2013. “CRISPR Interference ( CRISPRi ) for Sequence-Specific Control of Gene Expression.” Nature Protocols 8 (11): 2180–96. https://doi.org/10.1038/nprot.2013.132. Le, Tien, Zhiguo Liang, Heeren Patel, Marcus H. Yu, Gitanjali Sivasubramaniam, Matthew Slovitt, Guy Tanentzapf, et al. 2006. “A New Family of Drosophila Balancer Chromosomes with a W- Dfd-GMR Yellow Fluorescent Protein Marker.” Genetics 174 (4): 2255–57. https://doi.org/10.1534/genetics.106.063461. Leboy, Phoebe, Giovi Grasso-Knight, Marina D`Angelo, Susan Volk, Jane Lian, Hicham Drissi, Gary Stein, and Sherrill Adams. 2001. “Smad-Runx Interactions During Chondrocyte Maturation.” The Journal of Bone & Joint Surgery 83-A (S1): 15–22. Lee, Kyoung-Mi, Kwang Hwan Park, Ji Suk Hwang, Moses Lee, Dong Suk Yoon, Hyun Aae Ryu, Ho Sun Jung, et al. 2018. “Inhibition of STAT5A Promotes Osteogenesis by DLX5 Regulation.” Cell Death and Disease 9: 1–13. https://doi.org/10.1038/s41419-018-1184-7. Lee, Kyu-ho, Samuel Evans, Todd Y Ruan, and Andrew B Lassar. 2004. “SMAD-Mediated Modulation of YY1 Activity Regulates the BMP Response and Cardiac-Specific Expression of a GATA4/5/6- Dependent Chick 150  Nkx2.5 Enhancer.” Development 131: 4709–23. https://doi.org/10.1242/dev.01344. Lefebvre, Veronique;, Ping; Li, and Benoit De Crombrugghe. 1998. “A New Long Form of Sox5 (L-Sox5), Sox6 and Sox9 Are Coexpressed in Chondrogenesis and Cooperatively Activate the Type II Collagen Gene.” The EMBO Journal 17 (19): 5718–33. Lefebvre, Véronique, and Mona Dvir-Ginzberg. 2017. “SOX9 and the Many Facets of Its Regulation in the Chondrocyte Lineage.” Connect Tissue Res. 58 (1): 2–14. https://doi.org/10.1080/03008207.2016.1183667.SOX9. Lefebvre, Veronique, Wendong Huang, Vincent R Harley, Peter N Goodfellow, and Benoit D E Crombrugghe. 1997. “SOX9 Is a Potent Activator of the Chondrocyte-Specific Enhancer of the Proa1(II) Collagen Gene.” Molecular and Cellular Biology 17 (4): 2336–46. Lennartsson, Andreas, and Karl Ekwall. 2009. “Histone Modification Patterns and Epigenetic Codes.” Biochimica et Biophysica Acta 1790: 863–68. https://doi.org/10.1016/j.bbagen.2008.12.006. Leonard, Claire, Howard Fuld, Dorothy Frenz, Sherry Downie, Joan Massague, and Stuard Newman. 1991. “Role of Transforming Growth Factor-β in Chondrogenic Pattern Formation in the Embryonic Limb: Stimulation of Mesenchymal Condensation and Fibronectin Gene Expression by Exogenenous TGF-β and Evidence for Endogenous TGF-β-like Activity.” Developmental Biology 145: 99–109. Leung, Keith K H, Ling Jim Ng, Ken K Y Ho, Patrick P L Tam, and Kathryn S E Cheah. 1998. “Regulatory DNA Elements Mediate Developmental Stage- and Tissue-Specific Expression of the Human COL2A1 Gene in Transgenic Mice.” Journal of Cell Biology 141 (6): 1291–1300. Levine, Mike. 2010. “Transcriptional Enhancers in Animal Development and Evolution.” Current Biology 20 (17): R754–R763. https://doi.org/10.1016/j.cub.2010.06.070.Transcriptional. Li-Chin Yao, Ira L. Blitz*, Daniel A. Peiffer*, Sopheap Phin*, Ying Wang, Souichi Ogata, Ken W. Y. Cho, Kavita Arora† and Rahul Warrior. 2006. “Schnurri Transcription Factors from Drosophila and Vertebrates Can Mediate Bmp Signaling through a Phylogenetically Conserved Mechanism.” Development 133 (20): 4025–34. https://doi.org/10.1242/dev.02561. Li, Bo, and Colin N Dewey. 2011. “RSEM: Accurate Transcript Quantification from RNA-Seq Data with or without a Reference Genome.” BMC Bioinformatics 12 (323): 1–16. https://doi.org/10.1186/1471-2105-12-323. Li, Heng, and Richard Durbin. 2009. “Fast and Accurate Short Read Alignment with Burrows- Wheeler Transform.” Bioinformatics 25 (14): 1754–60. https://doi.org/10.1093/bioinformatics/btp324. Li, Ian M H, Ke Liu, Alice Neal, Peter D Clegg, Sarah De Val, and George Bou-gharios. 2018. “Differential Tissue Specific, Temporal and Spatial Expression Patterns of the Aggrecan Gene Is Modulated by Independent Enhancer Elements.” Nature Scientific Reports 8 (950): 1–12. https://doi.org/10.1038/s41598-018-19186-4. Li, Xuelin, Shuyi Nie, Chenbei Chang, Tao Qiu, and Xu Cao. 2006. “Smads Oppose Hox Transcriptional Activities.” Experimental Cell Research 312 (6): 854–64. https://doi.org/10.1016/j.yexcr.2005.12.002. Liang, Hsiao-lan, Mu Xu, Yi-chun Chuang, and Christine Rushlow. 2012. “Response to the BMP Gradient Requires Highly Combinatorial Inputs from Multiple Patterning Systems in the Drosophila Embryo.” Development 139: 1956–64. https://doi.org/10.1242/dev.079772. Lien, Ching-ling, John Mcanally, James A Richardson, and Eric N Olson. 2002. “Cardiac-Specific Activity of an Nkx2 – 5 Enhancer Requires an Evolutionarily Conserved Smad Binding Site.” Developmental Biology 244: 257–66. https://doi.org/10.1006/dbio.2002.0603. Liew, Chong Wee, Jeremie Boucher, Jit Kong Cheong, Cecile Vernochet, Ho-Jin Koh, Cristina Mallol, Kristy Townsend, et al. 2013. “Ablation of TRIP-Br2, a Novel Regulator of Fat Lipolysis, Thermogenesis and Oxidative Metabolism, Prevents Diet-Induced Obesity and Insulin Resistance.” Nature Medicine 19 (2): 217–26. https://doi.org/10.1038/nm.3056.Ablation. 151  Liu, Aimin, and Lee A Niswander. 2005. “Bone Morphogenetic Protein Signaling and Vertebrate Nervous System Development.” Nature Reviews Neuroscience 6 (December): 945–55. https://doi.org/10.1038/nrn1805. Liu, Chia-feng, Marco Angelozzi, Abdul Haseeb, and Veronique Lefebvre. 2018. “SOX9 Is Dispensable for the Initiation of Epigenetic Remodeling and the Activation of Marker Genes at the Onset of Chondrogenesis.” Development 145: 1–15. https://doi.org/10.1242/dev.164459. Liu, Chia-Feng, and Véronique Lefebvre. 2015. “The Transcription Factors SOX9 and SOX5 / SOX6 Cooperate Genome-Wide through Super-Enhancers to Drive Chondrogenesis.” Nucleic Acids Research 43 (17): 8183–8203. https://doi.org/10.1093/nar/gkv688. Liu, Chia-Feng, William E Samsa, Guang Zhou, and Veronique Lefebvre. 2017. “Transcriptional Control of Chondrocyte Specification and Differentiation.” Seminars in Cell and Developmental Biology 62: 34–49. https://doi.org/10.1016/j.semcdb.2016.10.004.Transcriptional. Liu, Jiancheng, Xiwei Wu, Heying Zhang, Gerd P Pfeifer, and Qiang Lu. 2017. “Dynamics of RNA Polymerase II Pausing and Bivalent Histone H3 Methylation during Neuronal Differentiation in Brain Development.” CellReports 20: 1307–18. https://doi.org/10.1016/j.celrep.2017.07.046. Long, Fanxin, and David M Ornitz. 2013. “Development of the Endochondral Skeleton.” Cold Spring Harb Perspect Biol 5: 1–20. Long, Hannah K, Sara L Prescott, and Joanna Wysocka. 2016a. “Ever-Changing Landscapes: Transcriptional Enhancers in Development and Evolution.” Cell 167 (5): 1170–87. https://doi.org/10.1016/j.cell.2016.09.018. ———. 2016b. “Review Ever-Changing Landscapes : Transcriptional Enhancers in Development and Evolution.” Cell 167 (5): 1170–87. https://doi.org/10.1016/j.cell.2016.09.018. López-Rovira, Teresa, Elisabet Chalaux, Joan Massagué, Jose Luis Rosa, and Francesc Ventura. 2002. “Direct Binding of Smad1 and Smad4 to Two Distinct Motifs Mediates Bone Morphogenetic Protein-Specific Transcriptional Activation of Id1 Gene.” Journal of Biological Chemistry 277 (5): 3176–85. https://doi.org/10.1074/jbc.M106826200. Lorberbaum, David S, Andrea I Ramos, Kevin A Peterson, Brandon S Carpenter, David S Parker, Sandip De, Lauren E Hillers, et al. 2016. “An Ancient yet Flexible Cis- Regulatory Architecture Allows Localized Hedgehog Tuning by Patched / Ptch1.” ELife 5: 1–22. https://doi.org/10.7554/eLife.13550. Lorda-Diez, Carlos I, Juan A Montero, Carmen Martinez-Cue, Juan A Garcia-Porrero, and Juan M Hurle. 2009. “Transforming Growth Factors β Coordinate Cartilage and Tendon Differentiation in the Developing Limb Mesenchyme.” Journal of Biological Chemistry 284 (43): 29988–96. https://doi.org/10.1074/jbc.M109.014811. Lorzadeh, Alireza, Misha Bilenky, Colin Hammond, David J H F Knapp, Luolan Li, Paul H Miller, Annaick Carles, et al. 2016. “Nucleosome Density ChIP-Seq Identifies Distinct Chromatin Modification Signatures Associated with MNase Accessibility.” CellReports 17: 2112–24. https://doi.org/10.1016/j.celrep.2016.10.055. Love, Michael I, Wolfgang Huber, and Simon Anders. 2014. “Moderated Estimation of Fold Change and Dispersion for RNA-Seq Data with DESeq2.” Genome Biology 15 (550): 1–21. https://doi.org/10.1186/s13059-014-0550-8. Luger, Karollin; Mader A; Richmond, Robin K;, David F; Sargent, and Timothy J; Richmond. 1997. “Crystal Structure of the Nucleosome Core Particle at 2.8A Resolution.” Nature 7: 251–60. Lui, Julian C, Presley Garrison, Quang Nguyen, Michal Ad, Chithra Keembiyehetty, Weiping Chen, Youn Hee Jee, et al. 2016. “EZH1 and EZH2 Promote Skeletal Growth by Repressing Inhibitors of Chondrocyte Proliferation and Hypertrophy.” Nature Communications 7: 1–12. https://doi.org/10.1038/ncomms13685. Ma, Chi, Ying Zhang, Yu Qing Li, Cheng Chen, Wei Cai, and Yue Lin Zeng. 2015. “The Role of PPARγ in Advanced Glycation End Products-Induced Inflammatory Response in Human Chondrocytes.” PLoS ONE 10 152  (5): 1–12. https://doi.org/10.1371/journal.pone.0125776. Mackie, E J, L Tatarczuch, and M Mirams. 2011. “The Skeleton : A Multi-Functional Complex Organ . The Growth Plate Chondrocyte and Endochondral Ossification.” Journal of Endocrinology 211: 109–21. https://doi.org/10.1530/JOE-11-0048. Manikandan, Muthurangan, Sarah Abuelreich, Mona Elsafadi, Hussain Alsalman, Hassan Almalak, Abdulaziz Siyal, Jamil Amjad Hashmi, et al. 2018. “NR2F1 Mediated Down-Regulation of Osteoblast Differentiation Was Rescued by Bone Morphogenetic Protein-2 (BMP-2) in Human MSC.” Differentiation 104: 36–41. https://doi.org/10.1016/j.diff.2018.10.003. Marqués, Guillermo. 2005. “Morphogens and Synaptogenesis in Drosophila .” Journal of Neurobiology . Hoboken : Wiley Subscription Services, Inc., A Wiley Company . https://doi.org/10.1002/neu.20165. Marqués, Guillermo, Theodore E Haerry, M Lisa Crotty, Mingshan Xue, Bing Zhang, and Michael B O Connor. 2003. “Retrograde Gbb Signaling through the Bmp Type 2 Receptor Wishful Thinking Regulates Systemic FMRFa Expression in Drosophila,” 5457–70. https://doi.org/10.1242/dev.00772. Maruyama, Kenta, Masahiro Fukasaka, Alexis Vandenbon, Tatsuya Saitoh, Takumi Kawasaki, Takeshi Kondo, Kazunari K. Yokoyama, et al. 2012. “The Transcription Factor Jdp2 Controls Bone Homeostasis and Antibacterial Immunity by Regulating Osteoclast and Neutrophil Differentiation.” Immunity 37 (6): 1024–36. https://doi.org/10.1016/j.immuni.2012.08.022. Marzaioli, Viviana, Jason P Mcmorrow, Hannes Angerer, Alyssa Gilmore, Daniel Crean, Davide Zocco, Peadar Rooney, et al. 2012. “Histamine Contributes to Increased RANKL to Osteoprotegerin Ratio Through Altered Nuclear Receptor 4A Activity in Human Chondrocytes” 64 (10): 3290–3301. https://doi.org/10.1002/art.34554. Massague, Joan, Stacy Blain, and Roger Lo. 2000. “TGFb Signaling in Growth Control , Cancer , and Heritable Disorders.” Cell 103: 295–309. Massagué, Joan, Joan Seoane, and David Wotton. 2005. “Smad Transcription Factors.” Genes & Development 19: 2783–2810. https://doi.org/10.1101/gad.1350705.embryo. Mathelier, Anthony, Wenqiang Shi, and Wyeth W Wasserman. 2015. “Identification of Altered Cis -Regulatory Elements in Human Disease.” Trends in Genetics 31 (2): 67–76. https://doi.org/10.1016/j.tig.2014.12.003. Mccabe, Brian D, Sabrina Hom, Hermann Aberle, Richard D Fetter, Guillermo Marques, Theodore E Haerry, Hong Wan, Michael B O Connor, Corey S Goodman, and a Pejmun Haghighi. 2004. “Highwire Regulates Presynaptic BMP Signaling Essential for Synaptic Growth.” Neuron 41 (6): 891–905. https://doi.org/10.1016/S0896-6273(04)00073-X. Mehler, Mark F, Peter C Mabie, Damin Zhang, and John A Kessler. 1997. “Bone Morphogenetic Proteins in the Nervous System.” Trends in Neurosciences 20 (7): 309–17. https://doi.org/10.1016/S0166-2236(96)01046-6. Metsalu, Tauno, and Jaak Vilo. 2015. “ClustVis: A Web Tool for Visualizing Clustering of Multivariate Data Using Principal Component Analysis and Heatmap.” Nucleic Acids Research 43: 566–70. https://doi.org/10.1093/nar/gkv468. Mi, Huaiyu, Anushya Muruganujan, Dustin Ebert, Xiaosong Huang, and D Thomas. 2019. “PANTHER Version 14: More Genomes, a New PANTHER GO-Slim and Improvements in Enrichment Analysis Tools.” Nucleic Acids Research 47: 419–26. https://doi.org/10.1093/nar/gky1038. Miguel-aliaga, Irene, Douglas W Allan, Stefan Thor, and Wit Gbb. 2004. “Independent Roles of the Dachshund and Eyes Absent Genes in BMP Signaling, Axon Pathfinding and Neuronal Specification.” Development 131: 5837–48. https://doi.org/10.1242/dev.01447. Minina, Eleonora, Hans Markus Wenzel, Conny Kreschel, Seth Karp, William Gaffield, Andrew P Mcmahon, and Andrea Vortkamp. 2001. “BMP and Ihh / PTHrP Signaling Interact to Coordinate Chondrocyte Proliferation 153  and Differentiation” 4534: 4523–34. Miyazono K, Maeda S and Imamura T. 2006. “Smad Transcriptional Co-Activators and Co-Repressors.” In Smad Signal Transduction, edited by Peter ten Dijke and Carl-Henrik Heldin, 5:277–93. Dordrecht: Springer Netherlands. Miyazono, Kohei, Yuto Kamiya, and Masato Morikawa. 2010. “Bone Morphogenetic Protein Receptors and Signal Transduction.” J. Biochem. 147 (1): 35–51. https://doi.org/10.1093/jb/mvp148. Morikawa, Masato, Daizo Koinuma, Shuichi Tsutsumi, Eleftheria Vasilaki, Yasuharu Kanki, Carl Henrik Heldin, Hiroyuki Aburatani, and Kohei Miyazono. 2011. “ChIP-Seq Reveals Cell Type-Specific Binding Patterns of BMP-Specific Smads and a Novel Binding Motif.” Nucleic Acids Research 39 (20): 8712–27. https://doi.org/10.1093/nar/gkr572. Morimoto, Riyo, Akitsugu Yamamoto, Yoshihiro Akimoto, and Akiko Obinata. 2011. “Homeoprotein Hex Is Expressed in Mouse Developing Chondrocytes.” The Journal of Biochemistry 150 (1): 61–71. https://doi.org/10.1093/jb/mvr039. Morrell, Nicholas W, Donald B Bloch, Peter Dijke, Marie-jose T H Goumans, Jim Smith, Paul B Yu, and Kenneth D Bloch. 2016. “Targeting BMP Signalling in Cardiovascular Disease and Anaemia.” Nat Rev Cardiol. 13 (2): 106–20. https://doi.org/10.1038/nrcardio.2015.156.Targeting. Mu, Yabing, Shyam Kumar Gudey, and Maréne Landström. 2012. “Non-Smad Signaling Pathways.” Cell Tissue Res 327: 11–20. https://doi.org/10.1007/s00441-011-1201-y. Mueller, Thomas D. 2015. Mechanisms of BMP – Receptor Interaction and Activation. Bone Morphogenic Protein. 1st ed. Vol. 99. Elsevier Inc. https://doi.org/10.1016/bs.vh.2015.06.003. Mukhopadhyay, Krish;, Veronique; Lefebvre, Guang; Zhou, Silvio; Garofalo, James H; Kimura, and Benoit De Crombrugghe. 1995. “Use of a New Rat Chondrosarcoma Cell Line to Delineate a 119-Base Pair Chondrocyte-Specific Enhancer Element and to Define Active Promoter Segments in the Mouse Pro-A1(II) Collagen Gene.” Journal of Biological Chemistry 270 (46): 27711–19. Mulloy, Barbara, and Chris C Rider. 2015. The Bone Morphogenetic Proteins and Their Antagonists. Bone Morphogenic Protein. 1st ed. Vol. 99. Elsevier Inc. https://doi.org/10.1016/bs.vh.2015.06.004. Murakami, Shunichi, Mikio Kan, Wallace L Mckeehan, and Benoit De Crombrugghe. 2000. “Up-Regulation of the Chondrogenic Sox9 Gene by Fibroblast Growth Factors Is Mediated by the Mitogen-Activated Protein Kinase Pathway.” Proc Natl Acad Sci USA 97 (3): 1113–18. Murre, Cornelis, Patrick Schonleber McCaw, and David Baltimore. 1989. “A New DNA Binding and Dimerization Motif in Lmmunoglobulin Enhancer Binding, Daughterless, MyoD, and Myc Proteins.” Cell 56: 777–83. Nakade, K., J. Pan, A. Yoshiki, H. Ugai, M. Kimura, B. Liu, H. Li, et al. 2007. “JDP2 Suppresses Adipocyte Differentiation by Regulating Histone Acetylation.” Cell Death and Differentiation 14 (8): 1398–1405. https://doi.org/10.1038/sj.cdd.4402129. Nakahiro, Takeshi, Hisanori Kurooka, Kentaro Mori, Kazuo Sano, and Yoshifumi Yokota. 2010. “Identification of BMP-Responsive Elements in the Mouse Id2 Gene.” Biochemical and Biophysical Research Communications 399 (3): 416–21. https://doi.org/10.1016/j.bbrc.2010.07.090. Nakamura, Yukio, Koji Yamamoto, Xinjun He, Bungo Otsuki, Youngwoo Kim, Hiroki Murao, Tsunemitsu Soeda, et al. 2011. “Wwp2 Is Essential for Palatogenesis Mediated by the Interaction between Sox9 and Mediator Subunit 25.” Nature Communications 2 (251): 1–23. https://doi.org/10.1038/ncomms1242.Wwp2. Nawachi, Kumiko, Miho Inoue, Satoshi Kubota, Takashi Nishida, Gen Yosimichi, Tohru Nakanishi, Manabu Kanyama, et al. 2002. “Tyrosine Kinase-Type Receptor ErbB4 in Chondrocytes: Interaction with Connective Tissue Growth Factor and Distribution in Cartilage.” FEBS Letters 528: 109–13. https://doi.org/10.1016/S0014-5793(02)03263-5. 154  Nichols, James T, Luyuan Pan, Cecilia B Moens, and Charles B Kimmel. 2013. “Barx1 Represses Joints and Promotes Cartilage in the Craniofacial Skeleton.” Development 140: 2765–75. https://doi.org/10.1242/dev.090639. Nie, Yumin, Hongde Liu, and Xiao Sun. 2013. “The Patterns of Histone Modifications in the Vicinity of Transcription Factor Binding Sites in Human Lymphoblastoid Cell Lines.” PloS One 8 (3): 1–11. https://doi.org/10.1371/journal.pone.0060002. Nordin, Kara, and Carole LaBonne. 2014. “Sox5 Is a DNA Binding Co-Factor for BMP R-Smads That Directs Target Specificity during Patterning of the Early Ectoderm.” Developmental Cell 31 (3): 374–82. https://doi.org/10.1016/j.devcel.2014.10.003.Sox5. Norrie, Jacqueline L, Qiang Li, Swanie Co, Bau-lin Huang, Ding Ding, Jann C Uy, Zhicheng Ji, et al. 2016. “PRMT5 Is Essential for the Maintenance of Chondrogenic Progenitor Cells in the Limb Bud.” Stem Cells and Regeneration 143: 4608–19. https://doi.org/10.1242/dev.140715. Oh, Chun-do, Yue Lu, Shoudan Liang, Yuko Mori-akiyama, Di Chen, and Benoit De Crombrugghe. 2014. “SOX9 Regulates Multiple Genes in Chondrocytes , Including Genes Encoding ECM Proteins, ECM Modification Enzymes , Receptors , and Transporters.” PloS ONE 9 (9). https://doi.org/10.1371/journal.pone.0107577. Ohba, Shinsuke, Xinjun He, Hironori Hojo, Andrew P Mcmahon, and Edythe Broad-cirm. 2016. “Distinct Transcriptional Programs Underlie Sox9 Regulation of the Mammalian Chondrocyte.” Cell Reports 12 (2): 229–43. https://doi.org/10.1016/j.celrep.2015.06.013.Distinct. Ohlen, Tonia Von, Derek Lessing, Roel Nusse, and Joan E. Hooper. 1997. “Hedgehog Signaling Regulates Transcription through Cubitus Interruptus, a Sequence-Specific DNA Binding Protein.” PNAS 94: 2404–9. Ong, Chin-Tong and, and Victor Corces. 2011. “Enhancer Function: New Insights into the Regulation of Tissue- Specific Gene Expression.” Nature Reviews Genetics 12 (4): 283–93. https://doi.org/10.1038/nrg2957.Enhancer. Ono, Noriaki, Wanida Ono, Takashi Nagasawa, and Henry M Kronenberg. 2014. “A Subset of Chondrogenic Cells Provides Early Mesenchymal Progenitors in Growing Bones.” Nat Cell Biol. 16 (12): 1157–67. https://doi.org/10.1038/ncb3067.A. Oughtred, Rose, Chris Stark, Bobby-joe Breitkreutz, Jennifer Rust, Lorrie Boucher, Christie Chang, Nadine Kolas, et al. 2019. “The BioGRID Interaction Database : 2019 Update.” Nucleic Acids Research 47: 529–41. https://doi.org/10.1093/nar/gky1079. Park, Peter J. 2009. “ChIPseq: Advantages and Challenges of a Maturing Technology.” Nature Reviews Genetics 10: 669–80. Parker, David S., Michael A. White, Andrea I. Ramos, Barak A. Cohen, and Scott Barolo. 2011. “The Cis-Regulatory Logic of Hedgehog Gradient Responses: Key Roles for Gli Binding Affinity, Competition, and Cooperativity.” Sci Signal 4 (176): 1–16. https://doi.org/10.1126/scisignal.2002077.The. Patro, Rob, Geet Duggal, Michael I Love, Rafael A Irizarry, and Carl Kingsford. 2017. “Salmon: Fast and Bias-Aware Quantification of Transcript Expression Using Dual-Phase Inference.” Nature Methods 14 (4): 417–19. https://doi.org/10.1038/nmeth.4197.Salmon. Pepe, Anna Elena, Qingzhong Xiao, Anna Zampetaki, Zhongyi Zhang, Akira Kobayashi, Yanhua Hu, and Qingbo Xu. 2010. “Crucial Role of Nrf3 in Smooth Muscle Cell Differentiation From Stem Cells.” Circulation Research 106: 870–79. https://doi.org/10.1161/CIRCRESAHA.109.211417. Pertea, Mihaela, Daehwan Kim, Geo Pertea, Jeffrey T Leek, and L Steven. 2016. “Transcript-Level Expression Analysis of RNA-Seq Experiments with HISAT, StringTie, and Ballgown.” Nature Protocols 11 (9): 1650–67. https://doi.org/10.1038/nprot.2016.095.Transcript-level. Petersen, Britt-Sabina, Broder Fredrich, Marc P Hoeppner, David Ellinghaus, and Andre Franke. 2017. 155  “Opportunities and Challenges of Whole-Genome and -Exome Sequencing.” BMC Genetics 18 (14): 1–13. https://doi.org/10.1186/s12863-017-0479-5. Pfeiffer, Barret D, Arnim Jenett, Ann S Hammonds, Teri-t B Ngo, Sima Misra, Christine Murphy, Audra Scully, et al. 2008. “Tools for Neuroanatomy and Neurogenetics in Drosophila.” PNAS 105 (28): 9715–20. Postberg, Jan, Sakeh Forcob, Wei-jen Chang, and Hans J Lipps. 2010. “The Evolutionary History of Histone H3 Suggests a Deep Eukaryotic Root of Chromatin Modifying Mechanisms.” BMC Evolutionary Biology 10 (259): 1–13. https://doi.org/10.1186/1471-2148-10-259. Pouponnot, Celio; Jayaraman, Lata; and Massague, Joan. 1998. “Physical and Functional Interaction of SMADs and P300/CBP.” Journal of Biological Chemistry 273 (36): 22865–69. Pradeepa, Madapura M., Graeme R. Grimes, Yatendra Kumar, Gabrielle Olley, Gillian C.A. Taylor, Robert Schneider, and Wendy A. Bickmore. 2016. “Histone H3 Globular Domain Acetylation Identifies a New Class of Enhancers.” Nature Genetics 48 (6): 681–86. https://doi.org/10.1038/ng.3550. Pratap Singh, Pratik Narendra, Upendra Yadav Singh, Kimi Azad, Pooja Goswami, Veena Kinare, and Amitabha Bandyopadhyay. 2018. “NFIA and GATA3 Are Crucial Regulators of Embryonic Articular Cartilage Differentiation.” Development 145: 1–12. https://doi.org/10.1242/dev.156554. Preitschoph, ANDREA, KATHARINA SCHORGHOFER, DAVID KINSLECHNER, BIRGIT SCHUTZ, HANNES ZWICKL, MARGIT ROSNER, JOZSEF GABOR JOO, STEFAN NEHRER, MARKUS HENGSTSCHLAGER, and MARIO MIKULA. 2016. “Rapamycin-Induced Hypoxia Inducible Factor 2A Is Essential for Chondrogenic Differentiation of Amniotic Fluid Stem Cells.” Fetal and Neonatal Stem Cells 5: 580–90. Pundhir, Sachin, Panayiota Poirazi, and Jan Gorodkin. 2015. “Emerging Applications of Read Profiles towards the Functional Annotation of the Genome.” Frontiers in Genetics 6: 1–11. https://doi.org/10.3389/fgene.2015.00188. Pyrowolakis, George, Britta Hartmann, Bruno Mu, Konrad Basler, and Markus Affolter. 2004. “Widespread BMP-Induced Repression during Drosophila Development.” Developmental Cell 7: 229–40. Qiryaqoz, Zeena, Suraj Timilsina, Daniel Czarnowski, Paul H Krebsbach, and Luis G Villa-diaz. 2019. “Identification of Biomarkers Indicative of Functional Skeletal Stem Cells.” Orthod Craniofac Res 22: 192–98. https://doi.org/10.1111/ocr.12260. Rada-Iglesias, Alvaro, Ruchi Bajpai, Tomek Swigut, Samantha A Brugmann, A Ryan, and Joanna Wysocka. 2011. “A Unique Chromatin Signature Uncovers Early Developmental Enhancers in Humans.” Nature 470 (7333): 279–83. https://doi.org/10.1038/nature09692.A. Ramirez, Fidel, Devon P Ryan, Bjorn Gruning, Vivek Bhardwaj, Fabian Kilpert, Andreas S Richter, Steffen Heyne, Friederike Dundar, and Thomas Manke. 2016. “DeepTools2: A next Generation Web Server for Deep-Sequencing Data Analysis.” Nucleic Acids Research 44: 160–65. https://doi.org/10.1093/nar/gkw257. Ramos, Andrea;, and Scott Barolo. 2013. “Low-Affinity Transcription Factor Binding Sites Shape Morphogen Responses and Enhancer Evolution.” Phil Trans R Soc B 386. Ranger, Ann M, Louis C Gerstenfeld, Jinxi Wang, Tamiyo Kon, Hyunsu Bae, Ellen M Gravallese, Melvin J Glimcher, and Laurie H Glimcher. 2000. “The Nuclear Factor of Activated T Cells (NFAT) Transcription Factor NFATp (NFATc2) Is a Repressor of Chondrogenesis.” Journal of Experimental Medicine 191 (1): 9–21. Raouf, Afshin, and Arun Seth. 2000. “Ets Transcription Factors and Targets in Osteogenesis.” Oncogene 19: 6455–63. Rawson, Joel M, Michael Lee, Eric L Kennedy, and Scott B Selleck. 2003. “Drosophila Neuromuscular Synapse Assembly and Function Require the TGF-β Type I Receptor Saxophone and the Transcription Factor Mad.” 156  Journal of Neurobiology 55 (2): 134–50. https://doi.org/10.1002/neu.10189. Rebeiz, M and Williams, T. 2011. “Experimental Approaches to Evaluate the Contributions of Candidate Cis-Regulatory Mutations to Phenotypic Evolution.” In Molecular Methods for Evolutionary Genetics, 351–75. Rentsendorj, Otgonchimeg;, Andrea; Nagy, Iidiko; Sinko, Andreea Daraba, Endre Barta, and Ibolya Kiss. 2005. “Highly Conserved Proximal Promoter Element Harbouring Paired Sox9-Binding Sites Contributes to the Tissue- and Developmental Stage-Specific Activity of the Matrilin-1 Gene.” Biochemical Journal 398: 705–16. Retting, Kelsey N, Buer Song, Byeong S Yoon, and Karen M Lyons. 2009. “BMP Canonical Smad Signaling through Smad1 and Smad5 Is Required for Endochondral Bone Formation.” Development 1104: 1093–1104. https://doi.org/10.1242/dev.029926. Reynard, Louise N, and John Loughlin. 2013. “Insights from Human Genetic Studies into the Pathways Involved in Osteoarthritis.” Nature Publishing Group 9 (10): 573–83. https://doi.org/10.1038/nrrheum.2013.121. Rice, S. J., K. Cheung, L. N. Reynard, and J. Loughlin. 2019. “Discovery and Analysis of Methylation Quantitative Trait Loci (MQTLs) Mapping to Novel Osteoarthritis Genetic Risk Signals.” Osteoarthritis and Cartilage 27 (10): 1545–56. https://doi.org/10.1016/j.joca.2019.05.017. Robinson, James T., Wendy Thorvaldsdóttir, Helga Winckler, Mitchell Guttman, Eric S. Lander, Gad Getz, and Jill P. and Mesirov. 2011. “Integrative Genomics Viewer.” Nature Biotechnology 29 (1): 24–26. https://doi.org/10.1038/nbt.1754.Integrative. Robinson, Mark D, Davis J Mccarthy, and Gordon K Smyth. 2010. “EdgeR: A Bioconductor Package for Differential Expression Analysis of Digital Gene Expression Data.” Bioinformatics 26 (1): 139–40. https://doi.org/10.1093/bioinformatics/btp616. Rodrigo, Isabel, Paola Bovolenta, Baljinder S Mankoo, Kenji Imai, Rodrigo E T Al, and M O L C E L L B Iol. 2004. “Meox Homeodomain Proteins Are Required for Bapx1 Expression in the Sclerotome and Activate Its Transcription by Direct Binding to Its Promoter.” Molecular and Cellular Biology 24 (7): 2757–66. https://doi.org/10.1128/MCB.24.7.2757. Rodrigo, Isabel, Robert E Hill, Rudi Balling, Andrea Münsterberg, and Kenji Imai. 2003. “Pax1 and Pax9 Activate Bapx1 to Induce Chondrogenic Differentiation in the Sclerotome.” Development 130: 473–82. https://doi.org/10.1242/dev.00240. Ross, Sarah, and Caroline S Hill. 2008. “How the Smads Regulate Transcription.” IJBCB 40: 383–408. https://doi.org/10.1016/j.biocel.2007.09.006. Rudnicki, Julie A, and Anthony M C Brown. 1997. “Inhibition of Chondrogenesis by Wnt Gene Expression in Vivo and in Vitro.” Developmental Biology 118: 104–18. Rushlow, Christine, Pamela F Colosimo, Meng-chi Lin, Mu Xu, and Nikolai Kirov. 2001. “Transcriptional Regulation of the Drosophila Gene Zen by Competing Smad and Brinker Inputs.” Gene & Developent 15: 340–51. https://doi.org/10.1101/gad.861401.Medea. Salazar, Valerie S, Satoshi Ohte, Luciane P Capelo, Laura Gamer, and Vicki Rosen. 2016. “Specification of Osteoblast Cell Fate by Canonical Wnt Signaling Requires Bmp2.” Stem Cells and Regeneration 143: 4352–67. https://doi.org/10.1242/dev.136879. Santos-Rosa, Helena, Robert Schneider, Andrew J Bannister, Julia Sherriff, Bradley E Bernstein, N C Tolga Emre, Stuart L Schreiber, Jane Mellor, and Tony Kouzarides. 2002. “Active Genes Are Tri-Methylated at K4 of Histone H3.” Nature 419: 407–11. https://doi.org/10.1038/nature01071.Published. Schmidt, Dominic, Michael D Wilson, Benoit Ballester, Petra C Schwalie, Gordon D Brown, Aileen Marshall, Claudia Kutter, et al. 2010. “Five-Vertebrate ChIP-Seq Reveals Transcription Factor Binding.” Science (New York, N.Y.) 328 (May): 1036–40. https://doi.org/10.1126/science.1186176. 157  Schmierer, Bernhard, and Caroline S Hill. 2007. “TGFβ – SMAD Signal Transduction: Molecular Specificity and Functional Flexibility.” Nature Reviews Molecular Cell Biology 8: 970–82. https://doi.org/10.1038/nrm2297. Schubert, Thomas, Simone Kaufmann, Ann-kathrin Wenke, Susanne Gra, and Anja-katrin Bosserhoff. 2009. “Role of Deleted in Colon Carcinoma in Osteoarthritis and in Chondrocyte Migration.” Rheumatology 48: 1435–41. https://doi.org/10.1093/rheumatology/kep245. Seki, Kenji, and Akiko Hata. 2004. “Indian Hedgehog Gene Is a Target of the Bone Morphogenetic Protein Signaling Pathway.” Journal of Biological Chemistry 279 (18): 18544–49. https://doi.org/10.1074/jbc.M311592200. Sekiya, Ichiro, Tao Tang, Masaya Hayashi, Toshiyuki Morito, Young-jin Ju, Tomoyuki Mochizuki, and Takeshi Muneta. 2009. “Periodic Knee Injections of BMP-7 Delay Cartilage Degeneration Induced by Excessive Running in Rats.” Journal of Orthopaedic Research 27: 1088–92. https://doi.org/10.1002/jor.20840. Semba, Ichiro, Kazuaki Nonaka, Ichiro Takahashi, Katsu Takahashi, Ralph Dashner, Lillian Shum, Glen H Nuckolls, and Harold C Slavkin. 2000. “Positionally-Dependent Chondrogenesis Induced by BMP4 Is Co-Regulated by Sox9 and Msx2.” Developmental Dynamics 217: 401–14. Sensiate, Lucimara Aparecida, Debora Sobreira, Fernanda Cristina Da Veiga, Denner Jefferson Peterlini, Angelica Vasconcelos Pedrosa, Thais Rirsch, Paulo Pinto Joazeiro, et al. 2014. “Dact Gene Expression Profiles Suggest a Role for This Gene Family in Integrating Wnt and TGF- b Signaling Pathways During Chicken Limb Development.” Developmental Dynamics 243: 428–39. https://doi.org/10.1002/dvdy.23948. Shen, Ming, Takeshi Kawamoto, Weiqun Yan, Kazuko Nakamasu, Mami Tamagami, Yasuhiko Koyano, Mitsuhide Noshiro, and Yukio Kato. 1997. “Molecular Characterization of the Novel Basic Helix-Loop-Helix Protein DEC1 Expressed in Differentiated Human Embryo Chondrocytes.” Biochemical and Biophysical Research Communications 236: 294–98. https://doi.org/10.1006/bbrc.1997.6960. Shi, Xinge, Hui Ye, Xuedong Yao, and Yanzheng Gao. 2017. “The Involvement and Possible Mechanism of NR4A1 in Chondrocyte Apoptosis during Osteoarthritis.” Am J Transl Res 9 (2): 746–54. Shi, Xingming;, Xiangli; Yang, Di; Chen, Zhijie; Chang, and Xu Cao. 1999. “Smad1 Interacts with Homeobox DNA-Binding Proteins in Bone Morphogenetic Protein Signaling.” Journal of Biological Chemistry 274 (19): 13711–17. Shi, Yigong and, and Joan Massague. 2003. “Mechanisms of TGF-Beta Signaling from Cell Membrane to the Nucleus.” Cell 113: 685–700. Shi, Yigong, Yan-fei Wang, Lata Jayaraman, Haijuan Yang, and Joan Massague. 1998. “Crystal Structure of a Smad MH1 Domain Bound to DNA : Insights on DNA Binding in TGF-β Signaling.” Cell 94 (2): 585–94. Shlyueva, D., G. Stampfel, and and Stark A. 2014. “Transcriptional Enhancers: From Properties to Genome-Wide Predictions.” Nature Reviews Genetics 15: 272–86. Shukunami, Chisa, Kiyoto Ishizeki, Tadao Atsumi, Yoshiyuki Ohta, Fujio Suzuki, and Yuji Hiraki. 1997. “Cellular Hypertrophy and Calcification of Embryonal Carcinoma-Derived Chondrogenic Cell Line ATDC5 In Vitro.” Journal of Bone and Mineral Research 12 (8): 1174–88. Siepel, Adam, Gill Bejerano, Jakob S Pedersen, Angie S Hinrichs, Minmei Hou, Kate Rosenbloom, Hiram Clawson, et al. 2005. “Evolutionarily Conserved Elements in Vertebrate, Insect, Worm, and Yeast Genomes.” Genome Research 15: 1034–50. https://doi.org/10.1101/gr.3715005. Silva, Iris A L, Natércia Conceição, Édith Gagnon, Helena Caiado, Jacques P Brown, Fernando Gianfrancesco, Laëtitia Michou, and M Leonor Cancela. 2018. “Effect of Genetic Variants of OPTN in the Pathophysiology of Paget ’ s Disease of Bone.” BBA - Molecular Basis of Disease 1864: 143–51. https://doi.org/10.1016/j.bbadis.2017.10.008. Silva, Susana da, and Fan Wang. 2011. “Retrograde Neural Circuit Specification by Target-Derived Neurotrophins 158  and Growth Factors.” Curr Opin Neurobiol. 21 (1): 61–67. https://doi.org/10.1016/j.conb.2010.07.007.Retrograde. Sivasankaran, Rajeev, M Alessandra Vigano, Bruno Mu, Markus Affolter, and Konrad Basler. 2000. “Direct Transcriptional Control of the Dpp Target Omb by the DNA Binding Protein Brinker.” The EMBO Journal 19 (22): 6162–72. Slattery, Matthew, Todd Riley, Peng; Liu, Namiko; Abe, Pilar; Gomez-Alcala, Iris; Dror, Tianyin; Zhou, et al. 2011. “Cofactor Binding Evokes Latent Differences in DNA Binding Specificity between Hox Proteins.” Cell 147 (6): 1270–82. https://doi.org/10.1016/j.cell.2011.10.053.Cofactor. Small, Kerrin S, Åsa K Hedman, Elin Grundberg, Alexandra C Nica, Gudmar Thorleifsson, Augustine Kong, Unnur Thorsteindottir, et al. 2011. “Identification of an Imprinted Master Trans-Regulator at the KLF14 Locus Related to Multiple Metabolic Phenotypes.” Nature Genetics 43 (6): 561–64. https://doi.org/10.1038/ng.833.Identification. Smallwood, Andrea, and Bing Ren. 2013. “Genome Organization and Long-Range Regulation of Gene Expression by Enhancers.” Current Opinion in Cell Biology 25: 387–94. https://doi.org/10.1016/j.ceb.2013.02.005. Smith, Edwin, and Ali Shilatifard. 2014. “Enhancer Biology and Enhanceropathies.” Nature Structural and Molecular Biology 21 (3): 210–2019. Smith, James Oliver, Alexander Aarvold, Edward R Tayton, Douglas G Dunlop, and Richard O C Oreffo. 2011. “Skeletal Tissue Regeneration: Current Approaches, Challenges, and Novel Reconstructive Strategies for an Aging Population.” Tissue Engineering Part B: Reviews 17 (5): 307–20. https://doi.org/10.1089/ten.teb.2011.0143. Smits, Patrick, Peter Dy, Srijeet Mitra, and Véronique Lefebvre. 2004. “Sox5 and Sox6 Are Needed to Develop and Maintain Source, Columnar, and Hypertrophic Chondrocytes in the Cartilage Growth Plate.” Journal of Cell Biology 164 (5): 747–58. https://doi.org/10.1083/jcb.200312045. Smits, Patrick, Ping Li, Jennifer Mandel, Zhaoping Zhang, Jian Ming Deng, Richard R Behringer, Benoit De Crombrugghe, and Véronique Lefebvre. 2001. “The Transcription Factors L-Sox5 and Sox6 Are Essential for Cartilage Formation.” Developmental Cell 1: 277–90. Song, Buer, Kristine Estrada, and Karen Lyons. 2009. “Smad Signaling in Skeletal Development and Regeneration.” Cytokine and Growth Factor Reviews 20 (5–6): 379–88. https://doi.org/10.1016/j.cytogfr.2009.10.010.Smad. Song, Lin, Lillian Chau, Yoshitaka Sakamoto, Juichiro Nakashima, Masafumi Koide, and Rocky S. Tuan. 2004. “Electric Field-Induced Molecular Vibration for Noninvasive, High-Efficiency DNA Transfection.” Molecular Therapy 9 (4): 607–16. https://doi.org/10.1016/j.ymthe.2004.01.017. Song, Zhuoyue, Xiaolei Lian, Yang Wang, Yong Xiang, and Guangheng Li. 2017. “KLF15 Regulates in Vitro Chondrogenic Differentiation of Human Mesenchymal Stem Cells by Targeting SOX9.” Biochemical and Biophysical Research Communications 493 (2): 1082–88. https://doi.org/10.1016/j.bbrc.2017.09.078. Spaapen, Frank, Guus G H van den Akker, Marjolein M J Caron, Peggy Prickaerts, Celine Rofel, Vivian E H Dahlmans, Don A M Surtel, et al. 2013. “The Immediate Early Gene Product EGR1 and Polycomb Group Proteins Interact in Epigenetic Programming during Chondrogenesis.” PloS One 8 (3): 1–16. https://doi.org/10.1371/journal.pone.0058083. Spitz Francois and Furlong Eileen. 2012. “Transcription Factors: From Enhacer Binding to Developmental Control.” Nature Reviews Genetics 13: 613–26. Sun, Cheng, Hanning Liu, Ke Si, Yaru Wu, Kun Zhao, Ruixia Xu, Zhou Zhou, and Zhe Zheng. 2019. “Meis2 Represses the Osteoblastic Transdifferentiation of Aortic Valve Interstitial Cells through the Notch1/Twist1 Pathway.” Biochemical and Biophysical Research Communications 509: 455–61. https://doi.org/10.1016/j.bbrc.2018.12.040. 159  Suryamohan, Kushal, and Marc S Halfon. 2015. “Identifying Transcriptional Cis -Regulatory Modules” 4 (April). https://doi.org/10.1002/wdev.168. Swanson, Christina I.; Schwimmer, David B.; and Barolo, Scott. 2011. “Rapid Evolutionary Rewiring of a Structurally Constrained Eye Enhancer.” Current Biology 21 (14): 1186–96. https://doi.org/10.1016/j.cub.2011.05.056.Rapid. Szklarczyk, Damian, Annika L Gable, David Lyon, Alexander Junge, Stefan Wyder, Jaime Huerta-cepas, Milan Simonovic, et al. 2019. “STRING V11 : Protein – Protein Association Networks with Increased Coverage, Supporting Functional Discovery in Genome-Wide Experimental Datasets.” Nucleic Acids Research 47: 607–13. https://doi.org/10.1093/nar/gky1131. Takahashi, K, G H Nuckolls, I Takahashi, K Nonaka, M Nagata, T Ikura, H C Slavkin, and L Shum. 2001. “Msx2 Is a Repressor of Chondrogenic Differentiation in Migratory Cranial Neural Crest Cells.” Developmental Dynamics 222: 252–62. https://doi.org/10.1002/dvdy.1185. Takeda, Shu, Jean-Pierre Bonnamy, Michael J Owen, Patricia Ducy, and Gerard Karsenty. 2001. “Continuous Expression of Cbfa1 in Nonhypertrophic Chondrocytes Uncovers Its Ability to Induce Hypertrophic Chondrocyte Differentiation and Partially Rescues Cbfa1-Deficient Mice.” Genes & Development 15: 467–81. https://doi.org/10.1101/gad.845101.Hypertrophic. Taylor, Gillian C.A., Ragnhild Eskeland, Betül Hekimoglu-Balkan, Madapura M. Pradeepa, and Wendy A. Bickmore. 2013. “H4K16 Acetylation Marks Active Genes and Enhancers of Embryonic Stem Cells, but Does Not Alter Chromatin Compaction.” Genome Research 23 (12): 2053–65. https://doi.org/10.1101/gr.155028.113. Teramura, T., J. Hasei, H. Asahara, and M.K. Lotz. 2015. “Krüppel-like Factors (KLFS), Novel Transcriptional Regulators of Articular Chondrocytes That Are Abnormally Expressed in Osteoarthritic Cartilage.” Osteoarthritis and Cartilage 23 (2015): A101. https://doi.org/10.1016/j.joca.2015.02.814. Thomas, Paul D, Anish Kejariwal, Nan Guo, Huaiyu Mi, Michael J Campbell, Anushya Muruganujan, Betty Lazareva-ulitsky, et al. 2006. “Applications for Protein Sequence – Function Evolution Data: MRNA / Protein Expression Analysis and Coding SNP Scoring Tools.” Nucleic Acids Research 34: 645–50. https://doi.org/10.1093/nar/gkl229. Tie, Feng, Rakhee Banerjee, Carl A Stratton, Jayashree Prasad-sinha, Vincent Stepanik, Andrei Zlobin, Manuel O Diaz, Peter C Scacheri, and Peter J Harte. 2009. “CBP-Mediated Acetylation of Histone H3 Lysine 27 Antagonizes Drosophila Polycomb Silencing.” Development 3141: 3131–41. https://doi.org/10.1242/dev.037127. Tolkachov, Alexander, Cornelius Fischer, Thomas H Ambrosi, Melissa Bothe, Chung-ting Han, Matthias Muenzner, Susanne Mathia, et al. 2018. “Loss of the Hematopoietic Stem Cell Factor GATA2 in the Osteogenic Lineage Impairs Trabecularization and Mechanical Strength of Bone.” Molecular and Cellular Biology 38 (12): 1–16. Tong, Qiang, Gokhan Dalgin, Haiyan Xu, Chao-Nan Ting, Jeffrey Leiden, and Gokhan Hotamisligil. 2000. “Function of GATA Transcription Factors in Preadipocyte- Adipocyte Transition.” Science 290: 134–39. Tong, Qiang, Judy Tsai, Guo Tan, Gokhan Dalgin, and Gokhan Hotamisligil. 2005. “Interaction between GATA and the C/EBP Family of Transcription Factors Is Critical in GATA-Mediated Suppression of Adipocyte Differentiation.” Molecular and Cellular Biology 25 (2): 706–15. https://doi.org/10.1128/MCB.25.2.706. Topol, Lilia, Wen Chen, Hai Song, Timothy F Day, and Yingzi Yang. 2009. “Sox9 Inhibits Wnt Signaling by Promoting β-Catenin Phosphorylation in the Nucleus.” Journal of Biological Chemistry 284 (5): 3323–33. https://doi.org/10.1074/jbc.M808048200. Tsai, Sophia Y., and Ming J. Tsai. 1997. “Chick Ovalbumin Upstream Promoter-Transcription Factors (COUP-TFs): Coming of Age.” Endocrine Reviews 18 (2): 229–40. https://doi.org/10.1210/er.18.2.229. 160  Tsuda, Takeshi, Dessislava Markova, Hui Wang, Lucia Evangelisti, Te-cheng Pan, and Mon-li Chu. 2004. “Zinc Finger Protein Zac1 Is Expressed in Chondrogenic Sites of the Mouse.” Developmental Dynamics 229: 340–48. https://doi.org/10.1002/dvdy.10439. Tsuji, Kunikazu, Amitabha Bandyopadhyay, Brian D Harfe, Karen Cox, Sanjeev Kakar, Louis Gerstenfeld, Thomas Einhorn, Clifford J Tabin, and Vicki Rosen. 2006. “BMP2 Activity , Although Dispensable for Bone Formation , Is Required for the Initiation of Fracture Healing.” Nature Genetics 38 (12): 1424–29. https://doi.org/10.1038/ng1916. Tuli, Richard, Suraj Tuli, Sumon Nandi, Xiaoxue Huang, Paul A Manner, William J Hozack, Keith G Danielson, David J Hall, and Rocky S Tuan. 2003. “Transforming Growth Factor-β-Mediated Chondrogenesis of Human Mesenchymal Progenitor Cells Involves N-Cadherin and Mitogen- Activated Protein Kinase and Wnt Signaling Cross-Talk.” Journal of Biological Chemistry 278 (42): 41227–36. https://doi.org/10.1074/jbc.M305312200. Tzchori, Itai, Timothy F Day, Peter J Carolan, Yangu Zhao, Christopher A Wassif, Liqi Li, Mark Lewandoski, et al. 2009. “LIM Homeobox Transcription Factors Integrate Signaling Events That Control Three-Dimensional Limb Patterning and Growth.” Development 136: 1375–85. https://doi.org/10.1242/dev.026476. Ueda, Shogo, Ingrid Rosenburg Cordeiro, Yuuta Moriyama, Chika Nishimori, Kei-ichi Kai, Reiko Yu, Ryoichiro Nakato, Katsuhiko Shirahige, and Mikiko Tanaka. 2019. “Cux2 Refines the Forelimb Field by Controlling Expression of Raldh2 and Hox Genes.” Biology Open 8: 1–10. https://doi.org/10.1242/bio.040584. Underhill, T Michael, Helen J Dranse, and Lisa M Hoffman. 2014. “Analysis of Chondrogenesis Using Micromass Cultures of Limb Mesenchyme BT -.” In Skeletal Development and Repair: Methods and Protocols in Molecular Biolo, edited by Matthew J Hilton, 251–65. Totowa, NJ: Humana Press. https://doi.org/10.1007/978-1-62703-989-5_19. Urist, Marshall R. 1965. “Bone: Formation by Autoiduction.” Science 150 (3698): 893–99. Valenta, Tomas, Jan Lukas, Lenka Doubravska, Bohumil Fafilek, and Vladimir Korinek. 2006. “HIC1 Attenuates Wnt Signaling by Recruitment of TCF-4 and b -Catenin to the Nuclear Bodies.” The EMBO Journal 25 (11): 2326–37. https://doi.org/10.1038/sj.emboj.7601147. Verschueren, Kristin, Jacques E Remacle, Clara Collart, Harry Kraft, Betty S Baker, Przemko Tylzanowski, Luc Nelles, et al. 1999. “SIP1 , a Novel Zinc Finger/Homeodomain Repressor, Interacts with Smad Proteins and Binds to 5`-CACCT Sequences in Candidate Target Genes.” Journal of Biological Chemistry 274 (29): 20489–98. Veverytsa, Lyubov, and Douglas W Allan. 2011. “Retrograde BMP Signaling Controls Drosophila Behavior through Regulation of a Peptide Hormone Battery.” Development 3157: 3147–57. https://doi.org/10.1242/dev.064105. Vortkamp, Andrea, Kaechoong Lee, Beate Lanske, Gino V Segre, Henry M Kronenberg, and Clifford J Tabin. 1996. “Regulation of Rate of Cartilage Differentiation by Indian Hedgehog and PTH-Related Protein.” Science 273 (5275): 613–22. https://doi.org/10.1126/science.273.5275.613. Vuilleumier, Robin, Tianshun Lian, Stephane Flibotte, Zaynah N Khan, Alisa Fuchs, George Pyrowolakis, and Douglas W Allan. 2018. “Retrograde BMP Signaling Activates Neuronal Gene Expression through Widespread Deployment of a Conserved BMP-Responsive Cis -Regulatory Activation Element.” Nucleic Acids Research, 1–21. https://doi.org/10.1093/nar/gky1135. Walsh, Christopher M, and Sean B Carroll. 2007. “Collaboration between Smads and a Hox Protein in Target Gene Repression.” Development 134: 3585–92. https://doi.org/10.1242/dev.009522. Wang, Lai, Yvonne Y. Shao, and R. Tracy Ballock. 2005. “Peroxisome Proliferator Activated Receptor-γ (PPARγ) Represses Thyroid Hormone Signaling in Growth Plate Chondrocytes.” Bone 37 (3): 305–12. https://doi.org/10.1016/j.bone.2005.04.031. 161  Wang, Liguo, Shengqin Wang, and Wei Li. 2012. “RSeQC: Quality Control of RNA-Seq Experiments.” Bioinformatics 28 (16): 2184–85. https://doi.org/10.1093/bioinformatics/bts356. Wang, Qing, Xiaochao Wei, Tianhui Zhu, Ming Zhang, Run Shen, Lianping Xing, Regis J O Keefe, and Di Chen. 2007. “Bone Morphogenetic Protein 2 Activates Smad6 Gene Transcription through Bone-Specific Transcription Runx2.” Journal of Biological Chemistry 282 (14): 10742–48. https://doi.org/10.1074/jbc.M610997200. Wang, Richard N, Jordan Green, Zhongliang Wang, Jixing Ye, Zhengjian Yan, Sahitya Denduluri, Olumuyiwa Idowu, et al. 2014. “ScienceDirect Bone Morphogenetic Protein ( BMP ) Signaling in Development and Human Diseases.” Genes & Diseases 1 (1): 87–105. https://doi.org/10.1016/j.gendis.2014.07.005. Wang, Weiguang, Na Lian, Lingzhen Li, Heather E Moss, Weixi Wang, Daniel S Perrien, Florent Elefteriou, and Xiangli Yang. 2009. “Atf4 Regulates Chondrocyte Proliferation and Differentiation during Endochondral Ossification by Activating Ihh Transcription.” Development 136: 4143–53. https://doi.org/10.1242/dev.043281. Wang, Weiguang, Buer Song, Teni Anbarchian, Anna Shirazyan, Joshua E Sadik, and Karen M Lyons. 2016. “Smad2 and Smad3 Regulate Chondrocyte Proliferation and Differentiation in the Growth Plate.” PLoS Genetics 12 (10): 1–25. https://doi.org/10.1371/journal.pgen.1006352. Wang, Yi, Sika Ristevski, and Vincent R Harley. 2006. “SOX13 Exhibits a Distinct Spatial and Temporal Expression Pattern During Chondrogenesis, Neurogenesis, and Limb Development.” The Journal of Histochemistry & Cytochemistry 54 (12): 1327–33. https://doi.org/10.1369/jhc.6A6923.2006. Wang, Zhao Jun, Hai Bin Zhang, Cheng Chen, Hao Huang, and Jian Xia Liang. 2018. “Effect of PPARG on AGEs-Induced AKT/MTOR Signaling-Associated Human Chondrocytes Autophagy.” Cell Biology International 42 (7): 841–48. https://doi.org/10.1002/cbin.10951. Wang, Zhibin, Chongzhi Zang, Jeffrey A Rosenfeld, Dustin E Schones, Suresh Cuddapah, Kairong Cui, Tae-young Roh, Weiqun Peng, Michael Zhang, and Keji Zhao. 2008. “Combinatorial Patterns of Histone Acetylations and Methylations in the Human Genome.” Nature Genetics 40 (7): 897–903. https://doi.org/10.1038/ng.154.Combinatorial. Warner, Dennis R, Justin P Wells, Robert M Greene, and M Michele Pisano. 2013. “Gene Expression Changes in the Secondary Palate and Mandible of Prdm16−/− Mice.” Cell Ti 351 (3): 445–52. https://doi.org/10.1007/s00441-012-1525-2.Gene. Wasserman, Wyeth W, and Albin Sandelin. 2004. “Applied Bioinformatics for the Identification of Regulatory Elements.” Nature Reviews Genetics 5 (April): 276. https://doi.org/10.1038/nrg1315. Watanabe, Naoko, Yoko Tezuka, Kenji Matsuno, Seiji Miyatani, Naoko Morimura, Masafumi Yasuda, Ryoji Fujimaki, et al. 2003. “Suppression of Differentiation and Proliferation of Early Chondrogenic Cells by Notch.” Journal of Bone and Mineral Metabolism 21: 344–52. https://doi.org/10.1007/s00774-003-0428-4. Weake Vikki and Workman Jerry. 2010. “Inducible Gene Expression: Diverse Regulatory Mechanisms.” Nature Reviews Genetics 11: 426–37. Weiss, Alexander; Charbonnier, Enrica; Ellertsdottir, Elin;, Aristotelis; Tsirigos, Christian; Wolf, Reinhard Schuh, George; Pyrowolakis, and Ma Affolter. 2010. “A Conserved Activation Element in BMP Signaling during Drosophila Development.” Nature Structural and Molecular Biology. 17: 69–77. Weiszmann, Richard, Ann S Hammonds, and Susan E Celniker. 2009. “Determination of Gene Expression Patterns Using High-Throughput RNA in Situ Hybridization to Whole-Mount Drosophila Embryos.” Nature Protocols 4 (5): 605–18. https://doi.org/10.1038/nprot.2009.55. Weston, Andrea D, Roshantha A S Chandraratna, Joseph Torchia, and T Michael Underhill. 2002. “Requirement for RAR-Mediated Gene Repression in Skeletal Progenitor Differentiation.” Journal of Cell Biology 158 (1): 39–162  51. https://doi.org/10.1083/jcb.200112029. Whetstine, Johnathan R. 2010. “Histone Methylation: Chemically Inert but Chromatin Dynamic.” In Handbook of Cell Signaling, Second Edi, 3:2389–97. Elsevier Inc. https://doi.org/10.1016/B978-0-12-374145-5.00287-4. White, Michael A, Davis S Parker, Scott Barolo, and Barak A Cohen. 2012. “A Model of Spatially Restricted Transcription in Opposing Gradients of Activators and Repressors.” Molecular Systems Biology 8 (614): 1–9. https://doi.org/10.1038/msb.2012.48. Whitman, Malcolm. 1998. “Smads and Early Developmental Signaling by the TGF ␤ Superfamily.” Gene & Developent 12: 2445–62. Woods, Anita, Claudine G James, Guoyan Wang, Holly Dupuis, and Frank Beier. 2009. “Control of Chondrocyte Gene Expression by Actin Dynamics: A Novel Role of Cholesterol / Ror-a Signalling in Endochondral Bone Growth.” J. Cell. Mol. Med. 13 (9): 3497–3516. https://doi.org/10.1111/j.1582-4934.2009.00684.x. Wozney, John M, Vicki Rosen, Anthony J Celeste, Lisa M Mitsock, J Matthew, Ronald W Kriz, Rodney M Hewick, et al. 1988. “Novel Regulators of Bone Formation : Molecular Clones and Activities Linked References Are Available on JSTOR for This Article : Novel Regulators of Bone Formation : Molecular Clones and Activities.” Science 242 (4885): 1528–34. Wu, Jinghai, Seetha V. Srinivasan, Jon C. Neumann, and Jerry B. Lingrel. 2005. “The KLF2 Transcription Factor Does Not Affect the Formation of Preadipocytes but Inhibits Their Differentiation into Adipocytes.” Biochemistry 44 (33): 11098–105. https://doi.org/10.1021/bi050166i. Xi, Yang, Wanjing Shen, Lili Ma, Ming Zhao, Jiachen Zheng, Shizhong Bu, Shinjiro Hino, and Mitsuyoshi Nakao. 2016. “HMGA2 Promotes Adipogenesis by Activating C/EBPb-Mediated Expression of PPARg.” Biochemical and Biophysical Research Communications 472: 617–23. https://doi.org/10.1016/j.bbrc.2016.03.015. Xie, Wei-fen, X I N Zhang, Shinji Sakano, Veronique Lefebvre, and Linda J Sandell. 1999. “Trans-Activation of the Mouse Cartilage-Derived Retinoic Acid-Sensitive Protein Gene by Sox9.” Journal of Bone and Mineral Research 14 (5): 757–63. Xu, Lan. 2006. “Regulation of Smad Activities.” Biochim Biophys Acta. 1759 (11–12): 503–13. Xu, Mu, Nikolai Kirov, and Christine Rushlow. 2005. “Peak Levels of BMP in the Drosophila Embryo Control Target Genes by a Feed-Forward Mechanism.” Development, 1637–47. https://doi.org/10.1242/dev.01722. Xu, Pin-Xian, Jane Cheng, Jonathan Epstein, and Richard Maas. 1997. “Mouse Eya Genes Are Expressed during Limb Tendon Development and Encode a Transcriptional Activation Function.” Proc Natl Acad Sci USA 94: 11974–79. Xu, Susan C, Marie A Harris, John L R Rubenstein, Gregory R Mundy, and Stephen E Harris. 2001. “Bone Morphogenetic Protein-2 (BMP-2) Signaling to the Col2a1 Gene in Chondroblasts Requires the Homeobox Gene Dlx2.” DNA and Cell Biology 20 (6): 359–65. Xu, Xiaolei, Zhizhang Yin, John B Hudson, Edwin L Ferguson, and Manfred Frasch. 1998. “Smad Proteins Act in Combination with Synergistic and Antagonistic Regulators to Target Dpp Responses to the Drosophila Mesoderm.” Gene & Developent 12: 2354–70. Yamamoto, Shiori, Yuji Uchida, Tomomi Ohtani, Erina Nozaki, Chunyang Yin, Yoshihiro Gotoh, Nayuta Yakushiji-Kaminatsui, et al. 2019. “Hoxa13 Regulates Expression of Common Hox Target Genes Involved in Cartilage Development to Coordinate the Expansion of the Autopodal Anlage.” Development, Growth & Differentiation 61: 228–51. https://doi.org/10.1111/dgd.12601. Yamashita, Satoshi, Kensuke Katao, Hiroto Yamamoto, Tomoko Kato, Satoshi Hara, Katsushi Yamaguchi, Claire Renard-Guillet, et al. 2019. “Comparative Analysis Demonstrates Cell Type-Specific Conservation of SOX9 Targets between Mouse and Chicken.” Scientific Reports 9: 1–15. https://doi.org/10.1038/s41598-019-48979-163  4. Yammani, Raghunatha R. and Loeser, Richard. 2014. “Stress Inducible Gene Nuclear Protein-1 (NUPR1/P8) Regulates MMP-13 Expression in Human Articular Chondrocytes.” Arthritis Rheumatol. 66 (5): 1266–71. https://doi.org/10.1002/art.38391.Stress. Yan, Jianyun, Jun Li, Jun Hu, Lu Zhang, Chengguo Wei, Nishat Sultana, Xiaoqiang Cai, Weijia Zhang, and Chen-leng Cai. 2018. “Smad4 Deficiency Impairs Chondrocyte Hypertrophy via the Runx2 Transcription Factor in Mouse Skeletal Development.” Journal of Biological Chemistry 293 (24): 9162–75. https://doi.org/10.1074/jbc.RA118.001825. Yanagino, Takuya, Keizo Yuasa, Masami Nagahama, Yoshiko Matsuda, and Akihiko Tsuji. 2009. “Transcriptional Regulation of Fibrillin-2 Gene by E2F Family Members in Chondrocyte Differentiation.” Journal of Cellular Biochemistry 106: 580–88. https://doi.org/10.1002/jcb.22029. Yanez-Cuna, Omar, Evgeny Kvon, and Alexander Stark. 2013. “Deciphering the Transcriptional Cis- Regulatory Code.” Cell Trends in Genetics 29 (1): 11–22. https://doi.org/10.1016/j.tig.2012.09.007. Yang, Tian, Alexander Bassuk, and Bernd Fritzsch. 2013. “Prickle1 Stunts Limb Growth through Alteration of Cell Polarity and Gene Expression.” Dev Dyn. 242 (11): 1293–1306. https://doi.org/10.1002/dvdy.24025.Prickle1. Yap, Kai Lee, Polina Sysa-shah, Brad Bolon, Ren-chin Wu, Min Gao, Alice L Herlinger, Francesco Faiola, et al. 2013. “Loss of NAC1 Expression Is Associated with Defective Bony Patterning in the Murine Vertebral Axis.” PloS ONE 8 (7): 1–16. https://doi.org/10.1371/journal.pone.0069099. Yella, Venkata Rajesh, Devesh Bhimsaria, Debostuti Ghoshdastidar, Jose Rodriguez-Martinez, Aseem Ansari, and Manju Bansal. 2018. “Flexibility and Structure of Flanking DNA Impact Transcription Factor Affinity for Its Core Motif.” Nucleic Acids Research 14;46(22): 11883–97. https://doi.org/10.1093/nar/gky1057. Yoon, Byeong S, Dmitry A Ovchinnikov, Isaac Yoshii, Yuji Mishina, Richard R Behringer, and Karen M Lyons. 2005. “Bmpr1a and Bmpr1b Have Overlapping Functions and Are Essential for Chondrogenesis in Vivo.” PNAS 102 (14): 1–6. Yoshida, Carolina A., and Toshihisa Komori. 2005. “Role of Runx Proteins in Chondrogenesis.” Critical ReviewsTM in Eukaryotic Gene Expression 15 (3): 243–54. 10.1615/CritRevEukarGeneExpr.v15.i3.60. You, Di, Cheng Yang, Jun Huang, Haoli Gong, Mingming Yan, and Jiangdong Ni. 2019. “Long Non-Coding RNA MEG3 Inhibits Chondrogenic Differentiation of Synovium-Derived Mesenchymal Stem Cells by Epigenetically Inhibiting TRIB2 via Methyltransferase EZH2.” Cellular Signalling 63: 1–10. https://doi.org/10.1016/j.cellsig.2019.109379. Yuan, Yuan, Honglue Tan, and Pengyi Dai. 2017. “Krüppel-Like Factor 2 Regulates Degradation of Type II Collagen by Suppressing the Expression of Matrix Metalloproteinase (MMP)-13.” Cellular Physiology and Biochemistry 42 (6): 2159–68. https://doi.org/10.1159/000479991. Zawel, Leigh, Jia Le Dai, Phillip Buckhaults, Shibin Zhou, Kenneth W Kinzler, Bert Vogelstein, and Scott E Kern. 1998. “Human Smad3 and Smad4 Are Sequence-Specific Transcription Activators.” Mol Cell 1: 611–17. Zeng, Kefeng, Hong Qi Zhang, Yong Chen, and Qile Gao. 2016. “Estradiol via Estrogen Receptor Beta Inhibits Chondrogenesis of Mouse Vertebral Growth Plate in Vitro.” Child’s Nervous System 32: 461–65. https://doi.org/10.1007/s00381-015-2973-2. Zeng, Li, Herve Kempf, L Charles Murtaugh, Mie Elissa Sato, and Andrew B Lassar. 2005. “Shh Establishes an Nkx3.2/Sox9 Autoregulatory Loop That Is Maintained by BMP Signals to Induce Somitic Chondrogenesis.” Genes & Development 16: 1990–2005. https://doi.org/10.1101/gad.1008002.tion. Zentner, Gabriel E, Paul J Tesar, and Peter C Scacheri. 2011. “Epigenetic Signatures Distinguish Multiple Classes of Enhancers with Distinct Cellular Functions.” Genome Research 21: 1273–83. https://doi.org/10.1101/gr.122382.111.21. 164  Zhang, Bing, Zhiyong Li, Wei Wang, Jichao Guo, Shufeng Kang, Shizhao Liu, Hongzhu Li, Depeng Wang, and Xiangbei Qi. 2018. “NCOA3 Loss Disrupts Molecular Signature of Chondrocytes and Promotes Posttraumatic Osteoarthritis Progression.” Cellular Physiology and Biochemistry 49: 2396–2413. https://doi.org/10.1159/000493839. Zhang, Cui, Fan-fan Hong, Cui-cui Wang, Liang Li, Jian-ling Chen, Fei Liu, Ren-fu Quan, and Jin-Fu Wang. 2017. “TRIB3 Inhibits Proliferation and Promotes Osteogenesis in HBMSCs by Regulating the ERK1/2 Signaling Pathway.” Scientific Reports 7: 1–13. https://doi.org/10.1038/s41598-017-10601-w. Zhang, Hailan, Michael Levine, and Hilary L Ashe. 2001. “Brinker Is a Sequence-Specific Transcriptional Repressor in the Drosophila Embryo.” Gene & Developent 15: 261–66. https://doi.org/10.1101/gad.861201.GENES. Zhang, Jianfei, Wenbin Zhang, Jun Shi, Jiewen Dai, and Steve Guofang Shen. 2018. “Dlx2 Overexpression Enhanced Accumulation of Type II Collagen and Aggrecan by Inhibiting MMP13 Expression in Mice Chondrocytes.” Biochemical and Biophysical Research Communications 503: 528–35. https://doi.org/10.1016/j.bbrc.2018.05.066. Zhang, Wei, Peichun Hsu, Biao Zhong, Shang Guo, Chi Zhang, Yukai Wang, Congfeng Luo, Yulin Zhan, and Changqing Zhang. 2018. “MiR-34a Enhances Chondrocyte Apoptosis , Senescence and Facilitates Development of Osteoarthritis by Targeting DLL1 and Regulating PI3K/AKT Pathway.” Cellular Physiology and Biochemistry 48: 1304–16. https://doi.org/10.1159/000492090. Zhang, Ying E. 2009. “Non-Smad Pathways in TGF-β Signaling.” Cell Research 19 (1): 128–39. https://doi.org/10.1038/cr.2008.328.Non-Smad. Zhang, Ying, Rong-lin Xie, Jonathan Gordon, Kimberly Leblanc, Janet L Stein, Jane B Lian, Andre J Van Wijnen, and Gary S Stein. 2012. “Control of Mesenchymal Lineage Progression by MicroRNAs Targeting Skeletal Gene Regulators Trps1 and Runx2.” The Journal of Biological Chemistry 287 (26): 21926–35. https://doi.org/10.1074/jbc.M112.340398. Zhang, Yong, Tao Liu, Clifford A Meyer, Jérôme Eeckhoute, David S Johnson, Bradley E Bernstein, Chad Nusbaum, et al. 2008. “Model-Based Analysis of ChIP-Seq (MACS).” Genome Biology 9 (9): R137. https://doi.org/10.1186/gb-2008-9-9-r137. Zhao, Chen, Wei Jiang, Nian Zhou, Junyi Liao, Mingming Yang, Ning Hu, Xi Liang, et al. 2017. “Sox9 Augments BMP2-Induced Chondrogenic Differentiation by Downregulating Smad7 in Mesenchymal Stem Cells (MSCs).” Genes & Diseases 4: 229–39. https://doi.org/10.1016/j.gendis.2017.10.004. Zhao, Haixia, Wenrong Zhou, Zhengju Yao, Yong Wan, Jingjing Cao, Lingling Zhang, Jianzhi Zhao, et al. 2015. “Foxp1/2/4 Regulate Endochondral Ossification as a Suppresser Complex.” Developmental Biology 398 (2): 242–54. https://doi.org/10.1016/j.ydbio.2014.12.007.Foxp1/2/4. Zhao, Yingming, and Benjamin A Garcia. 2015. “Comprehensive Catalog of Currently Documented Histone Modifications.” Cold Spring Harb Perspect Biol 7: 1–22. Zhou, Bo, Lihong Chen, Xing Wu, Jing Wang, Yinliang Yin, and Guang Zhu. 2008. “MH1 Domain of SMAD4 Binds N-Terminal Residues of the Homeodomain of Hoxc9.” Biochimica et Biophysica Acta MH1 1784: 747–52. https://doi.org/10.1016/j.bbapap.2008.01.021. Zhou, Guang, Qiping Zheng, Feyza Engin, Elda Munivez, Yuqing Chen, Eiman Sebald, Deborah Krakow, and Brendan Lee. 2006. “Dominance of SOX9 Function over RUNX2 during Skeletogenesis.” Proc Natl Acad Sci USA 103 (50): 19004–9. Zhou, Xin, Klaus Von Der Mark, Stephen Henry, William Norton, Henry Adams, and Benoit De Crombrugghe. 2014. “Chondrocytes Transdifferentiate into Osteoblasts in Endochondral Bone during Development , Postnatal Growth and Fracture Healing in Mice.” PLoS Genetics 10 (12). https://doi.org/10.1371/journal.pgen.1004820. 165  Zhu, Haitao;, Peter; Kavsak, and Gerald Abdollah, Shirin; Wrana, Jerrey and Thomsen. 1999. “A SMAD Ubiquitin Ligase Targets the BMP Pathway and Affects Embryonic Pattern Formation.” Nature 400: 687–94. Zhu, Lirong, Jian Lan Peng, Karine Harutyunyan, Monica Garcia, Monica Justice, and John Belmont. 2007. “Craniofacial, Skeletal, and Cardiac Defects Associated with Altered Embryonic Murine Zic3 Expression Following Targeted Insertion of a PGK-NEO Cassette.” Frontiers in Bioscience 12: 1680–90. https://doi.org/10.2741/2180. Zou, Weiguo, Xi Chen, Jae Shim, Zhiwei Huang, Nicholas Brady, Dorothy Hu, Kirsten Sigrist, Laurie H Glimcher, and Dallas Jones. 2011. “The E3 Ubiquitin Ligase Wwp2 Regulates Craniofacial Development through Monoubiquitination of Goosecoid.” Nat Cell Biol. 13 (1): 59–65. https://doi.org/10.1038/ncb2134.The. Zuscik, Michael J, Matthew J Hilton, Xinping Zhang, Di Chen, and Regis J O Keefe. 2008. “Review Series Regulation of Chondrogenesis and Chondrocyte Differentiation by Stress.” The Journal of Clinical Investigation 118 (2): 429–38. https://doi.org/10.1172/JCI34174.of. 166   Appendix  A. B. C. Figure S1: SnapGene Vector Maps Vector maps of (A) MCS-TOPO Gateway plasmid, (B) pGL3-Gateway-mpId3-eGFP Destination plasmid and (C) pGL2-Basic-mpId3-eGFP Negative control plasmid as exported from the SnapGene software (https://www.snapgene.com/)   167  Figure S2: Previously identified Col10a1 and Msx2 BMP-responsive regions  A) The UCSC Genome Browser (GRCg6a/galGal6) screenshot displays the conservation of the region upstream of the G.gallus COL10A1 gene locus. While the region upstream of the COL10A1 gene is relatively well conserved in other bird species, there is very little conservation between the chicken and vertebrate species.  B) This is the sequence of the 560bp BMP-responsive fragment upstream of the Msx2 gene; blue highlight indicates the 52bp region within 560bp fragment that was also shown to be BMP-responsive (Brugger et al. 2004). Yellow highlight indicates the region we cloned and used in our reporter experiment in PLM cells. Green highlight indicates the conserved BMP-SE we identified with our homer motif search.      A)       B) 560bp BMP-responsive fragment upstream of the Msx2 gene chr13:53,475,962-53,476,524 GATCTGCCTTATCCCTCCTTCCCACAGCGCACAGGTAAAAAAAGAAAATGAGCTCAGAGAGGTGCCATCTTTTGCCCGAAGTCACACAGCGAATGTCCACGGATTGGAGGGCAGTGGTGGAATTCCTGGCGGCCCTTGGCGCCCATTTGTCTGCCCGCTTCTGATACCCGGGTTCGGAGAATAGGCCTCTAACAAGCGGCCCATTAGAAGGAATTGTCACTCCTCCGGGAGTGAGGTTGTCCCATTAGGGCGAATTGTCATTCCTCCTGGAGCGAGGTTGTCCTGGCTCCGCGAAGGCTGAGTGCCGGGCCGAGAGCAATTAACGCGGCTCCGGCGCGGGCAGCCGCCTCTGCCCCGGGCAGCGGGGGCGGGGCGCCCGGCGCGGCTGGAGCCGGTCACCCGGCGCAGCCCCTTCCCCCGGAGCCCGCCTTTCATCTCCCCGCGCCTGGCGCCTACCCGCAGCCCTGGCCCGTCTACAGCCTTCTCTGCCCCTCCCCCTGCCCCTCCTAGCCTCTGGGATTCCGGCGATCCGTCCTCCGGGTTGCAAACGCCCAGGTCTGT        168                 -1697 -2383 UCSC Genome Browser GRCm38/mm10: chr9:64,022,728-64,023,411  -1911       GGCTCTCCCTCCTGGCGTCCTAGGCTCCTCTCAAGGCCAGGCGCTTTCTTCTCCCTTCGAGGATTTACCTTATGCCCCTTCGAGAGTGATAGAGGGGCGTGCTTTCCCCAGTTATGTTTCACTATAAAAACGGGGGCAAGTTGGGGACAGACTTGGGGACCCGGCATGGGAAGGGTTAAGGCCGCTCCTGCCCTGGAGCCGGCGCGGCGGGGTACCTCCTCGGGCGTGCTTGCTAGTCCGAGGGTGGCAAAAGTTCAAGCTTGTAAAGTCAGGATTGGCCCCGCGCGGAGGGGAAAAAAGGCCTGGTTCCCCCTCCCCCTCCCGGGCGTGGCGCCGATTGGCCCAGCCGCAGGGCCGGCCCCGCCCTCCCCTGGCGCGCCGCGGCCCCACCCCCGGCGCAGCACCACGCCCCTTTCCGGGCAGGAGCGGGGCCCCGAGCTCAATGAAAGAAATTCCGCTACTCTCATTGGCCTGGGGCTTGACGCCGCACCCCGCGGGAGGCGGAGTGGGGAGGATGGGGCGCGGACTTGTGATCCCCTTTAGGATACCTGTGCTCTGGACTGCTGTAGCGAAGGTGGTGTGCAAGTCCAGGAACGGGTCTGGAAGTGCGCTGGGGGAAGGGTCGTGTGTGTAGGAAAGATACAGACTACTTCCCTGAGTCTTACCTGGAGCCAGGTACTCCAGT Figure S3: Representation and sequence of Smad6 gene regulatory region  The illustration represents the upstream region of the Smad6 gene containing two previously identified BMP-responsive regulatory regions (marked and cited). The sequence between -1697 to -1911 was identified by Ishida et al. as a BMP-responsive regulatory region.  Bold indicates the 28bp sequence studied by Ishida et al. with multiple potential GC-rich Smad-binding sites. Bold underlined indicates the palindromic GC-rich binding site identified by mutation analysis to be the most important site for BMP-responsiveness in the identified sequence. Grey highlighted sequence represents a region identified via deletional analysis by Ishida et al. to have significant luciferase activity. (Specifically, it is the sequence between the D and E fragments studied in Figure 1B). This area was not further tested by Ishida et al. for a Smad-binding element. Bold, red and underlined sequence is a highly conserved putative BMP-CRE, predicted through our bioinformatic analyses (Table 3.10).   169                 B.  A.  Figure S4: Alcian Blue stain indicates hypertrophic development of PLM cells when treated with BMP4 on Day 4 post seeding.  A) BMP4 and NOG stimulate or block, respectively, cartilage formation in a dose-dependent manner. Alcian blue is used to stain acidic polysaccharides, which can be found in the extracellular matrix of mature chondrocytes. Here we verified the doses to use for our RNA-seq experiments. B) Cultures from each of the experiments used for RNA-seq (Exp1., Exp2. And Exp3.) and ChIP-seq (Exp4.) were stained on Day 4 post-seeding.  170            Figure S5: RNA-seq Gene Body Coverage Plot indicates uniform coverage profile along the gene body with no 5`/3` bias.  171     Figure S6: RNA-seq Sample Correlation map reveals good correlation between samples from the same time points and with the same treatment.  172  173          Figure S7: Heatmap of top 30 upregulated and top 30 downregulated genes from RNA-seq.  We selected the top 1500 genes from the RNA-seq list with the highest confidence score (median ranking from the 10 differential expression pipelines). We sorted them based on their |fold change| and selected the top 30 upregulated and top 30 downregulated genes with the highest change (|fold change|>3). We took the CPM counts from the Kallisto pipeline and normalized each sample with their respective control (Control 0h). For the upregulated top30, we ended up using only the top27 genes, as the following genes were eliminated from the list for having 0 CPM counts as per Kallisto: Gm7265, BC006965 and 2600014E12Rik. Using the ClustVis web tool we plotted these genes with a heatmap. No scaling was applied, and both rows and columns were clustered using correlation distance and complete linkage. Samples with the same treatment are clustered, but we can also observe samples from the same biological experiments clustering (slight bias). STRING Molecular Function GO terms were used to annotate all the genes in the heatmap. In the upregulated list, there are more genes with “extracellular matrix structural constituent” (8 genes) and “transcription factor/coregulator activity” (5 genes) annotations vs in the downregulated list, there are more genes with “signaling receptor binding” (6 genes) and “growth factor/cytokine activity” (5 genes) annotations.      174    Figure S8: Putative BMP-CREs. Screenshots of the USCS browser indicating putative BMP-CREs in red highlight for the genes Id1, Id3, Bambi and Smad7. Blue highlights indicate the previously reported BMP-responsive regions (as per description in Table3.10)  Id1 BMP-CREs Id3 BMP-CREs Bambi BMP-CREs Smad7 BMP-CREs 175                             Figure S9: PLM cell transfection with GFP reporter plasmids containing the Msx2 BMP-SE or Dlx2 BMP-AE reveals a BMP4-regulated GFP expression.  176     Figure S10: The unusual linker region and flanking sequence of Jdp2 BMP-AE resulted in the generation of an additional Jdp2 BMP-SE binding site with the generation of the Smad1/5/9 mutant.  We mutated the SMAD1/5/9 binding site to the usual ‘tgatga’ sequence; however, due to the unusually high GC-rich sequence in the linker site and the conserved ‘GTCT’ sequence right next to the Smad4 binding site, we generated a new “almost perfect” BMP-CRE binding site with a canonical n=5 linker. This could explain why did not see any reporter-expression loss with the mutant Jdp2-AEmut enhancer fragment.  Bold letters denote the bipartite BMP-CRE. Yellow highlight indicates the SMAD1/5/9 binding site, green highlight the Smad4 binding site and grey underlined the additional Smad4 binding site. Lowercase indicates the mutated base pairs. Red denotes non-conserved base pairs. Bold and underlined indicates the new BMP-CRE binding site we generated with the SMAD1/5/9 site mutation. 177  Table S1: Identified BMP-AE2 motifs in the Drosophila genome with an average PhastCons score > 0.55 * Red indicates enhancers tested and functionalized in vivo with reporter assays in transgenic flies   178   Table S2: Strict BMP4-upregulated DEG list                             2610035D17Rik Cacna1a Dlx3 Gm17315 Kctd4 Nkx3-2 Ptch1 Sort1 Tspoap12810025M15Rik Cacna1h Dlx5 Gm43196 Kif26b Nmu Ptch2 Sox10 Tuba4a4930402H24Rik Cadm3 Dlx6 Gm45837 Klf15 Nol4l Ptger2 Sox13 Tunar5730508B09Rik Cadps Dnajb11 Gm7265 Klf3 Notum Ptger4 Sox5 Txndc115930412G12Rik Calml3 Doc2a Gm7902 Klf4 Npc1 Pthlh Sox9 Uap1l1G630022F23Rik Calml4 Dock6 Gnb4 Klhl21 Npnt Rab11fip4 Sp7 Unc80AW551984 Camsap3 Dok4 Gng2 Krt8 Nr4a3 Rab28 Spata1 VdrAbtb2 Capn6 Dusp15 Gnptg Krtcap3 Nrarp Rab32 Sphk1 Vgll2Acan Ccdc157 Ebf4 Golga5 Lamb2 Nrxn1 Rab33b Spon1 Vps37bAcvr1 Ccndbp1 Efcab1 Golim4 Limd1 Ntng1 Rag1 Srgap1 Wfikkn2Adamts3 Cd9 Efna1 Gp1bb Litaf Nup210 Rep15 Srprb Wipi1Adamtsl2 Cdc42ep3 Efna3 Gramd1a Lpar4 Nupr1 Rftn2 Ssbp2 Yipf5Adcy9 Cdh11 Ehd3 Grem1 Lrrc1 Nxpe3 Rgcc Ssc4d Zc3h12dAdgra3 Cdh19 Emp2 Grhl1 Lrrc55 Optn Rhbdd1 Ssh3 Zdhhc12Adgrb3 Cdkl4 Enc1 Grhl2 Lrrc8b Otogl Rnf11 St3gal4 Zfp28Ahr Chadl Enpp2 Gria4 Lrrc8d Pappa2 Rnf144a St8sia4 Zfyve21Amt Chrdl1 Entpd1 Grid2 Lsr Papss1 Robo2 Stk32a Zfyve27Angel1 Chrdl2 Entpd7 Grid2ip Maf Pde10a Rorc Stra6lAnk2 Chrna7 Epb41l4b Grik3 Man2a1 Pde4a Rrbp1 Strip2Ankrd50 Chsy3 Eps8l2 Gsg1l Map1a Pde4dip Runx3 Stxbp1Anks1 Clcn5 Epyc Gulo Map6 Peli2 Rxfp1 Stxbp2Ap5z1 Cldn19 Ets2 Hapln3 Map7d2 Penk S100a1 Sumf1Apmap Clec14a Ezh1 Harbi1 Matn1 Pex11a Samd9l Surf4Arhgef1 Cmtm8 Fam19a5 Hcn1 Matn3 Phactr1 Scamp2 Sytl4Arhgef16 Col11a2 Fam20b Hdlbp Mbnl1 Phldb1 Sdf4 Tacc1Arhgef5 Col15a1 Fam63a Hey2 Mboat1 Phospho1 Sec14l2 Tcaf1Arl4a Col17a1 Farp1 Hip1r Mboat2 Pik3ip1 Sec23a Tcaf2Art3 Col20a1 Fat4 Hopx Mbp Pim3 Sec31a Tcp11l2Asap1 Col4a1 Fbxl20 Hoxa11os Mef2d Pir Sec61a1 Tex2Asap2 Comp Fermt2 Hoxc13 Meltf Plac8 Sec62 TgfaAspg Cpt1b Fgf15 Hs3st3a1 Mest Plagl1 Selenok ThrbAtf4 Cracr2a Fgfr4 Hyou1 Mfsd1 Plcl1 Sema3g Thsd4Atf6 Crip2 Fhdc1 Icos Mgat5 Pld2 Sema6d Tlr3Atp1b2 Ctdspl Ficd Ifitm10 Miat Plec Serpini1 Tm9sf2Atp6v0a2 Cux2 Foxf2 Igf2os Mical1 Plekhb1 Setdb2 Tmem151aB3galt5 Cyb5d2 Frs2 Igsf1 Micall1 Plekhm3 Sh3tc2 Tmem158B3gat1 Cyb5r1 Fryl Il16 Mid2 Plod2 Shc4 Tmem2BC006965 Cyp26b1 Fut9 Il6ra Mrgpre Plod3 Sipa1l1 Tmem246BC064078 Cyp2j9 Btbd3 Impdh1 Msx2 Plpp6 Slc25a13 Tmem266BC065397 Dcaf12l2 Gab1 Isl2 Mtss1l Plxnb1 Slc25a45 Tmem40Baiap2l1 Ddb1 Gadd45b Ism1 Mxra8 Pou3f3 Slc26a2 Tmem59Baiap2l2 Ddhd1 Gal3st1 Itga1 Myh14 Ppargc1a Slc35d1 Tmem64Bbs9 Dennd3 Gata4 Itga10 Myo1e Ppm1e Slc37a4 Tnfrsf11bBcam Depdc7 Gcnt1 Itpr2 Nacc2 Ppp1r16b Slc39a14 Tob2Bcl2 Dgke Gfpt1 Jph1 Ndrg4 Prex1 Slc39a7 Tox2Bmf Dhh Gfpt2 Kcnj12 Necab3 Prickle1 Slc45a4 Trank1Bmp5 Dhx35 Ginm1 Kcnj3 Nfe2l1 Prkag3 Slc4a2 Trim25Bsn Dhx57 Gja3 Kcnk5 Nim1k Prkcz Slc6a4 Trim46C030037D09Rik Dlc1 Gli1 Kcnmb4 Nin Proser2 Slc9a3r1 Trim62C130050O18Rik Dlx2 Gm10561 Kcns1 Nkain4 Psen1 Smox Tspan11BMP4-upregulated DEGs 179  Table S3: Strict BMP4-downregulated DEG list                            1190002N15Rik Car12 Dtx4 Gpr4 Micalcl Rab30 Tcf201700025A08Rik Ccna2 Dusp16 Gprc5a Mki67 Rad1 Tcf242610307P16Rik Ccnb1 Ebf2 Gpsm2 Mmaa Rad51ap1 Tead32810021J22Rik Cd38 Ebf3 Gsc Mpp3 Ran Tek2810408I11Rik Cd47 Edn1 Gsg2 Mpped1 Ranbp6 Tjp12810474O19Rik Cdc25c Efemp1 H2afy Mrps22 Rapgef5 Tlr26820408C15Rik Cdc7 Efna5 Has3 Msh6 Rarb Tmem164A730020E08Rik Cdca2 Efnb3 Hdac7 Mycn Rasef Tmem200bB130024G19Rik Cdca3 Egfem1 Hdgfrp3 Myo1b Rassf2 Tmem51C130021I20Rik Cdca7 Elmod1 Hic1 Ncapg Raver2 Tnfsf13bD030056L22Rik Cdca8 Emilin3 Hmgb1 Ncaph2 Ret Tpm4D930015E06Rik Cdh2 Eml4 Hpse2 Ncbp1 Rhobtb2 Tpx2AI314180 Celsr1 Emx2 Hunk Ncoa3 Rnd3 TrioAbca12 Cenpf Enox1 Id3 Nectin2 Rnf128 Ttc27Abcd2 Cenpn Epb41l2 Igfbp2 Nefl Rnf150 TtkAbhd6 Cenpv Epb41l3 Il1rap Neu3 Rpap3 TtpaAdm Cenpw Epha3 Ipo5 Nhs Sae1 Tuba1bAfap1 Cep19 Epha4 Irx3 Nlrp10 Scrt1 Tuba1cAfdn Chst2 Ephb2 Islr2 Nop16 Sema3f Tube1Ahsa1 Ckap2 Ermp1 Jak1 Notch4 Senp1 TwistnbAldh1a2 Ckap2l Esm1 Jazf1 Nrip1 11-Sep Txnrd1Amigo1 Cldn1 Etv4 Kalrn Ntf3 06-Sep Uba2Amigo2 Clec2l Etv6 Kars Nuf2 Sesn3 Usp1Ank3 Cntnap2 Exoc6 Kat6b Nxnl2 Sf3a3 Vav3Anln Col3a1 Ezr Kbtbd11 Nxpe2 Sfxn5 Vcam1Anp32a Colec12 Fam107b Kcnab1 Nxph4 Sgpl1 VclApcdd1 Cped1 Fam184b Kif11 Osr1 Sh2d3c Vgll3Apela Crabp1 Fdps Kif23 Otud7b Shb Vps13aArhgap10 Crabp2 Fgf16 Kif2c Pbk Shisa3 Wdr12Arhgap18 Crispld2 Fgf18 Kif4 Pcdh18 Shmt1 Wwc1Arhgap19 Csn3 Fhl3 Klhl14 Pcdh9 Six1 Ybx3Arhgap26 Ctsk Filip1l Kpna2 Pcsk5 Slc22a23 Zbtb44Arhgap28 Cttnbp2 Flrt1 Kpna3 Pdgfb Slc30a10 Zcchc12Arhgef39 Cxcl12 Flt4 L1td1 Phf6 Slco2a1 Zfp395Aspm Cxcl14 Fzd7 Lgi1 Pip4k2b Slco4a1 Zfp422Atp2b1 Cxcl5 Gatc Lgr6 Pla2g16 Smad1 Zfp512bAurka Dchs2 Gdf6 Lmnb1 Plk2 Snrnp40 Zfp608Avpr1a Dgkd Gkap1 Lpar1 Pls1 Sox12 Zfp618BC037039 Dhcr7 Gli3 Lsm3 Pmf1 Spata5 Zfp90BC055324 Dio3 Glis3 Magee2 Polr2m Spred2Bard1 Diras2 Gm1000 Magi1 Prkcb Stambpl1Blnk Dkk2 Gm16216 Man1a Prkce Stard10Btbd10 Dleu2 Gm21596 Map2k6 Prlr Stip1Bub1b Dlg1 Gm27010 Mcm10 Prnd StomC1s1 Dmd Gm42752 Mcoln3 Prokr1 Stox2C1s2 Dock5 Gm5620 Mcph1 Prom1 Stra6Cacna2d2 Dpf3 Gm5860 Megf11 Ptchd1 StrbpCacna2d3 Dph3 Gpi1 Mex3b Ptn Stx1aCadm1 Dsc2 Gpr161 Mfhas1 Ptprj Stxbp6Camk1d Dsc3 Gpr21 Mgst1 Pttg1 Tcerg1lBMP4-downregulated DEGs 180  Table S4: Strict NOG-upregulated DEG list              Table S5: Strict NOG-downregulated DEG list                            NOG-upregulated DEGs 9430020K01Rik Ecscr Nuak2Abca12 Eml4 Otud7bAdcyap1r1 Enox1 Oxnad1Anks1b Fjx1 Pcdhgc3Apba2 Fzd7 Pik3r3Apitd1 Gimap9 Prickle2Arhgap10 Gm10125 Ptchd4Arhgap20 Gm26507 Rab30Astn1 Gpi1 Rcc1lBegain Gpr4 Rrp1bBmp6 Gsc Rtkn2C130021I20Rik Hunk Ruvbl1Cacna1e Idh3b Sdk1Cacna2d3 Itga8 Sept6Cd1d1 Kalrn Slc9a9Cd38 Lgi1 Spred2Cdh2 Lgr6 StomCebpa Lpar1 Stox2Dennd5b Lrch1 Stra6Dicer1 Macc1 Suclg2Dock5 Magee2 Tacr1Dsc2 Mfhas1 Tpd52Dtnb Mpp2 TrioEbf2 Nek6 Tuba1cEbf3 Nfkbia ZyxF730043M19Rik NsdhlNOG-downregulated DEGs 1500009L16Rik Gal Lmna Sh3tc22310030G06Rik Gata2 Maf Shc4A730089K16Rik Gcnt1 Maoa Slc1a3Arhgef5 Gfpt2 Matn1 Slc9a2Arl4a Gli1 Mical1 Sod2B230354K17Rik Gm10561 Mmp15 Ssbp2Cacna1a Gm17315 Msantd1 Susd6Ccdc186 Gm20632 Msx2 TgfaCd9 Gm28309 Nebl Timp4Chrna7 Gm44250 Optn Tmem151aClec14a Gm7265 Penk Tmem246Col11a2 Gria4 Pim3 Tmem30aCol15a1 Grik3 Pitpnm1 Tmem64Crip1 Hapln3 Plscr4 TunarCyth1 Hcn1 Pnp Upf3bCyyr1 Hecw2 Pou3f3 Wrnip1Dlx3 Hey1 Prickle1 YwhahEbf4 Icos Ptch1 Zmpste24Eno2 Id4 Ptch2Enpp2 Itgb3 Ptger4Ets2 Jph1 Rasl11bFaah Kcnj13 RhofFam46a Kcnmb4 Rspo2Fgfr4 Kdsr Runx3Fhdc1 Klhl21 Serpini1181   # Gene Name RNAseq log2FoldChange chr peak start peak end peak name Number of motifs in peak Distance of peak to TSS 1 4930412O13Rik 2.76 2 9880396 9881889 H3K27ac_BMP12_vs_Noggin12_up.bed2255 2 0 2 Tmeff1 0.67 4 48584499 48585519 H3K27ac_BMP12_vs_Noggin12_up.bed2889 2 0 3 Chst3 0.64 10 60219167 60219440 H3K27ac_BMP12_vs_Noggin12_up.bed406 1 0 4 Ckap4 0.55 10 84533808 84534508 H3K27ac_BMP12_vs_Noggin12_up.bed477 1 0 5 Id4 1.3 13 48261212 48261820 H3K27ac_BMP12_vs_Noggin12_up.bed1132 1 0 6 Nin 0.77 12 70111612 70111952 H3K27ac_BMP12_vs_Noggin12_up.bed928 1 0 7 Ppp3ca 0.59 3 136669969 136670207 H3K27ac_BMP12_vs_Noggin12_up.bed2779 1 0 8 Smad9 0.92 3 54755187 54757017 H3K27ac_BMP12_vs_Noggin12_up.bed2631 1 0 9 Slc44a1 0.58 4 53440430 53440662 H3K27ac_BMP12_vs_Noggin12_up.bed2896 1 18 10 Slc2a13 1.1 15 91572853 91573231 H3K27ac_BMP12_vs_Noggin12_up.bed1558 1 30 11 Ctnnbip1 0.46 4 149518299 149518688 H3K27ac_BMP12_vs_Noggin12_up.bed3065 1 64 12 Grin3a 1.27 4 49845881 49846298 H3K27ac_BMP12_vs_Noggin12_up.bed2891 1 138 13 Ubxn4 0.34 1 128244175 128244460 H3K27ac_BMP12_vs_Noggin12_up.bed178 1 212 14 Rab2a 0.34 4 8535864 8536295 H3K27ac_BMP12_vs_Noggin12_up.bed2825 1 221 15 Klf4 1.54 4 55532736 55533483 H3K27ac_BMP12_vs_Noggin12_up.bed2902 1 271 16 Bambi 1.25 18 3507017 3507674 H3K27ac_BMP12_vs_Noggin12_up.bed1948 1 283 17 Jdp2 2.36 12 85599452 85600633 H3K27ac_BMP12_vs_Noggin12_up.bed964 1 348 18 P4hb 0.37 11 120572614 120572866 H3K27ac_BMP12_vs_Noggin12_up.bed831 1 387 19 Fgfr3 1.83 5 33722245 33722514 H3K27ac_BMP12_vs_Noggin12_up.bed3123 1 572 20 Cadps 2.25 14 12822014 12822468 H3K27ac_BMP12_vs_Noggin12_up.bed1288 1 611 21 Id4 1.3 13 48262081 48263250 H3K27ac_BMP12_vs_Noggin12_up.bed1133 1 655 22 Gpc6 0.76 14 116926044 116928027 H3K27ac_BMP12_vs_Noggin12_up.bed1447 1 748 23 Smad7 0.86 18 75368310 75369071 H3K27ac_BMP12_vs_Noggin12_up.bed2078 1 782 24 Nr3c1 1.18 18 39489949 39490470 H3K27ac_BMP12_vs_Noggin12_up.bed2012 1 831 25 4930412O13Rik 2.76 2 9882688 9882993 H3K27ac_BMP12_vs_Noggin12_up.bed2256 1 1437 26 Smad7 0.86 18 75369232 75374268 H3K27ac_BMP12_vs_Noggin12_up.bed2079 3 1704 27 Nog 4.61 11 89298858 89299427 H3K27ac_BMP12_vs_Noggin12_up.bed723 1 2905 28 Msx2 1.7 13 53476318 53477248 H3K27ac_BMP12_vs_Noggin12_up.bed1158 1 3245 29 Dlx2 3.77 2 71542550 71543147 H3K27ac_BMP12_vs_Noggin12_up.bed2375 1 3607 30 Cyp26b1 1.24 6 84587495 84588389 H3K27ac_BMP12_vs_Noggin12_up.bed3459 1 5519 31 Wwp2 2.21 8 107424775 107425000 H3K27ac_BMP12_vs_Noggin12_up.bed3958 1 11398 32 Copa 0.34 1 172063533 172064153 H3K27ac_BMP12_vs_Noggin12_up.bed247 1 18376 33 Met 0.68 6 17490863 17491153 H3K27ac_BMP12_vs_Noggin12_up.bed3361 1 27064 34 Stx5a 0.45 19 8713505 8714110 H3K27ac_BMP12_vs_Noggin12_up.bed2131 3 27303 35 Cnn1 0.99 9 22071009 22071379 H3K27ac_BMP12_vs_Noggin12_up.bed4028 1 27902 36 Dnajc12 0.94 10 63338426 63339013 H3K27ac_BMP12_vs_Noggin12_up.bed416 1 43430 37 Tmeff1 0.67 4 48540393 48540670 H3K27ac_BMP12_vs_Noggin12_up.bed2888 1 44504 38 Cdr2 0.44 7 121034783 121035175 H3K27ac_BMP12_vs_Noggin12_up.bed3763 2 52472 39 Dlx3 2.02 11 95182329 95182664 H3K27ac_BMP12_vs_Noggin12_up.bed753 1 62211 40 Spryd3 0.47 15 102203677 102203912 H3K27ac_BMP12_vs_Noggin12_up.bed1605 1 67444 41 Tmem59 0.37 4 107253207 107253420 H3K27ac_BMP12_vs_Noggin12_up.bed2973 1 74809 42 Dlx2 3.77 2 71628560 71628907 H3K27ac_BMP12_vs_Noggin12_up.bed2376 2 81807 43 Gal3st4 0.81 5 138187177 138187408 H3K27ac_BMP12_vs_Noggin12_up.bed3315 1 85432 44 Ypel2 1.16 11 87086483 87086717 H3K27ac_BMP12_vs_Noggin12_up.bed712 2 92777 45 Ptger4 1.23 15 5143892 5144120 H3K27ac_BMP12_vs_Noggin12_up.bed1466 3 100067 46 Atp1b2 1.55 11 69766018 69766261 H3K27ac_BMP12_vs_Noggin12_up.bed665 2 160077 47 Tpd52l1 1.28 10 31608849 31609160 H3K27ac_BMP12_vs_Noggin12_up.bed355 1 162929 48 Mitf 1.23 6 97981287 97981861 H3K27ac_BMP12_vs_Noggin12_up.bed3499 1 174236 49 Mpz 2.53 1 171329226 171329529 H3K27ac_BMP12_vs_Noggin12_up.bed244 1 178516 Table S6: BMP-CRE motifs within H3K27ac peaks (12hrs) up to 1Mb from upregulated DEGs                            182                                          50 Gm53 1.19 11 96064636 96065357 H3K27ac_BMP12_vs_Noggin12_up.bed758 1 185748 51 Fgfr2 0.99 7 132931238 132931513 H3K27ac_BMP12_vs_Noggin12_up.bed3798 1 191837 52 Nxph3 1 11 95712916 95713279 H3K27ac_BMP12_vs_Noggin12_up.bed755 1 198347 53 Frs2 0.86 10 117376573 117376960 H3K27ac_BMP12_vs_Noggin12_up.bed524 1 228100 54 Neat1 0.72 19 5609785 5609986 H3K27ac_BMP12_vs_Noggin12_up.bed2111 1 235492 55 Depdc7 1.12 2 104494040 104494766 H3K27ac_BMP12_vs_Noggin12_up.bed2427 2 248112 56 Rab33b 0.67 3 51225023 51225503 H3K27ac_BMP12_vs_Noggin12_up.bed2618 1 258417 57 Vps37b 0.62 5 123748985 123749376 H3K27ac_BMP12_vs_Noggin12_up.bed3291 1 282894 58 Col9a3 2.36 2 180273187 180273406 H3K27ac_BMP12_vs_Noggin12_up.bed2564 2 324384 59 Kdelr2 0.41 5 143731888 143732173 H3K27ac_BMP12_vs_Noggin12_up.bed3321 1 328051 60 Sost 4 11 102296039 102296777 H3K27ac_BMP12_vs_Noggin12_up.bed771 1 329025 61 Slc7a3 0.81 X 101420021 101420329 H3K27ac_BMP12_vs_Noggin12_up.bed4276 1 334002 62 Lrrc75b 1.29 10 75212112 75212361 H3K27ac_BMP12_vs_Noggin12_up.bed442 1 347969 63 B3gnt8 1.73 7 25248113 25248410 H3K27ac_BMP12_vs_Noggin12_up.bed3621 1 379214 64 Adam17 0.29 12 21783962 21784165 H3K27ac_BMP12_vs_Noggin12_up.bed867 1 410331 65 Plekhm3 0.66 1 64532965 64533242 H3K27ac_BMP12_vs_Noggin12_up.bed107 1 423582 66 Col9a3 2.36 2 180171689 180171906 H3K27ac_BMP12_vs_Noggin12_up.bed2563 1 425884 67 Ccdc71l 0.71 12 31949613 31950179 H3K27ac_BMP12_vs_Noggin12_up.bed889 1 428525 68 Gm20633 1.39 3 96697077 96697386 H3K27ac_BMP12_vs_Noggin12_up.bed2704 1 450698 69 Pcyt1b 0.81 X 94122087 94122544 H3K27ac_BMP12_vs_Noggin12_up.bed4267 1 467225 70 Vstm2l 3.03 2 157424298 157424576 H3K27ac_BMP12_vs_Noggin12_up.bed2522 1 490077 71 Cux2 1.55 5 121545719 121545932 H3K27ac_BMP12_vs_Noggin12_up.bed3286 1 504170 72 Rag1 2.3 2 102185754 102186315 H3K27ac_BMP12_vs_Noggin12_up.bed2424 2 536254 73 Wwp2 2.21 8 106893004 106893582 H3K27ac_BMP12_vs_Noggin12_up.bed3954 1 542816 74 Gdf10 1.35 14 34502997 34503909 H3K27ac_BMP12_vs_Noggin12_up.bed1339 1 579411 75 Npnt 1.18 3 133544432 133544671 H3K27ac_BMP12_vs_Noggin12_up.bed2771 1 594142 76 Slc6a4 2.37 11 77607205 77607408 H3K27ac_BMP12_vs_Noggin12_up.bed686 1 608603 77 Rhou 0.46 8 123041909 123042238 H3K27ac_BMP12_vs_Noggin12_up.bed3988 1 611691 78 Grik3 2.03 4 126102828 126103690 H3K27ac_BMP12_vs_Noggin12_up.bed3011 1 612129 79 Ankrd6 0.64 4 32304661 32305401 H3K27ac_BMP12_vs_Noggin12_up.bed2860 1 645440 80 Zbtb8a 0.46 4 128727497 128727802 H3K27ac_BMP12_vs_Noggin12_up.bed3015 1 650314 81 Dock9 0.56 14 122452544 122454499 H3K27ac_BMP12_vs_Noggin12_up.bed1462 1 654811 82 Sox6 1.17 7 115381543 115382113 H3K27ac_BMP12_vs_Noggin12_up.bed3745 1 656631 83 Adam17 0.29 12 22030484 22030839 H3K27ac_BMP12_vs_Noggin12_up.bed868 1 656853 84 Trp53inp2 0.42 2 156065298 156065575 H3K27ac_BMP12_vs_Noggin12_up.bed2518 1 684240 85 Aspg 1.44 12 111377450 111377681 H3K27ac_BMP12_vs_Noggin12_up.bed1001 1 729002 86 Ncmap 1.65 4 136143413 136143784 H3K27ac_BMP12_vs_Noggin12_up.bed3031 1 745187 87 Ncmap 1.65 4 136144387 136144703 H3K27ac_BMP12_vs_Noggin12_up.bed3033 1 746161 88 Ptx3 0.51 3 66985694 66985918 H3K27ac_BMP12_vs_Noggin12_up.bed2659 1 765785 89 Zeb2 0.52 2 44270901 44272354 H3K27ac_BMP12_vs_Noggin12_up.bed2319 1 845041 90 Stard8 1.1 X 99853381 99853685 H3K27ac_BMP12_vs_Noggin12_up.bed4274 1 850134 91 Serinc5 0.83 13 93499462 93499794 H3K27ac_BMP12_vs_Noggin12_up.bed1223 1 888325 92 Ncmap 1.65 4 136310437 136310724 H3K27ac_BMP12_vs_Noggin12_up.bed3037 1 912211 93 Wnk4 1.3 11 100319997 100320334 H3K27ac_BMP12_vs_Noggin12_up.bed764 1 940233 94 Ankrd6 0.64 4 31963980 31964225 H3K27ac_BMP12_vs_Noggin12_up.bed2855 2 986616 183                               # Gene Name RNAseq log2FoldChange chr peak start peak end peak name Number of motifs in peak Distance of peak to TSS 1 Vcl -0.53 14 20929391 20929971 H3K27ac_BMP12_vs_Noggin12_down.bed751 1 0 2 Gas1 -1.7 13 60175594 60177184 H3K27ac_BMP12_vs_Noggin12_down.bed677 1 181 3 Hmga2 -0.75 10 120475013 120476237 H3K27ac_BMP12_vs_Noggin12_down.bed248 1 232 4 Bmp2 -3.29 2 133552460 133552770 H3K27ac_BMP12_vs_Noggin12_down.bed1415 1 302 5 Lmx1b -1.04 2 33639416 33640027 H3K27ac_BMP12_vs_Noggin12_down.bed1328 1 484 6 1190002N15Rik -0.8 9 94537217 94537552 H3K27ac_BMP12_vs_Noggin12_down.bed2555 1 529 7 Sema3f -0.9 9 107709371 107709899 H3K27ac_BMP12_vs_Noggin12_down.bed2573 2 576 8 Pard3 -0.47 8 127064650 127064943 H3K27ac_BMP12_vs_Noggin12_down.bed2419 1 758 9 Bcl11a -1.13 11 24075843 24076597 H3K27ac_BMP12_vs_Noggin12_down.bed279 1 1459 10 Zfp608 -0.92 18 54987767 54988489 H3K27ac_BMP12_vs_Noggin12_down.bed1163 1 1691 11 1700021F13Rik -1.56 5 119797779 119798342 H3K27ac_BMP12_vs_Noggin12_down.bed1902 1 9527 12 Tbx15 -0.71 3 99254037 99256813 H3K27ac_BMP12_vs_Noggin12_down.bed1573 1 13657 13 Dnph1 -0.8 17 46555320 46555761 H3K27ac_BMP12_vs_Noggin12_down.bed1077 1 58532 14 Podnl1 -1.29 8 84197466 84198547 H3K27ac_BMP12_vs_Noggin12_down.bed2351 1 71478 15 Spry1 -1.41 3 37723995 37724554 H3K27ac_BMP12_vs_Noggin12_down.bed1496 1 84049 16 Mthfd1 -0.42 12 76370519 76370758 H3K27ac_BMP12_vs_Noggin12_down.bed544 3 115288 17 Nek6 -0.47 2 38348135 38348343 H3K27ac_BMP12_vs_Noggin12_down.bed1337 1 163300 18 Gpsm2 -0.5 3 108911018 108911356 H3K27ac_BMP12_vs_Noggin12_down.bed1588 1 188710 19 Gas1 -1.7 13 59971788 59972033 H3K27ac_BMP12_vs_Noggin12_down.bed673 1 205332 20 Myo16 -0.83 8 9977819 9978173 H3K27ac_BMP12_vs_Noggin12_down.bed2271 1 294399 21 H2afy -0.64 13 55831695 55832003 H3K27ac_BMP12_vs_Noggin12_down.bed660 1 304358 22 Podnl1 -1.29 8 84662240 84662820 H3K27ac_BMP12_vs_Noggin12_down.bed2357 1 536252 23 H2afy -0.64 13 55512458 55512774 H3K27ac_BMP12_vs_Noggin12_down.bed655 1 623587 24 Gemin5 -0.45 11 58949180 58949384 H3K27ac_BMP12_vs_Noggin12_down.bed331 1 780642 25 Podnl1 -1.29 8 84976055 84976427 H3K27ac_BMP12_vs_Noggin12_down.bed2361 1 850067 26 Esrrb -1.64 12 85485347 85486063 H3K27ac_BMP12_vs_Noggin12_down.bed567 2 875054 Table S7: BMP-CRE motifs within H3K27ac peaks (12hrs) up to 1Mb from downregulated DEGs 184                               Table S8: BMP-CRE motifs within H3K27ac peaks (24hrs) up to 1Mb from upregulated DEGs  # Gene Name RNAseq log2FoldChange chr peak start peak end peak name Number of motifs in peak Distance of peak to TSS 1 4930412O13Rik 2.38 2 9880669 9881876 H3K27ac_BMP24_vs_Noggin24_up.bed2661 2 0 2 G630022F23Rik 1.96 5 33722934 33723580 H3K27ac_BMP24_vs_Noggin24_up.bed3775 1 89 3 Smad9 0.52 3 54755685 54756307 H3K27ac_BMP24_vs_Noggin24_up.bed3183 1 104 4 Jdp2 2.26 12 85599372 85599799 H3K27ac_BMP24_vs_Noggin24_up.bed1125 1 268 5 Ldlrad3 0.6 2 102185753 102186099 H3K27ac_BMP24_vs_Noggin24_up.bed2877 2 286 6 Dlx2 4.3 2 71544092 71546357 H3K27ac_BMP24_vs_Noggin24_up.bed2805 1 397 7 Dlx6 2.18 6 6863842 6864712 H3K27ac_BMP24_vs_Noggin24_up.bed4066 1 509 8 Slc6a8 0.48 X 73674575 73674837 H3K27ac_BMP24_vs_Noggin24_up.bed5146 1 1426 9 Dlx2 4.3 2 71543195 71543990 H3K27ac_BMP24_vs_Noggin24_up.bed2804 1 2764 10 Nog 4.28 11 89298845 89299398 H3K27ac_BMP24_vs_Noggin24_up.bed809 1 2934 11 Msx2 1.52 13 53476420 53477204 H3K27ac_BMP24_vs_Noggin24_up.bed1317 1 3347 12 Frs2 1.07 10 117152251 117152509 H3K27ac_BMP24_vs_Noggin24_up.bed581 1 3778 13 Foxf2 0.96 13 31619826 31620208 H3K27ac_BMP24_vs_Noggin24_up.bed1239 2 5608 14 Plec 0.86 15 76208629 76209049 H3K27ac_BMP24_vs_Noggin24_up.bed1783 1 23525 15 Atp1a2 1.53 1 172326591 172327178 H3K27ac_BMP24_vs_Noggin24_up.bed317 1 28528 16 Col11a1 2.65 3 113949656 113950022 H3K27ac_BMP24_vs_Noggin24_up.bed3316 1 80518 17 Dlx2 4.3 2 71628523 71629696 H3K27ac_BMP24_vs_Noggin24_up.bed2806 2 81770 18 Synpo 0.71 18 60748329 60748627 H3K27ac_BMP24_vs_Noggin24_up.bed2381 1 88188 19 Chd7 0.69 4 8535835 8536136 H3K27ac_BMP24_vs_Noggin24_up.bed3428 1 154270 20 Trim62 0.55 4 128727599 128727878 H3K27ac_BMP24_vs_Noggin24_up.bed3658 1 155702 21 Mitf 0.73 6 97980981 97981949 H3K27ac_BMP24_vs_Noggin24_up.bed4231 1 173930 22 Mmp14 0.38 14 54631406 54631717 H3K27ac_BMP24_vs_Noggin24_up.bed1598 1 199795 23 Myo10 0.3 15 25364339 25364620 H3K27ac_BMP24_vs_Noggin24_up.bed1722 1 257905 24 Abca1 0.7 4 53440439 53440661 H3K27ac_BMP24_vs_Noggin24_up.bed3511 1 280545 25 Arhgef1 0.97 7 25247748 25248423 H3K27ac_BMP24_vs_Noggin24_up.bed4350 1 344837 26 Tm9sf2 0.28 14 122452972 122454254 H3K27ac_BMP24_vs_Noggin24_up.bed1701 1 345935 27 Ptprn2 1.98 12 116047698 116047924 H3K27ac_BMP24_vs_Noggin24_up.bed1184 1 437796 28 Anxa8 2.15 14 34549059 34549282 H3K27ac_BMP24_vs_Noggin24_up.bed1562 1 463079 29 Fbxl20 0.28 11 97628072 97628305 H3K27ac_BMP24_vs_Noggin24_up.bed889 1 522098 30 Tmem200c 1.13 17 69383481 69383786 H3K27ac_BMP24_vs_Noggin24_up.bed2201 1 546346 31 Dbx2 2.46 15 96286910 96287500 H3K27ac_BMP24_vs_Noggin24_up.bed1852 1 630951 32 Ankrd6 0.76 4 32304641 32304878 H3K27ac_BMP24_vs_Noggin24_up.bed3469 1 645963 33 Sox6 1.66 7 115381582 115382106 H3K27ac_BMP24_vs_Noggin24_up.bed4526 1 656638 34 Fbln2 1.03 6 92091450 92091747 H3K27ac_BMP24_vs_Noggin24_up.bed4217 1 878996 35 Myo5b 0.52 18 75370742 75371486 H3K27ac_BMP24_vs_Noggin24_up.bed2423 1 929807 185   # Gene Name RNAseq log2FoldChange chr peak start peak end peak name Number of motifs in peak Distance of peak to TSS 1 Atp2b1 -0.78 10 98914990 98915193 H3K27ac_BMP24_vs_Noggin24_down.bed903 2 0 2 Rai14 -0.32 15 10714011 10714696 H3K27ac_BMP24_vs_Noggin24_down.bed2828 1 0 3 Wee1 -0.52 7 110122056 110122855 H3K27ac_BMP24_vs_Noggin24_down.bed7441 1 0 4 Mpped2 -1.11 2 106693072 106694078 H3K27ac_BMP24_vs_Noggin24_down.bed4721 1 0 5 Lmx1b -1.26 2 33639398 33640584 H3K27ac_BMP24_vs_Noggin24_down.bed4553 1 0 6 Tnfrsf21 -1.15 17 43016555 43016897 H3K27ac_BMP24_vs_Noggin24_down.bed3638 1 1 7 Vcl -0.77 14 20929468 20930308 H3K27ac_BMP24_vs_Noggin24_down.bed2527 1 36 8 Gas1 -2.03 13 60174484 60177318 H3K27ac_BMP24_vs_Noggin24_down.bed2317 1 47 9 Nectin2 -0.61 7 19748946 19749502 H3K27ac_BMP24_vs_Noggin24_down.bed7146 1 71 10 Id3 -0.66 4 136143102 136143404 H3K27ac_BMP24_vs_Noggin24_down.bed5954 1 93 11 Fam60a -0.43 6 148944066 148946370 H3K27ac_BMP24_vs_Noggin24_down.bed7071 1 97 12 Smad3 -0.48 9 63757619 63757888 H3K27ac_BMP24_vs_Noggin24_down.bed8278 1 106 13 Aes -0.86 10 81559670 81559956 H3K27ac_BMP24_vs_Noggin24_down.bed827 1 177 14 Gsc -2.37 12 104472798 104473152 H3K27ac_BMP24_vs_Noggin24_down.bed1946 1 178 15 Pdlim7 -0.8 13 55511691 55513491 H3K27ac_BMP24_vs_Noggin24_down.bed2257 1 185 16 Skida1 -0.58 2 18047643 18048840 H3K27ac_BMP24_vs_Noggin24_down.bed4429 1 211 17 Bnc2 -0.83 4 84674307 84675012 H3K27ac_BMP24_vs_Noggin24_down.bed5710 1 263 18 Uck2 -0.36 1 167283722 167285041 H3K27ac_BMP24_vs_Noggin24_down.bed478 1 279 19 Dusp7 -0.5 9 106368969 106369873 H3K27ac_BMP24_vs_Noggin24_down.bed8420 1 338 20 Bmp2 -2.75 2 133552503 133554390 H3K27ac_BMP24_vs_Noggin24_down.bed4804 1 345 21 1190002N15Rik -0.86 9 94536566 94537730 H3K27ac_BMP24_vs_Noggin24_down.bed8377 1 351 22 Id3 -0.66 4 136143873 136145119 H3K27ac_BMP24_vs_Noggin24_down.bed5955 1 377 23 Cacna1g -0.51 11 94473433 94473728 H3K27ac_BMP24_vs_Noggin24_down.bed1481 2 470 24 Sema3f -0.83 9 107709368 107709978 H3K27ac_BMP24_vs_Noggin24_down.bed8433 2 497 25 Tead3 -0.46 17 28349856 28350268 H3K27ac_BMP24_vs_Noggin24_down.bed3567 1 537 26 Sox12 -0.5 2 152396765 152397425 H3K27ac_BMP24_vs_Noggin24_down.bed4841 1 638 27 Zfp608 -0.77 18 54985966 54989404 H3K27ac_BMP24_vs_Noggin24_down.bed3998 1 776 28 Aebp2 -0.47 6 140623664 140624043 H3K27ac_BMP24_vs_Noggin24_down.bed7050 1 1002 29 Ubtf -0.4 11 102318198 102318739 H3K27ac_BMP24_vs_Noggin24_down.bed1552 1 1003 30 Mex3b -0.87 7 82868342 82870121 H3K27ac_BMP24_vs_Noggin24_down.bed7348 1 1010 31 Cul1 -0.33 6 47454488 47454755 H3K27ac_BMP24_vs_Noggin24_down.bed6714 1 1091 32 Bcl11a -1.11 11 24075736 24076326 H3K27ac_BMP24_vs_Noggin24_down.bed1122 1 1730 33 Ubtf -0.4 11 102317064 102317784 H3K27ac_BMP24_vs_Noggin24_down.bed1551 1 1958 34 Bcl2l11 -0.66 2 128128268 128128547 H3K27ac_BMP24_vs_Noggin24_down.bed4777 1 2231 35 Anp32a -0.26 9 62344123 62344555 H3K27ac_BMP24_vs_Noggin24_down.bed8267 2 2831 36 Id3 -0.66 4 136148001 136148822 H3K27ac_BMP24_vs_Noggin24_down.bed5956 1 4505 37 Sae1 -0.26 7 16400332 16401407 H3K27ac_BMP24_vs_Noggin24_down.bed7122 2 12536 38 Tpx2 -0.44 2 152830229 152831343 H3K27ac_BMP24_vs_Noggin24_down.bed4857 1 16621 39 Notch4 -1.18 17 34586098 34587456 H3K27ac_BMP24_vs_Noggin24_down.bed3613 1 21831 40 Ubtf -0.4 11 102297306 102297558 H3K27ac_BMP24_vs_Noggin24_down.bed1549 1 22184 41 Nxnl2 -2.11 13 51202622 51203025 H3K27ac_BMP24_vs_Noggin24_down.bed2202 1 31598 42 Tcf15 -1.43 2 152105499 152105735 H3K27ac_BMP24_vs_Noggin24_down.bed4840 1 37826 43 Crabp2 -1.51 3 87905949 87906164 H3K27ac_BMP24_vs_Noggin24_down.bed5248 1 42502 44 Six1 -0.44 12 73112350 73113149 H3K27ac_BMP24_vs_Noggin24_down.bed1819 1 58464 45 Fblim1 -0.46 4 141537877 141538381 H3K27ac_BMP24_vs_Noggin24_down.bed5979 2 67715 46 Aifm2 -0.64 10 61783521 61784393 H3K27ac_BMP24_vs_Noggin24_down.bed723 1 68259 47 Cdon -0.48 9 35342177 35342542 H3K27ac_BMP24_vs_Noggin24_down.bed8129 1 78586 48 2810021J22Rik -0.37 11 58948996 58949293 H3K27ac_BMP24_vs_Noggin24_down.bed1239 1 81781 49 Ccnf -0.36 17 24168980 24169421 H3K27ac_BMP24_vs_Noggin24_down.bed3518 1 81988                            Table S9: BMP-CRE motifs within H3K27ac peaks (24hrs) up to 1Mb from downregulated DEGs 186    50 Ddx47 -0.24 6 134926890 134928501 H3K27ac_BMP24_vs_Noggin24_down.bed7024 1 83111 51 Npm3 -0.47 19 45659773 45660227 H3K27ac_BMP24_vs_Noggin24_down.bed4329 1 89364 52 Ets1 -0.48 9 32540202 32543034 H3K27ac_BMP24_vs_Noggin24_down.bed8114 1 93187 53 Ttc28 -0.46 5 110978711 110979545 H3K27ac_BMP24_vs_Noggin24_down.bed6325 1 98909 54 Eln -2.18 5 134638830 134639277 H3K27ac_BMP24_vs_Noggin24_down.bed6461 1 108046 55 Efnb3 -0.68 11 69671370 69672225 H3K27ac_BMP24_vs_Noggin24_down.bed1320 1 111166 56 Sept11 -0.43 5 93205755 93206250 H3K27ac_BMP24_vs_Noggin24_down.bed6281 1 112319 57 Otud7b -0.57 3 96219874 96220732 H3K27ac_BMP24_vs_Noggin24_down.bed5322 2 115348 58 Barx1 -1.88 13 48546145 48546454 H3K27ac_BMP24_vs_Noggin24_down.bed2188 2 116544 59 Otud7b -0.57 3 96221128 96221697 H3K27ac_BMP24_vs_Noggin24_down.bed5323 1 116602 60 Gm13781 -1.29 6 30692928 30693749 H3K27ac_BMP24_vs_Noggin24_down.bed6655 1 116894 61 Birc5 -0.44 11 117968714 117968946 H3K27ac_BMP24_vs_Noggin24_down.bed1650 1 119464 62 Glrb -0.52 3 81036669 81038445 H3K27ac_BMP24_vs_Noggin24_down.bed5204 1 123010 63 Cdkn3 -0.62 14 46884142 46884488 H3K27ac_BMP24_vs_Noggin24_down.bed2621 1 123602 64 Hmgb1 -0.41 5 149052387 149052957 H3K27ac_BMP24_vs_Noggin24_down.bed6549 1 131532 65 Ikzf3 -3.13 11 98682708 98683039 H3K27ac_BMP24_vs_Noggin24_down.bed1509 2 136678 66 Tcirg1 -0.45 19 3768139 3768471 H3K27ac_BMP24_vs_Noggin24_down.bed4171 1 138662 67 Larp1b -1.27 3 40800005 40800647 H3K27ac_BMP24_vs_Noggin24_down.bed5088 1 149707 68 Ecscr -1.67 18 35562386 35562925 H3K27ac_BMP24_vs_Noggin24_down.bed3935 1 159431 69 Tnnt3 -0.67 7 142659110 142659388 H3K27ac_BMP24_vs_Noggin24_down.bed7615 1 160275 70 Stx1a -0.85 5 135187335 135188048 H3K27ac_BMP24_vs_Noggin24_down.bed6465 1 163854 71 Ecscr -1.67 18 35889148 35889832 H3K27ac_BMP24_vs_Noggin24_down.bed3943 1 166793 72 Cdca3 -0.37 6 125009255 125009514 H3K27ac_BMP24_vs_Noggin24_down.bed6989 1 179709 73 Zbtb44 -0.94 9 31211443 31211751 H3K27ac_BMP24_vs_Noggin24_down.bed8109 1 180800 74 Gypc -0.67 18 32377285 32377573 H3K27ac_BMP24_vs_Noggin24_down.bed3919 1 182461 75 Hmgb2 -0.44 8 57328381 57328714 H3K27ac_BMP24_vs_Noggin24_down.bed7776 1 183129 76 Smyd4 -0.43 11 75531688 75531901 H3K27ac_BMP24_vs_Noggin24_down.bed1368 1 183256 77 Esr2 -1.61 12 76370197 76371001 H3K27ac_BMP24_vs_Noggin24_down.bed1843 3 192939 78 Nfib -0.96 4 82503769 82505240 H3K27ac_BMP24_vs_Noggin24_down.bed5685 1 200510 79 Cdc25c -0.4 18 34953908 34954322 H3K27ac_BMP24_vs_Noggin24_down.bed3932 1 202376 80 Acy3 -1.08 19 4192247 4192487 H3K27ac_BMP24_vs_Noggin24_down.bed4174 1 205587 81 Saal1 -0.31 7 46919696 46920023 H3K27ac_BMP24_vs_Noggin24_down.bed7242 1 209017 82 Npr1 -1.79 3 90248199 90248836 H3K27ac_BMP24_vs_Noggin24_down.bed5291 1 217030 83 Cep85l -0.89 10 53596919 53597967 H3K27ac_BMP24_vs_Noggin24_down.bed685 2 217069 84 Prkcb -1.22 7 122067489 122067972 H3K27ac_BMP24_vs_Noggin24_down.bed7477 2 220779 85 Zkscan4 -0.96 13 21715022 21716135 H3K27ac_BMP24_vs_Noggin24_down.bed2027 1 236116 86 Fzd8 -0.71 18 9449689 9450004 H3K27ac_BMP24_vs_Noggin24_down.bed3854 1 236834 87 Itgb6 -1.88 2 60962755 60963220 H3K27ac_BMP24_vs_Noggin24_down.bed4608 1 240113 88 Zkscan4 -0.96 13 21722283 21722929 H3K27ac_BMP24_vs_Noggin24_down.bed2031 1 243377 89 Rrp1b -0.33 17 32284197 32284687 H3K27ac_BMP24_vs_Noggin24_down.bed3602 2 248098 90 Zfp618 -0.56 4 63215417 63215728 H3K27ac_BMP24_vs_Noggin24_down.bed5653 1 249845 91 Ctps -0.55 4 120825315 120825536 H3K27ac_BMP24_vs_Noggin24_down.bed5860 1 255040 92 Mtss1 -0.92 15 58823038 58823471 H3K27ac_BMP24_vs_Noggin24_down.bed2913 2 258555 93 Prokr1 -1.25 6 87849995 87850967 H3K27ac_BMP24_vs_Noggin24_down.bed6864 1 259253 94 Smad3 -0.48 9 64020969 64021997 H3K27ac_BMP24_vs_Noggin24_down.bed8291 2 262976 95 Plekha2 -1.02 8 25368342 25368835 H3K27ac_BMP24_vs_Noggin24_down.bed7698 1 266149 96 Tns1 -0.41 1 74392110 74392576 H3K27ac_BMP24_vs_Noggin24_down.bed184 1 267662 97 Slco2a1 -1.45 9 102717155 102718366 H3K27ac_BMP24_vs_Noggin24_down.bed8407 1 270346 98 Gm13781 -1.29 6 30304138 30304445 H3K27ac_BMP24_vs_Noggin24_down.bed6636 1 271590 99 Begain -0.87 12 108793728 108794121 H3K27ac_BMP24_vs_Noggin24_down.bed1969 1 274096 100 Nsmce2 -0.38 15 59648461 59649920 H3K27ac_BMP24_vs_Noggin24_down.bed2918 1 274264 187    101 Zkscan4 -0.96 13 21754158 21754795 H3K27ac_BMP24_vs_Noggin24_down.bed2039 1 275252 102 Pmf1 -0.47 3 88685796 88686363 H3K27ac_BMP24_vs_Noggin24_down.bed5270 1 275466 103 H2afy -0.71 13 55837212 55837457 H3K27ac_BMP24_vs_Noggin24_down.bed2269 1 298904 104 Megf11 -1.62 9 64084810 64085162 H3K27ac_BMP24_vs_Noggin24_down.bed8293 1 300464 105 Pdlim7 -0.8 13 55212697 55213089 H3K27ac_BMP24_vs_Noggin24_down.bed2250 1 300587 106 Snrnp40 -0.26 4 130663627 130664440 H3K27ac_BMP24_vs_Noggin24_down.bed5904 1 303496 107 H2afy -0.71 13 55831656 55831866 H3K27ac_BMP24_vs_Noggin24_down.bed2265 1 304495 108 Zkscan4 -0.96 13 21787604 21788271 H3K27ac_BMP24_vs_Noggin24_down.bed2047 1 308698 109 H2afy -0.71 13 55826222 55826464 H3K27ac_BMP24_vs_Noggin24_down.bed2262 1 309897 110 Ncbp1 -0.27 4 46451124 46451644 H3K27ac_BMP24_vs_Noggin24_down.bed5605 3 312512 111 Col26a1 -0.42 5 136566207 136566483 H3K27ac_BMP24_vs_Noggin24_down.bed6472 1 316726 112 Spag5 -0.4 11 77983488 77984505 H3K27ac_BMP24_vs_Noggin24_down.bed1394 1 317024 113 Zkscan4 -0.96 13 21809353 21810123 H3K27ac_BMP24_vs_Noggin24_down.bed2048 1 330447 114 Hic1 -5.64 11 74837918 74838337 H3K27ac_BMP24_vs_Noggin24_down.bed1348 1 331182 115 Hic1 -5.64 11 74837360 74837765 H3K27ac_BMP24_vs_Noggin24_down.bed1347 2 331754 116 0610007P14Rik -0.33 12 85485125 85486032 H3K27ac_BMP24_vs_Noggin24_down.bed1906 2 338518 117 Zkscan4 -0.96 13 21833767 21834832 H3K27ac_BMP24_vs_Noggin24_down.bed2054 1 354861 118 Cbln1 -2.42 8 87834225 87834451 H3K27ac_BMP24_vs_Noggin24_down.bed7900 1 361634 119 Wnt5b -1.6 6 119175343 119175659 H3K27ac_BMP24_vs_Noggin24_down.bed6965 1 368688 120 Fndc5 -0.45 4 129513789 129514013 H3K27ac_BMP24_vs_Noggin24_down.bed5893 2 376791 121 Iah1 -0.43 12 20920644 20921199 H3K27ac_BMP24_vs_Noggin24_down.bed1737 1 395193 122 Xrcc6 -0.4 15 81585603 81586024 H3K27ac_BMP24_vs_Noggin24_down.bed3020 1 401811 123 Sult1a1 -1.2 7 126272967 126273194 H3K27ac_BMP24_vs_Noggin24_down.bed7490 1 403238 124 Fmo2 -3.62 1 163314691 163314900 H3K27ac_BMP24_vs_Noggin24_down.bed466 1 415966 125 Gpi1 -0.51 7 34653672 34654022 H3K27ac_BMP24_vs_Noggin24_down.bed7208 1 423384 126 Gpi1 -0.51 7 34654505 34655670 H3K27ac_BMP24_vs_Noggin24_down.bed7209 2 424217 127 Dync2li1 -0.54 17 84184101 84185840 H3K27ac_BMP24_vs_Noggin24_down.bed3805 1 440656 128 Kctd1 -0.49 18 14682809 14683246 H3K27ac_BMP24_vs_Noggin24_down.bed3887 1 468200 129 Spon2 -2.94 5 33692376 33692667 H3K27ac_BMP24_vs_Noggin24_down.bed6146 1 473922 130 Stmn1 -0.41 4 133967557 133967976 H3K27ac_BMP24_vs_Noggin24_down.bed5929 1 500344 131 Pced1b -0.77 15 96709117 96709702 H3K27ac_BMP24_vs_Noggin24_down.bed3060 2 537405 132 Clec2l -1.3 6 39206174 39206529 H3K27ac_BMP24_vs_Noggin24_down.bed6701 1 543106 133 Kcnab1 -1.35 3 65665826 65666123 H3K27ac_BMP24_vs_Noggin24_down.bed5170 1 556443 134 Zkscan4 -0.96 13 22035905 22037038 H3K27ac_BMP24_vs_Noggin24_down.bed2056 1 556999 135 Scube3 -0.47 17 27556765 27557124 H3K27ac_BMP24_vs_Noggin24_down.bed3557 1 585192 136 Fam78b -0.97 1 166409525 166409798 H3K27ac_BMP24_vs_Noggin24_down.bed477 1 591619 137 Ttc28 -0.46 5 110269409 110269820 H3K27ac_BMP24_vs_Noggin24_down.bed6321 1 609983 138 Stox2 -0.59 8 46740537 46741204 H3K27ac_BMP24_vs_Noggin24_down.bed7746 2 611144 139 Slco4a1 -1.26 2 179842765 179842984 H3K27ac_BMP24_vs_Noggin24_down.bed5005 1 613261 140 Efna5 -0.71 17 63499782 63500137 H3K27ac_BMP24_vs_Noggin24_down.bed3707 1 618466 141 Acy3 -1.08 19 4615085 4615486 H3K27ac_BMP24_vs_Noggin24_down.bed4179 1 628425 142 Myl4 -0.79 11 105181968 105182565 H3K27ac_BMP24_vs_Noggin24_down.bed1575 1 631306 143 Notch4 -1.18 17 33889859 33890595 H3K27ac_BMP24_vs_Noggin24_down.bed3608 1 673673 144 Stip1 -0.27 19 6363927 6364539 H3K27ac_BMP24_vs_Noggin24_down.bed4200 1 675487 145 Myf5 -0.69 10 108161644 108161987 H3K27ac_BMP24_vs_Noggin24_down.bed916 1 675511 146 Tcf7 -0.69 11 51606030 51606908 H3K27ac_BMP24_vs_Noggin24_down.bed1209 1 676106 147 Flt4 -1.83 11 50292056 50292261 H3K27ac_BMP24_vs_Noggin24_down.bed1204 1 682794 148 Gatc -0.46 5 116024062 116024409 H3K27ac_BMP24_vs_Noggin24_down.bed6357 1 682885 149 Nuf2 -0.36 1 170214957 170215432 H3K27ac_BMP24_vs_Noggin24_down.bed493 1 683494 150 Oxnad1 -0.45 14 31385867 31386079 H3K27ac_BMP24_vs_Noggin24_down.bed2596 1 699295 151 Rpia -0.64 6 71493941 71494696 H3K27ac_BMP24_vs_Noggin24_down.bed6811 1 701710 152 Ankle1 -0.59 8 70698978 70700277 H3K27ac_BMP24_vs_Noggin24_down.bed7820 1 705733 188       153 Stmn1 -0.41 4 133752722 133753314 H3K27ac_BMP24_vs_Noggin24_down.bed5925 1 715006 154 Etv4 -1.26 11 101063567 101064184 H3K27ac_BMP24_vs_Noggin24_down.bed1540 1 721187 155 Pole4 -0.33 6 83456506 83457010 H3K27ac_BMP24_vs_Noggin24_down.bed6846 1 751142 156 Bcl11a -1.11 11 23307054 23307344 H3K27ac_BMP24_vs_Noggin24_down.bed1120 2 770712 157 Ube2c -0.35 2 163995177 163995716 H3K27ac_BMP24_vs_Noggin24_down.bed4915 1 774182 158 Nfib -0.96 4 83486085 83486328 H3K27ac_BMP24_vs_Noggin24_down.bed5697 1 780336 159 Rtp4 -1.22 16 24393509 24394349 H3K27ac_BMP24_vs_Noggin24_down.bed3247 1 783591 160 Ccna2 -0.39 3 35754122 35754497 H3K27ac_BMP24_vs_Noggin24_down.bed5070 2 817653 161 Npm3 -0.47 19 44931166 44931521 H3K27ac_BMP24_vs_Noggin24_down.bed4321 2 818070 162 Tnfaip6 -1.48 2 52857771 52858617 H3K27ac_BMP24_vs_Noggin24_down.bed4588 3 819763 163 0610007P14Rik -0.33 12 84996703 84996908 H3K27ac_BMP24_vs_Noggin24_down.bed1898 1 827642 164 Fam65b -1.05 13 23746762 23748051 H3K27ac_BMP24_vs_Noggin24_down.bed2101 1 834138 165 Snx9 -0.32 17 4995209 4995560 H3K27ac_BMP24_vs_Noggin24_down.bed3464 1 845768 166 Dgkd -0.24 1 88701734 88702032 H3K27ac_BMP24_vs_Noggin24_down.bed227 1 848448 167 Jazf1 -0.75 6 52203935 52204159 H3K27ac_BMP24_vs_Noggin24_down.bed6735 1 864472 168 Notch4 -1.18 17 33684827 33685067 H3K27ac_BMP24_vs_Noggin24_down.bed3605 1 879201 169 Rtkn2 -0.76 10 67096532 67096832 H3K27ac_BMP24_vs_Noggin24_down.bed738 1 882738 170 Irx3 -1.38 8 90907940 90908188 H3K27ac_BMP24_vs_Noggin24_down.bed7917 1 893728 171 Fam65b -1.05 13 23684267 23685765 H3K27ac_BMP24_vs_Noggin24_down.bed2096 1 896424 172 Slc14a2 -1.05 18 79109456 79109838 H3K27ac_BMP24_vs_Noggin24_down.bed4138 1 900363 173 Btbd11 -1.53 10 86295626 86296287 H3K27ac_BMP24_vs_Noggin24_down.bed858 1 908813 174 Fam65b -1.05 13 23621479 23622502 H3K27ac_BMP24_vs_Noggin24_down.bed2094 1 959687 175 Vgll3 -1.12 16 64852578 64852874 H3K27ac_BMP24_vs_Noggin24_down.bed3375 1 962759 176 Pdgfb -2.26 15 79028347 79028895 H3K27ac_BMP24_vs_Noggin24_down.bed3003 2 985913 177 Fam65b -1.05 13 23585286 23585895 H3K27ac_BMP24_vs_Noggin24_down.bed2090 1 996294 189    Table S10: Differentially regulated RNA-seq genes with endothelial-related GO term GO:0010594 regulation of endothelial cell migrationGO:0001936 regulation of endothelial cell proliferationGO:0003158 endothelium developmentGO:0001938 positive regulation of endothelial cell proliferationGO:0010595 positive regulation of endothelial cell migrationGO:0045446 endothelial cell differentiationGO:0043542 endothelial cell migrationAmot Ecm1 Ctnnb1 Ecm1 Amot Ctnnb1 AmotItgb3 Itgb3 Afdn Itgb3 Itgb3 Afdn Rab13Slit2 Flt1 Rhoa Agtr1a Prkca Heg1 Slit2Klf4 Agtr1a Heg1 Pgf Gata2 Tmem100 RhoaPrkca Pgf Tmem100 Prkca Map3k3 Acvrl1 Hmgb1Pparg Prkca Acvrl1 Hmgb2 Smoc2 Plod3 Paxip1Gata2 Pparg Plod3 Gata2 Ets1 Col23a1 Cyp1b1Map3k3 Hmgb2 Col23a1 Aplnr Hdac7 Tnmd FapRhoa Gata2 Tnmd Cxcl12 Fgf18 Notch4 Emp2Smoc2 Aldh1a2 Notch4 Acvrl1 Gata3 Myadm Nr4a1Ets1 Aplnr Myadm Aggf1 Nrp1 Tjp1 Nrp1Acvrl1 Cxcl12 Tjp1 Nr4a1 Flt4 Nr2f2 Loxl2Hdac7 Acvrl1 Bmp4 Flt4 Fgf16 Ezr KdrCar10 Aggf1 Nr2f2 Fgfr3 Plpp3 Cldn1 Pik3r3Fgf18 Nr4a1 Ezr Adora2b Adora2b Dll1 Grem1Gata3 Tnmd Foxc2 Bmp4 Bmp4 Acvr1,Tgfb1 Clec14aEmp2 Flt4 Cldn1 Vegfc Sparc Prox1 Lemd3Dnaja4 Fgfr3 Dll1 Bmp2 Foxc2 Arhgef26 Efnb2Nrp1 Adora2b Acvr1,Tgfb1 Apela Vegfc Kdr Stard13Glul Bmp4 Prox1 Nrarp Plk2 Stc1Flt4 Sparc Arhgef26 Igf2 Igf2 Hey2Fgf16 Nr2f2 Kdr Prox1 Calr CNMDPlpp3 Foxc2 Stc1 Ccl2 Prox1 FasBmper Vegfc Hey2 Kdr Lgmn Ppp1r16bAdora2b Bmp2 CNMD Stat5a Kdr Rapgef3Bmp4 Apela Fas Dysf Stat5a TekSparc Nrarp Ppp1r16b Igf1 Ccbe1 Hey1Adgrb1 Igf2 Rapgef3 Ppp1r16b Igf1 Col18a1Nr2f2 Prox1 Tek Apln Tek DmdFoxc2 Ccl2 Stard13 Bmp6 Tgfb1Vegfc Kdr Hey1 Vegfd Anxa3Plk2 Stat5a Col18a1 Pdgfb PdgfbIgf2 Ptprm Dmd Sema5a Sema5aCalr CNMD PdgfbProx1 DysfLgmn Igf1Kdr Ppp1r16bStat5a Mef2cPtprm RgccCcbe1 AplnStc1 TekEfna1 Bmp6Igf1 VegfdMef2c PdgfbRgcc Sema5aTekStard13Tgfb1Adamts9Anxa3Meox2PdgfbSema5a


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items