Sile Fiona Yang

Functional Characterisation and Therapeutic Potential of the Zinc Finger Protein ZNF827

Sile Fiona Yang

A thesis submitted to the University of Sydney in fulfillment of the requirements for the degree of Doctor of Philosophy

Children’s Medical Research Institute

Sydney Medical School Faculty of Medicine and Health The University of Sydney

February 2019

Statement of Originality

This is to certify that to the best of my knowledge, the content of this thesis is my own work. This thesis has not been submitted for any degree or other purposes.

I certify that the intellectual content of this thesis is the product of my own work and that all the assistance received in preparing this thesis and sources have been acknowledged.

Sile Fiona Yang

Acknowledgements

This thesis would not have been possible without the support of these amazing people.

First and foremost, I would like to thank my supervisors Associate Professor Hilda A. Pickett and Dr Alexander P. Sobinoff for their ongoing invaluable guidance and advice throughout my PhD. I could not have asked for better mentors. Thank you both for always making time for me to answer my questions, and for your encouragement when I am unsure. Hilda, your passion and dedication to science always inspires me and you are my role model in science. Alex, thank you for being so down-to-earth that you are like a very knowledgeable peer that I can always go to for advice on my experiments and career.

To the past and current members of TLRU I have met during my PhD – Monica, Michael, Chris T, Josh, Chris N, Vivian and Robert, thank you all for the helpful discussions and troubleshooting for my project, and the memorable Christmas parties. Special thanks to Monica, for patiently teaching me lab techniques, and helping me settle in the lab when I first started. Michael, thank you for always being willing to help with a laid-back attitude, and all the tech support. To Jane, Eugene and Dale whom we shared the same lab space; Jane, thank you for being an excellent lab manager to ensure the lab functions orderly and we have all we need for experiments, and for our fun conversations. Eugene, thank you for being the other pharmacist in the lab who I can have pharmacists’ chats with. Dale, thank you for the Friday afternoon chats in tissue culture and your expertise with Incucyte.

I would like to thank all the members of the other telomere labs and all the other great people I have met at CMRI. Special thanks to Joyce, for your mental support and our destressing karaoke sessions in the lab. Melissa and Omesha, thank you for introducing me to delicious Indonesian and Sri Lankan food. Sam and Natsuki, thank you for our morning tea chats and the emotional support. Mingqi and Irene, thank you for the company when I had to work late and for the conversations and dinners we shared.

During the last 4 years, not only I have developed the essential skills as a scientist, I have also gained precious friendships that I am very grateful for.

I would also like to thank all my friends outside of CMRI, especially Eva, Hannah, Hanna, Gloria, Jenny and Rosemary, for being very supportive friends all these years.

To my family, I would like to thank my parents for the resilience and determination I have learned. Mum, thank you for your tenacity and doggedness that brought us to Australia which is a big turning point in my life. Dad, thank you for always being strong and caring, and trusting in everything I do. Fong, thank you for your philosophical insights and always being supportive to my decisions. My dear grandma, without you I would not have been the person I am today. You have always been dearly missed.

To the very important person in my life, Eddie, thank you for always being there for me, for keeping me sane when I am stressed, and for making me happy when I am down.

This research project was supported by the Australian Government Research Training Program scholarship provide by the University of Sydney and the CMRI scholarship. I would like to also acknowledge the CMRI facilities including ATAC, the Bioinformatic unit and VGEF for their support to my project.

iii

Publications and Abstracts

Publications

Yang, S. F., Sun, A. A., Shi, Y., Li, F., and Pickett, H. A. (2018) Structural and functional characterization of the RBBP4-ZNF827 interaction and its role in NuRD recruitment to telomeres. Biochemical Journal. 475, 2667-2679

Harvey R.P., Asili N., Xaymardan M., Forte M., Waardenberg A.J., Cornwell J., Janbandhu V., Kesteven S., Chandrakanthan V., Malinowska H., Reinhard H., Yang S.F., Pickett H. A., … , Feneley M. (2018) PDGFRα signaling in cardiac stem and stromal cells modulates quiescence, metabolism and self-renewal, and promotes anatomical and functional repair. BioRxiv. doi: https://doi.org/10.1101/225979

Sobinoff A.P., Allen J.A., Neumann A.A., Yang S.F., Walsh M.E., Henson J.D., Reddel R.R., Pickett H.A. (2017) BLM and SLX4 play opposing roles in recombination- dependent replication at human telomeres. EMBO J., 36(19):2907-2919.

Lee, M., Napier, C. E., Yang, S. F., Arthur, J. W., Reddel, R. R., Pickett, H. A. (2016) Comparative analysis of whole genome sequencing-based telomere length measurement techniques. Methods., 114:4-15.

Oral presentations

Yang, S. F., Sobinoff, A. P., Pickett, H. A. Functional characterisation and therapeutic potential of the zinc finger protein ZNF827. University of Sydney Postgraduate and Early Career Researchers Cancer Research Symposium, the University of Sydney, November 2018, Sydney NSW, Australia.

Poster presentations

Yang, S. F., Sobinoff, A. P., Pickett, H. A. Functional characterisation and therapeutic potential of the zinc finger protein ZNF827. KCA Early Career Showcase, Kids Cancer Alliance, November 2018, Sydney NSW, Australia.

Yang, S. F., Sobinoff, A. P., Pickett, H. A. ZNF827 - a molecular target for telomere maintenance in cancer. Lorne Genome Conference, February 2018, Lorne VIC, Australia.

Yang, S. F., Conomos, D., Sobinoff, A. P., Pickett, H. A. ZNF827 - a molecular target for telomere maintenance in cancer. KCA Early Career Showcase, Kids Cancer Alliance, December 2017, Sydney NSW, Australia.

Yang, S. F., Conomos, D., Sobinoff, A. P., Pickett, H. A. ZNF827 - a molecular target for telomere maintenance in cancer. Australian Society for Medical Research, June 2017, Sydney NSW, Australia.

Yang, S. F., Conomos, D., Sobinoff, A. P., Pickett, H. A., ZNF827 - a molecular target for telomere maintenance in cancer. Cell cycle, DNA Damage Response and Telomeres, Australian Cell Cycle Meeting, March 2017, Sydney NSW, Australia.

Yang, S. F., Conomos, D., Sobinoff, A. P., Pickett, H. A., ZNF827 – a potential molecular target for anticancer therapies targeting ALT telomeres. Sydney Cancer Conference 2016, Sydney NSW, Australia.

Yang, S. F., Conomos, D., Sobinoff, A. P., Pickett, H. A., ZNF827 – a potential molecular target for anticancer therapies targeting ALT telomeres. Westmead Association Hospital Week Research Symposium 2016, Westmead NSW, Australia

Yang, S. F., Conomos, D., Sobinoff, A. P., Pickett, H. A., Characterisation of ZNF827 as a molecular target for telomere maintenance in cancer. Australian Society for Medical Research 2016, Sydney NSW, Australia.

Yang, S. F., Conomos, D., Sobinoff, A. P., Pickett, H. A., Characterisation of ZNF827 as a molecular target for telomere maintenance in cancer. NSW and ACT Cell & Developmental Biology Meeting 2016, Sydney NSW, Australia

Yang, S. F., Conomos, D., Sobinoff, A. P., Pickett, H. A., Characterisation of ZNF827 as a molecular target for telomere maintenance in cancer. Lorne Genome Conference 2016, Lorne VIC, Australia.

Thesis Abstract

The ends of linear eukaryotic chromosomes resemble double strand breaks, susceptible to the vigilant cellular DNA damage response machinery. Telomeres are specialised nucleoprotein structures found at the ends of human chromosomes that function as a protective cap safeguarding the integrity of the genome. In normal human somatic cells, telomeres shorten with each replicative cell division. Eventually, shortened telomeres trigger cellular senescence and apoptosis. Cancer cells overcome this proliferative barrier by acquiring a telomere maintenance mechanism. Alternative Lengthening of Telomeres (ALT) is a homologous recombination (HR) mediated telomere maintenance mechanism utilised by approximately 10-15% of all cancers. ALT status in certain cancer subtypes, such as osteosarcomas and astrocytomas, often predicts aggressive nature and poor prognosis. Therefore, the ALT pathway presents an attractive avenue for developing novel cancer therapeutics. To date, ALT specific therapeutic targets are lacking.

Our lab has previously identified the zinc finger protein ZNF827 as a promising molecular target in the ALT pathway. ZNF827 recruits the nucleosome remodelling and histone deacetylase (NuRD) complex to ALT telomeres, and collaboratively facilitates multifaceted functions to promote HR-mediated telomere synthesis. However, the precise molecular mechanisms underlying the important roles of ZNF827 at ALT telomeres have not been fully elucidated. Prior to the study by our lab, ZNF827 was a protein of unknown function.

This thesis has focused on the functional characterisation of ZNF827, with an ultimate goal of validating it as a therapeutic target in cancers utilising the ALT pathway. We have characterised ZNF827 from several perspectives. First, we have explored the biological functions of ZNF827 as a transcription factor and discovered novel roles of ZNF827 in cellular processes including embryonic development and immune response. Second, we have gained new insights into the recruitment of ZNF827 to ALT telomeres, characterised structurally and functionally the interaction interface between ZNF827 and the NuRD complex, and implicated ZNF827 SUMOylation in promoting ALT activity. We have also discovered that ZNF827 is novel ssDNA binding protein implicated in HR-directed repair of replication-associated DNA damage at telomeres as well as genome-wide, and that its role as a DNA damage response and

vi repair protein is likely to be mediated through the ATR-mediated DNA signalling pathway. Finally, we have provided preliminary evidence that supports synthetic lethality between ZNF827 inhibition and the topoisomerase I inhibitor topotecan, implicating ZNF827 as a potential molecular target for ATR inhibition. Taken together, our findings have substantially expanded our knowledge on the zinc finger protein ZNF827 and provided further support for its therapeutic potential in the development of novel cancer therapies.

vii

Authorship Attribution Statement

This thesis contains material published in the manuscript ‘Structural and functional characterization of the RBBP4-ZNF827 interaction and its role in NuRD recruitment to telomeres’ in the Biochemical Journal. This is section 4.3 to 4.7 in Chapter 4; figures 4.3 to 4.7. I contributed to the design of the study with the co-authors, interpretation of the analysis and writing of the drafts of the manuscript. I performed the molecular and functional studies in section 4.6 to 4.7.

In addition to the statements above, in cases where I am not the corresponding author of a published item, permission to include the published material has been granted by the corresponding author.

Sile Fiona Yang

February 2019

As supervisor for the candidature upon which this thesis is based, I can confirm that the authorship attribution statements above are correct.

Associate Professor Hilda A. Pickett

February 2019

viii

Abbreviations

53BP1 TP53-binding protein 1 aa amino acid ABDIL antibody dilution buffer ACKR2 atypical chemokine receptor 2 AGRF Australian Genome Research Facility ALT alternative lengthening of telomeres Alt-NHEJ alternative non-homologous end joining APB ALT-associate PML Bodies ASF1 anti-silencing factor 1 ATM ataxia telangiectasia mutated ATR ataxia telangiectasia and Rad-3 related ATRIP ATR interacting protein atRA all trans retinoic acid ATRX α-thalassaemia/mental retardation syndrome X linked BCA assay bicinchoninic acid assay BCLL11A B-cell lymphoma/leukemia 11A BCLL11B B-cell lymphoma/leukemia 11B BioID proximity-dependent biotin identification BIR break induced replication BITS break-induced telomere synthesis BLM Bloom syndrome protein BrdU 5‐bromo‐2'‐deoxyuridine BrdC 5‐bromo‐2'‐deoxycytidine BRIT1 BRCT-repeat inhibitor of hTERT expression BRCA1 breast cancer type 1 susceptibility protein BSA bovine serum albumin BTK Bruton tyrosine kinase C2H2 cysteine2histidine2 C-BERST dCas9–APEX2 biotinylation at genomic elements by restricted spatial tagging

CHD3 chromodomain helicase DNA‐binding protein 3 CHD4 chromodomain helicase DNA‐binding protein 3 ChIP chromatin immunoprecipitation ChIP-seq ChIP sequencing CHK1 serine/threonine-protein kinase 1 CHK2 serine/threonine-protein kinase 2 CMRI Children’s Medical Research Institute CO-FISH chromosome orientation fluorescence in situ hybridisation Co-IP co-immunoprecipitation COUP-TF II COUP transcription factor 2 CRISPR clustered regularly interspaced short palindromic repeats CST complex CTC1-STN1-TEN1 complex CTIP DNA endonuclease RBBP8 DAPI 4',6‐diamidino‐2‐phenylindole dihydrochloride DAXX death-domain-associated protein DDR DNA damage response DECIPHER DatabasE of genomiC varIation and Phenotype in Humans using Ensembl Resources DMEM Dulbecco’s Modified Eagle Medium DMSO dimethyl sulfoxide DNA-PKcs DNA-dependent protein kinase catalytic subunit ds double-stranded DSB double strand breaks ECTR extrachromosomal telomeric repeats EDTA ethylenediamine tetraacetic acid EGF epidermal growth factor EME1 essential meiotic endonuclease 1 EMSA electrophoretic mobility shift assay EMT epithelial to mesenchymal transition ERK mitogen-activated protein kinase ESC embryonic stem cells EXO1 exonuclease 1 FANCM Fanconi anaemia group M

FANCA Fanconi anaemia group A FANCD2 Fanconi anaemia group D2 FCS foetal calf serum FEN1 flap endonuclease 1 FISH fluorescence in situ hybridisation FOG1 friend of GATA protein 1 FUCCI fluorescence ubiquitination cell cycle indicator GATAD2A transcriptional repressor p66-alpha GATAD2B transcriptional repressor p66-beta GTEx Genotype-Tissue Expression GWAS genome-wide association studies H2AX histone H2AX HDAC1 histone deacetylase 1 HDAC2 histone deacetylase 2 HIC1 hypermethylated in cancer 1 hnRNPA1 heterogeneous nuclear ribonucleoprotein A1 HOP2 homologous-pairing protein 2 homolog HR homologous recombination hTERT telomerase reverse transcriptase catalytic subunit hTR telomerase RNA template IF immunofluorescence IFN- α interferon α IL10RA interleukin-10 receptor subunit alpha IL1RN interleukin-1 receptor antagonist protein IRF7 interferon regulatory factor 7 IPA ingenuity pathway analysis ITC isothermal titration calorimetry LBD ligand-binding domain MAPK1 mitogen-activated protein kinase 1 MBD2 methyl-CpG-binding domain protein 2 MBD3 methyl-CpG-binding domain protein 3 MEFs mouse embryonic fibroblasts

Meta-TIF metaphase- telomere dysfunction-induced foci MND1 meiotic nuclear division protein 1 homolog MRN MRE11-RAD50-NBS1 MS mass spectrometry MTA1 metastasis-associated protein 1 MTA2 metastasis-associated protein 2 MTA3 metastasis-associated protein 3 MUS81 crossover junction endonuclease MUS81 NFkB nuclear factor NF-kappa-B p100 subunit NHEJ non-homologous end joining NuRD nucleosome remodelling and histone deacetylase complex NKX2-3 homeobox protein NK-2 homolog 3 PCNA proliferating cell nuclear antigen PHF6 plant homeodomain finger 6 PICh proteomics of isolated chromatin segments PLA proximity ligation assay PML-NB promyelocytic leukaemia nuclear bodies PMSF phenylmethylsulphonyl fluoride POT1 protection of telomeres 1 PTGER2 prostaglandin E2 receptor EP2 subtype PTM posttranslational modification qRT-PCR quantitative reverse transcription PCR RAD51 DNA repair protein RAD51 homolog 1 RAP1 repressor activator protein 1 RAR/RXR retinoic acid receptors and retinoid X receptors RBBP4 retinoblastoma-binding protein 4 RBBP7 retinoblastoma-binding protein 7 RFC replication factor C RNA-seq RNA sequencing RPA replication protein A RTEL1 regulator of telomere elongation helicase 1 SALL1 sal-like protein 1

xii

SALL4 sal-like protein 4 SCE sister chromatid exchange SIRT1 NAD-dependent protein deacetylase sirtuin-1 SLX1 structure-specific endonuclease subunit 1 SLX4 structure-specific endonuclease subunit 4 SMARCAL1 SWI/SNF-related matrix-associated actin-dependent regulator of chromatin subfamily A-like protein 1 SMC5-6 structural maintenance of chromosome 5-6 SNAI1 protein snail homolog 1 SNP single nucleotide polymorphism ss single-stranded ssDNA single-stranded DNA STAT1 signal transducer and activator of transcription 1-alpha/beta STING stimulator of interferon genes protein SUMO small ubiquitin-like modifiers TERRA telomeric repeat-containing RNA TIN2 TRF1-interacting nuclear factor 2 TIF telomere dysfunction-induced foci TNF tumour necrosis factor TOPBP1 DNA topoisomerase 2-binding protein 1 TPP1 TINT1-PTOP-PIP1 TR4 testicular receptor 4 TRF1 telomeric repeat-binding factor 1 TRF2 telomeric repeat binding factor 2 TRIM24 tripartite motif-containing protein 24 TLS translesion DNA synthesis TMM telomere maintenance mechanism T-SCEs telomere-sister chromatid exchanges TZAP telomere zinc finger-associated protein VGEF Vector and genome editing facility WRN Werner syndrome ATP-dependent helicase ZMYND8 zinc finger MYND domain-containing protein 8

xiii

List of Tables

Table 1. Biological functions of RRK-motif containing NuRD binding zinc finger transcription factors ...... 47 Table 2.1 List of mammalian cell lines used in this study ...... 52 Table 2.2 List of plasmid constructs used in the overexpression experiments ...... 57 Table 2.3 siRNAs used in this study ...... 58 Table 2.4 List of qRT-PCR primers ...... 61 Table 2.5 A list of proteins purified by HaloTag® ...... 65 Table 2.6 EMSA probes used in this study ...... 67 Table 2.7 RNA-seq experimental design ...... 71 Table 2.8 ChIP-seq experimental design ...... 72 Table 2.9 List of primary antibodies used in this study ...... 82 Table 2.10 List of secondary antibodies used in this study ...... 83 Table 3.1 Mammalian ZNF827 orthologs with over 90% sequence similarities to human ZNF827 ...... 88 Table 3.2 Clinical phenotypes associated with ZNF827 gain and loss from the DECIPHER database ...... 91 Table 3.3 ZNF827 SNPs and associated traits in GWAS studies ...... 91 Table 4.1 Data collection and refinement statistics of ZNF8271-14 bound to RBBP4 ...... 123

Table 4.2 Binding affinity between RBBP4 and ZNF8271-14 peptide measured by ITC ...... 127

xiv

List of Figures

Figure 1.1 Composition and function of the mammalian telomere...... 2 Figure 1.2 The mechanistic details of telomere shortening...... 5 Figure 1.3 Graphical representation of telomere length dynamics in somatic cells over cell divisions from physiological state to cancer state...... 9 Figure 1.4 Graphical representation of phenotypic markers of ALT...... 20 Figure 1.5 DNA double strand break (DSB) repair pathways...... 22 Figure 1.6 Characteristics of telomeric chromatin and DNA at ALT telomeres...... 30 Figure 1.7 ALT-mediated telomere synthesis...... 36 Figure 1.8 The multifaceted role of ZNF827-NuRD at ALT telomeres ...... 42 Figure 1.9 ZNF827 is a C2H2 zinc finger protein...... 48 Figure 3.1 ZNF827 gene expression in human tissues from public databases...... 89 Figure 3.2 ZNF827 have distinct transcriptional roles in ALT and telomerase cells. 95 Figure 3.3 Multiple transcriptional roles of ZNF827...... 100 Figure 3.4 STING and SNAI1 expressions are affected by ZNF827...... 100 Figure 3.5 ZNF827 may be a self-regulatory transcription factor...... 103 Figure 4.1 ZNF827 does not directly interact with COUP-TF II or TR4...... 118 Figure 4.2 ZNF827 may be involved in transcriptional suppression at telomeres. . 119 Figure 4.3 The N terminus of ZNF827 interacts with RBBP4...... 122 Figure 4.4 Interaction between RBBP4 and the ZNF827 N-terminal peptide...... 126 Figure 4.5 Structural comparison of the RBBP4-ZNF827 complex with RBBP4-FOG1 and RBBP4-PHF6 complexes. a ...... 128 Figure 4.6 RBBP4 mutants disrupt binding to ZNF827 and telomeres in vivo...... 131 Figure 4.7 Exogenous expression of RBBP4 mutants does not affect ALT activity. a ...... 133 Figure 4.8 ZNF827 SUMOylation is required for ALT activity and suppression of telomeric DNA damage...... 136 Figure 5.1 ZNF827 binds to ss G-rich telomeric DNA and the binding requires zinc fingers 1-3...... 145 Figure 5.2 Validation of ZNF827 and ZNF827 mutants purified by HaloTag®...... 147 Figure 5.3 ZNF827 binds to ss non-telomeric pentaprobes and binding is dependent on zinc fingers 1-3...... 150 Figure 5.4 ZNF827 zinc finger mutants display aberrant nuclear localisation...... 151

Figure 5.5 ZNF827 directly interacts with RPA...... 154 Figure 5.6 RPA induction by topotecan and ZNF827-telomere colocalisations...... 156 Figure 5.7 ZNF827 depletion exacerbates RPA foci induction by topotecan...... 160 Figure 5.8 ZNF827 is involved in the DNA damage and repair pathway, independent of ATM...... 164 Figure 5.9 ZNF827 interacts with ATR and ZNF827 depletion suppresses CHK1 and RPA activation...... 167 Figure 5.10 ZNF827 is at replication forks and plays a role in cellular response to replication stress. a...... 170 Figure 5.11 ZNF827 depletion induces replication defects and suppresses SCEs. 174 Figure 5.12 ZNF827 affects cell cycle progression and ZNF827 depletion causes G1 to early S arrest. a...... 178 Figure 5.13 ZNF827 depletion leads to abnormal mitotic entry, upregulation of p21 and suppression of the ATR-CHK1 pathway. a ...... 180 Figure 5.14 ZNF827 depletion inhibits cell growth and sensitises cancer cells to topotecan treatment...... 184 Figure 5.15 ZNF827 depletion increases NuRD components...... 185

xvi

TABLE OF CONTENTS

Introduction ...... 1

1.1 Telomeres protect chromosome ends from a DNA damage response ...... 1

1.2 Replication at telomeres – intrinsic limitations and challenges ...... 4

1.3 Telomere dysfunction and tumorigenesis ...... 7

1.4 Chemotherapeutic potential of targeting telomere maintenance mechanisms ... 11

1.5 ALT – Persistent replication stress and DNA damage drive HR-mediated telomere synthesis ...... 13 Phenotypic markers of ALT ...... 14 DSB repair pathways – NHEJ and HR ...... 20 ALT is an HR-mediated telomere maintenance mechanism ...... 23 Dysfunctional telomeres are processed differently in ALT cells ...... 25 ALT telomeres have significantly elevated replication stress ...... 27

1.6 The ALT proteome – interplay of proteins promoting and repressing HR ...... 31 Sequential steps in ALT-mediated telomere synthesis ...... 32 Factors that resolve replication stress repress ALT ...... 37

1.7 The expanded ALT proteome may provide more potential ALT-specific molecular targets ...... 39

1.8 ZNF827 in ALT-mediated telomere synthesis ...... 40

1.9 ZNF827 is a promising therapeutic target in ALT cancers ...... 43

1.10 Probable roles of ZNF827 beyond ALT telomeres ...... 44

1.11 The NuRD-binding motif in ZNF827 is conserved in other zinc finger proteins 46

1.12 ZNF827 and other C2H2 zinc finger proteins involved in the DDR...... 48

1.13 Characterising ZNF827 ...... 50

xvii

Materials and Methods ...... 52

2.1 Mammalian Cell Culture ...... 52 Growing cells ...... 52 Cryopreservation and thawing of cell lines ...... 52

2.2 Cell Line Manipulations: Transient Overexpression, siRNA Knockdown, and CRISPR knockout ...... 53 Generation of plasmid constructs ...... 53 Plasmid propagation ...... 55 Transient expression of plasmid constructs ...... 55 Gene knockdown with small interfering RNA ...... 58 ZNF827 knockout using Clustered Regularly Interspaced Short Palindromic Repeats/Caspase 9 (CRISPR/Cas9)...... 58

2.3 DNA Based Experiments ...... 59 Genomic DNA extraction ...... 59 C-circle assay ...... 60

2.4 RNA Based Experiments ...... 61 RNA extraction ...... 61 Quantitative reverse transcription PCR (qRT-PCR) ...... 61 TERRA analysis ...... 62

2.5 Protein Detection, Purification and Quantitation ...... 62 Protein extraction using Benzonase® ...... 62 Western Blot ...... 63 HaloTag® protein purification ...... 63 Protein quantitation by BCA (bicinchoninic acid) assay ...... 65 Mass spectometry ...... 65

2.6 Experiments for Protein-Protein Interactions ...... 65 Co-immunoprecipitation ...... 65 Protein colocalisations by immunofluorescence ...... 66 Proximity ligation assay (PLA) ...... 66

xviii

2.7 Experiments for Protein-DNA Interactions ...... 67 Electrophoretic Mobility Shift Assay (EMSA) ...... 67

2.8 Experiments for Chromatin Interactions ...... 68 Telomere - chromatin immunoprecipitation (ChIP) ...... 68

2.9 Next Generation Sequencing ...... 70 RNA-seq ...... 70 ChIP-seq ...... 71

2.10 Fixed Cell Fluorescence Microscopy ...... 72 Cell preparation ...... 72 Telomere - Fluorescence in situ hybridisation (telomere-FISH) ...... 73 Indirect immunofluorescence (IF) ...... 74 Combined indirect immunofluorescence with fluorescence in situ hybridisation (Combined IF with FISH) ...... 75 - Meta-TIFs (telomere dysfunction induced foci in metaphase), APBs (ALT associated PML bodies), telomere-ZNF827/RPA32 colocalisations ...... 75 Sister chromatid exchange (SCE) ...... 77 Comet assay ...... 78 Imaging and image analysis ...... 79

2.11 Time-lapse Live Cell Microscopy ...... 79 Fluorescence ubiquitination cell cycle indicator (FUCCI) live cell imaging ...... 79 Cell growth and apoptotic assay with IncuCyte® ...... 80

2.12 Cell Cycle Analysis ...... 80

2.13 Drug Treatment ...... 81

2.14 Antibodies – Primary and Secondary ...... 82

2.15 Statistical Analyses ...... 83

2.16 Mining ZNF827 Data in Public Databases ...... 83

xix

Unravelling the Biological Functions of ZNF827 ...... 84

3.1 Introduction ...... 84

3.2 ZNF827 orthologs and tissue expressions ...... 85

3.3 ZNF827 mutations and clinical phenotypes ...... 90

3.4 ZNF827 has distinct transcriptional roles in ALT and telomerase cells ...... 92

3.5 Multiple transcriptional roles of ZNF827 in the immune response, embryonic and organ development, cellular functions and cancer ...... 95

3.6 ZNF827 may be a self-regulatory transcription factor ...... 101

3.7 Discussion ...... 103

Defining the Molecular Interactions of ZNF827-NuRD and ZNF827-Nuclear Receptors, and the Role of ZNF827 SUMOylation at Telomeres ...... 110

4.1 Introduction ...... 110

4.2 ZNF827 recruitment to telomeres does not depend on direct interaction with TR4 and COUP-TF II, but may involve suppression of transcription at telomeres ...... 115

4.3 ZNF827 binds to the NuRD complex component RBBP4 ...... 120

4.4 Defining the binding interactions between ZNF827 and RBBP4 ...... 124

4.5 Structural comparisons of RBBP4-ZNF827, RBBP4-FOG1 and RBBP4-PHF6 127

4.6 RBBP4 mutants do not bind to ZNF827 or telomeres ...... 129

4.7 Disruption of the interaction between ZNF827 and RBBP4 is insufficient to alter ALT activity ...... 132

4.8 SUMOylation of ZNF827 is required for ALT activity and suppression of telomeric DNA damage ...... 134

4.9 Discussion ...... 136

ZNF827 is a Novel ssDNA Binding Protein Involved in the Replication Associated DNA Damage and Repair Pathway ...... 142

5.1 Introduction ...... 142

5.2 ZNF827 binds to single stranded (ss) telomeric and non-telomeric DNA via its zinc fingers 1-3 ...... 143

5.3 ZNF827 directly interacts with replication protein A (RPA) ...... 152

5.4 ZNF827 depletion exacerbates RPA induction following replication-associated DNA damage ...... 157

5.5 ZNF827 is involved in the DNA damage and repair pathway, independent of ATM-mediated NHEJ ...... 160

5.6 ZNF827 interacts with ATR and functions in the activation of ATR-CHK1 ...... 164

5.7 ZNF827 localises to replication forks and plays a role in the cellular response to replication stress ...... 168

5.8 ZNF827 depletion induces replication defects and impedes DNA repair by HR ...... 171

5.9 ZNF827 affects cell cycle progression and its depletion causes G1 to early S arrest ...... 175

5.10 ZNF827 depletion reduces cell growth and may sensitise cancer cells to topotecan treatment ...... 181

5.11 ZNF827 depletion increases NuRD components ...... 184

5.12 Discussion ...... 185

Chapter 6 Final Discussion ...... 193

xxi

References ...... 198

xxii

Introduction

1.1 Telomeres protect chromosome ends from a DNA damage response

The ends of linear chromosomes in eukaryotic cells resemble double strand

breaks (DSBs), which if left unprotected, are vulnerable to the vigilant cellular DNA

damage response (DDR) machinery. Telomeres provide the perfect solution, and have

evolved to counteract this problem in order to maintain genome stability. Telomeres are nucleoprotein structures at the termini of eukaryotic chromosomes that exert a

protective capping function against the cellular DDR pathways (de Lange, 2005,

Lazzerini-Denchi and Sfeir, 2016). In humans, telomeric DNA consists of long tandem

arrays of a hexanucleotide repeat sequence 5′-TTAGGG-3′, which are mostly double-

stranded (ds), but end in a G-rich single-stranded (ss) 3′ overhang (Meyne et al., 1989,

Makarov et al., 1997). The 3′ G-rich overhang strand invades the ds telomeric region, concealing the ends by forming a telomere-loop (t-loop) structure, facilitated by the six-subunit protein complex shelterin (Figure 1.1) (Griffith et al., 1999, de Lange, 2004).

By directly depositing onto the telomeric DNA, shelterin stabilises the t-loop and

shields telomeres from being processed as DSBs (de Lange, 2005, Lazzerini-Denchi

and Sfeir, 2016).

Proper capping function of telomeres fundamentally relies on the stability of the

t-loop and inherent residence of shelterin. Shelterin recognises 5′-TTAGGG-3′ repeats through the direct binding of its TRF1 and TRF2 subunits to the ds telomeric DNA

(Figure 1.1) (Chong et al., 1995, Broccoli et al., 1997). RAP1 (repressor activator

protein 1) is a conserved shelterin subunit that interacts with TRF2 (Sfeir et al., 2010).

TRF1 and TRF2 are bridged by TIN2 (TRF1-interacting nuclear factor 2), which

interacts with the TPP1 (discovered simultaneously by three different groups and

named as TINT1, PTOP or PIP1) and POT1 (protection of telomeres 1) heterodimer

(Ye et al., 2004a, Ye et al., 2004b, Liu et al., 2004, Houghtaling et al., 2004). POT1 is the third component of shelterin in direct contact with telomeric DNA. Through interaction with TPP1, POT1 is recruited to telomeres and covers the ss region of the

5′-TTAGGG-3′ repeats via its oligonucleotide/oligosaccharide binding (OB) folds

(Baumann and Cech, 2001, Loayza and de Lange, 2003). The formation of the t-loop is regulated by TRF2 (Doksani et al., 2013, Van Ly et al., 2018). Coordinated binding of these shelterin components is crucial for suppressing aberrant activation of DNA damage and repair pathways at telomeres (Figure 1.1).

Figure 1.1 Composition and function of the mammalian telomere. Telomeres are comprised of long tandem arrays of a hexanucleotide sequence TTAGGG, that are mostly double-stranded, but end with a single-stranded 3′ overhang (Meyne et al., 1989, Makarov et al., 1997). Telomeric DNA is shielded by the shelterin complex and

the 3′ overhang is tucked into the double-stranded region forming a telomere loop (t- loop). Shelterin is a six-subunit complex consisting of TRF1, TRF2, TIN2, RAP1, TPP1 and POT1. The shelterin complex together with the t-loop protect the chromosome ends from ATM and ATR DNA damage signalling pathways, and from DNA repair mechanisms by NHEJ and HR (de Lange, 2005). This figure was adapted from (Lazzerini-Denchi and Sfeir, 2016).

The two main DDR pathways are mediated by the signalling kinases ATM

(ataxia telangiectasia mutated) and ATR (ataxia telangiectasia and Rad-3 related),

which are activated respectively by the formation of DSBs and ss breaks (Shiloh,

2003). Usually, activation of a DDR pathway in response to DNA damage is crucial for

the maintenance of genome integrity by triggering cell cycle arrest to allow time for

DNA repair, or by eliminating irreparable cells by senescence and/or apoptosis.

However, aberrant activation of a DDR pathway at telomeres can cause catastrophic

end-to-end chromosome fusions and oncogenic genome instability (Artandi et al.,

2000). TRF2 is predominantly responsible for supressing the ATM-mediated DNA

damage signalling cascade and the detrimental end-to-end fusions caused by non- homologous end joining (NHEJ) at telomeres (Karlseder et al., 2004, Denchi and de

Lange, 2007). Depletion of TRF2 activates ATM and its downstream effectors resulting in the formation of telomere dysfunction-induced foci (TIFs) which are characterised by DNA damage markers 53BP1 and γH2AX associating with telomeres (Karlseder et al., 2004). The current mechanistic understanding of TRF2 in the suppression of a

DDR and NHEJ involves TRF2-mediated t-loop formation, which prevents ATM activation by concealing the chromosome end from detection by the DSB sensor

MRE11-RAD50-NBS1 (MRN) complex. This directly inhibits the ATM kinase as well as its key downstream effectors, such as the E3 ubiquitin ligase RNF168, and prevents

NHEJ by preventing Ku70-Ku80 heterodimerisation (Doksani et al., 2013, Karlseder

et al., 2004, Okamoto et al., 2013, Ribes-Zamora et al., 2013, Van Ly et al., 2018).

POT1 represses ATR-dependent DDR signalling by binding to telomeric

single-stranded DNA (ssDNA), subsequently inhibiting the loading of repair factors for

homologous recombination (HR) (Denchi and de Lange, 2007, Gong and de Lange,

2010, Wu et al., 2006). Inhibition of POT1 or its recruitment to telomeres by inhibiting

TPP1 or TIN2, induces the formation of ATR dependent TIFs (Denchi and de Lange,

2007, Kibe et al., 2010, Takai et al., 2011). It has been proposed that POT1 inhibits

ATR activation by occlusion of RPA at telomeric ssDNA by competition or via an

hnRNPA1 (heterogeneous nuclear ribonucleoprotein A1) mediated RPA-to-POT1

switch (Gong and de Lange, 2010, Flynn et al., 2011). It is particularly important to

ensure suppression of ATM and ATR signalling at telomeres during DNA replication

when the transiently unwound t-loop presents a substrate for DDR activation. During

telomere replication, TRF2 binding to telomeric DNA is essential to inhibit ATM,

whereas ATR inhibition involves the coordination of TPP1-POT1 and TRF1, the

shelterin component required for efficient replication of telomeres (Sfeir et al., 2009).

1.2 Replication at telomeres – intrinsic limitations and challenges

Other than the propensity of eliciting a DDR, there are intrinsic limitations and challenges associated with replication at telomeres which lead to gradual telomere attrition and inherently heightened replication stress. Human telomeres undergo gradual shortening at a rate of 40-200 base pairs for each cell division (Hastie et al.,

1990, Harley et al., 1990), partly attributed to the inability of DNA polymerases to copy the end of the lagging strand during DNA replication, known as the end replication

problem (Olovnikov, 1973). Additionally, nucleolytic processing of 5′ end resection

occurs at telomeres post replication to regenerate the 3’ G-rich overhangs essential

4 for t-loop formation (Dai et al., 2010). The leading and lagging strands are initially processed differently to generate overhangs, by Apollo nuclease-mediated blunt end resection and removal of the RNA primer from the terminal Okazaki fragment, respectively (Wu et al., 2010, Lam et al., 2010). The ends of both strands are further resected by exonuclease 1 (EXO1), followed by C-strand fill-in facilitated by the CST

(CTC1-STN1-TEN1) complex (Figure 1.2) (Wu et al., 2012, Dai et al., 2010). This tightly regulated process is thought to also contribute considerably to the rate of telomere shortening (Huffman et al., 2000).

Figure 1.2 The mechanistic details of telomere shortening. Lagging-strand DNA synthesis leaves a short 3′ overhang after the removal of the terminal RNA primer.

Leading-strand DNA synthesis can be fully completed to the end, which is processed by the nuclease Apollo to regenerate the 3′ overhang, essential for t-loop formation. Both the leading and lagging ends are further resected by EXO1 to create extended overhangs, which are then filled in by the CST complex.

Telomere replication poses an intrinsic challenge to the DNA replication

machinery in several ways. While the terminal t-loop structure is pivotal in suppressing

an aberrant DDR, it presents an inherent barrier to DNA replication (Uringa et al., 2012,

Vannier et al., 2013, Sarek et al., 2015). Telomeric DNA has the tendency to form

stable secondary structures including G-quadruplexes, D-loops, T-loops, DNA-RNA hybrids and Holliday junctions, all of which pose an impediment to the DNA replication machinery (Pickett and Reddel, 2015, Sundquist and Klug, 1989). The repetitive

nature of telomeric DNA is likely to induce stalling of replication forks and

chromosomal breakage, making telomeres resemble common fragile sites (Sfeir et al.,

2009, Verdun and Karlseder, 2006, Martinez et al., 2009). Unresolved impediments at

replication forks can exacerbate fork stalling into fork collapse, compromising the

fidelity of replication. Multiple proteins have been shown to assist in alleviating replication stress at telomeres to warrant faithful replication. TRF1 plays a pivotal role in counteracting telomere fragility and preventing stalled replication forks (Sfeir et al.,

2009). Loss of TRF1 activates a strong DDR in S phase that is dependent on ATR, the major regulator of the replication stress response (Flynn and Zou, 2011), but not

ATM. Helicases BLM and RTEL1 have been shown to suppress the fragile telomere

phenotype. Together these data suggest that TRF1 specifically functions in S phase

to facilitate telomere replication, and suppresses ATR through cooperation with BLM

and RTEL1, which unwind secondary structures such as G-quadruplexes (Sfeir et al.,

2009).

In addition, TRF2 has been shown to unwind t-loops with the help of RTEL1, allowing the progression of replication forks through telomeres (Sarek et al., 2015). In cooperation with Apollo nuclease, TRF2 relieves topological stress during replication of telomeres (Ye et al., 2010). Another RecQ helicase, WRN, has been shown to mediate telomere lagging strand synthesis via a TRF1-independent mechanism

(Crabbe et al., 2004). The CST complex that binds to telomeric ssDNA has also been implicated in assisting telomere replication by promoting replication fork restart at stalled forks (Stewart et al., 2012). Furthermore, circumstantial evidence suggests that

BRCA2 and RAD51 are important in preventing telomere fragility independently of

TRF1, by promoting HR-mediated repair of stalled forks (Badie et al., 2010). It appears that efficient telomere replication is regulated by numerous factors, but the precise interaction dynamics between these factors in assisting replication fork progression through telomeres remain to be elucidated. Since replication stress is a prominent source of genome instability and greatly implicated in tumorigenesis (Gaillard et al.,

2015), understanding how cells resolve replication stress associated with telomeres will help increase our understanding of genome stability and its contribution to cancer development.

1.3 Telomere dysfunction and tumorigenesis

Telomere dysfunction can be caused by replication defects, depletion or inhibition of shelterin binding, and gradual telomere attrition accompanied with cell division. Because of gradual telomere attrition, telomere length is the major determinant of the replicative lifespan of most human somatic cells. After a limited number of cell divisions, telomeres eventually become too short to maintain chromosome end protection. Telomere shortening or telomere deprotection eventuate in loss of telomere function and subsequent activation of a DDR, which arrests cell

7 proliferation and induces replicative senescence in primary human cells (Figure 1.3)

(d'Adda di Fagagna et al., 2003, Kaul et al., 2012). Senescent primary human cells exhibit elevated levels of γH2AX, a typical marker of the DDR, globally and specifically at telomeres. This is accompanied by activated DNA damage checkpoint kinases

CHK1 and CHK2 – downstream targets of ATR and ATM respectively, indicative of activation of both DNA damage signalling pathways (d'Adda di Fagagna et al., 2003,

Kaul et al., 2012). Dysfunctional telomeres therefore elicit a DDR analogous to DSBs.

It has been demonstrated that five dysfunctional telomeres are sufficient to induce senescence (Kaul et al., 2012). Telomere-initiated replicative senescence is mediated by either p53 or p16/RB tumour suppressor pathways (Smogorzewska and de Lange,

2002). Thus, the natural telomere shortening process is thought to function as a tumour suppression mechanism.

Cells that have a muted DDR or cell cycle arrest phenotype, especially those lacking functional p53 or RB tumour suppressor pathways, can bypass senescence, leading to extended growth, which ultimately results in telomere crisis (Figure 1.3)

(Shay et al., 1991, Artandi and DePinho, 2010). In crisis, the accumulation of critically short dysfunctional telomeres causes chromosome end-to-end fusions and large- scale chromosomal rearrangements, resulting in extensive genome instability (Artandi et al., 2000, Maciejowski et al., 2015). Genome instability in cells going through crisis is thought to promote tumorigenesis by inducing widespread genomic alterations including aneuploidy, gene loss and amplification, translocations, whole-genome reduplication, chromothripsis and kataegis, which are characteristics commonly found in cancers (Maciejowski and de Lange, 2017). Even though most cells in the crisis state succumb to apoptosis due to chromosome breakage and additional telomere deprotection during prolonged mitotic arrest (Hayashi et al., 2012, Hayashi et al.,

2015), rare incipient cancer cells harbouring favourable mutations escape crisis by activation of a telomere maintenance mechanism (TMM), which counteracts the genome instability resulting from dysfunctional telomeres. Thus, telomere crisis exerts a selective pressure permissive to cancer development. The two-sided consequences of dysfunctional telomeres, either a trigger of tumour suppressive replicative senescence, or a driver of tumour promoting genome instability in crisis, are greatly dependent on the reliability of the initial DDR and the integrity of subsequent signalling pathways for cell fate.

Figure 1.3 Graphical representation of telomere length dynamics in somatic cells over cell divisions from physiological state to cancer state. In somatic cells, telomeres shorten gradually with each cell division. Eventually short telomeres lose end protection function and trigger a DDR-mediated growth arrest called senescence. Some cells that have lost or acquired mutations in the p53 or RB tumour suppressor pathways bypass senescence and continue to divide, resulting in telomere erosion that causes chromosome fusions and extensive genome instability. This stage is

termed telomere crisis, where most cells undergo apoptosis. Rarely, a few cells can escape the crisis state by activating a telomere maintenance mechanism (TMM), which allows them to replenish the telomeres and turn into an unlimited proliferative state or cancer. The majority (85-90%) of cancer cells activate telomerase. The remaining 10-15% of cancer cells activate the alternative lengthening of telomeres (ALT) pathway.

Activation of a TMM is an inevitable solution for cancer cells to resolve telomere

dysfunction. The majority of human cancers (~90%) activate telomerase to heal dysfunctional telomeres in order to attain unlimited proliferative capacity (Shay and

Bacchetti, 1997). Telomerase is an enzyme capable of directly adding telomeric

repeats to chromosome ends via reverse transcription using its core components: the

reverse transcriptase (hTERT), the internal RNA template (hTR), and several RNA

binding proteins which confer stability (Morin, 1989, Cohen et al., 2007). Whole genome sequencing studies in melanomas and neuroblastomas reveal telomerase activation is highly correlated with TERT promoter mutations and TERT loci rearrangements (Huang et al., 2013, Horn et al., 2013, Peifer et al., 2015, Valentijn et al., 2015). A telomerase-independent TMM, referred to as alternative lengthening of telomeres (ALT) is utilised by a small but significant proportion of cancers (~10%)

(Heaphy et al., 2011b, Bryan et al., 1995). ALT is a homologous recombination (HR)

–dependent TMM that uses intra-, inter- or extra-chromosomal telomeric DNA as a copy template for telomere extension (Dunham et al., 2000, Muntoni et al., 2009,

Natarajan et al., 2003). ALT has been associated with mutations in the chromatin remodeler α-thalassaemia/mental retardation syndrome X linked (ATRX) both in vitro

and in human cancers (Lovejoy et al., 2012, Heaphy et al., 2011a). ATRX suppression

10 favours the activation of ALT in immortalised cell lines, even though the key steps underlying ALT activation in cancers remain elusive (Napier et al., 2015).

1.4 Chemotherapeutic potential of targeting telomere maintenance mechanisms

On the premise that all cancers require a TMM to sustain unlimited proliferation,

TMMs present an attractive avenue for anticancer therapeutics, and there has been a growing interest in targeting TMMs as a strategy to develop anticancer therapies. The strategy targeting TMMs has been focused on direct inhibition of the telomere lengthening process by inhibiting the proteins involved. One typical example is targeting telomerase in cancers employing telomerase as a TMM. Increasing structural and functional knowledge of telomerase has led to the development of telomerase inhibitors, one of which is imetelstat, a synthetic oligonucleotide that inhibits telomerase activity by antagonising the RNA template hTR (Ruden and Puri, 2013).

Even though its promising anticancer effects in inhibiting telomerase activity and cell proliferation have been shown in a number of cancers by preclinical and clinical studies, leading to its advancement to Phase II clinical trials, the use of imetelstat in clinical settings has been limited by its off-target effects causing pronounced haematological toxicity and inhibition of mesenchymal stem cells (Schrank et al., 2018).

Cancers require an extended period of treatment to allow for sufficient telomere shortening, but the toxicities associated with imetelstat necessitate frequent drug free intervals, which could allow restoration of telomere length (Mender et al., 2015).

Moreover, in cancers with long telomeres, the prolonged duration required to cause sufficient telomere shortening may render the effect of inhibiting telomerase, or in a broader sense, TMM inhibition in treating cancers clinically irrelevant. Thus, one would argue that the clinical importance of TMM inhibition lies in preventing the reoccurrence of secondary malignancies by preventing the activation of TMMs in incipient cancer

11 cells. Another proposal for a much more effective strategy of targeting telomeres in cancers is to disrupt the integrity of the protective t-loop; the ensuing telomere uncapping will effectively cause a DDR and chromosomal fusions, consequently resulting in growth arrest and cell death within a couple of cell divisions.

Nevertheless, a successful anti-TMM strategy is reliant on finding molecular targets of TMMs differential in cancer cells for targeted therapies. This essentially depends on a complete understanding of the TMMs. While there is a clear molecular target in telomerase cancers, cancers utilising ALT lack suitable therapeutic targets for anticancer strategies. For several reasons, development of anticancer therapies targeting ALT is of significant clinical value. ALT has a higher prevalence in a subset of cancers developed from mesenchymal origins such as glioblastoma multiforme, osteosarcomas and various types of soft tissue sarcomas that are often associated with poor prognosis (Bryan et al., 1997, Hakin-Smith et al., 2003, Henson et al., 2005,

Costa et al., 2006). The ALT mechanism therefore presents a prime target for diagnosis, prognosis and treatment of these aggressive cancers. Unsurprisingly, telomerase inhibitors are ineffective in treating cancers or preventing new cancers using ALT as the TMM. Furthermore, there are concerns that telomerase inhibition will promote activation of ALT in some cancers (Hu et al., 2012). Because inhibiting either

TMM exerts selection pressure for activation of the other, and most tumours display heterogeneity in TMMs, it may be anticipated that successful anticancer treatment strategies targeting TMMs involve the use of both telomerase and ALT inhibitors to prevent drug resistance.

In a recent study, ALT cancer cells were found to be hypersensitive to ATR inhibitors, which suggests increased ATR signalling at ALT telomeres possibly due to elevated replication stress and a persistent DDR (Flynn et al., 2015). However, it was

shown subsequently in another study that ATR inhibition does not universally kill ALT

cancer cells (Deeg et al., 2016), suggestive of a more complex mechanism in ALT that

involves multiple DNA damage signalling pathways. Currently, the causal genetic or

epigenetic changes leading to ALT activation have not been identified, while our

knowledge of ALT-mediated telomere synthesis has been increasing. A thorough

understanding of the ALT mechanism, the implicated DDR pathways, and the ALT-

specific proteome will improve our understanding of the aggressive cancers that utilise

ALT, and is fundamental in the pursuit of finding therapeutic targets differential to ALT.

1.5 ALT – Persistent replication stress and DNA damage drive HR-mediated telomere synthesis

It is generally recognised that ALT is an HR-dependent TMM (Pickett and

Reddel, 2015). Activation of ALT in cancer cells enables sufficient telomere replenishment to attain unlimited proliferation, but does not fully suppress telomere dysfunction. In fact, ALT cancer cells display multiple characteristics of telomere dysfunction. Chromosome ends with undetectable telomere signals as well as a high level of spontaneous telomeric DDR, indicative of eroded and deprotected telomeres, are characteristic of ALT cancer cells (Henson et al., 2002, Cesare et al., 2009,

Lovejoy et al., 2012). Also, there are ongoing genome rearrangements and instability in ALT cells, likely to be induced by telomere dysfunction (Scheel et al., 2001, Lovejoy et al., 2012). Telomere dysfunction is strongly linked to ALT-associated telomere recombination (Cesare and Reddel, 2008). It has been hypothesised that ALT is a consequence of accrued telomere dysfunction in conjunction with de-repression of a telomere-specific HR mechanism, although the exact mechanistic details leading to

ALT activation remain to be solved (Cesare and Reddel, 2008, Pickett and Reddel,

2015). Recent studies have corroborated that telomere dysfunction induced by

replication stress and telomeric DSBs drive HR-mediated repair and ALT-associated

telomere synthesis (O'Sullivan and Almouzni, 2014, Cho et al., 2014, Dilley et al., 2016,

Doksani and de Lange, 2016).

Phenotypic markers of ALT

By definition, ALT is telomere maintenance via a telomerase independent

mechanism (Bryan et al., 1995, Bryan et al., 1997). The most definitive proof of ALT

in an immortal or tumour cell line is therefore the maintenance of telomere lengths

over hundreds of population doublings in the absence of telomerase activity. However,

this approach is not practical due to the length of time required for the cells to pass

through numerous population doublings, its inability to measure the level of ALT

activity, and impracticalities in examining human tumours (Henson and Reddel, 2010).

Fortunately, ALT is associated with several phenotypic markers that are widely used

to identify ALT status and measure ALT activity in cancer cell lines and tumour

samples (Figure 1.4). These markers are correlated with either HR or elevated DNA damage at ALT telomeres.

1.5.1.1 Telomere length heterogeneity

The average telomere length in mortal human cells ranges from 15kb in sperm

and embryonic cells to 4kb in cells at crisis (Harley, 1991). Distribution of telomere length across telomerase-positive and ALT immortalised cell lines and tumours has been well studied and is one of the characteristics useful for differentiation of the TMMs.

In contrast to the comparatively homogenous telomere lengths in telomerase-positive tumours and cell lines (de Lange et al., 1990, Counter et al., 1992), ALT cell lines and tumours exhibit telomere length heterogeneity shown by telomere restriction fragment

(TRF) analysis, with length distributions ranging from very short (<3kb) to extremely long (>50kb) averaging at a mean length of ~20kb (Figure 1.4 a) (Murnane et al., 1994,

Bryan et al., 1995, Bryan, 1997). Telomere lengths can also be investigated via a

visualisation method using fluorescence in situ hybridisation (FISH), which

corroborated telomere length heterogeneity in ALT cells with FISH signals ranging

from very pronounced to undetectable (Perrem et al., 2001). This heterogeneity can

be explained by rapid and random telomere length fluctuations observed in ALT cells

(Murnane et al., 1994). The rate of change in telomere lengths varied from gradual

telomere erosion of 50-100bp to rapid increases of up to 20kb in both critically short

and long telomeres (Murnane et al., 1994). The long heterogeneous telomere length

pattern is a well-established hallmark for ALT and may be used to help identify ALT

tumours by direct assessment of telomere lengths (Meeker et al., 2002).

1.5.1.2 ALT-associated PML Bodies (APBs)

The occurrence of APBs is a typical hallmark of ALT cells. APBs are the subset

of ubiquitous promyelocytic leukaemia nuclear bodies (PML-NBs) that contain

telomeric DNA (Figure 1.4 b) (Yeager et al., 1999). PML-NBs are dynamic, functionally

heterogeneous, ring-like structures comprised of different isoforms of PML protein that bring various other nuclear proteins close together in the nucleus (Bernardi and

Pandolfi, 2007, Jensen et al., 2001, Lang et al., 2010). In addition to telomeric DNA,

APBs contain telomere binding proteins amidst DNA repair, replication and recombination proteins (Wu et al., 2003, Nabetani et al., 2004, Henson et al., 2005).

It has long been proposed that APBs are sites of ALT-associated telomeric recombination, based on multiple lines of evidence. The HR-associated MRN complexes (MRE11-RAD50-NBS1) and structural maintenance of chromosome 5

(SMC5)-SMC6 complexes present in APBs are required for ALT activity, while the

MRN complex is essential in APB formation (Jiang et al., 2005, Potts and Yu, 2007,

Zhong et al., 2007, Wu et al., 2003). When ALT is inhibited by knockdown or

sequestration of these complexes, APBs decrease or disappear (Jiang et al., 2005,

Zhong et al., 2007, Potts and Yu, 2007). Depletion of PML-NBs results in reductions

in telomere length and extrachromosomal telomeric repeats (Osterwald et al., 2015).

The appearance of APBs in immortalised cell lines coincides with ALT activation and increases in frequency in G2 phase, when telomeres are most recombinogenic

(Yeager et al., 1999, Grobelny et al., 2000).

Cell cycle regulated telomere clustering occurs within APBs, and promotes telomere-telomere interactions between heterologous chromosome ends (Draskovic et al., 2009). APBs can be further induced by DNA damage in ALT cells, and predominantly harbour linear telomeric DNA, suggesting a role of APBs in sequestering linear chromosome ends for HR repair (Fasching et al., 2007). Recent studies have corroborated that DSBs initiated specifically at telomeres induce APB associated telomere clustering events between heterologous chromosomes, leading to HR-mediated telomere synthesis in ALT cells (Cho et al., 2014, Dilley et al., 2016).

Direct evidence confirming APBs as sites of HR-mediated ALT telomere synthesis have emerged only recently. Zhang et al. has developed an assay called ATSA (ALT telomere DNA synthesis in APBs) and directly demonstrated telomere synthesis in

APBs in the natural cellular environment in the absence of induced DNA damage

(Zhang et al., 2019a). Paradoxically, there is some evidence inconsistent with APBs being platforms of telomere recombination in ALT cells. APBs have been reported to be absent in two cell lines utilising ALT (Fasching et al., 2005). Telomerase-positive cells with long telomeres also contain APBs without evidence of ALT-mediated telomere maintenance (Pickett et al., 2009).

1.5.1.3 C-circles ALT cells have a high abundance of linear or circular extrachromosomal

telomeric repeats (ECTRs). The most commonly used ECTR as a marker for ALT

status and activity is C-circles. C-circles are extrachromosomal telomeric DNA circles

that are mostly double-stranded with a partially ss C-rich region (Henson et al., 2009).

High levels of C-circles are consistently found across all the ALT cell lines tested

(Henson et al., 2009). C-circle occurrence strongly correlates with ALT activation during cellular immortalisation, and rapidly declines within 24 hours following ALT inhibition (Henson et al., 2009). Telomerase-positive cell lines and mortal cell lines contain 750-fold less C-circles than ALT cells (Henson et al., 2009). These observations establish C-circles as a robust marker for ALT activity. C-circles consist of complete telomeric C-rich strands and incomplete G-rich strands allowing them to be readily detected by modified rolling circle amplification assays (Figure 1.4 c)

(Henson et al., 2009). Despite the widespread use of the C-circle assay for the identification of ALT status in cancer cell lines and tumour samples, the origin of C- circles has remained elusive. A very recent study has shed some light on this phenomenon. Zhang et al. has shown that C-circles can arise from break-induced replication fork collapse on the lagging C-rich strand, implicating C-circle generation in the replication stress response at ALT telomeres (Zhang et al., 2019b). Interestingly,

C-circles have also been observed in a telomerase-positive cell line incorporated with non-canonical telomeric variant repeats and human embryonic stem cells with long telomeres (Conomos et al., 2012, Rivera et al., 2017), which supports that C-circles may be a by-product of replication stress resolution. Overall, C-circles are a practical and specific ALT marker, although it remains to be fully elucidated how C-circles are initiated in ALT cells and its relation to the ALT mechanism.

1.5.1.4 Telomere-sister chromatid exchanges (T-SCEs)

T-SCEs are post-replicative telomeric exchange events between sister

chromatids, which are highly elevated in ALT cells in comparison to telomerase-

positive cells (Londono-Vallejo et al., 2004, Bechter et al., 2003). The increased

exchange events are specific to telomeric regions as the frequencies of sister chromatid exchanges (SCEs) and HR events at non-telomeric sequences are similar

between telomerase-positive and ALT cells (Londono-Vallejo et al., 2004, Bechter et

al., 2003). T-SCEs can be visualised as fluorescent signals using Chromosome

Orientation Fluorescence in situ hybridisation (CO-FISH), which involves incorporation

of 5‐bromo‐2'‐deoxyuridine (BrdU) into the newly synthesised DNA strand, followed

by its destruction by UV radiation leaving the template strand intact for FISH with

telomere strand-specific probes that are able to show if a post-replicative cross-over

event has occurred between a leading and a lagging strand telomere (Figure 1.4 d)

(Londono-Vallejo et al., 2004).

It has been found that increased T-SCEs, resulting from overexpression of the

Holliday junction resolvase SLX4 in ALT cells, paradoxically coincide with telomere shortening and a reduction of telomere extension events, suggesting that T-SCE events result from a cross-over resolution process to prevent hyperextension of telomeres in ALT cells (Wilson et al., 2013, Sarkar et al., 2015, Sobinoff et al., 2017).

These findings suggest that even though T-SCEs appear to have a good correlation to ALT activity, they may only reflect increased recombination events in ALT cells rather than increased telomere extension directly.

1.5.1.5 Telomere Dysfunction-Induced Foci (TIFs)

Typically, ALT cancer cells have an elevated level of spontaneous DDR at telomeres compared to telomerase positive cancer cells (Cesare et al., 2009, Lovejoy

et al., 2012). Telomeric DNA damage can be visualised and measured by

fluorescence microscopy as colocalisations of 53BP1 or γH2AX DNA damage foci and

telomeric foci. 53BP1 or γH2AX positive telomeric foci are named telomere

dysfunction-induced foci (TIFs) (Figure 1.4 e) (Cesare et al., 2009). Accordingly, the

detection of TIFs in metaphase is termed meta-TIFs. Both interphase TIF and meta-

TIF assays are used to quantitatively measure the extent of DNA damage at telomeres.

Figure 1.4 Graphical representation of phenotypic markers of ALT. a. Telomere length can be measured by terminal restriction fragment (TRF) analysis (left) and visualised by telomere-FISH (right). ALT cancer cells typically depict long heterogenous telomere lengths in comparison to the relatively shorter and homogenous telomere lengths in telomerase (Tel+) cells. b. APBs (ALT-associated PML bodies), visualised by the colocalisation of telomeric DNA and PML (promyelocytic leukaemia) nuclear bodies, is a typical phenotypic marker of ALT cancer cells. c. ALT cancer cells contain extrachromosomal telomeric DNA circles – C-circles that are self-priming and usually detected by a modified rolling circle amplification assay (Henson et al., 2009). The C-circle assay image was adapted from (Henson et al., 2009). d. T-SCEs (Telomere-sister chromatid exchanges) are telomere recombination events that frequently occur in ALT cancer cells and correlate well with ALT activity. T-SCEs are visualised as dual fluorescent signals using Chromosome Orientation Fluorescence in situ hybridisation (CO-FISH). e. ALT cancer cells have an elevated level of spontaneous TIFs (Telomere dysfunction-induced foci) visualised by the colocalisation of telomeric DNA and γH2AX.

DSB repair pathways – NHEJ and HR DSBs are detrimental DNA lesions that can potentially lead to chromosome fragmentation and cell death if not repaired properly (Aparicio et al., 2014). Faithful repair of DSBs is essential for safeguarding genome integrity and cell survival. Repair of DNA DSBs predominantly relies on the two major DNA repair pathways – classical

NHEJ (c-NHEJ) and HR. c-NHEJ involves ligation of two broken blunt ends independent of sequence homology, and remains active throughout the cell cycle with dominance in G0/G1 (Lieber, 2008). Initiation of NHEJ requires loading of the Ku 70/80 heterodimer onto blunt or near blunt ends, which activates the DNA-dependent protein kinase catalytic subunit (DNA-PKcs) that subsequently mediates a signalling cascade to facilitate the repair process (Figure 1.5 a) (Lieber, 2008). Despite the mutagenicity of c-NHEJ resulting from removal or random addition of a few nucleotides at the re-

20 joining sites, c-NHEJ has a critical role in restoring genome integrity by preventing chromosomal loss, particularly in G0/G1 when sister-chromatids are not present

(Chiruvella et al., 2013). An alternative pathway of NHEJ (alt-NHEJ) has been described in more recent years. In contrast to c-NHEJ, alt-NHEJ requires initial end resection to generate microhomologies for the subsequent ligation, facilitated by a distinct set of proteins including CtIP, the XRCC1/DNA ligase III complex and PARP1

(Figure 1.5 b) (Ceccaldi et al., 2016). DSB repair by alt-NHEJ is more mutagenic because of increased loss of nucleotides and predilection of inter-chromosome fusions that can lead to chromosomal translocations and rearrangements (Wang et al., 2006,

Iliakis et al., 2015). Alt-NHEJ is generally considered as a backup repair pathway when c-NHEJ is deficient (Wang et al., 2006, Iliakis et al., 2015).

HR is an error-free repair pathway, as it utilises a homologous sister chromatid as a template for DNA repair. HR predominates during S and G2 phases in mitotic cells when sister chromatids are available as homologous templates, and nucleases mediating end resection are more active (Shrivastav et al., 2008, Ira et al., 2004, Aylon et al., 2004). DNA end resection at DSBs prevents c-NHEJ and primes the ends for

HR. 5′ end resection is initiated by the MRN complex in collaboration with CtIP, and is then extended by the nucleases EXO1 and DNA2. Replication protein A (RPA) initially coats the ssDNA overhangs generated by end resection, forming RPA-ssDNA intermediates to protect the nascent ssDNA from degradation or self-annealing (Chen et al., 2013). Subsequently, multiple mediators including RAD52, BRCA2 and PALB2 promote displacement of RPA and RAD51 loading onto ssDNA (Figure 1.5 c). The formation of RAD51-ssDNA nucleofilaments facilitates homology searches and strand invasion, resulting in double Holliday junctions, that are processed by either a resolvase into non-crossover or crossover products, or dissolved by BLM-TOP3A

21 mediated branch migration and dissolution resulting in exclusively non-crossover products (Li and Heyer, 2008).

As explained previously, DSB DNA damage and repair pathways are actively repressed at normal telomeres. Genome instability caused by aberrant repair of shortened telomeres, on the background of loss of tumour suppressor function contribute to tumorigenesis. While dysfunctional telomeres are exposed as DSBs that can potentially unleash all these DNA repair pathways, cancer cells utilising ALT have evolved a HR-dominant mechanism to repair the shortened telomeres.

Figure 1.5 DNA double strand break (DSB) repair pathways. The two major DSBs repair pathways are NHEJ and HR. a. c-NHEJ is initiated by the binding of Ku 70/80 at DSBs that blocks end resection, followed by activation of DNA-PKcs which triggers

a signalling cascade that mediates end ligation. It is a relatively accurate repair pathway responsible for the repair of majority of DSBs with minor deletions. b. Alt- NHEJ is a minor NHEJ repair pathway that requires some extent of end resection to generate microhomologies for end ligation. DNA repair by alt-NHEJ is mutagenic because of large deletions and insertions and the propensity of causing chromosome rearrangement. c. HR is a dominant repair pathway for homology-directed repair when sister chromatids are available. HR repair requires extended end resection for subsequent RAD51-dependent strand invasion. It is considered an error-free repair pathway. This figure was adapted from (Ceccaldi et al., 2016).

ALT is an HR-mediated telomere maintenance mechanism

Several studies agree that HR-mediated telomere lengthening by ALT is the

synthesis of new telomeric DNA using another telomeric sequence as a copy template.

The first direct evidence that ALT involves HR arises from the observation of intertelomeric tag copying, in which a DNA tag incorporated into the telomeres of a single chromosome in ALT cells was copied onto to another, demonstrating intertelomeric templating (Dunham et al., 2000). The DNA template for ALT is not necessarily another telomere, the DNA tag can also be copied within the same telomere indicative of the ability of telomeres to use themselves, or a telomere on a sister chromatid as a copy template (Muntoni et al., 2009). Also, extrachromosomal telomeric DNAs C-circles, used as a typical marker of ALT, have been demonstrated to act as an effective template for rolling circle amplification in vitro (Henson et al.,

2009). Therefore, it is plausible that the ALT mechanism involves copying of any telomeric sequences, either extrachromosomal or chromosomal. Further evidence of an HR-mediated mechanism comes from studies that illustrate the necessity of HR

proteins including the MRN complex, SMC5-SMC6 heterodimer, FANCA, FANCD2,

MMS21, MUS81, and FEN1 in telomere maintenance in human ALT cells (Cesare and

Reddel, 2010). Many other HR proteins including RAD51, RAD52, RPA, BLM, SLX4,

MRN, BRIT1 and BRCA1 are also present at ALT telomeres (Wilson et al., 2013, Wan

et al., 2013, Bhattacharyya et al., 2009, Sobinoff et al., 2017, Conomos et al., 2014).

In ALT cells, HR-directed repair that is usually restricted to cells in S and G2

phases becomes dysregulated, and can occur throughout the cell cycle even in G1

and during mitosis (Cho et al., 2014, Min et al., 2017). Consistent with global HR

pathways, HR-mediated telomere synthesis in ALT inherently involves 5′ end resection, invasion of the 3′ overhang into a DNA template forming an HR intermediate, extension by template copying and dissolution of the recombination intermediate (Pickett and

Reddel, 2015). The unique structure of telomeres give rise to some complexities.

Telomeres intrinsically have a 3′ telomeric overhang shielded from irregular recombination processes by the shelterin component POT1. In ALT-associated HR, this overhang must somehow be released from POT1 and loaded with RAD51 in order to facilitate invasion into an adjacent telomeric DNA template to initiate HR (Pickett and Reddel, 2015). It is not known how this process occurs. Due to the terminal one- ended nature of telomeres, HR intermediates formed following strand invasion are D- loops instead of double Holliday junctions. The processes of HR-mediated telomere extension and dissolution of D-loops have been extensively examined in recent studies. It has been demonstrated that the mechanism of telomere synthesis in ALT is analogous to the process of one-ended DSB repair, called break-induced replication

(McEachern and Haber, 2006). Current evidence elucidating the molecular steps and proteins involved in HR-mediated telomere synthesis in ALT is reviewed in a later section.

Dysfunctional telomeres are processed differently in ALT cells

Dysfunctional telomeres, detected as DSBs by the DDR, can elicit either NHEJ

or HR, depending on how the telomere becomes dysfunctional. Telomere

dysfunction/deprotection due to gradual telomere attrition or induced by shelterin loss predominantly causes c-NHEJ-mediated chromosome end-to-end fusions (Counter et al., 1992, Dimitrova et al., 2008). Interestingly, there is substantial evidence to support alt-NHEJ activity at dysfunctional telomeres. Chromosome end-to-end fusions following gradual telomere attrition continue to occur even when DNA ligase 4, DNA-

PK and 53BP1, the core components of c-NHEJ, are deleted in mouse cells, suggestive of active alt-NHEJ at naturally shortened telomeres (Wong et al., 2007, Rai et al., 2010). Sequence analysis of telomere fusions in human cancers exposed frequent microhomologies and extensive deletions in the fusion sites which are defining features of alt-NHEJ (Lin et al., 2010, Simpson et al., 2015, Letsolo et al.,

2010, Mateos-Gomez et al., 2015). These findings implicate the involvement of alt-

NHEJ in processing dysfunctional telomeres in the early stages of tumorigenesis. The highest alt-NHEJ activity is observed after the complete loss of the shelterin complex in Ku70/80 deficient mouse cells, suggesting that alt-NHEJ is redundantly repressed by Ku70/80 and the shelterin complex at normal telomeres (Sfeir and de Lange, 2012,

Rai et al., 2010).

In recent years, a new experimental tool using fusion of the nuclease Fok1 to

TRF1 (TRF1-Fok1) has been developed to induce DSBs specifically at telomeres (Cho et al., 2014). Interestingly, a recent study has shown that both HR and PARP1/Ligase

III dependent alt-NHEJ contribute to the repair of TRF1-FokI induced DSBs at internal telomeres of mouse embryonic fibroblasts (MEFs) during S phase (Doksani and de

Lange, 2016). Intriguingly, HR-directed repair of TRF1-FokI induced DSBs leads to

the induction of ALT hallmarks including telomere length heterogeneity, increased T-

SCEs, and the accumulation of extrachromosomal telomeric repeats (ECTRs)

(Doksani and de Lange, 2016). It is surprising that alt-NHEJ, but not c-NHEJ,

contributes to internal telomeric DSB repair, since genomically alt-NHEJ is an error-

prone repair pathway that functions as a backup pathway to c-NHEJ (Iliakis et al.,

2015). These findings show that dysfunctional telomeres induced by internal telomeric

DSBs in non-ALT cells are repaired by a combination of HR and NHEJ pathways. In

contrast, dysfunctional telomeres in ALT cells predominantly undergo HR (Dunham et

al., 2000). TRF1-Fok1 induced DSBs at telomeres in ALT cells initiate RAD51- dependent long-range homology-directed searches resulting in telomere clustering and HR-mediated telomere synthesis (Cho et al., 2014). Depletion of HR (NBS1,

SMC5, BRCA2 and RAD51) or NHEJ factors (53BP1) shows that ALT telomere movement and telomere clustering are facilitated by the HR machinery, independently of NHEJ promoting protein 53BP1. Conversely, 53BP1 is essential for the mobility of deprotected telomeres in facilitating NHEJ in telomerase-positive cells (Dimitrova et al., 2008). The induction of recombination and ALT-like phenotypes in MEFs and telomerase positive cells after induced DSBs and chromatin damage at telomeres respectively, support the speculation that a constant source of telomeric damage may

be essential to drive HR-directed repair in ALT (O'Sullivan et al., 2014, Doksani and

de Lange, 2016). ALT cells may have evolved mechanisms solely favouring HR over

NHEJ, as NHEJ will not be a viable option for telomere maintenance without

telomerase. Therefore, the processing of dysfunctional telomeres at ALT telomeres is

mechanistically distinct from naturally shortened telomeres and telomerase positive

telomeres, highlighting a dependence on HR and possible active repression of c-/alt-

NHEJ in ALT cells. In line with that, there is a prominent accumulation of HR proteins at ALT telomeres.

ALT telomeres have significantly elevated replication stress

As eluded to previously, telomere damage may be beneficial in promoting ALT.

This is supported by the fact that ALT telomeres possess ongoing and spontaneous

DNA damage. Recent studies have shed light on how different DNA repair pathways drive telomere synthesis at ALT telomeres in response to DNA damage. The main source of DNA damage at ALT telomeres arises from elevated replication stress.

Replication stress results from any condition that transiently slows down DNA synthesis and/or stalls replication forks, and is characteristic of precancerous and cancerous cells (Gaillard et al., 2015). In comparison to normal telomeres that have heightened replication stress as discussed in the earlier section, replication stress at

ALT telomeres is even more substantial. This is thought to be the result of an altered heterochromatic state, and structural deficiencies associated with ALT telomeres that create several sources of replication stress.

1.5.5.1 Altered chromatin landscape

The chromatin of ALT telomeres deviates from typical telomeric chromatin.

Normally, telomeric DNA is tightly arranged into nucleosomes with a linker DNA segment of ~160 bp, approximately 40 bp shorter than in bulk chromatin (Tommerup et al., 1994). Enrichment of heterochromatic marks H3K9me3, H4K20me3, HP1 and histone hypoacetylation are characteristic of the compacted telomeric chromatin

(Schoeftner and Blasco, 2009). At ALT telomeres, the heterochromatinisation and chromatin compaction become deregulated. A direct comparison between telomeric chromatin of ALT and telomerase-positive cells revealed that ALT telomeres display a more open and accessible chromatin structure, with decreased nucleosome density

27 and reduced levels of the heterochromatic mark H3K9me3, combined with the sub- telomeric region of ALT telomeres being hypomethylated (Figure 1.6 a) (Episkopou et al., 2014, Lovejoy et al., 2012). These features correlate with increased transcription of telomeric DNA generating an abnormally high level of long noncoding telomeric repeat-containing RNAs (TERRA) (Figure 1.6 a) (Lovejoy et al., 2012, Arora et al.,

2014). Increased TERRA induces a high incidence of DNA-RNA hybrids, also known as R-loops, which originate from TERRA invading duplex telomeric DNA, and contribute significantly to the elevated replication stress at ALT telomeres (Arora et al.,

2014). It has been demonstrated in a recent study that the accumulation of R-loops after the depletion of RNaseH1, an endonuclease that removes DNA-RNA hybrids, led to increased replication stress in ALT cells and increased ALT activity, marked by elevated APBs and C-circles (Arora et al., 2014).

1.5.5.2 Interspersed variant repeats and nuclear receptor occupancy

ALT telomeres contain a considerable proportion of different variant telomeric repeats amidst the canonical TTAGGG repeat arrays (Figure 1.6 b), which may be a consequence of HR-directed copying and propagation of the proximal telomeric regions that contain many variant repeats, or mutations incorporated by the low-fidelity polymerase η (Conomos et al., 2012, Lee et al., 2014, Garcia-Exposito et al., 2016).

Replication stress at ALT telomeres is exacerbated by the occurrence of noncanonical variant repeats interspersed throughout ALT telomeres (Conomos et al.,

2012, Lee et al., 2014). The effects of these variant repeats are two-fold. First, they cause displacement of TRF1/TRF2, due to reduced binding affinity of TRF2 and TRF1 to non-canonical repeats (Conomos et al., 2012). Reduced shelterin loading at ALT telomeres leads to deprotected ends and an elevated DDR (Cesare et al., 2009). Loss of TRF1 also induces telomere fragility due to replication fork impediment (Sfeir et al.,

2009, Martinez et al., 2009). Second, variant repeats, especially TCAGGG create high affinity binding motifs for a group of NR2C/F orphan nuclear receptors including TR2,

TR4, COUP-TF I, COUP-TF II and EAR2, found specifically at ALT telomeres (Figure

1.6 b) (Conomos et al., 2012, Dejardin and Kingston, 2009). These nuclear receptors drive anomalous interactions between telomeres and genomic regions resulting in interstitial telomeric insertions (Marzec et al., 2015). These telomeric insertions potentially create more common fragile sites in the genome, which have been proposed as a significant contributing factor for extensive genome instability in ALT cancers (Marzec et al., 2015). Nuclear receptors are transcription factors; the effects of their recruitment on transcriptional activity at ALT telomeres and whether they in turn attract transcriptional machinery to ALT telomeres needs to be investigated, as this may present another source of replication stress. Another study by our lab found that the orphan nuclear receptors TR4 (NR2C2) and COUP-TF II (NR2F2) recruit the zinc finger protein ZNF827 to ALT telomeres, which in turn directly interacts with the nucleosome remodelling and histone deacetylase (NuRD) chromatin remodelling complex (Figure 1.6 b). ZNF827-NuRD further displaces shelterin binding, promotes telomere-telomere interactions, and increases replication stress and telomere synthesis (Conomos et al., 2014).

Figure 1.6 Characteristics of telomeric chromatin and DNA at ALT telomeres. a. ALT telomeres display an open and relaxed chromatin landscape with reduced heterochromatic marks, decreased H3K9me3 and reduced nucleosome density. The open chromatin may be attributed to the increased transcription of TERRAs. b. ALT telomeres contain a high proportion of interspersed variant repeats e.g. C-type TCAGGG repeats. These variant repeats create high affinity binding sites for orphan nuclear receptors such as COUP-TF II (NR2F2) which recruit ZNF827-NuRD. ZNF827-NuRD contributes to the desaturation of TRF2 at ALT telomeres. This figure was adapted from (O'Sullivan and Almouzni, 2014).

1.5.5.3 Loss of ATRX chromatin remodeller

Mutations in the ATRX/DAXX (death-domain-associated protein) chromatin

remodelling complex and histone H3.3 that correlate strongly with ALT cancers, cause

defects in nucleosome assembly, which can intensify replication stress at telomeres

(Heaphy et al., 2011a, Lovejoy et al., 2012). In addition, in vitro studies demonstrate

that ATRX can directly unwind G-quadruplexes (Law et al., 2010), and the absence of

ATRX may impair unwinding of secondary structures during replication. Direct

evidence for ATRX as an ALT repressor is provided by recent studies that have shown

suppression of ALT activity measured by decreased telomere length, decreased APB

formation, decreased T-SCEs and reduced C-circle levels by ATRX re-introduction

(Napier et al., 2015, Clynes et al., 2015). ATRX-deficient ALT cells are characterised

by elevated levels of RPA at telomeres, indicative of stalled replication forks (Wong et

al., 2010, Flynn et al., 2015). ALT repression by ATRX corresponds with a decrease

in replication forks and increased deposition of histone H3.3, which suggests that

ATRX represses ALT by helping resolve replication stress and rectifying nucleosome

assembly (Clynes et al., 2015). Further evidence implicating replication stress in

promoting ALT comes from a recent study that demonstrates depletion of the histone

chaperone ASF1 (anti-silencing factor 1), which is essential for histone incorporation

30 into chromatin during DNA replication, leads to the induction of ALT phenotypes in telomerase-positive cells (O'Sullivan et al., 2014). The ALT phenotypes, specifically

C-circles and APBs, are drastically decreased by ATR inhibitors, suggesting ATR activation after ASF depletion and replication stress sensing by ATR are required for the induction of ALT.

These studies suggest that altered chromatin dynamics, loss of ATRX, and structural abnormalities including elevated R-loops, desaturation of shelterin and the presence of variant repeats at ALT telomeres, all contribute to enhanced replication stress and the promotion of recombination at ALT telomeres. Accordingly, replication stress is paradoxically advantageous in ALT and any event that reduces replication stress may repress ALT.

1.6 The ALT proteome – interplay of proteins promoting and repressing HR

The HR-mediated ALT mechanism is most likely to be analogous to the process of HR-directed DNA repair, which fundamentally involves 5′ end resection, invasion of the 3’ overhang into a DNA template forming HR intermediates, extension by template copying, dissolution of the recombination intermediates and possibly lagging strand fill-in (Pickett and Reddel, 2015). It is well established that ALT telomeres are enriched with HR proteins, some of which are shown to be essential for ALT activity, such as the MRN complex and SMC5-SMC6 (Cesare and Reddel, 2010). However, the mechanistic details of the interaction dynamics of these proteins in contributing to HR- mediated telomere maintenance are not fully defined. Recent studies have advanced our understanding in the molecular mechanisms of some of the HR proteins, as well as identified new ALT-associated proteins in the sequential steps of HR-mediated telomere synthesis. ALT-mediated telomere synthesis entails an intricate interplay of

DNA damage and repair pathways facilitated by a myriad of proteins (Sobinoff and

Pickett, 2017).

Sequential steps in ALT-mediated telomere synthesis In HR-directed repair, the MRN complex is responsible for the detection of

DSBs and the recruitment of ATM kinase to initiate 5′ to 3′ end resection (Lee and

Paull, 2007). The MRN complex has been shown to be crucial for telomere elongation in ALT as its depletion led to decreased overall telomere length (Jiang et al., 2005,

Zhong et al., 2007). Thus, it seems reasonable to speculate that MRN functions to

detect telomeric DSBs and recruit ATM to ALT telomeres. However, a recent study

reveals PCNA (proliferating cell nuclear antigen) loading by RFC (replication factor C)

as the initial DNA damage sensor, preceding MRE11 post DSB induction at ALT

telomeres (Dilley et al., 2016). 5′-3′ end resection is the critical step to prime the DSB

for HR-mediated repair, and to simultaneously inhibit NHEJ (Figure 1.5) (Ceccaldi et

al., 2016). The nucleases required for end resection to generate an extended telomeric

3’ overhang necessary for strand invasion have not been clearly defined. The RecQ

helicases BLM and WRN have been shown to influence recombination at ALT

telomeres, possibly through promoting 5′ to 3′ end resection by recruiting the

exonucleases EXO1 and DNA2 (Mendez-Bermudez et al., 2012, Sturzenegger et al.,

2014, O'Sullivan et al., 2014, Nimonkar et al., 2011).

The 3′ telomeric overhang is usually protected by POT1, but to facilitate strand

invasion, POT1 must be first replaced by ssDNA binding protein RPA. A recent study

has shed light on how this process occurs. The dynamics between TERRA and

heterogeneous nuclear ribonucleoprotein A1 (hnRNPA1) have been demonstrated to

orchestrate a cell-cycle dependent switch between RPA and POT1 binding at ss

telomeric DNA (Flynn et al., 2011). RPA is subsequently displaced to facilitate BRCA2-

mediated RAD51 loading onto ss DNA forming a RAD51 nucleofilament, which

initiates homology-directed strand invasion and HR (Sung and Klein, 2006). It has

recently been reported that TRF1-Fok1 induced telomeric DSBs at ALT telomeres

trigger a RAD51-dependent homology searching mechanism in which telomeres

undergo long-range directional movements to reach distant telomeres resulting in

telomere-telomere clustering, increased C-circles and telomere synthesis (Cho et al.,

2014). Depletion of RAD51 abrogates telomere clustering and long-range movement, indicating the long-range homology search is RAD51 dependent. The acute telomere

damage induces elevated levels of ss telomeric DNA, in concurrence with RPA and

RAD51 colocalisations with damaged telomeres. The HOP2-MND1 heterodimer,

which is important for chromosome pairing in meiotic recombination, is found to be essential for RAD51-mediated recombination between non-sister chromatids at ALT telomeres (Figure 1.7 a). Another recent study utilising the same TRF1-Fok1 system

shows that HOP2 recruitment to telomeres after acute damage is dependent on ATR-

CHK1 signalling (Figure 1.7 a) (Dilley et al., 2016). The same study found that telomere breaks induce long tract telomeric DNA synthesis ranging from 5 to 70 kb and this process can occur via a RAD51-HOP2 and ATR-independent mechanism devoid of long-range homology searches (Dilley et al., 2016). Depletion of RAD51 and

HOP2 increases nascent telomere synthesis, hinting at a RAD51-independent break- induced replication mechanism at ALT telomeres that possibly entails intratelomeric recombination or uses ECTRs as homology templates (Figure 1.7 a). Together, these studies have illustrated the potential existence of two pathways for ALT-associated telomere synthesis – a RAD51-dependent mechanism and a RAD51-independent mechanism.

In yeast, the ALT pathway has been well understood to engage break-induced

replication (BIR) – a repair pathway for one-ended DSBs and collapsed DNA

replication forks, that requires Pol32, a subunit of DNA polymerase δ for telomere

extension (Bosco and Haber, 1998, McEachern and Haber, 2006, Lydeard et al.,

2007). Recent studies have confirmed that the HR-mediated DNA repair mechanism

driving telomere extension in human ALT cancer cells is analogous to BIR with some

mechanistic distinctions (Dilley et al., 2016, Roumelioti et al., 2016). This mechanism

underlying ALT telomere maintenance is defined as break-induced telomere synthesis

(BITS). BIR in human cells is reliant on two subunits of polymerase δ, POLD3, the

human ortholog of yeast Pol32, and POLD4 (Costantino et al., 2014). Similarly,

polymerase δ is necessary for BITS in ALT cancer cells, with this function being

dependent on the POLD3 subunit, evident by depletion of POLD3 significantly reducing telomere synthesis and telomere length (Dilley et al., 2016). POLD3 directly

interacts with PCNA, a DNA clamp that acts as a processivity factor of polymerases

(Strzalka and Ziemienowicz, 2011), to recruit polymerase δ to damaged ALT

telomeres (Figure 1.7 b). Telomere colocalisations of POLD3-PCNA require the clamp

loader RFC1, which loads PCNA onto DNA. The study also reports that PCNA and

RFC1 act as the initial sensor of DNA damage at telomeres, preceding ATM and ATR

(Dilley et al., 2016). Even though polymerase α-primase and polymerase ε have also

been implicated in BIR in yeast (Lydeard et al., 2010), their involvement in BITS in

human ALT cancer cells has been excluded, as were the translesion synthesis

polymerases, including polymerase ζ, of which POLD3 is a component, and

polymerase η (Dilley et al., 2016). These findings reflect the assembly of a noncanonical replisome that drives BITS at ALT telomeres. While polymerase η is not essential for BITS, another study has proposed that in the context of restarting stalled

replication forks, it may play a role in initiating telomere synthesis after strand invasion,

which is then taken over by polymerase δ to facilitate long tract synthesis (Figure 1.7

b) (Garcia-Exposito et al., 2016).

Processing of recombination intermediates after telomere extension is critical

in averting over-extension and ensuring the proper completion of telomere synthesis.

Recombination intermediates in BITS, like BIR, adopt a D-loop structure. Through D-

loop metabolism, ALT-mediated telomere extension can be regulated. D-loops are processed either by resolution, producing equal numbers of crossover and non-

crossover products, or by dissolution, resulting in only non-crossover products

(Sobinoff and Pickett, 2017). Resolution of D-loops at ALT telomeres is facilitated by the molecular scaffold protein SLX4, which is recruited to telomeres by TRF2, and in turn recruits the endonucleases SLX1 and MUS81-EME1 (Wan et al., 2013, Wilson et al., 2013). BLM and TOP3A, core components of the BLM-TOP3A-RMI1-RMI2 (BTR) dissolution complex, are also recruited by TRF2 to telomeres, and have been found to

be essential for telomere maintenance (Temime-Smaali et al., 2008). T-SCEs are crossover products of telomere recombination, indicative of resolution events at the telomeres. It has been reported that T-SCEs were substantially decreased by SLX4 depletion, while BLM depletion caused a significant increase, indicative of a balancing act between dissolution and resolution by BLM and SLX4 (Sarkar et al., 2015). The opposing roles of BLM and SLX4 at telomeres have been further explored in ALT cancer cells in a recent study. BLM was found to promote POLD3 colocalisations at

ALT telomeres and induce POLD3-dependent long tract telomere synthesis, while the converse was observed for SLX4 (Sobinoff et al., 2017). Furthermore, BLM and SLX4 exerted opposite effects on intertelomeric recombination, with BLM overexpression increasing intertelomeric tag copying. These findings collectively suggest that SLX4

hinders ALT-mediated telomere extension by D-loop resolution during BITS, while

BLM promotes long tract telomere extension and post-extension dissolution of D-loops

(Figure 1.7 c, d). The molecular details of the final lagging strand fill-in step and the polymerase accountable remain enigmatic.

Figure 1.7 ALT-mediated telomere synthesis. a. An extended 3′ overhang resulting from end resection coated by RPA activates the ATR-CHK1 signalling pathway to facilitate RAD51/HOP2-MND1-mediated homology-directed strand invasion and intertelomeric recombination. The resected 3′ overhang can also invade itself through intratelomeric recombination by a mechanism yet to be defined. b. Telomere extension following strand invasion occurs via a mechanism termed break-induced telomere synthesis (BITS) orchestrated by a noncanonical replisome consisting of Pol δ-PCNA- RFC1. BITS can also be initiated by Pol η, then taken over by Pol δ for long tract telomere synthesis. c. BITS is inhibited by SLX4-mediated resolvases, which resolves the recombination intermediates prematurely, resulting in telomere exchange events without telomere extension. d. The BLM-TOP3A-RMI1/2 complex promotes branch

migration and telomere synthesis followed by dissolution with no exchange of telomeric DNA. This figure was adapted from (Sobinoff and Pickett, 2017).

Factors that resolve replication stress repress ALT

Given that replication stress is a common source of DNA damage at ALT

telomeres that potentially drives HR-mediated telomere synthesis, it is possible that reducing replication stress may repress ALT, or vice versa. This proposition is supported by several lines of evidence. Depletion of FANCM (Fanconi anaemia protein M), the stalled fork remodeler, leads to activated ATR signalling shown by increased phosphorylated CHK1 and RPA at ALT telomeres, suggestive of increased replication fork stalling, accompanied by the recruitment of the BLM helicase and the

E3 ubiquitin ligase BRCA1 (Pan et al., 2017). As discussed earlier, BLM plays essential roles in facilitating HR-mediated telomere synthesis. BRCA1 has previously

been reported to localise at ALT telomeres as well as to facilitate the unwinding of

telomeric substrates by BLM (Conomos et al., 2014, Acharya et al., 2014). Therefore,

it is likely that increased stalled replication forks induced by FANCM depletion lead to

recruitment of HR proteins to promote recombination at ALT telomeres. In other words,

FANCM may function as an ALT repressor.

FANCD2, another Fanconi anaemia protein involved in stalled fork remodelling,

has been demonstrated to repress telomere recombination in ALT cells (Root et al.,

2016). Loss of FANCD2 leads to a drastic increase in ECTRs, thought to be by-

products of recombination, and the increase in ECTRs is suppressed by BLM

depletion (Root et al., 2016). It is proposed in the study that FANCD2 acts in opposition

to BLM-mediated telomere recombination by promoting resolution of stalled forks

(Root et al., 2016). SMARCAL1, a DNA annealing helicase that remodels stalled

replication forks to promote fork restart, is enriched at ALT telomeres (Cox et al., 2016).

Loss of SMARCAL1 results in irreversibly stalled replication forks, which eventually

collapse into DSBs, subsequently recruiting RPA and RAD51 to the damaged

telomeres. This was accompanied by an increase in C-circles, indicative of increased

ALT activity (Cox et al., 2016). DSBs at ALT telomeres have been shown to facilitate

RAD51-dependent telomere clustering and consequently homology-directed DNA repair to salvage collapsed replication forks (Cox et al., 2016, Cho et al., 2014). Taken together, these studies suggest that FANCM, FANCD2 and SMARCAL1 are recruited

to ALT telomeres by elevated replication stress, and may function as ALT repressors

by promoting restart of stalled replication forks to warrant faithful telomere replication, thus supressing break-induced HR-mediated telomere synthesis.

Further evidence is provided by a recent study has found enrichment of the

translesion DNA synthesis (TLS) proteins FANCJ, RAD18 and polymerase η at ALT

telomeres (Garcia-Exposito et al., 2016). TLS is a DNA damage tolerance mechanism

that prevents disruption to DNA replication by bypassing DNA lesions, facilitated by

specialised DNA polymerases capable of accommodating damaged DNA such as

polymerase η (Sale, 2013). The study has implicated these TLS proteins in alleviating

replication stress at ALT telomeres. Specifically, depletion of polymerase η

significantly enhances aphidicolin-induced telomere fragility, leads to increased ALT

hallmarks evidenced by increased APBs, C-circles and T-SCEs, and stimulates mitotic

DNA synthesis at ALT telomeres (Garcia-Exposito et al., 2016). These results again

support the notion that removal of replication stress, by polymerase η in this case, may

suppress ALT activity. Paradoxically, increased ALT hallmarks coincide with increased

senescence in ALT cells. This may suggest a mechanism in place at ALT telomeres to modulate the levels of replication stress required for HR-mediated telomere

38 synthesis, whereby excessive replication stress may become detrimental by exhausting the repair pathways.

1.7 The expanded ALT proteome may provide more potential ALT-specific molecular targets

There is a plethora of proteins found at ALT telomeres. This includes a large group of proteins involved in various DNA damage and repair pathways. Many of them have well characterised genomic functions, which have been inferred and proved or redefined experimentally at ALT telomeres. As discussed above, increased knowledge of the functions of these proteins at ALT telomeres has considerably advanced our understanding of the HR-mediated ALT mechanism. Nonetheless, there are several proteins with genomically defined functions and implications at ALT telomeres that lack molecular and mechanistic details, warranting further investigation.

Recently, the ALT proteome has expanded drastically due to advances in protein labelling, purification and detection techniques, specifically PICh (proteomics of isolated chromatin segments), BioID (proximity-dependent biotin identification), and

C-BERST (dCas9–APEX2 biotinylation at genomic elements by restricted spatial tagging), which have enabled endeavours to profile the protein composition at ALT telomeres (Dejardin and Kingston, 2009, Garcia-Exposito et al., 2016, Gao et al.,

2018b). PICh involves isolation of a targeted region of DNA with its associated proteins using a specific nucleic acid probe, followed by protein identification by mass spectrometry (MS) (Dejardin and Kingston, 2009). BioID exploits the promiscuous mutant biotin ligase BirA fused to a protein of interest that, when ectopically expressed in cells, triggers biotinylation of endogenous proteins in proximity upon addition of biotin, followed by purification of biotinylated proteins and identification using MS

(Roux et al., 2013). C-BERST is another MS-based technique that enables efficient

mapping of the proteome associated with a specific genomic loci by combining

dSpyCas9, a nuclease-dead Cas9 that can be targeted to any genomic region, with

APEX2, an engineered ascorbate peroxidase that allows rapid biotinylation of proteins

near the defined genomic loci (Gao et al., 2018b). Telomere PICh, TRF1-BioID and C-

BERST have led to the identification of a large group of novel ALT-associated proteins

that can largely be divided into two types - proteins with defined functions whose roles

in ALT are unknown, and proteins with little or no known functions hitherto (Dejardin and Kingston, 2009, Garcia-Exposito et al., 2016, Gao et al., 2018b). Most ALT-

associated proteins with defined functions fall into the categories of chromatin

remodelling, DNA replication, transcription, various repair pathways including HR,

mismatch repair, single strand annealing, and TLS.

On the other hand, proteins whose functions have not been identified besides

their associations at ALT telomeres present great opportunities for novel protein

discoveries, particularly for identifying new players in DNA damage and repair given

the dominance of the DDR at ALT telomeres. Furthermore, they represent potential

ALT-specific molecular targets for novel therapeutics targeting ALT cancers.

1.8 ZNF827 in ALT-mediated telomere synthesis

The zinc finger protein ZNF827 was an uncharacterised protein on the list of

ALT-specific proteins pulled down in the telomere-PICh study (Dejardin and Kingston,

2009). The presence of ZNF827 at ALT telomeres was reinforced in both the TRF1-

BioID and the C-BERST studies (Garcia-Exposito et al., 2016, Gao et al., 2018b). A previous study by our lab has shed some light on the function of ZNF827 in ALT cancer cells, providing evidence to support it as a molecular target for anticancer therapies

(Conomos et al., 2014). Evidence supports a role for ZNF827 in establishing the altered chromatin dynamics, changed telomeric nucleoprotein structure and telomere-

telomere interactions, to promote HR-mediated telomere synthesis (Conomos et al.,

2014); however, the precise molecular function of ZNF827 is currently not defined.

Nuclear receptors COUP-TF II (NR2F2) and TR4 (NR2C2) bound to variant

repeats interspersed within ALT telomeres recruit the zinc finger protein ZNF827,

which subsequently, through its N-terminal RRK motif, recruits the NuRD chromatin

remodelling complex to ALT telomeres (Conomos et al., 2014). ZNF827 in concert

with NuRD exacerbates shelterin desaturation, evidenced by a significant decrease in

TRF2 loading at ALT telomeres following ZNF827 overexpression, which

unexpectedly concurs with a reduction in spontaneous DDR signalling (Figure 1.8 a)

(Conomos et al., 2014). One plausible explanation is that the NuRD complex,

containing the ATP-dependent chromodomain helicase and histone deacetylase

subunits, facilitates hypoacetylation at ALT telomeric chromatin, resulting in chromatin

compaction that may be compensatory to the shelterin desaturation effect (Figure 1.8

b) (Conomos et al., 2014).

Furthermore, ZNF827 appears to be a positive regulator of ALT by promoting

telomere synthesis (Figure 1.8). Specifically, ZNF827 overexpression increases ALT

hallmarks, including APBs, C-circles and T-SCEs, which are reduced by ZNF827

depletion (Conomos et al., 2014). The spontaneous DDR at ALT telomeres measured

by γH2AX telomere-ChIP is significantly reduced following overexpression of ZNF827.

The ZNF827-NuRD interaction enhances telomere-telomere interactions in the form of telomere bridges, recruits the HR proteins BRIT1 and BRCA1 and increases telomere synthesis measured by BrdU incorporation, creating an HR-permissive environment for ALT-mediated telomere synthesis (Figure 1.8 c, d, e) (Conomos et al.,

2014). Nonetheless, the effect of ZNF827 on telomere length is not known. It is interesting to note that increased ALT activity coincides with a reduction in telomeric

DNA damage following ZNF827 overexpression. This is in direct contrast to the observations in other studies (Cox et al., 2016, Garcia-Exposito et al., 2016, O'Sullivan et al., 2014, Dilley et al., 2016), which have shown that increased ALT phenotypes and telomere synthesis are necessitated by enhanced DNA damage. Collectively these findings may reflect that the role of ZNF827 is mainly on the repair end of the

ALT damage and repair equilibrium.

Figure 1.8 The multifaceted role of ZNF827-NuRD at ALT telomeres (Conomos et al., 2014). ZNF827-NuRD recruitment is dependent on nuclear receptors TR4 and COUP-TF II. a. ZNF827 in concert with NuRD triggers displacement of the shelterin component TRF2. b. The histone deacetylase subunit in NuRD facilitates histone hypoacetylation and chromatin compaction, thought to be compensatory to TRF2 displacement. c. ZNF827-NuRD recruits HR proteins BRIT1 and BRCA1 to ALT telomeres. d. ZNF827-NuRD facilitates telomere-telomere interactions. e. ZNF827- NuRD creates a molecular scaffold to promote ALT-mediated telomere synthesis.

1.9 ZNF827 is a promising therapeutic target in ALT cancers

Given that ALT is an HR-dependent mechanism, targeting proteins involved in

HR-mediated telomere synthesis may seem to be the most direct strategy to treat ALT cancers. Most HR proteins that associate with ALT telomeres also have critical functions in HR-mediated DNA repair in normal cells to maintain genome integrity, rendering them unsuitable as therapeutic targets. The majority of other well- characterised ALT associated factors are also essential components in normal cellular processes. Thus far, therapeutic targets for ALT cancers are lacking. Successful anti-

ALT therapeutic strategies are likely to arise from targeting a deficiency specific to the

ALT pathway, specifically proteins with an essential role in ALT but not in normal cells, or a structural defect at ALT telomeres. ZNF827 seems to fulfil these considerations.

No other reports of ZNF827 function have been published to date. ZNF827 plays a multifaceted role in promoting HR-mediated telomere synthesis in ALT cells

(Conomos et al., 2014), while it is likely dispensable in normal cells. ZNF827 potentially worsens the structural defects in ALT telomeres by TRF2 displacement and chromatin remodelling. Importantly, there is evidence that ZNF827 is required for ALT cell viability. Specifically, ZNF827 depletion causes elevated levels of senescence and apoptosis in ALT cell lines (Conomos et al., 2014). Therefore, ZNF827 represents a promising therapeutic target in the subset of cancers utilising ALT as their TMM. The recruitment of HR proteins, increased telomere-telomere interactions and increased telomere synthesis all suggest ZNF827 functions in the HR repair pathway. This leads to the speculation that inhibition of HR by targeting ZNF827 may induce synthetic lethality with PARP inhibitors or replication stress inducing agents such as hydroxyurea, topoisomerase inhibitors and G-quadruplex stabilising ligands.

1.10 Probable roles of ZNF827 beyond ALT telomeres

The extrapolation of ZNF827 having a wider genomic role is based on its interaction with the chromatin remodeler NuRD and the involvement in HR, a universal

DNA repair pathway. However, the scarcity of literature and reports about ZNF827 in relation to disease may suggest its physiological role is redundant. A recent study has found that NR2C/F nuclear receptors specifically found at ALT telomeres tether telomeres to regular NR2C/F binding sites within genomic DNA and drive insertions of interstitial telomeric repeats in the genome, creating common fragile sites that may contribute to persistent genome instability in ALT cancers (Marzec et al., 2015). Even though the study failed to identify the mediator of these nuclear-receptor-mediated telomeric insertions, it is likely to be ZNF827-NuRD, as their recruitment has been shown in a study by our lab to be dependent on the NR2C/F nuclear receptors

(Conomos et al., 2014).

NuRD is a multi-subunit ATP-dependent chromatin remodelling complex that can interact with a myriad of transcription factors bringing it to different genomic locations to facilitate its canonical functions in transcriptional regulation (Lai and Wade,

2011). Numerous zinc finger proteins act as transcription factors for the NuRD complex to mediate normal developmental processes. For instance, BCL6, FOG1,

Ikaros and BCL11B recruit NuRD by direct binding, which subsequently facilitates transcriptional regulation of multiple lineage-specific genes during haematopoiesis

(Lai and Wade, 2011). By analogy, ZNF827 is predicted to be a transcription factor.

Taken together, it is reasonable to infer that ZNF827 can associate with genomic regions. It remains to be determined whether ZNF827 binds to genomic regions and whether recruitment of NuRD by ZNF827 affects the transcriptional regulation of specific genes that may or may not be related to ALT.

Undoubtedly ZNF827-NuRD partakes in mediating HR in the context of ALT

telomeres, which is a unique scenario of DNA damage and HR-mediated repair

ultimately driving telomere maintenance. Consistent with this, ZNF827 has been found

to be enriched in the same spatial interaction network with 53BP1 and BRCA1 by a proteomic study that has mapped the protein networks of endogenously-tagged

53BP1 and BRCA1 using APEX-based proximity labelling and mass spectrometry

(Gupta et al., 2018). Furthermore, emerging evidence has implicated NuRD in the

DNA damage and repair response. Multiple studies have shown that NuRD is recruited to DSBs to promote repair by recruiting HR proteins including BRCA1 and facilitating transcriptional repression (Polo et al., 2010, Chou et al., 2010, Smeenk et al., 2010,

Larsen, 2010). This circumstantial evidence supports the hypothesis that ZNF827 is a

novel player in the DNA damage response and repair pathway, recruited to ALT telomeres by the associated intrinsic DNA damage.

Recently, ZNF827 gene deletions and mutations have been associated with

developmental delay, Sotos syndrome, and Hirschsprung’s disease in two case

reports (Vlaikou et al., 2014, Sio et al., 2017). Both Sotos syndrome and

Hirschsprung’s disease are congenital disorders. Sotos syndrome is an overgrowth

disorder characterised by distinctive facial features, excessive bone growth,

developmental delay and intellectual disability (Sio et al., 2017). Hirschsprung’s

disease is a developmental disorder that causes intestinal obstruction or chronic

constipation due to the absence of the enteric ganglia along the intestine (Amiel et al.,

2008). Additionally, ZNF827 has been identified as a genetic factor that can influence

concentrations of plasma liver enzymes in two GWAS studies, whereas a ZNF827

single nucleotide polymorphism (SNP) has been linked to increased risk of Alzheimer’s

disease (Kim et al., 2011, Chambers et al., 2011, Jiang et al., 2015).

1.11 The NuRD-binding motif in ZNF827 is conserved in other zinc finger proteins

The NuRD complex is comprised of six core subunits including CHD3/4,

HDAC1/2, MTA1/2/3, RBBP4/7, MBD2/3 and GATAD2A/B, but the precise composition varies in different cell types to direct the localisation and activity of NuRD in a multitude of cellular roles (Allen et al., 2013, Bowen et al., 2004). Apart from its established role in transcriptional regulation during haematopoietic stem cell development, NuRD is involved in non-transcriptional activities including chromatin assembly, cell cycle progression, the DNA damage and repair response, and DNA replication (Lai and Wade, 2011). These multifaceted roles of NuRD have implicated it in both promoting and suppressing tumorigenesis. The CHD and HDAC proteins are the two catalytic subunits in NuRD conferring its ATP-dependent chromatin remodelling and histone deacetylase activities, respectively, while the remaining subunits are involved in interactions with DNA and other proteins directing the functional specificity of NuRD (Allen et al., 2013).

The transcriptional activity of NuRD is mediated through various transcription factors, most of which are zinc finger proteins. ZNF827 interacts with the NuRD complex through its conserved N-terminal RRKQXXP domain (Conomos et al., 2014).

The NuRD-binding RRK motif is conserved in other zinc finger proteins including

BCL11B, BCL11A, FOG1, FOG2, SALL1, SALL4 and ZNF296 (Cismasiu et al., 2005,

Topark-Ngarm et al., 2006, Hong et al., 2005, Lauberth and Rauchman, 2006, Kloet et al., 2018, Moody et al., 2018, Gao et al., 2013) (Figure 1.9 a). The N-terminal conserved motif in BCL11B has been shown to interact with MTA1/2/3 (Dubuissez et al., 2016, Cismasiu et al., 2005), while the conserved motif in FOG1 interacts with

MTA1 and RBBP4 (Lejon et al., 2011). RBBP4 interacts with the N-terminal motif in

BCL11A (Moody et al., 2018). The N-terminal domains of SALL1 and SALL4 have been demonstrated to be essential for the interaction with NuRD, even though the precise subunit involved has not been identified (Lauberth and Rauchman, 2006, Gao et al., 2013). FOG2 and ZNF296 have been reported to interact with the NuRD

complex (Garnatz et al., 2014, Kloet et al., 2018). These zinc finger proteins share the

same NuRD binding motif with ZNF827, and are transcription factors that recruit NuRD

to facilitate the transcriptional regulation of various developmental processes including haematopoiesis, organ development and embryonic stem cell pluripotency and

differentiation (Table 1). Interestingly, transcription factors that are essential in the

early stages of development are often suppressed later in life, and their dysregulated

expression may contribute to tumorigenesis. For example, SALL4 expression is down-

regulated and silenced in fully differentiated cells (Zhang et al., 2015). Aberrant

reactivation of SALL4 has been strongly correlated to many cancers such as

haematological malignancies, germ cell tumours, various epithelial cancers, breast

cancer, lung cancer and glioma (Tatetsu et al., 2016).

Table 1. Biological functions of RRK-motif containing NuRD binding zinc finger transcription factors Proteins Functions Refs (Avram and BCL11B T cell development, function and survival Califano, 2014) (Liu et al., 2003, BCL11A Lymphopoiesis and haemoglobin switching Sankaran et al., 2009) (Welch et al., FOG1 Erythropoiesis 2004) (Garnatz et al., FOG2 Heart development 2014) (Botzenhart et SALL1 Organogenesis al., 2007) (Tatetsu et al., SALL4 Embryonic stem cells pluripotency and self-renewal 2016) (Kloet et al., ZNF296 Embryonic stem cell differentiation 2018)

Figure 1.9 ZNF827 is a C2H2 zinc finger protein. a. The RRK motif is conserved in other C2H2 zinc finger transcription factors that interact with the NuRD complex. b. Schematic of ZNF827 showing the positions of the RRK motif, and the two zinc finger clusters. c. Amino acid positions and sequences of the nine zinc finger domains in ZNF827. d. Schematic of the site-specific DNA recognition by three classical C2H2 zinc finger domains (adapted from (Fedotova et al., 2017)). A classical C2H2 domain forms a finger-like structure comprised of an α helix and two β strands, in which the two cysteines and two histidines coordinate a zinc ion. Amino acids at positions -1, 2, 3, 6 are involved in site specific DNA-recognition (Wolfe et al., 2000).

1.12 ZNF827 and other C2H2 zinc finger proteins involved in the DDR

ZNF827 is a C2H2 zinc finger protein that contains nine C2H2 zinc finger (ZnF) domains positioned in two clusters, with three tandem repeating units located towards the midpoint and the remaining six proximal to the C terminus of the protein (Uniprot)

(Figure 1.9 b, c). This is the largest class of putative human transcription factors consisting of over 700 C2H2 ZnF proteins (Weirauch and Hughes, 2011, Vaquerizas

et al., 2009). The classical C2H2 ZnF domain, first discovered in transcription factor

IIIA from Xenopus laevis, consists of a 30-amino acid repeat folded into two β strands

and an α helix that tetrahedrally coordinate a zinc ion via the conserved two cysteine

and two histidine residues, forming a finger-like structure (Figure 1.9 d) (Miller et al.,

1985, Lee et al., 1989, Pavletich and Pabo, 1991, Parraga et al., 1988). Each C2H2

ZnF contacts three or more DNA bases, with the sequence specificities conferred

largely by the four amino acid residues at positions -1, 2, 3 and 6 of the DNA-contacting

α helix (Wolfe et al., 2000) (Figure 1.9 d). As transcription factors, C2H2 ZnF proteins are known to display sequence-specific DNA binding ability. Ironically, for the majority of the C2H2 ZnF proteins, it remains to be shown whether they bind DNA or have a

specific DNA binding motif (Emerson and Thomas, 2009, Jolma et al., 2013). C2H2

ZnF domains can also bind RNA, proteins and other ligands such as DNA-RNA

hybrids (Brown, 2005, Brayer and Segal, 2008, Iuchi, 2001).

In addition to transcriptional regulation, C2H2 ZnF proteins play a role in a

range of cellular processes such as development, the DDR, genome integrity and

tumour suppression. In a recent DNA damage recruitment screen, transcription factors

including ZnF proteins have been recognised as a large class of DDR factors. 34 out

of 120 proteins localised to DNA damage sites contained ZnF domains including C2H2

motifs (Izhar et al., 2015). In addition to ZNF827, two other C2H2 ZnF proteins have

been identified at mammalian telomeres to play an important role in telomere length

regulation. First, ZSCAN4 has been shown to be crucial for telomere extension during

early embryogenesis, a process essential for embryonic stem cell self-renewal. A

recent study suggests that ZSCAN4 facilitates telomerase-mediated telomere

elongation as well as telomere recombination by promoting genome-wide DNA

demethylation in mouse embryonic stem cells (Dan et al., 2017). Second, ZBTB48,

49 also known as TZAP, has been found to bind directly to canonical telomeric repeats with 3 of its 11 C2H2 ZnFs. TZAP binds hyper-elongated telomeres with a low density of shelterin to promote telomere trimming, thereby setting an upper limit to telomere length (Li et al., 2017).

Interestingly, studies have found the MYND-type ZnF protein ZMYND8 and the

C2H2 ZnF protein HIC 1 to be involved in DSB repair, in concert with the NuRD complex. ZMYND8 recruits NuRD to DSBs via its MYND domain, which interacts with the GATAD2A subunit (Spruijt et al., 2016). The recruitment of NuRD by ZMYND8 to

DSBs promotes repression of active transcription and DNA repair through the HR pathway. ZMYND8 also binds to three other ZnF proteins including ZNF532, ZNF592 and ZNF687, collectively called the Z3 module which is involved in regulating the

ZMYND8-NuRD interaction at DNA damage sites (Spruijt et al., 2016). HIC1

(hypermethylated in cancer 1) is a tumour suppressor gene frequently silenced in human cancers that has been shown to be essential for etoposide-induced apoptosis

(Chen et al., 2005). A study found that depletion of HIC1 impairs repair of etoposide induced DSBs, which increases HIC1 SUMOylation and its SUMOylation-dependent interaction with the NuRD subunit MTA1 (Dehennaut et al., 2013).

1.13 Characterising ZNF827

This thesis focuses on the functional characterisation of ZNF827. Current evidence has illustrated ZNF827 as a promising molecular target in the ALT pathway, which supports further investigation into ZNF827 in several aspects. To further characterise ZNF827 as a novel molecular target in ALT, we aim to investigate its biological functions as a transcription factor and the differential effects it has on ALT and telomerase cells. To provide the molecular basis for ZNF827 inhibition, it is imperative to delineate the molecular details underlying its ALT-promoting function

50 including ZNF827 recruitment, the protein and DNA interactions involved, its functions in relation to NuRD, and how it promotes HR at ALT telomeres. Further, we aim to explore its potential role as a novel protein involved in DNA damage and repair pathways. The overarching aim is to validate ZNF827 as a novel therapeutic target in cancers utilising the ALT pathway, as well as discovering new functions and expanding knowledge of this poorly understood protein.

Materials and Methods

2.1 Mammalian Cell Culture

Growing cells All cell lines used in this study are immortal, either tumour derived, or

immortalised with the simian virus 40 large T antigen (SV40 LgT) as described in Table

2.1. Cell lines were cultured in Dulbecco’s Modified Eagle Medium (DMEM; D5796,

Sigma-Aldrich) supplemented with 10% foetal calf serum (FCS) at 37°C with 10% CO2.

Cells were grown in monolayers in single use tissue culture flasks (BD Falcon) and

subcultured at confluency using standard cell culture passaging protocols. Briefly,

following media aspiration and one warm PBS wash, cells were dislodged from the

flasks with pre-warmed 0.05% trypsin/EDTA (GIBCO) at 37°C for 5 minutes. At least

four times the volume of warm media was added to deactivate trypsin and

homogenously resuspend the cells before transferring appropriate fractions into fresh

flasks with sufficient fresh DMEM supplemented with 10% FCS. Cells were typically

passaged in a 1:4 to 1:16 split ratio depending on the growth rate of the cell line.

Table 2.1 List of mammalian cell lines used in this study Telomere Method of Cell Line Cellular Origin Maintenance p53 status Immortalisation Mechanism U-2 OS Osteosarcoma Tumour ALT Wildtype Foetal lung WI38-VA13/2RA SV40 LgT ALT SV40 LgT fibroblast HT1080 Fibrosarcoma Tumour Telomerase Wildtype HT1080 6TG Fibrosarcoma Tumour Telomerase Mutated HEK 239T Embryonic kidney SV40 LgT Telomerase SV40 LgT Glioblastoma T98G Tumour Telomerase Mutated multiforme

Cryopreservation and thawing of cell lines

For cryopreservation, cells were pelleted by centrifugation at 200 x g for 5

minutes, resuspended in freezing media (90% FCS, 10% DMSO (dimethyl sulfoxide)),

and aliquoted into cryogenic vials for storage at -80°C. Vials were transferred to liquid

nitrogen for long term storage. To thaw cryopreserved cells, the frozen vial was thawed

in a 37°C water bath, promptly transferred to a 15 mL falcon tube containing 5mL

DMEM supplemented with 10% FCS, then centrifuged at 200 x g for 5 minutes. The

cell pellet was resuspended in fresh DMEM supplemented with 10% FCS, transferred

into a new tissue culture flask, and incubated at 37°C with 10% CO2. Media was

replaced with fresh DMEM supplemented with 10% FCS to remove non-viable cells

after 24 hours.

2.2 Cell Line Manipulations: Transient Overexpression, siRNA Knockdown, and

CRISPR knockout

Generation of plasmid constructs

Details of all the plasmid constructs used in overexpression experiments are

summarised in Table 2.2.

2.2.1.1 ZNF827 full length and mutants

pCMV6 Myc-DDK-tagged ZNF827 full length (RC221405), and pCMV6-Entry

(PS100001) empty vector were purchased from OriGene Technologies. pHTN

HaloTag® CMV-neo empty vector was purchased from Promega. pHTN HaloTag®

CMV-neo ZNF827 was generated by SgfI-NotI (New England BioLabs Inc.) restriction

enzyme digest of the donor pCMV6 Myc-DDK-tagged ZNF827 plasmid to isolate the

ZNF827 insert, followed by gel purification of the ZNF827 insert and ligation into the recipient pHTN HaloTag® CMV-neo empty vector using T4 DNA ligase (New England

BioLabs Inc.). Likewise, pCMV6 Myc-DDK-tagged ZNF827 SUMOylation mutant and pHTN HaloTag® CMV-neo ZNF827 SUMOylation mutant constructs were generated by ligation of the mutant insert isolated by SgfI-NotI digest from a pUC57 ZNF827

SUMOylation mutant construct obtained commercially from Genscript with the

respective vectors. The pCMV6 Myc-DDK-tagged ZNF827 ΔRRK mutant was

generated by a previous lab member using site directed mutagenesis (Agilent

Technologies) (Conomos et al., 2014).

ZNF827 zinc finger mutant plasmid constructs were generated by Genscript, or

restriction enzyme digest and ligation. pCMV6 Myc-DDK-tagged ZNF827 ZnF1-3 deleted and pHTN HaloTag® CMV-neo ZNF827 ZnF1-3 deleted mutant constructs were made commercially by Genscript. To generate pCMV6 Myc-DDK-tagged

ZNF827 ZnF4-9 deleted mutant construct, ZnF4-9 was excised with BsmI and MluI-

HF restriction enzymes (New England BioLabs Inc.) from the full length ZNF827 construct and replaced by a filler fragment, designed with SnapGene and purchased from Integrated DNA Technologies, by ligation using T4 DNA ligase (New England

BioLabs Inc.). Similarly, pCMV6 Myc-DDK-tagged ZNF827 ZnF1-9 deleted mutant construct was generated by excision of ZnF 4-9 from the ZNF827 ZnF1-3 deleted construct and ligation with the filler fragment. All three pHTN HaloTag® CMV-neo

ZNF827 ZnF mutant constructs were made by isolation of the respective mutant insert from the pCMV6 plasmids and ligation with the pHTN HaloTag® CMV-neo vector, as described above.

2.2.1.2 RBBP4 full length and mutants

pCMV6 Myc-DDK-tagged RBBP4 construct (RC208761) was purchased from

OriGene Technologies. pCMV6 Myc-DDK-tagged E126N128E179AAA, pCMV6 Myc-

DDK-tagged Y181A, and pCMV6 Myc-DDK-tagged P43S73AA mutant constructs

were made by In-Fusion cloning using the CloneAmp™ HiFi PCR Premix kit (Clontech

Laboratories). Briefly, all three RBBP4 mutant inserts were amplified from pFastBac

HT B RBBP4 mutant constructs, gifted by Professor Yunyu Shi, with primers designed

in SnapGene, and cloned into the pCMV6 vector according to manufacturer’s protocol.

Primer sequences are forward 5′-GAC GCG TAC GCG GCC ATG GCC GAC AAG

GAA GCA G-3′ and reverse 5′-TTT CTG CTC GCG GCC GGA CCC TTG TCC TTC

TGG ATC C-3′.

2.2.1.3 Plasmid verification

All plasmid constructs were confirmed by Sanger sequencing at the Australian

Genome Research Facility (AGRF).

Plasmid propagation

To propagate plasmid constructs for experiments, plasmids were transformed

in One Shot® Stbl3™ chemically competent E. coli, One Shot® TOP10 chemically competent E. coli (ThermoFisher Scientific) or STELLAR chemically competent E. coli

(Clontech) according to the manufacturers’ instructions. Transformed cells were plated onto agar plates containing the appropriate selection antibiotic, kanamycin 25 µg/mL or ampicillin 100 µg/mL (see Table 2.2), followed by incubation at 37°C for 16-18 hours.

Aseptically, each single colony was selected with a sterile pipette tip, placed into a

mini culture (5mL LB broth with selection antibiotic), and incubated with vigorous

shaking at 37°C for 8 hours. Turbid mini cultures were transferred into maxi cultures

(100 mL LB broth with selection antibiotic), and incubated with vigorous shaking at

37 °C overnight. Plasmid DNA was then extracted from bacteria using the QIAGEN

Plasmid Maxi Kit (Qiagen) according to the manufacturer’s instructions. All plasmid

DNA was stored in 10 mM Tris-Cl, pH 8.0 at -20°C. Glycerol stocks were prepared at

a 1:1 ratio of bacteria to glycerol and stored at -80°C.

Transient expression of plasmid constructs

Overexpression of plasmid constructs was achieved by reverse transfection

with FuGENE-6 transfection reagent (Promega). For a 75 cm2 transfection, 1 mL

55 reduced serum medium Opti-MEM (Gibco) was combined with 60 µL FuGENE-6 transfection reagent and incubated at room temperature for 5 minutes. 20 µg of plasmid DNA was then added into the mixture (3:1 FuGENE:plasmid ratio), vortexed briefly and incubated at room temperature for 15 minutes. Empty vector plasmid was included as a control. During incubation, cells to be transfected were trypsinised, resuspended in DMEM supplemented with 10% FCS, and counted using a Beckman

Coulter cell counter. The prepared transfection mixture was transferred into a 75 cm2 flask and swirled gently to cover the flask. 0.8 - 1x106 cells in 2 mL DMEM supplemented with 10% FCS were directly pipetted onto the transfection mixture in the flask and shaken gently to mix. Additional DMEM supplemented with 10% FCS was added to the flask to a total volume of 15 mL. Cells were incubated at 37°C for 48 hours prior to harvesting. To harvest, transfected cells were washed once with warm

PBS, trypsinised for 5 minutes at 37°C, resuspended in DMEM supplemented with 10%

FCS, and counted using a Beckman Coulter cell counter. Known numbers of cells were collected in pellets by centrifugation at 200 x g at 4°C, flash frozen in liquid nitrogen and stored at -80°C for qRT-PCR and Western Blot validation of overexpression, and subsequent experiments.

Table 2.2 List of plasmid constructs used in the overexpression experiments

Bacterial Mammalian Cell Size Gene/Insert Vector Detection Tag Mutation/Deletions Resistance Selection (kb) Empty Vector 4.9 n/a

ZNF827 full length 8.1 n/a ZNF827 SUMOylation 8.1 K466R, K523R, K597R, K673R, K704R mutant ZNF827 ΔRRK 8.1 R3G, R4G, K5A

ZNF827 ZnF1-3 deleted 7.9 Zinc Fingers 1 to 3 deleted pCMV6 Kanamycin Neomycin (G418) Myc-DDK ZNF827 ZnF4-9 deleted Entry 7.5 Zinc Fingers 4 to 9 deleted ZNF827 ZnF1-9 deleted 6.4 Zinc Fingers 1 to 9 deleted

RBBP4 full length 6.2 n/a

RBBP4 Y181A 6.2 Y181A

RBBP4 P43S73AA 6.2 P43A, S73A

RBBP4 E126N128E179AAA 6.2 E126A, N128A, E179A

Empty Vector 6.1 n/a

ZNF827 full length 9.3 n/a ZNF827 SUMOylation pHTN 9.3 K466R, K523R, K597R, K673R, K704R mutant HaloTag® Ampicillin Neomycin (G418) HaloTag ZNF827 ZnF1-3 deleted CMV-neo 9.1 Zinc Fingers 1 to 3 deleted

ZNF827 ZnF4-9 deleted 8.7 Zinc Fingers 4 to 9 deleted

ZNF827 ZnF1-9 deleted 7.6 Zinc Fingers 1 to 9 deleted Cell cycle chromobody® pCCC- n/a, commercially available from Kanamycin Neomycin (G418) TagRFP 5.1 (PCNA) plasmid TagRFP ChromoTek

Gene knockdown with small interfering RNA

siRNAs used in this study were purchased from ThermoFisher Scientific and

details listed in Table 2.3. Cells were transfected using Lipofectamine RNAiMAX

transfection reagent (ThermoFisher Scientific) with an siRNA concentration of 20 µM.

For transfection in a 75 cm2 flask, 3 mL reduced serum medium Opti-MEM (Gibco) was combined with 15 µL of 20 µM siRNA, mixed gently to cover the surface area of the flask and incubated at room temperature for 5 minutes. 30 µL of Lipofectamine

RNAiMAX transfection reagent was then added to the mixture, and mixed and incubated for a further 15 minutes at room temperature. Meanwhile, cells to be transfected were trypsinised, resuspended in DMEM supplemented with 10% FCS, and counted using a Beckman Coulter cell counter. 0.8 - 1x106 cells in DMEM

supplemented with 10% FCS were directly pipetted onto the transfection mixture in

the flask, and shaken gently to mix. Additional DMEM supplemented with 10% FCS was added to the flask up to a total volume of 15 mL. Cells were harvested 72 hours post-transfection for knockdown validation and subsequent experiments.

Table 2.3 siRNAs used in this study siRNA Details and Catalogue Number

siScrambled Stealth RNAi™ Negative Control Med GC Duplex #2 (12935‐112, Invitrogen)

siZNF827 Stealth RNAi™ siRNA ZNF827HSS135819 Duplex Oligoribonucleotides (Invitrogen) 5’ – GGG CAG UCU UCU GGC UGA GAA AUC A - 3’ 5’ – UGA UUU CUC AGC CAG AAG ACU GCC C – 3’

ZNF827 knockout using Clustered Regularly Interspaced Short Palindromic

Repeats/Caspase 9 (CRISPR/Cas9)

CRISPR/Cas9 genome editing was performed by the Vector and Genome

Engineering Facility (VGEF) at Children’s Medical Research Institute to knockout the

ZNF827 gene in U-2 OS, HT1080 and HEK 293T cells. 70-90% confluent cells were

provided in 6-well plates to the VGEF facility for the CRISPR/Cas9 experiments. The

complete CRISPR procedure, including single guide RNA design, plasmid preparation

and transfection, clone selection by PCR, expansion and confirmation of knockout

clones by pGEM sequencing, was carried out by VGEF. The guide RNA sequence

was GTCTCTGGAGGACCGGATCCAGG (the last three nucleotides are the PAM

sequence) which targets exon 2 of the ZNF827 gene. One complete ZNF827 knockout

clone was obtained from U-2 OS while no viable clones were obtained from HT1080

and HEK 293T.

2.3 DNA Based Experiments

Genomic DNA extraction

Cell pellets of 1 x 106 cells were resuspended in 500 µL lysis buffer (50 mM

Tris–Cl, 100 mM NaCl, 50 mM ethylenediamine tetraacetic acid (EDTA), 0.5% SDS, pH 8). Samples were then digested with 50 µg/mL RNase A for 20 minutes at room temperature followed by digestion with 400 µg/mL Proteinase K overnight at 55°C.

DNA was extracted by adding equal volumes of phenol/chloroform/isoamyl alcohol

(25:24:1, Sigma) to the samples, and mixing by inversion in MaXtract High Density tubes (Qiagen), which separate the aqueous DNA phase after centrifugation at 2000 rpm for 10 minutes at room temperature. This step was completed twice to achieve high extraction efficiency. The top DNA-containing layer was transferred carefully to a fresh tube containing 0.1 x volume of sodium acetate and 2.5 x volume of 100% ethanol and left to precipitate at -20 °C for at least 3 hours. Precipitated DNA was pelleted by centrifugation at 17,000 x g for 60 minutes at 4°C, briefly washed with 70% ethanol, and centrifuged again at 17,000 x g for 30 minutes at 4°C. The DNA pellet was air dried for 30 minutes and dissolved in 10 mM Tris-Cl, pH 8.0 before being

quantitated on the NanoDrop ND-1000 Spectrophotometer (ThermoFisher Scientific).

Genomic DNA was stored at -20°C until analysis.

C-circle assay

For the C-circle assay, genomic DNA was extracted as described above and

diluted to 5 ng/µL with 10 mM Tris-Cl pH 8.0. Each reaction was prepared as follows:

10 μL of DNA was added to 10 μL 200 μg/mL bovine serum albumin (BSA), 0.1%

Tween-20, 1 mM dATP, 1 mM dGTP, 1mM dTTP, 1× Φ29 buffer (New England Biolabs)

and 7.5 U Φ29 DNA polymerase (New England Biolabs). The reactions were

incubated for 8 hours at 30°C allowing for rolling circle amplification, with the reaction

terminated by incubation for 20 minutes at 65°C. The reaction products were diluted

with 100 µL 2 x SSC, and the total volume was dot blotted onto a 2 x SSC-soaked

Biodyne B 0.45 µm nylon membrane (Pall). Membranes were air-dried for 30 minutes

and the DNA was UV-cross-linked onto the membrane with 240 mJ using a

Stratalinker at 254 nm. The membrane was pre-hybridised in PerfectHyb Plus

hybridisation buffer (Sigma-Aldrich) for 1 hour at 37°C and. Following the addition of

32 γ P-(CCCTAA)3 telomeric probe end-labelled using T4 polynucleotide kinase (New

England Biolabs), the membrane was incubated overnight at 37°C. The following day,

the membrane was washed three times in 0.1% SDS and 0.5 × SSC for 15 minutes shaking at room temperature and allowed to air dry. The membrane was then exposed to a PhosphorImager screen overnight and scanned on a Typhoon FLA9500 Imager

(GE Healthcare Life Sciences). The telomere C-circle assay signal was quantified using the array analysis function on the ImageQuant TL software (GE Healthcare Life

Sciences).

2.4 RNA Based Experiments

RNA extraction

Total RNA was extracted from cell pellets using the RNeasy Mini Kit (Qiagen)

with on-column DNAse I digestion (Qiagen) according to manufacturer’s protocol. RNA

was eluted in 30 µL of RNase free water, and stored at -80°C until analysis.

Quantitative reverse transcription PCR (qRT-PCR)

After RNA quantitation using the NanoDrop ND-1000 Spectrophotometer

(ThermoFisher Scientific), 2 µg of total RNA was used for cDNA synthesis using

SuperScript III RT (Invitrogen) according to manufacturer’s protocol. A reverse transcriptase negative control (RT-) was included in each cDNA synthesis. cDNA was digested with RNase H (Invitrogen) for 20 minutes at 37°C, and diluted 1:5 with RNase free water. In the subsequent qRT-PCR reactions, Platinum™ SYBR™ Green qPCR

SuperMix (ThermoFisher Scientific) was used as per the manufacturer’s instructions with 3 µL of cDNA template and 0.5 µM forward and reverse primers. The reactions were loaded into a 96-well PCR plate and run on the LightCycler® 96 (Roche) with

PCR conditions 95˚C for 10 minutes, 40 cycles of 95˚C for 15 seconds, 60˚C for 60 seconds and 72°C for 30 seconds, followed by melt curve analysis. RT- and water

controls were included. Matched PCR efficiencies for the primer sets were confirmed

by standard curve comparison. Analysis was performed with LightCycler® 96 system

software, using the ∆∆Ct method with GAPDH as the reference gene. PCR primers

used in this study are listed in Table 2.4.

Table 2.4 List of qRT-PCR primers Primer Name Primer sequence (5’-3’) GAPDH Forward ACC CAC TCC TCC ACC TTT G GAPDH Reverse CTC TTG TGC TCT TGC TGG G ZNF827 Forward GGC TCA ACT CAG GAC AGT GG ZNF827 Reverse CCG GCA CTT GTA CTC CAT CTT

RBBP4 Forward GAC GCA GTG GAA GAA CGA GT RBBP4 Reverse TCC CCA GGA CAA GTC GAT GA HDAC1 Forward GGA AAT CTA TCG CCC TCA CA HDAC1 Reverse AAC AGG CCA TCG AAT ACT GG MTA1 Forward TGC TCA ACG GGA AGT CCT ACC MTA1 Reverse GGG CAT GTA GAA CAC GTC ACC COUP-TF II Forward TGT TTG TGT TGA ATG CGG CG COUP-TF II Reverse CGC CTT GAG CTT CTC CAC TT TR4 Forward TCA TTG AGG TTG AAG GCC CC TR4 Reverse GAA AGG CTG GGA TTG ACC GA SNAI1 Forward CTC AAG ATG CAC ATC CGA AGC SNAI2 Reverse ATC TGA GTG GGT CTG GAG GT SIRT1 Forward TCG CAA CTA TAC CCA GAA CAT AGA CA SIRT1 Reverse CTG TTG CAA AGG AAC CAT GAC A STING Forward CAC CTG TGT CCT GGA GTA CG STING Reverse AGT GTC CGG CAG AAG AGT TT

TERRA analysis

For TERRA analysis, 2 µg of RNA was redissolved in 200 µL 2 x SSC and dot blotted onto a 2 x SSC-soaked Biodyne B 0.45 µm nylon membrane (Pall) ensuring an RNase free environment. Membranes were air-dried for 30 minutes and the DNA was UV-cross-linked onto the membrane with 240 mJ using a Stratalinker at 254 nm.

The membrane was then treated as described in the telomere-ChIP assay in Section

32 2.8.1, using the DNA oligonucleotide probe 5′-(CCCTAA)3-3′ end labelled with [γ- P]-

ATP to detect the single strand 5′ -(UUAGGG)n-3′ TERRA strand.

2.5 Protein Detection, Purification and Quantitation

Protein extraction using Benzonase®

Equal loading of protein was achieved by protein extraction from equal numbers of cells in each sample. Protein was extracted from 106 cells per sample by vigorous pipetting and incubation at room temperature for 30 minutes in 100 μL of EDTA-free 4 x LDS buffer (106 mM Tris-Cl, 141 mM Tris-Base, 40% Glycerol, 2% LDS, 0.075%

SERVA Blue G50) supplemented with 2.5% Benzonase® nuclease 25 U/μL (Novagen,

Merck Millipore) and 2% beta-mercaptoethanol (Sigma).

Western Blot

Following protein extraction, all samples were denatured at 90°C for 15 minutes before resolving on NuPAGE® Novex® 4-12% Bis-Tris Mini Gels (Invitrogen) with 1 x

105 cells (10 μL) loaded per lane in NuPAGE® MES SDS running buffer (Invitrogen)

at 200 V for 50 minutes. Gels were transferred onto a PVDF membrane (Immobilon-

P, Millipore) in transfer buffer (190 mM glycine, 25 mM Tris and 10% methanol) at 100

V for 60 minutes. Membranes were stained with Ponceau-S for 30 minutes, de-stained

then blocked for 1 hour at room temperature in 5% skim milk dissolved in PBS with

0.1% Tween-20 (PBST) or TBS with 0.1% Tween-20 (TBST) for phosphorylated proteins. Membranes were cut into appropriate segments based on the size markers, with each segment incubated with antibodies for proteins of interest diluted in 0.5% skim milk in PBST/TBST or 5% BSA overnight at 4°C (Refer to Table 2.9 in Section

2.14 for the list of antibodies used in this study). Blots were then washed for 3 x 5 minutes in PBST/TBST followed by incubation for 1 hour at room temperature with appropriate secondary antibodies (See Table 2.10 ) diluted 1:5000 in 0.5% skim milk or 5% BSA in PBST/TBST. Blots were washed for 3 x 5 minutes in PBST/TBST before incubation with SuperSignal West Pico Chemiluminescent Substrate (ThermoFisher

Scientific) for 5 minutes at room temperature. Blots were then exposed to a

Luminescent Image Analyser (LAS-4000, Fujifilm). Quantitation was conducted using

Multigauge software.

HaloTag® protein purification

A list of purified proteins is shown in Table 2.5. Protein purification was

performed in HEK 293T cells overexpressing HaloTag® ZNF827 full length, ZNF827

ZnF1-3 deleled (ZnF1-3 del), ZNF827 ZnF4-9 deleted (ZnF4-9 del), ZNF827 ZnF1-9

deleted (ZnF1-9 del), ZNF827 SUMOylation mutant (SUMO Δ) or empty vector control

using the HaloTag® Protein Detection and Purification System (Promega) based on

the manufacturer’s protocol with minor modifications. Briefly, cell pellets collected in

overexpression experiments, were gradually thawed on ice, resuspended

homogenously in Mammalian Lysis Buffer (Promega) supplemented with 1 x Protease

Inhibitor Cocktail (Promega) to a concentration of 2-6 x 107 cells/mL, and incubated at

4°C for 1 hour on a rotator. Cell lysates were centrifuged at 10,000 x g for 30 minutes

at 4°C to collect the supernatant, which was then diluted 1:3 by adding HaloTag®

Protein Purification Buffer. 180 μL/mL lysis buffer of HaloLink™ Resin slurry was equilibrated by five 5-minute washes in 5 mL of HaloTag® Purification Buffer followed by removal of supernatant after centrifugation at 1500 x g for 5 minutes between washes. Cell lysates were added to the equilibrated resin, mixed well by inverting the tubes, and incubated on an rotator at 4°C overnight. The following day, samples were centrifuged at 1500 x g for 5 minutes. Supernatants were transferred to another tube as flowthrough fraction for binding efficiency analysis as required. The resin bound with proteins was washed with 5 mL HaloTag® Protein Purification Buffer on a rotator for 3 x 10 minutes followed by removal of supernatant after centrifugation at 1500 x g for 5 minutes between washes. For protein elution, HaloTEV Protease cleavage solution (9 μL of HaloTEV Protease in 291 μL HaloTag® Protein Purification Buffer per 900 μL resin slurry) was added to the settled resin, mixed well and incubated on a rotator at 4°C overnight. The following day, the supernatant (eluate) was collected by centrifugation at 1500 x g for 5 minutes. To remove residual resin, the eluate was transferred to a spin column in a 1.5 mL low bind microcentrifuge tube and collected by centrifugation at 10,000 x g for 15 seconds. Eluates were mixed with 10% glycerol to maintain protein stability, aliquoted and stored at -80°C until use.

Table 2.5 A list of proteins purified by HaloTag® Protein Purification Method Protein ZNF827 full length HaloTag® ZNF827 ZnF1-3 deleted (The tag was cleaved during ZNF827 ZnF4-9 deleted elution resulting in untagged ZNF827 ZnF1-9 deleted proteins) ZNF827 SUMOylation mutant Empty vector

Protein quantitation by BCA (bicinchoninic acid) assay

ZNF827 full length and mutant proteins purified by HaloTag® described above

were quantitated by the BCA assay using the Pierce™ BCA Protein Assay Kit according to the manufacturer’s instructions. Briefly, BSA was prepared at a range of concentrations to construct a standard curve. All samples were prepared in duplicate reactions and measured at 562 nm using a VERSAMax microplate reader (Bio- strategy). Sample concentrations were calculated using the standard curve.

Mass spectrometry

ZNF827 pull-downs were performed in WI38-VA13/2RA and HEK 293T cells

overexpressing HaloTag ZNF827 or empty vector control using the HaloTag® Protein

Pull-Down and Labelling System (Promega) according to the manufacturer’s protocol.

Pull-downs were subjected to mass spectrometry analysis to identify potential ZNF827 interacting partners. Sample processing, mass spectromtery and downstream analysis were provided by the Biomedical Proteomics Facility at CMRI.

2.6 Experiments for Protein-Protein Interactions

Co-immunoprecipitation

Cell pellets of 15-20 x 106 cells, collected in overexpression experiments, were

thawed on ice, and resuspended homogenously in 1 mL lysis buffer (20mM HEPES-

KOH pH 7.9, 200 mM NaCl, 2 mM MgCl2, 10% glycerol, 0.1% Triton X-100, 1 mM

dithiothreitol (DTT), 1× complete protease inhibitor (CPI), and 1 mM

phenylmethylsulphonyl fluoride (PMSF)). Resuspended cells were lysed by incubation

with rotation for 1 hour at 4°C. Lysates were subjected to centrifugation at 13,000 rpm

for 40 minutes at 4°C to remove cell debris. The supernatant was then transferred to

a fresh cold low bind 1.5mL tube for subsequent immunoprecipitation. For

immunoprecipitation, protein lysates were incubated with antibody for the protein of

interest (10 μg per mL lysate) or IgG control (Refer to Table 2.9 in Section 2.14 for the

list of antibodies used in this study), preincubated with Dynabeads Protein G, as per

manufacturer’s instructions (Life Technologies), with rotation overnight at 4°C. The

following day, beads were washed three times in cold lysis buffer and separated on a

magnetic rack. Proteins were then eluted from the beads by resuspension in 50 mM

Glycine pH 2.8 and sodium dodecyl sulphate (SDS) sample buffer (0.2 M Tris-Cl, pH

6.8, 28% glycerol, 13.5% β-mercaptoethanol, 6% SDS, 6 mM EGTA, and 0.07%

bromophenol blue) for 10 minutes at 70°C and subjected to Western blot analysis, as

described above, to detect various protein interactions.

Protein colocalisations by immunofluorescence

Refer to Section 2.10.3 for the protein immunofluorescence protocol.

Colocalisations in this study were defined by at least 50% overlap of two protein

immunofluorescence signals. Quantitated analysis was conducted by automation with

CellProfiler (Broad Institute).

Proximity ligation assay (PLA)

Cells at ~70%-80% confluency were pre-extracted with 0.2% Triton X-100 in

PBS for 5 minutes on ice followed by fixation at -20˚C in 70% methanol and 30% ethanol for 30 minutes. Cells were further permeabilised in 1% Triton X-100 in PBS on ice for 10 minutes followed by blocking in antibody dilution buffer (ABDIL - 20 mM Tris-

Cl, pH 7.5, 2% BSA, 0.2% fish gelatine, 150 mM NaCl, 0.1% Triton X-100 and 0.1%

sodium azide) for 1 hour at room temperature. Cells were then incubated with primary

antibodies for ZNF827, ATR and RPA (Refer to Table 2.9 in Section 2.14 for antibody

details). Following primary antibody incubations PLA was performed using the

Duolink® PLA kit as per manufacturer’s instructions.

2.7 Experiments for Protein-DNA Interactions

Electrophoretic Mobility Shift Assay (EMSA)

Proteins used in EMSA experiments were purified as described in section 2.5.3

and quantitated as described in section 2.5.3. Purified protein samples of equal

amounts were incubated in binding buffer (20 mM HEPES-KOH, pH7.9, 100 mM KCl,

0.8 mM ZnCl2, 0.2 mM EDTA, 5% glycerol, 0.5 mM DTT and 0.5 mM PMSF) with 6.25

μg/mL poly (dI:dC) and 50 ng/μL BSA on ice for 20 minutes. ~0.12 nM (~2.5 fmol) γ-

32P labelled DNA oligonucleotide (see Table 2.6) probes were added to samples and

incubated on ice for another 30 minutes. Binding reactions were loaded onto an 8%

native acrylamide/bisacrylamide (19:1, Biorad) gel pre-run at 100V in 0.5 x TB buffer

(89 mM Tris, 89 mM boric acid) for 20 minutes. Gels were electrophoresed at 150 V

for 150 minutes at 4°C, dried at 65°C for 45 minutes, and then exposed to a Phosphor

screen overnight. Gel images were obtained by scanning the screens with a Typhoon

FLA9500 Imager (GE Life Sciences).

Double stranded (ds) oligonucleotides were generated by annealing the

complementary strands. Single stranded (ss) and ds probes were end labelled with γ-

32P using T4 polynucleotide kinase (New England Biolabs) according to the

manufacturer’s instructions.

Table 2.6 EMSA probes used in this study # Probe Sequence (5’-3’) 1 ss telomeric G rich oligo TTA GGG TTA GGG TTA GGG 2 ss telomeric C rich oligo CCC TAA CCC TAA CCC TAA

3 ss scrambled G rich oligo TGA GTG TGA GTG TGA GTG 4 ss scrambled C rich oligo CAC TCA CAC TCA CAC TCA 5 ss pentaprobe 3 CGC TCT ATT CTA CTG TCC TGT GCA TTC AAT CGT TGA GTT CGA TCT AGT CTC GTC TAA CCC TCC CCT GCT CCG CTG GTC TGG CCT CGC CTA TCC TAC CCA T 6 ss pentaprobe 9 ATG GGT AGG ATA GGC GAG GCC AGA CCA GCG GAG CAG GGG AGG GTT AGA CGA GAC TAG ATC GAA CTC AAC GAT TGA ATG CAC AGG ACA GTA GAA TAG AGC G 7 ds telomeric oligo Annealed probe 1 and 2 8 ds scrambled oligo Annealed probe 3 and 4 9 ds pentaprobe 3-9 Annealed probe 5 and 6

2.8 Experiments for Chromatin Interactions

Telomere - chromatin immunoprecipitation (ChIP)

Subconfluent cells were harvested by trypsinisation followed by centrifugation at 200 x g for 5 minutes. Cell crosslinking, lysis and nuclei preparation were conducted following the truChIP™ Chromatin Shearing Kit protocol (Covaris). Briefly, 1 x 107 cells were washed in cold PBS and crosslinked with 1% methanol-free formaldehyde

(28906, Pierce™) for 5 minutes at room temperature on a rotating platform. Each crosslinking reaction was quenched by adding 87 μL of Quenching Buffer E (truChIP™

Chromatin Shearing Kit, Covaris) for 5 minutes at room temperature on a rotating platform. Cells were collected by centrifugation at 500 x g for 5 minutes, washed twice with cold PBS and then collected by centrifugation at 200 x g for 5 minutes at 4°C. To lyse the plasma membrane, crosslinked cells were resuspended in 1 mL Lysis Buffer

B with 1 x protease inhibitors and incubated for 10 minutes on a rocker at 4°C. Intact nuclei were collected by centrifugation at 1,700 x g for 5 minutes at 4°C, and subsequently resuspended in Wash Buffer C with 1 x protease inhibitor and incubated for 10 minutes on a rocker at 4°C. Each nuclei pellet was collected again by centrifugation at 1,700 x g for 5 minutes at 4°C, rinsed twice with Shearing Buffer D3 containing 1 x protease inhibitor and then resuspended in 1 mL Shearing Buffer D3 containing 1 x protease inhibitor. Resuspended nuclei pellets were then transferred to

1 mL AFA (Adaptive Focused Acoustics) tubes for subsequent chromatin shearing.

Optimal chromatin shearing was achieved by subjecting each AFA tube to 15-20

minutes of sonication depending on the cell line in the Covaris E220 AFA focused-

ultrasonicator. Following sonication, the sheared chromatin was transferred to

microcentrifuge tubes and centrifuged at 13,000 x g for 10 minutes at 4 °C, and the supernatant containing the solubilised chromatin collected into a fresh tube.

For each immunoprecipitation reaction, 112 μL of chromatin was added to 225

μL of ice-cold IP buffer (20 mM HEPES-KOH, pH 8.0, 150 mM KCl, 1.5 mM EDTA, 1%

Triton X-100, 1 x CPI, 1 mM PMSF), and antibody for protein of interest (5-10 μg/IP) was added. Samples were rotated for 1 hour at 4°C followed by the addition of 40 μL

BSA (nuclease-free)-blocked protein G agarose beads (Roche), and further incubated with rotation overnight at 4°C. The following day, beads were collected by centrifugation for 30 seconds at 17,000 x g at room temperature, discarding the

supernatant. The beads were washed twice with 1 mL of room temperature buffer A

(20 mM HEPES-KOH, pH 7.9, 2 mM MgCl2, 300 mM KCl, 1 mM EDTA, 10% glycerol

and 0.1% Triton X-100), with centrifugation for 30 seconds at 17 000 x g at room

temperature discarding the supernatant each wash, followed by another wash with 1 mL of TE (10 mM Tris-Cl, 1 mM EDTA, pH 8.0). The immunoprecipitated DNA was decrosslinked from the proteins by incubating samples with 200 µg of proteinase K

(Roche) in 100 µL of elution buffer (50 mM NaHCO3, 1% SDS) at 55 °C for 5 hours.

Samples were then purified using the QIAquick PCR Purification Kit (Qiagen), and the

DNA eluted in 140 µL TE.

Purified DNA was denatured by the addition of 900 µL of 0.46 M NaOH and

incubation at 95 °C for 5 minutes. Samples were then cooled on ice, split equally in

half and loaded onto two Hybond XL (GE Healthcare Life Sciences) membranes

soaked in 2 x SSC. Membranes were air dried for 30 minutes and UV-crosslinked with

240 mJ using a Stratalinker at 254 nm followed by pre-hybridisation in Church buffer

(1% BSA, 1 mM EDTA, 500 mM Na2HPO4, pH 7.2, 7% SDS) for 60 minutes at 55°C.

32 One membrane was hybridised overnight with γ- P-(CCCTAA)3 telomeric probe, the other with γ-32P-ALU probe labelled using the DECAprimeTM II Random Primed DNA

Labelling Kit (Ambion) according to the manufacturer’s instructions. The following day, membranes were washed 3 times for 10 minutes in 2 x SSC at room temperature and air dried before being exposed to a PhosphorImager screen overnight. The Phosphor screen was imaged using a Typhoon FLA9500 Imager (GE Healthcare Life Sciences).

The image was quantitated using the array analysis function on the ImageQuant software (GE Healthcare Life Sciences). The amount of telomeric DNA immunoprecipitated was calculated for each sample based on the standard curve generated from a series of input telomeric DNA. The ALU blot was used to control for any non-specific pull-down.

2.9 Next Generation Sequencing

RNA-seq

Experimental design for RNA-seq in this study is outlined in Table 2.7. RNA extractions were carried out at the same time in batches for all three biological replicates as described in section 2.4.1. Extracted RNAs were subjected to quality check using the Bioanalyzer (Agilent Technologies) prior to sample submission to

AGRF. RNA-sequencing of samples from three biological replicates was performed by

AGRF on the HiSeq 2500 Illumina NGS sequencing platform in technical duplicates to generate single end 50 bp short reads at a sequencing depth of 12-15 million reads per sample.

Data analysis was conducted by Erdahl Teber in the Bioinformatics Unit at

CMRI. The Bioconductor Rsubread package (Liao et al., 2013) in the R statistical

language version 3.0.3 was employed to map the reads to an unmasked human

reference (hg19) and count total reads using the hg19 NCBI RefSeq gene model

annotations. The Picard package (Broad Institute) was used to provide statistics on

the quality of alignment of reads. The Voom program from the Bioconductor limma

package (Law et al., 2014) was used to transform counts to log-counts per million with

associated precision weights. Differential expression analysis was performed using

limma (Smyth, 2005) only on genes which had greater than 1 read per million mapped

reads in at least one third of the sample libraries.

Pathway analysis of the differentially expressed genes was performed using

IPA (Ingenuity Pathway Analysis, Qiagen).

Table 2.7 RNA-seq experimental design Biological Cell line Manipulation Sample Replicates WI38- Empty Vector transfection Control 3 VA13/2RA ZNF827 overexpression Treatment 3 (ALT) ZNF827 ΔRRK overexpression Treatment (mutant) 3 Empty Vector Control 3 293T ZNF827 overexpression Treatment 3 (Telomerase) ZNF827 ΔRRK overexpression Treatment (mutant) 3

ChIP-seq

The ChIP part of this experiment including chromatin preparation, sonication,

immunoprecipitation and DNA decrosslinking was carried out as described in section

2.8.1. Each decrosslinked sample was purified using the QIAquick PCR Purification

Kit (Qiagen) with the DNA being eluted in 50 µL of filtered TE. Eluted DNA was quantitated using the Qubit Fluorometer (ThermoFisher Scientific) according to the manufacturer’s instructions. The experimental design is outlined in Table 2.8. For each replicate, several ZNF827 IPs were pooled together to obtain sufficient DNA for

sequencing. DNAs passed the initial quality and quantity check on Bioanalyzer (Agilent)

71 before submission to AGRF. Paired-end 150 bp ChIP-sequencing was performed by

AGRF on the HiSeq 2500 Illumina NGS sequencing platform. Analysis was conducted by Emilie Wilkie in the Bioinformatics Unit at CMRI. Read alignment and quality control for each replicate were performed using STAR with the human reference genome

(h38). Disconcordant pairs, reads aligning to mitochondria as well as duplicated reads were removed. Peak calling was conducted with MACS2.

Table 2.8 ChIP-seq experimental design Cell line Manipulation ZNF827 ChIP Replicates WI38-VA13/2RA Empty Vector transfection Endogenous 2 (ALT) ZNF827 overexpression Endogenous + Exogenous 2 293T Empty Vector Endogenous 2 (Telomerase) ZNF827 overexpression Endogenous + Exogenous 2

2.10 Fixed Cell Fluorescence Microscopy

Cell preparation

Metaphase spreads

To prepare metaphase spreads, sub-confluent cells were treated with 100 ng/mL colcemid (Gibco) in DMEM supplemented with 10% FCS for 4 hours to arrest cells in metaphase prior to harvest. For overexpression experiments, colcemid was added at 44 hours post transfection, and at 68 hours post transfection for siRNA knockdown experiments. As mitotic cells are less adherent, media was collected along with the adherent cells harvested routinely by trypsinisation and centrifugation.

Pelleted cells were then resuspended in hypotonic solution (0.2% potassium chloride and 0.2% tri-sodium citrate) for 10 minutes at 37°C before being further processed for indirect immunofluorescence (IF) and fluorescence in situ hybridisation (FISH).

Interphase nuclei

For interphase IF and FISH, cells were cultured as usual on sterile Alcian blue

stained-glass coverslips (#0117580, Marienfeld-Superior) in 6 or 12 well tissue culture

plates (Corning) before further processing.

Telomere - Fluorescence in situ hybridisation (telomere-FISH)

Metaphase spreads – telomere fragility

Following the hypotonic solution incubation in section 2.10.1 above, swollen

cells were fixed by gradually adding 1 mL of fresh ice-cold fixative (methanol/acetic acid 3:1), mixed by inversion and incubated on ice for 5 minutes. Cells were then

collected by centrifugation at 1200 rpm for 8 minutes. 10 mL ice cold fixative was

added to resuspend the cells, followed by 5 minutes incubation on ice, and

centrifugation at 1200 rpm for 8 minutes. This fixing step was repeated another two

times. Fixed cells were then resuspended in 500 – 1000 µL of ice-cold fixative, and

dropped onto clean, dry microscope slides (HDS Surefrost) with 50 - 100 µL cell

solution per slide. To drop chromosomes, a clean dry slide was held over a 75°C water

bath, and the cell solution was dropped from a pipette onto the slide, which was quickly

flipped and held close to the surface of water bath for 5 seconds. After leaving to dry

for 2-3 days in the dark, slides were then blocked with ABDIL containing 100 µg/mL

DNase-free RNase A (Sigma-Aldrich) for 30 minutes at 37°C, rinsed in PBS and then

fixed in 4% formaldehyde in PBS at room temperature for 10 minutes. Following a quick rinse in deionised water, slides were dehydrated by a graded ethanol series

(70% for 3 minutes, 90% for 3 minutes and 100% for 3 minutes), and then air dried.

Dehydrated slides were then overlaid with a fluorophore conjugated centromere- telomere PNA probe containing 0.3 µg/mL CENT-Cy3-OO-

(AAACTAGACAGAAGCAT) and 0.3 µg/mL AF488–OO-(CCCTAA)3 (Panagene),

denatured for 5 minutes at 80°C and left to hybridise overnight at room temperature in

the dark in a humidified chamber. The following day, slides were washed in PNA wash

A (70% formamide and 10 mM Tris-Cl, pH7.5) for 3 x 5 minutes, and in PNA wash B

(50 mM Tris-Cl, pH 7.5, 150 mM NaCl and 0.08% Tween-20) for 3 x 5 minutes. Slides

were then incubated with 50 ng/mL DAPI in PBS for 15 minutes followed by 2 x 5

minutes washes in PBST and a quick rinse in deionised water. After airdrying, slides

were mounted in Prolong Gold Antifade (Invitrogen) and stored at 4°C until microscope

analysis.

Indirect immunofluorescence (IF)

Interphase nuclei – ZNF827-RPA32/ATR/γH2AX colocalisations

After PBS wash, cells cultured on coverslips as described in section 2.10.1

were fixed in 2% paraformaldehyde in PBS at room temperature for 10 minutes. Cells

fixed on coverslips were rinsed with deionised water, and then permeabilised using

KCM buffer (120 mM potassium chloride, 20 mM NaCl, 10 mM Tris, pH 7.5, 0.1%

Triton X-100) at room temperature for 10 minutes. Coverslips were blocked with

antibody diluent buffer (ABDIL - 20 mM Tris-Cl, pH 7.5, 2% BSA, 0.2% fish gelatine,

150 mM NaCl, 0.1% Triton X-100, 0.1% sodium azide) containing 100 µg/mL RNase

A at room temperature for one hour. Primary antibodies diluted in ABDIL were added

to the coverslips, and incubated at room temperature with gentle rocking for 1 hour or

at 4°C overnight (refer to Table 2.9 in Section 2.14 for the list of antibodies used in this

study). Coverslips were washed in PBST three times for 10 minutes with shaking, and

then incubated with appropriate fluorophore conjugated secondary antibodies diluted

1:500 in ABDIL (See Table 2.10) at room temperature with gentle rocking in the dark

for 1 hour. Following three 10-minute PBST washes with shaking, coverslips were

incubated with 50 ng/mL DAPI in PBS for 15 minutes followed by 2 x 5 minutes washes

in PBST and a quick rinse in deionised water. After airdrying, slides were mounted in

Prolong Gold Antifade (Invitrogen) and stored at 4°C until microscope analysis.

Combined indirect immunofluorescence with fluorescence in situ hybridisation (Combined IF with FISH) - Meta-TIFs (telomere dysfunction induced foci in metaphase), APBs (ALT associated PML bodies), telomere-ZNF827/RPA32 colocalisations

Combined IF with FISH on metaphase spreads – Meta-TIF assay

Following hypotonic solution incubation in section 2.10.1 above, swollen cells

were cytocentrifuged onto SuperFrost Plus glass slides (Menzel-Glaser) at 2000 rpm for 10 minutes in a Shandon Cytospin 4 (ThermoFisher Scientific) set at medium acceleration. Slides were incubated in KCM buffer (120 mM potassium chloride, 20 mM NaCl, 10 mM Tris, pH 7.5, 0.1% Triton X-100) at room temperature for 10 minutes for permeabilisation, rinsed in PBS, and then fixed in 4% formaldehyde in PBS at room temperature for 10 minutes. After a quick rinse in deionised water, slides were blocked in a humidified chamber with antibody dilution buffer (ABDIL - 20 mM Tris-Cl, pH 7.5,

2% BSA, 0.2% fish gelatine, 150 mM NaCl, 0.1% Triton X-100, 0.1% sodium azide) containing 100 µg/mL RNase A for 30 minutes at 37°C or at room temperature for one hour. Slides were then overlaid with primary antibody anti-γH2AX mouse monoclonal

(JBW301, Millipore) diluted 1 in 200 in ABDIL and incubated at room temperature for

1 hour or overnight at 4°C. Following three 5-minute PBST washes with shaking, slides were incubated with secondary antibody goat anti-mouse-Alexa Fluor 594 (Invitrogen) diluted 1:500 in ABDIL for 1 hour at room temperature in a humidified chamber. Slides were washed for 3 x 5-minutes in PBST and fixed in 4% formaldehyde in PBS for 10 minutes at room temperature, in order to fix the antibodies in place during the subsequent FISH steps. A graded ethanol series was used to dehydrate the slides

(70% for 3 minutes, 90% for 3 minutes and 100% for 3 minutes), followed by airdrying

the slides in the dark. For the subsequent telomere-FISH, airdried slides were overlaid

with 0.3 µg/mL Alexa Fluor 488-conjugated telomere PNA probe AF488–OO-

(CCCTAA)3 (Panagene), denatured at 80°C for 3 minutes and hybridised at room temperature overnight in the dark in a humidified chamber. The following day, slides were washed in PNA wash A (70% formamide and 10 mM Tris-Cl, pH 7.5) for 3 x 5 minutes, then in PNA wash B (50 mM Tris-Cl, pH 7.5, 150 mM NaCl and 0.08% Tween-

20) for 3 x 5 minutes. Slides were then incubated with 50 ng/mL DAPI in PBS for 15 minutes, followed by a quick rinse in deionised water and a graded ethanol series as described above. After airdrying, slides were mounted in Prolong Gold Antifade

(Invitrogen) and stored at 4°C until microscope analysis.

Combined IF with FISH on interphase spreads – APBs and telomere-

ZNF827/RPA32 colocalisations

See section 2.10.3 IF - interphase nuclei for the IF experimental procedures.

Following incubation with secondary antibodies, coverslips were fixed again with 2% paraformaldehyde in PBS for 10 minutes at room temperature to fix antibodies in place for the subsequent FISH steps. Coverslips were then rinsed twice with deionised water, dehydrated by a graded ethanol series (70% ethanol for 3 minutes, 90% ethanol for 3 minutes, and 100% ethanol for 3 minutes). Coverslips were removed from 100% ethanol and allowed to air dry. For telomere FISH, 0.3 µg/mL Alexa Fluor 488- conjugated telomere PNA probe AF488–OO-(CCCTAA)3 was added to clean slides,

and the coverslips were placed cell side down onto the PNA solution. Slides were

denatured at 80°C for 10 minutes and placed in a humidified chamber for hybridisation at room temperature overnight. The following day, slides were washed in PNA wash

A (70% formamide and 10 mM Tris-Cl, pH 7.5) for 2 x 10 minutes with shaking, then

in PNA wash B (50 mM Tris-Cl, pH 7.5, 150 mM NaCl and 0.08% Tween-20) for 3 x

10 minutes, with 50 ng/ml DAPI added in the second wash. Following a quick rinse in

deionised water, coverslips were dehydrated again in a graded ethanol series as

described above. After airdrying, slides were mounted in Prolong Gold Antifade

(Invitrogen) and stored at 4°C until microscope analysis.

Sister chromatid exchange (SCE)

20-25% confluent cells were cultured in fresh media supplemented with 7.5 μM

5‐bromo‐2′‐deoxycytidine (BrdU) and 2.5 μM BrdC (BrdU:BrdC 3:1 ratio; Sigma‐

Aldrich) for 32–48 hours depending on the mitotic index of the cell line. Cell cultures

were treated with 100 ng/mL colcemid for the last 4 hours of incubation to accumulate

mitotic cells. Cells were harvested by trypsinisation and centrifugation, and then

incubated in hypotonic buffer for 10 minutes at 37°C. Swollen cells were fixed by gradually adding 1 mL of fresh ice-cold fixative (methanol/acetic acid 3:1), mixing by inversion and incubating on ice for 5 minutes. Cells were then collected by centrifugation at 1200 rpm for 8 minutes. 10 mL ice cold fixative was added to resuspend the cells followed by 5 minutes incubation on ice and centrifugation at 1200 rpm for 8 minutes. This fixing step was repeated another two times. Fixed cells were then resuspended in 500 – 1000 µL of ice-cold fixative, and dropped onto clean, dry microscope slides (HDS Surefrost, 50 - 100 µL cell solution per slide). To drop chromosomes, a clean dry slide was held over a 75°C water bath, and the cell solution was dropped from a pipette onto the slide, which was quickly flipped and held close to

the surface of water bath for 5 seconds. Slides were left to dry for 2 to 3 days and then

treated with 100 μg/ml DNase‐free RNase A (Sigma) in 2 x SSC for 30 minutes at

37°C, rinsed in PBS, and postfixed in 4% formaldehyde in PBS at room temperature

for 10 minutes. Following a quick rinse in deionised water, slides were dehydrated in

a graded ethanol series (70% for 3 minutes, 90% for 3 minutes, and 100% for

3 minutes) and allowed to airdry. Slides were then stained in 0.5 μg/mL Hoechst 33258

(Sigma‐Aldrich) in 2 x SSC for 15 minutes at room temperature, rinsed in dH2O, and

air‐dried. Slides were then flooded with 200 μL 2 x SSC and exposed to long‐wave

( 365 nm) UV light (Stratalinker 1800 UV irradiator; Agilent Technologies) for

45∼ minutes. The BrdU/BrdC‐substituted DNA strands were then digested in 10 U/μL

Exonuclease III solution (New England Biolabs) in the supplied buffer at 37°C for

30 minutes. After a quick rinse in deionised water, slides were incubated with 50 ng/mL

DAPI in PBS for 15 minutes, washed twice in PBST for 5 minutes, rinsed in deionised

water, and airdried. Airdried slides were mounted in Prolong Gold Antifade (Invitrogen)

and stored at 4°C until microscope analysis.

Comet assay

Comet assays were performed using the CometAssay® kit (Trevigen)

according to the manufacturer’s protocol with minor modifications. Briefly, cells were

counted and resuspended in PBS. 20 μl of cells at 1 x 105/mL were combined with molten LMAgarose (Trevigen) at 37 °C. The cell mixture was pipetted onto the sample area of a prewarmed slide (CometSlide™, Trevigen). Slides were cooled at 4 °C in the dark for 25 minutes, and then immersed in ice-cold lysis solution (CometAssay® kit,

Trevigen) at 4 °C overnight. On the following day, slides were removed from lysis

solution and gently immersed in ice-cold 1 x neutral electrophoresis buffer (100 μM

Tris base and 300 μM sodium acetate) for 30 minutes. Slides were then

electrophoresed in 1 x neutral electrophoresis buffer (100 μM Tris base and 300 μM

sodium acetate) for 45 minutes at 1 V/cm. Following electrophoresis, slides were

immersed in DNA precipitation solution (1 M ammonium acetate in 95% ethanol) for

30 minutes at room temperature. Slides were then transferred and immersed in 70%

ethanol for 30 minutes at room temperature. Slides were then dried for 15 minutes at

37 °C, incubated with 1 μM YOYO stain (Invitrogen) for 5 minutes, and washed for 2

x 5 minutes in deionised water. After airdrying, slides were mounted in Prolong Gold

Antifade (Invitrogen) and stored at 4°C until microscope analysis.

Imaging and image analysis

Interphase cells were imaged manually on an Axio-Imager M1 (Carl Zeiss)

microscope, and analysed either by manual scoring with ZEN (Carl Zeiss) software or

by semi-automated counting in CellProfiler (Broad Institute). APBs were imaged by automation on the MetaSystems Metafer Scanning Platform (Carl Zeiss) microscope and analysed by a built-in function in Metafer 4 (MetaSystems). Metaphase spreads were imaged by automation on the MetaSystems Metafer Scanning Platform (Carl

Zeiss) microscope, and analysed using the Isis software (Metasystems).

2.11 Time-lapse Live Cell Microscopy

Fluorescence ubiquitination cell cycle indicator (FUCCI) live cell imaging

HT1080 FUCCI cells were seeded in glass bottom 12-well plates (MakTek

Corporation) the day before imaging at 30% confluency, to ensure that cells were

actively cycling for the duration of the experiment. Media was replaced with phenol red

free DMEM (Gibco) supplemented with 10% FCS and drug treatment immediately

before transferring the plate to the microscope chamber. Imaging was performed on

the Cell Observer Widefield Microscope (Zeiss) using a 20 x objective at 37°C with 10%

CO2 in a XLmulti S1 full enclosure chamber. Cells were incubated in the XLmulti S1 full enclosure chamber for 2 hours prior to imaging. Three to four positions per well were captured with AxioCam (Zeiss) 506 Mono using ZEN software (Carl Zeiss) with images taken every 6 minutes for 60 hours. More than 15 cells were analysed at each position for each condition to confirm reproducibility.

Cell growth and apoptotic assay with IncuCyte®

Cells were seeded at 20-30% in a 96-well black plate with clear bottom (3603,

Corning) and left to adhere at 37°C with 10% CO2 for at least 3 hours. The media was

aspirated and replaced with phenol red free media supplemented 2 µM Nucview® 488

Caspase-3 substrate (Biotium) and placed into the IncuCyte® for baseline imaging

without drug treatment. Topotecan 2µg/mL or DMSO control was added to the wells,

and incubated at 37°C with 10% CO2 for 1 hour. The media was then aspirated, cells

were washed carefully once with warm PBS, and the media replaced with phenol red

free media supplemented 2 µM Nucview® 488 Caspase-3 substrate (Biotium). The

plate was then placed back into IncuCyte®, imaged immediately post drug treatment,

and then every 4 hours for 72 to 84 hours in the phase and green channels. Image

analysis was performed with the IncuCyte® Zoom software. Cell growth was

measured as % confluency over time, and apoptotic activity as green object

confluence/area over time.

2.12 Cell Cycle Analysis

Cells were trypsinised, counted and collected by centrifugation at 1000 rpm for

5 minutes. 1 × 106 cells were resuspended in 0.5 mL PBS, then fixed in 5 mL of ice-

cold 70% ethanol drop by drop whilst being gently vortexed. The cells were incubated

on ice for 30 minutes then stored at -20°C for 2-7 days until propidium iodide staining.

Prior to propidium iodide (Sigma) staining, the ethanol fixed cells were centrifuged at

300 × g for 6 minutes to remove ethanol, washed once with PBS, and resuspended in

0.5mL PBS per 1 × 106 cells. The fixed cells were treated with 0.5 mg of RNase

(Sigma) and 25 µg of propidium iodide at 37oC for 30 minutes, and then allowed to

return to room temperature in the dark for 10 minutes. Labelled cells were analysed using the BD FACSCanto Flow Cytometer (BD Biosciences) containing a blue air

80 cooled 488 nm argon laser to excite propidium iodide. Between 10 000 -15 000 events were collected at an approximate flow rate of 200 events/second. The forward scatter

(FSC, size) and side scatter (SSC, internal granularity) of each cell was recorded. To discriminate and eliminate cell debris and doublets, the pulse area (FL2-A) was plotted against the pulse width (FL2-W). Doublets identified as cells with 4N DNA content and increasing pulse width were eliminated. A histogram displaying the cell counts against the FL2-A was used to calculate the percentage of cells in each cell cycle phase, as well as the percentage of non-viable cells. Data analysis was conducted using BD

FACS Diva software (BD Bioscience).

2.13 Drug Treatment

Topotecan and aphidicolin were used in this study. For experiments where they were used to cause replication stress and DNA damage, cells were treated with topotecan 2 μg/mL for either 1 hour or 24 hours, or aphidicolin 1 μM for 20 hours. For replication fork stalling induction in the PCNA colocalisation experiments, cells were treated with aphidicolin 0.4 μM for 6-8 hours. Topotecan and aphidicolin were dissolved in DMSO. DMSO was used as a control.

2.14 Antibodies – Primary and Secondary

Table 2.9 List of primary antibodies used in this study

MW Primary antibodies Source Catalog No. Dilution Applications Host (kD) 1 in 200 (WB/IF); sc-249818, 1 μg/100 μL WB, IF, IP, ZNF827 119 Santa Cruz Goat T20 lysate (IP); 15 PLA, ChIP μL/110 μL (ChIP) 1 in 1000 (WB); 1 Myc-Tag n/a Cell Signaling Technologies 2276, 9B11 WB, IF, ChIP Mouse in 4000 (IF); 1 in 1000 (WB); 6 RBBP4 48 Novus NB500-123 WB, ChIP Mouse μL/110 μL (ChIP) 6 μL/110 μL RBBP7 46 Novus NB120-3535 ChIP Rabbit (ChIP) 8 μL/110 μL GATAD2B/p66 beta 65 Novus NBP1-87358 ChIP Rabbit (ChIP) CHD4 218 Novus NB100-57521 1 in 2000, 10 WB, ChIP Rabbit HDAC1 62 Cell Signaling Technologies 5356, 10E2 1 in 1000 WB mouse MTA1 81 Cell Signaling Technologies 5646, D17G10 1 in 1000 WB Rabbit COUP-TF II 45 R&D systems H7147 1 in 1000 WB Mouse TR4 67 R&D systems H0107B 1 in 500 WB Mouse TR2 65 Santa Cruz M85, sc9087 1 in 500 WB mouse FLAG-Tag n/a Origene OTI4C5 1 in 2000 WB Mouse Actin 42 Sigma A2066 1 in 5000 WB Rabbit Vinculin 124 Sigma V9131 1 in 5000 WB Mouse pCHK1 (S345) 56 Cell Signaling Technologies 2348, 133D3 1 in 1000 WB Rabbit CHK1 56 Cell Signaling Technologies 2360, 2G1D5 1 in 1000 WB Mouse pRPA2 (S33) 32 Bethyl lab A300-246A 1 in 2000 WB Rabbit 1 in 500 (WB/IF); RPA2 32 Abcam ab2175 1 μg/100 μL WB, IF, IP Mouse lysate (IP) ATR 309 Abcam ab2905 1 in 500 IF, PLA Rabbit ATR 309 Cell Signaling Technologies 2790 1 in 1000 WB Rabbit p21 21 Cell Signaling Technologies 2947, 12D1 1 in 1000 WB Rabbit p53 43 Santa Cruz D01 1 in 1000 WB Mouse 05-636, p-γH2AX (S139) 17 Millipore 1 in 1000 WB Mouse JW301 pCHK2 (Thr68) 62 Cell Signaling Technologies 2661 1 in 1000 WB Rabbit 1 μg/100 μL Normal IgG n/a Cell Signaling Technologies 2729 lysate (IP); 5 IP, ChIP Rabbit μL/110 μL (ChIP) 1 μg/100 μL Normal IgG n/a R&D Systems AB108C lysate (IP); 5 IP, ChIP Goat μL/110 μL (ChIP) TRF2 55.5 Novus Biologicals NB110‐57130 5 μL/110 μL ChIP Rabbit

Table 2.10 List of secondary antibodies used in this study Antibody Host Source Catalog No. Applications Dilution HRP anti-Mouse Polyclonal Goat Dako P0447 WB 1:5000 IgG HRP anti-Rabbit IgG Polyclonal Goat Dako P0448 WB 1:5000 HRP anti-Goat IgG Polyclonal Rabbit Dako P0449 WB 1:5000 Donkey anti-goat Invitrogen A-11055 IF 1:500 Alexa Fluor 488 Donkey anti-mouse Invitrogen A-21202 IF 1:500 Donkey anti-rabbit Invitrogen A-21206 IF 1:500 Donkey anti-goat Invitrogen A-11058 IF 1:500 Alexa Fluor 594 Donkey anti-mouse Invitrogen A-21203 IF 1:500 Donkey anti-rabbit Invitrogen A-21207 IF 1:500 Donkey anti-mouse Invitrogen A-31571 IF 1:500 Alexa Fluor 647 Donkey anti-goat Invitrogen A-21447 IF 1:500

2.15 Statistical Analyses

Microsoft Excel and GraphPad Prism 7 were used to generate graphs and perform statistical analysis by calculating P values using unpaired Student’s two‐tailed t‐tests.

2.16 Mining ZNF827 Data in Public Databases

ZNF827 was searched in public databases to gain some insights into its biological significance. ZNF827 orthologs were obtained from EnsemblCompara

(EnsemblCompara, 2018). Tissue expression data were obtained from the GTEx

(Genotype-Tissue Expression) portal and the Expression Atlas (The_Genotype-

Tissue_Expression_Project, 2017, Expression_Atlas, 2018). ZNF827 single nucleotide polymorphism (SNP) data was obtained from the GWAS Catalog

(GWAS_Catalog, 2018). Clinical phenotypes associated with ZNF827 mutations were obtained from DECIPHER (DatabasE of genomiC varIation and Phenotype in Humans using Ensembl Resources) (DECIPHER, 2018).

Unravelling the Biological Functions of ZNF827

3.1 Introduction

ZNF827 is enriched at the telomeres of cancer cells that utilise the ALT pathway,

where it fulfils a vital functional role by recruiting the NuRD complex (Conomos et al.,

2014). The N-terminal RRK motif in ZNF827 is essential for the recruitment of NuRD

to ALT telomeres, and this interaction is abolished when the RRK motif is mutated

(Conomos et al., 2014). Whilst the NuRD complex plays important roles in multiple

cellular processes, including transcriptional regulation in haematopoiesis, cell cycle

progression, chromatin assembly, DNA replication and DNA repair (Lai and Wade,

2011), no other biological functions of ZNF827 have been reported. This makes

ZNF827 a promising novel therapeutic target for cancers utilising ALT. In order to

characterise ZNF827 as a therapeutic target in ALT, it is imperative to first find out

whether it has other vital non-telomeric biological functions in cellular processes, as

successful and specific targeting relies on ZNF827 having a distinct function in the

ALT pathway.

ZNF827 belongs to the superfamily of C2H2 zinc finger proteins, which exist in

high abundance in eukaryotic organisms (Laity et al., 2001). Though evolutionarily

conserved across different species, the C2H2 zinc finger gene family has undergone

extensive lineage-specific expansions resulting in wide structural and functional

diversity, suggestive of an important role in adaptive evolution (Emerson and Thomas,

2009, Tadepally et al., 2008). On the primate lineage, rapid expansion of this gene family has occurred, giving rise to a substantial proportion of human C2H2 zinc finger genes with no mouse orthologs (Emerson and Thomas, 2009). Since their discovery in 1985, C2H2 zinc finger proteins have been best characterised as transcription factors with DNA recognition specificity involved in transcriptional activation and

84 repression (Miller et al., 1985, Lee et al., 1989, Pavletich and Pabo, 1991, Wolfe et al.,

2000). This is the largest class of transcription factors in humans and constitutes approximately 700 C2H2 zinc finger proteins (Vaquerizas et al., 2009). Given that

ZNF827 is a C2H2 zinc finger protein that interacts with the NuRD complex, and that

NuRD is known to interact with various transcription factors to affect transcriptional regulation, it is likely that ZNF827 may function as a transcriptional regulator involved in certain cellular processes.

The conservation of orthologous genes across different species indicates conserved biological functions (Emerson and Thomas, 2009). To date, published literature on ZNF827 is lacking, and ZNF827 has not been identified as a major determinant in any diseases. To gain a glimpse into the possible biological significance of ZNF827, we examined a range of reliable public databases to examine ZNF827 gene orthology, ZNF827 gene expression in human tissues, and any clinical phenotypes associated with ZNF827 mutations. To investigate the impact of ZNF827 on transcriptional regulation, we conducted RNA sequencing (RNA-seq) analysis to determine global gene expression patterns following ZNF827 overexpression and identify candidate genes and pathways regulated by ZNF827. By including the RRK mutant in the RNA-seq experiment, we compared the differences in gene expression patterns in the presence and absence of NuRD recruitment. We performed ChIP followed by next generation sequencing (ChIP-seq) to identify ZNF827 binding sites to non-telomeric regions of the genome.

3.2 ZNF827 orthologs and tissue expressions

To gain insight into the evolutionary progression and significance of the ZNF827 gene, the EnsemblCompara ortholog database was searched for ZNF827 orthologs.

The human ZNF827 gene has 138 orthologs evolutionarily conserved in different

85 species of vertebrates from fish to higher order mammals (EnsemblCompara

(EnsemblCompara, 2018)). In mammals, 59 orthologs share over 90% sequence similarity with the human ZNF827 ortholog, ranging from 90.17% in tree shrews, to

100% match in bonobos (see Table 3.1).

ZNF827 gene expression data in human tissues was obtained from the GTEx

(Genotype-Tissue Expression) project portal which is an open access public database providing gene expression data from RNA-sequencing experiments of over 10,000 samples collected from 53 non-diseased tissue sites (GTEx portal, (The_Genotype-

Tissue_Expression_Project, 2017)). Expression of TERF2, the gene name for the telomere-associated protein TRF2, was included for relative comparison. The ZNF827 gene is universally expressed at low levels across the 53 different tissues, with the lowest expression in blood and liver, and the highest expression in the prostate gland and the cerebellar hemisphere (Figure 3.1 a). Notably, ZNF827 expression is comparatively higher in tissues from the reproductive system (Figure 3.1 a).

In addition to adult human tissues, ZNF827 is expressed in foetal tissues, with a distinct pattern emerging from comparison between foetal tissues at early and later stages along the developmental timeline. Notably, foetal ZNF827 expression analysed in the RIKEN FANTOM 5 project was collectively low in the 22 samples collected from

20 different human foetal tissues at developmental stages between 16 weeks up to 40 weeks, whereas remarkably higher expression of ZNF827 was reported by the NIH

Roadmap Epigenomics Mapping Consortium in 53 samples covering 19 different tissues from human foetuses with congenital defects at earlier developmental ages of between 85 and 120 days (~12 weeks to 17 weeks) (Figure 3.1 b). These expression data suggest a development dependent expression pattern of ZNF827. Further expression data of ZNF827 with respect to development was accessed from the

Expression Atlas powered by the EMBL-EBI (European Molecular Biology Laboratory

- European Bioinformatics Institute), with search results filtered by developmental stage (Expression_Atlas, 2018). Consistently, ZNF827 gene expression was significantly higher in foetal samples compared to the juvenile samples based on the

RNA-seq data collected from a study that examined the transcriptional profiles of intestines during development (Figure 3.1 c) (Kraiczy et al., 2019). Together, these data suggest that ZNF827 is a developmentally regulated gene, highly expressed at early developmental stages, but silenced at later stages.

Table 3.1 Mammalian ZNF827 orthologs with over 90% sequence similarities to human ZNF827 Sequence Sequence alignment to alignment to Species Species human human ZNF827 ZNF827 Bonobo (Pan paniscus) 100.00% Degu (Octodon degus) 96.22% Naked mole-rat female Gorilla (Gorilla gorilla gorilla) 99.91% 96.11% (Heterocephalus glaber) Chimpanzee (Pan troglodytes) 99.91% Brazilian guinea pig (Cavia aperea) 96.00% Vervet-AGM (Chlorocebus 99.54% Bushbaby (Otolemur garnettii) 95.98% sabaeus) Crab-eating macaque (Macaca Squirrel (Ictidomys 99.54% 95.75% fascicularis) tridecemlineatus) Pig-tailed macaque (Macaca 99.54% Elephant (Loxodonta africana) 95.50% nemestrina) Drill (Mandrillus leucophaeus) 99.54% Rabbit (Oryctolagus cuniculus) 95.50% Mouse Lemur (Microcebus Gibbon (Nomascus leucogenys) 99.54% 94.76% murinus) Olive baboon (Papio anubis) 99.54% Pig (Sus scrofa) 94.47% Orangutan (Pongo abelii) 99.53% Panda (Ailuropoda melanoleuca) 94.26% Angola colobus (Colobus Ma's night monkey (Aotus 99.35% 94.25% angolensis palliatus) nancymaae) Bolivian squirrel monkey (Saimiri 98.70% Microbat (Myotis lucifugus) 94.17% boliviensis boliviensis) Capuchin (Cebus capucinus 98.52% Ryukyu mouse (Mus caroli) 93.88% imitator) Black snub-nosed monkey 98.32% Cow (Bos taurus) 93.81% (Rhinopithecus bieti) Golden snub-nosed monkey 98.06% Mouse (Mus musculus) 93.78% (Rhinopithecus roxellana) Coquerel's sifaka (Propithecus 97.78% Algerian mouse (Mus spretus) 93.60% coquereli) Tarsier (Carlito syrichta) 97.46% Shrew mouse (Mus pahari) 93.31% Tasmanian devil (Sarcophilus Macaque (Macaca mulatta) 97.32% 93.08% harrisii) Lesser Egyptian jerboa (Jaculus 97.29% Prairie vole (Microtus ochrogaster) 93.03% jaculus) Panda (Ailuropoda melanoleuca) 97.22% Degu (Octodon degus) 92.92% Horse (Equus caballus) 96.87% Rat (Rattus norvegicus) 92.86% Cat (Felis catus) 96.77% Dolphin (Tursiops truncatus) 92.85% Northern American deer mouse Leopard (Panthera pardus) 96.77% 92.82% (Peromyscus maniculatus bairdii) Chinese hamster CHOK1GS Tiger (Panthera tigris altaica) 96.77% 92.69% (Cricetulus griseus) Upper Galilee mountains blind Chinese hamster CriGri (Cricetulus 96.76% 92.69% mole rat (Nannospalax galili) griseus) Dog (Canis lupus familiaris) 96.66% Sheep (Ovis aries) 92.64% Armadillo (Dasypus 96.62% Megabat (Pteropus vampyrus) 91.71% novemcinctus) Long-tailed chinchilla (Chinchilla 96.49% Brazilian guinea pig (Cavia aperea) 91.53% lanigera) Ferret (Mustela putorius furo) 96.40% Tree Shrew (Tupaia belangeri) 90.17% Damara mole rat (Fukomys 96.30% damarensis)

Figure 3.1 ZNF827 gene expression in human tissues from public databases. a. ZNF827 gene expression data in 53 different tissue types obtained from RNA-seq experiments from the GTEx project, and presented as TPM (transcripts per million) in a heatmap. TERF2 gene expression data included as a reference. b. Foetal expression data of ZNF827 from the RIKEN FANTOM 5 Consortium in comparison to ZNF827 expression in foetuses with congenital defects from the NIH Roadmap Epigenomics Mapping Consortium. Data presented as TPM in a bar graph. c. ZNF827

gene expression in foetal and juvenile intestinal tissues from a study (Kraiczy et al., 2019) accessed at the Expression Atlas. Data presented as TPM in a bar graph.

3.3 ZNF827 mutations and clinical phenotypes

To gauge the physiological significance of ZNF827, the DECIPHER (DatabasE of genomiC varIation and Phenotype in Humans using Ensembl Resources) database and the GWAS (Genome-Wide Association Studies) catalogue were searched for

ZNF827 genomic variants and the associated clinical phenotypes. Out of 16 patient records, 8 had recorded phenotypes associated with both heterozygous gain and loss of ZNF827. All phenotypes involved developmental and growth defects, including intellectual disability, deformities of facial features, and cardiac structural malformations, as summarised in Table 3.2. As a result, ZNF827 is categorised as a gene likely to exhibit high haploinsufficiency, meaning heterozygous loss of function mutations are not tolerated (DECIPHER, (DECIPHER, 2018)). Therefore, ZNF827 is likely to play a role in developmental processes at early stages of life.

In the GWAS accessed from the GWAS Catalog by EMBL-EBI, single nucleotide polymorphisms (SNPs) of ZNF827 has been annotated to be linked to a range of different traits including liver enzyme levels, susceptibility to dental cavities, heel bone densities, behavioural disinhibition, immune response induced interferon α secretion, coronary artery diseases, prostate-specific antigen levels, response to carboplatin in endothelial ovarian cancers, and breast cancer (Table 3.3)

(GWAS_Catalog, 2018) . All these associations are circumstantial and unvalidated, but point to a possible role of ZNF827 in physiological processes and the development of certain diseases.

Table 3.2 Clinical phenotypes associated with ZNF827 gain and loss from the DECIPHER database Genotype / Patient Sex Size Inheritance Phenotypes Class Deeply set eye; Downslanted Heterozygous palpebral fissures; Heterochromia 249541 unknown 2.39 Mb Unknown Deletion iridis; Intellectual disability; Nasal speech; Protruding ear Abnormality of the thumb; Frontal Heterozygous De novo 248279 other 10.11 Mb bossing; Low-set ears; Pulmonic Deletion constitutive stenosis; Ventricular septal defect Inherited from Heterozygous 251350 46XY 3.09 Mb parent with similar Intellectual disability Duplication phenotype to child Heterozygous De novo Atrial septal defect; Atrioventricular 253743 46XX 4.88 Mb Deletion constitutive canal defect; Blue sclerae Heterozygous 257894 46XX 17.42 Mb Unknown Ectopic anus; Intellectual disability Duplication Maternally Heterozygous inherited, Generalized-onset seizure; Mild 276281 46XY 22.58 Mb Duplication constitutive in global developmental delay mother Aplasia/Hypoplasia of the Heterozygous De novo thumb; Global developmental 285956 46XX 57.82 Mb Duplication constitutive delay; Growth delay; Pulmonary arterial hypertension; Strabismus Heterozygous De novo 290268 unknown 100.16 kb Intellectual disability; Psychosis Deletion constitutive

Table 3.3 ZNF827 SNPs and associated traits in GWAS studies Reported Traits SNP ID SNP type GWAS studies Kanai M, Kim YJ, Chambers Liver enzyme levels (AST, GGT and rs4835265 JC (Kanai et al., 2018, Kim et intron variant ALT) rs4547811 al., 2011, Chambers et al., 2011) Immune response to smallpox (secreted (Kennedy et al., 2012) rs2048161 intron variant interferon α) Kennedy RB rs11100904 intergenic (Zeng et al., 2013, Shaffer et Dental caries rs723794 variant al., 2013) Zeng Z, Shaffer JR rs10022648 (Kim, 2018, Kemp et al., 2017) Heel bone mineral density intron variant rs6816078 Kim SK, Kemp JP (van der Harst and Verweij, rs4345206 Coronary artery disease intron variant 2018) (Verweij et al., 2017) rs35879803 Van der Harst P, Verweij N Non-substance use behavioural disinhibition including symptoms of rs1027841 intron variant (McGue et al., 2013) McGue M conduct disorder and aggression intergenic (Hoffmann et al., 2017) Prostate-specific antigen levels rs56935123 variant Hoffmann TJ Response to carboplatin in epithelial rs17806780 intron variant (Gao et al., 2018a) Gao B ovarian cancer synonymous (Michailidou et al., 2017) Breast cancer rs6818581 variant Michailidou K 91

3.4 ZNF827 has distinct transcriptional roles in ALT and telomerase cells

The data mining that we conducted on ZNF827 using public databases hinted

that ZNF827 may have other non-telomere related functions. However, there have been no other studies on ZNF827 thus far, besides the study by our lab demonstrating its roles in concert with NuRD at ALT telomeres. Here, we investigated whether

ZNF827 plays a role in transcriptional regulation by using RNA sequencing analysis following overexpression of either wild-type ZNF827 or the ΔRRK mutant that is deficient in NuRD recruitment. Exogenous ZNF827 and the ΔRRK mutant were transiently expressed in the ALT cell line WI38-VA13/2RA, and the telomerase positive cell line 293T. Exogenous expression was validated by qRT-PCR and western blot

analysis (Figure 3.2 a, b).

Surprisingly, differential gene analysis showed ZNF827 overexpression in the

ALT cell line WI38-VA13/2RA had little effect on the cellular transcriptional profile, with

only one single gene differentially expressed, the non-functional pseudogene FAAHP1

(fatty acid amide hydrolase pseudogene 1) (Figure 3.2 c). In stark contrast, a total of

573 genes were affected following ZNF827 overexpression in the telomerase positive cell line 293T, with 313 upregulated genes and 260 downregulated genes (Figure 3.2 c). It has previously been shown that exogenously expressed ZNF827 colocalises predominantly at telomeres in ALT cells, but displays pan-nuclear distribution in telomerase positive cells (Conomos et al., 2014). This intriguing observation suggests specific and differential binding of ZNF827 in ALT and telomerase-positive cells. Our gene expression data support ZNF827 having telomeric predominance in ALT cells, with potentially a broader role in transcriptional regulation in non-ALT cells.

Interestingly, in the ALT cell line WI38-VA13/2RA, overexpression of the ΔRRK

mutant resulted in 114 differentially expressed genes, contrasting the marginal effect

of wild-type ZNF827 overexpression (Figure 3.2 c). One plausible explanation for this

notable difference may be dominant negative effects of the ΔRRK mutant, which is

retained at ALT telomeres but is deficient in NuRD recruitment, thereby disrupting the

sequestration of NuRD to telomeres and allowing it to perform transcriptional activity

at non-telomeric regions of the genome. Differentially expressed genes from the 293T

samples with ZNF827 and ΔRRK mutant overexpression as well as the WI38-

VA13/2RA samples with ΔRRK mutant overexpression were subjected to canonical

pathway analysis in the Ingenuity Pathway Analysis (IPA) program (Qiagen). The

WI38-VA13/2RA samples with ZNF827 overexpression were omitted due to the lack

of differentially expressed genes. Distinct pathways were affected between the ALT

and telomerase-positive cell lines (Figure 3.2 d), in concordance with cell type specific

transcriptional activity of the NuRD complex (Allen et al., 2013, Bowen et al., 2004).

While there were some common canonical pathways shared in ZNF827 and ΔRRK

mutant samples in 293T, the impact on the transcriptional landscape was significantly

more profound with ZNF827 overexpression, demonstrated by the greater number of

affected pathways compared to samples with ΔRRK mutant overexpression (Figure

3.2 d). These differences may be due to the inability of the ΔRRK mutant to recruit

NuRD, and the fundamental differences in transcriptional regulation by NuRD-ZNF827 vs ZNF827. Taken together, these data demonstrate that ZNF827 influences gene expression profiles disparately in different cell types, with a more prominent effect in telomerase positive cells, and that the effects are partly dependent on the NuRD complex.

Figure 3.2 ZNF827 have distinct transcriptional roles in ALT and telomerase cells. a. qRT-PCR analysis confirming ZNF827 overexpression following 48-hour transfection of wild-type ZNF827 and the ΔRRK mutant in WI38-VA13/2RA (ALT) and 293T (telomerase). ZNF827 expression normalised to internal control. Data presented as mean ± SD in a bar graph from three biological replicates. **** P<0. 0001 by two tailed t test. b. Western blot validation of ZNF827 protein overexpression following 48- hour transfection of wild-type ZNF827 and the ΔRRK mutant in WI38-VA13/2RA (ALT) and 293T (telomerase). Representative image from three biological replicates. c. Differentially expressed genes from RNA-seq analysis presented as volcano plots in WI38-VA13/2RA (ALT) and 293T (telomerase) with ZNF827 or ΔRRK mutant overexpression. Analysis performed by Erdahl Teber on technical duplicates of three biological replicates (n=6), with empty vector as control. Green dots are significantly differentially expressed genes with absolute value of fold change (FC) ≥ 1.5 (Log2FC ≤ - 0.585 or ≥ 0.585) and -log10(adjusted p value) > 1.3 (adjusted p value <0.05). Tables show the numbers of genes upregulated or downregulated. d. Canonical pathway analysis on the differentially expressed genes in WI38-VA13/2RA (ALT) with ΔRRK mutant overexpression, and 293T (telomerase) with ZNF827 or ΔRRK mutant overexpression presented as a heatmap. Significant pathways have -log(p value) > 1.3 and darker colour shows greater -log(p value). White spaces indicate pathways not involved in the corresponding treatment.

3.5 Multiple transcriptional roles of ZNF827 in the immune response, embryonic and organ development, cellular functions and cancer

Canonical pathway analysis by IPA identified significantly affected pathways using a significance cut-off of -log(p value) > 1.3 based on the differentially expressed genes, but the activity pattern – whether the pathway is up or downregulated, for most pathways was not able to be determined. Most of the canonical pathways affected in the ALT cell line WI38-VA13/2RA with overexpressed ΔRRK mutant were related to the immune response. Notably the pathway most significantly affected was interferon signalling, which coincides with a GWAS that reported an association between

ZNF827 and the secretion of interferon α (IFN-α) (Figure 3.3 a). In fact, the top 10 activated upstream regulators consisted of cytokines, including the interferon α group, other interferon molecules, as well as transcription regulators IRF7 (interferon regulatory factor 7), IRF3 and STAT1 that are involved in the transcriptional regulation of immune response related genes (Figure 3.3 b). The most inhibited upstream regulators were more functionally diverse, including transcription regulators NKX2-3,

TRIM24 and SIRT1, kinases MAPK1 and BTK, transmembrane receptors PTGER2,

IL10RA and ACKR2 and the cytokine IL1RN (Figure 3.3 b). These data implicate

ZNF827 in cellular immune response pathways. Recently, it has been reported that

ALT cells have a defective interferon response to extrachromosomal DNAs due to muted STING expression (Chen et al., 2017). We found that ZNF827 knockdown further suppressed STING expression (Figure 3.4 a). STING expression can be activated by STAT1 (Nishikawa et al., 2018), an upstream regulator that may be activated by ZNF827 as shown in the RNA-seq data. Taken together, ZNF827 may have a role in the STING-mediated cytosolic DNA sensing pathway in ALT cells.

In the telomerase positive 293T cell line, ZNF827 significantly influenced a diverse range of pathways, including embryonic stem cell development and pluripotency, organ development, cellular proliferation and development, immune response related pathways and cancer signalling (Figure 3.3 c). Interestingly in the

ΔRRK mutant counterpart, the transcriptional impact on most of these pathways was abrogated, suggestive of its NuRD dependent transcriptional regulation (Figure 3.3 c).

A similar pattern was observed in the activated or inhibited upstream regulators

(Figure 3.3 d).

Multiple canonical pathways linked ZNF827 to transcriptional regulation in embryonic stem cells and during organ development, which resonates with the

96 developmental abnormalities associated with ZNF827 mutations (Table 3.2).

Signalling pathways related to the immune system, particularly the inflammatory response, were also significantly affected, as well as those for hepatic diseases, cellular proliferation and cancer. All these pathways agree with the phenotypic traits associated with ZNF827 SNPs in the GWAS (Figure 3.3 c, Table 3.3). Exogenous expression of both ZNF827 and the ΔRRK mutant led to the activation of major pleiotropic transcriptional regulators involved in various cellular processes such as

NFkB, TNF, ERK and EGF, reinforcing the role of ZNF827 as a transcription factor

(Figure 3.3 d). Similarly, the predicted top diseases and functions associated with

ZNF827, based on the differentially expressed genes, involved immune responses, organ development, cellular proliferation and survival, developmental disorders and cancers (Figure 3.3 e).

Although ZNF827 seems to have several transcriptional functions, evidence from the RNA-seq analysis and data mining most strongly implicate ZNF827 in the transcriptional regulatory network in early life development, likely in cooperation with the NuRD complex. In support of ZNF827 playing an important role in early development, ZNF827 has been identified in a PhD thesis as a novel regulator of epithelial to mesenchymal transition (EMT), a vital process in normal embryonic development (Sahu, 2017). Our data showed that SNAI1 a key regulator of EMT, was significantly upregulated by ZNF827 overexpression in 293T with a fold change of 2.39, and the upregulation was validated by qRT-PCR (Figure 3.4 b). These data further strengthen a role for ZNF827 in embryonic development. The transcriptional role of

ZNF827 in ALT cells seem to be less important. The fact that ZNF827 appears to have distinct roles in different cell types suggests that it is unlikely to be essential in

97 fundamental cellular processes, and perhaps more likely to exert a temporal function in transcriptional regulation.

Figure 3.3 Multiple transcriptional roles of ZNF827. a. Canonical pathway analysis in WI38-VA13/2RA overexpressing ΔRRK mutant with significance threshold -log (p value) > 1.3. b. Top 10 upstream regulators activated (z score > 2) and inhibited (z score <-2) in WI38-VA13/2RA overexpressing ΔRRK mutant. c. Top 20 canonical pathways involved in 293T overexpressing ZNF827 with -log (p value) > 1.3, compared to the corresponding pathways in 293T overexpressing ΔRRK mutant. Bars below the threshold indicates insignificance. d. Top 20 activated upstream regulators (z score >2) and inhibited upstream regulators (z score < -2) involved in 293T overexpressing ZNF827, in comparison to the corresponding values in 293T overexpressing ΔRRK mutant. * indicates z score below threshold, grey spaces indicate regulators not involved in the mutant samples. e. Top 5 affected diseases and functions ranked according to the significance score calculated by IPA in WI38-VA13/2RA overexpressing the RRK mutant, and 293T cells overexpressing ZNF827 or the RRK mutant.

Figure 3.4 STING and SNAI1 expressions are affected by ZNF827. a. qRT-PCR analysis confirming ZNF827 knockdown following 72 hr siRNA transfection in WI38- VA13/2RA and U-2 OS (ALT) (left). qRT-PCR analysis of STING expression following 72 hr siRNA knockdown of ZNF827 in WI38-VA13/2RA and U-2 OS (ALT) (right). ZNF827 expression normalised to internal control. Data presented as mean ± SD in a

100 bar graph from three biological replicates. **** P<0. 0001 by two tailed t test. b. qRT- PCR analysis of SNAI1 following 48 hr ZNF827 overexpression in WI38-VA13/2RA (ALT) and 293T (telomerase). * P<0. 05, ** P<0. 005 by two tailed t test.

3.6 ZNF827 may be a self-regulatory transcription factor

ChIP sequencing (ChIP-seq) analysis was used to investigate whether ZNF827 binds to non-telomeric regions of the genome, and to identify potential binding motifs.

To be consistent with the RNA-seq experiment, ChIP-seq was performed in WI38-

VA13/2RA and 293T in two biological replicates following transient ZNF827 overexpression (Figure 3.5 a). The ZNF827 ChIP-seq assay was validated by telomere-ChIP which showed increased binding of ZNF827 to telomeres accompanied by decreased TRF2 binding, consistent with previous findings by our lab (Figure 3.5 b) (Conomos et al., 2014). Unfortunately, ChIP-seq analysis was limited by poor read depth and quality. All the 293T samples were excluded after quality filtering. Despite limitations with the sequencing read quality, peak analysis was conducted on the

WI38-VA13/2RA samples that passed quality filtering. There were 15 and 10 peaks identified in the samples with endogenous ZNF827 and overexpressed ZNF827, respectively. Interestingly, all the peaks except one were located in the intronic regions of the ZNF827 gene and the neighbouring loci RP11-18K12.1/2 (Figure 3.5 c). This indicates that ZNF827 binds to its own regulatory regions and may be a self-regulatory transcription factor. We were not able to determine the binding motifs because the limited number of peaks were insufficient for robust motif analysis.

101

102

Figure 3.5 ZNF827 may be a self-regulatory transcription factor. a. Western blot analysis confirming ZNF827 overexpression following 48 hr transient transfection in WI38-VA13/2RA and 293T. b. Telomere-ChIP against ZNF827, TRF2 (positive control), and IgG (negative control) in WI38-VA13/2RA and 293T 48 hr after ZNF827 overexpression (top panel). Quantitation of telomere-ChIP data (bottom panel). Data are mean ± range, n = 2 independent experiments. c. Peaks summary from ZNF827 ChIP-seq analysis of the WI38-VA13/2RA samples demonstrating ZNF827 binding to the introns of the ZNF827 gene and its neighbour RP11-181K12.1/2. There are 15 overlapped peaks between the two replicates with endogenous, and 10 overlapped peaks between the two replicates with ZNF827 overexpression, with IDR (irreproducible discovery rate) <0.05. ChIP-seq analysis was performed by Emilie Wilkie.

3.7 Discussion

The zinc finger protein ZNF827 is a promising novel molecular target in cancer cells utilising the ALT pathway, as demonstrated previously by our lab (Conomos et al., 2014). In support of characterising ZNF827 as a therapeutic target for the development of novel anticancer strategies, this study has expanded our knowledge of this poorly known zinc finger protein by describing its biological functions as a transcription factor. Through RNA-sequencing and ChIP-sequencing studies, augmented by data gathered from public databases, we have revealed that ZNF827 acts as a transcription factor, and is involved in transcriptional regulation during embryonic and early organ development, and as part of the immune response. Our data suggest that ZNF827 plays a critical transcriptional role in developmental processes at early stages of life, but becomes less functionally important later in development, supporting ZNF827 as a therapeutic target in cancers. We have also demonstrated that the role of ZNF827 at ALT telomeres is independent of its transcriptional function.

103

The RNA-seq data implicating ZNF827 in vital developmental processes at early stages of life show great concordance with data obtained from public databases.

Specifically, tissue expression data indicate that ZNF827 displays an early development dependent expression pattern, with mRNA levels being much higher in foetal tissues compared to adult tissues. Clinical phenotypes associated with ZNF827 mutations entail a range of developmental disorders including eye, ear and skull malformation, heart defects, global developmental delay and intellectual disability, which confirm the vital function of ZNF827 in early life development. In contrast, SNP data collected from the adult population only weakly associate ZNF827 with mild medical concerns such as liver enzyme levels, immune response to small pox, and dental caries, suggesting that ZNF827 is unlikely to have vital functions in normal somatic cells.

The transcriptional roles of ZNF827 that we identified are in remarkable parallel with those of other RRK motif containing zinc finger transcription factors, which include cell and organ development, embryonic stem cell pluripotency and differentiation

(Table 1 in the Introduction). It is not uncommon that transcription factors that are essential in early life development are transcriptionally suppressed later in life. For example, the RRK motif-containing transcription factor SALL4, which is pivotal in the regulation of embryonic stem cell pluripotency, becomes silenced in fully differentiated

cells (Tatetsu et al., 2016). Similarly, the low expression of ZNF827 in adult tissues

may be attributed to coordinated transcriptional silencing to ensure normal

development. Considering that all the reported developmental disorders are

associated with heterozygous ZNF827 mutations, it is very likely that homozygous

mutations or loss of ZNF827 may confer embryonic lethality. Our data have provided

sufficiently strong evidence implicating ZNF827 in embryonic and organ development

104 to warrant future in vivo knockout studies to delineate the precise functions of ZNF827 in embryonic development.

Despite a high amount of poor quality reads that led to the exclusion of more than half of the samples from the ChIP-seq analysis, , the remaining data showed that

ZNF827 binds to the intronic regions of the ZNF827 gene, suggesting autoregulation, which is a common characteristic of transcription factors involved in development

(Bateman, 1998, Ngondo and Carbon, 2014, Crews and Pearson, 2009). The universally low expression of ZNF827 in adult tissues and cell lines may be a result of negative autoregulation.

For this study, we mainly focused on the role of ZNF827 at telomeres, and the potential utility of ZNF827 in cancer therapeutics. Dysregulated expression of SALL4 as well as many other C2H2 zinc finger proteins have been correlated to tumorigenesis

(Jen and Wang, 2016, Tatetsu et al., 2016). Thus far there have been no reports or literature that link ZNF827 to the development of cancer, apart from the publication from our lab demonstrating the involvement of ZNF827 at telomeres in ALT cancers.

ZNF827 gene expression was relatively low and variable amongst the small panel of

ALT and telomerase-positive cell lines tested, and there was no clear correlation with

TMM status (Conomos, 2014).

Our RNA-seq data analysis showed some evidence that may implicate ZNF827 in cancer development. Specifically, the most activated upstream regulators in 293T cells included key mediators of major signalling pathways for cell proliferation and survival, including NFkB, TNF, ERK and EGF, that have been strongly correlated with different types of cancers. In addition, colorectal and ovarian cancer signalling pathways were significantly impacted by ZNF827 overexpression. We identified

105 upregulation of SNAI1, the master regulator of epithelial to mesenchymal transition

(EMT), by ZNF827. This finding was consistent with data presented in a recent PhD thesis that identified ZNF827 as an important mediator of the alternative splicing of numerous genes during EMT in normal development and cancer metastasis (Sahu,

2017). By driving EMT, SNAI1 exerts a vital function in embryonic development and cancer metastasis (Wang et al., 2013). These associations support future studies to

explore the effect of ZNF827 on cancer cell growth and metastasis.

The minor impact that ZNF827 overexpression had on the transcriptional profile

in WI38-VA13/2RA indicates that the function of ZNF827 at ALT telomeres, and in ALT

cells, is not facilitated through its transcriptional role. This is consistent with the

distinctly predominant localisation of ZNF827 at telomeres in ALT cancer cells

(Conomos et al., 2014). In contrast, the marked impact of the dominant negative RRK

mutant suggests the transcriptional roles of ZNF827 are dependent on the NuRD

complex. It is also possible that the RRK mutant recruits other transcriptional

regulators with different functional domains even though it is deficient in NuRD binding.

The majority of the pathways affected by the RRK mutant are related to the immune

response, corresponding to activated upstream regulators that include multiple

interferons and STAT1, a transcriptional activator essential in mediating interferon

signalling (Lim and Cao, 2006).

A recent study demonstrated that induction of extrachromosomal telomeric

repeat (ECTR) DNA in human fibroblasts activates the cGAS-STING (cyclic GMP-

AMP synthase-stimulator of interferon genes) cytosolic DNA sensing pathway, subsequently triggering an antiproliferative interferon response (Chen et al., 2017). It was also found that this cytosolic DNA sensing pathway is defective in ALT cancer cells due to repressed STING mRNA and protein expression, allowing ALT cells to

106 evade ECTR-induced growth defects. Interestingly, we have demonstrated here that

ZNF827 depletion led to further suppression of STING mRNA expression. Another recent study found that STING gene expression is positively regulated by STAT1 directly binding to the STING promoter region (Nishikawa et al., 2018). Our RNA-seq data identified STAT1 as an upstream regulator potentially regulated by ZNF827.

Taken together, ZNF827 may be involved in an immune response signalling network with STING and STAT1, and may potentially contribute to the regulation of the cytosolic DNA sensing pathway in ALT cells. Our RNA-seq analysis also identified

SIRT1 as an upstream regulator inhibited by RRK mutant overexpression in WI38-

VA13/2RA. SIRT1 is a histone deacetylase that can interact and activate HDAC1, the histone deacetylase subunit in the NuRD complex (Dobbin et al., 2013). SIRT1 associates with telomeres in mouse embryonic fibroblasts, and promotes homologous

(HR) recombination events at telomeres and throughout the genome (Palacios et al.,

2010). Given these associations between ZNF827 and SIRT1, it would be of interest to explore further whether SIRT1 affects the roles of ZNF827-NuRD at telomeres.

The major caveat of this study was the unequal transient ZNF827

overexpression between the cell lines, with ZNF827 expression more than 50-fold

higher in 293T cells compared to WI38-VA13/2RA cells. There was no significant difference between the expression levels of ZNF827 and the RRK mutant in 293T cells, while the RRK mutant was expressed approximately 3-fold higher than ZNF827 in

WI38-VA13/2RA cells. The low transcriptional impact of ZNF827 observed in WI38-

VA13/2RA cells may be partly attributed to the lower ZNF827 overexpression. To eliminate bias caused by unequal protein expression, the transfection protocol could be optimised to try to attain similar levels of protein expression between different cell

107 lines. It would also be interesting to see if higher ZNF827 expression would lead to a greater transcriptional impact.

To validate ZNF827 as a promising therapeutic target, it is of fundamental importance to demonstrate tolerance of ZNF827 inhibition in normal somatic cells. In this study, the RNA-seq experiment was conducted with ZNF827 overexpression instead of knockdown due to the consideration that ZNF827 is expressed at very low baseline levels. Retrospectively, the study would have been more complete and informative if ZNF827 knockdown samples and a mortal cell line were included. To support ZNF827 as a viable molecular target, the circumstantial evidence in this study will need to be further reinforced by future studies investigating the effect of ZNF827 inhibition in both mortal and cancer cell lines, such as RNA-seq experiments and cell proliferation studies with ZNF827 depletion. Nevertheless, this study has provided valuable insights into the biological functions of ZNF827 that justify further investigation to validate its physiological roles in cellular processes, such as its role as a novel regulator in embryonic development.

Unfortunately, the ChIP-seq experiment was not able to provide insight into the binding motifs of ZNF827. The low sequencing read quality may be due to technical issues. In retrospect, the low read quality may be attributable to later findings that

ZNF827 preferentially binds to ssDNA rather than dsDNA (Chapter 5). This has

significant implications in the design of a ChIP-seq experiment. Our ChIP-seq experiment was based on a library preparation of dsDNA, which means ssDNA was excluded. Therefore, it may be pertinent to revisit the ChIP-seq approach in the future using a ssDNA-specific library preparation.

108

In summary, in this chapter we have gained insights into the biological functions of ZNF827 as a transcription factor. We have provided evidence to support that

ZNF827 plays a critical role in embryonic and organ development as well as being involved in immune response-related pathways, cancer cell growth and metastasis.

Further, we conclude that the functions of ZNF827 at ALT telomeres are not related to its role as a transcription factor.

109

Defining the Molecular Interactions of ZNF827- NuRD and ZNF827-Nuclear Receptors, and the Role of ZNF827 SUMOylation at Telomeres

Sections 4.3 to 4.7 in this chapter are adapted from our manuscript titled ‘Structural

and functional characterization of the RBBP4-ZNF827 interaction and its role in NuRD recruitment to telomeres’ published in Biochemical Journal (Yang et al., 2018) with minor changes. The published work is a collaborative effort with Professor Yunyu Shi’s group. For the completeness of the chapter, the structural data from Professor Shi’s group have been included. Aiai Sun and Fudong Li from Shi’s group performed the structural biology experiments described in sections 4.3 to 4.5. I, the PhD candidate and first author of the published manuscript conducted the molecular biology and functional experiments described in sections 4.6 and 4.7.

4.1 Introduction

As a transcription factor, ZNF827 has been revealed by this study to be involved in several cellular processes, most notably in embryonic and organ development, and immune response related pathways. The role of ZNF827 in telomere biology, and specifically the ALT pathway, appears to be separate from these cellular processes.

In support of this, the predominant localisation of ZNF827 at telomeres has no transcriptional impact on ALT cells. Through its specific association with telomeres,

ZNF827 plays a unique and multifaceted role in regulating multiple mechanistic elements of ALT, making it a promising molecular target for the development of therapeutics for the 10-15% cancers utilising ALT (Conomos et al., 2014). In the

110

interest of characterising ZNF827 as a molecular target, it is critical to fully understand

the molecular and mechanistic details of ZNF827 biology, including the recruitment of

ZNF827 to telomeres by nuclear receptors COUP-TF II and TR4, the recruitment of

NuRD by ZNF827, and the posttranslational modifications of ZNF827 required for its

recruitment and activity at telomeres, before successful strategies targeting ZNF827

in ALT cancers can be designed.

NuRD is a specialised chromatin remodelling complex comprised of six core

subunits (HDAC1/2, CHD3/4, MTA1/2/3, RBBP4/7, MBD2/3 and GATAD2A/B) (Allen

et al., 2013), with the HDAC and CHD components being the two catalytic subunits

enabling its histone deacetylase and ATP-dependent chromatin remodelling activities, respectively (Denslow and Wade, 2007, Tong et al., 1998, Xue et al., 1998, Wade et al., 1998, Zhang et al., 1998). The other subunits are responsible for directing the functional specificity of NuRD by facilitating direct interactions with DNA and other cofactors (Allen et al., 2013). The diverse compositions of NuRD give rise to broad cellular roles including transcription, chromatin assembly, cell cycle progression, embryonic stem cell (ESC) regulatory pathways, DNA replication and the DNA damage response (DDR), which paradoxically implicate NuRD in both the suppression and promotion of tumorigenesis (Lai and Wade, 2011, Hu and Wade, 2012).

Our lab has identified ZNF827 as a novel interactor of NuRD and linked NuRD to ALT telomeres (Conomos et al., 2014). The multifaceted role of ZNF827 in the ALT pathway is mainly mediated through the action of NuRD. NuRD is recruited to ALT telomeres by ZNF827, where it remodels the telomeric chromatin by displacing shelterin binding, and counteracts the resulting chromatin decompaction through

HDAC-mediated histone hypoacetylation (Conomos et al., 2014). NuRD then recruits

HR proteins, such as BRIT1 and BRCA1, which facilitate homology-directed telomere

111

synthesis. Preferential association of NuRD with ALT telomeres during the G1/S phase

of the cell cycle promotes telomere-telomere associations, visualised in cytogenetic

preparations as telomere bridges (Conomos et al., 2014). Importantly, depletion of

ZNF827 inhibits NuRD localisation to telomeres and causes loss of cell viability

specifically in cancer cell lines that use the ALT pathway (Conomos et al., 2014). This makes the interaction between ZNF827 and NuRD a compelling focus for the future development of ALT cancer therapeutics.

ZNF827 is a C2H2 zinc finger protein that recruits NuRD through a conserved

N-terminal RRKQXXP domain that it shares with other zinc finger proteins such as

FOG1, SALL1 and BCL11B (Hong et al., 2005, Lauberth and Rauchman, 2006,

Cismasiu et al., 2005, Topark-Ngarm et al., 2006). Similarly, conserved domains within

histone H3 and plant homeodomain finger 6 (PHF6) also interact with NuRD (Todd and Picketts, 2012, Liu et al., 2015, Schmitges et al., 2011). It has previously been

shown that the N-terminal domain of FOG1 can interact with both MTA1 and RBBP4

components of the NuRD complex (Lejon et al., 2011). The potential for multi-subunit

interactions, combined with the sequence similarity of the FOG1 and ZNF827 N-

terminal domains, precipitated the proposition that telomere tethering is achieved by

multiple components of the NuRD complex simultaneously interacting with ZNF827

molecules bound at different telomeres (Conomos et al., 2013). The interacting

interface between ZNF827 and NuRD is yet to be defined. Based on the similarity between ZNF827 and FOG1, we hypothesise that ZNF827 interacts with the RBBP4 or MTA1 subunit of the NuRD complex.

The recruitment of ZNF827 to ALT telomeres is dependent on orphan nuclear receptors COUP-TF II and TR4, which bind to TCAGGG variant repeats that are interspersed throughout ALT telomeres (Conomos et al., 2014, Conomos et al., 2012).

112

However, it is not known how COUP-TF II and TR4 recruit ZNF827 to ALT telomeres

and whether ZNF827 binds directly to these receptors. COUP-TF II is implicated in

cancer development regardless of the context of ALT. COUP-TF II is highly expressed in the mesenchyme of various organs and has been shown to play important roles in the development of blood and lymphatic vessels, and is involved in metastatic tumour

progression (Pereira et al., 2000, Xu et al., 2015). TR4 is found in erythroid cells, liver,

spermatocytes and some regions of the brain, and has functions in lipid and lipoprotein

regulation, spermatogenesis and the development of the central nervous system (Lee

et al., 2002). There are data that suggest TR4 plays distinct roles in prostate cancer.

Specifically, TR4 can both inhibit initiation of prostate cancer via regulation of the ATM

DNA repair pathway, and conversely promote metastasis of prostate cancer once it

has developed (Lin et al., 2015). Interestingly, our data analysis presented in Chapter

3 identified that ZNF827 has the highest expression in prostate glands.

COUP-TF II and TR4 belong to a subfamily of the nuclear receptor super family

known as the orphan receptors, because specific endogenous ligands to these

proteins have yet to be identified (Kliewer et al., 1999). Like other nuclear receptors,

they contain a highly conserved DNA binding domain containing two zinc finger motifs

and a putative ligand-binding domain (LBD) (Lee et al., 2002, Pereira et al., 2000).

Both COUP-TF II and TR4 bind efficiently as homodimers to DNA response elements

containing two direct repeats of the RGGTCA consensus motif with spacing of one or

more nucleotides (Cooney et al., 1992, Hirose et al., 1995). The RGGTCA motif shares

high sequence similarity to the C-type telomeric variant repeat TCAGGG, to which

COUP-TF II and TR4 display high binding affinity (Conomos et al., 2012).

Crystal structures of the LBD of COUP-TF II and TR4 share high similarities,

and reveal a charge clamp pocket for coactivator binding as well as a ligand binding

113

pocket (Kruse et al., 2008, Zhou et al., 2011). These studies have shown that both

ligand binding and coactivator binding are required for full activity of COUP-TF II and

TR4, and that their activation is driven by sequential binding of a ligand and a

coactivator, with the former inducing a conformational change exposing the coactivator

binding pocket (Kruse et al., 2008, Zhou et al., 2011). Coactivators interact with the

receptors at LxxLL motifs. This motif is also present in ZNF827 (Kruse et al., 2008,

Zhou et al., 2011). This has led to our hypothesis that ZNF827 directly binds to COUP-

TF II and TR4 through its LxxLL motif(s) and acts as a coactivator. If this hypothesis

is correct, the interface of these protein-protein interactions can be exploited in the

development of ZNF827 targeting strategies to prevent ZNF827 recruitment to

telomeres.

ZNF827 can undergo SUMOylation at five lysine residues, according to data

published in two recent MS-based proteomic screens that completed site-specific

mapping of the global SUMOylation networks in human cells (Hendriks et al., 2014,

Hendriks et al., 2015). SUMOylation is a type of reversible posttranslational

modification (PTM) of lysine residues in proteins by small ubiquitin-like modifiers

(SUMOs), which are involved in a range of cellular processes encompassing cell cycle progression, transcriptional regulation, chromatin remodelling and DNA repair (Flotho and Melchior, 2013, Hickey et al., 2012, Ulrich and Walden, 2010). SUMOs are found to be enriched in PML nuclear bodies (Hendriks et al., 2014, Hendriks et al., 2015). In relation to telomere biology, evidence suggests that SUMOylation of the shelterin proteins TRF1, TRF2, TIN2 and RAP1 is necessary for the recruitment of telomeres to PML nuclear bodies, and thus the formation of APBs, which are thought to be sites of telomere recombination and extension (Potts and Yu, 2007). ZNF827 is shown to be present in APBs, where it promotes telomere-telomere interactions in a cell cycle

114

dependent manner (Conomos et al., 2014). We therefore also hypothesise that

SUMOylation of ZNF827 is required for its ALT promoting activity.

To address the hypotheses proposed above, in this chapter we have shown

that recruitment of ZNF827 to telomeres does not involve a direct interaction between

ZNF827 and nuclear receptors, and may instead involve suppression of transcriptional

activity at telomeres. In collaboration with Professor Yunyu Shi’s group, we have

defined the interaction interface between ZNF827 and NuRD by resolving the X-ray

crystal structure of the complex formed between the NuRD component RBBP4 and

ZNF827. We have also identified the precise residues responsible for binding

specificity, and performed functional studies to demonstrate that disruption of these

residues directly impedes RBBP4 binding to ZNF827 both in vitro and in vivo, and

inhibits RBBP4 binding to telomeres. Lastly, we have used mutational studies to

demonstrate that SUMOylation of ZNF827 is required to promote ALT activity.

4.2 ZNF827 recruitment to telomeres does not depend on direct interaction with

TR4 and COUP-TF II, but may involve suppression of transcription at telomeres

Given that the recruitment of ZNF827 to telomeres is dependent on the two orphan nuclear receptors COUP-TF II and TR4, and that ZNF827 contains four putative LxxLL nuclear receptor coactivator binding motifs (Figure 4.1 a), we decided to investigate whether there is a direct interaction between ZNF827 and the nuclear receptors. To achieve this, we used ZNF827 co-immunoprecipitation as well as Halo-

Tag ZNF827 pulldown, following overexpression of ZNF827 in the ALT cell line WI38-

VA13/2RA. As demonstrated in Figure 4.1 b, COUP-TF II and TR4 did not coimmunoprecipitate or pull down with ZNF827, indicating no direct interaction between ZNF827 and the two nuclear receptors.

115

It has previously been shown that full activity of COUP-TF II and TR4 requires

sequential binding of the ligand and the coactivator, in which conformational changes

induced by ligand binding expose the binding pocket for the recruitment of the

coactivator (Kruse et al., 2008, Zhou et al., 2011). Even though endogenous ligands

for COUP-TF II and TR4 have not been identified, exogenous retinoic acid is able to

activate COUP-TF II and TR4 and recruit coactivators (Zhou et al., 2011, Kruse et al.,

2008). We then tested whether ZNF827 binding to COUP-TF II and TR4 as a

coactivator required initial ligand binding. Co-immunoprecipitation (co-IP) experiments

were performed with the addition of all-trans retinoic acid (atRA) (Figure 4.1 c). Protein

bands were detected in the Myc-ZNF827 IP eluates, however, they were proved to be

false positives (antibody bands) (Figure 4.1 c). Thus, no direct interaction was

detected, even in the presence of retinoic acid, following ZNF827 co-IP with either a

direct antibody or a Myc-tag antibody.

Despite the absence of a direct physical interaction, we investigated whether the association between ZNF827 and the nuclear receptors COUP-TF II and TR4 was at the level of transcriptional regulation. Previously, it has been shown that TR4 depletion caused a remarkable increase in ZNF827 expression and vice versa, whereas COUP-TF II exerted no effect (Conomos, 2014). Here, ZNF827

overexpression in both an ALT and a telomerase-positive cell line led to a significant

increase in TR4 expression, whilst COUP-TF II was downregulated (Figure 4.2 a).

Together these data suggest a feedback mechanism between TR4 and ZNF827, and that ZNF827 is a negative regulator of COUP-TF II.

Studies have depicted TR4 as a negative feedback modulator of the retinoid- mediated signalling pathways in which TR4 protein expression was increased by retinoic acid, and in turn suppressed retinoic acid induced transactivation and its own

116

protein expression (Lee et al., 1998, Inui et al., 1999). Likewise, COUP-TF II has been

shown to be a repressor of retinoic acid induced transactivation by the retinoic acid

receptors and retinoid X receptors (RAR/RXR) (Tran et al., 1992, Kliewer et al., 1992,

Cooney et al., 1993). Here, we show that in the ALT cell line WI38-VA13/2RA, atRA

clearly induced a steady increase in TR4 protein levels over time, peaking at 48 hours

followed by a slight decline in the last 48 hours, analogous to the reported negative

feedback control of TR4 on its own expression (Figure 4.2 b) (Lee et al., 1998). In

contrast, there was no change in COUP-TF II protein expression, although a similar

decline was observed in the last 48 hours. No changes were observed in ZNF827

gene expression either. Notably, in the telomerase positive glioblastoma cell line T98G,

atRA treatment resulted in a delayed but gradual increase in protein expression of TR4

as well as COUP-TF II from 48 hours onwards, coinciding with a sustained reduction

in ZNF827 gene expression, which is in support of a negative feedback mechanism

between ZNF827 and the nuclear receptors (Figure 4.2 b).

Since retinoic acid can increase protein expression of TR4 and COUP-TF II,

which bind specifically to ALT telomeres and recruit ZNF827, the effect of retinoic acid

on the telomere binding of these proteins was explored by telomere-ChIP. Intriguingly,

atRA treatment led to an observable increase of COUP-TF II binding to telomeres,

which was accompanied by an increase in binding of the NuRD component RBBP7

(Figure 4.2 c). This agrees with the previous study that found COUP-TF II

overexpression increased NuRD recruitment to telomeres (Conomos et al., 2014). The

increase in TR4 and ZNF827 binding post atRA treatment was only slightly above

background. There was an unexpectedly marked increase in TRF2 binding to

telomeres. These observations may suggest a collaborative effort by the nuclear

receptors, TRF2, the NuRD complex and possibly ZNF827 to suppress retinoic acid

117

induced transactivation at telomeres. As predicted, atRA treatment did not cause increased transcriptional activity at telomeres measured by TERRA levels (Figure 4.2 d). Intriguingly, depletion of ZNF827 led to a slight increase in TERRA, which was significantly enhanced in conjunction with atRA treatment (Figure 4.2 d). The increase in TERRA levels after ZNF827 depletion suggests that ZNF827 plays a role in suppressing transactivation at telomeres. Collectively, our data show that ZNF827 recruitment to telomeres is not via direct binding to nuclear receptors, but may be associated with the transcriptional activity or TERRA levels at telomeres.

Figure 4.1 ZNF827 does not directly interact with COUP-TF II or TR4. a. ZNF827 protein sequence containing four LxxLL coactivator binding motifs (highlighted in green). b. Co-IP with direct ZNF827 antibody in WI38-VA13/2RA overexpressing ZNF827, followed by western blot analysis of ZNF827, COUP-TF II and TR4 (left panel). Halo-Tag pulldown in WI38-VA13/2RA overexpressing Halo-Tag-ZNF827, followed by western blot analysis of ZNF827, COUP-TF II and TR4 (right panel). c. Direct ZNF827 co-IP in the presence of 20μM atRA in WI38-VA13/2RA overexpressing

118

ZNF827, followed by western blot analysis of ZNF827, COUP-TF II and TR4 (left panel). Co-IP with a Myc antibody in the presence of 20μM atRA in WI38-VA13/2RA overexpressing Myc-ZNF827, followed by western blot analysis of ZNF827, COUP-TF II and TR4 (right panel). Eluates were reanalysed by western blotting with Myc antibody controls to check for false positives (gray box, right panel).

Figure 4.2 ZNF827 may be involved in transcriptional suppression at telomeres. a. Real-time RT-PCR showing TR4 (left) and COUP-TF II (right) transcript levels relative to empty vector control in WI38-VA13/2RA and 293T overexpressing ZNF827 (mean ± SD; n = 3 technical replicates; *P<0. 05, ** P<0. 005 by two tailed t test). b. Western blot analysis of COUP-TF II and TR4 protein levels in WI38-VA13/2RA and

119

T98G incubated with 20μM atRA over a 96-hour time course (left panel). Bar graphs below protein bands show protein expressions relative to untreated samples. The protein bands were quantitated by densitometry using Multi Gauge software (Fujifilm). Real-time RT-PCR showing ZNF827 transcript levels relative to DMSO control at corresponding time points in the 96hr time course experiment (right panel, mean ± SD; n = 3 technical replicates; *P<0. 05, ** P<0. 005 by two tailed t test). c. Telomere‐ChIP against TRF2, ZNF827, COUP-TF II, TR4, and RBBP7 in WI38-VA13/2RA (top panel) post 24 hours incubation with 20μM atRA or DMSO control. Quantitation of the percentage of telomeric DNA pulled down (bottom panel). Data are mean ± range, n = 2 independent experiments. d. Quantitative RT‐PCR confirming ZNF827 knockdown 48 hours post‐transfection with siRNA in WI38-VA13/2RA (left panel, mean ± SD; n = 3 technical replicates; ***P<0. 0002 by two tailed t test). TERRA analysis of WI38- VA13/2RA following ZNF827 knockdown and 24 hr 20μM atRA or DMSO incubation. TERRA levels are quantitated relative to untreated sample by ImageQuant TL (GE). Data are presented as mean ± S.E.M.; n = 3 technical replicates (bottom panel).

4.3 ZNF827 binds to the NuRD complex component RBBP4

Next, we set out to define the molecular details of NuRD recruitment to ALT telomeres by ZNF827. Previously our lab has demonstrated that the N-terminal RRK motif of ZNF827 is essential for NuRD binding (Conomos et al., 2014). However, the exact subunit of the NuRD complex involved in the interaction and the interaction interface between ZNF827 and NuRD have not been elucidated. To define the interaction at the structural and molecular level, we collaborated with the structural biology group led by Professor Yunyu Shi. Results in sections 4.3 to 4.7 in this chapter are from a manuscript we have published in collaboration with Professor Shi’s group.

The structural biology experiments described in 4.3 to 4.5 were performed by Aiai Sun and Fudong Li from Professor Shi’s group. I carried out all the molecular biology and functional studies in 4.6 to 4.7.

120

The N-terminal region of ZNF827 contains an RRK motif (Figure 4.3 a), which

shows high sequence similarity to sequences in other proteins, including FOG1 and

PHF6, that have previously been reported to interact with RBBP4. To determine

whether RBBP4 binds directly to the N-terminus of ZNF827, we conducted isothermal

titration calorimetry (ITC) using a synthetic ZNF8271-14 peptide to titrate full-length

RBBP4. The binding assay showed that RBBP4 interacts with the N-terminal ZNF827

peptide with stoichiometry of approximately 1:1 and a dissociation constant (KD) of 1.6

± 0.1 μM (Figure 4.3 b).

We further solved the complex structure of RBBP4 bound to ZNF8271-14 via X-

ray crystallography at a resolution of 1.9 Å (Table 4.1). Among the 14 amino acids in

the ZNF827 peptide, 10 amino acids (residues 3–12) are visible in the final model

(Figure 4.3 c). The structure of RBBP4 (residues 7-410) consists of seven blades

forming a 'velcro' closure β-propeller structure with an additional α-helix at the N-

terminus (Figure 4.3 c). The N-terminal 6 residues, loop region residues 90-111, and

C-terminal residues 411-425 of RBBP4 are not visible in the final electron density

maps. Overall, the structure of the complex shows that the ZNF8271-14 peptide

contacts RBBP4 with a largely extended conformation across the axis channel of the

smaller surface, mainly between blades 7 and 1, but also involves residues from the

loop region of all other blades (Figure 4.3 c).

121

Figure 4.3 The N terminus of ZNF827 interacts with RBBP4. a. Schematic representation of the human ZNF827 protein, including the N-terminal residues 1-14

(ZNF8271-14) and nine C2H2 zinc-finger domains. The C2H2 zinc fingers are shown as gray bars and the N-terminal sequence is expanded as single letter amino acids. b. Isothermal titration calorimetry (ITC) experiment showing the binding affinity of the

ZNF8271-14 peptide with RBBP4. Data were fitted to a one-site binding model using MicroCal PEAQ-ITC Analysis Software, and the calculated binding parameters were H= -57.0 ± 0.4 kJ/mol, and -T S= 23.8 kJ/mol (Table 4.2). c. Two orthogonal views of△ RBBP4 bound to the ZNF827△1-14 peptide. The ZNF8271-14 peptide is shown in green, with RBBP4 in gray. This figure is from our publication in collaboration with Professor Shi’s group (Yang et al., 2018), and the structural biology experiments described in this figure were performed by our collaborators Aiai Sun and Fudong Li.

122

Table 4.1 Data collection and refinement statistics of ZNF8271-14 bound to RBBP4 (Data in this table are from our publication in collaboration with Professor Shi’s group (Yang et al., 2018), and were collected by our collaborators Aiai Sun and Fudong Li.) Data collection

Space group P 21 Cell dimensions a, b, c (Å) 75.96, 59.78, 102.56 α, β, γ (°) 90.00, 94.97, 90.00 Wavelength (Å) 0.9792 Resolution (Å) 50.00 - 1.90 (2.00 -1.90)* Rsym (%) 10.5 (45.8) Rpim (%) 4.5(19.7) Rrim (%) 11.5(50.0) I / σI 9.7(3.3) Completeness (%) 86.20 (80.80) Redundancy 6.0 (6.1) Refinement Resolution (Å) 40.70 - 1.90 No. reflections 62331 (3088) Rwork / Rfree (%) 17.25/21.20 No. of atoms 6602 Protein 6000 Peptide 174 Water 428 B-factors (Å2) 35.21 Protein 34.84 Peptide 46.39 Water 35.88 R.M.S. deviations Bond length (Å) 0.013 Bond angles (°) 1.233 Ramachandran values Most favored (%) 97.26 Additional allowed (%) 2.61 Outliers (%) 0.13 *Values in parentheses are for highest-resolution shell

123

4.4 Defining the binding interactions between ZNF827 and RBBP4

The binding interface of RBBP4 is aspartate and glutamate rich, forming a

highly negatively charged pocket that specifically binds the ZNF827 N-terminal region

(Figure 4.4 a). A network of electrostatic interactions, together with hydrogen bonds, van der Waals and hydrophobic contacts, anchor the ZNF8271-14 peptide into its

binding cleft at the smaller surface of RBBP4 (Figure 4.4 b, c). The side chain of Arg-

4 is sandwiched by Tyr-181 and Phe-321 to form cation-π interactions, whereby the

guanidinium group on Arg-4 forms electrostatic contacts with Glu-231, and a water-

mediated hydrogen bond with Asn-277, together with its main chain carbonyl group

interacting with the Lys-376 side chain in RBBP4. In the neighbouring channel, the

side chain of Lys-5 in ZNF827 is surrounded by the side chains of Leu-45, Tyr-181

and His-71, with its ε-group specifically forming contacts with Glu-126, Asn-128 and

Glu-179 in RBBP4. The side chain and backbone amide of Gln-6 forms a hydrogen

bond and a water-mediated hydrogen bond with the main chain and side chain of Glu-

395 in RBBP4, respectively. Moreover, Pro-9 in ZNF827 inserts into a hydrophobic

pocket encompassed by His-71, Pro-43, Ser-73 and Trp-42 in RBBP4. The backbone

carbonyl and amide group of Lys-10 forms a hydrogen bond and a water-mediated

hydrogen bond with the side chain amide and carbonyl group of Asn397 in RBBP4,

respectively. Arg-11 sits on the small panel formed by Trp-42 and Glu-41 and interacts with Glu-75 and Glu-41 in RBBP4 by electrostatic interactions.

To elucidate the crucial residues responsible for binding specificity, we designed site-directed mutants in both the ZNF8271-14 peptide and full-length RBBP4, and conducted further in vitro ITC assays (Table 4.2). The binding assays showed that

Arg-4, Lys-5 and Pro-9 in ZNF827 are crucial for binding to RBBP4, since the affinities between ZNF8271-14 peptides with mutations R4A, K5A, and P9A and RBBP4 were

124

reduced by ~120-, ~180- and ~40-fold, respectively, compared with that of the wild-

type ZNF8271-14 peptide. Moreover, the binding affinities between R4K and K5R mutants in ZNF827 were reduced by ~15- and ~50-fold, respectively, revealing that the binding of Arg-4 and Lys-5 are unique and cannot be substituted.

E126N128E179AAA and E126N128AA in RBBP4 showed weak binding (>

1mM) to the ZNF8271-14 peptide, indicative of the importance of these amino acids in

ZNF827 binding (Table 4.2). Y181A and P43S73AA in RBBP4 caused ~40- and ~150-

fold reduction in binding affinity, respectively, compared with that of wild-type RBBP4.

E231N277AA and E395A caused a > 10-fold reduction in binding to the ZNF8271-14

peptide.

125

Figure 4.4 Interaction between RBBP4 and the ZNF827 N-terminal peptide. a. Electrostatic surface potential representation of the RBBP4 binding pocket with the

ZNF8271-14 peptide (shown in green stick model). b. A simulated annealing omit map

(light gray) contoured at 1.0σ shows the electron density for the ZNF8271-14 peptide bound to RBBP4. RBBP4 residues (labelled in gray) are shown in gray stick model. Hydrogen bonds and salt bridge interactions are delineated by black dashed lines. c. Schematic representation of the interactions observed between RBBP4 and the

ZNF8271-14 peptide. Residues in RBBP4 and the ZNF8271-14 peptide engaged in recognition are shown in gray and green, respectively. Waters involved in the interaction are shown as red discs. Hydrogen bonds and salt bridge interactions are delineated by black dashed lines. This figure is from our publication in collaboration with Professor Shi’s group (Yang et al., 2018), and the structural biology experiments described in this figure were performed by our collaborators Aiai Sun and Fudong Li.

126

Table 4.2 Binding affinity between RBBP4 and ZNF8271-14 peptide measured by ITC (Data in this table are from our publication in collaboration with Professor Shi’s group (Yang et al., 2018), and were collected by our collaborators Aiai Sun and Fudong Li.)

KD △H △G -T△S Protein Peptide N µM kJ/mol kJ/mol kJ/mol WT WT 1.25 1.6±0.1 -57.0±0.4 -33.2 23.8 WT R4A 1* 200.0±14.9 -75.0±4.3 -21.1 53.9 WT R4K 1.47 24.3±1.0 -41.2±1.3 -26.4 14.9 WT K5A 1* 303.0±23.9 -62.9±4.4 -20.1 42.8 WT K5R 1* 92.2±4.8 -72.9±2.1 -23.1 49.9 WT K5E N.D. WT Q6A 1.07 11.6±0.2 -56.1±0.6 -28.2 27.9 WT P9A 1.17 74.8±9.1 -52.3±9.4 -23.6 28.7 WT R11A 1.28 95.9±3.4 -44.1±2.8 -23.0 21.2 E126N128E179AAA WT - > 1 mM - - - E126N128AA WT - > 1 mM - - - -88.3± P43S73AA WT 1* 242.0±32.9 -20.7 67.7 11.3 E231N277AA WT 1.12 21.4±1.6 -69.3±3.1 -26.7 42.6 E395A WT 1.00 15.9±1.1 -44.9±1.7 -27.4 17.5 Y181A WT 0.89 62.7±6.9 -54.2±4.0 -24.0 30.2 N.D, No detectable interaction under the experimental conditions.

* Due to low binding affinity, some titration curves were fitted with “N” values (binding stoichiometry of

the reaction) fixed to 1 to give more reasonable KD values. Note that in these fittings, ∆H might not be well determined (Turnbull and Daranas).

4.5 Structural comparisons of RBBP4-ZNF827, RBBP4-FOG1 and RBBP4-PHF6

The interaction between RBBP4 and ZNF827 is reminiscent of the previously characterized binding interfaces of RBBP4-FOG1 and RBBP4-PHF6 (Liu et al., 2015,

Lejon et al., 2011). Structural superimposition of the RBBP4 moieties of these three complexes revealed that ZNF827 binds to the same groove of RBBP4 as the FOG1 and PHF6 peptides, in a largely extended conformation (Figure 4.5 a). Arg-4, Lys-5 and Pro-9 in ZNF827 interact with the same residues of RBBP4 as Arg-4, Lys-5 and

Pro-9 of FOG1 and Arg-163, Lys-164 and Pro-167 of PHF6 (Figure 4.5 b). However,

127

ZNF827 and FOG1 show differences in the interaction between Arg-3 and RBBP4.

Specifically, electron densities for the side chain of Arg-3 in ZNF827 (Figure 4.4 b)

were not observed, while the side chain of Arg-3 in FOG1 showed clear electron

densities, and formed substantial hydrogen bonds with RBBP4 (Lejon et al., 2011).

We observed that the side chains of Arg-3 had alternative orientations in two copies

of FOG1 in the asymmetric unit, and their interaction with RBBP4 varied. These

observations suggest that the first arginine (Arg-3) in the “RRK” motif is dispensable

for interaction with RBBP4. The equivalent amino acid in PHF6 is a serine, further

supporting the dispensability of this amino acid for RBBP4 binding (Figure 4.5 c).

Figure 4.5 Structural comparison of the RBBP4-ZNF827 complex with RBBP4-

FOG1 and RBBP4-PHF6 complexes. a. Structural superimposition of RBBP4 moieties of RBBP4-ZNF827, RBBP4-FOG1 and RBBP4-PHF6 complexes. Only the RBBP4 molecule of the RBBP4-ZNF827 complex is shown. ZNF827, FOG1 and PHF6 peptides are coloured green, yellow and blue, respectively. b. Crucial residues responsible for RBBP4 binding are shown as a stick model. The RBBP4 molecule of RBBP4-ZNF827 is shown in gray surface representation. c. Sequence alignment of

128

RBBP4 bound peptides of human ZNF827, FOG1 and PHF6. The three key residues responsible for RBBP4 interaction are highlighted in yellow. This figure is from our publication in collaboration with Professor Shi’s group (Yang et al., 2018), and the structural biology experiments described in this figure were performed by our collaborators Aiai Sun and Fudong Li.

4.6 RBBP4 mutants do not bind to ZNF827 or telomeres

The effect of E126N128E179AAA, Y181A and P43S73AA mutations in RBBP4

on the interaction between full-length ZNF827 and RBBP4 was investigated further in vivo. E126N128E179AAA, Y181A and P43S73AA mutations were cloned into a mammalian RBBP4 expression vector. Wild-type RBBP4, and E126N128E179AAA,

Y181A and P43S73AA RBBP4 mutants were exogenously expressed in the ALT- positive human cancer cell line WI38-VA13/2RA (Figure 4.6 a), and the ZNF827-

RBBP4 interaction was examined by co-immunoprecipitation (co-IP) with a verified antibody against ZNF827.

Consistent with previous findings, we found that ZNF827 interacts with endogenous RBBP4 (Conomos et al., 2014), as well as exogenously expressed wild- type RBBP4 (Figure 4.6 b). In contrast to wild-type RBBP4, E126N128E179AAA,

Y181A and P43S73AA RBBP4 mutants completely abolished binding to ZNF827 without affecting the binding of ZNF827 to endogenous RBBP4, demonstrating the significance of these residues in the interaction (Figure 4.6 b). Complete disruption of

ZNF827 binding to all three mutants in vivo, despite varying levels of decreased binding affinity in vitro, suggests that even a moderate decrease in binding affinity

(~40-fold for Y181A) is detrimental to maintaining a stable interaction between full length ZNF827 and RBBP4.

129

The RBBP4 mutants were then explored in the context of telomere binding

using telomere-chromatin IP (telomere-ChIP). Interestingly, exogenous expression of

wild-type RBBP4 consistently resulted in a substantial increase in binding of both

ZNF827 and RBBP4 to telomeres; however, this increase was not seen following

exogenous expression of the three mutants (Figure 4.6 c). Our data indicate that none of the three RBBP4 mutants bind to telomeres (Figure 4.6 c). Overall, the increase in

ZNF827 and RBBP4 telomere binding following exogenous RBBP4 expression did not affect the binding of other NuRD components to telomeres. The observed increase in

RBBP4 binding at telomeres following wild-type RBBP4 overexpression, without changes in other NuRD components, suggests that either there are insufficient additional NuRD components to be recruited to telomeres, or RBBP4 alone is not able to facilitate an increase in the formation of the multi-subunit NuRD complex.

Alternatively, excess RBBP4 present at the telomeres may function independently of

NuRD, either alone or in complex with other proteins, given the prolific role of RBBP4 in several other protein complexes.

130

Figure 4.6 RBBP4 mutants disrupt binding to ZNF827 and telomeres in vivo. a. Western blot analysis of WI38-VA13/2RA cell extracts showing expression of endogenous RBBP4, and exogenously expressed FLAG-tagged RBBP4 and mutant RBBP4 proteins. The numbers above show relative exogenous expression of wild- type RBBP4 and mutants normalized to loading control. The numbers below show relative endogenous RBBP4 expression normalized to loading control. The protein bands were quantitated by densitometry using Multi Gauge software (Fujifilm). b. Co- IP with ZNF827 antibody in WI38-VA13/2RA cells overexpressing wild-type RBBP4 or the respective mutant (E126N128E179AAA, Y181A or P43S73AA) RBBP4 and ZNF827, followed by western blot analysis of endogenous RBBP4 and exogenous FLAG-tagged RBBP4 and RBBP4 mutants. c. Telomere-ChIP against ZNF827 and NuRD components in WI38-VA13/2RA cells 48 hours after overexpression of ZNF827

131

and wild-type RBBP4 or the respective mutant (E126N128E179AAA, Y181A or P43S73AA) RBBP4 (top panel). Quantitation of telomere-ChIP data (bottom panel). Data are mean ± range, n = 2 independent experiments. This figure is from our publication in collaboration with Professor Shi’s group (Yang et al., 2018). I performed the molecular biology and functional studies described in this figure.

4.7 Disruption of the interaction between ZNF827 and RBBP4 is insufficient to

alter ALT activity

ZNF827-mediated recruitment of NuRD to telomeres has previously been

shown to be integral to ALT activity. We have demonstrated that the

E126N128E179AAA, Y181A and P43S73AA RBBP4 mutants do not bind directly to

ZNF827 and are not recruited to telomeres. We analysed C-circles and ALT-

associated PML bodies (APBs), two widely utilised markers of ALT activity, to

investigate the impact of RBBP4 mutant expression on ALT activity. C-circles are extrachromosomal partially single stranded circles of C-rich telomeric DNA that can be amplified by rolling circle amplification (Henson et al., 2009), and APBs are nuclear foci containing PML protein as well as telomeric DNA and telomere binding proteins

(Yeager et al., 1999).

Neither exogenous expression of wild-type RBBP4 nor the three mutants affected C-circles levels compared to empty vector control (Figure 4.7 a). In addition, there were no significant changes in the number of APBs in WI38-VA13/2RA cells exogenously expressing wild-type RBBP4 or the three mutants (Figure 4.7 b). Overall, these data indicate that exogenous expression of wild-type and mutant RBBP4 proteins have no effect on ALT activity. This may be explained by saturation of NuRD binding at ALT telomeres achieved by endogenous RBBP4, or functional redundancy between NuRD subunits.

132

Figure 4.7 Exogenous expression of RBBP4 mutants does not affect ALT activity. a. C-circle analysis of WI38-VA13/2RA cells 48 hours after overexpression of ZNF827 and wild type RBBP4 or the respective mutant (E126N128E179AAA, Y181A or P43S73AA) RBBP4 (top panel). C-circle amplification is quantitated relative to empty vector as mean ± S.E.M.; n = 3 independent experiments (bottom panel). b. Quantitation of APBs in WI38-VA13/2RA cells 48 hours after overexpression of ZNF827 and wild type RBBP4 or the respective mutant (E126N128E179AAA, Y181A or P43S73AA) RBBP4. Data are mean ± S.E.M.; n = 3 independent experiments with 300 nuclei per replicate. This figure is from our publication in collaboration with Professor Shi’s group (Yang et al., 2018). I performed the molecular biology and functional studies described in this figure.

133

4.8 SUMOylation of ZNF827 is required for ALT activity and suppression of

telomeric DNA damage

We hypothesise that the ALT-promoting activity of ZNF827 is regulated by

SUMOylation. To study the impact of ZNF827 SUMOylation on ALT activity, a ZNF827

SUMOylation mutant construct was generated with all five reported SUMO sites at

amino acid positions 466, 523, 694, 673 and 704 (Hendriks et al., 2014, Hendriks et

al., 2015) mutated from lysine to arginine (K to R) (Figure 4.8 a). The SUMO mutant

was overexpressed at a comparable level to wild-type ZNF827 (Figure 4.8 b). The

ZNF827 SUMO mutant retained the ability to localise at telomeres with no observed

difference to wild-type ZNF827 (Figure 4.8 b). The effects of the SUMO mutant on ALT

activity were measured by quantitation of the widely used ALT phenotypic markers C-

circles and APBs, following transient overexpression (Figure 4.8 d, e). Exogenous expression of the ZNF827 SUMO mutant resulted in a significant decline in the levels of C-circles and APBs, even to a greater extent than the ΔRRK mutant, indicative of

suppression of ALT activity (Figure 4.8 d, e). It has previously been demonstrated by

telomere-ChIP that exogenous expression of wild-type ZNF827 reduced the level of

spontaneous γH2AX at ALT telomeres (Conomos et al., 2014). Here we used the

meta-TIF assay to quantify DNA damage occurrence specifically at telomeres by

measuring γH2AX and telomere colocalisations on metaphase spreads. Exogenous

expression of the ZNF827 SUMO mutant greatly increased the number of meta-TIFs

compared to ZNF827 wild-type overexpression, demonstrating that the role of ZNF827

in the suppression of telomeric DNA damage is dependent on SUMOylation (Figure

4.8 f).

134

135

Figure 4.8 ZNF827 SUMOylation is required for ALT activity and suppression of telomeric DNA damage. a. Schematic of SUMOylation mutated residues in ZNF827. b. Western blot analysis showing overexpression of wild-type ZNF827, the ZNF827 SUMO mutant and the ZNF827 ΔRRK mutant following 48 hr transfection. c. Representative interphase nuclei of the WI38/VA13.2RA overexpressing wild-type ZNF827 and the SUMO mutant, stained with Myc-tagged ZNF827 (red) immunofluorescence, telomere FISH (green) and DAPI (blue) showing colocalisations of both ZNF827 and the SUMO mutant with telomeres. d. Quantitation of APBs in WI38/VA13.2RA 48 hours after overexpression of wild-type ZNF827, the ZNF827 SUMO mutant and the ZNF827 ΔRRK mutant. Data are mean ± S.E.M.; n = 3 independent experiments with at least 200 nuclei per replicate. e. C-circle analysis of WI38/VA13.2RA cells 48 hours after overexpression of wild-type ZNF827, the ZNF827 SUMO mutant and the ZNF827 ΔRRK mutant (top panel). C-circle amplification is quantitated as mean ± S.E.M.; n = 3 independent experiments (bottom panel). f. Meta- TIF assay with γH2AX immunofluorescence (red) and telomere-FISH (green) in WI38/VA13.2RA 48 hours after overexpression of wild-type ZNF827 and the SUMO mutant (left panel). Quantitation of the number of meta-TIF in WI38/VA13.2RA 48 hours after overexpression of wild-type ZNF827 and the SUMO mutant (right panel). Data are mean ± S.E.M.; n = 3 independent experiments with 50 metaphase spreads per replicate.

4.9 Discussion

This study focused on defining the mechanistic details of ZNF827 recruitment

to telomeres in ALT cancer cells. To achieve this, we explored the mechanism of

ZNF827 recruitment by nuclear receptors, characterised the ZNF827-NuRD

interaction interface, and described the important role of ZNF827 SUMOylation in

promoting ALT activity. These data provide a structural and functional platform for the

development of ALT-specific anticancer strategies targeting ZNF827.

136

The important link bridging ZNF827-NuRD to ALT telomeres is the unique

presence of nuclear receptors TR4 and COUP-TF II at ALT telomeres. Variant repeats

interspersed throughout the telomere sequences in ALT cancer cells provide binding

sites for TR4 and COUP-TF II, which are required for ZNF827-NuRD recruitment to

ALT telomeres (Conomos et al., 2012, Lee et al., 2014, Conomos et al., 2014,

Conomos, 2014). Depletion of TR4 and COUP-TF II have been reported to impair the recruitment of the ZNF827-NuRD complex (Conomos, 2014). Therefore, understanding the molecular mechanism of ZNF827 recruitment by TR4 and COUP-

TF II is key to disrupt the localisation of ZNF827-NuRD at ALT telomeres.

Our data exclude ZNF827 as a coactivator for COUP-TF II and TR4 and suggest a form of transcriptional feedback loop between ZNF827 and the nuclear receptors. We have demonstrated that ZNF827 recruitment does not involve direct interactions with COUP-TF II or TR4, even in the presence of retinoic acid, a ligand shown to activate COUP-TF II and TR4 for coactivator binding (Kruse et al., 2008,

Zhou et al., 2011). Previously, our lab found that TR4 negatively regulates ZNF827 gene expression (Conomos, 2014). In this study we found that ZNF827 upregulates

TR4, but downregulates COUP-TF II, suggesting a regulatory role of ZNF827 in both

TR4 and COUP-TF II expression, and a feedback loop between TR4 and ZNF827.

The ligand retinoic acid induced protein expression of TR4 and COUP-TF II, consistent with previous reports (Inui et al., 1999), and increased COUP-TF II recruitment to telomeres. Increased COUP-TF II binding at telomeres was accompanied by a marked increase in TRF2 binding, as well as an observable increase in the NuRD component RBBP7 and ZNF827. Increased COUP-TF II binding following treatment with retinoic acid failed to facilitate transcriptional activity at telomeres, measured by TERRA levels. This may be attributed to an increase in

137

telomeric chromatin compaction resulting from increased ZNF827-NuRD as well as

TRF2 binding (Conomos et al., 2014), in response to retinoic acid-induced

transactivation at telomeres. Consistent with this, retinoic acid caused an increase in

TERRA levels in ZNF827 depleted cells. Together, these data suggest that the

recruitment of NuRD-ZNF827 may be associated with the transcription of TERRAs,

possibly modulated by TR4 and COUP-TF II at ALT telomeres. Besides TR4 and

COUP-TF II, other nuclear receptors including TR2, COUP-TF I and EAR2 are also found at ALT telomeres (Dejardin and Kingston, 2009, Marzec et al., 2015). Given that nuclear receptors are transcription factors, it is likely that their presence at ALT telomeres has an impact on TERRA transcription. This study has not directly examined the effects of nuclear receptors on TERRA transcription, and how TERRA may subsequently affect ZNF827 recruitment. Further investigations are warranted to address these questions.

The NuRD complex is recruited specifically to telomeres in cells that utilise the

ALT pathway of telomere maintenance. NuRD recruitment to telomeres is mediated by ZNF827, which bridges the interaction between telomeric DNA and the NuRD complex (Conomos et al., 2014). The NuRD–ZNF827 complex then functions as a

molecular platform for chromatin remodelling, telomere clustering, and the recruitment

of HR proteins, which promote homology-directed repair. Telomeric localisation of

NuRD is vital to ALT activity, and disruption of the N-terminal domain of ZNF827 inhibits NuRD recruitment to telomeres, reducing the viability of ALT cancer cells.

Consequently, the intricacies of the interaction between NuRD and ZNF827 are relevant to understanding the mechanism of ALT-mediated telomere maintenance, and may provide potential avenues for the development of ALT-specific cancer therapeutics.

138

In collaboration with Professor Shi’s group, we demonstrated that RBBP4 binds to ZNF827 with a stoichiometry of 1 : 1, and solved the crystal structure of RBBP4 bound to the N-terminal domain of∼ ZNF827 at a resolution of 1.9 Å. The RBBP4 interface comprises a β-propeller structure consisting of seven blades, which form a highly negatively charged cleft that binds the N-terminal domain of ZNF827 through a network of electrostatic interactions. The RBBP4–ZNF827 interaction resembles the previously characterised binding interfaces of RBBP4–FOG1 and RBBP4–PHF6 (Liu et al., 2015, Lejon et al., 2011), with variation in the binding dynamics of the first arginine in the ‘RRK’ motif, indicative of the first arginine being dispensable for the interaction. In addition to its ability to bind to RBBP4, FOG1 can also form pairwise interactions between its N-terminal region and the NuRD component MTA1. The level of amino acid conservation between the N-terminal RRKQXXP domains of FOG1 and

ZNF827 implicates a similar mode of binding between ZNF827 and MTA1, which may independently contribute to the telomeric localisation of NuRD.

Consistent with previous reports (Conomos et al., 2014), we identified that Arg-

4, Lys-5, and, to a lesser extent, Pro-9 in ZNF827 are vital for binding to RBBP4. By targeting residues in RBBP4 predicted to interact with ZNF827, we demonstrated loss of ZNF827 binding to the triple RBBP4 mutant E126N128E179AAA and to the double mutant E126N128AA, and decreased affinity of ZNF827 binding to Y181A and

P43S73AA RBBP4 mutants. By analysing these mutants in vivo, we demonstrated that the E126N128E179AAA, Y181A, and P43S73AA RBBP4 mutants do not bind to full-length ZNF827, nor are they recruited to telomeres. This supports the crystal structure and deduced importance of these residues from our structural analyses.

We observed an increase in telomeric binding of both ZNF827 and RBBP4 following overexpression of wild-type RBBP4, indicative of RBBP4 overexpression

139

stabilising the interaction between RBBP4–ZNF827 and telomeres. Interestingly, the

recruitment of other NuRD components to telomeres was not affected by enhanced

RBBP4–ZNF827 binding, suggesting that there is insufficient additional endogenous

NuRD subunits to be recruited to telomeres, or that RBBP4 can bind to telomeres independently of NuRD, either autonomously or in complex with other proteins.

Consistent with this observation, overexpression of RBBP4 mutants that were unable to bind to telomeres did not disrupt constitutive NuRD binding.

Our data comprehensively characterise the binding interface between RBBP4 and the N-terminal domain of ZNF827, and demonstrate the in vitro and in vivo binding capabilities of RBBP4 to ZNF827 and to telomeres. Nevertheless, using our experimental system that involved overexpression of wild-type and mutant RBBP4, we found no change in ALT activity, as detected using ALT biomarkers. This is consistent with there being no change in NuRD recruitment to telomeres, and can be explained by the presence of endogenous RBBP4 protein in the WI38-VA13/2RA cell line, which appears to be sufficient to facilitate NuRD–ZNF827 binding to telomeres. It is also

possible that functional redundancy between the multiple subunits that comprise

NuRD may confound interpretation of complex recruitment to telomeres. The major

limitation of the RBBP4 mutant experiments is that they were conducted in the

presence of endogenous RBBP4 which may have, to some extent, masked the effects

of the RBBP4 mutants. To obtain more definitive results, an improved experimental

approach would be to first define the effects of depleting endogenous RBBP4 protein

by RNA interference or CRISPR on NuRD recruitment and ALT activity, then

performing rescue experiments in RBBP4-depleted cells by overexpressing wild-type

RBBP4 and mutants.

140

We have also demonstrated that SUMOylation of ZNF827 is required to

mediate its function in promoting ALT activity and suppress telomeric DNA damage.

This is supported by previous findings that demonstrate SUMOylation of telomeric

proteins is essential for their recruitment to APBs, which have recently been shown to

be sites for telomeric recombination (Potts and Yu, 2007, Zhang et al., 2019a). The

induction of meta-TIFs following expression of the SUMO mutant reinforced the role

of ZNF827 in suppressing telomeric DNA damage. Colocalisation of the SUMO mutant

to telomeres indicates that SUMOylation is not essential for ZNF827 recruitment. This

suggests that SUMOylation of ZNF827 is required for subsequent protein interactions,

which are yet to be defined. The SUMO mutant generated in this study is based on

the 5 SUMO sites in ZNF827 identified in a previous proteomic study (Hendriks et al.,

2014). Interestingly, a more recent study of a large-scale SUMO proteome analysis identified a total of 26 SUMOylation sites in ZNF827 (Hendriks et al., 2017). It remains to be investigated whether loss of more SUMO sites will cause a greater effect, and which specific SUMO sites are essential for mediating ALT activity.

In summary, this study has defined important aspects of the molecular mechanisms driving the functions of ZNF827 at ALT telomeres. Our data provide structural and functional information regarding the RBBP4–ZNF827 interaction and the mechanism of ZNF827-NuRD recruitment to ALT telomeres, present insights into the association between ZNF827 and nuclear receptors, and illustrate the importance of posttranslational modifications such as SUMOylation in mediating the functions of

ZNF827. These findings support ZNF827 as a therapeutic target in cancers utilising

ALT, and help build the groundwork for the future development of ZNF827 modulators.

141

ZNF827 is a Novel ssDNA Binding Protein Involved in the Replication Associated DNA Damage and Repair Pathway

5.1 Introduction

ZNF827 recruitment to ALT telomeres is dependent on TR4 and COUP-TF II

(Conomos et al., 2014); however, as demonstrated in the previous chapter, this

recruitment does not rely on direct interactions between ZNF827 and the nuclear

receptors. This implicates the involvement of other proteins and/or telomeric DNA in

ZNF827 recruitment to telomeres, yet the mechanistic details are still to be defined.

ZNF827 belongs to the superfamily of C2H2 zinc finger proteins, which are known for

their DNA binding capabilities (Wolfe et al., 2000). ZNF827 contains nine C2H2 zinc

finger domains spatially arranged in two separate clusters, with the first three zinc

fingers (ZnF 1-3) towards the mid-point of the protein and the other six zinc fingers

(ZnF 4-9) proximal to the C-terminal. In this chapter, we aim to define the mechanistic details of ZNF827 recruitment to telomeres by investigating the binding capabilities of

ZNF827 to telomeric DNA and non-telomeric DNA.

Thus far, the best-known function of ZNF827 is the multifaceted role it plays in concert with the NuRD complex at ALT telomeres (Conomos et al., 2014). ALT

telomeres constantly pose challenges to the DNA damage and repair pathways,

evident by persistent replication stress, a spontaneous DNA damage response, and

the constant need for HR-mediated repair, specifically at telomeres. The roles of

ZNF827 in suppressing TIFs, facilitating hypoacetylation, and recruiting the HR

proteins BRIT1 and BCRA1 all contribute to overcoming these challenges at ALT

telomeres. ZNF827 has been identified in the same interaction network as DNA

damage and repair proteins 53BP1 and BRCA1 (Gupta et al., 2018). Furthermore,

142

there have been increasing data demonstrating a role for NuRD in mediating the DNA

damage response and repair (Polo et al., 2010, Chou et al., 2010, Smeenk et al., 2010,

Spruijt et al., 2016). This evidence prompted us to explore the role of ZNF827 as a

novel player in the DNA damage and repair pathway.

Here, we investigate the functions of ZNF827 in DNA damage and repair, both

genome-wide as well as at telomeres, with the aim of defining the interactions of

ZNF827 with replication protein A (RPA) and ATR. We further demonstrate the

involvement of ZNF827 in the activation of the ATR-CHK1 DNA damage signalling

pathway, replication stress and cell cycle progression.

5.2 ZNF827 binds to single stranded (ss) telomeric and non-telomeric DNA via

its zinc fingers 1-3

To investigate whether the recruitment of ZNF827 to telomeres involves direct

interaction with telomeric DNA, we performed electrophoretic mobility shift assays

(EMSAs) with ZNF827 recombinant proteins purified with the HaloTag® protein purification system and telomeric DNA oligos. To identify which zinc fingers are required for DNA binding activity, zinc finger mutants with either one or both zinc finger

clusters deleted were also generated and purified (Figure 5.1 a). Purified proteins were

validated by HaloTag® TMR ligand detection and western blot analysis (Figure 5.2 a,

b). Mass spectrometry (MS) analysis also confirmed the specificity of the HaloTag

protein purification method (Figure 5.2 c). HaloTag® empty vector and the ZNF827

SUMOylation mutant (SUMO Δ) were included in the binding assays as negative and

positive controls, respectively, as SUMOylation should not affect DNA binding activity,

given that SUMO Δ can associate with telomeres as shown in Chapter 4. Double-

stranded (ds) telomeric EMSA probes were obtained by radiolabelling annealed

143

complementary 18-mer G-rich and C-rich oligonucleotides digested by Exo VII to remove any residual single-stranded (ss) regions.

ZNF827 and the mutant proteins were allowed to bind to radiolabelled ds and ss telomeric DNA substrates as shown in Figure 5.1 b. Unexpectedly, the results demonstrated that neither ZNF827 or any of the mutant proteins bound to ds telomeric probes, but that ZNF827 clearly interacted with the ss G-rich telomeric DNA, as indicated by the shifted bands (Figure 5.1 b). Further binding experiments between ss telomeric G-rich oligos and ZNF827 or the mutant proteins, along with empty vector and the SUMO Δ mutant, confirmed binding of ZNF827 to ss G-rich telomeric DNA

(Figure 5.1 c). ZNF827, the SUMO Δ protein and ZnF4-9 del mutant protein exhibited binding activities to the ss G-rich telomeric 18-mer oligos, while the interactions were abolished with the ZnF1-3 del and ZnF1-9 del mutant proteins (Figure 5.1 c). These data demonstrate that zinc fingers 1-3 (ZnF1-3) are required for binding to ss G-rich telomeric DNA. Surprisingly, ZNF827 did not bind to ss C-rich telomeric DNA, but showed binding to a scrambled sequence of the ss G-rich telomeric repeats (Figure

5.1 d). The additional bands observed with the scrambled G-rich sequence may suggest a different binding stoichiometry to that of telomeric G-rich repeats. Again, the

ZnF4-9 del mutant retained binding activities analogous to ZNF827, whereas the

ZnF1-3 del mutant protein was completely devoid of binding to either telomeric repeats or the scrambled sequences (Figure 5.1 d). These results indicate that ZNF827 directly interacts with telomeres by preferentially binding to G-rich ss regions that are prevalent in telomeric repeats, and that the binding is dependent on zinc fingers 1-3.

144

Figure 5.1 ZNF827 binds to ss G-rich telomeric DNA and the binding requires zinc fingers 1-3. a. A schematic of ZNF827 mutants generated for the EMSA binding assays. b. EMSA assay showing ZNF827 or mutant proteins unable to bind ds telomeric DNA (no band shifts), while ZNF827 can bind ss G-rich telomeric repeats (band shifts). c. EMSA assay showing binding of ss G-rich telomeric DNA to ZNF827 and the mutants. Binding is present with wild-type ZNF827, SUMO mutant and the ZnF4-9 del mutant, but absent with the ZnF1-3 del and ZnF1-9 del mutants. d. EMSA

145

assays showing binding of wild-type ZNF827, ZnF4-9 del mutant and ZnF1-3 del mutant to ss G- and C-rich telomeric repeats and scrambled ss G- and C-rich telomeric repeats. Wild-type ZNF827 and ZnF4-9 del mutant binds ss G-rich telomeric and scrambled repeats, but not ss C-rich telomeric or scrambled repeats. ZnF1-3 del mutant has no binding activity.

146

Figure 5.2 Validation of ZNF827 and ZNF827 mutants purified by HaloTag®. a. Detection of ZNF827 and the mutants by HaloTag® TMR ligand following 48 hr overexpression of the respective HaloTag® plasmid constructs. b. Western blot analysis of ZNF827 and the mutants after protein purification by HaloTag®. c. Mass spectrometry analysis of ZNF827 and the mutants pulled down by HaloTag®.

147

The ability of ZNF827 to bind to scrambled as well as canonical ss telomeric

sequences suggests that ZNF827 does not bind to specific sequence motifs, but rather

displays binding affinity to a wide range of G-rich ss DNA sequences. Pentaprobe was

employed as a ssDNA substrate to test this hypothesis. Pentaprobe is a DNA

sequence of minimum length designed to contain all possible 5 base pair sequence

motifs for testing the DNA binding activity of proteins (Kwan et al., 2003). This

pentaprobe sequence was made into six pairs of overlapping complementary 100-mer

ss oligonucleotides. PP3 and PP9 were one of the complementary pairs gifted by Dr

Joel Mackay to be used in this study. We identified that ZNF827 as well as the ZnF4-

9 del mutant were able to bind to both ssPP3 and ssPP9, but not the ds counterpart

(Figure 5.3 a, c). The binding was consistently abrogated when zinc fingers 1-3 were deleted in the ZnF1-3 del mutant, corroborating that zinc fingers 1-3 are essential for the binding of ZNF827 to ss DNA (Figure 5.3 b, c). Taken together, these data demonstrate that ZNF827 binds via zinc fingers 1-3 to diverse ssDNA sequences with no stringent sequence specificity.

Previously, it has been shown that ZNF827 predominantly colocalises with telomeres in ALT cells. Given the diversity of ssDNA sequences, including both telomeric and non-telomeric sequences, that ZNF827 can bind to in vitro, we speculated that a protein-protein interaction may be responsible for targeting ZNF827 specifically to telomeres. While zinc fingers 4-9 are dispensable for ssDNA binding, they may be required for recruiting ZNF827 to the telomere. We further examined the

role of the two zinc finger clusters for ZNF827 recruitment to the telomere in vivo using

ZNF827 immunofluorescence (IF) and telomere FISH following overexpression of

ZNF827 and ZNF827 mutants in a U-2 OS ZNF827 CRISPR knockout (KO) cell line.

The U-2 OS ZNF827 KO cell line was generated using CRISPR/Cas9 genome editing

148

technology by the Vector and Genome Engineering Facility at Children’s Medical

Research Institute (CMRI). ZNF827 KO was validated by pGEM sequencing which

confirmed all the ZNF827 alleles have out-of-frame deletions. We were not able to

confirm ZNF827 KO by western blot due to the very low baseline endogenous ZNF827

protein level.

For ZNF827 and telomere colocalisations, the RRK and SUMO mutants were included as positive controls as they were shown to remain at telomeres ((Conomos

et al., 2014) and Figure 4.8 c). ZNF827 and the mutants were overexpressed at comparable levels (Figure 5.4 a). Interestingly, in contrast to wild-type ZNF827, and

the RRK and SUMO mutants, which all show distinct punctate nuclear foci and

prominent colocalisations with telomeric DNA, the ZnF1-3 del mutant depicted strong

and dispersed cytoplasmic as well as nuclear staining with no clear foci (Figure 5.4 b).

This suggests that zinc fingers 1-3 are crucial for retaining ZNF827 in the nucleus and for the formation of ZNF827 nuclear foci; probably by binding to ssDNA. In support of this, the ZnF4-9 del mutant with zinc fingers 4-9 deleted but retaining zinc fingers 1-3 displayed solely nuclear localisation, but staining was dispersed throughout the nucleus with no discrete colocalisations at the telomere. Thus, zinc fingers 4-9 appear to play an important role in bringing ZNF827 to the telomere, possibly by facilitating a protein-protein interaction. The ZnF1-9 del mutant with all zinc fingers deleted showed a similar staining pattern to the ZnF1-3 del mutant. Collectively, these findings indicate that both zinc finger clusters are functionally indispensable for recruiting ZNF827 to the telomere, with zinc fingers 1-3 necessary for DNA binding and zinc fingers 4-9 likely important for protein-protein interactions.

149

Figure 5.3 ZNF827 binds to ss non-telomeric pentaprobes and binding is dependent on zinc fingers 1-3. a. EMSA assays showing binding of ZNF827 to ss pentaprobes PP3 (left) and PP9 (right). b. EMSA assays showing binding of ZNF827 and the zinc finger mutants to ss pentaprobes PP3 (left) and PP9 (right). Binding present with wild-type ZNF827 and ZnF4-9 del mutant, but absent with ZnF1-3 del and ZnF1-9 del mutants. c. EMSA assays showing no binding between ZNF827 and the zinc finger mutants and the ds pentaprobe PP3-PP9. Binding to ss pentaprobes PP3 (left) and PP9 (right) used as positive controls.

150

Figure 5.4 ZNF827 zinc finger mutants display aberrant nuclear localisation. a. Western blot analysis of ZNF827 and mutants 48 hours post transfection. b. Representative images of colocalisations of Myc-tagged ZNF827 and the mutants labelled by IF (red) with telomeres labelled by FISH (green) following 48 hours overexpression of the Myc-tagged constructs in ZNF827 KO U-2 OS. Nuclei stained by DAPI in blue. The white arrows indicate colocalisations.

151

5.3 ZNF827 directly interacts with replication protein A (RPA)

RPA is the most well studied ssDNA binding protein with no sequence

specificity, known for its established roles in DNA replication, recombination and repair

(Zou et al., 2006). RPA is found at telomeres during S phase of the cell cycle (Verdun

and Karlseder, 2006), and is implicated in telomere maintenance in both yeast and

humans (Verdun et al., 2005, Kibe et al., 2007). Our data indicate that ZNF827 is a

novel ssDNA binding protein at the telomere, and plausibly at other genomic regions

with exposed or elevated levels of ssDNA. Considering the similarities between

ZNF827 and RPA, we explored whether there is a direct association between them.

Discrete RPA foci were observed in unchallenged U-2 OS cells, consistent with a

constant level of telomeric DNA damage and homologous recombination (HR)

mediated repair events in ALT cells (Figure 5.5 a). The topoisomerase I inhibitor topotecan greatly induced RPA foci as a result of increased replication associated

DNA damage, i.e. stalled or collapsed replication forks. ZNF827 strikingly associated with RPA, in both unchallenged and topotecan-treated U-2 OS cells (Figure 5.5 a).

Interestingly, in the topotecan-treated U-2OS cells, ZNF827 predominantly colocalised with the larger RPA foci. To ascertain whether ZNF827 association with RPA was attributable to the dominant presence of ZNF827 at ALT telomeres, we studied the association in two telomerase-positive cell lines, HT1080 and HT1080 6TG, in addition to the U-2 OS ALT cell line. In Figure 5.5, endogenous ZNF827 was examined with

RPA foci induced by topotecan since there were minimal RPA signals in unchallenged telomerase cells (Figure 5.6 a). A substantial proportion of endogenous ZNF827 was observed to associate with RPA as colocalisations in the two telomerase cell lines as well as the ALT cell line. Considering that ZNF827 rarely associates with telomeres in telomerase-positive cells (Conomos et al., 2014) (Figure 5.6 b), the colocalisations

152

observed in HT1080 and HT1080 6TG cells suggest that the association between

ZNF827 and RPA is not coincidental but linked to the inherent functions of the proteins.

Direct interaction between ZNF827 and RPA was further supported by co- immunoprecipitation (co-IP) experiments. RPA was immunoprecipitated with ZNF827 in both WI38-VA13/2RA and U-2 OS ALT cell lines and in the HT1080 telomerase- positive cell line, indicative of a direct interaction between the two proteins irrespective of telomeres (Figure 5.5 c). Co-IP with an RPA2 antibody also pulled down ZNF827, further strengthening the interaction. The ZNF827 and RPA interaction was also detected in topotecan-treated cells at similar levels. HDAC1 and MTA1 are NuRD components known to interact with ZNF827 (Conomos et al., 2014), and were therefore used as positive controls in the co-IP.

Additional evidence in support of a direct interaction between ZNF827 and RPA was demonstrated in a Duolink® proximity ligation assay (PLA). PLA assays allow in situ detection of protein interactions and involve labelling the proteins of interest with primary antibodies, which are subsequently bound by PLA probes (oligonucleotide- labelled secondary antibodies). The PLA probes hybridise only if the proteins are in close proximity, forming a self-priming circular DNA template, which is amplified by a rolling circle amplification reaction. The amplicon with repeating sequences attached to the proteins are then labelled by complementary oligonucleotides coupled to fluorophores, detected as discrete dots by immunofluorescence. The PLA assay was performed in HT1080 cells between endogenous ZNF827 and RPA proteins. In unchallenged HT1080 cells, there was minimal PLA signal between ZNF827 and RPA.

This may seem inconsistent with the co-IP results, but is likely to be attributable to the very low level of endogenous ZNF827. Notably, PLA foci of ZNF827 and RPA were significantly induced by topotecan treatment, indicating that replication associated

153

DNA damage accentuates the ZNF827 and RPA interaction (Figure 5.5 d). Taken together, our data show that ZNF827 directly interacts with RPA. Given the pivotal roles of RPA in DNA replication, damage and repair, we predict that ZNF827 plays important functions in these pathways.

Figure 5.5 ZNF827 directly interacts with RPA. a. Representative images of ZNF827 (red) and RPA (green) colocalisations by IF following 48 hr overexpression of ZNF827 and 24 hr incubation with 2 μg/mL topotecan or DMSO in U-2 OS. Nuclei stained by DAPI in blue. The white arrows indicate colocalisations. b. Representative images of colocalisations of endogenous ZNF827 (red) and RPA (green) by IF

154

following 24 hr incubation with 2 μg/mL topotecan in U-2 OS (ALT), HT1080 (telomerase) and HT1080 6TG (telomerase). Nuclei stained by DAPI in blue. The white arrows indicate colocalisations. c. Co-IP with direct ZNF827 or RPA2 antibody in WI38/VA13.2RA overexpressing ZNF827, followed by western blot analysis of RPA2 or ZNF827 (top panel). Co-IP with direct ZNF827 antibody in U-2 OS overexpressing ZNF827 and treated with 2μg/mL topotecan or DMSO for 24 hrs, followed by western blot analysis of RPA2, HDAC1 and MTA1 (middle panel). Co-IP with direct ZNF827 antibody in HT1080 overexpressing ZNF827, followed by western blot analysis of RPA2, HDAC1 and MTA1 (bottom panel). d. PLA assay showing the interaction of endogenous ZNF827 and RPA2 in HT1080 following 24 hr incubation with 2 μg/mL topotecan or DMSO. Representative images of ZNF827 and RPA2 PLA signals for ZNF827 ab only control, DMSO control and topotecan treatment (left panel). Nuclei stained by DAPI in blue. Quantitation of PLA dots from 100 nuclei per experimental condition by manual counting (right panel). Data presented as mean ± SEM; **** P<0. 0001 by two tailed t test. The PLA assay was performed by Christopher Nelson in our lab.

155

Figure 5.6 RPA induction by topotecan and ZNF827-telomere colocalisations. a. Representative images of RPA foci (green) induction by topotecan in U-2 OS (ALT), HT1080 (telomerase) and HT1080 6TG (telomerase) cells after 24 hr incubation with 2 μg/mL topotecan or DMSO. b. Representative images of ZNF827 (red) and telomere (green) colocalisations in HT1080 6TG (telomerase) and U-2 OS (ALT) cells 48 hr post ZNF827 overexpression.

156

5.4 ZNF827 depletion exacerbates RPA induction following replication- associated DNA damage

To begin to understand the mechanistic details of the interaction between

ZNF827 and RPA, we performed IF on U-2 OS ZNF827 CRISPR KO cells following exogenous expression of ZNF827 and all five ZNF827 mutants. Neither the RRK or

SUMO mutant had an impact on the colocalisation of ZNF827 and RPA, suggesting that their interaction is independent of both the NuRD complex and SUMOylation

(Figure 5.7 a). Amongst the zinc finger mutants, deletion of zinc fingers 1-3 or all zinc fingers consistently showed dispersed cytoplasmic and nuclear localisation with an absence of discrete foci, indicative of an inability for ZNF827 to be retained in the nucleus. Deletion of zinc fingers 4-9 displayed mainly nuclear localisation with widespread smaller indistinct foci (Figure 5.7 a). Following exogenous expression of the zinc finger mutants, RPA retained its staining pattern (in characteristic nuclear foci), and these foci appeared to be unaffected by the dispersed and aberrant ZNF827 localisation.

Large RPA foci were present in the U-2 OS ALT cell line in unchallenged conditions, in contrast to the HT1080 telomerase-positive cell line in which RPA foci only became visible upon topotecan treatment (Figure 5.5 a and Figure 5.7 b). There was also a remarkable induction of RPA foci in U-2 OS cells, indicative of robust induction of stalled or collapsed replication forks by topotecan. In concordance with a spontaneous and elevated level of DNA damage and repair at ALT telomeres, a considerable number of RPA foci were associated with telomeres in unchallenged U-

2 OS cells (Figure 5.7 b). Upon DNA damage induction by topotecan treatment, a significant proportion of induced RPA foci associated with telomeric DNA in HT1080 cells (Figure 5.7 b, c). Intriguingly, ZNF827 depletion significantly reduced the

157

proportion of RPA foci associated with telomeres in both unchallenged and topotecan

treated U-2 OS cells (Figure 5.7 c). This proportional decrease is likely to be due to

the substantial increase in the total number of nuclear RPA foci following ZNF827

depletion (Figure 5.7 c). The overall increase in RPA foci suggests that ZNF827

depletion triggers accumulation of ssDNA due to unresolved DNA damage, most likely

stalled or collapsed replication forks originating in non-telomeric regions. In topotecan-

treated U-2 OS cells, ZNF827 depletion did not cause an observable increase in overall RPA foci, perhaps due to the massive induction of RPA by topotecan alone

(Figure 5.7 c).

In unchallenged HT1080 cells, depletion of ZNF827 did not change either the

proportion of RPA at telomeres or the total number of nuclear RPA foci, possibly

because of the very low number of RPA foci observed under unchallenged conditions,

which are consistent with the very low spontaneous DNA damage burden in

telomerase-positive cells (Figure 5.7 c). Following DNA damage induction by

topotecan, ZNF827 depletion caused a significantly greater increase in overall nuclear

RPA foci as well as a proportional increase in RPA associating at telomeres (Figure

5.7 c). These results suggest that ZNF827 plays a role in the repair of DNA damage

specifically involving the generation of ssDNA, such as stalled or collapsed replication

forks and end resection during HR-mediated repair.

158

159

Figure 5.7 ZNF827 depletion exacerbates RPA foci induction by topotecan. a. Representative images of colocalisations of Myc-tagged ZNF827 and the mutants (red) with RPA2 (green) labelled by IF following 48 hr overexpression of the Myc-tagged ZNF827 and mutant constructs in ZNF827 KO U-2 OS cells. Nuclei stained by DAPI in blue. The white arrows indicate colocalisations. b. Representative images of RPA2 foci (red) labelled by IF and telomeres (green) labelled by FISH in U-2 OS and HT1080 following 72 hr siRNA knockdown of ZNF827 and 24 hr incubation with 2 μg/mL topotecan or DMSO. Nuclei stained by DAPI in blue. The white arrows indicate colocalisations. c. Data quantitation of the frequency of RPA2 colocalising with telomeres (left) and the total nuclear RPA foci (right) by automated counting using CellProfiler. Data presented as mean ± SEM from 200 nuclei per experimental condition in one replicate; *** P<0. 0002, **** P<0. 0001 by two tailed t test.

5.5 ZNF827 is involved in the DNA damage and repair pathway, independent of

ATM-mediated NHEJ

The data in this study so far, in combination with the previous findings that

ZNF827 recruits HR proteins BRIT1 and BRCA1 to telomeres and suppresses spontaneous telomeric DNA damage (Conomos et al., 2014), have implicated ZNF827 in the DNA damage and repair pathway. We further investigated the functions of

ZNF827 with respect to the DNA damage and repair pathway at telomeres and the whole genome. DNA DSBs were induced by irradiation (IR) which causes DSBs globally, or by topotecan that results in DSBs from stalled or collapsed replication forks.

Colocalisations of ZNF827 and the DNA damage marker γH2AX were clearly observed and increased significantly after IR, supporting the involvement of ZNF827 in the DNA damage response (Figure 5.8 a). The effect of ZNF827 depletion on DNA damage at a genomic level was explored using the Comet assay. Remarkably,

ZNF827 depletion caused DNA damage measured as tail moment to a similar extent as topotecan treatment, while ZNF827 depletion in conjunction with topotecan further

160

increased the amount of DNA damage, suggestive of a role of ZNF827 in suppressing

DNA damage (Figure 5.8 b). At telomeres, IR led to a rapid increase in ZNF827 and

telomere colocalisations which slowly declined to baseline over time, suggestive of

gradual resolution of IR-induced DNA damage (Figure 5.8 c). Taken together, these

data support ZNF827 as a novel player in the DNA damage and repair pathway at

telomeres and genomically.

DSBs are predominantly repaired by the two major DNA repair pathways – non-

homologous end joining (NHEJ) and homologous recombination (HR). Thus far our

data on ZNF827 including its binding to ssDNA, interaction with RPA, association with

replication associated DNA damage, and specificity to ALT telomeres that are

undergoing continuous HR-mediated BITS, all suggest that ZNF827 is involved in HR-

mediated DNA repair. Here we investigated whether ZNF827 plays a role in the other

DNA repair pathway NHEJ. TRF2 depletion at telomeres activates the ATM DDR pathway and triggers telomere-telomere fusions by NHEJ (Karlseder et al., 2004,

Denchi and de Lange, 2007), which provides a model system to explore the role of

ZNF827 in NHEJ. The HT1080 6TG telomerase positive cell line with two efficiencies of stable TRF2 knockdown generated by shRNA transduction were obtained from Dr

Anthony Cesare at CMRI. TRF2 sh-F induced an ATM-mediated DDR at telomeres with no telomere-telomere fusions, i.e. intermediate-state telomeres, whereas TRF2 sh-G induced a more severe DDR at telomeres and NHEJ-mediated telomere- telomere fusions, i.e. uncapped telomeres (Cesare et al., 2013). The meta-TIF assay was performed to assess quantitatively the effect of ZNF827 on TRF2-knockdown induced DNA damage at telomeres. Consistent with the previous study, TRF2 sh-G induced a much greater number of DDR positive telomeres than TRF2 sh-F (Cesare et al., 2013). Overexpression of ZNF827 had little impact on the telomeric DNA

161

damage caused by TRF2 knockdown in either sh-F or sh-G cells (Figure 5.8 d). Given

that DNA damage induced by TRF2 knockdown is mediated by the ATM kinase and is predominantly processed by NHEJ (Karlseder et al., 2004, Denchi and de Lange,

2007), ZNF827 having no effect on this type of DNA damage suggests that ZNF827 is not involved in ATM-mediated DNA damage signalling and DNA repair by NHEJ.

162

163

Figure 5.8 ZNF827 is involved in the DNA damage and repair pathway, independent of ATM. a. Representative images of γH2AX (red) and ZNF827 (green) colocalisations labelled by IF in WI38-VA13/2RA cells overexpressing ZNF827, harvested 1 hr post IR exposure (left panel). Nuclei stained by DAPI in blue. Quantitation of γH2AX and ZNF827 colocalisations from 50 nuclei per experimental condition by manual counting (right panel). Data presented as mean ± SEM from one replicate; ** P<0. 005 by two tailed t test. The white arrows indicate colocalisations. b. Comet assay showing a representative image of comet tails from each experimental condition in U-2 OS following 72 hr siRNA knockdown of ZNF827 and 24 hr incubation with 2 μg/mL topotecan or DMSO (left panel). Quantitation of tail moment in 50 cells per experimental condition by a semi-automated approach using CellProfiler (right panel). Data presented as mean ± SEM from one replicate; * P<0. 05, ** P<0. 005 by two tailed t test. c. Quantitative data of ZNF827 and telomere colocalisations in WI38- VA13/2RA overexpressing ZNF827, harvested at 0.5, 1, 1.5, 2 and 3 hr post IR exposure. Data presented as mean ± SEM from 236 nuclei per experimental condition from one replicate; ** P<0. 005 by two tailed t test. d. Meta-TIF assay in HT1080 6TG with stable TRF2 knockdown by shRNA transduction with TRF sh-F and TRF sh-G. Representative images of metaphase spreads with γH2AX foci labelled in red by IF at telomeres (green) labelled by FISH (left panel). Chromosomes stained by DAPI in blue. Quantitation of meta-TIFs in 50 metaphases per experimental condition by manual counting. Data presented as mean in a bar graph from one replicate. **** P<0. 0001 by two tailed t test.

5.6 ZNF827 interacts with ATR and functions in the activation of ATR-CHK1

As we excluded ZNF827’s involvement in ATM-mediated NHEJ DNA repair, we

further investigated the role of ZNF827 in HR-mediated DNA repair. The ATR-

mediated DDR signalling pathway plays a critical role in HR-mediated DNA repair, especially at stalled or collapsed replication forks (Cimprich and Cortez, 2008, Flynn and Zou, 2011). The ATR kinase is activated by ssDNA structures coated by RPA, and is the master regulator of cellular responses to replication stress (Flynn and Zou,

164

2011). It has also been found that ATR plays a critical role in ALT cell viability (Flynn

et al., 2015), plausibly attributable to the elevated levels of ssDNA and replication

stress at ALT telomeres. Based on our finding that ZNF827 binds to ssDNA and

interacts with RPA, we explored the role of ZNF827 in the ATR DNA damage signalling

pathway. ZNF827 visibly colocalised with ATR in unchallenged U-2 OS cells (Figure

5.9 a). Upon replication associated DNA damage induced by topotecan, the majority

of ZNF827 foci colocalised with ATR (Figure 5.9 a). This indicates a physical

association between ZNF827 and ATR. However, co-IP experiments failed to detect

a direct interaction between ZNF827 and ATR (Figure 5.9 b). This may be due to their

interaction being weak or transient, meaning that it cannot be efficiently detected by

the co-IP method. To overcome this limitation, we further examined the interaction

between ZNF827 and ATR using the PLA assay, which permits the identification of

weak or transient individual interactions between two proteins in their native form. A

preliminary PLA assay demonstrated a substantial number of PLA signals between

ZNF827 and ATR, indicative of a direct interaction (Figure 5.9 c). However, the PLA

assay will need to be repeated to obtain sufficient data for quantitative analysis in order

to corroborate the interaction. Nevertheless, collectively these results suggest that

ZNF827 transiently interacts with the ATR kinase.

To investigate the precise function of ZNF827 in the ATR pathway, we examined the effect of ZNF827 depletion on the downstream effectors of ATR. To activate ATR, U-2 OS cells were treated with DNA damage and replication stress

inducing agents topotecan 2 μg/mL for 24 hours or aphidicolin 1 μM for 20 hours. As

expected, CHK1, the major downstream effector of ATR, became phosphorylated at

S345 upon topotecan or aphidicolin treatment, indicative of ATR activation (Figure 5.9

d, e). RPA, which is also a downstream target of ATR, was phosphorylated at S33

165

after topotecan treatment (Figure 5.9 d). Strikingly, both phosphorylation of CHK1

S345 and RPA S33 were markedly suppressed by ZNF827 depletion (Figure 5.9 d, e).

Interestingly even in unchallenged U-2 OS cells, ZNF827 depletion reduced baseline phosphorylation of CHK1 S345. These results suggest that ZNF827 is required for the activation of ATR. Given that ATR plays an essential role in DNA repair involving ssDNA, these data support the findings of our earlier experiments in which ZNF827 depletion triggered accumulation of ssDNA and unresolved DNA damage (Figure 5.7).

Phosphorylation of CHK2 was not affected by ZNF827 depletion, further supporting that the role of ZNF827 in DNA damage and repair is independent of ATM (Figure 5.9 d). Collectively, our data suggest that ZNF827 plays an important role in the activation of the ATR-CHK1 pathway and the repair of DNA damage involving ssDNA, likely stalled or collapsed replication forks.

166

Figure 5.9 ZNF827 interacts with ATR and ZNF827 depletion suppresses CHK1 and RPA activation. a. Representative images of ZNF827 (red) and ATR (green) colocalisations labelled by IF in U-2 OS cells following 48 hr ZNF827 overexpression and 24 hr incubation with 2 μg/mL topotecan or DMSO. Nuclei stained by DAPI in blue. The white arrows indicate colocalisations. b. Co-IP with direct ZNF827 antibody in WI38-VA13/2RA and U-2 OS cells overexpressing ZNF827, followed by western blot analysis of ATR and ZNF827. c. Representative images of PLA assay showing interaction between ZNF827 and ATR. The PLA assay was performed by Christopher Nelson in our lab. d. Western blot analysis of p-CHK1 (S345), total CHK1, p-RPA2 (S33), total RPA2 and pCHK2 (T68) in U-2 OS cells following 72 hr siRNA knockdown

167

of ZNF827 and 24 hr incubation with 2 μg/mL topotecan or DMSO. Vinculin used as a loading control. e. Western blot analysis of p-CHK1 (S345) and total CHK1 in U-2 OS cells following 72 hr siRNA knockdown of ZNF827 and 20 hr incubation with 1 μM aphidicolin or DMSO. Vinculin used as a loading control.

5.7 ZNF827 localises to replication forks and plays a role in the cellular response to replication stress

Considering the clear role we have identified for ZNF827 in the ATR-CHK1 pathway, ATR’s fundamental functions in responding to replication stress, and the inherent replication stress present at ALT telomeres, we were prompted to investigate the role of ZNF827 in replication stress genomically as well as at telomeres. We examined ZNF827 colocalisations with PCNA, a processivity factor for DNA polymerases that marks replication forks, and demonstrated that ZNF827 associates with replication forks. The nuclear dynamics of PCNA denote the different stages of S phase, with widespread dispersed foci indicating early S phase, and sporadic discrete foci likely to mark stalled replication forks in late S phase (Essers et al., 2005). In

Figure 5.10 a, ZNF827 was shown to colocalise with PCNA in two different nuclear localisation patterns in both U-2 OS and HT1080 cells treated with 0.4 μM aphidicolin for 20 hours to induce stalled replication forks. Particularly in U-2 OS cells, the colocalisations between ZNF827 and PCNA became clearer in cells with sporadic discrete PCNA foci, which is indicative of stalled replication forks that form after aphidicolin treatment (Figure 5.10 a). This implicates the involvement of ZNF827 in processing stalled replication forks.

To further determine whether ZNF827 is present at stalled replication forks, we examined the dynamics of ZNF287 at DNA damage sites specifically resulting from

168

stalled or collapsed replication forks induced by topotecan. Compared to unchallenged

HT1080 and HT1080 6TG cells, topotecan efficiently induced γH2AX foci, marking

either stalled or collapsed replication forks (Figure 5.10 b, c). This coincided with a striking induction of ZNF827, suggesting a rapid response of ZNF827 recruitment to replication-associated DNA damage (Figure 5.10 b, c). Additionally, ZNF827 colocalised with topotecan-induced γH2AX foci, indicating that ZNF827 is present at stalled or collapsed replication forks (Figure 5.10 b, c). Collectively, these data demonstrate that ZNF827 is a novel player in DNA damage and repair, with a specific role in replication-associated DNA damage.

169

Figure 5.10 ZNF827 is at replication forks and plays a role in cellular response to replication stress. a. Representative images of two PCNA (red) nuclear localisation patterns, and colocalisations between ZNF827 (green) and PCNA (red) labelled by IF in U-2 OS and HT1080 following 48 hr ZNF827 and PCNA chromobody

170

(Cell Cycle Chromobody®-RFP) overexpression and 6 hr incubation with 0.4 μM aphidicolin to induce stalled replication forks. Nuclei stained by DAPI in blue. The white arrows indicate colocalisations. b. Representative images of γH2AX (red) and endogenous ZNF827 (green) foci labelled by IF in HT1080 and HT1080 6TG following 24 hr incubation with 2 μg/mL topotecan or DMSO (left panel). Nuclei stained by DAPI in blue. The white arrows indicate colocalisations. c. Data quantitation of γH2AX foci induction (top left), ZNF827 foci induction (top right) and γH2AX and ZNF827 colocalisations (bottom) from 50 nuclei per experimental condition by manual counting. Data presented as mean ± SEM from one replicate; **** P<0. 0001 by two tailed t test.

5.8 ZNF827 depletion induces replication defects and impedes DNA repair by

To ascertain the role of ZNF827 in DNA replication dynamics, we examined the effects of ZNF827 depletion on replication by measuring phenotypic markers of replication defects at telomeres as well as genomically. This included quantitation of fragile telomeres, telomere signal free ends, chromosome breakage events and micronuclei. Telomeres are a good model for studying factors that modulate replication stress due to their inherent vulnerability. ALT telomeres, in particular, have elevated levels of replication stress due to structural aberrations. We discovered that ZNF827 depletion caused clear replication defects at telomeres, demonstrated by an observable increase in the frequency of fragile telomeres, although the difference did not attain statistical significance, and a significant increase in telomeres with signal free ends compared to control (Figure 5.11 a, b). Genomically, ZNF827 depletion induced a much greater occurrence of chromosome breakage events (Figure 5.11 c).

The proportion of micronuclei was not affected at normal conditions by ZNF827 depletion. Paradoxically, ZNF827 depletion suppressed topotecan-induced micronuclei (Figure 5.11 d). One possible explanation for this may be that depletion of

171

ZNF827 causes accumulation of DNA damage during replication that prevents cells

from entering mitosis. Taken together, these results illustrate that ZNF827 exerts a

vital function in DNA replication both at telomeres and genome-wide.

Given that ZNF827 is both associated with and induced by replication-

associated DNA damage, and that ZNF827 plays an important role in ensuring

successful DNA replication, we next explored the effects of ZNF827 depletion on the

repair of replication-associated DNA damage by HR. HR is an error free DNA repair

pathway predominantly active during replication that uses sister chromatids as templates to ensure faithful DNA repair. The sister chromatid exchange (SCE) assay was used to quantitatively measure HR-mediated repair after induction of DNA damage resulting from collapsed replication forks by topotecan. Topotecan caused a drastic induction of SCEs, which was significantly stunted in ZNF827 CRISPR KO U-

2 OS cells (Figure 5.11 e, f). A similar result was observed in HT1080 cells after

ZNF827 knockdown by siRNA. The suppression of SCEs in ZNF827 depleted cells suggests that ZNF827 promotes HR-mediated repair of replication-associated DNA damage, i.e. repair of DSBs from stalled or collapsed replication forks, which is consistent with our previous study that found ZNF827 can recruit HR proteins

(Conomos et al., 2014). Given the essential roles of the ATR-CHK1 pathway in safeguarding DNA replication and HR-mediated DNA repair (Syljuasen et al., 2005,

Flynn and Zou, 2011, Sorensen et al., 2005b), these findings may be explained by inhibition of the ATR-CHK1 pathway caused by ZNF827 depletion, as demonstrated in Figure 5.9.

172

173

Figure 5.11 ZNF827 depletion induces replication defects and suppresses SCEs. a. Representative images of fragile telomeres in HT1080 cells following 72 hr siRNA knockdown of ZNF827 (left panel). Quantitation of frequency of fragile telomeres per chromosome from 950 chromosomes (right panel). Data presented as mean ± SEM. The white arrows indicate examples of fragile telomeres. b. Representative images of telomere signal free ends in HT1080 cells following 72 hr siRNA knockdown of ZNF827 (left panel). Quantitation of frequency of telomere signal free ends per chromosome from 950 chromosomes per experimental condition (right panel). Data presented as mean ± SEM. * P<0. 05 by two tailed t test. The white arrows indicate examples of telomere signal free ends. c. Representative images of chromosome breakage in HT1080 cells following 72 hr siRNA knockdown of ZNF827 (left panel). Quantitation of frequency of chromosome breakage per chromosome in 950 chromosomes per replicate from two biological replicates (right panel). Data presented as mean ± SEM. *** P<0. 0002 by two tailed t test. The white arrows indicate examples of chromosome breakage. d. Representative images of the proportion of nuclei with micronuclei in U-2 OS cells following 72 hr siRNA knockdown of ZNF827 and 24 hr incubation with 2 μg/mL topotecan or DMSO (top panel). Quantitation of micronuclei in 500-1000 nuclei per replicate from three biological replicates (bottom panel). Data presented as mean ± SEM. * P<0. 05, ** P<0. 005 by two tailed t test. e. Representative images of SCEs in U-2 OS and U-2 OS ZNF827 CRISPR KO cells incubated with 2 μg/mL topotecan or DMSO for 1 hr. Green arrows indicate points of exchange events. f. Quantitative data of SCEs in U-2 OS and U-2 OS ZNF827 CRISPR KO cells incubated with 2 μg/mL topotecan or DMSO for 1 hr (left), and in HT1080 cells incubated with 2 μg/mL topotecan or DMSO for 1 hr at 24 hr within the 72 hr siRNA knockdown of ZNF827 (right). Data presented as mean ± SEM from 800- 1000 chromosomes per replicate from three biological replicates. * P<0. 05, *** P<0. 0002, **** P<0. 0001 by two tailed t test.

174

5.9 ZNF827 affects cell cycle progression and its depletion causes G1 to early S

arrest

Our findings implicate ZNF827 in mediating the ATR-CHK1 pathway as well as

DNA replication, leading us to explore further whether ZNF827 depletion influences

cell cycle progression. Using flow cytometry, we examined the effects of ZNF827

depletion on cell cycle progression in unchallenged and topotecan-treated HT1080

cells. Interestingly, in unchallenged cells, depletion of ZNF827 caused a subtle

horizontal shift that was most noticeable at the G1/S border. This was reflected in the

quantitation as a subtle increase in S phase cells accompanied by a reduction in G2/M

cells (Figure 5.12 a, b). Topotecan treatment triggered a distinct G2/M arrest, in line

with previous studies (Ohneseit et al., 2005, Feeney et al., 2003). ZNF827 depletion

in conjunction with topotecan treatment caused a conspicuous G1 arrest while

reducing the G2/M arrest (Figure 5.12 a, b). These data demonstrate that ZNF827 has

an impact on cell cycle progression, most likely by promoting G1/S transition.

To further characterise the impact of ZNF827 on the cell cycle, we conducted

live cell imaging with HT1080 FUCCI cells gifted by Dr Anthony Cesare at CMRI, to

monitor cell progression through the cell cycle in real-time following ZNF827 depletion

(Figure 5.12 c). The FUCCI (fluorescence ubiquitination cell cycle indicator) cells were stably transfected with fluorescence-labelled cell cycle-regulated proteins, geminin

(labelled by GFP) and Cdt1 (labelled by RFP), which are only expressed during specific phases of the cell cycle. Their expressions are used to classify cells by cell cycle phase in a live cell setting (Figure 5.12 c). Noticeably ZNF827 depletion led to a strong G1 arrest comparable to topotecan treatment, while ZNF827 depletion in conjunction with topotecan further exacerbated G1 arrest (Figure 5.12 d, e). In line with the flow cytometry results, ZNF827 depletion had no impact on G2/M by itself, but

175

resulted in a significant decrease in topotecan induced G2/M arrest, which coincided with the increased G1 arrest (Figure 5.12 f). These results suggest that ZNF827 acts predominantly during the G1/S phase of the cell cycle.

A significant proportion of cells arrested at G2/M by topotecan eventually bypassed mitosis and re-entered G1, consistent with previous studies (Figure 5.13 a)

(Ohneseit et al., 2005, Feeney et al., 2003). Depletion of ZNF827 also caused a significant proportion of cells to fail to go through mitosis. Intriguingly, more than half of these cells did not enter mitosis because they were not able to enter the S/G2-M phase at all (Figure 5.13 a). Analysis of FUCCI cells demonstrated that these cells were arrested at the G1 to early S transition, indicated by yellow cells, and then reversed back to red indicating re-entry into G1 following complete bypass of S/G2-M and mitosis. This suggests that the cells were prevented from undergoing DNA replication (Figure 5.12 c, Figure 5.13 a). ZNF827 depletion in conjunction with topotecan treatment significantly increased the proportion of cells that bypassed S/G2-

M and mitosis. These observations suggest that ZNF827 depletion impedes normal

DNA replication, in agreement with the replication defects in ZNF827-depleted cells identified in Figure 5.11. Interestingly, we observed that some ZNF827 depleted cells, despite bypassing S/G2-M, underwent mitosis in G1 (Figure 5.13 b). This surprising observation requires further investigation.

The G1 and early S arrest resulting from ZNF827 depletion coincided with an upregulation of p53 and p21 (Figure 5.13 c), which have been long established to play a crucial role in mediating G1 and S phase cell cycle arrest in response to DNA damage (Niculescu et al., 1998, Ogryzko et al., 1997, Radhakrishnan et al., 2004,

Gartel and Radhakrishnan, 2005, Waga et al., 1994). Strikingly, p21 became upregulated even in unchallenged ZNF827-depleted cells to a similar extent as

176

topotecan-treated cells, further reinforcing that ZNF827 depletion induces DNA damage (Figure 5.13 c). Topotecan-treated cells were harvested immediately after 1- hour drug treatment, or allowed to recover for 24 hours after drug removal. ZNF827 depletion in conjunction with topotecan triggered a further induction of p21, suggestive of persistent unresolved DNA damage. Corresponding to the topotecan induced G2/M cell cycle arrest, there was a robust induction of phosphorylation of CHK1 and RPA immediately after topotecan treatment, indicating activation of the ATR-CHK1 mediated G2/M DNA damage checkpoint (Liu et al., 2000). Interestingly, both CHK1 and RPA activation were almost completely diminished in ZNF827 depleted cells 24 hours after topotecan removal, while they were sustained in the scrambled control.

There were also marked increases in p53 and p21 induction 24 hours after topotecan removal in both scrambled control and ZNF827-depleted cells, confirming persistent unresolved DNA damage. Therefore, the drastic abolishment of CHK1 and RPA activation was not a result of DNA damage being resolved, but more likely due to

ZNF827 loss rendering cells unable to sustain ATR-CHK1 activation, confirming our findings in section 5.6 that ZNF827 depletion suppresses ATR-CHK1 activation. Loss of the ATR-CHK1 mediated G2/M cell cycle checkpoint may explain the occurrence of mitosis in G1 without DNA replication in ZNF827-depleted cells.

177

Figure 5.12 ZNF827 affects cell cycle progression and ZNF827 depletion causes G1 to early S arrest. a. Histogram of cell cycle analysis by flow cytometry displaying

178

the cell cycle profile of HT1080 cells following 72 hr siRNA knockdown of ZNF827 and 1 hr incubation with 2 μg/mL topotecan or DMSO. b. Data quantitation of cell cycle profile of a. c. Schematic of the fluorescent cellular changes associated with FUCCI cell cycle sensor. Cells in G1 express Cdt1-RFP (red). Cells in G2/M express Geminin- GFP (green). Cells in S phase appear yellow (both Geminin-GFP and Cdt1-RFP expressed). d. Quantitative data of live cell imaging showing G1 duration in HT1080 FUCCI cells following 72 hr siRNA knockdown of ZNF827 and 1 hr incubation with 2 μg/mL topotecan or DMSO. Data presented as mean ± SEM from 40 cells per replicate from three biological replicates; ** P<0. 005, *** P<0. 0002, **** P<0. 0001 by two tailed t test. e. Quantitative data of live cell imaging showing the proportion of cells with G1 duration > 6.8 hr in HT1080 FUCCI cells following 72 hr siRNA knockdown of ZNF827 and 1 hr incubation with 2 μg/mL topotecan or DMSO. Data presented as mean ± SEM from 40 cells per replicate from three biological replicates; * P<0. 05, ** P<0. 005, by two tailed t test. f. Quantitative data of live cell imaging showing S/G2-M duration in HT1080 FUCCI cells following 72 hr siRNA knockdown of ZNF827 and 1 hr incubation with 2 μg/mL topotecan or DMSO. Data presented as mean ± SEM from 40 cells per replicate from three biological replicates; *** P<0. 0002, **** P<0. 0001 by two tailed t test.

179

Figure 5.13 ZNF827 depletion leads to abnormal mitotic entry, upregulation of p21 and suppression of the ATR-CHK1 pathway. a. Quantitative data of live cell imaging showing the proportion of cells bypassing mitosis in HT1080 FUCCI cells following 72 hr siRNA knockdown of ZNF827 and 1 hr incubation with 2 μg/mL topotecan or DMSO. Data presented as mean ± SEM from 40 cells per replicate from three biological replicates. b. Quantitative data of live cell imaging showing the proportion of cells undergoing mitosis without a S/G2 phase in HT1080 FUCCI cells

180

following 72 hr siRNA knockdown of ZNF827 and 1 hr incubation with 2 μg/mL topotecan or DMSO. Data presented as mean ± SEM from 40 cells per replicate from three biological replicates; ** P<0. 005 by two tailed t test. c. Western blot analysis of p21, p53, p-CHK1 (S345), total CHK1, p-RPAs (S33) and total RPA2 in HT1080 cells following 72 hr siRNA knockdown of ZNF827 and 1 hr incubation with 2 μg/mL topotecan or DMSO, harvested immediately, or 24 hr after drug removal. Vinculin used as a loading control.

5.10 ZNF827 depletion reduces cell growth and may sensitise cancer cells to topotecan treatment

Considering the effects of ZNF827 depletion on DNA replication and the cell cycle, we examined the impact of ZNF827 on cell proliferation and cell viability. Live cell assays were performed on U-2 OS cells in the IncuCyte® to monitor real-time cell proliferation by confluency and cell viability using Nucview488, a fluorescent reporting dye for caspase 3 activity that allows quantitative measurement of apoptotic activity.

Cells were treated with topotecan or DMSO control for the first half of the assay and treatments were removed for the remaining time. Topotecan effectively inhibited cell growth and its inhibitory effects lasted till the end of experiment, even after drug removal (Figure 5.14 a). Consistent with G1 and early S phase arrest, ZNF827 depletion reduced cell proliferation, surprisingly to a similar efficacy as topotecan.

ZNF827 depletion in conjunction with topotecan completely abrogated cell proliferation, indicative of a synergistic effect (Figure 5.14 a). Furthermore, ZNF827 depletion alone caused significantly higher apoptotic activity compared to scrambled control (Figure

5.14 b). Topotecan triggered much greater apoptotic activity as expected, which was exceeded markedly by the combination of ZNF287 depletion and topotecan towards the end of the drug-on period till the end of experiment (Figure 5.14 b). This suggests

181

that ZNF827 inhibition confers sensitisation to topotecan treatment, perhaps by muting

ATR-CHK1 activation that promotes DNA repair and cell survival. The antiproliferative

and apoptotic effects of ZNF827 depletion coincided with p21 induction, suggesting

p21 dependent growth arrest and apoptosis (Figure 5.14 c). Together, these data

suggest that ZNF827 loss is detrimental to cell proliferation and may sensitise cancer

cells to DNA damaging chemotherapeutics such as topotecan. These findings may be

exploited for synthetic lethality.

Towards the goal of exploiting ZNF827 inhibition as a novel therapeutic strategy, it is of fundamental importance to define the structure and function of ZNF827. To gain insight into the structural component of ZNF827 involved in mediating cell proliferation, we performed a cell growth rescue assay in ZNF827 CRISPR KO U-2 OS cells by re- introducing wild-type and ZNF827 mutants. Consistently, cell proliferation was inhibited in ZNF827 KO cells compared to normal U-2 OS cells. Wild-type ZNF827 did not fully rescue cell growth, while the most recovery was observed following exogenous expression of the RRK and SUMO mutants (Figure 5.14 d). Comparatively, reintroduction of the three zinc finger mutants displayed lower recue efficiencies overall, with the ZnF1-3 del mutant showing the least growth recovery until 96 hours onwards, consistent with the role of this zinc finger cluster in mediating DNA binding.

Nevertheless, the differences were not significant enough to make any conclusions.

182

183

Figure 5.14 ZNF827 depletion inhibits cell growth and sensitises cancer cells to topotecan treatment. a. Growth curves by confluency obtained from IncuCyte® live cell assays of U-2 OS cells with ZNF827 depleted by siRNA and incubation with 2 μg/mL topotecan or DMSO for the first 42 hours. Data presented as mean ± SEM from three biological replicates. b. Apoptotic activity assay quantitatively measures caspase 3 activity by Nucview488 fluorescent signals over time in IncuCyte® in U-2 OS cells with ZNF827 depleted by siRNA and incubation with Nucview 488 and 2 μg/mL topotecan or DMSO for the first 42 hours. Data presented as mean ± SEM from three biological replicates. c. Western blot analysis of p21 following 72 hr siRNA knockdown of ZNF827 and 24 hr incubation with 2 μg/mL topotecan or DMSO. Vinculin used as a loading control. d. 5-day growth curves of U-2 OS, U-2 OS ZNF827 CRISPR knockout and U-2 OS ZNF827 CRISPR knockout cells overexpressing wild-type ZNF827 and the ZNF827 mutants, harvested daily for 5 days.

5.11 ZNF827 depletion increases NuRD components

Finally, we briefly explored the effects of ZNF827 depletion on the NuRD

complex. ZNF827 depletion induced two NuRD components HDAC1 and MTA1 at the

protein levels as well as transcriptionally in both U-2 OS and HT1080 cells (Figure

5.15 a, b, c). The induction was observed in both unchallenged and topotecan or aphidicolin treated cells, and therefore independent of induced DNA damage (Figure

5.15 a, b). It has not been examined whether the other NuRD components are also affected. Whether the induction of NuRD components partly accounts for the findings associated with ZNF827 depletion remains to be determined. The NuRD complex is known to be involved in many vital cellular processes, some of which include DNA replication, cell cycle progression and cell proliferation. Given the interaction between

ZNF827 and NuRD, it will be prudent to further investigate whether NuRD is involved in these cellular processes in concert with ZNF827.

184

Figure 5.15 ZNF827 depletion increases NuRD components. a. Western blot analysis of MTA1 and HDAC1 in U-2 OS and HT1080 cells following 72 hr siRNA knockdown of ZNF827 and 24 hr incubation with 2 μg/mL topotecan or DMSO. Vinculin used as a loading control. b. Western blot analysis of HDAC1 in U-2 OS and HT1080 cells following 72 hr siRNA knockdown of ZNF827 and 20 hr incubation with 1 μM aphidicolin or DMSO. Vinculin used as a loading control. c. Quantitative RT-PCR showing transcript levels relative to empty vector control of ZNF827, HDAC1 and MTA in U-2 OS cells and of MTA in HT1080 cells following 72 hr siRNA knockdown of ZNF827 and 24 hr incubation with 2 μg/mL topotecan or DMSO. (mean ± SD; n = 3 technical replicates; *P<0. 05, ** P<0. 005, **** P<0. 0001 by two tailed t test).

5.12 Discussion

In this study, we have identified and characterised ZNF827 as a novel ssDNA binding protein involved in the activation of the ATR-CHK1 signalling pathway that responds to replication-associated DNA damage. This work extends the role of

185

ZNF827 beyond its very specific function at ALT telomeres. We provide several lines

of evidence to support the role of ZNF827 as a novel DNA damage and repair protein

both at telomeres and genome-wide.

We have established the DNA binding capabilities of ZNF827. Specifically,

ZNF827 shows preferential binding affinity to ss G-rich telomeric DNA and non-

telomeric ssDNA, conferred by its first cluster of zinc fingers, i.e. zinc fingers 1-3. This

explains the predominant localisation of ZNF827 at ALT telomeres, as ALT cells

characteristically have high levels of ssDNA originating from intrinsic G-rich overhangs

and ssDNA intermediates that form from stalled replication forks and during HR.

Interestingly, the binding of ZNF827 to random ssDNA sequences, in addition to ss G- rich telomeric DNA, indicates a non-specific binding affinity; however, the absence of binding to ss C-rich telomeric DNA suggests a certain level of binding specificity that remains to be determined. Essentially, our data demonstrate that ZNF827 is a novel ssDNA binding protein capable of binding to telomeric and non-telomeric ssDNA sequences.

The main eukaryotic ssDNA binding protein RPA binds ssDNA in a non-specific manner and is required for almost all cellular processes involving the formation of ssDNA, such as DNA replication, recombination, the DNA damage checkpoint response, and the major DNA repair pathways including nucleotide excision repair

(NER), base excision repair (BER), DNA mismatch repair (MMR), and HR-directed

DSB repair (Zou et al., 2006). We have shown that ZNF827 directly interacts with RPA independently of TMM, and that ZNF827 depletion intensifies RPA foci following induction of DNA damage during replication, suggesting a role for ZNF827 in the cellular response to replication-associated DNA damage. It has previously been proposed that RPA is required to prevent G-quadruplex formation during replication of

186

lagging-strand telomeres (Audry et al., 2015). Considering our findings that ZNF827

preferentially binds to G-rich ss telomeric DNA and directly interacts with RPA, it is

possible that ZNF827 is involved in suppressing G-rich secondary structures during

replication at telomeres as well as other genomic regions.

hnRNP-A1 (heterogenous nuclear ribonucleoprotein A1) has been implicated

in telomere maintenance and can also bind to G-rich ss telomeric DNA as well as the

RNA component of telomerase (Fiset and Chabot, 2001, Zhang et al., 2006). Recent

evidence has revealed that hnRNP-A1 plays a critical role in orchestrating cell cycle-

regulated RPA displacement from newly synthesised G-rich ss telomeric DNA to facilitate the RPA-to-POT1 switch post DNA replication, hence preventing unnecessary activation of ATR-mediated DNA damage signalling pathways at telomeres (Flynn et al., 2011). This process has not been explored at the telomeres of

ALT cells, which we have demonstrated to have a high baseline level of RPA foci

indicative of persistently active ATR-mediated DNA damage signalling. Given the

common ssDNA substrates and interconnections between ZNF827, RPA and hnRNP-

A1, it may be worthwhile to explore whether the RPA-to-POT1 switch is deficient at

ALT telomeres and whether ZNF827 is involved in this process.

In line with the ZNF827 and RPA interaction, our data have further implicated

ZNF827 in the ATR DNA damage signalling pathway. Specifically, we demonstrated that depletion of ZNF827 led to a significant reduction in topotecan-induced CHK1 and

RPA phosphorylation, indicative of inhibition of the ATR signalling pathway. These data suggest that ZNF827 plays a role in the activation of the ATR pathway; however, the precise function of ZNF827 is unknown. We described some evidence to suggest a weak direct interaction between ZNF827 and the ATR kinase. However, the preliminary PLA data were not definitive due to the lack of appropriate controls and

187

quantitative analysis. Further PLA experiments with improved design need to be conducted to ascertain a direct interaction between ZNF827 and ATR. Complete activation of the ATR pathway requires recruitment of the ATR-ATRIP complex by

RPA-coated ssDNA to DNA damage sites and stressed replication forks, and the sequential activation of the ATR-ATRIP complex by its regulators including the 9-1-1 complex and TOPBP1 (Flynn and Zou, 2011). The details of the involvement of

ZNF827 in these processes remain to be defined. Nevertheless, the key functions of the ATR signalling pathway in the cellular response to replication stress (Flynn and

Zou, 2011), the involvement of ZNF827 in this pathway, together with its interaction with RPA and ssDNA strongly implicate ZNF827 in the ATR-mediated replication stress response and the repair of replication-associated DNA damage.

We have presented further evidence to support ZNF827 playing a role in the cellular response to DNA damage. Notably, endogenous ZNF827 was greatly induced by replication-associated DNA damage. ZNF827 localised at DNA damage sites, and localisations were noticeably increased by IR or topotecan induced DNA damage in both ALT and telomerase-positive cells. Furthermore, ZNF827 depletion induced global DNA damage to a similar extent as topotecan. Consistent with a previous report from our lab that ZNF827 suppresses DNA damage at telomeres (Conomos et al.,

2014), IR rapidly increased ZNF827 colocalisations with telomeres, which gradually regressed to baseline over time.

Direct evidence linking ZNF827 to replication stress and repair of replication- associated DNA damage has been provided in this study. Robust colocalisations of

ZNF827 with aphidicolin-induced PCNA foci in both ALT and telomerase-positive cells indicate that ZNF827 associates with stalled replication forks genome-wide. The induction of replication defects at telomeres and genomically by ZNF827 depletion

188

depicts a critical role for ZNF827 in faithful DNA replication globally. Further, depletion

of ZNF827 significantly reduced SCEs, indicating inhibition of HR-mediated DNA

repair. The effect of ZNF827 depletion on replication and HR-mediated repair may be

explained by concomitant inhibition of the ATR-CHK1 signalling pathway, as described

previously. In addition to regulating the cellular response to replication stress, both

ATR and CHK1 are required for HR-mediated repair (Wang et al., 2004, Sorensen et

al., 2005a, Buisson et al., 2017, Flynn and Zou, 2011). Based on our data, we propose

that ZNF827 plays an important role in HR-mediated DNA repair at stalled or stressed replication forks by regulating the ATR-CHK1 pathway to ensure faithful DNA repair and replication.

Replication defects observed in ZNF827-depleted cells coincided with

disrupted cell cycle progression, implicating ZNF827 in the cell cycle checkpoint

response. We observed a prominent G1/S arrest accompanied by upregulation of p21

following ZNF827 depletion alone, suggesting p53-p21 dependent cell cycle arrest

triggered by the accumulation of ZNF827 depletion-induced replicative DNA damage.

Remarkably, topotecan induced activation of the ATR-CHK1 pathway was efficiently

inhibited in ZNF827 depleted cells after initial activation, despite persistent DNA

damage. This further supports ZNF827 as a positive regulator of the ATR-CHK1

pathway. In response to DNA damage sustained during replication, the ATR-CHK1 pathway activates the G2/M DNA damage checkpoint to ensure efficient DNA repair and promote cell survival (Kastan and Bartek, 2004, Liu et al., 2000). Accordingly, loss of ATR-CHK1 activation in ZNF827 depleted cells inflicted with DNA damage suggests a defective G2/M DNA damage checkpoint, which may explain the decreased G2/M arrest observed in ZNF827-depleted cells, and the proportion of ZNF827 depleted cells undergoing mitosis by bypassing S phase. Effectively, cells undergoing mitosis

189

with accumulated DNA damage arrest in the following G1 phase in a p53-p21

dependent manner (Kastan and Bartek, 2004), accounting for ZNF827 depletion

induced G1/S arrest and persistent p53 and p21 induction despite loss of ATR-CHK1 activation. Therefore, the prominent G1/S arrest observed after ZNF827 depletion is likely to be a consequence of a defective G2/M DNA damage checkpoint resulting from

ZNF827 depletion induced inactivation of the ATR-CHK1 pathway.

ATR-CHK1 mediated DNA damage checkpoint signalling plays a critical role in promoting efficient DNA repair, thus safeguarding genome stability and cell survival.

Deletion of ATR or CHK1 results in lethality in human cells and embryonic lethality in mice (Marechal and Zou, 2013, Liu et al., 2000). Consistent with ZNF827 playing a role in the ATR-CHK1 pathway and cell cycle arrest, we have further demonstrated that ZNF827 depletion led to a significant reduction in the proliferation of the osteosarcoma cell line U-2 OS, comparable to topotecan treatment, whereas topotecan in conjunction with ZNF827 depletion almost completely abolished cell growth. In addition, there was increased apoptotic activity in ZNF827 depleted cells, which was markedly enhanced in conjunction with topotecan, suggestive of a synergistic effect of ZNF827 depletion and topotecan treatment. These antiproliferative and proapoptotic effects associated with ZNF827 depletion alone and the synergistic action with topotecan are accompanied by p21 upregulation, and likely to be linked to inactivation of the ATR-CHK1 pathway.

ATR inhibition has been under vigorous investigation for its therapeutic potential in cancer therapy on its own and in combination with various chemotherapies.

The proposed rationale is that cancer cells with intrinsically elevated levels of replication stress are particularly susceptible to ATR inhibition (Fokas et al., 2014, Zou,

2018). It has been proposed that cancer cells utilising the ALT pathway are

190

hypersensitive to ATR inhibitors, likely attributed to the high level of replication stress

at ALT telomeres (Flynn et al., 2015). Furthermore, disruption of the DNA damage

response and repair pathway by ATR inhibition enhances the cytotoxic effects of

chemotherapeutic drugs in cancer cells (Fokas et al., 2014, Zou, 2018).

ATR inhibition has been shown to sensitise cancer cells to topoisomerase

inhibitors such as camptothecin and topotecan (Cliby et al., 2002, Flatten et al., 2005).

Very recently, a successful phase I clinical study of combination therapy with ATR

inhibitor M6620 and topotecan in patients with advanced solid tumours has been

reported (Thomas et al., 2018). Our data have demonstrated an important role for

ZNF827 in the activation of the ATR-CHK1 pathway, and provided preliminary evidence for the synergistic action of ZNF827 depletion in combination with topotecan in cancer cells. These findings strongly support further studies to define the precise functions of ZNF827 in the ATR-CHK1 pathway and explore the effects of ZNF827 depletion in a panel of mortal cell strains and cancer cell lines, in order to validate

ZNF827 as a novel therapeutic target in the ATR-CHK1 pathway for synthetic lethality with DNA damaging chemotherapeutics.

There are several limitations in this study. First, although the results were significant, some experiments have not yet been repeated. To improve the robustness and statistical power of the data, reproducibility must be demonstrated in biological replicates. Second, the CRISPR ZNF827 KO U-2 OS cell line was only validated by sequencing initially because endogenous ZNF827 cannot be detected by Western blot.

Due to the possibility of ZNF827 reactivation in cancer cells, it is imperative to ascertain the knockout status of ZNF827 periodically during subculturing. It is possible that the inconclusive results from the growth recovery experiment may have been confounded by some residual ZNF827. If further recovery experiments are to be

191

performed in the ZNF827 KO U-2 OS cell line, the absence of endogenous ZNF827 should be confirmed by experimental techniques other than Western blot, such as IF or PCR with primers designed to amplify the deleted regions. Lastly, the experiment that demonstrated topotecan sensitisation to ZNF827 depletion was only performed in one cancer cell line with one concentration of topotecan. Even though the results are promising and bear great implications for the validation of ZNF827 as a therapeutic target in anticancer therapies, they are limited to one cell line. Further experiments must be designed to test a range of topotecan concentrations with ZNF827 depletion in a panel of representative cell lines and cell strains.

Overall, this study has extended our knowledge of ZNF827 beyond its specific function at ALT telomeres. We have defined ZNF827 as a novel ssDNA binding protein involved in the cellular response to replication-associated DNA damage. We have demonstrated a direct interaction between ZNF827 and RPA, and provided evidence to support the function of ZNF827 as a positive regulator of the ATR-CHK1 DNA damage signalling pathway, the ATR-mediated DNA damage checkpoint, and HR- mediated DNA repair, in response to replication stress. We propose that ZNF827 is a novel molecular target for inhibition of the ATR pathway and that ZNF827 inhibition confers synthetic lethality with DNA damaging chemotherapies.

192

Chapter 6 Final Discussion

This thesis has focused on the functional characterisation of the poorly understood zinc finger protein ZNF827. ZNF827 was first reported by our lab as an

ALT-associated protein involved in promoting the HR-mediated ALT TMM in cancer cells, in concert with the NuRD complex (Conomos et al., 2014). The specificity and multifaceted functional roles of ZNF827 at ALT telomeres led us to hypothesise that

ZNF827 is a promising molecular target in the 10-15% of cancers that utilise the ALT pathway, for which suitable therapeutic targets are currently lacking. This hypothesis forms the basis of this research project.

Towards the ultimate goal of validating ZNF827 as a therapeutic target in cancers, this thesis has contributed considerable novel knowledge on the C2H2 zinc finger protein ZNF827, which provides supporting evidence and a foundation for the future development of anticancer therapies targeting ZNF827. The functional characterisation of ZNF827 was conducted from several perspectives. We have explored the biological functions of ZNF827 as a transcription factor and shed light on an important role for ZNF827 in embryonic development, and that the function of

ZNF827 at ALT telomeres is not dependent on its transcriptional impact. We have gained new insights into the recruitment of ZNF827 to ALT telomeres, characterised structurally and functionally the interaction interface between ZNF827 and the NuRD complex, and implicated ZNF827 SUMOylation in promoting ALT activity. We have also revealed that ZNF827 is novel ssDNA binding protein implicated in HR-directed repair of replication-associated DNA damage at telomeres as well as genome-wide, and that its role as a DNA damage response and repair protein is likely to be mediated through the ATR-mediated DNA signalling pathway. Finally, we have provided

193

preliminary evidence that supports synthetic lethality between ZNF827 inhibition and the topoisomerase I inhibitor topotecan.

The most compelling discovery presented in this thesis is that ZNF827 is a novel ssDNA binding protein and a novel player in the ATR-mediated DNA damage signalling pathway involved in HR-mediated repair of replication-associated DNA damage. The precise mechanistic details regarding how ZNF827 is involved in the

ATR pathway need to be defined, but we speculate that it is dependent on the ssDNA and RPA binding capability of ZNF827, given that RPA coated ssDNA structures are the key mediator of ATR activation (Flynn and Zou, 2011, Cimprich and Cortez, 2008).

It seems likely that the role of ZNF827 in embryonic development and its specific association with ALT telomeres may be mechanistically linked to its function in ATR-mediated repair of replication-associated DNA damage. There are some intriguing similarities between embryonic stem cells and ALT telomeres. Both ESCs and ALT telomeres display an elevated level of endogenous replication stress due to the high proliferative rate in early embryogenesis (Ahuja et al., 2016) and structural aberrations at ALT telomeres, respectively. In mouse ESCs, the high level of replication stress has been associated with constitutive DDR activation and ssDNA accumulation (Chuykin et al., 2008, Ahuja et al., 2016), which are also characteristics of ALT telomeres. The constitutive γH2AX appearance in mouse ESCs is dependent on the ATR kinase, but not ATM, consistent with ATR being the key mediator of the cellular response to replication stress. Despite displaying marked features of replication stress, ESCs display a very low level of replication defects. It is not well understood how ESCs cope with this endogenous replication stress.

This thesis has depicted an important role for ZNF827 both in embryonic development that is correlated with haploinsufficiency, and in the repair of ATR-

194

mediated replication associated DNA damage. Loss of ATR results in embryonic

lethality (Brown and Baltimore, 2000), but it has not been fully elucidated how ATR

protects ESCs against replication stress. Together with our data, it is reasonable to

speculate that ZNF827 may be involved in a replication stress coping mechanism

mediated by ATR in early embryogenesis. Some ALT cancer cells are found to be

hypersensitive to ATR inhibition, indicating the importance of the ATR pathway at ALT

telomeres, likely attributable to its role in counteracting replication stress (Flynn et al.,

2015). This thesis provides strong support for further investigations to elucidate the

role of ZNF827 in ATR-mediated repair of replication-associated DNA damage in the

context of embryonic development and ALT telomeres.

The mechanistic details of ZNF827 recruitment to ALT telomeres have not been

clearly defined. Based on the findings in this thesis, two mechanisms may be proposed.

First, ZNF827 may be recruited by direct binding to ssDNA. ALT telomeres contain a

high abundance of ssDNA that can potentially be bound directly by ZNF827. The fact

that the ZNF827 ZnF4-9 del mutant retains ssDNA binding ability but does not localise at ALT telomeres, suggests the involvement of another factor in mediating the binding.

Second, TERRA may be involved in nuclear receptor-mediated ZNF827 recruitment.

Our data exclude a direct interaction between nuclear receptors and ZNF827, and implicate TERRA as the intermediator. Given that ZNF827 binds to ssDNA species, it is likely that ZNF827 also binds RNA. In support of this hypothesis, ZNF827 has been identified in a PhD thesis as an important regulator in the alternative splicing of genes involved in EMT (Sahu, 2017). Our findings warrant future studies to examine the binding ability of ZNF827 to RNAs, and whether ZNF827 binding to ssDNA or RNA is involved in the recruitment of ZNF827 to telomeres.

195

In this thesis, we have provided structural and functional analysis of ZNF827

regarding the NuRD-binding RRK motif and the zinc finger clusters. We have

confirmed that the RRK motif directly binds to the RBBP4 subunit of the NuRD complex,

and identified the specific amino acid residues required for binding. Since the

interaction between ZNF287 and NuRD is essential for the ALT-promoting functions

of ZNF827, the structural data in this thesis will potentially be useful to guide the

development of modulators targeting ZNF827-NuRD in ALT cancers.

ZNF827-NuRD promotes telomere-telomere interactions in ALT cancer cells

(Conomos et al., 2014). Considering the RRK-motif can interact with multiple subunits

of NuRD, as demonstrated by FOG1 (Lejon et al., 2011), we hypothesise that telomere

tethering is mediated by ZNF827 molecules bound at separate telomeres interacting

with multiple subunits in the same NuRD complex. This hypothesis remains to be

tested in future studies. The first zinc finger cluster in ZNF827 ZnF1-3 is required for

ssDNA binding, while the other cluster ZnF4-9 is dispensable. This suggests that

ZnF4-9 is likely to be involved in protein-protein interactions. Given our finding that

ZNF827 directly interacts with RPA, it will be pertinent to further investigate whether

the ZnF4-9 zinc finger cluster is essential for RPA binding. The structural and

functional insights on ZNF827 in this thesis will provide guidance for further structural

studies to determine the structure of full-length ZNF827, and define at the molecular

level its DNA and protein binding domains.

Our data also demonstrate that ZNF827 is not only a specific molecular target in ALT, but also a potential molecular target in the ATR-mediated DNA damage signalling pathway. Essentially the functions of ZNF827 in ALT may be correlated to its role in the ATR pathway. We have described preliminary evidence to support synthetic lethality between ZNF827 inhibition and topotecan, which is consistent with

196

synthetic lethality between ATR inhibition and chemotherapeutic agents such as

topotecan (Zou, 2018). This implicates ZNF827 as a potential therapeutic target for

ATR inhibition, extending its therapeutic potential in anticancer therapies beyond ALT

cancers.

In conclusion, this thesis has substantially expanded our knowledge on the zinc

finger protein ZNF827. ZNF827 is more than an ALT-associated protein. We have shed light on the functional roles of ZNF827 as a novel regulator in embryonic development, as an important player in the ATR signalling pathway and as a potential therapeutic target for novel anticancer therapies. Our findings provide directions for future studies to further elucidate the biological significance of ZNF827.

197

References ACHARYA, S., KAUL, Z., GOCHA, A. S., MARTINEZ, A. R., HARRIS, J., PARVIN, J. D. & GRODEN, J. 2014. Association of BLM and BRCA1 during telomere maintenance in ALT cells. PLoS ONE, 9, e103819. AHUJA, A. K., JODKOWSKA, K., TELONI, F., BIZARD, A. H., ZELLWEGER, R., HERRADOR, R., ORTEGA, S., HICKSON, I. D., ALTMEYER, M., MENDEZ, J. & LOPES, M. 2016. A short G1 phase imposes constitutive replication stress and fork remodelling in mouse embryonic stem cells. Nat Commun, 7, 10660. ALLEN, H. F., WADE, P. A. & KUTATELADZE, T. G. 2013. The NuRD architecture. Cell Mol Life Sci, 70, 3513-24. AMIEL, J., SPROAT-EMISON, E., GARCIA-BARCELO, M., LANTIERI, F., BURZYNSKI, G., BORREGO, S., PELET, A., ARNOLD, S., MIAO, X., GRISERI, P., BROOKS, A. S., ANTINOLO, G., DE PONTUAL, L., CLEMENT-ZIZA, M., MUNNICH, A., KASHUK, C., WEST, K., WONG, K. K., LYONNET, S., CHAKRAVARTI, A., TAM, P. K., CECCHERINI, I., HOFSTRA, R. M., FERNANDEZ, R. & HIRSCHSPRUNG DISEASE, C. 2008. Hirschsprung disease, associated syndromes and genetics: a review. J Med Genet, 45, 1-14. APARICIO, T., BAER, R. & GAUTIER, J. 2014. DNA double-strand break repair pathway choice and cancer. DNA Repair (Amst), 19, 169-75. ARORA, R., LEE, Y., WISCHNEWSKI, H., BRUN, C. M., SCHWARZ, T. & AZZALIN, C. M. 2014. RNaseH1 regulates TERRA-telomeric DNA hybrids and telomere maintenance in ALT tumour cells. Nat. Commun, 5, 5220. ARTANDI, S. E., CHANG, S., LEE, S. L., ALSON, S., GOTTLIEB, G. J., CHIN, L. & DEPINHO, R. A. 2000. Telomere dysfunction promotes non-reciprocal translocations and epithelial cancers in mice. Nature, 406, 641-645. ARTANDI, S. E. & DEPINHO, R. A. 2010. Telomeres and telomerase in cancer. Carcinogenesis, 31, 9-18. AUDRY, J., MAESTRONI, L., DELAGOUTTE, E., GAUTHIER, T., NAKAMURA, T. M., GACHET, Y., SAINTOME, C., GELI, V. & COULON, S. 2015. RPA prevents G- rich structure formation at lagging-strand telomeres to allow maintenance of chromosome ends. EMBO J, 34, 1942-1958. AVRAM, D. & CALIFANO, D. 2014. The multifaceted roles of Bcl11b in thymic and peripheral T cells: impact on immune diseases. J Immunol, 193, 2059-65. AYLON, Y., LIEFSHITZ, B. & KUPIEC, M. 2004. The CDK regulates repair of double- strand breaks by homologous recombination during the cell cycle. EMBO J, 23, 4868-75. BADIE, S., ESCANDELL, J. M., BOUWMAN, P., CARLOS, A. R., THANASOULA, M., GALLARDO, M. M., SURAM, A., JACO, I., BENITEZ, J., HERBIG, U., BLASCO, M. A., JONKERS, J. & TARSOUNAS, M. 2010. BRCA2 acts as a RAD51 loader to facilitate telomere replication and capping. Nat. Struct. Mol. Biol, 17, 1461- 1469. BATEMAN, E. 1998. Autoregulation of eukaryotic transcription factors. Prog Nucleic Acid Res Mol Biol, 60, 133-68. BAUMANN, P. & CECH, T. R. 2001. Pot1, the putative telomere end-binding protein in fission yeast and humans. Science, 292, 1171-1175. BECHTER, O. E., ZOU, Y., SHAY, J. W. & WRIGHT, W. E. 2003. Homologous recombination in human telomerase-positive and ALT cells occurs with the same frequency. EMBO Rep, 4, 1138-1143.

198

BERNARDI, R. & PANDOLFI, P. P. 2007. Structure, dynamics and functions of promyelocytic leukaemia nuclear bodies. Nat. Rev. Mol. Cell Biol, 8, 1006-1016. BHATTACHARYYA, S., KEIRSEY, J., RUSSELL, B., KAVECANSKY, J., LILLARD- WETHERELL, K., TAHMASEB, K., TURCHI, J. J. & GRODEN, J. 2009. Telomerase associated protein 1, HSP90 and topoisomerase IIα associate directly with the BLM helicase in immortalized cells using ALT and modulate its helicase activity using telomeric DNA substrates. J. Biol. Chem, 284, 14966- 14977. BOSCO, G. & HABER, J. E. 1998. Chromosome break-induced DNA replication leads to nonreciprocal translocations and telomere capture. Genetics, 150, 1037- 1047. BOTZENHART, E. M., BARTALINI, G., BLAIR, E., BRADY, A. F., ELMSLIE, F., CHONG, K. L., CHRISTY, K., TORRES-MARTINEZ, W., DANESINO, C., DEARDORFF, M. A., FRYNS, J. P., MARLIN, S., GARCIA-MINAUR, S., HELLENBROICH, Y., HAY, B. N., PENTTINEN, M., SHASHI, V., TERHAL, P., VAN MALDERGEM, L., WHITEFORD, M. L., ZACKAI, E. & KOHLHASE, J. 2007. Townes-Brocks syndrome: twenty novel SALL1 mutations in sporadic and familial cases and refinement of the SALL1 hot spot region. Hum Mutat, 28, 204-5. BOWEN, N. J., FUJITA, N., KAJITA, M. & WADE, P. A. 2004. Mi-2/NuRD: multiple complexes for many purposes. Biochim Biophys Acta, 1677, 52-7. BRAYER, K. J. & SEGAL, D. J. 2008. Keep your fingers off my DNA: protein-protein interactions mediated by C2H2 zinc finger domains. Cell Biochem Biophys, 50, 111-31. BROCCOLI, D., SMOGORZEWSKA, A., CHONG, L. & DE LANGE, T. 1997. Human telomeres contain two distinct Myb-related proteins, TRF1 and TRF2. Nat. Genet, 17, 231-235. BROWN, E. J. & BALTIMORE, D. 2000. ATR disruption leads to chromosomal fragmentation and early embryonic lethality. Genes Dev, 14, 397-402. BROWN, R. S. 2005. Zinc finger proteins: getting a grip on RNA. Curr Opin Struct Biol, 15, 94-8. BRYAN, T. M. 1997. An alternative mechanism for telomere lengthening in immortal human cells. PhD, The University of Sydney. BRYAN, T. M., ENGLEZOU, A., DALLA-POZZA, L., DUNHAM, M. A. & REDDEL, R. R. 1997. Evidence for an alternative mechanism for maintaining telomere length in human tumors and tumor-derived cell lines. Nat. Med, 3, 1271-1274. BRYAN, T. M., ENGLEZOU, A., GUPTA, J., BACCHETTI, S. & REDDEL, R. R. 1995. Telomere elongation in immortal human cells without detectable telomerase activity. EMBO J., 14, 4240-4248. BUISSON, R., NIRAJ, J., RODRIGUE, A., HO, C. K., KREUZER, J., FOO, T. K., HARDY, E. J., DELLAIRE, G., HAAS, W., XIA, B., MASSON, J. Y. & ZOU, L. 2017. Coupling of Homologous Recombination and the Checkpoint by ATR. Mol Cell, 65, 336-346. CECCALDI, R., RONDINELLI, B. & D'ANDREA, A. D. 2016. Repair Pathway Choices and Consequences at the Double-Strand Break. Trends Cell Biol, 26, 52-64. CESARE, A. J., HAYASHI, M. T., CRABBE, L. & KARLSEDER, J. 2013. The telomere deprotection response is functionally distinct from the genomic DNA damage response. Mol. Cell, 51, 141-155. CESARE, A. J., KAUL, Z., COHEN, S. B., NAPIER, C. E., PICKETT, H. A., NEUMANN, A. A. & REDDEL, R. R. 2009. Spontaneous occurrence of telomeric DNA

199

damage response in the absence of chromosome fusions. Nat. Struct. Mol. Biol, 16, 1244-1251. CESARE, A. J. & REDDEL, R. R. 2008. Telomere uncapping and alternative lengthening of telomeres. Mech. Ageing Dev, 129, 99-108. CESARE, A. J. & REDDEL, R. R. 2010. Alternative lengthening of telomeres: models, mechanisms and implications. Nat. Rev. Genet, 11, 319-330. CHAMBERS, J. C., ZHANG, W., SEHMI, J., LI, X., WASS, M. N., VAN DER HARST, P., HOLM, H., SANNA, S., KAVOUSI, M., BAUMEISTER, S. E., COIN, L. J., DENG, G., GIEGER, C., HEARD-COSTA, N. L., HOTTENGA, J. J., KUHNEL, B., KUMAR, V., LAGOU, V., LIANG, L., LUAN, J., VIDAL, P. M., MATEO LEACH, I., O'REILLY, P. F., PEDEN, J. F., RAHMIOGLU, N., SOININEN, P., SPELIOTES, E. K., YUAN, X., THORLEIFSSON, G., ALIZADEH, B. Z., ATWOOD, L. D., BORECKI, I. B., BROWN, M. J., CHAROEN, P., CUCCA, F., DAS, D., DE GEUS, E. J., DIXON, A. L., DORING, A., EHRET, G., EYJOLFSSON, G. I., FARRALL, M., FOROUHI, N. G., FRIEDRICH, N., GOESSLING, W., GUDBJARTSSON, D. F., HARRIS, T. B., HARTIKAINEN, A. L., HEATH, S., HIRSCHFIELD, G. M., HOFMAN, A., HOMUTH, G., HYPPONEN, E., JANSSEN, H. L., JOHNSON, T., KANGAS, A. J., KEMA, I. P., KUHN, J. P., LAI, S., LATHROP, M., LERCH, M. M., LI, Y., LIANG, T. J., LIN, J. P., LOOS, R. J., MARTIN, N. G., MOFFATT, M. F., MONTGOMERY, G. W., MUNROE, P. B., MUSUNURU, K., NAKAMURA, Y., O'DONNELL, C. J., OLAFSSON, I., PENNINX, B. W., POUTA, A., PRINS, B. P., PROKOPENKO, I., PULS, R., RUOKONEN, A., SAVOLAINEN, M. J., SCHLESSINGER, D., SCHOUTEN, J. N., SEEDORF, U., SEN-CHOWDHRY, S., SIMINOVITCH, K. A., SMIT, J. H., SPECTOR, T. D., TAN, W., TESLOVICH, T. M., TUKIAINEN, T., UITTERLINDEN, A. G., VAN DER KLAUW, M. M., VASAN, R. S., WALLACE, C., WALLASCHOFSKI, H., WICHMANN, H. E., WILLEMSEN, G., WURTZ, P., XU, C., YERGES-ARMSTRONG, L. M., et al. 2011. Genome-wide association study identifies loci influencing concentrations of liver enzymes in plasma. Nat Genet, 43, 1131-8. CHEN, H., LISBY, M. & SYMINGTON, L. S. 2013. RPA coordinates DNA end resection and prevents formation of DNA hairpins. Mol Cell, 50, 589-600. CHEN, W. Y., WANG, D. H., YEN, R. C., LUO, J., GU, W. & BAYLIN, S. B. 2005. Tumor suppressor HIC1 directly regulates SIRT1 to modulate p53-dependent DNA-damage responses. Cell, 123, 437-48. CHEN, Y. A., SHEN, Y. L., HSIA, H. Y., TIANG, Y. P., SUNG, T. L. & CHEN, L. Y. 2017. Extrachromosomal telomere repeat DNA is linked to ALT development via cGAS-STING DNA sensing pathway. Nat Struct Mol Biol, 24, 1124-1131. CHIRUVELLA, K. K., LIANG, Z. & WILSON, T. E. 2013. Repair of double-strand breaks by end joining. Cold Spring Harb Perspect Biol, 5, a012757. CHO, N. W., DILLEY, R. L., LAMPSON, M. A. & GREENBERG, R. A. 2014. Interchromosomal homology searches drive directional ALT telomere movement and synapsis. Cell, 159, 108-121. CHONG, L., VAN STEENSEL, B., BROCCOLI, D., ERDJUMENT-BROMAGE, H., HANISH, J., TEMPST, P. & DE LANGE, T. 1995. A human telomeric protein. Science, 270, 1663-1667. CHOU, D. M., ADAMSON, B., DEPHOURE, N. E., TAN, X., NOTTKE, A. C., HUROV, K. E., GYGI, S. P., COLAIACOVO, M. P. & ELLEDGE, S. J. 2010. A chromatin localization screen reveals poly (ADP ribose)-regulated recruitment of the

200

repressive polycomb and NuRD complexes to sites of DNA damage. Proc. Natl. Acad. Sci. U. S. A, 107, 18475-18480. CHUYKIN, I. A., LIANGUZOVA, M. S., POSPELOVA, T. V. & POSPELOV, V. A. 2008. Activation of DNA damage response signaling in mouse embryonic stem cells. Cell Cycle, 7, 2922-8. CIMPRICH, K. A. & CORTEZ, D. 2008. ATR: an essential regulator of genome integrity. Nat. Rev. Mol. Cell Biol, 9, 616-627. CISMASIU, V. B., ADAMO, K., GECEWICZ, J., DUQUE, J., LIN, Q. & AVRAM, D. 2005. BCL11B functionally associates with the NuRD complex in T lymphocytes to repress targeted promoter. Oncogene, 24, 6753-6764. CLIBY, W. A., LEWIS, K. A., LILLY, K. K. & KAUFMANN, S. H. 2002. S phase and G2 arrests induced by topoisomerase I poisons are dependent on ATR kinase function. J Biol Chem, 277, 1599-606. CLYNES, D., JELINSKA, C., XELLA, B., AYYUB, H., SCOTT, C., MITSON, M., TAYLOR, S., HIGGS, D. R. & GIBBONS, R. J. 2015. Suppression of the alternative lengthening of telomere pathway by the chromatin remodelling factor ATRX. Nat. Commun, 6, 7538. COHEN, S. B., GRAHAM, M. E., LOVRECZ, G. O., BACHE, N., ROBINSON, P. J. & REDDEL, R. R. 2007. Protein composition of catalytically active human telomerase from immortal cells. Science, 315, 1850-1853. CONOMOS, D. 2014. Alternative lengthening of telomeres: Remodelling the telomere architecture. Doctor of Philosophy, University of Sydney. CONOMOS, D., PICKETT, H. A. & REDDEL, R. R. 2013. Alternative lengthening of telomeres: remodeling the telomere architecture. Front. Oncol, 3, 27. CONOMOS, D., REDDEL, R. R. & PICKETT, H. A. 2014. NuRD-ZNF827 recruitment to telomeres creates a molecular scaffold for homologous recombination. Nat. Struct. Mol. Biol, 21, 760-770. CONOMOS, D., STUTZ, M. D., HILLS, M., NEUMANN, A. A., BRYAN, T. M., REDDEL, R. R. & PICKETT, H. A. 2012. Variant repeats are interspersed throughout the telomeres and recruit nuclear receptors in ALT cells. J. Cell Biol, 199, 893-906. COONEY, A. J., LENG, X., TSAI, S. Y., O'MALLEY, B. W. & TSAI, M. J. 1993. Multiple mechanisms of chicken ovalbumin upstream promoter transcription factor- dependent repression of transactivation by the vitamin D, thyroid hormone, and retinoic acid receptors. J Biol Chem, 268, 4152-60. COONEY, A. J., TSAI, S. Y., O'MALLEY, B. W. & TSAI, M. J. 1992. Chicken ovalbumin upstream promoter transcription factor (COUP-TF) dimers bind to different GGTCA response elements, allowing COUP-TF to repress hormonal induction of the vitamin D3, thyroid hormone, and retinoic acid receptors. Mol Cell Biol, 12, 4153-63. COSTA, A., DAIDONE, M. G., DAPRAI, L., VILLA, R., CANTU, S., PILOTTI, S., MARIANI, L., GRONCHI, A., HENSON, J. D., REDDEL, R. R. & ZAFFARONI, N. 2006. Telomere maintenance mechanisms in liposarcomas: association with histologic subtypes and disease progression. Cancer Res, 66, 8918-8924. COSTANTINO, L., SOTIRIOU, S. K., RANTALA, J. K., MAGIN, S., MLADENOV, E., HELLEDAY, T., HABER, J. E., ILIAKIS, G., KALLIONIEMI, O. P. & HALAZONETIS, T. D. 2014. Break-induced replication repair of damaged forks induces genomic duplications in human cells. Science, 343, 88-91. COUNTER, C. M., AVILION, A. A., LEFEUVRE, C. E., STEWART, N. G., GREIDER, C. W., HARLEY, C. B. & BACCHETTI, S. 1992. Telomere shortening

201

associated with chromosome instability is arrested in immortal cells which express telomerase activity. EMBO J, 11, 1921-1929. COX, K. E., MARECHAL, A. & FLYNN, R. L. 2016. SMARCAL1 Resolves Replication Stress at ALT Telomeres. Cell Rep, 14, 1032-40. CRABBE, L., VERDUN, R. E., HAGGBLOM, C. I. & KARLSEDER, J. 2004. Defective telomere lagging strand synthesis in cells lacking WRN helicase activity. Science, 306, 1951-1953. CREWS, S. T. & PEARSON, J. C. 2009. Transcriptional autoregulation in development. Curr Biol, 19, R241-6. D'ADDA DI FAGAGNA, F., REAPER, P. M., CLAY-FARRACE, L., FIEGLER, H., CARR, P., VON ZGLINICKI, T., SARETZKI, G., CARTER, N. P. & JACKSON, S. P. 2003. A DNA damage checkpoint response in telomere-initiated senescence. Nature, 426, 194-198. DAI, X., HUANG, C., BHUSARI, A., SAMPATHI, S., SCHUBERT, K. & CHAI, W. 2010. Molecular steps of G-overhang generation at human telomeres and its function in chromosome end protection. EMBO J, 29, 2788-27801. DAN, J., ROUSSEAU, P., HARDIKAR, S., VELAND, N., WONG, J., AUTEXIER, C. & CHEN, T. 2017. Zscan4 Inhibits Maintenance DNA Methylation to Facilitate Telomere Elongation in Mouse Embryonic Stem Cells. Cell Rep, 20, 1936-1949. DE LANGE, T. 2004. T-loops and the origin of telomeres. Nat. Rev. Mol. Cell Biol, 5, 323-329. DE LANGE, T. 2005. Shelterin: the protein complex that shapes and safeguards human telomeres. Genes Dev, 19, 2100-2110. DE LANGE, T., SHIUE, L., MYERS, R. M., COX, D. R., NAYLOR, S. L., KILLERY, A. M. & VARMUS, H. E. 1990. Structure and variability of human chromosome ends. Mol. Cell. Biol, 10, 518-527. DECIPHER 2018. Mapping the clinical genome - gene:ZNF827. Wellcome Sanger Institute. DEEG, K. I., CHUNG, I., BAUER, C. & RIPPE, K. 2016. Cancer Cells with Alternative Lengthening of Telomeres Do Not Display a General Hypersensitivity to ATR Inhibition. Front Oncol, 6, 186. DEHENNAUT, V., LOISON, I., DUBUISSEZ, M., NASSOUR, J., ABBADIE, C. & LEPRINCE, D. 2013. DNA double-strand breaks lead to activation of hypermethylated in cancer 1 (HIC1) by SUMOylation to regulate DNA repair. J Biol Chem, 288, 10254-64. DEJARDIN, J. & KINGSTON, R. E. 2009. Purification of proteins associated with specific genomic loci. Cell, 136, 175-186. DENCHI, E. L. & DE LANGE, T. 2007. Protection of telomeres through independent control of ATM and ATR by TRF2 and POT1. Nature, 448, 1068-1071. DENSLOW, S. A. & WADE, P. A. 2007. The human Mi-2/NuRD complex and gene regulation. Oncogene, 26, 5433-5438. DILLEY, R. L., VERMA, P., CHO, N. W., WINTERS, H. D., WONDISFORD, A. R. & GREENBERG, R. A. 2016. Break-induced telomere synthesis underlies alternative telomere maintenance. Nature. DIMITROVA, N., CHEN, Y. C., SPECTOR, D. L. & DE LANGE, T. 2008. 53BP1 promotes non-homologous end joining of telomeres by increasing chromatin mobility. Nature, 456, 524-528. DOBBIN, M. M., MADABHUSHI, R., PAN, L., CHEN, Y., KIM, D., GAO, J., AHANONU, B., PAO, P. C., QIU, Y., ZHAO, Y. & TSAI, L. H. 2013. SIRT1 collaborates with

202

ATM and HDAC1 to maintain genomic stability in neurons. Nat Neurosci, 16, 1008-15. DOKSANI, Y. & DE LANGE, T. 2016. Telomere-Internal Double-Strand Breaks Are Repaired by Homologous Recombination and PARP1/Lig3-Dependent End- Joining. Cell Rep, 17, 1646-1656. DOKSANI, Y., WU, J. Y., DE LANGE, T. & ZHUANG, X. 2013. Super-resolution fluorescence imaging of telomeres reveals TRF2-dependent T-loop formation. Cell, 155, 345-356. DRASKOVIC, I., ARNOULT, N., STEINER, V., BACCHETTI, S., LOMONTE, P. & LONDONO-VALLEJO, A. 2009. Probing PML body function in ALT cells reveals spatiotemporal requirements for telomere recombination. Proc. Natl. Acad. Sci. U. S. A, 106, 15726-15731. DUBUISSEZ, M., LOISON, I., PAGET, S., VORNG, H., AIT-YAHIA, S., ROHR, O., TSICOPOULOS, A. & LEPRINCE, D. 2016. Protein Kinase C-Mediated Phosphorylation of BCL11B at Serine 2 Negatively Regulates Its Interaction with NuRD Complexes during CD4+ T-Cell Activation. Mol Cell Biol, 36, 1881- 98. DUNHAM, M. A., NEUMANN, A. A., FASCHING, C. L. & REDDEL, R. R. 2000. Telomere maintenance by recombination in human cells. Nat. Genet, 26, 447- 450. EMERSON, R. O. & THOMAS, J. H. 2009. Adaptive evolution in zinc finger transcription factors. PLoS Genet, 5, e1000325. ENSEMBLCOMPARA 2018. Gene: ZNF827 ENSG00000151612 - Orthologues. October 2018 ed.: EMBL-EBI. EPISKOPOU, H., DRASKOVIC, I., VAN BENEDEN, A., TILMAN, G., MATTIUSSI, M., GOBIN, M., ARNOULT, N., LONDONO-VALLEJO, A. & DECOTTIGNIES, A. 2014. Alternative Lengthening of Telomeres is characterized by reduced compaction of telomeric chromatin. Nucleic Acids Res, 42, 4391-4405. ESSERS, J., THEIL, A. F., BALDEYRON, C., VAN CAPPELLEN, W. A., HOUTSMULLER, A. B., KANAAR, R. & VERMEULEN, W. 2005. Nuclear dynamics of PCNA in DNA replication and repair. Mol. Cell. Biol, 25, 9350-9359. EXPRESSION_ATLAS 2018. ensg00000151612 (ZNF827) Homo sapiens zinc finger protein 827. EMBL-EBI. FASCHING, C. L., BOWER, K. & REDDEL, R. R. 2005. Telomerase-independent telomere length maintenance in the absence of alternative lengthening of telomeres-associated promyelocytic leukemia bodies. Cancer Res, 65, 2722- 2729. FASCHING, C. L., NEUMANN, A. A., MUNTONI, A., YEAGER, T. R. & REDDEL, R. R. 2007. DNA damage induces alternative lengthening of telomeres (ALT) associated promyelocytic leukemia bodies that preferentially associate with linear telomeric DNA. Cancer Res, 67, 7072-7077. FEDOTOVA, A. A., BONCHUK, A. N., MOGILA, V. A. & GEORGIEV, P. G. 2017. C2H2 Zinc Finger Proteins: The Largest but Poorly Explored Family of Higher Eukaryotic Transcription Factors. Acta Naturae, 9, 47-58. FEENEY, G. P., ERRINGTON, R. J., WILTSHIRE, M., MARQUEZ, N., CHAPPELL, S. C. & SMITH, P. J. 2003. Tracking the cell cycle origins for escape from topotecan action by breast cancer cells. Br J Cancer, 88, 1310-7. FISET, S. & CHABOT, B. 2001. hnRNP A1 may interact simultaneously with telomeric DNA and the human telomerase RNA in vitro. Nucleic Acids Res, 29, 2268- 2275.

203

FLATTEN, K., DAI, N. T., VROMAN, B. T., LOEGERING, D., ERLICHMAN, C., KARNITZ, L. M. & KAUFMANN, S. H. 2005. The role of checkpoint kinase 1 in sensitivity to topoisomerase I poisons. J Biol Chem, 280, 14349-55. FLOTHO, A. & MELCHIOR, F. 2013. Sumoylation: a regulatory protein modification in health and disease. Annu Rev Biochem, 82, 357-85. FLYNN, R. L., CENTORE, R. C., O'SULLIVAN, R. J., RAI, R., TSE, A., SONGYANG, Z., CHANG, S., KARLSEDER, J. & ZOU, L. 2011. TERRA and hnRNPA1 orchestrate an RPA-to-POT1 switch on telomeric single-stranded DNA. Nature, 471, 532-536. FLYNN, R. L., COX, K. E., JEITANY, M., WAKIMOTO, H., BRYLL, A. R., GANEM, N. J., BERSANI, F., PINEDA, J. R., SUVA, M. L., BENES, C. H., HABER, D. A., BOUSSIN, F. D. & ZOU, L. 2015. Alternative lengthening of telomeres renders cancer cells hypersensitive to ATR inhibitors. Science, 347, 273-277. FLYNN, R. L. & ZOU, L. 2011. ATR: a master conductor of cellular responses to DNA replication stress. Trends Biochem Sci, 36, 133-40. FOKAS, E., PREVO, R., HAMMOND, E. M., BRUNNER, T. B., MCKENNA, W. G. & MUSCHEL, R. J. 2014. Targeting ATR in DNA damage response and cancer therapeutics. Cancer Treat Rev, 40, 109-17. GAILLARD, H., GARCIA-MUSE, T. & AGUILERA, A. 2015. Replication stress and cancer. Nat. Rev. Cancer, 15, 276-289. GAO, B., LU, Y., NIEUWEBOER, A. J. M., XU, H., BEESLEY, J., BOERE, I., DE GRAAN, A. M., DE BRUIJN, P., GURNEY, H., C, J. K., CHIEW, Y. E., JOHNATTY, S. E., BEALE, P., HARRISON, M., LUCCARINI, C., CONROY, D., MATHIJSSEN, R. H. J., P, R. H., BALLEINE, R. L., CHENEVIX-TRENCH, G., MACGREGOR, S. & DE FAZIO, A. 2018a. Genome-wide association study of paclitaxel and carboplatin disposition in women with epithelial ovarian cancer. Sci Rep, 8, 1508. GAO, C., DIMITROV, T., YONG, K. J., TATETSU, H., JEONG, H. W., LUO, H. R., BRADNER, J. E., TENEN, D. G. & CHAI, L. 2013. Targeting transcription factor SALL4 in acute myeloid leukemia by interrupting its interaction with an epigenetic complex. Blood, 121, 1413-21. GAO, X. D., TU, L. C., MIR, A., RODRIGUEZ, T., DING, Y., LESZYK, J., DEKKER, J., SHAFFER, S. A., ZHU, L. J., WOLFE, S. A. & SONTHEIMER, E. J. 2018b. C- BERST: defining subnuclear proteomic landscapes at genomic elements with dCas9-APEX2. Nat Methods, 15, 433-436. GARCIA-EXPOSITO, L., BOURNIQUE, E., BERGOGLIO, V., BOSE, A., BARROSO- GONZALEZ, J., ZHANG, S., RONCAIOLI, JUSTIN L., LEE, M., WALLACE, CALLEN T., WATKINS, SIMON C., OPRESKO, PATRICIA L., HOFFMANN, J.- S. & O’SULLIVAN, RODERICK J. 2016. Proteomic Profiling Reveals a Specific Role for Translesion DNA Polymerase η in the Alternative Lengthening of Telomeres. Cell Reports, 17, 1858-1871. GARNATZ, A. S., GAO, Z., BROMAN, M., MARTENS, S., EARLEY, J. U. & SVENSSON, E. C. 2014. FOG-2 mediated recruitment of the NuRD complex regulates cardiomyocyte proliferation during heart development. Dev Biol, 395, 50-61. GARTEL, A. L. & RADHAKRISHNAN, S. K. 2005. Lost in transcription: p21 repression, mechanisms, and consequences. Cancer Res, 65, 3980-5. GONG, Y. & DE LANGE, T. 2010. A Shld1-controlled POT1a provides support for repression of ATR signaling at telomeres through RPA exclusion. Mol. Cell, 40, 377-387.

204

GRIFFITH, J. D., COMEAU, L., ROSENFIELD, S., STANSEL, R. M., BIANCHI, A., MOSS, H. & DE LANGE, T. 1999. Mammalian telomeres end in a large duplex loop. Cell, 97, 503-514. GROBELNY, J. V., GODWIN, A. K. & BROCCOLI, D. 2000. ALT-associated PML bodies are present in viable cells and are enriched in cells in the G2/M phase of the cell cycle. J. Cell Sci, 113, 4577-4585. GUPTA, R., SOMYAJIT, K., NARITA, T., MASKEY, E., STANLIE, A., KREMER, M., TYPAS, D., LAMMERS, M., MAILAND, N., NUSSENZWEIG, A., LUKAS, J. & CHOUDHARY, C. 2018. DNA Repair Network Analysis Reveals Shieldin as a Key Regulator of NHEJ and PARP Inhibitor Sensitivity. Cell, 173, 972-988 e23. GWAS_CATALOG 2018. Gene: ZNF827. EMBL-EBI. HAKIN-SMITH, V., JELLINEK, D. A., LEVY, D., CARROLL, T., TEO, M., TIMPERLEY, W. R., MCKAY, M. J., REDDEL, R. R. & ROYDS, J. A. 2003. Alternative lengthening of telomeres and survival in patients with glioblastoma multiforme. Lancet, 361, 836-838. HARLEY, C. B. 1991. Telomere loss: mitotic clock or genetic time bomb? Mutat. Res, 256, 271-282. HARLEY, C. B., FUTCHER, A. B. & GREIDER, C. W. 1990. Telomeres shorten during ageing of human fibroblasts. Nature, 345, 458-460. HASTIE, N. D., DEMPSTER, M., DUNLOP, M. G., THOMPSON, A. M., GREEN, D. K. & ALLSHIRE, R. C. 1990. Telomere reduction in human colorectal carcinoma and with ageing. Nature, 346, 866-868. HAYASHI, M. T., CESARE, A. J., FITZPATRICK, J. A., LAZZERINI-DENCHI, E. & KARLSEDER, J. 2012. A telomere-dependent DNA damage checkpoint induced by prolonged mitotic arrest. Nat. Struct. Mol. Biol, 19, 387-394. HAYASHI, M. T., CESARE, A. J., RIVERA, T. & KARLSEDER, J. 2015. Cell death during crisis is mediated by mitotic telomere deprotection. Nature, 522, 492- 496. HEAPHY, C. M., DE WILDE, R. F., JIAO, Y., KLEIN, A. P., EDIL, B. H., SHI, C., BETTEGOWDA, C., RODRIGUEZ, F. J., EBERHART, C. G., HEBBAR, S., OFFERHAUS, G. J., MCLENDON, R., RASHEED, B. A., HE, Y., YAN, H., BIGNER, D. D., OBA-SHINJO, S. M., MARIE, S. K., RIGGINS, G. J., KINZLER, K. W., VOGELSTEIN, B., HRUBAN, R. H., MAITRA, A., PAPADOPOULOS, N. & MEEKER, A. K. 2011a. Altered telomeres in tumors with ATRX and DAXX mutations. Science, 333, 425. HEAPHY, C. M., SUBHAWONG, A. P., HONG, S. M., GOGGINS, M. G., MONTGOMERY, E. A., GABRIELSON, E., NETTO, G. J., EPSTEIN, J. I., LOTAN, T. L., WESTRA, W. H., SHIH, I. M., IACOBUZIO-DONAHUE, C. A., MAITRA, A., LI, Q. K., EBERHART, C. G., TAUBE, J. M., RAKHEJA, D., KURMAN, R. J., WU, T. C., RODEN, R. B., ARGANI, P., DE MARZO, A. M., TERRACCIANO, L., TORBENSON, M. & MEEKER, A. K. 2011b. Prevalence of the alternative lengthening of telomeres telomere maintenance mechanism in human cancer subtypes. Am. J. Pathol, 179, 1608-1615. HENDRIKS, I. A., D'SOUZA, R. C., CHANG, J. G., MANN, M. & VERTEGAAL, A. C. 2015. System-wide identification of wild-type SUMO-2 conjugation sites. Nat. Commun, 6, 7289. HENDRIKS, I. A., D'SOUZA, R. C., YANG, B., VERLAAN-DE, V. M., MANN, M. & VERTEGAAL, A. C. 2014. Uncovering global SUMOylation signaling networks in a site-specific manner. Nat. Struct. Mol. Biol, 21, 927-936.

205

HENDRIKS, I. A., LYON, D., YOUNG, C., JENSEN, L. J., VERTEGAAL, A. C. & NIELSEN, M. L. 2017. Site-specific mapping of the human SUMO proteome reveals co-modification with phosphorylation. Nat Struct Mol Biol, 24, 325-336. HENSON, J. D., CAO, Y., HUSCHTSCHA, L. I., CHANG, A. C., AU, A. Y., PICKETT, H. A. & REDDEL, R. R. 2009. DNA C-circles are specific and quantifiable markers of alternative-lengthening-of-telomeres activity. Nat. Biotechnol, 27, 1181-1185. HENSON, J. D., HANNAY, J. A., MCCARTHY, S. W., ROYDS, J. A., YEAGER, T. R., ROBINSON, R. A., WHARTON, S. B., JELLINEK, D. A., ARBUCKLE, S. M., YOO, J., ROBINSON, B. G., LEAROYD, D. L., STALLEY, P. D., BONAR, S. F., YU, D., POLLOCK, R. E. & REDDEL, R. R. 2005. A robust assay for alternative lengthening of telomeres in tumors shows the significance of alternative lengthening of telomeres in sarcomas and astrocytomas. Clin. Cancer Res, 11, 217-225. HENSON, J. D., NEUMANN, A. A., YEAGER, T. R. & REDDEL, R. R. 2002. Alternative lengthening of telomeres in mammalian cells. Oncogene, 21, 598-610. HENSON, J. D. & REDDEL, R. R. 2010. Assaying and investigating Alternative Lengthening of Telomeres activity in human cells and cancers. FEBS Lett., 584, 3800-3811. HICKEY, C. M., WILSON, N. R. & HOCHSTRASSER, M. 2012. Function and regulation of SUMO proteases. Nat. Rev. Mol. Cell Biol, 13, 755-766. HIROSE, T., APFEL, R., PFAHL, M. & JETTEN, A. M. 1995. The orphan receptor TAK1 acts as a repressor of RAR-, RXR- and T3R-mediated signaling pathways. Biochem Biophys Res Commun, 211, 83-91. HOFFMANN, T. J., PASSARELLI, M. N., GRAFF, R. E., EMAMI, N. C., SAKODA, L. C., JORGENSON, E., HABEL, L. A., SHAN, J., RANATUNGA, D. K., QUESENBERRY, C. P., CHAO, C. R., GHAI, N. R., AARONSON, D., PRESTI, J., NORDSTROM, T., WANG, Z., BERNDT, S. I., CHANOCK, S. J., MOSLEY, J. D., KLEIN, R. J., MIDDHA, M., LILJA, H., MELANDER, O., KVALE, M. N., KWOK, P. Y., SCHAEFER, C., RISCH, N., VAN DEN EEDEN, S. K. & WITTE, J. S. 2017. Genome-wide association study of prostate-specific antigen levels identifies novel loci independent of prostate cancer. Nat Commun, 8, 14248. HONG, W., NAKAZAWA, M., CHEN, Y. Y., KORI, R., VAKOC, C. R., RAKOWSKI, C. & BLOBEL, G. A. 2005. FOG-1 recruits the NuRD repressor complex to mediate transcriptional repression by GATA-1. EMBO J, 24, 2367-2378. HORN, S., FIGL, A., RACHAKONDA, P. S., FISCHER, C., SUCKER, A., GAST, A., KADEL, S., MOLL, I., NAGORE, E., HEMMINKI, K., SCHADENDORF, D. & KUMAR, R. 2013. TERT promoter mutations in familial and sporadic melanoma. Science, 339, 959-961. HOUGHTALING, B. R., CUTTONARO, L., CHANG, W. & SMITH, S. 2004. A dynamic molecular link between the telomere length regulator TRF1 and the chromosome end protector TRF2. Curr. Biol, 14, 1621-1631. HU, G. & WADE, P. A. 2012. NuRD and pluripotency: a complex balancing act. Cell Stem Cell, 10, 497-503. HU, J., HWANG, S. S., LIESA, M., GAN, B., SAHIN, E., JASKELIOFF, M., DING, Z., YING, H., BOUTIN, A. T., ZHANG, H., JOHNSON, S., IVANOVA, E., KOST- ALIMOVA, M., PROTOPOPOV, A., WANG, Y. A., SHIRIHAI, O. S., CHIN, L. & DEPINHO, R. A. 2012. Antitelomerase therapy provokes ALT and mitochondrial adaptive mechanisms in cancer. Cell, 148, 651-663.

206

HUANG, F. W., HODIS, E., XU, M. J., KRYUKOV, G. V., CHIN, L. & GARRAWAY, L. A. 2013. Highly recurrent TERT promoter mutations in human melanoma. Science, 339, 957-959. HUFFMAN, K. E., LEVENE, S. D., TESMER, V. M., SHAY, J. W. & WRIGHT, W. E. 2000. Telomere shortening is proportional to the size of the G-rich telomeric 3'- overhang. J. Biol. Chem, 275, 19719-19722. ILIAKIS, G., MURMANN, T. & SONI, A. 2015. Alternative end-joining repair pathways are the ultimate backup for abrogated classical non-homologous end-joining and homologous recombination repair: Implications for the formation of chromosome translocations. Mutat. Res. INUI, S., LEE, Y. F., HAAKE, A. R., GOLDSMITH, L. A. & CHANG, C. 1999. Induction of TR4 orphan receptor by retinoic acid in human HaCaT keratinocytes. J Invest Dermatol, 112, 426-31. IRA, G., PELLICIOLI, A., BALIJJA, A., WANG, X., FIORANI, S., CAROTENUTO, W., LIBERI, G., BRESSAN, D., WAN, L., HOLLINGSWORTH, N. M., HABER, J. E. & FOIANI, M. 2004. DNA end resection, homologous recombination and DNA damage checkpoint activation require CDK1. Nature, 431, 1011-1017. IUCHI, S. 2001. Three classes of C2H2 zinc finger proteins. Cell Mol Life Sci, 58, 625- 35. IZHAR, L., ADAMSON, B., CICCIA, A., LEWIS, J., PONTANO-VAITES, L., LENG, Y., LIANG, A. C., WESTBROOK, T. F., HARPER, J. W. & ELLEDGE, S. J. 2015. A Systematic Analysis of Factors Localized to Damaged Chromatin Reveals PARP-Dependent Recruitment of Transcription Factors. Cell Rep, 11, 1486- 500. JEN, J. & WANG, Y. C. 2016. Zinc finger proteins in cancer progression. J Biomed Sci, 23, 53. JENSEN, K., SHIELS, C. & FREEMONT, P. S. 2001. PML protein isoforms and the RBCC/TRIM motif. Oncogene, 20, 7223-7233. JIANG, S., YANG, W., QIU, Y., CHEN, H. Z. & ALZHEIMER'S DISEASE NEUROIMAGING, I. 2015. Identification of novel quantitative traits-associated susceptibility loci for APOE epsilon 4 non-carriers of Alzheimer's disease. Curr Alzheimer Res, 12, 218-27. JIANG, W. Q., ZHONG, Z. H., HENSON, J. D., NEUMANN, A. A., CHANG, A. C. & REDDEL, R. R. 2005. Suppression of alternative lengthening of telomeres by Sp100-mediated sequestration of MRE11/RAD50/NBS1 complex. Mol. Cell. Biol, 25, 2708-2721. JOLMA, A., YAN, J., WHITINGTON, T., TOIVONEN, J., NITTA, K. R., RASTAS, P., MORGUNOVA, E., ENGE, M., TAIPALE, M., WEI, G., PALIN, K., VAQUERIZAS, J. M., VINCENTELLI, R., LUSCOMBE, N. M., HUGHES, T. R., LEMAIRE, P., UKKONEN, E., KIVIOJA, T. & TAIPALE, J. 2013. DNA-binding specificities of human transcription factors. Cell, 152, 327-39. KANAI, M., AKIYAMA, M., TAKAHASHI, A., MATOBA, N., MOMOZAWA, Y., IKEDA, M., IWATA, N., IKEGAWA, S., HIRATA, M., MATSUDA, K., KUBO, M., OKADA, Y. & KAMATANI, Y. 2018. Genetic analysis of quantitative traits in the Japanese population links cell types to complex human diseases. Nat Genet, 50, 390-400. KARLSEDER, J., HOKE, K., MIRZOEVA, O. K., BAKKENIST, C., KASTAN, M. B., PETRINI, J. H. & DE LANGE, T. 2004. The telomeric protein TRF2 binds the ATM kinase and can inhibit the ATM-dependent DNA damage response. PLoS Biol, 2, E240.

207

KASTAN, M. B. & BARTEK, J. 2004. Cell-cycle checkpoints and cancer. Nature, 432, 316-23. KAUL, Z., CESARE, A. J., HUSCHTSCHA, L. I., NEUMANN, A. A. & REDDEL, R. R. 2012. Five dysfunctional telomeres predict onset of senescence in human cells. EMBO Rep, 13, 52-59. KEMP, J. P., MORRIS, J. A., MEDINA-GOMEZ, C., FORGETTA, V., WARRINGTON, N. M., YOULTEN, S. E., ZHENG, J., GREGSON, C. L., GRUNDBERG, E., TRAJANOSKA, K., LOGAN, J. G., POLLARD, A. S., SPARKES, P. C., GHIRARDELLO, E. J., ALLEN, R., LEITCH, V. D., BUTTERFIELD, N. C., KOMLA-EBRI, D., ADOUM, A. T., CURRY, K. F., WHITE, J. K., KUSSY, F., GREENLAW, K. M., XU, C., HARVEY, N. C., COOPER, C., ADAMS, D. J., GREENWOOD, C. M. T., MAURANO, M. T., KAPTOGE, S., RIVADENEIRA, F., TOBIAS, J. H., CROUCHER, P. I., ACKERT-BICKNELL, C. L., BASSETT, J. H. D., WILLIAMS, G. R., RICHARDS, J. B. & EVANS, D. M. 2017. Identification of 153 new loci associated with heel bone mineral density and functional involvement of GPC6 in osteoporosis. Nat Genet, 49, 1468-1475. KENNEDY, R. B., OVSYANNIKOVA, I. G., PANKRATZ, V. S., HARALAMBIEVA, I. H., VIERKANT, R. A. & POLAND, G. A. 2012. Genome-wide analysis of polymorphisms associated with cytokine responses in smallpox vaccine recipients. Hum Genet, 131, 1403-21. KIBE, T., ONO, Y., SATO, K. & UENO, M. 2007. Fission yeast Taz1 and RPA are synergistically required to prevent rapid telomere loss. Mol. Biol. Cell, 18, 2378- 2387. KIBE, T., OSAWA, G. A., KEEGAN, C. E. & DE LANGE, T. 2010. Telomere protection by TPP1 is mediated by POT1a and POT1b. Mol. Cell. Biol, 30, 1059-1066. KIM, S. K. 2018. Identification of 613 new loci associated with heel bone mineral density and a polygenic risk score for bone mineral density, osteoporosis and fracture. PLoS One, 13, e0200785. KIM, Y. J., GO, M. J., HU, C., HONG, C. B., KIM, Y. K., LEE, J. Y., HWANG, J. Y., OH, J. H., KIM, D. J., KIM, N. H., KIM, S., HONG, E. J., KIM, J. H., MIN, H., KIM, Y., ZHANG, R., JIA, W., OKADA, Y., TAKAHASHI, A., KUBO, M., TANAKA, T., KAMATANI, N., MATSUDA, K., CONSORTIUM, M., PARK, T., OH, B., KIMM, K., KANG, D., SHIN, C., CHO, N. H., KIM, H. L., HAN, B. G., LEE, J. Y. & CHO, Y. S. 2011. Large-scale genome-wide association studies in East Asians identify new genetic loci influencing metabolic traits. Nat Genet, 43, 990-5. KLIEWER, S. A., LEHMANN, J. M. & WILLSON, T. M. 1999. Orphan nuclear receptors: shifting endocrinology into reverse. Science, 284, 757-760. KLIEWER, S. A., UMESONO, K., HEYMAN, R. A., MANGELSDORF, D. J., DYCK, J. A. & EVANS, R. M. 1992. Retinoid X receptor-COUP-TF interactions modulate retinoic acid signaling. Proc Natl Acad Sci U S A, 89, 1448-52. KLOET, S. L., KAREMAKER, I. D., VAN VOORTHUIJSEN, L., LINDEBOOM, R. G. H., BALTISSEN, M. P., EDUPUGANTI, R. R., PORAMBA-LIYANAGE, D. W., JANSEN, P. & VERMEULEN, M. 2018. NuRD-interacting protein ZFP296 regulates genome-wide NuRD localization and differentiation of mouse embryonic stem cells. Nat Commun, 9, 4588. KRAICZY, J., NAYAK, K. M., HOWELL, K. J., ROSS, A., FORBESTER, J., SALVESTRINI, C., MUSTATA, R., PERKINS, S., ANDERSSON-ROLF, A., LEENEN, E., LIEBERT, A., VALLIER, L., ROSENSTIEL, P. C., STEGLE, O., DOUGAN, G., HEUSCHKEL, R., KOO, B. K. & ZILBAUER, M. 2019. DNA

208

methylation defines regional identity of human intestinal epithelial organoids and undergoes dynamic changes during development. Gut, 68, 49-61. KRUSE, S. W., SUINO-POWELL, K., ZHOU, X. E., KRETSCHMAN, J. E., REYNOLDS, R., VONRHEIN, C., XU, Y., WANG, L., TSAI, S. Y., TSAI, M. J. & XU, H. E. 2008. Identification of COUP-TFII orphan nuclear receptor as a retinoic acid-activated receptor. PLoS Biol, 6, e227. KWAN, A. H., CZOLIJ, R., MACKAY, J. P. & CROSSLEY, M. 2003. Pentaprobe: a comprehensive sequence for the one-step detection of DNA-binding activities. Nucleic Acids Res, 31, e124. LAI, A. Y. & WADE, P. A. 2011. Cancer biology and NuRD: a multifaceted chromatin remodelling complex. Nat. Rev. Cancer, 11, 588-596. LAITY, J. H., LEE, B. M. & WRIGHT, P. E. 2001. Zinc finger proteins: new insights into structural and functional diversity. Curr Opin Struct Biol, 11, 39-46. LAM, Y. C., AKHTER, S., GU, P., YE, J., POULET, A., GIRAUD-PANIS, M. J., BAILEY, S. M., GILSON, E., LEGERSKI, R. J. & CHANG, S. 2010. SNMIB/Apollo protects leading-strand telomeres against NHEJ-mediated repair. EMBO J, 29, 2230-2241. LANG, M., JEGOU, T., CHUNG, I., RICHTER, K., MUNCH, S., UDVARHELYI, A., CREMER, C., HEMMERICH, P., ENGELHARDT, J., HELL, S. W. & RIPPE, K. 2010. Three-dimensional organization of promyelocytic leukemia nuclear bodies. J. Cell Sci, 123, 392-400. LARSEN, A. 2010. Characterizing the organization within alternative lengthening of telomere associated-promyelocytic leukemia nuclear bodies. Master, University of Toronto, Canada. LAUBERTH, S. M. & RAUCHMAN, M. 2006. A conserved 12-amino acid motif in Sall1 recruits the nucleosome remodeling and deacetylase corepressor complex. J. Biol. Chem, 281, 23922-23931. LAW, C. W., CHEN, Y., SHI, W. & SMYTH, G. K. 2014. Voom: precision weights unlock linear model analysis tools for RNA-seq read counts. Genome Biol, 15, R29. LAW, M. J., LOWER, K. M., VOON, H. P., HUGHES, J. R., GARRICK, D., VIPRAKASIT, V., MITSON, M., DE GOBBI, M., MARRA, M., MORRIS, A., ABBOTT, A., WILDER, S. P., TAYLOR, S., SANTOS, G. M., CROSS, J., AYYUB, H., JONES, S., RAGOUSSIS, J., RHODES, D., DUNHAM, I., HIGGS, D. R. & GIBBONS, R. J. 2010. ATR-X syndrome protein targets tandem repeats and influences allele-specific expression in a size-dependent manner. Cell, 143, 367-378. LAZZERINI-DENCHI, E. & SFEIR, A. 2016. Stop pulling my strings - what telomeres taught us about the DNA damage response. Nat Rev Mol Cell Biol. LEE, J. H. & PAULL, T. T. 2007. Activation and regulation of ATM kinase activity in response to DNA double-strand breaks. Oncogene, 26, 7741-7748. LEE, M., HILLS, M., CONOMOS, D., STUTZ, M. D., DAGG, R. A., LAU, L. M., REDDEL, R. R. & PICKETT, H. A. 2014. Telomere extension by telomerase and ALT generates variant repeats by mechanistically distinct processes. Nucleic Acids Res, 42, 1733-1736. LEE, M. S., GIPPERT, G. P., SOMAN, K. V., CASE, D. A. & WRIGHT, P. E. 1989. Three-dimensional solution structure of a single zinc finger DNA-binding domain. Science, 245, 635-7.

209

LEE, Y. F., LEE, H. J. & CHANG, C. 2002. Recent advances in the TR2 and TR4 orphan receptors of the nuclear receptor superfamily. J Steroid Biochem Mol Biol, 81, 291-308. LEE, Y. F., YOUNG, W. J., BURBACH, J. P. & CHANG, C. 1998. Negative feedback control of the retinoid-retinoic acid/retinoid X receptor pathway by the human TR4 orphan receptor, a member of the steroid receptor superfamily. J Biol Chem, 273, 13437-43. LEJON, S., THONG, S. Y., MURTHY, A., ALQARNI, S., MURZINA, N. V., BLOBEL, G. A., LAUE, E. D. & MACKAY, J. P. 2011. Insights into association of the NuRD complex with FOG-1 from the crystal structure of an RbAp48.FOG-1 complex. J. Biol. Chem, 286, 1196-1203. LETSOLO, B. T., ROWSON, J. & BAIRD, D. M. 2010. Fusion of short telomeres in human cells is characterized by extensive deletion and microhomology, and can result in complex rearrangements. Nucleic Acids Res, 38, 1841-1852. LI, J. S., MIRALLES FUSTE, J., SIMAVORIAN, T., BARTOCCI, C., TSAI, J., KARLSEDER, J. & LAZZERINI DENCHI, E. 2017. TZAP: A telomere- associated protein involved in telomere length control. Science, 355, 638-641. LI, X. & HEYER, W. D. 2008. Homologous recombination in DNA repair and DNA damage tolerance. Cell Res, 18, 99-113. LIAO, Y., SMYTH, G. K. & SHI, W. 2013. The Subread aligner: fast, accurate and scalable read mapping by seed-and-vote. Nucleic Acids Res, 41, e108. LIEBER, M. R. 2008. The mechanism of human nonhomologous DNA end joining. J Biol Chem, 283, 1-5. LIM, C. P. & CAO, X. 2006. Structure, function, and regulation of STAT proteins. Mol Biosyst, 2, 536-50. LIN, S. J., YANG, D. R., LI, G. & CHANG, C. 2015. TR4 Nuclear Receptor Different Roles in Prostate Cancer Progression. Front Endocrinol (Lausanne), 6, 78. LIN, T. T., LETSOLO, B. T., JONES, R. E., ROWSON, J., PRATT, G., HEWAMANA, S., FEGAN, C., PEPPER, C. & BAIRD, D. M. 2010. Telomere dysfunction and fusion during the progression of chronic lymphocytic leukaemia: evidence for a telomere crisis. Blood, 116, 1899-1907. LIU, D., SAFARI, A., O'CONNOR, M. S., CHAN, D. W., LAEGELER, A., QIN, J. & SONGYANG, Z. 2004. PTOP interacts with POT1 and regulates its localization to telomeres. Nat. Cell Biol, 6, 673-680. LIU, P., KELLER, J. R., ORTIZ, M., TESSAROLLO, L., RACHEL, R. A., NAKAMURA, T., JENKINS, N. A. & COPELAND, N. G. 2003. Bcl11a is essential for normal lymphoid development. Nat Immunol, 4, 525-32. LIU, Q., GUNTUKU, S., CUI, X. S., MATSUOKA, S., CORTEZ, D., TAMAI, K., LUO, G., CARATTINI-RIVERA, S., DEMAYO, F., BRADLEY, A., DONEHOWER, L. A. & ELLEDGE, S. J. 2000. Chk1 is an essential kinase that is regulated by Atr and required for the G(2)/M DNA damage checkpoint. Genes Dev, 14, 1448- 1459. LIU, Z., LI, F., ZHANG, B., LI, S., WU, J. & SHI, Y. 2015. Structural basis of plant homeodomain finger 6 (PHF6) recognition by the retinoblastoma binding protein 4 (RBBP4) component of the nucleosome remodeling and deacetylase (NuRD) complex. J Biol Chem, 290, 6630-8. LOAYZA, D. & DE LANGE, T. 2003. POT1 as a terminal transducer of TRF1 telomere length control. Nature, 424, 1013-1018.

210

LONDONO-VALLEJO, J. A., DER-SARKISSIAN, H., CAZES, L., BACCHETTI, S. & REDDEL, R. R. 2004. Alternative lengthening of telomeres is characterized by high rates of telomeric exchange. Cancer Res, 64, 2324-2327. LOVEJOY, C. A., LI, W., REISENWEBER, S., THONGTHIP, S., BRUNO, J., DE LANGE, T., DE, S., PETRINI, J. H., SUNG, P. A., JASIN, M., ROSENBLUH, J., ZWANG, Y., WEIR, B. A., HATTON, C., IVANOVA, E., MACCONAILL, L., HANNA, M., HAHN, W. C., LUE, N. F., REDDEL, R. R., JIAO, Y., KINZLER, K., VOGELSTEIN, B., PAPADOPOULOS, N. & MEEKER, A. K. 2012. Loss of ATRX, genome instability, and an altered DNA damage response are hallmarks of the Alternative Lengthening of Telomeres pathway. PLoS Genet., 8, e1002772. LYDEARD, J. R., JAIN, S., YAMAGUCHI, M. & HABER, J. E. 2007. Break-induced replication and telomerase-independent telomere maintenance require Pol32. Nature, 448, 820-823. LYDEARD, J. R., LIPKIN-MOORE, Z., SHEU, Y. J., STILLMAN, B., BURGERS, P. M. & HABER, J. E. 2010. Break-induced replication requires all essential DNA replication factors except those specific for pre-RC assembly. Genes Dev, 24, 1133-44. MACIEJOWSKI, J. & DE LANGE, T. 2017. Telomeres in cancer: tumour suppression and genome instability. Nat Rev Mol Cell Biol, 18, 175-186. MACIEJOWSKI, J., LI, Y., BOSCO, N., CAMPBELL, P. J. & DE LANGE, T. 2015. Chromothripsis and Kataegis Induced by Telomere Crisis. Cell, 163, 1641-54. MAKAROV, V. L., HIROSE, Y. & LANGMORE, J. P. 1997. Long G tails at both ends of human chromosomes suggest a C strand degradation mechanism for telomere shortening. Cell, 88, 657-666. MARECHAL, A. & ZOU, L. 2013. DNA damage sensing by the ATM and ATR kinases. Cold Spring Harb Perspect Biol, 5. MARTINEZ, P., THANASOULA, M., MUNOZ, P., LIAO, C., TEJERA, A., MCNEES, C., FLORES, J. M., FERNANDEZ-CAPETILLO, O., TARSOUNAS, M. & BLASCO, M. A. 2009. Increased telomere fragility and fusions resulting from TRF1 deficiency lead to degenerative pathologies and increased cancer in mice. Genes Dev, 23, 2060-2075. MARZEC, P., ARMENISE, C., PEROT, G., ROUMELIOTI, F. M., BASYUK, E., GAGOS, S., CHIBON, F. & DEJARDIN, J. 2015. Nuclear-receptor-mediated telomere insertion leads to genome instabiltiy in ALT cancers. Cell, 160, 913- 927. MATEOS-GOMEZ, P. A., GONG, F., NAIR, N., MILLER, K. M., LAZZERINI-DENCHI, E. & SFEIR, A. 2015. Mammalian polymerase θ promotes alternative NHEJ and suppresses recombination. Nature, 518, 254-257. MCEACHERN, M. J. & HABER, J. E. 2006. Break-induced replication and recombinational telomere elongation in yeast. Annu. Rev. Biochem, 75, 111- 135. MCGUE, M., ZHANG, Y., MILLER, M. B., BASU, S., VRIEZE, S., HICKS, B., MALONE, S., OETTING, W. S. & IACONO, W. G. 2013. A genome-wide association study of behavioral disinhibition. Behav Genet, 43, 363-73. MEEKER, A. K., GAGE, W. R., HICKS, J. L., SIMON, I., COFFMAN, J. R., PLATZ, E. A., MARCH, G. E. & DE MARZO, A. M. 2002. Telomere length assessment in human archival tissues: combined telomere fluorescence in situ hybridization and immunostaining. Am. J. Pathol, 160, 1259-1268.

211

MENDER, I., GRYAZNOV, S. & SHAY, J. W. 2015. A novel telomerase substrate precursor rapidly induces telomere dysfunction in telomerase positive cancer cells but not telomerase silent normal cells. Oncoscience, 2, 693-695. MENDEZ-BERMUDEZ, A., HIDALGO-BRAVO, A., COTTON, V. E., GRAVANI, A., JEYAPALAN, J. N. & ROYLE, N. J. 2012. The roles of WRN and BLM RecQ helicases in the Alternative Lengthening of Telomeres. Nucleic Acids Res, 40, 10809-10820. MEYNE, J., RATLIFF, R. L. & MOYZIS, R. K. 1989. Conservation of the human telomere sequence (TTAGGG)n among vertebrates. Proc. Natl. Acad. Sci. U. S. A, 86, 7049-7053. MICHAILIDOU, K., LINDSTROM, S., DENNIS, J., BEESLEY, J., HUI, S., KAR, S., LEMACON, A., SOUCY, P., GLUBB, D., ROSTAMIANFAR, A., BOLLA, M. K., WANG, Q., TYRER, J., DICKS, E., LEE, A., WANG, Z., ALLEN, J., KEEMAN, R., EILBER, U., FRENCH, J. D., QING CHEN, X., FACHAL, L., MCCUE, K., MCCART REED, A. E., GHOUSSAINI, M., CARROLL, J. S., JIANG, X., FINUCANE, H., ADAMS, M., ADANK, M. A., AHSAN, H., AITTOMAKI, K., ANTON-CULVER, H., ANTONENKOVA, N. N., ARNDT, V., ARONSON, K. J., ARUN, B., AUER, P. L., BACOT, F., BARRDAHL, M., BAYNES, C., BECKMANN, M. W., BEHRENS, S., BENITEZ, J., BERMISHEVA, M., BERNSTEIN, L., BLOMQVIST, C., BOGDANOVA, N. V., BOJESEN, S. E., BONANNI, B., BORRESEN-DALE, A. L., BRAND, J. S., BRAUCH, H., BRENNAN, P., BRENNER, H., BRINTON, L., BROBERG, P., BROCK, I. W., BROEKS, A., BROOKS-WILSON, A., BRUCKER, S. Y., BRUNING, T., BURWINKEL, B., BUTTERBACH, K., CAI, Q., CAI, H., CALDES, T., CANZIAN, F., CARRACEDO, A., CARTER, B. D., CASTELAO, J. E., CHAN, T. L., DAVID CHENG, T. Y., SENG CHIA, K., CHOI, J. Y., CHRISTIANSEN, H., CLARKE, C. L., COLLABORATORS, N., COLLEE, M., CONROY, D. M., CORDINA- DUVERGER, E., CORNELISSEN, S., COX, D. G., COX, A., CROSS, S. S., CUNNINGHAM, J. M., CZENE, K., DALY, M. B., DEVILEE, P., DOHENY, K. F., DORK, T., DOS-SANTOS-SILVA, I., DUMONT, M., DURCAN, L., DWEK, M., ECCLES, D. M., EKICI, A. B., ELIASSEN, A. H., ELLBERG, C., ELVIRA, M., et al. 2017. Association analysis identifies 65 new breast cancer risk loci. Nature, 551, 92-94. MILLER, J., MCLACHLAN, A. D. & KLUG, A. 1985. Repetitive zinc-binding domains in the protein transcription factor IIIA from Xenopus oocytes. EMBO J, 4, 1609- 14. MIN, J., WRIGHT, W. E. & SHAY, J. W. 2017. Alternative Lengthening of Telomeres Mediated by Mitotic DNA Synthesis Engages Break-Induced Replication Processes. Mol Cell Biol, 37. MOODY, R. R., LO, M. C., MEAGHER, J. L., LIN, C. C., STEVERS, N. O., TINSLEY, S. L., JUNG, I., MATVEKAS, A., STUCKEY, J. A. & SUN, D. 2018. Probing the interaction between the histone methyltransferase/deacetylase subunit RBBP4/7 and the transcription factor BCL11A in epigenetic complexes. J Biol Chem, 293, 2125-2136. MORIN, G. B. 1989. The human telomere terminal transferase enzyme is a ribonucleoprotein that synthesizes TTAGGG repeats. Cell, 59, 521-529. MUNTONI, A., NEUMANN, A. A., HILLS, M. & REDDEL, R. R. 2009. Telomere elongation involves intra-molecular DNA replication in cells utilizing Alternative Lengthening of Telomeres. Hum. Mol. Genet, 18, 1017-1027.

212

MURNANE, J. P., SABATIER, L., MARDER, B. A. & MORGAN, W. F. 1994. Telomere dynamics in an immortal human cell line. EMBO J, 13, 4953-4962. NABETANI, A., YOKOYAMA, O. & ISHIKAWA, F. 2004. Localization of hRad9, hHus1, hRad1 and hRad17, and caffeine-sensitive DNA replication at the alternative lengthening of telomeres-associated promyelocytic leukemia body. J. Biol. Chem, 279, 25849-25857. NAPIER, C. E., HUSCHTSCHA, L. I., HARVEY, A., BOWER, K., NOBLE, J. R., HENDRICKSON, E. A. & REDDEL, R. R. 2015. ATRX represses alternative lengthening of telomeres. Oncotarget, 6, 16543-16558. NATARAJAN, S., GROFF-VINDMAN, C. & MCEACHERN, M. J. 2003. Factors influencing the recombinational expansion and spread of telomeric tandem arrays in Kluyveromyces lactis. Eukaryot. Cell, 2, 1115-1127. NGONDO, R. P. & CARBON, P. 2014. Transcription factor abundance controlled by an auto-regulatory mechanism involving a transcription start site switch. Nucleic Acids Res, 42, 2171-84. NICULESCU, A. B. R., CHEN, X., SMEETS, M., HENGST, L., PRIVES, C. & REED, S. I. 1998. Effects of p21Cip1/Waf1 at both the G1/S and the G2/M cell cycle transitions: pRb is a critical determinant in blocking DNA replication and in preventing endoreduplication. Mol. Cell. Biol, 18, 629-643. NIMONKAR, A. V., GENSCHEL, J., KINOSHITA, E., POLACZEK, P., CAMPBELL, J. L., WYMAN, C., MODRICH, P. & KOWALCZYKOWSKI, S. C. 2011. BLM- DNA2-RPA-MRN and EXO1-BLM-RPA-MRN constitute two DNA end resection machineries for human DNA break repair. Genes Dev, 25, 350-62. NISHIKAWA, Y., MATSUZAKI, Y., KIMURA, K., ROKUNOHE, A., NAKANO, H. & SAWAMURA, D. 2018. Modulation of Stimulator of Interferon Genes (STING) Expression by Interferon-gamma in Human Keratinocytes. Biochem Genet, 56, 93-102. O'SULLIVAN, R. J. & ALMOUZNI, G. 2014. Assembly of telomeric chromatin to create ALTernative endings. Trends Cell Biol, 24, 675-685. O'SULLIVAN, R. J., ARNOULT, N., LACKNER, D. H., OGANESIAN, L., HAGGBLOM, C., CORPET, A., ALMOUZNI, G. & KARLSEDER, J. 2014. Rapid induction of alternative lengthening of telomeres by depletion of the histone chaperone ASF1. Nat. Struct. Mol. Biol, 21, 167-174. OGRYZKO, V. V., WONG, P. & HOWARD, B. H. 1997. WAF1 retards S-phase progression primarily by inhibition of cyclin-dependent kinases. Mol. Cell. Biol, 17, 4877-4882. OHNESEIT, P. A., PRAGER, D., KEHLBACH, R. & RODEMANN, H. P. 2005. Cell cycle effects of topotecan alone and in combination with irradiation. Radiother Oncol, 75, 237-45. OKAMOTO, K., BARTOCCI, C., OUZOUNOV, I., DIEDRICH, J. K., YATES, J. R., III & DENCHI, E. L. 2013. A two-step mechanism for TRF2-mediated chromosome-end protection. Nature, 494, 502-505. OLOVNIKOV, A. M. 1973. A theory of marginotomy. The incomplete copying of template margin in enzymic synthesis of polynucleotides and biological significance of the phenomenon. J. Theor. Biol, 41, 181-190. OSTERWALD, S., DEEG, K. I., CHUNG, I., PARISOTTO, D., WORZ, S., ROHR, K., ERFLE, H. & RIPPE, K. 2015. PML induces compaction, TRF2 depletion and DNA damage signaling at telomeres and promotes their alternative lengthening. J. Cell Sci, 128, 1887-1900.

213

PALACIOS, J. A., HERRANZ, D., DE BONIS, M. L., VELASCO, S., SERRANO, M. & BLASCO, M. A. 2010. SIRT1 contributes to telomere maintenance and augments global homologous recombination. J. Cell Biol, 191, 1299-1313. PAN, X., DROSOPOULOS, W. C., SETHI, L., MADIREDDY, A., SCHILDKRAUT, C. L. & ZHANG, D. 2017. FANCM, BRCA1, and BLM cooperatively resolve the replication stress at the ALT telomeres. Proc Natl Acad Sci U S A, 114, E5940- E5949. PARRAGA, G., HORVATH, S. J., EISEN, A., TAYLOR, W. E., HOOD, L., YOUNG, E. T. & KLEVIT, R. E. 1988. Zinc-dependent structure of a single-finger domain of yeast ADR1. Science, 241, 1489-92. PAVLETICH, N. P. & PABO, C. O. 1991. Zinc finger-DNA recognition: crystal structure of a Zif268-DNA complex at 2.1 A. Science, 252, 809-17. PEIFER, M., HERTWIG, F., ROELS, F., DREIDAX, D., GARTLGRUBER, M., MENON, R., KRAMER, A., RONCAIOLI, J. L., SAND, F., HEUCKMANN, J. M., IKRAM, F., SCHMIDT, R., ACKERMANN, S., ENGESSER, A., KAHLERT, Y., VOGEL, W., ALTMULLER, J., NURNBERG, P., THIERRY-MIEG, J., THIERRY-MIEG, D., MARIAPPAN, A., HEYNCK, S., MARIOTTI, E., HENRICH, K. O., GLOCKNER, C., BOSCO, G., LEUSCHNER, I., SCHWEIGER, M. R., SAVELYEVA, L., WATKINS, S. C., SHAO, C., BELL, E., HOFER, T., ACHTER, V., LANG, U., THEISSEN, J., VOLLAND, R., SAADATI, M., EGGERT, A., DE WILDE, B., BERTHOLD, F., PENG, Z., ZHAO, C., SHI, L., ORTMANN, M., BUTTNER, R., PERNER, S., HERO, B., SCHRAMM, A., SCHULTE, J. H., HERRMANN, C., O'SULLIVAN, R. J., WESTERMANN, F., THOMAS, R. K. & FISCHER, M. 2015. Telomerase activation by genomic rearrangements in high- risk neuroblastoma. Nature. PEREIRA, F. A., TSAI, M. J. & TSAI, S. Y. 2000. COUP-TF orphan nuclear receptors in development and differentiation. Cell Mol Life Sci, 57, 1388-98. PERREM, K., COLGIN, L. M., NEUMANN, A. A., YEAGER, T. R. & REDDEL, R. R. 2001. Coexistence of alternative lengthening of telomeres and telomerase in hTERT-transfected GM847 cells. Mol. Cell. Biol, 21, 3862-3875. PICKETT, H. A., CESARE, A. J., JOHNSTONE, R. L., NEUMANN, A. A. & REDDEL, R. R. 2009. Control of telomere length by a trimming mechanism that involves generation of t-circles. EMBO J, 28, 799-809. PICKETT, H. A. & REDDEL, R. R. 2015. Molecular mechanisms of activity and derepression of alternative lengthening of telomeres. Nat Struct Mol Biol, 22, 875-80. POLO, S. E., KAIDI, A., BASKCOMB, L., GALANTY, Y. & JACKSON, S. P. 2010. Regulation of DNA-damage responses and cell-cycle progression by the chromatin remodelling factor CHD4. EMBO J, 29, 3130-3139. POTTS, P. R. & YU, H. 2007. The SMC5/6 complex maintains telomere length in ALT cancer cells through SUMOylation of telomere-binding proteins. Nat. Struct. Mol. Biol, 14, 581-590. RADHAKRISHNAN, S. K., FELICIANO, C. S., NAJMABADI, F., HAEGEBARTH, A., KANDEL, E. S., TYNER, A. L. & GARTEL, A. L. 2004. Constitutive expression of E2F-1 leads to p21-dependent cell cycle arrest in S phase of the cell cycle. Oncogene, 23, 4173-4176. RAI, R., ZHENG, H., HE, H., LUO, Y., MULTANI, A., CARPENTER, P. B. & CHANG, S. 2010. The function of classical and alternative non-homologous end-joining pathways in the fusion of dysfunctional telomeres. EMBO J, 29, 2598-2610.

214

RIBES-ZAMORA, A., INDIVIGLIO, S. M., MIHALEK, I., WILLIAMS, C. L. & BERTUCH, A. A. 2013. TRF2 interaction with Ku heterotetramerization interface gives insight into c-NHEJ prevention at human telomeres. Cell Rep, 5, 194-206. RIVERA, T., HAGGBLOM, C., COSCONATI, S. & KARLSEDER, J. 2017. A balance between elongation and trimming regulates telomere stability in stem cells. Nat Struct Mol Biol, 24, 30-39. ROOT, H., LARSEN, A., KOMOSA, M., AL-AZRI, F., LI, R., BAZETT-JONES, D. P. & STEPHEN MEYN, M. 2016. FANCD2 limits BLM-dependent telomere instability in the alternative lengthening of telomeres pathway. Hum Mol Genet, 25, 3255- 3268. ROUMELIOTI, F. M., SOTIRIOU, S. K., KATSINI, V., CHIOUREA, M., HALAZONETIS, T. D. & GAGOS, S. 2016. Alternative lengthening of human telomeres is a conservative DNA replication process with features of break-induced replication. EMBO Rep, 17, 1731-1737. ROUX, K. J., KIM, D. I. & BURKE, B. 2013. BioID: a screen for protein-protein interactions. Curr Protoc Protein Sci, 74, Unit 19 23. RUDEN, M. & PURI, N. 2013. Novel anticancer therapeutics targeting telomerase. Cancer Treat. Rev, 39, 444-456. SAHU, K. S. 2017. Deciphering gene regulatory circuitry governing cell fate changes. Doctor of Natural Sciences, Johannes Gutenberg University, Mainz. SALE, J. E. 2013. Translesion DNA synthesis and mutagenesis in eukaryotes. Cold Spring Harb Perspect Biol, 5, a012708. SANKARAN, V. G., XU, J., RAGOCZY, T., IPPOLITO, G. C., WALKLEY, C. R., MAIKA, S. D., FUJIWARA, Y., ITO, M., GROUDINE, M., BENDER, M. A., TUCKER, P. W. & ORKIN, S. H. 2009. Developmental and species-divergent globin switching are driven by BCL11A. Nature, 460, 1093-7. SAREK, G., VANNIER, J. B., PANIER, S., PETRINI, J. H. & BOULTON, S. J. 2015. TRF2 recruits RTEL1 to telomeres in S phase to promote T-loop unwinding. Mol. Cell, 57, 1-14. SARKAR, J., WAN, B., YIN, J., VALLABHANENI, H., HORVATH, K., KULIKOWICZ, T., BOHR, V. A., ZHANG, Y., LEI, M. & LIU, Y. 2015. SLX4 contributes to telomere preservation and regulated processing of telomeric joint molecule intermediates. Nucleic Acids Res, 43, 5912-5923. SCHEEL, C., SCHAEFER, K. L., JAUCH, A., KELLER, M., WAI, D., BRINKSCHMIDT, C., VAN VALEN, F., BOECKER, W., DOCKHORN-DWORNICZAK, B. & POREMBA, C. 2001. Alternative lengthening of telomeres is associated with chromosomal instability in osteosarcomas. Oncogene, 20, 3835-3844. SCHMITGES, F. W., PRUSTY, A. B., FATY, M., STUTZER, A., LINGARAJU, G. M., AIWAZIAN, J., SACK, R., HESS, D., LI, L., ZHOU, S., BUNKER, R. D., WIRTH, U., BOUWMEESTER, T., BAUER, A., LY-HARTIG, N., ZHAO, K., CHAN, H., GU, J., GUT, H., FISCHLE, W., MULLER, J. & THOMA, N. H. 2011. Histone methylation by PRC2 is inhibited by active chromatin marks. Mol Cell, 42, 330- 41. SCHOEFTNER, S. & BLASCO, M. A. 2009. A 'higher order' of telomere regulation: telomere heterochromatin and telomeric RNAs. EMBO J, 28, 2323-2336. SCHRANK, Z., KHAN, N., OSUDE, C., SINGH, S., MILLER, R. J., MERRICK, C., MABEL, A., KUCKOVIC, A. & PURI, N. 2018. Oligonucleotides Targeting Telomeres and Telomerase in Cancer. Molecules, 23. SFEIR, A. & DE LANGE, T. 2012. Removal of shelterin reveals the telomere end- protection problem. Science, 336, 593-597.

215

SFEIR, A., KABIR, S., VAN OVERBEEK, M., CELLI, G. B. & DE LANGE, T. 2010. Loss of Rap1 induces telomere recombination in the absence of NHEJ or a DNA damage signal. Science, 327, 1657-1661. SFEIR, A., KOSIYATRAKUL, S. T., HOCKEMEYER, D., MACRAE, S. L., KARLSEDER, J., SCHILDKRAUT, C. L. & DE LANGE, T. 2009. Mammalian telomeres resemble fragile sites and require TRF1 for efficient replication. Cell, 138, 90-103. SHAFFER, J. R., FEINGOLD, E., WANG, X., LEE, M., TCUENCO, K., WEEKS, D. E., WEYANT, R. J., CROUT, R., MCNEIL, D. W. & MARAZITA, M. L. 2013. GWAS of dental caries patterns in the permanent dentition. J Dent Res, 92, 38-44. SHAY, J. W. & BACCHETTI, S. 1997. A survey of telomerase activity in human cancer. Eur. J. Cancer, 33, 787-791. SHAY, J. W., PEREIRA-SMITH, O. M. & WRIGHT, W. E. 1991. A role for both RB and p53 in the regulation of human cellular senescence. Exp. Cell Res, 196, 33-39. SHILOH, Y. 2003. ATM and related protein kinases: safeguarding genome integrity. Nat. Rev. Cancer, 3, 155-168. SHRIVASTAV, M., DE HARO, L. P. & NICKOLOFF, J. A. 2008. Regulation of DNA double-strand break repair pathway choice. Cell Res, 18, 134-147. SIMPSON, K., JONES, R. E., GRIMSTEAD, J. W., HILLS, R., PEPPER, C. & BAIRD, D. M. 2015. Telomere fusion threshold identifies a poor prognostic subset of breast cancer patients. Mol. Oncol, 9, 1186-1193. SIO, C. A., JUNG, K., KIM, J. H., CHEONG, H. S., SHIN, E., JANG, H., YOON, M., JANG, H. & SHIN, H. D. 2017. Sotos syndrome associated with Hirschsprung's disease: a new case and exome-sequencing analysis. Pediatr Res, 82, 87-92. SMEENK, G., WIEGANT, W. W., VROLIJK, H., SOLARI, A. P., PASTINK, A. & VAN ATTIKUM, H. 2010. The NuRD chromatin-remodeling complex regulates signaling and repair of DNA damage. J. Cell Biol, 190, 741-749. SMOGORZEWSKA, A. & DE LANGE, T. 2002. Different telomere damage signaling pathways in human and mouse cells. EMBO J, 21, 4338-4348. SMYTH, G. K. 2005. Limma: Linear models for microarray data. In: GENTLEMAN, R., CAREY, V., DUDOIT, S., IRIZARRY, R. & HUBER, W. (eds.) Bioinofrmatics and Computational Biology Solutions using R and Bioconductor. New York: Springer. SOBINOFF, A. P., ALLEN, J. A., NEUMANN, A. A., YANG, S. F., WALSH, M. E., HENSON, J. D., REDDEL, R. R. & PICKETT, H. A. 2017. BLM and SLX4 play opposing roles in recombination-dependent replication at human telomeres. EMBO J, 36, 2907-2919. SOBINOFF, A. P. & PICKETT, H. A. 2017. Alternative Lengthening of Telomeres: DNA Repair Pathways Converge. Trends Genet, 33, 921-932. SORENSEN, C. S., HANSEN, L. T., DZIEGIELEWSKI, J., SYLJUASEN, R. G., LUNDIN, C., BARTEK, J. & HELLEDAY, T. 2005a. The cell-cycle checkpoint kinase Chk1 is required for mammalian homologous recombination repair. Nat Cell Biol, 7, 195-201. SORENSEN, C. S., HANSEN, L. T., DZIEGIELEWSKI, J., SYLJUASEN, R. G., LUNDIN, C., BARTEK, J. & HELLEDAY, T. 2005b. The cell-cycle checkpoint kinase Chk1 is required for mammalian homologous recombination repair. Nat. Cell Biol, 7, 195-201. SPRUIJT, C. G., LUIJSTERBURG, M. S., MENAFRA, R., LINDEBOOM, R. G., JANSEN, P. W., EDUPUGANTI, R. R., BALTISSEN, M. P., WIEGANT, W. W., VOELKER-ALBERT, M. C., MATARESE, F., MENSINGA, A., POSER, I., VOS,

216

H. R., STUNNENBERG, H. G., VAN ATTIKUM, H. & VERMEULEN, M. 2016. ZMYND8 Co-localizes with NuRD on Target Genes and Regulates Poly(ADP- Ribose)-Dependent Recruitment of GATAD2A/NuRD to Sites of DNA Damage. Cell Rep, 17, 783-798. STEWART, J. A., WANG, F., CHAIKEN, M. F., KASBEK, C., CHASTAIN, P. D., WRIGHT, W. E. & PRICE, C. M. 2012. Human CST promotes telomere duplex replication and general replication restart after fork stalling. EMBO J, 31, 3537- 3549. STRZALKA, W. & ZIEMIENOWICZ, A. 2011. Proliferating cell nuclear antigen (PCNA): a key factor in DNA replication and cell cycle regulation. Ann Bot, 107, 1127- 40. STURZENEGGER, A., BURDOVA, K., KANAGARAJ, R., LEVIKOVA, M., PINTO, C., CEJKA, P. & JANSCAK, P. 2014. DNA2 cooperates with the WRN and BLM RecQ helicases to mediate long-range DNA end resection in human cells. J Biol Chem, 289, 27314-26. SUNDQUIST, W. I. & KLUG, A. 1989. Telomeric DNA dimerizes by formation of guanine tetrads between hairpin loops. Nature, 342, 825-829. SUNG, P. & KLEIN, H. 2006. Mechanism of homologous recombination: mediators and helicases take on regulatory functions. Nat. Rev. Mol. Cell Biol, 7, 739-750. SYLJUASEN, R. G., SORENSEN, C. S., HANSEN, L. T., FUGGER, K., LUNDIN, C., JOHANSSON, F., HELLEDAY, T., SEHESTED, M., LUKAS, J. & BARTEK, J. 2005. Inhibition of human Chk1 causes increased initiation of DNA replication, phosphorylation of ATR targets, and DNA breakage. Mol. Cell. Biol, 25, 3553- 3562. TADEPALLY, H. D., BURGER, G. & AUBRY, M. 2008. Evolution of C2H2-zinc finger genes and subfamilies in mammals: species-specific duplication and loss of clusters, genes and effector domains. BMC Evol Biol, 8, 176. TAKAI, K. K., KIBE, T., DONIGIAN, J. R., FRESCAS, D. & DE LANGE, T. 2011. Telomere protection by TPP1/POT1 requires tethering to TIN2. Mol. Cell, 44, 647-659. TATETSU, H., KONG, N. R., CHONG, G., AMABILE, G., TENEN, D. G. & CHAI, L. 2016. SALL4, the missing link between stem cells, development and cancer. Gene, 584, 111-9. TEMIME-SMAALI, N., GUITTAT, L., WENNER, T., BAYART, E., DOUARRE, C., GOMEZ, D., GIRAUD-PANIS, M. J., LONDONO-VALLEJO, A., GILSON, E., AMOR-GUERET, M. & RIOU, J. F. 2008. Topoisomerase IIIα is required for normal proliferation and telomere stability in alternative lengthening of telomeres. EMBO J, 27, 1513-1524. THE_GENOTYPE-TISSUE_EXPRESSION_PROJECT 2017. Gene expression for ZNF827 (ENSG00000151612.11). 2017 ed.: Broad Institute. THOMAS, A., REDON, C. E., SCIUTO, L., PADIERNOS, E., JI, J., LEE, M. J., YUNO, A., LEE, S., ZHANG, Y., TRAN, L., YUTZY, W., RAJAN, A., GUHA, U., CHEN, H., HASSAN, R., ALEWINE, C. C., SZABO, E., BATES, S. E., KINDERS, R. J., STEINBERG, S. M., DOROSHOW, J. H., ALADJEM, M. I., TREPEL, J. B. & POMMIER, Y. 2018. Phase I Study of ATR Inhibitor M6620 in Combination With Topotecan in Patients With Advanced Solid Tumors. J Clin Oncol, 36, 1594- 1602. TODD, M. A. & PICKETTS, D. J. 2012. PHF6 interacts with the nucleosome remodeling and deacetylation (NuRD) complex. J Proteome Res, 11, 4326-37.

217

TOMMERUP, H., DOUSMANIS, A. & DE LANGE, T. 1994. Unusual chromatin in human telomeres. Mol. Cell. Biol, 14, 5777-5785. TONG, J. K., HASSIG, C. A., SCHNITZLER, G. R., KINGSTON, R. E. & SCHREIBER, S. L. 1998. Chromatin deacetylation by an ATP-dependent nucleosome remodelling complex. Nature, 395, 917-21. TOPARK-NGARM, A., GOLONZHKA, O., PETERSON, V. J., BARRETT, B., JR., MARTINEZ, B., CROFOOT, K., FILTZ, T. M. & LEID, M. 2006. CTIP2 associates with the NuRD complex on the promoter of p57KIP2, a newly identified CTIP2 target gene. J. Biol. Chem, 281, 32272-32283. TRAN, P., ZHANG, X. K., SALBERT, G., HERMANN, T., LEHMANN, J. M. & PFAHL, M. 1992. COUP orphan receptors are negative regulators of retinoic acid response pathways. Mol Cell Biol, 12, 4666-76. TURNBULL, W. B. & DARANAS, A. H. 2003. On the value of c: can low affinity systems be studied by isothermal titration calorimetry? J Am Chem Soc, 125, 14859-66. ULRICH, H. D. & WALDEN, H. 2010. Ubiquitin signalling in DNA replication and repair. Nat. Rev. Mol. Cell Biol, 11, 479-489. URINGA, E. J., LISAINGO, K., PICKETT, H. A., BRIND'AMOUR, J., ROHDE, J. H., ZELENSKY, A., ESSERS, J. & LANSDORP, P. M. 2012. RTEL1 contributes to DNA replication, repair and telomere maintenance. Mol. Biol. Cell, 23, 2782- 2792. VALENTIJN, L. J., KOSTER, J., ZWIJNENBURG, D. A., HASSELT, N. E., VAN SLUIS, P., VOLCKMANN, R., VAN NOESEL, M. M., GEORGE, R. E., TYTGAT, G. A., MOLENAAR, J. J. & VERSTEEG, R. 2015. TERT rearrangements are frequent in neuroblastoma and identify aggressive tumors. Nat Genet, 47, 1411-4. VAN DER HARST, P. & VERWEIJ, N. 2018. Identification of 64 Novel Genetic Loci Provides an Expanded View on the Genetic Architecture of Coronary Artery Disease. Circ Res, 122, 433-443. VAN LY, D., LOW, R. R. J., FROLICH, S., BARTOLEC, T. K., KAFER, G. R., PICKETT, H. A., GAUS, K. & CESARE, A. J. 2018. Telomere Loop Dynamics in Chromosome End Protection. Mol Cell, 71, 510-525 e6. VANNIER, J. B., SANDHU, S., PETALCORIN, M. I., WU, X., NABI, Z., DING, H. & BOULTON, S. J. 2013. RTEL1 is a replisome-associated helicase that promotes telomere and genome-wide replication. Science, 342, 239-242. VAQUERIZAS, J. M., KUMMERFELD, S. K., TEICHMANN, S. A. & LUSCOMBE, N. M. 2009. A census of human transcription factors: function, expression and evolution. Nat Rev Genet, 10, 252-63. VERDUN, R. E., CRABBE, L., HAGGBLOM, C. & KARLSEDER, J. 2005. Functional human telomeres are recognized as DNA damage in G2 of the cell cycle. Mol. Cell, 20, 551-561. VERDUN, R. E. & KARLSEDER, J. 2006. The DNA damage machinery and homologous recombination pathway act consecutively to protect human telomeres. Cell, 127, 709-720. VERWEIJ, N., EPPINGA, R. N., HAGEMEIJER, Y. & VAN DER HARST, P. 2017. Identification of 15 novel risk loci for coronary artery disease and genetic risk of recurrent events, atrial fibrillation and heart failure. Sci Rep, 7, 2761. VLAIKOU, A. M., MANOLAKOS, E., NOUTSOPOULOS, D., MARKOPOULOS, G., LIEHR, T., VETRO, A., ZIEGLER, M., WEISE, A., KRESKOWSKI, K., PAPOULIDIS, I., THOMAIDIS, L. & SYRROU, M. 2014. An interstitial

218

4q31.21q31.22 microdeletion associated with developmental delay: case report and literature review. Cytogenet Genome Res, 142, 227-38. WADE, P. A., JONES, P. L., VERMAAK, D. & WOLFFE, A. P. 1998. A multiple subunit Mi-2 histone deacetylase from Xenopus laevis cofractionates with an associated Snf2 superfamily ATPase. Curr Biol, 8, 843-6. WAGA, S., HANNON, G. J., BEACH, D. & STILLMAN, B. 1994. The p21 inhibitor of cyclin-dependent kinases controls DNA replication by interaction with PCNA. Nature, 369, 574-578. WAN, B., YIN, J., HORVATH, K., SARKAR, J., CHEN, Y., WU, J., WAN, K., LU, J., GU, P., YU, E. Y., LUE, N. F., CHANG, S., LIU, Y. & LEI, M. 2013. SLX4 assembles a telomere maintenance toolkit by bridging multiple endonucleases with telomeres. Cell Rep, 4, 861-869. WANG, H., WANG, H., POWELL, S. N., ILIAKIS, G. & WANG, Y. 2004. ATR affecting cell radiosensitivity is dependent on homologous recombination repair but independent of nonhomologous end joining. Cancer Res, 64, 7139-7143. WANG, M., WU, W., WU, W., ROSIDI, B., ZHANG, L., WANG, H. & ILIAKIS, G. 2006. PARP-1 and Ku compete for repair of DNA double strand breaks by distinct NHEJ pathways. Nucleic Acids Res, 34, 6170-6182. WANG, Y., SHI, J., CHAI, K., YING, X. & ZHOU, B. P. 2013. The Role of Snail in EMT and Tumorigenesis. Curr Cancer Drug Targets, 13, 963-972. WEIRAUCH, M. T. & HUGHES, T. R. 2011. A catalogue of eukaryotic transcription factor types, their evolutionary origin, and species distribution. Subcell Biochem, 52, 25-73. WELCH, J. J., WATTS, J. A., VAKOC, C. R., YAO, Y., WANG, H., HARDISON, R. C., BLOBEL, G. A., CHODOSH, L. A. & WEISS, M. J. 2004. Global regulation of erythroid gene expression by transcription factor GATA-1. Blood, 104, 3136-47. WILSON, J. S., TEJERA, A. M., CASTOR, D., TOTH, R., BLASCO, M. A. & ROUSE, J. 2013. Localization-dependent and -independent roles of SLX4 in regulating telomeres. Cell Rep, 4, 853-860. WOLFE, S. A., NEKLUDOVA, L. & PABO, C. O. 2000. DNA recognition by Cys2His2 zinc finger proteins. Annu Rev Biophys Biomol Struct, 29, 183-212. WONG, K. K., MASER, R. S., SAHIN, E., BAILEY, S. T., XIA, H., JI, H., MCNAMARA, K., NAYLOR, M., BRONSON, R. T., GHOSH, S., WELSH, R. & DEPINHO, R. A. 2007. Diminished lifespan and acute stress-induced death in DNA-PKcs- deficient mice with limiting telomeres. Oncogene, 26, 2815-2821. WONG, L. H., MCGHIE, J. D., SIM, M., ANDERSON, M. A., AHN, S., HANNAN, R. D., GEORGE, A. J., MORGAN, K. A., MANN, J. R. & CHOO, K. H. 2010. ATRX interacts with H3.3 in maintaining telomere structural integrity in pluripotent embryonic stem cells. Genome Res, 20, 351-360. WU, G., JIANG, X., LEE, W. H. & CHEN, P. L. 2003. Assembly of functional ALT- associated promyelocytic leukemia bodies requires Nijmegen breakage syndrome 1. Cancer Res, 63, 2589-2595. WU, L., MULTANI, A. S., HE, H., COSME-BLANCO, W., DENG, Y., DENG, J. M., BACHILO, O., PATHAK, S., TAHARA, H., BAILEY, S. M., DENG, Y., BEHRINGER, R. R. & CHANG, S. 2006. Pot1 deficiency initiates DNA damage checkpoint activation and aberrant homologous recombination at telomeres. Cell, 126, 49-62. WU, P., TAKAI, H. & DE LANGE, T. 2012. Telomeric 3' overhangs derive from resection by Exo1 and Apollo and fill-in by POT1b-associated CST. Cell, 150, 39-52.

219

WU, P., VAN OVERBEEK, M., ROONEY, S. & DE LANGE, T. 2010. Apollo contributes to G overhang maintenance and protects leading-end telomeres. Mol. Cell, 39, 606-617. XU, M., QIN, J., TSAI, S. Y. & TSAI, M. J. 2015. The role of the orphan nuclear receptor COUP-TFII in tumorigenesis. Acta Pharmacol Sin, 36, 32-6. XUE, Y., WONG, J., MORENO, G. T., YOUNG, M. K., COTE, J. & WANG, W. 1998. NURD, a novel complex with both ATP-dependent chromatin-remodeling and histone deacetylase activities. Mol Cell, 2, 851-61. YANG, S. F., SUN, A. A., SHI, Y., LI, F. & PICKETT, H. A. 2018. Structural and functional characterization of the RBBP4-ZNF827 interaction and its role in NuRD recruitment to telomeres. Biochem J, 475, 2667-2679. YE, J., LENAIN, C., BAUWENS, S., RIZZO, A., SAINT-LEGER, A., POULET, A., BENARROCH, D., MAGDINIER, F., MORERE, J., AMIARD, S., VERHOEYEN, E., BRITTON, S., CALSOU, P., SALLES, B., BIZARD, A., NADAL, M., SALVATI, E., SABATIER, L., WU, Y., BIROCCIO, A., LONDONO-VALLEJO, A., GIRAUD- PANIS, M. J. & GILSON, E. 2010. TRF2 and apollo cooperate with topoisomerase 2α to protect human telomeres from replicative damage. Cell, 142, 230-242. YE, J. Z., DONIGIAN, J. R., VAN OVERBEEK, M., LOAYZA, D., LUO, Y., KRUTCHINSKY, A. N., CHAIT, B. T. & DE LANGE, T. 2004a. TIN2 binds TRF1 and TRF2 simultaneously and stabilizes the TRF2 complex on telomeres. J. Biol. Chem, 279, 47264-47271. YE, J. Z., HOCKEMEYER, D., KRUTCHINSKY, A. N., LOAYZA, D., HOOPER, S. M., CHAIT, B. T. & DE LANGE, T. 2004b. POT1-interacting protein PIP1: a telomere length regulator that recruits POT1 to the TIN2/TRF1 complex. Genes Dev, 18, 1649-1654. YEAGER, T. R., NEUMANN, A. A., ENGLEZOU, A., HUSCHTSCHA, L. I., NOBLE, J. R. & REDDEL, R. R. 1999. Telomerase-negative immortalized human cells contain a novel type of promyelocytic leukemia (PML) body. Cancer Res, 59, 4175-4179. ZENG, Z., SHAFFER, J. R., WANG, X., FEINGOLD, E., WEEKS, D. E., LEE, M., CUENCO, K. T., WENDELL, S. K., WEYANT, R. J., CROUT, R., MCNEIL, D. W. & MARAZITA, M. L. 2013. Genome-wide association studies of pit-and- fissure- and smooth-surface caries in permanent dentition. J Dent Res, 92, 432- 7. ZHANG, J. M., YADAV, T., OUYANG, J., LAN, L. & ZOU, L. 2019a. Alternative Lengthening of Telomeres through Two Distinct Break-Induced Replication Pathways. Cell Rep, 26, 955-968 e3. ZHANG, Q. S., MANCHE, L., XU, R. M. & KRAINER, A. R. 2006. hnRNP A1 associates with telomere ends and stimulates telomerase activity. RNA, 12, 1116-1128. ZHANG, T., ZHANG, Z., SHENGZHAO, G., LI, X., LIU, H. & ZHAO, Y. 2019b. Strand break-induced replication fork collapse leads to C-circles, C-overhangs and telomeric recombination. PLoS Genet, 15, e1007925. ZHANG, X., YUAN, X., ZHU, W., QIAN, H. & XU, W. 2015. SALL4: an emerging cancer biomarker and target. Cancer Lett, 357, 55-62. ZHANG, Y., LEROY, G., SEELIG, H. P., LANE, W. S. & REINBERG, D. 1998. The dermatomyositis-specific autoantigen Mi2 is a component of a complex containing histone deacetylase and nucleosome remodeling activities. Cell, 95, 279-89.

220

ZHONG, Z. H., JIANG, W. Q., CESARE, A. J., NEUMANN, A. A., WADHWA, R. & REDDEL, R. R. 2007. Disruption of telomere maintenance by depletion of the MRE11/RAD50/NBS1 complex in cells that use alternative lengthening of telomeres. J. Biol. Chem, 282, 29314-29322. ZHOU, X. E., SUINO-POWELL, K. M., XU, Y., CHAN, C. W., TANABE, O., KRUSE, S. W., REYNOLDS, R., ENGEL, J. D. & XU, H. E. 2011. The orphan nuclear receptor TR4 is a vitamin A-activated nuclear receptor. J. Biol. Chem, 286, 2877-2885. ZOU, L. 2018. Ataxia Telangiectasia-Mutated and Rad3-Related Inhibition and Topoisomerase I Trapping Create a Synthetic Lethality in Cancer Cells. J Clin Oncol, 36, 1628-1630. ZOU, Y., LIU, Y., WU, X. & SHELL, S. M. 2006. Functions of human replication protein A (RPA): from DNA replication to DNA damage and stress responses. J. Cell. Physiol, 208, 267-273.

221