UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Transcriptional silencing of endogenous retroviruses : interplay between histone H3K9 methylation and… Leung, Danny Chi Yeu 2011

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

Item Metadata

Download

Media
24-ubc_2012_spring_leung_danny.pdf [ 9.01MB ]
Metadata
JSON: 24-1.0072395.json
JSON-LD: 24-1.0072395-ld.json
RDF/XML (Pretty): 24-1.0072395-rdf.xml
RDF/JSON: 24-1.0072395-rdf.json
Turtle: 24-1.0072395-turtle.txt
N-Triples: 24-1.0072395-rdf-ntriples.txt
Original Record: 24-1.0072395-source.json
Full Text
24-1.0072395-fulltext.txt
Citation
24-1.0072395.ris

Full Text

TRANSCRIPTIONAL SILENCING OF ENDOGENOUS RETROVIRUSES: INTERPLAY BETWEEN HISTONE H3K9 METHYLATION AND DNA METHYLATION  by  Danny Chi Yeu Leung B.Sc., University College London, 2004 M.Sc., Imperial College London, 2005  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY in The Faculty Of Graduate Studies (Medical Genetics)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) November 2011  © Danny Chi Yeu Leung, 2011  Abstract Endogenous retroviruses (ERVs) are found in genomes of all higher eukaryotes. As retrotransposition is deleterious, pathways have evolved to repress these retroelements. While DNA methylation transcriptionally represses ERVs in differentiated cells, this epigenetic mark is dispensable for maintaining proviral silencing during early stages of mouse embryogenesis and in embryonic stem cells (mESCs). Studies in diverse species have found histone H3K9 methylation and DNA methylation to function together to repress retrotransposons. However, until recently, little was known about the role of this histone modification in proviral silencing in mESCs. Interestingly, our analysis of mESCs lacking G9a, an H3K9-specific lysine methyltransferase (KMTase) revealed that although ERVs lost H3K9 di-methylation (me2) and DNA methylation, they remained silent. Strikingly, the levels of H3K9 tri-methylation (me3) remained unaltered, suggesting that this mark may instead be responsible for maintaining these parasitic elements transcriptionally inactive. The first stage of my research focused on identifying the enzyme depositing H3K9me3 at ERVs and on determining its role in proviral silencing. I discovered that Setdb1, another H3K9-specific KMTase, was indeed depositing H3K9me3 at a subset of ERVs and was required for maintaining transcriptional repression. Interestingly, this silencing pathway operated independently of DNA methylation. Through collaboration, we also discovered that this pathway played a diminished role in differentiated cells. Taken together, these findings indicate the existence of a DNA methylation-independent proviral silencing pathway in mESCs. The second stage of my research focused on the establishment of transcriptional repression of newly integrated proviruses. By employing an exogenous retroviral construct, I discovered a dramatic silencing defect in mESCs lacking G9a, which phenocopied cells depleted ii  of the de novo DNA methyltransferases. Furthermore, efficient DNA methylation of proviruses required G9a-mediated H3K9me2. These findings reveal that histone modifications and DNA methylation function in concert to defend the genome against invading retroviral elements in mESCs. Taken together with discoveries regarding the mechanism of DNA demethylation in early embryos, I propose that cells undergoing DNA methylation reprogramming predominantly employ histone modification-based pathways to maintain these parasitic elements in a silent state; however, the establishment of transcriptional repression for newly integrated elements also requires de novo DNA methylation.  iii  Preface Collaborators: Dr. Toshiyuki Matsui, Dr. Makoto Tachibana and Dr. Yoichi Shinkai Shinkai Lab Experimental Research Center for Infectious Diseases Institute for Virus Research Kyoto University  Dr. Irina Maksakova and Dr. Dixie Mager Mager Lab Terry Fox laboratory BC Cancer Agency University of British Columbia  Dr. Bernhard Lehnertz and Dr. Fabio Rossi Rossi Lab Biomedical Research Centre University of British Columbia  Dr. Misha Bilenky and Dr. Martin Hirst Hirst Lab BC Cancer Agency Canada’s Michael Smith Genome Sciences Centre, University of British Columbia iv  A version of Chapter 3 has been published. Toshiyuki Matsui*, Danny Leung*, Hiroki Miyashita, Hitoshi Miyachi, Hiroshi Kimura, Makoto Tachibana, Matthew C. Lorincz, Yoichi Shinkai (2010) Proviral silencing in embryonic stem cells require the histone methyltransferase ESET. Nature 464, 927-931. I conducted all experiments involving the exogenous retroviral vector and wrote most of the manuscript. Cells used in this study were derived by Toshiyuki Matsui (Shinkai lab), who also carried out northern blotting and a proportion of the ChIP and bisulfite sequencing analyses of ERVs described in the manuscript. *- Indicates co-first author  A version of Chapter 4 has been published. Danny Leung, Kevin Dong, Irina A. Maksakova, Preeti Goyal, Ruth Appanah, Sandra Lee, Makoto Tachibana, Yoichi Shinkai, Bernhard Lehnertz, Dixie L. Mager, Fabio M.V. Rossi and Matthew C. Lorincz (2011) Lysine methyltransferase G9a is required for de novo DNA methylation and the establishment but not maintenance of proviral silencing. PNAS 2011: 1014660108v1-201014660. I conducted most of the experiments and wrote most of the manuscript. Analysis of the MFG vector was done by Kevin Dong. Methylation sensitive digest and Southern blotting were conducted by Irina Maksakova (Mager Lab). Mouse embryonic fibroblasts were derived by Bernhard Lehnertz (Rossi Lab) and CKO ES cells were derived by Dr. Makoto Tachibana (Shinkai Lab).  A version of Chapter 5 has been published. Danny Leung and Matthew C. Lorincz (2011) Proviral silencing - why histone marks take center stage? Trends in Biochemical Sciences. (Submitted). I wrote this opinion article with guidance from Dr. Matthew Lorincz. Check the first pages of these chapters to see footnotes with similar information.  v  Table of Contents Abstract......................................................................................................................................... ii Preface .......................................................................................................................................... iv Table of Contents ........................................................................................................................ vi List of Tables ............................................................................................................................... ix List of Figures............................................................................................................................... x List of Illustrations..................................................................................................................... xii List of Abbreviations ................................................................................................................ xiii Acknowledgements .................................................................................................................... xx 1. Introduction .............................................................................................................................. 1 1.1 Endogenous retroviruses ................................................................................................................. 2 1.1.1 Proviral structure and lifecycle ................................................................................................... 2 1.1.2 Mouse endogenous retrovirus ..................................................................................................... 6 1.1.3 ERV expression in mouse and role of ERVs in development .................................................... 8 1.1.4 Mutagenic mechanisms of ERVs ................................................................................................ 9 1.1.5 Human endogenous retroviruses upregulation in disease ......................................................... 10 1.2 DNA methylation ............................................................................................................................ 12 1.2.1 DNA methylation and DNA methyltransferases ...................................................................... 12 1.2.2 Mechanisms of DNA methylation-mediated silencing ............................................................. 14 1.2.3 Role of DNA methylation in ERV silencing ............................................................................ 14 1.2.4 DNA methylation reprogramming in development .................................................................. 15 1.3 Chromatin biology ......................................................................................................................... 19 1.3.1 Covalent histone modifications ................................................................................................ 19  vi  1.3.2 Histone H3K9 methyltransferases ............................................................................................ 21 1.4 Interplay between epigenetic mechanisms ................................................................................... 26 1.5 Role of epigenetics in proviral silencing in mESCs ..................................................................... 28 1.6 Thesis objectives ............................................................................................................................. 31  2. Materials and methods .......................................................................................................... 33 2.1 Cell lines and culturing conditions ............................................................................................... 34 2.2 Generation of cell lines................................................................................................................... 34 2.3 Viral infections and flow cytometry ............................................................................................. 35 2.4 Generation of mESCs harbouring silent XRV ............................................................................ 36 2.5 Sequencing of reactivated MusD elements .................................................................................. 36 2.6 siRNA-mediated knockdown......................................................................................................... 37 2.7 Genomic DNA isolation and DNA methylation analysis ............................................................ 38 2.8 Combined-bisulfite-restriction-analysis (COBRA) ..................................................................... 38 2.9 Nuclear and whole-cell extractions and western blotting analysis ............................................ 39 2.10 RNA extraction and RT-PCR ..................................................................................................... 40 2.11 Native chromatin immunoprecipitation ..................................................................................... 40 2.12 Cross-linked chromatin immunoprecipitation .......................................................................... 42  3. Role of H3K9 methylation in maintenance of proviral silencing ...................................... 46 3.1 Introduction .................................................................................................................................... 47 3.2 Results ............................................................................................................................................. 48 3.2.1 Identification of the H3K9 KMTase maintaining proviral silencing ........................................ 48 3.2.2 Setdb1 is required for maintenance of class I and II ERV silencing ........................................ 49 3.2.3 Decreased levels of H3K9me3 and Setdb1 at class I and II ERVs upon Setdb1 deletion........ 51 3.2.4 H4K20me3 is reduced at class I and II ERVs upon Setdb1 depletion ..................................... 55 3.2.5 No dramatic change in DNA methylation at ERVs upon Setdb1 deletion ............................... 56 3.2.6 XRV reactivation concomitant with DNA demethylation in Setdb1 CKO cells ...................... 59  vii  3.2.7 Younger ERVs are more prevalently derepressed in Setdb1 CKO mESCs ............................. 63 3.2.8 Depletion of Kap-1 phenocopies Setdb1 .................................................................................. 65 3.3 Discussion ........................................................................................................................................ 68  4. Role of H3K9 methylation in establishment of exogenous retroviral silencing ............... 71 4.1 Introduction .................................................................................................................................... 72 4.2 Results ............................................................................................................................................. 73 4.2.1 G9a is required for the establishment of silencing of an MLV-based vector ........................... 73 4.2.2 Introduction of a catalytically active G9a transgene rescues the silencing defect, while Suv39h1/h2 are not required for silencing of newly integrated proviruses.............................. 76 4.2.3 The silencing defect in G9a-/- mESCs phenocopies that observed in de novo DNA methyltransferase mutants ........................................................................................................ 78 4.2.4 H3K9me2 is decreased in the 5’LTR/promoter region of the MSCV-GFP provirus in G9a-/mESCs ...................................................................................................................................... 80 4.2.5 G9a is required for efficient DNA methylation of MLV-based XRV ...................................... 83 4.2.6 Maintenance of proviral silencing in mESCs is not dependent upon G9a ............................... 86 4.3 Discussion ........................................................................................................................................ 88  5. Discussion and concluding remarks ..................................................................................... 91 5.1 DNA methylation-independent proviral silencing ...................................................................... 92 5.2 Setdb1/H3K9me3 is required for silencing of class I and II ERVs............................................ 94 5.3 Role of H3K9me2 in establishment of proviral silencing ........................................................... 96 5.4 Role of other histone marks in proviral silencing ....................................................................... 98 5.5 A model explaining the requirement for DNA methylation-independent proviral silencing pathways in mESCs...................................................................................................................... 101  Bibliography ............................................................................................................................. 107  viii  List of Tables Table 1 Primers list................................................................................................................... 44 Table 2 KMTases involved in silencing of class I-III ERVs.................................................. 101  ix  List of Figures  Figure 1 ERVs remain silent despite reduction of DNA methylation and H3K9me2 in G9a -/mESCs..........................................................................................................................30 Figure 2 Setdb1 is required for XRV silencing...........................................................................49 Figure 3 Class I and II ERVs are reactivated upon Setdb1 depletion.........................................51 Figure 4 Reduction of H3K9me3 at class I and II ERVs upon Setdb1 depletion.......................53 Figure 5 Suv39h1/h2 DKO cells show reduction of H3K9me3 at major satellites but not at IAP LTRs.............................................................................................................................54 Figure 6 H3K9me2 is not deposited by Setdb1 in vivo...............................................................55 Figure 7 Setdb1 directly targets ERV 5’LTRs in mESCs...........................................................55 Figure 8 Reduction of H4K20me3 at MLV and IAP LTRs in Setdb1 deleted mESCs..............56 Figure 9 Modest to no change in DNA methylation patterns at ERV LTRs in Setdb1 CKO cells...............................................................................................................................58 Figure 10 H3K9me3 on proviruses is deposited independently of DNA methylation................59 Figure 11 Maintenance of XRV silencing requires Setdb1.........................................................61 Figure 12 Reduction of H3K9me3 and H4K20me3 at MSCV LTR in Setdb1 deleted mESCs..........................................................................................................................62 Figure 13 Reactivation of XRV is coupled with DNA demethylation........................................63 Figure 14 Setdb1 deleted cells express a greater diversity of MusD elements than wildtype cells...............................................................................................................................64 Figure 15 Depletion of Kap-1 phenocopies depletion of Setdb1.................................................66 Figure 16 Setdb1 recruitment to proviral 5’LTR regions is dependent on Kap-1.......................67 Figure 17 Retroviral vector and infection schema.......................................................................74 x  Figure 18 G9a-/- mESCs show a defect in silencing of XRV.....................................................75 Figure 19 Introduction of wildtype but not catalytically inactive G9a transgene rescues silencing defect............................................................................................................................77 Figure 20 Dnmt3a/3b-/- mESCs phenocopy silencing defect of G9a null cells..........................79 Figure 21 H3K9me2 at MSCV LTR is reduced in G9a-/- mESCs..............................................82 Figure 22 The rate of proviral de novo DNA methylation is reduced in G9a-/- mESCs.............85 Figure 23 G9a is not required for the maintenance of proviral silencing in mESCs...................87 Figure 24 ERVs derepressed in Setdb1 KO versus Kdm1a KO mESCs......................................99  xi  List of Illustrations  Illustration 1 The structure of an integrated provirus....................................................................4 Illustration 2 Process of reverse transcription...............................................................................5 Illustration 3 Structure of the DNA methyltransferases in mice.................................................13 Illustration 4 Mouse genome undergoes two waves of DNA methylation reprogramming......................................................................................................16 Illustration 5 Methylation of lysine residues on histone tails can confer different transcriptional states......................................................................................................................20 Illustration 6 Structure of the Suv39 sub-family of SET domain containing KMTases.............22 Illustration 7 The stages of early mouse embryonic development, Tet protein expression and DNA methylation dynamics................................................................................103 Illustration 8 Chromatin modifying enzymes involved in establishment and maintenance of proviral silencing and a hypothetical pathway for turnover of DNA methylation at these elements in mESCs....................................................................................105  xii  List of Abbreviations 4-OHT  4-hydroxytamoxifen  5caC  5-carboxylcytosine  5hmC  5-hydroxymethylcytosine  5mC  5-methylcytosine  AID  Activation-induced deaminase  AKR  Aldo-keto reductase  APOBEC  Apolipoprotein B mRNA-editing enzyme catalytic polypeptide-like  Ash2l  Ash2 (absent, small or homeotic 2)- like  Avy  Agouti viable yellow  BER  Base excision repair  cDNA  Complementary DNA  ChIP  Chromatin immunoprecipitation  ChIP-seq  Chromatin immunoprecipitation – massive-parallel sequencing  CKO  Conditional knockout  CMT3  Chromomethylase 3  COBRA  Combined bisulfite restriction analysis  CpG  Cytosine-guanine dinucleotide  Cre-ER  Cre recombinase- oestrogen receptor  CSF1R  Colony-stimulating factor 1 receptor gene  Cys  Cystine  Dazl  Deleted in azoospermia-like  DDM1  Deficient in DNA methylation 1 xiii  DIM-2  Defective in DNA methylation-2  DIM-5  Defective in DNA methylation-5  DKO  Double knockout  DMEM  Dulbecco’s modified eagle’s medium  DNA  Deoxyribonucleic acid  Dnmt  DNA methyltransferase  Dnmt1  DNA methyltransferase 1  Dnmt3a  DNA methyltransferase 3a  Dnmt3b  DNA methyltransferase 3b  Dnmt3l  DNA methyltransferase 3-like  Dot1l  Disruptor of telomeric silencing 1- like  EC  Embryonic carcinoma cells  EDTA  Ethylenediaminetetraacetic acid  env  Envelope  ESCs  Embryonic stem cells  ETn  Early transposon  Ezh2  Enhancer of zeste homolog 2  FACS  Fluorescence-activated cell sorting  G9a  H3K9-specfc lysine methyltransferase  gag  Group-specific antigen  Gapdh  Glyceraldehyde-3-phosphate dehydrogenase  gDNA  Genomic DNA  Gfp  Green fluorescent protein  xiv  GLN  murine retrovirus using glutamine tRNA  Glp  G9a- like protein  H2A.Z  Variant histone H2A.Z  H3K27  Histone 3 lysine 27  H3K27ac  Histone 3 lysine 27 acetylation  H3K27me3  Histone 3 lysine 27 tri-methylation  H3K36  Histone 3 lysine 36  H3K4  Histone 3 lysine 4  H3K4me3  Histone 3 lysine 4 tri-methylation  H3K64  Histone 3 lysine 64  H3K64me3  Histone 3 lysine 64 tri-methylation  H3K79  Histone 3 lysine 79  H3K9  Histone 3 lysine 9  H3K9ac  Histone 3 lysine 9 acetylation  H3K9me1  Histone 3 lysine 9 mono-methylation  H3K9me2  Histone 3 lysine 9 di-methylation  H3K9me3  Histone 3 lysine 9 tri-methylation  H4K20  Histone 4 lysine 20  H4K20me3  Histone 4 lysine 20 tri-methylation  HCl  Hydrochloric acid  Hdac  Histone deacetylase  HEPES  4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid  HERV  Human endogenous retrovirus  xv  HERV-E  Human endogenous retrovirus type E  HERV-K  Human endogenous retrovirus type K  HERV-K (HML)  Human endogenous retrovirus type K (human MMTV like)  HERV-K18  Human endogenous retrovirus type K subfamily 18  HERV-W  Human endogenous retrovirus type W  HLA  Human leukocyte antigen  Hp1  Heterochromatin protein 1  IAP  Intracisternal A-type particle  IAPEz  Intracisternal A-type particle subfamily Ez  Kap1  KRAB-associated protein 1  KCl  Potassium chloride  KD  Knockdown  KMTase  Lysine methyltransferase  KO  Knockout  KOH  Potassium hydroxide  KRAB-ZFP  Krüppel-associated box domain- zinc finger protein  KRY  Kryptonite  LiCl  Lithium chloride  LIF  Leukemia inhibitory factor  LINE  Long interspersed nucleotide elements  LSH  Lymphoid-specific helicase  LTR  Long terminal repeat  Mage-a2  Melanoma-associated antigen family A- 2  xvi  MaLR  Mammalian apparent LTR retrotransposons  Mbd  Methyl binding domain  MEF  Mouse embryonic fibroblast  MERV-L  Mouse endogenous virus type L  mESCs  Mouse embryonic stem cells  MET1  Methyltransferase 1  MLV  Moloney murine leukemia virus  MMTV  Mouse mammary tumor virus  mRNA  Messenger RNA  MSCV  Mouse stem cell virus  MTA  Mouse transposon A  MusD  Mouse type D retrotransposons  Na2HPO4  Di-sodium hydrogen phosphate  NaCl  Sodium chloride  NLS  Nuclear localization signal  NP-40  Tergitol-type NP-40  Np9  HERV-K 5q33.3 provirus Np9 protein  Nsd1  Nuclear receptor binding SET domain protein  NuRD  Nucleosome remodeling and histone deacetylase complex  Oct4  Octamer-binding transcription factor 4  PBS  Primer binding site  PCR  Polymerase chain reaction  Peg13  Paternally expressed 13  xvii  PGC  Primordial germ cells  PI  Post-infection  PIC  Protease inhibitory cocktail  PMSF  Phenylmethanesulfonylfluoride  pol  Polymerase  PPT  Poly-purine tract  pro  Protease  PVDF  Polyvinylidene fluoride  PWWP  Proline-tryptophan-tryptophan-proline motif  qPCR  Quantitative polymerase chain reaction  qRT-PCR  Quantitative reverse transcriptase polymerase chain reaction  R  Repeat sequence of the LTR  Rec  HERV-K 7q22.1 provirus Rec protein  REPBASE  Repetitive DNA database  RNA  Ribonucleic acid  RNA-seq  RNA – massive-parallel sequencing  RNAi  RNA interference  RT  Reverse transcriptase  SDS-PAGE  Sodium dodecyl sulfate- polyacrylamide gel electrophoresis  SET  Su(var)3-9, Enhancer-of-zeste, Trithorax domain  Setdb1  SET domain bifurcated 1  Setdb2  SET domain bifurcated 2  Sin3a  Sin3 homolog A  xviii  SINE  Short interspersed nucleotide elements  siRNA  Short interfering RNA  Suv39h1  Suppressor of variegation 3-9 homolog 1  Suv39h2  Suppressor of variegation 3-9 homolog 2  Suv420h1  Suppressor of variegation 4-20 homolog 1  Suv420h2  Suppressor of variegation 4-20 homolog 2  SUV5  Suppressor of variegation 3-9 homolog 5  SUV6  Suppressor of variegation 3-9 homolog 6  Taq  Thermus aquaticus  Tdg  Thymine-DNA glycosylase  TET  Ten-eleven translocation  Tet1  Ten-eleven translocation 1  Tet2  Ten-eleven translocation 2  Tet3  Ten-eleven translocation 3  TG  Transgene  THE1B  MaLR-like retrovirus  TKO  Triple knockout  tRNA  Transfer RNA  U3  Unique sequence to the 3’end of LTR  U5  Unique sequence to the 5’end of LTR  Wdr5  WD repeat domain 5  XRV  Exogenous retrovirus  xix  Acknowledgements My sincere thanks go to my supervisor, Matthew Lorincz, who has provided me guidance on the path to becoming an independent researcher. I also thank all the members of my lab, Yoichi Shinkai, Dixie Mager, Fabio Rossi and Carolyn Brown who gave me the help and support I needed in order to be productive.  I thank my parents and sisters for their encouragement and support as I make this uncertain journey to fulfill my dreams. Lastly, I thank my dear Athena who has given me unending support and motivation through the years and made it possible for me to pursue my career unreservedly.  xx  1. Introduction  1  1.1 Endogenous retroviruses Transposable elements have colonized the genomes of all higher eukaryotes. They are vastly diverse in sequence and constitute significant fractions of host genomes. These repetitive elements can be divided into DNA transposons and retrotransposons. The latter replicates through reverse transposition and amplify via a “copy-and-paste” process. Retrotransposons are further subdivided into non-long terminal repeat (LTR) retrotransposons and LTR retrotransposons. Non-LTR retrotransposons include the long interspersed nuclear elements (LINE) and short interspersed nuclear elements (SINE) and are the most abundant elements in mammalian genomes. LTR retrotransposons, including endogenous retroviruses (ERVs), constitute approximately 8 and 10% of the human and mouse genome, respectively [1, 2] and are divided into three classes, which will be described in greater detail below [3]. Retrotransposition of a subset of ERVs is responsible for up to 10% of all spontaneous mutations in mice [4] and therefore can have detrimental effects on the host fitness.  1.1.1  Proviral structure and lifecycle  ERVs are relics of exogenous retroviruses (XRV) that have infected and integrated into the host germ line and became “endogenized”. They can transmit vertically to offspring or horizontally to neighboring cells via production of infectious particles [5-7]. Upon integration, the retroviral element is known as a provirus and is treated as a part of the host genome. The provirus contains 5’ and 3’ LTRs flanking the retroviral genes: gag, pro, pol and env (Illustration 1). The gag gene encodes for the group specific antigen  2  proteins, which include the viral matrix, capsid and nucleoproteins. The pro gene encodes the protease, which functions to cleave the gag polyprotein precursor. The pol gene encodes the reverse transcriptase and integrase proteins and the env gene encodes the surface glycoproteins and transmembrane polyproteins. The group antigen proteins are the structural components of the retroviral particle, consisting of the RNA genome binding proteins and core nucleoproteins. The reverse transcriptase is an RNA-dependent DNA polymerase and possesses RNase H activity required during retrotransposition. The integrase protein introduces breaks in the host chromosomal DNA to mediate integration of proviral DNA into the genome. These enzymes are essential for the reverse transcription and integration process and in turn, the amplification of the provirus. The envelop proteins are the surface proteins that bind to host receptors on the cell membrane and therefore determine the range of host cells the retrovirus is able to infect (viral tropism). In addition to the retroviral genes, the provirus also contains the primer-binding site (PBS) and the poly-purine tract (PPT), which play essential functions in the retrotransposition and transcription of the element respectively [8]. The flanking LTRs contain the U5 (unique to 5’ end), R (repeated sequence) and U3 (unique to 3’ end) sequences, which are necessary for regulation of proviral transcription. Throughout evolution, many ERVs in mammalian genomes have acquired mutations where segments of the provirus may be duplicated or deleted. In fact, various ERV classes lack the gagpro-pol-env retroviral genes and thousands are even reduced to solitary LTRs [9].  3  Illustration 1. The structure of an integrated provirus (not to scale). 5’ and 3’ long terminal repeats (LTR) flank the primer binding site (PBS) and retroviral genes. The gag encodes for the matrix proteins, capsid proteins and nucleoproteins. The pro gene encodes for the protease. The pol gene encodes for the reverse transcriptase and integrase enzymes. The env gene encodes for the surface and transmembrane glycoproteins  In order for retroviruses to propagate and expand, they must replicate through retrotransposition. For XRVs or ERVs capable of producing functional retroviral particles, the glycoproteins of the virion bind to the host cell membrane receptors leading to fusion of the membranes, allowing the viral RNA genome entry into the cytoplasm. The viral genome then reverse transcribes into a double stranded DNA copy (Illustration 2). The DNA is transported to the nucleus where the integrase protein introduces double stranded breaks, facilitating the integration of viral DNA into the host genome. Upon integration, the provirus becomes a part of the host cell’s genome and is transcribed by endogenous RNA polymerase II and replicated along with the host chromosomes by the cellular machinery. The resulting transcripts are capped at the 5’ end and polyadenylated and therefore behave as do the host mRNAs.  4  Illustration 2. Process of reverse transcription. (I) The retroviral PBS is recognized by host complementary tRNA molecules that function as a primer for extension, catalyzed by reverse transcriptase using the viral RNA genome (yellow) as template. (II) Viral R and U5 sequences are degraded by RNase H. (III) The newly generated DNA (green) hybridizes to the 3’ LTR of the RNA and primes the extension through the viral sequence. (IV) RNase H removes most of viral genome and the remaining RNA sequence is used as primer for extension towards the 3’ tRNA end. (V) The tRNA is removed along with remaining hybridized RNA. The newly generated strand hybridizes to complementary PBS sequence. (VI) Extension of both strands completes production of a double stranded DNA molecule of the proviral genome.  As previously mentioned, ERVs are the descendants of XRVs, which entered the germ line of hosts. Through evolution, many proviral elements have lost their viral gene  5  sequences and/or their capacity for transcription, possibly due to non-homologous recombination between proviral copies and accumulation of mutations in the promoter/enhancer sequences respectively. Therefore these ERVs may be incapable of retrotransposition. In contrast, others ERVs are capable of producing full-length transcripts containing all the viral genes and thus can retrotranspose. A few elements are even able to form virions capable of infecting neighboring cells. Interestingly, subsets of ERVs have accumulated mutations that enhance their capacity for intracellular retrotransposition, facilitating amplification to high copy numbers [10].  1.1.2 Mouse endogenous retrovirus In the mouse, ERVs comprise ~10% of the genome and their reverse transposition activity have been reported to account for 10% of all spontaneous germ line mutations [1, 4]. These retroelements are divided into three classes based on the sequence of their pol genes and range greatly in copy number from several to tens of thousands of copies per genome [8]. They also vary between different mouse strains, as some elements are more prevalent in one strain versus another [9]. Class I endogenous retroviruses/gammaretroviruses are classified as type C based on virion morphology [11] and constitute ~0.7% of the mouse genome [1]. Members of this ERV class are grouped together based on their similarity to the Moloney Murine Leukemia Virus (MoMLV or MLV) [8], which was identified in leukemic cells of the AKR mutant mouse strain [12, 13]. MLV integration into the mouse germ line dates to approximately 1.5 million years ago and is relatively “younger” than members of the  6  other classes of ERVs [14]. Vectors based on the MLV XRVs have been extensively used in molecular biology and also in gene therapy trials [15-18]. Depending on the mouse strain, there are 25 to 70 copies of MLV elements in the genome [8, 14]. Class I ERVs are further subdivided based on their range of target host cells, as determined by the env gene sequence. A few members of this class, such as GLN, have the capacity to produce functional virions [8, 19]. Class II ERVs/betaretroviruses are classified as Type B and D based on their viral particle morphology[20] and comprise of ~3% of the mouse genome [1]. This class consists of many more members capable of producing functional retroviral particles as compared to class I [21, 22]. Class II ERVs show some sequence similarity to the Mouse Mammary Tumor Virus (MMTV) [8], first identified as a factor capable of causing mammary cancers in mice, manifesting in vertical transmission from mother to offspring via viral particles released into the milk [23]. Another member of this class includes the well-studied non-infectious family of Intracisternal A Particles (IAPs) [24, 25], which is present in approximately 2000 copies in the mouse genome [9]. Other class II ERVs include MusD and the closely related Early Transposon (ETn) elements, which lack the retroviral genes but utilize the gene products of MusD elements to retrotranspose [26, 27]. The LTR sequences of these two families are virtually identical and are present at a combined ~400 copies in the mouse genome [9]. MusD are among the previously mentioned subsets of ERVs that contain mutations in the gag gene, allowing for intracellular maturation of virions and thus enhance their ability to retrotranspose within the host cell [10]. Class III ERVs, also known as spumavirus-like elements, are the most numerous 7  of all three ERV classes, comprising of ~5.4% of the mouse genome [1]. These elements consist largely of Murine ERV-L (MERV-L) [28] and the non-autonomous MaLR elements [29], the latter of which is the most abundant ERV family in the mouse genome. These retroelements are relatively old and members are found in all placental mammals indicating their integration at more than 70 million years ago [28]. Class III ERVs are amongst the most transcriptionally active [4] and some elements have even been co-opted by the host for use as promoters during specific stages of early embryogenesis [30, 31].  1.1.3 ERV expression in mouse and role of ERVs in development Given the possible detrimental consequences of ERV retrotransposition, which will be further discussed below, the host cells must tightly regulate the transcription of these elements. Interestingly, unlike most differentiated cell types, particular families of ERVs are expressed in cells of the early embryo and placenta [31-34]. As such elements have a greater chance to amplify and integrate into the germ line, they have the highest copy numbers. Amongst the most active ERVs are the IAP and MusD/ETn elements [24, 33, 35, 36]. IAP expression has been observed in several mouse tumors and cell lines [25] and ETn expression in undifferentiated embryonic carcinoma (EC) and embryonic stem cells (ESCs) [35, 37]. Both families are highly expressed during early embryogenesis, but are silenced as development progresses. Interestingly, Peaston et al. discovered that class III ERVs such as Mouse Transposons A (MTA) and MERV-L have been co-opted to affect the expression of endogenous genes during early stages of mouse embryonic development [30, 31]. The  8  LTRs of these elements are used as alternative promoters for genes in the oocyte and two-cell stages. Many chimaeric transcripts begin from MTA LTRs and splice into neighboring genes. Such transcripts disappear by the eight-cell stage, where MTA elements are no longer expressed. These findings indicate that the host cells have evolved to utilize specific ERV subfamilies at precise developmental periods to direct genic transcription profiles by exploiting proviral regulatory sequences.  1.1.4 Mutagenic mechanisms of ERVs The aberrant expression of ERVs can cause damage to the host genome through several mechanisms including insertion into gene introns leading to premature polyadenylation, improper splicing and transcriptional deregulation of nearby genes [4]. ETn elements have been reported to most commonly cause aberrant splicing and introduce ectopic polyadenylation signal into genes by integrating in introns [4]. Whereas, IAP elements have been found to more frequently affect the expression of neighboring genes [4]. For instance, in the viable yellow agouti (Avy) mouse model, an IAP element integrated in intron 1 of the agouti gene in some mouse strains can function as an alternate promoter [38]. Ectopic transcription starts from the IAP cryptic promoter instead of the mouse haircycle specific promoter and expression of the provirus leads to ubiquitous expression of the gene, which manifests as a phenotype of altered yellow coat colour and obesity [38]. Further illustrating the potentially deleterious effects of deregulation of these elements, multiple studies in Drosophila, mice and humans, have reported that upregulation of retrotransposons can lead to an increase in double stranded breaks in the genome as newly reversed transcribed retroviral DNA are integrated [39-43]. Therefore, 9  transcriptional silencing of ERVs is important for maintaining the integrity of both the genome and transcriptome.  1.1.5 Human endogenous retroviruses upregulation in disease The upregulation of ERVs has been observed in many human diseases including neurological disorders, autoimmune diseases and cancers. Human endogenous retroviruses (HERVs) much like their mouse counterparts, constitute a significant proportion of the genome (~8%) [2] and their transcriptional states are tightly controlled. In neurological disorders such as multiple sclerosis (MS), increased expression of HERVW elements is found in autopsied brains of patients and the upregulation in MS plaques correlates with increased demyelination and inflammation [44, 45]. Other studies have found correlations between the production of HERV proteins and the pathogenesis of autoimmune disorders such as systemic lupus erythematosus and rheumatoid arthritis, potentially via molecular mimicry [46-49]. HERV-E proteins are reported to cross-react with human HLA class I proteins resulting in autoimmunity against endogenous proteins [46]. HERVs have also been reported to form superantigens in autoimmune disease patients. The env proteins of HERV-K18, a MMTV-related element [50], interacts with the Vβ chain of T-cell receptors and activate the T-cells in autoimmune disorders such as rheumatoid arthritis and type I diabetes [48, 51]. These findings demonstrate that the deregulation of HERVs may contribute to immunological response observed in specific autoimmune disorders. Deregulation of HERVs has also been reported in various cancer types. Studies 10  have found increased transcription of HERVs in multiple cancers and cancer cell lines [52]. For example, HERV-K proteins have been detected in melanomas [53], testicular germ cell tumors [54] and ovarian cancers [55], whereas HERV-W proteins have been found in breast cancers [56], endometrial carcinoma tissues [57] and neuroblastoma cell lines [45]. HERV derepression could lead to oncogenesis via several potential mechanisms. Proviruses encode oncogenic proteins, possibly resulting from erroneous incorporation of host genes into the retroviral genome (oncogene capture). Np9 and Rec are two proteins with oncogenic potential, encoded by HERV-K (HML) elements [58, 59]. Transcripts of these genes are detected specifically in malignant tissue samples [60]. HERVs may also induce oncogenesis via inactivation of tumor suppressor genes by insertional mutagenesis. HERV-K (HML) elements possess intact retroviral genes and could potentially retrotranspose, leading to introduction of mutations. Unlike the other diseases described above, although many studies have found HERVs up-regulated in cancers, evidence supporting the aforementioned mechanisms of proviral deregulation promoting tumorigenesis is lacking. A recent study has shed light on this relationship by revealing that the survival of malignant cells in Hodgkin’s lymphoma depends on the aberrant expression of the colony-stimulating factor 1 receptor (CSF1R) gene, which initiates from the adjacent upregulated THE1B element, a MaLR-like HERV [61]. Therefore, the derepression of this HERV is directly involved in the pathogenesis of this cancer. Given that derepression of ERVs generally has deleterious consequences, it is essential for host cells to employ mechanisms for maintaining these intracellular parasites in a silent state.  11  1.2 DNA methylation 1.2.1 DNA methylation and DNA methyltransferases Epigenetic modifications are covalent modifications of DNA or associated proteins that affect the expression of genes without altering the underlying coding sequence. Of the different epigenetic pathways, DNA methylation is the best-characterized transcriptional silencing mechanism and is evolutionarily conserved in a wide range of organisms, including fungi, plants and animals. In the context of transcriptional repression, DNA methylation involves the catalytic addition of a methyl group to the 5th carbon of the cytosine base (5mC). In mammals, methylation occurs mostly in the context of cytosineguanine dinucleotides (CpG). In mice, approximately 40% of promoters contain or are proximal to regions containing high frequencies of CpGs known as CpG islands [62]. However, paradoxically, CpGs within CpG islands generally remain unmethylated, while CpGs outside of such islands are typically densely methylated. Methylation patterns are maintained during DNA replication and cell divisions and are vital in the regulation of gene expression, as well as other important processes such as X chromosome inactivation [63] and genomic imprinting [64]. While the enzymes catalyzing DNA methylation, the DNA methyltransferases (Dnmts), were cloned and characterized over a decade ago, the mechanisms responsible for targeting of the Dnmts to specific genomic regions have yet to be elucidated. DNA methylation in mammals is established by the de novo Dnmts, Dnmt3a and Dnmt3b, and maintained by the maintenance Dnmt, Dnmt1 (Illustration 3). During early stages of embryonic and germ line development, new methylation patterns are established by Dnmt3a and Dnmt3b, which use unmethylated CpGs as their substrates. These de 12  novo Dnmts are highly expressed in embryos, ESCs and germline cells [65], and are essential for establishing methylation patterns in these cells [66]. Dnmt3a and Dnmt3b double knockout (DKO) mouse embryos reveal abnormal morphology at E8.5 and E9.5 as well as embryonic lethality before E11.5 [67]. The mESCs derived from these blastocysts show global DNA hypomethylation. Dnmt3l, a related protein with sequence homology to Dnmt3a and 3b, lacks catalytic activity, but is essential for de novo methylation in the germ line [68]. Dnmt1, on the other hand, assures the maintenance of established methylation patterns. Dnmt1 is associated with replication foci in S phase [69], where it recognizes hemimethylated DNA and methylates the cytosine on the newly synthesized strand. Dnmt1 KO embryos die during mid-gestation (before E10.5), indicating its importance in embryonic development [70]; whereas Dnmt1 KO mESCs are viable but show severe DNA hypomethylation [70].  13  Illustration 3. Structure of the DNA methyltransferases in mice. The maintenance Dnmt, Dnmt1 contains a nuclear localization signal (NLS), a replication foci targeting sequence, cystine-rich domain and a methyltransferase domain. The de novo Dnmts, Dnmt3a and Dnmt3b both have PWWP domain, cystine-rich domains and the catalytic methyltransferase domain. Dnmt3l contains the cystine-rich domain but lacks the catalytic domain; however, this protein is also necessary for de novo DNA methylation in the germ line.  1.2.2 Mechanisms of DNA methylation-mediated silencing Two principle mechanisms of DNA methylation-mediated inhibition of gene expression have been proposed, including the direct and indirect models. In the direct inhibition model, CpG methylation at promoter or enhancer regions inhibits binding of critical transcription factors and therefore prevents active transcription. In the indirect model, the methylated cytosine is the ligand for binding of recognition proteins such as MeCP2 or MBD2, which subsequently recruit Sin3a and NuRD complexes, respectively [71-73]. Both complexes contain histone deacetylases (Hdacs), which remove the active acetylation mark on histone tails and ultimately lead to chromatin compaction and inhibition of transcription.  1.2.3 Role of DNA methylation in ERV silencing Dr. Tim Bestor and colleagues have suggested that DNA methylation evolved primarily as a defense mechanism against transposable elements [74]. Indeed, in addition to the many important regulatory roles that DNA methylation plays in the cell, this mark is essential for ERV repression in specific cell types and tissues [75, 76]. In most tissues, ERV LTRs are heavily DNA methylated and this mark is required for repression of IAP elements in late stage embryos, embryonic fibroblasts (MEFs), neuronal cells and 14  epiblast-derived stem cells [75, 76]. Analysis of Dnmt1 KO mice reveals that null embryos die before day E8.5. To allow for further analysis of its role in embryonic development Dnmt1N, a targeted Dnmt1 allele was employed, reducing the protein level by 95% and the genome-wide DNA methylation down to ~30% of wildtype. These mutant embryos survive for an additional two days and embryos at day E9.5 reveal a 50 to 100-fold upregulation of IAP expression [75]. More recently, another study looked at the expression levels of IAP in various cell types in a Dnmt1 KO genetic background [76]. A >100-fold increase of IAP transcripts is detected upon differentiation of Dnmt1 null mESCs. Intriguingly, mESCs devoid of DNA methylation do not show proviral silencing defects [76]. This observation will be a central theme in my thesis, as discussed below. High levels of IAP mRNA and proteins are detected in Dnmt1 null epiblast derived stem cells and also Dnmt1 conditional KO (CKO) MEFs, neurons and glial cells [76]. Interestingly, DNA methylation also has a critical function in transcriptional silencing of repetitive elements and their relics in filamentous fungi [77] and plants [78], indicating evolutionary conservation of defense mechanisms against these retroelements. Taken together, these studies illustrate the importance of DNA methylation in ERV silencing in various cell types.  1.2.4 DNA methylation reprogramming in development During early development in mice, the genome undergoes two waves of DNA methylation reprogramming (Illustration 4) [79, 80]. In the first wave, fertilized oocytes undergo asymmetrical demethylation, in which the paternal genome becomes rapidly demethylated while the maternal genome loses methylation gradually until approximately 15  the morula stage. The genome remains hypomethylated until approximately the blastocyst stage where DNA methylation patterns are then re-established [79]. The second wave occurs when primordial germ cells (PGCs) migrate into the genital ridge. PGCs undergo dramatic DNA methylation reprogramming at ~E11.5-13.5, in preparation for germ line development, including establishment of gender-specific DNA methylation patterns [81]. The loss and re-establishment of DNA methylation patterns are crucial for normal development [80]. Although the kinetics of DNA methylation reprogramming is well described, the mechanism responsible for DNA demethylation remains controversial [82]. Recently, several different pathways have been characterized, which shed some light on the inner workings of this process.  16  Illustration 4. Mouse genome undergoes two waves of DNA methylation reprogramming. Soon after fertilization, non-imprinted paternal (blue) and maternal (red) genomes undergo asymmetrical global de-methylation. Repetitive elements (grey) also become hypomethylated; however, DNA methylation at imprinted genes (green) remains unchanged. DNA methylation patterns are re-established at approximately the blastocyst stage. The second wave of DNA methylation occurs in primordial germs cells as they migrate into the genital ridge (Figure modified from Wu and Zhang [82]).  One such mechanism involves the repair of single stranded breaks by the base excision repair pathway (BER). This mechanism has been implicated in active removal of 5mC from the paternal genome of early stage zygotes and also in PGCs. Studies found that deamination of 5mC is catalyzed by the enzymes activation-induced deaminase (AID) or apolipoprotein B mRNA-editing enzyme catalytic polypeptide-like protein (APOBEC), leading to a transition mutation resulting in a T-G mismatch [83, 84]. In turn, the mismatch triggers BER to replace the T with an unmethylated C and thus effectively removing the 5mC base. This is supported by the evidence that demethylation of both zygote and PGCs are mechanistically linked to the introduction of single stranded DNA breaks and activation of BER [85]. However, the function of AID and deamination may only be a part of the DNA reprogramming machinery, as PGCs do not express the required deaminases in vivo [85]. Thus other DNA demethylation pathways likely exist. Recently, the discovery and characterization of the Tet family of proteins have provided new insight into the biochemical basis of this process [86]. The three members of this family, Tet1, Tet2 and Tet3, all catalyze oxidation reactions converting 5mC to 5hydroxylmethyl-cytosine  (5hmC)  or  further  oxidation  products,  including  5-  carboxylcytosine (5caC) [87, 88]. These proteins have been implicated in the DNA methylation reprogramming process. Iqbal et al. reported that Tet3 converts 5mC to  17  5hmC in the paternal pronucleus at stages immediately following fertilization and proposed that 5hmC is important in the DNA methylation reprogramming process in that it is an intermediate between 5mC and unmethylated cytosine [89]. This proposed function of Tet proteins is consistent with previous observations that DNA demethylation begins soon after fertilization and the genome remains hypomethylated until after the blastocyst stage. Intriguingly, Tet1 and Tet2 are expressed at high levels in mESCs [90, 91] and in blastocysts [92], where 5hmC is relatively abundant. Furthermore, these enzymes play an important role in the maintenance of the pluripotent state [90, 93]. 5hmC may represent an intermediate step in either active or passive pathways leading to the loss of DNA methylation [89, 94]. Importantly, unlike 5mC, 5hmC itself is likely to be non-repressive, as none of the MBD proteins tested thus far are capable of binding to this mark [95, 96]. Thus, Tet proteins, via their catalytic activity, may function in part to counteract the repressive function of the de novo Dnmts, which are also highly expressed in mESCs [97]. This conjecture is supported by genome-wide studies, which reveal that Tet1 and 5hmC are enriched in the promoter regions and gene bodies respectively, of many actively transcribing genes [90, 98]. Taken together, it appears multiple pathways function in concert during particular stages of early embryonic development to reprogram the DNA methylation patterns. Therefore, given the potentially detrimental effects of deregulating expression of genes and repetitive elements during these periods, we hypothesized that other methods must exist to modulate transcriptional control of proviruses.  18  1.3 Chromatin biology 1.3.1 Covalent histone modifications In eukaryotes, nucleosomes form the basic building blocks of chromatin. 147bp of DNA is wrapped around core histones and condensed to form chromatin, which allows or restricts interactions with various nuclear molecules. Therefore, the regulation of chromatin structure is a critical step for transcriptional control. An important mechanism for epigenetic regulation of chromatin structure is the post-translational modification of histones. Covalent modifications, predominantly of N-terminal histone tails, including methylation [99], phosphorylation [100], acetylation [101, 102], ubiquitylation [103] and sumoylation [100], promote changes in chromatin state, affecting maintenance of genomic integrity and regulation of transcription [104]. Lysine methylation involves the deposition of one (me1), two (me2) or three (me3) methyl groups on the ε-amino group of lysine residues [105]. Lysine methylation marks are found at multiple positions on the histone tail, including at H3K4, H3K9, H3K27, K3K36, H3K79 and H4K20, which are thought to promote different transcriptional states (Illustration 5).  19  Illustration 5. Methylation of lysine residues on histone tails can confer different transcriptional states. In mouse, methylation of histone H3K4, H3K36 and H3K79 by Ash2l/Wdr5, Ezh2 and Dot1l respectively is correlated with transcriptional activation. Contrastingly, methylation of H3K9, H3K27 and H4K20 by Suv39h1/h2, Nsd1 and Suv420h1/h2 respectively is correlated with transcriptional repression. Enzymes other than the ones indicated above may catalyze covalent histone modifications on the same residue.  The focus of my thesis work is on H3K9 methylation, which is generally regarded as a mark of transcriptional silencing or heterochromatic regions. Many studies have established the enrichment of H3K9me2 and/or H3K9me3 at specific genomic regions, including telomeres, satellite repeats, ERVs and imprinted genes [106, 107]. For example, Peg13, an imprinted gene, is marked by H3K9me3 on the silent allele and by H3K4me3, a modification associated with activation, on the expressed allele, consistent with mono-allelic expression pattern [108]. H3K9 methylation is thought to exert a repressive effect by creating binding sites for other proteins including members of the heterochromatin protein 1 (Hp1) family of repressors. In mammals, there are three Hp1 isoforms: Hp1α, Hp1β and Hp1γ. They contain chromo domains, which bind to  20  methylated H3K9 and play an important role in H3K9me3-mediated silencing [109-111]. Interestingly, a recent study conducted by Dr. Irina Maksakova from the Lorincz lab found that in mouse, the transcriptional silencing conferred by H3K9 methylation can also occur independently of Hp1 proteins [112]. Thus, it is particularly important to identify and understand the enzymes responsible for deposition of these histone marks.  1.3.2 Histone H3K9 methyltransferases In mammals, there are six H3K9 specific lysine methyltransferases (KMTases) belonging to the Suv39 subfamily of SET domain-containing proteins. Five of these, including: Suv39h1/KMT1A,  Suv39h2/KMT1B,  G9a/KMT1C,  Glp/KMT1D  and  Setdb1/KMT1E/ESET are bona fide KMTases. Setdb2/KMT1F, the sixth family member, shares sequence homology with Setdb1 but little is known about the activity and function of this protein. Members of this subfamily share similarities in protein sequence and function. Unlike other KMTases, such as Ezh2 [113] or Suv420h1/h2 [114], which deposit H3K27me3 and H4K20me3 respectively, the H3K9 KMTases preferentially methylate histone octamers over nucleosomal histones [115-118]. All members contain the catalytic SET domain, pre- and post-SET domains, which function to stabilize the SET domain in order for efficient deposition of H3K9 methylation (Illustration 6) [104, 119].  21  Illustration 6. Structure of the Suv39 sub-family of SET domain containing KMTases. All members of this sub-family contain the pre-SET, the post-SET and the catalytic SET domains. In addition, Suv39h1 and Suv39h2 both contain chromo domains. G9a and Glp also contain the 7X and 8X ankryin repeats, respectively. Setdb1 contains 2 tudor domains, a methyl binding domain (MBD) and a signature bifurcated SET domain, containing a large insertion within the catalytic domain (indicated by the black box). Setdb2 sequence is highly homologous to Setdb1 and also contains the MBD, the pre-set, the post-set and the bifurcated SET domains.  The first H3K9-specific KMTase characterized was identified in a modifier screen for the position effect variegation phenotype in Drosophila melanogaster [120]. The enzyme was named Suv3-9 and is essential in repression of heterochromatic sequences. Its ortholog in fission yeast, Clr4, is necessary for silencing centromeres and mating type loci [121], showing an evolutionarily conserved function for this KMTase. The mouse orthologs, Suv39h1 and Suv39h2, show high degrees of sequence similarity to one another, and are required for the deposition of H3K9me3 at pericentromeric regions, including major satellite repeats [122, 123]. Suv39h1 is ubiquitously expressed, whereas Suv39h2 is specific to germ line cells and early embryonic cells [122, 123]. KO of both 22  genes leads to dramatically reduced viability [124] and loss of H3K9me3 and DNA methylation at pericentric heterochromatin [124-126]. However, DNA methylation patterns in euchromatic regions are unaffected, indicating that other Suv39 subfamily members may function in other genomic compartments [127]. G9a, another H3K9 KMTase, was originally discovered as a gene mapping to the human major histocompatibility complex locus [128] and was later characterized as a KMTase based on its conserved SET domain sequence [129]. G9a and the closely related G9a-like protein (Glp), form a heterodimeric complex and together mono- and dimethylate H3K9 [118]. The activity of each protein is dependent upon the presence of the other. The G9a/Glp complex is dispersed across euchromatic regions but excluded from heterochromatin [130, 131]. In G9a depleted mESCs, H3K9me2 is only found in heterochromatic regions, further showing that G9a and Glp are the KMTases that dimethylate euchromatic regions of the genome [130]. Interestingly, the G9a/Glp complex interacts with both maintenance and de novo DNA methyltransferases, suggesting that it may act in concert with DNA methylation [132, 133]. G9a and Glp stability also depends on their interaction with Wiz, a zinc finger protein [134]. Wiz acts as a bridge between G9a/Glp and other proteins such as CtBP, which in turn interacts with Hdac1, Hdac2 and Lsd1/Kdm1a among other proteins [134]. Wiz also appears to have DNA binding activity and is responsible for recruiting G9a/Glp to particular loci [134]. G9a null mice show severe growth and developmental defects, in addition to lethality between day E8.5 and E12.5 [130] and G9a-/- mESCs exhibit substantial change in DNA methylation patterns coupled with deregulation of many genes [135]. Interestingly, G9a appears to have an important role in silencing pluripotency associated genes, as indicated by a reversal of  23  differentiated G9a-/- cells to a pluripotent state [133, 136]. Consistent with this observation, MEFs are induced to become pluripotent by treatment with BIX01294, a G9a-specific methyltransferase inhibitor [137, 138]. Genome-wide survey of G9amediated H3K9me2 has revealed large domains of this mark across thousands of loci in the genome [139]. Taken together, G9a/Glp plays vital roles in epigenetic regulation in development. Setdb1/ESET, another euchromatic KMTase, contains a bifurcated SET domain, capable of depositing one, two or three methyl groups on histone H3K9 [140]. The Setdb1 gene was first mapped to human chromosome 1q21, a region commonly duplicated in cancers [141]. Previous studies have reported interactions between Setdb1 and other repressor proteins, including Hp1, Dnmt3a and Kap-1 in various cell types to establish H3K9me3-mediated silencing [142, 143]. Kap-1 is a co-repressor protein required for transcriptional repression in early embryos [144] and is a member of a silencing complex including the KRAB-ZFP protein ZFP809, necessary for silencing of MLV  retroelements  [145].  Co-immunoprecipitation  and  immunofluorescence  experiments found that Kap-1 also functions to recruit Setdb1 to specific loci in both heterochromatic and euchromatic regions [142, 146, 147]. Setdb1 can also be recruited by Mbd1 to form a complex with chromatin assembly factor 1 (Caf1) and is required for silencing by H3K9me3 deposition at gene promoter regions [148], indicating the presence of multiple mechanisms for recruiting this KMTase to target sequences. Setdb1 plays a vital role in early development, as Setdb1 KO mice show embryonic lethality between days E3.5 to E5 [149]. Maternal Setdb1 transcripts are abundant in oocytes and expression of the zygotic gene is high in blastocysts [149], 24  suggesting that this KMTase may be important in transcriptional regulation during early developmental stages. Intriguingly, Drosophila harboring mutations in dSetdb1 show a reduction of H3K9me2 and Hp1 binding on chromosome 4, which is particularly enriched  for  heterochromatic  sequences,  concomitant  with  chromosome-wide  transcriptional derepression [150]. These heterozygous mutants (homozygous mutants are not viable) also have severe defects in ovarian development in females [151], indicating that dSetdb1 is necessary for proper germ cell development [152]. Additionally, met2, the ortholog of Setdb1 in Caenorhabditis elegans has been reported to act with Hp1 and the NuRD complex for silencing genes essential for proper development [153], suggesting that Setdb1 possesses important evolutionarily conserved roles. Taken together, the above studies indicate that Setdb1 is an important transcriptional silencer, which is required during early embryonic development across species. However, until recently, little was known about the role of Setdb1 in ERV silencing.  25  1.4 Interplay between epigenetic mechanisms Histone modifications and DNA methylation can act in concert to confer transcriptional silencing. This interplay can be found across diverse eukaryotic species. In Arabidopsis thaliana, DNA methylation and H3K9 methylation are tightly coupled for silencing of specific genes and repetitive elements. Jackson et al. first identified the H3K9 KMTase KRYPTONITE (KYP), later shown to primarily deposit H3K9me2 [78, 154, 155]. Intriguingly, KYP mutants phenocopy mutants of the DNMT CHROMOMETHYLASE 3 (CMT3), which shows a loss of CpNpG trinucleotide methylation, a DNA methylation mark found in plants, concomitant with reactivation of endogenous retrotransposons [78]. CMT3 interacts directly with the Arabidopsis homolog of Hp1, which binds H3K9me marked histones [78]. Thus H3K9me regulates DNA methylation via directing Hp1 proteins that in turn interact with CMT3. In another study, genome-wide analysis confirmed that CMT3 and KYP-mediated DNA methylation and H3K9me mainly target transposable elements, which are dispersed through out the genome [156]. Interestingly, DNA methylation at constitutive heterochromatin, which is concentrated at chromocenters and pericentromeric repeats, is maintained in a distinct fashion, as it precedes the deposition of H3K9me [154] and requires DDM1 and MET1, homologs of the mammalian LSH and DNMT1, respectively. This indicates that different sequences employ distinct combinations of epigenetic pathways and in particular hierarchical orders to regulate chromatin structure and in turn transcriptional state. Further supporting this idea, another study found that two other KMTases related to KYP, SUV5 and SUV6, function with KYP in different combinations to establish DNA methylation and silencing at specific loci [157].  26  Another example of histone modifications acting in concert with DNA methylation can be found in Neurospora crassa, where the enzyme DIM-5, an H3K9specific KMTase, is required for the DNA methylation, normal growth and full fertility [77, 158]. DNA methylation in Neurospora crassa is catalyzed by the DNMT DIM-2, which requires the catalytic activity of DIM-5 to be maintained, as indicated by the DNA methylation defect observed in the DIM-5 mutant strain [77]. The H3K9me3 deposited by DIM-5 is recognized and bound by the Neurospora homolog of the HP1 proteins [159]. As in Arabidopsis, the HP1 protein in Neurospora is required for DNA methylation and is enriched at heterochromatic sequences [159]. HP1 functions as an adaptor molecule between the H3K9me3 mark and the DNMT, providing a link between the two pathways [159]. The evolutionary conservation of the interplay between these epigenetic pathways for repetitive elements silencing demonstrates the general importance of protecting the host genome against expansion of these intracellular parasites.  27  1.5 Role of epigenetics in proviral silencing in mESCs As previously mentioned, DNA methylation has been reported to play essential roles in proviral silencing in differentiated cell types and tissues [75]; however, strikingly, during stages of development when DNA methylation is reprogrammed, proviral silencing functions independently of this epigenetic mark, as a recent study has found that DNA methylation is dispensable for proviral silencing in mESCs [76]. Hutnick et al. showed that in differentiated cells, DNA methylation plays a vital role in silencing of IAP ERVs, the same is not true in mESCs. The authors found no increase in IAP mRNA and protein levels in Dnmt1 null, OCT4 positive mESCs. However, a subtle derepression was detected in OCT4 and Dnmt1 double negative cells, suggesting that previous studies reporting derepression of ERVs in DNA hypomethylated mESCs [160] may have been confounded by a sub-population of cells undergoing differentiation. Consistently, Dnmt1 null mESCs induced to differentiate by withdrawal of leukemia inhibitory factor showed a >100-fold upregulation of IAP expression [76]. Also, Dnmt1 conditional knockout (CKO) MEFs, neuronal cells and epiblast-derived stem cells all showed de-repression of IAP upon Dnmt1 depletion [76]. The authors also eliminated RNA interference-based silencing pathway as a candidate mechanism of repressing IAP elements in mESCs by demonstrating wildtype levels of IAP in Dnmt1 and Dicer double KO mESCs [76]. These findings are consistent with previously reported DNA methylation independent mechanism for silencing of introduced exogenous retroviral constructs [161]. Taken together, this suggests that during the stage in early embryonic development, from which mESCs are derived, proviral silencing is maintained by a DNA methylation independent  28  mechanism; however, until recently, this alternate ERV silencing pathway remained elusive. Genome-wide studies, including ChIP-seq-based approaches, have revealed the enrichment levels of multiple histone modifications across the genome in mESCs [108]. Interestingly, class I and II ERV LTR sequences are enriched with H3K9me2 and/or H3K9me3 in mESCs [108]. However, the specific KMTases targeting these elements had not been identified. As H3K9 methylation is an important silencing mechanism of repetitive elements in other organisms including yeast [162], plants [78], filamentous fungi [77] and animals [127], we decided to investigate its role on proviral silencing in mESCs. Previous studies revealed that deletion of Suv39h1 and Suv39h2 had no effect on H3K9 methylation, DNA methylation or expression of MLV and IAP elements [125, 163]. As the majority of ERVs integrate into euchromatin, we initially focused our studies on G9a and its partner Glp, which are widely dispersed in the euchromatic compartment [130, 131]. Shinkai and colleagues reported a dramatic decrease of H3K9me1 and H3K9me2 coupled with change in expression of ~300-400 genes in mESCs when either gene is deleted [135]. However, experiments, which aim to address the role of G9a in ERV repression and DNA methylation, had not been conducted. We utilized G9a-/- mESCs to investigate the role played by this KMTase in deposition of H3K9me2 at proviral elements and on proviral silencing per se. We found that G9a is required for the deposition of H3K9me2 and DNA methylation on class I and II ERVs (Figure 1A and C) [164]. Surprisingly, although proviruses show substantial loss of both of these repressive epigenetic marks, no increase in transcription was observed in G9a-/- mESCs (Figure 1B) [164]. 29  Dong et al. 2008 Figure 1. ERVs remain silent despite reduction of DNA methylation and H3K9me2 in G9a-/- mESCs. (A) Bisulfite sequencing revealed a reduction of DNA methylation at MLV and IAP LTR regions in G9a-/- mESCs. (B) Quantitative RT-PCR (qRT-PCR) found no change in mRNA levels of these elements in G9a-/- mESCs. The modest  30  increase in IAP expression detected in Dnmt1-/- mESCs may be due to differentiation of a subpopulation of cell in culture. Error bars indicate standard deviation between three technical replicates of one sample. C) H3K9me2 levels at ERVs were reduced in G9a-/mESCs, while H3K9me3 and Hp1α enrichments were retained at wildtype levels, implicating a potential role in proviral silencing. Error bars indicate standard deviation between three technical replicates of one sample.  Intriguingly, consistent with a previous genome-wide analysis, we found significant enrichment of H3K9me3 at the LTRs of class I and II ERVs [108]. Importantly, no decrease in H3K9me3 was detected in G9a KO cells [164], leaving open the possibility that this mark, rather than H3K9me2, is required for proviral silencing in mESCs (Figure 1C). Interestingly, H3K9me3 at these LTRs is absent in differentiated cells such as MEFs and neuroprogenitor cells, in which DNA methylation has been reestablished [108]. Taken together, histone modifications, such as H3K9me3, may be essential to proviral silencing in cell types where DNA methylation is being reprogrammed or is not acting as a robust silencing mechanism, thus necessitating the employment of alternate repression pathways. Therefore, we next set out to identify the KMTase depositing this mark and to determine the role if any that it plays in ERV silencing.  1.6 Thesis objectives Given the potentially deleterious effects of ERV activation, multiple parallel pathways have evolved to transcriptionally silence these elements. While proviral silencing in differentiated cells requires DNA methylation, alternative mechanisms may be utilized in cells undergoing DNA methylation reprogramming, such as mESCs. The main objective 31  of my thesis is to dissect the relationship between histone H3K9 methylation and DNA methylation in the context of proviral silencing. Our previous work demonstrated that despite a loss of H3K9me2 and DNA methylation, mESCs lacking G9a do not show reactivation of class I and II ERVs. Interestingly, enrichment of H3K9me3 was retained at wildtype levels in these cells, suggesting that this mark may function in the maintenance of silencing. In Chapter 3 of this thesis, I performed an in-depth study to identify the enzyme depositing H3K9me3 at these retroelements and determine its role in maintaining ERVs in a silent state. I also performed analysis of the DNA methylation state of such elements, to delineate the relationship between these epigenetic regulatory pathways in transcriptional repression of ERVs. This work sheds light on a previously unknown process, whereby host cells employ alternative mechanisms for maintaining these parasitic sequences in a silent state during stages of embryonic development when DNA methylation-mediated silencing is not sufficient. In chapter 4 of this thesis, I further investigate the role of H3K9 methylation in the establishment of silencing of newly integrated proviral elements. By utilizing MLVbased retroviral constructs, I was able to elucidate the epigenetic mechanisms necessary to initiate the repression of invading retroelements in mESCs. The findings further our knowledge of how mESCs defend against XRVs and provide insight into how histone modifications function in concert with de novo DNA methylation to direct transcriptional repression of specific sequences.  32  2. Materials and methods  33  2.1 Cell lines and culturing conditions J1 wildtype (129S4/SvJae), Dnmt1-/- [65], Dnmt3a-/- (clone 6aa), Dnmt3b-/- (clone 8bb) Dnmt3a and Dnmt3b DKO (Dnmt3a/b-/-) [67], Dnmt TKO (Dnmt1-/-,3a/b-/-) [160], TT2 wildtype (C57BL/6xCBA), G9a-/- (clones 2-3 and 22-10), G9a-/- with a wildtype (G9a-/Tg) [130] and a catalytic inactive transgene (G9a-/-Tg C1168A) [130], G9a CKO [165], R1 wildtype (129X1/SvJ x 129S1) and Suv39h1 and Suv39h2 DKO (Suv39h1/2-/-) [124], Setdb1 CKO and Setdb1 CKO with a wildtype transgene (Setdb1CKO Tg) [130, 166] mESCs were maintained in Dulbecco’s modified Eagle’s Medium (DMEM) supplemented with 15% fetal bovine serum (HyClone), 20mM HEPES, 0.1mM nonessential amino acids, 0.1mM 2-mercaptoethanol, 0.05mM penicillin, 0.05mM streptomycin, leukemia inhibitory factor (LIF) and 2mM glutamine on gelatinized plates. Cells were passaged every 48-72 hours. For treatment with 4-hydroxytamoxifen (4OHT), cells were cultured in media for 4 days with 800nM of 4-OHT and further cultured without the drug for an additional 2 days.  2.2 Generation of cell lines For this project, Setdb1 CKO mESCs were established by the Shinkai lab via standard gene targeting procedures, as summarized in Matsui and Leung et al.[166]. To generate the Setdb1 CKO mESC line, Cre recombinase and oestrogen receptor (Cre-ER) fusion gene was introduced into a clone containing targeted Setdb1 CKO and KO alleles.  34  2.3 Viral infections and flow cytometry For retroviral infections, approximately 1.5×106 Phoenix A packaging cells were transfected with plasmid DNA via calcium phosphate-mediated transfection, as previously described [166]. 9µg of retroviral construct plasmid DNA, 250ng of vesicular stomatitis virus glycoprotein plasmid DNA and 61µl of 2M CaCl2 was mixed with 2X HEPES buffered saline (50mM HEPES pH 7.05, 10mM KCl, 12mM Dextrose, 280mM NaCl and 1.5mM Na2HPO4) pH 7 ± 0.05 and applied to the media of Phoenix A cells. Cells were incubated at 370C for 9hrs before washing. 48hrs post-transfection, 500µl1100µl of viral supernatant was added to target cells supplemented with 4µg/ml of polybrene. Viral particles were concentrated by centrifugation for 45minutes at 3000rpm in a Heraeus Labofuge 400 centrifuge and cultured at 37oC. Infected cells were cultured for an additional 72 hours and subsequently analyzed by flow cytometry. Proviral copy number was determined by quantitative real-time PCR (qPCR) using genomic DNA isolated from infected pools and primers specific for the GFP region of the provirus (All primer sequences included in Table 1), in parallel with a control cell line harboring a single provirus (as determined by Southern blotting). All samples were normalized to the endogenous β-major globin gene. Trypsinized cells were resuspended in 500µl of PBS supplemented with 2% bovine calf serum and 1ug/ml propidium iodide and analyzed by flow cytometry using a BD LSRII flow cytometer. Data on at least 10,000 viable cells (as determined by electronic gating in the forward and side scatter channels) were collected for each sample and analyzed using FlowJo software (Treestar).  35  2.4 Generation of mESCs harbouring silent XRV To generate a population of mESCs harboring a silent XRV construct, cells were infected with the Mouse Stem Cell Virus (MSCV) - GFP retrovirus as described above. Infected cells were analyzed by flow cytometry at day 4 post-infection (PI) for confirmation of efficient transduction as determined by percentage of viable cells expressing the GFP reporter gene. Cells were further cultured to allow silencing of the construct. At day 14 PI, GFP negative cells were isolated by fluorescent activated cell sorting (FACS). The sorted pool of cells was enriched with cells harbouring a silent copy of the XRV. The copy number of the construct was determined by qPCR as described above.  2.5 Sequencing of reactivated MusD elements RNA was isolated as mentioned below from wildtype and Setdb1 CKO mESCs 6 days post-4-OHT treatment. RNA was then converted to cDNA using Superscript II according to manufacturers protocol (Fermentas). PCR was conducted on cDNA samples with primers specific for the MusD 5'LTR and 3'proximal region. The PCR product was resolved on an agarose gel and subsequently extracted with the Qiagen gel extraction kit. The purified product was TA cloned into pGEM-T vector (Promega) and sequenced. Sequences were aligned and phylogenetic trees constructed using MEGA4 software [167]  36  2.6 siRNA-mediated knockdown For knockdown experiments, mESCs were trypsinized, diluted in antibiotic free mESC media, seeded on a 6-well-plate (2x105 per well) and cultured overnight at 370C. On the day of the transfection, stock solutions (20µM) of siRNAs targeting Setdb1, Kap1, Mbd1, HDac1, Suv39h1, Suv39h2 and/or Glp (Dharmacon siGENOME SMARTpool or siGENOME siRNAs) were diluted to 2µM with siRNA buffer (Dharmacon). In two separate tubes, 50µl of 2µM siRNA was mixed with 50µl of OPTI-MEM (tube 1) and 4µl of DharmaFECT transfection reagent 1 was mixed with 96µl of OPTI-MEM (tube 2). Both tubes were incubated for 5 minutes at room temperature. Tube 1 and 2 were then mixed and incubated at room temperature for 20 minutes to allow for the lipid-siRNA complex to form. Fresh Antibiotic free media (1.8ml) was then mixed with the transfection reagent-siRNA solution and added to the well following removal of growth medium. Transfected cells were passaged 24 hours post-transfection. A second transfection was conducted with the same reagents on the following day. The efficiency of siRNA-mediated knockdown was analyzed on day 5 post-initial transfection by qRT-PCR with primers specific to the siRNA-targeted gene. The effects on reactivation of XRV and ERVs were analyzed on day 5 post-transfection via flow cytometry and qRT-PCR respectively.  37  2.7 Genomic DNA isolation and DNA methylation analysis Genomic DNA was isolated with the Bradley’s protocol [164]. Trypsinized cells were washed twice in PBS and resuspended in 500µl of PBS, 500µl of 2X Bradley’s solution and 5 units of Proteinase K and incubated overnight at 550C for cell lysis. The genomic DNA was subsequently precipitated by ethanol precipitation and pelleted by centrifugation. The DNA pellets were resuspended in TE with RNase A. To analyze the DNA methylation status, 200µg of genomic DNA was subjected to sodium bisulfite conversion using the EZ DNA Methylation-Gold kit (Zymo Research). Specific primers were employed in nested or semi-nested PCR reactions. PCR products were cloned via T/A cloning using the pGEM-T easy kit (Promega) and individual inserts were sequenced using the Big Dye® Terminator v3.1 Cycle Sequencing kit. Sequencing data was analyzed using Sequencher software (Gene Codes). The mean number of methylated CpGs (mCpGs)/molecule sequenced is presented for each set of samples.  2.8 Combined-bisulfite-restriction-analysis (COBRA) Genomic DNA was harvested from mESCs by the Bradley’s genomic DNA isolation method, as mentioned above. DNA was subjected to bisulfite conversion with the EZ DNA Methylation-Gold kit (Zymo Research). Specific primers targeting the MSCV 5’LTR were used to amplify the region of interest, which contains five restriction sites for TaqI endonuclease. The PCR product was digested with TaqI (restriction site: TCGA). Due to protection of methylated cytosine from bisulfite conversion to thymine, the enzyme would only cut methylated sites, as unmethylated sites no longer contained  38  cytosines in the sequence. The digested products were resolved on a 0.8-2% agarose gel. gDNA isolated from infected mESCs, with heavily methylated MSCV LTRs (as determined by bisulfite sequencing) is included as a positive control for the TaqI digest.  2.9 Nuclear and whole-cell extractions and western blotting analysis Nuclear extracts were prepared as previously described [168]. Cells were trypsinized and resuspended in 1.5ml of cold PBS. The cells were then pelleted for 10 seconds and resuspended in 400µl of cold Buffer A (10mM HEPES-KOH, 1.5mM MgCl2, 10mM KC1, 0.5mM dithiothreitol and 0.2mM PMSF (pH 7.8)). The cells were allowed to swell on ice for 10 minutes, and subsequently vortexed for 10 seconds. Samples were centrifuged for 10 seconds at ~16000 g, and the supernatant fraction was discarded. The pellet was resuspended in 20-100µl of cold Buffer C (20mM HEPES-KOH, 25% glycerol, 420mM NaCl, 1.5mM MgCl2, 0.2mM EDTA, 0.5mM dithiothreitol and 0.2mM PMSF (pH 7.8)) and incubated on ice for 20 min for high-salt extraction. Cellular debris was removed by centrifugation for 2 minutes at 40C and the supernatant fraction containing the DNA binding proteins, was stored at -800C. Whole cell extracts were prepared as described previously [169]. Trypsinized cells were resuspended in 2X Laemmli buffer and incubated at 100 °C for 10 min. Cells were then homogenized through a 25-gauge needle syringe for 10-15 repetitions. Extracts were stored at -800C. 50-100µg of extracts were loaded and run on 8-10%SDS-PAGE gels (Bio-Rad Mini Gels) at 80 to 120V then transferred to polyvinylidene fluoride (PVDF) membranes.  39  Blots were incubated facedown on primary antibodies at 40C overnight without agitation. Blots were subsequently incubated with secondary antibodies conjugated with fluorophores and analyzed with the Odyssey Infrared Imaging System (LI-COR Biosciences), according to the manufacturer’s protocol. Antibodies used include: G9a (1:2000; PPMX, A8620A), TFII-I (1:1000; kind gift of Ivan Sadowski), Dnmt1 (1:500; Imgenex, IMG-261A), β-actin (1:2000; MP Biomedical), H3K9me2 (1:200 Upstate, 07441) and H3K9me3 (1:200 Active Motif, 39161).  2.10 RNA extraction and RT-PCR RNA was isolated using RNeasy kit (Qiagen) according to the manufacturer’s protocol. DNaseI treated RNA was subject to first-strand cDNA synthesis using RevertAid H Minus kit (Fermentas) in the presence or absence of reverse transcriptase. qRT-PCR using primers of interest, or β-actin specific primers as an internal control were conducted with EvaGreen dye (Biotium) on an Opticon 2 thermal cycler (Bio-Rad). Relative expression levels were determined by normalizing to the β-actin gene.  2.11 Native chromatin immunoprecipitation Native ChIP was conducted as previously described [166]. This method is optimized for investigating the enrichments of histone proteins and shows lower false positive rates relative to cross-linked ChIP. To generate chromatin, 1x107 cells were harvested and washed in PBS. Cells were resuspended in 250µl of douncing buffer (10mM Tris-HCl  40  (pH7.5) 4mM MgCl2, 1mM CaCl2, 1X Protease inhibitory cocktail (PIC)) and homogenized through a 25-gauge needle syringe for 25 repetitions. Subsequently, 1.25µl of 50U/ml of MNase was added to the nuclei and incubated at 370C for 7 minutes. The reaction was quenched by 0.5M EDTA and incubated on ice for 5 minutes. Hypotonic lysis  buffer  (1ml)  (0.2mM  EDTA  (pH8.0),  0.m1M  benzamidine,  0.1mM  phenylmethylsulfonyl fluoride, 1.5mM dithiothreitol, 1X PIC) was added and incubated for 1 hour on ice. Cellular debris was pelleted and the supernatant was recovered. To generate pre-blocked beads for purification of immunoprecipitated material, 300µl of protein A and protein G Sepharose beads were mixed and washed twice with 1ml of IP buffer (10mM Tris-HCl (pH8.0), 1% Triton X-100, 0.1% Deoxycholate, 0.1% SDS, 90mM NaCl, 2mM EDTA, 1X PIC). Beads were blocked with 300µg of sonicated salmon sperm DNA and 750µg of BSA and rotated at 40C for 3 hours. Beads were then washed once more with IP buffer and finally resuspended in 1X volume of IP buffer. To pre-clear chromatin, 100µl of the blocked protein A/G beads were added to the digested chromatin fractions and rotated at 4oC for 2 hours. 100µl of the pre-cleared chromatin was purified by phenol-chloroform extraction and DNA fragment sizes were analyzed on a 1.5% agarose gel. Digested chromatin was divided into 1x106 cells equivalents per IP tube and the volume adjusted to 325µl with IP buffer. Antibodies specific for: H3K4me2 (3µg; Abcam, ab7766), H3K9ac (5µg; Upstate, 06-599), H3K9me2 (5µg; Abcam, ab1220), H3K9me3 (5µg; Active Motif, 39161), H4K20me3 (5µl; Active Motif, 39180) and H2A.Z (10µl, gift from Dr. Luc Gaudreau) were added to each tube and rotated at 4oC for  41  1 hour. The antibody-protein-DNA complex is precipitated by adding 20µl of the blocked protein A/G beads and rotated at 4oC overnight. The immunoprecipitated complex was washed twice with 400µl of ChIP Wash buffer (20mM Tris-HCl (pH 8.0), 0.1% SDS, 1% Triton X-100, 2mM EDTA, 150mM NaCl, 1X PIC), followed by a single wash with ChIP Final Wash Buffer (20mM Tris-HCl (pH 8.0), 0.1% SDS, 1% Triton X-100, 2mM EDTA, 500mM NaCl, 1X PIC). The protein-DNA complex was eluted by incubating the beads in 200µl of elution buffer (100mM NaHCO3, 1% SDS) at 68oC for 2 hours. IPed material was purified using the QIAquick PCR Purification Kit in 50 µl of elution buffer, according to the manufacturers protocol. Purified DNA was diluted 1:4 and analyzed by qPCR with EvaGreen dye and hot-start Taq polymerase (Fermentas) using 2 µl of template.  2.12 Cross-linked chromatin immunoprecipitation A cross-linked ChIP protocol was used to determine the enrichment of Setdb1. This method is optimized for investigating the enrichments of non-histone proteins, which bind to chromatin at lower affinity as compared to histone proteins. 1X107 mESCs suspended in 1ml of PBS with 10% fetal calf serum were fixed with in 1% formaldehyde at 370C for 5 minutes. The reaction was quenched by the addition of 100µl of 2.5M glycine and incubated on ice for 1 minute. Cells were then collected by centrifugation and washed twice in lysis buffer (10mM NaCl, 10mM Tris-HCl, 0.5% NP-40 and 1X PIC). The cell pellets were resuspended in SDS lysis buffer (50mM Tris-HCl, 10mM EDTA, 1%SDS and 1X PIC). After 10 minutes of incubation on ice, 400µl of dilution  42  buffer (150mM NaCl, 15mM Tris-HCl, 1mM EDTA, 1% Triton X-100 and 1X PIC) was added before the chromatin, contained in the lysate, was sonicated to fragment sizes of 300-1000bp using the Bioruptor sonicator (Diagenode). After insoluble materials were removed by centrifugation, chromatin was divided into equal volumes and the antibodies of interest were added. Anti-Setdb1 antibody (5µl Upstate 07-378) was used along with IgG antibody (Sigma) as a negative control. Samples were incubated overnight at 40C. Antibodies were subsequently collected with Dynabeads (Invitrogen). The beads are removed by using a magnetic bar, which attracts them to the side of the tube allowing for washing in 500µl of low salt buffer (150mM NaCl, 20mM Tris-HCl, 2mM EDTA, 0.1% SDS and 1% Triton X-100), high salt buffer (500mM NaCl, 20mM Tris-HCl, 2mM EDTA, 0.1% SDS and 1% Triton X-100) and then with LiCl buffer, a denaturing agent (250mM LiCl, 10mM Tris-HCl, 1mM EDTA, 1% sodium deoxycholate and 1% NP-40). Final washes were conducted with Tris-EDTA before elution buffer was added. The eluted Ab-DNA complex was incubated at 650C for reversal of crosslinks before treatment with RNase and Proteinase K. Remaining DNA is extracted by phenolchloroform and purified by ethanol precipitation. The resulting DNA was analyzed by qPCR with specific primers.  43  Table 1. Primers list Primer Target RT-PCR  Primer Sequence  β-actin  CTGGCTCCTAGCACCATGAAGATC TGCTGATCCACATCTGCTGG  MusD  GTGGTATCTCAGGA(G/A)GAGTGCC GGGCAGCTCCTCTATCTGAGTG  IAP  AAGCAGCAATCACCCACTTTGG CAATCATTAGATGTGGCTGCCAAG  MTA  ATGTTTTGGGGAGGACTGTG AGCCCCAGCTAACCAGAA  MLV  CTCGACCTCACCACCGATTACT AGCAGGGCCAGAGTTAATGA  Dazl  TACAAAATGCCCGCAGAAATAG CCGGACTCAACCTTCTCAATG  Dnmt1  TCGGCTGAACAACCCCGGCACCAC CTTCAGCACCATGGAGCGTCTGTAGG  Dnmt3a  GTCCGCAGCGTCACACAGAAGC TCTTTGGCGTCAATCATCACGG  Dnmt3b  AGGTTTATATGAGGGCACAGGAAGGC CATGTTGGACACGTCCGTGTAGTGAGC  Gfp  GGCGGATCTTGAAGTTCACC ACTACAACAGCCACAACGTCTATATCA  Setdb1  CAGCTCGCAGAGCTGGAGAC CTGTGAGTCTTTGGTGAAGGATTTC  Kap1  CGGAAATGTGAGCGTGTTCTC CGGTAGCCAGCTGATGCAA  Mbd1  GGATCCTTCTCATGGGTCCAC CCAGGACTCAGCCATAGATGCT  Hdac1  GTGGCCCTGGACACAGAGAT GCTTGAAATCCGGTCCAAAGT  Mage-a2  AAGGGAGGTCTCCATGCTGT TCTCCCATCTCAGGCTTCTC  44  Primer Target Primer Sequence Bisulfite analysis (Y= C or T, R= A or G) MLV 5’LTR  1st round: 2nd round:  IAP 5’LTR  1st round: 2nd round:  MusD 5’LTR  1st round: 2nd round:  MSCV 5’LTR  1st round: 2nd round:  Dazl  1st round: 2nd round:  TATTTTGTAAGGTATGAAAAAGTATTAGAGT AAATCRATAATCCCTAAACAAAAATCTCCA TAAATTTGTGTGTTTGTTAATGTTTTGATT AAATCRATAATCCCTAAACAAAAATCTCCA GGYGTTGATAGTTGTGTTTTAAGTGGTAAAT ATTCTAATTCTAAAATAAAAAATCTTCCTTA GATAGTTGTGTTTTAAGTGGTAAATAAATA ATTCTAATTCTAAAATAAAAAATCTTCCTTA AAATTTGAGTTTTGATTAGTATGAAATTGT AATCTAATATTTCTTCTTCCTTAAACCATA AAATTTGAGTTTTGATTAGTATGAAATTGT AACTTTAAACCCTTTCTTCTTCCACCTAAA TCTCRTCTCCTACCAAAACCACATATCCTA TAGGTTTGGTAAGTTAGTTTAAGTAAYGTT TTTGTAAGGTATGGAAAATATATAATTG TAGGTTTGGCAAGTTAAGTAAYGTT GGTTYGAGTTTTATTGATAGATAGATGGAT AACACCCTACAACTCAACTCTACTATAA GATTTTTGTTATTTTTTAGTTTTTTTAGGAT AAAATTCTCTCAACTAACCTAACTTATTTCT  ChIP analysis MSCV  AACCATCAGATGTTTCCAGGGTG TTCGGATGCAAACAGCAAGAGGC  IAP  CTCCATGTGCTCTGCCTTCC CCCCGTCCCTTTTTTAGGAGA  MTA  ATGTTTTGGGGAGGACTGTG AGCCCCAGCTAACCAGAAC  MLV  ATAAAGCCTCTTGCTGTTTGCATC TGGGCAGGGGTCTCCAAATCT  Major satellite  GACGACTTGAAAAATGACGAAATC CATATTCCAGGTCCTTCAGTGTGC  β-Major globin  CTGCTCACACAGGATAGAGAGGG GCAAATGTGAGGAGCAACTGATC  Gapdh  ATCCTGTAGGCCAGGTGATG AGGCTCAAGGGCTTTTAAGG  MusD  CCCTTCCTTCATAACTGGTGTCGCA TAGCATCTCTCTGCCATTCTTCAGG  Mage-a2  TTGGTGGACAGGGAAGCTAGGGGA CGCTCCAGAACAAAATGGCGCAGA  45  3. Role of H3K9 methylation in maintenance of proviral silencing#  #  - The results in this chapter were published in:  Toshiyuki Matsui*, Danny Leung*, Hiroki Miyashita, Hitoshi Miyachi, Hiroshi Kimura, Makoto Tachibana, Matthew C. Lorincz**, Yoichi Shinkai** (2010) Proviral silencing in embryonic stem cells require the histone methyltransferase ESET. Nature 464, 927-931 *-indicates co-first authorship; **-indicates co-corresponding authors  46  3.1 Introduction As discussed in detail in the introduction, the transcriptional silencing of ERVs is vital for maintaining the integrity of the host genome and transcriptome and therefore, host fitness. DNA methylation is essential for silencing of such elements in differentiated cell types [75, 76]. Interestingly however, proviral silencing in mESCs may involve alterative pathways [75, 76] such as covalent histone modifications. In distantly related eukaryotes, H3K9 methylation plays a critical role in transcriptional repression of retrotransposons [75, 76]. Taken together with studies in mESCs that reveal enrichment of H3K9me2 and/or me3 at class I and II ERV LTRs [108], these chromatin marks may function in maintaining silencing of such retroelements. In mammals, there are six H3K9-specific KMTases: Suv39h1, Suv39h2, G9a, Glp, Setdb1 and Setdb2. Other than Suv39h1/h2, which are dispensable for maintaining silencing and DNA methylation of IAP elements [125], the role that other H3K9-specific KMTases play has not been investigated. We have previously demonstrated that while mESCs lacking G9a or Glp show decreased H3K9me2 and DNA methylation at class I and class II ERVs. These elements retain wildtype levels of H3K9me3 and remain transcriptionally silent [164], indicating that deposition of the H3K9me3 mark by an alternative H3K9 KMTase may play a role in the silencing of ERVs in mESCs. However, mESCs lacking Suv39h1 and Suv39h2 (Suv39h1/h2 DKO) [125], which are responsible for H3K9me3 of pericentromeric heterochromatin, also show no reduction of H3K9me3 at IAP elements [127], implicating an alternative KMTase in this process. Therefore, my first focus was to identify the KMTase responsible for deposition of H3K9me3 at class I and II ERVs and subsequently to determine its role in proviral silencing. 47  3.2 Results 3.2.1 Identification of the H3K9 KMTase maintaining proviral silencing To identify the enzyme maintaining proviral silencing in mESCs, I first infected wildtype mESCs with MSCV-GFP (a Mouse Stem Cell Virus (MSCV) LTR/promoter driving expression of a Gfp reporter gene) and generated a pool of cells in which this exogenous retrovirus (XRV) was silenced by sorting GFP- cells. Subsequently, I conducted a siRNA-mediated knockdown (KD) screen. GFP- cells were treated with siRNAs targeting the known H3K9 KMTases: Setdb1, Setdb2, Glp, Suv39h1 and Suv39h2 (G9a was excluded as our previous findings revealed that this KMTase is not required for maintaining proviral silencing) [108]. Reactivation of Gfp was detected by flow cytometry 5 days post-siRNA treatment. The results revealed an upregulation of the silent XRV only upon Setdb1 KD, indicating that Setdb1 is required for maintaining XRV silencing (Figure 2A). The efficiencies of individual knockdowns were validated by qRT-PCR (Figure 2B).  48  Figure 2. Setdb1 is required for XRV silencing. (A) siRNA-mediated KD screen revealed that Setdb1 is the only H3K9 KMTase required for maintaining XRV transcriptionally silent. (B) Efficient KDs of all targeted KMTases were confirmed by qRT-PCR. Expression of all genes are normalized to β-actin levels and compared to the mean expression of the other non-targeted KMTase genes. Error bars indicate standard deviation between three technical replicates of one sample.  3.2.2 Setdb1 is required for maintenance of class I and II ERV silencing To determine whether Setdb1 is indeed the KMTase responsible for the silencing of class I and II ERVs, I utilized a conditional knockout (CKO) mESC line generated by our collaborators in Dr. Yoichi Shinkai’s laboratory [166]. Setdb1 CKO cells show a growth defect upon treatment with 4-hydroxytamoxifen (4-OHT), which induces Cre-ER recombinase-mediated deletion of Setdb1. Consistent with previous reports, Setdb1 CKO cells exhibit a lethality phenotype by 9 days post-4-OHT treatment [166]. Strikingly, qRT-PCR analysis of Setdb1 depleted cells at day 6 and 7 post-4-OHT treatment, in parallel with Suv39h/h2 DKO [125] or Dnmt1, Dnmt 3a and Dnmt 3b triple knockout (Dnmt TKO) [160] mESCs, revealed that transcription of class I (MLV) and Class II (IAP and MusD) ERVs were significantly induced upon Setdb1 deletion, while transcription levels of the class III ERV MTA (previously shown not to be marked by 49  H3K9me3) remained unchanged (Figure 3). Proviral reactivation following Setdb1 deletion was validated by northern analysis of this and additional Setdb1 CKO mESC clones derived from Setdb1 CKO embryos by Dr. Toshiyuki Matsui in Dr. Shinkai’s Lab [166]. In contrast, consistent with our observations, no increase in ERV expression was observed in the Suv39h1/h2 DKO line (Figure 3) [164]. Interestingly, only a weak induction of the analyzed ERVs was detected in Dnmt TKO cells, indicating that rather than employing DNA methylation-mediated repression, maintenance of proviral silencing in mESCs depend predominantly on Setdb1 (Figure 3).  50  Figure 3. Class I and II ERVs are reactivated upon Setdb1 depletion. Upregulation of MLV, IAP and MusD ERVs were detected by qRT-PCR in Setdb1 CKO mESCs 6 days and 7 days post-4-OHT treatment (grey bars). No reactivation of these elements was observed in Suv39h1/h2 DKO cells and only a subtle increase was seen in Dnmt1-/- and Dnmt TKO cells. MTA expression was not substantially induced in any of the lines analyzed. Error bars indicate standard deviation between three technical replicates of one sample.  3.2.3 Decreased levels of H3K9me3 and Setdb1 at class I and II ERVs upon Setdb1 deletion To determine whether Setdb1 is directly targeting class I and II ERVs and whether it is  51  the KMTase responsible for deposition of H3K9me3 at such elements, I conducted chromatin immunoprecipitation (ChIP) in Setdb1 CKO cells.  Following 4-OHT  treatment, I found a reduction of H3K9me3 at the LTRs of MLV, IAP and MusD elements, as well as at the downstream retroviral genes, which is rescued to wildtype levels with the introduction of a full-length wildtype Setdb1 transgene (Setdb1 CKOTG). Gapdh, a transcriptionally active gene, was used as a negative control amplicon, as it is not marked by H3K9me3 in these cells (Figure 4).  52  Figure 4. Reduction of H3K9me3 at class I and II ERVs upon Setdb1 depletion. ChIP experiments analyzed by qPCR showed that mESCs lacking Setdb1 show a reduction in H3K9me3 at 5’LTR and downstream regions of MLV, IAP and MusD. Gapdh functioned as a negative control amplicon, showing no enrichment of H3K9me3 at the promoter region of this expressed gene. Error bars indicate standard deviation between three technical replicates of one sample.  53  Figure 5. Suv39h1/h2 DKO cells show reduction of H3K9me3 at major satellites but not at IAP LTRs. Error bars indicate standard deviation between three technical replicates of one sample.  Suv39h1/h2 DKO mESCs were also analyzed by ChIP and showed no reduction of H3K9me3 at ERVs, indicating that Setdb1 is the only bona fide H3K9 tri-methylase depositing H3K9me3 at these elements (Figure 5). Consistent with our previous observations, H3K9me2 at these ERVs is either unchanged or slightly increased in the absence of Setdb1, revealing that Setdb1 is not responsible for deposition of H3K9me2 in vivo (Figure 6) [164]. Along with the loss of H3K9me3, I also found an increase in H2A.Z levels, a histone variant associated with transcriptional activation, at these elements (Figure 6). Furthermore, I confirmed that the LTRs of class I and II ERVs are direct targets of Setdb1, as KD resulted in a loss of enrichment of this KMTase (Figure 7). The enrichment levels of IgGs are indicative of background non-specific binding and minor changes of the low levels are often observed between different samples.  54  Figure 6. H3K9me2 is not deposited by Setdb1 in vivo. In mESCs lacking Setdb1, no change or increased levels of H3K9me2 was observed at MLV, IAP and MusD 5’LTR regions, indicating that this mark is deposited in a Setdb1-independent manner. The reactivation of these elements is concomitant with an increase in H2A.Z enrichment levels. Error bars indicate standard deviation between three technical replicates of one sample.  Figure 7. Setdb1 directly targets ERV 5’LTRs in mESCs. ChIP experiments demonstrated enrichment of Setdb1 at the 5’LTR regions of MLV and IAP elements, which was reduced to background levels upon depletion of Setdb1 by siRNA KD. Gapdh functioned as a negative control amplicon, which showed no enrichment of Setdb1 above background levels. Error bars indicate standard deviation between three technical replicates of one sample.  3.2.4 H4K20me3 is reduced at class I and II ERVs upon Setdb1 depletion Consistent with previous reports [108, 127], ERVs marked by H3K9me3 were also  55  enriched for H4K20me3 in mESCs. Strikingly, I found that this mark was significantly reduced after Setdb1 deletion, indicating that Setdb1 acts upstream of the KMTases Suv420h1 and Suv420h2, which are responsible for deposition of the H4K20me3 mark [117] (Figure 8). H3K9me3 may generally be required for H4K20me3 deposition, as Suv39h1/2 act upstream of Suv420h1/h2 at major satellite repeats [117, 170]. In contrast, while enrichment of H4K20me3 was lost at ERVs in Suv420h1/h2 DKO cells, H3K9me3 levels remained unperturbed and no increase in ERV expression was detected, indicating that Setdb1-mediated proviral silencing does not depend upon Suv420h1/h2 or H4K20me3.  Figure 8. Reduction of H4K20me3 at MLV and IAP LTRs in Setdb1 deleted mESCs. Error bars indicate standard deviation between three technical replicates of one sample.  3.2.5 No dramatic change in DNA methylation at ERVs upon Setdb1 deletion As mentioned above, surprisingly, ERVs are derepressed to a far lesser extent in Dnmt TKO than Setdb1 CKO mESCs. Furthermore, H3K9me3 enrichment at these elements is maintained in Dnmt TKO cells, suggesting that Setdb1 functions independently of DNA  56  methylation in this context. To determine if Setdb1 is required for DNA methylation of ERVs, I analyzed the DNA methylation states of the ERV LTR regions via bisulfite sequencing. The percentage of methylated CpGs was unchanged for IAP elements (100% versus 98.2%) and moderately decreased for MLV elements (100% versus 76.8%) following Setdb1 deletion (Figure 9). Thus, Setdb1 is not essential for DNA methylation of ERVs per se. However, as only a subset of these elements, of which there are ~2500 and ~70 copies in the mouse genome respectively, are potentially active (a relatively high GC->AT transition rate [171] reduces the transcription efficiency of “older” elements [172]), it remained a formal possibility that the DNA demethylation observed occurs specifically at “young” ERVs that are derepressed upon Setdb1 depletion. However, determining the DNA methylation states solely of the proviruses that are derepressed is technically challenging.  57  Figure 9. Modest to no change in DNA methylation patterns at ERV LTRs in Setdb1 CKO cells. Bisulfite sequencing analysis of the 5’LTR regions revealed a subtle or no change in DNA methylation levels for MLV and IAP elements respectively, in Setdb1 depleted relative to wildtype mESCs.  58  3.2.6 XRV reactivation concomitant with DNA demethylation in Setdb1 CKO cells To address whether proviral reactivation leads to change of DNA methylation status, I employed the MSCV-GFP construct described above, an MLV-based XRV with a strong enhancer/promoter that nevertheless is frequently silenced in mESCs early after infection via a DNA methylation-independent pathway [173]. The use of this XRV vector circumvented the problem encountered in studying ERVs by controlling the proviral load to an average of one copy of the XRV per cell. ChIP analysis of J1 wildtype and Dnmt TKO mESCs revealed significant enrichment of H3K9me3 at the MSCV provirus in both lines at day 4 PI, demonstrating that this mark is deposited at newly integrated XRVs in a DNA methylation-independent manner (Figure 10A). Furthermore, the provirus shows minimal de novo DNA methylation in WT J1 cells at this time point, as determined by COBRA analysis (Figure 10B) [174].  59  Figure 10. H3K9me3 on proviruses is deposited independently of DNA methylation. (A) XRV (Day 4 PI) and IAP LTRs are enriched for H3K9me3 at similar levels between J1 wildtype and Dnmt TKO mESCs. Error bars indicate standard deviation between three technical replicates of one sample. (B) COBRA analysis found that XRV LTRs (Day 4 PI) are DNA hypomethylated. Genomic DNA from pre-deletion Setdb1 CKO cells with heavily methylated MSCV LTR regions (as determined by bisulfite sequencing) was included as a positive control for TaqI restriction digestion.  To determine whether Setdb1 is required for silencing of the MSCV provirus, I infected Setdb1 CKO and Setdb1 CKOTG cells with the MSCV-GFP XRV, and isolated GFP negative cells at day 14 PI. This population of mESCs, which includes both uninfected cells and cells harbouring a silent provirus (with a mean proviral copy number of 0.86 copies per cell (Figure 11B)) was expanded for further analyses. Following Setdb1 deletion in this pool, the percentage of viable GFP+ Setdb1 CKO cells increased dramatically, as did the level of proviral RNA (Figure 11A, C-D). Consistent with the observed changes in ERVs, H3K9me3 and H4K20me3 enrichment across the MSCV provirus was concomitantly reduced with Setdb1 deletion, while H3K9me2 remained unchanged and H2A.Z levels were increased (Figure 12).  60  Figure 11. Maintenance of XRV silencing requires Setdb1. (A) Setdb1 CKO mESCs harbouring a silent XRV, showed proviral reactivation upon 4-OHT treatment, which resulted in a gradual increase in the percentage of GFP positive cells. No reactivation was observed in untreated Setdb1 CKO cells and Setdb1 CKOTG cells ± 4-OHT. (B) Copy number qPCR determined that the Setdb1 CKO cells harbour an average of 0.86 copy of the XRV per cell. Error bars indicate standard deviation between three technical replicates of one sample. (C) Contour plots of flow cytometry analyses showing reactivation of XRV in Setdb1 depleted cells. (D) qRT-PCR validating the observations by flow cytometry. An increase in Gfp mRNA was detected in Setdb1 CKO cells 7 days post-4-OHT treatment. Error bars indicate standard deviation between three technical replicates of one sample.  61  Figure 12. Reduction of H3K9me3 and H4K20me3 at MSCV LTR in Setdb1 deleted mESCs. Consistent with observations in ERVs, ChIP analysis revealed a reduction of H3K9me3 and H4K20me3 at the MSCV LTR and downstream Gfp gene in Setdb1 deleted cells along with a slight increase of H3K9me2 and H2A.Z enrichment. Gapdh was analyzed as a negative control amplicon. Error bars indicate standard deviation between three technical replicates of one sample.  To determine whether derepression of the MSCV provirus was accompanied by DNA demethylation, “wildtype” untreated and GFP+ cells sorted at day 6 post-4-OHT treatment were analyzed by bisulfite sequencing. While the LTR was densely methylated before 4-OHT treatment, the sorted/GFP+ population showed a significantly lower level of DNA methylation in this region, indicating that reactivation of proviral expression following Setdb1 depletion is accompanied by loss of DNA methylation (Figure 13).  62  Figure 13. Reactivation of XRV is coupled with DNA demethylation. Bisulfite sequencing analyses found that GFP+ sorted mESCs, reflecting XRV reactivation, revealed a reduction of DNA methylation (100% versus 34.2%) at the MSCV 5’LTR region.  3.2.7 Younger ERVs are more prevalently derepressed in Setdb1 CKO mESCs ERVs exist in high copy numbers in the mouse genome, yet not all elements of a subfamily are identical. Older elements may acquire more mutations and lose their capacity for transcriptional activation. Sequence analysis of the MusD elements expressed in Setdb1 CKO cells before and after Setdb1 depletion revealed a higher degree of sequence diversity in the latter, indicating that additional MusD ERVs were reactivated upon Setdb1 deletion (Figure 14A). However, comparison to the genomic sequences of all MusD elements reveals that in both cases, the expressed MusD elements show highest sequence similarity to relatively recent insertions, consistent with the observation that 63  mutations in the LTR region of older retroelements reduces their transcription potential (Figure 14B) [172].  Figure 14. Setdb1 deleted cells express a greater diversity of MusD elements than wildtype cells. (A) A phylogenetic tree of MusD cDNA expressed in wildtype and Setdb1 deleted cells is shown, with the addition of coding-competent MusD elements [10] as well as two recently retrotransposed MusD elements (Dac1J and Dac2J) [175] as reference sequences. 86 percent of the MusD sequences originating from wildtype are highly similar in sequence (boxed) and cluster with the recently transposed copies. While the cDNA clones expressed in Setdb1 KO cells show greater sequence diversity, most are relatively closely related to one or the other of the coding competent elements. (B) A  64  MusD phylogenetic tree was constructed for the region amplified using the neighbourjoining algorithm. The coding-competent MusD elements are also shown (stars). The subset of MusD copies expressed in wildtype and to a lesser extent in Setdb1 KO cells is clearly biased towards relatively “young” genomic elements. (Construction of phylogenetic trees was assisted by Dr. Irina Maksakova)  3.2.8 Depletion of Kap-1 phenocopies Setdb1 Setdb1 is a member of a large co-repressor complex that includes Kap-1 [146], Hp1 proteins [147] and subunits of a NuRD-like complex [176]. To determine whether Kap-1 or Mbd1, both of which have been reported to recruit Setdb1 to genes in other cell types, are required for proviral silencing in mESCs, I utilized siRNAs-mediated KD of Setdb1, Kap-1, Mbd1 or Hdac1, a subunit of the NuRD-like complex, to determine whether these proteins are required for proviral repression. Strikingly, as observed for Setdb1 KD, Kap1 KD yielded a dramatic increase in expression of class I and II ERVs and the XRV elements (Figure 15A-B). Concomitant with transcriptional induction, a decrease in Setdb1 binding and H3K9me3 was detected at these elements (Figure 16). In contrast, depletion of Mbd1 or Hdac1 yielded no increase in proviral reactivation, despite efficient KD (Figure 15B-C). Taken together, these results indicate that Kap-1 functions to recruit Setdb1 to the retroviral LTR targets.  65  66  Figure 15. Depletion of Kap-1 phenocopies depletion of Setdb1. (A) mESCs harboring silent MSCV-GFP proviruses were transfected with siRNAs specific for Setdb1, Kap-1, Mbd1 and Hdac1 and analyzed by flow cytometry, which revealed XRV reactivation in Setdb1 and Kap-1 KD. Scrambled siRNA was used as a negative control. (B) RNA was isolated from transfected cells and analyzed by qRT-PCR with primers specific for the MSCV, MLV and MusD 5’LTRs. A significant increase in proviral expression was detected in the Setdb1 and Kap-1 KD samples only. Primers specific for the endogenous Dazl gene were used as a negative control. No proviral reactivation was observed in Mbd1 and Hdac1 KD. Error bars indicate standard deviation between three technical replicates of one sample. (C) Efficient KDs were confirmed by qRT-PCR. Error bars indicate standard deviation between three technical replicates of one sample.  Figure 16. Setdb1 recruitment to proviral 5’LTR is dependent on Kap-1. (A) NativeChIP showed a reduction of H3K9me3 enrichment at MSCV, MLV, MusD and IAP LTR regions upon siRNA KD of Kap-1. No reduction of H3K9me3 was detected at major satellite repeats, which are deposited in a Setdb1-independent manner. Gapdh served as a negative control amplicon. (B) Setdb1 enrichment levels, as determined by cross-linked ChIP, at both ERV and XRV LTR regions were reduced after depletion of Kap-1.  67  3.3 Discussion As our previous experiments revealed, the lost of H3K9me2 and DNA methylation in G9a-/- mESCs did not lead to transcriptional reactivation of class I and II ERVs. Here, I conducted a siRNA-mediated KD screen targeting the remaining possible KMTase candidates and discovered that only Setdb1/Eset is necessary for maintaining silencing of introduced retroviral constructs. Further analysis was carried out using Setdb1 CKO mESCs (generated by our collaborator Dr. Yoichi Shinkai), which confirmed that Setdb1 is indeed the KMTase depositing H3K9me3 at class I and II ERVs. Interestingly, the loss of H3K9me3 at ERVs in Setdb1 depleted cells was coupled with a reduction of H4K20me3, another histone modification correlated with transcriptional repression. This observation suggests that this chromatin mark, deposited by Suv420h1/h2, functions in the same silencing pathway as Setdb1. However, analysis of Suv420h1/h2 DKO cells reveals that this mark is dispensable for proviral silencing. Surprisingly, proviral silencing occurs independently of DNA methylation, as indicated by the high levels of ERV expression in Setdb1 CKO mESCs versus the subtle derepression in Dnmt TKO cells [166]. Also, bisulfite analysis revealed no apparent change in DNA methylation at ERV LTRs between wildtype and Setdb1 CKO cells [166]. However, due to technical challenges of surveying only reactivated ERVs, the lack of DNA methylation defect could be explained by the number of copies of individual ERV types present in the genome. To overcome this issue, I utilized the MSCV-GFP retroviral vector to generate Setdb1 CKO cells harbouring a silent XRV. By only analyzing cells that become GFP positive upon Setdb1 deletion, indicative of XRV  68  reactivation, I discovered that the upregulation of the MSCV provirus occurs concomitant with a reduction in DNA methylation. Taken together these results indicate that Setdb1deposited H3K9me3 functions independently of DNA methylation for maintenance of proviral silencing. To address the mechanism of recruitment of Setdb1 to ERV LTRs, I treated mESCs with siRNA targeting Kap-1. Kap-1, which interacts directly with Setdb1 and KRAB Zinc-finger proteins in various cell types [142] represses genes in early mouse embryos [144] and is a member of a complex necessary for silencing of MLV retroelements [145] in mESCs. Depletion of Kap-1 resulted in derepression of ERVs and XRVs, phenocopying depletion of Setdb1. This increase in proviral expression is coupled with a reduction in Setdb1 and H3K9me3 enrichment at target LTRs, therefore showing the involvement of Kap-1 in recruitment of Setdb1 and proviral silencing in mESCs. This observation is consistent with that reported by Rowe et al. [177]. Consistent with previous studies, we found that the mechanism of ERV repression in mESCs was strikingly different in differentiated cells. Setdb1-deficient MEFs isolated from Setdb1 CKO embryos by our collaborator Dr. Toshiyuki Matsui, demonstrated no significant increase in ERV transcription [166]. This finding was not unexpected, given that Mikklesen et al. had found that both class I and II ERVs were devoid of H3K9me3 in differentiated cells such as MEFs and neuronal progenitor cells [108]. Interestingly, studies have found that DNA methylation is essential for proviral silencing in mESCdifferentiated cells, MEFs and in later stages of embryogenesis [76]. Taken together, these data indicate that mESCs employ Setdb1 to maintain proviral silencing, which is then superseded by other pathways upon cellular differentiation. 69  In summary, these findings reveal that Setdb1 is a critical downstream effector of a Kap-1-dependent silencing pathway that acts on ERVs as well as XRVs in mESCs. The reason why Setdb1 is required for proviral silencing specifically in mESCs remains a mystery, particularly in light of the fact that ERVs are generally densely DNA methylated in these cells. DNA methylation may not efficiently mediate silencing of retroelements in mESCs as a result of high turnover, or the absence of co-repressors, such as Mbd proteins. Regardless, ERVs are variably demethylated early in embryogenesis and in primordial germ cells, in conjunction with the genome-wide demethylation that occur at these stages [178, 179]. Given that Setdb1 is highly expressed in oocytes and preimplantation embryos [149] and is required for deposition of the H3K9me3 mark at class I and II ERVs, I propose that this KMTase plays a critical role in silencing of ERVs specifically during those stages in development when DNA methylation is reprogrammed.  70  4. Role of H3K9 methylation in establishment of exogenous retroviral silencing#  #  - The results in this chapter were published in:  Danny Leung, Kevin Dong, Irina A. Maksakova, Preeti Goyal, Ruth Appanah, Sandra Lee, Makoto Tachibana, Yoichi Shinkai, Bernhard Lehnertz, Dixie L. Mager, Fabio M.V. Rossi and Matthew C. Lorincz. Lysine methyltransferase G9a is required for de novo DNA methylation and the establishment but not maintenance of proviral silencing. PNAS 2011: 1014660108v1-201014660.  71  4.1 Introduction MLV-based retroviral vectors have been widely used in both the laboratory and the clinical settings, primarily due to their ability to efficiently infect target cells and in turn to effectively deliver their genetic payload. Many versions of these vectors have been engineered to optimize expression, including the MSCV vector, which harbours mutations that allow for improved expression in mESCs [180]. However, despite such sequence modifications, this vector is readily silenced [181]. Therefore, understanding the mechanism for silencing of newly integrated proviruses may have implications across multiple disciplines. In addition to transcriptional repression, XRVs are frequently de novo DNA methylated and bound by Mbd protein [174, 182], which indicates the involvement of the de novo Dnmts in establishing silencing of these elements. However, other epigenetic pathways may also function in concert with DNA methylation. As mentioned previously, we have shown that in mESCs lacking the “euchromatic” KMTase G9a [164], ERVs show reduced H3K9me2 and DNA methylation. Although these epigenetic marks are neither necessary nor sufficient for maintaining transcriptional repression of class I and II ERVs, little is known about the role of H3K9me2 or de novo DNA methylation in the establishment of silencing of newly integrated proviruses in mESCs. Taken together with previously reported interactions between G9a and de novo Dnmts [133], the two epigenetic silencing pathways may function together to repress these invading retroviral sequences. In this study, I delineated the role played by G9a in the establishment of proviral silencing and de novo DNA methylation, utilizing the MSCV retroviral vector as a model.  72  4.2 Results 4.2.1 G9a is required for the establishment of silencing of an MLV-based vector To determine whether G9a plays a role in repressing the expression of newly integrated proviral elements, wildtype (TT2) and G9a-/- mESCs [130] were transduced with the MLV-based retroviral vector MSCV-GFP and passaged for further analysis. This vector harbors mutations in the LTR as well as a tRNAGln in place of the tRNApro PBS, the latter of which ablates binding of the stem cell-specific repressor complex [18, 183] to the overlapping silencer element and thus is “optimized” for expression (Figure 17A). The absence of G9a expression in the G9a-/- line was confirmed by western blotting [184]. Following infection, the percentage of cells expressing GFP was analyzed on a weekly basis by flow cytometry (Figure 17B).  73  Figure 17. Retroviral vector and infection schema. (A) Map of the 5’ region of the MSCV-GFP retroviral vectors. The stem cell-specific silencer binds to the MLV/wildtype PBS but not the MSCV PBSs (sequence differences are underlined). (B) mESCs were infected with MSCV-GFP retrovirus and passaged for further analyses.  While the MSCV vector was expressed in a significant number of cells early after infection, a dramatic decrease in the percentage of GFP+ cells was observed over time in culture in wildtype cells, as expected (Figure 18A-B). Interestingly, the MSCV vector was not efficiently silenced over time in G9a-/- cells (Figure 18A-B). Consistent with the flow cytometry analyses, qRT-PCR analysis at day 15 PI reveals a significantly higher 74  level of GFP expression in G9a-/- cells than in the wildtype parent line (Figure 18C). Both pools of infected cells harbor a mean of ~1 proviral copy per cell, ruling out the possibility that the observed difference in expression was the result of a difference in proviral content (Figure 18D).  Figure 18. G9a-/- mESCs show a defect in silencing of XRV. (A) TT2 and G9a-/mESCs were infected with MSCV-GFP and analyzed by flow cytometry at multiple time points PI. Uninfected cells were also analyzed as a negative control. The histograms shown are of TT2 and G9a-/- cells at days 4 and 24 PI. (B) Analysis of GFP expression 75  by flow cytometry at successive time points PI. More than 50% of GFP+ wildtype cells at day 4 PI become GFP- by day ~15 PI, indicating silencing of the XRV provirus. Contrastingly, the percentage of GFP+ G9a-/- cells remain generally unchanged at all analyzed time points PI. (C) Proviral expression was independently determined at day 15 PI by qRT-PCR in the presence or absence of reverse transcriptase (+/-RT). Error bars indicate standard deviation between three technical replicates of one sample. (D) Proviral copy number was determined by qPCR using primers specific for the GFP gene or the endogenous -major globin gene and normalizing to a control cell line harboring a single copy GFP transgene. Error bars indicate standard deviation between three technical replicates of one sample.  4.2.2 Introduction of a catalytically active G9a transgene rescues the silencing defect, while Suv39h1/h2 are not required for silencing of newly integrated proviruses To verify whether the deletion of G9a per se or its catalytic activity is responsible for the observed phenotype, I infected G9a-/- lines stably expressing a wildtype (G9a-/-Tg) or a catalytically inactive G9a transgene (G4) [130] with MSCV-GFP virus, as described above. Importantly, expression of exogenous G9a “rescues” the silencing defect observed in G9a-/- mESCs, confirming that this KMTase is indeed required for the establishment of proviral silencing (Figure 19A). Whereas, the catalytic mutant behaves similarly to the G9a-/- mESCs, thus indicating that the catalytic activity of G9a is required for this process (Figure 19A). To determine whether mESCs deficient in the “heterochromatic” H3K9 KMTases, Suv39h1 and Suv39h2, show a similar silencing defect, Suv39h1/2 DKO mESCs were infected with MSCV-GFP, as above, and analyzed at successive days PI. In contrast to the G9a-/- line, and consistent with a previous report showing no decrease in DNA methylation of endogenous MLV [163], Suv39h1/2 DKO cells show no defect in silencing of introduced MLV provirus, relative to the wildtype R1 parent line (Figure 19B). In fact, proviral silencing is somewhat more efficient in the Suv39h1/2  76  DKO than the wildtype line, perhaps due to the re-localization of Hp1 proteins [185] and Dnmt3b [163] away from the pericentromeric compartment in these cells. Regardless, taken together, these data reveal that G9a plays a role in silencing of MLV-based proviruses, while Suv39h1/h2 are not required for silencing of such XRVs in mESCs.  Figure 19. Introduction of wildtype but not catalytically inactive G9a transgene rescues silencing defect. (A) TT2 (G9a+/+) and G9a-/- cells expressing a G9a wildtype (G9a-/- Tg) or catalytically inactive (G9a-/- Tg C1168) transgene were infected in parallel with MSCV-GFP and analyzed for GFP expression by flow cytometry at successive time points PI. The silencing defect is rescued only in the G9a-/- Tg line. Mean proviral copy number/cell for each infected population was determined by qPCR as above. (B) R1 (Suv39h1/2+/+) and Suv39h1/2-/- mESCs were infected in parallel with MSCV-GFP and GFP expression was analyzed by flow cytometry at successive time points PI. No silencing defect was detected in cells lacking these KMTases. Mean proviral copy number/cell for each infected population was determined as above.  77  4.2.3 The silencing defect in G9a-/- mESCs phenocopies that observed in de novo DNA methyltransferase mutants In wildtype mESCs, DNA methylation of MLV-based vectors increases with prolonged passage in culture [186] and dense DNA methylation is sufficient for proviral silencing in somatic and EC cells [182, 186, 187]. To determine whether mESCs deficient in Dnmt3a and Dnmt3b have a silencing defect similar to that observed in the G9a-/- line, wildtype (J1) mESCs were infected with VSV-G pseudotyped MSCV-GFP virus in parallel with Dnmt3a and Dnmt3b DKO (Dnmt3a/b-/-) cells. Consistent with the observations in wildtype TT2 mESCs, proviral expression in J1 mESCs decreased substantially in the first two weeks post-infection (Figure 20A). In contrast, Dnmt3a/b DKO cells demonstrated a silencing defect similar to that observed in G9a-/- mESCs (Figure 20A). To determine which of the de novo Dnmt is required for silencing, I infected Dnmt3a-/(6aa) and Dnmt3b-/- (8bb) mESCs and analyzed them by flow cytometry at successive days PI, as above. Intriguingly, whereas Dnmt3a-/- cells demonstrated a similar silencing defect to that observed in G9a-/- cells, Dnmt3b-/- cells showed a relatively modest silencing defect (Figure 20B). qPCR analysis of the relative proviral copy number showed that the differences in the percentages of GFP+ cells in each line cannot be explained by differences in proviral load (Figure 20B). Taken together with the observation that Dnmt1-deficient cells have only a modest defect in de novo DNA methylation of newly integrated proviruses [65], these results indicate that both G9a and Dnmt3a2 (the predominant isoform of Dnmt3a in mESCs (34)) are required for the establishment of proviral silencing in mESCs.  78  Figure 20. Dnmt3a/3b-/- mESCs phenocopy silencing defect of G9a null cells. A) J1 (Dnmt3a/3b+/+) and Dnmt3a/3b-/- mESCs were infected in parallel with MSCV-GFP and GFP expression was analyzed by flow cytometry at successive time points PI. Silencing defect similar to that of G9a-/- cells was detected. Mean proviral copy number/cell for each infected population was determined as above. B) J1, Dnmt3a-/- and Dnmt3b-/- mESCs were infected in parallel with MSCV-GFP and GFP expression was analyzed by flow cytometry at successive time points PI. Dnmt3a-/- cells show a silencing defect as seen in the DKO cells, whereas Dnmt3b-/- shows a modest phenotype. Mean proviral copy number/cell for each infected population was determined as above.  79  4.2.4 H3K9me2 is decreased in the 5’LTR/promoter region of the MSCV-GFP provirus in G9a-/- mESCs In G9a-/- cells bulk levels of H3K9me2, as measured by western blotting analyses, were reduced as expected, while Suv39h1/2-/- cells showed no decrease in this epigenetic mark (Figure 21A) [125]. To determine whether the silencing defect observed in G9a-/- cells is associated with a reduction in proviral H3K9 methylation, native ChIP experiments were conducted on chromatin isolated from wildtype TT2 and G9a-/- cells at day 22 PI, using antibodies raised against H3K9me2 or H3K9me3.  The specificity of these  antibodies for ChIP studies was confirmed using primers specific for the Mage-a2 gene promoter and major satellite repeats, respectively. Consistent with previous observations [130, 164], the Mage-a2 gene shows a high level of G9a-dependent H3K9me2 enrichment relative to the IgG control, but very low H3K9me3 enrichment (Figure 21B). In contrast, major satellite repeats, previously found to be marked by H3K9me3 independent of G9a [125, 164], show a high level of H3K9me3 in wildtype and G9a-/cells (Figure 21D). Strikingly, analysis of the MSCV LTR in the same chromatin isolates reveals a >3-fold decrease in H3K9me2 enrichment in the G9a-/- line at day 22 PI (Figure 21E). In contrast, the level of H3K9me3 enrichment in the same region, while greater than that observed with control IgG, is not significantly different between the two lines, indicating that an alternative KMTase is responsible for H3K9me3 of proviral elements (Figure 21E). This observation is consistent with my previous findings that Setdb1 is responsible for deposition of H3K9me3 at proviruses in mESCs [166]. In contrast, analysis of the GFP region of the provirus reveals high levels of both H3K9me2 and H3K9me3 in both  80  cells lines, indicating that the transcribed region of the provirus is marked by H3K9me2/3 independent of G9a catalytic activity (Figure 21E). The presence of H3K9me2 and/or H3K9me3 in gene bodies has been reported previously [188, 189]. ChIP analysis using antibodies specific for H3K4me2 and H3K9ac, marks typically associated with transcriptionally active genes, revealed the opposite pattern, with a higher levels of enrichment in the Mage-a2 and proviral 5’LTR promoter regions in the G9a-/- line than the wildtype (Figure 21F). The presence of these active marks in the proviral LTR in the wildtype line at levels significantly above background presumably reflects the presence of constitutively expressing proviral integrants in this pool of infected cells. Taken together, these results indicate that G9a contributes to proviral silencing in mESCs by acting directly on the proviral promoter region.  81  Figure 21. H3K9me2 at the MSCV LTR is reduced in G9a-/- mESCs. (A) Depletion of H3K9me2 in the G9a-/- line was confirmed by western blotting using extract isolated from TT2 and G9a-/- cells. In contrast, depletion of H3K9me2 was not observed in Suv39h1/2-/- mESCs relative to the wildtype R1 parent line. (B) Chromatin isolated from MSCV-GFP infected TT2 and G9a-/- mESCs was subject to ChIP using H3K9me2- and H3K9me3-specific antibodies. Non-specific IgG was used as a control. qPCR analyses  82  showed a loss of H3K9me2 in cells lacking G9a and low enrichment of H3K9me3 at the Mage-a2 locus. Error bars indicate standard deviation between three technical replicates of one sample. (C) The same samples were amplified with primers specific for the Gapdh gene, which functioned as a negative control amplicon and (D) major satellite repeats, which showed a reduction of H3K9me2 but not H3K9me3 enrichment in G9a-/- cells. Error bars indicate standard deviation between three technical replicates of one sample. (E) The same samples were amplified with primers specific for MSCV 5’LTR and GFP regions of the introduced proviruses, revealing a decrease in H3K9me2 at the 5’LTR in the G9a-/- line. Error bars indicate standard deviation between three technical replicates of one sample. (F) ChIP was also conducted with H3K4me2- and H3K9ac-specific antibodies and analyzed by qPCR with primers specific for the MSCV 5’LTR and Magea2 gene. An increase of both active marks was observed in G9a-/- mESCs. Error bars indicate standard deviation between three technical replicates of one sample.  4.2.5 G9a is required for efficient DNA methylation of MLV-based XRV Previously, we showed by bisulfite sequencing that the 5’LTR/PBS region of the class I and II ERVs are hypomethylated in G9a-/- relative to wildtype cells [164]. To determine whether DNA methylation of the MSCV vector is also dependent upon G9a, genomic DNA was isolated from MSCV-GFP infected TT2 and G9a-/- cells at days 4 and 18 PI. Bisulfite sequencing analysis was conducted using primers specific for the MSCV 5’LTR/PBS region. While an increase in proviral DNA methylation was detected in both TT2 and G9a-/- cells with passage in culture, a significantly lower level of DNA methylation, ~3-fold, was detected in the G9a-/- cells compared to the wildtype cells at both time points (Figure 22A). Strikingly, the level of proviral DNA methylation in the G9a-/- cells at day 18 PI is similar to that of the wildtype cells at day 4 PI, indicating that DNA methylation accumulates at a much slower rate in the absence of G9a (Figure 22A). To determine whether, this defect in efficient DNA methylation is dependent on the catalytic activity of G9a, the MSCV 5’LTR region was analyzed by bisulfite sequencing in the G9a-/-Tg cells and the G4 catalytically inactive mutant line at days 4  83  and 18 PI. The methylation patterns reveal a similar defect in DNA methylation efficiency in the G4 line as seen in the G9a-/- mESCs, indicating that the de novo DNA methylation of newly integrated provirus requires the catalytic activity of G9a (Figure 22B). Taken together, these results reveal that G9a is required for H3K9me2 and efficient de novo DNA methylation of MLV-based proviral elements in mESCs.  84  Figure 22. The rate of proviral de novo DNA methylation is reduced in G9a-/mESCs. A) Genomic DNA was isolated from TT2 and G9a-/- mESCs infected with MSCV-GFP virus at day 4 and day 18 PI and analyzed by bisulfite sequencing. The results showed that cells lacking G9a were incapable of efficiently de novo DNA methylate XRV LTRs (B) Bisulfite analysis was conducted in parallel on genomic DNA isolated from MSCV-GFP infected G9a-/- mESCs with a wildtype (G9a-/-Tg) or catalytically inactive (G9a-/-TgC1668A) transgene at day 4 and 18 PI. Similar to G9a-/cells, DNA hypomethylation of LTRs was observed in only G9a-/-TgC1668A cells, indicating that the catalytic activity of G9a is required for efficient DNA methylation of proviruses. The mean percentage of mCpGs/molecule sequenced is shown for each data set.  85  4.2.6 Maintenance of proviral silencing in mESCs is not dependent upon G9a Having shown that G9a is required for the establishment of proviral silencing, we next wished to determine whether this KMTase also plays a role in the maintenance of XRV silencing in mESCs. Towards this end, I infected G9a CKO mESCs [165] with MSCVGFP and isolated cells harbouring silenced provirus (GFP- cells) (Figure 23A). As a positive control, Setdb1 CKO cells, shown above to be necessary for maintenance of proviral silencing, [166] were infected in parallel with the same viral supernatant (Figure 23A-B). Treatment with 4-hydroxytamoxifen (4-OHT) induces Cre-ER recombinasemediated deletion of G9a and Setdb1, respectively. Deletion of G9a yielded no increase in the percentage of GFP+ cells (Figure 23C). In contrast, deletion of Setdb1 resulted in a 5-fold increase in the percentage of GFP+ cells, consistent with my previous findings (Figure 23C) [166]. Depletion of G9a and Setdb1 RNA in the 4-OHT treated CKO lines was confirmed by qRT-PCR with primers specific for the deleted exons (Figure 23D). A modest decrease of Setdb1 RNA was observed in the G9a CKO but not to the extent observed in the Setdb1 CKO (Figure 23D). While expression of the Mage-a2 gene, shown previously to be induced in G9a-/- mESCs [130], was clearly induced following deletion of G9a, no change in proviral GFP expression was observed in these cells (Figure 23D). In contrast, a 12-fold increase in proviral GFP expression was observed in the Setdb1 CKO line (Figure 23D). Taken together, these results reveal that while Setdb1 plays a critical role in the maintenance of proviral silencing in mESCs, G9a is dispensable for this process.  86  Figure 23. G9a is not required for the maintenance of proviral silencing in mESCs. (A) A schematic representing the generation of cells harboring silent XRVs is shown. G9a and Setdb1 CKO cells were infected with MSCV-GFP and GFP- cells were isolated by FACS. (B) FACS analysis of the percentage of GFP+ G9a and Setdb1 CKO cells day 4 PI showed similar proportion of XRV infected cells between the two lines. (C) Cells harbouring silent XRVs were treated with 4-OHT to induce G9a or Setdb1 deletion. Treated cells were analyzed by flow cytometry 7 days post-treatment. Reactivation of MSCV-GFP was detected by flow cytometry in only Setdb1 depleted cells (D) RNA was  87  isolated from untreated and 4-OHT treated G9a and Setdb1 CKO cells and the mean foldenrichment in expression of G9a, Setdb1, Mage-a2 and the MSCV-GFP provirus (normalized to β-actin) relative to the untreated WT parent lines were determined by qRT-PCR. Reduction of G9a and Setdb1 mRNA levels were detected in the G9a and Setdb1 depleted cells respectively. Whereas Mage-a2 reactivation was only detected in 4OHT treated G9a CKO cells and consistent with the flow cytometry analyses, MSCVGFP expression was only induced in 4-OHT treated Setdb1 CKO cells. Error bars indicate standard deviation between three technical replicates of one sample.  4.3 Discussion Methylation of H3K9 acts upstream of DNA methylation in plants and filamentous fungi and plays a critical role in transcriptional silencing of transposable elements and their relics in these distantly related eukaryotes [190, 191]. Recently, we reported that G9a is required for DNA methylation of ERVs in mESCs, indicating that a similar pathway exists in mammalian cells [164]. Indeed, genome-wide studies reveal that ERVs are marked by H3K9me2 and/or H3K9me3 in murine cells [108, 125, 127]. However, the role that specific H3K9 KMTases play in DNA methylation and silencing of introduced retroviral elements had not been addressed. My results reveal that G9a is required for efficient DNA methylation and establishment of silencing of MLV-based vectors in mESCs, suggesting that this KMTase acts upstream of the de novo DNA methylation machinery in a proviral-silencing pathway. Here, establishment of proviral silencing refers to the initial repression of newly integrated elements, which upon entering the germ line or becoming “endogenized” is maintained by alternative mechanisms, such as Setdb1-mediated silencing. Previously, we showed that in mESCs lacking G9a, Dnmt3a recruitment to these ERVs is perturbed [164]. Consistent with these observations, Epsztejn-Litman et al.  88  reported that G9a interacts directly with Dnmt3a and Dnmt3b in mESCs [133], and suggested that G9a promotes de novo DNA methylation by these enzymes [133]. Intriguingly, Dnmt1 was also reported to interact directly with G9a, albeit in a differentiated cell type [132, 192]. In a recent publication, Dnmt3l, a cofactor required for de novo DNA methylation, was shown to be necessary for the establishment of silencing of MSCV in mESCs [193]. However, how the Dnmts are directed to the proviruses remained unknown. Here, I show that G9a is required for efficient de novo DNA methylation of introduced retroviruses in mESCs and taken together with previous findings, indicate that G9a indeed directs the de novo Dnmts to proviral sequences. Furthermore, using mESCs bearing a conditional knock out allele of G9a, I found that G9a is dispensable for the maintenance of proviral DNA methylation in mESCs, which is carried out by Setdb1 and Dnmt1. Similarly, our collaborators have shown that in MEFs, where DNA methylation of IAP elements is dependent upon Dnmt1 and Dnmt3b, but not Dnmt3a [194, 195], G9a is not required to maintain these elements in a densely methylated state. Taken together, these results indicate that G9a plays a more important role in the establishment than the maintenance of DNA methylation. Based on the evidence provided, G9a appears to function primarily in the establishment of proviral silencing, by depositing the repressive dimethyl mark on lysine 9 and by promoting recruitment of de novo DNMT activity to the newly integrated proviruses. Alternatively, G9a or the mark it deposits may inhibit the binding of an as yet unidentified DNA demethylase to repetitive elements, and may in turn protect these regions against DNA demethylation. Interestingly, it was recently shown that the chromatin remodeler LSH interacts with G9a to establish silencing of many genes in mESCs [196]. Since LSH has  89  been implicated in proviral silencing in germ line cells [197], it is possible that G9a establishes proviral silencing in a pathway that also involves LSH. Once established, silencing of XRVs is sustained by DNA methylation and/or H3K9me3, which are maintained in a G9a-independent manner by Dnmt1 and Setdb1, respectively. This model is consistent with the observation that knockout of Setdb1 leads to reactivation of a proviral reporter, while knockout of G9a does not. As ERVs are also marked by H3K9me3 in mESCs, it is not surprising that deletion of G9a does not induce expression of these elements. As discussed in the previous chapter, we have found that Setdb1 is responsible for maintaining the transcriptionally silent state of proviruses in mESCs, independent of G9a [166]. Given the importance of maintaining such parasitic elements in a silent state, it is not surprising that multiple proviral silencing mechanisms that act at the transcriptional level, including DNA methylation and covalent histone modification-based pathways, have evolved in metazoans specifically for this purpose. We have shown that H3K9 methylation of proviral elements is mediated by two KMTases, G9a and Setdb1, which play distinct roles in the establishment and maintenance of proviral silencing, respectively.  90  5. Discussion and concluding remarks#  #  - The illustrations, figures and segments of the discussion in this chapter can be found in a submitted manuscript:  Danny Leung and Matthew Lorincz (2011) Proviral silencing - why histone marks take center stage? (Submitted)  91  Retrotransposons, including LTR and non-LTR classes, are found in the genomes of all higher eukaryotes. LTR retrotransposons, including ERVs, account for approximately 10% of the mouse genome and are divided into class I (~0.7%), class II (~3%) and class III (~5.4%) based on the sequence of the reverse transcriptase genes [1]. Retrotransposition of a subset of class I and II ERVs is responsible for up to 10% of all spontaneous mutations in mice [4]. To minimize their impact on host fitness, numerous pathways have evolved to restrict each step of the retroviral life cycle, including at the transcriptional level. DNA methylation, the best characterized transcriptional silencing mechanism, involves the catalytic addition of a methyl group to the 5th carbon of the cytosine base (5mC) by the de novo Dnmts, Dnmt3a and Dnmt3b and the maintenance Dnmt, Dnmt1. This epigenetic mark is essential during embryogenesis, as Dnmt1 and Dnmt3b KO mice are embryonic lethal [67, 75]. DNA methylation is required for repression of IAP ERVs in late stage embryos [75], MEFs, neuronal cells and epiblastderived stem cells [76]. Intriguingly however, during early embryonic development, when DNA methylation is reprogrammed, silencing of introduced retroviral vectors occurs independent of DNA methylation [161].  5.1 DNA methylation-independent proviral silencing Consistent with previous observations, recent reports reveal that DNA methylation is dispensable for silencing of ERVs in mESCs [76, 164, 166], pluripotent cells derived from the inner cell mass of blastocysts. Although OCT4+ Dnmt1 null mESCs show no increase in IAP expression relative to the wildtype parent line, a >100-fold increase in expression of this class II ERV is observed following differentiation induced by leukemia 92  inhibitory factor withdrawal [76]. Taken together, these results indicate that in mESCs, in contrast to somatic cells, proviral silencing is maintained by a DNA methylation independent mechanism. Interestingly, interdependence between DNA methylation and H3K9 methylation has been documented in distantly related eukaryotes. In Neurospora crassa, the H3K9 KMTase DIM5 is required for CpG methylation [77], while in Arabidopsis thaliana, the KMTase KRYPTONITE is required for CpNpG methylation [78]. Similarly, in mice, the related H3K9 KMTases Suv39h1 and Suv39h2 are required for DNA methylation of pericentromeric repeats [163]. In contrast, DNA methylation is not required for H3K9me of this heterochromatic region. Thus, while H3K9me generally acts upstream of DNA methylation, the role of this covalent histone mark in silencing of ERVs, which frequently integrate into euchromatin in mammalian cells [198], had until recently not been addressed. H3K9me in the mouse is catalyzed by KMTases belonging to the Suv3-9 family of SET domain containing proteins, of which there are five members with bona fide catalytic activity towards H3K9, including Suv39h1, Suv39h2, G9a, Glp and Setdb1. To determine whether H3K9me is required for proviral silencing in mESCs, we initially investigated the role of G9a, which localizes to the “euchromatic” compartment [130, 131], and is highly expressed in these cells. Surprisingly, while we found that this KMTase is required for the deposition of H3K9me2 and DNA methylation on class I and II ERVs, no increase in transcription of these elements was observed in G9a-/- mESCs [164]. Intriguingly, we found significant enrichment of H3K9me3 at the LTRs of many class I and II ERVs, consistent with a previous genome-wide analysis [108]. Importantly, 93  no decrease of this mark was detected in G9a KO mESCs [164, 184], leaving open the possibility that this mark, rather than H3K9me2, is required for maintaining proviral silencing in mESCs.  5.2 Setdb1/H3K9me3 is required for silencing of class I and II ERVs To determine which KMTase deposits H3K9me3 and whether the enzyme is responsible for maintaining proviral silencing, I conducted a siRNA screen targeting the remaining possible KMTase candidates. I found that only Setdb1, a KMTase that catalyzes the addition of one, two or three methyl groups to H3K9 and also localizes in the euchromatic compartment, was necessary for maintaining XRV silencing. As the data presented in chapter 3 demonstrates, further analysis with Setdb1 CKO mESCs revealed dramatic up-regulation of class I and II ERVs, and a concomitant decrease in H3K9me3 at these elements [166]. Consistent with this observation, depletion of the Setdb1interacting protein Kap-1, previously shown to be required for silencing of MLV-based XRVs in both mESCs and EC cells [145], also led to dramatic up-regulation of the same elements in mESCs [166, 177]. Importantly, another independent study echoed these findings in ex vivo cultured blastocysts [166, 177], indicating this pathway described in mESCs indeed functions during the developmental stage from which the cells are derived. Upon Kap-1 KD, proviral derepression was coupled with a reduction in Setdb1 enrichment and H3K9me3 at target LTRs, indicating that Setdb1 recruitment to proviral sequences in mESCs required Kap-1. Taken together, these data indicate that a Setdb1/Kap-1 complex plays an essential role in ERV repression. Interestingly, the loss of H3K9me3 at ERVs in Setdb1 depleted mESCs was accompanied by a reduction in 94  H4K20me3[166], a histone modification previously associated with transcriptional repression.  However,  no  increase  in  proviral  expression  was  observed  in  Suv420h1/KMT5B and Suv420h2/KMT5C DKO cells, indicating that although deposition of this mark at ERVs is dependent upon Setdb1, it is not necessary for proviral repression [166]. The observation that H3K9me3 is required for proviral silencing in mESCs raised the question of the role of DNA methylation in this process. Strikingly, bisulfite sequencing analysis revealed only a subtle or no change in DNA methylation at the LTRs of several reactivated ERV families in Setdb1 CKO mESCs [166]. Consistent with this observation, deletion of all three Dnmts generally does not perturb proviral silencing [76], indicating that DNA methylation is neither necessary nor sufficient for proviral silencing in mESCs. Interestingly, the IAPEz family of ERVs appears to be an exception. Using siRNA-mediated knockdown (KD) of Setdb1 and Dnmt1 alone or in combination, Dr. Irina Maksakova, a postdoctoral fellow in the Lorincz lab, found that Setdb1 acts synergistically with DNA methylation to silence this particular family of repeats [199]. Similarly, Rowe et al. also observed a synergistic effect on proviral derepression in Kap1 KO mESCs treated with 5-azacytidine [166, 177]. Although the influence of 5azacytidine treatment on epigenetic marks is more complex than originally believed [200], these observations indicate that unlike many of the other ERV subfamilies, robust silencing of IAPEz ERVs is dependent upon both H3K9me3 and DNA methylation. Consistent with the aforementioned studies, the mechanism of ERV repression in mESCs is strikingly different in differentiated cells. Setdb1-deficient MEFs isolated from Setdb1 CKO embryos demonstrated no significant increase in ERV transcription [166]. 95  Similarly, Kap-1 null MEFs exhibit no proviral silencing defect [177]. Indeed, class I and II ERVs are devoid of H3K9me3 in differentiated cells such as MEFs and neuronal progenitor cells [108]. In contrast, IAP elements are highly upregulated in Dnmt1 null embryos at E9.5 and Dnmt1 depleted MEFs [75, 76]. The culmination of these data indicate that while mESCs predominantly employ Setdb1 to maintain class I and II ERV silencing, this function is superseded by DNA methylation following cellular differentiation.  5.3 Role of H3K9me2 in establishment of proviral silencing Upon discovering the mechanism employed by mESCs to maintain ERV repression, I next focused on addressing the mechanism for establishing transcriptional silencing of newly integrated proviruses. MLV-based retroviral vectors have been used extensively in the laboratory setting for gene transfer studies and in the clinic to deliver genes for therapeutic purposes [201]. Despite the fact that MLV preferentially integrates within or near the promoter regions of genes [202], DNA methylation and associated transcriptional silencing of such vectors are frequently observed [203], explaining in part why they have shown limited efficacy in clinical applications. Here, I show that in addition to DNA methylation, G9a plays a role in the establishment of proviral silencing in mESCs. As described in chapter 4, by analyzing the kinetics of XRV silencing, I found that mESCs devoid of G9a demonstrate a pronounced silencing defect, concomitant with a dramatic reduction in H3K9me2 enrichment on the XRV LTR region. Strikingly, G9a  96  null mESCs also fail to efficiently de novo DNA methylate the newly integrated provirus and the failure to silence phenocopies that of Dnmt3a/3b DKO mESCs. This suggests that G9a interacts with and/or recruits the de novo Dnmts to the XRVs, which agree with previously published reports [133, 164]. Also consistent with aforementioned studies, I found that G9a is not required for maintaining XRV silencing. Intriguingly, ChIP experiments revealed that newly integrated proviruses are also marked by H3K9me3, albeit at low enrichment levels. This chromatin mark may also have a role in establishing silencing, as elements enriched for H3K9me3 at day 4 PI are likely to already be silenced. However, the technical difficulties posed by the cell lethality of Setdb1 KO cells precluded further analysis of this H3K9 tri-methylase. Nevertheless, I found that a potentially distinct proportion of the proviruses are silenced over a longer period of time (~day 4 to day 15 PI). These elements show a much higher enrichment level of H3K9me2 versus H3K9me3 in wildtype cells and remain transcriptionally active and DNA hypomethylated in G9a deficient mESCs. Interestingly, a third distinct population of proviruses, namely those that are constitutively expressed, was also detected. Even in wildtype cells, about 50% of GFP+ cells at day 4 PI remain so by day 35 PI, indicating that certain proviruses escape silencing all together. This phenomenon is most likely a function of the proviral integration site in the host genome, where repressive complexes and proteins, including G9a and Setdb1 may be excluded. I propose that upon integration, proviruses are either silenced quickly, possibly by Setdb1mediated H3K9me3, gradually, by G9a-deposited H3K9me2 and DNA methylation or avoid transcriptional repression altogether. Based on these results, G9a-mediated  97  H3K9me2 may function to protect against DNA demethylation of proviruses in mESCs to establish transcriptional repression, which is maintained by Setdb1 and/or Dnmt1.  5.4 Role of other histone marks in proviral silencing Several other histone modifications play roles in proviral silencing in mESCs (summarized in Table 2). Employing whole transcriptome analyses, Macfarlan et al. observed significant upregulation of class III ERVs, in particular MERV-L elements, in mESCs deficient in the H3K4 demethylase Kdm1a/Lsd1. The increase in MERV-L expression was accompanied by increased H3K4me3 and H3K27Ac and reduced H3K9me2 at these elements [204]. Further analysis revealed that Kdm1a-mediated repression is dependent upon the catalytic activity of this demethylase. Interestingly, DNA methylation analyses of MERV-L elements revealed that Kdm1a KO mESCs retain levels of DNA methylation similar to wildtype, a result resembling the modest DNA methylation changes observed at upregulated class I and II ERVs in Setdb1 KO mESCs. Importantly, MERV-L elements are not derepressed upon Setdb1 depletion in mESCs [199] and are not marked by H3K9me3 in these cells [108], indicating that different ERV classes are regulated by distinct histone modifying activities. Interestingly, the authors also discovered an interaction between Kdm1a and Kap-1. Taken together with the previously reported upregulation of MERV-L elements in Kap-1 KO mESCs [177], these observations indicate that Kap-1 interacts with both Setdb1 and Kdm1a and may function through both enzymes to mediate proviral silencing. Intriguingly, a relatively small number of ERV families appear to be repressed by both pathways (Figure 24). Although  98  the focus of their study was on class III ERVs, analysis of the RNA-seq data generated by Macfarlan et al. [204] and our own study of Setdb1 KO cells [199] revealed that GLN/RLTR1B, MLV/RLTR4 and in particular ETn ERVs are upregulated in both Kdm1a KO and Setdb1 KO cells. While we cannot rule out the possibility that distinct elements within each family are upregulated in Kdm1a KO and Setdb1 KO mESCs, these observations suggest that these epigenetic pathways may cooperate to silence a subset of ERVs.  Figure 24. ERVs derepressed in Setdb1 KO versus Kdm1a KO mESCs. For all Repbase [205] annotated ERV sub-families (413 in total), reads per kilobase per million mapped reads (RPKM) [199] were generated from previously published data for Setdb1 KO [199], Kdm1a KO[204] and their corresponding wildtype parent mESC lines. To calculate RPKM values, the sum of unique and multi-matched reads aligned to all copies of an annotated ERV subfamily was normalized to the total number of exonic reads and the agglomerated length of all copies of each subfamily. Subsequently, Z-scores, which reflect changes in expression of each ERV subfamily in the mutant lines, were calculated 99  using the RPKM of KO versus wildtype for each pair, as previously described [199]. Black data points- ERVs families with highest increase in expression in Setdb1 KO cells. Red data points- ERVs families with highest increase in expression in Kdm1a KO cells. Green data points - ERVs families with increased expression in both cell lines. The results indicate that while a number of proviruses are reactivated exclusively in one or the other KO line, others are clearly derepressed in both KO lines. (Bioinformatic analysis was conducted by Dr. Mohammad Karimi)  In another study, Leeb et al. found that polycomb repressive complex 2, which deposits H3K27me3, is involved in the silencing of IAP elements in mESCs [206]. However, as these elements are not marked by H3K27me3 [108], the role that polycombgroup proteins play in proviral silencing is likely indirect. Interestingly, H3K64me, a modification within the globular domain of histone H3, was recently discovered by mass spectrometry [207]. In an independent study, H3K64me3 was shown to be enriched at heterochromatin and at IAP elements in mESCs [208], as is H3K9me3. Also similar to H3K9me3, this modification appears to mark these elements at specific developmental stages, as it is lost in MEFs. However, the KMTase responsible for deposition of H3K64me3 and the role of this mark in proviral silencing remain to be determined. Regardless, taken together, these experiments reveal that histone modification-mediated transcriptional regulation may be particularly important for silencing of ERVs in mESCs and in the blastocysts from which they are derived.  100  Involved in silencing of ERV class: KMTase Setdb1  I  II  III  ++  ++  -  G9a Glp  -  -  ? ? -  Suv39h1/h2 Kdm1a  +  +  ++  Ezh2  -  +  -  Reference Matsui et al. 2010 [166], Karimi et al. 2011 [199] Dong et al. 2008 [164] Dong et al. 2008 [164] Matsui et al. 2010 [166] Macfarlan et al. 2011 [204] Leeb et al. 2010 [206]  H3K64me3 KMTasea Kap-1b  ?  ?  ?  Daujat et al. 2009 [208]  +  +  +  Matsui et al. 2010 [166], Rowe et al. 2010 [177] a The KMTase responsible for H3K64me3 has not yet been identified. b Kap-1, a corepressor that interacts with Setdb1 and Kdm1a and is required for proviral silencing, is also shown. Table 2. KMTases involved in silencing of class I-III ERVs. The role of different KMTase in silencing of individual ERV classes and corresponding references are summarized.  5.5 A model explaining the requirement for DNA methylation-independent proviral silencing pathways in mESCs The studies described above reveal that factors which either add repressive covalent histone marks or remove active covalent histone marks, such as G9a, Setdb1 and Kdm1a, are particularly important for proviral silencing during at least one of the developmental stages when DNA methylation is reprogrammed (the roles of these enzymes in the germ line remain to be explored). Although the kinetics of DNA methylation reprogramming  101  are well described, the mechanism responsible for DNA demethylation remains enigmatic [82]. Recent discovery and characterization of the Tet family of proteins have provided intriguing insights into this process [86]. Tet family consists of Tet1, Tet2 and Tet3 proteins, which all catalyze oxidation of 5mC to 5hmC [89]. Tet1 and Tet2 are highly expressed in mESCs and blastocysts [90-92], and play an important role in the maintenance of the pluripotent state [90, 209]. Conversion of 5mC to 5hmC may be the initial step to DNA demethylation [89, 94] as suggested by a recent study showing that Tet2 is capable of further oxidizing 5hmC to 5caC, a modification readily excised by Thymine-DNA glycosylase (Tdg) and ultimately may lead to replacement of the original modified base with unmodified C, perhaps via the BER pathway [210]. Regardless, unlike 5mC, 5hmC appears to be non-repressive, as Mbd proteins are incapable of interacting with this mark [95, 96]. Also, enrichment of Tet1 is found at unmethylated promoter regions of transcribed genes [90, 98]. Interestingly, the expression levels of both Tet1 and Tet2 are dramatically reduced upon cellular differentiation. Therefore, it is likely that Tet1 and Tet2 proteins function to negate the repressive effect of 5mC at specific developmental stages, effectively rendering DNA methylation non-repressive at specific genomic sites, or removing this mark altogether (Illustration 7) [98].  102  Illustration 7. The stages of early mouse embryonic development, Tet protein expression and DNA methylation dynamics. Relevant stages of early mouse embryonic development are shown, along with the time points for derivation of mESCs, epiblast derived stem cells, embryonic germ cells and MEFs. The kinetics of Tet genes expression are also shown, along with the dynamics of DNA methylation reprogramming, illustrated by a gradient where higher levels of methylation are indicated by darker shading. High levels of 5hmC [211-213] and expression of all three Tet proteins [214] have also been detected in various brain tissues. Chromatin-based proviral silencing may be of particular importance in each of the cell types where one or more of the Tet proteins are expressed. (Figure modified from Wu and Zhang [82]).  Curiously, as discussed above, although proviral elements are not repressed by DNA methylation in mESCs, they appear to be heavily DNA methylated in their 5’LTR regions when analyzed by bisulfite sequencing or Southern analysis with methylation sensitive restriction enzyme digestion [164, 166, 204]. However, as neither bisulfite sequencing nor methyl-sensitive digestion can discriminate between 5mC and 5hmC [215], results previously obtained with these techniques in mESCs must be interpreted with caution. The revelation that 5mC may be converted to 5hmC in these cells suggests an intriguing alternative model, namely, that the “DNA methylation” detected at specific  103  genomic loci using these methods may predominantly be the non-repressive 5hmC, a potential intermediate in the pathway to DNA demethylation. The presence of 5hmC in these regions would explain the necessity of alternative silencing mechanisms, such as those based on altering histone modifications, specifically in cells expressing relatively high levels of one or more of the TET proteins, such as mESCs. I propose that at a subset of ERVs, a fraction of “DNA methylation” detected by bisulfite sequencing and/or Southern blotting in mESCs is actually 5hmC. As 5hmC, unlike 5mC, apparently does not promote transcriptional repression, any ERVs capable of promoting deposition of this mark in their promoter regions, via recruitment of one or more of the Tet proteins, could effectively evade silencing by DNA methylation. ERVs that exploit deposition of 5hmC in germ cells or early in embryonic development to maximize their expression would likely retrotranspose at a higher rate than those elements that are maintained in a silent state by DNA methylation during these stages. Conversely, given the deleterious effects associated with retrotransposition, at these developmental stages, the establishment of alternative proviral silencing pathways would confer a selective advantage to the host. Chromatin-based silencing pathways may have evolved to serve this purpose (Illustration 8). While hmeDIP-seq analysis of an agglomeration of all ERVs [98] and hmeDIP analysis of a specific IAP element [216] have shown only low levels of 5hmC enrichment in mESCs, systematic analyses of the distribution of this mark at specific ERV subfamilies, particularly those that are constitutively expressed or maintained in a silent state by Setdb1 in these cells, such as ETn and MusD elements, have yet to be conducted.  104  Illustration 8. Chromatin modifying enzymes involved in establishment and maintenance of proviral silencing and a hypothetical pathway for turnover of DNA methylation at these elements in mESCs. The H3K4 demethylase Kdm1a and the H3K9 KMTase Setdb1, which act either alone or in combination (depending upon the ERV subfamily) to maintain these elements in a silent state in mESCs, are shown [199, 204]. Also, G9a functions independently to establish H3K9me2/de novo DNA methylation-mediated silencing for newly integrated proviruses, which potentially protects against DNA “demethylation” by the Tet proteins. A hypothetical DNA methylation-demethylation cycle at these elements is also shown. I propose that a subset of elements recruit Tet proteins and in turn, promote catalysis of 5mC to 5hmC or 5caC as intermediates in a pathway to DNA demethylation via Tdg-initiated BER, thus evading repression by DNA methylation. In response, DNA methylation-independent, chromatinbased silencing pathways, such as those involving Kdm1a or Setdb1 and Kap-1 have evolved in the host to maintain these parasitic elements in a silent state. Newly integrated proviruses are unable to recruit Tet proteins, as G9a/H3K9me2 perturbs this process and in turn renders such elements silent.  105  A prediction of my model is that the silencing pathways that affect covalent modifications on histones independent of DNA methylation would not be employed in cell types in which the expression of Tet proteins is either low or absent. Indeed, this is precisely what is observed for H3K9me3, which is lost at ERVs in differentiated cells [108, 166]. Conversely, this model also predicts that other cell types in which DNA methylation levels are “insufficient” for ERV repression would employ such alternative silencing pathways. Primordial germ cells (PGCs) for example, undergo dramatic DNA methylation reprogramming at ~E11.5-13.5, in preparation for germ line development, including establishment of gender-specific DNA methylation patterns (Illustration 7) [81]. Similar to mESCs, these cells also express Tet1 and Tet2 proteins at high levels [85]. Intriguingly, IAP elements, which show a significant decrease in DNA methylation levels at this stage in development [75, 85, 179, 217], are apparently maintained in a silent state via DNA methylation independent mechanisms, as indicated by the lack of IAP derepression in PGCs isolated from Dnmt1 null E9.5 embryos [75]. Clearly, establishing whether the histone modifications discussed above play a role in proviral silencing in PGCs would shed light on this question. Furthermore, determining 5hmC and 5mC levels at such ERVs in mESCs as well as in primary tissues where DNA methylation is reprogrammed, such as in the early embryo and in PGCs, will further our understanding of the ongoing arms race between the host and this highly successful class of intracellular parasites.  106  Bibliography 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18.  Waterston, R.H., et al., Initial sequencing and comparative analysis of the mouse genome. Nature, 2002. 420(6915): p. 520-62. Lander, E.S., et al., Initial sequencing and analysis of the human genome. Nature, 2001. 409(6822): p. 860-921. Stocking, C. and C.A. Kozak, Murine endogenous retroviruses. Cell Mol Life Sci, 2008. 65(21): p. 3383-98. Maksakova, I.A., et al., Retroviral elements and their hosts: insertional mutagenesis in the mouse germ line. PLoS Genet, 2006. 2(1): p. e2. Alamgir, A.S., et al., Precise identification of endogenous proviruses of NFS/N mice participating in recombination with moloney ecotropic murine leukemia virus (MuLV) to generate polytropic MuLVs. J Virol, 2005. 79(8): p. 4664-71. Belshaw, R., et al., Long-term reinfection of the human genome by endogenous retroviruses. Proc Natl Acad Sci U S A, 2004. 101(14): p. 4894-9. Hook, L.M., et al., Genetics of mouse mammary tumor virus-induced mammary tumors: linkage of tumor induction to the gag gene. J Virol, 2000. 74(19): p. 8876-83. Boeke, J.D. and J.P. Stoye, Retrotransposons, Endogenous Retroviruses, and the Evolution of Retroelements. 1997. Zhang, Y., et al., Genome-wide assessments reveal extremely high levels of polymorphism of two active families of mouse endogenous retroviral elements. PLoS Genet, 2008. 4(2): p. e1000007. Ribet, D., et al., Murine MusD retrotransposon: structure and molecular evolution of an "intracellularized" retrovirus. J Virol, 2007. 81(4): p. 1888-98. D'Souza, V., et al., NMR structure of the 101-nucleotide core encapsidation signal of the Moloney murine leukemia virus. J Mol Biol, 2004. 337(2): p. 427-42. Gross, L., "Spontaneous" leukemia developing in C3H mice following inoculation in infancy, with AK-leukemic extracts, or AK-embrvos. Proc Soc Exp Biol Med, 1951. 76(1): p. 27-32. Steffen, D., et al., Identification of DNA fragments carrying ecotropic proviruses of AKR mice. Proc Natl Acad Sci U S A, 1979. 76(9): p. 4554-8. Stoye, J.P. and J.M. Coffin, Polymorphism of murine endogenous proviruses revealed by using virus class-specific oligonucleotide probes. J Virol, 1988. 62(1): p. 168-75. Solaiman, F., et al., Modular retro-vectors for transgenic and therapeutic use. Mol Reprod Dev, 2000. 56(2 Suppl): p. 309-15. Tai, C.K. and N. Kasahara, Replication-competent retrovirus vectors for cancer gene therapy. Front Biosci, 2008. 13: p. 3083-95. Colicelli, J. and S.P. Goff, Isolation of a recombinant murine leukemia virus utilizing a new primer tRNA. J Virol, 1986. 57(1): p. 37-45. Barklis, E., R.C. Mulligan, and R. Jaenisch, Chromosomal position or virus mutation permits retrovirus expression in embryonal carcinoma cells. Cell, 1986. 47(3): p. 391-9.  107  19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37.  Ribet, D., et al., The GLN family of murine endogenous retroviruses contains an element competent for infectious viral particle formation. J Virol, 2008. 82(9): p. 4413-9. Gifford, R., et al., Evolution and distribution of class II-related endogenous retroviruses. J Virol, 2005. 79(10): p. 6478-86. Baillie, G.J., et al., Multiple groups of endogenous betaretroviruses in mice, rats, and other mammals. J Virol, 2004. 78(11): p. 5784-98. McCarthy, E.M. and J.F. McDonald, Long terminal repeat retrotransposons of Mus musculus. Genome Biol, 2004. 5(3): p. R14. Czarneski, J., J.C. Rassa, and S.R. Ross, Mouse mammary tumor virus and the immune system. Immunol Res, 2003. 27(2-3): p. 469-80. Chase, D.G. and L. Piko, Expression of A- and C-type particles in early mouse embryos. J Natl Cancer Inst, 1973. 51(6): p. 1971-5. Wivel, N.A. and G.H. Smith, Distribution of intracisternal A-particles in a variety of normal and neoplastic mouse tissues. Int J Cancer, 1971. 7(1): p. 16775. Mager, D.L. and J.D. Freeman, Novel mouse type D endogenous proviruses and ETn elements share long terminal repeat and internal sequences. J Virol, 2000. 74(16): p. 7221-9. Ribet, D., M. Dewannieux, and T. Heidmann, An active murine transposon family pair: retrotransposition of "master" MusD copies and ETn trans-mobilization. Genome Res, 2004. 14(11): p. 2261-7. Benit, L., et al., ERV-L elements: a family of endogenous retrovirus-like elements active throughout the evolution of mammals. J Virol, 1999. 73(4): p. 3301-8. Smit, A.F., Identification of a new, abundant superfamily of mammalian LTRtransposons. Nucleic Acids Res, 1993. 21(8): p. 1863-72. Peaston, A.E., B.B. Knowles, and K.W. Hutchison, Genome plasticity in the mouse oocyte and early embryo. Biochem Soc Trans, 2007. 35(Pt 3): p. 618-22. Peaston, A.E., et al., Retrotransposons regulate host genes in mouse oocytes and preimplantation embryos. Dev Cell, 2004. 7(4): p. 597-606. Dupressoir, A. and T. Heidmann, Germ line-specific expression of intracisternal A-particle retrotransposons in transgenic mice. Mol Cell Biol, 1996. 16(8): p. 4495-503. Loebel, D.A., et al., Restricted expression of ETn-related sequences during postimplantation mouse development. Gene Expr Patterns, 2004. 4(4): p. 467-71. Svoboda, P., et al., RNAi and expression of retrotransposons MuERV-L and IAP in preimplantation mouse embryos. Dev Biol, 2004. 269(1): p. 276-85. Brulet, P., et al., Early differential tissue expression of transposon-like repetitive DNA sequences of the mouse. Proc Natl Acad Sci U S A, 1983. 80(18): p. 5641-5. Brulet, P., H. Condamine, and F. Jacob, Spatial distribution of transcripts of the long repeated ETn sequence during early mouse embryogenesis. Proc Natl Acad Sci U S A, 1985. 82(7): p. 2054-8. Baust, C., et al., Structure and expression of mobile ETnII retroelements and their coding-competent MusD relatives in the mouse. J Virol, 2003. 77(21): p. 1144858.  108  38. 39. 40. 41. 42. 43. 44.  45. 46.  47. 48. 49. 50. 51. 52. 53. 54.  Morgan, H.D., et al., Epigenetic inheritance at the agouti locus in the mouse. Nat Genet, 1999. 23(3): p. 314-8. Chen, Y., A. Pane, and T. Schupbach, Cutoff and aubergine mutations result in retrotransposon upregulation and checkpoint activation in Drosophila. Curr Biol, 2007. 17(7): p. 637-42. Klattenhoff, C., et al., Drosophila rasiRNA pathway mutations disrupt embryonic axis specification through activation of an ATR/Chk2 DNA damage response. Dev Cell, 2007. 12(1): p. 45-55. Theurkauf, W.E., et al., rasiRNAs, DNA damage, and embryonic axis specification. Cold Spring Harb Symp Quant Biol, 2006. 71: p. 171-80. Belgnaoui, S.M., et al., Human LINE-1 retrotransposon induces DNA damage and apoptosis in cancer cells. Cancer Cell Int, 2006. 6: p. 13. Farkash, E.A. and E.T. Luning Prak, DNA damage and L1 retrotransposition. J Biomed Biotechnol, 2006. 2006(1): p. 37285. Mameli, G., et al., Brains and peripheral blood mononuclear cells of multiple sclerosis (MS) patients hyperexpress MS-associated retrovirus/HERV-W endogenous retrovirus, but not Human herpesvirus 6. J Gen Virol, 2007. 88(Pt 1): p. 264-74. Ruprecht, K., et al., Regulation of human endogenous retrovirus W protein expression by herpes simplex virus type 1: implications for multiple sclerosis. J Neurovirol, 2006. 12(1): p. 65-71. Piotrowski, P.C., S. Duriagin, and P.P. Jagodzinski, Expression of human endogenous retrovirus clone 4-1 may correlate with blood plasma concentration of anti-U1 RNP and anti-Sm nuclear antibodies. Clin Rheumatol, 2005. 24(6): p. 620-4. Freimanis, G., et al., A role for human endogenous retrovirus-K (HML-2) in rheumatoid arthritis: investigating mechanisms of pathogenesis. Clin Exp Immunol, 2010. 160(3): p. 340-7. Colmegna, I. and R.F. Garry, Role of endogenous retroviruses in autoimmune diseases. Infect Dis Clin North Am, 2006. 20(4): p. 913-29. Balada, E., J. Ordi-Ros, and M. Vilardell-Tarres, Molecular mechanisms mediated by human endogenous retroviruses (HERVs) in autoimmunity. Rev Med Virol, 2009. 19(5): p. 273-86. Ono, M., et al., Nucleotide sequence of human endogenous retrovirus genome related to the mouse mammary tumor virus genome. J Virol, 1986. 60(2): p. 58998. Marguerat, S., et al., Association of human endogenous retrovirus K-18 polymorphisms with type 1 diabetes. Diabetes, 2004. 53(3): p. 852-4. Ruprecht, K., et al., Endogenous retroviruses and cancer. Cell Mol Life Sci, 2008. 65(21): p. 3366-82. Buscher, K., et al., Expression of human endogenous retrovirus K in melanomas and melanoma cell lines. Cancer Res, 2005. 65(10): p. 4172-80. Seniuta, N.B., et al., [HERV-K-associated carcinogenesis: co-expression of viral and cellular proteins in the development of human germ-cell tumors]. Vopr Virusol, 2009. 54(2): p. 21-6.  109  55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72.  Wang-Johanning, F., et al., Expression of multiple human endogenous retrovirus surface envelope proteins in ovarian cancer. Int J Cancer, 2007. 120(1): p. 81-90. Bjerregaard, B., et al., Syncytin is involved in breast cancer-endothelial cell fusions. Cell Mol Life Sci, 2006. 63(16): p. 1906-11. Strick, R., et al., Proliferation and cell-cell fusion of endometrial carcinoma are induced by the human endogenous retroviral Syncytin-1 and regulated by TGFbeta. J Mol Med (Berl), 2007. 85(1): p. 23-38. Boese, A., et al., Human endogenous retrovirus protein cORF supports cell transformation and associates with the promyelocytic leukemia zinc finger protein. Oncogene, 2000. 19(38): p. 4328-36. Denne, M., et al., Physical and functional interactions of human endogenous retrovirus proteins Np9 and rec with the promyelocytic leukemia zinc finger protein. J Virol, 2007. 81(11): p. 5607-16. Armbruester, V., et al., A novel gene from the human endogenous retrovirus K expressed in transformed cells. Clin Cancer Res, 2002. 8(6): p. 1800-7. Lamprecht, B., et al., Derepression of an endogenous long terminal repeat activates the CSF1R proto-oncogene in human lymphoma. Nat Med, 2010. 16(5): p. 571-9, 1p following 579. Mohn, F., et al., Lineage-specific polycomb targets and de novo DNA methylation define restriction and potential of neuronal progenitors. Mol Cell, 2008. 30(6): p. 755-66. Panning, B., X-chromosome inactivation: the molecular basis of silencing. J Biol, 2008. 7(8): p. 30. Li, E., C. Beard, and R. Jaenisch, Role for DNA methylation in genomic imprinting. Nature, 1993. 366(6453): p. 362-5. Lei, H., et al., De novo DNA cytosine methyltransferase activities in mouse embryonic stem cells. Development, 1996. 122(10): p. 3195-205. Chen, T., et al., Establishment and maintenance of genomic methylation patterns in mouse embryonic stem cells by Dnmt3a and Dnmt3b. Mol Cell Biol, 2003. 23(16): p. 5594-605. Okano, M., et al., DNA methyltransferases Dnmt3a and Dnmt3b are essential for de novo methylation and mammalian development. Cell, 1999. 99(3): p. 247-57. Hata, K., et al., Dnmt3L cooperates with the Dnmt3 family of de novo DNA methyltransferases to establish maternal imprints in mice. Development, 2002. 129(8): p. 1983-93. Tatematsu, K.I., T. Yamazaki, and F. Ishikawa, MBD2-MBD3 complex binds to hemi-methylated DNA and forms a complex containing DNMT1 at the replication foci in late S phase. Genes Cells, 2000. 5(8): p. 677-88. Li, E., T.H. Bestor, and R. Jaenisch, Targeted mutation of the DNA methyltransferase gene results in embryonic lethality. Cell, 1992. 69(6): p. 91526. Cowley, S.M., et al., The mSin3A chromatin-modifying complex is essential for embryogenesis and T-cell development. Mol Cell Biol, 2005. 25(16): p. 69907004. Curradi, M., et al., Molecular mechanisms of gene silencing mediated by DNA methylation. Mol Cell Biol, 2002. 22(9): p. 3157-73.  110  73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93.  Fujita, N., et al., Methyl-CpG binding domain 1 (MBD1) interacts with the Suv39h1-HP1 heterochromatic complex for DNA methylation-based transcriptional repression. J Biol Chem, 2003. 278(26): p. 24132-8. Yoder, J.A., C.P. Walsh, and T.H. Bestor, Cytosine methylation and the ecology of intragenomic parasites. Trends Genet, 1997. 13(8): p. 335-40. Walsh, C.P. and T.H. Bestor, Cytosine methylation and mammalian development. Genes Dev, 1999. 13(1): p. 26-34. Hutnick, L.K., et al., Repression of retrotransposal elements in mouse embryonic stem cells is primarily mediated by a DNA methylation-independent mechanism. J Biol Chem, 2010. 285(27): p. 21082-91. Tamaru, H. and E.U. Selker, A histone H3 methyltransferase controls DNA methylation in Neurospora crassa. Nature, 2001. 414(6861): p. 277-83. Jackson, J.P., et al., Control of CpNpG DNA methylation by the KRYPTONITE histone H3 methyltransferase. Nature, 2002. 416(6880): p. 556-60. Santos, F. and W. Dean, Epigenetic reprogramming during early development in mammals. Reproduction, 2004. 127(6): p. 643-51. Surani, M.A., Reprogramming of genome function through epigenetic inheritance. Nature, 2001. 414(6859): p. 122-8. Hajkova, P., et al., Epigenetic reprogramming in mouse primordial germ cells. Mech Dev, 2002. 117(1-2): p. 15-23. Wu, S.C. and Y. Zhang, Active DNA demethylation: many roads lead to Rome. Nat Rev Mol Cell Biol, 2010. 11(9): p. 607-20. Bhutani, N., et al., Reprogramming towards pluripotency requires AID-dependent DNA demethylation. Nature, 2010. 463(7284): p. 1042-7. Rai, K., et al., DNA demethylation in zebrafish involves the coupling of a deaminase, a glycosylase, and gadd45. Cell, 2008. 135(7): p. 1201-12. Hajkova, P., et al., Genome-wide reprogramming in the mouse germ line entails the base excision repair pathway. Science, 2010. 329(5987): p. 78-82. Tahiliani, M., et al., Conversion of 5-methylcytosine to 5-hydroxymethylcytosine in mammalian DNA by MLL partner TET1. Science, 2009. 324(5929): p. 930-5. He, Y.F., et al., Tet-Mediated Formation of 5-Carboxylcytosine and Its Excision by TDG in Mammalian DNA. Science, 2011. Ito, S., et al., Tet Proteins Can Convert 5-Methylcytosine to 5-Formylcytosine and 5-Carboxylcytosine. Science, 2011. Iqbal, K., et al., Reprogramming of the paternal genome upon fertilization involves genome-wide oxidation of 5-methylcytosine. Proc Natl Acad Sci U S A, 2011. 108(9): p. 3642-7. Ito, S., et al., Role of Tet proteins in 5mC to 5hmC conversion, ES-cell selfrenewal and inner cell mass specification. Nature, 2010. 466(7310): p. 1129-33. Koh, K.P., et al., Tet1 and Tet2 regulate 5-hydroxymethylcytosine production and cell lineage specification in mouse embryonic stem cells. Cell Stem Cell, 2011. 8(2): p. 200-13. Ruzov, A., et al., Lineage-specific distribution of high levels of genomic 5hydroxymethylcytosine in mammalian development. Cell Res, 2011. Wu, H., et al., Dual functions of Tet1 in transcriptional regulation in mouse embryonic stem cells. Nature, 2011.  111  94. 95. 96. 97. 98. 99. 100. 101. 102. 103. 104. 105. 106. 107. 108. 109. 110. 111. 112.  Wossidlo, M., et al., 5-Hydroxymethylcytosine in the mammalian zygote is linked with epigenetic reprogramming. Nat Commun, 2011. 2: p. 241. Frauer, C., et al., Recognition of 5-Hydroxymethylcytosine by the Uhrf1 SRA Domain. PLoS One, 2011. 6(6): p. e21306. Valinluck, V., et al., Oxidative damage to methyl-CpG sequences inhibits the binding of the methyl-CpG binding domain (MBD) of methyl-CpG binding protein 2 (MeCP2). Nucleic Acids Res, 2004. 32(14): p. 4100-8. Xu, Y., et al., Genome-wide Regulation of 5hmC, 5mC, and Gene Expression by Tet1 Hydroxylase in Mouse Embryonic Stem Cells. Mol Cell, 2011. 42(4): p. 45164. Ficz, G., et al., Dynamic regulation of 5-hydroxymethylcytosine in mouse ES cells and during differentiation. Nature, 2011. 473(7347): p. 398-402. Cheung, P. and P. Lau, Epigenetic regulation by histone methylation and histone variants. Mol Endocrinol, 2005. 19(3): p. 563-73. Shiio, Y. and R.N. Eisenman, Histone sumoylation is associated with transcriptional repression. Proc Natl Acad Sci U S A, 2003. 100(23): p. 1322530. Verdone, L., M. Caserta, and E. Di Mauro, Role of histone acetylation in the control of gene expression. Biochem Cell Biol, 2005. 83(3): p. 344-53. Eberharter, A. and P.B. Becker, Histone acetylation: a switch between repressive and permissive chromatin. Second in review series on chromatin dynamics. EMBO Rep, 2002. 3(3): p. 224-9. Weake, V.M. and J.L. Workman, Histone ubiquitination: triggering gene activity. Mol Cell, 2008. 29(6): p. 653-63. Martin, C. and Y. Zhang, The diverse functions of histone lysine methylation. Nat Rev Mol Cell Biol, 2005. 6(11): p. 838-49. Taverna, S.D., et al., How chromatin-binding modules interpret histone modifications: lessons from professional pocket pickers. Nat Struct Mol Biol, 2007. 14(11): p. 1025-40. Beck, H.C., et al., Quantitative proteomic analysis of post-translational modifications of human histones. Mol Cell Proteomics, 2006. 5(7): p. 1314-25. Garcia, B.A., et al., Pervasive combinatorial modification of histone H3 in human cells. Nat Methods, 2007. 4(6): p. 487-9. Mikkelsen, T.S., et al., Genome-wide maps of chromatin state in pluripotent and lineage-committed cells. Nature, 2007. 448(7153): p. 553-60. Bannister, A.J., et al., Selective recognition of methylated lysine 9 on histone H3 by the HP1 chromo domain. Nature, 2001. 410(6824): p. 120-4. Motamedi, M.R., et al., HP1 proteins form distinct complexes and mediate heterochromatic gene silencing by nonoverlapping mechanisms. Mol Cell, 2008. 32(6): p. 778-90. Mosch, K., et al., HP1 recruits activity-dependent neuroprotective protein to H3K9me3 marked pericentromeric heterochromatin for silencing of major satellite repeats. PLoS One, 2011. 6(1): p. e15894. Maksakova, I.A., et al., H3K9me3 binding proteins are dispensable for SETDB1/H3K9me3-dependent retroviral silencing. Epigenetics Chromatin, 2011. 4(1): p. 12.  112  113. 114. 115. 116. 117. 118. 119. 120. 121. 122. 123. 124. 125. 126. 127. 128. 129.  Cao, R. and Y. Zhang, SUZ12 is required for both the histone methyltransferase activity and the silencing function of the EED-EZH2 complex. Mol Cell, 2004. 15(1): p. 57-67. Fang, J., et al., Purification and functional characterization of SET8, a nucleosomal histone H4-lysine 20-specific methyltransferase. Curr Biol, 2002. 12(13): p. 1086-99. Ogawa, H., et al., A complex with chromatin modifiers that occupies E2F- and Myc-responsive genes in G0 cells. Science, 2002. 296(5570): p. 1132-6. Wang, H., et al., mAM facilitates conversion by ESET of dimethyl to trimethyl lysine 9 of histone H3 to cause transcriptional repression. Mol Cell, 2003. 12(2): p. 475-87. Schotta, G., et al., A silencing pathway to induce H3-K9 and H4-K20 trimethylation at constitutive heterochromatin. Genes Dev, 2004. 18(11): p. 125162. Tachibana, M., et al., Histone methyltransferases G9a and GLP form heteromeric complexes and are both crucial for methylation of euchromatin at H3-K9. Genes Dev, 2005. 19(7): p. 815-26. Shinkai, Y., Regulation and function of H3K9 methylation. Subcell Biochem, 2007. 41: p. 337-50. Reuter, G. and P. Spierer, Position effect variegation and chromatin proteins. Bioessays, 1992. 14(9): p. 605-12. Allshire, R.C., Elements of chromosome structure and function in fission yeast. Semin Cell Biol, 1995. 6(2): p. 55-64. O'Carroll, D., et al., Isolation and characterization of Suv39h2, a second histone H3 methyltransferase gene that displays testis-specific expression. Mol Cell Biol, 2000. 20(24): p. 9423-33. Aagaard, L., et al., Functional mammalian homologues of the Drosophila PEVmodifier Su(var)3-9 encode centromere-associated proteins which complex with the heterochromatin component M31. EMBO J, 1999. 18(7): p. 1923-38. Peters, A.H., et al., Loss of the Suv39h histone methyltransferases impairs mammalian heterochromatin and genome stability. Cell, 2001. 107(3): p. 323-37. Peters, A.H., et al., Partitioning and plasticity of repressive histone methylation states in mammalian chromatin. Mol Cell, 2003. 12(6): p. 1577-89. Garcia-Cao, M., et al., Epigenetic regulation of telomere length in mammalian cells by the Suv39h1 and Suv39h2 histone methyltransferases. Nat Genet, 2004. 36(1): p. 94-9. Martens, J.H., et al., The profile of repeat-associated histone lysine methylation states in the mouse epigenome. Embo J, 2005. 24(4): p. 800-12. Milner, C.M. and R.D. Campbell, The G9a gene in the human major histocompatibility complex encodes a novel protein containing ankyrin-like repeats. Biochem J, 1993. 290 ( Pt 3): p. 811-8. Tachibana, M., et al., Set domain-containing protein, G9a, is a novel lysinepreferring mammalian histone methyltransferase with hyperactivity and specific selectivity to lysines 9 and 27 of histone H3. J Biol Chem, 2001. 276(27): p. 25309-17.  113  130. 131. 132. 133. 134. 135. 136. 137. 138. 139. 140. 141. 142.  143. 144. 145. 146.  Tachibana, M., et al., G9a histone methyltransferase plays a dominant role in euchromatic histone H3 lysine 9 methylation and is essential for early embryogenesis. Genes Dev, 2002. 16(14): p. 1779-91. Rice, J.C., et al., Histone methyltransferases direct different degrees of methylation to define distinct chromatin domains. Mol Cell, 2003. 12(6): p. 15918. Esteve, P.O., et al., Direct interaction between DNMT1 and G9a coordinates DNA and histone methylation during replication. Genes Dev, 2006. 20(22): p. 3089-103. Epsztejn-Litman, S., et al., De novo DNA methylation promoted by G9a prevents reprogramming of embryonically silenced genes. Nat Struct Mol Biol, 2008. 15(11): p. 1176-83. Ueda, J., et al., Zinc finger protein Wiz links G9a/GLP histone methyltransferases to the co-repressor molecule CtBP. J Biol Chem, 2006. 281(29): p. 20120-8. Sampath, S.C., et al., Methylation of a histone mimic within the histone methyltransferase G9a regulates protein complex assembly. Mol Cell, 2007. 27(4): p. 596-608. Feldman, N., et al., G9a-mediated irreversible epigenetic inactivation of Oct-3/4 during early embryogenesis. Nat Cell Biol, 2006. 8(2): p. 188-94. Kubicek, S., et al., Reversal of H3K9me2 by a small-molecule inhibitor for the G9a histone methyltransferase. Mol Cell, 2007. 25(3): p. 473-81. Shi, Y., et al., A combined chemical and genetic approach for the generation of induced pluripotent stem cells. Cell Stem Cell, 2008. 2(6): p. 525-8. Wen, B., et al., Large histone H3 lysine 9 dimethylated chromatin blocks distinguish differentiated from embryonic stem cells. Nat Genet, 2009. 41(2): p. 246-50. Yang, L., et al., Molecular cloning of ESET, a novel histone H3-specific methyltransferase that interacts with ERG transcription factor. Oncogene, 2002. 21(1): p. 148-52. Lin, W.M., et al., Modeling genomic diversity and tumor dependency in malignant melanoma. Cancer Res, 2008. 68(3): p. 664-73. Sripathy, S.P., J. Stevens, and D.C. Schultz, The KAP1 corepressor functions to coordinate the assembly of de novo HP1-demarcated microenvironments of heterochromatin required for KRAB zinc finger protein-mediated transcriptional repression. Mol Cell Biol, 2006. 26(22): p. 8623-38. Li, H., et al., The histone methyltransferase SETDB1 and the DNA methyltransferase DNMT3A interact directly and localize to promoters silenced in cancer cells. J Biol Chem, 2006. 281(28): p. 19489-500. Wiznerowicz, M., et al., The Kruppel-associated box repressor domain can trigger de novo promoter methylation during mouse early embryogenesis. J Biol Chem, 2007. 282(47): p. 34535-41. Wolf, D., et al., Primer binding site-dependent restriction of murine leukemia virus requires HP1 binding by TRIM28. J Virol, 2008. 82(9): p. 4675-9. Schultz, D.C., et al., SETDB1: a novel KAP-1-associated histone H3, lysine 9specific methyltransferase that contributes to HP1-mediated silencing of  114  147.  148. 149. 150. 151. 152. 153. 154. 155. 156. 157. 158. 159. 160. 161. 162.  euchromatic genes by KRAB zinc-finger proteins. Genes Dev, 2002. 16(8): p. 919-32. Ryan, R.F., et al., KAP-1 corepressor protein interacts and colocalizes with heterochromatic and euchromatic HP1 proteins: a potential role for Kruppelassociated box-zinc finger proteins in heterochromatin-mediated gene silencing. Mol Cell Biol, 1999. 19(6): p. 4366-78. Sarraf, S.A. and I. Stancheva, Methyl-CpG binding protein MBD1 couples histone H3 methylation at lysine 9 by SETDB1 to DNA replication and chromatin assembly. Mol Cell, 2004. 15(4): p. 595-605. Dodge, J.E., et al., Histone H3-K9 methyltransferase ESET is essential for early development. Mol Cell Biol, 2004. 24(6): p. 2478-86. Tzeng, T.Y., et al., Epigenetic regulation of the Drosophila chromosome 4 by the histone H3K9 methyltransferase dSETDB1. Proc Natl Acad Sci U S A, 2007. 104(31): p. 12691-6. Yoon, J., et al., dSETDB1 and SU(VAR)3-9 sequentially function during germline-stem cell differentiation in Drosophila melanogaster. PLoS One, 2008. 3(5): p. e2234. Rangan, P., et al., piRNA Production Requires Heterochromatin Formation in Drosophila. Curr Biol, 2011. Andersen, E.C. and H.R. Horvitz, Two C. elegans histone methyltransferases repress lin-3 EGF transcription to inhibit vulval development. Development, 2007. 134(16): p. 2991-9. Soppe, W.J., et al., DNA methylation controls histone H3 lysine 9 methylation and heterochromatin assembly in Arabidopsis. EMBO J, 2002. 21(23): p. 6549-59. Jackson, J.P., et al., Dimethylation of histone H3 lysine 9 is a critical mark for DNA methylation and gene silencing in Arabidopsis thaliana. Chromosoma, 2004. 112(6): p. 308-15. Tran, R.K., et al., Chromatin and siRNA pathways cooperate to maintain DNA methylation of small transposable elements in Arabidopsis. Genome Biol, 2005. 6(11): p. R90. Ebbs, M.L. and J. Bender, Locus-specific control of DNA methylation by the Arabidopsis SUVH5 histone methyltransferase. Plant Cell, 2006. 18(5): p. 116676. Tamaru, H., et al., Trimethylated lysine 9 of histone H3 is a mark for DNA methylation in Neurospora crassa. Nat Genet, 2003. 34(1): p. 75-9. Freitag, M., et al., HP1 is essential for DNA methylation in neurospora. Mol Cell, 2004. 13(3): p. 427-34. Tsumura, A., et al., Maintenance of self-renewal ability of mouse embryonic stem cells in the absence of DNA methyltransferases Dnmt1, Dnmt3a and Dnmt3b. Genes Cells, 2006. 11(7): p. 805-14. Pannell, D., et al., Retrovirus vector silencing is de novo methylase independent and marked by a repressive histone code. Embo J, 2000. 19(21): p. 5884-94. Nakayama, J., et al., Role of histone H3 lysine 9 methylation in epigenetic control of heterochromatin assembly. Science, 2001. 292(5514): p. 110-3.  115  163. 164. 165. 166. 167. 168. 169. 170. 171. 172. 173. 174. 175. 176.  177. 178. 179. 180.  Lehnertz, B., et al., Suv39h-mediated histone H3 lysine 9 methylation directs DNA methylation to major satellite repeats at pericentric heterochromatin. Curr Biol, 2003. 13(14): p. 1192-200. Dong, K.B., et al., DNA methylation in ES cells requires the lysine methyltransferase G9a but not its catalytic activity. Embo J, 2008. 27(20): p. 2691-701. Tachibana, M., et al., Functional dynamics of H3K9 methylation during meiotic prophase progression. Embo J, 2007. 26(14): p. 3346-59. Matsui, T., et al., Proviral silencing in embryonic stem cells requires the histone methyltransferase ESET. Nature, 2010. 464(7290): p. 927-31. Tamura, K., et al., MEGA4: Molecular Evolutionary Genetics Analysis (MEGA) software version 4.0. Mol Biol Evol, 2007. 24(8): p. 1596-9. Tachibana, M., et al., G9a/GLP complexes independently mediate H3K9 and DNA methylation to silence transcription. EMBO J, 2008. 27(20): p. 2681-90. Nickerson, J.A., et al., Immunolocalization in three dimensions: immunogold staining of cytoskeletal and nuclear matrix proteins in resinless electron microscopy sections. Proc Natl Acad Sci U S A, 1990. 87(6): p. 2259-63. Kourmouli, N., et al., Heterochromatin and tri-methylated lysine 20 of histone H4 in animals. J Cell Sci, 2004. 117(Pt 12): p. 2491-501. Meunier, J., et al., Homology-dependent methylation in primate repetitive DNA. Proc Natl Acad Sci U S A, 2005. 102(15): p. 5471-6. Maksakova, I.A. and D.L. Mager, Transcriptional regulation of early transposon elements, an active family of mouse long terminal repeat retrotransposons. J Virol, 2005. 79(22): p. 13865-74. Pannell, D., et al., Retrovirus vector silencing is de novo methylase independent and marked by a repressive histone code. Embo J, 2000. 19(21): p. 5884-94. Niwa, O., et al., Independent mechanisms involved in suppression of the Moloney leukemia virus genome during differentiation of murine teratocarcinoma cells. Cell, 1983. 32(4): p. 1105-13. Kano, H., H. Kurahashi, and T. Toda, Genetically regulated epigenetic transcriptional activation of retrotransposon insertion confers mouse dactylaplasia phenotype. Proc Natl Acad Sci U S A, 2007. 104(48): p. 19034-9. Schultz, D.C., J.R. Friedman, and F.J. Rauscher, 3rd, Targeting histone deacetylase complexes via KRAB-zinc finger proteins: the PHD and bromodomains of KAP-1 form a cooperative unit that recruits a novel isoform of the Mi-2alpha subunit of NuRD. Genes Dev, 2001. 15(4): p. 428-43. Rowe, H.M., et al., KAP1 controls endogenous retroviruses in embryonic stem cells. Nature, 2010. 463(7278): p. 237-40. Kim, S.H., et al., Differential DNA methylation reprogramming of various repetitive sequences in mouse preimplantation embryos. Biochem Biophys Res Commun, 2004. 324(1): p. 58-63. Lane, N., et al., Resistance of IAPs to methylation reprogramming may provide a mechanism for epigenetic inheritance in the mouse. Genesis, 2003. 35(2): p. 8893. Hawley, R.G., et al., Versatile retroviral vectors for potential use in gene therapy. Gene Ther, 1994. 1(2): p. 136-8.  116  181. 182. 183. 184. 185. 186. 187. 188. 189. 190. 191. 192. 193. 194. 195. 196. 197. 198.  Cherry, S.R., et al., Retroviral expression in embryonic stem cells and hematopoietic stem cells. Mol Cell Biol, 2000. 20(20): p. 7419-26. Lorincz, M.C., D. Schubeler, and M. Groudine, Methylation-mediated proviral silencing is associated with MeCP2 recruitment and localized histone H3 deacetylation. Mol Cell Biol, 2001. 21(23): p. 7913-22. Hawley, R.G., et al., The HOX11 homeobox-containing gene of human leukemia immortalizes murine hematopoietic precursors. Oncogene, 1994. 9(1): p. 1-12. Leung, D.C., et al., Lysine methyltransferase G9a is required for de novo DNA methylation and the establishment, but not the maintenance, of proviral silencing. Proc Natl Acad Sci U S A, 2011. 108(14): p. 5718-23. Lachner, M., et al., Methylation of histone H3 lysine 9 creates a binding site for HP1 proteins. Nature, 2001. 410(6824): p. 116-20. Yao, S., et al., Retrovirus silencing, variegation, extinction, and memory are controlled by a dynamic interplay of multiple epigenetic modifications. Mol Ther, 2004. 10(1): p. 27-36. Swindle, C.S., H.G. Kim, and C.A. Klug, Mutation of CpGs in the murine stem cell virus retroviral vector long terminal repeat represses silencing in embryonic stem cells. J Biol Chem, 2004. 279(1): p. 34-41. Appanah, R., et al., An unmethylated 3' promoter-proximal region is required for efficient transcription initiation. PLoS Genet, 2007. 3(2): p. e27. Vakoc, C.R., et al., Histone H3 lysine 9 methylation and HP1gamma are associated with transcription elongation through mammalian chromatin. Mol Cell, 2005. 19(3): p. 381-91. Zhou, Y., E.B. Cambareri, and J.A. Kinsey, DNA methylation inhibits expression and transposition of the Neurospora Tad retrotransposon. Mol Genet Genomics, 2001. 265(4): p. 748-54. Lindroth, A.M., et al., Requirement of CHROMOMETHYLASE3 for maintenance of CpXpG methylation. Science, 2001. 292(5524): p. 2077-80. Sharif, J., et al., The SRA protein Np95 mediates epigenetic inheritance by recruiting Dnmt1 to methylated DNA. Nature, 2007. 450(7171): p. 908-12. Ooi, S.K., et al., Dynamic instability of genomic methylation patterns in pluripotent stem cells. Epigenetics Chromatin, 2010. 3(1): p. 17. Jackson-Grusby, L., et al., Loss of genomic methylation causes p53-dependent apoptosis and epigenetic deregulation. Nat Genet, 2001. 27(1): p. 31-9. Dodge, J.E., et al., Inactivation of Dnmt3b in mouse embryonic fibroblasts results in DNA hypomethylation, chromosomal instability, and spontaneous immortalization. J Biol Chem, 2005. 280(18): p. 17986-91. Myant, K., et al., LSH and G9a/GLP complex are required for developmentally programmed DNA methylation. Genome Res, 2011. 21(1): p. 83-94. De La Fuente, R., et al., Lsh is required for meiotic chromosome synapsis and retrotransposon silencing in female germ cells. Nat Cell Biol, 2006. 8(12): p. 1448-54. Bushman, F.D., Targeting survival: integration site selection by retroviruses and LTR-retrotransposons. Cell, 2003. 115(2): p. 135-8.  117  199. 200. 201. 202. 203. 204. 205. 206. 207. 208. 209. 210. 211. 212. 213. 214. 215. 216. 217.  Karimi, M.M., et al., DNA Methylation and SETDB1/H3K9me3 Regulate Predominantly Distinct Sets of Genes, Retroelements, and Chimeric Transcripts in mESCs. Cell Stem Cell, 2011. 8(6): p. 676-87. Komashko, V.M. and P.J. Farnham, 5-azacytidine treatment reorganizes genomic histone modification patterns. Epigenetics, 2010. 5(3). Thomas, C.E., A. Ehrhardt, and M.A. Kay, Progress and problems with the use of viral vectors for gene therapy. Nat Rev Genet, 2003. 4(5): p. 346-58. Wu, X., et al., Transcription start regions in the human genome are favored targets for MLV integration. Science, 2003. 300(5626): p. 1749-51. Ellis, J., Silencing and variegation of gammaretrovirus and lentivirus vectors. Hum Gene Ther, 2005. 16(11): p. 1241-6. Macfarlan, T.S., et al., Endogenous retroviruses and neighboring genes are coordinately repressed by LSD1/KDM1A. Genes Dev, 2011. 25(6): p. 594-607. Jurka, J., et al., Repbase Update, a database of eukaryotic repetitive elements. Cytogenet Genome Res, 2005. 110(1-4): p. 462-7. Leeb, M., et al., Polycomb complexes act redundantly to repress genomic repeats and genes. Genes Dev, 2010. 24(3): p. 265-76. Hake, S.B., et al., Expression patterns and post-translational modifications associated with mammalian histone H3 variants. J Biol Chem, 2006. 281(1): p. 559-68. Daujat, S., et al., H3K64 trimethylation marks heterochromatin and is dynamically remodeled during developmental reprogramming. Nat Struct Mol Biol, 2009. 16(7): p. 777-81. Wu, H., et al., Dual functions of Tet1 in transcriptional regulation in mouse embryonic stem cells. Nature. He, Y.F., et al., Tet-Mediated Formation of 5-Carboxylcytosine and Its Excision by TDG in Mammalian DNA. Science, 2011. Globisch, D., et al., Tissue distribution of 5-hydroxymethylcytosine and search for active demethylation intermediates. PLoS One, 2010. 5(12): p. e15367. Kriaucionis, S. and N. Heintz, The nuclear DNA base 5-hydroxymethylcytosine is present in Purkinje neurons and the brain. Science, 2009. 324(5929): p. 929-30. Song, C.X., et al., Selective chemical labeling reveals the genome-wide distribution of 5-hydroxymethylcytosine. Nat Biotechnol, 2010. 29(1): p. 68-72. Szwagierczak, A., et al., Sensitive enzymatic quantification of 5hydroxymethylcytosine in genomic DNA. Nucleic Acids Res, 2010. 38(19): p. e181. Nestor, C., et al., Enzymatic approaches and bisulfite sequencing cannot distinguish between 5-methylcytosine and 5-hydroxymethylcytosine in DNA. Biotechniques, 2010. 48(4): p. 317-9. Williams, K., et al., TET1 and hydroxymethylcytosine in transcription and DNA methylation fidelity. Nature, 2011. 473(7347): p. 343-8. Kuramochi-Miyagawa, S., et al., MVH in piRNA processing and gene silencing of retrotransposons. Genes Dev, 2010. 24(9): p. 887-92.  118  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

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"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0072395/manifest

Comment

Related Items