UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

HP1-mediated transcriptional silencing of ERVs and genes in mouse embryonic stem cells Jensen, Kristoffer Nyquist 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_jensen_kristoffer.pdf [ 10.4MB ]
JSON: 24-1.0389977.json
JSON-LD: 24-1.0389977-ld.json
RDF/XML (Pretty): 24-1.0389977-rdf.xml
RDF/JSON: 24-1.0389977-rdf.json
Turtle: 24-1.0389977-turtle.txt
N-Triples: 24-1.0389977-rdf-ntriples.txt
Original Record: 24-1.0389977-source.json
Full Text

Full Text

HP1-mediated transcriptional silencing of ERVs and genes in mouse embryonic stem cells  by  Kristoffer Nyquist Jensen  B.Sc., The University of British Columbia, 2016  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  MASTER OF SCIENCE in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Medical Genetics)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  April 2020  © Kristoffer Nyquist Jensen, 2020   ii The following individuals certify that they have read, and recommend to the Faculty of Graduate and Postdoctoral Studies for acceptance, a thesis entitled:  HP1-mediated transcriptional silencing of ERVs and genes in mouse embryonic stem cells  submitted by Kristoffer Nyquist Jensen in partial fulfillment of the requirements for the degree of Master of Science in Medical Genetics  Examining Committee: Matthew C. Lorincz Supervisor  Pamela Hoodless Supervisory Committee Member  Fabio Rossi Supervisory Committee Member Ivan J. Sadowski Additional Examiner     iii Abstract The three mammalian Heterochromatin Protein 1 (HP1) proteins are considered hallmarks of H3K9me3-marked heterochromatin, and are essential for establishing this transcriptionally silent chromatin state genome-wide. They also have a proposed involvement in regulating H3K9me3 levels, based on their interaction with histone lysine methyltransferases. However, they have individually been shown to not be integral for silencing of certain classes of endogenous retroviruses (ERVs). I show that HP1 isoforms in mESCs are functionally redundant and only upon deletion of all three isoforms is there a loss of ERV silencing. I also show that although there are some minor effects on H3K9me3 levels genome-wide following HP1 protein depletion, there are minimal effects on this mark over genes and ERVs, unlike the effects seen on the H3K9me3- and HP1-dependent mark H4K20me3. I also investigate two reported HP1-interacting proteins, AHDC1 and CHAMP1, for their impact on gene regulation and pluripotency maintenance in mESCs. I identify gene promoters where HP1 isoforms are bound independently of H3K9me3, and therefore hypothesize that AHDC1 and CHAMP1 are responsible for HP1 protein recruitment to these promoters via their putative DNA-binding motifs. Both AHDC1 and CHAMP1 have associated severe neurodevelopmental phenotypes when mutated in humans, which could be caused by disruption of HP1-mediated gene silencing and aberrant stem cell differentiation patterns. However, I find no significant effect on gene regulation upon disruption of these two genes in mESCs. I also find no observed effect on cell growth or differentiation potentials of Ahdc1 and Champ1 KO cells. The observed neurodevelopmental phenotypes in humans can therefore not be explained by disruption of HP1-mediated gene silencing in mESCs, although it is still possible they are caused by failure of this mechanism in differentiated cells at later developmental stages.   iv Lay Summary The Heterochromatin Protein 1 family includes three proteins that help package the DNA strands by tethering together DNA-bound proteins called histones. DNA regions to be compacted are marked by a chemical modification to a lysine residue on these histones (H3K9me3), which act as binding sites for HP1 proteins, ultimately preventing transcription of such regions. I used mouse embryonic stem cells lacking all three HP1 proteins in conjunction, and investigated the effects on transcription of certain “parasitic” DNA sequences that should remain un-transcribed in order to protect the DNA from disruptive mutations. I also used CRISPR/Cas9-mediated gene editing and high throughput sequencing technology to assess if HP1 proteins can influence where H3K9me3 is deposited. Finally, I studied two proteins interacting with HP1 proteins that could be responsible for HP1 protein recruitment to regions without H3K9me3, to see if losing these had effects on gene transcription in stem cells.     v Preface The experiments reported in this thesis were designed in collaboration with my supervisor, Dr. Matthew C. Lorincz, and conducted by me. I also performed the analysis of publicly available and self-generated data. Additionally, I supervised two undergraduate students, Rebecca Ruthven and William Poon, who helped in the design and transfection of CRISPR/Cas9 short guide RNAs, tissue culturing, and screening for knock-out cells by PCR and Western blot (presented in Chapter 4). All figures were generated by me, except the Western blot in Figure 4.4, which was performed by William Poon. The CRISPR/Cas9 vector backbone used for cloning sgRNA plasmids was generated by Dr. Louis Lefebvre. The HP1 conditional triple knock-out cell line used to generate ChIP-seq data and CRISPR/Cas9-edited cell lines was a kind gift from Dr. Haruhiko Koseki. Dr. Lorincz, Dr. Julie Brind’Amour, Aaron Bogutz, and Julien Richard Albert helped with editing and helpful comments during thesis writing.    vi Table of Contents  Abstract ......................................................................................................................................... iii	Lay Summary ............................................................................................................................... iv	Preface .............................................................................................................................................v	Table of Contents ......................................................................................................................... vi	List of Tables ................................................................................................................................ ix	List of Figures .................................................................................................................................x	List of Abbreviations .................................................................................................................. xii	Acknowledgements .................................................................................................................. xviii	Dedication ................................................................................................................................. xviii	Chapter 1: Introduction ................................................................................................................1	1.1	 Functions and organization of the mammalian genome ................................................. 1	1.2	 DNA is organized into protein-associated chromatin ..................................................... 3	1.3	 Post-translational modifications of histone tails ............................................................. 5	1.4	 The “writers” of H3K9me3 ............................................................................................. 6	1.5	 H3K9me3-mediated transcriptional silencing ................................................................ 7	1.5.1	 Transposable elements require H3K9me3 for their silencing ................................. 8	1.6	 H3K9me3 marks are “read” by HP1 proteins ............................................................... 11	1.7	 HP1 proteins: structure and conservation ..................................................................... 13	1.8	 HP1 proteins have varied chromatin-associated functions ........................................... 15	1.9	 Proteins interacting with HP1 isoforms ........................................................................ 17	1.10	 Thesis goals ................................................................................................................... 20	  vii Chapter 2: Materials and Methods ............................................................................................23	2.1	 Cell lines ....................................................................................................................... 23	2.2	 Generation of CRISPR KO plasmid ............................................................................. 23	2.3	 Whole-cell protein extraction and western blot ............................................................ 25	2.4	 Immunofluorescence cell imaging ................................................................................ 25	2.5	 Genomic DNA isolation and genotyping ...................................................................... 26	2.6	 RNA extraction, RT-qPCR, and RNA sequencing ....................................................... 27	2.7	 Native Chromatin immunoprecipitation (ChIP) ........................................................... 28	2.8	 Construction of ChIP sequencing libraries and bioinformatic analysis ........................ 30	2.9	 Primer sequences .......................................................................................................... 31	Chapter 3: HP1 proteins are indispensable for ERV silencing ...............................................33	3.1	 Background ................................................................................................................... 33	3.2	 Results ........................................................................................................................... 36	3.2.1	 Loss of HP1 isoforms cause abnormal mESC growth .......................................... 36	3.2.2	 HP1 proteins act redundantly to silence ERVs in mESCs .................................... 38	3.2.3	 H3K9me3 levels are not affected by loss of HP1 proteins ................................... 40	3.3	 Discussion ..................................................................................................................... 45	Chapter 4: HP1 proteins and their interactors in gene regulation .........................................49	4.1	 Background ................................................................................................................... 49	4.1.1	 HP1-interacting candidate #1: AHDC1 ................................................................ 55	4.1.2	 HP1-interacting candidate #2: CHAMP1 ............................................................. 58	4.2	 Results ........................................................................................................................... 61	4.2.1	 mESCs harboring a deletion of Ahdc1 show a normal growth phenotype ........... 61	  viii 4.2.2	 Deletion of Ahdc1 does not significantly affect gene expression ......................... 62	4.2.3	 Ahdc1 KO mESCs have a normal differentiation potential .................................. 65	4.2.4	 mESCs harboring a deletion of Champ1 show a normal growth phenotype ........ 67	4.2.5	 Deletion of Champ1 does not significantly affect gene expression ...................... 68	4.2.6	 Champ1 KO mESCs have a normal differentiation potential ............................... 71	4.3	 Discussion ..................................................................................................................... 72	Chapter 5: Concluding remarks and future directions ............................................................77	5.1	 Major findings ............................................................................................................... 77	5.2	 Relevance of studying HP1 mediated transcriptional silencing ................................... 79	5.3	 Future directions and outstanding questions ................................................................. 80	5.4	 Conclusion .................................................................................................................... 83	Bilbliography ................................................................................................................................84	   ix List of Tables  Table 1-1 Reported phenotypes upon in vivo deletion of HP1 isoforms in mice ......................... 16	Table 1-2 Interactors of HP1a.. .................................................................................................... 19	Table 2-1 Target sequences for CRISRP/Cas9 sgRNAs. ............................................................. 24	   x List of Figures  Figure 1-1 Structure of the nucleosome. ......................................................................................... 3	Figure 1-2 Comparison of the three isoforms of HP1 in mice. ..................................................... 14	Figure 1-3 Germ cell and early embryonic development expression of HP1 isoforms. ............... 16	Figure 2-1 Map of CRISPR/Cas9 vector. ..................................................................................... 25	Figure 2-2 Alignment of sequenced PCR results for CRISPR/Cas9-mediated KO validation .... 27	Figure 3-1 Validation of deletion of HP1 genes in mESCs. ......................................................... 38	Figure 3-2 Expression of ERV families in HP1 protein-depleted mESCs ................................... 39	Figure 3-3 RNA- and ChIP-seq analysis of ERV families in HP1 TKO cells ............................. 41	Figure 3-4 Differential ChIP-seq analysis of H4K20me3 in HP1 WT and TKO cells. ................ 43	Figure 3-5 Differential ChIP-seq analysis of H3K9me3 in HP1 WT and TKO cells ................... 44	Figure 4-1 Interactomes of HP1 proteins ...................................................................................... 51	Figure 4-2 HP1 isoforms bind genes independent of H3K9me3 .................................................. 54	Figure 4-3 The Ahdc1 locus and deletion strategy ........................................................................ 57	Figure 4-4 The Champ1 locus and deletion strategy .................................................................... 60	Figure 4-5 Growth curves of Ahdc1 WT and KO mESCs. ........................................................... 62	Figure 4-6 Gene expression analysis of Ahdc1 KO mESCs ......................................................... 64	Figure 4-7 Expression and ChIP-seq analysis of genes in Ahdc1 KO mESCs ............................. 65	Figure 4-8 Analysis of differentiation potential following Ahdc1 KO in mESCs ........................ 66	Figure 4-9 Growth curve of Champ1 WT and KO mESCs. ......................................................... 68	Figure 4-10 Gene expression analysis of Champ1 KO mESCs .................................................... 70	Figure 4-11 Expression and ChIP-seq analysis of genes in Champ1 KO mESCs ........................ 71	  xi Figure 4-12 Analysis of differentiation potential following Champ1 KO in mESCs ................... 72	     xii List of Abbreviations ˚C – degrees Celsius µL – microliter  2i – GSK3b and MEK1/2 inhibitors 4-OHT – 4-hydroxytamoxifen aa – amino acid ADNP – Activity-dependent Neuroprotective Protein  AHDC1 – AT-hook DNA-binding Motif Containing 1  aka – also known as  AurkB – Aurora kinase B bp – base pair BSA – Bovine Serum Albumin BWA – Burrows-Wheeler Aligner C2H2 motif – Cystein2/Histidine2 motif CAF1 – Chromatin Assembly Factor 1  Cas9 – CRISPR-associated protein 9 CBX1/2/3/5/6/7 – Chromobox 1/2/3/5/6/7 cDNA – complementary DNA CDYL2 – Chromodomain Y Like 2 CHAMP1 – Chromosome Alignment-maintaining Phosphoprotein 1 CHD - chromodomain CHD4 – Chromodomain Helicase DNA binding protein 4 ChIP – chromatin immunoprecipitation   xiii ChIP-seq – chromatin immunoprecipitation sequencing Chr – chromosome CRISPR – clustered regularly interspaced short palindromic repeats CSD – chromoshadow domain CSF1R – Colony-stimulating Factor 1 Receptor C-terminal – carboxy-terminal cTKO – conditional triple knockout DAPI – 4',6-diamidino-2-phenylindole
 DEPC – diethyl pyrocarbonate DNA – deoxyribonucleic Acid DNAme – DNA methylation DTT – Dithiothreitol  E1.5-19.5 – embryonic day 1.5 – 19.5 EB – embryoid body EDTA - Ethylenediaminetetraacetic acid ERV – endogenous retrovirus ESC – embryonic stem cell  ETn/MusD – Early transposon/Mus musculus type D EtOH – ethanol fc – fold-change fl - floxed g – gravities  GATA4 – GATA binding protein 4   xiv GFP – Green Fluorescent Protein GLP – G9A-like Protein GO – gene ontology h – hour H3K4me3 – Histone 3 Lysine 4 tri-methylation H3K9me1/2/3 – Histone 3 Lysine 9 mono-/di-/tri-methylation H3K27me3 – Histone 3 Lysine 27 tri-methylation HCl – hydrochloric acid  HP1 – Heterochromatin Protein 1  IAP – Intracisternal A--particle ID – intellectual disability KAP1 – KRAB-associated Protein 1  k – kilo / 1000 kb – kilobase  kDa – kilo-Dalton KDM1A/LSD1 – Lysine-specific histone demethylase 1A KMT – lysine methyl transferase KO – knock-out KRAB-ZFP – Krüppel-associated Box Domain – Zinc Finger Protein L – leucine  LIF – Leukemia Inhibitory Factor LINE – Long Interspersed Nuclear Element LTR – long terminal repeat   xv mag. – magnification  MAPQ – mapping quality score MEF – mouse embryonic fibroblast MeOH – methanol mESCs – mouse embryonic stem cells MPP8 – M-phase Phosphoprotein 8 NaCl – sodium chloride NaHCO3 – sodium bicarbonate  NaOH – sodium hydroxide NHEJ – non-homologous end joining  N-terminal – amino-terminal  OCT4 – Octamer-binding Transcription Factor 4 ORC – origin recognition complex ORF – open reading frame P – proline P2 – post-natal day 2 PBS – Phosphate-buffered saline PBST – Phosphate-buffered saline + Tris PCR – polymerase chain reaction PGC – primordial germ cell PIC – proteinase inhibitor cocktail PMSF - phenylmethylsulfonyl fluoride POGZ – Pogo Transposable Element-derived Protein with Zinc Finger Domain   xvi PPIA – Peptidylpropyl Isomerase A PRC1/2 – Polycomb Repressive Complex 1/2 PTM – post-translational modification PVDF – polyvinylidine difluoride RIPA buffer – Radioimmunoprecipitation Assay Buffer RNA – ribonucleic acid RNA-seq – RNA sequencing RPKM – reads per kilobase per million RT-qPCR – reverse transcriptase quantitative polymerase chain reaction SETDB1 – SET domain-containing Box 1 SDS – sodium dodecyl sulfate sgRNA – short guide RNA SINE – Short Interspersed Nuclear Element SOX2/17 – SRY-box Transcription Factor 17 SSIV – SuperScript IV SUV39H1/2 – Suppressor of variegation 3.9. h1/2 SUV420H1/2 – Suppressor of variegation 4.20. h1/2 SuVar – Suppressor of Variegation TBS – Tris-buffered saline TE – transposable elements TKO – triple knockout TSS – transcription start site UBC BRC – University of British Columbia – Biomedical Research Centre   xvii ULI – ultra-low input V – valine  WT - wildtype ZFP / ZNF – zinc finger-containing protein / zinc finger   xviii Acknowledgements Firstly, I would like to thank all members of the Lorincz Lab, past and present, for their help and guidance. Julien, Aaron, Julie, Kenjiro, Kentaro, Carol, Irina, Peter, Matt and Preeti, your hard work and dedication to your craft inspired me to do my best every day. In particular, I would like to thank Julie for her patience in answering my never-ending list of questions, for her comforting words and guidance when science experiments grow frustrating, and for the love and care she shows for those around her. Also, a big thank you to my supervisor, Dr. Matt Lorincz, for allowing me time and space to explore my passion for genetics. Your passion for genetics and science is inspiring, and will continue to shape me as a scientist in the years to come. And thank you to my supervising committee, Dr. Pamela Hoodless and Dr. Fabio Rossi, for your time and guidance.  Secondly, I want to thank Julien for putting up with me both at work and as a roommate, and a huge thank you also to Meaghan. How you’ve remained such a beam of sunshine in the face of two grumpy roommates writing their graduate theses is beyond my comprehension. It has been a privilege to share a home with you both for these few years, and my diet of Coke Zero and pasta would have been the end of me if not for your insistence that I eat an occasional vegetable. I will, however, never concede that ketchup is not a tomato sauce! Thank you, as well, to my fellow students in the Molecular Epigenetics Group, Amanda, Sam, Brad, Tom, Bronwyn, Roula, Riley, Sam, Fagun, Thomas, James, Grace, Oscar, Ben, Kristina, Nicole and Mackenzie, for your support and friendship.    xix I also want to express my love and gratitude to my friends for putting up with my ups and downs. Joaquin, Tyler, Angus, Taylor, Alex and Kai, you are my chosen family and my emotional support. Thank you for the countless hours of drinks, sports, and general merriment, and for reminding me to exercise my body in addition to my mind. My grad school experience would have been much sadder without you. To Kent and Brandon, thank you for cheering me on, checking in on me, and allowing me to spend time with you, and your dogs, when school stress ran high. Love and gratitude also to Jill, Ian, Hannah, and Francesca, who’s friendship I value immensely, and who’s late night and early morning counseling, both in person and via text messaging was vital for my sanity. I could not have done any of this without any of you!  Lastly to my family: Mamma, Pappa, Farmor, Andreas, Johanne, Erlend, Iver, Ingvild, Sverre, Mirjam and tante ABO. Thank you for allowing me to travel and explore the world with the knowledge that you’ll love and support me in whatever crazy adventure I chose next. Home will always be wherever you all are.   xx Dedication To mom, who taught me to  always ask questions. And to dad, who taught me to always have patience when answering them.    1 Chapter 1: Introduction 1.1 Functions and organization of the mammalian genome The information necessary for a cell to develop, differentiate, and sustain itself is encoded in a sequence of deoxy-ribonucleic acid (DNA) residues strung into a genome. This DNA sequence encodes all protein-coding genes, which are transcribed to ribonucleic acid (RNA) sequences, then translated into proteins that carry out cellular functions and homeostasis. Although we still have much to learn about the genome, in the 20 years since the first human genome was sequenced, researchers have shown that the dogma of genome function is a lot more complex than “gene to RNA to protein”. For example, one totipotent cell can differentiate into every cell type of the adult organism, each carrying the same genome. Each cell must therefore not only contain the blueprint for each protein needed, but also the instructions for which genes to express, when and how much protein to produce. In their review, Shapiro and Sternberg (2005) list functions required of a genome that go beyond encoding proteins:  (1) Regulating timing and extent of coding sequence expression. (2) Organizing coordinated expression of protein and RNA molecules that function together. (3) Packaging DNA appropriately within the cell. (4) Replicating the genome in synchrony with the cell division cycle. (5) Transmitting replicated DNA accurately to progeny cells at cell division. (6) Detecting and repairing errors and damage to the genome. (7) Restructuring the genome when necessary (as part of the normal life cycle or in response to a critical selective challenge).     2 These functions are, in large part, achieved by complexing DNA strands with other molecules in the nucleus, e.g. the transcriptional machinery, and histone proteins involved in packaging DNA.   Surprisingly, although proteins carry out most required cellular functions, the complement of protein coding genes comprises only 1.5% of the mouse genome, and close to 50% of the mammalian genome is made up of repetitive DNA sequences (Slotkin & Martienssen, 2007).When the genomes of higher eukaryotes were first sequenced, these long stretches of repetitive, non-genic DNA, mainly consisting of satellite DNAs and transposable elements (TEs), were labeled ‘junk DNA,’ as they littered the genome with no known function (Ohno, 1972). However, we now know that this interpretation is misleading, and many of the functions proposed by Shapiro and Sternberg rely heavily on these regions (Shapiro & Sternberg, 2005). For example, the centromeric regions of chromosomes are made up of repetitive sequences tightly packaged into protein-DNA structures called heterochromatin, which aids in transmitting replicated DNA accurately to progeny cells during cell division. Since some repetitive sequences, like TEs, can be transcriptionally active, they may also act as regulatory sequences for nearby genes, and thereby organize coordinated gene expression as well as regulate timing and extent of transcription. Understanding the mechanisms involved in transcriptional silencing of repetitive sequence can therefore help us understand gene regulation and the process of cell differentiation and development. The work presented here will focus on how the genome prevents aberrant transcription of TEs by forming compacted protein-DNA structures around repetitive sequences, and how, in some cases, this strategy is also utilized to regulate the expression of genes for proper development and cellular differentiation.     3 1.2 DNA is organized into protein-associated chromatin Given the size of the mouse genome (~3 billion bases) and its need to fit into the roughly 10µm wide nucleus, it is necessary to compact and organize the DNA strands. In mice, the 19 autosomal chromosomes and two sex chromosomes are packaged into protein-DNA complexes called chromatin by wrapping DNA around histone proteins. This forms the nucleosomes, the basic unit of chromatin, with 146-147 base pairs (bps) wrapped around histone octamers made up of two copies each of histones H2A, H2B, H3 and H4 (Luger, Mäder, Richmond, Sargent, & Richmond, 1997) (Figure 1.1).    Figure 1-1 Structure of the nucleosome. 146-1477bp of DNA is wrapped around a core of two copies each of H2A, H2B, H3 and H4. The histone tails protrude from the nucleosome core and are subject to PTMs. Some commonly methylated lysine residues on H3 and H4 are highlighted, including H3K9 and H4K20.   Nucleosomes can in turn be bound together to form higher order chromatin structures, with the highest level of compaction being the X-shaped condensed chromosomes seen during metaphase. Compaction of DNA into chromosomes is vital for efficiently segregating DNA strands equally into daughter cells during cell division (Dernburg, Sedat, & Hawley, 1996). Additionally, chromatin compaction has important implications in the regulation of transcription, since DNA    4 wrapped around histones is largely inaccessible to the transcription machinery (reviewed in Cutter & Hayes, 2015). Thus, the chromatin structure must be stable and resilient enough to protect the genome integrity throughout the cell cycle and development, as well as plastic enough to be easily adapted to changing transcriptional requirements as the cell develops and differentiates. To that end, some regions, including the long stretches of repetitive DNA in telomeres and centromeres, are constitutively condensed, whilst genic regions may be densely packaged in some cell types and at specific developmental times, but de-condensed and open in other cell types. The differential packaging density of genomic regions have led to two distinct chromatin classifications; euchromatin – which is gene rich, loosely packaged, and more available for transcriptional activity, and heterochromatin – the densely packaged and largely transcriptionally silenced chromatin state.   Although heterochromatin and euchromatin are commonly described as transcriptionally silenced or active regions, respectively, it is in fact the underlying gene density and sequence that seem most closely correlated with the transcriptional activity of each chromatin state  (Gilbert, et al., 2004). Furthermore, based on the persistence of the chromatin structure in different cells and at different developmental stages, heterochromatin can be sub-classified into two classes, constitutive and facultative. As the name suggests, constitutive heterochromatin is established early in development and remains condensed throughout life. In contrast, facultative heterochromatin is more dynamic and can be established and removed in different cell types and at different developmental stages. Both regions share some structural characteristics, however, each of these regions can also be distinguished by distinct DNA-interacting proteins, covalent post-translational modifications (PTMs) of histone tails, and the “readers” of these PTMs.    5  1.3 Post-translational modifications of histone tails As previously discussed, chromatin state is, in large part, correlated with gene density and the make-up of the underlying DNA sequence. One key finding shows that each chromatin state is enriched for distinct PTMs of the N-terminal tail of histones (reviewed in Beisel & Paro, 2011). Characterized histone PTMs include methylation, phosphorylation, acetylation, ubiquitination and sumoylation, found at over 60 distinct amino acid residues (reviewed in Bannister & Kouzarides, 2011). A plethora of enzymes have since been identified that can “write” these PTMs at specific regions of the genome, generally guided by DNA sequences or factors binding in a DNA sequence-specific manner. The presence of PTMs in a region can in turn be recognized and bound by “reader” enzymes that exert various downstream functions. The combination of these histone PTMs, and the “writers” and “readers” affecting them, make up the Histone Code (Strahl & Allis, 2000), promoting or inhibiting critical nuclear pathways such as transcription or DNA repair. Throughout embryonic development and adult life, epigenetic changes – including, but not limited to, changing histone PTMs, the availability of histone PTM “readers”, and associated changes in chromatin compaction – are believed to be the key events allowing cells to maintain their cellular identity, or differentiate.   The most well-characterized histone PTMs are the lysine methyl-marks on the histone H3 tail. While some PTMs, such as tri-methylation of histone H3 at lysine 4 (H3K4me3) are considered “active” marks (H3K4me3 is found over actively transcribing gene promoters), other marks such as H3K9me3 and H3K27me3 are recognized as “repressive” marks, as they are found in transcriptionally silent regions of the genome (Beisel & Paro, 2011). Some histone tail    6 modifications, such as acetylation, are thought to change the DNA-histone binding affinity, thereby altering the chromatin compaction and DNA accessibility. Unlike acetylation, methylation of histone tails does not significantly alter the histone protein charge and DNA affinity. Rather, these marks are thought to exert their function through recruitment or exclusion of chromatin-modifying and transcriptional machineries. This thesis will mainly focus on H3K9me3, proteins that bind to this mark, and the influence of these interactions on transcription of genes and/or repetitive DNA elements.  1.4 The “writers” of H3K9me3 H3K9me3 is considered a silencing histone mark, and is a hallmark of heterochromatin regions. In mice, there are many lysine methyl transferases (KMTs), but only 5 are known to “write” histone lysine 9 methylation; G9A, GLP (G9A-like protein), SUV39H1 (Suppressor of variegation 39h1), SUV39H2, and SETDB1 (SET domain-containing Box 1). All of these contain the conserved catalytic SET domain, which facilitates methylation of lysine 9 (Leung & Lorincz, 2012), and a number of other domains that may be unique or shared among them. While they differ in where in the genome they “write” this methyl mark, none of these KMTs include domains that bind to DNA. Rather it is their interactions with other chromatin factors, including DNA-binding factors which may recognize specific sequence motifs, that likely guides these epigenetic modifiers to specific genomic regions.   G9A and GLP are closely related KMTs which form a heterodimer to write mono- and di-methylation of lysine 9 (Tachibana, et al., 2002). In embryonic stem cells (ESCs) and somatic cells, G9A/GLP deposit H3K9me1/2 over large megabase-sized domains predominantly in    7 euchromatic regions (Peters, et al., 2003; Rice, et al., 2003). In contrast, SUV39H1/2 are paralogs that catalyze H3K9 di- and tri-methylation in the pericentromere to establish constitutive heterochromatin (Aagaard, Schmid, Warburton, & Jenuwein, 2000; Peters, et al., 2001). SUV39H1/2 also contain a domain that can recognize H3K9 methylation, and can therefore act as both a “writer,” and a “reader” to reinforce and help spread this mark (Mozzetta, Boyarchuk, Pontis, & Ait-Si-Ali, 2015). Finally, SETDB1 can write H3K9 mono-, di- and tri-methylation, but predominantly lays down H3K9me3 over specific classes of TEs in euchromatic regions (Karimi, et al., 2011; Liu, et al., 2014; Matsui, et al., 2010).   1.5 H3K9me3-mediated transcriptional silencing Due to their targeting of distinct genomic regions, and the differences seen between H3K9me2 and me3 patterns, it is not surprising that disruption of each of these K9 KMTs results in distinct phenotypes in vivo. G9A/GLP knock-out (KO) mice die from severe growth restrictions by embryonic day 9.5 (E9.5) (Tachibana, et al., 2002; 2005), while SUV39H1/2 KOs show reduced viability in utero. Surviving SUV39H1/2 KO embryos show growth retardation at birth (Peters, et al., 2001). Finally, SETDB1 KO embryos die at the peri-implantation stage, around E3.5-5.5 (Dodge, Kang, Beppu, Lei, & Li, 2004).  In mouse ESCs (mESCs), disruption of the K9 KMTs impacts transcription of distinct genes and TEs, and the cell’s ability to divide and differentiate. For example, disruption of G9A/GLP leads to global loss of H3K9me2 in euchromatic regions, up-regulation of a class of genes with H3K9me2 marks in their promoter, as well as increased expression of one class of TEs, Class III endogenous retroviruses (ERVs) (Maksakova, et al., 2013; Tachibana, et al., 2005; Yokochi, et    8 al., 2009). In contrast, depletion of SUV39H1/2 leads to loss of H3K9me3 over pericentromeric heterochromatin, with resulting chromosome instability, disrupted cell division, and increased transcription of major satellite DNA elements from the pericentromere (Lehnertz, et al., 2003). As opposed to the up-regulation of class III ERVs in G9a/GLP KOs, depletion of SETDB1 results in a significant up-regulation of class I and II ERVs (Matsui, et al., 2010).   1.5.1 Transposable elements require H3K9me3 for their silencing As described by Shapiro and Sternberg (2005), the presence of TEs in the mammalian genome influences many genome functions. Some TEs retain their transcriptional activity following genome insertion, meaning they contain regulatory elements that can affect transcription of the TE itself, as well as surrounding sequences, including genes (Brind'Amour, et al., 2018). TEs are also of interest because of their ability to transpose, i.e. copy themselves, or “jump”, to new locations in the genome. This can be either detrimental or beneficial to the host genome. Much as any other mutational event, TE insertion or excision may lead to disruption of coding sequence or regulatory sequence. In fact, roughly 10% of all germline mutations in mice are caused by novel TE insertions (Gagnier, Belancio, & Mager, 2019). However, most TE insertions must be either neutral, or even beneficial to the host, either as a source of new regulatory sequence, or simply as a provider of mutations to drive evolution and speciation, considering their high abundance in the mouse and human genomes (Zhang, Romanish, & Mager, 2011) (Thompson, Macfarlan, & Lorincz, 2016).   ERVs, also called Long Terminal Repeat (LTR) retrotransposons due to the presence of repeat sequences at each end of the element, are a class of TEs originating from ancient retroviral    9 infection of the host germline genome. Over time, ERV insertions have come to make up over 10% of the total mouse genome (Stocking & Kozak, 2008). To mobilize in the host genome, retrotransposons employ reverse transcription to generate viral “genomic copies” for retrotransposition, essentially expanding their numbers in the genome by a “copy-paste” mechanism. In contrast, non-LTR retrotransposons (long and short interspersed nuclear elements [LINEs and SINEs]) are the most abundant interspersed repeats in the mouse genome, constituting 27.4% (Stocking & Kozak, 2008).  Full-length ERVs are autonomous TEs, meaning they contain all the sequences required for their retrotransposition. These elements are made up of LTRs flanking an open reading frame of viral genes. However, like the rest of the host genome, TEs will accumulate deleterious mutations over time, and most ERVs are therefore transcriptionally silent (Mager & Stoye, 2015). Further, due to the identical repetitive sequences in the 5’ and 3’ LTR regions, ERVs are prone to homologous recombination, where the LTR sequences align, leading to excision of the viral genes between them and leaving behind a solo LTR sequence (Mager & Stoye, 2015) (Nellåker, et al., 2012). These solo LTRs are the most abundant ERV-derived sequence in mammalian genomes (Belshaw, et al., 2007). Intriguingly, since the LTR region contains transcription factor (TF) binding sites, solo LTRs can retain their transcriptional potential and may function as cis regulatory elements of downstream host genes. In fact, the majority of regulatory sequences in primates are thought to be derived from TEs (Jacques, Jeyakani, & Bourque, 2013), and aberrant expression from such TEs can influence the expression of nearby genes (Brind'Amour, et al., 2018). This further highlights the strong need for the genome to control transcription of ERV sequences even when they are not capable of retrotransposition.    10  Since some ERV families retain their transcriptional potential, and a subset of these the ability to retrotranspose (Stocking & Kozak, 2008), the genome has developed numerous ways to repress TE expression. These include transcriptional silencing by methylation of cytosine in the DNA strands (DNA methylation [DNAme]), and/or PTMs, which promote chromatin compaction (reviewed in (Slotkin & Martienssen, 2007)). DNAme is closely linked to transcriptional silencing by H3K9 methylation and is important in processes such as genomic imprinting. Although essential, and much studied, DNAme is beyond the scope of this thesis, and will thus not be further discussed here. TEs can also be silenced post-transcriptionally by RNA interference and the PIWI/piRNA pathway in the germline.   H3K9me3 is the major silencing histone PTM responsible for ERV silencing, as specific classes of ERVs show increased transcriptional activity upon loss of any one of the K9 methyl transferases (Lehnertz, et al., 2003; Maksakova, et al., 2013; Matsui, et al., 2010; Tachibana, et al., 2005). In particular, SETDB1-dependent silencing of ERVs, which has been shown to be dependent on KAP1 (KRAB-associated Protein 1), has been much studied. In mESCs, SETDB1 is believed to interact with KAP1, which is a co-repressor recruited to ERVs by KRAB-ZFPs (Krüppel-associated box domain-zinc finger proteins) recognizing sequence motifs in specific classes of ERVs (Rowe, et al., 2010). Although a direct interaction between KAP1 and SETDB1 has been difficult to confirm, KAP1 and SETDB1 are clearly both required for proper silencing of ERVs in euchromatic regions, since loss of either leads to loss of H3K9me3, and increased transcriptional activity of overlapping sets of ERVs both in mESCs and in vivo (Liu, et al., 2014; Matsui, et al., 2010; Rowe, et al., 2010). Since both KAP1 and SETDB1 have been reported to    11 interact with HP1 proteins (Hauri, et al., 2016), it is also possible that SETDB1 and KAP1 do not directly bind each other, but rather that their interaction with HP1 proteins mediates their recruitment to genomic locations for silencing of ERVs, although this has not yet been investigated.    1.6 H3K9me3 marks are “read” by HP1 proteins Since histone tail methylation does not alter the DNA-binding affinity of histones to change chromatin packaging, H3K9me3 likely affects chromatin packaging and transcription via proteins binding, or “reading”, this histone mark. The most well characterized K9me3 “readers” are the HP1 family of proteins, which, alongside H3K9me3, are hallmarks of heterochromatic regions. Recent work showed that, in vitro, HP1 proteins will dimerize, with one subunit binding H3K9me3-marks on adjacent histones, essentially forming a bridge between histones and resulting in chromatin compaction (Machida, et al., 2018). In mice, HP1 proteins and H3K9me3 mark large heterochromatin domains covering the telomere and centromeric regions, but is also found as smaller domains throughout the euchromatic regions of the genome, such as over some gene promoters and over specific classes of ERVs (Bickmore & van Steensel, 2013). The presence of H4K20me3 at such ERVs further implicates HP1 proteins in this silencing process, since HP1 isoforms are required for the recruitment of the H4K20 methyltransferases SUV420H1/2 (Schotta, et al., 2004). Previous work in the Lorincz Lab concluded that individual HP1 isoforms were not required for silencing of class I & II ERV families, and only a moderate up-regulation of class III ERVs was seen upon HP1 isoform depletion (Maksakova, et al., 2011). However, the possibility that HP1 isoforms are highly redundant with respect to silencing of ERVs remained to be addressed. I therefore chose to revisit the link between different HP1    12 isoforms and ERV silencing in greater detail, exploiting a recently developed HP1 triple knockout (TKO) mESC line.  As previously discussed, heterochromatin can also be further subdivided into two categories; constitutive and facultative heterochromatin. Both categories are transcriptionally silenced, and show a high degree of chromatin compaction. Constitutive telomeric and centromeric heterochromatic regions are bound by, predominantly, HP1a and b isoforms and marked by H3K9me3, and remain condensed throughout the cell cycle and in all investigated cell types. In contrast, facultative heterochromatin usually targets genes in what is normally considered euchromatic regions in a subset of cell types, at specific time points in development. While there is still much discussion over what exactly characterizes facultative heterochromatin, it is clear that facultative heterochromatin exists in regions both marked by, or lacking H3K9me3. For example, regions marked by H3K27me3 are, by many, also considered facultative heterochromatin (Beisel & Paro, 2011). X inactivation, the process in which one X chromosome is silenced in female somatic cells to regulate gene dosage, relies on H3K27me3, as deposited by the Polycomb Repressive Complex 2 (PRC2), as well as DNAme and a “coating” of the chromatin strand by the long non-coding RNA Xist (reviewed in Chow & Heard, 2009). Facultative heterochromatin can also form in regions lacking both H3K9me3 and H3K27me3, such as in the promoter of genes marked by the ChAHP complex. ChAHP is a three-protein complex of the sequence-specific DNA-binding protein ADNP (Activity-dependent Neuroprotective Protein), HP1g, and the chromatin remodeller CHD4 (Chromodomain Helicase DNA binding protein 4), which silences genes essential for cell lineage differentiation in mESCs (Ostapcuk, et al., 2018) (Further discussed below). The work presented here will focus on the    13 role of the readers of H3K9me3, namely HP1 proteins, in establishing facultative heterochromatin for transcriptional repression of ERVs (Chapter 3), and on transcriptional repression of genes marked by HP1 isoforms in absence of H3K9me3 (Chapter 4).  1.7 HP1 proteins: structure and conservation The Heterochromatin Protein 1 family are the canonical “readers” of H3K9me3. While orthologues of mammalian HP1 proteins were first identified and characterized in Drosophila melanogaster (Eissenberg, et al., 1990; James & Elgin, 1986), HP1 homologues are found in species ranging from fission yeast (Saccharomyces pombe) to human, with some species containing one (S. pombe and Neurospora) or two (C. elegans) isoforms. Three isoforms; HP1a (aka Cbx5), HP1b (aka Cbx1) and HP1g (aka Cbx3), are encoded in mice and humans, whilst Drosophila similarly have three isoforms (HP1a, b and c), plus two tissue-specific isoforms (HP1d in ovaries and HP1e in testes). Early studies in Drosophila revealed that HP1 proteins act as suppressors of position effect variegation (SuVar/SUV), and as an integral part of the condensed structure of heterochromatin ((Eissenberg, et al., 1990; James & Elgin, 1986)), illustrating the critical role of readers in directing the functional “readout” of histone PTMs.  The amino acid sequence in mice share ~50% identity with that of Drosophila HP1 isoforms, and the protein domains and organization is well conserved throughout eukaryotes (Singh, et al., 1991)(Figure 1.2). The domains and organization are also conserved between isoforms, with each isoform containing a chromodomain (CHD), a chromoshadow domain (CSD), and a hinge region connecting the two. The CHD has been shown in vitro to bind directly to histone    14 H3K9me2 and me3, with a strong preference for H3K9me3 (Bannister, et al., 2001; Lachner, O'Carroll, Rea, Mechtler, & Jenuwein, 2001).   Figure 1-2 Comparison of the three isoforms of HP1 in mice. The chromodomain (CHD) is responsible for recognizing and binding H3K9me3. The Chromoshadow domain (CSD) is involved in protein dimerization and other protein interactions.   Most of the variation between HP1 isoforms is found in the length and sequence of the hinge region. In both mice and humans, all three isoforms are approximately 20kDa in mass, only differing by 18 amino acids from the largest (HP1a) to the smallest (HP1g) isoform. Notably, the hinge region is subject to post-translational modifications, including phosphorylation, which is thought to cause isoform-specific localization and protein interactions (Badugu, Yoo, Singh, & Kellum, 2005; Koike, Maita, Taira, Ariga, & Iguchi-Ariga, 2000; Smothers & Henikoff, 2001; Zhao, Heyduk, & Eissenberg, 2001).   The chromoshadow (CSD) domain is responsible for the formation of both homo- and HP1 heterodimers, as well as the interaction with a multitude of other proteins with functions including chromatin modification, DNA replication and repair, and nuclear architecture (Brasher, et al., 2000; Cowieson, Partridge, Allshire, & McLaughlin, 2000). Many other proteins known to    15 bind and regulate chromatin structure also contain a CSD, such as the lysine methyltransferase Suv39h1. The CSD domain recognizes a short aa motif, P*V*L (where * is any amino acid), which is found in many HP1-interacting proteins (Thiru, et al., 2004). However, HP1 isoforms have also been shown to interact with proteins lacking this motif (Nozawa, et al., 2010; Smothers & Henikoff, 2000), revealing that there are likely multiple sequence features within HP1 proteins that promote their interactions with other chromatin factors. Indeed, thus far, over 200 HP1-interacting proteins have been identified, a subset of which are discussed below.  1.8 HP1 proteins have varied chromatin-associated functions Due to sequence and domain similarities, HP1 isoforms appear to have many redundant and overlapping functions (reviewed in Kwon & Workman, 2008). However, deletions of individual isoforms in the mouse cause unique phenotypes, indicating the isoforms must also have isoform-specific functions. Whilst HP1a KO mice are viable and fertile (Allan, et al., 2012), HP1b KO mice die from presumed respiratory failure by post-natal day 2 (P2) (Aucott, et al., 2008). HP1g KO males are infertile due to a failure of spermatogenesis, and infertility has also been reported in HP1g KO females (See Table 1.1) (Brown, et al., 2010; Takada, et al., 2011). It is not yet known what confers the isoform-specific functions of HP1 proteins, but their expression patterns may play a role. As seen in Figure 1.3, HP1a and b show similar temporal and spatial expression patterns in germ cell- and pre-implantation development, whereas HP1g is the major isoform expressed following fertilization.    16 Table 1-1 Reported phenotypes upon in vivo deletion of HP1 isoforms in mice  Figure 1-3 Germ cell and early embryonic development expression of HP1 isoforms. Isoforms show similar expression patterns in germ cells, whereas HP1g is the predominant isoform in early development. Data from Dr. Brind’Amour and DBTMEE.hgc.jp. PGC – primordial germ cell. ESC – embryonic stem cell. MEF – mouse embryonic fibroblast.    17 HP1 isoforms have also been reported to have functions beyond heterochromatin formation for transcriptional silencing. For example, in Drosophila, HP1 isoforms act as telomere capping proteins, and HP1 mutant cells show fusion and abnormal chromosome configurations during metaphase (Fanti, Giovinazzo, Berloco, & Pimpinelli, 1998). Smallwood et al. also postulate that HP1 isoforms are involved in recruitment of the splicing machinery, to efficiently process transcribed RNA (2012). In addition, HP1 isoforms has reported functions in DNA replication and repair, by interactions with the origin recognition complex (ORC) and chromatin assembly factor 1 (CAF1) respectively (reviewed in Kwon & Workman, 2008). CAF1-association is believed to help re-establish heterochromatin following the necessary displacement of histones and their histone marks during DNA replication. There is still much to learn about the role of HP1 proteins and the multitude of HP1-interacting proteins in these nuclear pathways. Further, it is of interest to determine whether these activities operate independent of H3K9me3.  1.9 Proteins interacting with HP1 isoforms Another potential explanation for differences in HP1 isoform function is differences in interacting partners, which can aid or prevent specific targeting of HP1 isoforms and/or modify their functions. Several proteomics studies have investigated the interacting partners of HP1 proteins (Nozawa, et al., 2010) and (Hauri, et al., 2016). So far, over 200 distinct proteins have been reported to interact with one, or more, HP1 isoform (Table 1.2). Intriguingly, many of the identified HP1 interactors have reported associations with human disease. E.g. POGZ (Pogo Transposable Element-derived Protein with Zinc Finger Domain) was shown to be associated with White-Sutton Syndrome, a form of ID and autism spectrum disorder (White, et al., 2016; Stessman, et al., 2016). In 2010, Nozawa, et al. showed that HP1a interacts with POGZ in    18 human cell lines (Nozawa, et al., 2010), an interaction reported to promote de-stabilization of HP1a binding to chromosome arms during mitosis, allowing its removal and activation of the kinase Aurora B (AurkB). Nozawa, et al. showed that loss of POGZ resulted in an accumulation of HP1a on chromosome arms, rather than in centromeric regions, and cells failed to properly align chromosomes and undergo normal mitosis.   Ostapcuk, et al. (2018) also identified HP1 isoforms as an integral part of the ChAHP protein complex. This complex can recruit HP1 proteins to specific genomic locations in the absence of H3K9me3. Ostapcuk, et al. showed that the DNA-binding protein ADNP binds HP1g (though it can also bind HP1b and a isoforms, in preferred order) along with CHD4. ADNP recognizes a short sequence motif in the promoter region of specific genes in mESCs, and forms the ChAHP complex to silence transcription of these genes. Loss of either complex subunit, or their interaction, results in ESCs that spontaneously differentiate into the endoderm cell lineage, so HP1-mediated silencing is necessary to prevent premature cellular differentiation. In humans, mutations of ADNP have been implicated as the cause of Helsmoortel – Van der Aa Syndrome, a rare syndrome causing autism spectrum disorders as well as facial dysmorphia and disruption of other organ functions (Helsmoortel, et al., 2014).    19 Table 1-2 Interactors of HP1a. Interactors of HP1a were identified by immunoprecipitation experiments using HP1a protein as bait. Categories denote interacting strength and which region of HP1a the protein interacts with. Adapted from Nozawa, et al. (2010).   Strikingly, the list of proteins interacting with HP1 isoforms contains a large number of proteins with putative DNA-binding domains, like a number of zinc finger proteins (ZFPs). Many are also implicated in human diseases similar to the syndrome resulting from ADNP mutations, such as AHDC1 (AT-hook DNA-binding Motif Containing 1) (Xia et al., 2014), POGZ (Stessman, et al., 2016) and CHAMP1 (Chromosome Alignment-containing Phosphoprotein 1) (Hempel, et al., 2015) (Figure 1.4). We therefore hypothesize that HP1 isoforms can use various sequence-specific DNA-binding proteins as recruiters to genic loci for transcriptional repression, similar to    20 what is seen in the ADNP-HP1g-CHD4 interaction. Using data from several proteomics studies of the HP1 protein interactome as a starting point, in Chapter 4, I investigate the function of two additional HP1-interacting proteins, AHDC1 and CHAMP1. Ahdc1 and Champ1 show similar phenotypes to that of Adnp when mutated in humans, making them potential co-regulators of HP1-mediated gene silencing in mESCs.   1.10 Thesis goals HP1 proteins are known as “readers” of H3K9me3, binding this histone mark and establishing a transcriptionally silent chromatin state. The deposition of H3K9me3 over ERVs in euchromatic regions is attributed to a complex of KRAB-ZFPs, KAP1 and SETDB1, with HP1 recognizing this lysine methylation. However, HP1 isoforms’ role in establishment and maintenance of H3K9me3 at ERVs and over regions flanking these repeats has not been explored in detail. Based on the background presented in Chapter 1, I hypothesize that HP1 isoforms, acting redundantly, are required for transcriptional silencing of specific ERV families in mESCs, and that their presence aid in reinforcing the silencing histone lysine methylation mark over these ERVs. In contrast, HP1 has also been shown to regulate expression of specific genes in mESCs, in an H3K9me3-independent manner. Based on published HP1 protein interactomes, I therefore hypothesize that HP1 proteins use DNA-binding proteins as interchangeable protein complex subunits to facilitate their recruitment to gene promoters in absence of H3K9me3. The general objective of my thesis work was to investigate HP1-mediated transcriptional silencing mechanisms in mESCs.      21 Chapter 2 describes the materials and methods employed in generating the data presented in Chapters 3 & 4, including cell lines used, and culturing conditions. I also adapted and optimized a native Chromatin Immunoprecipitation sequencing (ChIP-seq) protocol, performed mESC growth assays, immunofluorescence cell imaging, and generated KO mESC cell lines through CRISPR/Cas9-mediated gene editing. Chapter 2 also describes the bioinformatic analyses used on newly generated RNA- and ChIP-seq datasets, as well as mined publicly available data.   The aim of Chapter 3 was to show that the three HP1 protein family members act redundantly to silence transcription of specific classes of ERVs. First, I mined publicly available RNA-seq datasets and discovered that, although HP1a, b and g single KO cells showed only minor up-regulation of ERV transcription, HP1 TKO cells showed upregulation of specific Class II ERV families, indicating high functional redundancy between the HP1 isoforms. The upregulated ERV families were direct targets of HP1 protein binding, and followed the H3K9m3-HP1 recruitment dogma, as shown by ChIP-seq data. Secondly, I utilized an inducible HP1 TKO mESC line, generated by our collaborator Dr. Haruhiko Koseki’s lab, to generate H3K9me3 ChIP-seq data, and showed that deposition and maintenance of this mark is not affected by the loss of HP1 proteins. I also confirmed that H4K20me3 deposition, which was known to depend on HP1 protein recruitment to H3K9me3, was disrupted in HP1 TKO cells.  In Chapter 4, I aimed to characterize two out of 200+ reported interactors of HP1 proteins, AHDC1 and CHAMP1, with respect to their HP1-mediated transcriptional silencing potential. I hypothesized that these proteins were responsible for recruiting HP1 isoforms to some gene promoters, in absence of H3K9me3. I characterized the resulting growth phenotype and looked    22 for gross morphological changes following CRISPR/Cas9-mediated gene deletion, and found that neither candidate had any adverse effects on growth or cell morphology. Nor did gene disruption affect either cell lines’ ability to differentiate into embryoid bodies containing all three germ layers. Transcriptome analysis by RNA-seq also confirmed that neither candidate was responsible for HP1-mediated transcriptional silencing.   The experimental results in Chapter 3 and 4 are summarized and discussed in Chapter 5.     23 Chapter 2: Materials and Methods 2.1 Cell lines All cell lines generated in the following experiments originated from an HP1 conditional knock-out cell line kindly gifted to us by Dr. Haruhiko Koseki. These mESCs carry HP1a/b/gfl/fl; Rosa26-Cre/Ert2 transgenes. Cre, which recognizes the LoxP sites flanking all HP1 loci, is sequestered in the cytoplasm and released upon treatment with 4-hydroxytamoxifen (4-OHT) to induce translocation to the nucleus, where it collapses the LoxP sites and deletes all three HP1 genes.   mESCs were cultured in ES media containing serum and 2i, as described in Ostapcuk, et al., (2018). This media contains DMEM High Glucose supplemented with 15% FBS (HyClone Laboratories, Logan, UT, USA), 1 mM L-glutamine, 1 mM non-essential amino acids, 1mM sodium pyruvate, 0.1 mM betamercaptoethnol, 100 U/mL penicillin-streptomycin, recombinant leukemia inhibitory factor (LIF), 3µM GSK (CHIR-99021 HCl, Selleckchem.com, Houston, TX, USA), and 1µM MSK inhibitors (PD0325901 [Mirdametinib], Selleckchem.com, Houston, TX, USA). Cells were plated on gelatinized plates and passaged every 2-3 days.   2.2 Generation of CRISPR KO plasmid Short guide RNA (sgRNA) were designed in silico using the “CRISPR guides” tool on Benchling.com, against the RefSeq annotation of Ahdc1 and Champ1. Two guides were designed per gene, one guide on each side of the coding exon (exon 6 for Ahdc1, exon 2 for Champ1). sgRNA sequences are listed in Table 2.1. The guides were cloned into vectors designed by our collaborator, Dr. Louis Lefebvre, containing the Cas9 endonuclease and puromycin resistance    24 (Figure 2.1). Vectors were checked for sgRNA insertion by Sanger sequencing. HP1 conditional TKO cells (not induced with 4-OHT) were transfected using Lipofectamine 3000 (Invitrogen, Carlsbad, CA, USA) as per the manufacturer’s protocol. Cells were grown under puromycin selection for 48h, then 42 individual colonies picked and seeded at low density. 10 clones of each gene KO were genotyped and Sanger sequenced to verify gene deletion. WT control cells were HP1 conditional TKO cells transfected with vectors lacking the sgRNA sequence, then subjected to the same puromycin selection and colony picking process as CRISPR KO cells.   Table 2-1 Genomic target sequences for CRISRP/Cas9 sgRNAs. Two sgRNAs were used per gene, flanking the ORF of each gene to induce NHEJ between the cut sites and deletion of the interstitial exon. Ahdc1 5’ sgRNA  TGGCAGGTTCTGGTCCCGGC Ahdc1 3’ sgRNA CGAGGTTTTGTTCCTGGGTA Champ1 5’ sgRNA GCTTACTCTCTATAGCACGT Champ1 3’ sgRNA GGAGGGGCCACCATTACATG  Cloning and expression of sgRNA from pPuro2-hU6-gRNA CBh-Cas9Here is the structure of our pPuro2-hU6-gRNA CBh-Cas9 vector and the sequence of the cloning site for new sgRNAs:!The plasmid was originally obtained from Rupesh Amin in Mark Groudine’s lab. It was modified in the Lefebvre Lab by insertion of a PGK-Puro-pA cassette for transient enrichment of transformants.   25 Figure 2-1 Map of CRISPR/Cas9 vector. The vector was generated by Dr. Louis Lefebvre, and contain the Cas9 protein coding sequence, and a puromycin selectable marker.   2.3 Whole-cell protein extraction and western blot Protein extraction for western blot was done by lysing 1 million cells in RIPA buffer (50mM Tris pH 8.0, 150mM NaCl, 1% NP-40, 0.25% deoxycholate, 0.1% SDS), then sonication for 5 minutes, 30 seconds on/30 seconds off. Lysate was centrifuged, and 15µL was boiled for 5 minutes with 5µL NuPAGETM LDS Sample Buffer (4X) (Invitrogen, Carlsbad, CA, USA).   Western blot gels were loaded with 10µL sample, and run time and voltage optimized to fit target protein size (~15 min at 100V, then 45 min at 120V). Proteins were transferred to a PVDF membrane overnight at 30V in transfer buffer (25mM Tris, 190mM glycine, 20% methanol, pH 8.3). Membranes were then incubated in blocking buffer (TBS, 3% BSA) for 30 minutes. 1˚ antibody incubation was done in 5mL blocking buffer for 1h at room temperature. 1˚ antibodies used included Champ1 (AbClonal A11690), HP1a (Cell Signaling #2616), HP1b (Cell Signaling #2613), and HP1g (Cell Signalling #2619). The membrane was washed briefly in TBS-T, then incubated with LI-COR IRDye 2˚ antibodies for 1h at room temperature, washed again, then imaged on an Odyssey Imaging System (LI-COR, Lincoln, NE, USA).  2.4 Immunofluorescence cell imaging Cells were grown on microscope cover slips for 24h, then fixed for 10 minutes at room temperature in 4% paraformaldehyde in PBS, pH 7.4. Following three washes in ice cold PBS, slides were incubated in PBS with 0.2% Triton X-100 for 10 minutes at room temperature, then    26 washed again in ice cold PBS. Blocking was done in blocking buffer (3% BSA, 22.52 mg/mL glycine, 0.1% PBST) for 30 minutes. 1˚ antibody (Novus Bioloicals  NBP1-84238) was diluted in 1% BSA-0.1% PBST and added to slides in a humidified chamber at room temperature for 1h. Following another wash in ice cold PBS, slides were counter-stained in DAPI as well as fluorescent 2˚ antibody was added in 3% BSA-0.1% PBST, and slides were incubated for 1h in a humidified dark chamber. Slides were then mounted on microscope slides and sealed, and imaged.  2.5 Genomic DNA isolation and genotyping  Genomic DNA was extracted using the HotSHOT method (Truett, et al., 2018). 500k cells were lysed and boiled in 75µL HotSHOT I alkaline buffer (25mM NaOH, 0.2mM EDTA) for 30 minutes, then neutralized in HotSHOT II buffer (40mM Tris-HCl pH 5.0). 2µL DNA was used per polymerase chain reaction (PCR) PCR primers against HP1 isoforms were designed by a previous graduate student in the lab. Primers against Ahdc1 and Champ1 were designed to flank each of the sgRNA sites (short PCR product), and for the two primers on the outside of the excised region to be able to produce a PCR product upon CRISPR deletion (long PCR product). PCR products were verified in WT and KO cells by Sanger sequencing and aligning to the expected nucleotide sequence in silico (Figure 2.2).    27  Figure 2-2 Alignment of sequenced PCR results for CRISPR/Cas9-mediated KO validation – A. PCR products from the Ahdc1 Long PCR reaction were gel purified and Sanger sequenced, and aligned to the genomic Ahdc1 locus to show successful gene deletion. B. PCR products from the Champ1 Long PCR reaction were gel purified and Sanger sequenced, and aligned to the genomic Champ1 locus to show successful gene deletion.  2.6 RNA extraction, RT-qPCR, and RNA sequencing RNA was extracted by TRI-reagent by adding 400µL TRI-reagent to 500,000 cells and mixing well, then incubated for 5 minutes at room temperature. 80µL chloroform was added, and the solution mixed and incubated for 3 minutes, then centrifuged at 12,000g at 4˚C for 15 minutes. Aqueous phase was transferred to a new tube, and 200µL isopropanol added to precipitate RNA. The sample was incubated for 10 minutes at room temperature, then spun down for 10 minutes at 12,000g in 4˚C. The pellet was washed briefly in 75% ethanol, then centrifuged at 7,500g at 4˚C for 5 minutes and allowed to dry. RNA was re-suspended in 40µL RNase-free water, incubated at 55˚C for 10 min, then stored at -70˚C until further use.  To generate cDNA, RNA was converted with SuperScript IV reverse transcriptase, as per manufacturers recommendation. RNA was converted by mixing 50µM Oligo d(T)20 primer, 10µM dNTP mix and RNA sample in DEPC water. Samples were mixed and centrifuged briefly,    28 then incubated at 65˚C for 5 minutes. 5X SSIV buffer, 100mM DTT, Ribolock RNase inhibitor (Thermo Scientific, Waltham, MA, USA) and SuperScript IV reverse transcriptase (Invitrogen, Carlsbad, CA, USA) was mixed and added to the annealed RNA template, then incubated at 53˚C for 10 minutes and 80˚C for 10 minutes. qPCR experiments were done using primers of interest and primers to the Ppia housekeeping gene, with an EvaGreen dye (Biotium) on a QuantStudio 3 Real-Time PCR system (Thermo Fischer Scientific, Waltham, MA, USA). Relative expression was calculated by normalization to Ppia expression.   RNA samples were submitted to the UBC BRC sequencing core for sequencing and performed as per their standard protocol. Sample quality was verified on the Agilent 2100 Bioanalyzer, and prepped following the standard protocol for the NEBnext Ultra II Stranded mRNA (New England Biolabs, Ipswich, MA, USA). Sequencing was done on the Illumina NextSeq 500 with paired-end 42bp x 42bp reads.   2.7 Native Chromatin immunoprecipitation (ChIP)  ChIP-seq was performed using an ultra-low input (ULI-) ChIP sequencing protocol (Brind’Amour, et al., 2015), adapted for a larger cell number. Cells were harvested at 70% confluency, and aliquoted into 500,000 cell/sample. Each aliquot was flash frozen in liquid nitrogen, and stored at -80˚C until further use. Each sample (HP1 TKO and HP1 cTKO [WT]) was thawed and diluted in 20µL 1X PBS. For each sample, 200k cells (8µL) in duplicate was lysed in 50µL EZ Nuclear lysis buffer (LB) (Sigma-Aldrich, St. Louis, MO, USA) + 1X proteinase inhibitor cocktail (PIC) (Sigma-Aldrich, St. Louis, MO, USA) and spun down at top speed for 15 seconds. All but 5µL of the buffer was removed, the pellet resuspended in 20µL of    29 fresh LB+PIC. DNA was digested by addition of 20µL MNase master mix (76.5µL ddH2O, 20µL 10X MNase buffer, 1.85µL 200mM DTT, 2µL MNase [New England Biolabs, Ipswich, MA, USA]) and incubated at 37˚C for 8 min. Digestion was stopped by addition of 4µL 200mM EDTA and 410µL ChIP buffer (20mM Tris-HCl pH 8.0, 2mM EDTA pH 8.0, 150mM NaCl, 0.1% Triton X-100, 1mM PMSF, 1X PIC). Samples were sonicated on low for 15 seconds to increase solubility, and topped up with 100µL ChIP buffer. Pre-clearing was done using a mix of Protein A and G Dynabeads (Life Technologies) for 2.5 hours at 4˚C with rotation. Bead-Antibody complexes were generated by washing 150µL Protein A + 150µL Protein G Dynabeads in 600µL ChIP buffer, then resuspended in 300µL fresh ChIP buffer. 50µL of bead mix was mixed with 5µL anti-H3K9me3 antibody (Active Motif 39161) or 10µL anti-H4K20me3 antibody (Active Motif 39180) and 1000µL ChIP buffer, then aliquoted 210µL per reaction condition, and incubated for 3h at 4˚C with rotation. 220µL of precleared chromatin was added to each of the bead-antibody complex samples, as well as 22µL set aside as input samples. Chromatin + bead-antibody complex was incubated over-night at 4˚C with rotation. The remaining chromatin was checked for proper MNase digestion on an Agilent TapeStation. Following over-night incubation, bead-antibody complex samples were washed twice in 200µL low-salt wash buffer (20mM Tris-HCl pH 8.0, 2mM EDTA pH 8.0, 150mM NaCl, 1% Triton X-100, 0.1% SDS, 1X PIC), followed by two washes with 200µL high-salt wash buffer (20mM Tris-HCl pH 8.0, 2mM EDTA pH 8.0, 500mM NaCl, 1% Triton X-100, 0.1% SDS, 1X PIC). Protein-DNA complexes were eluted from the magnetic beads by incubation in 25µL elution buffer (100mM NaHCO3, 1% SDS) and Proteinase K at 55˚C for 20 minutes, then 65˚C for 90 minutes. Prior to library preparation, samples were cleaned up using AMPure XP beads (Beckman Coulter, Indianapolis, IN, USA).    30 2.8 Construction of ChIP sequencing libraries and bioinformatic analysis Library construction and data processing was performed as by our lab’s standard protocol, published by (Brind’Amour, et al., 2015). ChIP and Input DNA was end repaired and A-tailed using NEBNext UltraTM II End Repair/dA-Tailing Module (New England Biolabs, Ipswich, MA, USA). Adaptors were ligated with Quick T4 DNA ligase (New England Biolabs, Ipswich, MA, USA). Following adaptor ligation, the libraries were amplified by 14 PCR cycles and size selected (200-700bp) on a 2% EX Gel (Life Technologies, Thermo Fischer Scientific, Waltham, MA, USA). Amplified libraries were quantified by QuBit and Agilent Tape station, and sequenced by the UBC BRC sequencing core, on an Illumina Hi-seq 2000 as per the recommended protocol. Sequenced reads were aligned to the mm10 genome using the Burrows-Wheeler Aligner (BWA) with default parameters, excluding multi-mapped reads and PCR duplicates from downstream calculations. Genome coverage counts were calculated from BED files, using a minimum MAPQ of > 5. All subsequent calculation and analyses were done using VisR software (Younesy, Möller, Lorincz, Karimi, & Jones, 2015).  RNA samples were submitted to the UBC BRC sequencing core for sequencing. Sample quality was verified on the Agilent 2100 Bioanalyzer, and prepped following the standard protocol for the NEBnext Ultra II Stranded mRNA (New England Biolabs, Ipswich, MA, USA). Sequencing was done on the Illumina NextSeq 500 with paired-end 42bp x 42bp reads. Sequenced reads were aligned to mm10 using the STAR aligner with default parameters, and PCR duplicates and multi-mapped reads removed from further analysis.      31 Transcription levels and histone mark enrichment was quantified by calculating Reads Per Kilobase per Million mapped reads (RPKM), per the formula  𝑅𝑃𝐾𝑀% = 𝑛 𝐿×1,000,000 𝑁%  (n= number of reads aligned to region, L= region length in kilobases, Nx= total number of aligned reads). To compare two samples, I calculated a Log2(fold change)  log1[ 𝑅𝑃𝐾𝑀3 + 1056) (𝑅𝑃𝐾𝑀9 + 1056 ],  where RPKMa and RPKMb are values in a region of interest in sample A and B, and a Log10(mean expression) log;<[ 𝑅𝑃𝐾𝑀3 + 𝑅𝑃𝐾𝑀9 2 + 1056].  Samples were also compared by calculating a Z-score 𝑍 = (𝑅𝑃𝐾𝑀3 − 𝑅𝑃𝐾𝑀9 + 1056) (𝑅𝑃𝐾𝑀3 + 𝑅𝑃𝐾𝑀9 + 1056).  Publicly available datasets were downloaded and processed in the same way as self-generated data. I utilized RNA- and ChIP-seq data from Ostapcuk, et al. (2018), available in the GEO database under accession GSE97045.  2.9 Primer sequences Dnmt3b-qRT-F  GTT AAT GGG AAC TTC AGT GAC CA Dnmt3b-qRT-R  CTG CGT GTA ATT CAG AAG GCT Hoxa1-qRT-F   AGC CAC CAA GAA GCC TGT CGT T Hoxa1-qRT-R   TTG ACC CAC GTA GCC GTA CTC T Nes-qRT-F   CTG CAG GCC ACT GAA AAG TT Nes-qRT-R   TTC CAG GAT CTG AGC GAT CT     32 Hand1-qRT-F   TCA AAA AGA CGG ATG GTG GT Hand1-qRT-R   GCG CCC TTT AAT CCT CTT CT Gata4-qRT-F   CTG GAA GAC ACC CCA ATC TC Gata4-qRT-R   CAC AGG CAT TGC ACA GGT AG HP1a-cKO-F   TTT GTC TCT CCC ATG AAT AGA HP1a-R   CAT ACA TGC ACA TAC ACA TAC TAA CAA AT HP1a-KO-F   GTA CTC TCT GTG GTG GAG CAG HP1b-cKO-F   GCG CAA AGC TGA TTC TGA TTC HP1b-R   TAT AAG AGC AAG CCC CAA ATC  HP1b-KO-R   TGT TCG AGA ATA GCG GAG TGG HP1g-cKO-F   TCA GTG CTT CAT TTC TTA CAG  HP1g-cKO-R   CGT AAA ACG TAA TGC TCT TCT HP1g-KO-F   ACT CAG CTC TCC ATT GTT GTC HP1g-KO-R   AGT GGT ATT TGC GGT ATT ATC Ahdc1-del2-F    GGG TCT TCC CAA CCT TCT TC Ahdc1-del2-R   TAC TTG GGC TCT CGG AGG TA Ahdc1-short del2-Rev  CCA GCC CTG TTC AGT TTT TC Ahdc1-short del2-F  GCT GCC CAC TCC TTA GTG AC Champ1-short del1-F  CTG CAT GGT AAG GAG CAG GT Champ1-short del1-R  AAG GGC CTT TCT GAG CCT AC Champ1-short del2-R  CGC CAT TGC AAA CAA CTA GA Ppia-qRT-F   CGC GTC TCC TTC GAG CTG TTT G Ppia-qRT-R   TGT AAA GTC ACC ACC CTG GCA CAT    33 Chapter 3: HP1 proteins are indispensable for ERV silencing 3.1 Background Due to their potential for causing insertional mutagenesis, and influencing gene expression regulation, TEs are a threat to genome integrity and their transcription must therefore be tightly regulated. ERVs are a type of TE stemming from insertions of viral genomes into the host, and in mice there are three classes of ERVs based on their reverse transcriptase gene or their similarity to exogenous retroviruses. Although most copies of ERVs are incapable of transcription and/or retrotransposition due to the accumulation of inactivating mutations over time, numerous ERVs in the mouse genome retain their ability to initiate transcription and transpose, and are therefore considered active TEs that must be silenced to protect genome integrity (reviewed in (Gagnier, et al., 2019)). Indeed, as many as 10-12% of all germline mutations in mice are due to ERV insertions, of which 50% are insertions of the Class II Intracisternal A-particle (IAP) group (Maksakova, et al., 2006). Interestingly, the majority of ERV copies exist as solo LTRs, where the flanking terminal repeats have undergone homologous recombination, causing deletion of the interstitial retroviral genes (Belshaw, et al., 2007; Nellåker, et al., 2012). The LTR region contains the regulatory motifs and transcription initiation sites for the ERV, so solo LTRs can remain transcriptionally active and act as transcription start sites (TSSs) or putative enhancers and regulatory sequences, and therefore need to be epigenetically silenced. Such silencing is achieved, in large part, by H3K9me3 and DNAme, depending on the tissue type (reviewed in (Leung & Lorincz, 2012)). In mESCs, H3K9me3 is required for silencing of Class I and II ERVs, and is deposited by the HMT SETDB1, as shown by deletion of Setdb1 in mESCs causing massive up-regulation of, in particular, two specific Class II ERV families; IAPs and    34 ETn/MusDs (Matsui, et al., 2010). In contrast, Class III ERVs lack H3K9me3 and are silenced mainly by the lysine demethylase KDM1A/LSD1 (Macfarlan, et al., 2011).  Although H3K9me3 is closely associated with transcriptional silencing, especially of repetitive DNA sequences and ERVs, H3K9 methylation is not the main cause of silencing in and of itself. Rather, these PTM, like other histone-methyl marks, recruit chromatin-modifying proteins to compact and silence the methylated region. The most widely studied of the H3K9me3 “readers” are the HP1 family of proteins. The dogma for HP1-mediated transcriptional silencing is via binding to H3K9me3, as shown by Bannister, et al., (2001) and Lachner, et al., (2001), and bridging adjacent nucleosomes to compact the chromatin structure (Machida, et al., 2018). The ability of HP1 proteins to silence transcription has been shown in various models. In Schizosaccharomyces pombe, the HP1 homologues CHP2 and SWI6 are both required for formation of a transcriptionally silent chromatin state (Sadaie, et al., 2008). In Drosophila, mutating HP1a or HP1c causes up-regulation of two families of transposons (Kwon, et al., 2010). And in Neurospora, DNA methylation and silencing of transposable elements depend on HP1 binding to H3K9me3 (Tamaru & Selker, 2001), showing the conserved involvement of HP1 proteins in transcriptional silencing across species.   However, previous studies in the Lorincz lab investigated the impact of individual depletion of two HP1 isoforms, HP1a and HP1b, on ERV transcription in mESCs (Maksakova, et al., 2011). This work revealed only a minimal effect on ERV silencing, indicating that HP1 proteins individually are not critical players in ERV silencing. Although single HP1 isoforms were not found to affect ERV transcription (Maksakova, et al., 2011), the sequence similarities between    35 each HP1 isoform indicates a high level of redundancy in function between all three isoforms. Additionally, the redundancy of HP1 protein isoforms in silencing of genic transcription further suggests that one remaining HP1 isoform may functionally compensate for the loss of two other HP1 homologues in transcriptional silencing of ERVs (Ostapcuk, et al., 2018). I therefore chose to further investigate the role of HP1 proteins in ERV silencing in mESCs, taking advantage of a combinatorial HP1 mutant line.   Although HP1 proteins are “readers” of H3K9me3, placing them downstream of HMTs and the methyl mark in the silencing pathway, proteomics work has also revealed direct interactions between HP1 isoforms and the three HMTs responsible for writing H3K9me3: G9A/GLP, SUV39h1/2 and SETDB1 (Hauri, et al., 2016; Nozawa, et al., 2010). Further, tethering of HP1 to pericentromeric heterochromatin in cells lacking H3K9me3 in such regions due to deletion of SUV39h1/2 promoted re-establishment of this mark (Kourmouli, et al., 2005). HP1 proteins might therefore be involved in a feedback loop, recruiting further HMTs to sites of H3K9me3 to help maintain, and possibly spread this histone mark (Kourmouli, et al., 2005). Paradoxically, HP1 proteins have also been reported to be antagonistic to H3K9me3 deposition. In fission yeast, an orthologue of HP1 has been shown to constrain H3K9me3 spreading of H3K9me3 across natural heterochromatin borders (Stunnenberg, et al., 2015). It is still unclear whether such mechanisms operate in pluripotent mammalian cells, highlighting the need for further understanding of the role and redundancy of HP1 proteins in silencing of ERVs, and the impact of HP1 proteins on H3K9me3 dynamics.      36 Furthermore, deposition of H4K20me3 genome-wide was shown to depend on H3K9me3 and HP1 proteins (Schotta, et al., 2004). Schotta, et al. showed that HP1 proteins bind H3K9me3 to recruit the HMT SUV420h1/2, which “writes” the H4K20me3 mark. H4K20me3 is also a silencing mark, shown to be antagonistic to the active histone PTM H4K16 acetylation (Nishioka, et al., 2002). I therefore also wanted to investigate whether loss of HP1 proteins affects H3K9me3 or H4K20me3 levels over ERVs in mESCs, either by aiding in its deposition and maintenance, or by regulating its spreading.   3.2 Results 3.2.1 Loss of HP1 isoforms cause abnormal mESC growth To investigate the effect of loss of HP1 proteins on overall cell morphology and growth, I utilized an inducible HP1 TKO cell line generated by our collaborator Dr. Haruhiko Koseki. This cell line carries LoxP sites, a 34bp nucleotide sequence recognized by the bacteriophage-derived protein Cre, flanking both loci of all three HP1 genes (see chapter 2 for details). Upon treatment with 4-OHT, Cre fusion proteins sequestered in the cytoplasm translocate to the nucleus where Cre causes recombination between the two LoxP sites and excision of the DNA sequence in between (Feil, et al., 1996) (Zhang, et al., 1996). I first verified the protocol established by Dr. Koseki’s lab for excision of all three HP1 isoforms by PCR. Induction of HP1 gene excision by 800nm 4-OHT for 48h was sufficient to achieve complete loss of all three HP1 genes and protein (Figure 3.1 A & B). Upon HP1 protein depletion, mESCs showed signs of impeded cell division and growth rate, and increased cell death, as seen by brightfield microscopy of the cell cultures (Figure 3.1 C). In one experiment, the TKO cells failed to re-attach to the culture dish following passaging. Overall, the HP1 TKO cells showed reduced growth rates and a high incidence of cell    37 death, although the cells could be propagated for up to one week after withdrawal of 4-OHT. Since mESCs lacking one or two HP1 isoforms do not show such severe effects on cell growth (Maksakova, et al., 2011), the presence of at least one HP1 isoform is sufficient and necessary for proper ESC growth and proliferation, indicating a high level of redundancy in function between the HP1 proteins.       38 Figure 3-1 Validation of deletion of HP1 genes in mESCs – A. Gel image of PCR samples from a time course of 4-OHT treated cells shows that 48h of 4-OHT treatment is sufficient for homozygous deletion of all three HP1 genes. Only results for the HP1a locus are shown. PCR primers were previously designed to distinguish loci with and without LoxP sites. KO primers only amplified a product upon successful Cre-mediated excision. B. Western blot of protein extracts from HP1 TKO cells treated with 4-OHT for 48h showed complete depletion of HP1 proteins. C. Brightfield microscopy images of HP1 TKO mESCs following 48h of 4-OHT diluted in 100% EtOH showed abnormal cell growth and colony morphology, and increased cell death rates compared to WT cells treated with an equal volume 100% EtOH.   3.2.2 HP1 proteins act redundantly to silence ERVs in mESCs To investigate whether HP1 proteins can functionally replace each other to enforce transcriptional silencing of ERVs, I mined publicly available RNA-seq data from HP1a KO, HP1b KO, HP1g KO and an independently generated HP1 TKO line (Ostapcuk, et al., 2018), and compared ERV expression in the single versus TKO lines. Since we hypothesized HP1-mediated silencing is a major regulator of ERV transcription on a genome-wide scale, I assessed the expression of aggregated ERV families, rather than individual elements. I also limited the analysis to ERV families with >100 copies in the mouse genome, since families with fewer copies genome-wide are more sensitive to outliers artificially inflating the magnitude of transcription level change. Expression of ERVs was not significantly increased in HP1a or b KO cells, as previously reported by Maksakova, et al. (2011), nor in HP1g KOs (Figure 3.2 A, B & C). However, specific ERV families, in particular Class II and Class III families, were up-regulated upon deletion of all three HP1 isoforms in conjunction (Figure 3.2 D). This shows that HP1 isoforms do indeed silence transcription of specific ERV families in mESCs, and that the isoforms play a redundant role in such silencing.    39  Figure 3-2 Expression of ERV families in HP1 protein-depleted mESCs – A. Expression analysis of RNA-seq data from HP1a KO cells. Each data point represents the average expression of one ERV family. B. Expression analysis of RNA-seq data from HP1b KO cells. C. Expression analysis of RNA-seq data from HP1g KO cells. D. Expression analysis of RNA-seq data from HP1 TKO cells. Deletion of individual HP1 isoforms caused only minimal up-regulation of ERV families. In contrast, expression of ERV families such as the Class II IAP and RLTR families was significantly up-regulated in absence of all three HP1 isoforms in conjunction.     40 To further link HP1 isoforms to the observed transcriptional up-regulation of ERV families, I utilized published ChIP-seq data from mESCs of endogenously FLAG-tagged HP1 isoforms as well as H3K9me3 (Ostapcuk, et al., 2018). As expected, the up-regulated MERVK10C, IAP and RLTR families show strong enrichment for H3K9me3 and all three HP1 isoforms (Figure 3.3), indicating that they are likely direct targets of HP1-mediated transcriptional silencing. These ERV families are also significantly up-regulated in SETDB1 KO ESCs (Karimi, et al., 2011; Maksakova, et al., 2011), indicating that H3K9me3 is required for their silencing. Thus, transcriptional regulation of specific Class II ERV families is HP1-mediated, requiring H3K9me3 and at least one isoform of HP1 for successful silencing. Paradoxically, the most highly up-regulated families, including Class III MT2_Mm/MERVL and MER89 families, show relatively low levels of enrichment of both H3K9me3 and HP1 isoforms (Figure 3.3). Their dysregulation is therefore likely due to indirect effects rather than resulting from failure of HP1-mediated transcriptional silencing in absence of HP1 proteins.   3.2.3 H3K9me3 levels are not affected by loss of HP1 proteins To study the impact of HP1 proteins on H3K9me3 levels in mESCs, I generated ChIP-seq data of H3K9me3 in HP1 WT and TKO cells. I investigated the H4K20me3 mark, since H4K20me3 deposition is reported to be dependent on HP1 proteins binding H3K9me3 and recruiting the H4K20 methyltransferases SUV420h1/2 (Schotta, et al., 2004). HP1 TKO cells were harvested following 48h of induction in 800nM 4-OHT in 100% EtOH. HP1 WT cells were treated with an equal volume 100% EtOH for the same amount of time. ChIP was done using a modified version of the ULI-ChIP protocol (Brind’Amour, et al., 2015), with antibodies specific for H3K9me3 (ActiveMotif 39161) and H4K20me3 (AcitveMotif 39180).    41  Figure 3-3 RNA- and ChIP-seq analysis of ERV families in HP1 TKO cells – A. Expression analysis of aggregated ERV families in HP1 TKO coloured by the level of H3K9me3 ChIP-seq enrichment. Each data point represents the average expression of one ERV family. Most up-regulated ERV families are enriched for the H3K9 methyl mark, except the Class III families MERVL, MER89 and MT2_Mm. H3K9me3 RPKM was calculated as ChIP signal – input signal, and the colouring was clamped at 3 RPKM to reveal differences in H3K9 methylation on a smaller scale. RNA-seq and ChIP-seq data from Ostapcuk, et al. (2018). B. Expression analysis of ERV families in HP1 TKO cells, coloured by the ChIP-seq enrichment of any HP1 isoform (value of highest isoform shown). Up-regulated families are modestly enriched for HP1 isoforms. HP1 RPKM was calculated as ChIP signal – input signal, and the colouring was clamped at 1 RPKM. RNA-seq and ChIP-seq data from Ostapcuk, et al. (2018).   As expected, H4K20me3 deposition required the presence of at least one HP1 isoform, as shown by the significant loss of H4K20me3 genome-wide in the HP1 TKO cell line (Figure 3.4 A). A similar reduction in H4K20me3 was also observed over genes (+/- 1kb of the TSS), and over aggregated ERV families (Figure 3.4 B & C).     42 In contrast, ChIP-seq analysis revealed only modest changes of H3K9me3 following HP1 protein ablation, with a slight trend towards loss of H3K9me3 observed in 2kb genome-wide bins (Figure 3.5 A). Out of ~59,000 2kb bins on a randomly selected chromosome, approximately 3,000 bins showed a loss of H3K9me3 after filtering steps, compared to 1500 gaining H3K9me3. These results suggest that HP1 proteins may perform different functions in distinct genomic regions, helping to reinforce the methyl mark in some regions, while preventing its aberrant spreading in others. However, when investigating H3K9 methylation over gene promoters, or over aggregated ERVs families, no significant loss of H3K9me3 was observed (Figure 3.5 B & C). Additionally, the scale of change of H3K9me3 in genome-wide bins was minimal in comparison to that of H4K20me3. It is therefore unlikely that HP1 proteins play a significant role in regulating H3K9me3 levels over genes or ERVs for the purpose of transcriptional regulation. The loss of the H4K20me3 silencing mark over genes and ERVs suggests some increase in transcription of these regions. However, the genes and ERV families showing loss of H4K20me3 do not overlap with the regions showing increased transcription in the HP1 TKOs cells, suggesting that H4K20me3 is just one layer of a set of redundant transcriptional silencing mechanisms acting on these regions.       43  Figure 3-4 Differential ChIP-seq analysis of H4K20me3 in HP1 WT and TKO cells - A. Change in ChIP-seq enrichment of H4K20me3 over 2kb genomic bins in HP1 WT and TKO cells. Bins on a randomly chosen chromosome are shown (Chr 14). B. Change in H4K20me3 ChIP-seq enrichment over all TSS +/- 1kb. C. Change in H4K20me3 ChIP-seq enrichment over aggregated ERV families also show reduced levels of H4K20me3 upon loss of all three HP1 isoforms. D. Transcriptional and differential ChIP-seq analysis of aggregated ERV families in HP1 TKO and WT cells. Dot colour corresponds to change in enrichment (z-score) of H4K20me3 upon HP1 protein    44 depletion. Darker dots have the greatest loss of H4K20me3, and show that ERV families up-regulated in HP1 TKO cells do not see a significant loss of H4K20me3 upon HP1 protein depletion.    Figure 3-5 Differential ChIP-seq analysis of H3K9me3 in HP1 WT and TKO cells – A. Change in H3K9me3 ChIP-seq enrichment over 2kb genome-wide bins upon deletion of all HP1 proteins in mESCs. Graph shows bins on a randomly selected chromosome. There was no clear trend towards loss or gain of H3K9me3 over 2kb bins. B.    45 Change in H3K9me3 levels over gene promoters (TSS +/- 1kb) following loss of HP1 proteins. Only 4 and 11 genes show a z-score of above or below 0.75 and -0.75 respectively. C. Change in H3K9me3 ChIP-seq enrichment in HP1 TKO versus WT cells over aggregated ERV families. HP1 TKO cause only minimal change in H3K9me3 enrichment, with no families reaching a z-score above or below 0.75 or -0.75. D. Transcriptional and differential ChIP-seq analysis of H3K9me3 over aggregated ERV families in HP1 TKO and WT cells. Dot colour corresponds to change in enrichment (z-score) of H3K9me3 upon HP1 protein depletion. Darker pink dots show a gain in H3K9me3-enrichment upon HP1 protein depletion.  3.3 Discussion The results presented here show that the presence of at least one functional HP1 isoform is essential for growth and transcriptional regulation in mESCs. Upon loss of all three HP1 proteins in conjunction, mESCs showed abnormal cell growth and colony morphologies, and increased cell death. It is likely that this observed phenotype is due to an amalgamation of failed chromatin-associated processes, considering the many reported mechanisms where HP1 proteins are involved. For example, loss of HP1 proteins could cause abnormal cell growth through increased expression of lineage-specifying genes such as Gata4 or Sox17 (Ostapcuk, et al., 2018). Similarly, HP1 proteins are essential for pericentromeric heterochromatin formation, which have important structural roles in chromosome segregation during the cell cycle as well as in silencing of transcription from repetitive DNA (Peters, et al., 2001). Future work should investigate cell differentiation and chromosome segregation during the cell cycle to fully characterize the HP1 TKO mESC phenotype, for example by checking the ploidy of HP1 TKO cells at high passaging numbers. Understanding the full complexity of the HP1 TKO phenotype will eventually aid in teasing apart the many distinct and overlapping functions of the HP1 protein isoforms.    46 Contrary to previous reports that individual HP1 isoforms were dispensable for silencing of ERVs (Maksakova, et al., 2011), the work presented here showed the need for at least one HP1 isoform for proper ERV silencing. Specifically, certain Class II ERV families are subject to HP1-mediated transcriptional silencing, and are direct targets of HP1 protein binding via the H3K9me3 histone PTM. Only upon loss of all three HP1 paralogues is there a significant loss in ERV silencing, further establishing that HP1 proteins have high functional redundancy between the three isoforms. However, although the magnitude of the ERV up-regulation in the HP1 TKO cells is significant, the effects seen can only be said to be modest in comparison to the degree of ERV dysregulation observed upon loss of H3K9me3 following deletion of  Setdb1 (Maksakova, et al., 2011; Matsui, et al., 2010). Whereas HP1 isoforms are reported to bind exclusively to the histone PTM H3K9me3, there are several other H3K9me3 readers in the genome, including CBX2, CBX7, CDYL2, and MPP8 (Bernstein, et al., 2006; Fischle, Franz, Jacobs, Allis, & Khorasanizadeh, 2008; Kokura, Sun, Bedford, & Fang, 2010). The relatively higher dysregulation of ERVs seen in the SETDB1 KOs compared to the HP1 TKOs may therefore be explained by H3K9me3 being involved in additional silencing pathways independent of HP1 proteins. Investigations of the roles of the remaining H3K9me3 readers in transcriptional silencing are necessary to further understand what causes the difference in dysregulation of ERVs between Setdb1 KOs and HP1 TKOs. Maksakova, et al. (2011) also proposed that H3K9me3 may be antagonistic to binding of the transcription initiation machinery or prohibitive of the establishment of other epigenetic marks needed for transcriptional activation, so further investigations could focus on the inherent transcription-preventing properties of H3K9me3 itself.     47 Although several studies report an interaction between HP1 proteins and HMT “writers” of H3K9me3, I find no significant effect of loss of HP1 proteins on H3K9me3 enrichment levels genome-wide. Contrary to the proposed hypothesis (Kourmouli, et al., 2005; Maison & Almouzni, 2004), there appears to be no feedback loop of HP1 proteins binding H3K9me3 and recruiting HMTs to maintain and help spread this histone mark in mESCs. However, it is still possible that H3K9me3 levels change in regions not investigated here, such as over pericentromeric heterochromatin boundaries, as reported in fission yeast (Stunnenberg, et al., 2015). It is also possible that loss of H3K9me3 occurs at individual ERVs, but this change is masked when analyzing high throughput sequencing data at the level of aggregated ERV families. An investigation of smaller scale regions, as opposed to the 2kb genomic bins investigated here, and of individual ERV elements should be conducted before concluding HP1 proteins have no impact on H3K9me3 deposition and maintenance.  RNA- and ChIP-seq data revealed that the most highly dysregulated ERV families in the HP1 TKO cell line are not highly enriched for HP1 proteins, nor H3K9me3. Their up-regulation is therefore likely due to indirect effects of HP1 protein depletion. One possible explanation is that loss of HP1 skews the ratio of 2 cell-like cells to inner cell mass-like cells in the “mESC” pool. 2 cell-like cells were first identified as a transient cell population in mESC cultures by Macfarlan, et al. (2012). Macfarlan, et al. showed that some mESCs lack expression of the blastocyst cell markers Oct4, Sox2 and Nanog, and have increased expression of 2 cell-specific transcripts. These cells also show activation of MERVL/MT2_Mm ERV families, as seen in the HP1 TKO cell lines. However, Oct4 and Sox2 are not dysregulated in the HP1 TKO cells (z-score of 0.6 and -0.1 respectively), though Nanog expression is significantly decreased (z-score of -3.3).    48 Single-cell RNA-seq analysis, or the generation of a MERVL-reporter line, would help to further characterize any heterogeneity in cellular differentiation states in the HP1 TKO cells, and would confirm or refute whether the increase in MERVL/MT2_Mm expression is due to moderate up-regulation in all cells or a dramatic up-regulation restricted to a small number of 2 cell-like cells.   Altogether, the results reported in this chapter help refine our understanding of HP1 protein function and their relationship with the H3K9me3 histone mark, highlighting the high level of redundancy in function between HP1 isoforms, and providing further evidence for the importance of HP1 proteins in protecting the genomic integrity of embryonic stem cells from the potentially detrimental effect of ERV transcription.    49 Chapter 4: HP1 proteins and their interactors in gene regulation  4.1 Background HP1 proteins are widely recognized as hallmarks of transcriptionally silent heterochromatin. Founded on numerous studies in model organisms, the current dogma of HP1 protein function implicates an essential role for recruitment to the histone PTM H3K9me3, where binding causes compaction of the chromatin structure and in turn prevents transcription. However, as explored in the previous chapters, the last decade of research has revealed that this is not the full story. For example, HP1 proteins have functions in DNA replication via interaction with ORC (Auth, Kunkel, & Grummt, 2006), and postulated involvement in splicing and processing of transcribed RNA (Smallwood, et al., 2012). More recently, Ostapcuk, et al. (2018) showed that HP1g can be recruited to gene promoters by binding the sequence-specific DNA-binding protein ADNP (further discussed below), in the absence of H3K9me3. This novel mechanism for HP1 protein recruitment to DNA suggests that there is still more to be learned about how HP1 isoforms function, and specifically how they are targeted to chromatin to exert their silencing function independent of H3K9me3.   Protein recruitment to chromatin can be achieved in several ways: via binding directly to histone PTMs, e.g. HP1 protein binding to H3K9me3; by direct, sequence-specific binding to DNA motifs; or by indirect recruitment via interactions with other DNA- or histone PTM-binding proteins. To date, HP1 proteins have not been shown to bind directly to DNA, and do not contain any recognized DNA-binding domains. Furthermore, in vitro studies of HP1 isoforms reveal that they bind no histone PTMs other than H3K9 methylation (Bannister, et al., 2001). The most likely mechanism for HP1 recruitment in the absence of H3K9me3 is therefore through    50 interaction with other chromatin-interacting proteins, either DNA-binding factors or histone PTM-binding proteins. Of note, proteomics studies have identified more than 200 proteins that interact with one or more of the HP1 isoforms, many of which have putative DNA-binding domains (Hauri, et al., 2016; Nozawa, et al., 2010; Ostapcuk, et al., 2018) (Figure 4.1). Nozawa and colleagues classified putative HP1a interactors based on their binding mechanism (Nozawa, et al., 2010). Most interactors contain P*V*L motifs (where * is any amino acid), which are recognized by the CSD of HP1. In contrast, associations of some HP1a interactors are disrupted following mutation of the H3K9me3-binding CHD. Still other proteins require both the CSD and the hinge region for their stable protein interaction with HP1 proteins, or do not contain the P*V*L motif but lose HP1 interaction upon CSD disruption alone.     51 Figure 4-1 Interactomes of HP1 proteins - A. Nozawa, et al. (2010) identified and classified interactors of HP1a, based on the which parts of HP1a they appear to interact with. Class I proteins lose their HP1 interaction upon disruption of the CSD. Class II proteins are sensitive to disruption of the H3K9me3-binding CD. Class III proteins require both the CSD and the hinge region of HP1a for their proper HP1 interaction, whilst Class IV proteins bind the CSD but do not have P*V*L motifs. B. Hauri, et al. (2016) identified interactors of all three HP1 isoforms. Only a randomly selected subset of HP1 interactors are shown, including a number of putative and known DNA-binding proteins. Proteins in blue share no identifiable traits, whilst proteins in yellow are known zinc finger-containing proteins. Genes chosen for CRISPR/Cas9-mediated deletion are highlighted in both panels.  Notably, proteins containing zinc fingers (ZFPs in mice, ZNFs in human) are enriched in the class of HP1 interactors that contain P*V*L motifs. ZFPs are putative DNA-binding proteins, and were first discovered and characterized by Diakun, et al. (1986). Zinc finger domains are also a common component of DNA-binding domains of many eukaryotic transcription factors (Porteus & Carroll, 2005). The most common type of zinc fingers are strings of 30 amino acids (aa) containing a Cys2/His2 (C2H2) motif that coordinates one zinc molecule, which aids in binding to three distinct nucleotides in the target DNA (Porteus & Carroll, 2005). Considering the numerous ZFPs and other DNA-binding HP1-interacting proteins identified by (Hauri, et al., 2016; Nozawa, et al., 2010), we hypothesize that HP1 proteins are recruited to DNA by specific DNA-binding interactors in a sequence-specific manner, and in turn regulate transcription in cis.   One HP1-interacting protein, the P*V*L motif-containing ADNP, was studied in depth by Dr. Mark Bühler and colleagues for its role in transcriptional regulation (Ostapcuk, et al., 2018). Mutation of ADNP has been implicated as a cause of the neurodevelopmental syndrome Helsmoortel – van der Aa Syndrome, with phenotypes including intellectual disability, autism    52 spectrum disorder, facial dysmorphias, and failure to thrive (Helsmoortel, et al., 2014). Ostapcuk and colleagues went on to show that ADNP recognizes DNA sequence motifs in some gene promoters in mESCs, and in turn recruits HP1g as well as the chromatin remodeller CHD4 in a complex the authors named ChAHP (Ostapcuk, et al., 2018). ADNP is responsible for this complex binding to gene promoters. This recruitment promotes formation of a condensed region specifically over the proximal gene promoter, inhibiting transcription. ChAHP appears to silence the expression of genes important in the endodermal cell lineage, and cells spontaneously differentiate into the endodermal lineage upon loss of ADNP in ESCs. Presumably, because these cells show aberrant expression of endoderm lineage markers, they cannot be induced towards a neural progenitor cell lineage, providing a potential explanation for the neurodevelopmental phenotype observed in patients with ADNP mutations (Ostapcuk, et al., 2018). The authors also investigated gene dysregulation in an HP1 TKO cell line, and found that ChAHP-bound genes are up-regulated in the absence of HP1 isoforms and connecting this protein complex with HP1-mediated transcriptional silencing. This suggests that HP1-mediated silencing, in the absence of H3K9me3, is essential for preventing aberrant endoderm-specific gene expression, and protecting ESCs from premature differentiation during development.   Considering that ADNP is only one in a long list of proteins with putative DNA-binding domains that interact with HP1 isoforms, I hypothesized that HP1 recruitment by a DNA sequence-specific binding protein like ADNP is not a unique pathway, but rather one of many ways HP1 can impart silencing on various groups of genes by interacting with distinct DNA binding factors. By mining HP1 TKO and ADNP KO RNA-seq data, and HP1 isoform ChIP sequencing (ChIP-seq) data (Ostapcuk, et al., 2018), I found that recruitment of HP1 proteins to gene    53 promoters cannot be entirely explained by ADNP binding to TSSs lacking H3K9me3 in mESCs (Figure 4.2). Indeed, there are a significant number of genes with HP1 binding in their promoter region that lack H3K9me3 and ADNP binding yet are nevertheless dysregulated upon HP1 depletion. Combining the results from 2 proteomics studies of HP1-interacting proteins (Hauri, et al., 2016; Nozawa, et al., 2010), I generated a list of proteins with putative DNA-binding activity that were independently identified as interactors with one, or more HP1 isoform. From this curated list, I selected two candidates for further investigation, based on a thorough literature search of each gene (Figure 4.1). The candidates I selected, AHDC1 and CHAMP1, have not previously been associated with transcriptional silencing in mESCs, but both have been implicated in human neurodevelopmental syndromes, like that observed for ADNP mutations. Using a CRISPR/Cas9-mediated gene deletion approach in mESCs, I investigated the gene-regulatory potential of Ahdc1 and Champ1 and describe the effects observed on gene transcription as well as cell growth and differentiation potential.    54   Figure 4-2 HP1 isoforms bind genes independent of H3K9me3 - A. Expression analysis of genes in mESCs lacking “writers” of H3K9me3 compared to HP1 TKO and ADNP KO cells. Genes dysregulated in HP1 TKO ESCs do not correspond to those dysregulated in KOs of H3K9me3 “writers.” HP1 TKO ESCs also show unique genes dysregulated compared to ADNP KOs, indicating an unknown mechanism for HP1 protein recruitment to gene promoters. Gene lists were filtered for genes longer than 1,000bp. Only genes with an RPKM > 1 in at least one sample (KO or WT) were investigated. A z-score above 0.75 was considered up-regulated. RNA-seq and ChIP-seq data from Chen, et al., (2018), Karimi, et al., (2011) and Ostapcuk, et al., (2018). B. Scatter plots of H3K9me3    55 ChIP-seq data and HP1 isoform ChIP-seq reveal that not all HP1 isoforms co-localize with H3K9me3. A significant number of 2kb genome-wide bins show enrichment of at least one HP1 isoform in absence of H3K9me3. C. HP1 isoforms can also bind in 2kb bins centered on TSSs, in absence of H3K9me3.   4.1.1 HP1-interacting candidate #1: AHDC1 Ahdc1 (AT-hook DNA-binding domain containing 1) is a 7-exon gene on mouse chromosome 4, with its entire ORF encoded in exon 6 (Figure 4.3 A). Ahdc1 is expressed at the highest levels in adult adrenal tissues, as well as the ovary and testes (Yue, et al., 2014). The gene produces a 1594 aa protein, with orthologues found in the superclass Tetrapoda, including humans, mice, rats, chickens and frogs (Bult, et al., 2019). Protein alignments show that AHDC1 contains two conserved regions, with two AT-hook motifs in conserved region 1 (N-terminal end) (Figure 4.3 B)(Xia, et al., 2014). The AT-hook domains can recognize and bind AT-rich DNA sequences, although no specific target sequence motif has been identified. AHDC1 also contains two P*V*L motifs, located at aa 120, ~200 aa N-terminal of Conserved region 1, and at aa 1132, at the N-terminal end of Conserved region 2. These motifs are recognized and bound by the chromoshadow domain present in all three mammalian HP1 proteins. Conserved region 2 of AHDC1 contains a putative PDZ binding domain, believed to be involved in interactions with other proteins (Xia, et al., 2014).  In humans, mutations in AHDC1 were identified as causative for Xia-Gibbs Syndrome, a rare neurodevelopmental syndrome associated with intellectual disability (ID) (Xia, et al., 2014; Yang, et al., 2015). Interestingly, patients with Xia-Gibbs Syndrome share many similarities in phenotype with those suffering from Helsmoortel-Van der Aa Syndrome, caused by mutations in    56 ADNP, including ID and global developmental delay, failure to thrive, and distinct facial features. Xia, et al. and Yang, et al. identified 9 truncating mutations in AHDC1 in Xia-Gibbs patients. 7 of these do not disrupt the AT-hook domains but do disrupt parts of conserved region 1 and/or all of conserved region 2, and the remaining two mutations were found at the N-terminal end, disrupting most of Conserved Region 1, including one or both of the AT-hook domains (Yang, et al., 2015). The first P*V*L motif is not disrupted in any of the identified mutations, as opposed to the second P*V*L motif which is disrupted in all but one patient. The authors hypothesized that the severe phenotype is due to loss of protein interactions essential for neuron and brain development (Yang, et al., 2015), but the role of AHDC1 in gene regulation or chromatin function has not been addressed.     57  Figure 4-3 The Ahdc1 locus and deletion strategy - A. The mouse Ahdc1 locus has 7 exons, with the entire coding sequence contained in exon 6. Two CRISPR/Cas9 sgRNAs were designed flanking the ORF, and PCR primers were designed around the sgRNA target sites. B. The AHDC1 protein have two conserved regions (light blue). The two A-T hook domains give the protein DNA-binding capabilities, and the PDZ domain is involved in protein    58 interactions. Yellow bars denote the site of mutations reported in humans with Xia-Gibbs syndrome (Xia, et al., 2014) (Yang, et al., 2015). C. Verification of successful gene and protein deletion by CRISPR/Cas9-mediated gene editing. PCR experiments confirmed two clones as homozygous Ahdc1 KOs. IF shows loss of AHDC1 protein following gene deletion. RNA-seq data shows no reads aligning to the Ahdc1 exon 6.  4.1.2 HP1-interacting candidate #2: CHAMP1 Based on its recurrent presence in HP1 interactome studies (Hauri, et al., 2016; Nozawa, et al., 2010; Ostapcuk, et al., 2018), its putative DNA-binding domain, and its similarities to Adnp and Ahdc1, I also chose to study the role of Chromosome Alignment-maintaining Phosphoprotein 1 (CHAMP1) (aka ZFP828) in HP1-mediated transcriptional silencing. In mice, Champ1 has two exons, with the ORF contained in exon 2 and encoding an 802 aa protein (812 aa in humans) (Figure 4.4 A). CHAMP1 has five C2H2 zinc-finger domains, two and three in the N- and C-terminus of the protein, respectively. The gene is conserved throughout vertebrates (Itoh, et al., 2011), and amongst the conserved regions are three distinct repeat motifs: the FPE motifs, WK motifs and SPE motifs (Figure 4.4 B). Itoh, et al. showed that the FPE-containing region and the C-terminal end of the protein are responsible for protein localization to spindle fibers during mitosis, whilst the C-terminus is also involved in chromosome localization (Itoh, et al., 2011). HP1 interaction was also shown to be mediated by the C-terminal region of CHAMP1 (Isidor, et al., 2016). The FPE region on its own binds the kinetochore. The SPE region contains 22 serine residues, and was shown to be involved in functional regulation via phosphorylation of these residues (Itoh, et al., 2011). The WK motifs are involved in CHAMP1’s interaction with MAD2L2, which is an important component of the spindle assembly checkpoint. Of note, in    59 interphase, CHAMP1 localizes to constitutive heterochromatin regions, a localization that is believed to be due to its interaction with HP1 proteins (Isidor, et al., 2016).   Itoh and colleagues first identified CHAMP1 in a human cell line in 2011, as a regulator of kinetochore-microtubule attachment (Itoh, et al., 2011), and like ADNP and AHDC1, it has also been implicated in human disease. In 2015, Hempel and colleagues, utilizing whole exome sequencing data, identified five patients with de novo frameshift or nonsense mutations in CHAMP1 (Hempel, et al., 2015). A further 11 cases were identified by Isidor, et al. (2016) and Tanaka, et al. (2016) the following year, all with similar reported phenotypes, including general developmental delays, intellectual disability, speech impediment, and subtle facial dysmorphias. Importantly, all identified mutations in humans cause truncated proteins lacking, at minimum, the three C-terminal zinc fingers (Figure 4.4 B) (Hempel, et al., 2015; Isidor, et al., 2016), indicating that the HP1 interaction may be essential for proper CHAMP1 function. Although the molecular mechanism for how mutations in CHAMP1 cause the observed disease phenotype has not yet been elucidated, it is speculated that loss of CHAMP1 leads to failure of proper neuronal differentiation during embryogenesis.    60  Figure 4-4 The Champ1 locus and deletion strategy - A. Champ1 is a two-exon gene with the entire ORF embedded in exon 2. CRISPR/Cas9 sgRNAs were designed to flank the ORF, and PCR primers were designed verify gene excision. B. CHAMP1 contains five zinc finger motifs with putative DNA-binding ability, as well as a number of SPE, WK, and FPE repeat motifs involved in various protein functions. Reported mutations with a human disease phenotype are reported in yellow bars (Hempel, et al., 2015) (Isidor, et al., 2016). C. Successful CRISPR/Cas9-mediated gene editing was confirmed by PCR and Western blot. Two clones showed gene deletion,    61 whilst only one clone showed successful protein ablation. Some RNA-seq reads aligned to the Champ1 exon 2, but there were no splicing events from exon 1, indicating no gene transcripts.  4.2 Results 4.2.1 mESCs harboring a deletion of Ahdc1 show a normal growth phenotype To investigate the role of AHDC1 in transcriptional regulation, I derived Ahdc1 KO mESC lines using a CRISPR/Cas9-mediated deletion approach. Cas9-mediated deletion was targeted to exon 6 to induce non-homologous end joining (NHEJ) and excision of the entire Ahdc1 coding sequence (Figure 4.3 A). I verified the deletion using PCR followed by Sanger sequencing, and confirmed protein depletion by immunofluorescence (Figure 4.3 C), with a total of 2 KO clones validated. Cells were allowed to recover under normal growth conditions for at least two weeks following gene deletion. Based on the adverse effects from mutations in vivo, and Ahdc1’s similarities with Adnp, I expected adverse effects on cell growth upon deletion. Surprisingly, Ahdc1 KO cells did not show altered growth rates compared to WT controls (Figure 4.5). Indeed, the cells appeared normal and healthy, indicating that AHDC1 does not play a significant role in regulating processes affecting mES cell growth.    62  Figure 4-5 Growth curves of Ahdc1 WT and KO mESCs. Growth curve of two Ahdc1 KO ESCs lines show no altered growth rate compared to that of Ahdc1 WT cells. 100k WT and KO cells were grown in standard growth conditions for 48h, then counted and re-seeded at the starting density. Generation doublings were calculated as log2(cell count/seeding cell count).  4.2.2 Deletion of Ahdc1 does not significantly affect gene expression To investigate the transcriptional profile of cells depleted for AHDC1, I harvested cells from one Ahdc1 KO clone, and performed RNA-seq. I first confirmed gene deletion using RNA-seq data, by verifying the absence of sequencing reads over the deleted exon (Figure 4.4 C).  Comparing the RPKM values of the Ahdc1 KO cell line to the profile of the WT control revealed only modest dysregulation of genes in the KO cells (Figure 4.6 A). After filtering, only 409 genes had a Z-score above 0.75. Of these, only 46 had a log2(fold change) above 1 (Figure 4.6 B). Intersecting this list with the list of genes upregulated in HP1 TKOs revealed only 5 genes in common between the two (Figure 4.6 B). Gene ontology (GO) analysis of the 46 upregulated genes using the DAVID Functional Annotation Tool (Huang, Sherman, & Lempicki, 2009a; 2009b) showed no significant enrichment of any GO term associated with cell differentiation or    63 neuronal development, which might explain the observed clinical phenotype resulting from AHDC1 mutations in vivo. Interestingly, some imprinted genes were amongst the genes dysregulated. However, GO analysis did not identify imprinted genes as enriched in the list of dysregulated genes. Furthermore, analysis of ChIP-seq data for HP1 isoforms as well as H3K9me3 (Ostapcuk, et al., 2018) revealed that none of the dysregulated genes, imprinted or not, were bound by HP1 proteins or enriched for H3K9me3, indicating that such upregulated genes are unlikely to be direct targets of HP1-mediated transcriptional silencing (Figure 4.7). Similarly, 32 genes showed a Z-score below -0.75 and a log2(fold change) below -1 (Figure 4.6 B). Only two downregulated genes in Ahdc1 KOs were also down-regulated in HP1 TKOs, none of which were enriched for H3K9me3 or HP1 isoform binding (Figure 4.7). Again, the DAVID Functional Annotation Tool returned no significantly enriched GO terms. Taken together, these data indicate no significant transcriptional changes indicative of an altered differentiation program, or gene sets that could explain the observed human phenotype upon loss of functional AHDC1.     64  Figure 4-6 Gene expression analysis of Ahdc1 KO mESCs - A. Scatter plots of gene expression in Ahdc1 KO versus Ahdc1 WT mESCs reveal only a limited number of genes with a log2(fc) above 1 in Ahdc1 KO mESCs. In comparison, HP1 TKO mESCs show a large number of genes with a log2(fc) above 1.  B. Venn diagrams of genes up- or down-regulated in Ahdc1 KO and HP1 TKO mESCs revealed only 7 genes dysregulated in both cell lines, indicating AHDC1 is not a co-repressor of transcription along HP1 isoforms. Up- or down-regulation was called by filtering for genes longer than 1,000bp, with a log2(fc) larger than +/- 1 and a z-score larger than +/- 0.75.     65  Figure 4-7 Expression and ChIP-seq analysis of genes in Ahdc1 KO mESCs – A. H3K9me3 ChIP-seq and HP1 TKO RNA-seq data (Ostapcuk, et al., 2018) show that genes dysregulated in Ahdc1 KOs are not marked by H3K9me3 in ESCs. B. HP1 isoform ChIP-seq and HP1 TKO RNA-seq data (Ostapcuk, et al., 2018) show that genes dysregulated in Ahdc1 KO cells are not bound by an HP1 isoform, as would be expected if they were subject to HP1-mediated transcriptional silencing.    4.2.3 Ahdc1 KO mESCs have a normal differentiation potential The phenotype observed in patients with mutations in AHDC1 is likely reflective of a failure of proper neuronal development. I therefore hypothesized that loss of AHDC1 might affect the differentiation potential of ESCs. To investigate this, I cultured Ahdc1 KO and WT mESCs in the absence of 2i and LIF to allow embryoid body (EB) differentiation. EBs were passaged and sub-sampled at 7 days following removal of 2i/LIF, and imaged and harvested at 14 days. I found no observable differences in size or morphology between Ahdc1 KO and WT EBs following differentiation (Figure 4.8 A), indicating overtly normal EB differentiation in the absence of AHDC1.     66  Figure 4-8 Analysis of differentiation potential following Ahdc1 KO in mESCs - A. Brightfield microscopy images of day 14 Ahdc1 KO and WT EBs. Ahdc1 KO mESCs were differentiated by withdrawal of 2i and LIF from the growth media for up to 14 days. Resulting EBs showed no observable differences in size or morphology compared to WT cells. B. RT-qPCR of cDNA from Ahdc1 KO and WT EBs. EBs were passaged and sub-sampled at 7 days, and harvested at 14 days, for RNA extraction. RT-qPCR of specific cell lineage marker genes revealed no change in their expression, indicating no change in differentiation potential upon loss of Ahdc1.   Differentiation of ESCs into EBs yields colonies that consist of many different differentiated cells types. It is therefore possible that although the EBs look similar under a microscope, they might be structurally different, made up of different ratios of germ layers. To investigate whether the EBs contained the expected complement of differentiated cells, I performed RT-qPCR on day 7 and day 14 EBs using a panel of lineage-specific gene amplicons, including Dnmt3b (Epiblast), Hoxa1 (early somatic cells), Nes (Ectoderm), Hand1 (Mesoderm), and Gata4 (Endoderm). Ahdc1 KO EBs showed the same expression pattern of these lineage-specific genes as WT cells (Figure 4.8 B), indicating that all major differentiation programs are activated as normal. I    67 therefore conclude that AHDC1 does not play a significant role in regulation of lineage-specifying genes in ESCs.   4.2.4 mESCs harboring a deletion of Champ1 show a normal growth phenotype To investigate whether CHAMP1 can regulate gene transcription and ESC differentiation, I derived several Champ1 KO ESC clonal lines, using the same approach as for deletion of Ahdc1. Champ1 also has its entire ORF encoded in a single exon, so I designed CRISPR/Cas9 guide RNAs flanking this exon to induce excision of the exon by NHEJ. Clones were screened for deletion by PCR and Sanger sequencing (Figure 4.4), and CHAMP1 depletion confirmed by Western blot in 1 clone (Figure 4.4). Cells were allowed to recover for two weeks before being assessed for growth rate. The KO cells retained their normal ESC morphology, and did not show any changes in growth rate compared to WT controls (Figure 4.9), indicating no adverse effects on cell division or growth upon loss of Champ1.      68  Figure 4-9 Growth curve of Champ1 WT and KO mESCs. Growth curves of two technical replicates of a Champ1 KO clone showed no altered growth potential of mESCs upon loss of Champ1. 100k cells were seeded and grown under normal growth conditions for 48h, then sampled and re-seeded at the initial seeding density. Generation doublings were calculated as log2(cell count / seeding cell count).   4.2.5 Deletion of Champ1 does not significantly affect gene expression To determine if any genes are aberrantly expressed in the absence of CHAMP1, I extracted RNA from Champ1 KO and control WT ESCs and performed RNA-seq. Sequenced reads were aligned, normalized to library depth and gene length, and compared to the same WT control line as for Ahdc1 KO cells (Figure 4.10 A). After filtering, 1,379 genes had a z-score above 0.75 and 419 of these had a log2(fc) above 1 (Figure 4.10 B). 38 genes upregulated in Champ1 KOs were also upregulated in HP1 TKOs, but none were enriched for H3K9me3 or HP1 binding following intersection with ChIPseq datasets (Figure 4.11). GO analysis showed no enrichment for terms involved in nervous system development or cell differentiation. Similarly, 1,343 genes had a z-score below -0.75, and 168 of these had a log2(fc) below -1 (Figure 4.10 B). 25 genes    69 downregulated in Champ1 KOs showed an overlap with those downregulated in HP1 TKOs, but none of these showed H3K9me3 binding or HP1 protein enrichment (Figure 4.11), indicating that they are not direct targets of HP1-mediated silencing. GO functional annotation clustering of the 168 down-regulated genes identified no GO terms involved in maintaining mESC pluripotency, nor genes involved in differentiation of mESCs to specific cell lineages, such as neuronal precursors. As for Ahdc1 KO mESCs, Champ1 KOs showed dysregulation of some imprinted genes. However, these were also not enriched for HP1 protein binding, indicating they are not subject to HP1-mediated silencing mechanisms. Gene dysregulation due to loss of Champ1 function in ESCs can therefore not be exploited to gain insight into the neurodevelopmental phenotype observed in patients with mutations in this gene.    70  Figure 4-10 Gene expression analysis of Champ1 KO mESCs - A. Scatter plots of RNA-seq data of Champ1 KO mESCs reveal only a limited number of genes dysregulated upon loss of Champ1 in comparison to loss of all three HP1 isoforms. B. Venn diagrams of genes dysregulated in Champ1 KO and HP1 TKO mESCs reveal that only a minority of the genes dysregulated in Champ1 KOs are also dysregulated in HP1 TKOs, indicating that most are not subject to HP1-mediated transcriptional silencing. Up- or down-regulation was called by filtering for genes longer than 1,000bp, with a FC larger than +/- 1 and a z-score larger than +/- 0.75.     71  Figure 4-11 Expression and ChIP-seq analysis of genes in Champ1 KO mESCs – A. Integrating RNA-seq data with publicly available ChIP-seq data of H3K9me3 and HP1 isoforms in ESCs (Ostapcuk, et al., 2018) revealed that genes dysregulated in Ahdc1 KO mESCs are not bound by H3K9me3. B. Genes dysregulated in Champ1 KO cells are not bound by HP1 isoforms, indicating they are not direct targets of HP1-mediated transcriptional silencing mechanisms.   4.2.6 Champ1 KO mESCs have a normal differentiation potential To test whether loss of Champ1 affects the ability of mESCs to properly differentiate, I grew Champ1 KO and WT cells in the absence of 2i and LIF for 14 days to allow for EB differentiation. Cells were passaged and subsampled at day 7. As for Ahdc1 KO cells, I did not observe any overt morphological differences between Champ1 KO and WT EBs at day 14 (Figure 4.12 A). RT-qPCR of KO and WT EBs at days 7 and 14 showed no changes in expression of lineage-specific markers upon loss of CHAMP1 (Figure 4.12 B). Therefore, loss of CHAMP1 does not seem to affect the differentiation potential of ESCs.      72  Figure 4-12 Analysis of differentiation potential following Champ1 KO in mESCs - A. Brightfield microscopy of day 14 EBs differentiated from Champ1 KO mESCs show EBs that appear normal in size and morphology. B. RT-qPCR of cDNA from day 14 Champ1 WT and KO EBs. Cells were harvested at days 7 and 14 during EB differentiation for RNA extraction and RT-qPCR. RT-qPCR of Champ1 KO EBs show regulation of cell lineage-specific genes comparable to that of WT EBs, indicating no change in differentiation potential upon Champ1 loss.   4.3 Discussion I hypothesized that HP1 exerts gene silencing functions in ESCs by interacting with various DNA-binding proteins for sequence-specific recruitment. I therefore chose two such DNA-binding proteins, Ahdc1 and Champ1, and investigated their effect on genic transcription in stem cells, and subsequent effects on cell differentiation. Upon loss of either protein, I observed no morphological changes to the cell population, nor did I see any effects on the differentiation potential of ESCs. It is still possible that these proteins play a role in HP1-mediated gene regulation, however, these functions might be exerted in differentiated cells, further along in embryonic development. For example, Champ1 is highly expressed in the central nervous system    73 from E11.5 onward (Yue, et al., 2014), so further experiments should investigate the role of CHAMP1 and HP1 isoforms specifically in neuronal progenitor cells, at mid to late embryonic development.  RNA-seq revealed several imprinted genes dysregulated in both the Ahdc1 and Champ1 KO cell lines, although these were not marked by HP1 isoforms or H3K9me3, as shown by ChIP-seq data. Imprinted genes are transcribed from only a single allele, based on whether that allele was inherited from the maternal or paternal germline. Silenced alleles are generally marked with DNA methylation, which encompasses a regulatory region called the imprinting control region, creating a differentially methylated region between the two alleles. The connection between AHDC1, CHAMP1, HP1 isoforms, and these imprinted regions is not immediately obvious, although it seems unlikely that both proteins somehow affects expression of imprinted genes. Comparing genes up-regulated in both Ahdc1 and Champ1 KO cells, only one imprinted gene, Grb10, is shared between the two cell lines. It is therefore more likely that stochastic changes to imprinted gene expression occurred due to the cell culturing conditions, or the process of CRISPR/Cas9-mediated deletion and clonal selection. Following transfection with CRISPR/Cas9 plasmids, both WT and KO cell lines were grown in puromycin for selection of cells containing the plasmid, and individual colonies were picked and seeded at a low density, creating a bottleneck effect. Either the KO cell lines, or the WT cell line that the KO expression was compared to, could therefore have undergone loss of imprinting, frequently reported as a random event in in vitro cell culture of ESCs (Humpherys, et al., 2001) (Dean, et al., 1998). Further studies are required to conclusively determine whether AHDC1 and/or CHAMP1 are involved in regulation of imprinted genes in ESCs.    74  With the experiments performed here, I was also unable to verify whether the dysregulated genes were direct targets of AHDC1 or CHAMP1, or simply indirect effects. To establish a direct link between the dysregulated genes, AHDC1 or CHAMP1, and HP1, it would be necessary to perform ChIP-seq experiments for both AHDC1 and CHAMP1, either by developing ChIP-optimized primary antibodies, or by endogenously tagging each protein and using already developed antibodies recognizing the protein tag. Considering the only minor effects seen on gene transcription following loss of AHDC1 and CHAMP1, and the lack of any gross morphological changes in the KO cell lines, it is hard to justify performing these experiments for AHDC1 and CHAMP1 in ESCs. However, use of a neuronal differentiation approach might be informative in this regard.  Further, the transcriptional analysis performed here only investigated the expression in one replicate of each KO cell line. We chose to sequence one replicate as a first pass experiment, and to sequence further replicates only if the initial analysis indicated disruption of HP1-mediated transcriptional silencing upon gene KO, which it did not. Although the transcriptional analysis of only one RNA-seq replicate cannot determine unequivocally that Ahdc1 and Champ1 do not cause statistically significant changes in gene regulation upon gene deletion, we believe it is sufficient evidence to conclude that these genes do not play a major role in HP1-mediated silencing. The observed gene dysregulation may instead be explained by disruption of other functions of the investigated proteins. For example, human AHDC1 was shown to interact with CBX6, a component of the Polycomb Repressive Complex 1 (PRC1) which silences the transcription of many developmentally important genes, including members of Hox clusters    75 (Vandamme, Völkel, Rosnoblet, Le Faou, & Angrand, 2011). Intersecting the list of genes up-regulated in Ahdc1 KO cells with those up-regulated in a Ring1A/B dKOs, which are essential components of PRC1, shows some overlap between the two (12 out of 46 genes up-regulated in Ahdc1 KO cells are up-regulated in Ring1A/B dKOs). However, loss of CBX6 severely impacts stem cell differentiation (Santanach, et al., 2017), unlike what I observed for Ahdc1 KO cells, suggesting that dysregulation of CBX6 targets is not occurring to a significant degree in Ahdc1 KO mESCs. Similarly, CHAMP1 also have other reported functions that could affect gene expression, as it was first identified as a regulator of kinetochore-microtubule attachment and proper chromosome segregation during mitosis (Itoh, et al., 2011). CHAMP1 also interacts with the mitotic control protein MAD2L2 and aids in its localization to mitotic spindles (Hempel, et al., 2015). Disruption of these processes could very well lead to the gene dysregulation observed in my KO clones. That said, the data presented here of gross Champ1 KO cell morphology and differentiation potential does not indicate any disruption of mitosis.   Taken together, these results show that AHDC1 and CHAMP1 are not involved in HP1-mediated gene silencing for the purpose of preventing premature or aberrant cellular differentiation of ESCs, and Ahdc1 and Champ1 are not essential genes for ESC survival and growth. Further work should be done trying to connect AHDC1 and CHAMP1 function in differentiated cells, such as neuronal progenitors, to the molecular mechanisms of the observed human phenotype upon mutations in these genes. Future studies should also examine some of the other 200 + identified HP1-interacting proteins that can be screened for involvement in HP1-mediated gene silencing. In this study, I chose HP1-interacting candidates based on reported clinical phenotypes, and similarities with the already well-studied HP1 protein interactor Adnp.    76 However, there is no reason why HP1-mediated transcriptional silencing should result in a similar phenotype following disruption of distinct HP1 interacting proteins, so future studies should refine the candidate search criteria to include other potential detrimental effects due to loss of HP1-mediated silencing.     77 Chapter 5: Concluding remarks and future directions The main objectives of this thesis, as laid out in Chapter 1, were to investigate three functions of HP1 proteins: their involvement in silencing of endogenous retroviruses, their role in regulating H3K9me3 levels, and their gene regulatory functions, as mediated by two binding partners with putative DNA-interacting domains.   In Chapter 2, I detailed the experimental protocols employed to answer these questions. Assays I employed ranged from classical molecular biology techniques (PCR, Western blots) to emerging high-throughput sequencing experiments (RNA-seq & ChIP-seq) followed by bioinformatic analysis of resulting data integrated with publicly available data.   5.1 Major findings In Chapter 3, my analysis of publicly available RNA-seq data of HP1 TKO cells, as well as publicly available and self-generated ChIP-seq data of HP1 isoforms, H3K9me3 and H4K20me3, revealed that transcriptional silencing of Class II ERVs, specifically the highly abundant RLTR and IAP families, is dependent upon HP1 proteins. The three isoforms act redundantly, and can functionally replace each other to enforce this silencing. These Class II ERVs were direct targets of HP1 binding, and their silencing is consistent with the dogma that HP1 proteins play a role in chromatin compaction, in particular when recruited to H3K9me3 marked regions. The observation that deletion of all three HP1 proteins leads to ERV upregulation stands in contrast to the previous observation that depletion of individual HP1 isoforms does not lead to upregulation of ERV expression (Maksakova, et al., 2011). While Maksakova, et al. did not address the possibility of redundancy between HP1 isoforms, I show    78 that HP1 isoforms can functionally replace each other, and only upon the loss of all three isoforms is ERV silencing disrupted. This redundancy in function is not surprising considering the potential detrimental effects of ERV transcription (Maksakova, et al., 2006; Babaian & Mager, 2016; Gagnier, et al., 2019), and highlights the essential role of HP1 proteins in H3K9me3-mediated ERV silencing.   While several reports have shown that HP1 proteins can regulated H3K9me3 levels, either by reinforcing the deposition of this mark through recruitment of further HMTs (Maison & Almouzni, 2004), or by limiting the spreading of H3K9me3 across heterochromatin boundaries (Stunnenberg, et al., 2015), I found only minor effects on H3K9me3 levels following HP1 protein ablation. Indeed, my ChIP-seq data revealed minimal changes in H3K9me3 enrichment over genes and ERVs in HP1 TKO cells, and changes in only a small number of genomic bins, unlike the effects seen for the H3K9me3- and HP1 protein-dependent H4K20me3 mark. A caveat to these findings is that they were generated by focusing on a “coarse” scale, studying ERVs only as aggregated families, and investigating relatively large genomic bins (2kb each), so it is possible that HP1 proteins may still regulate H3K9me3 on scale smaller than 2kb, for example over a few hundred base pairs at the flanks of heterochromatin regions, or over individual TEs. Regardless, HP1 proteins do not appear to be a major player in H3K9me3 dynamics regulation genome-wide.  To measure HP1 proteins’ roles in the regulation of euchromatic gene promoters, as presented in Chapter 4, I used CRISPR/Cas9-mediated gene editing in mESCs to delete Ahdc1 and Champ1, two reported interactors of HP1. Transcriptional analysis of KO lines showed that these two    79 proteins are not responsible for HP1-mediated transcriptional silencing of genes, in contrast to what was shown for ADNP, another HP1 protein interactor. I hypothesized that these interactors were co-factors of HP1 proteins in gene silencing due to their putative DNA-binding domains and reported disease phenotypes upon mutations in vivo. However, only a few genes were dysregulated upon deletion of either Ahdc1 or Champ1, and these dysregulated genes were not bound by HP1 isoforms, as identified by ChIP-seq data. Deletion of Ahdc1 and Champ1 also did not affect cell growth or differentiation potential of mESCs. Though Ahdc1 and Champ1 KO did not show any effect in mESCs, it is still possible that they play a role in HP1-mediated transcriptional silencing in differentiated cells, such as neuronal progenitor cells, which could help explain their association with neurodevelopmental syndromes upon mutation in humans.   5.2 Relevance of studying HP1 mediated transcriptional silencing Firstly, the work presented here helps in our understanding of how endogenous retroviruses are silenced, shedding light on how organisms have evolved to protect genome integrity by limiting the mutational burden due to transcription and transposition. Further, since ERVs can also be co-opted to act as enhancers and regulatory sequence for gene expression (a process called exaptation, reviewed in (Maksakova, et al., 2006)), elucidating their regulation helps in our understanding of how organisms’ regulatory networks evolve. Characterizing retrovirus regulation can also help us understand human disease conditions, since their transcription is often associated with cancers (Babaian & Mager, 2016). In mice, failed silencing, and subsequent transposition of an IAP element to a region adjacent to the granulocyte-macrophage colony-stimulating factor (GM-CSF) gene cause upregulation of this proto-oncogene, and resulted in myelomonocytic leukemia (Leslie, Lee, & Schrader, 1991). Similarly, de-repression of THE1    80 ERVs, a human-specific sub-family of Class III ERVs, causes aberrant expression of CSF1R (Colony-stimulating Factor 1 Receptor), which is involved in cancer cell survival in Hodgkin’s lymphoma (Lamprecht, et al., 2010). Therefore, understanding the factors regulating ERV expression, such as HP1 proteins, and the dynamics of H3K9me3, may contribute to the development of treatments for specific cancers and human disease (Cuellar, et al.,  2017; Muratani, et al., 2014; Nagarajan, et al., 2014; Thorsen, et al., 2011).   Secondly, understanding the full spectrum of HP1-mediated transcriptional silencing, not only of ERVs, but also of genes, can shed light on human disease. As demonstrated by (Ostapcuk, et al., 2018), HP1 proteins are essential in protecting stem cell pluripotency, and failure of HP1-mediated transcriptional silencing due to loss of HP1 interacting partners can have severe effects in vivo. The work in this thesis aids in our understanding of HP1 protein function by characterization of two interacting partners of HP1 proteins, and sets a framework for further studies of additional HP1-interacting proteins and their involvement in HP1-mediated transcriptional regulation.  5.3 Future directions and outstanding questions The results in Chapter 3 highlighted a major difficulty in studying HP1 proteins: functional redundancy between the three isoforms. Determining HP1 protein function is confounded by the three isoforms having both distinct and overlapping functions, and being involved in as widely different cellular processes as DNA damage repair, chromosome segregation in mitosis, and transcriptional silencing. With the availability of the HP1 TKO cell line, the challenge then becomes teasing apart which observed phenotype is due to disruption of what HP1 function.    81 Future work should focus on determining and characterizing the resulting phenotype from HP1 TKO in mESCs. For example, cells should be investigated for any cell cycle defects or chromosome abnormalities. It is necessary to know the cause of the growth phenotype and cell death in detail before evaluating the exact role of HP1 proteins in these processes. Ultimately, deletion of all three HP1 proteins should be conducted in in vivo systems to investigate the role of these proteins at later developmental stages, though it is necessary for such experiments to be done as conditional tissue-specific deletions, e.g. in neuron lineages, due to the lethality of the HP1b KO observed in mice (Aucott, et al., 2008).  The data presented in Chapter 4 demonstrated that there are numerous regions bound by an HP1 isoform without the presence of H3K9me3. Interestingly, these regions encompass many gene promoters, making it likely that these chromatin proteins also have a non-canonical function in gene regulation. Future work should be focused on identifying hallmarks of such regions, for example whether they share specific DNA sequence motifs potentially recognized by an interactor of HP1 proteins. Regions lacking H3K9me3 are also likely missing the redundancy in silencing afforded by H3K9me3 and its many “readers”, so it would also be interesting to know whether these regions are more or less likely to show altered levels of transcription compared to H3K9me3-marked regions upon loss of HP1 proteins. As the necessary RNA- and ChIP-seq data for such analyses are already at hand, only bioinformatic reanalysis are required to address this.  Many questions remain regarding HP1 isoforms and the breadth of their reported interacting proteins. What is the importance of the remaining 200+ HP1 protein interactors identified by proteomics studies? What are the recruiting factors responsible for HP1 protein localization to    82 regions lacking H3K9me3? And why does HP1 interact directly with all identified H3K9 methyl transferases, if not to reinforce this epigenetic mark over HP1-bound regions? To address these questions, future studies should continue with a large-scale genetic screen of HP1 protein interactors and their effect on transcriptional silencing in mESCs. KO studies employing CRISPR/Cas9 technology and homologous recombination to incorporate a reporter like GFP (Green Fluorescent Protein) and/or a FLAG tag in place of the deleted gene would allow for more rapid identification of true KO cells, and the ability to study a greater number of HP1 protein interactors. Utilizing the framework of experiments performed in this thesis for genomic characterization of KO cell lines will help standardize the data for easy comparison of the results. ChIP-seq of these proteins would further elucidate their function and involvement with HP1-mediated silencing mechanisms by uncovering potential genomic binding sites, and any co-localization with HP1 isoforms. Since many of the reported HP1-interacting proteins do not, as of yet, have antibodies raised against them, such an approach would benefit from an endogenous tagging strategy to allow for the use of antibodies against the protein tag rather than relying on antibodies raised against the protein itself.  Additionally, the two HP1 protein interactors identified in this study, AHDC1 and CHAMP1, should be investigated in differentiated cells such as neuronal progenitors, and ultimately in an in vivo system, to establish a mechanism for their involvement in the reported human neurodevelopmental syndromes. It is possible that although these genes do not affect gene regulation in mESCs, they could impact the ability of cells to differentiate properly into specific neuronal lineages later in mouse and human development. These proposed experiments will show the involvement of HP1-interacting proteins in gene regulation in both ESCs and    83 differentiated cells, highlight any implication of HP1 proteins in such mechanisms, and potentially connect HP1-mediated mechanisms to human disease states.   5.4 Conclusion HP1 proteins have been the subject of over 1900 studies (PubMed, 2020), and implicated in a wide variety of cellular processes, yet remain rather enigmatic factors. The work presented in this thesis has refined our knowledge of HP1 proteins as transcriptional silencers, and explored the functions of 2 of their over 200 reported interacting partners. Further elucidating H3K9me3-independent mechanisms of HP1 proteins, especially HP1-mediated gene silencing mechanisms in conjunction with HP1-interacing proteins, may ultimately lead to a better understanding of human disease states and developmental syndromes.    84 Bilbliography Aagaard, L., Schmid, M., Warburton, P., & Jenuwein, T. (2000). Mitotic phosphorylation of SUV39H1, a novel component of active centromeres, coincides with transient accumulation at mammalian centromeres. Journal of Cell Science, 113 ( Pt 5), 817–829. Allan, R. S., Zueva, E., Cammas, F., Schreiber, H. A., Masson, V., Belz, G. T., et al. (2012). An epigenetic silencing pathway controlling T helper 2 cell lineage commitment. Nature, 487(7406), 249–253. http://doi.org/10.1038/nature11173 Aucott, R., Bullwinkel, J., Yu, Y., Shi, W., Billur, M., Brown, J. P., et al. (2008). HP1-beta is required for development of the cerebral neocortex and neuromuscular junctions. The Journal of Cell Biology, 183(4), 597–606. http://doi.org/10.1083/jcb.200804041 Auth, T., Kunkel, E., & Grummt, F. (2006). Interaction between HP1alpha and replication proteins in mammalian cells. Experimental Cell Research, 312(17), 3349–3359. http://doi.org/10.1016/j.yexcr.2006.07.014 Babaian, A., & Mager, D. L. (2016). Endogenous retroviral promoter exaptation in human cancer. Mobile DNA, 7(1), 24–21. http://doi.org/10.1186/s13100-016-0080-x Badugu, R., Yoo, Y., Singh, P. B., & Kellum, R. (2005). Mutations in the heterochromatin protein 1 (HP1) hinge domain affect HP1 protein interactions and chromosomal distribution. Chromosoma, 113(7), 370–384. http://doi.org/10.1007/s00412-004-0324-2 Bannister, A. J., & Kouzarides, T. (2011). Regulation of chromatin by histone modifications. Cell Research, 21(3), 381–395. http://doi.org/10.1038/cr.2011.22 Bannister, A. J., Zegerman, P., Partridge, J. F., Miska, E. A., Thomas, J. O., Allshire, R. C., & Kouzarides, T. (2001). Selective recognition of methylated lysine 9 on histone H3 by the HP1 chromo domain. Nature, 410(6824), 120–124. http://doi.org/10.1038/35065138    85 Beisel, C., & Paro, R. (2011). Silencing chromatin: comparing modes and mechanisms. Nature Reviews Genetics, 12(2), 123–135. http://doi.org/10.1038/nrg2932 Belshaw, R., Watson, J., Katzourakis, A., Howe, A., Woolven-Allen, J., Burt, A., & Tristem, M. (2007). Rate of recombinational deletion among human endogenous retroviruses. Journal of Virology, 81(17), 9437–9442. http://doi.org/10.1128/JVI.02216-06 Bernstein, E., Duncan, E. M., Masui, O., Gil, J., Heard, E., & Allis, C. D. (2006). Mouse polycomb proteins bind differentially to methylated histone H3 and RNA and are enriched in facultative heterochromatin. Molecular and Cellular Biology, 26(7), 2560–2569. http://doi.org/10.1128/MCB.26.7.2560-2569.2006 Bickmore, W. A., & van Steensel, B. (2013). Genome architecture: domain organization of interphase chromosomes. Cell, 152(6), 1270–1284. http://doi.org/10.1016/j.cell.2013.02.001 Brasher, S. V., Smith, B. O., Fogh, R. H., Nietlispach, D., Thiru, A., Nielsen, P. R., et al. (2000). The structure of mouse HP1 suggests a unique mode of single peptide recognition by the shadow chromo domain dimer. The EMBO Journal, 19(7), 1587–1597. http://doi.org/10.1093/emboj/19.7.1587 Brind’Amour, J., Kobayashi, H., Richard Albert, J., Shirane, K., Sakashita, A., Kamio, A., et al. (2018). LTR retrotransposons transcribed in oocytes drive species-specific and heritable changes in DNA methylation. Nature Communications, 9(1), 3331. http://doi.org/10.1038/s41467-018-05841-x Brind’Amour, J., Liu, S., Hudson, M., Chen, C., Karimi, M. M., & Lorincz, M. C. (2015). An ultra-low-input native ChIP-seq protocol for genome-wide profiling of rare cell populations. Nature Communications, 6, 6033. http://doi.org/10.1038/ncomms7033 Brown, J. P., Bullwinkel, J., Baron-Lühr, B., Billur, M., Schneider, P., Winking, H., & Singh, P.    86 B. (2010). HP1γ function is required for male germ cell survival and spermatogenesis. Epigenetics & Chromatin, 3(1), 9. http://doi.org/10.1186/1756-8935-3-9 Bult, C. J., Blake, J. A., Smith, C. L., Kadin, J. A., Richardson, J. E., Mouse Genome Database Group. (2019). Mouse Genome Database (MGD) 2019. Nucleic Acids Research, 47(D1), D801–D806. http://doi.org/10.1093/nar/gky1056 Chen, C. C. L., Goyal, P., Karimi, M. M., Abildgaard, M. H., Kimura, H., & Lorincz, M. C. (2018). H3S10ph broadly marks early-replicating domains in interphase ESCs and shows reciprocal antagonism with H3K9me2. Genome Research, 28(1), 37–51. http://doi.org/10.1101/gr.224717.117 Chow, J., & Heard, E. (2009). X inactivation and the complexities of silencing a sex chromosome. Current Opinion in Cell Biology, 21(3), 359–366. http://doi.org/10.1016/j.ceb.2009.04.012 Cowieson, N. P., Partridge, J. F., Allshire, R. C., & McLaughlin, P. J. (2000). Dimerisation of a chromo shadow domain and distinctions from the chromodomain as revealed by structural analysis. Current Biology, 10(9), 517–525. http://doi.org/10.1016/S0960-9822(00)00467-X Cuellar, T. L., Herzner, A.-M., Zhang, X., Goyal, Y., Watanabe, C., Friedman, B. A., et al. (2017). Silencing of retrotransposons by SETDB1 inhibits the interferon response in acute myeloid leukemia. The Journal of Cell Biology, 216(11), 3535–3549. http://doi.org/10.1083/jcb.201612160  Cutter, A. R., & Hayes, J. J. (2015). A brief review of nucleosome structure. FEBS Letters, 589(20 Pt A), 2914–2922. http://doi.org/10.1016/j.febslet.2015.05.016 Dean, W., Bowden, L., Aitchison, A., Klose, J., Moore, T., Meneses, J. J., et al. (1998). Altered imprinted gene methylation and expression in completely ES cell-derived mouse fetuses:    87 association with aberrant phenotypes. Development, 125(12), 2273–2282. Dernburg, A. F., Sedat, J. W., & Hawley, R. S. (1996). Direct evidence of a role for heterochromatin in meiotic chromosome segregation. Cell, 86(1), 135–146. http://doi.org/10.1016/s0092-8674(00)80084-7 Diakun, G. P., Fairall, L., & Klug, A. (1986). EXAFS study of the zinc-binding sites in the protein transcription factor IIIA. Nature, 324(6098), 698–699. http://doi.org/10.1038/324698a0 Dodge, J. E., Kang, Y.-K., Beppu, H., Lei, H., & Li, E. (2004). Histone H3-K9 methyltransferase ESET is essential for early development. Molecular and Cellular Biology, 24(6), 2478–2486. http://doi.org/10.1128/MCB.24.6.2478-2486.2004 Eissenberg, J. C., James, T. C., Foster-Hartnett, D. M., Hartnett, T., Ngan, V., & Elgin, S. C. (1990). Mutation in a heterochromatin-specific chromosomal protein is associated with suppression of position-effect variegation in Drosophila melanogaster. Proceedings of the National Academy of Sciences, 87(24), 9923–9927. Fanti, L., Giovinazzo, G., Berloco, M., & Pimpinelli, S. (1998). The Heterochromatin Protein 1 Prevents Telomere Fusions in Drosophila. Molecular Cell, 2(5), 527–538. http://doi.org/10.1016/S1097-2765(00)80152-5 Feil, R., Brocard, J., Mascrez, B., LeMeur, M., Metzger, D., & Chambon, P. (1996). Ligand-activated site-specific recombination in mice. Proceedings of the National Academy of Sciences, 93(20), 10887–10890. http://doi.org/10.1073/pnas.93.20.10887 Fischle, W., Franz, H., Jacobs, S. A., Allis, C. D., & Khorasanizadeh, S. (2008). Specificity of the chromodomain Y chromosome family of chromodomains for lysine-methylated ARK(S/T) motifs. Journal of Biological Chemistry, 283(28), 19626–19635.    88 http://doi.org/10.1074/jbc.M802655200 Gagnier, L., Belancio, V. P., & Mager, D. L. (2019). Mouse germ line mutations due to retrotransposon insertions. Mobile DNA, 10(1), 15–22. http://doi.org/10.1186/s13100-019-0157-4 Gilbert, N., Boyle, S., Fiegler, H., Woodfine, K., Carter, N. P., & Bickmore, W. A. (2004). Chromatin Architecture of the Human Genome: Gene-Rich Domains Are Enriched in Open Chromatin Fibers. Cell, 118(5), 555–566. http://doi.org/10.1016/j.cell.2004.08.011 Hauri, S., Comoglio, F., Seimiya, M., Gerstung, M., Glatter, T., Hansen, K., et al. (2016). A High-Density Map for Navigating the Human Polycomb Complexome. CellReports, 17(2), 583–595. http://doi.org/10.1016/j.celrep.2016.08.096 Helsmoortel, C., Vulto-van Silfhout, A. T., Coe, B. P., Vandeweyer, G., Rooms, L., van den Ende, J., et al. (2014). A SWI/SNF-related autism syndrome caused by de novo mutations in ADNP. Nature Genetics, 46(4), 380–384. http://doi.org/10.1038/ng.2899 Hempel, M., Cremer, K., Ockeloen, C. W., Lichtenbelt, K. D., Herkert, J. C., Denecke, J., et al. (2015). De Novo Mutations in CHAMP1 Cause Intellectual Disability with Severe Speech Impairment. American Journal of Human Genetics, 97(3), 493–500. http://doi.org/10.1016/j.ajhg.2015.08.003 Huang, D. W., Sherman, B. T., & Lempicki, R. A. (2009a). Bioinformatics enrichment tools: paths toward the comprehensive functional analysis of large gene lists. Nucleic Acids Research, 37(1), 1–13. http://doi.org/10.1093/nar/gkn923 Huang, D. W., Sherman, B. T., & Lempicki, R. A. (2009b). Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nature Protocols, 4(1), 44–57. http://doi.org/10.1038/nprot.2008.211    89 Humpherys, D., Eggan, K., Akutsu, H., Hochedlinger, K., Rideout, W. M., Biniszkiewicz, D., et al. (2001). Epigenetic instability in ES cells and cloned mice. Science, 293(5527), 95–97. http://doi.org/10.1126/science.1061402 Isidor, B., Küry, S., Rosenfeld, J. A., Besnard, T., Schmitt, S., Joss, S., et al. (2016). De Novo Truncating Mutations in the Kinetochore-Microtubules Attachment Gene CHAMP1 Cause Syndromic Intellectual Disability. Human Mutation, 37(4), 354–358. http://doi.org/10.1002/humu.22952 Itoh, G., Kanno, S.-I., Uchida, K. S. K., Chiba, S., Sugino, S., Watanabe, K., et al. (2011). CAMP (C13orf8, ZNF828) is a novel regulator of kinetochore-microtubule attachment. The EMBO Journal, 30(1), 130–144. http://doi.org/10.1038/emboj.2010.276 Jacques, P.-É., Jeyakani, J., & Bourque, G. (2013). The majority of primate-specific regulatory sequences are derived from transposable elements. PLoS Genetics, 9(5), e1003504. http://doi.org/10.1371/journal.pgen.1003504 James, T. C., & Elgin, S. C. (1986). Identification of a nonhistone chromosomal protein associated with heterochromatin in Drosophila melanogaster and its gene. Molecular and Cellular Biology, 6(11), 3862–3872. http://doi.org/10.1128/MCB.6.11.3862 Karimi, M. M., Goyal, P., Maksakova, I. A., Bilenky, M., Leung, D., Tang, J. X., et al. (2011). DNA methylation and SETDB1/H3K9me3 regulate predominantly distinct sets of genes, retroelements, and chimeric transcripts in mESCs. Cell Stem Cell, 8(6), 676–687. http://doi.org/10.1016/j.stem.2011.04.004 Koike, N., Maita, H., Taira, T., Ariga, H., & Iguchi-Ariga, S. M. M. (2000). Identification of heterochromatin protein 1 (HP1) as a phosphorylation target by Pim-1 kinase and the effect of phosphorylation on the transcriptional repression function of HP1 1. FEBS Letters,    90 467(1), 17–21. http://doi.org/10.1016/S0014-5793(00)01105-4 Kokura, K., Sun, L., Bedford, M. T., & Fang, J. (2010). Methyl-H3K9-binding protein MPP8 mediates E-cadherin gene silencing and promotes tumour cell motility and invasion. The EMBO Journal, 29(21), 3673–3687. http://doi.org/10.1038/emboj.2010.239 Kourmouli, N., Sun, Y.-M., van der Sar, S., Singh, P. B., & Brown, J. P. (2005). Epigenetic regulation of mammalian pericentric heterochromatin in vivo by HP1. Biochemical and Biophysical Research Communications, 337(3), 901–907. http://doi.org/10.1016/j.bbrc.2005.09.132 Kwon, S. H., & Workman, J. L. (2008). The heterochromatin protein 1 (HP1) family: put away a bias toward HP1. Molecules and Cells, 26(3), 217–227. Kwon, S. H., Florens, L., Swanson, S. K., Washburn, M. P., Abmayr, S. M., & Workman, J. L. (2010). Heterochromatin protein 1 (HP1) connects the FACT histone chaperone complex to the phosphorylated CTD of RNA polymerase II. Genes & Development, 24(19), 2133–2145. http://doi.org/10.1101/gad.1959110 Lachner, M., O'Carroll, D., Rea, S., Mechtler, K., & Jenuwein, T. (2001). Methylation of histone H3 lysine 9 creates a binding site for HP1 proteins. Nature, 410(6824), 116–120. http://doi.org/10.1038/35065132 Lamprecht, B., Walter, K., Kreher, S., Kumar, R., Hummel, M., Lenze, D., et al. (2010). Derepression of an endogenous long terminal repeat activates the CSF1R proto-oncogene in human lymphoma. Nature Medicine, 16(5), 571–9– 1p following 579. http://doi.org/10.1038/nm.2129 Lehnertz, B., Ueda, Y., Derijck, A. A. H. A., Braunschweig, U., Perez-Burgos, L., Kubicek, S., et al. (2003). Suv39h-Mediated Histone H3 Lysine 9 Methylation Directs DNA Methylation    91 to Major Satellite Repeats at Pericentric Heterochromatin. Current Biology, 13(14), 1192–1200. http://doi.org/10.1016/S0960-9822(03)00432-9 Leslie, K. B., Lee, F., & Schrader, J. W. (1991). Intracisternal A-type particle-mediated activations of cytokine genes in a murine myelomonocytic leukemia: generation of functional cytokine mRNAs by retroviral splicing events. Molecular and Cellular Biology, 11(11), 5562–5570. http://doi.org/10.1128/mcb.11.11.5562 Leung, D. C., & Lorincz, M. C. (2012). Silencing of endogenous retroviruses: when and why do histone marks predominate? Trends in Biochemical Sciences, 37(4), 127–133. http://doi.org/10.1016/j.tibs.2011.11.006 Liu, S., Brind’Amour, J., Karimi, M. M., Shirane, K., Bogutz, A., Lefebvre, L., et al. (2014). Setdb1is required for germline development and silencing of H3K9me3-marked endogenous retroviruses in primordial germ cells. Genes & Development, 28(18), 2041–2055. http://doi.org/10.1101/gad.244848.114 LTR retrotransposons transcribed in oocytes drive species-specific and heritable changes in DNA methylation. (2018). LTR retrotransposons transcribed in oocytes drive species-specific and heritable changes in DNA methylation. Nature Communications, 9(1), 3331. http://doi.org/10.1038/s41467-018-05841-x Luger, K., Mäder, A. W., Richmond, R. K., Sargent, D. F., & Richmond, T. J. (1997). Crystal structure of the nucleosome core particle at 2.8 A resolution. Nature, 389(6648), 251–260. http://doi.org/10.1038/38444 Macfarlan, T. S., Gifford, W. D., Driscoll, S., Lettieri, K., Rowe, H. M., Bonanomi, D., et al. (2012). Embryonic stem cell potency fluctuates with endogenous retrovirus activity. Nature, 487(7405), 57–63. http://doi.org/10.1038/nature11244    92 Macfarlan, T. S., Gifford, W. D., Agarwal, S., Driscoll, S., Lettieri, K., Wang, J., et al. (2011). Endogenous retroviruses and neighboring genes are coordinately repressed by LSD1/KDM1A. Genes & Development, 25(6), 594–607. http://doi.org/10.1101/gad.2008511 Machida, S., Takizawa, Y., Ishimaru, M., Sugita, Y., Sekine, S., Nakayama, J.-I., et al. (2018). Structural Basis of Heterochromatin Formation by Human HP1. Molecular Cell, 69(3), 385–397.e8. http://doi.org/10.1016/j.molcel.2017.12.011 Mager, D. L., & Stoye, J. P. (2015). Mammalian Endogenous Retroviruses. Microbiology Spectrum, 3(1), MDNA3–0009–2014. http://doi.org/10.1128/microbiolspec.MDNA3-0009-2014 Maison, C., & Almouzni, G. (2004). HP1 and the dynamics of heterochromatin maintenance. Nature Reviews Molecular Cell Biology, 5(4), 296–304. http://doi.org/10.1038/nrm1355 Maksakova, I. A., Goyal, P., Bullwinkel, J., Brown, J. P., Bilenky, M., Mager, D. L., et al. (2011). H3K9me3-binding proteins are dispensable for SETDB1/H3K9me3-dependent retroviral silencing. Epigenetics & Chromatin, 4(1), 12. http://doi.org/10.1186/1756-8935-4-12 Maksakova, I. A., Romanish, M. T., Gagnier, L., Dunn, C. A., van de Lagemaat, L. N., & Mager, D. L. (2006). Retroviral Elements and Their Hosts: Insertional Mutagenesis in the Mouse Germ Line. PLoS Genetics, 2(1), e2–10. http://doi.org/10.1371/journal.pgen.0020002 Maksakova, I. A., Thompson, P. J., Goyal, P., Jones, S. J., Singh, P. B., Karimi, M. M., & Lorincz, M. C. (2013). Distinct roles of KAP1, HP1 and G9a/GLP in silencing of the two-cell-specific retrotransposon MERVL in mouse ES cells. Epigenetics & Chromatin, 6(1), 15. http://doi.org/10.1186/1756-8935-6-15 Matsui, T., Leung, D., Miyashita, H., Maksakova, I. A., Miyachi, H., Kimura, H., et al. (2010).    93 Proviral silencing in embryonic stem cells requires the histone methyltransferase ESET. Nature, 464(7290), 927–931. http://doi.org/10.1038/nature08858 Mozzetta, C., Boyarchuk, E., Pontis, J., & Ait-Si-Ali, S. (2015). Sound of silence: the properties and functions of repressive Lys methyltransferases. Nature Reviews Molecular Cell Biology, 16(8), 499–513. http://doi.org/10.1038/nrm4029 Muratani, M., Deng, N., Ooi, W. F., Lin, S. J., Xing, M., Xu, C., et al. (2014). Nanoscale chromatin profiling of gastric adenocarcinoma reveals cancer-associated cryptic promoters and somatically acquired regulatory elements. Nature Communications, 5(1), 4361–14. http://doi.org/10.1038/ncomms5361 Nagarajan, R. P., Zhang, B., Bell, R. J. A., Johnson, B. E., Olshen, A. B., Sundaram, V., et al. (2014). Recurrent epimutations activate gene body promoters in primary glioblastoma. Genome Research, 24(5), 761–774. http://doi.org/10.1101/gr.164707.113 Nellåker, C., Keane, T. M., Yalcin, B., Wong, K., Agam, A., Belgard, T. G., et al. (2012). The genomic landscape shaped by selection on transposable elements across 18 mouse strains. Genome Biology, 13(6), R45. http://doi.org/10.1186/gb-2012-13-6-r45 Nishioka, K., Rice, J. C., Sarma, K., Erdjument-Bromage, H., Werner, J., Wang, Y., et al. (2002). PR-Set7 is a nucleosome-specific methyltransferase that modifies lysine 20 of histone H4 and is associated with silent chromatin. Molecular Cell, 9(6), 1201–1213. http://doi.org/10.1016/s1097-2765(02)00548-8 Nozawa, R.-S., Nagao, K., Masuda, H.-T., Iwasaki, O., Hirota, T., Nozaki, N., et al. (2010). Human POGZ modulates dissociation of HP1alpha from mitotic chromosome arms through Aurora B activation. Nature Cell Biology, 12(7), 719–727. http://doi.org/10.1038/ncb2075 Ohno, S. (1972). So much “junk” DNA in our genome. Brookhaven Symposia in Biology, 23,    94 366–370. Ostapcuk, V., Mohn, F., Carl, S. H., Basters, A., Hess, D., Iesmantavicius, V., et al. (2018). Activity-dependent neuroprotective protein recruits HP1 and CHD4 to control lineage-specifying genes. Nature, 557(7707), 739–743. http://doi.org/10.1038/s41586-018-0153-8 Peters, A. H. F. M., Kubicek, S., Mechtler, K., O'Sullivan, R. J., Derijck, A. A. H. A., Perez-Burgos, L., et al. (2003). Partitioning and plasticity of repressive histone methylation states in mammalian chromatin. Molecular Cell, 12(6), 1577–1589. Peters, A. H. F. M., O'Carroll, D., Scherthan, H., Mechtler, K., Sauer, S., Schöfer, C., et al. (2001). Loss of the Suv39h histone methyltransferases impairs mammalian heterochromatin and genome stability. Cell, 107(3), 323–337. Porteus, M. H., & Carroll, D. (2005). Gene targeting using zinc finger nucleases. Nature Biotechnology, 23(8), 967–973. http://doi.org/10.1038/nbt1125 Rice, J. C., Briggs, S. D., Ueberheide, B., Barber, C. M., Shabanowitz, J., Hunt, D. F., et al. (2003). Histone methyltransferases direct different degrees of methylation to define distinct chromatin domains. Molecular Cell, 12(6), 1591–1598. Rowe, H. M., Jakobsson, J., Mesnard, D., Rougemont, J., Reynard, S., Aktas, T., et al. (2010). KAP1 controls endogenous retroviruses in embryonic stem cells. Nature, 463(7278), 237–240. http://doi.org/10.1038/nature08674 Sadaie, M., Kawaguchi, R., Ohtani, Y., Arisaka, F., Tanaka, K., Shirahige, K., & Nakayama, J.-I. (2008). Balance between distinct HP1 family proteins controls heterochromatin assembly in fission yeast. Molecular and Cellular Biology, 28(23), 6973–6988. http://doi.org/10.1128/MCB.00791-08 Santanach, A., Blanco, E., Jiang, H., Molloy, K. R., Sansó, M., LaCava, J., et al. (2017). The    95 Polycomb group protein CBX6 is an essential regulator of embryonic stem cell identity. Nature Communications, 8(1), 1235–11. http://doi.org/10.1038/s41467-017-01464-w Schotta, G., Lachner, M., Sarma, K., Ebert, A., Sengupta, R., Reuter, G., et al. (2004). A silencing pathway to induce H3-K9 and H4-K20 trimethylation at constitutive heterochromatin. Genes & Development, 18(11), 1251–1262. http://doi.org/10.1101/gad.300704 Shapiro, J. A., & Sternberg, R. (2005). Why repetitive DNA is essential to genome function. Biological Reviews, 80(2), 227–250. http://doi.org/10.1017/S1464793104006657 Singh, P. B., Miller, J. R., Pearce, J., Kothary, R., Burton, R. D., Paro, R., et al. (1991). A sequence motif found in a Drosophila heterochromatin protein is conserved in animals and plants. Nucleic Acids Research, 19(4), 789–794. Slotkin, R. K., & Martienssen, R. (2007). Transposable elements and the epigenetic regulation of the genome. Nature Reviews Genetics, 8(4), 272–285. http://doi.org/10.1038/nrg2072 Smallwood, A., Hon, G. C., Jin, F., Henry, R. E., Espinosa, J. M., & Ren, B. (2012). CBX3 regulates efficient RNA processing genome-wide. Genome Research, 22(8), 1426–1436. http://doi.org/10.1101/gr.124818.111 Smothers, J. F., & Henikoff, S. (2000). The HP1 chromo shadow domain binds a consensus peptide pentamer. Current Biology, 10(1), 27–30. Smothers, J. F., & Henikoff, S. (2001). The hinge and chromo shadow domain impart distinct targeting of HP1-like proteins. Molecular and Cellular Biology, 21(7), 2555–2569. http://doi.org/10.1128/MCB.21.7.2555-2569.2001 Stessman, H. A. F., Willemsen, M. H., Fenckova, M., Penn, O., Hoischen, A., Xiong, B., et al. (2016). Disruption of POGZ Is Associated with Intellectual Disability and Autism Spectrum    96 Disorders. American Journal of Human Genetics, 98(3), 541–552. http://doi.org/10.1016/j.ajhg.2016.02.004 Stocking, C., & Kozak, C. A. (2008). Endogenous retroviruses. Cellular and Molecular Life Sciences, 65(21), 3383–3398. http://doi.org/10.1007/s00018-008-8497-0 Strahl, B. D., & Allis, C. D. (2000). The language of covalent histone modifications. Nature, 403(6765), 41–45. http://doi.org/10.1038/47412 Stunnenberg, R., Kulasegaran-Shylini, R., Keller, C., Kirschmann, M. A., Gelman, L., & Bühler, M. (2015). H3K9 methylation extends across natural boundaries of heterochromatin in the absence of an HP1 protein. The EMBO Journal, 34(22), 2789–2803. http://doi.org/10.15252/embj.201591320 Tachibana, M., Sugimoto, K., Nozaki, M., Ueda, J., Ohta, T., Ohki, M., et al. (2002). G9a histone methyltransferase plays a dominant role in euchromatic histone H3 lysine 9 methylation and is essential for early embryogenesis. Genes & Development, 16(14), 1779–1791. http://doi.org/10.1101/gad.989402 Tachibana, M., Ueda, J., Fukuda, M., Takeda, N., Ohta, T., Iwanari, H., et al. (2005). Histone methyltransferases G9a and GLP form heteromeric complexes and are both crucial for methylation of euchromatin at H3-K9. Genes & Development, 19(7), 815–826. http://doi.org/10.1101/gad.1284005 Takada, Y., Naruse, C., Costa, Y., Shirakawa, T., Tachibana, M., Sharif, J., et al. (2011). HP1  links histone methylation marks to meiotic synapsis in mice. Development, 138(19), 4207–4217. http://doi.org/10.1242/dev.064444 Tamaru, H., & Selker, E. U. (2001). A histone H3 methyltransferase controls DNA methylation in Neurospora crassa. Nature, 414(6861), 277–283. http://doi.org/10.1038/35104508    97 Tanaka, A. J., Cho, M. T., Retterer, K., Jones, J. R., Nowak, C., Douglas, J., et al. (2016). De novo pathogenic variants in CHAMP1 are associated with global developmental delay, intellectual disability, and dysmorphic facial features. Cold Spring Harbor Molecular Case Studies, 2(1), a000661. http://doi.org/10.1101/mcs.a000661 Thiru, A., Nietlispach, D., Mott, H. R., Okuwaki, M., Lyon, D., Nielsen, P. R., et al. (2004). Structural basis of HP1/PXVXL motif peptide interactions and HP1 localisation to heterochromatin. The EMBO Journal, 23(3), 489–499. http://doi.org/10.1038/sj.emboj.7600088 Thompson, P. J., Macfarlan, T. S., & Lorincz, M. C. (2016). Long Terminal Repeats: From Parasitic Elements to Building Blocks of the Transcriptional Regulatory Repertoire. Molecular Cell, 62(5), 766–776. http://doi.org/10.1016/j.molcel.2016.03.029 Thorsen, K., Schepeler, T., Øster, B., Rasmussen, M. H., Vang, S., Wang, K., et al. (2011). Tumor-specific usage of alternative transcription start sites in colorectal cancer identified by genome-wide exon array analysis. BMC Genomics, 12(1), 505–14. http://doi.org/10.1186/1471-2164-12-505 Truett, G. E., Heeger, P., Mynatt, R. L., Truett, A. A., Walker, J. A., & Warman, M. L. (2018). Preparation of PCR-Quality Mouse Genomic DNA with Hot Sodium Hydroxide and Tris (HotSHOT). Doi.org, 29(1), 52–54. http://doi.org/10.2144/00291bm09 Vandamme, J., Völkel, P., Rosnoblet, C., Le Faou, P., & Angrand, P.-O. (2011). Interaction proteomics analysis of polycomb proteins defines distinct PRC1 complexes in mammalian cells. Molecular & Cellular Proteomics : MCP, 10(4), M110.002642. http://doi.org/10.1074/mcp.M110.002642 White, J., Beck, C. R., Harel, T., Posey, J. E., Jhangiani, S. N., Tang, S., et al. (2016). POGZ    98 truncating alleles cause syndromic intellectual disability. Genome Medicine, 8(1), 3–11. http://doi.org/10.1186/s13073-015-0253-0 Xia, F., Bainbridge, M. N., Tan, T. Y., Wangler, M. F., Scheuerle, A. E., Zackai, E. H., et al. (2014). De novo truncating mutations in AHDC1 in individuals with syndromic expressive language delay, hypotonia, and sleep apnea. American Journal of Human Genetics, 94(5), 784–789. http://doi.org/10.1016/j.ajhg.2014.04.006 Yang, H., Douglas, G., Monaghan, K. G., Retterer, K., Cho, M. T., Escobar, L. F., et al. (2015). De novo truncating variants in the AHDC1 gene encoding the AT-hook DNA-binding motif-containing protein 1 are associated with intellectual disability and developmental delay. Cold Spring Harbor Molecular Case Studies, 1(1), a000562. http://doi.org/10.1101/mcs.a000562 Yokochi, T., Poduch, K., Ryba, T., Lu, J., Hiratani, I., Tachibana, M., et al. (2009). G9a selectively represses a class of late-replicating genes at the nuclear periphery. Proceedings of the National Academy of Sciences of the United States of America, 106(46), 19363–19368. http://doi.org/10.1073/pnas.0906142106 Younesy, H., Möller, T., Lorincz, M. C., Karimi, M. M., & Jones, S. J. M. (2015). VisRseq: R-based visual framework for analysis of sequencing data. BMC Bioinformatics, 16 Suppl 11(Suppl 11), S2–14. http://doi.org/10.1186/1471-2105-16-S11-S2 Yue, F., Cheng, Y., Breschi, A., Vierstra, J., Wu, W., Ryba, T., et al. (2014). A comparative encyclopedia of DNA elements in the mouse genome. Nature, 515(7527), 355–364. http://doi.org/10.1038/nature13992 Zhang, Y., Riesterer, C., Ayrall, A. M., Sablitzky, F., Littlewood, T. D., & Reth, M. (1996). Inducible site-directed recombination in mouse embryonic stem cells. Nucleic Acids Research, 24(4), 543–548. http://doi.org/10.1093/nar/24.4.543    99 Zhang, Y., Romanish, M. T., & Mager, D. L. (2011). Distributions of transposable elements reveal hazardous zones in mammalian introns. PLOS Computational Biology, 7(5), e1002046. http://doi.org/10.1371/journal.pcbi.1002046 Zhao, T., Heyduk, T., & Eissenberg, J. C. (2001). Phosphorylation site mutations in heterochromatin protein 1 (HP1) reduce or eliminate silencing activity. Journal of Biological Chemistry, 276(12), 9512–9518. http://doi.org/10.1074/jbc.M010098200  


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