Relative Roles of UBF and RRN3 in the Transcription of the Ribosomal RNA Genes and Ribosome Biogenesis Determined Using in Vivo Mouse Models

Relative roles of UBF and RRN3 in the transcription of the ribosomal RNA genes and ribosome biogenesis determined using in vivo mouse models

Thèse

Chelsea Herdman

Doctorat en biologie cellulaire et moléculaire Philosophiae doctor (Ph.D.)

Québec, Canada

Relative roles of UBF and RRN3 in the transcription of the ribosomal RNA genes and ribosome biogenesis determined using in vivo mouse models

Thèse

Chelsea Herdman

Sous la direction de :

Tom Moss, directeur de recherche

Résumé

La biogenèse des ribosomes, aussi appelée la synthèse ribosomale, est un processus cellulaire important se déroulant dans le nucléole et implique la transcription par les trois ARN polymérases nucléaires. L’étape initiale et limitante de ce processus est la transcription des ARNs ribosomaux catalytiques, 28S, 18S and 5.8S, sous la forme d’un long précurseur d’ARN ribosomal (pre-ARNr/47S) par l’ARN polymérase I (RPI). RPI possède un ensemble de facteurs de transcription généraux responsables de son activation. Ces facteurs sont la protéine architecturale UBF, le facteur SL1 qui contient TBP, le facteur d’initiation RRN3 et le facteur de terminaison TTF1. La synthèse de l’ARN ribosomale est finement régulée et correspond à 30-50% de l’ensemble de la transcription de la cellule. De plus, ce processus est lié à la croissance cellulaire, la transformation, la prolifération et à l’activité des facteurs suppresseurs de tumeurs et des oncogènes. UBF et RRN3 sont notamment activés par plusieurs voies de signalisation de croissance cellulaire. Dans les cellules de mammifère, il existe ~200 copies d’ADNr par génome haploïde. Les fragments répétés d’ADN ribosomal sont arrangés en répétition en tandem sur les bras courts des chromosomes acrocentriques. De façon intéressante, dans les cellules somatiques, seulement la moitié des copies d’ADNr sont actives, alors que les autres sont maintenues dans une forme inactive par les modifications épigénétiques et la formation d’hétérochromatine. La raison pour laquelle le génome contient autant de copies et la régulation de leur activité ne sont pas bien comprises.

Cette thèse présente l’analyse de l’importance in vivo d’UBF et de RRN3 pour la régulation de la transcription de l’ARNr et pour le maintien de la structure chromatinienne de l’ADNr. Nous avons précédemment analysé la perte de fonction de UBF dans les fibroblastes embryonnaires de souris en utilisant le système de perte de fonction conditionnelle dépendante du tamoxifène. Puisque l’un de nos objectifs était de comparer la fonction de RRN3 dans un modèle similaire, nous avons réanalysé la perte de fonction de RRN3 chez la souris et généré des lignées cellulaires comme préalablement réalisées avec la perte de fonction d’UBF. Nous avons déterminé que RRN3 est essentiel à la préimplantation et le développement est arrêté à E3.5, ce qui contredit les résultats obtenus par un autre groupe qui avait obtenu un arrêt du développement beaucoup plus tardif, à E9.5. Une lignée de fibroblastes embryonnaires de souris inductible au tamoxifène a été créée pour RRN3 de façon similaire à ce qui avait été fait pour UBF. La perte de fonction d’UBF ou de RRN3 inhibe la transcription par RPI. Par contre, nous démontrons que UBF est responsable du recrutement à l’ADNr des autres facteurs associés à RPI et du maintient de l’état ouvert de la chromatine. En comparaison, RRN3 est requis simplement pour le recrutement de RPI. Dans cette étude, nous avons également identifié une région frontalière en amont de l’ADNr formée de H2A.Z, TTF1, CTCF et des modifications d’histones activatrices. Nous avons également découvert que la perte d’UBF entraine une mort cellulaire synchronisée par apoptose, indépendamment de p53 et ce spécifiquement dans les lignées cellulaires transformées. Ce résultat suggère

iii

qu’il pourrait être possible de cibler UBF dans le traitement contre le cancer puisque la perte de UBF dans les lignées cellulaires primaires cause un arrêt de prolifération sans entrainer l’apoptose. Finalement, nous avons observé que le niveau d’activité de l’ADNr dans les cellules pluripotentes est différent que dans les cellules différenciées. Des lignées de cellules souches embryonnaires (ESCs) ont été générées à partir des souris conditionnelles pour UBF et RRN3 et nos résultats préliminaires suggèrent que la totalité des gènes de l’ADNr est active dans les cellules pluripotentes. Ce modèle est idéal pour étudier la régulation de l’ADNr ainsi que le rôle de UBF et RRN3 dans cette régulation après l’induction de la différentiation. En résumé, ces résultats permettront de clarifier le rôle in vivo de UBF et RRN3 dans la transcription de l’ARN ribosomal et dans le maintien de l’intégrité de l’ADNr.

Abstract

Ribosome biogenesis, or the synthesis of ribosomes, is an important cell process occurring in the nucleolus that utilizes transcription by all three nuclear RNA polymerases. The initial and rate-limiting step is the transcription of the catalytic ribosomal RNAs 28S, 18S and 5.8S in the form of a precursor ribosomal RNA (pre-rRNA/47S) by RNA polymerase I (RPI, also known as Pol1 and POLR1). RPI has a dedicated set of basal factors responsible for its activation. These are the architectural factor UBF, the TBP containing factor SL1, the initiation factor RRN3, and the termination factor TTF1. Ribosomal RNA synthesis is tightly regulated and accounts for 30-50% of total gene transcription. As such, this process is linked to cell growth, transformation, proliferation and the actions of tumour suppressors and oncogenes. Notably, UBF and RRN3 are activated by many of the same growth signaling pathways.

The human and mouse haploid genome contain ~200 copies of the ribosomal RNA genes, the ribosomal DNA (rDNA). These ribosomal DNA copies are arranged in tandem repeats on the short arms of acrocentric chromosomes. Interestingly, only a fraction of the rDNA copies are active, and a significant number are epigenetically silenced and heterochromatic. The reason for having so many copies and their regulation in vivo by silencing is not yet understood, though it has been connected with genome stability.

This thesis presents the analysis of the in vivo requirements for UBF and RRN3 in rRNA transcription and rDNA chromatin structure. We had previously analyzed the loss of UBF in mouse embryonic fibroblasts using tamoxifen-dependent conditional knockout. As we wanted to compare the loss of RRN3 in a similar model, we re-analyzed the RRN3 knockout mice and created cell lines as was performed for the UBF knockout. Importantly, we find that RRN3 is essential for preimplantation and its loss arrests development at E3.5, contrary to previous work that showed a late E9.5 developmental arrest. Using mouse embryonic fibroblast (MEF) cell lines conditional for UBF or RRN3, we found that the loss of either factor prevented RPI transcription. However, we found that UBF was essential for the recruitment of the other RPI transcription factors and the formation of the preinitiation complex, as well as to maintain an open rDNA chromatin structure, while RRN3 was required only for RPI recruitment. These studies allowed us to identify an upstream boundary element on the rDNA formed of H2A.Z, TTF1, CTCF and activating histone marks, which is independent of RPI activity. We also found that UBF loss, but not RRN3 loss, led to a synchronous and massive p53-independent apoptosis, specifically in oncogenically transformed cells. This strongly suggests that drug targeting UBF could be a viable cancer treatment. Finally, we have observed that the rDNA activity status in pluripotent cells differs from that of differentiated cells. Embryonic stem cells (ESCs) were also generated from the mice conditional for UBF and RRN3. Preliminary results indicate that, unlike somatic cells, all the rRNA genes in these and other pluripotent cell lines are potentially active. This makes ESCs and their

differentiation an ideal model in which to study the establishment of rDNA silencing and the role of UBF and/or RRN3 in this process. Together these data define the in vivo roles of UBF and RRN3 in ribosomal RNA transcription and suggest mechanisms by which they maintain rDNA integrity and may drive cell differentiation.

Table of Contents

Résumé ...... iii Abstract ...... v Table of Contents...... vii List of Tables ...... x List of Figures ...... xi Abbreviations ...... xiii Acknowledgements ...... xviii Preface ...... xx 1 Introduction ...... 1 1.1 The nucleolus: site of ribosome biogenesis ...... 1 1.1.1 The structure of the nucleolus ...... 1 1.1.2 Nucleolar organizer regions ...... 3 1.1.3 Functions of the nucleolus ...... 6 1.1.3.1 Ribosome biogenesis ...... 7 1.1.3.2 Cell cycle regulation ...... 8 1.1.3.3 Stress response ...... 9 1.1.3.4 Control of aging ...... 10 1.1.3.5 Other RNA maturation ...... 12 1.2 Ribosomal DNA ...... 12 1.2.1 Organization of the rDNA ...... 13 1.2.2 Epigenetic regulation of the rDNA ...... 15 1.2.2.1 rDNA activity status and chromatin ...... 15 1.2.2.2 Methylation of the rDNA ...... 18 1.2.2.3 Mechanism of rDNA silencing ...... 19 1.2.3 DNA damage and genomic instability of the rDNA...... 22 1.3 Ribosomal RNA transcription ...... 24 1.3.1 Formation of the pre-initiation complex ...... 24 1.3.1.1 UBF ...... 26 1.3.1.2 SL1 ...... 28 1.3.1.3 Pol I ...... 28 1.3.1.4 RRN3 ...... 29 1.3.1.5 Mechanism of PIC assembly ...... 30 1.3.2 Initiation ...... 31 1.3.3 Elongation ...... 31 1.3.4 Termination ...... 32 1.3.4.1 TTF1 ...... 32 1.3.5 Reinitiation ...... 33 1.3.6 Regulation of rRNA gene activity and transcription rates ...... 33 1.3.7 Cell growth, proliferation and rRNA transcription ...... 35

vii

1.3.7.1 Cell growth ...... 35 1.3.7.2 Cell cycle ...... 35 1.3.7.3 Mitogen activation of rRNA transcription ...... 36 1.3.8 rRNA transcription and apoptosis ...... 39 1.4 Ribosome biogenesis and disease ...... 41 1.4.1 Ribosomopathies ...... 41 1.4.2 Cancer ...... 42 1.4.2.1 c-Myc ...... 44 1.4.2.2 Nucleophosmin/B23 ...... 44 1.4.2.3 p53 ...... 45 1.4.2.4 RB ...... 45 1.4.2.5 ARF, MDM2 and TTF ...... 45 1.4.2.6 Ribosomal RNA transcription as a target for cancer treatment ...... 47 1.4.3 Viral infections ...... 49 1.4.4 Neurodegenerative disorders ...... 49 1.5 Mouse embryonic development...... 49 1.5.1 From totipotent to pluripotent; the first days of development ...... 50 1.5.2 Maternal-to-zygotic transition ...... 51 1.5.3 Chromatin regulation of maternal and paternal genomes ...... 53 1.5.4 The developing nucleolus ...... 54 Thesis objectives ...... 56 2 An enhancer adjacent chromatin boundary is maintained on the ribosomal RNA gene repeats even in the absence of the basal factors and active transcription...... 59 2.1 Preface ...... 59 2.2 Résumé ...... 60 2.3 Abstract ...... 61 2.4 Introduction ...... 62 2.5 Results ...... 63 2.5.1 A high-resolution map of basal factors across the rRNA genes...... 63 2.5.2 UBF binding is precisely delimited to the functional rDNA unit ...... 65 2.5.3 Mouse RRN3/TIF-1A is an essential protein ...... 65 2.5.4 Neither RRN3 nor RPI transcription is required to maintain the potentially active state of the rRNA genes68 2.5.5 UBF is essential for the recruitment of all components of the RPI transcription machinery ...... 70 2.5.6 UBF determines psoralen accessible and nucleosome exclusion on the active rRNA genes ...... 70 2.5.7 A chromatin boundary is found upstream of the spacer promoter ...... 72 2.5.8 TTF1 regulates spacer promoter transcription by arresting polymerase on active genes ...... 74 2.6 Discussion ...... 75 2.7 Materials and methods ...... 77 2.8 Acknowledgements ...... 80 2.9 Supplemental data ...... 80 3 Depletion of the cisplatin targeted HMGB-box factor UBF selectively induces p53 independent apoptotic death in transformed cells ...... 89 3.1 Preface ...... 89 3.2 Résumé ...... 90

viii

3.3 Abstract ...... 91 3.4 Introduction ...... 92 3.5 Results ...... 93 3.5.1 Cisplatin displaces UBF from the mouse rRNA genes and arrests their transcription ...... 94 3.5.2 UBF loss disrupts nucleolar functions in both primary and transformed MEFs ...... 96 3.5.3 Transformed iMEFs, but not primary MEFs, undergo synchronous apoptosis following Ubf inactivation... 96 3.5.4 Apoptosis is accompanied by the generation of a “nucleosomal ladder” of DNA cleavage ...... 99 3.5.5 UBF loss blocks proliferation and DNA replication, causing cell cycle arrest ...... 99 3.5.6 Apoptosis induced by UBF loss is p53 independent ...... 101 3.5.7 p53-independent apoptosis is a general response to UBF loss in oncogene stressed cells ...... 104 3.5.8 Oncogenic stress may induce apoptosis by aberrantly driving cells into S-phase ...... 104 3.6 Discussion ...... 106 3.7 Materials & Methods ...... 107 3.8 Acknowledgements ...... 109 3.9 References ...... 109 3.10 Supplemental data ...... 114 3.10.1 Supplemental data references ...... 121 4 The role of ribosomal DNA in maintaining or establishing pluripotency ...... 123 4.1 Preface ...... 123 4.2 Résumé ...... 124 4.3 Abstract ...... 125 4.4 Introduction ...... 126 4.5 Results ...... 127 4.5.1 A fraction of the rRNA genes are highly methylated in somatic cells but not in ES cells ...... 127 4.5.2 The rRNA genes are all psoralen-accessible in ES cells ...... 128 4.5.3 Embryonic stem cells display an upstream boundary of CTCF, TTF1 and a poised RNA polymerase I .. 129 4.5.4 Loss of UBF in ESCs leads to the loss of active rDNA repeats ...... 131 4.5.5 Using directed differentiation to study rDNA silencing throughout loss of pluripotency ...... 133 4.6 Conclusion ...... 133 4.7 Materials and Methods ...... 135 4.8 Acknowledgements ...... 138 4.9 Supplemental Data ...... 139 5 Discussion ...... 141 5.1 Development of the RRN3 conditional knockout mice and cell lines ...... 142 5.1.1 RRN3 is required for early embryonic development ...... 142 5.1.2 Tamoxifen-induced knockout of RRN3 in cell culture ...... 144 5.2 The role of UBF and RRN3 in rRNA transcription and rDNA chromatin structure ...... 146 5.3 The loss of UBF and genomic instability ...... 151 5.4 Tissue-specific loss of UBF and RRN3 ...... 152 5.5 Concluding remarks ...... 154 References ...... 155 Annexes ...... 177

List of Tables

Table 1.1 Identified phosphorylation sites of UBF ...... 38 Table 1.2 Identified phosphorylation sites of RRN3 ...... 39 Table 1.3 Anticancer drugs in use that inhibit rRNA synthesis ...... 48

List of Figures

Figure 1.1 Structure of the nucleolus as seen through an electron microscope ...... 2 Figure 1.2 The position of NORs on human and mouse chromosomes ...... 4 Figure 1.3 Mammalian acrocentric chromosome carrying a NOR ...... 5 Figure 1.4 UBF is found on active NORs throughout the cell cycle ...... 6 Figure 1.5 Ribosome biogenesis involves all three nuclear polymerases ...... 8 Figure 1.6 Nuceolar stress ...... 10 Figure 1.7 Model for eNoSC function in the nucleolus ...... 11 Figure 1.8 Ribosomal RNA is co-transcriptionally assembled ...... 13 Figure 1.9 Ribosomal DNA organization in mammals...... 14 Figure 1.10 The epigenetic readers, writers and erasers ...... 16 Figure 1.11 Psoralen crosslinking of active and inactive rDNA ...... 18 Figure 1.12 Model of the mechanism of rDNA silencing by NoRC ...... 20 Figure 1.13 Transcription of a lncRNA from the intergenic spacer ...... 21 Figure 1.14 TTF1 as a central regulator of rDNA activity status ...... 22 Figure 1.15 DNA damage response pathway ...... 23 Figure 1.16 Pre-initiation complex assembly in yeast and mammals ...... 25 Figure 1.17 The enhancesome model of UBF binding ...... 27 Figure 1.18 rDNA transcription is cell cycle dependant...... 36 Figure 1.19 rRNA transcription is linked to cell growth...... 37 Figure 1.20 Main apoptotic signaling pathways ...... 41 Figure 1.21 Oncogenes and tumor suppressor regulation of RPI basal factors ...... 43 Figure 1.22 c-Myc regulates multiple levels of ribosome biogenesis ...... 44 Figure 1.23 ARF is a negative regulator of ribosome biogenesis ...... 46 Figure 1.24 NPM and ARF regulate TTF1 localization and activity ...... 47 Figure 1.25 Preimplantation development in the mouse ...... 51 Figure 1.26 Gene expression regulation during early development ...... 52 Figure 1.27 UBF is essential for NPB formation in early embryos ...... 56 Figure 2.1 RNA polymerase I basal factor occupancy across the rDNA ...... 64 Figure 2.2 RRN3 is essential for RNA polymerase I recruitment to the rDNA ...... 67 Figure 2.3 UBF is necessary for RPI machinery recruitment ...... 69 Figure 2.4 Loss of UBF induces nucleosomal formation and shut down of rRNA genes ...... 71 Figure 2.5 A chromatin boundary element is found upstream of the rDNA unit ...... 73 Figure 2.6 Activating histone marks at the spacer promoter are increased after loss of UBF ...... 74 Figure 2.7 A stalled RNA polymerase I is found just downstream of the spacer promoter ...... 75 Figure 2.8 UBF occupancy correlates with GC-rich sequences in the rRNA gene body ...... 80 Figure 2.9 Conditional knockout models of RRN3 and UBF ...... 81 Figure 2.10 Deletion of the Rrn3/Tif1a gene arrests mouse development during early cleavage divisions ...... 83 Figure 2.11 qPCR analysis after RRN3 or UBF knockout ...... 85 Figure 2.12 TTF1 diminishes its presence through the gene body after knockout ...... 86 Figure 2.13 MNase digestion of UBF and RRN3 knockout MEFs ...... 87 Figure 3.1 Cisplatin treatment of Ubfwt/wt/Er-cre+/+/SvT iMEFS induces displacement of UBF from the nucleolus ..... 94 Figure 3.2 Cisplatin coordinately displaces UBF from the rRNA genes and arrests their transcription...... 95 Figure 3.3 UBF loss induces synchronous apoptotic cell death selectively in oncogenically transformed iMEFs ..... 97 Figure 3.4 UBF loss induces selective Caspase 3 cleavage in transformed iMEFs cells ...... 98

Figure 3.5 UBF loss arrests cell proliferation and leads to a cell cycle arrest ...... 100 Figure 3.6 Apoptosis of oncogenically transformed cells after Ubf gene inactivation is p53 independent ...... 102 Figure 3.7 p53 independent apoptosis is a general response to UBF loss in an oncogenic stress context ...... 103 Figure 3.8 Cell cycle distribution of p53-null cells during UBF depletion ...... 105 Figure 3.9 Cisplatin treatment of MEFs induces displacement of UBF from the nucleolus ...... 114 Figure 3.10 Analysis of UBF loss in primary MEFs ...... 115 Figure 3.11 Primary MEFs survive UBF loss while SV40Tt transformed iMEFs suffer cell death...... 117 Figure 3.12 TIF1A loss does not induce TUNEL positive apoptosis in SV40Tt transformed MEFs ...... 118 Figure 3.13 UBF loss leads to a cell cycle arrest and to a loss of mitotic cells ...... 119 Figure 3.14 p53-independent apoptosis is a general response to UBF loss in an oncogenic stress context ...... 120 Figure 4.1 rRNA genes in embryonic stem cells are fully unmethylated in the gene and control regions ...... 128 Figure 4.2 Psoralen crosslinking analysis of the rRNA genes...... 129 Figure 4.3 ChIP-seq analysis of the RPI machinery in ESCs ...... 130 Figure 4.4 CTCF and cohesin are found upstream of the spacer promoter in ESCs ...... 131 Figure 4.5 Preliminary 4-HT treatment of ES cells to induce CRE-mediated excision of Ubf and Rrn3 ...... 132 Figure 4.6 Knockout of UBF leads to the formation of a closed conformation of the rDNA ...... 132 Figure 4.7 Retinoic acid induced differentiation of embryonic stem cells ...... 133 Figure 4.8 Generating embryonic stem cell lines ...... 139

xii

Abbreviations

A adenine A adenosine a.a amino acid ActD actinomycin D AgNOR argyrophilic nucleolar organiser region ALU Arthrobacter luteus (element) AML acute myeloid leukemia AP alkaline phosphatase ARF alternative reading frame ARN acide ribonucléique ARNr ARN ribosomal ATM ataxia-telangiectasia mutated ATP adenosine triphosphate bp base pair BrUTP bromouridine-triphosphate C cytosine c-Myc avian myelocytomatosis virus oncogene cellular homolog CDK-2/4 cyclin-dependent kinase 2/4 CF core factor ChIP chromatin immunoprecipitation CK2 casein kinase II CpG cytosine-phosphate-guanine Cre causes recombination CSB Cockayne syndrome group B CTCF CCCTC-binding factor Ctrl control CV crystal violet DAPI 4′,6-diamidino-2-phenylindole DDR DNA-damage response DDX5 DEAD-box helicase 5 DFC dense fibrillar component DMEM Dulbecco’s modified eagle’s medium DNA deoxyribonucleic acid DNMT DNA (cytosine-5)-methyltransferase 1 DPC days post coitum DSB double strand breaks dsDNA double-stranded DNA E-Syt extended-synaptotagmin EDTA ethylenediaminetetraacetic acid EF3 elongation factor 3 EGTA ethylene glycol-bis(β-aminoethyl ether EM electron microscope eNoSC energy-dependent nucleolar silencing complex EPI epiblast EpiSCs epiblast stem cells ER-Cre estrogen receptor-Cre ERK extracellular signal–regulated kinases

xiii

ESC embryonic stem cell EtBr ethidium bromide ETS external transcribed sequences FACS fluorescence-activated cell sorting FBI fibrillarin FBS fetal bovine serum FC fibrillar center FCP1 TFIIF-associating component of CTD phosphatase FGF fibroblast growth factor FGFR2 FGF receptor FISH fluorescence in-situ hybridization Fob1p fork blocking less 1 protein G guanine GC granular component GC granular component gDNA genomic DNA GSK3 glycogen synthetase kinase 3 GTP guanosine-5'-triphosphate H histone h hour H2A.Z histone 2 A family member Z HDAC histone deacetylases HEAT Huntingtin, elongation factor 3, protein phosphatase 2A, TOR1 HMG high-mobility group HR homologous recombination ICM inner cell mass IF immunofluorescence IGS intergenic spacer IGV integrative genomics viewer iMEF immortalized MEF IP immunoprecipitation ITS internal transcribed sequences KLF Krüppel-like factor 2 KO knockout LBD ligand binding domain LIF leukemia inhibitory factor LINE long interspersed nuclear element lncRNA long non-coding RNA MAPK mitogen-activated protein kinase MDa mega dalton MDM2 murine E3 ligase murine double minute 2 me methylation MEF mouse embryonic fibroblast MEK MAPK/ERK kinase miRNA microRNA mRNA messenger RNA mTOR mammalian target of rapamycin MYMMP1A protein MYB binding protein 1a MZT maternal-to-zygotic transition NaCl sodium chloride

xiv

NAD nicotinamide adenine dinucleotide NBP nucleolar precursor bodies NCL nucleolin ncRNA non-coding RNA NHEJ non-homologous end-joining NML nucleomethylin NOR nucleolar organizer region NoRC nucleolar remodling complex NPM nucleophosmin nt.sec-1 nucleotides per second OCT4 octamer-binding transcription factor 4 PARP poly-(ADP-ribose) polymerase PAX6 paired box 6 PBS phosphate buffered saline PE primitive endoderm PFA paraformaldehyde pHT post 4-hydroxytamoxifen PI propidium iodide PI3K phosphatidyl inositol-3 kinase PIC pre-initiation complex PK proteinase K PMSF phenylmethylsulfonyl fluoride PP1 훾 protein phosphatase 1 isoform gamma PP2A subunit A of protein phosphatase 2A ppRb hyperohosphorylated Rb PTEN phosphatase and tensin homolog qPCR real-time polymerase chain reaction r-proteins ribosomal proteins r-proteins ribosomal proteins r.p.m. revolutions per minute Ras retrovirus-associated DNA sequences from murine sarcoma viruses RB retinoblastoma protein rDNA ribosomal DNA, a group term for the rRNA genes RFB replication fork barrier RNA ribonucleic acid RNP ribonucleoprotein particle RPI RNA polymerase I, also known as POL1 and POLR1 RPL ribosomal protein large subunit RRN3 regulation of RNA polymerase I rRNA ribosomal RNA s/sec second SDS-PAGE sodium dodecyl sulfate polyacrylamide gel electrophoresis SFES serum free embryonic stem cell media SINE short interspersed nuclear element Sir2p silent information regulator 2 protein SIRT1 NAD-dependent deacetylase sirtuin-1 SL1 selectivity factor SMAD1 mothers against decapentaplegic homolog 1 Smc1/3 structural maintenance of chromosomes protein 1/3 snoRNA small nucleolar RNA

snoRNPs small nucleolar ribonucleoprotein SNP single-nucleotide polymorphisms SOX (sex determining region Y)-box SpPr spacer promoter SRP signal recognition particle SSC saline-sodium citrate buffer SSU small subunit STAT signal transducer and activator of transcription 3 STS staurosporin SvT simian vacuolating virus 40 TAg T1 terminator site 1 TAFs TBP-associated factors TBP TATA binding protein TCS Treacher-Collins syndrome TE trophoectoderm TIF-IA transcription initiation factor IA TIF-IB transcription initiation factor IB TIP5 TTF1-interacting protein 5 Topo I topoisomerase I TOR1 target of rapamycin 1 TSS transcription start site TTF1 transcription termination factor 1 TUNEL terminal deoxynucleotidyl transferase-mediated dUTP nick end-labeling UAF upstream activating factor UBF upstream binding factor UCE upstream control element UV ultra violet VE visceral endoderm WGBS whole-genome bisulfite sequencing wt wild-type µCi microcurie ZGA zygotic genome activation

xvi

For Aiden

xvii

Acknowledgements

In the final steps of preparing this thesis, I have been able to step back and take the time to think about the five years I spent in the Moss Lab and realize just how grateful I am for this time in my life. So much of my experience in the lab and in Québec, in general, has shaped my professional and personal life.

Firstly, I would like to thank my thesis director Tom Moss for taking a chance on me and allowing me the opportunity to study in his lab. His enthusiasm for science is contagious and motivating. His patience, especially when I would become bogged down in details, and his encouragement to always keep pushing to find the answer were what helped me to persevere through the more challenging aspects of my Ph.D. Tom is a great mentor and a scientist to admire, especially in a field that is becoming obsessed with getting the end result quickly but not necessarily answering the “big questions”.

I would also like to thank my colleagues in the Moss Lab that I have worked with closely, Victor Stefanovsky, Nourdine Hamdane, Joël Boutin, Prakash Mishra and Jean-Clément Mars. I appreciated the discussion (scientific or not), the laughs, the crosswords and the drinks.

Je remercie Michel Tremblay en particulier, qui m’a enseigné la grosse majorité des protocoles en laboratoire. Il m’a aussi aidé avec plusieurs aspects de la vie de tous les jours qui sont normalement demandés aux parents. Du premier voyage au Costco, jusqu’à être témoin pour notre entrevue avec le greffier avant le mariage, Mike a toujours été là comme deuxième papa pendant mon séjour à Québec et je lui en serai toujours reconnaissante.

Of course, I must thank my husband, Gabriel, for his love, his support, his patience, his everything. His calm demeanor and his ability to encourage me at the same time as challenging me to do better has influenced how I approach research and many other aspects of life. Thank you as well for believing that I can be a student, a mother, a wife, and a future researcher.

I am very grateful for my family, both the old and the new. Thank you to my parents that have always supported me in my endeavors and are always just a phone call away. I owe my love of learning to them as they encouraged my love of reading and allowed me the space to pursue independent learning from a very young age. I also thank them for putting me in French Immersion even though no one in our family spoke the language. Merci aussi à ma nouvelle famille qui m’a accueilli à bras ouverts. Je vous aime beaucoup.

I very much appreciated my time at the research center and would like to thank the St. Patrick community at large. Whether it was to share protocols, borrow antibodies or discuss results, I always found a welcoming and collaborative environment. I’ve found the researchers very available and I appreciate the discussions about

xviii

career development and the reference letters, in particular from Dr. Jacques Côté, Dr. Lucie Jeannotte, Dr. Nicolas Bisson and Dr. Josée Lavoie. I also thank my good friends in the center and out, including Jana Krietsch, Niraj Joshi, and Jean-Clément Mars, for making these years pass too quickly.

I also thank the Canadian Institutes of Health Research for funding my Ph.D. studies and for the opportunity to participate in multiple international conferences.

Finally, I would like to thank my thesis committee, Dr. Marlene Oeffinger, Dr. Lucie Jeannotte, and Dr. Darren Richard for agreeing to read and evaluate this thesis.

xix

Preface

The work presented in this thesis is the product of nearly four years of work as a doctoral student in the laboratory of Tom Moss. The appendices contain work related to my first project in the lab, which was not sustainable as a Ph.D. project and therefore is not presented within the chapters of this dissertation. In total, my time in the Moss lab has led to the publication of one co-first-author paper, two second-author papers, one third-author paper and one first-author review manuscript as well as one more co-first author paper that has been submitted.

The Moss lab is interested in the regulation of cell growth and in particular, ribosome biogenesis. Ribosome biogenesis is activated by mitogenic pathways such as ERK/MAPK and therefore the lab also is interested in the regulation of these growth-activating pathways.

I arrived in the lab summer 2011 and began studying the role of mammalian Extended-Synaptotagmin (E-Syt), a factor that the lab had previously characterized in Xenopus. They found that E-Syt is a negative regulator of ERK activation through its capacity to block internalization of the FGF receptor at the plasma membrane (Jean et al. 2010). My role in this project was to investigate E-Syt in human cell lines and in mouse, and whether the phenotypes observed in Xenopus were conserved. Unfortunately, most of these phenotypes are not conserved in cell culture, and the knockout of one of the E-Syt family members (E-Syt2) gave no phenotype.

Though I continued to collaborate on the E-Syt project throughout the years, I soon shifted my focus to study the regulation of ribosomal RNA transcription. By 2012 I had started characterizing the knockout of RRN3 and I presented this project at my doctoral exam in 2013 when I transitioned from the M.Sc. to the Ph.D. program. I continued my work on RRN3 and UBF, another RNA Polymerase I transcription factor, the results of which are presented in Chapters 2, 3 and 4 of this thesis.

My principle work in comparing the loss of RRN3 to that of UBF and studying the role of these factors in transcription and chromatin regulation of the ribosomal DNA is included in this thesis as Chapter 2. This work entitled “An enhancer adjacent chromatin boundary is maintained on the ribosomal RNA gene repeats even in the absence of the basal factors and active transcription” (Herdman C*, Mars JC*, Stefanovsky VY, Tremblay MG, Sabourin-Felix M, and Moss T) has been submitted as a co-first author paper.

Chapter 3 of this dissertation is a second author manuscript published in Oncotarget in 2015: “Depletion of the cisplatin targeted HMGB-box factor UBF selectively induces p53-independent apoptotic death in transformed cells” (Hamdane N, Herdman C, Mars JC, Stefanovsky V, Tremblay MG, and Moss T). I played an integral part

in the preparation of this article, completing experiments leading to multiple figures and answering major revisions. This article in its published form is also presented as Annex 1.

The most recent aspect of my doctoral research, the development of a model by which to study rDNA transcription in pluripotent cells and preliminary results using these cells, is presented in Chapter 4.

Though my involvement in the E-Syt project was short-lived, it was productive in the end. Firstly I participated in the revision process of the manuscript “The endocytic adapter E-Syt2 recruits the p21 GTPase activated kinase PAK1 to mediate actin dynamics and FGF signalling” (Jean S, Tremblay MG, Herdman C, Guillou F and Moss T), which was published in Biology Open in 2012. In this publication, I was responsible for investigating the role of Xenopus E-Syt in cell migration which resulted in Figures 4b, 4c, and 5b. This publication is included in Annex 2.

Secondly, the negative results demonstrating the lack of phenotype in the mouse after the loss of E-Syt2 and E-Syt3 were published as a co-first-author paper in Cell Cycle in 2014. “Loss of Extended Synaptotagmins ESyt2 and ESyt3 does not affect mouse development or viability, but in vitro cell migration and survival under stress are affected” (Herdman C*, Tremblay MG*, Mishra PK, and Moss T) is included in Annex 3. All figures except 6c and 6d were produced by Michel Tremblay and myself. I participated in the writing and editing of the article.

The third publication that arose from this collaboration was a second author paper in the Journal of Biological Chemistry in 2015. Michel Tremblay did the majority of the work with contributions from the co-authors. I performed the immunofluorescence analysis that accompanied all of the mutants used in the immunoprecipitation experiments but these controls were not included in the paper. I also optimized an immunofluorescence protocol that analyzed endocytosis which translated into Figures 4e-g in the paper. This manuscript entitled “Extended Synaptotagmin Interaction with the Fibroblast Growth Factor Receptor Depends on Receptor Conformation, Not Catalytic Activity” (Tremblay MG, Herdman C, Guillou F, Mishra PK, Baril J, Bellenfant S and Moss T) is included in Annex 4.

Finally, I had the opportunity to write an invited review of the literature for Pharmacological Research in 2016. This paper, “Extended-Synaptotagmins (E-Syts); the extended story” (Herdman C and Moss T), is presented as Annex 5.

xxi

Chapter 1 Introduction

xxii

1 Introduction

Cell growth, or increase in cell mass, requires a significant number of ribosomes to produce the cell’s translational requirement. The biogenesis of ribosomes is therefore tightly linked to the cell’s capacity to grow and therefore proliferate (Lempiäinen & Shore 2009). In fact, 1-2 million ribosomes are produced every cell generation, utilizing all three RNA polymerases and a significant proportion of the cells translational activity. In growing cells, the ribosomal RNAs (rRNAs) account for 35% to 60% of all gene transcription (Moss & Stefanovsky 2002). As the manufacture of ribosomes is so energetically demanding, it is not surprising therefore that it is a major limiting task for cancer cell proliferation. In fact, changes in rRNA levels and the deregulation of many proteins involved in ribosome biogenesis have been linked with cancer and the nucleolus has long been used as a prognostic marker for tumor cells (Amsterdam et al. 2004, Derenzini et al. 2009, Drygin et al. 2010, Ruggero 2012, Ruggero & Pandolfi 2003).

The mammalian ribosome is a 4-MDa complex whose subunits are assembled in the nucleolus, the largest subnuclear organelle, from four rRNAs and ~82 ribosomal proteins (r-proteins). Several hundred other proteins and small RNAs are also needed to assemble the ribosome. Three rRNAs are transcribed from hundreds of nucleolar rDNA repeats by RNA polymerase I (RPI, also known as Pol I, POL1, POLR1), which has a set of dedicated transcription factors, namely UBF, RRN3/TIF-IA, SL1/TIF-IB and TTF1 (Moss & Stefanovsky 2002, Moss et al. 2007, Pederson 2011). These factors play a central role in the regulation of rRNA transcription, and two in particular, UBF and RRN3, have been linked to cell growth and proliferation. The roles these two proteins play in vivo is the primary focus of this thesis and hence will be introduced in the following chapter.

1.1 The nucleolus: site of ribosome biogenesis The nucleolus is the largest body in the nucleus where the machinery necessary for ribosome biogenesis is concentrated. As the nucleolus is visible in brightfield microscopy, it was identified in the early 19th-century (Pederson 2011). However, it was a century later before McClintock and Heitz independently associated the nucleolus with a particular chromosomal locus, thus named the nucleolar organizer region (NOR) (Heitz 1931, McClintock 1934) and still decades before the NOR was associated specifically with the rRNA genes or ribosomal DNA (rDNA) (Pederson 2011). We now know that the nucleolus is the site of ribosomal RNA (rRNA) transcription, maturation and assembly of the ribosomal subunits.

1.1.1 The structure of the nucleolus The nucleolus is formed of three visible compartments that were first described by their density of fibrils or granules observed using the electron microscope (EM). These are the fibrillar centers (FCs) of lower electron density surrounded by the dense fibrillar components (DFCs), which are then further surrounded by the

granular components (GCs) (Figure 1.1) (Farley et al. 2015, Pederson 2011). These compartments are thought to partition the different steps of ribosome biogenesis. For instance, the first step of rDNA transcription likely occurs at the interface between the FC and the DFC, whereas early processing of the rRNA precursor (pre-rRNA) happens in the DFC. The location of the actively transcribed rDNA is, however, still debated and some believe that transcription by RPI also occurs in the DFC while inactive factors are stored in the FC (Raška et al. 2006). Finally, ribosome subunit maturation takes place in the GC, and these subunits are then exported to the cytoplasm, where they undergo the final processing steps to render them translationally active. One may observe the different compartments of the nucleolus by immunolabelling specific ribosome biogenesis factors. For example, upstream binding factor (UBF), an RPI transcription factor, can indicate the FC and DFC regions whereas a late processing factor such as Nop52 can mark the GC (Savino et al. 2001). There can be multiple FCs per nucleolus, even up to two hundred as is the case with human fibroblast cell lines or mouse oocytes depending on their activity (Jordan & McGovern 1981, Mirre & Stahl 1981).

Figure 1.1 Structure of the nucleolus as seen through an electron microscope

Electron micrograph of a thin-sectioned nucleolus from a mouse cell. The fibrillar centers, dense fibrillary components and granular components are indicated with f, d and g respectively. The arrows indicate perinucleolar heterochromatin and the asterisk denotes a DFC clump within the FC. Reprinted with permission from Raska I., Trends in Cell Biology, 2003.

Though there are three visible components of the nucleolus, in fact, its architecture is not stable, but rather depends on a precise regulation of ongoing ribosome biogenesis and the interplay between the multitude of proteins that dynamically exchange in and out of the nucleolus. The nucleolus appears as a distinct body due to the proteins that associate with the rDNA that form a core structure that other nucleolar and non-nucleolar proteins can interact with stably or transiently, which in some cases can even be just for a few tenths of seconds. Furthermore, even the typical nucleolar proteins can migrate out of the nucleolus depending on the cell cycle stage or environmental cues (Pederson 2011). As the proteins that are present in the nucleolus are

so dynamic, the structure of the nucleolus appears to be a consequence of rRNA transcription and processing (Misteli 2001, Misteli & Phair 2000, Raška et al. 2006). For instance, inhibiting RPI transcription with Actinomycin D (ActD), a DNA intercalator with a particular specificity for rDNA due to its G/C content, causes a loss of nucleolar structure (Floutsakou et al. 2013, Sirri et al. 2008). Another study used 5,6-dichloro-1-ß-D- ribofuranosylbenzimidazole (DRB), a nucleoside analog that can indirectly and reversibly inhibit rRNA synthesis. They concluded that it is the production of rRNA transcripts that maintains the localization of rRNA processing factors in the nucleolus and not the rDNA (Louvet et al. 2005).

In addition to the multitude of proteins and the rDNA found in the nucleolus, there are, of course, many copies of the rRNA transcripts. It was previously shown that a small quantity of non-ribosomal RNA was also present in the FC (Thiry 1988). Recently, RPII transcripts from intronic Alu insertion elements were observed at high concentration in the human nucleolus and that they contribute to its structure through interaction with nucleolar protein nucleolin (NCL) (Caudron-Herger et al. 2015). Therefore, it is not only RNA produced from the rDNA that maintains the nucleolar structure but also noncoding RPII transcripts.

On the whole, the nucleolus is a stable entity but with a highly dynamic composition that responds to signaling pathways, drug treatments, cell cycle status, etc. So, to describe its structure as simply the three visible structural components, FC, DFC, and GC, is oversimplifying this complex multifunctional organelle.

1.1.2 Nucleolar organizer regions During mitosis, rRNA transcription arrests and the nucleolus disassembles in mammals and plants but not in yeast. At the end of mitosis, the nucleolus is reformed, remaining stable throughout interphase until the next cell cycle (Boisvert et al. 2007, Németh & Längst 2011, Raška et al. 2006). Nucleolar organizer regions (NORs) are the morphological sites where nucleoli form at the end of mitosis (McClintock 1934). These sites are the chromosomal loci where the rDNA arrays are found and are located on the short arms of acrocentric chromosomes (13, 14, 15, 21 and 22 in human cells), between the heterochromatic centromeric and telomeric DNA. In contrast, mice have all acrocentric chromosomes, but the NORs are still found on particular chromosomes (12, 15, 16, 18, 19) and are close to centromeric and telomeric DNA (Figure 1.2). Furthermore, the number of NORs varies between species, and the rDNA composition of NORs can vary between cells of an individual or between individuals of the same species (Stults et al. 2009, 2008). Notably, NORs are thought to contain only the rDNA and no other DNA sequences (Sakai et al. 1995).

Figure 1.2 The position of NORs on human and mouse chromosomes Diagram of the telocentric mouse chromosomes and the acrocentric human chromosomes indicating the position of the NORs. Modified and reprinted with permission from McStay B. and Grummt I., Annu Rev Cell Dev Biol, 2008.

Electron microscopy studies revealed that NORs appear as gaps on metaphase chromosomes, and are now referred to as secondary constrictions since centromeres are known as primary constrictions (Figure 1.3) (Heitz 1931, McClintock 1934). Heliot and colleagues discovered that these gaps were due to the low compaction of rDNA repeats in the active NOR. In fact, the satellite DNA repeats adjacent to the NORs referred to as perinucleolar heterochromatin, are ten times more condensed than active rDNA (Héliot et al. 1997). Even though NORs are decondensed, they have a tight link with heterochromatin. For example, heterochromatin from non-NOR chromosomes associates with nucleoli throughout the cell cycle. Also, the silent X-chromosome needs to temporarily associate with nucleoli during S phase to maintain its heterochromatic nature (Manuelidis & Borden 1988, Zhang et al. 2007). One reason for the proximity to heterochromatin could be to separate NORs from genes transcribed by RPII and RPIII (McStay 2016). Additionally, heterochromatin could prevent homologous recombination between rDNA repeats, therefore protect against genomic instability (McStay & Grummt 2008). The role of the rDNA and NORs in genomic instability and DNA damage will be discussed further at a later point in this chapter.

Figure 1.3 Mammalian acrocentric chromosome carrying a NOR Diagram of an acrocentric DAPI-stained human chromosome with the secondary constriction or NOR in red on the short arms. Reprinted with permission from McStay B. and Grummt I., Annu Rev Cell Dev Biol, 2008.

Another identifying feature of active NORs during metaphase is that they can be stained with silver nitrate, the AgNOR staining technique, due to the acidic or argyrophilic domains found in the RPI transcription machinery (Goodpasture & Bloom 1975, McClintock 1934, Ploton et al. 1994)). For example, RPI transcription factors UBF, SL1, and TTF1 have been identified as AgNOR proteins on mitotic NORs (Figure 1.4) (Jordan et al. 1996, Leung et al. 2004, Roussel et al. 1993, Sirri et al. 1999). These factors stay associated with the rDNA during mitosis, thereby allowing a rapid restart of transcription at the start of G1. In fact, UBF can induce an open chromatin conformation and therefore, may in part be responsible for inducing secondary constrictions on mitotic chromosomes (Chen et al. 2004). Furthermore, random insertions of large arrays of the 60/81 bp rDNA enhancer repeats from Xenopus known to bind UBF into human chromosomes induce visible secondary constrictions referred to as pseudo-NORs (Mais et al. 2005), and their formation depends on UBF (Grob et al. 2014). Interestingly, these pseudo-NORs, like natural NORs, associate with UBF and many other factors including RPI and even ribosome processing factors throughout the cell cycle and are stained with silver nitrate (Prieto & McStay 2007).

However, not all NORs are actively transcribed and chromosomes carrying such silent NORs do not display an associated secondary constriction, nor AgNOR staining. These are generally believed to be heterochromatic, and their inactivity appears to be inherited through many cell generations (Kurihara et al. 1994, Roussel et al. 1993).

The silver staining of metaphase spreads technique has been used to identify the number of active NORs in human cells. It is estimated that there is an average of eight active NORs and that this can vary from seven to ten (Héliot et al. 2000). This is curious since around 50% of the rDNA is heterochromatically silenced, which would mean that there should be five active NORs per cell (half of the total number of short arms on acrocentric chromosomes). This leads to the hypothesis that NORs must be in fact variable in their rDNA content and have both active and silent repeats (McStay 2016). Indeed, variation in rDNA content of NORs has been studied by digesting chromosomes using restriction enzymes not found in the repeat and separating

these products using pulsed field gel electrophoresis and then hybridizing using rDNA probes (Sakai et al. 1995). This technique has been done by numerous groups using EcoRV and has shown that most human NORs are made up of 70 copies of the rDNA but more recently the Pierce group found differences in NOR size all the way from just one repeat (40kb) to over 130 repeats (6Mb) in different human cell lines (Stults et al. 2009, 2008).

Figure 1.4 UBF is found on active NORs throughout the cell cycle Immuno-FISH labeling of the rDNA (green) and UBF (red) in HeLa cells during metaphase and interphase. The nuclei are stained with DAPI (blue). Reprinted with permission from McStay B. and Grummt I., Annu Rev Cell Dev Biol, 2008.

1.1.3 Functions of the nucleolus Over ten years ago, mass spectrometry studies had purified over 700 proteins from nucleoli of human cells and approximately 90% of human nucleolar proteins have yeast homologues, indicating the high level of conservation of the nucleolar proteome (Andersen et al. 2005, 2002; Scherl et al. 2002). Since 2014, more than 4500 proteins are named in the nucleolar protein database (http://lamondlab.com/NOPdb3.0/) (Quin et al. 2014). It is estimated that only around 30% of these proteins play a role in ribosome biogenesis, which is the most studied function of the nucleolus. However, it is less well known that there are in fact many other cellular processes that involve the nucleolus. These include cell cycle regulation, cellular stress response, control of aging, and the maturation of non-ribosomal RNAs and RNP complexes. One general function of the nucleolus is protein sequestration or release that can affect all the above listed processes (Emmott & Hiscox 2009).

1.1.3.1 Ribosome biogenesis Ribosome biogenesis is the main function of the nucleolus and the most well-studied. This process consumes much of the cell’s energy and is tightly linked to cell growth. In eukaryotic cells, assembly of a mature ribosome involves bringing together ~80 ribosomal proteins (r-proteins) and the four rRNAs (28S, 18S, 5.8S and 5S), which are the ribozymes in the ribosome core that carry out catalyzation of decoding and amino acid polymerization during translation (Lafontaine 2015, Tschochner & Hurt 2003). In fact, the cell utilizes all three RNA polymerases (Pol I, RPII and Pol III) to accomplish the task of generating the cell’s protein factory, the ribosome. RPI transcribes the precursor rRNA (pre-rRNA) which will become the mature 28S, 18S, and 5.8S, RPII produces the messenger RNAs (mRNAs) that encode the r-proteins and RPIII transcribes the 5S rRNA (Figure 1.5). The initial and rate-limiting step of ribosome biogenesis is the transcription of the rDNA to produce the pre-rRNA, named 35S in yeast and 47S in human (Mougey et al. 1993, Tschochner & Hurt 2003). In addition to the components that form the ribosomal subunits, ribosome biogenesis involves numerous other factors such as assembly factors and small nucleolar ribonucleoprotein complexes (snoRNPs) that contain the snoRNAs that target these complexes to the rRNA. These include methyltransferases and endo- or exonucleases that process the pre-rRNA, chaperones, and helicases to facilitate folding of the RNP subunits and ATP/GTPases to help assemble the complexes. In fact, the pre-rRNA undergoes hundreds of post- transcriptional modifications, namely methylation, and pseudouridylation. The early processing steps occur co- transcriptionally at the FC/DFC interface. These processing factors, the rRNAs, and the r-proteins come together to form the pre-90S ribosomal subunit in the GC of the nucleolus, which is then divided into the pre- 40S and pre-60S subunits. These pre-ribosome subunits undergo the last modifications in the nucleoplasm and are then exported to the cytoplasm as the mature 40S and 60S subunits ready to be assembled into the ribosome (Tschochner & Hurt 2003).

Figure 1.5 Ribosome biogenesis involves all three nuclear polymerases

Ribosomal biogenesis starts with transcription of the 47S rRNA precursor by RPI in the nucleolus (in brown). The pre-rRNA undergoes processing in the granular component of the nucleolus (light brown) by snoRNPs and processing factors produced from mRNAs transcribed by RPII. The mature rRNAs along with the 5S rRNA transcribed by RPIII are assembled into the ribosomal subunits and then exported from the nucleus.

1.1.3.2 Cell cycle regulation The nucleolus is a dynamic structure that disassembles during mitosis, which corresponds to a decrease in rDNA transcription, and reassembles at the start of the next cell cycle (Hein et al. 2012, Sirri et al. 2002). In fact, this reassembly and rDNA transcription are prerequisites for G1-S progression. The major role for the nucleolus in cell cycle regulation seems to be in the sequestration and release of factors that directly affect the cell cycle (Boisvert et al. 2007). For example, the tumor suppressor retinoblastoma protein’s (RB) location is temporally regulated throughout the cell cycle. Nucleolin/C23 (NCL) associates with active RB (pRB) during G1 and could have a role in RB’s regulation of the G1-S checkpoint (Angus et al. 2003, Bartek et al. 1996, Grinstein et al. 2006). Hyperphosphorylated RB (ppRB) is released from the nucleolus until late S or G2, at which point nucleophosmin/B23 (NPM) interacts with RB mediating its renewed retention in the nucleolus (Takemura et al. 2002). Though the mechanisms or functional purpose of the retention and release of certain factors is unclear, it does show that the nucleolus has a role in the cellular distribution of factors that are important for cell cycle regulation (Boisvert et al. 2007).

1.1.3.3 Stress response The protein components of the nucleolus are significantly altered under various stress conditions, and the release of nucleolar factors is an important characteristic of the tumor-suppressor protein p53-related stress response (Boulon et al. 2010, Moore et al. 2011, Rubbi & Milner 2003). In normal cells, p53 is continually transcribed but short-lived and found at low levels due to its ubiquitination by the murine E3 ligase murine double minute 2 (MDM2, HDM2 in human cells) marking it for nuclear export and proteasomal degradation (Olson 2004). Upon cellular stress, such as DNA damage, inhibition of ribosome biogenesis or oncogene activation, p53 is stabilized, leading to activation or repression of genes important for cell cycle regulation. Notably, the increased level of p53 activates p21, which in turn inhibits cyclin D1-CDK4 and cyclin E1-CDK2 leading to a G1 arrest (Prives 1998). Subsequently, if the stress or damage is not rectified, this process leads to apoptosis.

Two main types of cellular stress, in particular, induce a response involving ribosome biogenesis factors; oncogenic stress and nucleolar stress. Stress induced by oncogene activation leads to upregulation of the nucleolar factor murine p19ARF (p14ARF in human cells; alternative reading frame), which can sequester its binding partner MDM2 in the nucleolus. This sequestration favors the accumulation of p53 in the nucleus and subsequent cell cycle arrest (Weber et al. 2000, Wsierska-Gadek & Horky 2003). However, ARF levels are variable during the cell cycle and it is not found in all cell types, nor is it conserved in all species, therefore, other factors likely can replace ARF as MDM2 binding partners (David-Pfeuty & Nouvian-Dooghe 2002, Olson 2004, Zhang & Xiong 2001). Other nucleolar proteins that may have a role in stress-induced p53 stabilization include NPM and NCL, which both have been shown to interact directly with p53 (Colombo et al. 2002, Daniely & Borowiec 2000).

In the case of nucleolar stress, which is caused by defects in ribosome biogenesis, it appears that p53 is activated by r-proteins (Figure 1.6). Multiple r-proteins (RPS3, RPS5, RPS7, RPL5, RPL11, RPL23, RPL26), as well as the 5.8S and 5S rRNAs, have been shown to interact with MDM2 or p53 (Deisenroth & Zhang 2010, Fontoura et al. 1992, Riley & Maher 2007, Zhang & Lu 2009). Upon inhibition of ribosome biogenesis, a pool of uncomplexed ribosomal proteins could be available for interaction with other non-ribosomal partners. For example, in growing cells, normally L11 would be assembled into ribosomes and no longer present in the nucleolus. However, if ribosome biogenesis was inhibited, L11 would be in excess and free to bind MDM2 and inhibit the export of p53 (Zhang et al. 2003). Therefore, depending on the cellular stress, oncogenic or nucleolar, either ARF or r-proteins can stabilize p53 and induce a cell cycle arrest.

Figure 1.6 Nuceolar stress (A) In a normal cell, p53 levels are maintained by the interaction with MDM2 which ubiquitinates p53 targeting it for export from the nucleolus and proteosomal degradation. (B) Upon nucleolar stress, uncomplexed ribosomal proteins are released from the nucleolus and bind to MDM2, allowing the accumulation of p53 in the nucleus. Reprinted with permission from Quin J. et al., BBA – Molecular Basis of Disease, 2014.

1.1.3.4 Control of aging

Aging is the time-dependent decline of cellular functioning and is characterized by certain hallmarks including cellular senescence, stem cell exhaustion, genomic instability, telomere attrition, epigenetic alterations, and deregulated nutrient sensing (López-Otín et al. 2013). The nucleolus and ribosomal DNA are involved in many of these hallmarks and appear to be an important aspect of cellular aging, which has been demonstrated mostly through studies in yeast.

The main body of work involving the nucleolus and aging involves the sirtuins, a family of NAD+-dependent histone deacetylases that were linked to prolonging life span in yeast due to their role as silencers of the rDNA (Hein et al. 2012, Kennedy et al. 1995, 1997). Ribosomal gene arrays are very unstable due to their repetitive nature, which can lead to the loss of copies due to homologous recombination (HR). They are the most

unstable regions of the yeast genome (Kobayashi 2008) but there is a gene amplification system in place to regain lost copies. Fob1p blocks replication forks by binding to a replication fork barrier (RFB) downstream of the 35S coding sequence, leading to stalled forks and double strand breaks (DSBs) (Brewer et al. 1992, Burkhalter & Sogo 2004, Kobayashi et al. 1992, 2004; Weitao et al. 2003). Amplification occurs when the breaks are repaired by unequal sister chromatid exchange, due to transcripts originating from the promoter E- pro in the IGS. This transcription causes the displacement of the cohesin complex holding the sister chromatids in place and therefore promotes slippage and unequal crossovers (Kobayashi & Ganley 2005). A sirtuin family member and histone deacetylase, Sir2p, represses transcription from E-pro once amplification has restored any rDNA copy loss (Fritze et al. 1997, Saka et al. 2013). The deletion of Sir2p leads to hyper- recombination of the rDNA and subsequently a shorter lifespan (Kaeberlein et al. 1999). It is this balance of gene amplification and deleterious recombination regulated by Fob1p and Sir2p respectively that maintains the appropriate number of rRNA genes and their stability. These concepts constitute the ‘rDNA theory of aging’ where rDNA locus instability translates to genome-wide instability (Kobayashi 2008). This theory also describes the timing of senescence as being determined by the rDNA. As the rDNA forms the most unstable loci in the genome, the cell can likely discern this increasing instability throughout cell divisions or replicative stress and therefore, sense when the limit of divisions is being reached. Cells will enter senescence rather than continue to accumulate damage and risk becoming transformed (Kobayashi 2011a, 2014; MacInnes 2016).

Figure 1.7 Model for eNoSC function in the nucleolus A model depicting the proposed function of the energy-dependent nucleolar silencing complex (eNoSC). SIRT1, NML, and SUV39H1 form a complex on the rDNA and modify H3K9 methylation, which could influence silencing and heterochromatin formation. This complex is dependent on the energy status of the cell, as depicted by the high or low glucose situations. Reprinted with permission from Murayama A et al., Cell, 2008.

Though ribosomal DNA stability has been linked to aging and senescence in yeast, studies in mammals have just recently begun to show the link between ribosomal biogenesis and aging. Sirtuin 1 (SIRT1) is part of the eNoSC, or energy-dependent nucleolar silencing complex, that potentially epigenetically silences the rDNA based on the energy status of the cell (Figure 1.7) (Murayama et al. 2008). Interestingly, overexpression of SIRT1 in mouse brains prolongs life expectancy (Satoh et al. 2013). The nucleolus has been linked to senescence in mammalian cells as changes in nucleolar morphology are observed in senescent cells, which only have one large nucleolus as opposed to several smaller ones (Mehta et al. 2007). This typical change in nucleolar morphology in response to nucleolar stress often leads to cell cycle arrest. Senescent cells arrest at the G1-S checkpoint which is regulated in part by p53. As described in the previous section, some nucleolar factors, such as p19ARF mediate p53 stability and are therefore also involved in regulation of the senescent phenotype (Hein et al. 2012). The 5S ribonucleoprotein particle (RNP) which consists of RPL11, RPL5, and the 5S rRNA was recently shown to induce senescence upon oncogenic or replicative stress through p53 activation (Nishimura et al. 2015).

1.1.3.5 Other RNA maturation The nucleolus has also been implicated in the processing and maturation of RNA species other than the rRNA precursor. This includes the assembly of RNP complexes such as telomerase, spliceosomal small nuclear RNPs (snRNPs) and RNA modifications on tRNAs, RNAse P RNA, signal recognition particle (SRP) RNA and more recently some microRNAs (miRNAs) (Boisvert et al. 2007). The SRP complex, composed of six proteins and an RNA, are found in the nucleolus prior to their export to the cytoplasm which indicates a potential function for the nucleolus in the maturation of this complex (Jacobson & Pederson 1998). Other examples are RNA species transcribed by RPIII in the nucleus, which are then imported into the nucleolus, potentially for modification or maturation, similarly to the 5S rRNA. These include tRNA, RNase P RNA and the U6 spliceosomal snRNA (Ganot et al. 1999, Jacobson et al. 1997).

1.2 Ribosomal DNA In a cell, the r-proteins and the rRNA transcripts are equimolar and their amounts are tightly regulated to fill the cell’s requirement for ribosomes, but the number of rDNA copies that are transcribed to produce the rRNA is very different. The number of rDNA repeats differs between organisms, from less than 100 to more than 10,000 repeats (McStay & Grummt 2008). In Saccharomyces cerevisiae (S. cerevisiae) there are ~200 copies of a 9.1 kb repeat located on chromosome XII (Petes 1979) whereas Arabidopsis thaliana (A. thaliana) has 700-800 copies of a ~10 kb repeat found close to telomeres on chromosomes 2 and 4 (Copenhaver & Pikaard 1996). Drosophila melanogaster (D. melanogaster) has two NORs, one in the centromeric heterochromatin of the X chromosome and the other on the short arm of the heterochromatic Y chromosome. Each NOR contains

about 200 copies of an 11-17kb repeat, but copy number can vary greatly (Williams & Robbins 1992). Humans and mice each have around 200 rDNA repeats on the short arms of five acrocentric chromosomes (Britton- Davidian et al. 2012, Henderson et al. 1972), though in each genome, segments of rDNA are found inserted on other chromosomes and probably represent pseudogenes.

1.2.1 Organization of the rDNA Each mammalian rDNA repeat is quite large, spanning ~43 kb in humans and ~45 kb in mice and these are arranged in a head-to-tail manner (Gonzalez & Sylvester 1995, Grozdanov et al. 2003). The rDNA is the region in the cell that is the most actively transcribed and the pre-rRNA is assembled co-transcriptionally, meaning that the processing machinery and the transcription machinery are working closely together. This mechanism can be observed by electron microscopy using the Miller chromatin spreading technique (Miller spread) which produces "Christmas tree" structures (Moss et al. 2007). The rDNA forms the trunk of the tree and the multiple rRNA transcripts are the branches. The round, dark spots at the end of the nascent pre-rRNA transcripts are referred to as terminal knobs, which are thought to be made up of the SSU processome (Figure 1.8) (Mougey et al. 1993).

Figure 1.8 Ribosomal RNA is co-transcriptionally assembled Electron micrograph depicting a ‘Miller’ spread or Christmas tree from a mouse Ltk- cell. The rDNA is being transcribed by many closely packed polymerases. The inset shows the magnified rDNA, many rRNAs and the 3’ terminal knobs at the extremities. Reprinted with permission from Moss T., Cell Mol Life Sci, 2007.

The mouse and human rDNA repeat is composed of a long intergenic spacer (IGS) of ~30 kb and the coding sequences for the pre-rRNA (13-14 kb). The pre-rRNA or 47S is made up of the external transcribed sequences (ETS), internal transcribed sequences (ITS) and the rRNAs 18S, 5.8S, and 28S (Figure 1.9). In most species, the fourth rRNA, 5S, and the tRNAs are transcribed by RPIII at the nucleolar periphery, except

in yeast where the 5S sequence is linked to the rDNA and therefore transcribed in the nucleolus (Moss et al. 2007).

Figure 1.9 Ribosomal DNA organization in mammals A representation of the mammalian rDNA repeat (not to scale) is pictured with the transcribed region highlighted in yellow and the intergenic spacer in grey. The regulatory elements are indicated; terminator elements (T), spacer promoter (SP), promoter-proximal terminator (T0), promoter (P) formed of the upstream control element (UCE) and core promoter (Core), and the transcription start site (+1). Reprinted with permission from Russell J., Zomerdijk JCBM., Trends in Biochemical Sciences, 2005.

The IGS houses the regulatory elements such as promoters, spacer promoters, enhancer repeats, and terminators. The promoter is made up of two elements, the core promoter at the transcription start site (TSS) and the upstream control element (UCE), which lies ~100 nucleotides upstream (Haltiner et al. 1986, Learned et al. 1986). There are multiple terminator elements found at the 3’ end of the rDNA repeat (T1-T10) as well as one terminator (T0) 170 bp upstream of the promoter UCE in mouse (Figure 1.9) (Grummt et al. 1985, 1986a; Henderson & Sollner-Webb 1986, McStay & Reeder 1986, Moss 1983). Early studies in Xenopus demonstrated that the same sequence just upstream of the rRNA coding sequence was found repeated in the IGS (Moss 1983, Moss & Birnstiel 1979). This sequence was found to be the RPI promoter and its upstream repeats were therefore referred to as ‘spacer promoters’. These spacer promoters (SpPr) stimulate efficient pre-rRNA production and act in conjunction with adjacent arrays of short repeated enhancer sequences (Caudy & Pikaard 2002, De Winter & Moss 1986, 1987; Dunaway & Dröge 1989, Moss 1983, Osheim et al.

1996). Several functions of the T0 promoter-proximal terminator and IGS transcripts were originally suggested, mostly based on work done in Xenopus. These include a process called ‘readthrough enhancement’, by which polymerases recruited by the spacer promoter and transcribing the IGS are arrested at T0 and then handed over to the pre-rRNA promoter (Grimaldi et al. 1990, Längst et al. 1998, Moss 1983). If the T0 or the SpPr are mutated, termination of the IGS transcripts is impaired, and the amount of pre-rRNA transcripts produced from

the adjacent gene is reduced. Therefore, transcription from the IGS is clearly important for regulating the neighboring gene’s production (Grummt et al. 1985, 1986a; Henderson & Sollner-Webb 1986, McStay & Reeder 1986, Moss 1983). More recently, this region has been implicated in regulating the active chromatin state and gene silencing (Mayer et al. 2006, Santoro et al. 2009). The role of the T0 terminator and the transcripts originating from the SpPr in gene silencing will be discussed in more detail in another section of this chapter. The remainder of the IGS possesses many intermediate repetitive and transposable elements, and may potentially contain further regulatory elements, but to date, these appear to be of minor importance (McStay & Grummt 2008).

Due to the large number of rRNA genes and the repetitive nature of these sequences, there is not currently complete sequencing data available for the entire region. The coding region of the rDNA was sequenced and is available in GenBank for humans (Acc. No. U13369) and for mice (BK000964), however, there are many possible small variations in this sequence (Grozdanov et al. 2003, Kuo et al. 1996). The coding sequences for 18S, 5.8S and 28S are very highly conserved between species, but promoter and other rDNA sequences (ITS, ETS, IGS) are not. This is one reason why RPI machineries are incompatible between most animal species. For example, mouse and human systems are functionally incompatible (Moss et al. 1985, 2007).

1.2.2 Epigenetic regulation of the rDNA The epigenetic regulation of genes is manifested by the addition of epigenetic marks either on the DNA itself such as methylation or on the histones that make up the chromatin such as acetylation, methylation, ubiquitination, sumoylation, and phosphorylation. These modifications are performed by various enzymes such as histone acetyltransferases and methyltransferases called writers, are removed by erasers (e.g. histone deacetylases) and bound by readers (e.g. bromodomains, chromodomains) (Figure 1.10). Over the years, specific modifications have become associated either with active transcription or gene silencing however it is the complex pattern of these marks, the histone code, that can give an indication of gene activity (Wang et al. 2007). The rDNA’s transcriptional status is also demonstrated by epigenetic marks. Silenced genes are methylated and the affiliated histones have repressive marks, whereas the active genes lack DNA methylation and are bound by RPI and RPI-specific transcription factors (Németh & Längst 2011).

1.2.2.1 rDNA activity status and chromatin Growing cells actually have three populations of rDNA; (1) genes that are actively being transcribed and therefore are euchromatic and coupled with pre-rRNA molecules and the RPI machinery, (2) inactive genes that are in a potentially active state and still euchromatic and (3) silent genes that are methylated and are heterochromatic. The intergenic spacer of active, inactive or silent rRNA genes is nucleosomal but the presence of histones in the coding region of the active rDNA repeats is still debated. Some studies in yeast

have demonstrated that actively transcribed rDNA repeats are devoid of histones, while others have suggested the presence of histones or non-canoncial nucleosomes that are likely dynamic (Conconi et al. 1989, Dammann et al. 1993, Jones et al. 2007, Tongaonkar et al. 2005). One chromatin immunoprecipitation (ChIP) analysis performed in mouse cells demonstrated active histone marks at the promoter but none in the gene body, however, they did observe some repressive marks in the gene body (Zentner et al. 2011). Whether, histones are present or not, it is clear that active rDNA is open and non-nucleosomal in the classical sense and this is likely due to the presence of UBF, an RPI transcription factor that binds throughout the active rDNA.

Figure 1.10 The epigenetic readers, writers and erasers Epigenetic marks are specific modifications on amino acid residues of histone tails that are placed by writers such as histone acetyltransferases (HATs), histone methyltransferases (HMTs), protein arginine methyltransferases (PRMTs) and kinases. Readers containing bromodomains, chromodomains and Tudor domains bind these modifications. Erasers such as histone deacetylases (HDACs), lysine demethylases (KDMs) and phosphatases remove these epigenetic marks. Reprinted with permission from Falkenberg KJ., and Johnstone RW., Nat Rev Drug Discov, 2014.

If the active rDNA does contain dynamic nucleosomes in the gene body or in the promoter, chromatin remodelers would be necessary in order to make the gene accessible for the rapid transcription by RPI. These remodelers would also be necessary for the transition from inactive to active genes. One chromatin remodeler that is involved in rDNA regulation is Cockayne syndrome protein B (CSB), an ATPase that uses ATP

hydrolysis to break protein-DNA interactions (Beerens et al. 2005, Citterio et al. 2000). CSB is found in the nucleolus on active rRNA genes and it interacts with RPI machinery components (Bradsher et al. 2002). When CSB was overexpressed in CS1AN (CSB-defective human fibroblast) cells, rDNA transcription was activated but knockdown of CSB using siRNA in NIH3T3 cells impeded pre-initiation complex formation, hence inhibiting transcription. Upregulation of RPI transcription by CSB appears to require TTF1 binding to the T0 terminator just upstream of the promoter therefore, the authors suggest that TTF1 could target CSB to the rDNA, thereby helping RPI transcribe through chromatin due to its remodeling activity (Yuan et al. 2007).

Due to the open conformation of the active rDNA chromatin, it is possible to distinguish the proportion of active and inactive rDNA genes using the DNA intercalator psoralen. Psoralen intercalates in double-stranded DNA (dsDNA) and produces covalent crosslinks when exposed to UV irradiation. Active or potentially active genes that are euchromatic are accessible to psoralen crosslinking whereas silent, heterochromatic genes are not crosslinked. The active, crosslinked DNA migrates more slowly on an agarose gel and therefore one may identify the proportions of nucleosomal/silenced and non-nucleosomal/active rDNA (McStay & Grummt 2008). Somatic mouse cell lines have typically shown a 50:50 ratio of psoralen crosslinked or non-crosslinked rDNA (Figure 1.11). This proportion seems to be consistent from cell generation to cell generation (Conconi et al. 1989). As previously mentioned, over 50% of NORs are active in many mammalian cell lines tested. If only 50% of the rRNA genes are active, then this must mean that NORs contain a mixed population of active and silent genes. It is also possible that active rDNA from different NORs may associate with each other forming some sort of nucleolar ultrastructure, which would lead to a higher number of visibly active NORs (McStay & Grummt 2008).

Figure 1.11 Psoralen crosslinking of active and inactive rDNA (A) A diagram describing psoralen photocrosslinking technique. Cells are exposed to psoralen and irradiated with UVA. The DNA is then isolated, digested with restriction enzymes and evaluated by gel electrophoresis or electron microscopy. (B) Purified rDNA from psoralen crosslinked cells was separated by gel electrophoresis and stained with ethidium bromide (EtBr). Each band (a-active and i-inactive) was eluted and then spread for electron microscopy in denaturing conditions. The upper panel depicting the active band shows a heavily crosslinked double stranded DNA whereas the lower panel shows spaced single stranded DNA with bubbles typical of nucleosomal chromatin. (C) Yeast DNA was crosslinked, digested, separated by gel electrophoresis and blotted then hybridized with a rDNA-specific probe. Active rDNA is more open and thus incorporates more psoralen and migrates slower on a gel. Control is non- crosslinked DNA. (D) Transcription run-on (TRO) was performed in permeabilized yeast cells using radioactive RNA precursors. As psoralen can crosslink DNA-RNA hybrids, rDNA being actively transcribed is labelled. The left lane shows the rDNA probed as in (C) and the right lane was the dried gel exposed to film showing that rRNA transcripts are only crosslinked to one rDNA fraction, the active rDNA. Reprinted with permission from Toussaint M. et al., Biochem Cell Biol, 2005.

1.2.2.2 Methylation of the rDNA Methylation of cytosine residues adjacent to a guanosine in a CpG dinucleotide is a characteristic mark of silencing. Repression of genes by methylation is necessary for X-chromosome inactivation, imprinting, and transposon silencing. Methylation is inheritible in these mechanisms and therefore conserved throughout cell divisions however, methylation can also be a reversible mechanism for gene regulation. DNA methylation is executed by the DNA methyltransferases DNMT1, DNMT3a and DNMT3b (Klose & Bird 2006, Meehan et al.

2001). In mammals, multiple CpGs are commonly found concentrated around the promoters of genes, and referred to as CpG islands. They are usually maintained in an unmethylated state.

The methylation status of rat or mouse rDNA was studied using methylation-sensitive and methylation- insensitive restriction enzymes HpaII and MspI and a link was made between methylated rDNA and the inactive genes (Bird et al. 1981, Santoro & Grummt 2001). In vitro transcription experiments and transfection experiments in cells using mutant rDNA templates demonstrated that methylation of the CpG at position -133 in the UCE is enough to inhibit binding of UBF to the rDNA and therefore it remains nucleosomal, not allowing for assembly of the pre-initiation complex (PIC) (Santoro & Grummt 2001). In human cell lines, data suggest that a significant fraction of rDNA completely resists promoter methylation (Brown & Szyf 2008, Gagnon-Kugler et al. 2009, McGowan et al. 2008). When human DNMT1 and 3b genes were inactivated in HCT116 cells, or cells were treated with aza-dC to inhibit or reverse CpG methylation, a large portion of silent rDNA repeats were reactivated. This however did not lead to increased rRNA transcription but instead led to rRNA processing defects and increased genomic instability (Gagnon-Kugler et al. 2009).

Another factor that regulates the methylation landscape of the rDNA is Gadd45a, a demethylase that has been implicated in DNA repair, cell cycle, apoptosis and genomic stability (Barreto et al. 2007). It was found that Gadd45a is associated with the rDNA promoter and maintains a hypomethylated state on actively transcribing repeats. In fact, increased Gadd45a recruitment to the rDNA occurs in response to DNA damage, but the loss of Gadd45a results in an increased level of methylation. This mechanism was shown to involve the nucleotide excision repair (NER) pathway and the authors conclude that NER proteins and Gadd45a help to keep a portion of the rDNA repeats in an active state by regulating promoter methylation (Schmitz et al. 2009).

1.2.2.3 Mechanism of rDNA silencing Silencing and maintenance of heterochromatin at the rDNA is mediated by the nucleolar remodeling (NoRC) complex, which is composed of the ATPase SNF2h and the large scaffolding protein TIP5 (TTF1 interacting protein 5) (Strohner et al. 2001). This complex induces nucleosome sliding by recruiting enzymes such as DNA methyltransferases (DNMTs) and the Sin3 corepressor complex containing histone deacetylases (HDACs) and histone methyltransferases (HMTs) (Santoro & Grummt 2005, Santoro et al. 2002, Zhou et al. 2002). In this way NoRC induces the silencing of a portion of the rDNA repeats and this heterochromatic state is maintained throughout cell divisions. Santoro et al. have described a temporal model for rDNA silencing that is dependant on TTF1 binding to the rDNA regulatory regions (Santoro & Grummt 2005). Firstly, TTF1 bound to the promoter-proximal terminator T0 recruits TIP5 and consequently its partner SNF2h. The NoRC complex then recruits the Sin3 corepressor complex that remodels chromatin in the promoter region of the rDNA

(Figure 1.12). This complex is made up of HDAC1/HDAC2 that deacetylate histone H3 and H4 tails, DNMT1/DNMT3 that methylate CpGs on the rDNA and HMTs that methylate H3K9, H3K20 and H3K27.

Figure 1.12 Model of the mechanism of rDNA silencing by NoRC

(A) TTF1, bound at T0 recruits NoRC to the rDNA promoter. (B) NoRC then interacts with Sin3 corepressor complex and with histone methyltransferases. This leads to the deacetylation of histones H3 and H4 and methylation of H3K9, H3K20 and H3K27. (C) These modifications could signal the ATPase SNF2h to reposition the promoter-bound nucleosome downstream, thus blocking PIC formation. (D) This remodeling could also expose the CpG -133 allowing for methylation and blocking of UBF binding. Reprinted with permission from McStay B., Grummt I., Annu Rev Cell Dev Biol, 2008.

There is a promoter-proximal nucleosome found covering position -157 to the transcription start site (TSS) on potentially active genes. However, this nucleosome is found between positions -132 to +22 on silent rDNA copies (Li et al. 2006). The recruitment of the Sin3 complex and the resulting histone modifications could be a marker for the ATPase SNF2h to shift the nucleosome at the promoter 25 nucleotides (nt) downstream, thereby blocking binding of transcription factors and formation of the PIC. This could also expose the CpG at - 133, allowing for its methylation. As previously mentioned, when this position is methylated, UBF is not able to bind the rDNA and therefore not able to maintain an open rDNA conformation thereby allowing transcription (Li et al. 2006). The new nucleosomal landscape could also potentially signal other proteins, such as HP1, that could help spread heterochromatic marks across the gene (McStay & Grummt 2008). However, the full mechanism of silencing is likely more complicated and involves other factors to complete the transition from active to silent chromatin states.

Noncoding RNA (ncRNA) has been shown to play a role in many cell processes including gene expression, epigenetic regulation, and differentiation. MicroRNAs have been broadly implicated in the regulation of gene expression in mammals but other noncoding transcripts have more recently also been shown to play an important role in regulating gene expression (Rinn & Chang 2012). A ncRNA originating from the spacer

promoter 2 kb upstream of the TSS has been shown to regulate the heterochromatic state of the rDNA in mouse and play a role in nucleolar organization (Mayer et al. 2006, Moss et al. 1980, Savić et al. 2014). These long transcripts (IGS-rRNAs) are not stable in cells (Kuhn & Grummt 1992, Morgan et al. 1983, Paalman et al. 1995) and are likely processed into what is called the pRNA for promoter-associated RNA. The pRNAs are 150-300 nt long, processed RNAs that have a corresponding sequence to the rDNA promoter (Figure 1.13). They are stabilized through an interaction with TIP5 and are important for NoRC-mediated silencing as pRNA depletion using antisense oligonucleotides decreased rDNA methylation and activated rRNA production (Mayer et al. 2006). It seems as though the pRNA functions as a scaffold or physical regulator of NoRC recruitment to the rDNA as it forms a specific stem-loop secondary structure recognized by TIP5. When this structure was mutated, NoRC was no longer localized to the nucleolus and rDNA silencing did not occur (Mayer et al. 2008).

TTF1 binding of the T0 upstream of the promoter is clearly an important factor for the regulation of rDNA activity status as TTF1 interacts both with euchromatin-inducing factors (CSB) and the NoRC silencing complex. TTF1 therefore plays a central role in establishing the active or silent state of the rDNA (Figure 1.14). As the next section discusses, the production of pre-rRNA and the proportion of active to inactive rDNA repeats is essential for genomic stability and therefore TTF1 plays a role in maintaining healthy cell growth and avoiding cellular transformation (McStay & Grummt 2008).

Figure 1.13 Transcription of a lncRNA from the intergenic spacer

Transcripts from the intergenic spacer (IGS) are produced starting from the spacer promoter ~2 kb upstream of the 45S promoter. This long transcript is degraded or processed to give the pRNA, which is 150-300 nt long and matches the rDNA promoter. The mechanism of this processing is unknown. The pRNA binds TIP5, which is required for NoRC-mediated heterochromatin formation. Reprinted with permission from McStay B., Grummt I., Annu Rev Cell Dev Biol, 2008.

Figure 1.14 TTF1 as a central regulator of rDNA activity status

TTF1 binds the T0 upstream of the 47S promoter and recruits either Cockayne syndrome B protein (CSB with its partner G9a or TIP5 with its partner SNF2h (NoRC). It is suggested that the balance of CSB and NoRC on the rDNA regulates the proportion of active, euchromtic rRNA genes (blue) to heterochromatic and silenced rRNA genes (orange). The mechanism is yet unknown but the epigenetic state could be propagated throughout the rDNA copies. Reprinted with permission from McStay B., Grummt I., Annu Rev Cell Dev Biol, 2008.

1.2.3 DNA damage and genomic instability of the rDNA Cellular DNA is under constant stress from both endogenous (replication) and exogenous agents (radiation, reactive oxygen species) resulting in DNA damage. There is a wide variety of DNA damage types including mismatches, crosslinking, formation of abasic sites, or breaks. Each type of damage is repaired by an associated mechanism such as mismatch repair or base excision repair. There are two main types of DNA breaks; single stranded breaks (SSB) and double stranded breaks (DSB). The repair of double strand breaks is fundamental for maintaining genomic stability, as the accumulation of DNA damage can lead to apoptosis or cellular transformation (Dexheimer 2013).

Figure 1.15 DNA damage response pathway The DNA damage response (DDR) pathway starts with two principle damage sensors; the MRE11-RAD50- NBS1 (MRN) complex and replication protein A (RPA) and the RAD9-RAD1-HUS1 (9-1-1) complex. MRN senses double-strand breaks (DSBs) and activates apical kinase ataxia-telangiectasia mutated (ATM) whereas the 9-1-1 complex and RPA recruits ataxia-telangiectasia and Rad3-related (ATR) upon sensing single-stranded DNA. ATM and ATR phosphorylate various DNA damage signaling factors as well as CHK2 and CHK1 downstream kinases. These activate effectors such as p53 and the cell division cycle 25 (CDC25) phosphatases. The possible outcomes after DDR pathway activation are indicated at the bottom of the diagram. Reprinted with permission from Sulli G et al., Nat Reviews Cancer, 2012.

Two main mechanisms are used by the cell to repair DSBs, non-homologous end joining (NHEJ) and homologous recombination (HR). NHEJ is the process by which the DSB is repaired by excising a portion of the DNA and then joining both ends. HR uses the other strand of the DNA as a template to repair the lesion. Among the important proteins involved in this complex are MRN, Rad51, BRCA2 (Sancar et al. 2004). The DNA damage response involves a cascade of protein kinase activation and can be classified in three main steps (Figure 1.15). First, the detection of damage by sensors which in turn activate transducers of the signal. The transducers can then activate different mechanisms depending on the severity of the damage, from DNA repair to cell cycle arrest or apoptosis if the damage is too severe . The two main transducers are the serine/threonine kinases ATM and ATR. Both proteins have crucial non-redundant functions in the DNA damage response (DDR), but ATM will be introduced further due to its role in rDNA regulation. ATM is a key component of the DSB repair pathway. ATM is recruited to the site of DNA damage and becomes active. Once activated ATM phosphorylates a wide number of substrates such as Chk2, p53 and the histone variant H2AX, a commom marker of DSBs. The activation of the different substrates by ATM will play a major role in the cellular response to DNA damage (Maréchal & Zou 2013, Sulli et al. 2012).

A few studies have investigated the correlation between DNA damage, mainly DSBs, and RPI transcription. Global irradiation of the cell or nucleoli specific irradiation to introduce DSBs led to ATM mediated RPI transcription inhibition. This inhibition was specific to damaged nucleoli (Kruhlak et al. 2007). More recently, it has been demonstrated that laser-induced microirradiation in the nucleus led to global RPI transcription inhibition (Ciccia et al. 2014, Larsen et al. 2014). As these studies are not able to demonstrate the amount, nor the distribution of the DNA damage, recent studies have specifically induced DSBs in the rDNA using CRISPR/Cas9 or I-PpoI. I-PpoI is an endonuclease that was previously shown to induce DSBs in rDNA (Berkovich et al. 2007). Multiple studies have shown inhibition of RPI transcription upon rDNA DSBs and that this is ATM-dependent (Harding et al. 2015, van Sluis & McStay 2015, Warmerdam et al. 2016). The nucleolus was restructured upon RPI silencing and rDNA moved to nucleolar caps. Conflicting studies show HR (van Sluis & McStay 2015, Warmerdam et al. 2016) as well as NHEJ (Harding et al. 2015) as the repair methods used for these DSBs. Since treating cells with ActD results in a similar nucleolar reorganization, McStay’s group proposes that it is the inhibition of RPI transcription by ATM activation that causes the reorganization and not the DSBs themselves (van Sluis & McStay 2015). This is yet again an example of how rRNA transcription is essential for the chromosomal context of the rDNA and likely contributes to genomic stability.

1.3 Ribosomal RNA transcription As stated previously, transcription of the rDNA by RNA polymerase I (RPI) to produce the pre-rRNA occurs at the interface between the fibrillar centre and dense fibrillar component of the nucleolus. The product of transcription, the pre-rRNA is very large, measuring 6.9 kb in yeast named 35S and ~13 kb in mammals named 47S (Gonzalez & Sylvester 1995, Grozdanov et al. 2003, Petes 1979). This pre-rRNA is modified and processed to become the mature 18S, 5.8S and 28S rRNAs, forming two-thirds of the functional ribosome. Efficient transcription requires a set of Pol I-specific factors that regulate recruitment of the polymerase, initiation, promoter escape, elongation, termination and re-initiation (Moss et al. 2007). Regulation of transcription by activation or modulation of the key players involved in this important process is a key concept of this thesis. Each of these proteins will be discussed in regards to the various steps of transcription and then in regards to the pathways that modulate rRNA synthesis.

1.3.1 Formation of the pre-initiation complex Similar to the other RNA polymerases, RPI has a dedicated set of basal factors necessary for carrying out transcription. As RNA polymerases do not have high affinity for specific promoter sequences, a pre-initiation complex (PIC) is formed of transcription factors on the promoter elements to facilitate polymerase recruitment to the gene. In humans and mouse, the PIC is formed by the cooperative interaction of upstream binding factor (UBF) and the multi-component selectivity complex SL1/TIF-IB, which contains the TATA binding protein

(TBP) and TBP-associated factors (TAFs). RRN3/TIF-IA, an initiation factor, is believed then to be required for an active RPI to recognize the pre-initiation complex and be recruited to the promoter (Goodfellow & Zomerdijk 2012, Moss et al. 2007). Yeast however has two complexes that form on the promoter sequences; upstream activating actor (UAF) and the core factor (CF). The UAF is made up of TBP, Rrn5p, Rrn9p, Rrn10p, UAF30, and histones H3 and H4 and the CF is formed of TBP, Rrn6p, Rrn7p, and Rrn11p (Figure 1.16). The core factor is likely related to SL1, however the UAF seems to have no mammalian counterpart. There is still some debate and many questions yet unanswered about the order and timing of PIC formation, as well as the precise role of some of these factors in initiation.

Figure 1.16 Pre-initiation complex assembly in yeast and mammals (A) The pre-initiation complex in yeast in portrayed. Upstream activating factor (UAF) (in green) binds the upstream element (UE). The TBP-containing core factor (CF) binds the core promoter (Core). Rrn3p binds A43 subunit of RPI/PolI and recruits the polymerase to the promoter. (B) UBF (in purple) binds the upstream control element (UCE) and the core 47S promoter (Core). SL1 formed of TBP and the TAFs is recruited to the promoters as well. RRN3 then recruits the polymerase to the rDNA. Multiple interactions between these factors are indicated with arrows. Reprinted with permission from Moss T., Stefanovsky V., Cell, 2002.

1.3.1.1 UBF UBF is an abundant nucleolar protein that is found on active NORs during metaphase (Roussel et al. 1993, Wright et al. 2006). UBF has been shown to be involved in the regulation of RPI transcription in many ways; transcription activation, inhibition of transcription repression, elongation rates, and chromatin formation (McStay 2016). This being said, exactly how UBF is specifically involved in initiation in vivo has been long debated. In vitro transcription assays with the human PIC factors showed that UBF first binds the rDNA promoter and that this is important for SL1 recruitment (Bell et al. 1988, 1990, 1989). However, in mice, SL1 is capable of binding the promoter without UBF. In contrast DNA footprinting has also shown a bigger complex of both factors on the promoter so they likely bind together (Bell et al. 1990, Kuhn & Grummt 1992, Schnapp & Grummt 1991). Multiple studies have demonstrated a role for UBF in stimulating rRNA transcription by forming stable pre-initiation complexes (Bell et al. 1988, McStay et al. 1991b, Schnapp & Grummt 1991, Smith et al. 1990). Finally, another study using rat factors showed that UBF was not required for transcription at all (Smith et al. 1993).

Furthermore, until recently, whether UBF was even required in vivo was unknown as significant knockdown by siRNA of this factor did not affect transcription or cell viability (Sanij et al. 2008). These authors did show however that UBF levels were important for determining the proportion of active rRNA genes. We now know that UBF is an essential factor for development as UBF-null embyros arrest pre-implantation. This study revealed that UBF is essential in vivo for rRNA transcription and maintaining nucleolar integrity (Hamdane et al. 2014).

One way that UBF is likely involved in rRNA transcription is through facilitating PIC formation as it binds to the upstream and core promoter regions, creating a suitable chromatin environment for SL1 and eventual polymerase binding. UBF is part of a class of proteins containing high mobility group (HMG) boxes, a motif known to bend DNA. UBF has six of these HMG boxes and a C-terminal acidic tail (Jantzen et al. 1990, McStay et al. 1991a). These boxes allow UBF to loop nucleosome-free rDNA into a single turn every ~140 bp, forming a nucleosome-like structure named the enhanceosome (Figure 1.17). UBF can bind the rDNA core promoter and the upstream control element as a dimer and it was proposed that by creating in-phase turns it would bring the two promoter elements closer together and therefore present binding sites for SL1 on the same exposed surface (Bazett-Jones et al. 1994, Putnam et al. 1994, Stefanovsky et al. 2001a).

Figure 1.17 The enhancesome model of UBF binding (A) The domain structure of mammalian UBF. The six HMG boxes are indicated as well as the N-terminal dimerization domain and the C-terminal acidic domain in yellow. Two ERK phosphorylation sites (denoted by an asterix) are shown in HMG boxes 1 and 2. (B) Diagram depicting the structure of the enhancesome. The left image depicts a single enhancesome as determined by electron spectroscopic imaging. The right image represents the possible folding of two adjacent enhancesomes by UBF binding in order to bring the two promoters together. Reprinted with permission from Moss T., Cell Mol Life Sciences, 2007.

The C-terminal acidic domain of UBF recruits SL1 through an interaction with two of its subunits; TAFI48 and TBP, thus stabilizing the PIC (Beckmann et al. 1995, Bell et al. 1990, Friedrich et al. 2005, Hempel et al. 1996, Jantzen et al. 1992, Kihm et al. 1998, Tuan et al. 1999). UBF can also interact with the polymerase’s PAF53 and PAF49 domains, aiding in polymerase recruitment (Hanada et al. 1996-b; Seither et al. 1997). UBF not only binds the promoter region but actually associates with the rDNA along the entire transcribed region, and has been shown to play a role in the decondensation of rDNA chromatin by the displacement of DNA binding proteins (e.g. histone H1) (Chen et al. 2004, Kuhn & Grummt 1992, Kuhn et al. 1993, Mais et al. 2005, O'Sullivan et al. 2002, Sanij et al. 2008). Interestingly, Hmo1p, a yeast HMG-box protein that is a most likely candidate for a UBF analog, also binds along the rDNA repeat and interacts with the RPI subunit A49 to help activate transcription (Gadal et al. 2002, Hall et al. 2006, Kasahara et al. 2007). Though UBF seems to

specifically bind the rDNA more than any other region in the genome, it does not have a consensus binding sequence but rather seems to prefer the rDNA structure (Copenhaver et al. 1994). Therefore, UBF is an activator of rRNA synthesis through PIC stabilization and the formation of a suitable promoter environment for initiation.

1.3.1.2 SL1 An essential component of the pre-initiation complex is SL1 (also called TIF-IB in mouse), which is a ~300 kDa complex formed of the TATA-box binding protein (TBP) and at least three TBP associated factors (TAFs):

TAFI95, TAFI68 and TAFI48 (TAFI110/TAF1C, TAFI63/TAF1B, and TAFI48/TAF1A in human) (Comai et al.

1992, Eberhard et al. 1993, Heix et al. 1997, Learned et al. 1986, Zomerdijk et al. 1994). TAFI41 (TAF1D) and

TAFI12 are two more TAFs that have more recently been implicated in initiation (Denissov et al. 2007, Gorski et al. 2007). In yeast, the core factor (CF), composed of TBP, Rrn6p, Rrn7p, and Rrn11p, appears to be the functional equivalent of SL1, however, there is little sequence homology between the yeast and mammalian factors (Figure 1.16) (Goodfellow & Zomerdijk 2012).

SL1 has been shown to bind the core promoter, thereby stabilizing UBF and aiding recruitment of Pol I-RRN3 to the start site (Cavanaugh et al. 2002, Friedrich et al. 2005, Miller et al. 2001). It is not TBP that confers selectivity for the rDNA promoter but rather TAFI95, TAFI68, and TAFI48 that all make direct contact with the rDNA (Beckmann et al. 1995, Rudloff et al. 1994). They also interact with other PIC members UBF, RRN3, and RPI thereby regulating PIC assembly. In fact, one study demonstrated that the interaction between the C- terminal acidic domain of UBF and SL1 was necessary for the latter to specifically bind the rDNA promoter elements (Tuan et al. 1999). A regulator of this interaction is serine/threonine kinase Casein kinase 2 (CK2) that localizes to the rDNA promoter and phosphorylates TAFI110 and UBF in human cells, stabilizing the PIC though the mechanism remains to be elucidated (Lin et al. 2006b, Panova et al. 2006).

SL1 is involved in the organization of the rDNA regulatory elements by enabling interactions between the promoters, enhancers, and terminators (Denissov et al. 2007). SL1 could also be implicated in the maintenance of promoter hypomethylation due to Gadd45a recruitment by TAFI12 (Schmitz et al. 2009). As previously discussed, Gadd45a has been implicated in keeping the active rDNA gene promoters unmethylated by recruiting the NER machinery to replace methylated cytosines by unmethylated cytosines (Barreto et al. 2007).

1.3.1.3 Pol I RNA polymerase I has 14 subunits in yeast, where ten of these subunits make up the catalytic core of RPI and are either shared or have paralogous partners in RPII and RPIII. The four Pol I-specific subunits form two

peripheral heterodimers; A14/A43 and A49/34.5 (Kuhn et al. 2007, Werner et al. 2009). These subunits are important for RPI functioning in recruitment, promoter escape, and elongation. The A49/34.5 heterodimer can dissociate from the main RPI complex, is able to bind DNA and promote RNA cleavage (Kuhn et al. 2007). A49/A34.5 has been shown to be important for nucleolar formation and regulation of polymerase loading (Albert et al. 2011). Using deletion studies in yeast, these same subunits were shown to be important for Rrn3p recruitment and its dissociation from the polymerase at the start of elongation (Beckouet et al. 2008). In mouse, polymerase associated factor 53 (PAF53, A49 homolog) was shown to interact with UBF, supporting PIC formation (Hanada et al. 1996). However, a study using human UBF and RPI proteins demonstrated that human PAF49 (A34.5 homolog) interacts with UBF even though mouse PAF49 was shown to not interact (Panov et al. 2006b, Yamamoto et al. 2004). Instead, mouse PAF49 was shown to regulate transcription by interacting with TAFI48 (Yamamoto et al. 2004). Whether these discrepancies between human and mouse factor-binding are true in vivo or not is unknown but, it is clear that SL1 and UBF together are important for RPI recruitment.

There are two populations of RPI in mammalian cells, RPIα and RPIβ, which are both capable of transcribing RNA in vitro however it is only RPIβ that can interact with other members of the PIC to form an initiation- competent complex in vivo (Milkereit & Tschochner 1998, Miller et al. 2001). RRN3 interacts directly with the A43 subunit of RPIβ, which represents less than 10% of total RPI in cells. The A49/A34.5 heterodimer is also important for this interaction though it may or may not be directly interacting with RRN3. The interaction of RRN3 to RPI is likely what makes this fraction of the enzymes initiation-competent (Beckouet et al. 2008, Bodem et al. 2000, Cavanaugh et al. 2002, Dundr et al. 2002, Peyroche et al. 2000). RPIα is catalytically active and is able to initiate transcription from DNA ends without specificity, therefore could have a role in elongation (Miller et al. 2001).

Some factors that modulate rRNA transcription have been found to specifically associate with RPIβ but not the RPIα complexes. One example is the serine/threonine kinase casein kinase 2 (CK2) that localizes to the rDNA promoter and phosphorylates TAFI110 and UBF (Lin et al. 2006b, Panova et al. 2006). A second factor that displays specificity to RPIβ, or initiation-competent, complexes is topoisomerase IIα, which is recruited through CK2 though its role is yet unknown (Panova et al. 2006).

1.3.1.4 RRN3 Recruitment of RNA Polymerase I to the UBF-SL1 complex at the promoter requires Rrn3p in yeast and the conserved mammalian counterpart RRN3/TIF-IA (Bodem et al. 2000, Moorefield et al. 2000, Schnapp et al. 1993, Yamamoto et al. 1996). RRN3 is not an overly abundant protein and surprisingly, only a very small

portion of the polymerases are actually bound by RRN3. It is these complexes that are initiation competent, containing RPIβ (Miller et al. 2001).

RRN3 bridges the gap between RPI and SL1, interacting with the A43 subunit and the associated factor

PAF67 of RPI and the TAFI68 and TAFI95 subunits of SL1. This is conserved from yeast to mammals, as Rrn3p also interacts with SL1 equivalent, the core factor, and the polymerase in yeast (Bodem et al. 2000, Cavanaugh et al. 2002, Gorski et al. 2007, Miller et al. 2001, Moorefield et al. 2000, Peyroche et al. 2000, Yuan et al. 2002). Interestingly, yeast Rrn3p is essential for rDNA transcription and viability (Yamamoto et al. 1996), however the RRN3/TIF-IA knockout in mouse was shown to develop to embryonic stage E9.5 (9.5 days post coitum) (Yuan et al. 2005). This was surprising as mouse embryos require a large production of rRNA by this mid-gestational stage that necessitates functional embryonic factors. Maternal contribution of RNA and protein from the oocyte does assist embryonic development until zygotic genome activation occurs at the two- cell stage at which point the majority of maternal mRNA is degraded (Li et al. 2010, Piko & Clegg 1982). It is therefore astonishing that maternal mRNA of RRN3 could support development all the way to E9.5 as the authors claim (Yuan et al. 2005). In fact, the loss of other factors involved in ribosome biogenesis arrests development at blastocyst stage. This has been demonstrated for UBF (Hamdane et al. 2014), the processing factor Fibrillarin (Newton et al. 2003) and the RPA135 subunit of RPI (Chen et al. 2008).

1.3.1.5 Mechanism of PIC assembly The mechanism of PIC assembly has been subject to much debate. Whether the factors assemble in the traditional step-wise manner on the promoter or whether there is the formation of a holoenzyme complex or subcomplexes that are recruited to the rDNA is still not fully known in vivo. FRAP (fluorescence recovery after photobleaching) was used to study the dynamics of various RPI factors (Dundr et al. 2002). The authors tagged PIC subunits with GFP and observed the assembly and deassembly of transcription complexes. They found that RPI complexes are disassembled after each round of transcription and new complexes are assembled on the rDNA promoter, a so-called hit-and-run mechanism. However, in yeast, it has been shown that at least parts of the RPI complex remain stable throughout multiple rounds of transcription (Schneider & Nomura 2004). This could mean that preassembled complexes of RPI factors are present and ready to load on the rDNA promoter. A mouse holoenzyme has been proposed as Pol I, RRN3, SL1 and UBF were co- immunoprecipitated together as an initiation competent complex (Seither et al. 1998). In contrast, a study using affinity purification and gel filtration, also in murine cells, demonstrated a holoenzyme of Pol I, SL1 and TTF1 along with factors related to DNA repair (Hannan et al. 1999). This complex did not contain UBF but was stimulated by it in vitro. This is a more likely model as UBF is required for maintaining the open chromatin state of the rDNA promoter and therefore would not be found in a holoenzyme complex. These complexes have been shown to exist independently of DNA but they have not yet been shown to initiate transcription in vivo.

1.3.2 Initiation Once the initiation-competent PIC has been formed, the opening of the rDNA promoter and incorporation of the first nucleotides of the RNA chain can commence. In order for transcription to continue, the polymerase must escape from the promoter, dissociating from RRN3, SL1 and UBF (Aprikian et al. 2001, Hirschler- Laszkiewicz et al. 2003, Milkereit & Tschochner 1998-b). This step is essential for efficient transcription of the pre-rRNA. In fact, the Grummt Lab has shown that covalent tethering of RRN3 to its interaction partner, the A43 subunit of Pol I, inhibits pre-rRNA transcription. They found that the Ser170 and Ser172 of RRN3, which are conserved in Xenopus, chicken, mouse and human, require phosphorylation by CK2 in order for RRN3 to dissociate from Pol I, allowing elongation to occur. In their model, RRN3 would be then be dephosphorylated by FCP1 (TFIIF-associating component of CTD phosphatase), a phosphatase responsible for RPII C-terminal dephosphorylation and recycling, allowing for a new PIC to be formed and initiate a new round of transcription (Bierhoff et al. 2008, Cho et al. 1999). Contrary to this, a study in yeast expressed a RPI-Rrn3p fusion protein that could not dissociate in strains lacking Rrn3 and A43, but there was no transcription defect in this case (Laferté et al. 2006). This inconsistency could reflect different regulatory mechanisms of the PIC in yeast and mammals as it has been shown that these two species have different phosphorylation requirements for the RPI-RRN3 interaction (Bierhoff et al. 2008, Cavanaugh et al. 2002, Fath et al. 2001).

As previously mentioned, UBF interacts with the RPI heterodimer PAF49/PAF53. This could potentially activate transcription through changes in the DNA or polymerase conformation, though it is still unclear (Hanada et al. 1996-b). In yeast, A49/A34.5 has been shown to be important for nucleolar formation and regulation of polymerase loading (Albert et al. 2011). Using deletion studies, these same subunits were shown to be important for Rrn3p recruitment and its dissociation from the polymerase at the start of elongation (Beckouet et al. 2008). Hmo1p also interacts with A49 and helps with Rrn3p release, therefore playing a role in promoter escape (Beckouet et al. 2008, Gadal et al. 2002).

1.3.3 Elongation Transcription elongation by RPI must be efficient to produce the impressive quantity of rRNA transcripts required for ribosome biogenesis. It has been calculated that 100 polymerases on average are transcribing each active rDNA gene at a rate of ~95 nucleotides per second in mammalian cells (Dundr et al. 2002). In yeast, the elongation rate was calculated to be ~60 nucleotides per second and 50 polymerases per active gene on average (French et al. 2003). This remarkable production is regulated intrinsically by RPI’s subunits and by other factors. It is the A34.5/A49 complex that is functionally analogous to the RPII elongation factor TFIIF and is important for RPI elongation.

Notably, UBF has been shown to be important for elongation through phosphorylation-dependant remodeling of the rDNA chromatin. Two ERK phosphorylation sites on Thr117 and Thr201 of UBF HMG-boxes 1 and 2 were identified to have an activating role in rRNA production (Figure 1.17) (Stefanovsky et al. 2001b). This phosphorylation was shown to reduce UBF affinity for DNA. It was proposed that the enhanceosome structure formed by UBF could impede RPI advancement and therefore the phosphorylation by ERK would relax this structure and allow chromatin remodeling and RPI progression through the gene (Stefanovsky et al. 2006b). Interestingly, the UBF gene produces a splice variant, UBF2, that is missing 37 amino acids (a.a.) from HMG box 2. Though this form of UBF is expressed in equimolar amounts, it is less able to transactivate rRNA transcription (Kuhn et al. 1994, O'Mahony & Rothblum 1991).

Along this same vein, other factors that are involved in chromatin remodeling or help alleviate supercoiling and DNA damage are involved in elongation regulation. Topoisomerase IIα has been identified in RPIβ complexes in human cells and topoisomerase I was found with RPI in mouse and human cells (Hannan et al. 1999, Panova et al. 2006, Zhang et al. 1988). DNA damage leads to temporary repression of pre-rRNA transcription, partly due to ATM-mediated RPI displacement, and repair needs to occur before elongation ensues (Kruhlak et al. 2007).

As the pre-rRNA is cotranscriptionally processed, elongation is tightly related to the processing machinery. Transcriptional defects translate to processing defects and vice-versa (Granneman & Baserga 2005). The production of the pre-rRNA is tightly regulated and requires coordination of many factors, both RPI-specific and nonspecific.

1.3.4 Termination

Termination of RPI transcription requires binding of transcription termination factor (TTF1) to the T1-T10 terminator elements downstream of the rDNA gene, which causes the polymerase to pause (Bartsch et al. 1988, Grummt et al. 1985). RPI and transcript release factor PTRF then dissociate the transcription complex. In yeast, Reb1p is the TTF1 homolog that regulates transcription termination (Jansa & Grummt 1999).

1.3.4.1 TTF1 A key regulator of transcription termination and the epigenetic state of the rDNA is TTF1. TTF1 binds multiple sites in the rDNA spacer, the 3’ terminator sites (T1-T10) and two 5’ sites (TSP and T0) (Figure 1.9) (Grummt et al. 1985-b; Németh et al. 2008). These sites all contain conserved 18 bp SalI restriction sites and are therefore named Sal-boxes. Studies demonstrated that displacement of the T0 by even as few as two nucleotides negatively affected transcription and mutations in the SalI sequence in the T0 or T1-T10 boxes affected termination (Grummt et al. 1986b, McStay & Reeder 1990).

These Sal-box sites probably play a role in the topological organization of the rDNA chromatin. Chromosome conformation capture (3C) demonstrated interactions between the promoter and the terminator regions, which are mediated by TTF1 (Németh et al. 2008). These loops are mainly formed at active rDNA repeats according to the methylation status and transcription factor or histone occupancy at the promoter. This corresponds to the ‘ribomotor’ model in yeast that had previously suggested promoter-terminator looping in order to recycle polymerase from the terminator to the promoter in order to ensure rapid reinitiation (Kempers-Veenstra et al. 1986). This model would ensure the continued high density of polymerases on the rRNA gene in growing cells.

Evidently, the primary role of TTF1 at the terminators is arresting the polymerase but the upstream sites have been implicated in other processes of rRNA production regulation. Studies using mammalian cells or factors in vitro found TTF1 to be a transcriptional activator through chromatin modification. TTF1 binding to the T0 was implicated in recruiting ATP-dependent chromatin remodeling cofactors, displacing nucleosomes and activating transcription (Diermeier et al. 2013, Längst et al. 1998, Németh et al. 2004). Later it was found that

TTF1 bound to the T0 was able to recruit either the NoRC silencing complex or the ATP-dependent remodeling factor CSB, either playing a role in transcription activation or silencing, which has already been described in detail elsewhere in this chapter.

TTF1 was not only shown to be important for rRNA transcription termination but also for blocking the DNA replication machinery as the replication fork barrier (RFB) found at the 3’ end of the rDNA necessitates the terminator sites and TTF1 to function correctly (Gerber et al. 1997, Kobayashi et al. 1992). When RFB activity was reduced, RPI transcription inhibited replication fork progression therefore the RFB site is important for coordinating replication and transcription machineries on the rDNA (Akamatsu & Kobayashi 2015).

1.3.5 Reinitiation Once transcription is correctly terminated and the pre-rRNA transcript is released, RPI and the other factors can be recycled to form new pre-initiation complexes and reinitiate. It has been proposed that the TTF1 binding sites upstream of the promoter (T0), downstream of the spacer promoter (TSP) and downstream of the gene (T1-T10) could be involved in forming DNA loops in the promoter elements to facilitate reinitiation. SL1 and c-Myc have also been implicated in the formation of these loops (Nemeth 2008, Nemeth Langst 2011, Shiue 2009) though it is unclear whether these chromatin interactions are indeed necessary for transcription reinitiation.

1.3.6 Regulation of rRNA gene activity and transcription rates Due to the multitude of rRNA gene copies, two levels of regulation of transcription rates are possible. The polymerase itself may be regulated through various factors, modulating the rate of transcription from each

gene, or the number of genes that are being transcribed could change (McStay & Grummt 2008, Russell & Zomerdijk 2005). There is a large body of research supporting the first option, both through regulating the polymerase activity and post-translational modifications of Pol I-specific transcription factors. These points of control are especially relevant for short-term regulation of rRNA synthesis and will be discussed in detail in the following section. The number of active rRNA genes in mammals changes during development and differentiation, varying between different cell types (Haaf et al. 1991) and this level of regulation has more long-term effects.

In yeast, environmental conditions affect rRNA synthesis by modulating both the number of active genes and the rate of transcription per gene. As with many mammalian cell lines, yeast cells also display a 50:50 ratio of active to inactive genes during growth or log phase. However, during stationary phase, chromatin modifications, owing to the activity of the Rpd3-Sin3 HDAC complex, close the rDNA and the proportion of active genes is reduced (Dammann et al. 1993, Sandmeier et al. 2002). Even though this natural regulation of the number of rRNA genes occurs, there is also a level of regulation involving transcription rate. For example, a genetically engineered yeast model containing less than a third of the normal number of genes still produces equivalent amounts of rRNA transcripts (French et al. 2003). Furthermore, Rpd3-deleted strains that conserved a stable number of active genes still demonstrated a reduction in the quantity of rRNA transcripts upon entering stationary phase (Sandmeier et al. 2002). Therefore both levels of regulation are employed in yeast cells for regulation during the cell cycle.

Few studies in mammalian cells have concomitantly investigated the regulation of rRNA production and the proportion of active genes. Stefanovsky and Moss had demonstrated an upregulation of rRNA transcription in mouse and human cells upon growth factor stimulation or treatment with TSA, an HDAC inhibitor (Stefanovsky & Moss 2006). The proportion of active rRNA genes did not significantly change as observed by psoralen crosslinking experiments. It was later found that the augmentation of rRNA transcripts was not due to a higher polymerase loading on the gene but rather an accelerated elongation rate through mitogenic activation of UBF (Stefanovsky et al. 2006a). Another study investigated the role of UBF in rRNA transcription rate by siRNA- mediated knockdown of this factor (Sanij et al. 2008). The authors found that UBF plays a role in determining the number of active rRNA genes, however the transcription rate was unchanged, likely due to an upregulation of transcription from the remaining active genes. Therefore, UBF is an important regulator of both potential mechanisms of rRNA synthesis control.

1.3.7 Cell growth, proliferation and rRNA transcription

1.3.7.1 Cell growth Cell growth, or increase in cell mass, is essential for cell division and pre-rRNA levels and thus ribosome content echoes the cell’s capacity to grow and divide (Moss et al. 2007). As mentioned, ribosome biogenesis involves all three RNA polymerases and there must be a tight control of their activity in order to correctly transcribe the cell’s demand for ribosomes. It was proposed that a equimolar supply of rRNAs (produced by RPI and RPIII) and the ribosomal proteins (mRNAs produced by Pol II) was necessary for efficient ribosome assembly as a decrease in the level of rRNA transcripts lead to a decrease in the level of r-protein mRNAs and 5S rRNAs (Clarke et al. 1996, Powers & Walter 1999). Laferté and colleagues have shown that there is a link between the level of transcription of each of these polymerases in regards to ribosome production and that RPI transcription is the primary level of regulation (Laferté et al. 2006). Other studies claim that the number of ribosomal proteins can be regulated post-translationally rather than transcriptionally and the rapid turnover of r- proteins regulates the equilibrium between protein and rRNA ribosome components. Ribosomal proteins are synthesized in excess and rapidly imported into the nucleus and nucleolus. If required to produce ribosomal subunits, they will associate with rRNA and if not, they will be targeted for proteasomal degradation (Andersen et al. 2005). This corresponds with another group’s study implicating ubiquitin and the proteasome in ribosome biogenesis (Stavreva et al. 2006).

1.3.7.2 Cell cycle Ribosomal DNA transcription is cell-cycle related in mammals and starts after mitosis in G1 increasing to a maximum in S and G2 phases. Various epigenetic marks regulate the activity of the rDNA copies in interphase and these are maintained in order to re-establish the correct proportion of active and silent genes following mitosis (McStay 2016). Both SL1 and UBF remain associated with the rDNA during the disassembly of the nucleolus during mitosis and both factors are phosphorylated by cyclin dependant kinases (cdk)/cyclin complexes during the cell cycle. They could therefore have a role in arresting RPI at the start of mitosis and its reactivation in G1. The TAFI110 subunit of SL1 is phosphorylated during metaphase by cdc2-cyclin B and is therefore inactivated and cannot interact with UBF (Heix et al. 1998, Kuhn et al. 1998). In addition, TAFI68 is deacetylated by SIRT1 on the onset of mitosis, destabilizing SL1 binding to the rDNA promoter (Voit et al. 2015). The phosphorylation status of UBF is also regulated throughout the cell cycle. UBF is inactive and tightly bound to the rDNA in mitosis and then is activated in G1, phosphorylated first by cdk4-cyclin D1 and cdk2-cyclin E at Ser484 and then by cdk2-cyclin E and cdk2-cyclin A at Ser388 (Figure 1.18). This is required for UBF’s interaction with RPI and thus pre-initiation complex formation (Klein & Grummt 1999, Sirri et al. 1999, Voit & Grummt 2001, Voit et al. 2015). Cdc2-cyclin B also phosphorylates TTF1, which corresponds to a reduced chromatin-binding affinity of the latter to mitotic chromosomes (Sirri et al. 1999). TTF1 and UBF have

both been proposed to be involved in regulating rDNA chromatin architecture so these post-translational modifications could play a role in keeping the mitotic NORs inactive for a short time period but not epigenetically silenced (Russell & Zomerdijk 2005).

Figure 1.18 rDNA transcription is cell cycle dependant UBF is phosphorylated by Cdk4/cyclin D at serine 484 (S484) and by Cdk2/cyclin E or cyclin A at serine 388 (S388) during interphase, which activates rDNA transcription. At the start of mitosis, SL1 is phosphorylated on its TAFI110 subunit at threonine 852 (T852) by Cdk1/cyclin B, which inactivates SL1 therefore inhibiting rDNA transcription. Cdc14B then dephosphorylates SL1 at the end of mitosis, allowing rDNA transcription to re-initiate. Reprinted with permission from Drygin D., Annu Rev Pharmacol Toxicol, 2010.

1.3.7.3 Mitogen activation of rRNA transcription Ribosomal RNA transcription responds to extracellular signals that regulate cell growth and cell cycle control in order to produce enough ribosomes to fill the growing cell’s demand for protein synthesis. Mitogen-activated protein kinase (MAPK), phosphatidyl inositol-3 kinase (PI3K) and mammalian target of rapamycin (mTOR) pathways have been shown to converge on rRNA transcription, and more specifically the RPI factors, in response to nutrients, mitogens or growth factors (Figure 1.19) (Grummt 2003, Russell & Zomerdijk 2005). All of the RPI machinery seems to be affected by growth factor signaling pathways, though none so much as RRN3 and UBF (Tables 1.1 and 1.2). As shown in Table 1.2, direct phosphorylation of RRN3 at several sites is believed to be due to the ERK and RSK kinases, and mutation of these sites suppresses transcription. However in mammals it has been shown that UBF phosphorylation through the ERK/MAPK pathway is

necessary for transcription (Stefanovsky et al. 2001b, 2006a,b). There is also evidence that RRN3 can be regulated by the mTOR pathway and via the JNK pathway in stress situations (Mayer 2005, Mayer et al. 2004, Zhao et al. 2003). Treatment with rapamycin, an mTOR inhibitor, downregulates RPI transcription in yeast or mammalian cells as RRN3 is dissociated from the PIC and removed to the cytoplasm (Claypool et al. 2004, Mayer et al. 2004). However, Hannan et al. found that it is UBF that is affected by the TOR pathway-mediated repression of rRNA transcription (Hannan et al. 2003). As previously described, UBF is not only important for PIC formation and initiation but also for elongation, which is regulated by growth factor signaling. Furthermore, in yeast, extracts from growth-arrested cells do not contain active initiation competent RPI-RRN3 complexes (Milkereit & Tschochner 1998). Evidently, the mitogenic activation of rRNA transcription is regulated on multiple levels and understanding the difference between these essential factors and their potential levels of action is an important aspect of the comprehension of growth regulation.

Figure 1.19 rRNA transcription is linked to cell growth RNA polymerase I transcription is regulated in response to external stimuli. The bar diagrams show the relative response of RPI transcription to amino acid starvation (blue), to oxidative stress (orange) and to growth factor stimulation (purple). Reprinted with permission from Drygin D., Annu Rev Pharmacol Toxicol, 2010.

Table 1.1 Identified phosphorylation sites of UBF

Amino Acid Site Functional Data References Threonine 117 -phosphorylated by ERK1/2 (Stefanovsky et al. -T117A and T117D/E inhibit transcription activation and 2006b) abrogate immediate response to ERK1/2 {Stefanovsky:2006dk} -phosphorylation prevents interaction with DNA Threonine 201 -phosphorylated by ERK1/2 (Stefanovsky et al. -T201A and T201D/E inhibit transcription activation and 2006b) abrogate immediate response to ERK1/2 {Stefanovsky:2006dk -phosphorylation prevents interaction with DNA Serine 273 -identified in mass spectrometry study (no functional data) (Lin et al. 2006a) Serine 336 -identified in mass spectrometry study (no functional data) (Lin et al. 2006a) Serine 364 -identified in mass spectrometry study (no functional data) (Lin et al. 2006a) Serine 388 -after progression through G1 cdk2/cyclin E and A (Voit & Grummt phosphorylate S388 2001) -S388G UBF transcriptionally inactive -S388D enhances UBF activity -interaction with Pol I lost with S388G Serine 389 -identified in mass spectrometry study (Lin et al. 2006a) -S389A abrogated rRNA transcription in vitro and in vivo -S389E restored transcriptional activity -S389A caused loss of UBF-SL1 interaction in vitro -S389E partially restored UBF-SL1 interaction Serine 412 -identified in mass spectrometry study (no functional data) (Lin et al. 2006a) Serine 433 -identified in mass spectrometry study (no functional data) (Lin et al. 2006a) Serine 484 -identified in mass spectrometry study (Lin et al. 2006a) -phosphorylated by cdk4-cyclin D1 and cdk2-cyclin E (Voit et al. 1999) -G1 specific -mutation impairs rDNA transcription in vivo and in vitro Serine 546 -identified in mass spectrometry study (no functional data) (Lin et al. 2006a) Serine 584 -identified in mass spectrometry study (Lin et al. 2006a) -S584A reduced transcription in vivo but not in vitro Serine 638 -identified in mass spectrometry study (no functional data) (Lin et al. 2006a) Serine -multiple -phosphorylation upregulated with serum (Hannan et al. 2003) sites in -downregulated with rapamycin (mTOR dependant) (Kihm et al. 1998) acidic C- -phosphorylation important for Pol I transcription in vitro (Lin et al. 2006b) terminus -rapamycin does not inhibit rrn3 activity (Drakas et al. 2004) -CKII phosphorylated two of the eleven sites -CKI, Cdc2, and GSK each phosphorylated UBF but failed to enhance the TBP binding activity of rUBF -CKII phosphorylation helps stabilize interaction with SLI, promoting multiple rounds of transcription -UBF(∆675-765)-SL1 interaction lost -UBF(∆657-764)-TBP interaction lost -p110 subunit of PI3-K directly phosphorylates and activates UBF1, in complex with IRS-1 -mTOR stimulation of rRNA transcription requires S6K1 activation and phosphorylation of UBF C-terminal tail

Table 1.2 Identified phosphorylation sites of RRN3

Amino Acid Site Functional Data References Serine 44 -identified by mass spectrometry analysis (Mayer et al. 2004, -rapamycin treatment decreases phosphorylation at S44, Schlosser et al. 2002) downregulating TIF-IA activity indicating regulation through mTOR and likely S6K1 -S44A inactivates TIF-IA but does not cause loss of Pol I or SLI interaction Serines 170/172 -identified by mass spectrometry analysis (Schlosser et al. -phosphorylated by CK2 which releases TIF-IA from Pol I after 2002) initiation (Bierhoff et al. 2008) -dephosphorylation by FCP1 allows for a new round of transcription -inhibition of S170/172 sites inhibits rRNA transcription and causes nucleolar disruption Serine 199 -rapamycin treatment increases phosphorylation at S199, {Mayer:2004hk} downregulating TIF-IA activity -S199A has no effect -S199D TIF-IA is transcriptionally inactive and cannot interact with Pol I and SLI Threonine 200 -phosphorylation at T200 by JNK2 inhibits TIF-IA and rRNA (Bogoyevitch & Kobe synthesis under oxidative stress conditions 2006) {Mayer:2005et} Serine 633 -phosphorylated by ERK, activates TIF-IA {Zhao:2003ud} Serine 635 -phosphorylated by AMPK, impairs interaction with SLI {Hoppe:2009gi} Serine 649 -phosphorylated by RSK, activates TIF-IA {Zhao:2003ud} -S649A abolishes TIF-IA activity, impairs Pol I transcription, and retards cell growth

1.3.8 rRNA transcription and apoptosis Cell cycle arrest and senescence as a protective response to nucleolar stress was already discussed. However, impaired ribosome biogenesis, specifically impaired rRNA transcription, can also lead to apoptosis. Apoptosis is one of the main programmed cell death pathways in the cell. Apoptosis is a normal process that supports cellular homeostasis, development, and aging (Elmore 2007). Both external signals, from immune cells for example, or internal signals, such as activation of the DNA damage repair pathway, can activate apoptosis. Apoptosis leads to DNA fragmentation, reduction in cell size, blebbing of the membrane and ultimately phagocytosis by macrophages. This cell death mechanism involves a cascade of proteins but the cysteine protease caspases are among the main factors implicated in apoptosis and cleavage, or activation, of caspase-3 is regarded as a hallmark of apoptosis (Figure 1.20) (Elmore 2007). The guardian of the genome, p53, can trigger cell death upon cellular stress or the accumulation of DNA damage, though there are situations where DNA damage induces apoptosis in a p53-independent manner (Roos & Kaina 2006).

Multiple studies affecting both UBF and RRN3 have resulted in cell cycle arrest and apoptosis. UBF antibodies injected into human fibroblasts led to rRNA transcription inhibition, nucleolar perturbation, stabilization of p53 and apoptosis (Rubbi & Milner 2003). Therefore, inhibition of UBF led to cell death, however, knockdown of UBF in NIH3T3 cells did not lead to cell cycle arrest or apoptosis. In fact, rRNA transcription in cells treated with siRNA against UBF was only ~15% lower than the control cells (Sanij et al. 2008). The reduction in UBF led to the silencing of the rDNA but the remaining active copies increased their transcriptional output in order to compensate. However, more recently it was shown that the complete loss of UBF by knockout in mice does lead to a developmental arrest at E3.5 (Hamdane et al. 2014). The knockout of RRN3/TIF-IA also led to nucleolar disruption and p53-dependent apoptosis in embryos, though surprisingly at E9.5 (Yuan et al. 2005). Interestingly, a peptide corresponding to 22 amino acids in RRN3 on its binding interface with RPI can be used to inhibit rRNA transcription. This inhibition led to cell death but the authors characterized it as a necrotic cell death not consistent with apoptosis (Rothblum et al. 2014).

One recent study proposed a mechanism for the regulation of p53 upon reduction of RPI transcription factor levels and inhibition of rRNA synthesis. The authors observed an increase of acetylation on lysine 382 (K382) of p53 after depletion of the two largest RPI subunits, UBF or RRN3. P21 and pro-apoptotic BAX protein levels were increased, poly (ADP-ribose) polymerase (PARP) cleavage was enhanced and apoptosis ensued. They found that the K382 acetylation was dependent on nucleolar protein MYB binding protein 1a (MYBBP1A) which was translocated to the nucleoplasm. Contrary to this, when rRNA processing factors were depleted, MYBBP1A was retained in the nucleolus and though p53 was accumulated, its acetylation was not enhanced. This inhibition of ribosome biogenesis led to a RPL11-dependent G1 cell-cycle arrest and not apoptosis (Kumazawa et al. 2015). Therefore, p53 could mediate cell cycle arrest upon rRNA processing defects but apoptosis upon rRNA transcription defects.

Figure 1.20 Main apoptotic signaling pathways The intrinsic pathway is outlined in the center of the diagram. Cellular stress activates tumor-suppressor p53, which upregulates Puma and Noxa, leading to activation of Bax and Bak. Bax and Bak permeabilize the outer mitochondrial membrane, which releases cytochrome c, an integral part of the apoptosome complex. This activates the effector caspases 3, 6, and 7, which triggers apoptosis. The extrinsic pathway is activated by pro-apoptotic ligands produced by cytotoxic immune cells. These ligands activate DR4/DR5 receptors, recruiting Fas-associated death domain (FADD) and procaspases 8 and 10, which together form the death-inducing signaling complex (DISC). The cleaved caspases 8 and 10 activate the effector caspases 3, 6, and 7, leading to apoptosis. Reprinted with permission from Ashkenazi A., Nature Reviews Drug Discovery, 2008.

1.4 Ribosome biogenesis and disease

1.4.1 Ribosomopathies Ribosomopathies are a group of human diseases that are related to defects in ribosome biogenesis. Interestingly, even though ribosome biogenesis is an essential process in all cells, some ribosomopathies are tissue-specific (McCann & Baserga 2013). Mutations in factors from all levels of ribosome biogenesis have been shown to result in these diseases (Freed et al. 2010). Some ribosomopathies are caused by defects in rRNA transcription, like Treacher-Collins syndrome (TCS). TCS is due to autosomal dominant mutations in the TCOF1 gene and the disease seems to be due to haploinsufficiency. Treacle, the product of TCOF1, is a nucleolar phosphoprotein that colocalizes with UBF and RPI and plays a regulatory role in rRNA transcription

and processing (Freed et al. 2010, Hannan et al. 2013). It was more recently found that TCS can also be caused by mutations in RPI subunits RPA16 and RPA40 further linking this disease to rRNA synthesis (Yelick & Trainor 2015). Other ribosomopathies are due to mutations in small nucleolar ribonucleoproteins, the RNA/protein complexes that cleave, modify and fold the pre-rRNA. These include Dyskeratosis congenita which presents with a predisposition to cancer and bone marrow failure (Freed et al. 2010). A third category of ribosomopathies is due to mutations in various ribosomal proteins. The most studied of these is Diamond- Blackfan anemia (DBA) which is a congenital syndrome correlated with anemia, malformations and cancer. DBA is caused by mutations in ribosomal proteins, often RPS19, RPL5 and RPL11, leading to haploinsufficiency of the r-protein in question and defects in ribosome biogenesis (Danilova & Gazda 2015). This deficiency in ribosome biogenesis leads to an accumulation of free r-proteins, which can bind MDM2, thus liberating and activating p53, leading to cell cycle arrest and apoptosis (Narla & Ebert 2010). Although these diseases are caused by mutations in factors from various steps in ribosome biogenesis, they often present with similar symptoms; malformations, anemia, and a predisposition to cancer (Yelick & Trainor 2015).

The nucleolus may play a role in other diseases that are not directly related to ribosome biogenesis such as Bloom syndrome, Werner syndrome and Rothmund-Thomson syndrome. These diseases result from mutations in the BLM, WRN or RECQL4 DNA helicases, respectively, which are involved in DNA repair. Both the BLM and WRN helicases accumulate in the nucleolus during S phase and RECQL4 does so under oxidative stress conditions (Boisvert et al. 2007). WRN and BLM positively regulate RPI transcription though the mechanism is still unknown (Hannan et al. 2013). These diseases manifest with a predisposition to certain cancers and chromosomal instability so it is interesting to note that the nucleolus may play a role in resisting stress or maintaining genomic integrity by sequestering these helicases (Boisvert et al. 2007).

1.4.2 Cancer Ribosomal RNA transcription is significantly upregulated in cancer cells, therefore the nucleoli are larger and usually disorganized in human cancer cell lines (Farley et al. 2015). This morphological change has been used as a diagnostic marker for cancer aggressivity for over 100 years (Derenzini et al. 2009). Ribosome biogenesis must be upregulated to accommodate the requirement for increased cell growth and proliferation in cancer cells. The upregulation of rRNA synthesis may be due to overexpression or increased activity of components of the RPI machinery (Russell & Zomerdijk 2005). For example, UBF is overexpressed in 70% of hepatocellular carcinomas (Huang et al. 2002), however Rrn3 is not commonly mutated or amplified in most cancers and is not likely to be an oncogenic driver in general (Jin & Zhou 2016). In spite of this, it is upregulated or amplified in 16% of invasive breast cancer cases and could contribute to cellular transformation. In fact, invasive breast cancer patients that had mutations, amplifications or overexpression of

multiple RPI factors had a worse prognosis than patients with the same cancer but no alterations in these genes (Rossetti et al. 2016).

There is not only a link between ribosome biogenesis and cancer due to the increased production of protein necessary for cancer cell growth but also because ribosome biogenesis and, in particular, rRNA transcription is a target of many proteins implicated in cancer development. This process is regulated by growth factor signaling pathways such as Ras/MAPK, mTOR and PI3K, by tumor-suppressor proteins such as p53 and Rb, and by the proto-oncogene Myc. These factors are often mutated in precancerous cells, which can upregulate ribosome biogenesis leading to aberrant cell growth and division and potentially cellular transformation (Figure 1.21) (Drygin et al. 2010). A few of the best-described mechanisms relating rRNA transcription and cancer will be described here.

Figure 1.21 Oncogenes and tumor suppressor regulation of RPI basal factors

Oncogenes activate rRNA transcription either by stabilizing transcription factor interactions (protein-protein or protein-DNA) or by regulating the level of transcription factor expression (green arrows). Tumour suppressors inhibit rRNA transcription by perturbing interactions between factors involved in preinitiation complex formation. Reprinted with permission from Drygin D., Annu Rev Pharmacol Toxicol, 2010.

1.4.2.1 c-Myc Interestingly, the oncoprotein Myc has been shown to regulate all three RNA polymerases and therefore has a role in regulating ribosome production and thus, cell growth (Oskarsson & Trumpp 2005). It has been shown to regulate RPIII transcription through an interaction with TFIIIB (Gomez-Roman et al. 2003). C-Myc also has an influence on multiple RPII gene products including rRNA processing factors (Schlosser et al. 2003), UBF (Poortinga et al. 2004) and RRN3 (Poortinga et al. 2011, 2014). C-Myc associates with enhancer elements in the rDNA promoter and helps remodel chromatin by recruiting HATs, thereby activating RPI transcription (Arabi et al. 2005, Grandori et al. 2005). C-Myc interacts with SL1 and high levels of c-Myc have been shown to increase the occupancy of TBP on the promoter, which suggests positive regulation of rDNA transcription (Figure 1.22) (Grandori et al. 2005). Therefore, c-Myc can influence all three RNA polymerases in regards to ribosome biogenesis (Oskarsson & Trumpp 2005).

Figure 1.22 c-Myc regulates multiple levels of ribosome biogenesis c-Myc enhances transcription by all three RNA polymerases, playing a double role of regulating rRNA transcription by RPI/Pol I by modulating chromatin at the rDNA promoter and upregulating UBF and RRN3/TIFIA levels. Reprinted with permission from Oskarsson T. and Trumpp A., Nat Cell Biol, 2005.

1.4.2.2 Nucleophosmin/B23 Nucleophosmin (B23/NPM) is a phosphoprotein that is often overexpressed in cancer cells (Drygin et al. 2010). For example, ~30% of patients with acute myeloid leukemia have mutations in NPM that cause its translocation to the cytoplasm. Through its intrinsic histone chaperone and endonuclease activity, NPM plays a role in chromatin organization, DNA repair, pre-rRNA processing, and regulation of p53 and p19ARF activity

(Grisendi et al. 2006, Murano et al. 2008, Savkur & Olson 1998). Overexpression of NPM upregulates TAFI48 levels and stimulates cell division in transformed cells (Bergstralh et al. 2007, Grisendi et al. 2006). NPM also

interacts with c-Myc, regulating its localization and activation of rDNA transcription and coordinating proliferation and transformation induced by c-Myc (Li & Hann 2012, 2009).

1.4.2.3 p53 The guardian of the genome p53 responds to DNA damage or cellular stress by inducing a cell cycle arrest and repair, apoptosis or senescence. Many cancer cell lines lack a functional p53 and therefore bypass the normal cell checkpoints (Sigal & Rotter 2000). Cancer cells with deleted or mutated Rb and p53 have higher rRNA transcription rates and are usually more aggressive (Bates et al. 1998, Treré et al. 2004). P53 directly downregulates RPI through interaction with SL1, which interferes with PIC formation (Budde & Grummt 1999, Zhai & Comai 2000). Additionally, p53-null cells show increased RPI transcription (Zhai & Comai 2000).

1.4.2.4 RB The retinoblastoma susceptibility (Rb) gene is an important tumor suppressor and is mutated or inactive in the majority of cancer cells. RB is well known to interact with E2F transcription factors, inhibiting cell cycle progression through G1 (Dyson 1998). However, RB also interacts with UBF, dissociating it from the rDNA and therefore affecting PIC formation and rRNA transcription (Cavanaugh et al. 1995, Hannan et al. 2000a, Voit et al. 1997). RB downregulates RPI transcription in confluent cells that are cell cycle arrested (Hannan et al. 2000b), demonstrating again a role of rRNA synthesis regulation in the control of proliferation and cell growth.

1.4.2.5 ARF, MDM2 and TTF ARF (p19ARF in mouse, p14ARF in human) is the product of an alternative reading frame from the INK4a/ARF locus, which also encodes the p16INK4A tumor suppressor, and is inactivated in many cancers (Ko et al. 2016). As previously mentioned, ARF’s primary role is as a binding partner of MDM2, which regulates p53. Normally ARF is sequestered in the nucleolus but upon stress or oncogenic activation, it is released to the nucleoplasm and can, therefore, bind MDM2, which allows release and accumulation of p53. Activated p53 then induces a cell cycle arrest and potentially senescence or apoptosis. The overexpression of c-Myc and some other oncogenes leads to accumulation of ARF which helps counteract the hyper-proliferation of cells. Therefore, ARF has a tumor suppressor function through its role in p53 regulation.

Arf knockout mice show altered nucleoli morphology and increased protein synthesis and ribosome content (Apicelli et al. 2008). Using Arf-/- cells from knockout mice, Saporita et al. demonstrated that the DEAD-box RNA helicase DDX5 (also called p68) associates with UBF at the rDNA promoter, upregulating pre-rRNA synthesis. In wild-type cells, DDX5 is sequestered in the nucleus and occupancy on the rDNA is two-fold less than in the Arf-/- cells. In addition, when exogenous ARF was expressed in p53−/−Mdm2−/−Arf−/− triple

knockout MEFs, ribosome production was reduced, demonstrating the important role of ARF in the regulation of rRNA synthesis (Figure 1.23) (Saporita et al. 2011).

Figure 1.23 ARF is a negative regulator of ribosome biogenesis In wild-type MEFs, ARF interacts with DDX5 and sequesters it in the nucleoplasm however in Arf-/- MEFs, DDX5 locates to the rDNA promoter and upregulates rDNA transcription leading to an increase in ribosome production. When exogenous HA-ARF was expressed in p53-/-Mdm2-/-Arf-/- MEFs, ribosome production was reduced. Reprinted with permission from James A., Nucleus, 2014.

ARF not only regulates the location of DDX5 but a similar mechanism was also shown for TTF1. ARF interacts with TTF1 in the nucleoplasm, blocking the function of its nucleolar localization sequence (NoLS). This prevents TTF1 shuttling to the nucleolus and binding of the rDNA, consequently downregulating rRNA transcription (Lessard et al. 2010). ARF had previously been shown to interact with and degrade NPM, thereby balancing ribosome synthesis by inhibiting the NPM-mediated upregulation of rDNA transcription (Apicelli et al. 2008, Bertwistle et al. 2004, Itahana et al. 2003). ARF was implicated in NPM shuttling between the nucleoplasm and nucleolus (Brady et al. 2004). Lessard et al. have demonstrated that NPM is responsible for the nucleolar shuttling of TTF1 and necessary for its localization on the rDNA and that ARF can inhibit this process (Lessard et al. 2010). This study relates the ARF-NPM regulation of ribosome biogenesis directly to regulation of rDNA transcription. In a further study, it was found that in the absence of ARF, MDM2 ubiquitinylates TTF1 targeting it for degradation (Lessard et al. 2012). Therefore ARF and MDM2 regulate levels of TTF1 and the chaperone NPM is essential for the localization of TTF1 (Figure 1.24). Altogether these

results demonstrate an important role for ARF as a tumor suppressor as well as a regulator of ribosome biogenesis, therefore it is likely a major factor in coupling proliferation and cell growth.

Figure 1.24 NPM and ARF regulate TTF1 localization and activity (A) NPM is necessary for TTF1 transport into the nucleolus and could activate TTF1 by aiding its unfolding to reveal its NoLS. (B) ARF binds and inhibits the NoLS of TTF1 and blocks its entry into the nucleolus. Reprinted with permission from Lessard F. et al., Molecular Cell, 2010.

1.4.2.6 Ribosomal RNA transcription as a target for cancer treatment As many of the above examples demonstrate, multiple oncogenes and tumor suppressors converge on the rDNA transcription machinery. Therefore, there are many cancer drugs that act at least partly on rRNA synthesis (Table 1.2). Cisplatin is an interesting example as it is used to treat many types of cancer including sarcomas and lymphomas. It was originally thought that cisplatin-induced apoptosis through an excess of DNA damage caused by its crosslinks, however, cisplatin also has RPI transcription-specific activity (Jordan & Carmo-Fonseca 1998). It was found that UBF has a very high affinity for cisplatin-DNA adducts and that this can hijack UBF away from the rDNA (Chao et al. 1996, Treiber et al. 1994, Xiaoquan Zhai et al. 1998). Therefore, it may be that rRNA synthesis is primarily affected in these cells prior to the accumulation of damage.

There has been recent interest in developing drugs that target the RPI machinery specifically, even though it was long thought that ‘housekeeping processes’ should not be used as therapeutic targets as they would lack

specificity for transformed cells in a population. Cylene Pharmaceuticals developed a series of small molecules to target rRNA transcription and one, in particular, is promising (Quin et al. 2014). Bywater et al. reported that a small molecule inhibitor of RPI transcription, CX-5461, targeted B-cell lymphoma cells in vivo but did not affect the normal B-cells (Bywater et al. 2012). CX-5461, which inhibits rRNA transcription initiation by blocking the interface of SL1 that binds the promoter, was used in phase I clinical trials starting in 2013 (Quin et al. 2014). It is 300-400 times more specific for RPI than for RPII or RPIII and was anti-proliferative in numerous cancer cell lines (Rebello et al. 2016). In mouse models of lymphoma and leukemia, CX-5461 induced a rapid p53-dependent apoptosis specifically in cancer cells (Bywater et al. 2012). Furthermore, it was found that CX- 5461 can activate ATM/ATR-mediated cell cycle checkpoints in the absence of DNA damage and this is independent of p53 activation (Quin et al. 2016). Due to its rapid inhibition of RPI transcription, CX-5461 treatment leads to an open chromatin structure of previously active rDNA repeats that are devoid of RPI, which activates ATM signaling. Combining CX-5461 treatment with inhibition of the damage response pathway improved therapeutic efficacy in treatment p53-null lymphoma (Quin et al. 2016). Therefore it is possible to target RPI transcription specifically in transformed cells, independently of p53.

Table 1.3 Anticancer drugs in use that inhibit rRNA synthesis

Drug Target Mode of action Reference Actinomycin D CG-rich duplex Intercalates in GC-rich regions of rDNA and (Fetherston et DNA inhibits at low concentrations elongation of Pol al. 1984) I transcription Cisplatin Adjacent Forms crosslinks in DNA that possess high (Jordan & guanosines in DNA affinity for HMG-containing proteins. Hijacks Carmo-Fonseca UBF from its site of action, thus inhibiting Pol I 1998) transcription Irinotican/Topotican Topoisomerase I Traps Topo I to rDNA leading to DNA strand (Pondarré et al. breaks and inhibition of Pol I transcription 1997) Mitomycin C Guanosines in 5’- Inhibits Pol I transcription by alkylating (Rey et al. 1993) CpG-3’ motifs guanosines and inducing interstrand crosslinks in rDNA 5-Fluorouracil Thymidylate Incorporation of 5-FU in 47S pre-rRNA inhibits (Ghoshal & synthase processing of pre-rRNA Jacob 1997) Temsirolimus mTORC1 Inhibits rRNA synthesis by interfering with (Mahajan 1994) mTORC1 activity

Reproduced with permission from Drygin D et al., Annu Rev Pharmacol Toxicol, 2010.

1.4.3 Viral infections Many viral proteins have been found to localize in the nucleolus upon infection and even interact with nucleolar factors. Some viruses use the nucleolus as their site of replication, such as the Borna disease virus, or for processing of viral RNA, such as HIV-1 (Hiscox 2002). Viral infection can perturb the nucleolar structure and function, as is the case with the adenovirus, which causes the relocalization of nucleolin and nucleophosmin/B23 (Matthews 2001). Adenovirus and coronavirus infections lead to the displacement of fibrillarin as well and in the case of adenovirus, RPI transcription is affected (Castiglia & Flint 1983, Puvion- Dutilleul & Christensen 1993). It is not precisely known why viral proteins are targeted to the nucleolus but it could be to use the nucleolus as a site for viral replication or to affect cellular transcription, RNA processing or the cell cycle (Hiscox 2002).

1.4.4 Neurodegenerative disorders Nucleolar stress, or decreased rRNA transcription and destructuration of the nucleolus, is one of the first signs of aging and neurogenerative diseases. Recently, the nucleolus has been implicated in some human neurodegenerative disorders such as Alzheimer’s, Parkinson’s and Huntington’s. In general, the size and structure of the nucleolus are often found modified postmortem in these diseases (Hetman & Pietrzak 2012, Rieker et al. 2011). Evidence suggests a reduction in NOR activity (Dönmez et al. 2005) and increased silencing of the rDNA in the early stages of Alzheimer’s disease (Pietrzak et al. 2011). Recently, a more specific link between Parkinson’s and the nucleolus was observed. Parkin interacting substrate (PARIS/ZNF746), a zinc-finger protein that is accumulated in the human brain in Parkinson’s disease patients, was found to associate with the rDNA, colocalizing with RPI (Kang & Shin 2015). Kang et al. also demonstrated an interaction between PARIS and Myb-binding protein 1α (MYBBP1A), which interacts with RPI transcription machinery and represses ribosome biogenesis, leading to cell cycle arrest (Hochstatter et al. 2012). MYBBP1A was also shown to be important for p53 activation under nucleolar stress (Kumazawa et al. 2015, Kuroda et al. 2011, Ono et al. 2014). These authors hypothesized that PARIS could be implicated in rRNA production and nucleolar stress that is a factor of Parkinson’s disease (Kang & Shin 2015).

Neurodegenerative diseases that manifest by the presence of aggregates such as Huntington’s and spinocerebellar ataxia have been shown to demonstrate these aggregates not only in the cytoplasm but also in the nucleolus, and they involve interactions with nucleolin (De Rooij et al. 1996, Wills & Atkins 2006).

1.5 Mouse embryonic development This section will briefly describe early preimplantation mouse development and what is known about the nucleolus in these first days of development.

1.5.1 From totipotent to pluripotent; the first days of development Embryogenesis occurs when a sperm cell fuses with an oocyte, called fertilization, to form a totipotent zygote that gives rise to all tissues of the embryo (Clift & Schuh 2013). Prior to fertilization, oocytes are stocked in the ovary, where they have undergone replication and homologous chromosome recombination in preparation for meiosis. Oocytes are suspended in prophase I while they grow and accumulate factors necessary for the embryo. Once fully grown and upon hormonal activation, the oocytes enter meiosis, dividing and segregating half of the homologous chromosomes into a small cell named the polar body that remains within the zona pellucida, or extracellular matrix surrounding the mature oocyte. The maturation of an oocyte in the mouse takes around 12 hours to complete (Clift & Schuh 2013). This fully grown mature oocyte, or egg, now pauses in metaphase II to await fertilization in the oviduct. Once the sperm fertilizes the egg, forming the one-cell zygote, the remaining maternal sister chromatids segregate and half form the second polar body. The zygote is therefore formed of a haploid maternal pronucleus, a haploid paternal pronucleus from the sperm and the two polar bodies (Figure 1.25).

Replication occurs rapidly in the zygote after fertilization and the two pronuclei migrate towards each other and fuse, called syngamy, at the start of the first mitosis, occurring 16-20 hours after fertilization (Marikawa & Alarcón 2009). The first cell divisions in the zygote are called cleavage divisions, as the size of the zygote does not change but rather the material is cleaved to form smaller cells named blastomeres. The first cleavage division gives the two-cell zygote, which is totipotent, capable of giving rise to all tissues of the embryo. After this stage, the blastomeres start becoming polarized and cell fate specification is more established with each division. The following four cleavage divisions each take around 12 hours to complete. At the 8-cell stage, the morphology of the embryo changes and the blastomeres adhere together, establishing tight junctions, forming the morula or a compact ball of cells (Figure 1.25). These first divisions are not always synchronous in the mouse and therefore embryos can display an uneven number of cells, particularly after the morula stage (Marikawa & Alarcón 2009).

At 3.5 dpc (days post coitum) or E3.5, the 32-cell embryo is now known as the blastocyst and is made up of two distinct cell populations and a fluid-filled cavity called the blastocoel. Contrary to other species, the initial differentiation of embryonic cells does not immediately give the three germ layers in mammals, but rather produces two cell types; the trophectoderm (TE) cells and the inner cell mass (ICM). The trophectoderm (TE) will form the placenta and the inner cell mass (ICM) is made up of pluripotent cells that will undergo a second segregation to give the epiblast (EPI) and the primitive endoderm (PE) at E4.5, when the embryo will start implantation (Gasperowicz & Natale 2011). It is from the ICM that one may derive embryonic stem cells (ESCs) for cell culture. The EPI layer will give rise to the embryo itself and the extraembryonic mesoderm whereas the PE will give rise to the extraembryonic visceral endoderm (VE) and the parietal endoderm (PaE)

(Gasperowicz & Natale 2011). Post-implantation, the EPI layer differentiates into the three germ layers (endoderm, mesoderm, ectoderm) that will give rise to the embryo (Marikawa & Alarcón 2009). During these first hours of preimplantation development, many changes occur in the embryo, of which the development of the nucleolus and the regulation of chromatin will be discussed further.

Figure 1.25 Preimplantation development in the mouse (A) After ovulation occurs, eggs are fertilized in the oviduct to form the zygote. The zygote undergoes several mitotic divisions until becomes a morula after compaction of the cells. A cavity is formed in the embryo and cells differentiate to become the inner cell mass and the trophoectoderm. The embryo is now a blastocyst which will implant in the uterine wall. (B) The process of oocyte maturation and the first mitosis. Reprinted with permission from Clift D and Schuh M., Nat Rev Mol Cell Biol, 2013.

1.5.2 Maternal-to-zygotic transition An important aspect of preimplantation development is the maternal-to-zygotic transition (MZT), which is characterized by the degradation of a significant amount of maternal mRNA and the activation of zygotic transcription (Tadros & Lipshitz 2009). During oogenesis, the growing oocyte accumulates transcripts for the zygote and then remains transcriptionally and metabolically inactive until the zygote machinery takes over the cell’s processes. A portion of the maternal mRNA is degraded rapidly after fertilization and then another large quantity of transcripts is degraded at the two-cell stage (Alizadeh et al. 2005, Hamatani et al. 2004, Piko &

Clegg 1982). Some mRNAs can even be degraded to leave only ~10% remaining by the two-cell stage (Paynton et al. 1988, Schultz 1993).

In mouse, this degradation corresponds with the two waves of zygotic genome activation (ZGA), the first starting as early as the S/G2 phase of the one-cell stage and the second major activation occurring at the two- cell stage (Figure 1.26). Studies have demonstrated some transcription at the one-cell stage, either by BrUTP incorporation in the male pronucleus in one-cell zygotes (Aoki et al. 1997, Bouniol et al. 1995) or luciferase reporter gene activity (Nothias et al. 1996) but this is likely negligible (Hamatani et al. 2004). Therefore, embryos require their proper zygotic transcripts at the late two-cell stage, up to which point they are still dependent on the maternal mRNAs.

Interestingly, using a reporter gene under control of the rDNA promoter, RPI transcription was also shown to occur at the one-cell stage (Nothias et al. 1996). However, another study observed transcription of the rDNA only at the late two-cell stage by BrUTP incorporation in nucleolar precursor bodies (NPBs) by immunofluorescence (Zatsepina et al. 2003). Using AgNOR staining, Engel et al. also demonstrated a two-cell stage activation (Engel et al. 1977). However, treating embryos with actinomycin D (ActD), an RPI inhibitor arrested cleavage divisions from 1-cell onwards (Golbus et al. 1973). More recently, by specifically inhibiting RPI transcription with cancer drug CX-5461 (Bywater et al. 2012), Lin et al. observed a one-cell stage developmental arrest (Lin et al. 2014). Evidently, some zygotic activity at the one-cell stage is essential, in particular rRNA transcription, however, the major zygotic gene activation occurs at the two- to four-cell transition.

Figure 1.26 Gene expression regulation during early development

Figure legend on next page

During growth, mRNA levels are increased to sustain the oocyte during the transcriptionally silent phases of maturation and fertilization. The RNA-binding protein MYS2 stabilizes mRNAs and CPEB inhibits translation of maternal mRNAs by binding the 3’ untranslated region (UTR). During maturation, these mRNAs are either translated and/or degraded through mechanisms involving DAZL and the phosphorylation of MYS2 and CPE. Maternal mRNA levels are minimal by the two-cell stage. Zygotic genome activation initiates early on but has a robust activation at the 2-cell stage. This switch between maternal and zygotic regulation involves major chromatin remodelling by SWI/SNF and ISWI. Reprinted with permission from Clift D and Schuh M., Nat Rev Mol Cell Biol, 2013.

1.5.3 Chromatin regulation of maternal and paternal genomes Major changes in nucleus structure have been shown to occur during development, cellular reprogramming, and disease development. In fact, between fertilization and implantation, the mouse embryo undergoes a drastic nuclear reorganization, coinciding with epigenetic reprogramming and the first cell fate decisions for differentiation. While imaging embryos stained with DAPI, one may track these architectural changes that are largely due to the movement of AT-rich satellite sequences that stain strongly (Borsos & Torres-Padilla 2016).

Sperm DNA is not nucleosomal but rather surrounded by small DNA-binding proteins called protamines that condense and silence the DNA, hence it needs to be reassembled into chromatin so that development may progress. Upon fertilization, the sperm DNA is decondensed and then assembled by stored maternal histones prior to the formation of the pronuclei. The two pronuclei differ in their histone variants, histone marks and DNA methylation (Zhou & Dean 2015). For instance, the maternal histones that assemble on the male DNA are more acetylated than the histones in the maternal pronucleus (Adenot et al. 1997, Borsos & Torres-Padilla 2016, Nonchev & Tsanev 1990). The maternal and paternal DNA are differentially transcribed in mouse zygotes, and this is regulated by the different acetylation status (Ura et al. 1997). Interestingly, inducing hyperacetylation of H3 with trapoxin, a histone deacetylase (HDAC) inhibitor, prematurely increases transcription in 2-cell embryos (Aoki et al. 1997). Therefore, the regulation of ZGA appears to necessitate specific chromatin modifications during the first two cell divisions.

In fact, one major structural change that occurs in zygotic nuclei is centromere reorganization between the two-cell and four-cell stages. The chromatin associated with the AT-rich minor and major satellite repeats of the centromere is named centromeric and pericentric heterochromatin respectively. These can form chromocenters when pericentric and centromeric sequences between different chromosomes group together. These chromocenters are reorganized in mature oocytes when the pericentric repeats form rings around the nucleolar precursor bodies (NPBs), which will be discussed in the following section (Zuccotti et al. 2005).

As another example of chromatin regulation in the zygote, the histone variant H3.3, which is a marker of open chromatin, is commonly found in the paternal pronucleus (Ahmad & Henikoff 2002, Loppin et al. 2005, Torres-

Padilla et al. 2006, van der Heijden et al. 2005) and is important for maintaining the chromatin state during the first cleavage divisions in preimplantation mouse development (Lin et al. 2013). The maternal histone chaperone Hira (histone cell cycle regulation defective homolog A) is essential for male pronucleus formation by regulating sperm DNA H3.3 nucleosome assembly. While studying Hira mutants, these authors found that RPI transcription, but not RPII, was affected in these zygotes. Therefore it appears that Hira/H3.3-dependant transcription of the rDNA is essential for the first cleavage division (Lin et al. 2014). Another H3.3 chaperone, Daxx/ATRX, was implicated in centromere chromatin regulation, though the mechanism was not established (Fulka & Langerova 2014). There are still many details that remain to be investigated but it is clear that H3.3 is an important factor for the assembly of chromatin after fertilization and the reorganization of centromeric and pericentric chromatin in the embryo (Probst et al. 2010, Santenard et al. 2010).

Another major epigenetic factor in embryo viability is DNA methylation at CpG nucleotides, which is involved in gene silencing, dosage compensation, gene imprinting, X chromosome inactivation and silencing CpG-rich retrotransposons providing genome stability (Santos et al. 2002). Due to the high risk of spontaneous deamination of 5-methyl-cytosine (5mC) to thymidines, there is an underrepresentation of CpGs in the genome (Monk 2015). However, there exists CpG islands which are regions with a high percentage of CpGs but very little methylation, often associated with promoters of so-called housekeeping genes (Deaton & Bird 2011). In the early embryo, there is an important genome-wide demethylation starting in the zygote, reaching the low point in the blastocyst (Howlett & Reik 1991, Monk et al. 1987). About 85% of the sperm genome is methylated but then is actively demethylated in the zygote prior to the first cell division, whereas the oocyte genome has about 40% of CpG sites methylated, undergoes a more gradual passive demethylation (Kobayashi et al. 2012, Mayer et al. 2000, Oswald et al. 2000, Santos et al. 2002). The passive demethylation of the maternal DNA in the zygote is due to the lack of the maintenance methylase Dnmt1 (Carlson et al. 1992, Howlett & Reik 1991). This passive demethylation affects housekeeping genes and repeated sequences but imprinted genes maintain their methylation status. Post blastocyst implantation, methylation is reestablished by the de novo methylases Dnmt3a and -3b and preserved in the somatic cells (Monk et al. 1987, Okano et al. 1999).

As discussed in this section, the regulation of the epigenetic landscape in zygotes and embryonic stem cells is important for the establishment of pluripotency and subsequently cell-fate specification.

1.5.4 The developing nucleolus Unlike somatic cells, full-grown oocytes and early embryos do not have nucleoli as we know them but rather large round bodies named nucleolus precursor bodies (NPBs). NPBs are not tripartite like somatic nucleoli but are more compact, homogenous and mainly composed of proteins (Fléchon & Kopecny 1998, Fulka & Aoki 2016, Fulka et al. 2012, Shishova et al. 2015b). As growing oocytes have somatic-like nucleoli, the NPBs form

from nucleoli throughout oocyte maturation and this corresponds with the reorganization of the centromeric chromatin (Chouinard 1971). This also occurs in conjunction with the shutdown of rRNA transcription, with the last sites of RPI transcription occurring on the surface of the NPB in the maturing oocyte (Fulka & Aoki 2016). The NPB progressively changes to typical nucleoli in the zygote throughout the first few cleavage divisions up until the blastocyst stage. Interestingly, embryonic developmental potential can be somewhat predicted from the number and arrangement of NPBs in pronuclei (Gianaroli et al. 2003, Tesarik & Greco 1999).

Morphological changes to the NPB in the zygote commence around the same time as the degradation of maternal RNA occurs just after fertilization (Piko & Clegg 1982). Throughout the first cell divisions, the number of NPBs decreases and around the eight-cell stage the typical fibrillar and granular components of the somatic nucleoli start forming around this body (Aguirre-Lavin et al. 2012, Fléchon & Kopecny 1998). Ribosomal RNA production starts to occur on the periphery of the NPBs during these changes, though a recent study on the use of different fixatives on embryos during immunofluorescence protocols show contradictory results for the location of the rRNA and of some factors associated with rRNA transcription (Fléchon & Kopecny 1998, Shishova et al. 2015a).

Logically, it was assumed that the role of the NPB is specific to ribosome biogenesis, however, studies have elucidated other roles for this nuclear body. One technique used to study the function of the nucleolar precursor body is termed enucleolation, which is the microsurgical removal of the NPB (Fulka et al. 2003). This technique has shown that NPBs are coming from the oocyte and not sperm cells and they are essential for development as its removal cause embryos to arrest development at the two-cell stage (Ogushi et al. 2008). The simple explanation for this arrest is that RPI transcription does not occur and therefore neither does ribosome biogenesis. This is certainly the case in Xenopus, where a mutant lacking nucleoli is not able to produce de novo rRNA during embryogenesis (Brown & Gurdon 1964), however, in mouse it appears to be more complicated. In fact, neither pre-rRNA synthesis nor processing seems to be affected in zygotes developing from enucleolated oocytes, even though they arrest development at the two-cell stage. Fulka and Langerova have demonstrated that the oocyte does not accumulate nucleolar proteins nor their mRNAs (Fulka & Langerova 2014). Maternal nucleophosmin (NPM/B23), nucleolin (Ncl/C23), fibrillarin (Fbl) and UBF mRNA levels decreased throughout oocyte growth and were decreased even further after fertilization. The lowest level of transcripts was observed at the two-cell stage, when zygotic genome activation initiates. In addition, other than NPM, there was a decrease in protein level of these factors from early to mature (metaphase II) oocytes. Therefore, these authors postulate that the embryo does not need a stockpile of the maternal nucleolar proteins or mRNAs in the NPB, as was previously proposed.

Recent studies propose a role for the NPB in regulating the major and minor centromeric satellite sequences, whereas dysregulation of these sequences leads to chromocenter organization defects and developmental arrest in embryos (Fulka & Aoki 2016, Probst et al. 2010). Embryos developing from enucleolated oocytes, or nucleolus-less embryos, show a delay in their first cell division as compared to control embryos and have an altered chromatin structure (Fulka & Langerova 2014, Ogushi & Saitou 2010). It was found that there was a 20% decrease in the amount of major and minor satellite DNA during the first cell cycle in nucleolus-less embryos. Thus, using enucleolated oocytes, a role for the NPB in the organization and remodeling of pericentric and centromeric satellite repeats has been elucidated, though specific mechanisms have yet to be determined.

Interestingly, loss of UBF in mouse embryonic fibroblasts (MEFs) leads to a disorganization of the nucleolus and nucleolar factors, which either collapse into NPB-like foci (e.g. RPI, Fibrillarin) or are distributed to the nucleoplasm (e.g. NPM, MDM2) (Hamdane et al. 2014, 2016). The rDNA is also affected and is localized to the pericentromeric heterochromatin in UBF-null cells. Ubf-/- embryos arrest development prior to blastocyst stage and display irregular heterochromatin organization. These embryos lack a nucleolar precursor body, which could indicate that the presence of UBF or rRNA transcriptional activity is necessary for nuclear organization in embryonic development (Figure 1.27) (Hamdane et al. 2016). In accordance with this, another recent study followed UBF and processing factor Nopp140 throughout preimplantation development and found that RPI transcription was linked to the structural organization of the developing nucleus (Kone et al. 2016). As UBF is required for the formation of active rDNA chromatin in somatic cells, it could also have an essential role in regulating the heterochromatin architecture in early embryos (Hamdane et al. 2016).

Figure 1.27 UBF is essential for NPB formation in early embryos Ubf-null embryos arrest development prior to the 16-cell stage and fail to form nucleolar precursor bodies. Instead, heterochromatin is disorganized and the RPI machinery is coondensed into foci. Reprinted with permission from Hamdane N et al., Gene, 2016.

Thesis objectives

The main objective of my doctoral research was to investigate the role of RRN3 and UBF in ribosomal RNA transcription and chromatin regulation in vivo. Both these factors have been implicated in initiation and both are regulated by growth signaling pathways to activate ribosomal RNA transcription. Previous studies had shown that the knockout of Rrn3 in mouse led to an late developmental arrest at E9.5 (Yuan et al. 2005). This result was surprising as RRN3 is an essential gene in yeast and Ubf knockout in mouse arrests development preimplantation at E3.5 (Hamdane et al. 2014). Therefore, in order to better understand the role of these two factors in vivo, we felt it was important to analyze UBF and RRN3 in parallel. In order to accomplish this task, conditional knockout mice were generated in the laboratory and isogenetic cell lines were created to be able to compare these factors in as similar conditions as possible. Using these models, this thesis addresses the following objectives;

1. Confirm the loss of Rrn3 by reanalyzing the knockout mice and study the loss of Rrn3 in MEFs in order to compare with results previously obtained while studying the Ubf knockout. (Chapter 2)

2. Investigate the roles of RRN3 and UBF in pre-initiation complex formation in vivo as well as the role of UBF in the regulatory region of the rDNA. (Chapter 2)

3. As these factors are essential, study the ultimate effects of the loss of Ubf and Rrn3 on cell growth, division and viability. (Chapter 3)

4. Generate a model by which to study the role of UBF, RRN3 and more generally, rRNA transcription in pluripotency maintenance. (Chapter 4)

Chapter 2 An enhancer adjacent chromatin boundary is maintained on the ribosomal RNA gene repeats even in the absence of the basal factors and active transcription

Chelsea Herdman1,2,*, Jean-Clement Mars1,2,*, Victor Y. Stefanovsky1, Michel G. Tremblay1, Marianne Sabourin-Felix and Tom Moss1,2.

* Equal first authors. 1Laboratory of Growth and Development, St-Patrick Research Group in Basic Oncology, Cancer Division of the Quebec University Hospital Research Centre. 2Department of Molecular Biology, Medical Biochemistry and Pathology, Faculty of Medicine, Laval University.

Keywords RNA Polymerase I (RPI, PolI, POLR1), RPI associated factor RRN3/TIF1A, Upstream Binding Factor (UBF/UBTF), Ribosome Biogenesis, Ribosomal RNA (rRNA) genes,

2 An enhancer adjacent chromatin boundary is maintained on the ribosomal RNA gene repeats even in the absence of the basal factors and active transcription

2.1 Preface Until recently, little was understood about the regulation of the promoter and enhancer elements of the ribosomal RNA gene repeat in vivo. We have generated tools in the laboratory that allow the investigation of this region. A main aspect of my PhD was the generation and characterization of conditional knockout cell lines for RRN3 and their comparison to the UBF conditional cell lines generated by my colleague Nourdine Hamdane. Using the two cell lines, we were able to study the loss of UBF or RRN3 and directly compare the two. Importantly, with the development of a normalization protocol for chromatin immunoprecipitation (ChIP)- sequencing data by MSc student Marianne Sabourin-Félix, we were able to analyze pre-initiation complex formation and the chromatin landscape of the rDNA at a resolution not yet seen for the rDNA.

Prior to my arrival in the laboratory, Nourdine Hamdane, Michel Tremblay and Victor Stefanovsky had analyzed the loss of UBF in mouse and found it to be embryonic lethal at E3.5 even though UBF was previously thought to be non-essential. UBF was also found to be required for nucleolar structure and for PIC formation as shown by ChIP-qPCR (Hamdane et al. 2014). Further studies of the UBF knockout are described in Chapter 3 of this thesis.

In this study, we directly compared the roles of UBF and RRN3 using ChIP-sequencing and chromatin structure studies. To make this possible, I had first to reinvestigate the knockout of RRN3 in mice having a mixed background and after back-crossing to the BL/6 background. Michel Tremblay helped with mouse line maintenance and crosses. I performed the embryonic lethality study on the RRN3 heterozygous mice, discovering that, contrary to previous studies, Rrn3 was in fact essential for early embryonic development. Also, I analyzed crosses of UBF+/-RRN3+/- mice as many diseases involving ribosome biogenesis factors are due to haploinsufficiency. I created MEFs conditional for RRN3 using the same ER-Cre and p53 mouse lines that were used for UBF, hence generating a directly comparable model. After optimizing tamoxifen treatment conditions for gene excision, I examined the loss of RRN3 in the same way as for UBF. As co-first authors, Jean-Clément Mars and myself are responsible for the analysis of these cell lines by ChIP-seq, where he focused on UBF loss. Victor Stefanovsky is the author of Figure 4 and Supplementary Figure 6, where I helped with cell preparation for the experiments. Tom Moss and Marianne Sabourin-Félix did the bioinformatic analysis.

2.2 Résumé La transcription de l’ARN ribosomal, l’étape limitante de la biogénèse des ribosomes, est accomplie par l’ARN polymérase I (RPI) ainsi qu’un ensemble spécifique de facteurs généraux de transcription, UBF, SL1, RRN3 et TTF1. Nous avons établi des lignées de souris mutantes nous permettant de générer conditionnellement des mutants de perte de fonction de UBF et RRN3 afin d’étudier la régulation de la transcription de l’ADN ribosomal. Nous avons utilisé ces lignées pour comprendre la hiérarchie des facteurs de recrutement et des modifications de la chromatine au cours du transcription de l’ADNr. Nous avons démontré que UBF établit une chromatine spécialisée qui remplace les histones et les nucléosomes sur l’ensemble du gène de l’ADNr. Cette région est délimitée en amont par le “spacer promoter” et en aval par les sites de terminaison T1-T10 qui sont reconnus par TTF1. La perte de fonction de Rrn3 bloque le recrutement de RP1 et la transcription, mais n’affecte par les facteurs de pre-initiation SL1, UBF et TTF1. En comparaison, la perte de fonction de UBF empêche le recrutement de RP1, RRN3 et SL1, sans affecter TTF1. Une frontière en aval formée par H3K4me2, H3K4me3, H3K9ac, H2A.Z and H2A.Zac, se retrouve immédiatement en amont du «spacer promoter» et cette frontière n’est pas affectée par la perte de UBF étant indépendante de la transcription. En résumé, l’ADNr est délimité par une zone frontière constituée en amont de marques de la chromatine active, du «spacer promoter» et de TTF1 et d’un élément en aval formé par TTF1 qui est lié aux multiples sites de terminaison T1-T10.

2.3 Abstract Ribosomal RNA transcription, the initial and rate-limiting step of ribosome biogenesis, is accomplished by RNA polymerase I (RPI) and a set of specialized basal transcription factors, UBF, SL1, RRN3 and TTF1. We have established conditional knockout mice for UBF and RRN3 in order to study the regulation of rDNA transcription. As a preliminary to these studies, we found that, like UBF, the RRN3 gene is essential for mouse development beyond the morula stage. We then used cells derived from these mice to understand the hierarchy of factor recruitment and chromatin modification in rDNA activity. We have found that UBF establishes a specialized chromatin that replaces core histones and nucleosomes throughout the rDNA gene body. This region is delimited upstream by the spacer promoter and downstream by the T1-T10 termination sites and TTF1 binding at both these sites. Inactivation of Rrn3 prevents RPI recruitment and transcription, but has no effect on preinitiation factor SL1, UBF or TTF1 binding. Inactivation of Ubf prevents RPI, RRN3 and SL1 recruitment, but again has little effect on TTF1. An upstream boundary formed by H3K4me2, H3K4me3, H3K9ac, H2A.Z and H2A.Zac is found immediately upstream of the spacer promoter, which is occupied by a pseudo-stable arrested RPI complex. This upsteam boundary is unaffected by UBF loss and hence is also independent of RPI transcription. Thus, the functional rRNA gene domain is delimited by an upstream boundary consisting of active chromatin markers, the spacer promoter and TTF1, and a downstream boundary delimited by TTF1 at the multiple T1-T10 termination sites.

2.4 Introduction The initial step of ribosome biogenesis, transcription of the ribosomal RNA, is a highly regulated and essential process for cell growth. Synthesis of the rRNA accounts for over half of total RNA production in growing cells (Moss et al. 2007). The rRNA is transcribed from the rDNA, which is found as several hundred copies arranged in head-to-tail fashion in mammalian cells. The ~45 kb rDNA mouse repeat is formed of the rRNA coding sequence (~13-14 kb) and the intergenic spacer (IGS), which contains many repetitive sequences. The IGS also contains the regulatory elements of the rDNA, such as enhancer repeats, the 47S promoter (upstream control element (UCE) and core promoter) and the spacer promoter (SpPr) (Grozdanov et al. 2003, Kuhn & Grummt 1987, Moss et al. 2007). Only around half of the rDNA repeats are transcriptionally active within a given cell and these display low levels of CpG methylation and are potentially associated with active histone modifications H3K4me3 and H3K9ac in the promoter region. In contrast, the inactive portion of the rDNA genes are highly methylated and display repressive chromatin marks such as methylated H3K9, H3K27, and H3K20 (McStay & Grummt 2008).

The rDNA is transcribed by RNA polymerase I that has a set of dedicated transcription factors that form the pre-initiation complex (PIC) (Moss et al. 2007, Russell & Zomerdijk 2005). SL1 and UBF bind the promoter and allow RRN3 to recruit the active polymerase to the rDNA to initiate transcription (Moss et al. 2007). In vitro, SL1, formed of TBP and the TBP associated factors (TAFs), is essential for RPI transcription and it binds both the 47S as well as the spacer promoters. UBF binds the promoter regions, but also binds through the gene body (Diesch et al. 2015, Hamdane et al. 2014). However, the role of UBF in rRNA transcription was unsure as some studies found it to be non-essential for transcription (Smith et al. 1993), whereas others showed it was a transcriptional activator (Kuhn & Grummt 1992, Panov et al. 2006a). Interestingly, knockdown of this factor did not affect rRNA production but rather the number of active rDNA repeats (Sanij et al. 2008). However, recent studies have shown that UBF is indeed an essential protein for mouse development, rRNA transcription and for nucleolar structure (Hamdane et al. 2014). RRN3 is also an essential factor, as is its yeast homolog Rrn3p (Yamamoto et al. 1996, Yuan et al. 2005).

In addition to the proteins involved in initiation, TTF1 (transcription termination factor 1) is an important factor for RPI transcription. TTF1 arrests transcription downstream of the rRNA repeat at the T1-T10 terminator Sal- box sites, but it is also found at two upstream terminator sites (T0 and TSP). TTF1 has been implicated in rDNA silencing through recruitment of the nucleolar remodeling complex (NoRC) to this regulatory region via interaction with TIP5 (TTF1-interacting protein 5) and a non-coding RNA named pRNA (promoter homologous RNA) (Mayer et al. 2006, Savić et al. 2014, Schmitz et al. 2010).

Though the basal factors involved in RPI transcription are known, the relative requirements for UBF or SL1 in PIC formation and RPI-RRN3 recruitment are still debated. Quantitative PCR studies have identified RPI basal factor occupancy at specific locations throughout the rRNA gene sequence however detailed studies of the chromatin landscape and RPI factor occupancy of the rDNA are rare due to the difficulty of sequencing this repetitive region (Hamdane et al. 2014, Sanij et al. 2015, Zentner et al. 2011, 2014). Further, the rDNA repeat is not included in existing genome assemblies and so most publically available genome-wide analyses neglect to study the rDNA.

We have developed conditional mouse cell lines for UBF and RRN3 in order to study the requirement of these factors for rRNA transcription in vivo. Most recently, we have used these to understand the hierarchy of factor recruitment and chromatin modification in rDNA activity. We show that preinitiation complex formation absolutely requires UBF but not RRN3. We further define an upstream spacer promoter adjacent chromatin boundary that includes H2A.Z, H3K4me3 and TTF1 that is independent of rDNA transcriptional activity.

2.5 Results

2.5.1 A high-resolution map of basal factors across the rRNA genes We recently developed a normalization procedure for rDNA ChIP-Seq data that greatly reduces inherent noise in protein interaction maps and greatly enhances their quality (Sabourin-Félix M et al., in preparation). When applied to ChIP-Seq data for the mouse basal RPI transcription machinery, the procedure provided high- resolution maps of all the RPI transcription factors on the rDNA of Mouse Embryonic Fibroblasts (MEFs). The data clearly identified the promoter-bound pre-initiation complexes, engagement of both the initiation competent and elongating forms of RPI, and TTF1 bound at the multiple transcription termination sites (Figure 2.1). Engagement of RPI was found to be near uniform throughout the 47S transcribed region, attesting to the density of rDNA transcription. These data strongly suggest that in proliferating MEFs the 47S region contains no major sites of polymerase pausing (Figure 2.1A and B). In contrast, the RPI associated initiation factor RRN3/TIF-IA showed an interaction only over the first 2 to 3 kb of the transcribed region, consistent with in vitro data showing that RRN3 is released after initiation (Milkereit & Tschochner 1998, Schnapp et al. 1993). The RRN3-RPI interaction decayed exponentially with increasing distance from the initiation site, suggesting that its release is stochastic (Figure 2.1B). Assuming a mean elongation rate of 60 nt.sec-1 (Stefanovsky et al. 2006a), we estimate the half-life of the elongating RPI-RRN3 complex as around 15 s in rapidly proliferating MEFs. Thus, RRN3 acts mechanistically much like the Sigma factors of eubacteria that target the polymerase to promoters but are released during elongation (Bar-Nahum & Nudler 2001, Mukhopadhyay et al. 2001, Paget & Helmann 2003). Indeed, like bacterial Sigma, RRN3 was recently shown to contain a DNA interaction domain that is required for RPI initiation (Stepanchick et al. 2013).

1 RPI -

3 RRN3/

0 m TIF1A

E UBF

6 TAF68

- TTF1

IGS 18S 5.8S 28S IGS

5 RPI

0 RRN3/

n - TIF1A

2 UBF r 100

0 %GC

d l 40

6 F TAF68

0 TTF1

47S 18S 5.8S 28S Sp Pr/Tsp T0/47SPr T1-10 -2000 1 4000 8000 12000 bp

3 TAF95

e 6 TAF68

2 E TBP

l 0 Spacer Pr 47S Pr

2 UBF

Enhancer Repeats Sp Pr/Tsp T0/47SPr

Figure 2.1 RNA polymerase I basal factor occupancy across the rDNA (A) RPI, RRN3, UBF, TAF68 and TTF1 ChIP-seq performed in untreated UBF+/+ER-Cre+/+p53-/- MEFs was aligned to a concensus unit of the murine rRNA gene and visualized using the Integrative Genomics Viewer (IGV). The rDNA repeat is represented below the histograms and the scales are shown in fold enrichment over the input on the left hand side of the graphic. (B) Zoom of the functional rDNA unit with % of GC content added in grey over the UBF track. (C) Focus on the two promoters showing overlap of two TAFs (95 and 68) with TBP and UBF. 33

2.5.2 UBF binding is precisely delimited to the functional rDNA unit Various in vitro studies have implicated UBF in the formation of the RPI preinitiation complex (Goodfellow & Zomerdijk 2012, Grummt 2003, Moss & Stefanovsky 2002, Moss et al. 2007). But it has also been shown to bind over a wide region of the rDNA repeat, suggesting a role more in line with a component of chromatin than an initiation factor (Hamdane et al. 2014, O'Sullivan et al. 2002, Sanij et al. 2015, Zentner et al. 2014). Our data now show that in fact UBF performs both these roles. UBF was mapped throughout the functional rRNA gene unit and was bounded by two flanking DNA elements, upstream by the spacer promoter (SpPr) and downstream by the transcription termination sites (Sal-boxes) bound by Transcription Termination Factor 1 (TTF1). Consistent with its very low sequence selectivity (Bazett-Jones et al. 1994, Copenhaver et al. 1994, Stefanovsky & Bazett-Jones 1996), UBF bound almost continuously throughout the 47S transcribed region, but did display a variability that closely correlate with high G+C (above 50%) DNA content, (Figure 2.1B). This binding profile probably represented the true UBF occupancy and was not due to cross-linking or sequencing biases, since it was not observed for RPI. Further, the correlation with high G+C did not hold for the 47S and spacer promoter sequences. Both promoters displayed UBF binding that peaked in low G+C sequences and overlapped SL1/TIF-IB TBP-complex binding (TAF1B, -1C and TBP, Figure 2.1C). Further, no UBF at all was detected in the relatively G+C neutral intergenic spacer (IGS) (Supplementary Figure 2.8). Thus, it is very unlikely that recruitment of UBF to the rRNA genes is determined by DNA sequence selectivity, though once recruited, exact UBF positioning may be affected by underlying DNA sequences.

Since UBF is essential for rDNA activity (Hamdane et al. 2014), we sought to understand what, other than DNA sequence selectivity, could define the extent of its binding and hence the formation of active rDNA chromatin. To do this we first determined whether RPI transcription itself was necessary to maintain the active UBF-bound rDNA chromatin.

2.5.3 Mouse RRN3/TIF-1A is an essential protein RRN3/TIF-IA reversibly associates with RPI to form the initiation competent form of the polymerase (Milkereit & Tschochner 1998, Moss et al. 2007, Schnapp et al. 1993). Since its ortholog is essential for RPI transcription in yeast (Yamamoto et al. 1996), its elimination should also specifically abolish RPI transcription in mouse. However, previous data had suggested that mouse RRN3/TIF-IA might not in fact be essential in the early embryo (Yuan et al. 2005). Thus it was essential to resolve this question before RRN3 could be considered as a reliable target for specific RPI inactivation.

Deletion of the unique mouse Rrn3/TIF-IA gene was reported to arrest development at around E9.5, long after the normal onset of rRNA synthesis at the 2-cell stage (Yuan et al. 2005). In contrast, deletion of other RPI basal factors, Ubf and Polr1b (RPI subunit RPA135), or the early rRNA processing factor, Fbl (Fibrillarin),

arrested development at the morula or very early blastula stage (Chen et al. 2008, Hamdane et al. 2014, 2016; Newton et al. 2003). Thus, the existing data suggested that either RRN3 was not essential and was partly redundant with another factor. When we reanalyzed progeny from Rrn3/TIF1A+/- mice derived from mice carrying the original RRN3/TIF-IA floxed allele (Yuan et al. 2005), (Supplementary Figure 2.9A and 2.10D), we found that null embryos arrested during cleavage divisions as un-compacted morulae. The same result was obtained after extensive backcrossing to C57BL/6, the mouse strain used in the original publication (Supplementary Figure 2.10A and B). We conclude that RRN3, like UBF and RPI, is essential in mouse soon after the normal onset of rRNA gene activity (Engel et al. 1977). This strongly suggests that the mammalian RRN3 is indeed an essential part of the RPI transcription machinery.

As ribosomal protein genes are found to be haploinsufficient in many diseases (Danilova & Gazda 2015), so named ribosomopathies, we were curious whether heterozygous Rrn3+/- mice showed a phenotype. However, as is the case with Ubf+/- mice, no phenotype was observed. We even crossed the Ubf+/- with Rrn3+/- mice in order to obtain double heterozygotes Ubf+/-Rrn3+/-. The progeny of these double mutants were analyzed but no haploinsufficiency was found (Supplementary Figure 2.10C).

Since Rrn3 is a very early embryonic lethal gene, the RRN3 floxed mice were, therefore, crossed in order to generate Rrn3fl/fl/ER-Cre+/+/p53-/- and isogenic Rrn3+/+/ER-Cre+/+/p53-/- embryos from which conditional immortalized MEFs were isolated (Materials and Methods).

Figure 2.2 RRN3 is essential for RNA polymerase I recruitment to the rDNA (A) RRN3, RPI, UBF, TAF68, TBP and TTF1 ChIP-seq performed in untreated and tamoxifen treated Rrn3fl/flCreER+/+p53-/- MEFs was aligned to a concensus unit of the murine rRNA gene and visualized using the Integrative Genomics Viewer (IGV). Every second track indicates the knockout condition (Rrn3-). The rDNA repeat is represented below the histograms and the scales are shown in fold enrichment over the input on the left hand side of the graphic. (B) Zoom of the promoters and terminator regions.

2.5.4 Neither RRN3 nor RPI transcription is required to maintain the potentially active state of the rRNA genes We previously demonstrated that a short, low dose treatment of Ubffl/fl/ER-Cre+/+ MEFs with tamoxifen (4-HT, 50nM/4h) provided an efficient manner in which to specifically eliminate UBF in cell culture (Hamdane et al. 2014). Rrn3fl/fl/ER-Cre+/+/p53-/- and isogenic Rrn3+/+/ER-Cre+/+/p53-/- embryos were generated from immortalized MEFs conditional for RRN3. However, similar treatment of Rrn3fl/fl/ER-Cre+/+ cells provided only a partial gene inactivation, even if the treatment was longer and a higher dose (500nM/48h) (Supplementary Figure 2.10E). We found that this effect was inherent to the Rrn3fl/fl allele since the LoxP site sequences were verified (Supplementary Figure 2.10D) and it was independent of the source embryo, the ER-Cre transgene used and the p53 status, possibly explaining some of the difficulties encountered previously (Yuan et al. 2005). Even the use of the ER-CreT2 transgene, which has a 3 to 4-fold higher affinity for 4-HT (Feil et al. 1997, Indra et al. 1999), or the production of clonal cell lines did not result in efficient excision (Supplementary Figure 2.10F). To overcome this limitation, we developed a regime of repeated low dose 4-HT pre-treatments that allowed the complete inactivation of the Rrn3 gene and reduced the RRN3 protein level by ~90%, while having no effect on RRN3 levels in the Rrn3+/+ isogenic control cells (Supplementary Figure 2.9B, 2.10G and Materials and Methods). Concomitantly, de novo synthesis of the 47S precursor RNA, as determined by metabolic labelling (Stefanovsky & Moss 2016), was also strongly suppressed (Supplementary Figure 2.9F).

High resolution mapping of factor binding across the rDNA repeat in the Rrn3fl/fl/ER-Cre+/+/p53-/- isolated MEFs was indistinguishable from that seen in the wt MEFs (Figure 2.1). Near loss of RRN3 protein after 4-HT treatment of these MEFs essentially eliminated its engagement with the 5’ of the 47S transcribed region and strongly suppressed recruitment of RPI throughout the 47S region and at the spacer promoter (SpPr) (Figure 2.2A and B, and Supplementary Figure 2.11A). In contrast, RRN3 loss had no effect on UBF, its engagement remaining unaffected throughout the 47S region, the enhancer repeats and at the spacer promoter. Further, binding of the SL1 complex (TAFI68 and TBP) at both 47S and spacer promoters was unaffected by RRN3 loss, and TTF1 remained bound to the adjacent Tsp and T0 sites and to the T1-T10 47S termination motifs. Parallel 4-HT treatment of isogenic Rrn3+/+/ER-Cre+/+/p53-/- MEFs had no effect on either RRN3 or RPI recruitment, or indeed on any of the other factors in this study (data not shown). Thus, RRN3 and RPI engagement on the rDNA is not required for the establishment of the preinitiation complexes at both spacer and 47S promoters, or for the normal pattern of UBF binding.

These data demonstrated that the potentially active state of the rRNA genes is stably maintained over extended periods of gene inactivity. The long-term stability of UBF binding was particularly surprising given its low DNA binding constant (Kd ~ 10nM) (Leblanc et al. 1993) and high off-rate (t1/2 9 to 25 s) (Chen & Huang 2001) and its inability to compete directly with nucleosome formation. This suggested that a mechanism exists

to maintain the positioning of UBF and hence the potentially active state of the rRNA genes independently of RPI activity.

Figure 2.3 UBF is necessary for RPI machinery recruitment

(A) UBF, RPI, RRN3, TAF95, and TTF1 ChIP-seq performed in untreated and tamoxifen treated Ubffl/flCreER+/+p53-/- MEFs was aligned to a concensus unit of the murine rRNA gene and visualized using the Integrative Genomics Viewer (IGV). Every second track indicated the knockout condition (Ubf-). The rDNA repeat is represented below the histograms and the scales are shown in fold enrichment over the input on the left hand side of the graphic. (B) Zoom of the promoters and terminator regions. 69

2.5.5 UBF is essential for the recruitment of all components of the RPI transcription machinery Loss of UBF in SV40 transformed conditional MEFs was previously shown to eliminate RPI transcription (Hamdane et al. 2014), and this is also true for p53-/- immortalized (Ubffl/fl/ER-Cre+/+/p53-/-) MEFs (Supplementary Figure 2.9E and F). Using these MEFs, and the ChIP-Seq approach, we found that consistent with previous data (Hamdane et al. 2014), not only was RPI and RRN3 recruitment eliminated by UBF loss, but also SL1 (e.g. TAFI95) binding at both spacer and 47S promoters (Figure 2.3A and B). Further, TTF1 binding at the upstream Tsp and T0 sites was strongly suppressed and there was a small but clear shift in binding preference at the 47S termination region, binding at the T1 and T2 sites being reduced in favour of binding at the downstream sites. This suggested that the domain of UBF binding modulated accessibility of

TTF1 to the T1 and T2 sites. We also consistently observed a significant but low level of TTF1 recruitment throughout the 47S transcribed region that was considerably reduced on inactivation of RPI transcription either by loss of RRN3 or UBF (Supplementary Figure 2.12).

2.5.6 UBF determines psoralen accessible and nucleosome exclusion on the active rRNA genes Psoralen cross-linking accessibility has long been used to differentiate the active from the inactive rRNA genes based on the absence or presence of nucleosomal chromatin within the 47S transcribed region (Conconi et al. 1989). Suppression of RPI transcription by inactivation of the UBF gene in the Ubffl/fl/ER-Cre+/+/p53-/- MEFs eliminated the low mobility band corresponding to the active rDNA and enhanced the higher mobility inactive band (Figure 2.4A) consistent with previous data (Hamdane et al. 2014, Sanij et al. 2015). In contrast, suppression of transcription by inactivation of the Rrn3 gene in the Rrn3fl/fl/ER-Cre+/+/p53-/- MEFs did not significantly affect the rDNA banding pattern after psoralen crosslinking (Figure 2.4B and Supplementary Figure 2.9F). Thus, enhanced psoralen accessibility of the active rRNA genes corresponds with the binding of UBF and does not reflect RPI recruitment or active transcription.

We further asked if loss of UBF also led to the reformation of nucleosomal chromatin on the previously active rRNA genes. Before UBF deletion, micrococcal nuclease digestion of rDNA chromatin displayed a “ladder” of inter-nucleosomal DNA cleavage characteristic of nucleosomal chromatin within the non-transcribed IGS, but not within the transcribed 47S region (Figure 2.4D and Supplementary Figure 2.13B). Loss of UBF led to the establishment of the cleavage ladder of nucleosomal chromatin also within the 47S region. However, loss of RRN3 did not lead to a nucleosomal ladder, likely due to the continued presence of UBF on the rDNA (Figure 2.4E and Supplementary Figure 2.13C). Together with the psoralen accessibility data, this suggested that UBF binding was sufficient to exclude nucleosomes from the 47S region. Indeed, we previously showed that the nucleoprotein structure formed by UBF binding to DNA is incompatible with nucleosome formation

(Stefanovsky et al. 2001a). However, as mentioned above, the low binding constant and high off-rate of UBF make it very unlikely that it alone could exclude nucleosomes. The data, therefore, suggest a specific and transcription-independent mechanism for UBF deposition across the rDNA and the involvement of a chromatin remodeling complex.

Figure 2.4 Loss of UBF induces nucleosomal formation and shut down of rRNA genes The upper panels show typical electrophoretic separation of actively transcribed “a” and inactive “i” genes from psoralen crosslinking experiments of Ubffl/flERCre+/+p53-/- MEFs (A) and Rrn3 fl/flERCre+/+p53-/- MEFs (B). Ctrl is non-crosslinked DNA from MEFs. (C) Diagram depicting the rDNA unit and the positions of IGS and 47S probes. Genomic DNA was MNase digested and loaded on an agarose gel to investigate nucleosome occupancy of the rDNA of Ubffl/flERCre+/+p53-/- MEFs (D) and Rrn3 fl/flERCre+/+p53-/- MEFs (E). 71

2.5.7 A chromatin boundary is found upstream of the spacer promoter In order to investigate the potential mechanism of UBF deposition throughout the rDNA, we thought to analyze the chromatin environment across the rDNA repeat. We observed a decrease of canonical histone H3 throughout the regulatory and coding regions, supporting the lack of a nucleosomal ladder in these regions (Figure 2.4 and 2.5). We also observed strong peaks for histone variant H2A.Z and its acetylated form H2A.Zac, activating marks H3K4me2 and -me3 and insulator factor CTCF. Aligning these peaks, it appears that CTCF is just upstream of the UBF SpPr peak and H2A.Z and H3K4 are found adjacent and upstream of CTCF (Figure 2.5B). Indeed, CTCF was recently shown to interact with UBF and stimulate UBF binding to the rDNA (van de Nobelen et al. 2010). CTCF is the major chromatin insulator-binding transcription factor in mammals, and has been shown to act an activator and a repressor of transcription as well as a chromatin boundary (Burgess-Beusse et al. 2002, Cuddapah et al. 2009). Strong peaks of H3K4me2 and H3K4me3, associated with gene activation, were observed, however, we do not observe a peak of the repressive mark H3K9me3 (Figure 2.5A). In addition, we find H2A.Z, which has been associated with both active and inactive promoters, however H2A.Zac has been linked specifically to actively transcribed genes (Valdés-Mora et al. 2012). These marks are found upstream of the rDNA unit, but not at the downstream terminators (Figure 2.5).

We took advantage of our knockout model of Ubf in order to investigate whether these marks were dependent on gene activity and whether they were transient or stable. We observed no change for CTCF and H2A.Z occupancy after loss of UBF. Interestingly, we observed an increase of H3K4me3 and H2A.Zac signal in this region after loss of UBF (Figure 2.6). This is surprising as these marks are usually associated with gene activity. Further, we found these active marks mapped to three phased nucleosomes just upstream of the SpPr. Therefore, these results suggest the existence of an upstream chromatin boundary element formed by the activating marks H2A.Z, H2A.Zac, and H3K4, by CTCF and possibly TTF1 that delimits the rDNA unit upstream independent of its transcriptional activity.

Figure 2.5 A chromatin boundary element is found upstream of the rDNA unit (A) UBF, CTCF, H3K4me2, H3K4me3, H2A.Zac, H2A.Z, H3K9me3 and H3 ChIP-seq performed in untreated Ubffl/flCreER+/+p53-/- MEFs was aligned to a concensus unit of the murine rRNA gene and visualized using the Integrative Genomics Viewer (IGV). The rDNA repeat is represented below the histograms and the scales are shown in fold enrichment over the input on the left hand side of the graphic. (B) Zoom of the promoters and terminator regions.

Figure 2.6 Activating histone marks at the spacer promoter are increased after loss of UBF Further ChIP-seq analysis of chromatin factors on the rDNA show that there exist three phased nucleosomes upstream of the spacer promoter (track 8 and 9) and that H2A.Z is not affected by UBF loss. However, H3K4me3 and H2A.Zac relative occupancy appears to increase after UBF loss (tracks 4- 7).

2.5.8 TTF1 regulates spacer promoter transcription by arresting polymerase on active genes Unexpectedly, our ChIP-seq results also revealed a strong peak of RPI near the spacer promoter (Figure 2.1) that we also see at an analogous site on the human rDNA (data not shown). Promoter recruitment of RPI requires RRN3, which is released during early elongation, a process clearly evident in Figure 2.1. In contrast, the peak of RPI near the Spacer Promoter (SpPr) is strongly depleted of RRN3 compared with RPI at the 47S promoter, suggesting these RPI molecules are no longer initiation-competent and are arrested early in elongation. Fine mapping reveals that the peak of RPI is centered about 30 bp downstream of the published spacer promoter start site (+1) (Kuhn & Grummt 1987) and about 40bp upstream of a consensus TTF1 site, and abutts a peak of TTF1 binding (Figure 2.7). If the arrested polymerase occurs on the active rRNA genes, which is likely as RPI is lost on the SpPr after UBF knockout, this is consistent with the idea that TTF1 is preventing in cis silencing by suppressing synthesis of the lncRNA pRNA, which is transcribed starting from

the SpPr. Thus, TTF1 bound at this site may control synthesis of the pRNA and hence the potential for silencing. As TTF1 binding is hardly affected by the loss of Ubf or Rrn3, not only does it delimit the rDNA unit upstream and downstream, but is also an important regulator of the activity status of the rDNA.

RPI Rrn3 TAF1B TTF1

+1 +1

Spacer Transcript 47S rRNA

Spacer Prom. 47S Prom. Tsp T0

Figure 2.7 A stalled RNA polymerase I is found just downstream of the spacer promoter Diagram depicting the overlap of ChIP-seq signal for RPI, RRN3, TAF68 and TTF1 at the spacer promoter (left panel) and the 47S promoter (right panel) in MEFs. The Sal-box TTF1 binding sites (TSP and T0 are depicted in red and the promoters in blue boxes.

2.6 Discussion Using high-resolution mapping of the rDNA, we have characterized the binding of RPI transcription factors RRN3, UBF, SL1 and TTF through the gene body as well as the regulatory elements in the IGS. In analyzing the knockout of RRN3, we have shown that it is responsible for the recruitment of RPI, but SL1 recruitment and PIC assembly is unaffected by its loss. On the contrary, UBF is required for binding of SL1 and PIC formation at both spacer and pre-rRNA promoters, and for the recruitment of RPI-RRN3 to these promoters. In contrast, it is not required for TTF1 recruitment. An upstream boundary element is formed by activating histone marks such as H3K4me2/3, and the replacement histone H2A.Z and its acetylated form H2A.Zac. This boundary is sustained even in the complete absence of UBF and for tens of hours after transcription arrest, suggesting that it is a highly stable marker of potential rDNA activity.

Our data suggest that TTF1 is an important regulatory factor on both the active and the inactive rDNA genes.

Its binding to the TSP is required for recruitment of the NoRC complex in order to initiate silencing (McStay &

Grummt 2008). However, our results demonstrate that it also arrests RPI transcription from the spacer promoter, suggesting that its function here is to arrest synthesis of the pRNA. This would mean that, on one hand, TTF1 uses this pRNA to recruit the NoRC complex and silence the rDNA, while on the other hand it regulates synthesis of the lncRNA precursor of the pRNA. TTF1 would, therefore, be the master regulator of rDNA silencing. We are unsure however, whether TTF1 is also found on the heterochromatic silenced or the active rDNA fraction.

We found CTCF, H2A.Z and H3K4me2/3 upstream of the spacer promoter acting as a chromatin boundary. Both CTCF and H2A.Z are known to act as boundaries and are able to both activate or repress transcription from their associated genes (Billon & Côté 2012, Felsenfeld et al. 2004). CTCF was previously shown to bind the IGS, with a preference for non-methylated rDNA, and recruit UBF to the rDNA (van de Nobelen et al. 2010). In addition, these two factors were unchanged upon loss of UBF. Therefore, CTCF and this H2A.Z boundary may act as a placeholder on potentially active genes in order to deposit UBF.

CTCF may also act to separate active rDNA copies from adjacent silent copies. It has been proposed that NORs contain both active and inactive repeats (McStay 2016). Interestingly, CTCF has been implicated in X chromosome inactivation to separate the escape domains from those to be silenced and impede the spread of methylation or other silencing marks (Filippova et al. 2005). Similar to rDNA silencing, X inactivation also requires transcription of a lncRNA.

It was surprising to observe that the activating chromatin marks H2A.Zac and H3K4me3 were increased upon loss of UBF. However, the pattern of H2A.Zac correlating with H3K4me3 and not with heterochromatic marks is expected (Bernstein et al. 2005). A recent study suggested that poised genes could be reactivated upon H2A.Z acetylation and that it was this mark that was essential for upregulating genes in cancer cells (Valdés- Mora et al. 2012).

This suggests a possible model for the regulation of rDNA gene silencing where H2A.Z and CTCF act as a barrier or placeholder for active rDNA repeats, and it is the acetylation of H2A.Z that seperates the active genes from the inactive. This mechanism could depend on whether TTF1 recruits NoRC and pRNA to induce silencing or whether it arrests the polymerase from transcribing the IGS lncRNA, however this is yet to be shown experimentally. In conclusion, we have developed powerful tools with which to study the chromatin landscape of the rDNA and potentially uncovered a role for TTF1 as a master regulator of gene silencing.

2.7 Materials and methods Isolation and cultures of MEFs and iMEFs. The generation of conditional Ubffl/flER-Cre+/+ and control mouse lines was previously described (Hamdane et al. 2014). For Rrn3+/-, Rrn3fl/flER-Cre+/+ mice and the isogenetic Rrn3+/+ER-Cre+/+ control mice, the following mouse strains from Jackson Laboratory were used; B6.Cg-Tg(Sox2-cre)1Amc/J (#008454), B6;129- Gt(ROSA)26Sortm1(cre/ERT)Nat/J (#004847) and B6.Cg-Tg(UBC-cre/ERT2)1Ejb/2J (#008085). C57BL/6NCrl wild- type controls for backcrossing were bought from Charles River. The 129 Sox2-Cre mice used for the original crosses were kindly provided by Dr Lucie Jeannotte. The p53-null allele was introduced by crossing to strain 129-Trp53tm1Tyj/J (Jackson Laboratory #002080)(Hamdane et al. 2015). Primary mouse embryonic fibroblasts (MEFs) from E14.5 Ubf or Rrn3 floxed MEFs and corresponding p53-/- MEFs were prepared as previously described (Giroux et al. 1999, Hamdane et al. 2014). Cells were cultured in Dulbecco’s modified Eagle medium (DMEM)-high glucose (Life Technologies), supplemented with 10% fetal bovine serum (Wisent), L- glutamine (Life Technologies) and Antibiotic/Antimycotic (Wisent). RRN3 MEFs were immortalized by the introduction of the SV40 Tt antigens by transfection with the pBSV0.3T/t, a modification of the pBSV-early vector (Schaffner 1980) kindly provided by E. W. Khandjian. Clonal iMEF cells conditional for Rrn3 were generated by serial dilution and plating in a 96-well plate at <1 cell per well in pre-conditioned media (filtered media from Rrn3 iMEFs cultured for 3 days) and allowing for clonal expansion.

Embryo collection, culture and genotyping Heterozygous Rrn3+/- mice were inter-crossed and embryos isolated from pregnant female at 3.5 dpc by flushing uterine horns with PBS1X. Embryos were collected in 8-wells plate (Ibidi) and imaged by bright-field microscopy. DNA from the individual blastocysts was then amplified using the REPLI-g Mini kit (QIAGEN) and individual embryos were genotyped by PCR using the primers: A; 5'-GATCTTAATGGAGGGCAGCA , B; 5’- TGGATCCTGCAACTTTTTCC , C; 5’ TCCCAACCCTGACCTATCAC.

Inactivation of Ubf or Rrn3 in cell culture, and analysis of genotype, RNA and proteins. As previously described (Hamdane et al. 2014), cells were initially plated in 6 cm petri dishes (0.8x106 cells each) and cultured for 18 hours in DMEM, high glucose, 10% fetal bovine serum. To activate ER-Cre, 4- hydroxytamoxifen (4-HT) was added to a final concentration of 50nM, and after 4 hr incubation the medium replaced with fresh medium without 4-HT and cells harvested for analysis at various time points. In the case of Rrn3, cells were treated with 50nM 4-HT for a 4h period, passaged 1:2 and retreated with 50nM 4-HT. This double treatment protocol was repeated 16h later to ensure complete gene excision. Analyses of RNA, protein and genotype were systematically carried out on parallel cell cultures. Cells were genotyped by PCR before and after 4-HT treatment using the primers for Ubf: A; 5’TGATCCCTCCCTTTCTGATG, B; 5’TGGGGATAGGCCTTAGAGAGA, C; 5’CACGGGAAAACAAGGTCACT and for Rrn3: A; 5'- GATCTTAATGGAGGGCAGCA , B; 5’-TGGATCCTGCAACTTTTTCC , C; 5’ TCCCAACCCTGACCTATCAC. Metabolic labelling of RNA was carried out just before cell harvesting by addition of 10 Ci [3H]-uridine (PerkinElmer) to the culture medium and incubation for a further 3h. RNA was extracted with Trizol (Life Technologies) according to the manufacturer’s protocol and analyzed by gel electrophoresis, fluoroimaging (Enhance, PerkinElmer) and RNA species quantitated by scintillation counting as previously described (Stefanovsky & Moss 2016). For total protein, cells were washed with cold PBS, scraped into PBS, centrifuged 2 min at 2 000 r.p.m., then resuspended in sodium dodecyl sulphate (SDS) loading buffer. After fractionation on 8% SDS–polyacrylamide gel electrophoresis (SDS-PAGE), cell extracts were analysed by standard Western blotting procedures.

Psoralen crosslinking and Southern blot. Psoralen crosslinking was performed as described elsewhere (Conconi et al. 1989). Cells were grown in 60 mm petri dishes (0.8x106 cells/petri for NIH3T3, MEFs and L1210 cells, 2x106 cells/petri for ESCs) and treated as for RNA labeling, but at the end of the pulse labeling time the medium was changed for 1.5ml ice-cold serum-free medium and crosslinking immediately started. A 1/20 volume (75µl) of 200µg/ml Trioxsalen (4,5’,8- trimethylpsoralen, Sigma) in methanol was added and after 5 min incubation, the cells were irradiated for 5 min with a 366nm UV lamp (BlackRay model B-100A, 440W) placed at a distance of 6-7cm. Crosslinking was repeated three more times, each time adding fresh Trioxsalen. The crosslinking procedure was performed in a dark room and the cells were maintained throughout on ice. Subsequently, the cells were washed with PBS and lysed in-petri with 0.5 ml of lysis buffer (10mM Tris, pH 7.5, 50mM NaCl, 25mM EDTA, 2% SDS) and transferred to an Eppendorf tube. Petri dishes were rinsed with 0.5 ml TE, pH 7.5, this combined with the cell lysate, 1mg/ml Proteinase K (from 10mg/ml stock, Sigma) was added and the combined lysate incubated overnight at 50°C, adding fresh Proteinase K after the first 2 hrs. The genomic DNA was then deproteinized twice with Phenol/Chloroform, ethanol-precipitated and resuspended in 250µl TE, pH7.5, 0.1% SDS. 20µg of RNAse A (Sigma) was added and after 30 min at 37°C, 0.5mg/ml Proteinase K was also added and incubation continued for another 1h at 50°C. After twice extracting with phenol/chloroform treatmets the genomic DNA was ethanol precipitated, resuspended in 300 µl TE and quantified by EtBr staining on an agarose gel. About 10µg of genomic DNA was digested overnight either with BamHI or EcoRI, deproteinized, ethanol precipitated and finally resolved on a 1% TEA agarose gel at 2V.cm-1 for 18 hr in the absence of EtBr. The gel was subsequently EtBr-stained, photographed, the crosslinks reversed by UV irradiation of 4000J.cm-2 at 254nm in a UVC 500 crosslinker (Hoefer), transferred on a Biodyne B membrane (Pall) and the membrane crosslinked (70J.cm-2). Hybridization was performed overnight at 65°C with random-primed labeled DNA in 5xSSC, 0.5%SDS, 5xDenhardt’s, 100µg/ml salmon sperm DNA. Subsequently membranes were washed (15 min each time) with 6xSSC, 2xSSC and 0.1xSSC, 0.1% SDS. Data were analyzed by STORM 860 PhosphorImager. (GE Healthcare) and with ImageJ software (NIH).

Antibodies for Western Blot, Immunofluorescence and ChIP Rabbit antibodies against UBF, RPI large subunit (A194), TTF1, TAF68 and RRN3 (Supplementary Figure 2.9C) were generated in the laboratory. UBF, RPI and TTF1 have been previously described (Lessard et al. 2010). Rabbit antisera were raised against TAF68 aa 54-175 and RRN3 aa 464-656, expressed in E. coli and peptides were purified using the guanidinium chloride/urea denaturation method. Rabbit antibody against TAF95 was a gift from Ingrid Grummt. Anti-H2A.Z and anti-H2A.Zac were a gifts from Colyn Crane-Robinson. All other antibodies were obtained commercially; TBP (Abcam), anti-Tubulin (Sigma), anti-Fibrillarin (Covance), anti-CTCF (Millipore), anti-H3K4me3 (Abcam), anti-H2A.Z (Abcam), anti-H3K9me3 (Millipore), anti- H3 (Abcam).

Chromatin immunoprecipitation (ChIP). Cells were fixed with 1% formaldehyde for 8 min at room temperature. Nuclei were isolated using Lysis Buffer (10 mM Tris pH 7.5, 10 mM NaCl, 3 mM MgCl2, 0.5% NP-40), and the resulting chromatin sonicated (Bioruptor, Diagenode) for 2 cycles of 30 sec on/30 sec off for 15 min on high in Sonication Buffer (50 mM Tris- HCl pH 7.5, 150 mM NaCl, 2 mM EGTA, 4 mM EDTA, 0.1% SDS, 1% Triton X-100, 1% NP-40). Each immunoprecipitation was carried out on the equivalent of 50 x 106 cells in IP Buffer (150 mM NaCl, 50 mM Tris-HCl pH 7.5, 5 mM EDTA, 0.5% NP-40, 1% Triton X-100) overnight at 4°C. Antibody slurry was prepared the night before with 50 µl A/50 µl G Dynabeads and 60 µg/ml antibody per IP. Two or more biological

replicates were analyzed for each antibody. Samples were analyzed by qPCR before being sent for single-end sequencing on an Illumina Hi-seq 2000 (McGill University and Génome Québec Innovation Centre). qPCR analysis. Immunoprecipitated DNA was analyzed by qPCR/SYBR Green. Reactions (20 µl) were performed in triplicate with 2.5 µl of sample DNA, 20 pmol of each primer, and 10 µl of Quantitect SYBR Green PCR Master Mix (QIAGEN). Forty reaction cycles of 10 s at 95°C and 30 s at 58°C were carried out on a Multiplex 3005 Plus (Stratagene/Agilent). The amplicon coordinates relative to the 47S rRNA initiation site (BK000964) were as follows: IGS3, 42646– 42903; SpPr, 43089–43253; Tsp, 43267–43421; 47SPr, 45133–40; 47S, 159-320; ETS, 3078–3221; 28S, 10215–10411; T1– 3, 13412–13607. Data was analysed using the MxPro software (Agilent). The relative occupancy of each factor at each amplicon is given as % immunoprecipitation of the DNA input prior to ChIP. It was determined by comparison with a standard curve of amplification efficiency for each amplicon using a range of input DNA amounts and generated in parallel with each qPCR run. All primer pairs gave the similar amplification efficiencies (90–105%) as determined from the gradient of the curve fit. The curve fit correlation coefficient R2 was systematically between 0.99 and 1.0, demonstrating a near perfect fit. At least three biological replicates have been performed fro each antibody.

ChIP-seq analysis. Raw fastq.gz data was checked for quality using FastQC version 0.11.4 (Andrew 2010). The data was trimmed using Trimmomatic version 0.33 (Bolger et al. 2014) and the resulting quality filtered files were aligned to the mouse genome version MmGRCm38 to which a single copy of the mouse rDNA repeat sequence (GenBank BK000964v3) was added as an extra chromosome using Bowtie2 (Langmead & Salzberg 2012). For convenience, the origin of the rDNA repeat unit was displaced to the EcoRI site at 30,493 such that the pre- rRNA initiation site now fell at nucleotide 14,815. Reads were extended to 100bp, the coverage calculated using BEDtools (Quinlan Lab, University of Utah), and smoothed using a window of 25 bp. The coverage was normalized to reads per million (RPM) and the sample data was divided by the equivalent dataset for the input DNA. The BED files were converted to BEDgraph files and visualized in IGV (Integrative Genomics Viewer 2.3, Broad Institute).

MNase digestion of genomic DNA In these experiments between 3-9x107 cells were used. After trypsinizing and washing with PBS the cells were resuspended in buffer A (60 mM KCl, 15 mM NaCl, 15 mM Tris pH 7.6) plus 0.25 M sucrose, 1 mM EDTA, 0.1 mM EGTA, 0.5 mM spermine, 0.15 mM spermidine, 0.1 mM PMSF and protase inhibitors. TritonX100 was added to final concentarion 0.25% and the cells homogenized in a Dounce homogenizer using 10 piston strokes and centrifuged 5 min at 2000 rpm. The pellet was washed once by resuspending in buffer A plus 0.34M glucose and then centrifuged for 5 min at 2000 rpm. It was then resuspended in 1 ml buffer A without sucrose, recentrifuged and resuspended in buffer A containing 60 units DNAse per 107 cells. After incubating for 2 min on ice, 1 mM CaCl2 was added and incubation continued on ice. 200 µl aliquots were taken at 5', 10', 15', 20', and 30' and the reaction was stopped by adding EDTA to final concentration of 10 mM, SDS to 0.5% and Proteinase K to 1 mg/ml. After overnight incubation at 55°C, 0.1 mg/ml RNase A and 3 units RNase T1 were added and incubated for 1 h at 37°C. This was followed by the addition of 1 mg/ml Proteinase K for 2 h at 55°C. The DNA was then extracted with phenol/chloroform, precipitated and dissolved in 200-400 µl TE1X pH8.3. 10 µl DNA was mixed with 2X gel loading buffer containing 0.1% SDS and 0.1 mg/ml RNase A without dye. After 30 min incubation at 37°C the samples were resolved on 1.5% agarose gel, the gel

transferred onto a Biodyne B (Pall) membrane and then hybridized as described before. The 6.2 kb EcoRI/EcoRI IGS probe was obtained from plasmid 15 (Grozdanov et al. 2003)

Southern blot. Standard Southern blotting procedure was followed (Southern 2006). Briefly, 5µg of genomic DNA was digested with SpeI and loaded on a 0.8% agarose gel. After 16h migration at 30V, the gel was transferred by capillary to a Pall Biodyne membrane. The membrane was hybridized with a 3’ probe corresponding to a region overlapping exon 16 and intron 17 (Supplementary Figure 2.10E) overnight and the membrane was exposed on a phosphor screen for 16h and imaged on FujiFilm FLA-5100 Fluorescent Image Analyzer.

2.8 Acknowledgements We wish to thank Dr Lucie Jeannotte for providing the p53 null allele mice and advice on their use, Dr Ross Hannan for discussion at various stages of this work, and Drs Mark Robinson and Helen Lindsay for advice on bioinformatic analysis and for access to their computing facility. This work was funded by an operating grant from the Canadian Institutes of Health Research (CIHR, MOP12205) and a CIHR Frederick Banting and Charles Best Canada Graduate Scholarship Doctoral Award to Chelsea Herdman (CIHR CGS-D). The Research Centre of the Québec University Hospital Centre (CHU de Québec) is supported by the Fonds de recherche du Québec - Santé (FRQS).

2.9 Supplemental data Figure 2.8, related to Figure 2.1 Figure 2.9, related to Figures 2.2 and 2.3 Figure 2.10, related to Figure 2.2 Figure 2.11, related to Figures 2.2 and 2.3 Figure 2.12, related to Figures 2.1, 2.2 and 2.3 Figure 2.13, related to Figure 2.4

Figure 2.8 UBF occupancy correlates with GC-rich sequences in the rRNA gene body

UBF ChIP-seq results are portrayed in the upper histogram in blue and the % of GC DNA content in the lower histogram in grey.UBF seems to correlate with the signal for GC-rich (over 50%) sequences throughout the gene body but not in the promoter region.

Figure 2.9 Conditional knockout models of RRN3 and UBF

(A) to (C) Rrn3fl/flERCre+/+p53-/- MEFs and (D) to (E) Ubffl/flERCre+/+p53-/- MEFs were treated with 4-HT to induce the knockout of these factors. (A) and (D) depict the wildtype, conditional and mutant alleles for each factor. (B) and (E) show the genotyping by PCR and Western blot for UBF or RRN3 levels after tamoxifen treatment. (C) demonstrates an example of a Western blot of the RRN3 C-terminal antibody generated in the lab (F) shows the quantitative analysis of rRNA synthesis in both cell lines as measured by metabolic labeling with [3H]-uridine to follow synthesis and pool sizes.

Figure 2.10 Deletion of the Rrn3/Tif1a gene arrests mouse development during early cleavage divisions

E) Rrn3 / ER-Cr e+/+ SvT iMEFs Rrn3 tails

Days post 4-HT: Ctrl 4 8 12 / +/+ +/- (500nM/48h)

10000 s

8000 n

6000 g

5000

4000 p

S 3500

s F E F) iM vT S /+ + 2 T re -C R Rrn3 / ER-Cr eT2 +/+ SvT iMEFs /+ E 48h after 4-HT + 3 (50nM/4h) rn F1 F2 F3 F4 F5 F6 R

ox (626bp)

p wt (527bp)

y t

e G ∆ (245bp)

/ +/+ -/- G) Rrn3 ER-Cr e p53 MEFs % loading: 10 25 50 100

hours post 4-HT: 0 0 0 0 24 48 72 (3x50nM/4h)

n RRN3

P Tu b.

Figure 2.10 Deletion of the Rrn3/Tif1a gene arrests mouse development during early cleavage divisions

Figure legend on next page

It has been found that homozygous deletion of the mouse Fibrillarin (Fbl), RPI second largest subunit (Rpa135/Rpo1-2/Polr1b) or Upstream Binding Factor (Ubf/Ubtf) genes all cause developmental arrest during the cleavage divisions and well before the blastula stage (Chen et al. 2008, Hamdane et al. 2014, Newton et al. 2003). This is consistent with the formation of the nucleolar precursor and a requirement for the activation of the rRNA genes at or soon after the 2-cell stage (Engel et al. 1977, Hamdane et al. 2016). In contrast, it was reported that homozygous deletion of the gene for the RPI associated factor RRN3/TIF-IA (Rrn3/Tif1a) arrested development much later and clearly displayed axis formation, tissue differentiation and organogenesis (Yuan et al. 2005). While establishing Rrn3 conditional cell lines from the Rrn3flox mouse cell line created by Yuan et al., we also generated mice lacking a functional Rrn3 gene. By crossing these mice, we found that both mixed background C57BL/6J-129 and pure C57BL/6J background Rrn3-/- mice arrested development during the cleavage stages and failed to develop to blastula. Thus, loss of the RRN3 gene causes a very similar arrest in development to loss of Fibrillarin, RPI or UBF. Since our Rrn3+/- mouse line was extensively backcrossed to remove the transgenes used to recombine the Rrn3flox allele generated by Yuan et al. and we determine that no RRN3 protein was produced, we presently have no explanation for the discrepancy with the previous study. (A) Embryonic lethality study for RRN3 knockout. (B) Images of typical morphology for embryos of the different genotypes shown. (C) Embryonic lethality study for RRN3/UBF double heterozygous mice. (D) Alignment of LoxP sites from the original vector, UBF and RRN3 mice. (E) Southern blot analysis of RRN3 MEFs and mice tails using 3’ probe indicated in Figure 2.9. Fragment sizes are as follows; wild-type (+): 11973 bp, flox: 8895 bp and delta (-): 3988 bp. (F) PCR analysis after tamoxifen treatment of clonal cell lines produced from Rrn3fl/flERCreT2+/+sv40T iMEFs showing that even six different isolated floxed clones (F1-F6) undergo the same level of excision after one 4-HT treatment (G) Western blot analysis of RRN3 after knockout with serial dilution.

Figure 2.11 qPCR analysis after RRN3 or UBF knockout ChIP for UBF, RRN3, RPI, TTF1, TAF68, TBP, CTCF, H3, H2A.Z, H2A.Zac, H3K4me3 was analyzed by qPCR after knockout of RRN3 or UBF. Blue lines or bars represent the control, non- treated cells, whereas red represents three days after Cre induction. The primers are indicated on the X-axis and the percent enrichment on the Y-axis.

Figure 2.12 TTF1 diminishes its presence through the gene body after knockout ChIP-seq for TTF1 and RPI in UBF or RRN3 knockout cells. Black lines indicates that TTF1 occupancy density through the rRNA gene decreases after knockout and the arrest of transcription.

Figure 2.13 MNase digestion of UBF and RRN3 knockout MEFs

(A) to (C) Ethidium bromide and wild type controls for Figure 2.4. (D) Rrnfl/flERCre+/+p53-/- MEFs do not display a nucleosomal ladder after 4-HT treatment.

Chapter 3 Depletion of the cisplatin targeted HMGB-box factor UBF selectively induces p53-independent apoptotic death in transformed cells

Nourdine Hamdane1,2,*, Chelsea Herdman1,2, Jean-Clément Mars1,2, Victor Stefanovsky1, Michel G. Tremblay1, and Tom Moss1,2.

1Laboratory of Growth and Development, St-Patrick Research Group in Basic Oncology, Cancer Division of the Quebec University Hospital Research Centre. 2Department of Molecular Biology, Medical Biochemistry and Pathology, Faculty of Medicine, Laval University. *Present address; Inserm, U1110, Institute of Viral and Liver Diseases, Strasbourg, France; University of Strasbourg, Strasbourg, France.

Keywords Upstream Binding Factor (UBF/UBTF), Ribosome Biogenesis, Oncogenic Stress, Cell Death, Apoptosis, P53- independent, Cisplatin, Carboplatin.

3 Depletion of the cisplatin targeted HMGB-box factor UBF selectively induces p53 independent apoptotic death in transformed cells

3.1 Preface The primary characterization of the knockout of UBF was the doctoral project of Nourdine Hamdine and this was published in 2014 (Hamdane et al. 2014). This manuscript published in Oncotarget (see Annex 1) was the continued study of the loss of UBF in primary and transformed cell lines. Much of the work in primary MEFs and SV40 T immortalized MEFs was done by Nourdine Hamdane, Victor Stefanovsky and Michel Tremblay before I started my PhD project. I was involved in the cisplatin experiments and the experiments involving the transformed p53 null lines. My figures in this paper are as follows; Figures 3.1, 3.2d, 3.7 (in collaboration with Victor Stefanovsky), Supplementary Figures 3.11c, 3.11d, 3.12, 3.14b, and 3.14d. Jean-Clément Mars, Victor Stefanovsky and myself were responsible for the revision process.

NOTE: There has been a recent push in the field to use the human and yeast nomenclature for RRN3 in mouse as opposed to the original name, TIF-IA. This paper was published before this change and therefore RRN3 is referred to as TIF-IA/TIF1A. Please excuse the variability in nomenclature.

3.2 Résumé Les adduits cisplatine-ADN agissent comme leurres pour l’Upstream Binding Factor UBF (UBTF) et peuvent inhiber la transcription des gènes codants pour l’ARN ribosomal par l’ARN polymérase I. Cependant, l’importance de ces effets sur l’activité chimiothérapeutique de cis- ou carboplatin demeure inconnue. Nos résultats démontrent que le cisplatin induit un déplacement rapide de UBF des gènes d’ARN ribosomal et une forte inhibition de la synthèse d’ARNr, suggérant un rôle important dans la cytotoxicité de ce traitement. En étudiant la délétion conditionnelle de Ubf, nous avons démontré que UBF est un facteur essentiel pour la transcription des gènes codants pour l’ARN ribosomal ainsi que pour la synthèse des ribosomes. La perte d’Ubf cause la déstabilisation du noyau, entraine un arrêt de la prolifération cellulaire, et induit la mort cellulaire de façon robuste, rapide et synchronisée. Ces effets arrivent pincipalement dans les cellules soumises à un stress oncogénique. Cette apoptose n’est pas affectée par la perte de p53 et se déroule de façon similaire dans des lignées cellulaires transformées avec l’antigène SV40 T, avec Ras ou encore par une combinaison des oncogènes Ras et Myc. Nos résultats démontrent que l’inhibition de l’activité de UBF est un facteur majeur dans la cytotoxicité du cisplatin. En vertu de la toxicité des composés à base de platines, il est permis d’imaginer que des molécules affectant UBF pourraient s’avérer être une approche thérapeutique dans la lutte contre le cancer.

3.3 Abstract Cisplatin-DNA adducts act as strong decoys for the Upstream Binding Factor UBF (UBTF) and have been shown to inhibit transcription of the ribosomal RNA genes by RNA polymerase I. However, it is unclear if this plays a significant role in the chemotherapeutic activity of cis- or carboplatin. We find that cisplatin in fact induces a very rapid displacement of UBF from the ribosomal RNA genes and strong inhibition of ribosomal RNA synthesis, consistent with this being an important factor in its cytotoxicity. Using conditional gene deletion, we recently showed that UBF is an essential factor for transcription of the ribosomal RNA genes and for ribosome biogenesis. We now show that loss of UBF arrests cell proliferation and induces fully penetrant, rapid and synchronous apoptosis, as well as nuclear disruption and cell death, specifically in cells subjected to oncogenic stress. Apoptosis is not affected by homozygous deletion of the p53 gene and occurs equally in cells transformed by SV40 T antigens, by Ras or by a combination of Ras & Myc oncogenes. The data strongly argue that inhibition of UBF function is a major factor in the cytotoxicity of cisplatin. Hence, drug targeting of UBF may be a preferable approach to the use of the highly toxic platins in cancer therapy.

3.4 Introduction The commonly used chemotherapeutic drugs cisplatin and carboplatin are generally considered to exert their cytotoxicity by inducing DNA damage. These drugs interact with DNA to form intra- and inter-strand crosslinks, which must be repaired for the cell to proliferate [1]. Hence, cells that grow more rapidly or are limited in their capacity to repair DNA should disproportionately suffer cell death, which often occurs by apoptosis. Consequently, growth factor driven tumour growth and deficits in the ability to rapidly repair DNA both enhance the ability of cisplatin to induce cell death [1-5]. DNA-platin adducts are also aberrantly bound by a range of nuclear proteins, and this in general enhances cell death by delaying their repair [6, 7]. Important among these nuclear proteins are members of the High Mobility A and B families (HMGA and HMGB), which display elevated affinities for the bent DNA structure of the platin adducts via their HMGA-box and HMGB-box DNA binding domains [8-10]. Upstream Binding Factor (UBF) is an abundant multi-HMGB-box transcription factor that defines the active state of ribosomal RNA (rRNA) gene chromatin by replacing the core histones and is essential for transcription of these genes [11-13]. It has long been known that UBF has a particularly high affinity for cisplatin-DNA adducts, which may act as molecular decoys to attract this factor away from the rRNA genes and in so doing suppress their transcription [14-19]. Since transcription of the rRNA genes is the central event in the assembly of ribosomes, the protein factories of the cell, their activity is essential for cell growth and proliferation. The ability of cisplatin adducts to act as decoys for UBF binding could, therefore, enhance the drugs cytotoxicity either by inhibiting DNA repair or by inhibiting ribosome assembly, or both.

The rRNA genes are transcribed by RNA polymerase I (RPI/PolI), which is dedicated to this task. UBF is an HMGB-box DNA binding protein and one of the two essential RPI basal transcription factors [11, 20-23]. UBF is generally thought to mediate binding of the pre-initiation factor SL1/TIF1B and pre-initiation complex (PIC) assembly at the rRNA gene promoter. But UBF also forms a nucleosome-like structure that replaces histone chromatin throughout the transcribed regions of the rRNA genes and is able to regulate RPI transcription elongation in response to growth factor signalling [11, 24-28].

Ribosomal biogenesis is the process by which ribosomal RNA (rRNA) is transcribed, processed and assembled with the ribosomal proteins to create ribosomes [21, 29]. This energy consuming process is accomplished in the nucleolus and requires the action of the three RNA polymerases along with more than 200 different proteins and several hundred snoRNP complexes. Regulation of ribosome synthesis constitutes a major determinant of the increased protein synthesis needed for cell proliferation and, as such, its upregulation occurs in many cancers [30, 31]. An increased nucleolar volume reflects this increased ribosome synthesis, and is therefore a biomarker of cancer that was recognized already 80 years ago [32-34]. In fact, rRNA transcription is a common and probably an essential target of many oncogenes (Myc [35, 36], SV40-T

antigen [37, 38] and the Ras and mTOR signalling pathways [39-43]), and tumour suppressors (p53 [44], ARF [45-47], Rb [48, 49] and PTEN [50]).

Ribosomal biogenesis is such a central process in cell growth that it is also under the direct surveillance of the p53 pathway [51]. Defects in rRNA gene transcription [52], rRNA processing [53] or ribosome assembly [54] all cause p53 stabilization and arrest of cell proliferation. These findings have led to the investigation of small molecule inhibitors of ribosomal transcription as potential chemotherapeutic agents. Inhibition of the RPI pre- initiation factor SL1/TIF1B [55] or induced proteasome degradation of the RPI large subunit [56] both lead to arrest of rRNA synthesis and mediate cell death dependent on p53 function. However, the key to successful cancer therapy remains the selective targeting of cancer cells, and since p53 is often inactivated in human cancers, therapies that depend on functional p53 have limited application. Our data now suggest that inhibition the RPI basal transcription factor UBF (Upstream Binding Factor) represents a particularly valuable p53- independent target for cancer therapy.

Here we show that displacement of UBF and ablation of rRNA synthesis are very early effects of cisplatin treatment, and that in the absence of cisplatin elimination of UBF protein is sufficient to induce fully penetrant apoptotic cell death. Using cell cultures conditional for UBF expression, we find that complete loss of ribosome biogenesis induces synchronous and fully penetrant, p53-independent cell death by apoptosis specifically in cells transformed by known oncogenes. The data argue that a major factor in the cytotoxicity of cisplatin and similar drugs is their ability to inhibit the function of UBF. This suggests that UBF itself represents a preferred target for anticancer drug development.

3.5 Results Previous data has clearly indicated that cisplatin treatment of human cells leads to a partial or full displacement of human UBF and inhibition of rRNA synthesis [14, 15, 17, 18]. However, to what extent this plays a role in the selective cytotoxicity of cisplatin is not known. When the Mouse Embryonic Fibroblast (MEF) derived cell line NIH3T3 was treated for 4h with 30M cisplatin, a concentration calculated to be equivalent to the dose commonly used in therapy (e.g see [57, 58]), a large proportion of endogenous UBF was displaced from nucleoli and scattered throughout the nucleus at a large number of foci (Supplementary Figure 3.9). These foci were devoid of the other nucleolar proteins fibrillarin and RPI (data not shown), which remained together in dense nuclear bodies somewhat similar to the nucleolar precursor bodies forming on conditional deletion of the Ubf gene [11].

Figure 3.1 Cisplatin treatment of Ubfwt/wt/Er-cre+/+/SvT iMEFS induces displacement of UBF from the nucleolus iMEFs were treated with 30 µM cisplatin for 4 h in full medium or left untreated (0), then either fixed immediately or cultured in fresh medium lacking cisplatin overnight (22 h) as indicated in the timeline before fixing. The fixed samples were then subjected to indirect immunofluorescence analysis of UBF (green), fibrillarin (red) and DNA stained with DAPI (blue).

3.5.1 Cisplatin displaces UBF from the mouse rRNA genes and arrests their transcription To better understand the effect of cisplatin, we repeated and extended these studies using the independently isolated, iMEF cell line (Ubfwt/wt/Er-cre+/+/SvT) previously characterized by Hamdane et al. [11]. Already after 4h exposure of these cells to 30uM cisplatin, UBF was seen to coalesce from its normal specular distribution within nucleoli into more intense foci, while fibrillarin showed some degree of coalescence but was less affected (Figure 3.1). When these cells were cultured for a further ~18h in the absence of cisplatin, the UBF foci became more intense and UBF, but not fibrillarin, formed foci throughout the nucleus. The timing of the changes in UBF delocalization corresponded closely with changes in the interaction of UBF with the rRNA genes and with the transcription of these genes (Figure 3.2). After 4h of cisplatin treatment a mean reduction in UBF binding of 80% was observed across the 47S precursor rRNA coding region, and this corresponded with an 80% reduction in rRNA synthesis (Figure 3.2B and C). Due to its 5’ position in the 47S precursor, 18S rRNA synthesis was slightly less affected at 4h than the 28S rRNA, but nevertheless was reduced by over 60% after 4h cisplatin exposure (data not shown). 22 h after cisplatin exposure rRNA synthesis was no longer detectable. The effects of cisplatin on the activity of the rRNA genes also corresponded to an arrest of cell

proliferation, no increase in the viable cells count being detected after the 4h cisplatin treatment, and to a subsequent loss of viability (Figure 3.2D). These data suggest that the timeline of cisplatin cytotoxicity is consistent with its effects being mediated at least in part by disruption of UBF function, and the arrest of rRNA gene transcription and, hence, of ribosome biogenesis. Since cisplatin is a key chemotherapeutic agent that acts by inducing apoptotic cell death somewhat selectively in transformed cells (e.g. [3]), we sought to determine whether or not this activity could also be explained by the inhibition of UBF function.

Figure 3.2 Cisplatin coordinately displaces UBF from the rRNA genes and arrests their transcription (A) Timeline of cisplatin treatment and culture of Ubfwt/wt/Er-cre+/+/SvT iMEFs. (B) ChIP analyses of UBF occupancy across the rRNA gene 47S transcribed region. The positions of amplicons is indicated above the histogram showing the UBF occupancy normalized to that in the mock treated cells. (C) Synthesis rate of rRNA determined by [3H]-uridine metabolic labelling of mock treated cells and at the indicated times post cisplatin treatment. The upper panel displays a fluorogram of [3H]-rRNA, the central panel the corresponding EtBr stained total 18S rRNA, and the lower panel quantitation of [3H] incorporation into 47S rRNA performed in triplicate. (D) Live cell counts at indicated times following cisplatin treatment.

3.5.2 UBF loss disrupts nucleolar functions in both primary and transformed MEFs We previously generated mice conditional for the Ubf gene and demonstrated that loss of this gene arrested mouse development at the morula stage [11]. SV40Tt immortalized Mouse Embryonic Fibroblasts or iMEFs (Ubffl/fl/Er-cre+/+/SvT) generated from these mice allowed us to show that UBF was essential for transcription of the rRNA genes and for the existence of a functional nucleolus [11]. Not surprisingly, despite their limited proliferation potential, primary MEFs derived from these mice also require UBF for rRNA synthesis and for the maintenance of nucleoli (Supplementary Figure 3.10). Thus, UBF loss in primary MEFs recapitulated the effects observed in the transformed iMEFs.

3.5.3 Transformed iMEFs, but not primary MEFs, undergo synchronous apoptosis following Ubf inactivation Despite the apparently identical responses of the primary MEFs and the iMEFs to UBF loss, it became obvious from observing these cultures that the two cell types behaved very differently macroscopically. Inactivation of rRNA gene transcription in the Ubffl/fl/Er-cre+/+/SvT iMEFs induced changes in cell morphology soon after complete UBF depletion and the shutdown of rRNA synthesis. iMEFs became highly elongated and this presaged cell death as determined by plasma membrane failure (trypan blue), mitochondrial membrane depolarization (MitoTracker) and loss of clonal viability (Supplementary Figure 3.11A to D). Control Ubfwt/wt/Er- cre+/+/SvT iMEFs suffered none of these effects, clearly demonstrating that cell death was exclusively the result of inactivation of the Ubf gene. Interestingly, we detected no selective reduction of total cellular RNA in the Ubffl/fl/Er-cre+/+/SvT iMEFs relative to their wild type counterparts during UBF depletion that might suggest a role of ribosome depletion in the selective induction of apoptosis (data not shown). In contrast to the behavior of the Ubffl/fl/Er-cre+/+/SvT iMEFs, the primary Ubffl/fl/Er-cre+/+ MEFs showed no evidence of major morphological changes and survived in culture for many days following complete UBF loss, maintaining plasma membrane integrity and mitochondrial function (Supplementary Figure 3.11A to C).

To better understand the different responses of the transformed iMEFs and primary MEFs to UBF loss, we analyzed them for typical markers of cell death. TUNEL (terminal deoxynucleotidyl transferase-mediated dUTP nick end-labeling) analysis detects the single strand DNA cleavage that is characteristic of the early stages of apoptotic cell death. Ubffl/fl/Er-cre+/+/SvT iMEFs became TUNEL positive at 96h pHT, just 24h after complete shutdown of rRNA synthesis, while the control Ubfwt/wt/Er-cre+/+/SvT iMEFs remained TUNEL-negative throughout (Figure 3.3A). The TUNEL signal was fully penetrant and occurred synchronously, Ubffl/fl/Er- cre+/+/SvT iMEFs being TUNEL-negative at 72h pHT but all becoming TUNEL-positive at 96h pHT. In contrast, the Ubffl/fl/Er-cre+/+ primary MEFs remained TUNEL-negative at least until 144h pHT, (Figure 3.3B and data not shown).

Figure 3.3 UBF loss induces synchronous apoptotic cell death selectively in oncogenically transformed iMEFs (A) Ubffl/fl/Er-cre+/+/Sv-T and Ubfwt/wt/Er-cre+/+/Sv-T iMEFs and (B) Ubffl/fl/Er-cre+/+ and Ubfwt/wt/Er-cre+/+ primary MEFs were subjected to a TUNEL reaction immediately before, and at several time points after, treatment with 4-HT. In both cases, recombination and UBF protein levels were assayed in parallel and closely followed those shown in Supplementary Figure 3.10B and C.

Concomitant with the onset of TUNEL-positive apoptosis, the Ubffl/fl/Er-cre+/+/SvT iMEFs were also found to activate Caspase 3 from 96h pHT, as determined by the release of the 17kD peptide (p17) cleavage product (Figure 3.4A). In contrast, the control Ubfwt1wt/Er-cre+/+/SvT iMEFs displayed no significant cleavage of Caspase 3, consistent with the lack of a TUNEL signal. Further, Caspase 3 was not significantly activated in the primary MEFs (Figure 3.4B). Though a certain level of cleavage was detected in both Ubffl/fl and Ubfwt/wt

MEFs, this was much weaker than observed in the Ubffl/fl/Er-cre+/+/SvT iMEFs as can be seen by comparison with Staurosporin-treated iMEFs.

Interestingly, unlike the deletion of UBF, deletion of the essential RPI initiation factor TIF1A/Rrn3 did not induce apoptosis in SV40Tt transformed MEFs. 4-HT treatment of TIF1Afl/fl/Er-cre+/+/SvT MEFs resulted in complete depletion of TIF1A by 48h pHT, as observed for UBF, but did not lead to activation of Caspase 3, nor to a TUNEL signal (Supplementary Figure 3.12A and B). Thus, the induction of apoptosis in the SV40Tt transformed cells was not a general property of the arrest of rRNA gene transcription, suggesting it is specific to UBF depletion.

Given that the iMEFs were initially immortalized by the SV40 Tt oncogene (Sv-T), known to inactivate p53 [59, 60], it was not surprising to find the p53 levels in these cells were constitutively elevated and were not further induced by inactivation of the Ubf gene or by treatment with Staurosporin (Figure 3.4A). Thus, it was unclear whether or not p53 played a role in the apoptotic response in these cells. This question is directly addressed below using homozygous inactivation of the p53 gene. However, it should be noted that inactivation of the Ubf gene in the primary MEFs did not enhance the levels of p53 protein, which remained extremely low throughout (Figure 3.4B).

Figure 3.4 UBF loss induces selective Caspase 3 cleavage in transformed iMEFs cells Figure legend on next page

(A) Ubffl/fl/Er-cre+/+/Sv-T and Ubfwt/wt/Er-cre+/+/Sv-T iMEFs and (B) Ubffl/fl/Er-cre+/+ and Ubfwt/wt/Er-cre+/+ MEFs were assayed for activation (proteolytic cleavage) of Caspase 3 immediately before and at time points after treatment with 4-HT. In (B) “iMEF+Staurosporin” refers to the extract from iMEFs cells treated with 1µM Staurosporin used in (A), and allows a direct comparison of p17 and p53 levels in iMEFs with those in primary MEFs. (C) Electrophoretic fractionation on 1.5% agarose of genomic DNA recovered from Ubffl/fl/Er-cre+/+/Sv-T and Ubfwt/wt/Er-cre+/+/Sv-T iMEFs at different times post tamoxifen treatment (pHT). In (A) to (C), recombination and UBF protein levels were assayed in parallel with each analysis and closely followed those shown in Supplementary Figure 3.10B and C.

3.5.4 Apoptosis is accompanied by the generation of a “nucleosomal ladder” of DNA cleavage Apoptosis is often accompanied by inter-nucleosomal cleavage of genomic DNA to generate a “nucleosomal ladder” [61, 62], due to the result of the release of the nuclease EndoG from mitochondria [63, 64]. Beginning at or before 120h pHT we observed this characteristic nucleosomal fragmentation of genomic DNA in the apoptotic Ubffl/fl/Er-cre+/+/SvT but not in the control Ubfwt/wt/Er-cre+/+/SvT iMEFs (Figure 3.4C), nor in the corresponding primary MEFs (data not shown). Thus, three distinct markers; TUNEL signal, Caspase 3 cleavage and a nucleosomal ladder, indicated that on UBF loss MEFs underwent classic apoptotic cell death after oncogenic transformation with SV40-T, while UBF loss in untransformed MEFs induced none of these markers.

3.5.5 UBF loss blocks proliferation and DNA replication, causing cell cycle arrest To better understand the mechanisms leading to apoptosis in the transformed iMEFs, we determined the effects of Ubf inactivation on cell cycle progression and cell division. Before tamoxifen treatment, the Ubffl/fl/Er- cre+/+/SvT iMEFs displayed a large (~50%) actively replicating S-phase population (Figure 3.5A). Their proliferation was near completely arrested by 48 pHT, corresponding with the elimination of UBF protein and with the near complete shutdown of rRNA synthesis (e.g. see Supplementary Figure 3.10C to E and [11]). By 72h pHT, iMEFs had also stopped active DNA replication and the G2 population abruptly increased at the expense of S-phase cells, while the fraction of G1/G0 cells remained constant (Figure 3.5A and Supplementary Figure 3.13A). Concomitantly, the mitotic index fell to zero as determined by the fraction of cells phosphorylated on serine 28 of histone H3 (H3-S28P) (Figure 3.5C and D and Supplementary Figure 3.13B). Together these data suggested that many apparently G2 iMEFs were unable to complete their passage through mitosis. Parallel analysis of Ubfwt/wt/Er-cre+/+/SvT iMEFs post tamoxifen treatment revealed none of these effects, DNA replication and cell proliferation continuing essentially unabated (Figure 3.5A, C & D and Supplementary Figure 3.13B).

Figure 3.5 UBF loss arrests cell proliferation and leads to a cell cycle arrest (A) Ubffl/fl/Er-cre+/+/Sv-T and Ubfwt/wt/Er-cre+/+/Sv-T iMEFs and (B) the corresponding primary MEFs were analyzed for proliferation and cell cycle distribution at the indicated times post 4-HT treatment. The left- most graphics give cell counts relative to day 0 and include those for Ubfwt/wt/Er-cre+/+ MEFs cultured in the absence of 4-HT (Mock), while to the right of these are shown the cell cycle distributions obtained from FACS analyses for active DNA replication (Click-iT® EdU) and G1 and G2 DNA content (propidium iodide, PI). (C) shows examples of mitotic staining, and (D) a derived graphic of the mitotic index for the Ubffl/fl/Er- cre+/+/Sv-T and Ubfwt/wt/Er-cre+/+/Sv-T iMEFs as determined by the fraction of H3-S28phospho positive cells. In A to D, Ubf recombination and UBF protein levels were assayed in parallel and closely followed those shown in Supplementary Figure 3.10B and C.

100

The situation was somewhat different in the primary Ubffl/fl/Er-cre+/+ and control Ubfwtlwt/Er-cre+/+ MEFs (Figure 3.5B). These cells proliferated more slowly than iMEFs, and only a small fraction (~20%) was ever actively engaged in DNA synthesis. Further, regardless of UBF status these cells gradually arrested DNA replication between 24h and 48h pHT and displayed a corresponding increase in G2 cells, that is up to 24h earlier than for the UBF-null iMEFs. Thus, the primary MEFs underwent a natural slowing or arrest of proliferation regardless of UBF status, while proliferation arrest in the iMEFs was a direct result of the loss of UBF protein. This suggested that the catastrophic cell death observed in the iMEF cultures was related to their inability to assume a quiescent state. In contrast, MEFs naturally arrested proliferation and became quiescent regardless of UBF status or 4-HT treatment (Figure 3.5B), and hence this may have protected them from cell death on inactivation of the Ubf gene. Essentially then, UBF loss specifically targeted the SV40-Tt transformed cells for apoptotic cell death, and what is more the effect was fully penetrant. This suggests that inhibition of UBF or of ribosome biogenesis might represent an ideal target for the development of cancer specific cytotoxic drugs.

3.5.6 Apoptosis induced by UBF loss is p53 independent P53 is often required for the induction of apoptosis, hence its inactivation in many cancers represents a serious limitation to the efficacy of chemo- and radiation therapies [65-67]. The SV40 Tt oncogene is known to inactivate p53 [59, 60], suggesting that apoptosis induced by UBF loss did not depend on functional p53. To directly evaluate the role of p53, we generated p53-null MEFs either wild type or conditional for UBF (Ubffl/fl/Er- cre+/+/p53-/-) (Figures 3.6 and Supplementary Figure 3.14A) and found that they were immortalized and hence could be passaged indefinitely. Despite this, they did not undergo apoptotic cell death on inactivation of the Ubf gene, and displayed neither a TUNEL signal nor caspase 3 cleavage (Figure 3.6A and B). In contrast, after transformation with the SV40 Tt-antigens (SV40-T), the resulting p53-null (Ubffl/fl/Er-cre+/+/p53-/-/Sv-T) iMEFs underwent synchronous and homogeneous TUNEL positive apoptosis two days after loss of UBF, exactly as observed for the p53 positive iMEFs (Figure 3.6C). Thus, even in the complete absence of p53 the loss of UBF was sufficient to induce apoptosis in the SV40-Tt transformed iMEFs. However, in this case no cleavage/activation of Caspase 3 was detected (Figure 3.6D).

101

Figure 3.6 Apoptosis of oncogenically transformed cells after Ubf gene inactivation is p53 independent

(A) Ubffl/fl/Er-cre+/+/p53-/- and Ubfwt/wt/Er-cre+/+/p53-/- MEFs and (C) Ubffl/fl/Er-cre+/+/Sv-T/p53-/- and Ubfwt/wt/Er- cre+/+/Sv-T/p53-/- iMEFs were subjected to a TUNEL reaction and (B) and (D) assayed for activation (proteolytic cleavage) of Caspase 3 immediately before and at several time points after treatment with 4- HT. P53-null iMEFs (Ubffl/fl/Er-cre+/+/Sv-T/p53-/-) displayed the same TUNEL positive cell death, but Caspase 3 cleavage was not detected in these cells. In (B) and (D) “Ctrl” refers to an extract from iMEFs cells treated with 1µM Staurosporin. Recombination of the Ubf gene and UBF protein levels were assayed in parallel and closely followed those shown in Supplementary Figure 3.10B and C.

102

Figure 3.7 p53 independent apoptosis is a general response to UBF loss in an oncogenic stress context (A) Ubffl/fl/Er-cre+/+/Sv-T/p53-/- and their counterpart (B) Ubffl/fl/Er-cre+/+/Ras/p53-/- and (C) Ubffl/fl/Er- cre+/+/Ras/Myc/p53-/- iMEFs cells were subjected to a TUNEL reaction immediately before and at several time points after treatment with 4-HT. All cells synchronously became TUNEL positive at 96h post 4-HT, while neither effect was observed 24h previously. Recombination of the Ubf gene and UBF protein levels were assayed in parallel and closely followed those shown in Supplementary Figure 3.10B and C.

103

3.5.7 p53-independent apoptosis is a general response to UBF loss in oncogene stressed cells It was striking that UBF loss induced fully penetrant apoptosis in SV40-Tt transformed MEFs even in the complete absence p53. To determine if this effect was specific to the SV40-Tt oncogene or occurred under other oncogenic stresses, we investigated UBF-loss in MEFs transformed by the Ras and Myc oncogenes, commonly correlated with human cancers [68]. Ubffl/fl/Er-cre+/+/p53-/- MEFs were transformed by introduction of the Ras oncogene or a combination of the Ras and Myc oncogenes and the effects of inactivation of the Ubf gene were followed. In each case UBF was essentially eliminated by 48h pHT (Supplementary Figure 3.14B) and we observed a synchronous and homogeneous onset of TUNEL-positive apoptosis 48h later, exactly as for SV40-Tt transformation (compare Figure 3.7A with 3.7B and C). Colony forming assays also showed that in each case cell death approached 100% (Supplementary Figure 3.14D). In the case of SV40-Tt and combined Ras/Myc transformation we also observed a “nucleosomal ladder” of apoptotic DNA cleavage starting at 96h pHT, that is at or just after the appearance of the TUNEL signal (Supplementary Figure 3.14C), though this cleavage was not detected in the cells transformed with Ras alone.

3.5.8 Oncogenic stress may induce apoptosis by aberrantly driving cells into S- phase When the untransformed p53-null cells (Ubffl/fl/Er-cre+/+/p53-/-) were analyzed by FACS, we were surprised to find that, quite unlike the SV40-Tt transformed (Ubffl/fl/Er-cre+/+/p53+/+/ Sv-T) iMEFs (Figure 3.5A), UBF depletion caused a significant accumulation of cells in G1 at the expense of the actively replicating S-phase cells (Figure 3.8A). The G2 cell population displayed only a small increase and this anyhow closely resembled that observed for the control Ubfwt/wt/Er-cre+/+/p53-/- cells. In contrast, the Sv-T , Ras and Ras/Myc transformed Ubffl/fl/Er-cre+/+/p53-/- cells displayed the same G2 phase accumulation as seen for the p53-positive iMEFs (compare Figure 3.8B with 3.5A). This suggested that transformation drives cells into and through S-phase regardless of their ability to generate a full complement of ribosomes. Such a situation would be likely to lead to gross replicative errors and hence could explain the highly penetrant apoptosis occurring in both the p53- positive and p53-null transformed MEFs, but not in the untransformed p53-null MEFs.

104

Figure 3.8 Cell cycle distribution of p53-null cells during UBF depletion

(A) Untransformed Ubffl/fl/Er-cre+/+/p53-/- and Ubfwt/wt/Er-cre+/+/p53-/- MEFs. (B) The same p53-null MEFs after transformation with SV40Tt, Ras or Ras plus Myc oncogenes. The graphics show the cell cycle distributions obtained from FACS analyses for active DNA replication (Click-iT® EdU) and G1 and G2 DNA content (propidium iodide, PI).

105

3.6 Discussion Our data suggest that the ability of cisplatin to cause the displacement of UBF from the nucleolus is a key mechanism by which this drug induces selective cell death, since the simple loss of UBF induces a rapid and highly penetrant apoptosis in oncogenically stressed cells. We have shown that conditional deletion of the Ubf gene induces apoptosis specifically in cells transformed by viral and cellular oncogenes. Apoptosis following UBF loss was observed not only in cells expressing SV40Tt, but also in cells expressing the oncogenes Ras and Myc. What is more, in each case apoptosis was found to be fully penetrant, all cells without exception underwent apoptotic cell death. Strikingly, the onset of apoptosis occurred synchronously in all cells two days following complete loss of UBF. Significantly, the induction of TUNEL-positive cell death was completely independent of p53, since it occurred with the same timing and penetrance even after homozygous deletion of the p53 gene. In contrast, before oncogenic transformation primary cell cultures survived complete loss of UBF for many days after the transformed cells entered apoptosis and never underwent apoptosis.

These data strongly suggest that the commonly used chemotherapeutic drug Cisplatin, and by analogy, Carboplatin exert their cytotoxicity in large part by hijacking UBF, displacing it from the nucleolus and inhibiting ribosome biogenesis. In fact, inhibition of ribosome biogenesis may be a more general property of the cytotoxic drugs used in chemotherapy than previously realized, including rapamycin analogs, 5-fluorouracyl and camptothecin [52, 69]. Azacytidine (Azacitidine, Vidaza) and deoxyazacytidine (Decitabine) are DNA methyltransferase inhibitors that have been shown to be active in treating myelodysplastic syndromes and acute myeloid leukemia (AML) [70-72]. The initial studies of azacytidine already showed that it strongly inhibits ribosome biogenesis, and almost certainly does so by preventing rRNA methylation [73, 74]. More recently, deoxyazacitidine was also shown to inhibit ribosome biogenesis by inhibiting rRNA processing, though the underlying mechanism of action is quite different and involves loss of rRNA gene silencing and aberrant RNA polymerase II transcription of these genes [13, 75]. Recent studies of small molecule inhibitors that target ribosome biogenesis have further shown this may be a very valid clinical approach to treating a range of cancers [55, 56, 76, 77]. However, while cell death was independent of p53 in the case of the GC-rich DNA interacting drug BMH-21 [56], it was found to be dependent on a functional p53 in the case of CX-5461, which is believed to target the pre-initiation complex factor SL1 [55]. Our data showing TIF1A/Rrn3-loss does not induce apoptosis even in the presence of p53 clearly excludes the explanation that the cytotoxicity of these drugs is simply a function of their ability to suppress rRNA synthesis. Why then inhibition of UBF can induce apoptotic cell death with such penetrance and in the complete absence of p53 is for the still a matter of conjecture. However, it is amost certainly related to the role of UBF in forming a specialized chromatin structure on the active rRNA genes [11]. Loss of this structure would yield the rRNA gene arrays highly

106

susceptible to damage, and given the GC-richness of the rRNA genes the same could be argued for both cisplatin and BMH-21 drugs.

The Nucleolar Organizer Regions (NORs) each encompass around 40 rRNA gene units on the short arms of the five human acrocentric chromosomes [78]. These loci are particularly susceptible to DNA breakage and are subject to high levels of inter- and intra-chromosomal recombination [79-81]. Indeed, Robertsonian translocations have long been known to predominantly involve exchanges between the short arms of human acrocentric chromatids that often create fusions with chromatids of a metacentric chromosome [82]. Recent data strongly suggests that these and similar chromosome translocations result from disruption of the active chromatin structure of the rRNA genes, which in turn affects chromosome pairing causing aberrant resolution of mitotic chiasmata and fusion between non-homologous chromatids [83]. Loss of UBF clearly disrupts the chromatin structure of the rRNA genes, leaving them at least transiently as naked DNA, and would necessarily leave these genes highly susceptible to DNA damage and breakage. Since transformed iMEFs continue replication during UBF depletion, the disruption of rRNA gene chromatin would exacerbate the effects of DNA breakage, probably inhibit homologous repair processes and hence destabilize the genome. Indeed such destabilization has recently been observed as a result of siRNA knockdown of UBF [84].

3.7 Materials & Methods Isolation and cultures of MEFs and iMEFs. The generation of conditional Ubffl/flEr-cre+/+ and control mouse lines was previously described [11]. The p53- null allele was introduced by crossing to strain 129-Trp53tm1Tyj/J (Jackson Laboratory Stock # 002080). Primary mouse embryonic fibroblasts (MEFs) from E14.5 Ubffl/fl/Er-cre+/+ and isogenetic Ubfwt/wtEr-cre+/+ MEFs and corresponding p53-/- MEFs were prepared as previously described [11, 85]. Cells were cultured in Dulbecco’s modified Eagle medium (DMEM)-high glucose (Life Technologies), supplemented with 10% fetal bovine serum (Wisent), L-glutamine (Life Technologies) and Antibiotic/Antimycotic (Wisent). Where indicated, Cisplatin (Sandoz) was added to the cell culture medium from a 100mM solution in DMSO to give a final concentration of 30µM and cells incubated for 4 hr at 37oC. The culture medium was then replaced with medium without cisplatin and cells incubated for a further 16h at 37oC, before processing for immunofluorescence as described below. MEFs were immortalized by the introduction of the SV40 Tt antigens by transfection with the pBSV0.3T/t, a modification of the pBSV-early vector [86] kindly provided by E. W. Khandjian. The Ras and Ras/Myc transformed MEFs were generated by transfection or co-transfection with the plasmids pWZL-Ras- hygro and pBabe-c-myc-puro (kind gifts from Gerardo Ferbeyre) into Ubffl/fl/ Er-cre+/+/p53-/- MEFs and subsequent hygromycin or double hygromycin/puromycin selection.

Inactivation of Ubf or Tif1a in cell culture, and analysis of genotype, RNA and proteins. As previously described [11], cells were initially plated in 6 cm petri dishes (0.8x106 cells each) and cultured for 18 hours in DMEM, high glucose, 10% fetal bovine serum. To activate ER-Cre, 4-hydroxytamoxifen (4-HT) was added to a final concentration of 50nM, and after 4 hr incubation the medium replaced with fresh medium without 4-HT and cells harvested for analysis at various time points. In the case of Tif1a, cells were treated with 50nM 4-HT, 0h, then this treatment was repeated at 9h, 24h and 33h later to ensure complete gene

107

excision. Analyses of RNA, protein and genotype were systematically carried out on parallel cell cultures. Cells were genotyped by PCR before and after 4-HT treatment using the primers: A; 5’TGATCCCTCCCTTTCTGATG, B; 5’TGGGGATAGGCCTTAGAGAGA, C; 5’CACGGGAAAACAAGGTCACT, (Figure S2B). Metabolic labelling of RNA was carried out just before cell harvesting by addition of 10 Ci [3H]- uridine (PerkinElmer) to the culture medium and incubation for a further 3h. RNA was extracted with Trizol (Life Technologies) according to the manufacturer’s protocol and analyzed by gel electrophoresis, fluoroimaging (ENHance, PerkinElmer) and RNA species quantitated by scintillation counting as previously described [39, 40]. For total protein, cells were washed with cold PBS, scraped into PBS, centrifuged 30 s at 14 000 r.p.m., then resuspended in sodium dodecyl sulphate (SDS) loading buffer. After fractionation on 8%, 12% or 5–15% gradient SDS–polyacrylamide gel electrophoresis (SDS-PAGE [87]), cell extracts were analysed by standard Western blotting procedures.

Chromatin Immunoprecipitations (ChIP) ChIP was performed as previously described [11, 88]. The amplicon coordinates relative to the 47S rRNA initiation site (BK000964) were as follows: 47SPr, 45133-40; ETS, 3078-3221; ITS1, 6258-6432; 28S, 10215- 10411; T1-3, 13412-13607.

Antibodies for Western Blot, Immunofluorescence and ChIP Rabbit antibodies against UBF, RPI large subunit (A194), TTF1 and TIF1A were generated in the laboratory. All other antibodies were obtained commercially; Anti-Caspase-3, -p53 and -H3S-28phospho (Cell Signalling), anti-Tubulin (Sigma) and anti-Fibrillarin (Covance).

Immunofluorescence Cells were washed with PBS, fixed in 4% paraformaldehyde /PBS for 15 minutes and permeabilized with 0.5 % Triton/PBS for 5 minutes. Incubation with primary antibody was performed for 1h in PBS-5% BSA or 5% goat serum and cells were stained with AlexaFluor 488/568 conjugated anti-rabbit or -mouse IgG (Molecular Probes) and counterstained with DAPI. After mounting in 50% glycerol/50% 0.2 M Na-glycine, 0.3 M NaCl, 3D epifluorescent image stacks were generated on a Leica DMI6000B microscope equipped with a 63x or 100x objective and an Orca C4742-80-12AG camera (Hamamatsu). Image stacks were then deconvoluted and analyzed using Volocity software (Perkin-Elmer Improvision). Alternatively, image stacks were generated on a Leica SP5-II confocal microscope equipped with a 63x objective and running in standard scanning mode, and analyzed using Volocity software (Perkin-Elmer Improvision).

FACs analysis and determination of Mitotic Index. Cells were stained for ongoing DNA synthesis using the Click-iT® EdU Alexa Fluor® 647 Flow Cytometry Assay Kit (Life Technologies) following the manufacturer's protocol and subsequently with propidium iodide (PI) immediately before analysis by the cytometry service of the CHU de Québec Research Centre using a FACSCanto II flow cytometer and FACSDiva 6.1.2 software (Becton Dickinson). Parallel cultures were stained with anti-H3S-28phospho antibody and DAPI and imaged by epifluorescence on the Leica DMI6000B microscope using 20 and 40x objectives. The Mitotic Index was calculated as the ratio of H3S-28phospho- positive to DAPI positive nuclei.

108

Tunnel Assays Tunnel assays were performed with a Click-It Tunnel assay kit, Alexa 488 Imaging System, (Life Technologies). Cells were seeded in 35mM petri dishes, fixed and processed according to the manufacturer’s protocol and visualized by epifluorescence on the Leica DMI6000 B microscope using a 20x objective.

Colony formation assays The SV40-T, Ras and Ras/Myc transformed Ubffl/fl/Er-cre+/+/p53-/- and the isogenic wild-type MEF cells cultured in 100mm petri dishes were treated with 50nM 4-HT (Sigma) on day 0. The medium was changed after four hours to remove 4-HT, and on day 2 each culture was replated in duplicate at dilutions of 10 000, 50 000, 100 000, and 200 000 cells per 60mm petri. On day 6 and day 12 petri dishes were fixed for 5mins with 4% paraformaldehyde/PBS and stained with 0.05% crystal violet in distilled water (filtered) for 30mins. Petri dishes were then washed 3 times with water and left inverted to dry before being photographed.

MitoTracker assays Cells were plated in ibidi 35mm thin bottom petri dishes for subsequent live cell microscopy and treated for 4h with 50nM 4-HT (Sigma) and further cultured as standard for 96h to induce UBF loss. Cells were then treated with 25nM MitoTracker DeepRedTM (Life Technologies) for 20mins at 37°C in DMEM minus serum. Petri dishes were washed once with DMEM minus serum and then incubated in FluoroBrite DMEM (Life Technologies). Finally, live image stacks were generated on the Leica SP5-II confocal microscope and analyzed using Volocity software (Perkin-Elmer Improvision).

3.8 Acknowledgements We wish to thank Dr Lucie Jeannotte for providing the p53 null allele mice and advice on their use, Dr Ross Hannan and Elaine Sanij for discussion at various stages of this work, and Dr A. Brunet of the Cytology Laboratory of the Research Centre of the Québec University Hospital Centre (RC-CHUQ) for FACS analyses. This work was funded by operating grants from the Canadian Institutes of Health Research (CIHR, MOP12205) and from the Cancer Research Society (CRS/SRC). The Research Centre of the Québec University Hospital Centre (CHU de Québec) is supported by the Fonds de recherche du Québec - Santé (FRQS).

3.9 References 1. Rabik CA and Dolan ME. Molecular mechanisms of resistance and toxicity associated with platinating agents. Cancer Treat Rev. 2007; 33(1):9-23. 2. Wang Z, Xu J, Zhou JY, Liu Y and Wu GS. Mitogen-activated protein kinase phosphatase-1 is required for cisplatin resistance. Cancer Res. 2006; 66(17):8870-8877. 3. Kim M, Yan Y, Kortum RL, Stoeger SM, Sgagias MK, Lee K, Lewis RE and Cowan KH. Expression of kinase suppressor of Ras1 enhances cisplatin-induced extracellular signal-regulated kinase activation and cisplatin sensitivity. Cancer Res. 2005; 65(10):3986-3992. 4. Cavallo F, Feldman DR and Barchi M. Revisiting DNA damage repair, p53-mediated apoptosis and cisplatin sensitivity in germ cell tumors. Int J Dev Biol. 2013; 57(2-4):273-280. 5. Jacobsen C and Honecker F. Cisplatin resistance in germ cell tumours: models and mechanisms. Andrology. 2014. 6. Boulikas T. DNA lesion-recognizing proteins and the p53 connection. Anticancer Res. 1996; 16(1):225-242.

109

7. Wozniak K and Blasiak J. Recognition and repair of DNA-cisplatin adducts. Acta Biochim Pol. 2002; 49(3):583-596. 8. Reeves R and Adair JE. Role of high mobility group (HMG) chromatin proteins in DNA repair. DNA Repair (Amst). 2005; 4(8):926-938. 9. Pil PM and Lippard SJ. Specific binding of chromosomal protein HMG1 to DNA damaged by the anticancer drug cisplatin. Science. 1992; 256:234-237. 10. Hoffmann JS, Locker D, Viliani G and Leng M. HMG1 protein inhibits the translesion synthesis of the major DNA cisplatin adduct by cell extracts. JMolBiol. 1997; 270:539-543. 11. Hamdane N, Stefanovsky VY, Tremblay MG, Nemeth A, Paquet E, Lessard F, Sanij E, Hannan R and Moss T. Conditional inactivation of Upstream Binding Factor reveals its epigenetic functions and the existence of a somatic nucleolar precursor body. PLoS Genet. 2014; 10(8):e1004505. 12. Merz K, Hondele M, Goetze H, Gmelch K, Stoeckl U and Griesenbeck J. Actively transcribed rRNA genes in S. cerevisiae are organized in a specialized chromatin associated with the high-mobility group protein Hmo1 and are largely devoid of histone molecules. Genes Dev. 2008; 22(9):1190-1204. 13. Gagnon-Kugler T, Langlois F, Stefanovsky V, Lessard F and Moss T. Loss of human ribosomal gene CpG methylation enhances cryptic RNA polymerase II transcription and disrupts ribosomal RNA processing. Mol Cell. 2009; 35(4):414-425. 14. Chao JC, Wan XS, Engelsberg BN, Rothblum LI and Billings PC. Intracellular distribution of HMG1, HMG2 and UBF change following treatment with cisplatin. Biochim Biophys Acta. 1996; 1307(2):213-219. 15. Delmani FA, Torreblanca J, Moreno J, Garcia-Herdugo G, Vilaplana R and Gonzalez-Viltchez F. In vitro expression and redistribution of nucleolar proteins following the treatment with cis-dichloro-1,2- propylenediamine-N,N,N',N'-tetraacetato ruthenium (III) (RAP). Oncology research. 2009; 17(9):425-435. 16. Zhai XQ, Beckmann H, Jantzen HM and Essigmann JM. Cisplatin-DNA adducts inhibit ribosomal RNA synthesis by hijacking the transcription factor human upstream binding factor. Biochemistry. 1998; 37:16307-16315. 17. Jordan P and Carmo-Fonseca M. Cisplatin inhibits synthesis of ribosomal RNA in vivo. Nucleic Acids Res. 1998; 26:2831-2836. 18. Codony-Servat J, Gimeno R, Gelpi C, Rodriguez-Sanchez JL and Juarez C. The two isoforms of the 90- kDalton nucleolus organizer region autoantigen (upstream binding factor) bind with different avidity to DNA modified by the antitumor drug cisplatin. Biochemical Pharmacology. 1996; 51:1131-1136. 19. Treiber DK, Zhai X, Jantzen HM and Essigmann JM. Cisplatin-DNA adducts are molecular decoys for the ribosomal RNA transcription factor hUBF (human upstream binding factor). ProcNatlAcadSciUSA. 1994; 91:5672-5676. 20. Grummt I. Life on a planet of its own: regulation of RNA polymerase I transcription in the nucleolus. Genes Dev. 2003; 17(14):1691-1702. 21. Moss T, Langlois F, Gagnon-Kugler T and Stefanovsky V. A housekeeper with power of attorney: the rRNA genes in ribosome biogenesis. Cell Mol Life Sci. 2007; 64(1):29-49. 22. Russell J and Zomerdijk JC. RNA-polymerase-I-directed rDNA transcription, life and works. Trends Biochem Sci. 2005; 30(2):87-96. 23. Sanij E, Poortinga G, Sharkey K, Hung S, Holloway TP, Quin J, Robb E, Wong LH, Thomas WG, Stefanovsky V, Moss T, Rothblum L, Hannan KM, McArthur GA, Pearson RB and Hannan RD. UBF levels determine the number of active ribosomal RNA genes in mammals. J Cell Biol. 2008; 183(7):1259-1274. 24. Bazett-Jones DP, Leblanc B, Herfort M and Moss T. Short-range DNA looping by the Xenopus HMG-box transcription factor, xUBF. Science. 1994; 264:1134-1137. 25. Stefanovsky VY, Pelletier G, Bazett-Jones DP, Crane-Robinson C and Moss T. DNA looping in the RNA polymerase I enhancesome is the result of non-cooperative in-phase bending by two UBF molecules. Nucleic Acids Res. 2001; 29(15):3241-3247. 26. Stefanovsky VY and Moss T. The splice variants of UBF differentially regulate RNA polymerase I transcription elongation in response to ERK phosphorylation. Nucleic Acids Res. 2008; 36(15):5093-5101.

110

27. Stefanovsky VY, Langlois F, Pelletier G, Bazett-Jones DP and Moss T. ERK modulates DNA bending and Enhancesome Structure by phosphorylating HMG1-boxes 1 and 2 of the RNA polymerase I transcription factor UBF. Biochemistry. 2006; 45(11):3626-3634. 28. O'Sullivan AC, Sullivan GJ and McStay B. UBF binding in vivo is not restricted to regulatory sequences within the vertebrate ribosomal DNA repeat. Mol Cell Biol. 2002; 22(2):657-668. 29. Woolford JL, Jr. and Baserga SJ. Ribosome biogenesis in the yeast Saccharomyces cerevisiae. Genetics. 2013; 195(3):643-681. 30. Derenzini M, Trerè D, Pession A, Montanaro L, Sirri V and Ochs RL. Nucleolar function and size in cancer cells. AmJPathol. 1998; 152:1291-1297. 31. Montanaro L, Trere D and Derenzini M. Nucleolus, ribosomes, and cancer. Am J Pathol. 2008; 173(2):301- 310. 32. Derenzini M, Montanaro L and Trere D. What the nucleolus says to a tumour pathologist. Histopathology. 2009; 54(6):753-762. 33. Guttman PH and Halpern S. Nuclear-Nucleolar Volume Ratio in Cancer. The American Journal of Cancer. 1935; 25(4):802-806. 34. Derenzini M, Trerè D, Pession A, Govoni M, Sirri V and Chieco P. Nucleolar size indicates the rapidity of cell proliferation in cancer tissues. J Pathol. 2000; 191(2):181-186. 35. Poortinga G, Hannan KM, Snelling H, Walkley CR, Jenkins A, Sharkey K, Wall M, Brandenburger Y, Palatsides M, Pearson RB, McArthur GA and Hannan RD. MAD1 and c-MYC regulate UBF and rDNA transcription during granulocyte differentiation. Embo J. 2004; 23(16):3325-3335. 36. Arabi A, Wu S, Ridderstrale K, Bierhoff H, Shiue C, Fatyol K, Fahlen S, Hydbring P, Soderberg O, Grummt I, Larsson L-G and Wright APH. c-Myc associates with ribosomal DNA and activates RNA polymerase I transcription. Nat Cell Biol. 2005; 7(3):303-310. 37. Zhai WG, Tuan JA and Comai L. SV40 large T antigen binds to the TBP-TAFI complex SL1 and coactivates ribosomal RNA transcription. GenesDev. 1997; 11:1605-1617. 38. Tuan JC, Zhai WG and Comai L. Recruitment of TATA-binding protein-TAF, complex SL1 to the human ribosomal DNA promoter is mediated by the carboxy-terminal activation domain of upstream binding factor (UBF) and is regulated by UBF phosphorylation. MolCell Biol. 1999; 19:2872-2879. 39. Stefanovsky V, Langlois F, Gagnon-Kugler T, Rothblum LI and Moss T. Growth factor signaling regulates elongation of RNA polymerase I transcription in mammals via UBF phosphorylation and r-chromatin remodeling. Mol Cell. 2006; 21(5):629-639. 40. Stefanovsky VY, Pelletier G, Hannan R, Gagnon-Kugler T, Rothblum LI and Moss T. An immediate response of ribosomal transcription to growth factor stimulation in mammals is mediated by ERK phosphorylation of UBF. Mol Cell. 2001; 8(5):1063-1073. 41. Zhao J, Yuan X, Frodin M and Grummt I. ERK-Dependent Phosphorylation of the Transcription Initiation Factor TIF-IA Is Required for RNA Polymerase I Transcription and Cell Growth. Mol Cell. 2003; 11(2):405- 413. 42. Mayer C, Zhao J, Yuan X and Grummt I. mTOR-dependent activation of the transcription factor TIF-IA links rRNA synthesis to nutrient availability. Genes Dev. 2004; 18(4):423-434. 43. Hannan KM, Brandenburger Y, Jenkins A, Sharkey K, Cavanaugh A, Rothblum L, Moss T, Poortinga G, McArthur GA, Pearson RB and Hannan RD. mTOR-dependent regulation of ribosomal gene transcription requires S6K1 and is mediated by phosphorylation of the carboxy-terminal activation domain of the nucleolar transcription factor UBF. Mol Cell Biol. 2003; 23(23):8862-8877. 44. Zhai W and Comai L. Repression of RNA polymerase I transcription by the tumor suppressor p53. Mol Cell Biol. 2000; 20(16):5930-5938. 45. Zhang Y, Saporita AJ and Weber JD. P19ARF and RasV(1)(2) offer opposing regulation of DHX33 translation to dictate tumor cell fate. Mol Cell Biol. 2013; 33(8):1594-1607. 46. Lessard F, Morin F, Ivanchuk S, Langlois F, Stefanovsky V, Rutka J and Moss T. The ARF tumor suppressor controls ribosome biogenesis by regulating the RNA polymerase I transcription factor TTF-I. Mol Cell. 2010; 38(4):539-550.

111

47. Sugimoto M, Kuo ML, Roussel MF and Sherr CJ. Nucleolar Arf tumor suppressor inhibits ribosomal RNA processing. Mol Cell. 2003; 11(2):415-424. 48. Cavanaugh AH, Hempel WM, Taylor LJ, Rogalsky V, Todorov G and Rothblum LI. Activity of RNA polymerase I transcription factor UBF blocked by Rb gene product. Nature. 1995; 374:177-180. 49. Pelletier G, Stefanovsky VY, Faubladier M, Hirschler-Laszkiewicz I, Savard J, Rothblum LI, Cote J and Moss T. Competitive recruitment of CBP and Rb-HDAC regulates UBF acetylation and ribosomal transcription. Mol Cell. 2000; 6(5):1059-1066. 50. Zhang C, Comai L and Johnson DL. PTEN represses RNA Polymerase I transcription by disrupting the SL1 complex. Mol Cell Biol. 2005; 25(16):6899-6911. 51. Bursac S, Brdovcak MC, Donati G and Volarevic S. Activation of the tumor suppressor p53 upon impairment of ribosome biogenesis. Biochim Biophys Acta. 2014; 1842(6):817-830. 52. Rubbi CP and Milner J. Disruption of the nucleolus mediates stabilization of p53 in response to DNA damage and other stresses. Embo J. 2003; 22(22):6068-6077. 53. Pestov DG, Strezoska Z and Lau LF. Evidence of p53-dependent cross-talk between ribosome biogenesis and the cell cycle: effects of nucleolar protein Bop1 on G(1)/S transition. Mol Cell Biol. 2001; 21(13):4246- 4255. 54. Fumagalli S, Di Cara A, Neb-Gulati A, Natt F, Schwemberger S, Hall J, Babcock GF, Bernardi R, Pandolfi PP and Thomas G. Absence of nucleolar disruption after impairment of 40S ribosome biogenesis reveals an rpL11-translation-dependent mechanism of p53 induction. Nat Cell Biol. 2009; 11(4):501-508. 55. Bywater MJ, Poortinga G, Sanij E, Hein N, Peck A, Cullinane C, Wall M, Cluse L, Drygin D, Anderes K, Huser N, Proffitt C, Bliesath J, Haddach M, Schwaebe MK, Ryckman DM, et al. Inhibition of RNA Polymerase I as a Therapeutic Strategy to Promote Cancer-Specific Activation of p53. Cancer Cell. 2012; 22(1):51-65. 56. Peltonen K, Colis L, Liu H, Trivedi R, Moubarek MS, Moore HM, Bai B, Rudek MA, Bieberich CJ and Laiho M. A Targeting Modality for Destruction of RNA Polymerase I that Possesses Anticancer Activity. Cancer Cell. 2014; 25(1):77-90. 57. Duenas-Gonzalez A, Cetina L, Coronel J and Cervantes-Madrid D. Emerging drugs for cervical cancer. Expert opinion on emerging drugs. 2012; 17(2):203-218. 58. Ho KF, Swindell R and Brammer CV. Dose intensity comparison between weekly and 3-weekly Cisplatin delivered concurrently with radical radiotherapy for head and neck cancer: a retrospective comparison from New Cross Hospital, Wolverhampton, UK. Acta Oncol. 2008; 47(8):1513-1518. 59. Mietz JA, Unger T, Huibregtse JM and Howley PM. The transcriptional transactivation function of wild-type p53 is inhibited by SV40 large T-antigen and by HPV-16 E6 oncoprotein. Embo J. 1992; 11(13):5013-5020. 60. Lin JY and Simmons DT. The ability of large T antigen to complex with p53 is necessary for the increased life span and partial transformation of human cells by simian virus 40. J Virol. 1991; 65(12):6447-6453. 61. Duke RC, Chervenak R and Cohen JJ. Endogenous endonuclease-induced DNA fragmentation: an early event in cell-mediated cytolysis. Proc Natl Acad Sci U S A. 1983; 80(20):6361-6365. 62. Wyllie AH. Glucocorticoid-induced thymocyte apoptosis is associated with endogenous endonuclease activation. Nature. 1980; 284(5756):555-556. 63. Cote J and Ruiz-Carrillo A. Primers for mitochondrial DNA replication generated by endonuclease G. Science. 1993; 261(5122):765-769. 64. Li LY, Luo X and Wang X. Endonuclease G is an apoptotic DNase when released from mitochondria. Nature. 2001; 412(6842):95-99. 65. Lowe SW, Bodis S, McClatchey A, Remington L, Ruley HE, Fisher DE, Housman DE and Jacks T. p53 status and the efficacy of cancer therapy in vivo. Science. 1994; 266(5186):807-810. 66. Lowe SW, Ruley HE, Jacks T and Housman DE. p53-dependent apoptosis modulates the cytotoxicity of anticancer agents. Cell. 1993; 74:957-967. 67. Weller M. Predicting response to cancer chemotherapy: the role of p53. Cell Tissue Res. 1998; 292(3):435- 445. 68. Markert EK, Levine AJ and Vazquez A. Proliferation and tissue remodeling in cancer: the hallmarks revisited. Cell death & disease. 2012; 3:e397.

112

69. Gibbons JJ, Abraham RT and Yu K. Mammalian target of rapamycin: discovery of rapamycin reveals a signaling pathway important for normal and cancer cell growth. Semin Oncol. 2009; 36 Suppl 3:S3-S17. 70. Datta J, Ghoshal K, Denny WA, Gamage SA, Brooke DG, Phiasivongsa P, Redkar S and Jacob ST. A new class of quinoline-based DNA hypomethylating agents reactivates tumor suppressor genes by blocking DNA methyltransferase 1 activity and inducing its degradation. Cancer Res. 2009; 69(10):4277-4285. 71. Flotho C, Claus R, Batz C, Schneider M, Sandrock I, Ihde S, Plass C, Niemeyer CM and Lubbert M. The DNA methyltransferase inhibitors azacitidine, decitabine and zebularine exert differential effects on cancer gene expression in acute myeloid leukemia cells. Leukemia. 2009; 23(6):1019-1028. 72. Tsai HC, Li H, Van Neste L, Cai Y, Robert C, Rassool FV, Shin JJ, Harbom KM, Beaty R, Pappou E, Harris J, Yen RW, Ahuja N, Brock MV, Stearns V, Feller-Kopman D, et al. Transient low doses of DNA- demethylating agents exert durable antitumor effects on hematological and epithelial tumor cells. Cancer Cell. 2012; 21(3):430-446. 73. Cihak A, Weiss JW and Pitot HC. Characterization of polyribosomes and maturation of ribosomal RNA in hepatoma cells treated with 5-azacytidine. Cancer Res. 1974; 34(11):3003-3009. 74. Cihak A. Biological effects of 5-azacytidine in eukaryotes. Oncology. 1974; 30(5):405-422. 75. Moss T. DNA methyltransferase inhibition may limit cancer cell growth by disrupting ribosome biogenesis. Epigenetics. 2011; 6(2):128 - 133. 76. Drygin D, Lin A, Bliesath J, Ho CB, O'Brien SE, Proffitt C, Omori M, Haddach M, Schwaebe MK, Siddiqui- Jain A, Streiner N, Quin JE, Sanij E, Bywater MJ, Hannan RD, Ryckman D, et al. Targeting RNA Polymerase I with an Oral Small Molecule CX-5461 Inhibits Ribosomal RNA Synthesis and Solid Tumor Growth. Cancer Res. 2011; 71(4):1418-1430. 77. Rothblum K, Hu Q, Penrod Y and Rothblum LI. Selective inhibition of rDNA transcription by a small-molecule peptide that targets the interface between RNA polymerase I and Rrn3. Mol Cancer Res. 2014; 12(11):1586- 1596. 78. McStay B and Grummt I. The epigenetics of rRNA genes: from molecular to chromosome biology. Annu Rev Cell Dev Biol. 2008; 24:131-157. 79. Tchurikov NA, Fedoseeva DM, Sosin DV, Snezhkina AV, Melnikova NV, Kudryavtseva AV, Kravatsky YV and Kretova OV. Hot spots of DNA double-strand breaks and genomic contacts of human rDNA units are involved in epigenetic regulation. J Mol Cell Biol. 2014. 80. Postepska-Igielska A, Krunic D, Schmitt N, Greulich-Bode KM, Boukamp P and Grummt I. The chromatin remodelling complex NoRC safeguards genome stability by heterochromatin formation at telomeres and centromeres. EMBO Rep. 2013; 14(8):704-710. 81. Ide S, Miyazaki T, Maki H and Kobayashi T. Abundance of ribosomal RNA gene copies maintains genome integrity. Science. 2010; 327(5966):693-696. 82. Therman E, Susman B and Denniston C. The nonrandom participation of human acrocentric chromosomes in Robertsonian translocations. Annals of human genetics. 1989; 53(Pt 1):49-65. 83. Stimpson KM, Sullivan LL, Kuo ME and Sullivan BA. Nucleolar organization, ribosomal DNA array stability, and acrocentric chromosome integrity are linked to telomere function. PLoS One. 2014; 9(3):e92432. 84. Sanij E, Diesch J, Lesmana A, Poortinga G, Lidgerwood G, Hein N, Cameron DP, Ellul J, Goodall GJ, Wong LH, Dhillon AS, Hamdane N, Rothblum LI, Pearson RB, Haviv I, Moss T, et al. A novel role for the Pol I transcription factor UBTF in maintaining genome stability through the regulation of highly transcribed Pol II genes. Genome Res. 2014. 85. Giroux S, Tremblay M, Bernard D, Cadrin-Girard JF, Aubry S, Larouche L, Rousseau S, Huot J, Landry J, Jeannotte L and Charron J. Embryonic death of Mek1-deficient mice reveals a role for this kinase in angiogenesis in the labyrinthine region of the placenta. Curr Biol. 1999; 9:369-372. 86. Schaffner W. Direct transfer of cloned genes from bacteria to mammalian cells. Proceedings of the National Academy of Sciences of the United States of America. 1980; 77(4):2163-2167. 87. Laemmli UK. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature. 1970; 227:680-685. 88. Nelson JD, Denisenko O and Bomsztyk K. Protocol for the fast chromatin immunoprecipitation (ChIP) method. Nat Protoc. 2006; 1(1):179-185.

113

3.10 Supplemental data Figure 3.9, related to Figures 3.1 and 3.2 Figure 3.10, related to Figures 3.3 – 3.7 Figure 3.11, related to Figures 3.3 – 3.5 Figure 3.12, related to Figures 3.3 and 3.4 Figure 3.13, related to Figure 3.5 Figure 3.14, related to Figures 3.6 – 3.8

Figure 3.9 Cisplatin treatment of MEFs induces displacement of UBF from the nucleolus NIH3T3 cultures were treated with 30M cisplatin for 4h in full medium or left untreated (0), then subjected to indirect immunofluorescence analysis of UBF (green), fibrillarin (red) and DNA stained with DAPI (blue). A) Single optical sections from image stacks of two untreated and two cisplatin treated cell nuclei are shown. B) 3D isosurface images generated from further examples of image stacks using Volocity (Perkin- Elmer), colours as in A).

114

Figure 3.10 Analysis of UBF loss in primary MEFs

Figure legend on next page

115

Figure 3.10 Analysis of UBF loss in primary MEFs (cont)

A) to C) Induction of Cre activity in Ubffl/fl/Er-cre+/+ cells using only a 4h treatment with 50 nM 4- hydroxytamoxifen (4-HT) induced the complete excision of the floxed Ubf exons by 24h post 4-HT treatment (pHT), and the UBF protein was no longer detectable at 48 h pHT. D) and E) Concomitantly, we observed a 90% reduction of rRNA synthesis, far in excess of that occurring as the control Ubfwt/wt /Er-cre+/+ primary MEFs naturally reached confluence [1, 2], and rRNA synthesis was abolished by 72h pHT. F) and G) Further, loss of UBF caused the formation of a typical dense nucleolar protein body previously described in Ubffl/fl/Er-cre+/+/SvT iMEFs [3]. Immunofluorescence analysis of F) RPI (large subunit, A194) and G) TTF1 relative to Fibrillarin and DNA stained with DAPI before and after Ubf gene inactivation. The right-most panels show enlargements of single nucleolar bodies. Thus, UBF loss in primary MEFs recapitulated the effects previously observed in the SvT transformed MEFs (iMEFs).

116

Figure 3.11 Primary MEFs survive UBF loss while SV40Tt transformed iMEFs suffer cell death A) UBF-loss causes morphological changes in Ubffl/fl/Er-cre+/+Sv-T iMEFs but not in the Ubffl/fl/Er-cre+/+ MEFs. B) Parallel analysis of iMEF and MEF viable cells during inactivation of the Ubf gene. The trypan blue negative cell count for each cell type has been normalized to that before activation of ER-Cre with tamoxifen (0h pHT). C) MitoTracker staining of unfixed cultures of conditional iMEFs and MEFs before (0h pHT) and after (96h pHT) UBF loss. D) Colony Forming Assays for Ubffl/fl/Er-cre+/+Sv-T and matched Ubfwt/wt/Er-cre+/+Sv-T iMEFs treated for 4 hours with 50nM Tamoxifen at zero time and cultivated for a further 144 hours before fixing and staining cells.

117

Figure 3.12 TIF1A loss does not induce TUNEL positive apoptosis in SV40Tt transformed MEFs A) Time course of 4-HT treatment of Tif1a fl/fl/Er-cre+/+/Sv-T and Tif1a wt/wt/Er-cre+/+/Sv-T MEFs derived from crosses of mice conditional for Tif1a [4]. The upper panels show the endogenous TIF1A levels and the central panel control Tubulin levels. The lower panel the status of Caspase 3 showing the p17 fragment is not detected. B) TUNEL assays of the same cells before and after TIF1A depletion. No TUNEL signal was detected unless cells were treated with DNase1 as control for the detection assay.

118

A) Post 4-HT: 0h 48h 72h 105

104

C P 3

A 10 S

D S SS

E 102 0 G1 G2 G1 G2 G1 G2 10-2 0 50 100 150 200 0 50 100 150 200 0 50 100 150 200 Propidium Iodide (PI) Ubf f /f /Er-cre+/+/Sv-T

B) Ubff /f /Er-cre+/+/Sv-T hours post 4-HT: 0 48 72 96 200um

Merge

H3-S28P

Ubfwt/wt/Er-cre+/+/Sv-T hours post 4-HT: 0 48 72 96 200um

Merge

H3-S28P

Figure 3.13 UBF loss leads to a cell cycle arrest and to a loss of mitotic cells A) UbfFigurfl/fle/Er S5-cre+/+/Sv-T iMEFs were analyzed for cell cycle distribution and B) mitotic index at the indicated times post 4-HT treatment. A) Shows an example of FACS analyses for active DNA replication (Click-iT® EdU) and DNA content (PI) at 0, 48 and 72h post 4-HT treatment. B) Shows examples of immunofluorescence staining for phospho-serine 28 of histone H3. The mitotic index shown in Figure 5D was determined as the fraction of H3-S28 phospho-positive cells.

119

Figure 3.14 p53-independent apoptosis is a general response to UBF loss in an oncogenic stress context A) Genotyping of p53-/- and p53+/+ iMEFs using the Jackson Laboratory protocol (http://jaxmice.jax.org/protocolsdb/f?p=116:2:0::NO:2:P2_MASTER_PROTOCOL_ID,P2_JRS_CODE:1432 3,002080) “M” indicates the DNA ladder size marker used. B) Time-course of UBF-loss in SV40Tt, Ras and Ras/Myc transformed p53-null UBF conditional MEFs treated with tamoxifen. C) Analysis of DNA degradation in the same cells and D) Colony forming capacity of these same iMEFs cultures before and 144h after tamoxifen treatment.

120

3.10.1 Supplemental data references

1. Eagle H and Levine EM. Growth regulatory effects of cellular interaction. Nature. 1967; 213(5081):1102- 1106. 2. Abercrombie M. Contact inhibition and malignancy. Nature. 1979; 281(5729):259-262. 3. Hamdane N, Stefanovsky VY, Tremblay MG, Nemeth A, Paquet E, Lessard F, Sanij E, Hannan R and Moss T. Conditional inactivation of Upstream Binding Factor reveals its epigenetic functions and the existence of a somatic nucleolar precursor body. PLoS Genet. 2014; 10(8):e1004505. 4. Yuan X, Zhou Y, Casanova E, Chai M, Kiss E, Grone HJ, Schutz G and Grummt I. Genetic Inactivation of the Transcription Factor TIF-IA Leads to Nucleolar Disruption, Cell Cycle Arrest, and p53-Mediated Apoptosis. Mol Cell. 2005; 19(1):77-87.

121

Chapter 4 The role of ribosomal DNA in maintaining or establishing pluripotency

Chelsea Herdman, Michel Tremblay, Jean-Clément Mars, Victor Stefanovsky, Marianne Sabourin-Félix, and Tom Moss

Laboratory of Growth and Development, St-Patrick Research Group in Basic Oncology, Cancer Division of the Quebec University Hospital Research Centre and Department of Molecular Biology, Medical Biochemistry and Pathology, Faculty of Medicine, Laval University, Edifice St Patrick, 9 rue McMahon, Québec, QC, G1R 3S3, Canada.

122

4 The role of ribosomal DNA in maintaining or establishing pluripotency

4.1 Preface There is a growing body of evidence suggesting that the level of rRNA transcription is important to maintain pluripotency, which led to our interest in using our mouse models to study this phenomenon. First, we were curious about the chromatin state of the rDNA and its regulation in pluripotent cells. We decided to use our deep-sequencing chromatin immunoprecipitation (ChIP) protocol to study the histone landscape and the RPI machinery on the rDNA in pluripotent and differentiated cells. Secondly, we wanted to use our conditional mice for UBF and RRN3 to create models to study the effect of the loss of these factors on stem cells.

This chapter describes the first experiments performed in the study of rRNA transcription in pluripotent cells, which was undertaken in the last six months of my PhD. With advice and support from Michel Tremblay, I isolated embryonic stem cell lines from our mutant mouse lines. Psoralen photo-crosslinking experiments were performed on these cells by Victor Stefanovsky. Jean-Clément Mars is continuing the project and Marianne Sabourin-Félix performed the bioinformatics analysis of the ChIP-seq data.

123

4.2 Résumé La synthèse ribosomale est un processus cellulaire essentiel et étroitement lié à la croissance cellulaire. Il a été démontré que plusieurs voies de signalisation, d’oncogènes ou de gènes suppresseurs de tumeur convergent vers la transcription de l’ARN ribosomal. Les gènes codant pour l’ARNr forment les régions organisatrices nucléolaire (NOR). Ces régions sont d’importants foyers de recombinaison de l’ADN et l’inhibition de l’ARN ribosomal pourrait être impliquée dans la maintenance de la stabilité génomique. En analysant le niveau de méthylation de l’ADN ribosomal dans des cellules souches embryonnaires, nous avons observé une perte des marques de méthylation CpG dans les éléments de régulation et la séquence codante pour l’ADNr. De façon surprenante, le marqueur associé à l’inhibition des régions promotrices, CpG-133, est également non méthylé dans les cellules souches. Nos résultats démontrent que la chromatine adopte une conformation ouverte aux répétitions d’ADNr dans les cellules souches pluripotentes et certaines lignées cellulaires cancéreuses. Ces résultats suggèrent que l’ensemble des gènes codant pour l’ARNr, dans ces cellules, serait actif alors que dans les cellules somatiques seulement la moitié de ces gènes sont actifs. Nous avons établi un modèle permettant d’étudier l’inhibition de l’ADNr et son implication dans l’établissement de la pluripotence. Des résultats préliminaires démontrent une activité différente de la polymérase dans les régions promotrices.

124

4.3 Abstract Ribosome biogenesis is an essential process in the cell and tightly linked to cell growth. In fact, many signaling pathways, oncogenes and tumor suppressors converge on the transcription of the ribosomal DNA (rDNA). The ribosomal RNA (rRNA) genes (rDNA) form nucleolar organizer regions (NORs). Active NORs are hotspots for DNA recombination and rDNA silencing may be involved in maintaining genomic stability. Upon analysis of the methylation status of the rDNA in embryonic stem cells (ESCs), we have observed a complete lack of CpG methylation in the rDNA regulatory and coding regions. The classical CpG at -133 that has been associated with promoter silencing is fully unmethylated in ESCs. Further investigations have demonstrated an open chromatin conformation for all the rDNA repeats in ESCs and in at least one cancer cell line. This indicates that the rRNA genes in these cells are either all transcriptionally active, or have the potential to be active. This contrasts with somatic cells in which only a fraction of rDNA is active. We have established a model in which the establishment of rDNA silencing during differentiation and its loss during the establishment of induced pluripotency can be studied. Preliminary results show a difference of polymerase activity in the promoter region.

125

4.4 Introduction The ribosomal DNA (rDNA) encodes the catalytically active ribosomal RNAs (rRNAs) essential for ribosome structure and function. Evidently, the production of these rRNAs, which make up two thirds of the ribosome, is tightly regulated and linked to cell growth. Various signaling pathways, oncogenes and tumor suppressors regulate transcription of the rDNA. Importantly, in differentiated mammalian cell lines, around half of the ~200 rDNA gene repeats are transcriptionally silent (Moss et al. 2007, Stefanovsky & Moss 2006, Stefanovsky et al. 2006a). However, the proportion of active genes is able to respond to growth stimuli and adjust transcription levels accordingly so the number of active repeats does not seem to be related to cell growth. Rather, rDNA silencing seems to be required for resisting DNA damage and protecting the genome from instability (Gagnon- Kugler et al. 2009, Kobayashi 2011b). As genomic instability is a hallmark of cancer and some genetic syndromes, the deregulation of rDNA silencing could be associated with disease development.

The rDNA repeats, organized in a head-to-tail arrangement on the short arms of acrocentric chromosomes, form the nucleolar organizer regions (NORs). Actively transcribed NORs give rise to the nucleoli, the largest subnuclear bodies, where ribosome biogenesis occurs (McStay 2016). These regions are particularly susceptible to enhanced DNA recombination and may be associated with chromosomal breaks leading to, for example, Robertsonian-related chromosomal translocations (Jarmuz-Szymczak et al. 2014). These translocations are causes of genetic syndromes such as Down’s and Uniparental Disomy and are associated with multiple cancers (Li et al. 2014, Welborn 2004). An increase in these translocations occurs after knockdown of factors involved in rDNA structure regulation (Stimpson et al. 2014). Thus, rDNA silencing is important for chromosomal stability and could be implicated in the development of cancer.

Methylation-induced silencing of the rDNA involves the nucleolar remodeling complex (NoRC) and transcription of a long non-coding RNA (lncRNA) from the rDNA intergenic spacer (IGS) (Mayer et al. 2006, Santoro et al. 2009). The use of lncRNAs in DNA silencing is not specific to the rDNA and is particularly well documented in the case of X-inactivation during early development (Kornienko et al. 2013). The rDNA lncRNA is also responsible for silencing other regions of the genome in addition to the rDNA itself (Guetg & Santoro 2012).

Interestingly, we observed that the control and coding regions of the rDNA lack CpG methylation in pluripotent cells and this corresponds with the vast majority of these genes displaying an open, psoralen-accessible chromatin normally correlated with transcriptional activity. Thus, ESCs have most or all of the their rRNA genes in an active or potentially active state. In contrast to somatic cells in which only a fraction of the rDNA is active, ESCs therefore provide the opportunity to determine the chromatin status uniquely of active rDNA.

126

Further, since differentiation of ESCs to the various somatic cell types, must necessarilly be accompagnied by silencing of a fraction of the rDNA, these cells provide a unique opportunity to study the onset of silencing.

4.5 Results

4.5.1 A fraction of the rRNA genes are highly methylated in somatic cells but not in ES cells A recent study from the Stunnenberg lab investigated the methylomes of the E14 mouse embryonic stem cell line under different culture conditions and demonstrated the hypomethylated state of naïve ESCs grown in 2i+LIF medium (Habibi et al. 2013). We thought to compare the methylation status of the rDNA in ESCs and somatic cells using whole genome bisulfite sequencing (WGBS) as was performed in their study. We sent genomic DNA (gDNA) from two somatic cell lines (MEFs and NIH3T3) and one pluripotent cell line (E14 mouse ESCs) grown in serum+LIF or 2i+LIF medium for WGBS (a collaboration with Ivo Gut, CNAG, Barcelona). The data revealed a very high density of methylation within the ribosomal RNA gene body and over the promoter and enhancer elements in the somatic cells (Figure 4.1a). Interestingly, the same regions in the ESCs, both from our data and the publically available data, are devoid of CpG methylation. The surrounding IGS regions do however display CpG methylation, which is comparable to somatic cells. Strikingly, even the CpG at position -133, which is characteristic of rDNA promoter silencing is unmethylated in ESCs (Figure 4.1b). Therefore, a lack of DNA methylation in the rDNA gene and promoter sequences may be a characteristic of pluripotent cells. Interestingly, the rDNA region lacking CpG methylation in ESCs is in MEFs flanked 3’ by the TTF1 sites T1-T10 and 5’ again by TTF1 as well as the transcription independent chromatin boundary we observed previously (see Chapter 2). As will be seen below, we have found that this is also the case in ESCs.

127

Figure 4.1 rRNA genes in embryonic stem cells are fully unmethylated in the gene and control regions

(A) Whole genome bisulfite sequencing (WGBS) was performed on DNA from NIH3T3, Ubfwt/wtER-Cre+/+p53-/- MEFs and E14 ES cells grown in 2i+LIF or serum+LIF medium. The rRNA gene repeat is shown below the histograms. (B) Zoom of the 47S promoter region showing that the level of methylation on the CpG at position - 133, characteristic of promoter silencing, is extremely low or non-existant in ES cells.

4.5.2 The rRNA genes are all psoralen-accessible in ES cells Using the same cell lines as used for WGBS, we performed psoralen crosslinking experiments to study the activity status of the rDNA. This technique uses the enhanced accessibility of non-nucleosomal chromatin to psoralen to identify the actively transcribed or potentially active rRNA genes. After genomic DNA is crosslinked and then digested by restriction enzymes, the active “a” and inactive “i” fractions of the rDNA are separated by gel electrophoresis. Southern analysis using a probe for the coding region identifies the proportion of accessible to inaccessible rRNA genes. As shown in Chapter 2, we found that psoralen accessibility of both mouse fibroblast lines studied demonstrate two bands with the lower or inactive band more prominent suggesting a higher proportion of inactive than active rDNA repeats in these cells (Figure 4.2). However, in E14 ESCs the upper band is much stronger than the barely visible lower band. Similarly, L1210 mouse lymphocytic leukemia cells show the majority, if not all, of the rRNA genes are in an open conformation. Based on this analysis, the vast majority of the rDNA repeats are either actively transcribed or in a potentially active UBF-bound conformation in ESCs and at least one model cancer cell line.

128

Figure 4.2 Psoralen crosslinking analysis of the rRNA genes The probe position is indicated in the upper diagram and the lower panel shows a typical electrophoretic separation of actively transcribed “a” and inactive “i” genes from non-crosslinked NIH3T3 cells (Ctrl), from crosslinked L1210 mouse lymphocytic leukemia cells, E14 mouse ES cells grown in serum+LIF or 2i+LIF medium, UBFwt/wtER-Cre+/+ MEFs, and NIH3T3 fibroblasts.

4.5.3 Embryonic stem cells display an upstream boundary of CTCF, TTF1 and a poised RNA polymerase I The binding of the RPI-specific transcription factors and epigenetic factors to the rDNA regulatory elements in the intergenic spacer (IGS) modulates the transcriptional output of rRNA. In studying the hierarchy of preinitiation complex formation in MEFs, we found an upstream arrested polymerase at the spacer promoter (SpPr) (see Chapter 2). The SpPr lies around 2 kb upstream of the 47S pre-rRNA promoter and transcription

start site (TSS) (Kuhn & Grummt 1987) and is adjacent to a terminator Sal-box site (Tsp) found 63 bp downstream (Németh et al. 2008). This arrested polymerase was flanked upstream by a chromatin boundary element formed of H2A.Z, CTCF, and activating histone marks and just downstream by the termination factor TTF1. Interestingly, we showed that in MEFs the presence of this chromatin boundary was independent of RPI transcription or preinitiation complex formation as loss of UBF or RRN3 did not eliminate TTF1, H2A.Z or CTCF binding (Chapter 2)(Hamdane et al. 2014). Since ChIP-seq analysis in MEFs gives a snapshot of the whole population of the rDNA, both the active and inactive copies, it is difficult to say for certain that this chromatin boundary is associated with the activity status of the rDNA though this appears likely. However, our recent data suggests that all the rRNA gene copies are active or potentially active in ESCs. Therefore, we thought to investigate whether the chromatin boundary previously observed in MEFs is present on the active portion of the rDNA. We isolated embryonic stem cells from ERCre+/+ wild-type and Ubffl/flERCre+/+ or Rrn3fl/flERCre+/+ conditional mice (Supplementary Figure 4.7 and Material & Methods). As a starting point, we performed ChIP-seq analysis in the wild-type ERCre+/+ and E14 ESCs. The histograms are displayed in Figure

129

4.3a for the wild-type ESCs, and though not shown, the E14 cell line gave the same profiles. We observed nearly identical profiles for UBF, TAF68 (SL1) and TTF1 between ESCs and MEFs (compare Figure 4.3 with Figure 2.1, Chapter 2). Interestingly, the strong upstream peak of RPI at the SpPr is conserved in these pluripotent cells. However, the relative proportion of elongating RPI through the gene body is much less than in the MEFs. In addition, the peak of RRN3 at the 47S promoter is higher and does not demonstrate the same gradual release into the gene body. These results suggest differences in RNA polymerase I dynamics or elongation rate in pluripotent cells as compared to differentiated cells, and preliminary elongation rate measurements performed by Victor Stefanovsky suggest this may be the case. At the spacer promoter, the peaks of RPI and RRN3 are slightly downstream of UBF, SL1 and the probable initiation site, but upstream of TTF1. This suggests that these transcription complexes are arrested early during elongation. Indeed, as in MEFs, RPI at the spacer promoter is associated with much less RRN3 than at the 47S promoter, again suggesting it is no longer initiation competent.

Figure 4.3 ChIP-seq analysis of the RPI machinery in ESCs (A) ChIP-seq for RPI, RRN3, UBF, TAF68 (SL1), TTF1, and CTCF in ER-Cre+/+ ESCs was aligned to the rDNA repeat, which is displayed below the histograms. (B) Zoom of the spacer promoter region. Dotted lines indicate the center of each peak.

130

As observed in MEFs, CTCF is also found immediately upstream of the SpPr (Figure 4.3b and 2.5a). To confirm this result, we re-aligned the publically available ChIP-seq data from ESCs onto the rDNA. The re- aligned data for CTCF and also for cohesin (Smc1, Smc3, Rad21), involved in sister chromatid segregation and the formation of chromatin boundaries (Dowen et al. 2013, Kagey et al. 2010), revealed that both were present immediately upstream of the spacer promoter in ESCs (Figure 4.4).

Together, these data suggest the presence of a chromatin boundary formed by CTCF and cohesin upstream of the preinitiation complex and a poised RNA polymerase I at the spacer promoter of the rDNA in pluripotent ESCs. As our results suggest that all rRNA gene copies are active or potentially active in ESCs, this boundary

may function in establishing or maintaining an open chromatin state of the rDNA.

Fold enrichment

Figure 4.4 CTCF and cohesin are found upstream of the spacer promoter in ESCs

(A) Association of cohesin complex subunits (Smc1, Smc3 and Rad21 and boundary factor CTCF aligned using publically available data from the V6.5 ES cell line (Kagey et al. 2010, Dowen et al. 2013). The rRNA gene and control regions are represented below the ChIP-seq histograms. (B) Zoom of the promoter, regulatory spacer promoter and intervening enhancer regions.

4.5.4 Loss of UBF in ESCs leads to the loss of active rDNA repeats The mice conditional for UBF and RRN3 provide a valuable tool with which we could study the inhibition of rDNA transcription in pluripotent and differentiated cells. Preliminary testing of the CRE-mediated excision of Ubf and Rrn3 in the embryonic stem cell lines we have isolated is shown in Figure 4.5. Knockout after a 4h dose of 50nM tamoxifen (4-HT) is complete in the UBF conditional ESC line, as is the case in MEFs. However, also similar to the situation in MEFs, RRN3 protein levels are reduced but not lost with this low dose and short treatment time. Further testing is necessary in order to achieve a complete and rapid knockout for RRN3.

131

Figure 4.5 Preliminary 4-HT treatment of ES cells to induce CRE-mediated excision of Ubf and Rrn3 Western blot analysis of wild-type ER-Cre+/+, Ubffl/flER-Cre+/+, and Rrn3fl/flER-Cre+/+ ESCs after tamoxifen (4-HT) treatment. UBF, RRN3 and Tubulin protein levels were checked.

As shown previously, the vast majority of the rDNA repeats in ESCs are active according to psoralen crosslinking analysis. In fibroblast cell lines, the loss of UBF leads to a closed conformation, no longer accessible to psoralen. Similarly, even though almost all the rRNA gene copies are active in ESCs to start with, the loss of UBF leads to a rapid conformational change in the rDNA. Just 24 hours after tamoxifen treatment and the proportion of active to inactive genes ressembles that of MEFs where more than half the genes are inactive (Figure 4.6). Subsequent days show a complete loss of active repeats corresponding toUBF ablation. In contrast, the ER-Cre+/+ control line is unaffected by tamoxifen treatment and the majority of rDNA repeats remain active.

Figure 4.6 Knockout of UBF leads to the formation of a closed conformation of the rDNA A typical electrophoretic separation of actively transcribed “a” and inactive “i” genes from non-crosslinked ER- Cre+/+ ESCs (Ctrl), and from crosslinked ER-Cre+/+ ESCs grown in serum+LIF treated with tamoxifen (4-HT) to induce LoxP site recombination.

132

4.5.5 Using directed differentiation to study rDNA silencing throughout loss of pluripotency While more work is to be done to fully characterise the epigenetic landscape of the rDNA in ESCs, one of the main purposes of this study was to understand the role of rDNA silencing in pluripotency maintenance. Hence, the capacity to efficiently and rapidly differentiate the ESCs in order to perform ChIP and other analyses was essential. The first protocol tested in the laboratory involves differentiation from ESCs grown in 2i+LIF to neuronal precursors using retinoic acid (RA) (Kim et al. 2009, Xu et al. 2012). The colonies rapidly changed morphology, growing into a monolayer ressembling neural progenitors and some neuronal-like projections could even be seen at Day 3 (Figure 4.7)(Lundqvist et al. 2013). Differentiation will be monitored in the future by comparing mRNA levels by RT-qPCR and protein levels by IF for the pluripotency markers Oct-4, Sox2 and Nanog and the neural stem cell markers Sox1, Pax6 and Nestin. These differentiated cells will then be used to study the occupancy of RPI factors on the rDNA, specifically at the spacer promoter, as well as histone modifications and boundary factors. The level of CpG methylation on the rDNA will also be studied by WGBS and rDNA transcription will be analyzed by metabolic labeling. In addition, other simple differentiation protocols to mesodermal or germ cells could be used to confirm results obtained in neuronal precursors (Chen et al. 2012, Tada et al. 2005, Torres et al. 2012).

Figure 4.7 Retinoic acid induced differentiation of embryonic stem cells Cells were plated on Day 0 at low density in 2i+LIF media on gelatin coated plates. Media was changed on Day 1 to N2B27 media + 1µM RA and was replaced each subsequent day. By Day 3 all cells were morphologically different. Arrowheads depict potential neuronal projections.

4.6 Conclusion Mammalian somatic cells contain hundreds of copies of the rRNA genes, however a large proportion are epigenetically silenced. Though the purpose of the multitude of silent rRNA gene copies and the regulation of their silencing is relatively unknown, it has been proposed that the silent rDNA copies are required for maintaining genomic stability during the aging process (Kennedy et al. 1997, Kobayashi 2011b). Others have suggested that the heterochromatic rDNA copies may have a role in maintaining chromosome topology or

133

even nuclear structure (Savić et al. 2014). In contrast to mammalian somatic cells where approximately half of rRNA gene repeats are silenced, we observed that the vast majority of the repeats in embryonic stem cells and at least one cancer cell line are either active or potentially active. We found that the rDNA of ESCs display a total lack of CpG methylation within the promoter and the gene body and elevated psoralen accessibility. This very high proportion of active genes may answer the cancer or pluripotent cell’s increased translational demands. The mostly active state of the rDNA in ESCs must mean that the establishment of inactivating histone and DNA modifications, and transcriptional silencing must occur sometime during differentiation, leading to the high level of inactive and heterochromatic rRNA genes found in somatic cells. Inversely, if the majoritarily active rDNA status observed in the L1210 leukemia cell line is common to cancer cell lines, then perhaps the reactivation of silent rDNA copies may be a process of malignant transformation. Interestingly, having more open rDNA repeats has been associated with the potential for increased genomic instability (Kobayashi 2011b), which is a well known hallmark of cancer cells and a mechanism that drives cell transformation. ESCs have also been shown to display elevated chromosomal instability, as for example indicated by the higher levels of Robertsonian translocations in these cells in comparison to somatic cells (Gaztelumendi & Nogués 2014, Greaves 2015, Hanahan & Weinberg 2011, Li et al. 2014).

We also have observed an upstream chromatin boundary formed in part by CTCF and a poised RPI complex at the spacer promoter in ESCs. This is similar to what we had previously shown in MEFs, except that the RPI and RRN3 peak of binding are stronger at the spacer promoter as compared to MEFs. This could suggest that more of this upstream spacer promoter RPI is initiation competent in ESCs as compared to MEFs (Chapter 2). Since transcription from the IGS starting from the spacer promoter has been associated with rDNA silencing and heterochromatin formation (Mayer et al. 2006, Savić et al. 2014), perhaps RPI is paused at this site in order to quickly and efficiently restart transcription of the IGS-RNA upon induction of differentiation signals.

Due to their lack of methylated and silent rDNA, ESCs provide the opportunity to determine the chromatin status uniquely of the active rDNA. We will use these cells to study the onsest of silencing through differentiation to the various somatic cell types, and through inhibition of RPI synthesis by using our knockout models. We will also study the reverse process of activation of silent respeats by inducing pluripotency in MEFs and potentially induced oncogenic transformation of MEFs. Using these tools to investigate the regulation of silencing will hopefully reveal information vital to the understanding of the role of the silent copies of the rDNA.

134

4.7 Materials and Methods Derivation of mouse embryonic stem cells. Derivation was performed as described elsewhere (Bryja et al. 2006). Embryos were flushed from uterine horns dissected at E3.5 using a 1 ml syringe with a 27.5 gauge needle. 2 ml of pre-warmed ES-DMEM was flushed through each horn under a dissection microscope. Embryos were pipetted into individual wells of a 12- well plate containing LIF-producing STO feeders using as small of a volume as possible with a 20 µl pipette and low-binding pipette tips. Once blastocysts attach to the feeder layer and hatch from the zona pellucida, the media is changed (usually day 2-3). The medium was changed every second day, unless the colour of the media started to become orange or yellow, at which point it was changed sooner. Around day 5-7 the inner cell mass has expanded enough to pick the colony. The expanded blastocysts were picked under the inverted microscope in a minimal amount of sterile PBS (<5 µl). The cells were put in round-bottom 96-well plates containing 25 µl of 0.05% Trypsin-EDTA (Gibco) and pipetted up and down to mechanically separate cells after 5 mins incubation at 37°C. The cell suspension was then put into a new 12-well plate well with feeders. ESCs then started forming small colonies and were subsequently passaged as normal with 0.05% Trypsin- EDTA. By passage 4-5 or around 2 weeks after dissection, the ESCs were robustly growing and a stock was frozen in ES-DMEM with 30% FBS.

Culture of mouse ESCs and adaptation to 2i+LIF media. All petris were treated with 0.1% gelatin for 30 mins prior to cell seeding. STO feeder cells that produce LIF, previously inactivated with mitomycin C (Sigma, 10 µg/ml final, 2 h treatment), were thawed and 0.2x106 cells were plated per 35 mm petri in DMEM (Gibco) with 2.2 g/L NaHCO3 (Sigma), 10% fetal bovine serum (FBS) (Wisent), 1% L-Glutamine (Gibco), 1% Pen/Strep (Gibco). Embryonic stem cells (ESCs) were thawed and 6 0.5x10 cells were plated per 35 mm petri in ES-DMEM (2.2 g/L NaHCO3, 10% ES-certified FBS (Hyclone), 1% L-Glutamine, 1% Pen/Strep, and 3.5 µl beta mercaptoethanol (β-MeOH) per 500 ml media). Cells were passaged every 2-3 days at a ratio of no more than 1:10. A new batch of feeders was thawed every 5-7 days. To culture the ESCs without feeders, they were split 1:3 into gelatin-coated petris without feeders and passaged 3-4 times to deplete the feeder cells in ES-DMEM + 10 ng/ml LIF (Sigma). Following this, the media was changed to Serum-Free ES (SFES) media with two smalle molecule kinase (Mek and GSK3) inhibitors (2i) and LIF and cells were passaged 3-4 times in the new conditions prior to experimentation. SFES is made up of a 50:50 ratio of neurobasal media with B27 supplement (Gibco) and DMEM/F12 media with N2 supplement (Gibco), 1% Pen/Strep, and 1% L-Glutamine. Prior to use, inhibitors PD03259010 (1 µM, Stemgent) and CHIR99021 (3 µM, Stemgent), and 10 ng/ml LIF was added to a SFES aliquot.

135

lysate, 1mg/ml Proteinase K (from 10mg/ml stock, Sigma) was added and the combined lysate incubated overnight at 50°C, adding fresh Proteinase K after the first 2 hrs. The genomic DNA was then deproteinized twice with Phenol/Chloroform, ethanol-precipitated and resuspended in 250µl TE, pH7.5, 0.1% SDS. 20µg of RNAse A (Sigma) was added and after 30 min at 37°C, 0.5mg/ml Proteinase K was also added and incubation continued for another 1h at 50°C. After twice extracting with phenol/chloroform treatmets the genomic DNA was ethanol precipitated, resuspended in 300 µl TE and quantified by EtBr staining on an agarose gel. About 10µg of genomic DNA was digested overnight either with BamHI or EcoRI, deproteinized, ethanol precipitated and finally resolved on a 1% TEA agarose gel at 2V.cm-1 for 18 hr in the absence of EtBr. The gel was subsequently EtBr-stained, photographed, the crosslinks reversed by UV irradiation of 4000J.cm-2 at 254nm in a UVC 500 crosslinker (Hoefer), transferred on a Biodyne B membrane (Pall) and the membrane crosslinked (70J.cm-2). Hybridization was performed overnight at 65°C with random-primed labeled DNA in 5xSSC, 0.5%SDS, 5xDenhardt’s, 100µg/ml salmon sperm DNA. Subsequently membranes were washed (15 min each time) with 6xSSC, 2xSSC and 0.1xSSC, 0.1% SDS. Data were analyzed by STORM 860 PhosphorImager. (GE Healthcare) and with ImageJ software (NIH).

Whole-genome bisulfite sequencing. Genomic DNA was prepared and sent to the lab of Ivo Gut at The Institute in Barcelona, Spain. For analysis of datasets the quality check was performed with FastQC version 0.11.4 (Andrew 2010). Datasets were trimmed using Trimmomatic version 0.33 (Bolger et al. 2014). Unpaired reads after trimming were removed from further analysis. The trimmed data were then aligned to the reference bisulfite genomes using Bismark (Krueger & Andrews 2011). Bismark_methylation_extractor script was used to extract the CpG methylation profiles.

Inactivation of Ubf or Rrn3 in cell culture, and analysis of protein. Cells were initially plated in 35 mm petri dishes (0.5x106 cells each) and cultured for 18 hours in DMEM-ES. To activate ER-Cre, 4-hydroxytamoxifen (4-HT) was added to a final concentration of 50nM, and after 4 hr incubation the medium replaced with fresh medium without 4-HT and cells harvested for analysis at various time points. Analyses of protein and genotype were systematically carried out on parallel cell cultures. Cells were genotyped by PCR before and after 4-HT treatment using the primers for Ubf: A; 5’TGATCCCTCCCTTTCTGATG, B; 5’TGGGGATAGGCCTTAGAGAGA, C; 5’CACGGGAAAACAAGGTCACT and for Rrn3: A; 5'-GATCTTAATGGAGGGCAGCA , B; 5’-TGGATCCTGCAACTTTTTCC , C; 5’ TCCCAACCCTGACCTATCAC (data not shown). For total protein, cells were washed with cold PBS, scraped into PBS, centrifuged 2 mins at 2000 r.p.m., then resuspended in sodium dodecyl sulphate (SDS) loading buffer. After fractionation on an 8% SDS–polyacrylamide gel electrophoresis (SDS-PAGE), cell extracts were analysed by standard Western blotting procedures.

Antibodies for Western Blot and Chromatin Immunoprecipitation. Rabbit antibodies against UBF, RPI large subunit (A194), SL1 subunit TAF68, RRN3, and TTF1 were generated in the laboratory. Other antibodies were obtained commercially; anti-Tubulin (Sigma) and anti- H2A.Z (Abcam).

136

HCl pH 7.5, 150 mM NaCl, 2 mM EGTA, 4 mM EDTA, 0.1% SDS, 1% Triton X-100, 1% NP-40). Each immunoprecipitation was carried out on the equivalent of 50 x 106 cells in IP Buffer (150 mM NaCl, 50 mM Tris-HCl pH 7.5, 5 mM EDTA, 0.5% NP-40, 1% Triton X-100) overnight at 4°C. Antibody slurry was prepared the night before with 50 µl A/50 µl G Dynabeads and 60 µg/ml antibody per IP. For the moment, only one biological replicate of E14 and ER-Cre+/+ ESCs was analyzed for each antibody; UBF, TTF, RPI, RRN3, TAF68. Samples were sent for single-end sequencing on an Illumina Hi-seq 2000 (McGill University and Génome Québec Innovation Centre) and qPCR was used to confirm the results. qPCR analysis. Immunoprecipitated DNA was analyzed by qPCR/SYBR Green. Reactions (20 µl) were performed in triplicate with 2.5 µl of sample DNA, 20 pmol of each primer, and 10 µl of Quantitect SYBR Green PCR Master Mix (QIAGEN). Forty reaction cycles of 10 s at 95°C and 30 s at 58°C were carried out on a Multiplex 3005 Plus (Stratagene/Agilent). The amplicon coordinates relative to the 47S rRNA initiation site (BK000964) were as follows: IGS3, 42646– 42903; SpPr, 43089–43253; Tsp, 43267–43421; 47SPr, 45133–40; 47S, 159-320; ETS, 3078–3221; 28S, 10215–10411; T1– 3, 13412–13607. Data was analysed using the MxPro software (Agilent). The relative occupancy of each factor at each amplicon is given as % immunoprecipitation of the DNA input prior to ChIP. It was determined by comparison with a standard curve of amplification efficiency for each amplicon using a range of input DNA amounts and generated in parallel with each qPCR run. All primer pairs gave the similar amplification efficiencies (90–105%) as determined from the gradient of the curve fit. The curve fit correlation coefficient R2 was systematically between 0.99 and 1.0, demonstrating a near perfect fit.

ChIP-seq analysis. Raw fastq.gz data was checked for quality using FastQC version 0.11.4 (Andrew 2010). The data was trimmed using Trimmomatic version 0.33 (Bolger et al. 2014) and the resulting quality filtered files were aligned using Bowtie2 (Langmead & Salzberg 2012) with -k 3 parameter, allowing sequences to map three times to the reference genome (MmGRCm38) which had been modified to include the rRNA gene consensus as an extra chromosome {vandeNobelen:2010es, Grozdanov:2003bh}. The rDNA mouse gene (GenBank BK000964 v3) was rearranged to place the regulatory regions (30,528 to 45,306) upstream the coding region (1 to 30,527). Aligned data was converted to BAM format, sorted and indexed using SAMtools (Li et al. 2009). The rDNA chromosome was extracted from the data set and converted to BED format. Reads were extended to 100bp each and the coverage was calculated to give a histogram using BEDtools (Quinlan Lab, University of Utah). The histogram was smoothed using a window of 25 bp averaging. The coverage was normalized to reads per million (RPM) and the sample (IP) was divided by the input. Finally, the BED files were converted to BEDgraph files and visualized in IGV (Integrative Genomics Viewer 2.3, Broad Institute).

Alkaline phosphatase staining. Alkaline phosphatase staining was performed as described in the protocol of the Millipore Alkaline Phosphatase Detection Kit. Briefly, cells were fixed in 4% paraformaldehyde in PBS for 90 secs and then washed with TBS-T (20mM Tris-HCl ph 7.5, 150mM NaCl, 0.05% Tween-20). The cells were then covered with staining solution (Fast Red Violet, Naphthol AS-BI phosphate solution, water in a 2:1:1 ratio) and incubated in the dark for 15 mins. Cells were once again washed with TBS-T and then covered in PBS while imaging. Images were taken on a Leica inverted contrasting microscope (Leica DM IL LED) using the 10X objective.

137

Retinoic acid differentiation. 1 x 105 ESCs were plated in SFES 2i+LIF medium in 35mm petris on Day 0. 24 hours later (Day1), the medium was changed to SFES + 1 µM retinoic acid (RA) for the experiment petris or replaced with fresh SFES 2i+LIF for the control cells. The media was changed each day in both cases and brightfield photos were taken. The petris were confluent on Day 5 and cells were then taken for primary analysis (data not shown).

4.8 Acknowledgements We wish to thank Dr Lucie Jeannotte, Dr Jean Charron and Dr Steve Bilodeau for advice on the generation and maintenance of ES cell lines, Dr Ivo Gut for the whole-genome bisulfite sequencing analysis and Dr Raffaella Santoro for helpful discussion. This work was funded by an operating grant from the Canadian Institutes of Health Research (CIHR, MOP12205) and a CIHR Frederick Banting and Charles Best Canada Graduate Scholarship Doctoral Award to Chelsea Herdman (CIHR CGS-D). The Research Centre of the Québec University Hospital Centre (CHU de Québec) is supported by the Fonds de recherche du Québec - Santé (FRQS).

138

4.9 Supplemental Data

Figure 4.8 Generating embryonic stem cell lines

(A) Blastocyst embryo isolated at E3.5 from Ubfflox/floxER-Cre+/+, Rrn3flox/floxER-Cre+/+ and the ER-Cre+/+ wild-type control mice lines and placed these in culture on mitotically-inactivated feeder fibroblasts in ESC medium. Arrowhead indicates inner cell mass (ICM). (B) Two days after the embryo is placed in culture, the ICM and the trophoblasts divide and form a colony. After the colony is picked and passaged at around day 5-6, the trophoblasts will die out as their growth is not supported in the ESC medium but the ESC will continue to grow and divide, forming multiple colonies. Colonies of ESCs were growing well in culture after 3-6 passages depending on the cell line, as shown in (C). (D) ESC colonies on feeder layer stained with the alkaline phosphatase (AP) staining kit showing the pluripotent ESC colonies staining red. (E) ESC colonies transitioned to serum+LIF ESC medium grown on gelatin-coated plates and stained using the AP staining kit. (F) ESC colonies transitioned to 2i+LIF ESC medium grown on gelatin-coated plates.

139

Chapter 5 Discussion

140

5 Discussion

This doctoral thesis presents the analysis of the in vivo requirements for UBF and RRN3 in rRNA transcription and regulation of the rDNA chromatin structure. UBF and RRN3 were known to both be essential factors for mouse development and cell growth, most probably due to their role in ribosome biogenesis and in maintaining rDNA integrity (Hamdane et al. 2014)(Chapters 2 & 3). Importantly, while reinvestigating the knockout of RRN3, I found that deletion of Rrn3 causes a very early embryonic lethality. This finding is contrary to the previous data on this subject (Yuan et al. 2005), but is consistent with homozygous deletion of the Ubf, Fbl, or Polr1b genes (Chen et al. 2008, Hamdane et al. 2014, 2016; Newton et al. 2003). This is discussed in detail in section 5.1.

Using our cell lines conditional for UBF and RRN3, we were able to investigate the hierarchy of preinitiation complex formation and the requirements for each factor. Our normalization protocol for high-resolution ChIP- sequencing data (Sabourin-Félix M et al., in preparation) has allowed an in-depth look at the occupancy of RPI transcription factors and chromatin elements on the rDNA. Though RRN3 is required for transcription of the rRNA precursor, its presence is not required to maintain the potentially active gene structure since UBF and SL1 binding is unaffected. In contrast, not only does UBF act as an initiation factor essential for formation of the pre-initiation complex on the spacer and 47S promoters but it also establishes a specialized chromatin throughout the rRNA gene body. It is required for binding of all other RPI factors except TTF1. We propose that TTF1 is likely important for delimiting UBF binding and therefore a functional rRNA gene unit. Upstream of the unit, there exists a chromatin boundary formed of activating histone variants/marks H2A.Z, H2A.Zac, H3K4me2 and -me3, which is likely involved in UBF loading onto the rDNA and maintenance of the open chromatin of active gene repeats. These results are also discussed in section 5.2.

In studying the regulation of rRNA transcription in various cell types, we have observed differences in rDNA accessibility and methylation between embryonic stem cells and differentiated cells. In MEFs, a large portion of the rRNA genes are silenced and heterochromatic however, we have found that the rDNA repeats are all potentially active in pluripotent embryonic stem cells. Where analyses in somatic cells give an overall view of active and inactive rDNA copies concurrently, ESCs provide an ideal model in which to study the status and chromatin regulation uniquely of the active rDNA. Inducing differentiation in these cells will allow the study of the establishment of silencing. We have also generated ES cell lines conditional for UBF and RRN3 in order to study the effect of the inhibition of pre-rRNA synthesis on rDNA silencing or differentiation. This is discussed in section 5.2.

141

The presence of UBF is essential for cell viability, probably due to its role in maintaining rDNA chromatin. We investigated the effect of UBF loss in cell culture and determined that oncogenically transformed MEFs (iMEFs) undergo a synchronous apoptosis after knockout, while primary MEFs arrest cell growth but do not undergo apoptosis. Further, this cell death is p53-independent and not specific to a given oncogenic transformation. RRN3 loss did not produce the same effect in cell culture so this may be UBF-specific, which leads to the possibility of targeting UBF as a cancer therapy. Targeting ribosome biogenesis in cancer treatment and the next steps of this project are discussed in section 5.3 and 5.4 respectively.

5.1 Development of the RRN3 conditional knockout mice and cell lines A major aspect of my doctoral work was developing a conditional knockout model for RRN3 that was directly comparable to the existing model for UBF. In order to accomplish this, I reanalyzed the RRN3 knockout mice that had previously been published (Yuan et al. 2005) and isolated cell lines conditional for RRN3. In this section I will discuss the analysis of the mice and the difficulties I encountered with the in-culture excision of Rrn3.

5.1.1 RRN3 is required for early embryonic development The Grummt lab previously published the knockout of RRN3/TIF-IA as leading to an arrest at E9.5, well after the start of gastrulation and significant tissue differentiation (Yuan et al. 2005). This was surprising as RRN3 was shown to be required for the recruitment of the initiation competent form of RPI in mouse cell lines and in yeast (Milkereit & Tschochner 1998, Schnapp et al. 1993), and so could only be explained by a store of maternal RRN3 protein or mRNA. Yeast Rrn3 is an essential gene (Yamamoto et al. 1996) and it was presumed that murine Rrn3 would be an early essential gene since it is necessary for rRNA transcription. As the late developmental arrest was surprising, we chose to reanalyze the loss of RRN3 in mouse. We received the same mice conditional for RRN3 produced by the Grummt lab and crossed these with a Sox2Cre deleter strain to generate mice heterozygous for Rrn3. We found that the Rrn3-/- mice arrest preimplantation at E3.5 as an uncompacted morula (Figure 2.10), and though this contradicts the previous work, it was to be expected as will be discussed below. In accordance with this, the knockouts of Ubf, Fbl and Polr1b arrested development at morula or early blastocyst (Chen et al. 2008, Hamdane et al. 2014, 2016; Newton et al. 2003).

The start of embryonic pre-rRNA synthesis occurs either at the one- or two-cell stage (Golbus et al. 1973, Lin et al. 2014, Nothias et al. 1996, Zatsepina et al. 2003), therefore RPI transcription factors must be available in sufficient quantities in order to initiate transcription. Yuan et al. suggested that the maternal contribution of rRNA and of RRN3 mRNA could explain the late arrest of the RRN3 mutants (Yuan et al. 2005) however studies have shown that the majority of the maternal mRNA stored in the oocyte is degraded in the zygote

142

(Farley & Ryder 2008). For example, UBF, nucleophosmin 1, nucleolin and fibrillarin maternal mRNAs all were significantly degraded after fertilization (Fulka & Langerova 2014). In addition, it has been shown that fully- grown oocytes do contain a store of nucleolar factors but actually have low levels of rRNA (Shishova et al. 2015b). The level of RRN3 mRNA in embryos has not yet been experimentally tested but Yuan et al. compared the level of RRN3 mRNA in oocytes to that found in fibroblasts (Yuan et al. 2005). There was a significantly higher amount of RRN3 mRNA in oocytes however this does not give any indication of whether that surplus will be present in the embryo or that it will be sufficient for 9 days of development. Another possibility suggested by these authors was that the mRNA of RRN3 was relatively stable in oocytes, having a half-life of 13-15 hours. Again, it is difficult to compare levels of mRNA or protein from oocytes with that of embryos due to the massive degradation that occurs prior to zygotic genome activation. It is unlikely that the maternal contribution of RRN3 mRNA or of rRNA transcripts is enough to permit embryonic development to E9.5.

One could imagine that in the embryos arresting at E9.5, an unknown factor was potentially able to compensate somewhat for the lack of RRN3, leading to the late arrest observed by (Yuan et al. 2005). For instance, a factor similar to TBP, TBP-like factor TBP2, was identified as having a specific role during development (Bártfai et al. 2004, Gazdag et al. 2007). Four RRN3 pseudogenes were identified in human but no functional data has been attributed to these (Martin et al. 2004, Strausberg et al. 2002).

Yuan et al. describe exons 12, 13 and 14 as being flanked by LoxP sites, however when we sequenced the RRN3 conditional mice (Chapter 2, Figure 2.10), we found that the LoxP sites were intact but actually flanking exons 13, 14, and 15. This is most probably a simple error in the RRN3 map in the original publication (Yuan et al. 2005).

One difference between our studies was that the Grummt lab used a CMV-Cre deleter mouse strain and we used a Sox2-Cre deleter mouse strain (Hayashi et al. 2003, Schwenk et al. 1995, Yuan et al. 2005). Both strains should lead to ubiquitous Cre recombination however mosaicism has been shown to occur even from ubiquitous lines (Nagy 2000). It is important to inbreed to remove the Cre transgene from the heterozygous line, as Cre can catalyze recombination between endogenous cryptic pseudo-LoxP sites, leading to unwanted genomic alterations (Schmidt et al. 2000). Not only did we cross to remove the Cre transgene in our heterozygous mice, we also redid the original excision by crossing the Rrn3fl/fl mice with a second Sox2-Cre mouse from a different genetic background. Our first embryonic lethality analysis was performed with Rrn3+/- mice derived from the original BL/6 Rrn3fl/fl mice crossed to a 129 Sox2-Cre deleter strain. In order to confirm these results in a pure genetic background, the conditional mice were crossed for a second time but with a BL/6 Sox2-Cre deleter strain. In both cases crosses were made to remove the Cre transgene and we

143

backcrossed these mice further with BL/6 wildtypes six or more times to generate a population that was >98% BL/6. We also have confirmed that the LoxP sites were intact by sequencing in the Rrn3fl/fl mice.

It is possible that the Cre transgene was not crossed out in the original study. If the RRN3 heterozygote colony used for the lethality analysis retained a floxed allele (Rrn3+/fl) instead of being true heterozygotes (Rrn3+/-), and these were crossed with mice that still carried the Cre transgene (Rrn3+/-Cre+), this could lead to some Rrn3-/flCre+ embryos. These embryos could undergo excision during development leading to their eventual arrest, but not before some zygotic RRN3 was made, potentially allowing development to a later stage prior to arrest.

In summary, contrary to the previously published data (Yuan et al. 2005 & Chapter 2), starting from the same conditional gene allele, we show that Rrn3-/- mice arrest development before E3.5. We have discussed the possible reasons for the previous observations of a much delayed developmental arrest, but as we did not participate in the original analysis, we cannot be sure of the exact reasons that led to the original erroneous finding.

5.1.2 Tamoxifen-induced knockout of RRN3 in cell culture As described in Chapter 2, the knockout of RRN3 in cell culture using the tamoxifen-inducible ER-CreT or ER- CreT2 was not a simple task. Contrary to our UBFfl/flER-Cre+/+ MEFs which require just one 4 hour treatment of low dose tamoxifen, the RRN3 conditional MEFs require multiple treatments and cell passaging between doses.

Cre recombinase from the bacteriophage P1 recognizes LoxP sites, which are 34 bp sequences formed of two palindromic 13 bp repeats surrounding an 8 bp spacer region. The 8 bp spacer region can be variable in sequence with certain mutations in this region leading to more efficient recombination (Missirlis et al. 2006). To create a conditional knockout, the Cre recombinase is fused to the ligand binding domain (LBD) of the estrogen receptor (ER) that has been mutated to be more selective for 4-hydroxytamoxifen (4-HT/4-OHT) and insensitive to endogenous estradiol (Feil et al. 1996, 1997; Metzger et al. 1995).

It is inadvisable to use high doses of 4-HT for treatments as activation of Cre recombinase can lead to random DNA damage in mammalian cells due to a degree of non-specific endonuclease activity. However, prolonged treatments at a lower dose can lead to recombination of the LoxP sites without toxicity (Loonstra et al. 2001). Even a very low dose of 50nM 4-HT, which is ten to twenty times lower than the typical dose used in the literature (e.g. Bhaskara et al. 2008), does show some effect on proliferation (Figure 3.5B). Knowing this, we were careful to develop a protocol for the knockout of Rrn3 that would maximize excision with the minimal treatment possible. The protocol that allowed for the maximum excision was the combination of multiple, short

144

treatments of 50nM 4-HT and passaging of the cells. This being said, we were still never able to obtain excision of the Rrn3 gene leading to a protein loss of more than 90%.

There are many advantages to Cre-induced recombination in mouse and in cells, but there are also limitations. Cre-mediated recombination can be efficient even over long genetic distances (Zheng et al. 2000) or in trans between different chromosomes (Smith et al. 1995). However in some loci, even small changes of the distance between LoxP sites can affect efficiency (Coppoolse et al. 2005, Wang et al. 2009). Tamoxifen-inducible ER- CreT has been previously shown to result in only partial excision in some cases (Danielian et al. 1998) and even when excision is complete, there are examples where this does not lead to protein loss (Turlo et al. 2010). Using ER-CreT conditional mutants for various genes, Vooijs et al. have shown that LoxP recombination frequency varies depending on the position of the LoxP sites in the genome (Vooijs et al. 2001). In addition, silenced genes with methylated promoters have shown reduced recombination upon Cre activation (Long & Rossi 2009).

The difficulty we encountered in order to achieve knockout in cells is likely specific to the floxed Rrn3 allele, as the ER-Cre+/+ mouse line used in crosses with the Ubffl/fl mice was the same that was used for the Rrn3 conditional mice. In that case, in-cell gene excision was certainly higher than 99%. Also, Rrn3 conditional MEF lines were made from multiple different litters and at different times, therefore it is not a problem specific to one embryo.

During troubleshooting experiments, treatments with tamoxifen were increased to 500nM 4-HT for 48h but PCR and Southern analysis showed a 50:50 ratio of floxed to ∆ alleles, which was what we had already observed with a 50nM, 4h treatment (Figure 2.10). It was possible that we were observing a mixed population of cells, with some percentage being fully knocked out and others at various stages along the process of excision. Another possibility was that one of the Rrn3 floxed alleles was imprinted and inaccessible to the Cre, resulting in only one allele being excised and a 50:50 ratio of flox to ∆. Interestingly, even when I crossed Rrn3+/∆ mice with Rrn3fl/flER-Cre+/+ mice to generate a heterozygous Rrn3fl/∆ER-Cre+/+ line and the Rrn3+/∆ER- Cre+/+ control line in parallel, the excision was still not complete. PCR analysis showed that around half of the floxed alleles became ∆, as was observed for the homozygous floxed alleles. This suggested that the problem was mosaisicism in the population of cells and not imprinting of one of the alleles. The crosses were redone using the original conditional Rrn3fl/fl mice and a ER-CreT2+/+ line, which has a 3 to 4-fold higher affinity for 4-HT (Feil et al. 1997, Indra et al. 1999), however this did not improve Rrn3 excision. I also created clonal cell lines by dilution plating in order to see if the mixed population of cells was affecting knockout but I did not detect clonal differences in recombination (Figure 2.10). The only protocol that led to a higher level of excision in the

145

population was to treat 50nM for 4h, passage the cells, treat again and repeat the same process the following day.

As we have determined that the Cre is not the problem and the sequences of the LoxP sites are intact, we propose that the chromatin environment specific to the floxed region of the Rrn3 allele must be less accessible. The efficiency of LoxP recombination can vary based on the chromatin environment (Vooijs et al. 2001). The chromatin surrounding the LoxP sites in the RRN3 gene could impede Cre recombination in some cells, potentially leading to an incomplete excision. In addition, RRN3 is not an abundant protein and therefore may be transcribed much less than UBF for example, which is a very abundant protein. As previously mentioned, the knockout of UBF in MEFs was easily obtained therefore the chromatin environment and accessibility of the floxed gene may be a factor in knockout efficiency.

Clearly, the ER-Cre-induced knockout of Rrn3 is not 100%, however I was able to achieve >90% gene excision, allowing us to compare a very significant level of RRN3 loss with UBF loss in a closely similar model situation. We were still able to use these MEFs to study the requirements for UBF and RRN3 in preinitiation complex formation.

For future studies, it would be interesting to create CRISPR-Cas9 cell lines that could be conditionally and reversibly mutated for RRN3, UBF and other RPI factors, however this system also has its limitations (Sander & Joung 2014). The possibility of rescuing the phenotype with an inducible system (e.g. doxycycline-regulated Cas9) would allow highly-regulated temporal analysis of RPI inhibition (Cao et al. 2016, Dow et al. 2015). Rescue experiments could also distinguish the in vivo roles for the two splice variants of UBF. UBF2 is missing 37 a.a. from HMG-box 2 whereas UBF1 is the full length protein (Kuhn et al. 1994). Both factors are ubiquitously expressed however, few studies have shown specific roles for the two isoforms (Sanij et al. 2015, Stefanovsky & Moss 2008).

5.2 The role of UBF and RRN3 in rRNA transcription and rDNA chromatin structure In this study, we have shown that UBF and RRN3 are both essential in early development and for rRNA transcription. Using our cell lines conditional for these factors, we have investigated the hierarchy of preinitiation complex formation as well as proposed limits of the potentially active rDNA unit. These results are important for understanding the in vivo regulation of rRNA synthesis.

The study of the chromatin landscape and transcription regulation of the rDNA is challenging as mammalian cells have hundreds of rDNA genes existing in three states found in different proportions depending on the cell type. These are meCpG silenced and heterochromatic, conditionally silenced or potentially active, and actively

146

transcribed. The chromatin structure of the rDNA is not fully understood but is likely somewhat unconventional when compared to RNA polymerase II transcribed genes. Histones are present on the rDNA, but are not necessarily nucleosomal and are potentially present in an unfolded state (Prior et al. 1983, Sanij et al. 2008). We have previously shown that the nucleoprotein structure formed by UBF binding to DNA is incompatible with nucleosome formation (Stefanovsky et al. 2001a). In addition, an increase of H1 linker occupancy on the rDNA after loss of UBF has been observed, indicating that UBF likely inhibits assembly of nucleosome formation on active genes, thereby establishing an open chromatin structure (Hamdane et al. 2014, Sanij et al. 2008).

In order to investigate the activity signature of the rDNA, we analyzed RPI transcription factors and chromatin factors’ occupancy on the rDNA at a resolution not yet seen before for the rDNA (Chapter 2). Using ChIP sequencing, we observed a very high density of RPI throughout the entire gene body attesting to the high level of transcription (Chapter 2, Figure 2.1). We have confirmed that RRN3 is essential for RPI recruitment to the promoter and that it is indeed released as RPI elongates, as previous studies had proposed (Chapter 2, Figure 2.2) (Beckouet et al. 2008). Confirming what we and others had observed in ChIP-qPCR (Hamdane et al. 2014, Sanij et al. 2008), our ChIP-seq results showed UBF binding throughout the regulatory elements and the coding region. UBF coverage throughout the whole rDNA unit was very consistent. Interestingly, we observed differences between the occupancy pattern through the gene body (GC-rich) versus in the enhancers (Chapter 2, Figure 2.1). This could reflect two different roles for UBF, these being in elongation and structural maintenance of the gene or in initiation and promoter regulation. If UBF is responsible for keeping the rDNA in an open chromatin state, its binding to the rDNA would need to be quite stable. However, UBF has a high off- rate (t1/2 9 to 25 s) (Chen & Huang 2001), therefore it is very probably in constant flux, coming on and off the rDNA. It is likely that another mechanism establishes or maintains the positioning of UBF on the rDNA. UBF is however clearly essential for SL1, RRN3 and RPI recruitment to the rDNA, but not for TTF1, which may be involved in regulating UBF binding.

We propose that TTF1 delimits the functional rDNA unit at the TSP upstream and T1-T10 downstream (Figure 2.1). Upstream, TTF1 appears to be associated with a boundary factor formed of H2A.Z, CTCF and Cohesin (Chapters 2 and 4). As this boundary does not depend on UBF, it could be the factor that defines the active status of the rDNA. Active chromatin marks (H3K4me2/3) but not heterochromatin marks (H3K9me3) were also found at this boundary element. As H2A.Z is known to separate functional chromatin domains (Meneghini et al. 2003), it may act as a placeholder on potentially active genes. Nemeth et al. and Hamdane et al. saw high levels of RPI, TBP, UBF and H2A.Z occupancy somewhere near the SpPr in their low resolution ChIP analysis of the rDNA, so they also conclude that it is an entry point or assembly site for the RPI machinery (Németh et al. 2008, Hamdane et al. 2014). Using DNA methylation-sensitive ChIP, Nemeth et al. showed that factor binding at the SpPr did not correlate with methylation and RPI transcription factors were associated with

147

non-methylated, active genes whereas H3 core histone and H3K9me2 repressive marks were associated with methylated genes (Németh et al. 2008). The factors present at the SpPr are likely a signature for whether the gene is active or not. The knockout of TTF1, which is currently being generated in the lab, will clarify the role of this factor in regulating gene activity.

One difficulty that we have encountered in this study is that publically available aligned ChIP-seq data sets do not include sequencing data on the rDNA as the repetitive sequences are not represented in the silicon genome. One of the few ChIP-seq studies of the rDNA investigated the chromatin regulation on the rDNA and found that H3K36me3, which is usually present within the bodies of active, elongating genes transcribed by RPII, is absent from the rDNA gene body (Zentner et al. 2011). Our realignment of available data for H3K36me3 confirms this, but also shows that it concentrates within the upstream chromatin boundary. This is likely due to the unique chromatin structure that UBF maintains. We have also not found typical active chromatin marks within the gene body, at the 47S promoter or at the enhancers in any of our high resolution studies (Figure 2.5). Rather all the known active chromatin markers concentrate immediately adjacent to the SpPr at the upstream chromatin boundary.

Using our knockout MEFs for RRN3 and for UBF, we can at least start to distinguish between these gene states. The loss of RRN3 leads to transcription arrest, however UBF and SL1 are still present on the rDNA, and the gene is still open chromatin. This situation represents the rDNA environment that likely occurs just prior to RPI recruitment and initiation, which would potentially occur just after growth factor pathway stimulation. The loss of UBF leads to the loss of all RPI transcription factors except TTF1 on the rDNA and these genes displayed psoralen inaccessibility, however were not methylated (Hamdane et al. 2014, Chapter 2). This suggests that the rDNA lacking UBF does not become epigenetically silenced and can be potentially reactivated once again, therefore UBF appears to maintain the open chromatin structure of the actively transcribing rDNA. Therefore, loss of UBF does not induce the classical epigenetic silencing and heterochromatin formation but rather a reversible poised state of the rDNA (Chapter 2, Hamdane et al. 2014, Sanij et al. 2008).

Another tool that we have developed in order to investigate the rDNA landscape is the embryonic stem cell lines generated in the lab. Interestingly, we have found that the vast majority if not all of the rRNA genes are active or potentially active in ESCs based on WGBS and psoralen crosslinking analysis of these cells. We have used our wild-type ERCre cells and a commonly used commercial stem cell line, E14, to perform ChIP- seq analysis of these pluripotent ESCs. These results demonstrated a very similar pattern for RPI transcription factor binding of the rDNA. One observed difference could be the occupancy of RPI is less pronounced throughout the gene body and the peak of RRN3 at the 47S promoter is sharp as opposed to the gradual

148

release seen in MEFs. This could indicate differences in transcription rate however, preliminary elongation rate experiments indicate that ESCs do not actually transcribe the rDNA at a faster rate than in differentiated cells.

In ESCs, we also observed strong peaks of RPI, TTF1, and CTCF at the SpPr similarly to what was shown in MEFs. Alignment of H3K4me3 from Encode data sets for ES cells on the rDNA shows that this factor is also found around the SpPr (Hamdane et al. 2014). Therefore, it appears that ESCs demonstrate the same chromatin boundary upstream of the functional rDNA unit and active chromatin marks. As the rRNA genes are all active in ESCs, we are therefore likely seeing the “active rDNA signature” for RPI factors in our ChIP-seq analysis of MEFs and the silenced genes are not distinguishable by the factors we have studied thus far. The chromatin boundary could define potentially active genes but whether they are being actively transcribed or not is unclear.

As discussed in Chapter 2, UBF is implicated in maintaining open chromatin on the active rDNA repeats and TTF1 is involved in delimiting the rDNA unit. We propose to investigate the regulation of rDNA chromatin and the mechanism of silencing by UBF and TTF1 in ESCs. Using directed differentiation (e.g. RA, Chapter 4) would induce silencing of a large portion of the rRNA genes in ESCs and we may see differences in occupancy of certain factors. Inversely, we propose to not only study the role of ribosomal RNA transcription during differentiation by inhibiting RPI transcription through knockout or using differentiation-inducing drugs, but also during reprogramming. We will induce our MEFs using the Yamanaka factors to become induced pluripotent stem cells (iPSCs) (Takahashi & Yamanaka 2006) and study the changes in rDNA chromatin and activity.

We will also study the occupancy of other chromatin marks that are typical of silent or heterochromatic genes in order to further analysis of an active or silent signature of the rDNA. For example, the typical repressive chromatin mark H3K27me3 is usually found over tissue-specific promoters, which remain accessible to transcription factor binding and a paused polymerase. By contrast, another mark of silencing, H3K9me3, marks repetitive constitutive heterochromatin (Becker et al. 2016). G9a, the histone methyltransferase that realizes mono- and dimethylation of H3K9, has been shown to interact with CSB, a chromatin remodeler of the rDNA. H3K9me2/3 and the chromodomain factor heterochromatin protein 1γ (HP1γ), even though they are typically marks associated with heterochromatin formation, have been found associated with active rRNA genes (Bannister et al. 2001, Lachner et al. 2001, Tachibana et al. 2002, Yuan et al. 2007). We will investigate these factors in our ESCs, both wild-type and knockout, and during the exit from pluripotency. H3K9me2/3 is often evenly distributed and its disposition has been attributed to elongation (Becker et al. 2016). Therefore, it is unlikely we would see strong peaks arise after differentiation but there may be differences in large regions, named large organized chromatin K9 modifications (LOCKs) (Wen et al. 2009). H3K9me2/3 enrichment

149

expands during the exit from pluripotency (Becker et al. 2016) therefore we may see differences on the rDNA during our directed differentiation or induced pluripotency assays. In addition, H4K20me3 was shown to be associated with methylated and silent rDNA repeats (Yuan et al. 2007) so it could be a candidate mark to differentiate the active or inactive repeats in our assay.

Surprisingly, even though methylation is a central aspect of gene regulation, there does not exist much data for methylation throughout differentiation. In addition, the rDNA is excluded from the few publically available studies. Therefore, our WGBS studies throughout inhibition of rRNA transcription and differentiation and realignment of what public data exists should clarify the establishment of methylation on the rDNA during development.

Further studies will analyze the rDNA landscape in different cancer cell lines. As we had observed a higher proportion of active rRNA genes in L1210 mouse lymphocytic leukemia cells, we will also investigate if this is a common property of cancer cell lines. Upregulated ribosome biogenesis has been long associated with cancer, though whether this is due to a reactivation of silent rDNA copies or increased transcriptional activity is unknown.

In continuing these studies, we hope to clarify the chromatin landscape on the rDNA and relate the activity status of the rDNA to a certain chromatin structure of the promoter. We will use our conditional knockout MEFs and ESCs, iPSCs generated from the MEFs, and cancer cell lines to continue analysis of the rDNA chromatin landscape. One technique that could help us to determine the chromatin landscape of the different rDNA gene states is sequential ChIP. The active genes could be immunoprecipitated by RPI and then this chromatin could be immunoprecipitated a second time with transcription factors, histones and meCpG antibodies. The same technique could be performed but with TTF1 as the initial ChIP, before and after UBF knockout.

Finally, multiple studies have linked ribosome biogenesis to malignant transformation or to differentiation (Goudarzi & Lindström 2016, Orsolic et al. 2015, Stępiński 2016, Takada & Kurisaki 2015) and several reports have recently proposed a correlation between repression of RPI transcription rate, the exit from pluripotency and differentiation (Hayashi et al. 2014, Watanabe-Susaki et al. 2014, Zhang et al. 2014). Using Activin A to induce differentiation of human ESCs, Woolnough et al. demonstrated a downregulation of rRNA transcription and a reduction in UBF association prior to pluripotency gene expression changes (Woolnough et al. 2016). This same study showed that the inhibition of RPI using CX-5461 induced differentiation. The upregulation of RPI transcription slows differentiation in Drosophila germline stem cells (Zhang et al. 2014) and rRNA transcription decreases during rat muscle cell differentiation (Larson et al. 1993). Not only does inhibition of RPI transcription induce differentiation but upregulation of rRNA transcription by serum-induction can be helpful for reprogrammed cells to overcome the epigenetic barrier to pluripotency (Zhao et al. 2016). The rDNA

150

itself is likely regulated by pluripotency factors as OCT4, SOX2, Nanog, KLF4, STAT3, SMAD1 and C-MYC have been shown to associate with the rDNA (Zentner et al. 2014). It also appears that the mechanism of rDNA silencing is involved in pluripotency regulation. Savic et al. found that processing of the IGS-rRNA to pRNA is developmentally regulated and inhibiting this process impairs TIP5 recruitment to the rDNA and heterochromatin formation though the mechanism of this regulation is still unknown (Savić et al. 2014).

We are interested in studying the role of UBF and RRN3 in pluripotency maintenance. Preliminary tests indicate that the loss of UBF in ESCs leads to morphological changes and hence, probably differentiation (data not shown). Since the knockdown of UBF reversibly reduced the number of active genes being transcribed in somatic cells without inducing cell death (Sanij et al. 2008), it would be interesting to repeat this knockdown experiment in ESCs where all the rRNA genes are potentially active.

5.3 The loss of UBF and genomic instability The rDNA array is unstable and a target for homologous recombination, which leads continously to the amplification or deletion of rDNA repeats. This has been particularly studied in yeast and one potential reason for having so many silenced copies of the rDNA is to stabilize these loci (Kobayashi 2008). When human HCT116 cells were DNMT1/3b inactivated or treated with aza-dC to inhibit or reverse CpG methylation, a large portion of silent rDNA repeats were reactivated. This however did not lead to increase rRNA transcription but instead led to increased genomic instability (Gagnon-Kugler et al. 2009). Indeed, rDNA repeats in human cells have been more recently shown to be highly sensitive to double strand breaks (DSBs) and aberrant homologous recombination leading to a loss of repeats and cell viability (Warmerdam et al. 2016). In addition, UBF appears to have a role in DSB identification or repair as it is found along with γ-H2AX at hotspots of DNA damage (Tchurikov et al. 2016). This is important as, in 54% of solid tumours, rDNA clusters are rearranged prior to the start of tumour expansion (Stults et al. 2009). The silencing of rDNA transcription upon DNA damage is mediated by Nijmegen breakage syndrome protein 1 (NBS1) and its partner TCOF1/Treacle, which activate the ATM-mediated repair pathway (Ciccia et al. 2014, Larsen et al. 2014). If the damage occurs on the rDNA, then that specific NOR is silenced by ATM and repair proteins are recruited to the nucleolar cap which is formed upon transcription shut-down. If the damage occurs in chromatin outside of the nucleolus, then ATM signals for global RPI transcription to stop (Larsen & Stucki 2016).

One interesting aspect of our study is that the loss of UBF in transformed MEFs leads to a synchronous and complete apoptosis (Chapter 3 & Hamdane et al. 2015). However, the loss of RRN3, though rDNA transcription is arrested, does not lead to the same synchronous apoptosis. We believe that the difference between these two situations lies in the different roles of these factors in vivo. In the case of RRN3, transcription is arrested but the genomic landscape of the rDNA is not significantly altered. However, after UBF

151

loss, the rDNA is open and subject to DNA damage, which could lead to genomic instability and cell death we observed. DNA damage factors, such as ATM, NBS1 and Treacle, will be analyzed by ChIP after RRN3 or UBF loss in order to see if it is the loss of transcription or the presence of open chromatin that leads to DNA damage. Indeed, DNA damage was observed after UBF1/2 knockdown but not after RRN3 knockdown (Sanij et al. 2015). Interestingly, these authors found that the splice variant UBF2 has a role in transcription of the histone genes and they believe that the genomic instability they observed following knockdown is due to the role of UBF2 external to ribosome biogenesis. I have proposed the development of inducible knockout cell lines using doxycycline-regulated Cas9 technology. This could be used to generate rescue models of the two splice variants of UBF1 and 2 which would allow further analysis of the individual roles of these factors.

We observe a massive cell death after UBF loss specifically in transformed cells. This could be because transformed cells lack certain checkpoints that normally would cause a cell cycle arrest and ribosome biogenesis is upregulated in these cells (Hanahan & Weinberg 2011, White 2008). This would lead to an open rDNA state and potential DNA damage, leading to genomic instability and death. In primary MEFs, these checkpoints are still in place and therefore apoptosis is not observed. As RRN3 loss inhibits RPI transcription but does not appear to affect the rDNA chromatin landscape, this is likely why the arrest in these cells is different. Again, using CRISPR/Cas9 technology, the effect of the loss of UBF or RRN3 could be analyzed in various cancer cell lines to discern whether this is a general feature of oncogenically transformed cell lines.

5.4 Tissue-specific loss of UBF and RRN3 Due to our discovery that UBF loss induces a synchronized and complete apoptosis in transformed MEFs (Hamdane et al. 2015), we wanted to investigate the potential for targeting early ribosome biogenesis factors for cancer treatment. Despite improvements in cancer treatment, recurrence after therapy and drug resistance are major problems, and tumor heterogeneity is known to contribute to therapy failure (Hanahan & Weinberg 2011). Notably, one type of tumour cells, cancer stem cells, have been shown to survive cancer treatment. Cancer stem cells possess certain characteristics of stem cells, such as the ability to self-renew, but also have the ability to drive tumour growth through aberrant assymetric cell division (Kreso & Dick 2014). Inactivation of the tumor suppressor p53 has been implicated in cancer stem cell differentiation. As we have shown that UBF loss leads to p53-independent cell death (Hamdane et al. 2015), we thought that there may be a possibility of targeting ribosomal RNA transcription factors in order to affect these cells specifically.

Interestingly, inhibiting ribosome biogenesis in the liver by targeted deletion of r-protein S6 has previously been shown to have no phenotype and therefore no toxicity. However, after hepatectomy, where hepatocytes must proliferate rapidly to regenerate tissue, the mice lacking S6 demonstrate cell cycle arrest (Volarevic et al. 2000). Similarly, deletion of dyskerin, a pseudouridine synthase that modifies the rRNA, leads to hepatocyte

152

survival but only in non-proliferative cells (Ge et al. 2010). We wished to develop models where we could specifically inhibit RPI transcription in different cell populations, such as those undergoing proliferation or differentiation. These are essential for cancer development and treatments often target pathways that regulate these processes. However, many cancer drugs have high toxicity for non-cancerous cells. Since we have shown that UBF deletion leads to cell death specifically in oncogenically transformed cells (Hamdane et al. 2015), testing the loss of UBF in various tissues in mouse and generating tumor models or investigating the recovery after damage would demonstrate whether it is only proliferative cells that are affected or if the effect is global.

As a starting point, I generated Ubffl/flAlbCre+ and Rrn3fl/flAlbCre+ mice and studied mouse viability after loss of UBF or RRN3 in the liver. AlbCre transgenic mice have the recombinase driven by the albumin gene promoter which is hepatocyte-specific and activated starting at E10.5 as hepatocytes differentiate during development (Weisend et al. 2009). Both Ubffl/flAlbCre+ and Rrn3fl/flAlbCre+ mice were viable up to 6 months of age and were fertile (data not shown). Livers from these mice appeared morphologically normal, both by analyzing gross structure as well as sections stained with hematoxylin and eosin (H&E). Genotyping revealed a heterogenous population, which could be due to mosaic expression of the target gene or inconsistent recombination that has been reported for many tissue-specific driver Cre mouse strains (Heffner et al. 2012). 80% of the adult mouse liver’s DNA content is found in hepatocytes, with the rest being found in endothelial, Kupffer and other cell types that do not express Albumin (Weisend et al. 2009) so the heterogeneneity could also be due to the mixed population. The mosaisism is not a problem however, because it will be interesting to study knockout cells and wild type cells in the same population, especially after induced regeneration post liver damage (by

CCl4 treatment) or tumor induction (by N-Nitrosodiethylamine treatment) (Ge et al. 2010, Kemp 1995). Indeed, UBF has been shown to be upregulated in hepatocellular carcinomas and is required for cancer cell survival (Huang et al. 2002).

Concurrently, another protocol was tested for tissue-specific knockout of Ubf, this time in the skin. We developed a topical treatment protocol for tamoxifen-induced Cre recombination in the skin of UBFfl/flERCre+/+ mice and achieved excision based on immunofluorescence staining of frozen sections (data not shown). This model will be useful to investigate the role of ribosome biogenesis in differentiation in vivo as the skin is a fast regenerating organ where stem cells differentiate up through the multiple layers starting from the proliferative basal layer (Fuchs 2008).

Unfortunately, due to the necessary rederivitization of the mice and moving of the animal house in 2015, these studies did not advance further than this stage for the time being. Therefore, they were not included in the body of this manuscript. Interestingly, both these models will be significant when they are restored because

153

they allow the possibility of analyzing inhibition of ribosome biogenesis in vivo, where the researcher may compare knockout cells next to non-induced cells in the same population. The ultimate goal would be to generate cancer models in conditional tissue-specific knockout mice for UBF and analyze the effect on tumor cell proliferation.

5.5 Concluding remarks Ribosome biogenesis is central to the regulation of growth control in the cell, and as such it is of fundamental importance for disease development. This process depends on the rate of transcription of the rDNA, which is regulated by the basal RPI factors. We have established models that allow the specific investigation of rRNA transcription, which has already and I believe will continue to lead to a better understanding of the regulation of rDNA activity.

154

References

Adenot PG, Mercier Y, Renard JP, Thompson EM. 1997. Differential H4 acetylation of paternal and maternal chromatin precedes DNA replication and differential transcriptional activity in pronuclei of 1-cell mouse embryos. Development. 124(22):4615–25 Aguirre-Lavin T, Adenot P, Bonnet-Garnier AL, Lehmann GT, Fleurot R, et al. 2012. 3D-FISH analysis of embryonic nuclei in mouse highlights several abrupt changes of nuclear organization during preimplantation development. BMC Developmental Biology. 12(1):1–1 Ahmad K, Henikoff S. 2002. The Histone Variant H3.3 Marks Active Chromatin by Replication-Independent Nucleosome Assembly. Molecular Cell. 9(6):1191–1200 Akamatsu Y, Kobayashi T. 2015. The Human RNA Polymerase I Transcription Terminator Complex Acts as a Replication Fork Barrier That Coordinates the Progress of Replication with rRNA Transcription Activity. Molecular and Cellular Biology. 35(10):1871–81 Albert B, Léger-Silvestre I, Normand C, Ostermaier MK, Pérez-Fernández J, et al. 2011. RNA polymerase I–specific subunits promote polymerase clustering to enhance the rRNA gene transcription cycle. J Cell Biol. 192(2):277–93 Alizadeh Z, Kageyama S-I, Aoki F. 2005. Degradation of maternal mRNA in mouse embryos: selective degradation of specific mRNAs after fertilization. Mol. Reprod. Dev. 72(3):281–90 Amsterdam A, Sadler KC, Lai K, Farrington S, Bronson RT, et al. 2004. Many Ribosomal Protein Genes Are Cancer Genes in Zebrafish. PLoS Biol. 2(5):e139–39 Andersen JS, Lam YW, Leung A, Ong SE, Lyon CE. 2005. Nucleolar proteome dynamics. Nature. 433:77–83 Andersen JS, Lyon CE, Fox AH, Leung AKL, Lam YW, et al. 2002. Directed proteomic analysis of the human nucleolus. Current Biology. 12(1):1–11 Andrew S. 2010. FastQC: A quality control tool for high throughput sequence data. Angus SP, Solomon DA, Kuschel L, Hennigan RF, Knudsen ES. 2003. Retinoblastoma Tumor Suppressor: Analyses of Dynamic Behavior in Living Cells Reveal Multiple Modes of Regulation. Molecular and Cellular Biology. 23(22):8172–88 Aoki F, Worrad DM, Schultz RM. 1997. Regulation of transcriptional activity during the first and second cell cycles in the preimplantation mouse embryo. Developmental Biology. 181(2):296–307 Apicelli AJ, Maggi LB, Hirbe AC, Miceli AP, Olanich ME, et al. 2008. A non-tumor suppressor role for basal p19ARF in maintaining nucleolar structure and function. Molecular and Cellular Biology. 28(3):1068–80 Aprikian P, Moorefield B, Reeder RH. 2001. New Model for the Yeast RNA Polymerase I Transcription Cycle. Molecular and Cellular Biology. 21(15):4847–55 Arabi A, Wu S, Ridderstråle K, Bierhoff H, Shiue C, et al. 2005. c-Myc associates with ribosomal DNA and activates RNA polymerase I transcription. Nat Cell Biol. 7(3):303–10 Bannister AJ, Zegerman P, Partridge JF, Miska EA, Thomas JO, et al. 2001. Selective recognition of methylated lysine 9 on histone H3 by the HP1 chromo domain. Nature. 410(6824):120–24 Bao J, Ma H-Y, Schuster A, Lin Y-M, Yan W. 2013. Incomplete cre-mediated excision leads to phenotypic differences between Stra8-iCre; Mov10l1 lox/loxand Stra8-iCre; Mov10l1 lox/Δmice. Genesis. 51(7):481–90 Bar-Nahum G, Nudler E. 2001. Isolation and Characterization of σ70-Retaining Transcription Elongation Complexes from Escherichia coli. Cell. 106(4):443–51 Barreto G, Schäfer A, Marhold J, Stach D, Swaminathan SK, et al. 2007. Gadd45a promotes epigenetic gene activation by repair-mediated DNA demethylation. Nature. 445(7128):671–75 Bartek J, Bartkova J, Lukas J. 1996. The retinoblastoma protein pathway and the restriction point. Current Opinion in Cell Biology. 8(6):805–14 Bartsch I, Schoneberg C, Grummt I. 1988. Purification and characterization of TTFI, a factor that mediates termination of mouse ribosomal DNA transcription. Molecular and Cellular Biology. 8(9):3891–97 Bates S, Phillips AC, Clark PA, Stott F, Peters G. 1998. p14ARF links the tumour suppressors RB and p53. Nature Bazett-Jones DP, Leblanc B, Herfort M, Moss T. 1994. Short-range DNA looping by the Xenopus HMG-box transcription factor, xUBF. Science. 264(5162):1134–37

155

Bártfai R, Balduf C, Hilton T, Rathmann Y, Hadzhiev Y, et al. 2004. TBP2, a Vertebrate-Specific Member of the TBP Family, Is Required in Embryonic Development of Zebrafish. Current Biology. 14(7):593–98 Becker JS, Nicetto D, Zaret KS. 2016. H3K9me3-Dependent Heterochromatin: Barrier to Cell Fate Changes. Trends in Genetics. 32(1):29–41 Beckmann H, Chen JL, O'Brien T, Tjian R. 1995. Coactivator and promoter-selective properties of RNA polymerase I TAFs. Science. 270(5241):1506–9 Beckouet F, Labarre-Mariotte S, Albert B, Imazawa Y, Werner M, et al. 2008. Two RNA Polymerase I Subunits Control the Binding and Release of Rrn3 during Transcription. Molecular and Cellular Biology. 28(5):1596– 1605 Beerens N, Hoeijmakers JHJ, Kanaar R, Vermeulen W, Wyman C. 2005. The CSB Protein Actively Wraps DNA. Journal of Biological Chemistry. 280(6):4722–29 Bell SP, Jantzen HM, Tjian R. 1990. Assembly of alternative multiprotein complexes directs rRNA promoter selectivity. Genes Dev. 4(6):943–54 Bell SP, Learned RM, Jantzen HM, Tjian R. 1988. Functional cooperativity between transcription factors UBF1 and SL1 mediates human ribosomal RNA synthesis. Science. 241(4870):1192–97 Bell SP, Pikaard CS, Reeder RH, Tjian R. 1989. Molecular mechanisms governing species-specific transcription of ribosomal RNA. Cell. 59(3):489–97 Bergstralh DT, Conti BJ, Moore CB, Brickey WJ, Taxman DJ, Ting JPY. 2007. Global functional analysis of nucleophosmin in Taxol response, cancer, chromatin regulation, and ribosomal DNA transcription. Experimental Cell Research. 313(1):65–76 Berkovich E, Monnat RJ, Kastan MB. 2007. Roles of ATM and NBS1 in chromatin structure modulation and DNA double-strand break repair. Nat Cell Biol. 9(6):683–90 Bernstein BE, Kamal M, Lindblad-Toh K, Bekiranov S, Bailey DK, et al. 2005. Genomic maps and comparative analysis of histone modifications in human and mouse. Cell. 120(2):169–81 Bertwistle D, Sugimoto M, Sherr CJ. 2004. Physical and Functional Interactions of the Arf Tumor Suppressor Protein with Nucleophosmin/B23. Molecular and Cellular Biology. 24(3):985–96 Bhaskara S, Chyla BJ, Amann JM, Knutson SK, Cortez D, et al. 2008. Deletion of Histone Deacetylase 3 Reveals Critical Roles in S Phase Progression and DNA Damage Control. Molecular Cell. 30(1):61–72 Bierhoff H, Dundr M, Michels AA, Grummt I. 2008. Phosphorylation by Casein Kinase 2 Facilitates rRNA Gene Transcription by Promoting Dissociation of TIF-IA from Elongating RNA Polymerase I. Molecular and Cellular Biology. 28(16):4988–98 Billon P, Côté J. 2012. Precise deposition of histone H2A.Z in chromatin for genome expression and maintenance. Biochimica et Biophysica Acta (BBA) - Gene Regulatory Mechanisms. 1819(3-4):290–302 Bird AP, Taggart MH, Gehring CA. 1981. Methylated and unmethylated ribosomal RNA genes in the mouse. Journal of Molecular Biology. 152(1):1–17 Bodem J, Dobreva G, Hoffmann-Rohrer U, Iben S, Zentgraf H, et al. 2000. TIF-IA, the factor mediating growth- dependent control of ribosomal RNA synthesis, is the mammalian homolog of yeast Rrn3p. EMBO Rep. 1(2):171–75 Bogoyevitch MA, Kobe B. 2006. Uses for JNK: the Many and Varied Substrates of the c-Jun N-Terminal Kinases. Microbiology and Molecular Biology Reviews. 70(4):1061–95 Boisvert F-M, van Koningsbruggen S, Navascués J, Lamond AI. 2007. The multifunctional nucleolus. Nat Rev Mol Cell Biol. 8(7):574–85 Bolger AM, Lohse M, Usadel B. 2014. Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics. 30(15):2114–20 Borsos M, Torres-Padilla M-E. 2016. Building up the nucleus: nuclear organization in the establishment of totipotency and pluripotency during mammalian development. Genes Dev. 30(6):611–21 Boulon S, Westman BJ, Hutten S, Boisvert F-M, Lamond AI. 2010. The nucleolus under stress. Molecular Cell. 40(2):216–27 Bouniol C, Nguyen E, Debey P. 1995. Endogenous Transcription Occurs at the 1-Cell Stage in the Mouse Embryo. Experimental Cell Research. 218(1):57–62

156

Bradsher J, Auriol J, de Santis LP, Iben S, Vonesch J-L, et al. 2002. CSB Is a Component of RNA Pol I Transcription. Molecular Cell. 10(4):819–29 Brady SN, Yu Y, Maggi LB, Weber JD. 2004. ARF impedes NPM/B23 shuttling in an Mdm2-sensitive tumor suppressor pathway. Molecular and Cellular Biology. 24(21):9327–38 Brewer BJ, Lockshon D, Fangman WL. 1992. The arrest of replication forks in the rDNA of yeast occurs independently of transcription. Cell. 71(2):267–76 Britton-Davidian J, Cazaux B, Catalan J. 2012. Chromosomal dynamics of nucleolar organizer regions (NORs) in the house mouse: micro-evolutionary insights. Heredity (Edinb). 108(1):68–74 Brown DD, Gurdon JB. 1964. Absence of Ribosomal RNA Synthesis in the Anucleolate Mutant of Xenopus Laevis. PNAS. 51(1):139–46 Brown SE, Szyf M. 2008. Dynamic epigenetic states of ribosomal RNA promoters during the cell cycle. Cell Cycle. 7(3):382–90 Bryja V, Bonilla S, Arenas E. 2006. Derivation of mouse embryonic stem cells. Nat Protoc. 1(4):2082–87 Budde A, Grummt I. 1999. p53 represses ribosomal gene transcription. Oncogene Burgess-Beusse B, Farrell C, Gaszner M, Litt M, Mutskov V, et al. 2002. The insulation of genes from external enhancers and silencing chromatin. PNAS. 99 Suppl 4(suppl 4):16433–37 Burkhalter MD, Sogo JM. 2004. rDNA Enhancer Affects Replication Initiation and Mitotic Recombination. Molecular Cell. 15(3):409–21 Bywater MJ, Poortinga G, Sanij E, Hein N, Peck A, et al. 2012. Inhibition of RNA polymerase I as a therapeutic strategy to promote cancer-specific activation of p53. Cancer Cell. 22(1):51–65 Cao J, Wu L, Zhang S-M, Lu M, Cheung WKC, et al. 2016. An easy and efficient inducible CRISPR/Cas9 platform with improved specificity for multiple gene targeting. Nucleic Acids Research. 44(19):e149–49 Carlson LL, Page AW, Bestor TH. 1992. Properties and localization of DNA methyltransferase in preimplantation mouse embryos: implications for genomic imprinting. Genes Dev. 6(12B):2536–41 Castiglia CL, Flint SJ. 1983. Effects of adenovirus infection on rRNA synthesis and maturation in HeLa cells. Molecular and Cellular Biology. 3(4):662–71 Caudron-Herger M, Pankert T, Seiler J, Németh A, Voit R, et al. 2015. Alu element-containing RNAs maintain nucleolar structure and function. The EMBO Journal. 34(22):2758–74 Caudy AA, Pikaard CS. 2002. Xenopus Ribosomal RNA Gene Intergenic Spacer Elements Conferring Transcriptional Enhancement and Nucleolar Dominance-like Competition in Oocytes. Journal of Biological Chemistry. 277(35):31577–84 Cavanaugh AH, Hempel WM, Taylor LJ, Rogalsky V, Todorov G, Rothblum LI. 1995. Activity of RNA polymerase I transcription factor UBF blocked by Rb gene product. Nature. 374(6518):177–80 Cavanaugh AH, Hirschler-Laszkiewicz I, Hu Q, Dundr M, Smink T, et al. 2002. Rrn3 Phosphorylation Is a Regulatory Checkpoint for Ribosome Biogenesis. Journal of Biological Chemistry. 277(30):27423–32 Chao JC, Wan XS, Engelsberg BN, Rothblum LI, Billings PC. 1996. Intracellular distribution of HMG1, HMG2 and UBF change following treatment with cisplatin. Biochimica et Biophysica Acta (BBA) - Gene Structure and Expression. 1307(2):213–19 Chen D, Belmont AS, Huang S. 2004. Upstream binding factor association induces large-scale chromatin decondensation. PNAS. 101(42):15106–11 Chen D, Huang S. 2001. Nucleolar Components Involved in Ribosome Biogenesis Cycle between the Nucleolus and Nucleoplasm in Interphase Cells. J Cell Biol. 153(1):169–76 Chen H, Li Z, Haruna K, Li Z, Li Z, et al. 2008. Early pre-implantation lethality in mice carrying truncated mutation in the RNA polymerase 1-2 gene. Biochemical and Biophysical Research Communications. 365(4):636–42 Chen W, Jia W, Wang K, Zhou Q, Leng Y, et al. 2012. Retinoic acid regulates germ cell differentiation in mouse embryonic stem cells through a Smad-dependent pathway. Biochemical and Biophysical Research Communications. 418(3):571–77 Cho H, Kim T-K, Mancebo H, Lane WS, Flores O, Reinberg D. 1999. A protein phosphatase functions to recycle RNA polymerase II. Genes Dev. 13(12):1540–52 Chouinard LA. 1971. A light- and electron-microscope study of the nucleolus during growth of the oocyte in the prepubertal mouse. Journal of Cell Science. 9(3):637–63

157

Ciccia A, Huang J-W, Izhar L, Sowa ME, Harper JW, Elledge SJ. 2014. Treacher Collins syndrome TCOF1 protein cooperates with NBS1 in the DNA damage response. Proc. Natl. Acad. Sci. U.S.A. 111(52):18631–36 Citterio E, Van Den Boom V, Schnitzler G, Kanaar R, Bonte E, et al. 2000. ATP-Dependent Chromatin Remodeling by the Cockayne Syndrome B DNA Repair-Transcription-Coupling Factor. Molecular and Cellular Biology. 20(20):7643–53 Clarke EM, Peterson CL, Brainard AV, Riggs DL. 1996. Regulation of the RNA Polymerase I and III Transcription Systems in Response to Growth Conditions. Journal of Biological Chemistry. 271(36):22189–95 Claypool JA, French SL, Johzuka K, Eliason K, Vu L, et al. 2004. Tor pathway regulates Rrn3p-dependent recruitment of yeast RNA polymerase I to the promoter but does not participate in alteration of the number of active genes. Mol. Biol. Cell. 15(2):946–56 Clift D, Schuh M. 2013. Restarting life: fertilization and the transition from meiosis to mitosis. Nat Rev Mol Cell Biol. 14(9):549–62 Colombo E, Marine J-C, Danovi D, Falini B, Pelicci PG. 2002. Nucleophosmin regulates the stability and transcriptional activity of p53. Nat Cell Biol. 4(7):529–33 Comai L, Tanese N, Tjian R. 1992. The TATA-binding protein and associated factors are integral components of the RNA polymerase I transcription factor, SL1. Cell. 68(5):965–76 Conconi A, Widmer RM, Koller T, Sogo J. 1989. Two different chromatin structures coexist in ribosomal RNA genes throughout the cell cycle. Cell. 57(5):753–61 Copenhaver GP, Pikaard CS. 1996. RFLP and physical mapping with an rDNA-specific endonuclease reveals that nucleolus organizer regions of Arabidopsis thaliana adjoin the telomeres on chromosomes 2 and 4. Plant J. 9(2):259–72 Copenhaver GP, Putnam CD, Denton ML, Pikaard CS. 1994. The RNA polymerase I transcription factor UBF is a sequence-tolerant HMG-box protein that can recognize structured nucleic acids. Nucleic Acids Research. 22(13):2651–57 Coppoolse ER, de Vroomen MJ, van Gennip F, Hersmus BJM, van Haaren MJJ. 2005. Size Does Matter: Cre- mediated Somatic Deletion Efficiency Depends on the Distance Between the Target lox-Sites. Plant Mol Biol. 58(5):687–98 Cuddapah S, Jothi R, Schones DE, Roh T-Y, Cui K, Zhao K. 2009. Global analysis of the insulator binding protein CTCF in chromatin barrier regions reveals demarcation of active and repressive domains. Genome Res. 19(1):24–32 Dammann R, Lucchini R, Koller T, Sogo JM. 1993. Chromatin structures and transcription of rDNA in yeast Saccharomyces cerevisiae. Nucleic Acids Research. 21(10):2331–38 Danielian PS, Muccino D, Rowitch DH, Michael SK, McMahon AP. 1998. Modification of gene activity in mouse embryos in utero by a tamoxifen-inducible form of Cre recombinase. Current Biology. 8(24):1323–26 Daniely Y, Borowiec JA. 2000. Formation of a Complex between Nucleolin and Replication Protein a after Cell Stress Prevents Initiation of DNA Replication. J Cell Biol. 149(4):799–810 Danilova N, Gazda HT. 2015. Ribosomopathies: how a common root can cause a tree of pathologies. Disease Models & Mechanisms. 8(9):1013–26 David-Pfeuty T, Nouvian-Dooghe Y. 2002. Human p14(Arf): an exquisite sensor of morphological changes and of short-lived perturbations in cell cycle and in nucleolar function. Oncogene. 21(44):6779–90 De Rooij KE, Dorsman JC, Smoor MA, Dunnen Den JT, Van Ommen GJ. 1996. Subcellular localization of the Huntington's disease gene product in cell lines by immunofluorescence and biochemical subcellular fractionation. Hum. Mol. Genet. 5(8):1093–99 De Winter RFJ, Moss T. 1986. Spacer promoters are essential for efficient enhancement of X. laevis ribosomal transcription. Cell. 44(2):313–18 De Winter RFJ, Moss T. 1987. A complex array of sequences enhances ribosomal transcription in Xenopus laevis. Journal of Molecular Biology. 196(4):813–27 Deaton AM, Bird A. 2011. CpG islands and the regulation of transcription. Genes Dev. 25(10):1010–22 Deisenroth C, Zhang Y. 2010. Ribosome biogenesis surveillance: probing the ribosomal protein-Mdm2-p53 pathway. Oncogene. 29(30):4253–60

158

Denissov S, van Driel M, Voit R, Hekkelman M, Hulsen T, et al. 2007. Identification of novel functional TBP-binding sites and general factor repertoires. The EMBO Journal. 26(4):944–54 Derenzini M, Montanaro L, Treré D. 2009. What the nucleolus says to a tumour pathologist. Histopathology. 54(6):753–62 Dexheimer TS. 2013. DNA Repair Pathways and Mechanisms. In DNA Repair of Cancer Stem Cells, pp. 19–32. Springer Netherlands Diermeier SD, Németh A, Rehli M, Grummt I, Längst G. 2013. Chromatin-Specific Regulation of Mammalian rDNA Transcription by Clustered TTF-I Binding Sites. PLoS Genet. 9(9):e1003786–12 Diesch J, Hannan RD, Sanij E. 2015. Genome wide mapping of UBF binding-sites in mouse and human cell lines. Genom Data. 3:103–5 Dow LE, Fisher J, O'Rourke KP, Muley A, Kastenhuber ER, et al. 2015. Inducible in vivo genome editing with CRISPR-Cas9. Nat. Biotechnol. 33(4):390–94 Dowen JM, Bilodeau S, Orlando DA, Hübner MR, Abraham BJ, et al. 2013. Multiple structural maintenance of chromosome complexes at transcriptional regulatory elements. Stem Cell Reports. 1(5):371–78 Dönmez H, Akalın H, Karaman Y, Demirtaş H, İmamoğlu N, Özkul Y. 2005. Evaluation of the Nucleolar Organizer Regions in Alzheimer’s Disease. Gerontology. 51(5):297–301 Drakas R, Tu X, Baserga R. 2004. Control of cell size through phosphorylation of upstream binding factor 1 by nuclear phosphatidylinositol 3-kinase. Proc. Natl. Acad. Sci. U.S.A. 101(25):9272–76 Drygin D, Rice WG, Grummt I. 2010. The RNA polymerase I transcription machinery: an emerging target for the treatment of cancer. Annu. Rev. Pharmacol. Toxicol. 50:131–56 Dunaway M, Dröge P. 1989. Transactivation of the Xenopus rRNA gene promoter by its enhancer. Nature. 341(6243):657–59 Dundr M, Hoffmann-Rohrer U, Hu Q, Grummt I, Rothblum LI, et al. 2002. A kinetic framework for a mammalian RNA polymerase in vivo. Science. 298(5598):1623–26 Dyson N. 1998. The regulation of E2F by pRB-family proteins. Genes Dev. 12(15):2245–62 Eberhard D, Tora L, Egly JM, Grummt I. 1993. A TBP-containing multiprotein complex (TIF-IB) mediates transcription specificity of murine RNA polymerase I. Nucleic Acids Research. 21(18):4180–86 Elmore S. 2007. Apoptosis: A Review of Programmed Cell Death. Toxicol Pathol. 35(4):495–516 Emmott E, Hiscox JA. 2009. Nucleolar targeting: the hub of the matter. EMBO Rep. 10(3):231–38 Engel W, Zenzes MT, Schmid M. 1977. Activation of mouse ribosomal RNA genes at the 2-cell stage. Hum. Genet. 38(1):57–63 Farley BM, Ryder SP. 2008. Regulation of Maternal mRNAs in Early Development. Critical Reviews in Biochemistry and Molecular Biology. 43(2):135–62 Farley KI, Surovtseva Y, Merkel J, Baserga SJ. 2015. Determinants of mammalian nucleolar architecture. Chromosoma. 124(3):323–31 Fath S, Milkereit P, Peyroche G, Riva M, Carles C, Tschochner H. 2001. Differential roles of phosphorylation in the formation of transcriptional active RNA polymerase I. PNAS. 98(25):14334–39 Feil R, Brocard J, Mascrez B, LeMeur M, Metzger D, Chambon P. 1996. Ligand-activated site-specific recombination in mice. Proc. Natl. Acad. Sci. U.S.A. 93(20):10887–90 Feil R, Wagner J, Metzger D, Chambon P. 1997. Regulation of Cre recombinase activity by mutated estrogen receptor ligand-binding domains. Biochemical and biophysical …. 237(3):752–57 Felsenfeld G, Burgess-Beusse B, Farrell C, Gaszner M, Ghirlando R, et al. 2004. Chromatin boundaries and chromatin domains. Cold Spring Harb. Symp. Quant. Biol. 69(0):245–50 Fetherston J, Werner E, Patterson R. 1984. Processing of the external transcribed spacer of murine rRNA and site of action of actinomydn D. Nucleic Acids Research. 12(18):7187–98 Filippova GN, Cheng MK, Moore JM, Truong J-P, Hu YJ, et al. 2005. Boundaries between Chromosomal Domains of X Inactivation and Escape Bind CTCF and Lack CpG Methylation during Early Development. Developmental Cell. 8(1):31–42 Fléchon JE, Kopecny V. 1998. The nature of the “nucleolus precursor body” in early preimplantation embryos: a review of fine-structure cytochemical, immunocytochemical and autoradiographic data related to nucleolar function. Zygote. 6(02):183–91

159

Floutsakou I, Agrawal S, Nguyen TT, Seoighe C, Ganley ARD, McStay B. 2013. The shared genomic architecture of human nucleolar organizer regions. Genome Res. 23(12):2003–12 Fontoura BM, Sorokina EA, David E, Carroll RB. 1992. p53 is covalently linked to 5.8S rRNA. Molecular and Cellular Biology. 12(11):5145–51 Freed EF, Bleichert F, Dutca LM, Baserga SJ. 2010. When ribosomes go bad: diseases of ribosome biogenesis. Mol Biosyst. 6(3):481–93 French SL, Osheim YN, Cioci F, Nomura M, Beyer AL. 2003. In Exponentially Growing Saccharomyces cerevisiae Cells, rRNA Synthesis Is Determined by the Summed RNA Polymerase I Loading Rate Rather than by the Number of Active Genes. Molecular and Cellular Biology. 23(5):1558–68 Friedrich JK, Panov KI, Cabart P, Russell J, Zomerdijk JCBM. 2005. TBP-TAF Complex SL1 Directs RNA Polymerase I Pre-initiation Complex Formation and Stabilizes Upstream Binding Factor at the rDNA Promoter. Journal of Biological Chemistry. 280(33):29551–58 Fritze CE, Verschueren K, Strich R, Esposito RE. 1997. Direct evidence for SIR2 modulation of chromatin structure in yeast rDNA. The EMBO Journal. 16(21):6495–6509 Fuchs E. 2008. Skin stem cells: rising to the surface. J Cell Biol. 180(2):273–84 Fulka H, Aoki F. 2016. Nucleolus Precursor Bodies and Ribosome Biogenesis in Early Mammalian Embryos: Old Theories and New Discoveries. Biology of Reproduction Fulka H, Langerova A. 2014. The maternal nucleolus plays a key role in centromere satellite maintenance during the oocyte to embryo transition. Development. 141(8):1694–1704 Fulka H, MARTINKOVA S, Kyogoku H, Langerova A, Fulka J. 2012. Production of giant mouse oocyte nucleoli and assessment of their protein content. J. Reprod. Dev. 58(3):371–76 Fulka J Jr., Moor RM, Loi P, Fulka J. 2003. Enucleolation of porcine oocytes. Theriogenology. 59(8):1879–85 Gadal O, Labarre S, Boschiero C, Thuriaux P. 2002. Hmo1, an HMG‐box protein, belongs to the yeast ribosomal DNA transcription system. The EMBO Journal. 21(20):5498–5507 Gagnon-Kugler T, Langlois F, Stefanovsky V, Lessard F, Moss T. 2009. Loss of human ribosomal gene CpG methylation enhances cryptic RNA polymerase II transcription and disrupts ribosomal RNA processing. Molecular Cell. 35(4):414–25 Ganot P, Jády BE, Bortolin M-L, Darzacq X, Kiss T. 1999. Nucleolar Factors Direct the 2′-O-Ribose Methylation and Pseudouridylation of U6 Spliceosomal RNA. Molecular and Cellular Biology. 19(10):6906–17 Gasperowicz M, Natale DRC. 2011. Establishing Three Blastocyst Lineages—Then What? Biology of Reproduction. 84(4):621–30 Gazdag E, Rajkovic A, Torres-Padilla M-E, Tora L. 2007. Analysis of TATA-binding protein 2 (TBP2) and TBP expression suggests different roles for the two proteins in regulation of gene expression during oogenesis and early mouse development. Reproduction. 134(1):51–62 Gaztelumendi N, Nogués C. 2014. Chromosome Instability in mouse Embryonic Stem Cells. Sci Rep. 4: Ge J, Rudnick DA, He J, Crimmins DL, Ladenson JH, et al. 2010. Dyskerin ablation in mouse liver inhibits rRNA processing and cell division. Molecular and Cellular Biology. 30(2):413–22 Gerber J-K, Gögel E, Berger C, Wallisch M, Müller F, et al. 1997. Termination of Mammalian rDNA Replication: Polar Arrest of Replication Fork Movement by Transcription Termination Factor TTF-I. Cell. 90(3):559–67 Ghoshal K, Jacob ST. 1997. An alternative molecular mechanism of action of 5-fluorouracil, a potent anticancer drug. Biochemical Pharmacology. 53(11):1569–75 Gianaroli L, Magli MC, Ferraretti AP, Fortini D, Grieco N. 2003. Pronuclear morphology and chromosomal abnormalities as scoring criteria for embryo selection. Fertility and Sterility. 80(2):341–49 Giroux S, Tremblay M, Bernard D, Cardin-Girard J-F, Aubry S, et al. 1999. Embryonic death of Mek1-deficient mice reveals a role for this kinase in angiogenesis in the labyrinthine region of the placenta. Current Biology. 9(7):369–72 Golbus MS, Calarco PG, Epstein CJ. 1973. The effects of inhibitors of RNA synthesis (α‐amanitin and actinomycin D) on preimplantation mouse embryogenesis. Journal of Experimental Zoology Part A: Ecological Genetics and Physiology. 186(2):207–16 Gomez-Roman N, Grandori C, Eisenman RN, White RJ. 2003. Direct activation of RNA polymerase III transcription by c-Myc. Nature. 421(6920):290–94

160

Gonzalez IL, Sylvester JE. 1995. Complete Sequence of the 43-kb Human Ribosomal DNA Repeat: Analysis of the Intergenic Spacer. Genomics. 27(2):320–28 Goodfellow SJ, Zomerdijk JCBM. 2012. Basic Mechanisms in RNA Polymerase I Transcription of the Ribosomal RNA Genes. In Epigenetics: Development and Disease, Vol. 61, ed. TK Kundu, pp. 211–36. Dordrecht: Springer Netherlands Goodpasture C, Bloom SE. 1975. Visualization of nucleolar organizer regions im mammalian chromosomes using silver staining. Chromosoma. 53(1):37–50 Gorski JJ, Pathak S, Panov K, Kasciukovic T, Panova T, et al. 2007. A novel TBP‐associated factor of SL1 functions in RNA polymerase I transcription. The EMBO Journal. 26(6):1560–68 Goudarzi KM, Lindström MS. 2016. Role of ribosomal protein mutations in tumor development (Review). Int. J. Oncol. 48(4):1313–24 Grandori C, Gomez-Roman N, Felton-Edkins ZA, Ngouenet C, Galloway DA, et al. 2005. c-Myc binds to human ribosomal DNA and stimulates transcription of rRNA genes by RNA polymerase I. Nat Cell Biol. 7(3):311–18 Granneman S, Baserga SJ. 2005. Crosstalk in gene expression: coupling and co-regulation of rDNA transcription, pre-ribosome assembly and pre-rRNA processing. Current Opinion in Cell Biology. 17(3):281–86 Greaves M. 2015. Cancer's Darwinian dilemma: an evolutionary tale in three acts. BMJ. 351:h6581 Grimaldi G, Fiorentini P, Di Nocera PP. 1990. Spacer promoters are orientation-dependent activators of pre-rRNA transcription in Drosophila melanogaster. Molecular and Cellular Biology. 10(9):4667–77 Grinstein E, Shan Y, Karawajew L, Snijders PJF, Meijer CJLM, et al. 2006. Cell Cycle-controlled Interaction of Nucleolin with the Retinoblastoma Protein and Cancerous Cell Transformation. Journal of Biological Chemistry. 281(31):22223–35 Grisendi S, Mecucci C, Falini B, Pandolfi PP. 2006. Nucleophosmin and cancer. Nat Rev Cancer. 6(7):493–505 Grob A, Colleran C, McStay B. 2014. Construction of synthetic nucleoli in human cells reveals how a major functional nuclear domain is formed and propagated through cell division. Genes Dev. 28(3):220–30 Grozdanov P, Georgiev O, Karagyozov L. 2003. Complete sequence of the 45-kb mouse ribosomal DNA repeat: analysis of the intergenic spacer☆. Genomics. 82(6):637–43 Grummt I. 2003. Life on a planet of its own: regulation of RNA polymerase I transcription in the nucleolus. Genes Dev. 17(14):1691–1702 Grummt I, Kuhn A, Bartsch I, Rosenbauer H. 1986a. A transcription terminator located upstream of the mouse rDNA initiation site affects rRNA synthesis. Cell. 47(6):901–11 Grummt I, Maier U, Öhrlein A, Hassouna N, Bachellerie J-P. 1985. Transcription of mouse rDNA terminates downstream of the 3′ end of 28S RNA and involves interaction of factors with repeated sequences in the 3′ spacer. Cell. 43(3):801–10 Grummt I, Rosenbauer H, Niedermeyer I, Maier U, Öhrlein A. 1986b. A repeated 18 bp sequence motif in the mouse rDNA spacer mediates binding of a nuclear factor and transcription termination. Cell. 45(6):837–46 Guetg C, Santoro R. 2012. Formation of nuclear heterochromatin. Epigenetics. 7(8):811–14 Haaf T, Hayman DL, Schmid M. 1991. Quantitative determination of rDNA transcription units in vertebrate cells. Experimental Cell Research. 193(1):78–86 Habibi E, Brinkman AB, Arand J, Kroeze LI, Kerstens HHD, et al. 2013. Whole-genome bisulfite sequencing of two distinct interconvertible DNA methylomes of mouse embryonic stem cells. Cell Stem Cell. 13(3):360–69 Hall DB, Wade JT, Struhl K. 2006. An HMG Protein, Hmo1, Associates with Promoters of Many Ribosomal Protein Genes and throughout the rRNA Gene Locus in Saccharomyces cerevisiae. Molecular and Cellular Biology. 26(9):3672–79 Haltiner MM, Smale ST, Tjian R. 1986. Two distinct promoter elements in the human rRNA gene identified by linker scanning mutagenesis. Molecular and Cellular Biology. 6(1):227–35 Hamatani T, Carter MG, Sharov AA, Ko MSH. 2004. Dynamics of Global Gene Expression Changes during Mouse Preimplantation Development. Developmental Cell. 6(1):117–31 Hamdane N, Herdman C, Mars J-C, Stefanovsky V, Tremblay MG, Moss T. 2015. Depletion of the cisplatin targeted HMGB-box factor UBF selectively induces p53-independent apoptotic death in transformed cells. Oncotarget. 6(29):27519–36

161

Hamdane N, Stefanovsky VY, Tremblay MG, Németh A, Paquet E, et al. 2014. Conditional Inactivation of Upstream Binding Factor Reveals Its Epigenetic Functions and the Existence of a Somatic Nucleolar Precursor Body. PLoS Genet. 10(8):e1004505 Hamdane N, Tremblay MG, Dillinger S, Stefanovsky VY, Németh A, Moss T. 2016. Disruption of the UBF gene induces aberrant somatic nucleolar bodies and disrupts embryo nucleolar precursor bodies. Gene Hanada K, Song CZ, Yamamoto K, Yano K, Maeda Y, et al. 1996. RNA polymerase I associated factor 53 binds to the nucleolar transcription factor UBF and functions in specific rDNA transcription. The EMBO Journal. 15(9):2217 Hanahan D, Weinberg RA. 2011. Hallmarks of Cancer: The Next Generation. Cell. 144(5):646–74 Hannan KM, Brandenburger Y, Jenkins A, Sharkey K, Cavanaugh A, et al. 2003. mTOR-Dependent Regulation of Ribosomal Gene Transcription Requires S6K1 and Is Mediated by Phosphorylation of the Carboxy-Terminal Activation Domain of the Nucleolar Transcription Factor UBF. Molecular and Cellular Biology. 23(23):8862– 77 Hannan KM, Hannan RD, Smith SD, Jefferson LS, Lun M, Rothblum LI. 2000a. Rb and p130 regulate RNA polymerase I transcription: Rb disrupts the interaction between UBF and SL-1. , Published online: 12 October 2000; | doi:10.1038/sj.onc.1203875. 19(43):4988–99 Hannan KM, Kennedy BK, Cavanaugh AH, Hannan RD, Hirschler-Laszkiewicz I, et al. 2000b. RNA polymerase I transcription in confluent cells: Rb downregulates rDNA transcription during confluence-induced cell cycle arrest. , Published online: 21 July 2000; | doi:10.1038/sj.onc.1203690. 19(31):3487–97 Hannan KM, Sanij E, Rothblum LI, Hannan RD, Pearson RB. 2013. Dysregulation of RNA polymerase I transcription during disease. Biochimica et Biophysica Acta (BBA) - Gene Regulatory Mechanisms. 1829(3- 4):342–60 Hannan RD, Cavanaugh A, Hempel WM, Moss T, Rothblum L. 1999. Identification of a mammalian RNA polymerase I holoenzyme containing components of the DNA repair/replication system. Nucleic Acids Research. 27(18):3720–27 Harding SM, Boiarsky JA, Greenberg RA. 2015. ATM Dependent Silencing Links Nucleolar Chromatin Reorganization to DNA Damage Recognition. 13(2):251–59 Hayashi S, Tenzen T, McMahon AP. 2003. Maternal inheritance of Cre activity in a Sox2Cre deleter strain. Genesis. 37(2):51–53 Hayashi Y, Kuroda T, Kishimoto H, Wang C, Iwama A, Kimura K. 2014. Downregulation of rRNA Transcription Triggers Cell Differentiation. PLoS ONE. 9(5):e98586–10 Heffner CS, Herbert Pratt C, Babiuk RP, Sharma Y, Rockwood SF, et al. 2012. Supporting conditional mouse mutagenesis with a comprehensive cre characterization resource. Nat Commun. 3:1218–19 Hein N, Ganley A, Sanij E, Quin J, Hannan KM. 2012. The Nucleolus and Ribosomal Genes in Aging and Senescence Heitz E. 1931. Die Ursache der gesetzmässigen Zahl, Lage, Form und Grösse pflanzlicher Nukleolen. Planta. 12(4):775–844 Heix J, Vente A, Voit R, Budde A, Michaelidis TM, Grummt I. 1998. Mitotic silencing of human rRNA synthesis: inactivation of the promoter selectivity factor SL1 by cdc2/cyclin B-mediated phosphorylation. The EMBO Journal. 17(24):7373–81 Heix J, Zomerdijk JCBM, Ravanpay A, Tjian R, Grummt I. 1997. Cloning of murine RNA polymerase I-specific TAF factors: conserved interactions between the subunits of the species-specific transcription initiation factor TIF- IB/SL1. PNAS. 94(5):1733–38 Hempel WM, Cavanaugh AH, Hannan RD, Taylor L, Rothblum LI. 1996. The species-specific RNA polymerase I transcription factor SL-1 binds to upstream binding factor. Molecular and Cellular Biology. 16(2):557–63 Henderson AS, Warburton D, Atwood KC. 1972. Location of ribosomal DNA in the human chromosome complement. PNAS. 69(11):3394–98 Henderson S, Sollner-Webb B. 1986. A transcriptional terminator is a novel element of the promoter of the mouse ribosomal RNA gene. Cell. 47(6):891–900 Hetman M, Pietrzak M. 2012. Emerging roles of the neuronal nucleolus. Trends in Neurosciences. 35(5):305–14

162

Héliot L, Kaplan H, Lucas L, Klein C, Beorchia A, et al. 1997. Electron tomography of metaphase nucleolar organizer regions: evidence for a twisted-loop organization. Mol. Biol. Cell. 8(11):2199–2216 Héliot L, Mongelard F, Klein C, O'Donohue MF, Chassery JM, et al. 2000. Nonrandom distribution of metaphase AgNOR staining patterns on human acrocentric chromosomes. J. Histochem. Cytochem. 48(1):13–20 Hirschler-Laszkiewicz I, Cavanaugh AH, Mirza A, Lun M, Hu Q, et al. 2003. Rrn3 Becomes Inactivated in the Process of Ribosomal DNA Transcription. Journal of Biological Chemistry. 278(21):18953–59 Hiscox JA. 2002. The nucleolus – a gateway to viral infection? Arch. Virol. 147(6):1077–89 Hochstatter J, Hölzel M, Rohrmoser M, Schermelleh L, Leonhardt H, et al. 2012. Myb-binding Protein 1a (Mybbp1a) Regulates Levels and Processing of Pre-ribosomal RNA. Journal of Biological Chemistry. 287(29):24365–77 Howlett SK, Reik W. 1991. Methylation levels of maternal and paternal genomes during preimplantation development. Development. 113(1):119–27 Huang R, Wu T, Xu L, Liu A, Ji Y, Hu G. 2002. Upstream binding factor up-regulated in hepatocellular carcinoma is related to the survival and cisplatin-sensitivity of cancer cells. The FASEB Journal. 16(3):293–301 Indra AK, Warot X, Brocard J, Bornert JM, Xiao JH, et al. 1999. Temporally-controlled site-specific mutagenesis in the basal layer of the epidermis: comparison of the recombinase activity of the tamoxifen-inducible Cre- ER(T) and Cre-ER(T2) recombinases. Nucleic Acids Research. 27(22):4324–27 Itahana K, Bhat KP, Jin A, Itahana Y, Hawke D, et al. 2003. Tumor suppressor ARF degrades B23, a nucleolar protein involved in ribosome biogenesis and cell proliferation. Molecular Cell. 12(5):1151–64 Jacobson MR, Cao LG, Taneja K, Singer RH, Wang YL, Pederson T. 1997. Nuclear domains of the RNA subunit of RNase P. Journal of Cell Science. 110(7):829–37 Jacobson MR, Pederson T. 1998. Localization of signal recognition particle RNA in the nucleolus of mammalian cells. PNAS. 95(14):7981–86 Jansa P, Grummt I. 1999. Mechanism of transcription termination: PTRF interacts with the largest subunit of RNA polymerase I and dissociates paused transcription complexes from yeast and mouse. Mol. Gen. Genet. 262(3):508–14 Jantzen H-M, Admon A, Bell SP, Tjian R. 1990. Nucleolar transcription factor hUBF contains a DNA-binding motif with homology to HMG proteins. Nature. 344(6269):830–36 Jantzen HM, Chow AM, King DS, Tjian R. 1992. Multiple domains of the RNA polymerase I activator hUBF interact with the TATA-binding protein complex hSL1 to mediate transcription. Genes Dev. 6(10):1950–63 Jarmuz-Szymczak M, Janiszewska J, Szyfter K, Shaffer LG. 2014. Narrowing the localization of the region breakpoint in most frequent Robertsonian translocations. Chromosome Res. 22(4):517–32 Jean S, Mikryukov A, Tremblay MG, Baril J, Guillou F, et al. 2010. Extended-Synaptotagmin-2 Mediates FGF Receptor Endocytosis and ERK Activation In Vivo. Developmental Cell. 19(3):426–39 Jin R, Zhou W. 2016. TIF-IA: An oncogenic target of pre-ribosomal RNA synthesis. Biochim. Biophys. Acta. 1866(2):189–96 Jones HS, Kawauchi J, Braglia P, Alen CM, Kent NA, Proudfoot NJ. 2007. RNA polymerase I in yeast transcribes dynamic nucleosomal rDNA. Nat Struct Mol Biol. 14(2):123–30 Jordan EG, McGovern JH. 1981. The quantitative relationship of the fibrillar centres and other nucleolar components to changes in growth conditions, serum deprivation and low doses of actinomycin D in cultured diploid human fibroblasts (strain MRC-5). Journal of Cell Science. 52:373–89 Jordan P, Carmo-Fonseca M. 1998. Cisplatin inhibits synthesis of ribosomal RNA in vivo. Nucleic Acids Research. 26(12):2831–36 Jordan P, Mannervik M, Tora L, Carmo-Fonseca M. 1996. In vivo evidence that TATA-binding protein/SL1 colocalizes with UBF and RNA polymerase I when rRNA synthesis is either active or inactive. J Cell Biol. 133(2):225–34 Kaeberlein M, McVey M, Guarente L. 1999. The SIR2/3/4 complex and SIR2 alone promote longevity in Saccharomyces cerevisiae by two different mechanisms. Genes Dev. 13(19):2570–80 Kagey MH, Newman JJ, Bilodeau S, Zhan Y, Orlando DA, et al. 2010. Mediator and cohesin connect gene expression and chromatin architecture. Nature. 467(7314):430–35 Kang H, Shin J-H. 2015. Repression of rRNA transcription by PARIS contributes to Parkinson's disease. 73:220–28

163

Kasahara K, Ohtsuki K, Ki S, Aoyama K, Takahashi H, et al. 2007. Assembly of Regulatory Factors on rRNA and Ribosomal Protein Genes in Saccharomyces cerevisiae. Molecular and Cellular Biology. 27(19):6686–6705 Kemp CJ. 1995. Hepatocarcinogenesis in p53‐deficient mice. Molecular Carcinogenesis. 12(3):132–36 Kempers-Veenstra AE, Oliemans J, Offenberg H, Dekker AF, Piper PW, et al. 1986. 3'-End formation of transcripts from the yeast rRNA operon. The EMBO Journal. 5(10):2703–10 Kennedy BK, Austriaco NR, Zhang J, Guarente L. 1995. Mutation in the silencing gene SIR4 can delay aging in S. cerevisiae. Cell. 80(3):485–96 Kennedy BK, Gotta M, Sinclair DA, Mills K, McNabb DS, et al. 1997. Redistribution of silencing proteins from telomeres to the nucleolus is associated with extension of life span in S. cerevisiae. Cell. 89(3):381–91 Kihm AJ, Hershey JC, Haystead TA, Madsen CS, Owens GK. 1998. Phosphorylation of the rRNA transcription factor upstream binding factor promotes its association with TATA binding protein. Proc. Natl. Acad. Sci. U.S.A. 95(25):14816–20 Kim M, Habiba A, Doherty JM, Mills JC, Mercer RW, Huettner JE. 2009. Regulation of mouse embryonic stem cell neural differentiation by retinoic acid. Developmental Biology. 328(2):456–71 Klein J, Grummt I. 1999. Cell cycle-dependent regulation of RNA polymerase I transcription: the nucleolar transcription factor UBF is inactive in mitosis and early G1. PNAS. 96(11):6096–6101 Klose RJ, Bird AP. 2006. Genomic DNA methylation: the mark and its mediators. Trends in Biochemical Sciences. 31(2):89–97 Ko A, Han SY, Song J. 2016. Dynamics of ARF regulation that control senescence and cancer. BMB Rep, Nov.,, pp. 598–606 Kobayashi H, Sakurai T, Imai M, Takahashi N, Fukuda A, et al. 2012. Contribution of Intragenic DNA Methylation in Mouse Gametic DNA Methylomes to Establish Oocyte-Specific Heritable Marks. PLoS Genet. 8(1):e1002440–14 Kobayashi T. 2008. A new role of the rDNA and nucleolus in the nucleus—rDNA instability maintains genome integrity. BioEssays. 30(3):267–72 Kobayashi T. 2011a. How does genome instability affect lifespan? Genes Cells. 16(6):617–24 Kobayashi T. 2011b. Regulation of ribosomal RNA gene copy number and its role in modulating genome integrity and evolutionary adaptability in yeast. Cellular and Molecular Life Sciences. 68(8):1395–1403 Kobayashi T. 2014. Ribosomal RNA gene repeats, their stability and cellular senescence. Proc. Jpn. Acad., Ser. B. 90(4):119–29 Kobayashi T, Ganley ARD. 2005. Recombination Regulation by Transcription-Induced Cohesin Dissociation in rDNA Repeats. Science. 309(5740):1581–84 Kobayashi T, Hidaka M, Nishizawa M, Horiuchi T. 1992. Identification of a site required for DNA replication fork blocking activity in the rRNA gene cluster in Saccharomyces cerevisiae. Mol. Gen. Genet. 233(3):355–62 Kobayashi T, Horiuchi T, Tongaonkar P, Vu L, Nomura M. 2004. SIR2 Regulates Recombination between Different rDNA Repeats, but Not Recombination within Individual rRNA Genes in Yeast. Cell. 117(4):441–53 Kone MC, Fleurot R, Chebrout M, Debey P, Beaujean N, Bonnet-Garnier A. 2016. Three-Dimensional Distribution of UBF and Nopp140 in Relationship to Ribosomal DNA Transcription During Mouse Preimplantation Development. Biology of Reproduction. 94(4):95–95 Kornienko AE, Guenzl PM, Barlow DP, Pauler FM. 2013. Gene regulation by the act of long non-coding RNA transcription. BMC Biol. 11(1):59 Kreso A, Dick JE. 2014. Evolution of the cancer stem cell model. Cell Stem Cell. 14(3):275–91 Krueger F, Andrews SR. 2011. Bismark: a flexible aligner and methylation caller for Bisulfite-Seq applications. Bioinformatics. 27(11):1571–72 Kruhlak M, Crouch EE, Orlov M, Montaño C, Gorski SA, et al. 2007. The ATM repair pathway inhibits RNA polymerase I transcription in response to chromosome breaks. Nature. 447(7145):730–34 Kuhn A, Grummt I. 1987. A novel promoter in the mouse rDNA spacer is active in vivo and in vitro. The EMBO Journal. 6(11):3487–92 Kuhn A, Grummt I. 1992. Dual role of the nucleolar transcription factor UBF: trans-activator and antirepressor. PNAS. 89(16):7340–44

164

Kuhn A, Stefanovsky V, Grummt I. 1993. The nucleolar transcription activator UBF relieves Ku antigen-mediated repression of mouse ribosomal gene transcription. Nucleic Acids Research. 21(9):2057–63 Kuhn A, Vente A, Dorée M, Grummt I. 1998. Mitotic phosphorylation of the TBP-containing factor SL1 represses ribosomal gene transcription. Journal of Molecular Biology. 284(1):1–5 Kuhn A, Voit R, Stefanovsky V, Evers R, Bianchi M, Grummt I. 1994. Functional differences between the two splice variants of the nucleolar transcription factor UBF: the second HMG box determines specificity of DNA binding and transcriptional activity. The EMBO Journal. 13(2):416 Kuhn C-D, Geiger SR, Baumli S, Gartmann M, Gerber J, et al. 2007. Functional Architecture of RNA Polymerase I. Cell. 131(7):1260–72 Kumazawa T, Nishimura K, Katagiri N, Hashimoto S, Hayashi Y, Kimura K. 2015. Gradual reduction in rRNA transcription triggers p53 acetylation and apoptosis via MYBBP1A. Sci Rep. 5:10854 Kuo BA, Gonzalez IL, Gillespie DA, Sylvester JE. 1996. Human Ribosomal RNA Variants from a Single Individual and Their Expression in Different Tissues. Nucleic Acids Research. 24(23):4817–24 Kurihara Y, Suh DS, Suzuki H, Moriwaki K. 1994. Chromosomal locations of Ag-NORs and clusters of ribosomal DNA in laboratory strains of mice. Mammalian Genome. 5(4):225–28 Kuroda T, Murayama A, Katagiri N, Ohta Y-M, Fujita E, et al. 2011. RNA content in the nucleolus alters p53 acetylation via MYBBP1A. The EMBO Journal. 30(6):1054–66 Lachner M, O'Carroll D, Rea S, Mechtler K, Jenuwein T. 2001. Methylation of histone H3 lysine 9 creates a binding site for HP1 proteins. Nature. 410(6824):116–20 Laferté A, Favry E, Sentenac A, Riva M, Carles C, Chédin S. 2006. The transcriptional activity of RNA polymerase I is a key determinant for the level of all ribosome components. Genes Dev. 20(15):2030–40 Lafontaine DLJ. 2015. Noncoding RNAs in eukaryotic ribosome biogenesis and function. Nat Struct Mol Biol. 22(1):11–19 Langmead B, Salzberg SL. 2012. Fast gapped-read alignment with Bowtie 2. Nat. Methods. 9(4):357–59 Larsen DH, Hari F, Clapperton JA, Gwerder M, Gutsche K, et al. 2014. The NBS1–Treacle complex controls ribosomal RNA transcription in response to DNA damage. Nat Cell Biol. 16(8):792–803 Larsen DH, Stucki M. 2016. Nucleolar responses to DNA double-strand breaks. Nucleic Acids Research. 44(2):538–44 Larson DE, Xie W, Glibetic M, O'Mahony D, Sells BH, Rothblum LI. 1993. Coordinated decreases in rRNA gene transcription factors and rRNA synthesis during muscle cell differentiation. PNAS. 90(17):7933–36 Längst G, Becker PB, Grummt I. 1998. TTF‐I determines the chromatin architecture of the active rDNA promoter. The EMBO Journal. 17(11):3135–45 Learned RM, Learned TK, Haltiner MM, Tjian RT. 1986. Human rRNA transcription is modulated by the coordinate binding of two factors to an upstream control element. Cell. 45(6):847–57 Leblanc B, Read C, Moss T. 1993. Recognition of the Xenopus ribosomal core promoter by the transcription factor xUBF involves multiple HMG box domains and leads to an xUBF interdomain interaction. The EMBO Journal. 12(2):513 Lempiäinen H, Shore D. 2009. Growth control and ribosome biogenesis. Current Opinion in Cell Biology. 21(6):855–63 Lessard F, Morin F, Ivanchuk S, Langlois F, Stefanovsky V, et al. 2010. The ARF Tumor Suppressor Controls Ribosome Biogenesis by Regulating the RNA Polymerase I Transcription Factor TTF-I. Molecular Cell. 38(4):539–50 Lessard F, Stefanovsky V, Tremblay MG, Moss T. 2012. The cellular abundance of the essential transcription termination factor TTF-I regulates ribosome biogenesis and is determined by MDM2 ubiquitinylation. Nucleic Acids Research. 40(12):5357–67 Leung AKL, Gerlich D, Miller G, Lyon C, Lam YW, et al. 2004. Quantitative kinetic analysis of nucleolar breakdown and reassembly during mitosis in live human cells. J Cell Biol. 166(6):787–800 Li H, Handsaker B, Wysoker A, Fennell T, Ruan J, et al. 2009. The Sequence Alignment/Map format and SAMtools. Bioinformatics. 25(16):2078–79 Li J, Längst G, Grummt I. 2006. NoRC-dependent nucleosome positioning silences rRNA genes. The EMBO Journal. 25(24):5735–41

165

Li L, Zheng P, Dean J. 2010. Maternal control of early mouse development. Development. 137(6):859–70 Li Y, Schwab C, Ryan SL, Papaemmanuil E, Robinson HM, et al. 2014. Constitutional and somatic rearrangement of chromosome 21 in acute lymphoblastic leukaemia. Nature. 508(7494):98–102 Li Z, Hann SR. 2012. Nucleophosmin is essential for c-Myc nucleolar localization and c-Myc-mediated rDNA transcription. Oncogene. 32(15):1988–94 Li Z, Hann SR. 2009. The Myc-nucleophosmin-ARF network: a complex web unveiled. Cell Cycle. 8(17):2703–7 Lin C-J, Conti M, Ramalho-Santos M. 2013. Histone variant H3.3 maintains a decondensed chromatin state essential for mouse preimplantation development. Development. 140(17):3624–34 Lin C-J, Koh FM, Wong P, Conti M, Ramalho-Santos M. 2014. Hira-Mediated H3.3 Incorporation Is Required for DNA Replication and Ribosomal RNA Transcription in the Mouse Zygote. Developmental Cell. 30(3):268–79 Lin CH, Platt MD, Ficarro SB, Hoofnagle MH, Shabanowitz J, et al. 2006a. Mass spectrometric identification of phosphorylation sites of rRNA transcription factor upstream binding factor. AJP: Cell Physiology. 292(5):C1617–24 Lin CY, Navarro S, Reddy S, Comai L. 2006b. CK2-mediated stimulation of Pol I transcription by stabilization of UBF-SL1 interaction. Nucleic Acids Research. 34(17):4752–66 Long MA, Rossi FMV. 2009. Silencing Inhibits Cre-Mediated Recombination of the Z/AP and Z/EG Reporters in Adult Cells. PLoS ONE. 4(5):e5435 Loonstra A, Vooijs M, Beverloo HB, Allak Al B, van Drunen E, et al. 2001. Growth inhibition and DNA damage induced by Cre recombinase in mammalian cells. Proc. Natl. Acad. Sci. U.S.A. 98(16):9209–14 Loppin B, Bonnefoy E, Anselme C, Laurençon A, Karr TL, Couble P. 2005. The histone H3.3 chaperone HIRA is essential for chromatin assembly in the male pronucleus. Nature. 437(7063):1386–90 Louvet E, Junéra HR, Le Panse S, Hernandez-Verdun D. 2005. Dynamics and compartmentation of the nucleolar processing machinery. Experimental Cell Research. 304(2):457–70 López-Otín C, Blasco MA, Partridge L, Serrano M, Kroemer G. 2013. The Hallmarks of Aging. Cell. 153(6):1194– 1217 Lundqvist J, EL Andaloussi-Lilja J, Svensson C, Gustafsson Dorfh H, Forsby A. 2013. Optimisation of culture conditions for differentiation of C17.2 neural stem cells to be used for in vitro toxicity tests. Toxicology in Vitro. 27(5):1565–69 MacInnes AW. 2016. The role of the ribosome in the regulation of longevity and lifespan extension. Wiley Interdiscip Rev RNA. 7(2):198–212 Mahajan PB. 1994. Modulation of transcription of rRNA genes by rapamycin. International Journal of Immunopharmacology. 16(9):711–21 Mais C, Wright JE, Prieto J-L, Raggett SL, McStay B. 2005. UBF-binding site arrays form pseudo-NORs and sequester the RNA polymerase I transcription machinery. Genes Dev. 19(1):50–64 Manuelidis L, Borden J. 1988. Reproducible compartmentalization of individual chromosome domains in human CNS cells revealed by in situ hybridization and three-dimensional reconstruction. Chromosoma. 96(6):397– 410 Maréchal A, Zou L. 2013. DNA damage sensing by the ATM and ATR kinases. Cold Spring Harbor Perspectives in Biology. 5(9):a012716–16 Marikawa Y, Alarcón VB. 2009. Establishment of trophectoderm and inner cell mass lineages in the mouse embryo. Mol. Reprod. Dev. 76(11):1019–32 Martin J, Han C, Gordon LA, Terry A, Prabhakar S, et al. 2004. The sequence and analysis of duplication-rich human chromosome 16. Nature. 432(7020):988–94 Matthews DA. 2001. Adenovirus Protein V Induces Redistribution of Nucleolin and B23 from Nucleolus to Cytoplasm. J. Virol. 75(2):1031–38 Mayer C. 2005. The nucleolus as a stress sensor: JNK2 inactivates the transcription factor TIF-IA and downregulates rRNA synthesis. Genes Dev. 19(8):933–41 Mayer C, Neubert M, Grummt I. 2008. The structure of NoRC‐associated RNA is crucial for targeting the chromatin remodelling complex NoRC to the nucleolus. EMBO Rep. 9(8):774–80 Mayer C, Schmitz K-M, Li J, Grummt I, Santoro R. 2006. Intergenic transcripts regulate the epigenetic state of rRNA genes. Molecular Cell. 22(3):351–61

166

Mayer C, Zhao J, Yuan X, Grummt I. 2004. mTOR-dependent activation of the transcription factor TIF-IA links rRNA synthesis to nutrient availability. Genes Dev. 18(4):423–34 Mayer W, Niveleau A, Walter J, Fundele R, Haaf T. 2000. Embryogenesis: Demethylation of the zygotic paternal genome. Nature. 403(6769):501–2 McCann KL, Baserga SJ. 2013. Mysterious Ribosomopathies. Science. 341(6148):849–50 McClintock B. 1934. The relation of a particular chromosomal element to the development of the nucleoli in Zea mays. Z.Zellforsch. 21(2):294–326 McGowan PO, Sasaki A, Huang TCT, Unterberger A, Suderman M, et al. 2008. Promoter-wide hypermethylation of the ribosomal RNA gene promoter in the suicide brain. PLoS ONE. 3(5):e2085 McStay B. 2016. Nucleolar organizer regions: genomic “dark matter” requiring illumination. Genes Dev. 30(14):1598–1610 McStay B, Frazier MW, Reeder RH. 1991a. xUBF contains a novel dimerization domain essential for RNA polymerase I transcription. Genes Dev. 5(11):1957–68 McStay B, Grummt I. 2008. The Epigenetics of rRNA Genes: From Molecular to Chromosome Biology. 24(1):131– 57 McStay B, Hu CH, Pikaard CS, Reeder RH. 1991b. xUBF and Rib 1 are both required for formation of a stable polymerase I promoter complex in X. laevis. The EMBO Journal. 10(8):2297 McStay B, Reeder RH. 1990. An RNA polymerase I termination site can stimulate the adjacent ribosomal gene promoter by two distinct mechanisms in Xenopus laevis. Genes Dev. 4(7):1240–51 McStay B, Reeder RH. 1986. A termination site for Xenopus RNA polymerase I also acts as an element of an adjacent promoter. Cell. 47(6):913–20 Meehan RR, Pennings S, Stancheva I. 2001. Lashings of DNA methylation, forkfuls of chromatin remodeling. Genes Dev. 15(24):3231–36 Mehta IS, Figgitt M, Clements CS, Kill IR, Bridger JM. 2007. Alterations to nuclear architecture and genome behavior in senescent cells. Ann. N. Y. Acad. Sci. 1100:250–63 Meneghini MD, Wu M, Madhani HD. 2003. Conserved Histone Variant H2A.Z Protects Euchromatin from the Ectopic Spread of Silent Heterochromatin. Cell. 112(5):725–36 Metzger D, Clifford J, Chiba H, Chambon P. 1995. Conditional site-specific recombination in mammalian cells using a ligand-dependent chimeric Cre recombinase. Proc. Natl. Acad. Sci. U.S.A. 92(15):6991–95 Milkereit P, Tschochner H. 1998. A specialized form of RNA polymerase I, essential for initiation and growth- dependent regulation of rRNA synthesis, is disrupted during transcription. The EMBO Journal. 17(13):3692– 3703 Miller G, Panov KI, Friedrich JK, Trinkle-Mulcahy L, Lamond AI, Zomerdijk JCBM. 2001. hRRN3 is essential in the SL1-mediated recruitment of RNA Polymerase I to rRNA gene promoters. The EMBO Journal. 20(6):1373– 82 Mirre C, Stahl A. 1981. Ultrastructural organization, sites of transcription and distribution of fibrillar centres in the nucleolus of the mouse oocyte. Journal of Cell Science. 48(1):105–26 Missirlis PI, Smailus DE, Holt RA. 2006. A high-throughput screen identifying sequence and promiscuity characteristics of the loxP spacer region in Cre-mediated recombination. BMC Genomics. 7(1):73 Misteli T. 2001. The concept of self-organization in cellular architecture. J Cell Biol. 155(2):181–86 Misteli T, Phair RD. 2000. High mobility of proteins in the mammalian cell nucleus. Nature. 404(6778):604–9 Monk D. 2015. Germline-derived DNA methylation and early embryo epigenetic reprogramming: The selected survival of imprints. The International Journal of Biochemistry & Cell Biology. 67:128–38 Monk M, Boubelik M, Lehnert S. 1987. Temporal and regional changes in DNA methylation in the embryonic, extraembryonic and germ cell lineages during mouse embryo development. Development. 99(3):371–82 Moore HM, Bai B, Boisvert F-M, Latonen L, Rantanen V, et al. 2011. Quantitative Proteomics and Dynamic Imaging of the Nucleolus Reveal Distinct Responses to UV and Ionizing Radiation. Mol Cell Proteomics. 10(10):M111.009241–41 Moorefield B, Greene EA, Reeder RH. 2000. RNA polymerase I transcription factor Rrn3 is functionally conserved between yeast and human. PNAS. 97(9):4724–29

167

Morgan GT, Reeder RH, Bakken AH. 1983. Transcription in cloned spacers of Xenopus laevis ribosomal DNA. PNAS. 80(21):6490–94 Moss T. 1983. A transcriptional function for the repetitive ribosomal spacer in Xenopus laevis. Nature. 302(5905):223–28 Moss T, Birnstiel ML. 1979. The putative promoter of a Xenopus laevis ribosomal gene is reduplicated. Nucleic Acids Research. 6(12):3733–43 Moss T, Boseley PG, Birnsteil ML. 1980. More ribosomal spacer sequences from Xenopus laevis. Nucleic Acids Research. 8(3):467–85 Moss T, Langlois F, Gagnon-Kugler T, Stefanovsky V. 2007. A housekeeper with power of attorney: the rRNA genes in ribosome biogenesis. Cellular and Molecular Life Sciences. 64(1):29–49 Moss T, Mitchelson K, de Winter R. 1985. The promotion of ribosomal transcription in eukaryotes. Oxf. Surv. Eukaryot. Genes. 2:207–50 Moss T, Stefanovsky VY. 2002. At the center of eukaryotic life. Cell. 109(5):545–48 Mougey EB, O'Reilly M, Osheim Y, Miller OL, Beyer A, Sollner-Webb B. 1993. The terminal balls characteristic of eukaryotic rRNA transcription units in chromatin spreads are rRNA processing complexes. Genes Dev. 7(8):1609–19 Mukhopadhyay J, Kapanidis AN, Mekler V, Kortkhonjia E, Ebright YW, Ebright RH. 2001. Translocation of σ70 with RNA Polymerase during Transcription. Cell. 106(4):453–63 Murano K, Okuwaki M, Hisaoka M, Nagata K. 2008. Transcription Regulation of the rRNA Gene by a Multifunctional Nucleolar Protein, B23/Nucleophosmin, through Its Histone Chaperone Activity. Molecular and Cellular Biology. 28(10):3114–26 Murayama A, Ohmori K, Fujimura A, Minami H, Yasuzawa-Tanaka K, et al. 2008. Epigenetic Control of rDNA Loci in Response to Intracellular Energy Status. Cell. 133(4):627–39 Nagy A. 2000. Cre recombinase: the universal reagent for genome tailoring. Genesis. 26(2):99–109 Narla A, Ebert BL. 2010. Ribosomopathies: human disorders of ribosome dysfunction. Blood. 115(16):3196–3205 Newton K, Petfalski E, Tollervey D, Caceres JF. 2003. Fibrillarin Is Essential for Early Development and Required for Accumulation of an Intron-Encoded Small Nucleolar RNA in the Mouse. Molecular and Cellular Biology. 23(23):8519–27 Németh A, Guibert S, Tiwari VK, Ohlsson R, Längst G. 2008. Epigenetic regulation of TTF-I-mediated promoter- terminator interactions of rRNA genes. The EMBO Journal. 27(8):1255–65 Németh A, Längst G. 2011. Genome organization in and around the nucleolus. Trends in Genetics. 27(4):149–56 Németh A, Strohner R, Grummt I, Längst G. 2004. The chromatin remodeling complex NoRC and TTF-I cooperate in the regulation of the mammalian rRNA genes in vivo. Nucleic Acids Research. 32(14):4091–99 Nishimura K, Kumazawa T, Kuroda T, Katagiri N, Tsuchiya M, et al. 2015. Perturbation of Ribosome Biogenesis Drives Cells into Senescence through 5S RNP-Mediated p53 Activation. 10(8):1310–23 Nonchev S, Tsanev R. 1990. Protamine–histone replacement and DNA replication in the male mouse pronucleus. Mol. Reprod. Dev. 25(1):72–76 Nothias JY, Miranda M, DePamphilis ML. 1996. Uncoupling of transcription and translation during zygotic gene activation in the mouse. The EMBO Journal. 15(20):5715 O'Mahony DJ, Rothblum LI. 1991. Identification of two forms of the RNA polymerase I transcription factor UBF. PNAS. 88(8):3180–84 O'Sullivan AC, Sullivan GJ, McStay B. 2002. UBF Binding In Vivo Is Not Restricted to Regulatory Sequences within the Vertebrate Ribosomal DNA Repeat. Molecular and Cellular Biology. 22(2):657–58 Ogushi S, Palmieri C, Fulka H, Saitou M, Miyano T, Fulka J. 2008. The maternal nucleolus is essential for early embryonic development in mammals. Science. 319(5863):613–16 Ogushi S, Saitou M. 2010. The nucleolus in the mouse oocyte is required for the early step of both female and male pronucleus organization. J. Reprod. Dev. 56(5):495–501 Okano M, Bell DW, Haber DA, Li E. 1999. DNA Methyltransferases Dnmt3a and Dnmt3b Are Essential for De Novo Methylation and Mammalian Development. Cell. 99(3):247–57 Olson MOJ. 2004. Sensing cellular stress: another new function for the nucleolus? Sci. STKE. 2004(224):pe10– pe10

168

Ono W, Hayashi Y, Yokoyama W, Kuroda T, Kishimoto H, et al. 2014. The Nucleolar Protein Myb-binding Protein 1A (MYBBP1A) Enhances p53 Tetramerization and Acetylation in Response to Nucleolar Disruption. Journal of Biological Chemistry. 289(8):4928–40 Orsolic I, Jurada D, Pullen N, Oren M, Eliopoulos AG, Volarevic S. 2015. The relationship between the nucleolus and cancer: Current evidence and emerging paradigms. Seminars in Cancer Biology. 1–15 Osheim YN, Mougey EB, Windle J, Anderson M, O'Reilly M, et al. 1996. Metazoan rDNA enhancer acts by making more genes transcriptionally active. J Cell Biol. 133(5):943–54 Oskarsson T, Trumpp A. 2005. The Myc trilogy: lord of RNA polymerases. Nat Cell Biol. 7(3):215–17 Oswald J, Engemann S, Lane N, Mayer W, Olek A, et al. 2000. Active demethylation of the paternal genome in the mouse zygote. Current Biology. 10(8):475–78 Paalman MH, Henderson SL, Sollner-Webb B. 1995. Stimulation of the mouse rRNA gene promoter by a distal spacer promoter. Molecular and Cellular Biology. 15(8):4648–56 Paget MS, Helmann JD. 2003. The σ 70 family of sigma factors. Genome Biol. 4(1):203 Panov KI, Friedrich JK, Russell J, Zomerdijk JCBM. 2006a. UBF activates RNA polymerase I transcription by stimulating promoter escape. The EMBO Journal. 25(14):3310–22 Panov KI, Panova TB, Gadal O, Nishiyama K, Saito T, et al. 2006b. RNA Polymerase I-Specific Subunit CAST/hPAF49 Has a Role in the Activation of Transcription by Upstream Binding Factor. Molecular and Cellular Biology. 26(14):5436–48 Panova TB, Panov KI, Russell J, Zomerdijk JCBM. 2006. Casein Kinase 2 Associates with Initiation-Competent RNA Polymerase I and Has Multiple Roles in Ribosomal DNA Transcription. Molecular and Cellular Biology. 26(16):5957–68 Paynton BV, Rempel R, Bachvarova R. 1988. Changes in state of adenylation and time course of degradation of maternal mRNAs during oocyte maturation and early embryonic development in the mouse. Developmental Biology. 129(2):304–14 Pederson T. 2011. The Nucleolus. Cold Spring Harbor Perspectives in Biology. 3(3):a000638–38 Petes TD. 1979. Yeast ribosomal DNA genes are located on chromosome XII. PNAS. 76(1):410–14 Peyroche G, Milkereit P, Bischler N, Tschochner H, Schultz P, et al. 2000. The recruitment of RNA polymerase I on rDNA is mediated by the interaction of the A43 subunit with Rrn3. The EMBO Journal. 19(20):5473–82 Pietrzak M, Rempala G, Nelson PT, Zheng J-J, Hetman M. 2011. Epigenetic Silencing of Nucleolar rRNA Genes in Alzheimer's Disease. PLoS ONE. 6(7):e22585 Piko L, Clegg KB. 1982. Quantitative changes in total RNA, total poly(A), and ribosomes in early mouse embryos. Developmental Biology. 89(2):362–78 Ploton D, Gilbert N, Menager M, Kaplan H, Adnet JJ. 1994. Three-dimensional co-localization of nucleolar argyrophilic components and DNA in cell nuclei by confocal microscopy. J. Histochem. Cytochem. 42(2):137–48 Pondarré C, Strumberg D, Fujimori A, Torres-León R, Pommier Y. 1997. In vivo sequencing of camptothecin- induced topoisomerase I cleavage sites in human colon carcinoma cells. Nucleic Acids Research. 25(20):4111–16 Poortinga G, Hannan KM, Snelling H, Walkley CR, Jenkins A, et al. 2004. MAD1 and c-MYC regulate UBF and rDNA transcription during granulocyte differentiation. The EMBO Journal. 23(16):3325–35 Poortinga G, Quinn LM, Hannan RD. 2014. Targeting RNA polymerase I to treat MYC-driven cancer. 1–10 Poortinga G, Wall M, Sanij E, Siwicki K, Ellul J, et al. 2011. c-MYC coordinately regulates ribosomal gene chromatin remodeling and Pol I availability during granulocyte differentiation. Nucleic Acids Research. 39(8):3267–81 Powers T, Walter P. 1999. Regulation of Ribosome Biogenesis by the Rapamycin-sensitive TOR-signaling Pathway in Saccharomyces cerevisiae. Mol. Biol. Cell. 10(4):987–1000 Prieto JL, McStay B. 2007. Recruitment of factors linking transcription and processing of pre-rRNA to NOR chromatin is UBF-dependent and occurs independent of transcription in human cells. Genes Dev. 21(16):2041–54 Prior CP, Cantor CR, Johnson EM, Littau VC, Allfrey VG. 1983. Reversible changes in nucleosome structure and histone H3 accessibility in transcriptionally active and inactive states of rDNA chromatin. Cell. 34(3):1033–42 Prives C. 1998. Signaling to p53: breaking the MDM2-p53 circuit. Cell. 95(1):5–8

169

Probst AV, Okamoto I, Casanova M, Marjou El F, Le Baccon P, Almouzni G. 2010. A Strand-Specific Burst in Transcription of Pericentric Satellites Is Required for Chromocenter Formation and Early Mouse Development. Developmental Cell. 19(4):625–38 Putnam CD, Copenhaver GP, Denton ML, Pikaard CS. 1994. The RNA polymerase I transactivator upstream binding factor requires its dimerization domain and high-mobility-group (HMG) box 1 to bend, wrap, and positively supercoil enhancer DNA. Molecular and Cellular Biology. 14(10):6476–88 Puvion-Dutilleul F, Christensen ME. 1993. Alterations of fibrillarin distribution and nucleolar ultrastructure induced by adenovirus infection. Eur. J. Cell Biol. 61(1):168–76 Quin J, Chan KT, Devlin JR, Cameron DP, Diesch J, et al. 2016. Inhibition of RNA polymerase I transcription initiation by CX-5461 activates non-canonical ATM/ATR signaling. Oncotarget. 7(31):49800–49818 Quin JE, Devlin JR, Cameron D, Hannan KM, Pearson RB, Hannan RD. 2014. Targeting the nucleolus for cancer intervention. BBA - Molecular Basis of Disease. 1842(6):802–16 Raška I, Shaw PJ, Cmarko D. 2006. Structure and function of the nucleolus in the spotlight. Current Opinion in Cell Biology. 18(3):325–34 Rebello RJ, Kusnadi E, Cameron DP, Pearson HB, Lesmana A, et al. 2016. The dual inhibition of RNA Pol I transcription and PIM kinase as a new therapeutic approach to treat advanced prostate cancer. Clin. Cancer Res. clincanres.0124.2016 Rey JP, Scott R, Müller H. 1993. Induction and removal of interstrand crosslinks in the ribosomal RNA genes of lymphoblastoid cell lines from patients with Fanconi anemia. Mutation Research/Fundamental and Molecular Mechanisms of Mutagenesis. 289(2):171–80 Rieker C, Engblom D, Kreiner G, Domanskyi A, Schober A, et al. 2011. Nucleolar Disruption in Dopaminergic Neurons Leads to Oxidative Damage and Parkinsonism through Repression of Mammalian Target of Rapamycin Signaling. J. Neurosci. 31(2):453–60 Riley KJ-L, Maher LJ. 2007. p53 RNA interactions: new clues in an old mystery. RNA. 13(11):1825–33 Rinn JL, Chang HY. 2012. Genome Regulation by Long Noncoding RNAs. Annu. Rev. Biochem. 81(1):145–66 Roos WP, Kaina B. 2006. DNA damage-induced cell death by apoptosis. Trends in Molecular Medicine. 12(9):440– 50 Rossetti S, Wierzbicki AJ, Sacchi N. 2016. Mammary epithelial morphogenesis and early breast cancer. Evidence of involvement of basal components of the RNA Polymerase I transcription machinery. Cell Cycle. 15(18):2515–26 Rothblum K, Hu Q, Penrod Y, Rothblum LI. 2014. Selective Inhibition of rDNA Transcription by a Small-Molecule Peptide That Targets the Interface between RNA Polymerase I and Rrn3. 12(11):1586–96 Roussel P, André C, Masson C, Géraud G, Hernandez-Verdun D. 1993. Localization of the RNA polymerase I transcription factor hUBF during the cell cycle. Journal of Cell Science. 104 ( Pt 2)(2):327–37 Rubbi CP, Milner J. 2003. Disruption of the nucleolus mediates stabilization of p53 in response to DNA damage and other stresses. The EMBO Journal. 22(22):6068–77 Rudloff U, Eberhard D, Tora L, Stunnenberg H, Grummt I. 1994. TBP-associated factors interact with DNA and govern species specificity of RNA polymerase I transcription. The EMBO Journal. 13(11):2611 Ruggero D. 2012. Revisiting the Nucleolus: From Marker to Dynamic Integrator of Cancer Signaling. Science Signaling. 5(241):pe38–pe38 Ruggero D, Pandolfi PP. 2003. Does the ribosome translate cancer? Nat Rev Cancer. 3(3):179–92 Russell J, Zomerdijk JCBM. 2005. RNA-polymerase-I-directed rDNA transcription, life and works. Trends in Biochemical Sciences. 30(2):87–96 Saka K, Ide S, Ganley ARD, Kobayashi T. 2013. Cellular senescence in yeast is regulated by rDNA noncoding transcription. Curr. Biol. 23(18):1794–98 Sakai K, Ohta T, Minoshima S, Kudoh J, Wang Y, et al. 1995. Human ribosomal RNA gene cluster: identification of the proximal end containing a novel tandem repeat sequence. Genomics. 26(3):521–26 Sancar A, Lindsey-Boltz LA, Ünsal-Kaçmaz K, Linn S. 2004. Molecular Mechanisms of Mammalian DNA Repair and the DNA Damage Checkpoints. Annu. Rev. Biochem. 73(1):39–85 Sander JD, Joung JK. 2014. CRISPR-Cas systems for editing, regulating and targeting genomes. Nat. Biotechnol. 32(4):347–55

170

Sandmeier JJ, French S, Osheim Y, Cheung WL, Gallo CM, et al. 2002. RPD3 is required for the inactivation of yeast ribosomal DNA genes in stationary phase. The EMBO Journal. 21(18):4959–68 Sanij E, Diesch J, Lesmana A, Poortinga G, Hein N, et al. 2015. A novel role for the Pol I transcription factor UBTF in maintaining genome stability through the regulation of highly transcribed Pol II genes. Genome Res. 25(2):201–12 Sanij E, Poortinga G, Sharkey K, Hung S, Holloway TP, et al. 2008. UBF levels determine the number of active ribosomal RNA genes in mammals. J Cell Biol. 183(7):1259–74 Santenard A, Ziegler-Birling C, Koch M, Tora L, Bannister AJ, Torres-Padilla M-E. 2010. Heterochromatin formation in the mouse embryo requires critical residues of the histone variant H3.3. Nat Cell Biol. 12(9):853–62 Santoro R, Grummt I. 2001. Molecular mechanisms mediating methylation-dependent silencing of ribosomal gene transcription. Molecular Cell. 8(3):719–25 Santoro R, Grummt I. 2005. Epigenetic Mechanism of rRNA Gene Silencing: Temporal Order of NoRC-Mediated Histone Modification, Chromatin Remodeling, and DNA Methylation. Molecular and Cellular Biology. 25(7):2539–46 Santoro R, Li J, Grummt I. 2002. The nucleolar remodeling complex NoRC mediates heterochromatin formation and silencing of ribosomal gene transcription. Nat. Genet. 32(3):393–96 Santoro R, Schmitz K-M, Sandoval J, Grummt I. 2009. Intergenic transcripts originating from a subclass of ribosomal DNA repeats silence ribosomal RNA genes in trans. EMBO Rep. 11(1):52–58 Santos F, Hendrich B, Reik W, Dean W. 2002. Dynamic Reprogramming of DNA Methylation in the Early Mouse Embryo. Developmental Biology. 241(1):172–82 Saporita AJ, Chang H-C, Winkeler CL, Apicelli AJ, Kladney RD, et al. 2011. RNA Helicase DDX5 Is a p53- Independent Target of ARF That Participates in Ribosome Biogenesis. Cancer Research. 71(21):6708–17 Satoh A, Brace CS, Rensing N, Cliften P, Wozniak DF, et al. 2013. Sirt1 Extends Life Span and Delays Aging in Mice through the Regulation of Nk2 Homeobox 1 in the DMH and LH. Cell Metabolism. 18(3):416–30 Savić N, Bär D, Leone S, Frommel SC, Weber FA, et al. 2014. lncRNA maturation to initiate heterochromatin formation in the nucleolus is required for exit from pluripotency in ESCs. Cell Stem Cell. 15(6):720–34 Savino TM, Gébrane-Younès J, De Mey J, Sibarita JB, Hernandez-Verdun D. 2001. Nucleolar assembly of the rRNA processing machinery in living cells. J Cell Biol. 153(5):1097–1110 Savkur RS, Olson MO. 1998. Preferential cleavage in pre-ribosomal RNA byprotein B23 endoribonuclease. Nucleic Acids Research. 26(19):4508–15 Schaffner W. 1980. Direct transfer of cloned genes from bacteria to mammalian cells. PNAS. 77(4):2163–67 Scherl A, Couté Y, Déon C, Callé A. 2002. Functional proteomic analysis of human nucleolus. Molecular biology of …. 13:4100–4109 Schlosser A, Bodem J, Bossemeyer D, Grummt I, Lehmann WD. 2002. Identification of protein phosphorylation sites by combination of elastase digestion, immobilized metal affinity chromatography, and quadrupole-time of flight tandem mass spectrometry. Proteomics. 2(7):911–18 Schlosser I, Hölzel M, Mürnseer M, Burtscher H, Weidle UH, Eick D. 2003. A role for c‐Myc in the regulation of ribosomal RNA processing. Nucleic Acids Research. 31(21):6148–56 Schmidt EE, Taylor DS, Prigge JR, Barnett S, Capecchi MR. 2000. Illegitimate Cre-dependent chromosome rearrangements in transgenic mouse spermatids. PNAS. 97(25):13702–7 Schmitz K-M, Schmitt N, Hoffmann-Rohrer U, Schäfer A, Grummt I, Mayer C. 2009. TAF12 Recruits Gadd45a and the Nucleotide Excision Repair Complex to the Promoter of rRNA Genes Leading to Active DNA Demethylation. Molecular Cell. 33(3):344–53 Schmitz KM, Mayer C, Postepska A, Grummt I. 2010. Interaction of noncoding RNA with the rDNA promoter mediates recruitment of DNMT3b and silencing of rRNA genes. Genes Dev. 24(20):2264–69 Schnapp A, Grummt I. 1991. Transcription complex formation at the mouse rDNA promoter involves the stepwise association of four transcription factors and RNA polymerase I. Journal of Biological Chemistry. 266(36):24588–95 Schnapp A, Schnapp G, Erny B, Grummt I. 1993. Function of the growth-regulated transcription initiation factor TIF- IA in initiation complex formation at the murine ribosomal gene promoter. Molecular and Cellular Biology. 13(11):6723–32

171

Schneider DA, Nomura M. 2004. RNA polymerase I remains intact without subunit exchange through multiple rounds of transcription in Saccharomyces cerevisiae. PNAS. 101(42):15112–17 Schultz RM. 1993. Regulation of zygotic gene activation in the mouse. BioEssays. 15(8):531–38 Schwenk F, Baron U, Rajewsky K. 1995. A cre-transgenic mouse strain for the ubiquitous deletion of loxP-flanked gene segments including deletion in germ cells. Nucleic Acids Research. 23(24):5080–81 Seither P, Coy JF, Pouska A, Grummt I. 1997. Molecular cloning and characterization of the cDNA encoding the largest subunit of mouse RNA polymerase I. Mol. Gen. Genet. 255(2):180–86 Seither P, Iben S, Grummt I. 1998. Mammalian RNA polymerase I exists as a holoenzyme with associated basal transcription factors. Journal of Molecular Biology. 275(1):43–53 Shishova KV, Khodarovich YM, Lavrentyeva EA, Zatsepina OV. 2015a. High-resolution microscopy of active ribosomal genes and key members of the rRNA processing machinery inside nucleolus-like bodies of fully- grown mouse oocytes. Experimental Cell Research. 337(2):208–18 Shishova KV, Lavrentyeva EA, Dobrucki JW, Zatsepina OV. 2015b. Nucleolus-like bodies of fully-grown mouse oocytes contain key nucleolar proteins but are impoverished for rRNA. Developmental Biology. 397(2):267– 81 Sigal A, Rotter V. 2000. Oncogenic Mutations of the p53 Tumor Suppressor: The Demons of the Guardian of the Genome. Cancer Research. 60(24):6788–93 Sirri V, Hernandez-Verdun D, Roussel P. 2002. Cyclin-dependent kinases govern formation and maintenance of the nucleolus. J Cell Biol. 156(6):969–81 Sirri V, Roussel P, Hernandez-Verdun D. 1999. The mitotically phosphorylated form of the transcription termination factor TTF-1 is associated with the repressed rDNA transcription machinery. Journal of Cell Science. 112 ( Pt 19):3259–68 Sirri V, Urcuqui-Inchima S, Roussel P, Hernandez-Verdun D. 2008. Nucleolus: the fascinating nuclear body. Histochemistry and Cell Biology. 129(1):13–31 Smith AJ, De Sousa MA, Kwabi-Addo B, Heppell-Parton A, Impey H, Rabbitts P. 1995. A site-directed chromosomal translocation induced in embryonic stem cells by Cre-loxP recombination. Nat. Genet. 9(4):376–85 Smith SD, O'Mahony DJ, Kinsella BT, Rothblum LI. 1993. Transcription from the rat 45S ribosomal DNA promoter does not require the factor UBF. Gene Expr. 3(3):229–36 Smith SD, Oriahi E, Lowe D, Yang-Yen HF, O'Mahony D, et al. 1990. Characterization of factors that direct transcription of rat ribosomal DNA. Molecular and Cellular Biology. 10(6):3105–16 Southern E. 2006. Southern blotting. Nat Protoc. 1(2):518–25 Stavreva DA, Kawasaki M, Dundr M, Koberna K, Müller WG, et al. 2006. Potential Roles for Ubiquitin and the Proteasome during Ribosome Biogenesis. Molecular and Cellular Biology. 26(13):5131–45 Stefanovsky V, Langlois F, Gagnon-Kugler T, Rothblum LI, Moss T. 2006a. Growth Factor Signaling Regulates Elongation of RNA Polymerase I Transcription in Mammals via UBF Phosphorylation and r-Chromatin Remodeling. Molecular Cell. 21(5):629–39 Stefanovsky V, Moss T. 2006. Regulation of rRNA synthesis in human and mouse cells is not determined by changes in active gene count. Cell Cycle. 5(7):735–39 Stefanovsky VY, Bazett-Jones DP. 1996. The DNA supercoiling architecture induced by the transcription factor xUBF requires three of its five HMG-boxes. Nucleic acids … Stefanovsky VY, Langlois F, Bazett-Jones D, Pelletier G, Moss T. 2006b. ERK Modulates DNA Bending and Enhancesome Structure by Phosphorylating HMG1-Boxes 1 and 2 of the RNA Polymerase I Transcription Factor UBF †. Biochemistry. 45(11):3626–34 Stefanovsky VY, Moss T. 2008. The splice variants of UBF differentially regulate RNA polymerase I transcription elongation in response to ERK phosphorylation. Nucleic Acids Research. 36(15):5093–5101 Stefanovsky VY, Moss T. 2016. Metabolic Labeling in the Study of Mammalian Ribosomal RNA Synthesis. In The Nucleolus, Vol. 1455, pp. 133–45. New York, NY: Springer New York Stefanovsky VY, Pelletier G, Bazett-Jones DP, Crane-Robinson C, Moss T. 2001a. DNA looping in the RNA polymerase I enhancesome is the result of non-cooperative in-phase bending by two UBF molecules. Nucleic Acids Research. 29(15):3241–47

172

Stefanovsky VY, Pelletier G, Hannan RD, Gagnon-Kugler T, Rothblum LI, Moss T. 2001b. An immediate response of ribosomal transcription to growth factor stimulation in mammals is mediated by ERK phosphorylation of UBF. Molecular Cell. 8(5):1063–73 Stepanchick A, Zhi H, Cavanaugh AH, Rothblum K, Schneider DA, Rothblum LI. 2013. DNA Binding by the Ribosomal DNA Transcription Factor Rrn3 Is Essential for Ribosomal DNA Transcription. Journal of Biological Chemistry. 288(13):9135–44 Stępiński D. 2016. Nucleolus-derived mediators in oncogenic stress response and activation of p53-dependent pathways. Histochemistry and Cell Biology. 1–21 Stimpson KM, Sullivan LL, Kuo ME, Sullivan BA. 2014. Nucleolar Organization, Ribosomal DNA Array Stability, and Acrocentric Chromosome Integrity Are Linked to Telomere Function. PLoS ONE. 9(3):e92432–13 Strausberg RL, Feingold EA, Grouse LH, Derge JG, Klausner RD, et al. 2002. Generation and initial analysis of more than 15,000 full-length human and mouse cDNA sequences. PNAS. 99(26):16899–903 Strohner R, Németh A, Jansa P, Rohrer UH, Santoro R, et al. 2001. NoRC—a novel member of mammalian ISWI‐ containing chromatin remodeling machines. The EMBO Journal. 20(17):4892–4900 Stults DM, Killen MW, Pierce HH, Pierce AJ. 2008. Genomic architecture and inheritance of human ribosomal RNA gene clusters. Genome Res. 18(1):13–18 Stults DM, Killen MW, Williamson EP, Hourigan JS, Vargas HD, et al. 2009. Human rRNA Gene Clusters Are Recombinational Hotspots in Cancer. Cancer Research. 69(23):9096–9104 Su N, Jin M, Chen L. 2014. Role of FGF/FGFR signaling in skeletal development and homeostasis: learning from mouse models. Bone Res. 2:14003 Sulli G, Di Micco R, d'Adda di Fagagna F. 2012. Crosstalk between chromatin state and DNA damage response in cellular senescence and cancer. Nat Rev Cancer. 12(10):709–20 Tachibana M, Sugimoto K, Nozaki M, Ueda J, Ohta T, et al. 2002. G9a histone methyltransferase plays a dominant role in euchromatic histone H3 lysine 9 methylation and is essential for early embryogenesis. Genes Dev. 16(14):1779–91 Tada S, Era T, Furusawa C, Sakurai H, Nishikawa S, et al. 2005. Characterization of mesendoderm: a diverging point of the definitive endoderm and mesoderm in embryonic stem cell differentiation culture. Development. 132(19):4363–74 Tadros W, Lipshitz HD. 2009. The maternal-to-zygotic transition: a play in two acts. Development. 136(18):3033–42 Takada H, Kurisaki A. 2015. Emerging roles of nucleolar and ribosomal proteins in cancer, development, and aging. Cellular and Molecular Life Sciences. 72(21):4015–25 Takahashi K, Yamanaka S. 2006. Induction of Pluripotent Stem Cells from Mouse Embryonic and Adult Fibroblast Cultures by Defined Factors. Cell. 126(4):663–76 Takemura M, Ohoka F, Perpelescu M, Ogawa M, Matsushita H, et al. 2002. Phosphorylation-Dependent Migration of Retinoblastoma Protein into the Nucleolus Triggered by Binding to Nucleophosmin/B23. Experimental Cell Research. 276(2):233–41 Tchurikov NA, Yudkin DV, Gorbacheva MA, Kulemzina AI, Grischenko IV, et al. 2016. Hot spots of DNA double- strand breaks in human rDNA units are produced in vivo. Sci Rep. 6:25866 Tesarik J, Greco E. 1999. The probability of abnormal preimplantation development can be predicted by a single static observation on pronuclear stage morphology. Human Reproduction. 14(5):1318–23 Thiry M. 1988. Study of RNA distribution in the nucleolar components of Ehrlich cell using RNase-gold method. Histochemistry. 89(3):231–36 Tongaonkar P, French SL, Oakes ML, Vu L, Schneider DA, et al. 2005. Histones are required for transcription of yeast rRNA genes by RNA polymerase I. PNAS. 102(29):10129–34 Torres J, Prieto J, Durupt FC, Broad S, Watt FM. 2012. Efficient differentiation of embryonic stem cells into mesodermal precursors by BMP, retinoic acid and Notch signalling. PLoS ONE. 7(4):e36405 Torres-Padilla M-E, Bannister AJ, Hurd PJ, Kouzarides T, Zernicka-Goetz M. 2006. Dynamic distribution of the replacement histone variant H3.3 in the mouse oocyte and preimplantation embryos. Int. J. Dev. Biol. 50(5):455–61 Treiber DK, Zhai X, Jantzen HM, Essigmann JM. 1994. Cisplatin-DNA adducts are molecular decoys for the ribosomal RNA transcription factor hUBF (human upstream binding factor). PNAS. 91(12):5672–76

173

Treré D, Ceccarelli C, Montanaro L, Tosti E, Derenzini M. 2004. Nucleolar Size and Activity Are Related to pRb and p53 Status in Human Breast Cancer. J. Histochem. Cytochem. 52(12):1601–7 Tschochner H, Hurt E. 2003. Pre-ribosomes on the road from the nucleolus to the cytoplasm. Trends in Cell Biology. 13(5):255–63 Tuan JC, Zhai W, Comai L. 1999. Recruitment of TATA-binding protein-TAFI complex SL1 to the human ribosomal DNA promoter is mediated by the carboxy-terminal activation domain of upstream binding factor (UBF) and is regulated by UBF phosphorylation. Molecular and Cellular Biology. 19(4):2872–79 Turlo KA, Gallaher SD, Vora R, Laski FA, Iruela-Arispe ML. 2010. When Cre-Mediated Recombination in Mice Does Not Result in Protein Loss. Genetics. 186(3):959–67 Ura K, Kurumizaka H, Dimitrov S, Almouzni G, Wolffe AP. 1997. Histone acetylation: influence on transcription, nucleosome mobility and positioning, and linker histone-dependent transcriptional repression. The EMBO Journal. 16(8):2096–2107 Valdés-Mora F, Song JZ, Statham AL, Strbenac D, Robinson MD, et al. 2012. Acetylation of H2A.Z is a key epigenetic modification associated with gene deregulation and epigenetic remodeling in cancer. Genome Res. 22(2):307–21 van de Nobelen S, Rosa-Garrido M, Leers J, Heath H, Soochit W, et al. 2010. CTCF regulates the local epigenetic state of ribosomal DNA repeats. Epigenetics Chromatin. 3(1):19 van der Heijden GW, Dieker JW, Derijck AAHA, Muller S, Berden JHM, et al. 2005. Asymmetry in Histone H3 variants and lysine methylation between paternal and maternal chromatin of the early mouse zygote. Mech. Dev. 122(9):1008–22 van Sluis M, McStay B. 2015. A localized nucleolar DNA damage response facilitates recruitment of the homology- directed repair machinery independent of cell cycle stage. Genes Dev. 29(11):1151–63 Voit R, Grummt I. 2001. Phosphorylation of UBF at serine 388 is required for interaction with RNA polymerase I and activation of rDNA transcription. Proc. Natl. Acad. Sci. U.S.A. 98(24):13631–36 Voit R, Hoffmann M, Grummt I. 1999. Phosphorylation by G1-specific cdk-cyclin complexes activates the nucleolar transcription factor UBF. The EMBO Journal. 18(7):1891–99 Voit R, Schäfer K, Grummt I. 1997. Mechanism of repression of RNA polymerase I transcription by the retinoblastoma protein. Molecular and Cellular Biology. 17(8):4230–37 Voit R, Seiler J, Grummt I. 2015. Cooperative Action of Cdk1/cyclin B and SIRT1 Is Required for Mitotic Repression of rRNA Synthesis. PLoS Genet. 11(5):e1005246–17 Volarevic S, Stewart MJ, Ledermann B, Zilberman F, Terracciano L, et al. 2000. Proliferation, But Not Growth, Blocked by Conditional Deletion of 40S Ribosomal Protein S6. Science. 288(5473):2045–47 Vooijs M, Jonkers J, Berns A. 2001. A highly efficient ligand-regulated Cre recombinase mouse line shows that LoxP recombination is position dependent. EMBO Rep. 2(4):292–97 Wang GG, Allis CD, Chi P. 2007. Chromatin remodeling and cancer, part I: covalent histone modifications. Trends in Molecular Medicine. 13(9):363–72 Wang S-Z, Liu B-H, Tao HW, Xia K, Zhang LI. 2009. A Genetic Strategy for Stochastic Gene Activation with Regulated Sparseness (STARS). PLoS ONE. 4(1):e4200 Warmerdam DO, van den Berg J, Medema RH. 2016. Breaks in the 45S rDNA Lead to Recombination-Mediated Loss of Repeats. Cell Reports. 14(11):2519–27 Watanabe-Susaki K, Takada H, Enomoto K, Miwata K, Ishimine H, et al. 2014. Biosynthesis of Ribosomal RNA in Nucleoli Regulates Pluripotency and Differentiation Ability of Pluripotent Stem Cells. 32(12):3099–3111 Weber JD, Kuo M-L, Bothner B, DiGiammarino EL, Kriwacki RW, et al. 2000. Cooperative Signals Governing ARF- Mdm2 Interaction and Nucleolar Localization of the Complex. Molecular and Cellular Biology. 20(7):2517–28 Weisend CM, Kundert JA, Suvorova ES, Prigge JR, Schmidt EE. 2009. Cre activity in fetal albCre mouse hepatocytes: Utility for developmental studies. Genesis. 47(12):789–92 Weitao T, Budd M, Hoopes LLM, Campbell JL. 2003. Dna2 Helicase/Nuclease Causes Replicative Fork Stalling and Double-strand Breaks in the Ribosomal DNA of Saccharomyces cerevisiae. Journal of Biological Chemistry. 278(25):22513–22 Welborn J. 2004. Acquired Robertsonian translocations are not rare events in acute leukemia and lymphoma. Cancer Genetics and Cytogenetics. 151(1):14–35

174

Wen B, Wu H, Shinkai Y, Irizarry RA, Feinberg AP. 2009. Large histone H3 lysine 9 dimethylated chromatin blocks distinguish differentiated from embryonic stem cells. Nat. Genet. 41(2):246–50 Werner M, Thuriaux P, Soutourina J. 2009. Structure–function analysis of RNA polymerases I and III. Current Opinion in Structural Biology. 19(6):740–45 White RJ. 2008. RNA polymerases I and III, non-coding RNAs and cancer. Trends in Genetics. 24(12):622–29 Williams SM, Robbins LG. 1992. Molecular genetic analysis of Drosophila rDNA arrays. Trends in Genetics. 8(10):335–40 Wills NM, Atkins JF. 2006. The potential role of ribosomal frameshifting in generating aberrant proteins implicated in neurodegenerative diseases. RNA. 12(7):1149–53 Woolnough JL, Atwood BL, Liu Z, Zhao R, Giles KE. 2016. The Regulation of rRNA Gene Transcription during Directed Differentiation of Human Embryonic Stem Cells. PLoS ONE. 11(6):e0157276 Wright JE, Mais C, Prieto J-L, McStay B. 2006. A role for upstream binding factor in organizing ribosomal gene chromatin. Biochemical Society Symposia. 73:77–84 Wsierska-Gadek J, Horky M. 2003. How the nucleolar sequestration of p53 protein or its interplayers contributes to its (re)-activation. Ann. N. Y. Acad. Sci. 1010:266–72 Xiaoquan Zhai, Holger Beckmann, Hans-Michael Jantzen A, John M Essigmann. 1998. Cisplatin−DNA Adducts Inhibit Ribosomal RNA Synthesis by Hijacking the Transcription Factor Human Upstream Binding Factor†. American Chemical Society Xie W, Ling Te, Zhou Y, Feng W, Zhu Q, et al. 2012. The chromatin remodeling complex NuRD establishes the poised state of rRNA genes characterized by bivalent histone modifications and altered nucleosome positions. PNAS. 109(21):8161–66 Xu J, Wang H, Liang T, Cai X, Rao X, et al. 2012. Retinoic acid promotes neural conversion of mouse embryonic stem cells in adherent monoculture. Mol. Biol. Rep. 39(2):789–95 Yamamoto K, Yamamoto M, Hanada K-I, Nogi Y, Matsuyama T, Muramatsu M. 2004. Multiple Protein-Protein Interactions by RNA Polymerase I-Associated Factor PAF49 and Role of PAF49 in rRNA Transcription. Molecular and Cellular Biology. 24(14):6338–49 Yamamoto RT, Nogi Y, Dodd JA, Nomura M. 1996. RRN3 gene of Saccharomyces cerevisiae encodes an essential RNA polymerase I transcription factor which interacts with the polymerase independently of DNA template. The EMBO Journal. 15(15):3964–73 Yelick PC, Trainor PA. 2015. Ribosomopathies: Global process, tissue specific defects. Rare Diseases. 3(1):e1025185–12 Yuan X, Feng W, Imhof A, Grummt I, Zhou Y. 2007. Activation of RNA Polymerase I Transcription by Cockayne Syndrome Group B Protein and Histone Methyltransferase G9a. Molecular Cell. 27(4):585–95 Yuan X, Zhao J, Zentgraf H, Hoffmann-Rohrer U, Grummt I. 2002. Multiple interactions between RNA polymerase I, TIF-IA and TAF(I) subunits regulate preinitiation complex assembly at the ribosomal gene promoter. EMBO Rep. 3(11):1082–87 Yuan X, Zhou Y, Casanova E, Chai M, Kiss E, et al. 2005. Genetic Inactivation of the Transcription Factor TIF-IA Leads to Nucleolar Disruption, Cell Cycle Arrest, and p53-Mediated Apoptosis. Molecular Cell. 19(1):77–87 Zatsepina O, Baly C, Chebrout M, Debey P. 2003. The Step-Wise Assembly of a Functional Nucleolus in Preimplantation Mouse Embryos Involves the Cajal (Coiled) Body. Developmental Biology. 253(1):66–83 Zentner GE, Balow SA, Scacheri PC. 2014. Genomic Characterization of the Mouse Ribosomal DNA Locus. G3 (Bethesda). 4(2):243–54 Zentner GE, Saiakhova A, Manaenkov P, Adams MD, Scacheri PC. 2011. Integrative genomic analysis of human ribosomal DNA. Nucleic Acids Research. 39(12):4949–60 Zhai W, Comai L. 2000. Repression of RNA Polymerase I Transcription by the Tumor Suppressor p53. Molecular and Cellular Biology. 20(16):5930–38 Zhang H, Wang JC, Liu LF. 1988. Involvement of DNA topoisomerase I in transcription of human ribosomal RNA genes. PNAS. 85(4):1060–64 Zhang L-F, Huynh KD, Lee JT. 2007. Perinucleolar Targeting of the Inactive X during S Phase: Evidence for a Role in the Maintenance of Silencing. Cell. 129(4):693–706

175

Zhang Q, Shalaby NA, Buszczak M. 2014. Changes in rRNA Transcription Influence Proliferation and Cell Fate Within a Stem Cell Lineage. Science. 343(6168):298–301 Zhang Y, Lu H. 2009. Signaling to p53: Ribosomal Proteins Find Their Way. Cancer Cell. 16(5):369–77 Zhang Y, Wolf GW, Bhat K, Jin A, Allio T, et al. 2003. Ribosomal Protein L11 Negatively Regulates Oncoprotein MDM2 and Mediates a p53-Dependent Ribosomal-Stress Checkpoint Pathway. Molecular and Cellular Biology. 23(23):8902–12 Zhang Y, Xiong Y. 2001. Control of p53 ubiquitination and nuclear export by MDM2 and ARF. Cell Growth Differ. 12(4):175–86 Zhao J, Yuan X, Frödin M, Grummt I. 2003. ERK-dependent phosphorylation of the transcription initiation factor TIF-IA is required for RNA polymerase I transcription and cell growth. Molecular Cell. 11(2):405–13 Zhao Q, Wu Y, Shan Z, Bai G, Wang Z, et al. 2016. Serum starvation-induced cell cycle synchronization stimulated mouse rDNA transcription reactivation during somatic cell reprogramming into iPSCs. Stem Cell Research & Therapy. 7(1):1–10 Zheng B, Sage M, Sheppeard EA, Jurecic V, Bradley A. 2000. Engineering mouse chromosomes with Cre-loxP: range, efficiency, and somatic applications. Molecular and Cellular Biology. 20(2):648–55 Zhou L-Q, Dean J. 2015. Reprogramming the genome to totipotency in mouse embryos. Trends in Cell Biology. 25(2):82–91 Zhou Y, Santoro R, Grummt I. 2002. The chromatin remodeling complex NoRC targets HDAC1 to the ribosomal gene promoter and represses RNA polymerase I transcription. The EMBO Journal. 21(17):4632–40 Zomerdijk JCBM, Beckmann H, Comai L, Tjian R. 1994. Assembly of transcriptionally active RNA polymerase I initiation factor SL1 from recombinant subunits. Science. 266(5193):2015–18 Zuccotti M, Garagna S, Merico V, Monti M, Alberto Redi C. 2005. Chromatin organisation and nuclear architecture in growing mouse oocytes. Molecular and Cellular Endocrinology. 234(1-2):11–17

176

Annexes

1. Hamdane N, Herdman C, Mars JC, Stefanovsky V, Tremblay MG, and Moss T. (2015) Depletion of the cisplatin targeted HMGB-box factor UBF selectively induces p53-independent apoptotic death in transformed cells. Oncotarget 6(29); 27519-36.

2. Jean S, Tremblay MG, Herdman C, Guillou F and Moss T. (2012) The endocytic adapter E-Syt2 recruits the p21 GTPase activated kinase PAK1 to mediate actin dynamics and FGF signalling. Biology Open 1; 731-738.

3. Herdman C, Tremblay MG, Mishra PK, and Moss T. (2014) Loss of Extended Synaptotagmins ESyt2 and ESyt3 does not affect mouse development or viability, but in vitro cell migration and survival under stress are affected. Cell Cycle 13(16); 2616-25.

4. Tremblay MG, Herdman C, Guillou F, Mishra PK, Baril J, Bellenfant S and Moss T. (2015) Extended Synaptotagmin Interaction with the Fibroblast Growth Factor Receptor Depends on Receptor Conformation, Not Catalytic Activity. Journal of Biological Chemistry 290(26); 16142-56.

5. Herdman C, and Moss T. (2016) Extended-Synaptotagmins (E-Syts); the extended story. Pharmalogical Research 107; 48-56.

177