Raymond2020 Redacted.Pdf (4.359Mb)

This thesis has been submitted in fulfilment of the requirements for a postgraduate degree (e.g. PhD, MPhil, DClinPsychol) at the University of Edinburgh. Please note the following terms and conditions of use:

This work is protected by copyright and other intellectual property rights, which are retained by the thesis author, unless otherwise stated. A copy can be downloaded for personal non-commercial research or study, without prior permission or charge. This thesis cannot be reproduced or quoted extensively from without first obtaining permission in writing from the author. The content must not be changed in any way or sold commercially in any format or medium without the formal permission of the author. When referring to this work, full bibliographic details including the author, title, awarding institution and date of the thesis must be given.

Uncovering non-canonical functions for the

E3 ubiquitin ligase UBR2 in the mouse

Eleanor Siân Raymond

Presented for the degree of Doctor of Philosophy Genetics and Molecular Medicine

The University of Edinburgh September 2019 Minor corrections accepted February 2020

Declaration

I declare that the work presented within this thesis is my own, unless explicitly stated. The work has not been submitted for any other degree or professional qualification.

Eleanor Siân Raymond 16th September 2019

Acknowledgements

This thesis was a labour of love – and sometimes frustration – and it would not have been possible without the innumerable people who supported me, who deserve enormous gratitude and respect.

Firstly, and most importantly, Ian Adams. He was a constant source of inspiration and motivation, and always had confidence in my ideas. Without him, this thesis would never have been conceived, and his constant support allowed me the freedom to take his idea and run with it.

Members of the Adams Family, both past and present, were all largely responsible for my day- to-day sanity. James Crichton and Karen Dobie, in particular, were always available for advice, complaints and emotional support. Simon Brown, Veronica Duffy, Jennifer Lawson, Issy

MacGregor and Apiwat Moolnangdeaw provided chats and snacks whenever required. Chris

Playfoot, Marie MacLennan and Soña Relovska gave me the grounding in my project from the beginning. And our dear friend David Read, with his Spotify Premium and sunshine breaks, who saw the fun in every day in the lab.

Thank you to every member of the Pub Club, whose empathy, PhD memes and pub quiz knowledge were vital in maintaining a sense of balance that we all sorely needed as we went through this experience together. To Frank, Hollie, those in the Constellation and everyone else – you allowed me an escape from the science, and you became a community in which I was incredibly included and supported throughout anything.

And finally, my family – those who have never faltered in their confidence in my abilities, even when my own confidence was lacking. My parents, Diane and David Raymond, who instilled a great sense of curiosity in me from a young age. For mathematical advice, Daniel Cunnah.

For IT and tech. support, Sean Casey. And for snacks and cheerleading, Charlotte Cunnah.

Thank you.

“Opening your eyes is all that is needed.

Look with your eyes. Hear with your ears. Taste with your mouth.

Smell with your nose. Feel with your skin.

Then comes the thinking, afterward, and in that way: knowing the truth.”

– George R. R. Martin

This work, and the research contained within it, are dedicated to Sean Casey. You kept me going, every morning and every night.

The final days of this process felt almost insurmountable without the support of Kate, Charly, Daniel, Sophie and so many more, who gave me a roof

over my head and a place to work.

Abstract

Ubiquitination is an essential cellular mechanism in which the small polypeptide ubiquitin is attached to proteins, so regulating protein function or targeting them for degradation via the proteasome. The transfer of ubiquitin to specific target proteins is regulated by the highly diverse group of E3 ubiquitin ligases. These require two key domains: a substrate recognition site, and a ubiquitin carrier E2 protein binding site, which allow close proximity of the two, so ubiquitin transfer can occur. One family of E3 ubiquitin ligases are the UBR proteins, which confer substrate specificity through the N-end Rule pathway – in which they, canonically, bind either Type I (basic, e.g. Arginine) or Type II (bulky hydrophobic, e.g. Phenylalanine) amino acids on the N-terminus of the substrate. There are seven known UBR proteins, and

UBR2 is one of four in this family that are thought to have overlapping roles in the N-end Rule pathway. Published murine phenotypes in a Ubr2-/- mouse include male infertility and female- specific embryonic lethality. The tissue-specificity of the Ubr2-/- phenotype has been attributed to tissue-specific expression of UBR proteins allowing critical functions of the N- end rule pathway to become detectable in specific cell types. These phenotypes also appear to overlap that of a Tex19.1-/- mouse, which has undetectable expression of the UBR2- interacting protein TEX19.1. This is specifically expressed in hypomethylated tissues: the testes, ovaries, placenta, and in a mouse embryonic stem cell model.

In addition to the male infertility observed in both mouse models, Tex19.1-/- phenotypes include placental defects associated with intrauterine growth restriction, and cohesin misregulation during female meiosis, in the oocyte. There is some evidence that UBR2’s interaction with

TEX19.1 regulates its Type II N-end Rule activity, and so I hypothesized that UBR2 also has roles in the other phenotypes of the Tex19.1-/- mouse, and that these act through the N-end Rule, which has thus far not been directly shown in any of the phenotypes present.

Firstly, I aimed to study a potential role for UBR2 during female meiosis. In oocytes, cohesin molecules hold homologous chromosomes in a tightly aligned conformation. This must be

7 maintained for long periods of time, between meiotic prophase in the embryonic ovary and resumption of meiosis prior to ovulation during adulthood. This maintains chromosome stability and prevents missegregation during the first meiotic division, leading to aneuploidy and reduced fertility. A longer arrest period, i.e. in oocytes isolated from older females, is associated with loss of cohesin and increased chance of aneuploidy. In absence of UBR2 in oocytes from young females, less cohesin is present on the chromosome arms at the first meiotic metaphase and homologs are less tightly held together, so implying a role for protection of chromosome cohesion over time.

Next, in accordance with previous observations in absence of UBR2, I identified an embryonic lethality present in our Ubr2-/- mice, which is associated with intrauterine growth restriction and placental defects. Although this is superficially similar to the Tex19.1-/- placental phenotype, there also appear to be TEX19.1-independent functions of UBR2 in this tissue.

Finally, I aimed to ascertain the mechanism by which UBR2 plays a role in any of these phenotypes. As UBR2 is thought to confer particular specificity towards Type II N-end Rule substrates, and TEX19.1 is thought to regulate this role of UBR2, I used CRISPR-Cas9 technology to generate an Ubr2T2/T2 mouse model, which has undetectable Type II substrate binding. There are no gross phenotypes in any of the tissues identified using the Ubr2-/- null mouse; both placental growth and meiotic cohesin appear to be normally regulated, and so these functions of UBR2 do not appear to be acting through the N-end Rule pathway.

Additionally, the well-established role for UBR2 during spermatogenesis does not appear to act through Type II N-end Rule substrate binding, as has been assumed. Instead, I conclude that, although UBR2 is an important part of the N-end Rule pathway in other contexts, it has key non-canonical functions in both male and female fertility, and in placental development, which cannot be compensated for by other UBR proteins.

Lay Abstract

Proteins are the functional products of genes, and they act in cells to carry out the processes required for cells to survive. In order to keep cells or tissues functioning optimally, individual protein levels need to be tightly controlled, and so an efficient recycling process exists in order to “tag” proteins that need to be turned over – this is called ubiquitination. A small molecule, called ubiquitin, is transferred to a target protein by an ubiquitin ligase enzyme and this “tags” the protein, directing it towards a recycling centre called the proteasome. There are over

10,000 proteins in any given cell, and so target specificity is very important in order to recycle the correct proteins and ensure cell survival. One ubiquitin ligase, called UBR2, uses a process called the N-end Rule to specifically bind its target proteins. There are four UBR enzymes that are all thought to have overlapping roles in the N-end Rule pathway, but in a Ubr2 mutant mouse, in which UBR2 is not present to recycle its target proteins, there are a specific subset of tissues affected. This implies that UBR2 has unique functions in these tissues that cannot be covered by other UBR enzymes. Of these unique functions, the best studied prior to this thesis is its role in male fertility. I hypothesized that UBR2 is also important for female fertility and that both of these roles are regulated by UBR2’s N-end Rule activity, which has thus far not been directly shown.

Firstly, I aimed to characterise poorly understood roles of UBR2 in female fertility. I have shown that UBR2 appears to regulate chromosome cohesion during egg (oocyte) production.

All oocytes are present in the ovary during embryogenesis, and do not complete their development until immediately prior to ovulation, when pairs of chromosomes are divided in half. This can be over a period of many months in a mouse model, and during this time, chromosomes need to be held tightly together in order to align them and prevent missegregation upon division – this chromosome “cohesion” is maintained by a protein molecule called cohesin. In absence of UBR2, less cohesin is present and chromosomes are less tightly held together, increasing the risk of having the wrong number of chromosomes in the

9 resulting egg – a condition called aneuploidy which is associated with subfertility, miscarriage and developmental disorders such as Down Syndrome. UBR2 therefore is a potential candidate for protection of chromosome cohesion over time, which is especially associated with older women and their declining fertility in the years prior to menopause.

While investigating aspects of chromosome behaviour during egg development, I noticed that

Ubr2 mutant embryos were small for their gestational age, a condition that arises in around

10% of human pregnancies and is often caused by defects in the placenta. By investigating the cause of this growth restriction, I have also been able to uncover a novel role of UBR2 in placental development in mice.

Secondly, I wanted to investigate the mechanism behind these unique functions of UBR2; both male and female fertility, as well as in placental development. Instead of removing the

Ubr2 gene, as in the mouse model used above, I introduced specific mutations in order to prevent N-end Rule target binding of UBR2, while preserving the enzyme itself inside cells. I generated this mutation in a mouse model using the CRISPR-Cas9 genome editing technology and hypothesized that, if UBR2 is acting through the N-end Rule, these mice would also have fertility problems and placental defects. Interestingly, although the N-end Rule is the main known functional pathway of UBR2, I did not observe any defects in the testis, ovaries or placenta when the ability of UBR2 to act in this pathway was specifically disrupted. This is the first time that a UBR protein has been identified as having functions outside the N-end

Rule in a mouse model, and potentially outside of ubiquitination and protein recycling altogether.

Table of Contents

Declaration ...... 1

Acknowledgements ...... 3

Abstract ...... 7

Lay Abstract ...... 9

Table of Contents ...... 11

Abbreviations and Definitions ...... 17

List of Figures ...... 21

List of Tables ...... 25

Chapter 1 - Literature Review………………………………………………………….……………………….……27 1.1 Ubiquitination and Protein Turnover ...... 29

1.1.1 The E1-E2-E3 Ubiquitination Cascade ...... 30

1.1.2 The N-end Rule Pathway for Protein Degradation ...... 33

1.1.2.1 Classes of N-terminal Degradation ...... 33

1.1.2.2 Type I or Type II N-terminal Modifications ...... 35

1.1.2.3 Evolutionary Conservation of the UBR Protein Family ...... 38

1.1.2.4 Substrate Binding Specificity in the Arg/N-end Rule ...... 43

1.1.3 E2/E3 Interaction Dynamics and the E3 RING Domain ...... 47

1.1.4 The 26S Proteasome ...... 50

1.1.4.1 Regulatory Functions of Proteasomal Degradation ...... 51

1.1.4.2 Misfolded or Orphan Protein Turnover ...... 52

1.1.5 Non-Proteasomal Functions of Ubiquitination ...... 54

1.2 UBR2 and TEX19.1 ...... 57

1.2.1 Similarities to Tex19.1-/- Phenotypes ...... 58

1.2.2 A Model for TEX19.1-UBR2 Complex Function ...... 60

1.3 Chromosome Cohesion in the Mitotic Cell Cycle ...... 63

1.3.1 The Cohesin Complex ...... 63

1.3.2 Cohesin Dynamics in the Mitotic Cell Cycle ...... 67

1.3.3 Cohesinopathies ...... 71

1.4 Meiosis and Genomic Stability during Reproduction ...... 73

1.4.1 Aneuploidy and Fertility...... 73

1.4.2 The Two Meiotic Divisions ...... 74

1.4.2.1 Prophase I: Synapsis and Recombination ...... 75

1.4.2.2 Metaphase I: Chiasmata Stability and Homolog Segregation ...... 78

1.4.2.3 Meiosis II: Segregation of Sister Chromatids ...... 81

1.4.3 Gendered Gametogenesis ...... 81

1.4.4 Hypotheses for Oocyte Aneuploidy ...... 82

1.4.4.1 Meiotic Chromosome Cohesion ...... 84

1.4.4.2 SAC Stringency ...... 86

1.4.4.3 Ovarian Tissue Environment ...... 87

1.4.4.4 Systemic Ageing ...... 88

1.5 Placentation and its Role in IUGR ...... 91

1.5.1 The Murine Placenta ...... 94

1.5.2 Roles for Tex19.1 in Placental Development ...... 97

1.5.2.1 Genomic Imprinting in the Placenta ...... 97

1.5.2.2 Retrotransposons and the Placental Immune Response...... 100

1.6 Summary ...... 103

1.7 Thesis Objectives ...... 105

Chapter 2 - Materials and Methods……………………………………………………………………………..107 2.1 Animal Work ...... 109

2.1.1 Dissection and tissue retrieval ...... 109

2.1.2 Embryo tissue retrieval ...... 109

2.1.3 CRISPR-HDR by zygotic injection ...... 110

2.1.4 Genotyping assays ...... 111

2.2 Tissue culture ...... 115

2.2.1 Cell storage and general culture conditions ...... 115

2.2.2 HEK293T cells ...... 115

2.2.3 Mouse Embryonic Stem Cells ...... 116

2.2.3.1 CRISPR-Cas9 HDR for UBR2-T1 mES cell generation ...... 116

2.2.3.2 mESC synchronisation and sorting ...... 118

2.3 Histology and tissue analysis ...... 119

2.3.1 Fixation and tissue preparation ...... 119

2.3.2 H&E tissue staining ...... 119

2.3.3 Periodic Acid Schiff-Haematoxylin (PAS-H) tissue staining ...... 120

2.4 Meiotic chromosome spreads ...... 120

2.4.1 Prophase I oocyte spreads ...... 120

2.4.2 Prometaphase I oocyte culture and spreads ...... 121

2.4.3 Metaphase II oocyte culture and spreads ...... 122

2.4.4 Immunofluorescent staining ...... 122

2.5 Microscopy and analysis ...... 125

2.5.1 Brightfield microscopy ...... 125

2.5.1.1 Tiled brightfield imaging and analysis ...... 125

2.5.2 Epifluorescence microscopy...... 125

2.5.2.1 Z-stack epifluorescent microscopy ...... 126

2.5.3 Image analysis and quantification ...... 127

2.6 Protein analysis ...... 129

2.6.1 Preparation from cell culture ...... 129

2.6.1.1 Chromatin preparation ...... 129

2.6.1.2 Co-immunoprecipitation ...... 129

2.6.1.3 Peptide pulldown assays ...... 130

2.6.2 Preparation from animal tissues ...... 132

2.6.2.1 Chromatin preparation ...... 132

2.6.2.2 Peptide pulldown ...... 133

2.6.3 Western blotting and quantification ...... 133

2.7 RNA analysis ...... 135

2.7.1 RNA preparation ...... 135

2.7.2 qRT-PCR ...... 135

2.8 DNA analysis ...... 139

2.8.1 Plasmids ...... 139

2.8.1.1 Bacterial culture and DNA extraction ...... 139

2.8.1.2 Site directed mutagenesis ...... 140

2.8.1.3 RNA guide cloning for CRISPR-Cas9 ...... 142

2.8.2 DNA sequencing ...... 144

Chapter 3 - Ubr2 in Oocyte Development and Chromosome Cohesion during Female Meiosis………………………………………………………………………………………………………………………..147 3.1 Introduction ...... 149

3.2 Results ...... 153

3.2.1 AcSMC3 is specifically downregulated in Ubr2-/- metaphase I oocytes...... 153

3.2.2 Loss of Ubr2 does not negatively affect cohesin functions in foetal prophase I ... 158

3.2.3 Ubr2-/- metaphase I oocytes exhibit defects in sister chromatid cohesion...... 164

3.2.3.1 Ubr2 in chromosome arm cohesion ...... 164

3.2.3.2 Ubr2 in pericentromeric cohesion ...... 169

3.2.4 Ubr2-/- oocytes are not aneuploid ...... 173

3.3 Discussion ...... 177

3.3.1 The TEX19.1-UBR2 complex in meiotic cohesin regulation ...... 177

3.3.2 Ubr2 in regulation of cohesin in the arrested oocyte ...... 181

3.3.3 Ubr2 in regulation of MII kinetochore distance...... 183

3.3.4 Summary ...... 185

Chapter 4 - The Tex19.1-associated Role of Ubr2 in Embryonic Growth and Placental Trophoblast-Derived Tissues……………………………………………………………………………………....187 4.1 Introduction ...... 189

4.2 Results ...... 195

4.2.1 Ubr2-/- females on a C57BL/6J background are small and under-represented at weaning ...... 195

4.2.2 Ubr2-/- embryos exhibit intrauterine growth restriction and a placental size defect ...... 200

4.2.3 The Ubr2-/- and Tex19.1-/-placental defects arise at a similar developmental timepoint ...... 207

4.2.5 Ubr2 has a role in trophoblast-derived layers of the placenta ...... 209

4.2.6 Placental Tex19.1-/- retrotransposon derepression does not act through UBR2 ... 215

4.2.7 Tex19.1-mediated repression of imprinted loci is independent of Ubr2 ...... 219

4.3 Discussion ...... 223

4.3.1 Strain- and sex-dependent penetrance of the Ubr2-/- phenotype ...... 224

4.3.2 Ubr2’s role in cohesin regulation ...... 227

4.3.3 Roles for UBR proteins in embryonic and placental development ...... 229

4.3.4 Summary...... 230

Chapter 5 - Functional Mutations in the UBR2 N-end Rule……………………………………………231 5.1 Introduction ...... 233

5.2 Results ...... 239

5.2.1 In vitro validation of the Ubr2D118A and Ubr2D233A point mutations ...... 239

5.2.2 Generation and validation of a Ubr2-T2D233A mouse line using CRISPR-Cas9 ... 247

5.2.3 Phenotyping of the Ubr2-T2D233A mouse line ...... 253

5.2.3.1 Ubr2’s role in trophoblast-derived placental development is not a Type II function ...... 253

5.2.3.2 There are no gross infertility phenotypes observed in a Ubr2T2/T2 male mouse ...... 260

5.2.3.3 The novel role of Ubr2 in cohesin regulation does not act through Type II N- end Rule ...... 264

5.2.3.4 Summary of the Ubr2T2/T2 mouse phenotype ...... 273

5.2.4 A stable Ubr2-T1D118A ESC line using CRISPR Cas9 ...... 273

5.2.4.1 Generation of Ubr2-T1D118A ESCs ...... 273

5.2.4.2 Ubr2-T1D118A ESCs exhibit no cohesin misregulation ...... 277

5.3 Discussion ...... 281

5.3.1 Lack of redundancy within the UBR family ...... 281

5.3.2 Limitations of this study ...... 285

5.3.3 Further study of the N-domain and UBR box structures ...... 286

5.3.4 Summary ...... 287

Chapter 6 - Discussion and Conclusions………………………………………………………………………..289 6.1 Separation of the TEX19.1 and UBR2 Functions ...... 291

6.1.1 Spermatogenesis ...... 293

6.1.2 Placental development ...... 294

6.1.3 The interaction between TEX19.1 and UBR2 proteins ...... 295

6.2 Redundancy in the UBR Protein Family ...... 297

6.4 Further Applications for UBR2 N-end Rule Point Mutations ...... 301

6.4.1 Identification of UBR2 Substrates and Interactors ...... 301

6.4.2 Further study of the UBR2 N-end Rule Functions ...... 303

6.5 Conclusions ...... 305

References………………………………..…………………………………………………………………………………307

Abbreviations and Definitions

Ac ––––––––– Acetyl protein modification Ac/N-end ––––––––– N-terminally acetylated substrates for ubiquitination ACA ––––––––– Anti-centromeric antibody (CREST patients) AcSMC3 ––––––––– Acetylated SMC3 APC/C ––––––––– Anaphase-promoting complex, or cyclosome Arg/N-end ––––––––– N-terminal degron sequence for substrate ubiquitination CdLS ––––––––– Cornelia de Lange Syndrome cDNA ––––––––– Complementary DNA CDS ––––––––– Coding sequence CE ––––––––– Central element (SC) Clustered regularly interspaced short palindromic repeats CRISPR-Cas9 ––––––––– with Cas9 CTCF ––––––––– CCCTC-binding factor dpp ––––––––– Days post-parturition, or post-partum DSBs ––––––––– Double-strand breaks DUB ––––––––– Deubiquitinase enzyme EtOH ––––––––– Ethanol (100%, unless stated) FACS ––––––––– Fluorescence-associated cell sorting FCS ––––––––– Foetal calf serum FLAG-tag ––––––––– FLAG™ epitope tag gDNA/gRNA ––––––––– Guide DNA/RNA for CRISPR-Cas9 GFP ––––––––– Green fluorescent protein GVBD ––––––––– Germinal vesicle breakdown GV-stage ––––––––– Germinal vesicle stage oocyte H&E ––––––––– Haematoxylin and eosin H2A ––––––––– Histone 2A H3K9 ––––––––– Histone 3 Lys-9 H3K9me3 ––––––––– Histone 3 Lys-9 trimethylation – repressive histone mark hCG ––––––––– Human chorionic gonadotrophin HDR ––––––––– Homology directed repair, a type of CRISPR-Cas9 IF ––––––––– Immunofluorescence IFN-β ––––––––– Interferon-β

IUGR ––––––––– Intrauterine growth restriction Jz ––––––––– Junctional zone kDa ––––––––– Kilodaltons Lb ––––––––– Labyrinth LE ––––––––– Lateral element (SC) LINE-1 ––––––––– Long interspersed nuclear element protein 1 Md ––––––––– Maternal decidua mESC/ESC ––––––––– (Mouse) embryonic stem cell Met I ––––––––– Metaphase I (during MI) Met II ––––––––– Metaphase I (during MII) Met-Φ ––––––––– Non-acetylated substrates of the Ac/N-end Rule pathway MI ––––––––– Meiosis I MII ––––––––– Meiosis II mRNA ––––––––– Messenger RNA mtDNA ––––––––– Mitochondrial DNA Orf ––––––––– Open reading frame PAS-H ––––––––– Periodic Acid Schiff-Haematoxylin PBE ––––––––– Polar body extrusion PBS ––––––––– Phosphate-buffered saline (also known as Dulbecco’s) PBS-T ––––––––– PBS with 0.1% Tween-20 detergent PFA ––––––––– Paraformaldehyde PMSG ––––––––– Pregnant mare’s serum gonadotrophin PP2A ––––––––– Protein phosphatase 2A qRT-PCR ––––––––– Quantitative real-time polymerase chain reaction RBR ––––––––– RING-between-RING domain RBS ––––––––– Robert’s Syndrome RING ––––––––– Really Interesting New Gene domain SA (1, 2, 3) ––––––––– Stromal antigen proteins (1, 2, 3) SAC ––––––––– Spindle assembly checkpoint SC ––––––––– Synaptonemal complex SGO1/2 ––––––––– Shugoshin protein 1/2 ST/DT ––––––––– Single/Double thymidine block TE ––––––––– Trophectoderm (placental tissue) TEs ––––––––– Transposable elements Tetra-ARMS ––––––––– Tetra-primer amplification-refractory mutation system

TEX19.1 ––––––––– Testis-expressed protein 19.1 TGCs ––––––––– Trophoblast giant cells TI ––––––––– Type I (N-end Rule, Arginine) TII ––––––––– Type II (N-end Rule, Phenylalanine) UAIN ––––––––– UBR auto-inhibitory domain Ub ––––––––– Ubiquitin UBL ––––––––– Ubiquitin-like proteins UBR2 ––––––––– UBR-domain protein 2 Ubr2-/- ––––––––– Ubr2 null mouse line Ubr2T2/T2 ––––––––– Ubr2 mouse line with the D233A point mutation UPS ––––––––– Ubiquitin proteasome system UV ––––––––– Ultraviolet radiation -ve ––––––––– Negative control WABS ––––––––– Warsaw Breakage Syndrome ZFP ––––––––– Zinc-finger protein

List of Figures

1.1 ––––––––– The E1-E2-E3 Ubiquitination Cascade –––– 32 1.2 ––––––––– Phylogeny of the UBR N-end Rule E3 Family –––– 37 1.3 ––––––––– The Mammalian UBR E3 Family –––– 46 1.4 ––––––––– The TEX19.1-UBR2 Complex –––– 62 1.5 ––––––––– The Cohesin Complex –––– 65 1.6 ––––––––– Cohesin Dynamics in the Mitotic Cell Cycle –––– 70 1.7 ––––––––– Meiotic Prophase I Progression –––– 77 1.8 ––––––––– Meiotic Cohesin Establishment –––– 80 1.9 ––––––––– Differences in Mammalian Placental Development –––– 93 1.10 ––––––––– Murine Trophoectoderm Lineage Development –––– 96 3.1 ––––––––– AcSMC3 is reduced in the Ubr2-/- metaphase I oocyte –––– 156 3.2 ––––––––– Total SMC3 is unaffected in the Ubr2-/- metaphase I –––– 158 oocyte 3.3 ––––––––– The synaptonemal complex has normal morphology –––– 161 in Ubr2-/- prophase I oocytes 3.4 ––––––––– Crossover frequency and position is unaffected in the –––– 163 Ubr2-/- pachytene oocyte 3.5 ––––––––– Ubr2-/- results in chiasmata terminalisation in –––– 168 metaphase I 3.6 ––––––––– Ubr2 does not regulate centromeric distance at –––– 171 metaphase I 3.7 ––––––––– Ubr2 specifically regulates centromeric distance at –––– 172 metaphase II 3.8 ––––––––– Absence of Ubr2 does not affect ploidy in the –––– 175 metaphase II oocyte 3.9 ––––––––– A model for activity of the TEX19.1-UBR2 complex in –––– 179 meiotic oocytes 4.1 ––––––––– Ubr2 and Tex19.1 are coexpressed in hypomethylated –––– 191 developmental contexts 4.2 ––––––––– Ubr2-/- female mice do not survive to adulthood at a –––– 198 Mendelian ratio

4.3 ––––––––– Surviving female Ubr2-/- mice have a growth –––– 200 retardation 4.4 ––––––––– E18.5 Ubr2-/- mice have an embryonic growth defect –––– 201 associated with a placental defect 4.5 ––––––––– Ubr2-/- mice exhibit embryonic growth defects earlier –––– 206 in development 4.6 ––––––––– Ubr2-/- placentas are specifically reduced in size in the –––– 212 trophectoderm-derived labyrinth layer 4.7 ––––––––– The Ubr2-/- placental defect can be characterised by –––– 214 cell-specific compositional differences in both the

labyrinth and junctional zone tissue layers 4.8 ––––––––– The Ubr2-/- placental defect does not entirely –––– 218 phenocopy the Tex19.1-/- placenta 4.9 ––––––––– Ubr2 does not have a role in maintaining placental –––– 221 imprinting 5.1 ––––––––– The conserved UBR1 and UBR2 UBR box and N- –––– 237 domain 5.2 ––––––––– Site-directed mutagenesis of the Ubr2 CDS –––– 240 5.3 ––––––––– The UBR2-T1 and UBR2-T2 mutations prevent –––– 243 detectable substrate binding 5.4 ––––––––– UBR2-T1D118A and UBR2-T2D233A proteins are –––– 246 functional outside the N-end Rule 5.5 ––––––––– Generation of a Ubr2D233A mouse line using CRISPR- –––– 249 Cas9 5.6 ––––––––– Genotyping of Ubr2D233A mouse colony using tetra –––– 251 ARMS-PCR 5.7 ––––––––– The Ubr2D233A mutation prevents detectable Type II N- –––– 253 end Rule substrate binding in vivo 5.8 ––––––––– Ubr2+/T2 crosses produce offspring in Mendelian ratios –––– 256 5.9 ––––––––– Ubr2T2/T2 E18.5 embryos do not exhibit a growth or –––– 259 placental defect 5.10 ––––––––– TEX19.1 protein is stable in the Ubr2T2/T2 testis –––– 260 5.11 ––––––––– Ubr2T2/T2 males have no overt defects in –––– 263 spermatogenesis

5.12 ––––––––– Mitotic cohesin regulation is not perturbed in the –––– 265 Ubr2T2/T2 mouse 5.13 ––––––––– UBR2’s role in meiotic AcSMC3 regulation is not –––– 267 perturbed by Ubr2T2/T2 oocytes 5.14 ––––––––– UBR2’s role in total meiotic cohesin regulation is not –––– 269 perturbed in Ubr2T2/T2 oocytes 5.15 ––––––––– Ubr2T2/T2 oocytes do not exhibit defects in –––– 272 chromosome are cohesion

D118A 5.16 ––––––––– Ubr2-T1 mutant ES cell line screening –––– 276 5.17 ––––––––– ES cell synchronization to a G2/M-enriched –––– 278 population 5.18 ––––––––– Chromatin-bound cohesin levels at G2/M are –––– 280 unaffected in Ubr2D118A ESCs 5.19 ––––––––– Ubiquitous coexpression of UBR proteins across –––– 284 mouse tissues

List of Tables

2.1 ––––––––– Primers for mouse colony genotyping –––– 113

2.2 ––––––––– PCR cycling conditions for mouse colony genotyping –––– 113 assays

2.3 ––––––––– Primer sequences for UBR2-T1-D118A mESC –––– 117 genotyping: ARMS and Sanger

2.4 ––––––––– Tetra-primer ARMS-PCR cycling conditions for UBR2- –––– 117 T1-D118A mESC genotyping

2.5 ––––––––– qRT-PCR primers –––– 137

2.6 ––––––––– Plasmids used in various molecular biology assays –––– 139

2.7 ––––––––– Site-directed mutagenesis primers –––– 140

2.8 ––––––––– PCR cycling conditions for site-directed mutagenesis –––– 141

2.9 ––––––––– CRISPR gRNA generation constructs –––– 142

2.10 ––––––––– Primers used in CRISPR gRNA generation –––– 143

2.11 ––––––––– PCR cycling conditions for T7 pX330 guide –––– 144 amplification

2.12 ––––––––– Antibodies used for various assays –––– 145

Chapter 1

Literature Review

1.1 Ubiquitination and Protein Turnover

Protein turnover is an integral part of cell function and is vital for homeostasis of the eukaryotic proteome. In addition, there is an intrinsic requirement for turnover of certain proteins in essential cellular pathways involved in the cell cycle, DNA replication and stress responses, and in response to various stimuli in the eukaryotic system (Sadowski and Sarcevic

2010; Bedford et al. 2010; Hicke 2001). One of the major signaling pathways involved in protein turnover is the ubiquitin-proteasome system (UPS). Ubiquitin is a small 9kDa protein that acts as a covalently-bound signalling molecule across a wide variety of functions in the cell, particularly in regulation of protein levels and turnover (Pickart and Eddins 2004). It is covalently conjugated to target substrates via Lysine residues, and polyubiquitin chains can be built up by ubiquitination of Lysines within the ubiquitin protein itself. There are seven

Lysine residues within the molecule, but the most common ubiquitin chains are built up along

K48, K33, K29, K11 or K63 residues (W Li and Ye 2008). Protein turnover is generally understood to act through recognition of K48 or K63 polyubiquitin chains (Ohtake et al. 2018;

Tsuchiya et al. 2017), but the variety of different polyubiquitin chains, or even monoubiquitin signalling, act to carry out a diverse array of ubiquitination functions (Sadowski et al. 2012).

In addition to polyubiquitin chains, combinations of different ubiquitin-lysine linkages can result in branched chains that perform specific proteostatic functions (H. T. Kim et al. 2007).

Quality control in protein synthesis is another vital function of the UPS, and there are also functions for ubiquitination outside of protein degradation, for example in protein regulation through interaction with proteins containing UBDs (ubiquitin binding domains) (Tsuchiya et al. 2017).

Ubiquitin was the first identified of a family of small ubiquitin-like proteins (Ubls), which share structural similarities and can be conjugated to target proteins for signalling purposes in a similar way to ubiquitin itself. These include SUMO and NEDD8, among many others.

They require similar but distinct conjugating enzyme cascades as for ubiquitination, but act to perform a variety of signalling functions in eukaryotic cells, primarily through promoting

29 the interaction of the target with SIM (SUMO-interacting motifs) or other UBL-binding proteins (Hochstrasser 2009).

1.1.1 The E1-E2-E3 Ubiquitination Cascade

There are three required steps to ubiquitination of a substrate: binding to the ubiquitin molecule itself, binding to substrate proteins, and transfer of ubiquitin onto the substrate. For each function, there exists an individual class of enzymes, and together they act in a cascade of interactions, called the E1-E2-E3 cascade (Fig. 1.1). For preparation of the ubiquitin molecule itself, the E1 ubiquitin-activating enzyme undergoes ATP hydrolysis (Schulman and

Wade Harper 2009) in order to bind specifically through a thioester bond with an active site cysteine residue, to the ubiquitin molecule (Fig. 1.1a). The activated ubiquitin is then transferred to an E2 ubiquitin-conjugating enzyme (Fig. 1.1b), again via a thioester linkage

(Olsen and Lima 2013). Substrate specificity in this pathway is conferred through interaction of the target protein with an E3 ubiquitin ligase, which is the largest and most diverse group of the ubiquitination enzymes. Physically, these can be considered bi-substrate enzymes as they bind both the substrate and the E2, bringing them into the correct conformation for ubiquitin transfer (Fig. 1.1c). There are three major classes of E3 ubiquitin ligase, defined by their interaction mode with the E2 enzyme: RING (Really Interesting New Gene), HECT

(Homologous to E6‐Associated Protein C-Terminus) or RBR (RING-between-RING)

(Berndsen and Wolberger 2014). An interaction between a ubiquitin charged E2 and the E3 brings the target protein in close enough proximity to the E2 for ubiquitin transfer to occur.

There are roles for monoubiquitination or multiubiquitination in protein regulation within the cell (Sadowski and Sarcevic 2010) (Fig. 1.1f), but a large proportion of ubiquitinated target proteins are polyubiquitinated, and this requires multiple subsequent interactions between the E3 and charged E2 enzymes to allow building of a ubiquitin chain on the substrate (Fig.

1.1d,e). Therefore, E2/E3 interactions must be low-affinity and transient, although long-lasting

30 enough for ubiquitin-transfer to occur. A non-ubiquitin-charged E2 has lower binding affinity to either the RING or HECT domains of an E3, and this aids in dissociation of the E2/E3 complex after ubiquitin transfer (Metzger et al. 2014; Sluimer and Distel 2018). A polyubiquitinated substrate is then recognised by intrinsic components of the 26S proteasomal lid complex such as Rpn10, Rpn13 or Sem1 (Paraskevopoulos et al. 2014;

Kragelund et al. 2016; Hamazaki, Hirayama, and Murata 2015) (Fig. 1.1g), and the target protein is deubiquitylated by a proteasome-associated DUB enzyme in order for it to be degraded within this complex (Bedford et al. 2010). Individual ubiquitin molecules and amino acids are then recycled back into the proteasome (Fig. 1.1h).

Figure 1.1 The E1-E2-E3 Ubiquitination Cascade Transfer of the ubiquitin (Ub) molecule to substrate proteins occurs in a cascade of three classes of ubiquitination enzymes. An E1 ubiquitin activator undergoes ATP hydrolysis (a) in order to charge a Ub molecule, and this is transferred to an E2 ubiquitin conjugating enzyme (b). An E3 ubiquitin ligase binds both the substrate and the Ub-charged E2 and Ub is transferred to the target protein (c). Ubiquitination is able to alter the function of a protein, either through monoubiquitination (d) or subsequent E2-E3 interactions (e) and construction of a Ub linkage chain (f). Either polyubiquitinated or monoubiquitinated proteins interact with UbRs (Ub receptors). The classical pathway of ubiquitination involves binding of polyubiquitinated substrates to proteasomal UbRs, and degradation in the 26S proteasome (g). The proteasome complex contains both deubiquitinases (DUBs) and proteases, which release the Ub molecule for further ubiquitination experiments and subsequently degrade the substrate protein. Released amino acids and Ub molecules are recycled for protein synthesis and further rounds of protein degradation (h). Taken from (Deshaies and Joazeiro 2009).

1.1.2 The N-end Rule Pathway for Protein Degradation

E3 ubiquitin ligases are by far the largest and most diverse group of enzymes within the E1-

E2-E3 cascade, encoded by over 600 mammalian genes (Wei Li et al. 2008), and it is these that provide substrate specificity for the UPS. In contrast, there are only 2 known mammalian E1 ubiquitin activators enzymes, UBA1 and UBA6 (Schulman and Wade Harper 2009; J. Jin et al.

2007), which are promiscuous for binding to the mammalian E2 enzymes, of which there are around 40 known (Stewart et al. 2016).

One of the earliest discovered E3 substrate binding pathways, the N-end Rule, depends on the N-terminal sequence of the substrate, which acts as a degron binding site for the E3 ubiquitin ligase to carry out ubiquitination (Bachmair, Finley, and Varshavsky 1986). This was first discovered in S. cerevisiae, with identification of the E3 substrate recognition protein

Ubr1 (Bartel, Wünning, and Varshavsky 1990). Multicellular eukaryotes encode a family of seven UBR proteins, of which UBR1 is the best characterised, which are expected to carry out

N-end recognition (T Tasaki et al. 2005). The N-end Rule pathway, as for most E3s, also requires at least one internal Lysine residue in the target protein to act as the site for initial ubiquitin attachment by the associated E2 enzyme.

1.1.2.1 Classes of N-terminal Degradation

N-terminal degron sequences can be classified into two main subcategories: those

that are recognised by the Arg/N-end Rule, and by the Ac/N-end Rule. Nascent

proteins are expressed with an N-terminal methionine, encoded by the start codon

ATG. However, N-terminal post-translational modifications are not uncommon, and

can result in the modified protein being targeted for degradation by N-end Rule

machinery.

Firstly, the Arg/N-end Rule depends on the specific modification of the N-terminus into either Type I or Type II N-terminal recognition sequences. These types are defined by the N-terminal residues on the target protein. Type I is generally basic residues, the most common of these being arginine (Arg) (T Tasaki et al. 2005), which gives the classification its name. The Type II Arg/N-end Rule is defined instead by bulky hydrophobic N-terminal residues, such as phenylalanine (Phe) or tryptophan

(Trp). The Arg/N-end Rule is the more widely studied of the two categories (J. M.

Kim and Hwang 2014), and will be considered the major pathway of N-end Rule ubiquitination for the remainder of this thesis.

The Ac/N-end Rule pathway for protein degradation acts on proteins containing an

N-terminal acetyl group; this is a cotranslational modification that occurs during translation of 90% of mammalian proteins (Nguyen et al. 2019; Polevoda, Arnesen, and Sherman 2009). This is a lesser understood category of the N-end Rule, and was only recently identified as a degradation signal, in 2010 (Hwang, Shemorry, and

Varshavsky 2010). As so many Ac/N-end target proteins exist, the acetylated N- terminus is often shielded by structural folding or interaction in a protein complex

(Shemorry, Hwang, and Varshavsky 2013). This gives a mechanism by which E3 ubiquitin ligases such as TEB4 (Park et al. 2015) can detect protein misfolding or deleterious conformational changes, so inducing degradation via the UPS and maintaining proteomic integrity. Acetylation at the N-terminus depends on the residue present in the second position, immediately after the N-terminal methionine.

A small second residue, such as alanine or leucine, results in cleavage of methionine and acetylation of this second residue. However, if the second residue is large or bulky, the N-terminal methionine is preserved and acetylated directly. These proteins are classed as Met-Φ and can be degraded by either the Ac/N-end Rule or, when not yet acetylated, by the Arg/N-end Rule as a level of redundancy to allow quick degradation

34 of misfolded proteins. This non-acetylated Met-Φ turnover depends on components of the Type II Arg/N-end Rule pathway, presumably binding to the second, bulky residue with a similar conformation as to the bulky hydrophobic Type II N-degron (J.

M. Kim and Hwang 2014; Nguyen et al. 2019).

1.1.2.2 Type I or Type II N-terminal Modifications

One of the first discovered, and most direct, ways in which N-terminal modification acts to generate an N-degron signal for the N-end Rule is through arginylation. This a hierarchical process, beginning with the cleavage of the N-terminal methionine and exposure of the second position residue, such as asparagine, glutamine and cysteine, which are tertiary N-terminal residues. These are modified by a group of amidase enzymes to generate aspartic acid, glutamic acid, or oxidised to generate CysO2(H) or CysO3(H) – which are substrates of the arginyltransferase ATE1 (Takafumi Tasaki and Kwon 2007; Yong Tae Kwon, Kashina, and Varshavsky 1999; Takafumi Tasaki et al. 2012; R.-G. Hu et al. 2005; M. J. Lee et al. 2005). This enzyme conjugates an arginine residue on the N-terminal end to generate an arginine Type I N-terminal degron sequence.

Other Type I, or Type II, N-terminal sequences can arise through other cleavage or N- terminal modifications. For example, there are also some roles for the N-end Rule in specific targeting of substrates under particular cellular conditions, and this can be achieved by generation of a neo-N-terminus through proteolysis. These include N- end Rule degradation of the cohesin kleisin fragments during mitosis of S. cerevisiae, which are cleaved by separase upon removal from chromatin during chromosome segregation (Kim Nasmyth 2011). The neo-N-terminus generated on the C-fragment cleavage product bears an N-terminal arginine, so being a Type I substrate for the N-

35 end Rule pathway (Rao et al. 2001). Interestingly, the meiotic REC8 kleisin is also cleaved by separase in this way, but this undergoes N-terminal acetylation and degradation by the Ac/N-end Rule pathway (Y. J. Liu et al. 2016). Other recently described roles for the N-end Rule in conditional protein regulation include the innate immune response, with the cleavage of the inflammasome component NLRP1B to generate a neo-N-terminal Type II degron, and DNA replication stress regulators such as SDE2, which has a neo-Type I N-degron upon stress-induced cleavage (H. Xu et al. 2019; Rageul et al. 2019). Another role for the N-end Rule has been identified in the regulation of hedgehog (Hh) signalling pathways in both D. melanogaster and mammalian systems, potentially through degradation of an N-terminally extended

CSPP1-L ciliary protein isoform, which encodes an N-degron sequence (Kinsella et al.

2016; Shearer et al. 2018). Ciliary function is vital for developmental patterning and morphogen signalling, including via the Hh pathway (Frikstad et al. 2019; He, Agbu, and Anderson 2017).

Figure 1.2 Phylogeny of the UBR N-end Rule E3 Family Phylogenetic tree of UBR proteins across eukaryotic systems. The RING family (yellow) is the largest, encompassing the originally discovered UBR protein, S. cerevisiae, and various UBR1, UBR2 and UBR3 proteins. UBR3 has the highest sequence homology with S. cerevisiae UBR2, but this does not have N-end Rule activity and so is not shown here. UBR5 (orange) is the only UBR protein to contain a HECT domain, and together with the RING family, are the only UBR proteins that act in canonical E2-E3 ubiquitination interactions through their domain structures. UBR4, UBR6 and UBR7 have all evolved separately from the other UBR proteins and are only related through their shared UBR box domain structure. Schematic taken from (T Tasaki et al. 2005). sc: S. cerevisiae, sp: S. pombe, m: M. musculus, d: D. melanogaster, c: C. elegans, a: A. thaliana

1.1.2.3 Evolutionary Conservation of the UBR Protein Family

Recognition of Type I or Type II N-end Rule substrates in mammalian systems is carried out by a domain called the UBR box, which is conserved between a family of seven E3 ubiquitin ligases called the UBR proteins (T Tasaki et al. 2005). Only four mammalian UBR proteins have been shown to actively bind substrates using the N- end Rule protein recognition system: UBR1, UBR2, UBR4 and UBR5 (T Tasaki et al.

2009).

The UBR proteins are highly conserved across different species (Fig. 1.2), and UBR proteins were the first N-end Rule system discovered, with the S. cerevisiae gene UBR1

(Bartel, Wünning, and Varshavsky 1990). This encodes UBR1 protein, which contains the highly conserved UBR box found in mammalian UBR proteins.

S. cerevisiae UBR1 is a homolog of mammalian genes Ubr1 and Ubr2, which are 47% sequence conserved, with 68% sequence similarity to each other (T Tasaki et al. 2005).

They each contain a UBR box, an N-domain, a RING domain which carries out binding to the E2 ubiquitin conjugating enzyme and a UAIN (UBR auto-inhibitory domain), Fig. 1.3. The S. cerevisiae UBR1 protein has substrate binding abilities in both the Type I and Type II N-end Rule pathways, as well as to an internal degron site in the transcriptional repressor CUP9 (G. C. Turner, Du, and Varshavsky 2000; Du et al.

2002). UBR1 recognition of CUP9 depends on the C-terminus proximal UAIN, which has also been noted in some literature as region IV (Zenker et al. 2005). The domain carries out auto-inhibition in S. cerevisiae UBR1, whereby it acts to bind the N-terminal region of the Ubr1 protein, blocking binding of UBR1 to CUP9 (Du et al. 2002) CUP9 functions as a transcriptional repressor of peptide import proteins such as PTR2

(Kawai et al. 2014), and so the addition of N-end Rule peptides acts to promote CUP9 degradation, so positively regulating their own uptake (G. C. Turner, Du, and

Varshavsky 2000). CUP9 has an internal degron sequence, which has been limited to

38 the C-terminal 33-residue region. However, the recognition domain for CUP9 within

UBR1 has not been identified, and so it is not understood how the UAIN acts to physically block CUP9 interaction (Du et al. 2002; Xia et al. 2008) (Du et al., 2002;

Xia et al., 2008). Although the protein sequence is conserved between species, study of the UAIN domain has not so far been published in mammalian UBR proteins.

Other conserved UBR proteins across different species include the fission yeast S. pombe proteins Ubr1 and Ubr11. These have high sequence and structural similarity to

S. cerevisiae Ubr1, but only S. pombe Ubr11 has any documented activity in the N-end

Rule (Fujiwara et al. 2013). Ubr11 is only required for Type II N-end Rule functions, and so it remains to be seen if a Type I N-end Rule E3 ligase exists in this species

(Kitamura 2014).

Ubr11 also appears to carry out some non-N-end Rule functions that are conserved with S. cerevisiae UBR1. For example, it is required for PTR2 expression, which is a peptide uptake transporter, and this depends on a functional UAIN domain, as in S. cerevisiae UBR1 (Kitamura and Fujiwara 2013). UBR1 is required in budding yeast for

PTR2 expression through degradation of CUP9 transcriptional repressor (Du et al.

2002), but there are as yet no known CUP9 homologs in fission yeast and so the Ubr11

N-end Rule substrate for this function in S. pombe is unknown (Kitamura and Fujiwara

2013; Kitamura 2014). Ubr11 is also required for degradation of misfolded kinetochore components via Hsp70 chaperone-assisted proteasomal degradation, so carrying out nuclear quality control and homeostasis (Kriegenburg et al. 2014; Kitamura 2014).

S. pombe Ubr1 is functional as an E3 ubiquitin ligase, acting to degrade hypoxic transcription factor Sre1 (C.-Y. S. Lee et al. 2011) and to regulate proteasomal nuclear localisation by turnover of Cut8 nuclear envelope protein (Takeda and Yanagida

2005). However, neither of these substrates are confirmed to be N-end Rule

39 substrates as both lack N-terminal degron sequences, and therefore S. pombe Ubr1 may not be active in the N-end Rule at all.

There are very limited studies on non-N-end Rule mammalian UPS functions for

UBR1 or UBR2. Studies in our group have shown that UBR2 is able to turn over the retrotransposon-expressed protein LINE-1 Orf1 through an unknown mechanism – presumably by binding to an internal degron, similarly to the CUP9 substrate of S. cerevisiae UBR1, as Orf1 does not contain a canonical N-degron sequence. The physical interaction domains between UBR2 and Orf1 have not been identified (Maclennan et al. 2017), and it is unknown whether this depends on activity of the UAIN domain, or if binding to N-end Rule sequences are required for LINE-1 Orf1 binding.

UBR1 is the most well-studied of the mammalian UBR family and has functions in a variety of roles; a Ubr1-/- mouse is both viable and fertile, but has severe growth defects in skeletal and adipose tissues (Y. T. Kwon et al. 2001). UBR2 has specific roles in male meiotic progression and female embryonic survival (Y T Kwon et al. 2003; Yang et al. 2010), and has been implicated in the regulation of the mitotic cohesin complex

(J Reichmann et al. 2017). Therefore, although they are highly conserved, UBR1 and

UBR2 do appear to have distinct roles within a mammalian system. However, there is redundancy in some of these functions, and a double null mouse is inviable (Jee

Young An et al. 2006); embryos do not progress past midgestation, around E11.5.

UBR1 has been shown to bind both Type I and Type II N-terminal degron peptide sequences in a pulldown assay, but a Ubr1-/- N-end Rule reporter cell line has only a partial rescue of both classes of reporters. UBR2 specifically binds Type II peptides, and a double null Ubr1-/-Ubr2-/- reporter cell line has a slightly stronger rescue, as compared to Ubr1-/- alone (T Tasaki et al. 2005). However, these cell lines still exhibit some N-end Rule activity, indicating that other N-end Rule E3 ubiquitin ligases may be present in these cells – potentially UBR4 or UBR5.

Mammalian UBR4 (also known as p600) is a very large 570kDa protein (Fig. 1.3) which shares highest sequence and domain homology with D. melanogaster protein

PUSHOVER (or d-UBR4) and A. thaliana protein BIG (a-UBR4, Fig. 1.2). There is no known S. cerevisiae homolog, although all three homologs do share the conserved UBR box domain with S. cerevisiae UBR1 (T Tasaki et al. 2005). Genetic null experiments in mouse cell lines show that mammalian UBR4 is essential for a wide variety of functions across the mouse. These include oncogenic upregulation and increased tumour invasiveness, co-option in viral infection for innate immune response suppression, both placental and cardiovascular embryonic development, and endosome maturation in lysosomal proteolysis (S. T. Kim et al. 2018; Nakaya et al.

2013; Sakai et al. 2011; Morrison et al. 2013). Only the latter is directly linked to proteolysis, although not through proteasomal degradation. Endosome maturation requires membrane bound UBR4 and it is currently not known whether this acts through ubiquitination. Neither BIG or PUSHOVER have any documented activity in the UPS in their respective systems (Gil et al. 2001; Richards, Hillman, and Stern

1996; Hearn et al. 2018), and there are no previously identified ubiquitination or E2 binding domains within the mammalian Ubr4 gene. However, mammalian UBR4 isolated from cell lysates is able to bind both Type I and Type II N-end Rule peptide substrates in an in vitro peptide pulldown assay, and knockdown of UBR4 in Ubr1-/-

Ubr2-/- double null cells impairs degradation of N-end Rule reporters, which was partially preserved in the double null alone (T Tasaki et al. 2005), strongly implying that UBR4 does have some activity in the N-end Rule UPS, although in vivo substrates have not been identified.

Before identification of the conserved UBR box, UBR5 was studied as an E3 ubiquitin ligase (called EDD/hHYD), active in, among other functions, proteasomal degradation of topoisomerase TopBP1, leading to cell cycle regulation. This acts via its HECT E2

41 binding domain, which recognises the protein’s C-terminal BRCA1 domain, i.e.

TopBP1 is not an N-end Rule substrate (Honda et al. 2002; Henderson et al. 2002;

Callaghan et al. 1998), Fig. 1.3. UBR5 is a homolog of the D. melanogaster Hyd

(hyperplastic discs protein, or d-UBR5, Fig. 1.2) which is a tumour suppressor and E3 ubiquitin ligase active in regulation of signalling morphogen Hh (Hedgehog) (G.

Wang et al. 2014; Kinsella et al. 2016). Again, there is no known S. cerevisiae homolog for mammalian UBR5, except for the sequence similarity of the UBR box substrate binding domain to Ubr1 (T Tasaki et al. 2005). Within the N-end Rule, UBR5 has only been shown to be active in Type I N-degron binding (T Tasaki et al. 2009).

It should be noted that the S. cerevisiae genome also encodes a UBR2 gene, which is unrelated to the mammalian Ubr2 gene, instead having sequence similarity to the mammalian Ubr3 gene; these are thought to be functional homologs (Takafumi Tasaki et al. 2012). Neither has activity in the N-end Rule (T Tasaki et al. 2005), although they both contain the sequence-conserved UBR box domain (Fig. 1.3). S. cerevisiae

UBR2 is functional in a complex with the protein MUB1 as a RING-type E3 ubiquitin ligase via non-N-end Rule pathways, for example in kinetochore protein regulation and homeostasis of proteasomal protein transcriptional abundance (Akiyoshi et al.

2013; L. Wang et al. 2004). The mammalian UBR3 protein is required for genomic stability and regulates the degradation of APE1, a DNA repair protein, again via a non-

N-end Rule pathway of the UPS. UBR3 binds E2 ubiquitin conjugating enzymes via a RING domain (T Tasaki et al. 2005), as does S. cerevisiae UBR2. Not much is known about the substrate binding domains present within the mammalian UBR3 protein

(Meisenberg et al. 2012), but this is presumably through an internal degron. Although the UBR box has been identified primarily in N-end Rule degron recognition, it is structurally very similar to RING-like zinc finger domains, which are able to carry out a wide variety of functions including DNA-protein and protein-protein

42 interactions, being implicated in ubiquitin-recognition or protein-recognition (Kaur and Subramanian 2015; Freemont 1993). Therefore, the UBR domain could be active in non-N-end Rule activities in the non-N-end Rule UBR proteins such as UBR3 and

S. cerevisiae UBR2, for example through recognition of non-N-terminal degron regions of substrate proteins.

UBR6 and UBR7 are also mammalian proteins that contain a UBR box domain but have no documented activity in the N-end Rule. The only other characterised domains within these proteins are an F-box (UBR6) or a PHD (plant homeodomain,

UBR7) (T Tasaki et al. 2005), see Fig. 1.2 and Fig. 1.3. These are both potential E2 interaction domains that would be required for activity within the N-end Rule and/or the UPS. However, neither of these genes have confirmed activity either in N-terminal degron binding or in ubiquitination (Juliana Muñoz-Escobar, Kozlov, and Gehring

2017; Adhikary et al. 2019; Zimmerman et al. 2014) and so cannot be assumed to act as N-end Rule E3 ubiquitin ligases.

Finally, recent research has found proteins with a UBR-like ZZ domain that appear to act in the N-end Rule, binding the N-terminus of target proteins and directing them for degradation via the autophagy-lysosome pathway. For example, p62 (also known as SQSTM1) is a mammalian E3 ubiquitin ligase that binds the arginylated N- terminus of ER (endoplasmic reticulum) chaperones such as BiP via its ZZ-domain

(Cha-Molstad et al. 2015). Through this pathway, Arg-BiP mediates the degradation of bound, misfolded proteins.

1.1.2.4 Substrate Binding Specificity in the Arg/N-end Rule

Substrate binding specificity for the UBR proteins that act as N-end Rule E3 ubiquitin ligases is regulated by two protein domains: the UBR box, a highly

43 conserved 70-residue zinc-finger-like motif which is shared across all UBR proteins

(E Matta-Camacho et al. 2010), and the N-domain, which has high homology with the bacterial ClpS domain and is common to both mouse proteins UBR2 and UBR1.

The UBR box alone is able to bind Type I, or arginine-containing, target peptides, but both this and the N-domain are required for Type II, or phenylalanine-containing, peptide binding by UBR1 and UBR2 (T Tasaki et al. 2009).

There is also evidence to suggest that, at least UBR1, is also able to bind to non- acetylated Met-Φ proteins, aiding in their degradation as a redundant role for the degradation of acetylated Met-Φ proteins by the Ac-N-end Rule pathway (J. M. Kim and Hwang 2014). This requires both the UBR box and the N-domain and, given that the second residue is often bulky, for example tryptophan or phenylalanine, appears to be most related to canonical Type II N-end Rule degradation pathways used by

UBR proteins.

Although UBR2 contains both a UBR box and an N-domain, the two protein domains required for both Type I (basic) and Type II (bulky hydrophobic/aromatic) substrate binding, it has strongest detectable binding to Type II substrates (T Tasaki et al.

2005). As an isolated protein domain, the UBR2 UBR box is able to bind Type I peptide substrate sequences in vitro, to a similar affinity as that observed for the isolated UBR1 UBR box (E Matta-Camacho et al. 2010). However, this affinity appears to be much reduced when the whole UBR2 protein is present (T Tasaki et al.

2005), indicating a potential structural conformation that favours Type II substrate binding. This could alternatively indicate a level of competitive binding between UBR proteins, as UBR1, UBR4 and UBR5 are all also present in these in vitro binding assays.

It is possible that in some contexts UBR2 also binds Type I substrates in vivo.

However, UBR1, UBR4 and UBR5 are all strongly detectable in Type I binding assays, are ubiquitously coexpressed in a large majority of tissues or cell types. Therefore, in

44 the majority of contexts, these are likely the primary Type I regulators of the N-end

Rule pathway.

These four UBR proteins, which all have similarities in their recognised substrate degron sites, are likely to exhibit some redundancy in their substrate binding ability.

This has been functionally demonstrated between the two most similar UBR proteins:

UBR1 and UBR2. Single null mutant mice for Ubr2-/- exhibit some developmental defects including a variably penetrant female-specific embryonic lethality (Y T Kwon et al. 2003). A Ubr1-/- mouse is viable but exhibits growth defects particularly in the skeletal or adipose tissue (Y. T. Kwon et al. 2001). However, in a double null mouse there is a more penetrant embryonic lethality, associated with defects in neurogenesis and cardiovascular development that are not linked to either of the single mutant phenotypes. This indicates one is able to compensate for the function of the other in multiple developmental pathways (Jee Young An et al. 2006).

An example of redundancy outside these two highly conserved proteins is in the knockdown of Ubr4 by RNAi in cultured fibroblasts that already exhibit a Ubr1-/-

Ubr2-/- phenotype. Alone, the double mutant has a slight increase in both Type I and

Type II N-end Rule reporter proteins, but additional knockdown of Ubr4 results in a loss of Type II N-end Rule turnover, although another component of the Type I N- end Rule persisted in these cells – potentially UBR5. This indicates that they at least partially compensate for turnover of the same substrates, but as this is a reporter system there is no in vivo evidence that UBR1/UBR2 have overlapping biological

substrates with UBR4 (T Tasaki et al. 2005).

kDa 1 - 1722 200 UBR1

200

213

570

300

Figure 1.3 The Mammalian UBR E3 Family The mammalian UBR family are a group of mammalian proteins that all contain the conserved UBR box (red), which undergoes direct interaction with the N-terminus of target proteins. Within the family, there is a high level of variation in the protein size, E2 binding and other domains. UBR1 and UBR2 are the most highly conserved within this group, each 200kDa and containing an N-domain (pale blue) for Type II N-end Rule activity. They each contain a RING domain for E2 binding activity (green-blue), and a UAIN (UBR auto-inhibitory domain, yellow). UBR3 has no confirmed activity as an N-end Rule ubiquitin ligase, instead acting through internal degron binding, but contains a UBR box, RING domain and UAIN. UBR4 is the largest of the UBR proteins, at 570kDa, and contains a CRD (cysteine-rich domain, orange). UBR5 is the final UBR protein known to be active in the N-end Rule, specifically in Type I binding. It contains the UBR box for Type I binding, and has a HECT E2 binding and ubiquitin transfer domain (purple). UBR6 and UBR7 are the smallest of the UBR family and neither have any documented roles in the N-end Rule or in ubiquitination. Schematic adapted from (T Tasaki et al. 2009).

1.1.3 E2/E3 Interaction Dynamics and the E3 RING Domain

There are three major classes for E3s: those containing either the RING, HECT or RBR domains. These domains determine specificity to the E2 ubiquitin-conjugating enzyme and dictate the mechanism by which ubiquitin is transferred to the target protein. E3 ligases within the UBR family, although they are conserved in their substrate binding domain, are variable in this instance and so are expected to interact with a varied array of E2 enzymes.

UBR5 contains a HECT domain, which undergoes temporary ubiquitination as a proximal transfer point for build-up of ubiquitin chains before transfer to the substrate (M. Wang and

Pickart 2005; Sluimer and Distel 2018). Additionally, UBR5 does not undergo canonical interaction with its associated E2 enzyme UBCH4 and is able to carry out E2-independent ubiquitin-transfer (Edna Matta-Camacho et al. 2012; Shen et al. 2018). There are no known

UBR proteins with an RBR domain, which is a highly specialised E2 binding domain that is also able to catalyse ubiquitin transfer in an E2-independent manner (Spratt, Walden, and

Shaw 2014).

UBR1 and UBR2 both contain a RING domain, which interacts with E2 enzymes through sequence-specific contacts at the RING domain surface (Deshaies and Joazeiro 2009). E2-E3 interactions are highly variable, but there are three known E2 proteins that have been shown to interact with UBR2 and act in its ubiquitination of substrates: UBE2A/B, USE1 (C. W. P.

Lee et al. 2011) and UBE2O (H. Xu et al. 2019). The coding regions of human UBE2A and UBE2B paralogues share 80% identity and the proteins have 96% amino acid identity (Koken et al.

1991); they are distinctly expressed and have different but overlapping functions in the mouse

(Roest et al. 2004, 1996), but are biochemically indistinguishable and so are assumed to be functionally indistinct. They are both documented in complex with E1 ubiquitin activator

UBA1 (Y T Kwon et al. 2003; C. W. P. Lee et al. 2011). USE1 is an E2 known to participate in a ubiquitination cascade with the E1 ubiquitin activator UBA6 (J. Jin et al. 2007; Groettrup et al. 2008). This E1 is present in mammalian cells, but not many lower eukaryotes such as S.

47 cerevisiae, and so the UBA6-USE1 cascade is not as well studied as the other major E1 protein

UBA1. Finally, UBE2O was most recently identified as a potential interactor of UBR2, both being required for the N-end Rule-mediated proteolysis of the NLRP1B inflammasome (H. Xu et al. 2019). UBE2O is an E2/E3 hybrid protein that has been implicated in the turnover of orphan proteins that are misfolded as a result of loss of their protein complex (Yanagitani,

Juszkiewicz, and Hegde 2017). This function appears to be independent of any canonical E3 ligase, but it appears that UBE2O is able to carry out the role of an E2 alongside the UBR2 E3 as well.

There are 300 human genes encoding RING domain-containing proteins, most of which are expected to be at least partially active as E3 ubiquitin ligases (Wei Li et al. 2008). In S. cerevisiae, there are around 50 RING-encoding genes (Finley et al. 2012), and this increase in absolute gene number between species is generally attributed to evolutionary conservation and duplication, generating families of orthologous proteins that have overlapping functions

(Deshaies and Joazeiro 2009). This gives a level of assumed redundancy within higher eukaryotes which indicates the importance of the UPS for proteome homeostasis and regulation in mammalian systems and could be extended to the existence of both UBR1 and

UBR2 as mammalian homologs of S. cerevisiae UBR1.The canonical RING domain is a cysteine- rich protein domain with conserved histidine residues that coordinate two zinc molecules in the core of their structure. This is similar, but structurally and functionally distinct, to the zinc finger DNA-binding structure (Freemont 1993; Freemont, Hanson, and Trowsdale 1991).

Outwith these core residues, the intervening amino acids in the RING domain sequence can be variable, although there are several conserved residues across a variety of RING domain containing proteins that, through structural and mutational studies, have been shown to be important for physical contacts to the E2 ubiquitin ligase structures. For example, bulky hydrophobic residue, such as W408 (Tryptophan) in c-Cbl, is often found on the exposed

48 surface of the RING domain and the sidechain of this residue sits in the binding cleft of the

E2 enzyme (C. W. P. Lee et al. 2011; Deshaies and Joazeiro 2009).

E2-E3 interaction tends to be highly dynamic and transient, allowing quick replacement with newly ubiquitin-charged E2 enzymes, and efficient extension of the substrate ubiquitin chain.

E3 enzymes often interact with more than one E2, which could depend on the substrate or the type of ubiquitination chain being constructed. Therefore, it is difficult to reliably predict

E2-E3 pairs. In addition, it is generally thought that E2 enzymes will only bind to the RING domain after thioesterification and ubiquitin-charging, and that the uncharged E2 competitively binds between the E1 enzyme and the E3 RING domain. Binding of the RING domain to the E2-Ub thioester complex promotes discharge of the Ub to an acceptor – i.e. lysine residues in the target protein, as well as other ubiquitin molecules. The E2 and the substrate for ubiquitination in complex with an E3 ligase are generally far apart when in complex with the RING E3 enzyme. Some RING proteins require additional modifications to stabilise the E2 interaction and promote ubiquitin transfer across this distance, for example linkage proteins (J. Liu and Nussinov 2010) or even autoubiquitination of either the E2 or E3 proteins themselves (Patel, Sibbet, and Huang 2019).

Given the highly variable dynamics of E2-E3 interactions, this gives an even higher level of variability in the UPS than even conferred by the large number and diversity of ubiquitin ligase-encoding genes (W Li and Ye 2008). Therefore, caution must be taken in the study of

RING-E3 and E2 protein pairs. Introduction of specific mutations into the RING domain in an attempt to prevent binding to E2’s and generate a functionally “catalytically dead” E3 protein, must take into account potential variation in binding conformation and at least some level of ubiquitination activity.

1.1.4 The 26S Proteasome

After substrate proteins are tagged via the E1-E2-E3 cascade of ubiquitination, they are targeted to the proteasome for degradation. This is a 26S protein complex that acts as the major eukaryotic protease to degrade proteins tagged with ubiquitin in both the cytosol and the nucleus. It is an AAA+ ATPase motor and acts to unravel the higher order structure of target proteins through ATP hydrolysis. It has very low selectivity, acting to deubiquitinate and degrade a wide variety of eukaryotic proteins that have been polyubiquitinated (Bedford et al. 2010).

The complex consists of a 20S hollow core, a base complex, and a lid complex. The base and lid together are a 19S protein complex and act as the ubiquitin-recognition site (Tanaka

2009). Within this complex, subunits Rpn10, Rpn13 or Sem1 act to bind ubiquitin-conjugated proteins, and are the first binding point for ubiquitinated substrates (Shi et al. 2016;

Rosenzweig et al. 2012; Hamazaki, Hirayama, and Murata 2015; Kragelund et al. 2016;

Paraskevopoulos et al. 2014). These intrinsic proteasomal ubiquitin receptors do not exhibit exclusive recognition for particular ubiquitin chains as a signal for degradation at the proteasome, although there is a bias towards K48-ubiquitin chains (Tsuchiya et al. 2017).

The Rpn11 subunit is an example of the stably associated deubiquitinases (DUBs) in the proteasome, which act to remove the poly-ubiquitin chain from substrates and allow unfolding and translocation of the substrate itself into the proteasome core. In addition to ubiquitin chains, substrates also require an unstructured initiation region which can act as a starting point for unfolding in the 20S core. Components of the proteasome avoid degradation while close to the AAA+ ATPase core by lacking this unstructured or flexible region (Bedford et al. 2010).

1.1.4.1 Regulatory Functions of Proteasomal Degradation

Given the high specificity of E3 substrate binding, the proteasomal protein degradation pathway can act to specifically degrade proteins that are required to be short-lived, for example in response to cellular conditions. One such example of this is the necessity for proteasomal activity in NLRP1 inflammasome activation in the innate immune response, which is a response to bacterial infection by B. anthracis and subsequent exposure to anthrax lethal toxin (LT) (Sandstrom et al. 2019; Chui et al.

2019). In fact, the N-end Rule E3 ubiquitin ligase UBR2 was recently identified in

NLRP1 inflammasome activation (H. Xu et al. 2019). LT exposure in mammalian macrophage cells induces cleavage of NLRP1B, generating a neo-Type II N-degron sequence on the C-terminal fragment. Caspase-1, which stimulates the immune response in these cells, is activated by the cleaved N-terminal fragment of NLRP1B, and this response depends on degradation of the C-terminal fragment by the N-end

Rule. siRNA knockdown of either UBR2 or the E2 ubiquitin conjugating enzyme

UBE2O results in a vulnerability of macrophages to the B. anthracis infection and a loss of the immune response, implying that UBR2 specifically binds the C-terminal fragment for proteasomal degradation.

Another recently identified role of the Arg/N-end Rule in a regulatory function of proteasomal degradation is the conditional degradation of the genome surveillance protein SDE2 (Rageul et al. 2019). This protein is endolytically cleaved in response to

DNA replication stress in order to negatively regulate monoubiquitination of PCNA

(proliferating cell nuclear antigen), which is a cell stress response that prevents progression of DNA replication forks. Thus, SDE2 acts to protect against cellular stress and promotes progression of DNA replication and S-phase (Jo et al. 2016). After cleavage of SDE2, the neo-N-terminus of the C-fragment encodes a Type I N-degron which has been shown to be turned over through proteolysis. Stable siRNA

51 knockdown of either N-end Rule E3-encoding genes Ubr1 or Ubr2 results in a significant stabilization of the SDE2 C-terminal fragment, and both of these are required for the DNA stress response to UV irradiation. UBR1 and UBR2 are redundant in this capacity, and both must be removed to fully stabilise the C-terminal

SDE2 protein (Rageul et al. 2019). The activity of the Type I Arg/N-end Rule in this capacity acts alongside CRL4CDT2 E3 ubiquitin ligase under stressed conditions, adding another layer of redundancy to an essential cellular process (Jo et al. 2016).

Interestingly, this is an in vivo Type I function of UBR2, although peptide pulldown assays do not show a particular affinity for Type I peptide sequences by UBR2 protein

(T Tasaki et al. 2005). This shows that there are in vivo roles for proteins such as UBR2 that cannot necessarily be predicted by in vitro pulldown or binding assays.t

1.1.4.2 Misfolded or Orphan Protein Turnover

As described, degradation of specific proteins can be important for the functioning of pathways under various conditions in the cell, but a much more diverse function of the UPS is in the recycling of misfolded or denatured proteins. This acts in partnership with the lysosomal mode of protein degradation, which in itself is vital for turnover of proteins (Jackson and Hewitt 2016). Partial denaturation of cellular proteins can be caused by a variety of conditions – either in the initial folding process, which has been estimated to result in misfolding errors from 5-30% of the time (U.

Schubert et al. 2000; Hartl and Hayer-Hartl 2002), or in destabilisation of complex components leading to orphan proteins that have exposed and vulnerable interaction sites (Juszkiewicz and Hegde 2018). Defects in either of these pathways has been associated with a large variety of disorders, including aggregation of misfolded proteins as in Parkinson’s or Alzheimer’s disease (Q. Zheng et al. 2016), and degradation of misfolded but still healthy proteins, such as the binding of CHIP

52 chaperones to immature CFTR in some mutational backgrounds of cystic fibrosis

(Meacham et al. 2001).

Misfolded proteins are targeted to the proteasome in a two-step process. First, molecular chaperones act to protect exposed areas of misfolded or orphan proteins, allowing them to refold without aggregation. At this point, chaperones can either commit to refolding, or degradation of the protein in question. One of the determining factors at this step is the co-chaperone protein CHIP, which has E3 ubiquitin ligase activity and binds to Hsp70 or Hsp90 (heat shock proteins, known protein folding chaperones) in order to target bound proteins to the UPS, recycling amino acid components for new protein synthesis (Marques et al. 2006).

However, in addition to the activity of E3 ligase CHIP, there is some implication for the N-end Rule in these pathways, which have historically been associated with sequence-specific binding at the N-terminus of target proteins (Tasaki et al., 2012, and section 1.1.2). As discussed in section 1.1.2.1, UBR proteins have a recently identified role as a compensatory pathway for the Ac/N-end Rule to turnover misfolded proteins, both in S. cerevisiae (Eisele 2008) and in mammalian systems (J. M. Kim and

Hwang 2014; Nguyen et al. 2019; Shemorry, Hwang, and Varshavsky 2013).

Outside of the N-end Rule, a non-canonical role in protein quality control has been identified for S. cerevisiae Ubr1 (Heck, Cheung, and Hampton 2010; Nillegoda et al.

2010; Eisele and Wolf 2008), as well as S. pombe Ubr11 (Kriegenburg et al. 2014). In both species, this is dependent on the interaction of the protein chaperone Hsp70 to misfolded proteins in the cytoplasm (Summers et al. 2013; Kriegenburg et al. 2014), similarly to mammalian CHIP E3 (Marques et al. 2006). No such pathway has thus far been identified for mammalian UBR proteins, but it should be noted that CHIP E3 ubiquitin ligase is not encoded in the S. cerevisiae genome (Heck, Cheung, and

Hampton 2010), and so UBR proteins may not be required for non-N-end Rule

chaperone-assisted cytoplasmic proteostasis in the mammalian system.

Additionally, the ZZ-domain N-end-like pathway is implicated in turnover of

misfolded proteins via the autophagy-lysosome pathway, potentially through K63

ubiquitination; p62 binds the arginylated N-terminus of ER chaperone BiP to mediate

turnover of misfolded ER proteins (Cha-Molstad et al. 2015).

1.1.5 Non-Proteasomal Functions of Ubiquitination

Although the best studied function of ubiquitination is in the turnover of proteins tagged with polyubiquitin chains, there are a variety of different ubiquitin configurations that encode different signaling functions in a eukaryotic cell. While a straight polyubiquitin chain is made up of a series of ubiquitin proteins typically ubiquitinated at the K48 residue, K33, K29, or

K63 are all also common ubiquitination sites on the ubiquitin molecule and this indicates roles for a variety of differently structured polyubiquitin chains (W Li and Ye 2008). K63 ubiquitin chains have been implicated in induction of autophagy through recognition by p62

(Seibenhener et al. 2004; Pankiv et al. 2007) and therefore an alternative protein degradation pathway through the autophagosome, which can be active in degradation of protein aggregates that are too large for regular proteolytic activity (Reggiori and Klionsky 2005).

K29 and K33 ubiquitin-linkages have been implicated in specialised polyubiquitin chains in the regulation of AMPK kinase activity, in balance with the deubiquitination enzyme USP9X

(Al-Hakim et al. 2008).

Aside from variants in polyubiquitin chains, monoubiquitination and multiubiquitination are the conjugation of single ubiquitin molecules across the surface of the substrate. These have been generally associated with regulation of protein interactions, acting to promote binding to proteins containing ubiquitin-binding domains (Hicke 2001). Monoubiquitination is

54 sufficient to promote internalization and lysosomal endocytosis of ubiquitinated membrane- bound substrate RTKs (receptor tyrosine kinases) (Haglund et al. 2003).

The specific combination of substrate protein, E2 ubiquitin conjugating enzyme and E3 ubiquitin ligase confers the ubiquitin modification that is carried out. RING-type E3 ubiquitin ligases bind to a variety of E2’s and these determine the ubiquitin chain formation

(Christensen, Brzovic, and Klevit 2007; H. T. Kim et al. 2007). HECT-type E3’s however usually confer Ub chain linkages directly via a variable N-terminal region of the domain (M.

Wang and Pickart 2005). There is also evidence to suggest that sequence-specific interactions take place between the catalytic core of the E2 and the surrounding residues in either the substrate or the ubiquitin molecule being ubiquitinated, which regulates the length of the ubiquitin chain and depends on both the substrate and the E2 present (Sadowski et al. 2012;

Sadowski and Sarcevic 2010). Without knowing the structural conformation of their three- way transient interaction or which individual E2/E3 pair is involved in a ubiquitination reaction, it is difficult to predict which ubiquitin modification is utilised. Therefore, it is possible that individual E2 or E3 proteins are able to carry out multiple different ubiquitination activities, depending on the substrates bound and E2/E3 interaction utilised, so targeting individual substrates for any number of degradation or regulatory functions in the cell.

The N-end Rule E3 ubiquitin ligase UBR proteins are generally implicated in protein turnover, whether through N-end Rule or non-N-end Rule substrate binding pathways. This is reflected in the known ubiquitin linkages they carry out. For example, UBR2 acts alongside either E2 ubiquitin conjugating enzymes USE1 or UBE2A in order to deposit K48-linked Ub chains on RGS4/5 substrates, which have an N-terminal Type I degron sequence (C. W. P.

Lee et al. 2011; M. J. Lee et al. 2005). UBR4 and UBR5 have been shown to recognise K63-Ub chains deposited by other E3’s such as ITCH, and deposit Ub at K48 in order to generate

55 forked Ub chains, which have also been associated with proteasomal degradation (Ohtake et al. 2018).

Additionally, UBR2 has been implicated in a regulatory role for ubiquitination of histones in meiotic spermatocytes, by monoubiquitination of histone H2A. In somatic cells, this is mediated by Polycomb complex component and E3 ubiquitin ligase RING1B (D. E. Cohen and

Lee 2002) and is a repressive chromatin mark which is important for X-inactivation or transcriptional silencing (D. E. Cohen and Lee 2002). The sex chromosomes of spermatocytes are only synapsed at the PAR during prophase I, and so they must be transcriptionally silenced in what is called meiotic sex chromosome inactivation (MSCI) (J. M. A. Turner 2007).

This is mediated by monoubiquitination at H2A, but instead of RING1B, UBR2 localises to the XY body in the pachytene nucleus, and acts to ubiquitinate H2A alongside the E2 ubiquitin conjugating enzyme HR6B (J Y An et al. 2010). There is no clear N-terminal degron sequence in histone H2A, and so it has been hypothesised that this acts through either non- canonical internal degron binding, potentially through the UAIN domain, or as-yet undefined

N-end Rule related pathways.

1.2 UBR2 and TEX19.1

The mammalian UBR family is highly diverse, only united by the common UBR box domain that is able to physically bind to the N-terminus of target proteins for substrate ubiquitination (E Matta-Camacho et al. 2010; J Muñoz-Escobar et al. 2017). The UPS is a versatile system and covers a wide variety of proteome homeostasis functions, as well as regulation of short-lived proteins in context-specific pathways. Furthermore, the ubiquitin signalling system is much more diverse than simple canonical UPS-related degradation

(Sadowski et al. 2012; W Li and Ye 2008; Hicke 2001). These revelations have implications in the study of UBR2’s known roles across male meiotic progression, female embryonic survival

(Y T Kwon et al. 2003) or its recently uncovered role in cohesin regulation (J Reichmann et al.

2017).

The N-end Rule E3 ubiquitin ligase UBR2 is encoded by the gene Ubr2, conserved between both human and mouse genomes. Mouse Ubr2 is a 46 exon gene that encodes a 200kDa protein, which is one of four E3 ubiquitin ligase UBR proteins that have been shown to be active in the N-end Rule (T Tasaki et al. 2005), and is particularly prominent in Type II N-end Rule binding, although some Type I ability has been identified in vivo genetic studies (T Tasaki et al. 2009; Takafumi Tasaki et al. 2012; Rageul et al. 2019). A wide variety of roles for the UBR2 protein have been identified, involving degradation of N-end Rule targets such as inflammasome NLRPB1 and stress regulator SDE2 (see section 1.1.4.1 and Rageul et al., 2019;

Xu et al., 2019). These have not been identified as prominent phenotypes in a Ubr2-/- context and the latter can also be regulated by UBR1. Given that there is so much potential for redundancy between the UBR proteins in the N-end Rule, discussed in detail in section 1.1.2.4, it is interesting that a Ubr2-/- mouse has a very specific subset of phenotypes, indicating essential functions of the gene in the testis and in embryonic development (Y T Kwon et al.

2003), which could act through as-yet unidentified ubiquitination or N-end Rule substrates for UBR2.

A Ubr2-/- embryonic stem cell (ESC) line also has a phenotype, indicating a role for the N-end

Rule in cohesin regulation in these cells, which may be related to the role in cohesin regulation in the mouse (Reichmann et al. 2017), as well as in retrotransposon derepression (unpublished data, Karen Dobie, MRC HGU – discussed in more detail in Chapter 5).

Tex19.1 is a gene primarily expressed in hypomethylated tissues, including the testes, ovaries and placenta (Ollinger et al. 2008). The TEX19.1 protein interacts with UBR2 (Maclennan et al. 2017), and its expression pattern broadly overlaps the tisues which exhibit phenotypes in the Ubr2-/- mouse (Ollinger et al. 2008; Y T Kwon et al. 2003; Yang et al. 2010). Ectopic expression of Tex19.1 in a Flp-In-293 cell line results in specific protection of GFP Type II N- end Rule reporters implying that its interaction with UBR2 may antagonise UBR2 N-end

Rule activities (J Reichmann et al. 2017).

1.2.1 Similarities between Ubr2-/- and Tex19.1-/- Phenotypes

There are two major phenotypes of a Ubr2-/- mouse model that appear to confer specific roles of the UBR2 protein that cannot be compensated for by other UBR proteins. The most well characterised of these is in male meiotic progression (Yang et al. 2010), and there is also an embryonic lethality that appears to be of variable penetrance depending on the genetic background (Y T Kwon et al. 2003). These wild type processes are described in detail in sections 1.4 and 1.5, respectively.

Meiosis is a continuous process in males, occurring asynchronously throughout adult life.

Recombination and crossing over of genetic material is a vital part of this, giving rise to genetic variation and diversity in the population (Cole, Keeney, and Jasin 2012). With a lack of Ubr2, there is an asynapsis of chromosomes during prophase I, which is associated with loss of recombination foci and defects in the homology search (Crichton et al. 2017). This leads to gross spermatocyte apoptosis and male infertility, which can be characterised by gross

58 morphological changes in the seminiferous tubules of the testis, low testis weight and a lack of mature sperm (Y T Kwon et al. 2003). This phenotype is very similar to that characterised in the infertile Tex19.1-/- male, which has an asynaptic arrest and reduction in recombination rates (Crichton et al. 2017). Indeed, TEX19.1 protein is destabilised in the Ubr2-/- testis (Yang et al. 2010), indicating that their interaction is vital for normal spermatogenesis. However, it remains unclear whether the phenotype is a result of lack a functional TEX19.1-UBR2 complex, perhaps through TEX19.1’s regulation of N-end Rule activity, or if UBR2 is simply required for TEX19.1 stabilisation in this tissue, producing such a stark phenocopy between these male mice. Aside from UBR2, the UBR protein with the highest Type II N-end Rule affinity is UBR1 (T Tasaki et al. 2005), and these are expressed to similar levels in the testes

(Fig. 5.1). However, there is a difference in the cell types that express UBR1 and UBR2 in this tissue; UBR1 is highly expressed in spermatogonia, whereas UBR2 is specifically expressed in spermatocytes (Y T Kwon et al. 2003). This implies that the Ubr2-/- testis may have lost a specific Type II N-end Rule function in spermatocytes that cannot be compensated by UBR1 redundancy, resulting in the observed phenotypes.

The embryonic lethality present in Ubr2-/- mice is also similar to that observed in the Tex19.1-/- mouse. TEX19.1 is expressed in the trophoblast-derived tissues of the placenta, and in a

Tex19.1-/- there are embryonic growth restrictions and a placental defect that arises around

E12.5 days of development (J Reichmann et al. 2013; Tarabay et al. 2013). This has been linked to derepression of retrotransposons, particularly LINE-1, as well as increased expression of placental imprinted genes (Reichmann et al., 2013 and Playfoot et al. 2019, manuscript submitted, MRC HGU). We are thus far unaware of any placental defects in a Ubr2-/- context, but embryonic growth restriction and partially penetrant female-biased lethality in these mice (Y T Kwon et al. 2003) leads to my hypothesis that TEX19.1 and UBR2 are working together in placental development and support of embryonic growth.

Finally, surviving Ubr2-/- females are reported to be subfertile, but due to female-biased embryonic lethality the role for UBR2 or the N-end Rule in female meiotic processes has not been extensively studied (Y T Kwon et al. 2003). Tex19.1-/- females have a subfertility associated with chromosome missegregation and aneuploidy in the oocyte. This results from a misregulation of an acetylated subunit of the cohesin complex, SMC3, and chromosome cohesion defects (J Reichmann et al. 2017). The dynamics of the cohesin complex are discussed further in section 1.3, but in brief: acetylation of the SMC3 cohesin subunit, present in every known cohesin complex, is associated with a “cohesive” subpopulation of the complex. SMC3 acetylation is most well understood in mitosis, and it acts to recruit cohesin maintenance factor sororin (Ben-Shahar et al. 2008) to “lock” cohesin onto the chromosome axis. There are key differences in this dynamic in meiotic cells, as sororin appears to localise to chromosome axes independently of cohesin or SMC3 acetylation (Gómez et al. 2016; Jordan et al. 2017). Additionally, Tex19.1 is not expressed in the majority of mitotic cell types, indicating that it acts in a meiosis-specific AcSMC3 regulatory pathway. Regardless,

AcSMC3 appears to be required for normal chromosome segregation (J Reichmann et al. 2017), and so study of how TEX19.1 protein is able to regulate this cohesin subpopulation, potentially through its interaction with UBR2, will give insights into the importance of SMC3 acetylation in the oocyte.

1.2.2 A Model for TEX19.1-UBR2 Complex Function

Although Ubr2 is ubiquitously expressed, the only prevalent phenotypes in a Ubr2-/- mouse at least partially phenocopy the Tex19.1 phenotypes in these tissues. Therefore, the TEX19.1-

UBR2 complex appears to be important for N-end Rule functions in the testis, the placenta and, potentially, in the meiotic oocyte.

UBR2 has recently been implicated in chromosomal stability and regulation of the cohesin complex in mitotic tissues (Jee Young An et al. 2012; J Reichmann et al. 2017) and so this could imply a previously unidentified role for TEX19.1-dependent N-end Rule regulation of cohesin

60 during oogenesis. The Ubr2-/- mouse has a TEX19.1-independent role in mitotic AcSMC3 regulation – in the spleen, in absence of Ubr2, chromosome-bound AcSMC3 is specifically increased, and other cohesin subunits including total SMC3 levels are unaffected (J

Reichmann et al. 2017). This implies that UBR2 is able to negatively regulate the acetylated cohesin complex, whether this acts through direct turnover via the N-end Rule or through an indirect pathway (Fig. 1.4A). Ubr2-/- ESCs also have a specific increase in AcSMC3

(unpublished data, Karen Dobie, MRC HGU). Tex19.1 is expressed in this cell type, and a

Tex19.1-/- ESC line has a reduction in chromatin-associated AcSMC3, similarly to that observed in the Tex19.1-/- oocyte. Therefore, our group has hypothesised that TEX19.1 protein binds to

UBR2 and acts to protect AcSMC3, potentially by inhibiting its role in the N-end Rule (Fig.

1.4B), and that this interaction is also key for meiotic oocyte AcSMC3 regulation.

Spleen UBR2 AcSMC3

ESCs UBR2 AcSMC3

and

Oocytes TEX19.1 protects Type II substrates

Figure 1.4 The TEX19.1-UBR2 Complex A. In the mitotic mouse spleen, in absence of TEX19.1, the E3 ubiquitin ligase UBR2 is able to specifically (but not necessarily directly) turn over AcSMC3, i.e. the acetylated subpopulation of cohesin. B. In presence of TEX19.1, in both mitotic mESCs (mouse embryonic stem cells) and meiotic oocytes, TEX19.1 binds to UBR2 and inhibits its turnover of AcSMC3, so protecting the cohesin complex and chromosome cohesion, potentially through inhibition of the Type II N-end Rule.

1.3 Chromosome Cohesion in the Mitotic Cell Cycle

Among many other molecular functions, the N-end Rule E3 ubiquitin ligase UBR2 has been shown to play a role in genomic stability (Ouyang et al. 2006) and our lab have recently shown that UBR2 regulates cohesin complexes in the mitotic spleen (J Reichmann et al. 2017).

Sister chromatid cohesion, and the protein complex that carries it out – cohesin – is essential for genomic stability during the DNA synthesis and mitotic phases of the cell cycle (K

Nasmyth and Haering 2009). The cohesin complex also plays a major role in interphase chromatin organisation, being instrumental alongside CCCTC-binding factor (CTCF) in the formation and maintenance of loops and TADs (topologically associated domains) that have been associated with chromatin compartmentalisation and long-range transcriptional regulation (Schwarzer et al. 2017; G. Wutz et al. 2017). Misregulation of the cohesin complex in a meiotic context, particularly in oocytes, has been associated with chromosome missegregation, aneuploidy and subfertility (Jessberger 2012; Herbert et al. 2015). Little is understood about meiotic cohesin, and so insights from the study of UBR2’s role in mitotic cohesin regulation may tell us more about the regulation of the complex in a meiotic context.

1.3.1 The Cohesin Complex

Cohesins are a group of protein complexes that associate with the chromosome axes during

S phase and both mitotic and meiotic cycles, acting to hold sister chromatids together until chromosome division in mitotic anaphase or meiotic anaphase II. There are also regulatory roles for the cohesin complex during interphase of both cell cycles. Structurally, the complex is made up of three subunits that make up a tripartite ring: two SMCs (structural maintenance of chromosomes) and a kleisin. Accessory proteins such as SA1 or SA2 and other interactors also bind and regulate the complex’s behaviour (Nasmyth, 2009), see Fig. 1.5.

SMC (structural maintenance of chromosomes) proteins form the core of the cohesin complex.

These are highly conserved large 110-170kDa proteins which consist of a globular ATPase, that encompasses both the N- and C-terminal domains, a long anti-parallel coiled-coil domain and a globular hinge domain, which interacts with the hinge domain of other SMC proteins

(Haering et al. 2002). The ATPase domain has recently been shown to mediate interaction between cohesin and its host DNA molecule (Çamdere, Carlborg, and Koshland 2018;

Chapard et al. 2019; Hirano 2002), acting to entrap two strands of DNA between the kleisin and SMC ATPase heads. This activity has been implicated in the interaction of the cohesin complex with chromatin motor proteins, for cohesin translocation in the loop extrusion model of chromatin organisation (Ladurner et al., 2016; Nuebler et al., 2018). However, there are no implications for roles of the ATPase in sister chromatid cohesion itself.

There are 6 known mammalian SMC proteins, and these form the basis of three major known complexes: cohesin (SMC1/3), condensin (SMC2/4) and the SMC5/6 complex (Skibbens

2019). Within cohesin, the canonical mitotic complex is made up of SMC3 and SMC1α. SMC3 is present in all known cohesin complexes, but SMC1β can form also complexes in a meiotic context (Anderson et al. 1999).

Kleisin proteins are a highly evolutionary conserved family, and the major mammalian cohesin kleisin is RAD21, which is homologous to the S. cerevisiae kleisin SCC1 (Richardson, Morell, and Faulkner 2014). There are meiosis-specific cohesin kleisins REC8 and RAD21L, which act to form a variety of different cohesin complexes in a meiotic context (Anderson et al. 1999).

The only conserved domains between these proteins, and the kleisins involved in the condensin complex, are the N- and C-terminal interaction domains with either SMC protein, and so is has been surmised that the major function of this protein class is to close the cohesin ring and hold the complex together, so holding cohesed DNA molecules together (Richardson,

Morell, and Faulkner 2014), as well as to recruit the final component of the complex, an SA

64 protein. These differences in structure may, however, result in differences in function between the various cohesin complexes.

Figure 1.5 The Cohesin Complex Cohesin is a large multiprotein complex that is made up of two SMC proteins (SMC3 and SMC1α), which each have an ATPase head and interact through their globular hinge domains. Canonical mitotic subunits are shown in bold, and meiotic alternatives in brackets. The canonical kleisin subunit is RAD21, and this links the ATPase heads of the two SMC subunits via its N- and C-terminal ends. The kleisin recruits accessory proteins SA1 or SA2, which in turn are the binding site for WAPL, a cohesin removal factor. Binding of sororin to WAPL-cofactor PDS5A/B inhibits this removal activity. Taken from (Losada 2014).

The accessory proteins SA1 or SA2, also known as STAG1 or STAG2 (stromal antigens 1 and

2) specifically bind to the kleisin subunit (Zhang et al. 2013). These are highly functionally redundant, but are the most implicated of all cohesin complex components in oncogenic mutations associated with chromosome cohesion (J. S. Smith et al. 2016). These are differently localised during meiosis; SA1 is required for telomeric stability, and depletion of the protein results in aneuploidy during mitotic cell cycle (Kostova et al. 2009). SA2 is required for maintenance of centromeric sister chromatid cohesion, as well as chromatin structure and replication fork integrity during S-phase (Canudas and Smith 2009; Mondal et al. 2019). SA2 is also important for cohesin-mediated DNA damage repair during S/G2 phase, although both

SA1 and SA2 are necessary for the S phase DNA damage checkpoint (Kong et al. 2014). Meiotic

SA3, or STAG3 (stromal antigen 3), is specifically expressed in meiotic cells and localises along the chromosome axes during meiosis (Prieto et al. 2001). Indeed, SA3 is required for formation of the axis and synapsis of meiotic homologs during prophase I (Winters, McNicoll, and Jessberger 2014). Both SA1 and SA2-cohesin complexes are also present along the chromosome axis during meiosis (Prieto et al. 2002); the variation in different cohesin complexes during this specialised cell division indicates the highly varied roles for different cohesins.

The interaction between the cohesin complex and chromatin is thought to be topological, meaning that the ring structure doesn’t directly interact with DNA through a specific sequence (Haering et al. 2002). There are two major models for how this interaction takes place: the ring model, or the handcuff model. The ring model utilises a single cohesin ring to hold two strands of DNA in close conformation; either pairs of sister chromatids during cell division, or cis-acting DNA strands in chromatin structural regulation. The handcuff model instead suggests that each DNA strand is held by a single cohesin ring, which then interact together in a dimeric complex to bring the strands together. This dimer may be mediated through anti-parallel interaction of two RAD21 kleisin subunits, the structure of which requires direct interaction of the kleisin with either SA1 or SA2 (Zhang et al. 2008, 2013). This

‘handcuff’ model would be advantageous in adaptability of cohesin-regulated interactions of regions, for example in loop formation, without unloading and reassembly of the cohesin complex.

1.3.2 Cohesin Dynamics in the Mitotic Cell Cycle

The mammalian mitotic cell cycle is made up of four stages. G1, or the first growth phase, is followed by S phase, when DNA synthesis occurs. After this, the cell enters a second growth phase, G2, before entry into mitosis (M phase). During this, chromosomal alignment and cell division occur. The cell cycle, and the dynamics of the cohesin complex through these stages, are summarised in Fig. 1.6.

After DNA replication in S-phase, pairs of sister chromatids need to be physically connected throughout G2 and early mitosis. Correct segregation of chromosomes into balanced euploid daughter cells is one of the most important processes of the cell cycle, and so this physical connection must be maintained up until separation at anaphase of the mitotic division. This, at least in part, is mediated by the cohesin complex, which holds pairs of sister chromatids together. Cohesin is first associated with chromatin during G1, which is a process which requires ATPase activity and the heterodimeric chaperone complex NIPBL-MAU2, homologous to the S. cerevisiae complex SCC2/SCC4 (Watrin et al. 2006; Visnes et al. 2014;

Rhodes et al. 2017). The complex assembles fully before association with chromatin (Losada

2014), and is loaded onto DNA through an ‘entry’ gate between the two SMC subunits (Fig.

1.5). Throughout G1, the cohesin complex is dynamically loaded and unloaded, as cohesin release factor WAPL associates with the kleisin subunit (Fig. 1.6) and catalyses dissociation of the kleisin with SMC3, opening an ‘exit’ gate (Tedeschi et al. 2013). Once S-phase commences, acetyltransferases ESCO1 and ESCO2 are able to acetylate the SMC3 subunit at two lysine residues towards the protein’s N-terminus (K105 and K106 in both mouse and human SMC3), which acts to ‘lock’ cohesin onto chromatin during DNA replication (Ben-

Shahar et al. 2008). This acetylation recruits the cohesin maintenance factor sororin (Fig. 1.6), which is antagonistic of WAPL activity, so maintaining the cohesin complex by blocking opening of the ‘exit’ gate (Nishiyama et al. 2010). The WAPL accessory proteins PDS5A or

PDS5B (Fig. 1.6) have been associated both with cohesin removal (G. Wutz et al. 2017;

Nishiyama et al. 2010) and with establishment of the cohesin complex potentially through promotion of SMC3 acetylation by the acetyltransferases ESCO1 or ESCO2 (Hons et al. 2016;

Carretero et al. 2013). This indicates an essential cohesin regulatory function of PDS5 proteins, beyond simply acting as WAPL cofactors.

There are two phases of cohesin dissociation during mitosis: in prophase, and during anaphase. Prophase dissociation occurs along the chromosome arms as sororin is phosphorylated by Aurora B and Cdk1 kinases (Borton et al. 2016; Nishiyama et al. 2013). This allows WAPL activity and opening of the ‘exit’ gate between the SMC3 and kleisin subunit.

Centromeric cohesin molecules are protected from this dissociation step by the centromeric localisation of SGO1 (Shugoshin) and PP2A (Protein Phosphatase 2A), which act to dephosphorylate sororin and maintain the WAPL antagonism that was established earlier in the cell cycle (Fig. 1.6). There is also evidence to suggest that SGO1 and WAPL act as a second layer of regulation, directly competing for binding sites on the SA1 or SA2 subunit and enhancing the protection of cohesin from WAPL-mediated removal at the centromere (Hara et al. 2014).

Centromeric cohesion must be maintained until the spindle fibres have properly attached to kinetochores at the centromeres, ensuring even division of chromosomes during segregation and cell division. Once the spindle assembly checkpoint (SAC) has been satisfied, anaphase commences, during which SGO1 and PP2A dissociate from the centromere (Fig. 1.6). This has been hypothesised to be in part due to the physical forces put onto bioriented sister chromatid kinetochores as they are pulled away from each other during spindle depolymerisation

(Nerusheva et al. 2014). SGO1 and PP2A inhibit the enzyme separase (Sun et al. 2009; Clift,

Bizzari, and Marston 2009; Holland et al. 2007; Lianga et al. 2018), and upon SGO1-PP2A dissociation, separase cleaves the RAD21 kleisin (Uhlmann, Lottspelch, and Nasmyth 1999).

Another mechanism for regulation of separase-mediated kleisin cleavage is in the APC/C

(anaphase-promoting complex, or cyclosome) itself, which is a component of the SAC.

Separase is bound by a chaperone protein called securin, which prevents its enzymatic activity, until the SAC is satisfied (Luo and Tong 2018; Holland et al. 2007). At this point,

APC/C phosphorylates securin (Toda et al. 2012), allowing activity of separase and so promoting cohesin dissociation and allowing the chromosomes to divide. Cohesin components are recycled and SMC3 is deacetylated by the histone deacetylase HDAC8 (Fig.

1.6), and these protein subunits can then form new cohesin complexes in the new G1 phase daughter cells (Dasgupta et al. 2016).

Aside from sister chromatid cohesin during the cell cycle, the interphase roles of cohesin also require association and disassociation with chromatin. With loss of CTCF, which is associated with cohesin in formation of TADs and TAD boundaries (Hansen et al. 2017), cohesin is still associated with interphase chromosomes (Wendt et al. 2008), implying that

CTCF acts to organise cohesin’s ability to restructure chromatin, but the loading and regulation of cohesin itself involves other mechanisms. This is at least in part regulated by the binding of WAPL cohesin removal factor, which has been associated with the size of chromatin loop structures and regulation of cohesin’s hypothesised ability in loop extrusion

(G. Wutz et al. 2017). PDS5 proteins appear to be involved in regulating CTCF and the boundaries of loop extrusion in a WAPL-independent manner (G. Wutz et al. 2017).

Maintenance of cohesin complexes via SMC3 acetylation has also been associated with interphase chromatin architecture (Kawasumi et al. 2017), and the role of PDS5 proteins in promoting SMC3 acetylation may be linked to this (K.-L. Chan et al. 2013).

Figure 1.6 Cohesin Dynamics in the Mitotic Cell Cycle Cohesin complexes are dynamically loaded and unloaded throughout G1-phase, via an entry gate between the hinge regions of the two SMC subunits and an exit gate between SMC3 and the kleisin subunit, as a result of WAPL activity. In S-phase, the SMC3 subunit is acetylated, recruiting cohesin maintenance factor sororin, which inhibits WAPL activity and maintains chromosome cohesion. Mitotic progression involves both a prophase and an anaphase dissociation step, which involve sororin phosphorylation and WAPL activity, and separase activity, respectively. Cohesin components are recycled into G1 of the daughter cells’ cycle. Taken from (Losada 2014).

1.3.3 Cohesinopathies

Somatic mutations in cohesin subunits or regulatory pathways result in a group of pathologies known as cohesinopathies. One of the most well-studied human cohesinopathies is Cornelia de Lange syndrome (CdLS), which is characterised by intellectual disability and distinctive craniofacial and limb developmental morphologies (Kline et al. 2018). This has been associated with human mutations in cohesin chaperone NIPBL, the SMC subunits and the kleisin RAD21 (Deardorff et al. 2012). Other cohesinopathies include Roberts syndrome

(RBS) and Warsaw Breakage syndrome (WABS). RBS patients have growth and mental retardation, limb deformities and craniofacial defects. These are thought to be a result of widespread translational errors, stemming from defects in chromatin architecture and pericentromeric heterochromatin morphology (Whelan et al. 2012), as well as widespread mitotic aneuploidies and cell apoptosis (Banerji, Skibbens, and Iovine 2017). These symptoms are associated with a loss of cohesin acetylation as a result of mutations in ESCO2, which is specifically localised to pericentromeres and acts to acetylate the SMC3 subunit (Dupont et al. 2014; Kawasumi et al. 2017), so promoting dissociation of sister chromatids and potential chromosome missegregation during the mitotic cell division. Both null Esco1 or Esco2 mice are embryonic lethal and there are no documented human ESCO1 mutations, possibly because this would also be embryonic lethal (Gordillo et al. 2008; Banerji, Skibbens, and Iovine 2017).

WABS is a rare cohesinopathy associated with congenital microcephaly and sensorineural hearing loss, alongside a wide range of developmental abnormalities (Alkhunaizi et al. 2019).

This is associated with mutations in the DNA helicase DDX11/ChIR1 required for DNA repair and replication fork stabilization. Lack of functional DDX11 results in chromosomal breakages and disorganization of heterochromatin (Pisani et al. 2018). DDX11 is required for establishment of chromosome cohesion during DNA replication by association with replication fork protection factors and has recently been shown to interact with cohesin

71 complexes in vitro (Cortone et al. 2018), indicating that it has a vital role in cohesin loading and functions during S phase.

Given the widely varied roles for cohesin in the mitotic cell, both in interphase and during mitosis, molecular consequences of these mutations can vary from global transcriptional changes as a result of chromatin remodeling, associated with CdLS, to mitotic cell cycle failure as a result of aneuploidy and apoptosis, which is associated with RBS (Banerji, Skibbens, and

Iovine 2017).

In addition to the wide array of developmental disorders associated with defects in the cohesin complex, there is also an association between cohesin defects during meiosis, particularly in oocytes from older women, and reduced fertility or increased incidence of miscarriage as a result of chromosome missegregation and aneuploidy (Herbert et al. 2015).

Meiosis and the importance of cohesin in this process are discussed in more detail in section

1.4.

1.4 Meiosis and Genomic Stability during Reproduction

1.4.1 Aneuploidy and Fertility

Aneuploidy is an imbalance in the number of chromosomes in a given cell type; aneuploidy arising during gametogenesis, as a result of chromosome missegregation during the two meiotic divisions, leads to a variety of reproductive or developmental disorders. The outcome of chromosome missegregation can be highly variable depending on the chromosomes involved in the aneuploidy. For example, one of the most common cases of aneuploidy documented in human gametes is trisomy 21: a third copy of chromosome 21, which leads to the developmental disorder Down Syndrome (Xing et al. 2016). However, the vast majority of embryonic aneuploidies involving whole chromosomes are incompatible with live birth and undergo either spontaneous early onset miscarriage or cannot lead to pregnancy at all, resulting in subfertility of the ovulating female. Subfertility or pregnancy complications as a result of meiotic aneuploidy arise in 10-30% of human pregnancies and, interestingly, over 95% of these are a result of meiotic errors during oogenesis – i.e. chromosome missegregation during female meiosis, as opposed to chromosome missegregation in the male meiotic cells

(Hassold and Hunt 2001).

The incidence of meiotic aneuploidy increases exponentially with maternal age; at the age of

20, women are likely to experience aneuploid pregnancies 5% of the time, but at the age of 40, this incidence increases to 35%. This rate is estimated to be higher than even this, but early spontaneous abortion is often missed and not always linked directly to aneuploidy (Hassold and Hunt 2001). This is exacerbated in society by the growing social trend that women wait until they are older to have children; in the USA, for example, there was an 8-fold increase in the number of women choosing to have their first child over the age of 35 between 1970 and

2006 (Matthews and Hamilton 2009).

In addition to meiotic segregation errors, mitotic chromosome instability can lead to chromosome missegregation in the first few mitotic divisions, giving rise to mosaic

73 aneuploidy in the early embryo (Daughtry and Chavez 2016; Magli 2000). A large proportion of these early embryos undergo spontaneous abortion before the blastocyst stage (Bielanska

2002). However, these are not significantly associated with maternal age, and instead are thought to arise as a result of relaxed checkpoint controls during the first few mitotic divisions (McCoy et al. 2015; Mantikou et al. 2012).

Meiotic aneuploidy is a leading cause of birth defects or miscarriage (Hassold and Hunt 2001), and although we are able to diagnose it during both natural pregnancies and IVF (in vitro fertilisation), potentially avoiding it by selecting for healthy oocytes in the latter, there are no current clinical strategies for prevention during oogenesis (Nagaoka, Hassold, and Hunt

2012). This means that study of aneuploidy is vital to improving fertility technology and increasing the ability of people to have children at a wider range of life stages, to align with their own social and reproductive priorities.

Laboratory mice are not often maintained beyond their standard reproductive lifespan (8 to

11 months), roughly equivalent to those of human women with increased aneuploidy incidence, and at this age there is not an equivalent aneuploidy rate recorded (Koehler et al.

2006). However, 16-month aged mice have been associated with increased chromosome missegregation rates (Sakakibara et al. 2015).

1.4.2 The Two Meiotic Divisions

Mammalian meiosis is a two-stage specialised cell division in which four copies of each chromosome, generated by DNA replication in S phase, are segregated into haploid gametes, first by separation of the pairs of homologous chromosomes in Meiosis I (MI), and the segregation of the two sister chromatids within each homolog pair in Meiosis II (MII).

Haploid gametes then combine during fertilisation to produce a diploid zygote, which will then proceed through development. These are largely similar between oogenesis and

74 spermatogenesis, but there are key differences between male and female reproductive systems that result in specific adaptations in either spermatogenesis or oogenesis.

1.4.2.1 Prophase I: Synapsis and Recombination

Prophase I is a stage of meiosis that encompasses the processes of homology searching,

chromosome synapsis and homolog recombination during MI (Soh et al. 2017). After

DNA replication in S phase, chromosomes enter what is called leptotene of prophase

I (Fig. 1.7). They undergo widespread double-strand break (DSB) formation by

topoisomerase-like SPO11 protein (Keeney 2008), and then 5’ DNA resection occurs

to yield 3’ single-stranded tails (Neale, Pan, and Keeney 2005; Neale and Keeney 2006).

This facilitates a homology search within the nucleus, and 3’ tails undergo strand

invasion of their intact homolog, leading to recombinant products that can resolve

into either crossover or non-crossovers of genetic material (Allers and Lichten 2001;

Baudat and de Massy 2007). These recombinant repair events are able to act either

with the homolog, or the sister chromatid – but in meiotic cells, there is thought to

be a bias towards homolog-repair, leading to genetic crossovers (Humphryes and

Hochwagen 2014).

During this process, chromosome synapsis occurs – and this depends on formation of

the synaptonemal complex (SC), a proteinaceous scaffold that forms an axis along

which the homologous chromosomes are able to align. The SC is made up of three

component parts. First, the lateral element made up of proteins such as SYCP2 and

SYCP3, which localise to chromosomes in the leptotene stage of prophase I (Fig. 1.7).

It forms individual axes by polymerisation (Baier, Alsheimer, and Benavente 2007;

Syrjänen, Pellegrini, and Davies 2014) along each homolog, and as they align during

homolog DSB repair, the cell enters zygotene (Fig. 1.7). Transverse filaments and the

central element of the SC such as the proteins SYCE1, SYCE2 and SYCP1 form a

ladder-like structure that holds the two homologs together in a long axis (Y. Costa et

75 al. 2005; Bolcun-Filas et al. 2007; Hernández-Hernández et al. 2016; Dunne and Davies

2019). The next and longest stage of prophase I, pachytene (Fig. 1.7), is defined as when all chromosomes in a prophase I nucleus are fully synapsed and the SC has

“zipped” together, i.e. the LE and CE are fully formed along the length of the SC. DSB repair occurs primarily in this stage, in which homologs recombine and either produce crossover or non-crossover events (Baudat and de Massy 2007). The mismatch repair heterodimer MLH1/MLH3 localises and carries out crossover repair

(Falque et al. 2007; Rogacheva et al. 2014). The final stage of prophase I is diplotene

(Fig. 1.7), in which the SC begins to “unzip” and homologous chromosomes are only held together at their recombination points. At this point, the SC dissociates and genetic crossovers mature into chiasmata, marking the onset of prometaphase I.

Chiasmata are physical connections between genetic material at crossover points that act to maintain bivalent structure without proteinaceous scaffolds such as the SC or cohesin complexes (Bascom-Slack, Ross, and Dawson 1997).

Throughout prophase I, sister chromatids are held together by the cohesin complex

(Fig. 1.8), which is a tripartite protein ring complex. The structure of this complex was discussed in further detail in section 1.3.1. There are multiple meiosis-specific cohesin subunits, including SMC1β and kleisin proteins such as REC8 and RAD21L

(Fig. 1.5). Given that all mitotic subunits are also present during meiosis, this gives a huge variability in the potential cohesin complexes that are present in the meiotic cell.

These are thought to carry out a variety of different functions in the organisation of chromatin and sister chromatid cohesion during MI. SMC1α and SMC1β-containing cohesins share a role in SC axial length regulation during meiosis, but SMC1β is specifically required to prevent a DNA damage response at telomeres (Hodges et al.

2005; Ekaterina Revenkova et al. 2004; Biswas, Stevense, and Jessberger 2018).

Additionally, REC8 and RAD21L-containing cohesin complexes are spatially distinct

along the chromosome axis in prophase I (Rong et al. 2016), and these have been

shown to be required for distinct roles in SC formation and chromosome synapsis

(Ishiguro et al. 2011; Herrán et al. 2011; Fukuda et al. 2014; Ward et al. 2016). Therefore,

it seems that dynamics of different cohesin complexes are tightly regulated

throughout meiosis, each carrying out specific functions.

Figure 1.7 Meiotic Prophase I Progression The four stages of meiotic Prophase I. Pairs of sister chromatids are aligned and held together by cohesin rings (black bars), and double strand breaks (DSBs) form in leptotene. Each pair of homologous chromosomes begins to line up as a result of the homology search in zygotene, and once the full lengths of the homologs are aligned and synapsed, the cell progresses into pachytene. DNA recombination and crossing over occur here, and then the synaptonemal complex begins to break down, allowing desynapsis except for the crossover points in diplotene. Schematic adapted from (Öllinger, Reichmann, and Adams 2010).

1.4.2.2 Metaphase I: Chiasmata Stability and Homolog Segregation

After crossover maturation into chiasmata, the two homologs of each chromosome are maintained in their bivalent complex via sister chromatid cohesion within each homolog, carried out by cohesion complexes (Brooker and Berkowitz 2014). Cohesin complexes downstream of the crossover hold together the original sister chromatids, so holding the two homologs together and stabilising the position of the chiasma (Fig.

1.8). Bivalents align on spindle fibres in metaphase I, and chiasmata act to provide tension against the spindle forces until chromosomes are fully aligned, so reducing the incidence of premature chromosome separation.

The majority of aneuploidies detected in oogenesis are a result of non-disjunction or missegregation events that occur during the first meiotic division (Pellestor et al.

2002), although there are missegregation events that arise during MII as well

(MacLennan et al. 2015). Maintenance of MI chromosome cohesion is vital for normal segregation at both these stages, and so it is key that chromosome cohesion is protected throughout Metaphase I to allow proper attachment of spindle fibres, so chromosome segregation can occur.

Kinetochores are large proteinaceous structures that form at the centromere of each sister chromatid and act, among many other regulatory functions, as attachment points for the spindle microtubules (Thomas, Renjith, and Manna 2017; Joglekar and

Kukreja 2017). Attachment of the spindle occurs in mitotic cells in a bipolar orientation, allowing segregation of the two sisters to either end of the cell. Meiotic kinetochores instead align in a monopolar orientation in MI, allowing segregation of the sisters within each homolog pair towards the same direction, and then in a bipolar orientation in MII, generating a haploid cell upon division (Sakuno, Tada, and

Watanabe 2009; MacLennan et al. 2015).

Oocytes are among very few cell types that lack a specific centriole structure for spindle nucleation, having disassembled it during prophase I, and instead have multiple aMTOCs (acentriolar microtubule-organizing centres), which are protein complexes essential for normal cell division and production of euploid cells

(Baumann et al. 2017; Szollosi, Calarco, and Donahue 1972). Via the RAN-GTPase pathway, spindle fibres extend towards the kinetochore centromeric structures of meiotic bivalent chromosomes, and attach. After arm cohesion has degraded, microtubules rapidly depolymerise in order to pull the homologous chromosomes apart, and centromeric cohesion holds each pair of sisters together. During oogenesis, the meiotic spindle is asymmetrically positioned by activity of actin filaments in order to carry out the unequal divisions and polar body extrusion (PBE) that maintain maternal cytoplasmic resources (Mogessie, Scheffler, and Schuh 2018).

To ensure maintenance of sister chromatid cohesion throughout formation of the spindle, so preventing premature separation and missegregation of chromosomes, eukaryotes have evolved the spindle assembly complex (SAC), in which the APC/C

(Polanski 2013) signals for removal of arm cohesion along each bivalent through cohesin complex cleavage and depolymerisation of the microtubule filaments, so pulling the homologous pairs apart and towards each end of the dividing cell

(MacLennan et al. 2015; Herbert et al. 2015).

EMBRYO: OOCYTE DEVELOPMENT ADULTHOOD: OVULATION

Young Mothers Short Dictyate Arrest

Prophase I Metaphase I

Maternal and Cohesin Older Mothers

Paternal Sister rings hold Long Dictyate Chromatids, held the chiasma Arrest together by in place Cohesin rings

Figure 1.8 Meiotic Cohesin Establishment The cohesin complex (red) holds sister chromatids together along their full axis length and, after crossovers occur and resolve into chiasmata during Prophase I, hold the bivalent structure of the two homologs (blue and yellow) together, downstream of the chiasma. Oocytes go into dictyate arrest and, upon ovulation in adulthood, resume meiosis and progress into pro-Metaphase I. After a short dictyate arrest, cohesin complexes are present in abundance on the chromosome arms, holding the chiasmata in place while the chromosomes align on spindle fibres in Metaphase I. However, a longer dictyate arrest (i.e. an older female) is associated with a reduction in the number of cohesin complexes present along the chromosome arms, which destabilises the Metaphase I bivalent structure and increases the chance of premature separation before proper attachment of spindle fibres (not shown).

1.4.2.3 Meiosis II: Segregation of Sister Chromatids

After the first meiotic division is complete, meiocytes then progress into MII. This

stage is shorter, as there is no need for further exchange of genetic material. Pairs of

sister chromatids are held together by centromeric cohesin complexes, and the

kinetochores are bi-orientated on spindle fibres, much like in Metaphase I.

Spermatocytes progress through cell division, but oocytes arrest at Metaphase II until

fertilisation with mature sperm, upon which time centromeric cohesion breaks down

(MacLennan et al. 2015; Perry and Verlhac 2008). Again, only in oocytes, asymmetric

positioning of the spindle is achieved by actin anchoring against the egg’s surface, this

time parallel to the oocyte membrane (Mogessie, Scheffler, and Schuh 2018). When

cell division occurs, the second polar body is extruded and half of each pair of sister

chromatids is left in the oocyte, giving a haploid female nucleus which can then fuse

with the haploid sperm nucleus to produce a diploid single-cell zygote.

1.4.3 Gendered Gametogenesis

The stark difference between male and female incidence of aneuploidy could be attributed to sexual dimorphism between male and female meiosis itself. One of the key differences between these is the timing of gametogenesis; spermatogenesis begins at puberty and occurs continuously throughout adulthood. However, oogenesis is not a continuous process and begins during embryonic development of the ovary, undergoing two arrests: dictyate, at the end of prophase I (MacLennan et al. 2015), and the metaphase II arrest (Perry and Verlhac

2008). Prophase I, involving synaptonemal complex formation and pairing of homologous chromosomes in an aligned conformation (see section 1.4.2.1), proceeds synchronously in the embryonic ovary between E14.5-E18.5 in the mouse embryo (Dietrich and Mulder 1983;

Reichman, Alleva, and Smolikove 2017). After crossovers have occurred, the cells go into an extended arrest phase, known as dictyate arrest (Fig. 1.8). Meiosis does not resume until

81 immediately prior to ovulation; in humans this is after puberty and could be anything from 15 to 40 years later, and in laboratory mice this is between 2 and 12 months later, assuming a normal reproductive lifespan. Therefore, the dictyate arrest needs to be stable and able to maintain crossovers over this long period of time, allowing sufficient tension between the spindle and chiasmata on the bivalent in metaphase I.

1.4.4 Hypotheses for Oocyte Aneuploidy

Increased maternal age has long been associated with a decline in fertility, long before the onset of menopause. The underlying molecular cause is not fully understood, although there are a few prominent hypotheses. One of the original hypotheses centered on a ‘production line’ of oocytes that entered foetal meiosis stages in the same order that they then exited the dictyate arrest at ovulation, and that the number of crossovers established during the foetal period decline as development progresses. Therefore, oocytes ovulated later in life would have been produced later in embryogenesis, have less crossovers and less chiasmata, and so chromosomes would be more likely to prematurely dissociate during meiotic progression, resulting in aneuploidy (Polani and Crolla 1991). However, this have been repeatedly disproved, as oocytes isolated from different foetal stages show no difference in crossover frequency (Rowsey et al. 2014; Polani and Crolla 1991; Tease and Fisher 1986; Speed and

Chandley 1983).

Aside from chromosome aneuploidies, another theory to explain the maternal age effect includes a decrease in mitochondrial DNA (mtDNA) copy number. Mitochondria are the powerhouse of the cell; they are highly specialised organelles that generate ATP to provide energy for essential cellular functions via the electron transport chain. The mitochondria encode their own mtDNA, of which there are between 1 and 15 copies per organelle. Control of this copy number is vital for normal cellular function and oocyte cell division is associated with an asymmetric division in mitochondria; the majority of the organelles are retained in

82 the oocyte nucleus to promote ATP production during preimplantation development. It has been shown that polar bodies isolated from older women contain mitochondria with lower copy numbers (C. C. W. Chan et al. 2006; Konstantinidis et al. 2014). However, surprisingly, early embryo and blastocyst mtDNA copy number are increased from older women, and so there appears to be a recovery of mitochondrial mtDNA replication in the early stages of development (Elpida Fragouli et al. 2015). Interestingly, lower mtDNA copy numbers at this stage of development are associated with embryonic survival (Diez-Juan et al. 2015), and higher copy number has been associated with lower oocyte quality and increased aneuploidy rate (Elpida Fragouli et al. 2015). This indicates that although mtDNA copy number may decline with age, this does not have a direct link to the increase in aneuploidy associated with older oocytes.

Of the currently debated hypotheses behind age-related oocyte aneuploidy, those with the most molecular basis are: an increased chance of chromosome missegregation as a result of failed maintenance of chromosome cohesion at crossovers, or lower stringency at the spindle assembly checkpoint (SAC), resulting in premature depolymerisation of the meiotic spindle before proper attachment of each chromosome. These are discussed in detail in sections 1.4.4.1 and 1.4.4.2.

Aside from chromosomal or molecular changes over time, the age and health of the surrounding tissue environment may play a role, and these are not necessarily mutually exclusive with some of the prevailing molecular theories. Indeed, it is likely that aneuploidy is actually a consequence of multiple overlapping factors, both on a molecular and systemic level, and none of these hypotheses can be considered in isolation.

1.4.4.1 Meiotic Chromosome Cohesion

Chromosome cohesion and regulation of the cohesin complex during meiosis shares some similarities with regulation of mitotic cohesin, although there are major adaptations for the differences in cell cycle progression. There are two meiotic divisions, and so cohesins are maintained on sister chromatid centromeres through the first division of homologous pairs, and then cohesin at sister centromeres is removed prior to segregation in the second division (Brooker and Berkowitz 2014).

Both chromosome arm and centromeric cohesion are established in S-phase and intact throughout Prophase I, and arm cohesion acts to stabilise chiasmata, holding the bivalent together through alignment of the spindle in metaphase I. At the first metaphase-anaphase transition, cohesin complexes are removed along the chromosome arms, but maintained at the pericentromeric region. Changes in chromosome arm cohesion have been linked to chromosome missegregation during

Meiosis I. Live imaging experiments on young and old mouse oocytes shows that bivalents separate into univalents during metaphase I more frequently in older oocytes, then bi-orient on the MI spindle, so being more susceptible to missegregation (Sakakibara et al. 2015).

Loss of chromosome cohesion at metaphase I is often linked to the loss of chromosome-bound cohesin complexes, which act to hold sister chromosomes together during prophase I homology repair and recombination. Senescence- accelerated mice (SAM), which have a natural premature ageing phenotype, show a marked decrease in cohesin subunits including REC8 and SMC1β on oocyte chromosomes and reduced cohesion of chromatids at MI and MII compared to wild type mice (L. Liu and Keefe 2008). Meiotic cohesin subunit REC8 is depleted with age – immunofluorescence shows a of chromosome-associated protein in aged wild type prometaphase I oocytes (isolated from 14 month old female mice), alongside a

84 reduction in chromosome arm cohesion (Lister et al. 2010; Chiang et al. 2010;

Tachibana-Konwalski et al. 2010) – and there is no apparent mechanism for replenishment of the protein after removal from the chromosome axes (Burkhardt et al. 2016). Similarly, Smc1β expression only in prophase I provides sufficient protein for normal sister chromatid cohesion and chromosome segregation in MI and MII (E

Revenkova et al. 2010), as opposed to a full Smc1β-/- which results in defects in MI oocyte sister chromatid cohesion (Hodges et al. 2005). Finally, the most recently identified kleisin subunit, RAD21L, is necessary for female fertility only with age, implying a dependence on these complexes when REC8-containing complexes are depleted in later life (Herrán et al. 2011). Loss of both chromosome cohesion and chromosome-associated cohesin complexes has also been observed in oocytes isolated from older human women; in particular, the subunits REC8 and SMC1β were decreased in the oocytes isolated from otherwise healthy women around the age of 40 years or over, as compared to 20 year old controls (Tsutsumi et al. 2014).

Together, these data point to the idea that cohesin is maintained without further gene expression in prophase I and later stages, resulting in a slow depletion of the complexes from chromosomes in dictyate arrest. This is illustrated in Fig. 1.8, in which a longer dictyate arrest (in an older woman) results in a decreased abundance of cohesin rings associated with the chromatin, allowing premature chromosome separation to occur upon commencement of Metaphase I.

As discussed above, it is thought that different meiosis-specific cohesin complexes carry out different functions (Ward et al. 2016; Anderson et al. 1999; Wolf et al. 2018;

Biswas, Stevense, and Jessberger 2018). Cohesin depletion with age occurs primarily with REC8-containing complexes, although RAD21L and SMC1β-containing complexes are necessary with age (Herrán et al. 2011; Biswas, Stevense, and Jessberger

2018), indicating that different complexes are able to at least partially carry out the function of REC8-containing complexes if required.

1.4.4.2 SAC Stringency

The spindle assembly checkpoint (SAC) operates prior to anaphase I and ensures that chromosomes are correctly attached to the meiotic spindle before segregation. It is thought that, unlike in spermatocytes, the oocyte SAC triggers the metaphase- anaphase transition when only a threshold of microtubule-kinetochore attachments are achieved, leaving room for error in a small proportion of sisters that may not be correctly attached as the chromosomes begin to separate (Polanski 2013; Kolano et al.

2012). Transcriptional analysis in the human pro-metaphase II oocyte shows an age- related decrease in spindle checkpoint-associated genes such as BUB1B (BubR1 protein) and CDK1 (D. J. Baker et al. 2004; E Fragouli et al. 2010; Riris, Webster, and

Homer 2014). In the GV-stage (germinal vesicle) mouse oocyte, expression of Atrx is decreased between old and young mice (Pan et al. 2008). This is a heterochromatin remodeler but is not required for spindle organisation until metaphase II (De La

Fuente et al. 2004). These transcriptional changes offer no clear explanation for the increasing incidence of aneuploidy as a woman gets older, as the majority of chromosome missegregation occurs in MI (Sakakibara et al. 2015) and there appear to be no functional consequences at this stage as a result of loss of expression. Timing of anaphase I in old (16-19 month) mouse oocytes is unchanged, indicating no perturbation of the SAC (Duncan et al. 2009) and, furthermore, there are no differences in the timing of polar body extrusion of germinal vesicle breakdown onset in old or young mice, even when aneuploidy rates are increased (Lister et al. 2010).

This indicates that, although there are changes in SAC regulators between old and young mice, there does not appear to be a particular perturbation of the SAC itself,

86 and so these changes are not sufficient to be solely causative of ageing oocyte missegregation.

1.4.4.3 Ovarian Tissue Environment

The traditional view of the ovarian tissue environment during the ageing process is that there is a generic oocyte pool that is slowly depleted over time (Warburton

2005), as dictyate arrested oocytes resume meiosis and are ovulated individually. This has been the prevailing theory of oogenesis since its conception in the 1950s

(Zuckerman 1951). However, new evidence implies that there may be a small ovarian stem cell pool, although the biological significance is as yet unknown (Johnson et al.

2004; Bukovsky et al. 2004; Telfer and Anderson 2019). This could point to a depletion of stem cell environment with age (H. Ye et al. 2017), contributing to a declining molecular environment and poor oocyte quality later in life. These cells can be cultured in vitro and differentiated into oocyte-like meiocytes, with the distinctive large cytoplasmic area and expression of germ cell-specific marker genes such as

DAZL and SYCP3 (Parte et al. 2014).

However, there is much debate surrounding the existence of this stem cell pool, and whether they differentiate during adulthood into oocytes for ovulation, which has not thus far been shown in vivo. Oocytes actively going through prophase I have only been detected in embryonic ovaries (Reichman, Alleva, and Smolikove 2017) and genetic experiments show that oocytes that are ovulated during adulthood are maintained from embryonic prophase I. Depletion of cohesin molecules during adulthood but before ovulation results in aneuploid oocytes, which are unable to correctly segregate their chromosomes during the meiotic divisions (Tachibana-Konwalski et al. 2010) –

87 indicating that there is no cohesin turnover or replacement after the initial loading period in embryonic prophase I.

Instead of acting to directly supplement the oocyte population, it has been suggested that this ovarian stem cell pool may play a role in maintaining the follicular cell environment (Eppig and Wigglesworth 2000; Kerr et al. 2006; Albertini 2004), or could exist as a quiescent ‘backup’ pool that is not functional under normal reproductive conditions (Yuan et al. 2013). Regardless, the majority of ovulated oocytes do not appear to be derived from this stem cell population, instead being maintained in a stable state since embryonic prophase I.

With reproductive age and repeated instances of ovulation, the ovarian tissue goes through many physical and molecular changes and there is a documented increase in tissue fibrosis and degradation of surrounding support cells over time (Briley et al.

2016). Young women with reproductive conditions such as PCOS (polycystic ovarian syndrome) that result in ovarian scarring or cysts do not exhibit an increased risk of meiotic aneuploidy (Weghofer et al. 2007), although they exhibit reduced fertility due to hormonal or ovulatory defects. Therefore, there is no direct link between the scarring prevalent in older women and aneuploidy or subfertility.

1.4.4.4 Systemic Ageing

The ovary is not an isolated tissue, and female meiosis is very responsive to systemic changes in hormonal balance and interaction of different tissues together (Nelson,

Telfer, and Anderson 2013). With age, pre-menopausal human female hormonal levels can fluctuate (Musey et al. 1987), and low AMH (anti-Mullerian hormone) and high

FSH (follicle stimulating hormone) in women have been associated with increased incidence of embryo aneuploidy (Thum, Abdalla, and Taylor 2008; X. Jiang et al. 2018).

Human studies on the long-term effects of hormonal contraceptive-induced ovulation

arrest are limited, but there is a potential link between this and increased fertility in

later life (Farrow et al. 2002; Ford and MacCormac 1995), implying that the reduced

number of ovulation events or menstrual cycles can play a role in ovarian health.

However, this has not been directly linked to oocyte aneuploidy rates, and there are

many other factors including uterine environment and maternal health that can play

a role in the balance of female fertility rates. Additionally, hormonal effect is not

clearly separated from wider issues of maternal age; high FSH alone cannot induce

meiotic aneuploidy (Thum, Abdalla, and Taylor 2008) and there are factors other than

hormonal levels at play in age-related aneuploidy and subfertility (Warburton 2005),

including depletion in oocyte number with age, depletion of chromosome-associated

cohesin complexes and decreasing stringency of the SAC.

There are multiple prevailing theories on the molecular origins of oocyte aneuploidy and it is important to consider them all in tandem. However, the hypothesis with the most direct evidence behind it is the changes in chromosome cohesion, linked to declining levels of cohesin complex components. Aside from chromosome cohesion, the cohesin complex has been linked to genomic stability and large-scale chromatin structure (G. Wutz et al. 2017;

Hansen et al. 2017), and so I would hypothesise that this is a causative molecular deficiency.

Even oxidative stress and loss of SAC stringency have been linked to decline in function of the cohesin complex (Miao et al. 2017; Perkins et al. 2016), and so investigation of oocyte aneuploidy should begin with investigation of the meiotic cohesin complex itself.

1.5 Placentation and its Role in IUGR

The Ubr2-/- mouse has been documented to exhibit a variable perinatal lethality across different genetic strain backgrounds. This is more penetrant in females than males andthe few surviving Ubr2-/- females are much smaller than control animals, indicating a potential growth restriction during either development or adulthood as a result of the loss of UBR2 function (Y

T Kwon et al. 2003).

Growth restriction during development, or intrauterine growth restriction (IUGR), has long been associated with defects in placental development, restricting the availability of oxygen and/or nutrients to the foetus. This is a developmental condition defined in the smallest 10% of human foetuses (Scifres and Nelson 2009), and has been linked extensively to development of adult diseases. The placenta is one of the first organs to develop during embryogenesis, and is vital for embryo growth and survival, acting as the conduit for nutrient delivery from the maternal blood stream (Rossant and Cross 2001). In humans, placental defects are associated with intrauterine growth restriction (IUGR), where the foetus does not receive enough nutrients for a normal growth rate (Scifres and Nelson 2009). This can result in lower birthweight and perinatal morbidity, impacting respiratory and cardiovascular health in neonatal stages. This also leads to health complications later in life, increasing the risk of hypertension, heart disease and respiratory disease. There has also been an association between placental insufficiency and neurodevelopmental disorders, among many other health risks and a shortened life span (Malhotra et al. 2019). Therefore, the study of genes involved in healthy placental development is important to understand this process.

One of the most highly adapted tissues between species is the placenta; humans have a large, extensively villus placenta, with many folds in the trophoblast labyrinth surface that provide a large surface area for maternal-embryonic nutrient transfer. The mouse is much smaller and has less contact space between maternal and embryonic blood supplies, although vilification is still present (Fig. 1.9A). Both humans and rodents have cotyledons, which are bundles of

91 villus tissue, but mice have only a single cotyledon (also known as a discoid placenta), whereas humans have multiple cotyledons, arranged into a larger cluster (Fig. 1.9B) (Rossant and Cross 2001; Woods, Perez-Garcia, and Hemberger 2018). Some of the major differences between human and murine placental morphology have been attributed to the requirement for mouse pregnancies to support, on average, 8-12 foetuses at the same time. This indicates an evolutionary need to limit maternal resource input and ensure equal support of all embryos, which can be observed in the limited maternal vascularisation in the murine placenta, as compared to the more extensive vasculature and of human decidua.

While the murine and human placentas are largely different in morphology and caution must be taken in the extrapolation of data between the species (Malassine, Frendo, and Evain-Brion

2003), animal models are an invaluable tool to study the largely conserved genetic and molecular events that take place during placental development. Mouse phenotyping data, gene expression profiles, and physical protein interaction networks give a huge array of insights into human placental development (Woods, Perez-Garcia, and Hemberger 2018).

Figure 1.9 Differences in Mammalian Placental Development Mammalian placental development is generally defined by the extent of vilification into maternal tissues, increasing the surface area of nutrient and oxygen transfer. A. The mouse placenta is a distinct tissue, with junctional zone (blue) and labyrinth (pink) trophoblast-derived tissue layers. The maternal blood supply extends into the space (white) around the villus labyrinth. Inset: Foetal blood vessels sit alongside the trophoblast-surrounded maternal blood supply, where nutrients pass across the syncytiotrophoblast-derived interhaemal membrane. B. The human placenta is made up of multiple coletydons, which are bundles of villus labyrinth tissue arranged together to form a high surface are contact to the maternal tissues. Inset: Foetal blood vessels sit within the maternal blood stream, allowing for maximal nutrient transfer without the need for a distinct interhaemal membrane. Taken from (Rossant and Cross 2001).

1.5.1 The Murine Placenta

The murine placenta is a layered tissue made up of maternal, extraembryonic and embryonic cell lineages which is highly specialised for its function; to act as a conduit for nutrients between the maternal and embryonic blood supplies. The maternal contribution to placental development makes up the maternal decidua (Md) tissue, which is a highly vascularised extension of the uterine decidual wall, providing a conduit for maternal blood flow and nutrient transfer. The two other cell layers are the junctional zone (Jz), which directly borders the Md, and the labyrinth (Lb) – these are made up of a combination of extraembryonic and embryonic components, and act as the connection point to the maternal blood stream. The Jz is vital for many of the endocrine functions of the placenta, and the Lb is adapted to be highly villus, increasing the surface area of contacts with the maternal blood stream. A majority of cells in both layers are from the extraembryonic trophoblast cell lineage, and successful signalling between the embryonic/foetal tissues, the trophoblast tissues and the maternal decidua and vascular tissues are vital for normal placental functioning, in which the trophoblast acts as an intermediary between the embryonic and maternal cells (Rossant and

Cross 2001; J. C. Cross et al. 2002).

During embryonic development, formation of trophoblast cells and the trophectodermal tissue layer is one of the first distinct specialisation events. It arises from the outer trophectoderm layer of the blastocyst (E3.5, Fig. 1.10A) and, around the time of implantation

(E6.0), differentiates into the ectoplacental cone and the extraembryonic ectoderm (Fig.

1.10B). Away from the inner cell mass, a layer of trophoblast giant cells develop through endoreduplication, which is a specialised DNA replication event that results in polyploid nuclei. It has been hypothesised that this allows increased gene expression and quick tissue growth for large maternal-embryonic contact surface areas, but this is not clearly understood

(Larkins et al. 2001; James C. Cross 2014). The ectoderm develops (among other structures) into the chorionic membrane and the allantois (E7.5, Fig. 1.10C), which undergo chorioallantoic fusion to develop into the villus Lb, and the ectoplacental cone develops into

94 a spongiotrophoblast layer that forms the bulk of the Jz (E12.5, Fig. 1.10D). The spongiotrophoblast and trophoblast giant cells (TGCs) are the first structure of the placenta proper to form, providing a framework upon which chorioallantoic cells can undergo extensive folding and villous branching to form the high surface area labyrinth structure

(Rossant and Cross 2001). This provides ample membrane surface area for transfer of nutrients through into the foetal vasculature. While the major function of trophoblast tissues is acting as a junction between maternal and foetal blood systems, they also act in a variety of hormonal and signalling contexts between maternal and embryonic tissues, which are essential for successful pregnancy (M. A. Costa 2016; Malassine, Frendo, and Evain-Brion

2003; Linzer and Fisher 1999). For example, Jz TGCs secrete lactogen into the maternal blood stream, where it acts to moderate maternal metabolic rate and divert an energy supply for foetal growth, as well as prepare mammary tissues for growth and milk production

(Malassine, Frendo, and Evain-Brion 2003; Nieder and Jennes 1990; Walker et al. 1991).

By E10.5, the full placental structure has formed, supporting the complex processes that happen in the latter half of murine development. The Jz and Lb layers (Fig. 1.10D) are made up of a variety of specific trophoblast-derived cell types that carry out specific functions in the placenta. For example, spongiotrophoblast glycogen-rich cells (SGCs) are present in the

Jz, and these migrate and invade the Md between E13.5 and E18.5, lysing to release glycogen stores as an energy source towards the end of development, when embryonic growth rate is the highest (P. M. M. Coan et al. 2006). TGCs in the Jz are responsible for hormone secretion, as discussed above (Malassine, Frendo, and Evain-Brion 2003) and also function to aid the Jz invasion of the Md during late development (D. Hu and Cross 2010). Labyrinth-specific syncytiotrophoblasts (SynTs) are required along villus surfaces in this tissue layer for formation of the ‘interhaemal membrane’, which is the membrane over which nutrients will pass between maternal and embryonic blood supplies (Woods, Perez-Garcia, and Hemberger

2018). Finally, embryonic epithelial cells in the labyrinth make up the distinct compartments for foetal and maternal blood vessels (James C. Cross 2000).

A B C D

Figure 1.10 Murine Trophoectoderm Lineage Development The trophoblast lineage (blue) is derived from the blastomeric trophectoderm (A) and develops into the ectoplacental cone and extraembryonic ectoderm, as well as trophoblast giant cells (B, C). The ectoplacental cone itself develops into the bulk of the placental junctional zone at E12.5 (Jz, blue), which borders the maternal decidua (Md, yellow) with a layer of trophoblast giant cells in D. This forms a scaffold on which the chorionic ectoderm and allantois (C), which developed from the extraembryonic ectoderm, can fuse (chorioallantoic fusion) and become villus, forming the high surface area labyrinth (Lb, pink) layer seen at E12.5 (D). Taken from (Rossant and Cross 2001).

1.5.2 Roles for Tex19.1 in Placental Development

The Tex19.1-/- placenta is small, preceding an embryonic growth restriction that arises between

E12.5 and E14.5 (J Reichmann et al. 2013; Tarabay et al. 2013 and Playfoot et al. 2019, manuscript submitted, MRC HGU). This is particularly characterised by a reduction in the size of the Jz tissue layer, alongside compositional differences in this tissue, including a reduction in the abundance of spongiotrophoblast cells and SGCs (J Reichmann et al. 2013). Previous published and unpublished work in our lab has shown that this gene has a range of known roles in the murine placental trophoblast tissues, such as retrotransposon inhibition (J

Reichmann et al. 2013) and transcriptional repression through histone H3K9 trimethylation.

The latter is a function of TEX19.1’s interaction with KAP1, an E3 ligase that interacts with a subset of ZFP proteins and is involved in recruitment of the histone methyltransferase

SETDB1 and transcriptionally repressive H3K9me3 histone methylation marks to multiple genomic locations including retrotransposons and imprinted genes (Playfoot et al. 2019, manuscript submitted, and unpublished work by Chris Playfoot, MRC HGU). The placental defects in Tex19.1-/- mice bear some resemblance to those seen in human pregnancies that exhibit IUGR, which has been independently associated with changes in methylation states and changes in expression levels of both imprinted genes and retrotransposon-encoded proteins (Frost and Moore 2010; Michels, Harris, and Barault 2011).

1.5.2.1 Genomic Imprinting in the Placenta

As a consequence of methylation reprogramming during reproduction and

development, the trophectoderm-derived components of the placenta are

hypomethylated (Crichton et al. 2014; Oda et al. 2013). Monoallelic gene expression,

leading to asymmetrical maternal and paternal genetic contributions, arises as a result

of methylation of single alleles at specific loci. This phenomenon was first discovered

97 using pronuclear oocyte transfer and generation of uniparental embryos, which exhibit growth defects and embryonic lethality whether paternal or maternal alleles are missing (Kaufman, Barton, and Surani 1977; Barton, Surani, and Norris 1984;

Surani, Barton, and Norris 1984). Allele imprinting, and a dependence on either the paternal or maternal allele of certain genes, is often associated with dose-dependent expression of genes regulating development and growth, and so is particularly important in the placenta to regulate growth and proliferation during development.

Paternal expression has been linked to maximal input of maternal resources in the placenta, whereas maternal expression is more associated with limiting maternal input, and a combination of both creates a delicate balance of resources and finely tunes nutrient availability and growth throughout development (Frost and Moore

2010). Indeed, IUGR is a major phenotype of the human imprinting disorder SRS

(Silver Russell syndrome), in which around half of cases have loss of expression of growth factors such as IGF2 as a result of imprinting misregulation (Tunster, Jensen, and John 2013). IGF2 is a paternally expressed gene that is thought to be important for angiogenesis and vascularisation in placental tissue (Harris et al. 2011; Chao and

D’Amore 2008), and so loss of this function could lead to a restricted ability to transfer oxygen or nutrients to the foetus. Derepression of imprinted alleles has also been linked to this disorder in humans, such as hypomethylation of the gene MEST, which is prevalent in 7-10% of SRS cases (Hannula et al. 2001). MEST, and the mouse homolog Mest, is a paternally expressed gene involved in angiogenesis in both the decidua and trophoblast tissues within the placenta (Mayer et al. 2000). Female mice deficient in Mest are growth restricted both pre- and postnatally (Lefebvre et al. 1998).

Interestingly, both upregulation and loss of gene expression as a result of imprinting defects can lead to placental defects and growth disorders. Both Jz increases and decreases in size have been linked to embryonic growth defects (Woods, Perez-

Garcia, and Hemberger 2018), as placental size alone is not a true indicator of placental function – cell localisation and abundance are important factors for placental development. Therefore, imprinting misregulation in either direction, as a result of hypo- or hypermethylation, can be deleterious to the developing embryo.

Imprinting disorders in the human placenta have also been associated with preeclampsia (Zadora et al. 2017; Carney 2017), indicating the importance of imprinting in this tissue and a need for a further understanding of the regulatory processes protecting the methylation of these loci in an otherwise widely hypomethylated context. There are widespread differences in the methylation of imprinted loci between human and murine placental epigenomes, reflecting the high level of variability between the placentation and pregnancy requirements of different mammalian species. However, most of these differences are additional paternally imprinted loci in the mouse, and subsequent maternal expression of these genes. This indicates a push towards limitation of maternal input into developmental energy expenditure, possibly as a result of the requirement to support 8-12 embryos during the same pregnancy (Monk et al. 2006). This is known as parental conflict theory, and involves the balance of maternal and paternal gene expression levels (Moore and

Reik 1996; Haig 1999). As a model for human placental imprinting disorders, the murine placenta is still useful; human genes such as MEST, and the mouse homolog,

Mest, have been shown to have similar roles in both species (Mayer et al. 2000).

Alongside the documented IUGR and placental growth defects, the Tex19.1-/- placentas exhibit reduced amounts of H3K9me3 at specific imprinted genes (Mest, Mcts2 and

Impact) and transcriptionally de-repress Mest and Mcts2 around 2-fold in the labyrinth layer (Playfoot et al. 2019, manuscript submitted, MRC HGU). This is thought to act through TEX19.1’s interaction with KAP1, which is a transcriptional corepressor involved in deposition of H3K9me3, but also implicates UBR2 in this tissue for a

99 potential interaction with the TEX19.1-KAP1 complex and a subsequent role in imprinting.

1.5.2.2 Retrotransposons and the Placental Immune Response

Transposable elements (TEs) are mobile elements which make up around 40% of the sequenced mammalian genome and were first described nearly 70 years ago

(McClintock 1956, 1950). There are only two classes of TEs: firstly, transposons which move around the genome in a “cut and paste” method and are quite rare, only taking up 3% of the genome. The other type, retrotransposons, make up the majority of mammalian TEs and act to “jump” around the genome in a “copy and paste” method, being transcribed as RNA and reintegrating into new loci by reverse transcription and recombination, inducing chromosomal deletions or rearrangements (Goodier

2016). Although this activity is generally deleterious and contributes to genomic instability, retrotransposons are maintained in the germline in order to contribute to a healthy mutation rate and genetic variation throughout multiple generations.

In the majority of somatic tissues, a wide variety of genetic silencing mechanisms are in place to ensure retrotransposons do not actively damage the genome. One of these is epigenetic modifications, which are able to directly repress retrotransposition through DNA methylation and histone marks (Slotkin and Martienssen 2007; Walter et al. 2016). However, within placental tissue, there are functional advantages to the expression of some retrotransposons (Garcia-Perez, Widmann, and Adams 2016).

Therefore, a balance in repression at some retrotransposon loci must be struck in order to limit mutational activity and maintain genomic integrity.

There are two classes of retrotransposons: LTRs (long terminal repeats, including

ERVs, endogenous retroviruses) and non-LTRs, which do not contain the repeat

100 sections. One of the most commonly found families of retrotransposons in the mammalian genome is the non-LTR-class element LINE-1 (long interspersed element

1), copies of which make up 17% of genetic material (Lander et al. 2001; Slotkin and

Martienssen 2007). LINE-1 encodes its own mobilisation machinery, made up of a 5’

UTR promoter region, two protein-coding regions Orf1 and Orf2, and a long 3’ poly-A tail. ORF1p has activity in RNA-binding, and ORF2p acts as an endonuclease and reverse transcriptase; both proteins are essential for retrotransposition of LINE-1, acting to regulate and reintegrate the expressed RNA into new genomic loci (Mita et al. 2018; Martin 2006; Feng et al. 1996). Placental and cord blood supply misregulation of LINE-1-expressed proteins Orf1 and Orf2 have been associated with changes in birth weight in humans, ranging from IUGR to higher than normal birth weights

(Michels, Harris, and Barault 2011).

Tex19.1 has previously been shown to regulate expression of different retrotransposon

RNAs in both testis tissue (Ollinger et al. 2008; Judith Reichmann et al. 2012) and trophoblast-derived placental tissue layers (J Reichmann et al. 2013; Tarabay et al.

2013), including LINE-1 Orf1 RNA, and TEX19.1 protein has been implicated in regulation of LINE-1 Orf1 protein levels as well (Maclennan et al. 2017), so indicating a dual mechanism to protect the genome against retrotransposition both at an RNA and a protein level. Both Tex19.1-/- and Ubr2-/- embryonic stem cells (ESCs) have an upregulation of LINE-1 RNA levels, indicating a role for both of these genes in retrotransposon repression (Maclennan et al., 2017 and unpublished data, Karen

Dobie, MRC HGU).

Finally, upregulated expression of retrotransposons has been associated with an interferon (IFN) immune response in a variety of tissue contexts (Yu et al. 2015;

Mavragani et al. 2016; Kosumi et al. 2019), and this is hypothesised to be a molecular response to the danger to genomic stability that retrotransposition poses, so acting

101 as an additional defence mechanism. As such, interferon responses have been shown to be a vital component of the placental immune system for both embryonic and maternal survival (Racicot et al. 2016; J. Y. Kwon et al. 2018). In particular, the IFNβ response acts in response to LINE-1 retrotransposition (Yu et al. 2015). IFNβ is an established cellular defence mechanism against viral infection (Taniguchi and

Takaoka 2002), which involves upregulation of a wide variety of interferon response genes via the JAK-STAT signalling pathway such as IFITMs (interferon-induced transmembrane proteins) that act to carry out viral endocytosis and inhibit various stages of the viral life cycle (Schneider, Chevillotte, and Rice 2014; Zhao et al. 2019).

Unpublished data in our group show that the Tex19.1-/- placenta exhibits an IFNβ response alongside derepression of LINE-1, so indicating a potential link between retrotransposition and an immune response in this genotype (Chris Playfoot, MRC

HGU).

102

1.6 Summary

The mammalian UBR family is highly diverse, only united by the common UBR box domain that is able to physically bind to the N-terminus of target proteins (E Matta-Camacho et al.

2010; J Muñoz-Escobar et al. 2017). They are historically implicated as E3 ubiquitin ligases in the N-end Rule pathway of the UPS (T Tasaki et al. 2005, 2009; Bartel, Wünning, and

Varshavsky 1990), using this substrate binding domain to target proteins for proteasomal degradation. However, there are roles for UBR box containing proteins outside of this canonical pathway.

Perturbation of UBR2 in a variety of tissues can lead to a plethora of defects, and so the study of these roles, and the molecular basis of UBR2 function, is vital to understanding these protein regulatory pathways and how they impact highly specialized tissue or cell types. A recently identified novel role for UBR2 in the regulation of the cohesin complex has implications in the vulnerability for misregulation of meiotic cohesins in the oocyte (J

Reichmann et al. 2017). Spermatocytes also depend on UBR2 for genetic recombination, chromosomal synapsis and meiotic progression (Akera et al. 2017; Crichton et al. 2017). In addition, UBR2 appears to be specifically important for embryonic survival (Y T Kwon et al.

2003), although the molecular basis of this has not yet been described.

UBR2 has a strong interaction with the protein TEX19.1, which is involved in cohesin complex regulation in meiotic oocytes. Both UBR2 and TEX19.1 are also important for embryonic survival, the latter through regulation of placental development. These phenotypes indicate a potential link between UBR2 and TEX19.1, a protein expressed in hypomethylated tissues that has implicated roles in genomic stability through histone methylation or cohesin complex regulation.

103

104

1.7 Thesis Objectives

I hypothesised that the TEX19.1-UBR2 complex has a role in regulation of both oocyte meiotic chromosome cohesion and placental development, and the first objective of this thesis was to investigate these processes using a Ubr2-/- mouse. There is evidence that TEX19.1 is able to regulate the Type II N-end Rule substrate binding ability of UBR2, and so the second objective of this work is to investigate the molecular basis of the known Ubr2-/- phenotypes, using a genetic approach to perturb the canonical N-end Rule pathway.

In this thesis, I will address the following research questions, across three chapters of independent research.

− Is Ubr2 important for oocyte development and chromosome cohesin during female

meiosis?

− Does the role for Ubr2 in embryonic growth act through a regulation of the placental

trophoblast-derived tissues?

− What is the molecular basis behind loss of Ubr2 in the major identified phenotypes?

105

106

Chapter 2

Materials and Methods

107

108

2.1 Animal Work

Animals were maintained in a heterozygous colony with Ubr2+/- x Ubr2+/- or Ubr2+/T2 x Ubr2+/T2 crosses. Unless stated, animals used for experiments were 2-6 months old, with age-matched wild type (Ubr2+/+) controls from the same colony. All animal work was carried out after ethical approval by the University of Edinburgh Animal Welfare and Ethical Review Body, under UK Home Office Project Licences PPL60/4424 or P3900/7F29. Husbandry and animal care followed local institutional guidelines (University of Edinburgh, Central Bioresearch

Services).

2.1.1 Dissection and tissue retrieval

All animals were culled by Schedule 1 termination, in accordance with UK Home Office regulations. Adults were culled by cervical dislocation, and embryos beyond developmental stage E14.5 were culled by decapitation.

2.1.2 Embryo tissue retrieval

For embryo experiments, heterozygous pairs (either Ubr2+/- x Ubr2+/- or Ubr2+/T2 x Ubr2+/T2) crosses were set up and checked for plugs every 24 hours. The day that the plug was found was designated E0.5, and development was measured as E10.5 (10 days after plug), E12.5, E18.5, and so on. As a secondary measure of development, to ensure timing from plug dates were accurately representative of developmental stages, I observed limb and craniofacial development; E10.5 was designated as embryos having a lens indentation around the eye primordia but no nasal pits, E12.5 as embryos with developed and webbed digits, and E18.5 with having fully developed but closed eyes, developed whiskers and nail primordia.

Parturition occurs around E19.5-20.5, upon which mouse age is designated in days post parturition (dpp), with the day of birth designated as 0dpp.

109

At the required embryonic stage, pregnant females were culled as in section 2.1.1, and the uterus was dissected. Prior to weighing, extraembryonic membranes (amnion and chorion) were dissected away from the embryo and placenta itself. Embryos post-E14.5 were immediately decapitated after dissection in compliance with UK Home Office regulations, then the head and body were weighed separately and combined to give the total embryo weight. Late stage embryos weighed in this manner will have lost some fluid weight, but should be comparable between genotypes treated in the same way in these experiments. For isolation of internal embryonic tissues at E18.5, embryos were opened across the abdomen in an antero-posterior direction under a dissection microscope. Embryonic ovaries were isolated and stored in 5% FCS (foetal calf serum) in PBS at 4°C until chromosome spreading (see section 2.4.1).

2.1.3 CRISPR-HDR by zygotic injection

Ubr2-/- mice were as described previously (Crichton et al. 2017; Maclennan et al. 2017); the design and implementation of which was carried out using CRISPR-Cas9 by Karen Dobie

(MRC HGU). This mouse has a premature STOP codon in exon 3 (at position Cys-121) of the

Ubr2 gene; UBR2 protein cannot be detected in either the brain or the testes (Crichton et al.

2017). For genotyping purposes, a silent XbaI restriction enzyme site (5’−T^CTAGA−3’, NEB

CutSmart, R0145) was also introduced in exon 3.

To generate a Ubr2T2/T2 mouse line, wild type Cas9 mRNA (0.25μg/μl, Tebu-Bio CleanCap Cas9

L-7206), both wild type (with mutant PAM) and point mutant (D233A) dsDNA repair templates (1.5 μg/μl), and an RNA guide construct (250ng/μl, cloned from Addgene pX330, see section 2.8.1.3) specific to the target site were injected together into a single-celled mouse zygote on the C57BL/6J background. These were then transferred into the uterus of a host mother on a CD-1 background, to ensure all pups born into the Ubr2T2/T2 were on the C57BL/6J

110 background and had been edited by CRISPR. Injections and embryo transfer were carried out by Bioresearch and Veterinary Services (BVS) at the University of Edinburgh. There were 9 pups produced; three were successfully edited and used for subsequent matings and generation of the Ubr2T2/T2 mouse line used in all experiments in Chapter 5.

2.1.4 Genotyping assays

All mice used in this thesis were genotyped by earclipping at weaning (3 weeks post-partum, pp) or, for embryo experiments, using tail tips. DNA was extracted using the DNAreleasy™ standard protocol (Anachem LS02), in which tissue was submerged in 20µl DNAreleasy™ reagent and incubated on a thermocycler at 75°C for 5 minutes, 96°C for 2 minutes and 20°C for 10 minutes. After one freeze-thaw cycle (-20°C), the tissue was spun down using a micro- centrifuge at maximum speed and the supernatant used in subsequent PCR experiments. For all genotyping assays, PCR products were observed using gel electrophoresis, as per standard protocols and as follows. 1.5% agarose gel in TBE (Tris/Borate/EDTA) buffer with 0.5 µg/ml ethidium bromide for visualisation (Electran®, VWR). Samples were combined with loading buffer (5X Orange G, Sigma Aldrich O3756, in 30% glycerol), loaded alongside either 1kb

(NEB, N3232L) or 100bp (NEB, N3231S) DNA ladder, as appropriate for expected sizes, and run at 150V/cm. DNA bands were visualised using a UV SynGene U:Genius 3 Geldoc system.

For all embryo experiments, sex genotyping was used to separate male and female embryos.

Sex genotyping was carried out by PCR against the X-encoded Ube1X gene and the Y-encoded paralog Ube1Y (Chuma and Nakatsuji 2001); primer sequences are noted in Table 2.1. For each

20µl PCR reaction: 2µl 10X Platinum Taq buffer, 0.8µl 50mM MgCl2, 0.16µl 25µM dNTP mix,

0.8µl each primer at 100µM, 0.16µl Platinum Taq DNA polymerase (Invitrogen, 10966034) and

13.92µl dH2O, with 2µl DNA. PCR protocol as in Table 2.2.

111

The Ubr2-/- mouse was genotyped using a two-step protocol: PCR amplification of Ubr2 exon

3, with primers as in Table 2.1, and digestion using XbaI restriction enzyme (NEB CutSmart,

R0145). This was then assayed by gel electrophoresis, see Table 2.1 for expected band sizes.

For each 50µl PCR reaction: 5µl 10X Plat. Taq buffer, 1.5µl 50mM MgCl2, 1µl 10µM dNTP mix,

1µl each primer at 100µM, 1µl Platinum Taq DNA polymerase (Invitrogen, 10966034) and

37.5µl dH2O, with 2µl DNA. PCR protocol as in Table 2.3. The majority of routine genotyping assays for the Ubr2-/- colony were carried out initially by David Read, Mary Taggart or Jennifer

Lawson (MRC HGU), and later outsourced to Transnetyx Inc.

The Ubr2T2/T2 mouse was initially genotyped by tetra-primer ARMS (amplification-refractory mutation system) PCR, which utilises two pairs of primers (inner and outer, sequences in

Table 2.2) to detect specific point mutations in the sequence (Medrano and De Oliveira 2014;

S. Ye et al. 2001). These products could then be assayed by gel electrophoresis, as outlined in

Table 2.2. For each 20µl PCR reaction: 2µl 10X Platinum Taq buffer, 0.8µl 50mM MgCl2, 0.4µl

10µM dNTP mix, 0.2µl each inner primer at 100µM, 1µl each outer primer at 100µM (1:5 ratio inner:outer primers), 0.2µl Platinum Taq DNA polymerase (Invitrogen, 10966034) and 14µl dH2O, with 0.2µl DNA. PCR protocol as in Table 2.3. The details of the ARMS-PCR are outlined in Chapter 5 (Fig. 5.7) (S. Ye et al. 2001; Medrano and De Oliveira 2014). In order to confirm mutation, Sanger sequencing was implemented as in section 2.8.2. Upon establishment of the Ubr2T2/T2 colony, routine genotyping of mice at weaning was outsourced to Transnetyx Inc.

112

Assay Primers Expected Results Notes Forward (Ube1XA): Male (XY): 5’–TGGTCTGGACCCAAACGCTGCCACA 2 bands: 220bp, 200bp Ube1X/Ube1Y -- Reverse (Ube1XB): Female (XX): 5’–GGCAGCCATCACATAATCCAGATG Single band, 220bp +/+: Single band at 310bp Forward: XbaI digest +/-: 5’–TCTGAGGTTGCAAGAGAATGT (NEB Ubr2-/- 3 bands: 310bp, 205bp, Reverse: CutSmart, 100bp 5’–GGCCACAGATCAGATCAGCTAAACC R0145) -/-: 2 bands: 205bp, 100bp Inner Forward: Control band at 476bp 5’–ACACCTACTGCATGCTGTTTAACGA present in all reactions, plus: Inner Reverse: Tetra-primer 5’– +/+: ARMS PCR Ubr2T2/T2 AAATGACTTGCTCATAGGTGTGAACTTTAG Single band at 242bp (S. Ye et al. Outer Forward: +/T2: 2001) 5’–AGGGCTACCAATACAGAGCACTCTGAT 2 bands: 242bp, 292bp Outer Reverse: T2/T2: 5’–TCTCTGTAAGTTCAAGGCCAGTATGGTC Single band at 292bp

Table 2.1 Primers for mouse colony genotyping

Step Ube1X/Ube1Y Ubr2-/- Ubr2T2/T2 1) Initial Denaturation 94°C, 7 minutes 95°C, 7 minutes 95°C, 2 minutes 2) Denaturation 94°C, 30 seconds 95°C, 30 seconds 95°C, 1 minute 3) Annealing 66°C, 30 seconds 65°C, 30 seconds 63°C, 1 minute 4) Extension 72°C, 30 seconds 72°C, 1 minute 72°C, 1 minute 40 cycles 40 cycles 35 cycles Cycles Return to (2) 39 times Return to (2) 39 times Return to (2) 34 times 5) Final Extension 72°C, 5 minutes 72°C, 2 minutes 72°C, 2 minutes

Table 2.2 PCR cycling conditions for mouse colony genotyping assays

113

114

2.2 Tissue culture

2.2.1 Cell storage and general culture conditions

Cells were stored at -150°C in a liquid nitrogen facility, in DMSO+serum supplemented freezing media (culture media for each cell line, as in sections 2.2.2 and 2.23, was supplemented with 20% DMSO and 25% FCS). For retrieval, cells were quickly thawed in a

37°C waterbath and added to 9ml of pre-warmed media, then pelleted by centrifugation (200g,

5 minutes) and resuspended in fresh pre-warmed media. They were seeded into either T25 or

T75 culture flasks (Corning, 25cm2 and 75cm2) at 2x105 cells/ml, unless stated; HEK293T cells were seeded on uncoated flasks, and for mESCs flasks were first coated in 0.2% gelatin (Sigma) in PBS (phosphate buffered saline), which was removed by aspiration prior to addition of cell suspension in media.

All cells were cultured in their respective media at 37°C and 5% CO2. Standard passaging protocol was used for all cell lines: a 1x pre-warmed (37°C) PBS wash, then trypsinisation using pre-warmed trypsin-EDTA (1% trypsin/0.4% EDTA, Sigma Aldrich T4174) in PBS for 5 minutes at 37°C, 5% CO2. Trypsin was inactivated using 10 volumes of the appropriate pre- warmed media, and cells transferred to a 15ml falcon tube to pellet by centrifugation (200g, 5 minutes). Cells were then resuspended in fresh pre-warmed media and reseeded as above.

2.2.2 HEK293T cells

Wild type HEK293T cells were used primarily as a vehicle for protein overexpression through transfection of plasmids. HEK293T cells were cultured in serum-supplemented DMEM

(Dulbecco’s Modified Eagle’s Medium, Gibco 41965-039) (10% foetal calf serum/FCS,

200µg/ml penicillin-streptomycin, P/S, 8mM L-Glutamine) and allowed to grow for 48-72 hours between passaging. For overexpression experiments, cells were seeded at 1x105cells/ml in 6-well plates (9.6cm2) and allowed to grow for 24 hours. Then, the plasmids outlined in

115

Table 2.5 were transfected using the Lipofectamine 2000 (LF2000, Invitrogen 11668-019) standard protocol, as follows. 1µg of plasmid (or 0.5µg each if cotransfecting two plasmids), as measured by Nanodrop spectrophotometry (ThermoFisher Scientific), was added to 50µl

Opti-MEM Reduced Serum Media (Invitrogen 31985-092) and a master mix of 3µl LF2000 per 50µl Opti-MEM was made up. Both were incubated at room temperature for 5 minutes, then combined (total volume 100µl), gently mixed and incubated again at room temperature for 5 minutes. This mixture was then added dropwise to each well of the 6-well plate and the cells incubated for 24 hours as normal. Cells were harvested at 24 hours post-transfection, by either trypsinisation or in-well cell lysis, as determined by experimental protocol.

2.2.3 Mouse Embryonic Stem Cells

Feeder-independent E14 mouse embryonic stem cells (mESCs) were used in this thesis as a cell culture model in which both UBR2 and TEX19.1 are expressed, primarily to investigate the role of UBR2 in cohesin regulation. mESCs were cultured on 0.2% gelatin-coated flasks in serum/LIF-supplemented GMEM (Glasgow Minimal Essential Medium, Gibco 21710-025)

(10% FCS, 100µg/ml P/S, 2mM Gibco L-Glutamine, 1mM Sigma Aldrich sodium pyruvate, 1x

Sigma MEM non-essential amino acids, 50μM Gibco β-mercaptoethanol and Leukaemia inhibitory factor, or LIF) and passaged at maximum every 48 hours.

2.2.3.1 CRISPR-Cas9 HDR for UBR2-T1 mES cell generation

I generated a stable UBR2-T1 line on the E14 background with the point mutation

D118A by CRISPR-Cas9 HDR. E14 mESCs were transfected with the dsDNA repair

oligo with 50bp homology arms encoding the mutant repair template containing both

the point mutation and the mutant PAM sequence, alongside a Cas9 expressing

plasmid and pX330 plasmid with the specific guide sequence cloned into it; see

116

section 2.8.1.3 and Fig. 5.16. Transfection and picking of clonal colonies was carried

out by Fiona Kilanowski (MRC HGU). To genotype successful clones of the UBR2-

T1 line, I designed tetra-primer ARMS PCR, as detailed in section 2.4.1 and in Chapter

5, and confirmed this by Sanger sequencing of the outer fragment alone; genotyping

primer sequences in Table 2.4. Wild type untransfected E14 mESCs were used as a

control comparison to the UBR2-T1 line.

Genotyping Assay Primers Expected Results Forward inner (C allele): Control band at 362bp present in 5’–TAACTTCACAGAGACTGTGCAGTCGC all reactions, plus: Forward outer: +/+: 5’–CTTACACATCACTTGGTGTCCATGAGAA Single band at 235bp ARMS-PCR Reverse inner (A allele): +/T1: 5’–ACTCCATGCATAAAACACAGGTGGAGT 2 bands: 180bp, 235bp Reverse outer: T1/T1: 5’–CTTGCCCTCACATCCATTAAAGCTTAAA Single band at 180bp Forward: +/+ GAC 5’–CTTACACATCACTTGGTGTCCATGAGAA Sanger sequencing +/T1 GNC Reverse: T1/T1 GCC 5’–CTTGCCCTCACATCCATTAAAGCTTAAA

Table 2.3 – Primer sequences for UBR2-T1-D118A mESC genotyping (both ARMS and Sanger)

Step UBR2-D118A-T1 1) Initial Denaturation 95°C, 2 minutes 2) Denaturation 95°C, 1 minute 3) Annealing 61°C, 1 minute 4) Extension 72°C, 1 minute 35 cycles Cycles Return to (2) 34 times 5) Final Extension 72°C, 2 minutes

Table 2.4 – Tetra-primer ARMS-PCR cycling conditions for UBR2-D118A-T1 mESC genotyping

117

2.2.3.2 mESC synchronisation and sorting

To synchronise UBR2-T1 or E14 mES cells into G2/M, cells were seeded at

0.5x105cells/ml in a T25 flask, and after 24 hours of growth, media was changed to fresh GMEM+serum/LIF (see section 2.2.3) supplemented with 1.25mM thymidine

(Sigma Aldrich T9250) to a final concentration of 1.25mM. Cells were cultured for 17 hours under thymidine treatment, then released for 5 hours and cultured in a second thymidine block at 1.25mM. A double thymidine block arrests the cell cycle at the

G1/S phase transition (Chen and Deng 2018), and so in order to collect cells enriched at G2/M, cells were released into growth media for 4 hours and then harvested for subsequent experiments.

FACS (fluorescence-associated cell sorting) analysis for cell cycle tracking was carried out as follows. Cells were seeded in a 6-well plate at 1x105cells/ml and, at various timepoints of the double thymidine block detailed above, harvested by trypsinisation (as in section 2.2.1) and the cell pellet washed in 1x PBS. The suspension was spun down at maximum speed in a tabletop centrifuge and resuspended in 50% FCS in PBS. Cells were fixed by dropwise addition of 450µl ice cold 70% EtOH over a vortex and incubated for a minimum 24 hours at 4°C. To stain for FACS, cells were centrifuged at 1000rcf at 4°C for 5 minutes, washed twice in ice cold PBS, then resuspended in 50µg/ml propidium iodide, 100µg/ml RNase A in PBS and incubated at room temperature for 1 hour, protected from the light. Cells were stored on ice, protected from light, until analysis on a BD Biosciences LSR Fortessa™

(MRC HGU Core FACS Facility).

118

2.3 Histology and tissue analysis

2.3.1 Fixation and tissue preparation

For histological analysis, isolated mouse tissues were fixed in either 4% paraformaldehyde

(PFA) or Bouins fixative solution overnight at 4°C under continuous movement. They were then washed twice overnight at 4°C in 1x PBS, and dehydrated through an alcohol (ethanol,

EtOH) increasing series into 70% EtOH. Tissues were embedded through 100% xylene and into 100% paraffin wax using the Tissue-Tek VIP vacuum infiltration processor (Sakura) and hand embedding into 100% paraffin wax molds using the Tissue-Tek Embedding Console

System (Sakura). Tissue was sectioned into 5µm sections using a microtome (Thermo Fisher), stretched by floating on a 47°C water bath, and baked onto glass slides overnight in a 65°C histology oven.

Prior to staining, tissue sections were dewaxed in 100% xylene for 5 minutes three times, and then rehydrated through a step-wise decreasing EtOH series into tap water, as has been previously described (Puchtler, Meloan, and Waldrop 1986).

2.3.2 H&E tissue staining

Haematoxylin and eosin staining is a broadly used tissue stain that is useful for identifying major morphological features in tissue sections. This was carried out on tissue section as described (Fischer et al. 2008), and summarized as follows. Slides were stained in haematoxylin for 4 minutes, washed in running tap water and differentiated in acid/alcohol

(1% HCl in 70% EtOH) for a few seconds. After another wash in running tap water, slides were blued by immersion in saturated lithium carbonate solution for 10 seconds, stained in eosin for 2 minutes then washed again. Slides were dehydrated by rinsing in 100% EtOH, then three subsequent 2 minute washes in 100% EtOH and then cleared in 3 washes of xylene for

5 minutes each. They were then mounted using DPX mountant solution (Sigma Aldrich) and allowed to set in a fume hood overnight before imaging.

119

2.3.3 Periodic Acid Schiff-Haematoxylin (PAS-H) tissue staining

Alternatively, in order to stain glycogen-rich tissue sections, Periodic acid Schiff staining was carried out on slides by oxidation in 0.5% periodic acid solution for 5 minutes. Slides were then washed in running tap water and stained in Schiff reagent for 15 minutes. After washing, slides were counterstained in Meyers haematoxylin for 1 minute, washed in running tap water then dehydrated in xylene and mounted as in section 2.3.2.

Both H&E and PAS-H stained tissue sections were visualised using brightfield microscopy, outlined in section 2.5.1.

2.4 Meiotic chromosome spreads

Chromosome spreads are a method by which cells can be spread on glass slides in order to count chromosomes and visualise chromatin-bound proteins by immunofluorescence. All culture media and wash steps were carried out on oocytes using documented embryo-safe solutions, acquired from the companies stated.

2.4.1 Prophase I oocyte spreads

In order to isolate oocytes at pachytene, I dissected out E18.5 ovaries as detailed in section

2.1.2, and incubated each pair of ovaries in hypotonic extraction buffer (30mM Tris, 50mM sucrose, 17mM trisodium citrate dehydrate, 5mM EDTA, 0.5mM DTT and 0.5mM PMSF at pH8.2) for 15-30 minutes. Pairs of ovaries were transferred into 20μl 500mM sucrose solution in a PCR tube, then pierced with a needle for approximately 5 minutes. During this time, tissue was macerated against the side of the tube in order to release oocytes from the ovarian tissue. Sterilised glass slides were pre-soaked in 1% PFA fixative (1% paraformaldehyde, 0.15%

Triton X-100 detergent at pH 9.2 – pH was set using titration of 1M boric acid) and 10µl of the sucrose-oocyte suspension was dropped onto the slide and zig-zagged down the slide in

120 order to spread the material out for ease of imaging. In this way, 2 glass slides could be generated for each pair of ovaries isolated from one embryo. Slides were incubated in a humid chamber overnight at room temperature, then allowed to air-dry the following day. Once dry, these slides were stored short-term at 4°C, or long-term at -80°C, before immunofluorescent staining as described in section 2.4.4.

2.4.2 Prometaphase I oocyte culture and spreads

In order to isolate oocytes at prometaphase I, pairs of age-matched female mice between 9-14 weeks old (2-3 months) were intraperitoneally injected, staggered one hour apart, with 0.1ml

5IU PMSG (pregnant mare serum gonadotrophin) in order to induce ovarian follicular development. After 42 hours, ovaries were dissected into warm (37°C) M2 media (Sigma

Aldrich). Within 30 minutes of mouse cull, oocytes were isolated from whole ovaries and transferred into pre-gassed M16 (Sigma) culture under embryo-safe filtered mineral oil

(Sigma Aldrich BioXtra, M3510) at 37°C and 5% CO2 – due to superovulation around 15-20 oocytes can be isolated from one pair of ovaries. Oocytes were cultured for 2 hours and those that developed through germinal vesicle breakdown (GVBD) in that time were separated and further cultured as above for 3 hours post-GVBD.

After culture, chromosome spreads were performed 3 hours post-GVBD. Oocytes were washed into fresh M2 media, then the zona pellucida (ZP) was removed using a 5-10 second wash in Acidic Tyrode’s solution (Sigma Aldrich, T1788) and the oocytes were washed through fresh M2 four times. Oocytes were individually incubated in 0.5% sodium citrate, continuously moving, for 2 minutes, then dropped onto a fixative-soaked glass slide (1% paraformaldehyde/PFA, 0.14% Triton X-100 detergent, 3mM dithiothreitol/DTT at pH 9.2 – pH was adjusted with 1M boric acid) and air-dried for around 1 hour. These slides were previously scored with a 1cm2 area in order to quickly locate the individual oocytes on each slide. After air-drying, 50µl of 0.4% Tween-20 was added to each slide and allowed to air-dry

121 overnight. All slides were stored at 4°C short-term or -80°C long-term until immunofluorescent staining, as described in section 2.4.4.

2.4.3 Metaphase II oocyte culture and spreads

For prometaphase II oocyte spreads, female mice were treated with PMSG as in section 2.4.2, but 48 hours after follicle stimulation, mice were intraperitoneally injected with 0.1ml 5IU hCG (human chorionic gonadotrophin) to induce ovulation. Oocytes were isolated from the ampulla 18 hours post-hCG treatment into 37°C M16 (Sigma Aldrich) media. They were treated in 0.5mg/ml hyaluronidase in M2 media (Sigma Aldrich) for 5 minutes to remove cumulus cells, and then washed three times in fresh M2. Chromosomes were spread in 1%

PFA fixative as described in section 2.4.2. All slides were stored at 4°C short-term or -80°C long-term until immunofluorescent staining, as described in section 2.4.4.

2.4.4 Immunofluorescent staining

Glass slides with fixed chromosome spreads were rehydrated in room temperature 1x PBS from either 4°C short-term or -80°C long-term storage. 50µl blocking solution (0.15% BSA,

0.1% Tween-20, 5% goat serum in PBS) was added to each slide in a humid chamber and a flexible plastic coverslip added to spread the block over the surface of the slide. Slides were blocked in a sealed humid chamber for 1 hour at room temperature. Primary antibodies were then added at the concentration described in Table 2.11, in 50µl blocking solution per slide.

Primary antibodies spread with a disposable coverslip again and incubated either overnight at 4°C or for 3 hours at room temperature, in a sealed humid chamber. After primary antibody incubation, slides were washed three times in PBS for 5 minutes and then incubated with secondary fluorescent (Alexa Fluor®, Invitrogen) antibodies (again, in 50µl blocking solution

122 at the indicated concentration, Table 2.11) and 20µg/ml DAPI (Biotium, 40043, stock at

10mg/ml) for 1 hour in a humid chamber, protected from the light.

Slides were then mounted in PD mountant solution (1mg/ml p-phenylene diamine in 90% glycerol), gently compressed under a glass coverslip for approximately 15 minutes, then sealed with clear nail varnish and stored at 4°C, protected from light, until imaging using epifluorescent microscopy, as detailed in section 2.5.2.

123

124

2.5 Microscopy and analysis

2.5.1 Brightfield microscopy

Brightfield images were acquired using a Micropublisher 5MP cooled colour CCD camera

(Qimaging, Surrey, BC, Canada) mounted on a Zeiss Axioplan II fluorescence microscope with Plan Apochromat objectives (Carl Zeiss, Cambridge, UK). Image capture was performed using scripts written by the IGMM-AIR (Advanced Imaging Resource) facility for

Micromanager (https://open-imaging.com/). Objectives and magnifications used were set on

Micromanager prior to image capture, so this was captured in the image metadata and scale could be directly applied to images during analysis.

2.5.1.1 Tiled brightfield imaging and analysis

Tiled whole tissue scanning for whole placenta measurements was carried out using

an Olympus BX51 upright microscope (Olympus KeyMed, Southend on Sea, UK)

with Marzhauser motorized stage (Marzhauser Wetzlar GmbH & Co, Wetzlar,

Germany). The system uses Olympus plan Apochromat and UplanS Apochromat

objective lenses with magnifications from 2x-40x. Images are acquired using an

Olympus XC10 colour CCD camera. Batch scanning of slides is carried out by the

attached SL50 slide loader unit. All hardware and tissue measurements were

controlled through the Olympus Dotslide VS-ASW software, maintained by the

IGMM-AIR facility.

2.5.2 Epifluorescence microscopy

After immunofluorescent staining, epifluorescent images were acquired using a Photometrics

Coolsnap HQ2 CCD camera and a Zeiss Axioplan II fluorescence microscope with Plan- neofluar objectives (Carl Zeiss, Cambridge, UK), a Mercury Halide fluorescent light source

125

(Exfo Excite 120, Excelitas Technologies) and Chroma #83000 triple band pass filter set

(Chroma Technology Corp., Rockingham, VT) with the excitation filters installed in a motorised filter wheel (Ludl Electronic Products, Hawthorne, NY). Image capture was performed using scripts written by the IGMM-AIR (Advanced Imaging Resource) facility for

Micromanager (https://open-imaging.com/). Objectives and magnifications used were set on

Micromanager prior to image capture, so this was captured in the image metadata and scale could be directly applied to images during analysis.

2.5.2.1 Z-stack epifluorescent microscopy

Where required, z-stack images were taken using a Photometrics Coolsnap HQ2

CCD camera and a Zeiss AxioImager A1 fluorescence microscope with a Plan

Apochromat 100x 1.4NA objective, a Lumen 200W metal halide light source (Prior

Scientific Instruments, Cambridge, UK) and Chroma #89014ET single excitation and

emission filters (Chroma Technology Corp., Rockingham, VT) with the excitation

and emission filters installed in Prior motorised filter wheels. A piezoelectrically

driven objective mount (PIFOC model P-721, Physik Instrumente GmbH & Co,

Karlsruhe) was used to control movement in the z dimension. Hardware control,

image capture and analysis were performed using Volocity (Perkinelmer Inc.,

Waltham, MA) or Micromanager.

Deconvolution, and subsequent 3-dimensional measurements between foci, were

carried out using a calculated PSF in either Volocity (Perkinelmer Inc., Waltham, MA)

or Imaris v9.0 (Bitplane, Oxford Instruments plc). Centromeric antibody signals were

designated as “spots” at which co-ordinates could be recorded, and 3-dimensional

Pythagorean Theorem was implemented in order to measure distances between these

pairs of co-ordinates.

126

2.5.3 Image analysis and quantification

Unless specified, all 2-dimensional image analysis and quantifications were carried out using

FIJI (Image J) with 64bit Java8. In order to avoid scoring bias, image names were randomized using a FIJI macro script written by James Crichton (MRC HGU), and analysed blind.

For axial distance measurements in prophase I images, 3-colour stack images were converted to 8bit and analysed using the NeuronJ FIJI plugin, developed by ImageScience (Meijering et al. 2004).

For antibody stain intensity quantification in stacked 3-colour prometaphase I cell images, each individual bivalent was thresholded by condensed DAPI signal, and fluorescent antibody signal intensity measured, then normalised against other protein levels within the same bivalent. Median bivalent protein intensity was calculated across each nucleus, then the mean of these medians was calculated for all nuclei from each mouse. For comparison between genotypes, these relative signal intensities were normalised again the mean for control animals processed on the same day within each repeat.

127

128

2.6 Protein analysis

2.6.1 Preparation from cell culture

2.6.1.1 Chromatin preparation

mESCs were cultured in 6-well plates and enriched at G2/M using double thymidine

treatment as in sections 2.2.3 and 2.2.3.2. At the DT+4 (G2/M) timepoint, cells were

washed in ice-cold PBS and harvested by scraping, then resuspended in 200μl Buffer

A+ (10mM HEPES, 10mM KCl, 1.5mM MgCl2, 0.34M sucrose, 0.1% Triton-X100, 1mM

DTT, 5mM sodium butyrate to prevent deacetylation of proteins, 1X Roche Complete

Mini protease inhibitor and 10% glycerol in dH2O) and lysed on ice for 5 minutes.

Nuclei were separated by centrifugation at 1300g for 4 minutes at 4°C, then washed

in 500μl Buffer A (as Buffer A+, but with no Triton-X100 detergent). Nuclei were

separated by centrifugation as above, and resuspended in 200μl Buffer B (3mM EDTA,

0.2mM EGTA, 1mM DTT, 5mM sodium butyrate to prevent deacetylation of proteins

and 1X Roche Complete Mini protease inhibitor in dH2O); this was incubated on ice

for 30 minutes in order to lyse nuclei. Chromatin was separated by centrifugation at

1700g for 4 minutes at 4°C, then washed in 500μl Buffer B and pelleted as above.

Chromatin was then resuspended in 100μl Laemmli 2x sample buffer, sonicated to

shear chromatin for 20 cycles (30 seconds on, 30 seconds off – BioRuptor®, Diagenode)

and then the protein sample was denatured at >95°C in a heat block for 10 minutes.

Protein was then used for Western blotting, or stored at -20°C until later use.

2.6.1.2 Co-immunoprecipitation

HEK293T cells were transformed with combinations of UBR2-FLAG, TEX19.1-GFP

or GFP expression plasmids, as described in section 2.2.2 and Table 2.6 and harvested

by trypsinisation 24 hours later (section 2.2.1). Anti-FLAG M2 beads (Millipore,

129

M8823) were provided in glycerol to preserve the antibody structure, and so were equilibrated in Dilution Buffer (DB: 10mM Tris pH7.5, 0.15M NaCl, 0.5mM EDTA,

1mM PMSF and 1% Roche Complete Mini protease inhibitor in dH2O) by washing

30μl beads per experiment three times in 500μl DB, with centrifugation at 13,000g between each wash, and then stored on ice until use.

Transfected HEK293T cells were washed in ice-cold PBS, then resuspended in 200μl

Lysis Buffer (as Dilution Buffer above, but with 0.5% NP40 detergent) and lysed on ice for 30 minutes – during this time, the cells were occasionally gently mixed to avoid settling and ensure full cell lysis. Cell debris was pelleted by centrifugation at 16,000g for 10 minutes at 4°C, and then the protein supernatant was transferred to a pre- cooled tube and diluted in 800μl DB to a total volume of 1ml. An aliquot of this was taken and added to Laemmli sample buffer to act as the input sample for each co- immunoprecipitation experiment. The diluted lysate was then combined with 30μl equilibrated FLAG-M2 beads and mixed by inversion at room temperature for 1 hour.

Beads, and bound proteins, were then kept on ice throughout three washes in 500μl ice-cold DB, then denatured and dissociated in Laemmli sample buffer at >95°C for 10 minutes. Protein samples were used immediately for Western blotting, or stored at -

20°C until later use.

2.6.1.3 Peptide pulldown assays

Peptides specific to the N-terminal sequence of Type I (Arginine) or Type II

(Phenylalanine) target proteins in the N-end Rule (see Chapter 5, Fig. 5.3A) were conjugated to resin via a C-terminal Cysteine using the Sulfolink® Peptide

Immobilisation Kit (Thermo Scientific, 44999) according to the manufacturer’s instructions.

130

Briefly, each peptide Type I (R), Type II (F) and -ve (G, negative control) diluted to

0.5mg/ml in 2ml of Sulfolink Coupling Buffer (50mM Tris pH 8.5, 5mM EDTA-Na) and reduced by addition of 100µl TCEP (0.5M Tris(2-carboxlethyl)phosphine,

Sulfolink BondBreaker®) and incubation at room temperature for 30 minutes. Resin was equilibrated by three washes in Sulfolink Coupling Buffer, then peptide solution added and incubated end-over-end at room temperature for 15 minutes, and upright for an additional 30 minutes to allow coupling of peptide to resin beads. Resin was washed four times in Sulfolink Wash Solution (1M NaCl, 0.05% NaN3), and two additional times in Sulfolink Coupling Buffer. Resin then incubated end-over-end in

7.9mg/ml L-Cysteine-HCl in Sulfolink Coupling Buffer for 15 minutes at room temperature, and for an additional 30 minutes upright. Resin washed twice with storage buffer (0.05% sodium azide in 1x PBS) and stored at 4°C upright in 2ml fresh storage buffer – this should be maintained at 50% resin, 50% storage buffer.

Peptide coupling efficiency was calculated by comparison of protein content of peptide solution before and after coupling (i.e. flowthrough), measured by Nanodrop spectrophotometry. Generally, ≥80% coupling efficiency was deemed sufficient for peptide pulldown assays.

Immediately prior to pulldown assay, resin was aliquoted – three 110µl aliquots of preclear and one 220µl aliquot of each Type I, Type II and -ve peptide per condition – and then prepared by removal of storage buffer and washed in 1ml ice cold dilution buffer (10mM Tris pH7.5, 150mM NaCl, 0.5mM EDTA, 1x Roche Protease Inhibitor in dH2O). All centrifugation of resin performed at 8200g and 4°C for 30 seconds. All resin was stored in 100µl dilution buffer on ice until use.

For peptide pulldowns, HEK293T cells were seeded at 1x105cells/ml in 3 wells of a 6- well plate and, after 24 hours were transfected for protein overexpression with either pCMV10-UBR2-T1-FLAG or pCMV10-UBR2-T2-FLAG, alongside the wild type

131

control plasmid pCMV10-UBR2-FLAG, as in section 2.2.2. At 48 hours, cells were

washed with ice cold 1x PBS and cells were lysed in-well with 200µl ice cold lysis

buffer (10mM Tris pH7.5, 150mM NaCl, 0.5mM EDTA, 0.5% NP-40, 1x Roche

Protease Inhibitor in dH2O) and incubated for 30 minutes at 4°C with continuous

movement. Lysates from three wells were combined and centrifuged at 16,000g for 10

minutes at 4°C to separate out cell debris. Supernatant was transferred into a new

tube and diluted to 3ml with ice cold dilution buffer to reduce the concentration of

NP-40 detergent. At this stage, a 50µl aliquot of each sample was taken to be the input

control - 50µl Laemmli sample buffer (Sigma-Aldrich, S3401) was added and the

sample heated at 95°C for 5 minutes prior to Western blotting. Each sample was split

(1ml x3) across 3 preclear resin tubes (after removal of excess dilution buffer) and

incubated end-over-end at 4°C for 1 hour. Supernatant was then recombined, and

then split again into Type I, Type II and -ve peptide-conjugated resin (1ml per tube).

Incubated end-over-end at 4°C overnight. Supernatant was removed and resin

washed 3 times in 750µl ice cold wash buffer (10mM Tris pH7.5, 150mM NaCl, 0.5mM

EDTA, 0.1% NP-40, 1x Roche Protease Inhibitor in dH2O). Beads were resuspended

in 50µl Laemmli sample buffer and heated at 95°C for 5 minutes while pulse-vortexing

to mix every 1-2 minutes. Finally, resin was then centrifuged and supernatant (in

Laemmli buffer) isolated for analysis by Western blotting.

2.6.2 Preparation from animal tissues

2.6.2.1 Chromatin preparation

Spleens were dissected from adult males and homogenised using razor blades and

filtered through a 100µm nylon cell strainer into 1.5ml PBS. Cells were pelleted at 100g

for 4 minutes at 4°C, washed in ice-cold PBS and then resuspended in 200μl Buffer

A+ (see section 2.6.1.1) – from this point, chromatin preparation carried out exactly

132

as in section 2.6.1.1. As spleen tissue has high cell number compared to cell pellet

isolated from cell culture, protein isolated from spleen chromatin was diluted 1:100 in

Laemmli after sonication and denaturation at >95°C in order to produce accurate

Western blots. Protein samples were then stored at -20°C until analysis by Western

blotting.

2.6.2.2 Peptide pulldown

For peptide pulldowns, whole testes were dissected from adult males and

homogenized using razor blades and 100µm nylon cell strainers, as above, into PBS.

Cells were spun down, and then incubated with continuous movement in lysis buffer

at 4°C, as in section 2.6.1.3. Peptide pulldowns carried out as above, and protein

samples analysed using Tris-Acetate Western blotting, described in section 2.6.3.

2.6.3 Western blotting and quantification

For most Western blotting experiments, samples were heated in Laemmli buffer >95°C for 5 minutes, then 15µl loaded into wells of a 4-12% Bis-Tris gel (Invitrogen NuPAGE system) alongside a broad range Color Protein Standard (NEB, P7719S) marker. Gels were run in SDS-

MOPS buffer (NuPAGETM, Invitrogen NP0001) at 150 volts (V/cm) in order to separate proteins by molecular weight. Gels were transferred onto nitrocellulose blotting membranes using the iBlot2 dry transfer system (ThermoFisher, IB21001). The membranes were washed in PBS-T (0.1% Tween-20 in 1x PBS); membrane block and all antibody incubations were in

5% milk (Marvel milk powder) in PBS-T. Primary and HRP-conjugated (horseradish peroxidase) secondary antibodies are all noted in Table 2.10.

For high molecular weight proteins (above 200kDa, i.e. UBR proteins), such as in peptide pulldown assays, Western blotting was modified to use 3-8% Tris-Acetate gels (Invitrogen

133

NuPAGE), which have a higher resolution of protein separation at high molecular weights and either broad range Color Protein Standard (NEB, P7719S) or high molecular weight

HiMark Protein Standard (Life Technologies, LC5699) marker in Tris-Acetate SDS running buffer (Invitrogen NuPAGE, LA0041). For these experiments, proteins were transferred onto

PVDF (polyvinylidene fluoride) membranes and blocking/antibody staining was carried out in 3% BSA (bovine serum albumin) in 1x PBS-T.

For visualisation of Western blotting, membranes were incubated in ECL (enhanced chemiluminescent reagent, Thermo Fisher SuperSignal system, 34577) and visualised using photosensitive film in a dark room (SLS Amersham Hyperfilm).

If quantification was required, developed film was scanned as a TIFF image file and protein band intensity quantified using the Gel Analyser tool in FIJI (ImageJ).

134

2.7 RNA analysis

2.7.1 RNA preparation

RNA was prepared from whole mouse tissue by homogenization in 500μl TRIzol™ reagent

(Invitrogen, 15596026), and dilution to 1ml of TRIzol. RNA was precipitated by chloroform:isopropanol, as per the manufacturer’s instructions, and then washed in EtOH and resuspended in RNase-free dH2O. Isolated nucleic acid was then subjected to DNase treatment to isolate only RNA, using the TURBO DNA-free kit (Invitrogen, AM1907), and then purified using the RNeasy Mini kit (Qiagen, 74104). For RNA preparation from cell culture samples, RNA was directly isolated and purified using the RNeasy Plus Mini kit

(Qiagen, 74134), which does not require a separate DNase step.

After RNA purification, cDNA synthesis was carried out using the SuperScript™ First-Strand

Synthesis System (Thermo Scientific, 11904018), in which 1µg purified RNA was added to 1µl dNTPs and 0.2µl random primers in a 13µl reaction, then incubated at 65°C for 5 minutes and chilled on ice for 2 minutes. Random primers were used to maximise transcript coverage and to ensure no genomic DNA contamination was present. The reverse transcription mix was set up as follows: 13μl prepared RNA, 4μl 5x First Strand Synthesis buffer, 1μl 0.1M DTT, 1μM

RNAse Inhibitor, 1μl Superscript III Reverse Transcriptase enzyme. This was incubated at

25°C for 5 minutes, 50°C for 60 minutes and 70°C for 15 minutes. For a negative control (no reverse transcriptase, NRT), reactions were performed without reverse transcriptase enzyme. cDNA samples were diluted 1:50 for use in qRT-PCR and stored at -20°C until use.

2.7.2 qRT-PCR

Quantitative real-time PCR (qRT-PCR) is an assay by which the level of expression of a gene in a cell or tissue can be measured by looking at the amplification rate of cDNA produced from the mRNA complement of that sample, so giving the level of mRNA transcripts present, and comparing to the amplification rate of known housekeeping genes that have constant

135 expression level. 1:50 cDNA samples were used, and a control sample was generated by combining a small amount of each of the neat samples and creating a dilution series 1:10, 1:50,

1:100 and 1:1000. This was used to generate a standard amplification curve for each primer pair, and so only primer pairs with an R2 value of >0.98 in their standard curve were used.

Primers, as noted in Table 2.4, were ordered from Sigma Aldrich with desalting purification, and were reconstituted to 100µM stocks, diluted to 1µM working solutions and stored at -

20°C. Each 20µl reaction was made up with 2.5µl of each 1μM primer, 10μl of SYBR™ qPCR master mix (Applied Biosystems, 4472908) and 5µl of 1:50 cDNA sample. A minimum of three biological repeat samples were used and, for every sample, three technical repeats were carried out. For controls, an NTC (no template control) was set up with 5µl dH2O instead of the cDNA sample, and an NRT (no reverse transcriptase control) was used as detailed in section 2.7.1. These reactions were carried out in 96-well opaque qRT-PCR plates (Applied

BioSystems) on a Roche Lightcycler 480. Samples were pre-incubated at 50°C for 2 minutes, then 95°C for 2 minutes. 50 PCR cycles were carried out, which consisted of: an amplification step at 95°C for 5 seconds, then 60°C for 30 seconds. qRT-PCR output is a Cp (crossing point) value, which indicates the number of cycles taken to reach a plateau level of transcript amplification. The quantity of transcripts present was determined using the slope of the standard curve calculated for each primer pair. This quantity was normalised to reference gene β-actin and then calculated for each genotype as relative to wild type control samples. A two tailed Student’s T-test was used to determine if gene expression changes between genotypes were statistically significant.

136

Target Gene/Locus Forward Reverse Notes Cdx2 5’–GGGTGGGGGTAGCAATACTT 5’–CCCTTCCTGATTTGTGGAGA Trophoblast Ctsq 5’–AGAACAGCTGGGGTAGACG 5’–GCCACATGCTTTCTTGTGAA Trophoblast Ifi44 5’–ATGGAGACCTGGTTCAGCAA 5’–ACTGTCATCCTTGGCCTTGA IFN response Ifitm2 5’–CTCTTCTTCAACGCCTGCTG 5’–TGGGGTGTTCTTTGTGCAAA IFN response Ifitm3 5’–TATGAGGTGGCTGAGATGGG 5’–TCACCCACCATCTTCCGATC IFN response Impact 5’–TTTATGGCGAGGAGTGGTGT 5’–GTTCGATAAGTCTGCGCGTT Imprinted Irf7 5’–GTGTACTGGGAGGTAGGCAG 5’–TGCCCAAAACCCAGGTAGAT IFN response LINE-1 ORF2 5’–GGAGGGACATTTCATTCTCATC 5’–GCTGCTCTTGTATTTGGAGCATAGA Retrotransposon Mcts2 5’–TGCAAATATCATGTGCCCGG 5’–TTCTCAATGCCGATCCCCTT Imprinted Mest 5’–AAGATCTGGGAAGGGCTGAC 5’–GGTTGATTCTGCGGTTCTGG Imprinted MMERVK10C 5’–AACTGGTCGCAGGAGCTG 5’–GGTAAAGTCTCCGAGGGTCA Retrotransposon Oasl1 5’–CTGCTCCAGGATCATGAGCT 5’–TTCTGCCACATTGTTGGTGG IFN response Oasl2 5’–CCGGCAATCTACGAGACTCT 5’–GCTCTCTGTACCCATCTCCC IFN response Peg3 5’–CCAAGAGAACTGCCTACCCA 5’–TCCCTTGCTCTTCCGGATTT Imprinted Prl8a8 5’–AAATTATGTGGGTGCCTGGA 5’–TCACGCAGAATTTGTCTGTTG Trophoblast Psg21 5’–ATAGGGACCCGTTGCTGTAC 5’–TGTCACCATCCAGTCACCAA Trophoblast Psg23 5’–TCCAGTCACCACAACACGTA 5’–ATCCTGTGTCCTCCTGTGTG Trophoblast Tmem173 5’–CCTCAGTTGGATGTTTGGCC 5’–CCTCAGTTGGATGTTTGGCC IFN response VL30 5’–CAGCCTTGGCCTGAGAGTTT 5’–CTTTCTGGGCTGAAGTCCCT Retrotransposon β-actin 5’–GGCTGTATTCCCCTCCATCG 5’–ACATGGCATTGTTACCAACTGG Housekeeping

Table 2.5 – qRT-PCR Primers

137

138

2.8 DNA analysis

2.8.1 Plasmids

Various plasmids were used for overexpression of proteins in cell culture (see section 2.2.2),

as well as conduction of CRISPR-Cas9 in ESCs (section 2.2.3) and in generation of a mutant

mouse (section 2.1.3); all are noted in Table 2.6.

Plasmid Features Antibiotic Resistance Source N-terminal 3xFLAG epitope tag p3xFLAG-CMV-10-UBR2 Ampicillin Adams lab (MRC HGU) Mouse Ubr2 CDS GFP epitope tag pCAAGS-TEX19.1 Mouse Tex19.1 CDS Kanamycin Adams lab (MRC HGU) CAG promoter pX330-U6-Chimeric_BB- WT CRISPR-Cas9 machinery Ampicillin Addgene CBh-hSpCas9 U6 promoter

Table 2.6 – Plasmids used in various molecular biology assays

2.8.1.1 Bacterial culture and DNA extraction

Plasmids were amplified by transforming into subcloning efficiency DH5α competent

E. coli cells (Invitrogen, 18265017) as per the manufacturer’s instructions. Cells were

stored at -80°C and thawed on ice prior to use. These were gently mixed and aliquoted

into 50µl, then incubated on one with 5-10ng of DNA for 30 minutes. The cells were

heat shocked for 20 seconds at 42°C, cooled on ice for 2 minutes and then 50µl

prewarmed SOC medium was added. Cells were incubated with 1g shaking at 37°C

for 1 hour, then mixed with 150µl LB () medium and plated onto pre-poured LB agar

plates (10g/L NaCl, 10g/L Bacto-tryptone, 5g/L, Yeast extract, 15g/L Difco Agar) with

either Ampicillin or Kanamycin antibiotic treatment for selection of successful

transformants. Plates were inverted and incubated overnight at 37°C then colonies

139

were picked the following morning. These were then outgrown over the following

night in 5ml LB medium with 50µg/L of the appropriate antibiotic (Ampicillin or

Kanamycin) at 1g. DNA was isolated using the MiniPrep column spin kit (Qiagen)

and yield was quantified using Nanodrop spectrophotometry (ThermoFisher

Scientific).

2.8.1.2 Site directed mutagenesis

In order to introduce the D118A and D233A point mutations into the pCMV10-UBR2

plasmid, I used Phusion Site-Directed Mutagenesis (Thermo Scientific F-541) as per

manufacturer’s instructions. Briefly, a forward and reverse primer pair were designed

to meet at their 5’ end (Table 2.6), with the intended point mutation in the centre of

primer F. They were ordered from Sigma Aldrich with 5’ phosphorylation in order to

aid direct circularisation of the linear PCR product, and PAGE (polyacrylamide gel

electrophoresis) purified.

Mutagenesis Forward Reverse Target UBR2-D118A 5’–[Phos]GACTGTGCAGTTGCCCCCACCTGTG 5’-[Phos]TCTGCAGGAGTATGTAGGTTCCCCCAC UBR2-D233A 5’–[Phos]GCTGTTTAATGCTGAGGTTCACACC 5’-[Phos]ATGCAGTAGTAGGTGTCACTCTTCT

Table 2.7 – Site-directed mutagenesis primers. The introduced point mutation is marked in red.

The intact 11kb plasmid was added to the following PCR reaction; 31.7µl dH2O, 10µl

5x Phusion buffer, 2.5µl of each primer at 10µM, 0.3µl pCMV10-UBR2 (10ng total

plasmid DNA), 1µl 10mM dNTPs, 1.5µl DMSO (final concentration at 3%) and 0.5µl

Phusion HotStart DNA polymerase. As pCMV10-UBR2-FLAG is a large (11kb) GC-

rich plasmid, DMSO was added outwith the standard Phusion protocol. The

140

extension time was calculated at 30 seconds/kb of plasmid. The PCR protocol was

followed as outlined in Table 2.7. A negative control with no Phusion polymerase was

also used to ensure no contamination was present in the sample and all plasmid was

incorporated into the PCR reaction.

Step D118A D233A 1) Initial Denaturation 98°C, 30 seconds 98°C, 30 seconds 2) Denaturation 98°C, 10 seconds 98°C, 10 seconds 3) Annealing 64°C, 30 seconds 70°C, 30 seconds 4) Extension 72°C, 6 minutes 72°C, 6 minutes 25 cycles 25 cycles Cycles Return to (2) 24 times Return to (2) 24 times 5) Final Extension 72°C, 10 minutes 72°C, 10 minutes

Table 2.8 – PCR protocols for site-directed mutagenesis.

After PCR amplification, the linear product was treated with 1µl FastDigest DpnI per

reaction and incubated at 37°C for 15 minutes to digest the methylated parental

plasmid DNA, and then the product was circularised using T4 DNA ligase; 2µl DpnI-

treated PCR product was combined with 2µl 5x T4 buffer, 5.5µl dH2O and 0.5µl T4

ligase, then incubated at room temperature for 5 minutes.

The mutated plasmids were then amplified using DH5α competent E. coli cells, as

outlined in section 2.8.1.1, and screened for successful clones by Sanger sequencing of

the whole ORF (open reading frame) to verify that no unintended mutation of the

Ubr2 CDS ORF had arisen.

141

2.8.1.3 RNA guide cloning for CRISPR-Cas9

In order to generate the CRISPR gRNA guide for injection into zygotes as detailed in section 2.1.3, and for generation of mESCs as in section 2.2.3.1, I introduced the guide sequence in Figures 5.5 or 5.16 into the plasmid pX330 (Table 2.5). This plasmid encodes the wild type CRISPR-Cas9 guide machinery required for binding of the guide to wtCas9 protein and successful CRISPR mutational activity.

Both sense and antisense guide sequences were ordered as 5’-phosphorylated single stranded DNA oligos (Table 2.9).

Sense 5’–[Phos]CACCGGCTGCATGCTGTTTAATGATG–3’ Antisense 5’–[Phos]AAACCATCATTAAACAGCATGCAGCC–3’

Table 2.9 CRISPR gRNA generation constructs. Additional bases indicated in red for generation of sticky ends required for cloning into pX330. [Phos] indicates a phosphate group on the 5’ end of the next base in the sequence.

The pX330 plasmid was digested with BbsI restriction enzyme (consensus 5’–

GAAGAC(N)2^ –3’) in a 150μl reaction, as follows: 6μg pX330 plasmid, 6U/ml BbsI

(NEB), 3U/ml Fast AP (NEB), 10X NEB RE CutSmart buffer and dH2O up to a total volume of 150µl. This was combined with alkaline phosphatase treatment (Fast AP,

NEB), incubated at 37°C for 30 minutes, and then the linearized plasmid was gel purified using the QIAquick column gel extraction kit (Qiagen). During this, the sense and antisense oligos were diluted in dH2O to 100µM, combined in a 10µl reaction, denatured at 95°C for 5 minutes and allowed to slowly cool to room temperature to anneal into a dsDNA structure. This was diluted 1:200 to a final

142

concentration of 0.5µM. The ligation reaction was carried out overnight at room

temperature using T4 ligase, as follows: 50ng linearized pX330, 1µl 0.5µM dsDNA

guide, 10X Roche ligation buffer, 1µl T4 DNA ligase (Roche), and dH2O to a total

volume of 20µl.

The ligation reactions were then used to transform into DH5α competent E. coli cells

for amplification, as outlined in section 2.8.1.1, and selected using Ampicillin

antibiotic resistance. Plasmid DNA was isolated using the MiniPrep column spin kit

(Qiagen) and verified using Sanger sequencing with a primer specific to the sequence

of the U6 promoter.

For CRISPR-Cas9 in mESCs, this plasmid could be direct transfected into cells, as

detailed in section 2.2.3.1. For injection into mouse zygotes, the gDNA construct

including the inserted guide sequence was amplified by PCR with a primer pair

specific to both the sequence of the T7 promoter and the guide, and this product was

used in in vitro RNA synthesis of the gRNA construct including the cloned guide

sequence using the T7 Quick High Yield RNA Synthesis kit (NEB), as per

manufacturer’s instructions. The RNA template was purified using the RNeasy Mini

column purification kit (Qiagen) and used in the injection mix described in section

2.1.3.

Primer Sequence Application U6 sequencing 5’–ACTATCATATGCTTACCGTA Sanger sequencing T7 Forward (guide) 5’–TGTAATACGACTCACTATAGGGCTGCATGCTGTTTAATGATG PCR amplification T7 Reverse (pX330 universal) 5’–AAAAGCACCGACTCGGTGCC PCR amplification

Table 2.10 Primers used in CRISPR gRNA generation. Inserted guide sequence shown in red.

143

Step T7-guide PCR 1) Initial Denaturation 95°C, 30 seconds 2) Denaturation 95°C, 30 seconds 3) Annealing 50°C, 60 seconds 4) Extension 68°C, 45 seconds 35 cycles Cycles Return to (2) 34 times 5) Final Extension 68°C, 5 minutes

Table 2.11 PCR cycling conditions for T7 pX330 guide amplification.

2.8.2 DNA sequencing

All DNA Sanger sequencing was carried out using either a 3130XL or 3730 Genetic Analyser

(Applied BioSystems, MRC HGU, Core DNA Sequencing Facility). This was used primarily for genotyping of point mutations in either CRISPR-Cas9 generated mice or cell lines, or for verification of plasmids and PCR products during molecular biology experiments.

144

Biological Target Level Application, Dilution Species Reference/Source ACA Primary Immunofluorescence, 1:50 Hu Antibodies, Inc. 15-235-0001 (anti-centromere) Immunofluorescence, 1:1,000 Gift from J. M. Peters AcSMC3 Primary Mo Western Blotting, 1:1,000 (Nishiyama et al. 2010) FLAG-M2 Primary Western Blotting, 1:5,000 Mo Millipore F1804 GFP Primary Western Blotting, 1:2,500 Rb Abcam, ab290 Histone H3 Primary Western Blotting, 1:10,000 Rb Abcam, ab1791 Lamin B1 Primary Western Blotting, 1:5,000 Rb Abcam ab16048 BD Pharmingen Mlh1 Primary Immunofluorescence, 1:50 Mo Biosciences, 551091 Immunofluorescence, 1:5,000 SMC3 Primary Rb Abcam, ab128919 Western Blotting, 1:1,000 1:500 Rb LS-Bio, LS-B175 SYCP3 Primary Immunofluorescence Santa Cruz Antibodies 1:200 Mo D-1 sc-74569 Gift from J. Wang TEX19.1 Primary Western Blotting, 1:5000 Rb (Yang et al. 2010) UBR2 Primary Western Blotting, 1:500 Gt Aviva, OAEB 00482 Alexa Fluor® 488 (green) anti-Hu Secondary Immunofluorescence, 1:500 Gt Invitrogen, A11013 488 (green) anti- Alexa Fluor® Secondary Immunofluorescence, 1:500 Gt Mo Invitrogen, A11017 Alexa Fluor® 488 (green) anti-Rb Secondary Immunofluorescence, 1:500 Gt Invitrogen, A11008 Alexa Fluor® 594 (red) anti-Mo Secondary Immunofluorescence, 1:500 Gt Invitrogen, A11020 Alexa Fluor® 594 (red) anti-Rb Secondary Immunofluorescence, 1:500 Gt Invitrogen, A11012 HRP anti-Mo Secondary Western Blotting, 1:5,000 Gt BioRad, 170-6516 HRP anti-Rb Secondary Western Blotting, 1:5,000 Gt Cell Signalling 7074S HRP anti-Gt Secondary Western Blotting, 1:5,000 Rb BioRad, 172-1034

Table 2.12 – Antibodies used for various assays. Rb – rabbit. Mo – mouse. Hu – human. Gt – goat. 594, 488 – fluorescent wavelength. HRP – horseradish peroxidase.

145

146

Chapter 3

Ubr2 in Oocyte Development and Chromosome

Cohesion during Female Meiosis

147

148

3.1 Introduction

Oocytes are particularly vulnerable to aneuploidy, as compared to the aneuploidy rates that arise as a result of spermatogenesis (Hunt and Hassold 2002). As discussed in Chapter 1, female meiosis is a much longer process than that of males, and this is thought to be a factor in the higher incidence of chromosome missegregation observed throughout oogenesis

(Hassold and Hunt 2001; Herbert et al. 2015). Meiosis begins in the embryonic ovary, and chromosomes align for homologous recombination to occur. After genetic crossover, the link between homologous chromosomes matures and is only held together by interhomolog conformation and protein complexes, rather than a physical DNA linkage (Bascom-Slack,

Ross, and Dawson 1997). This can be visualised cytologically as chiasmata, which hold each pair of homologs, each made up of a pair of sister chromatids, together in a bivalent structure.

Cohesins are a particularly important group of protein complexes during this process, as they associate with the chromosome axis and hold sister chromatids together, so stabilising the chiasma structure downstream of the crossover event (MacLennan et al. 2015). Meiosis arrests at this point, in what is known as the dictyate arrest phase, and oocytes do not resume meiosis until the ovary is mature, at the onset of puberty. Chiasmata maintenance through dictyate arrest is key, as physical links allow the two homologous chromosomes to behave as a single unit (the bivalent) on the metaphase I spindle. Homologs are then able to attach to opposite spindle poles during the first meiotic division. Premature chromosome separation at this stage is one of the leading causes of oocyte aneuploidy, as homologous chromosomes lacking chiasmata (achiasmate chromosomes) will behave independently on the metaphase I spindle and therefore have a higher chance of segregating incorrectly to the same spindle pole

(Lister et al. 2010). The length of the dictyate arrest in a mouse model can vary from around 2 months, at female sexual maturity, to over a year or longer. There is evidence to suggest that, with age, there are less cohesin complexes present along the chromosome axis, and this is assumed to lead to a loss of chromosome cohesion during the dictyate arrest period

(Cimadomo et al. 2018). This has also been observed in aged human oocytes (Tsutsumi et al.

149

2014), correlating with the increased incidence of aneuploidy with maternal age in humans.

In order to study the cohesin complex in a meiotic context, with a view to preventing aneuploidy and related fertility disorders, we must first understand the mechanisms that undergo cohesin maintenance, particularly around MI and the dictyate arrest. The cohesin complex has been most widely studied in mitotic cells, and so we can use our knowledge of the mitotic maintenance mechanisms in order to infer conclusions about meiotic cells. The structure of the cohesin complex, and its dynamics within both mitotic and meiotic cell cycle, are discussed in detail in section 1.3.

The cohesin subunit SMC3 is present in all cohesin complexes, in both mitotic and meiotic cells. During mitotic S phase, acetylation of SMC3 acts to recruit cohesin maintenance factors such as sororin, which antagonises the cohesin removal factor WAPL (Nishiyama et al. 2010;

Losada 2014), so maintaining cohesion throughout G2. In mitotic prophase, sororin phosphorylation allows removal of chromosome arm cohesin, and chromosomes are held together by centromeric cohesin until segregation at the metaphase-anaphase transition

(Brooker and Berkowitz 2014; Ladurner et al. 2016). Although there are many similarities between mitotic and meiotic cohesin complexes, arm cohesion must be maintained on the bivalent throughout prophase I and into metaphase I (MacLennan et al. 2015). Centromeric cohesion is maintained through the first metaphase-anaphase transition, only dissociating prior to the second meiotic division. SMC3 acetylation (formation of the AcSMC3 subunit) and sororin recruitment are active processes in the meiotic oocyte nucleus (Gómez et al. 2016;

Huang et al. 2016), and so AcSMC3 may also be important for cohesin maintenance in the oocyte. Given the elevated importance for prevention of aneuploidy in the germline, the mammalian oocyte may have evolved mechanisms to maintain acetylated cohesin function during the later stages of female meiosis, and study of these will give insights into the molecular basis of aneuploidy and subsequent subfertility.

150

Ubr2 has been previously implicated in genomic stability and homologous recombination in a mitotic context (Ouyang et al. 2006), and our group have shown that Ubr2 acts to specifically regulate the AcSMC3 subunit of chromosome-associated cohesin, again in a mitotic context

(J Reichmann et al. 2017). A major interactor of the UBR2 protein is TEX19.1, which is primarily expressed in hypomethylated tissues, including the ovaries, testes, and in embryonic stem cells (ESCs) in culture (Ollinger et al. 2008; Tarabay et al. 2013; Maclennan et al. 2017; J Reichmann et al. 2017). Our lab have outlined a role for Tex19.1 in protection of chromosome cohesion before the first meiotic metaphase: a Tex19.1-/- mouse has low levels of chromosome-associated AcSMC3, but not other cohesin subunits, and chromosome missegregation at the first meiotic division, resulting in increased incidence of aneuploidy after meiotic completion and subfertility. TEX19.1 and UBR2 have a 1:1 stoichiometric interaction, and there is evidence to suggest that TEX19.1 acts to regulate the activity of UBR2, which is an E3 ubiquitin ligase that is active in the N-end Rule pathway (J Reichmann et al. 2017).

Additionally, our lab have shown that, when ectopically expressed, TEX19.1 is able to regulate the Type II N-end Rule substrate binding activity of UBR2 – i.e. it regulates UBR2’s ability to target Type II substrates for N-end Rule ubiquitination (J Reichmann et al. 2017; T Tasaki et al. 2005). Therefore, I hypothesised that TEX19.1 acts to protect AcSMC3 by preventing UBR2 from turning over either AcSMC3 itself or an indirect substrate that has downstream effects to regulate AcSMC3 and the cohesin complex as a whole (Fig. 1.4 and section 1.2.2). One prediction of this model is that, if UBR2 negatively regulates AcSMC3 in mouse oocytes, then loss of UBR2 may result in increased levels of AcSMC3. This molecular role could potentially have implications for meiotic chromosome segregation.

In this chapter, I aimed to look at the role of Ubr2 in meiotic cohesin regulation using a

Ubr2-/- mouse and investigate whether this is antagonistic to the known cohesin regulator

Tex19.1. I showed that there is a reduction in chromosome cohesion in both Metaphase I and

Metaphase II, although there is no detected chromosome missegregation in the first meiotic

151 division. This correlates with a specific decrease in the acetylated subpopulation of chromosome-associated cohesin complexes. Unexpectedly, rather than generating an opposite phenotype to Tex19.1-/- oocytes, loss of Ubr2 phenocopies aspects of the Tex19.1-/- oocyte phenotype, which indicates that there is a cell-type specific modulation of the role of

UBR2 in regulating AcSMC3.

152

3.2 Results

3.2.1 AcSMC3 is specifically downregulated in Ubr2-/- metaphase I oocytes

The Ubr2 and Tex19.1 genes are coexpressed in the murine ovary, and a stoichiometric interaction has been observed between these proteins in vitro (J Reichmann et al. 2017; Pan et al. 2008; Celebi et al. 2012). Lack of Tex19.1 shows a vital role for chromosome cohesion and segregation in the murine oocyte, and this is linked to a specific reduction in axis-associated acetylated SMC3, but not in total cohesin levels (J Reichmann et al. 2017). As Ubr2 is known to negatively regulate AcSMC3 in mitotic tissues, I hypothesised that Ubr2 and Tex19.1 work antagonistically in the oocyte to regulate cohesin complexes containing the AcSMC3 subunit.

In order to investigate this, I first aimed to measure the level of chromosome axis associated

AcSMC3 in the oocyte. As meiotic oocytes make up a very small proportion of the ovarian tissue and any direct protein quantification using Western blotting would be obscured by levels of the mitotic cohesin complex, I isolated oocytes during the germinal vesicle stage (GV) from the ovaries of Ubr2-/- and Ubr2+/+ females 42 hours after PMSG hormone treatment, cultured them to 3 hours post-GVBD (germinal vesicle breakdown) and carried out metaphase I chromosome spreads on glass slides in 1% PFA fixative. As female mice are reported to lose cohesin from the chromosomes of their oocytes with age, young (2-3 months old) age-matched mice were used for these experiments. I carried out immunofluorescence against AcSMC3 in these chromosome spreads (Fig. 3.1A). Cohesin complexes, represented here by AcSMC3, are present along the chromosome axes that hold each pair of sister chromatids together and can be distinctly observed in the condensed bivalent chromosome structure (Fig. 3.1A, inset). Centromeres are immunostained with ACA (anti-centromeric antibody), which is isolated from patients with CREST autoimmune disease and is specific to centromeric kinetochore complex proteins (Fritzler, Kinsella, and Garbutt 1980; Blengini and Schindler 2018). DAPI is used to stain DNA and observe gross chromosomal morphology at metaphase I, which appears generally normal with lack of Ubr2. All chromosome spreads

153 for quantification were thresholded at 18 bivalents per nucleus, allowing for loss of 2 bivalents per nucleus as an artefact of the chromosome spreading process. When the intensity of

AcSMC3 immunofluorescent signal is quantified, normalised against intensity of the centromeric ACA signal per bivalent and calculated as a median intensity across all bivalents within each nucleus, Ubr2-/- oocytes have 61.6% (±4.2%) of wild type AcSMC3 intensity along the chromosome axes, which is a statistically significant decrease in the level of protein present (Fig. 3.1B, p≥0.001, Student’s T-test). Therefore, loss of Ubr2 results in a loss of chromosome axis-associated AcSMC3-containing cohesin complexes in the metaphase I oocyte.

In order to assay whether this observed phenotype indicates a specific role for Ubr2 in

AcSMC3 regulation, I also quantified total SMC3 protein in Metaphase I oocyte spreads (Fig.

3.2A), again by immunofluorescence with normalisation against ACA signal. As SMC3 is present in all cohesin complexes, this gives an indication of the level of total cohesin present along the chromosome axis within each bivalent structure. The mean level of SMC3 protein in Ubr2-/- chromosome spreads is 88.1% (±20.2%) of the wild type mean protein, but this is not a statistically significant change in protein signal intensity (Fig. 3.2B, p=0.59, Student’s T-test).

It should be noted that there is more variation in the quantification of total SMC3 as compared to the variation of AcSMC3 (Fig. 3.2A, B), due to the quality of antibody binding in these immunofluorescence experiments. ACA in particular was a variable signal in presence of the total SMC3 antibody, potentially due to interference between the two antibodies used in this experiment, and so SMC3 oocytes were quantified with a thresholded minimum of

ACA signal present.

Loss of Ubr2 results is a specific decrease in AcSMC3-containing cohesin complexes, which is not reflected in total SMC3 levels – i.e. there is no change in total axial cohesin at this stage of meiosis. Therefore, UBR2 appears to promote AcSMC3-containing cohesins along the metaphase I oocyte chromosome axis. Surprisingly, this decrease in AcSMC3 in absence of

154

Ubr2 is a phenocopy of the cohesin phenotype observed in the Tex19.1-/- oocyte, both showing a 40-50% reduction in AcSMC3 intensity (J Reichmann et al. 2017). However, my original hypothesis was that Ubr2 and Tex19.1 work antagonistically to regulate acetylated cohesin levels in the oocyte, and so I concluded that the TEX19.1-UBR2 interaction may instead be working together to promote AcSMC3, potentially promoting chromosome cohesion.

155

B AcSMC3 Antibody Signal

***

AcSMC3Signal Relativeto ACA Signal

156

Figure 3.1 (previous page) – AcSMC3 is reduced in the Ubr2-/- metaphase I oocyte

+/+ -/- A. Representative wild type (Ubr2 ) and Ubr2 prometaphase I oocyte chromosome spreads, immunostained for AcSMC3 in red and anti-centromeric antibody (ACA, or CREST) in green, with DAPI shown in blue to mark DNA. Inset images show an individual bivalent in 3-colour (left) and the AcSMC3 alone (greyscale, right) along chromosome axes. B. Mean AcSMC3 -/- immunofluorescence signal over 14 Ubr2+/+ (grey) and 13 Ubr2 (green) oocytes from 5 mice of each genotype. Median of signal within each chromosome normalised to ACA signal for each

+/+ oocyte, and means calculated for each mouse, then normalised to the Ubr2 same day control mean to control for day-to-day variation. n is the number of nuclei analysed. Statistical significance calculated using Student’s T-test, *** p<0.001

-/- Figure 3.2 (next page) – Total SMC3 is unaffected in the Ubr2 metaphase I oocyte

+/+ -/- A. Representative wild type (Ubr2 ) and Ubr2 prometaphase I oocyte chromosome spreads, immunostained for SMC3 in red and anti-centromeric antibody (ACA, or CREST) in green, with DAPI shown in blue to mark DNA. Inset images show an individual bivalent in 3-colour (left) and the SMC3 alone (greyscale, right) along chromosome axes. B. Mean SMC3 immunofluorescence

-/- signal over 14 Ubr2+/+ (grey) and 10 Ubr2 (green) oocytes from 4 mice of each genotype. Median of signal within each bivalent normalised to ACA signal for each oocyte, and means calculated for

+/+ each mouse, then normalised to the Ubr2 same day control mean to control for day-to-day variation. n is the number of nuclei analysed.

157

B SMC3 Antibody Signal

SMC3 Signal SMC3 to Relative ACA Signal

158

3.2.2 Loss of Ubr2 does not negatively affect cohesin functions in foetal prophase I

While the cohesin complex is vital for chiasmata maintenance during metaphase I, this phenotype is not often observed in a genetic model of cohesin null or mutant oocytes, as there are also vital roles during prophase I, including homolog synapsis, genetic recombination and formation of the chromosome axis (Anderson et al. 1999; Agostinho et al. 2016; Rong et al.

2016). In both male and female cohesin mutant meiocytes, it has been observed that the length of the synaptonemal complex at prophase I is altered with changes in cohesin levels, or the composition of different cohesin complexes present along the axis (Biswas, Stevense, and

Jessberger 2018). As Ubr2 has many wide-reaching roles across different tissues, it is possible that the AcSMC3 phenotype observed in metaphase I could be a result of cohesin defects during establishment of chromosome cohesion during prophase I.

Although no specific role for AcSMC3 has been identified in prophase I SC formation, and total cohesin levels are unaffected in a Ubr2-/- oocyte, it is prudent to ascertain that cohesin is functioning normally during prophase I in absence of Ubr2. In order to do this, I measured the lengths of chromosome axes, by immunofluorescence against SC lateral element protein

SYCP3 (Syrjänen, Pellegrini, and Davies 2014; Enguita-Marruedo et al. 2018), in fully synapsed prophase I Ubr2+/+ and Ubr2-/- oocytes. In females, prophase I is approximately synchronous during foetal ovary development, and the fully synapsed stage of prophase I (pachytene) can be isolated from E18.5 foetal ovaries. I carried out chromosome spreads from this tissue onto glass slides with 1% PFA fixative, performed immunofluorescent staining and measured the length of each fully synapsed SYCP3 axis (Fig. 3.3A). As there is variation in the length of each of the 20 murine chromosomes, and there is no reliable way to identify each individual chromosome in these chromosome spreads, I ranked these 20 axis measurements within each nucleus by length. The shortest chromosome was designated 1, and so on until the longest chromosome was designated 20 (Fig. 3.3B). I then grouped these into sets of 5 chromosomes with similar lengths within each nucleus, and showed that there is no significant difference

159 in the lengths of each of these 4 groups between the Ubr2+/+ and Ubr2-/- genotypes (shortest 1-

5: p=0.50, 6-10: p=0.49, 11-15 p=0.17, and longest 16-20 p=0.40, Student’s T-tests). Therefore, cohesin complexes are functional during foetal prophase I, and the cohesin defects observed at metaphase I are a result of misregulation of the acetylated cohesin complex before commencement of metaphase I, potentially during the dictyate arrest.

The cohesin complex also plays a role in regulation of genetic recombination rate

(Bhattacharyya et al. 2019; Covo et al. 2010) during foetal prophase I, and so I performed immunofluorescence in foetal prophase I chromosome spreads against MLH1. This is a mismatch repair protein that marks genetic crossovers, which occur during pachytene as a result of homologous recombination (S. M. Baker et al. 1996; Anderson et al. 1999) (Fig. 3.4A).

In order to assess the number of crossovers that occur in absence of Ubr2, I first counted MLH1 foci per nucleus. Pachytene is the longest stage of prophase I, and MLH1 focus formation and resolution occurs continuously throughout this stage (Zickler and Kleckner 1999). Therefore,

I thresholded these counts at a minimum of 15 MLH1 foci per nucleus in order to assess select nuclei at optimal stages of pachytene. The mean wild type number of MLH1 foci present in a pachytene oocyte is 20.35±0.67, and the mean number of foci present in a Ubr2-/- pachytene oocyte is 21.51±0.60; this is not a statistically significant difference, and the median number of foci for each genotype is 21 foci per nucleus (Fig. 3.4B, p=0.20, Student’s T-test). This indicates that recombination and genetic crossover rate are not perturbed in absence of Ubr2 at pachytene (prophase I).

160

Ubr2+/+ SYCP3 Ubr2-/- SYCP3

B Oocyte Prophase I Chromosome Lengths

m) µ 15

Axis Length ( Length Axis 10

+/+

5 -/-

0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 Chromosomes, Ranked by Size

-/- Figure 3.3 The synaptonemal complex has normal morphology in Ubr2 prophase I oocytes

+/+ -/ A. Representative Prophase I pachytene oocyte nuclei for Ubr2 and Ubr2 - genotypes immunostained for SYCP3 protein in grey. Scale bars are 5µm. B. Chromosome axis lengths during

+/+ -/- pachytene, ranked by increasing length, across 42 Ubr2 (grey) and 52 Ubr2 (green) nuclei isolated from the ovaries of three E18.5 foetuses for each genotype. Standard error is shown.

161

In order to assess the position of recombination and crossovers along the chromosome axis, I measured position of MLH1 foci relative to the centromeric end of each chromosome, as marked by higher DAPI intensity at centromeric heterochromatin (Fig. 3.4A). As crossover interference affects the position of neighbouring recombination points, I limited this analysis to chromosome axes containing only a single MLH1 focus (Foss et al., 1993; Hillers, 2004;

Anderson et al., 1999). The length of each axis was measured and divided into three segments: distal, interstitial, and proximal to the centromere. MLH1 foci occur at the distal end of the chromosome in 35.0% of Ubr2+/+ pachytene oocyte chromosomes, and in 32.6% of Ubr2-/- pachytene oocyte chromosomes. They occur in the interstitial segment in 40.5% and 42.6% of each, and proximal to centromeres in 26.4% and 22.6% of each genotype. The position of

MLH1 foci are not significantly changed between Ubr2+/+ and Ubr2-/- single-focus chromosomes during pachytene (Fig. 3.4C, p=0.87, χ2 test).

Taken together with Fig. 3.3, in which Ubr2 is shown to not be vital for chromosome axis length or SC formation, I conclude that Ubr2 does not have a specific role in cohesin regulation in prophase I and Ubr2-/- oocytes progress through foetal prophase I normally. Similarly, there is no perturbation of prophase I progression, SC axis length or recombination rates in a

Tex19.1-/- foetal prophase I oocyte (J Reichmann et al. 2017), indicating that the AcSMC3 phenotype observed in both genotypes results from defects that arise postnatally, either before or during metaphase I.

162

B Mlh1 Focus Counts C Mlh1 Focus Positions

100%

80%

60%

40%

20%

0% +/+ n=318 -/- n=358

Proximal Interstitial Distal

-/- Figure 3.4 Crossover frequency and position is unaffected in the Ubr2 pachytene oocyte A. Representative images of Prophase I oocytes in the fully synapsed pachytene stage, indicated by the synaptonemal complex component SYCP3, immunostained in red. MLH1 recombination marker is immunostained in green. DAPI is used to mark pericentromeric heterochromatin in +/+ -/- blue. Scale bars are 5µm. B. Number of MLH1 foci identified per Ubr2 (grey) or Ubr2 (green) pachytene nucleus, with a lower threshold of 15 foci to select out very early or very late pachytene nuclei. n is number of nuclei analysed, from three mice of each genotype. C. Positions of single MLH1 foci along the chromosome axis in pachytene, relative to the telocentric centromere – proximal is within the third of the axis closest to the centromere, interstitial is in the centre of the axis, and distal is within the furthest third of the axis from the centromere. n is number of single- crossover axes analysed, over a total of 31 nuclei from three mice of each genotype.

163

3.2.3 Ubr2-/- metaphase I oocytes exhibit defects in sister chromatid cohesion

The major function for cohesin when it binds to the chromosome axis during meiosis is in the cohesion of sister chromatids together throughout meiosis. Cohesion along sister chromatid arms holds the homologous bivalent together in MI, and sister chromatid cohesion is maintained until the second meiotic division by centromeric cohesion. Both centromeric and arm cohesion are intact throughout prophase I, and arm cohesion acts to stabilise chiasmata, holding the bivalent together through alignment of the spindle in metaphase I. At the first metaphase-anaphase transition, cohesin complexes are removed along the chromosome arms, but maintained at the pericentromeric region. Centromeric cohesion does not break down until the metaphase-anaphase transition of MII, so allowing sister chromatids to bi-orient and segregate in the second meiotic division.

3.2.3.1 Ubr2 in chromosome arm cohesion

In order to measure the extent of chromosome cohesion in the chromosome arms I

used the number and position of chiasmata as a measure of how stably the

chromosome arms are held together. One of the consequences of losing chromosome

arm cohesion is a concentration of chiasmata towards the telomeric end of

chromosome arms, and a subsequent reduction in chiasmata frequency, as chiasmata

are lost off the telomeric end. A possible reason why loss of arm cohesion can affect

the number and position of chiasmata is that arm cohesion on the telomeric side of

chiasmata may act to prevent movement of chiasmata towards the telomere, where

they could become terminal or potentially be lost (Fig. 3.5A). This is known as

chiasmata terminalisation (Jagiello and Fang 1979; Jessberger 2012), which is

associated with an increased risk of premature separation of the homologs before

alignment on the spindle. This has been linked to a loss of cohesin molecules bound

to the chromosome arms (E Revenkova et al. 2010; Garcia-Cruz et al. 2010).

164

To investigate this, I assessed the number and position of chiasmata in each chromosome bivalent, marked by DAPI staining, of metaphase I oocyte spreads at 3 hours post-GVBD isolated from young (2-3 months old) Ubr2+/+ and Ubr2-/- females, as in section 3.2.2 (Fig. 3.5B). In order to first assess the absolute number of chiasmata,

I counted the number of chiasmata per nucleus (Fig. 3.5C). The mean number of chiasmata present in a Ubr2+/+ metaphase I oocyte is 27.35, represented by a combination of single and multiple instances of chiasmata within each bivalent.

There is no statistically significant change in the mean number of chiasmata in a

Ubr2-/- oocyte, which have 28.24 chiasmata per nucleus (p=0.41, Student’s T-test), and so it does not appear that chiasmata are gained or lost in absence of Ubr2. However, a subtle movement or loss of chiasmata may not be detected this way, and so I also quantified the proportion of single to multiple (two or three, more than this is very uncommon) chiasmata within each bivalent (Fig. 3.5D). Of the 20 chromosomal bivalents within the Ubr2+/+ nucleus, 67.0% have a single chiasma, and 33.0% have two or more, whereas in the Ubr2-/- nucleus 58.3% of bivalent have a single chiasma and

41.7% have multiple. This indicates that there is no significant change between Ubr2+/+ and Ubr2-/- oocytes at metaphase I (p=0.24, Fisher’s exact test). This is not a statistically significant deviation from the wild type proportions, and so there is not a detectable change in the number of chiasmata present in absence of Ubr2.

However, when the position of these chiasmata are taken into account, there is a deviation from the wild type state. Mouse chromosomes are telocentric, meaning the centromere is at one end of the chromosome. I estimated the distance of chiasmata within each bivalent with a single chiasma from the centromeric end by dividing the length of the bivalent into approximately three sections: distal, interstitial or proximal to either pair of centromeres, and then judged the position of the chiasma by blind scoring. I restricted this analysis to chromosomes in which only one

165 recombination event has occurred, as crossover interference between multiple events on the same chromosome axis affects the position of neighbouring chiasmata (Foss et al. 1993; Broman et al. 2002; Hillers 2004). Chiasma in the centromeric-proximal third of the bivalent occur in 14.0% of Ubr2+/+, but in only 5.7% of Ubr2-/- bivalents. There is also a reduction in interstitial single chiasma this occurs in 48.1% of Ubr2+/+ bivalents and 38.2% of Ubr2-/- bivalents. Finally, there is an increase in the proportion of bivalents that have a single chiasma distal to the centromere, or at the telomeric of the chromosomes, which indicate a terminal association between homologs, rather than a true chiasma. This occurs in 37.9% of Ubr2+/+ bivalents and 56.1% of Ubr2-/- bivalents. Taken together, this is a significant deviation from the proximal- interstitial-distal proportions observed in the Ubr2+/+ metaphase I oocyte (p=0.0004,

χ2 test), indicating that there is a shift in chiasmata towards the telomeric end of each pair of homologous chromosomes in metaphase I, in absence of Ubr2.

Chiasmata positions in wild type Ubr2+/+ metaphase I oocytes were estimated based on the conformation of the single-chiasma bivalents. The proportion of proximal- interstitial-distal chiasma are not statistically different from the proportional position of single recombination events in prophase I cells, as more accurately measured by distance of MLH1 foci from the centromeric end of pachytene chromosomes in Fig. 3.4C (p=0.14, χ2 test), indicating that these methods are indeed comparable to measure crossover position across these different stages of meiosis.

The shift in chiasmata position in metaphase I Ubr2-/- oocytes could potentially reflect either an altered distribution of crossover recombination during foetal meiotic stages, or functions of Ubr2 in maintaining chiasmata position in post-natal oocytes. When taken together with the normal crossover positions observed as MLH1 foci in Ubr2-/- foetal pachytene (Fig. 3.4C), the skew of chiasmata towards the distal or telomeric

166

end of the chromosome observed in Fig. 3.5 indicates a loss of post-natal chromosome

cohesion in absence of Ubr2, resulting in chiasmata terminalisation. This is the first

indication that Ubr2 has a role in meiotic chromosome cohesion regulation,

potentially as a result of the loss of AcSMC3 at metaphase I, as described in section

3.2.1.

-/- Figure 3.5 (next page) – Ubr2 results in chiasmata terminalisation in metaphase I A. Schematic representing the movement of chiasmata as a result of cohesin (red) loss along the chromosome axis in a Metaphase I bivalent with one chiasma. Recombined homologs are represented in yellow and blue. B. Representative Metaphase I oocyte chromosome spreads

+/+ -/- isolated from Ubr2 and Ubr2 females. Inset: representative bivalents with a single chiasma or multiple (two) chiasmata, indicated by white arrows. C. Number of chiasmata present within the +/+ -/- 20-chromosome nucleus of Ubr2 (grey) and Ubr2 (green) Metaphase I oocytes. Mean values represented by black bars. n is the number of nuclei analysed over 5 female mice of each genotype. D. Proportion of single to multiple chiasmata identified within each 20-chromosome nucleus. n is

+/+ -/- the number of bivalents analysed over 25 Ubr2 and 21 Ubr2 nuclei from 5 female mice per genotype. Standard error is indicated. E. Position of chiasma relative to the centromeric end of +/+ -/- each bivalent. n is the number of single-chiasma bivalents analysed, from 10 Ubr2 and 7 Ubr2 female mice. Standard error is indicated. Statistical significance calculated using Student’s T-test (3.5C), Fischer’s exact test (3.5D) or χ2 test, * p<0.05, *** p<0.001 (3.5E).

167

A B

Single Chiasma Multiple Chiasmata

Loss of cohesin and movement of chiasmata +/+ Ubr2 DAPI

-/- Ubr2 DAPI

C Chiasmata Number per Nucleus D Proportions of Chiasmata E Chiasmata Positions

100% 100%

80% 80% *** ns 60% 60%

40% 40%

20% 20% * ns 0% n=500 n=420 0% * +/+ -/- +/+ n=284 -/- n=185 Ubr2 Genotype Ubr2 Genotype None Single Multiple Proximal Interstitial Distal

168

3.2.3.2 Ubr2 in pericentromeric cohesion

The second type of chromosome cohesion observed in the oocyte is pericentromeric cohesin. This is distinctly regulated in a meiosis-specific manner, as it must be maintained through the first metaphase-anaphase transition, holding sister chromatids together until separation in anaphase II, in MII. This is occurs through centromeric localisation of shugoshin 1 (SGO1) and protein phosphatase 2A (PP2A), which act at least in part to prevent separase interaction with cohesin by dephosphorylation of the meiotic Rec8 kleisin subunit (Kitajima, Kawashima, and

Watanabe 2004; Kitajima et al. 2006; Z. Xu et al. 2009). I measured the distance between the centromeres of each pair of sister chromatids in metaphase I, as indicated in Fig. 3.6A and B, using 3-dimensional z-stack images. The mean distance within each pair of sister chromatid centromeres at this stage of a Ubr2-/- oocyte is

0.50±0.01µm, which is not significantly changed from the mean of 0.53±0.03µm in a

Ubr2+/+ oocyte (Fig. 3.6C, p=0.17, Student’s T-test). This implies that centromeric cohesin protection mechanisms are intact at this stage of meiosis.

I also measured the distance between sister chromatid centromeres at Metaphase II, in which sister chromatids depend on pericentromeric cohesion (Fig. 3.7A and B). At this stage of meiosis, Ubr2-/- centromeres are 0.54±0.06µm apart, which is significantly more than the mean wild type centromeric distance measured at this stage,

0.19±0.02µm (Fig. 3.7C, p<0.001, Student’s T-test).

It should be noted that the mean sister centromeric distance in Ubr2+/+ is significantly smaller in metaphase II than in metaphase I (mean distances 0.19µm and 0.53µm, respectively: p<0.00001, Student’s T-test, Fig. 3.6B and Fig. 3.7B), indicating that there may be some change in kinetochore or centromeric structure between these stages, bringing the sister centromeres closer together in metaphase II. Sister kinetochores are mono-oriented in metaphase I, whereas they change in structure and become bi-

169 oriented in metaphase II, allowing for segregation of the two sisters to opposite poles of the oocyte in the second meiotic division (MacLennan et al. 2015). This structural change could attribute for the reduction in measurements between these stages, although the ACA/CREST (anti-centromeric antibody) signal used to mark centromeric foci is not specific for particular proteins within the kinetochore

(Blengini and Schindler 2018; Fritzler, Kinsella, and Garbutt 1980) and so I cannot infer what structural changes may have taken place. There is no significant difference between centromeric distances in the Ubr2-/- metaphase I and metaphase II oocytes

(mean distances 0.50µm and 0.54µm, respectively: p=0.52, Student’s T-test, Fig. 3.6B and Fig. 3.7B), indicating that perhaps the changes in kinetochore structure do not taken place in absence of Ubr2. Therefore, this could indicate a role for cohesin in kinetochore orientation between the two stages of meiosis.

Taken together, these data indicate that Ubr2 has a role in maintaining pericentromeric sister chromatid cohesion in the murine oocyte, potentially through regulation of the AcSMC3-containing subpopulation of cohesin complexes. It may also play a secondary role in centromeric or kinetochore structure, which is specialised for meiotic spindle attachment and sister chromatid orientation.

170

C Distance between sister centromeres at Met I

Figure 3.6 – Ubr2 does not regulate centromeric distance at metaphase I

+/+ -/- A. Representative Ubr2 and Ubr2 Metaphase I bivalents, white arrows indicate centromeres marked by anti-centromeric antibody (ACA, or CREST) in green; DAPI was used to mark condensed DNA, in blue. B. Schematic shows the pattern of cohesins along the chromosome axis in red, holding the recombined homologs together (yellow and blue). Pairs of sister chromatid centromeres are indicated by black brackets. C. Distances between sister chromatid centromere

+/+ -/- pairs in Ubr2 and Ubr2 Metaphase I (Met I) oocytes, as measured across three dimensions in z- stack images and calculated using the Pythagorean Theorem. Means are indicated as black bars.

171

C Distance between sister centromeres at Met II

***

Figure 3.7 – Ubr2 specifically regulates centromeric distance at metaphase II

+/+ -/- A. Representative Ubr2 and Ubr2 Metaphase II bivalents, white arrows indicate centromeres marked by anti-centromeric antibody (ACA, or CREST) in green; DAPI was used to mark condensed DNA, in blue. B. Schematic shows the pattern of cohesins along the chromosome axis in red, holding the sister chromatids together at the centromere (yellow and blue). Pairs of sister chromatid centromeres are indicated by black brackets. C. Distances between sister chromatid

+/+ -/- centromere pairs in Ubr2 and Ubr2 Metaphase II (Met II) oocytes, as measured across three dimensions in z-stack images and calculated using the Pythagorean Theorem. Means indicated as black bars. Statistical testing carried out using Student’s T-test, *** p<0.001

172

3.2.4 Ubr2-/- oocytes are not aneuploid

Defects in chromosome cohesion and misregulation of the cohesin complex during oogenesis are associated with chromosome missegregation and aneuploidy (E Revenkova et al. 2010; L.

Liu and Keefe 2008; Herbert et al. 2015). In order to ascertain whether the misregulation of cohesin and cohesion observed in absence of Ubr2 during metaphase I is sufficient for missegregation, I investigated a potential role for Ubr2 in ploidy. I isolated mouse oocytes at metaphase II from both Ubr2-/- and Ubr2+/+ females from the same colony 18 hours after treatment with hCG, which was administered 48 hours after PMSG for superovulation. As female mice are reported to lose cohesin and exhibit an increase in oocyte aneuploidy with age (L. Liu and Keefe 2008), I carried out these experiments on young (2-3 months old) age- matched mice. Oocytes were dissected from the ampulla and spread on glass slides in a 1%

PFA fixative, and then visualised with DAPI to stain DNA, and anti-centromeric antibody

(ACA) (Fig. 3.8A). At this stage, pairs of sister chromatids can be observed, held together at their centromeric end by cohesin molecules. These are classified as either euploid (an expected 40 sister chromatids, in pairs of 2), hypoploid (less than 40) or hyperploid (more than 40) and calculated the proportion of oocytes observed with each (Fig. 3.8B). 23.5% of

Ubr2+/+ oocytes are hypoploid, and this is likely a artefact of the nature of chromosome spreading, in which some chromosomes can be lost or not bind directly to the slide. I did not observe any hyperploid metaphase II oocytes isolated from Ubr2+/+, and so 76.5% of nuclei in this genotype were observed as being euploid, so having gone through no chromosomal missegregation during the first meiotic division. In the Ubr2-/- Metaphase II oocyte the incidence of hypoploidy is at a similar rate (21.4% of nuclei), and no hyperploidy indicative of bone fide chromosome missegregation was observed. There were no observed hyperploid cells;

78.6% of Ubr2-/- were observed as euploid, and so there is no significant change in the proportion of ploidies observed between these and the wild type control cells (p=0.74, Fisher’s exact test). No incidences of premature separation of sister chromatids in the metaphase II

173 spreads were observed (Fig. 3.8A), and therefore analysis of metaphase II oocytes suggests that loss of Ubr2 does not have a detectable effect on the fidelity of chromosome segregation during meiosis I.

Figure 3.8 (next page) Absence of Ubr2 does not affect ploidy in the metaphase II oocyte A. Representative Metaphase II oocyte chromosome spreads, immunostained for anti-centromeric antibody (ACA, or CREST) in green, and DAPI showing condensed DNA in blue. Scale bars are 5µm. B. Proportional counts of hypoploid, hyperploid or euploid nuclei identified in Metaphase II +/+ -/- oocytes. n is the number of nuclei analysed, over three Ubr2 and three Ubr2 female mice. Statistical testing was carried out using Fisher’s exact test.

174

175

176

3.3 Discussion

In this chapter I showed that Ubr2 is important in regulation of AcSMC3-containing cohesin complexes, and a mouse model lacking Ubr2 has a specific loss of chromosome-associated

AcSMC3, which is associated with cohesion defects in both metaphase I and metaphase II.

However, this does not result in reduced numbers of chiasmata or chromosome missegregation during the first meiotic division.

3.3.1 The TEX19.1-UBR2 complex in meiotic cohesin regulation

Both the Ubr2-/- and Tex19.1-/- oocytes have a 40-50% reduction in chromosome-associated

AcSMC3 at metaphase I, which is not observed in other cohesin subunits (J Reichmann et al.

2017). In both genotypes, there are chromosome cohesion defects: terminalisation of chiasmata, and an increase in centromeric distance at metaphase II. My original hypothesis for the role of the TEX19.1-UBR2 complex in the oocyte was that Ubr2 acts to, either directly or indirectly, remove AcSMC3-containing cohesin from the chromosome axis and that Tex19.1 antagonises this to protect the cohesin complex and prevent chromosome missegregation (Fig.

3.9A). Instead, TEX19.1 and UBR2 may be working together in this tissue to protect AcSMC3 from turnover (Fig. 3.9B). However, as UBR2 has been shown to negatively regulate AcSMC3 in mitotic tissues, both in presence and absence of TEX19.1, it is not clear how it could have switched so starkly from downregulation to upregulation. This could be due to a role in a meiosis-specific pathway of cohesin regulation that is not found in mitotic cells, but there are no current candidates for this and it will be necessary to study UBR2’s role in meiosis further, as discussed in sections 3.3.2 and 3.3.3. Finally, another explanation could be that TEX19.1 is antagonistic of UBR2’s role to regulate a cohesin regulatory mechanism – as was previously hypothesised. However, if UBR2 is able to also negatively regulate a redundant protein, for example another member of the UBR family which acts as an E3 ubiquitin ligase in the N-end

Rule pathway for ubiquitination and protein degradation, and if this is unable to be negatively

177 regulated by TEX19.1, the other UBR protein could act to negatively regulate AcSMC3, independently of TEX19.1. Further work will be required to ascertain whether there are any changes in levels of other UBR proteins in the oocyte, in absence of UBR2.

178

UBR2 AcSMC3

TEX19.1 protects Type II

N-end substrates

UBR2 AcSMC3 TEX19.1

UBR2 AcSMC3

TEX19.1 protects Type II N-end substrates Other UBR

Figure 3.9 – A model for activity of the TEX19.1-UBR2 complex in meiotic oocytes A. As discussed in Chapter 1, in presence of TEX19.1, in mitotic ESCs (embryonic stem cells) TEX19.1 binds to UBR2 and inhibits its turnover of AcSMC3, so protecting the cohesin complex and chromosome cohesion, potentially through inhibition of the Type II N-end Rule. B. In the oocyte, TEX19.1 and UBR appear to work together to promote AcSMC3, so acting as a cohesin- protection mechanism during meiotic arrest and progression. C. Another hypothesis is that UBR2 negatively regulates a protein redundant for its role in AcSMC3 removal in the oocyte, which is then upregulated in absence of UBR2. TEX19.1 is unable to inhibit the removal of AcSMC3 by this protein, which could be another component of the N-end Rule pathway, and so AcSMC3 is negatively regulated, even in presence of TEX19.1.

179

Although Ubr2-/- oocytes phenocopy the Tex19.1-/- loss of AcSMC3, there is a potential difference in the severity of chromosome cohesion defect between the oocytes of these genotypes, as there is aneuploidy detected in absence of Tex19.1 in both metaphase II oocytes and in the single-celled zygote after natural fertilisation, whereas I did not observe it at metaphase II in the Ubr2-/-. A Tex19.1-/- oocyte also has a more pronounced loss of cohesion during metaphase I, with loss of absolute chiasmata number per nucleus as well as the terminalisation also detected in Ubr2-/- cells (J Reichmann et al. 2017). The published data on

Tex19.1-/- metaphase I oocytes were from chromosome spreads that were analysed at 5 hours post-GVBD, whereas in my experiments Ubr2-/- metaphase I spreads were carried out at 3 hours post-GVBD. As discussed earlier, chiasma terminalisation is associated with loss of cohesin, which is assumed to occur during dictyate arrest, but chiasmata terminalisation may only occur once GVBD has taken place and the cell has exited quiescence (Jessberger 2012).

While this experimental difference may affect number of chiasmata, which may have terminalised but not yet been lost at 3 hours post-GVBD, this does not explain the increased incidence of missegregation at both the first and second meiotic divisions observed in Tex19.1-

/- genotype. Therefore, we must proceed with caution when interpreting the loss of AcSMC3 in absence of Ubr2. Although TEX19.1 and UBR2 interact in a 1:1 stoichiometric ratio in vitro (J

Reichmann et al. 2017), this does not indicate that the TEX19.1-UBR2 complex is as such in vivo, and there may be as-yet unidentified components of this complex that are vital for

AcSMC3 regulation and mediation of the TEX19.1-UBR2 interaction. In absence of one or the other protein, unidentified oocyte-specific interactors may act differently, so leading to variability in the phenotype observed.

180

3.3.2 Ubr2 in regulation of cohesin in the arrested oocyte

This chapter shows that UBR2 has a role in regulation of the cohesin complex between foetal prophase I and ovulatory metaphase I, specifically through regulation of the AcSMC3 subunit.

Between these stages of meiosis, the oocyte nucleus is in the dictyate arrest phase, in which chromosomes are quiescent and chromosome cohesion must be maintained for a long period of time (Herbert et al. 2015). Exit from dictyate arrest is modulated by steroidal hormone signalling (Sen and Caiazza 2013), and occurs in small groups of oocytes at ovulation. This stage of meiosis is very difficult to study as oocytes cannot be isolated from the ovary for chromosome spreads, and so the majority of work surrounding age-related cohesin loss focuses on either foetal prophase I, or immediately after dictyate arrest in metaphase I

(Herbert et al. 2015; E Revenkova et al. 2010; J. Lee et al. 2003).

Cohesin has been observed to be reduced at metaphase I after a longer dictyate arrest period, i.e. in oocytes isolated from older females, but the mechanism by which cohesin loss is mediated is unknown (Jessberger 2012). Selective pre-dictyate depletion of the cohesin kleisin subunit REC8 in mouse oocytes results in chromosome missegregation and aneuploidy (Burkhardt et al. 2016), indicating that chromosome-associated cohesin complexes are not turned over during the arrest. Genes encoding cohesin subunits are expressed during this time, although only at around 10% of the mRNA levels detected in prophase I (Hodges et al. 2005), and this is not sufficient to replace all chromosome- associated cohesin complexes that are loaded during prophase I (Burkhardt et al. 2016;

Tachibana-Konwalski et al. 2010). This low-level cohesin synthesis during dictyate arrest

(Hodges et al. 2005) may be sufficient to supplement depletion of cohesin in young oocytes, but potentially not over time. Additionally, artificial expression of separase enzyme during metaphase I demonstrates an increased vulnerability to centromeric cohesin removal with age

(Chiang, Schultz, and Lampson 2011), implying a decline in cohesin protection mechanisms over time. Separase is active in wild type oocytes during mitotic anaphase and acts to cleave the kleisin subunit of the cohesin complex (Brooker and Berkowitz 2014). It is negatively

181 regulated by both SGO1/PP2A and by the chaperon securin; upon satisfaction of the spindle assembly checkpoint (SAC), securin is phosphorylated and separase is able to cleave kleisin and remove the cohesin complex (Z. Xu et al. 2009; Clift, Bizzari, and Marston 2009; Holland et al. 2007; Luo and Tong 2018). Separase is thought to only be active during MII, prior to the second meiotic division, as sister centromere cohesion is important for sister chromatid cohesion during MI, preventing premature separation (Terret et al. 2003).

Other cohesin regulatory mechanisms that are active during various stages of meiosis include the ESCO1/2 acetyltransferases (Whelan et al. 2012; Brooker and Berkowitz 2014), the

WAPL/PDS5 complex (Nishiyama et al. 2010; Wolf et al. 2018), which is actively antagonised by SMC3 acetylation and recruitment of the cohesin maintenance factor sororin (Ladurner et al. 2016; Brieño-Enríquez et al. 2016), and HDAC8 deacetylase (Brooker and Berkowitz 2014).

None of these mechanisms have been shown to be perturbed or enhanced in the ageing oocyte, and so it is unclear whether these play a role in oocyte ageing and loss of chromosome cohesion. However, they would be likely targets for an oocyte-specific cohesin maintenance mechanism that may have adapted for the extended dictyate arrest period.

UBR2 is an E3 ubiquitin ligase, with demonstrated ability in protein degradation via the ubiquitin-proteasome system (UPS) (T Tasaki et al. 2005; Rageul et al. 2019; H. Xu et al. 2019).

Proteolysis is a vital component of all cells, as it carries out homeostasis of the proteome as well as regulation of protein levels under specific conditions (Tanaka 2009). Loss of the Ubr2 gene in the oocyte results in a specific decrease in the acetylated subpopulation of cohesin molecules (the AcSMC3 subunit), and so this implies that UBR2 acts to regulate a cohesin maintenance or removal mechanism, such as those described above, acting to control against cohesin depletion and protect AcSMC3 on chromosome axes.

It would be interesting to observe if there is any change to Ubr2 mRNA or protein levels over time in the ageing oocyte, which could indicate an essential role for the enzyme in prevention

182 of oocyte aneuploidy with age. Tex19.1 mRNA can be observed in foetal, perinatal and postnatal mouse oocytes (Celebi et al. 2012), and interestingly this expression is reduced in germinal vesicle oocytes from older mice, indicating a potential decline in the TEX19.1-UBR2 cohesin regulatory function over time (Pan et al. 2008). However, the only cohesin subunit thus far shown to decline with age is the kleisin REC8 (L. Liu and Keefe 2008; Chiang et al.

2010), and so it must be noted there is no direct evidence to suggest that AcSMC3 is a regulator of cohesin maintenance in an aged oocyte.

This is a different mechanism for the regulation of AcSMC3 as compared to mitotic tissues: chromatin-associated AcSMC3 is increased in both the Ubr2-/- mitotic spleen and in Ubr2-/-

ESCs (Reichmann et al. 2017, and unpublished data, Karen Dobie, MRC HGU). This indicates both a TEX19.1-independent and TEX19.1-dependent role, respectively, for UBR2 in AcSMC3 removal. The potentially different mechanism by which UBR2 acts in any of these three tissues is not understood. All known functions of UBR2 involve ubiquitination and protein turnover, and so identification of UBR2 ubiquitination substrates in these various tissues would indicate how UBR2 is able to impact cohesin levels, potentially indirectly through one of the cohesin regulatory mechanisms, and whether there is an oocyte-specific substrate or cofactor that acts to change its regulatory function in the meiotic cell. These hypotheses could be tested using single cell oocyte mass spectrometry, or protein interaction assays coupled with proteasomal inhibition.

3.3.3 Ubr2 in regulation of MII kinetochore distance

A secondary potential role for Ubr2 in the oocyte identified in this chapter is in the regulation of centromeric or kinetochore structure (section 3.2.3.2). The kinetochore is a large protein complex localised to the sister chromatid centromeres and, among other functions, acts as an attachment point for spindle microtubules during the two meiotic divisions. The meiotic kinetochore is specialised for meiotic spindle attachment and sister chromatid orientation,

183 which is functionally different from the mitotic kinetochore. In the first meiotic division, the kinetochore mono-orients, allowing sister chromatids to segregation to the same spindle pole.

Metaphase II involves kinetochore remodelling to bi-orient sister chromatids, so allowing segregation in opposite directions at the second meiotic division to generate a haploid oocyte pronucleus (MacLennan et al. 2015). Interestingly, the distance between sister chromatid centromeres, as marked by ACA, is reduced between metaphase I and metaphase II (Fig. 3.6B and Fig. 3.7B), implying that this remodelling for sister chromatid orientation can be observed by immunofluorescence. Lack of Ubr2 results in an increase in distance between kinetochores at metaphase II, as compared to metaphase II in presence of Ubr2, which appears to be maintained from the distances measured at metaphase I in the Ubr2-/- oocyte.

Cohesin regulatory mechanisms involving separase and the SGO2/PP2A complex have been implicated in regulating kinetochore structure and attachment to microtubule spindle fibres, in both mouse and D. melanogaster oocyte meiosis (L. I. Wang, Das, and McKim 2019; Rattani et al. 2013; Tang et al. 2016). This could indicate that either cohesin regulation is essential for kinetochore function, or that these complexes have a secondary role in regulation of kinetochore proteins directly. The kinetochore structure has been implicated in maintenance of pericentromeric cohesin complexes (Weber et al. 2004; Schalch and Steiner 2017; Miller et al. 2012), so I would hypothesise that SGO2/PP2A are essential for this cohesin-kinetochore function. Furthermore, the effect of Ubr2 loss on kinetochore distance at metaphase II implies that UBR2 protein acts to promote kinetochore remodelling and reorientation, potentially through regulation of cohesin or a cohesin-regulatory mechanism such as SGO2/PP2A. One flaw in this hypothesis is that there is not currently any identified role for AcSMC3 specifically in meiotic centromeric cohesin, and it is not linked to SGO2 or PP2A. However,

AcSMC3 has been shown to recruit sororin in mitotic cells (K Nasmyth and Haering 2009) and, although sororin appears to have some AcSMC3-independent roles in meiotic arm cohesin, it also has a potential function in centromeric cohesin, which requires PP2A

184 phosphorylation activity (Gómez et al. 2016), although this is SGO2-independent. Therefore, this role could involve AcSMC3 regulation, which I have shown here requires UBR2.

No aneuploidy at metaphase II was identified in the Ubr2-/- oocyte – i.e. in the first meiotic division. However, if UBR2 is able to regulate centromeric cohesin or kinetochore structure, this could result in defects in spindle microtubule attachment during MII, and potentially missegregation during the second meiotic division, which was not tested in this study.

Further work to assess the ploidy of single-celled zygotes, after the second division, could lead to insights into UBR2’s role in oocyte meiosis and kinetochore-spindle attachment. This could explain the subfertility previously identified in Ubr2-/- female mice (Y T Kwon et al.

2003), which does not correlate with the lack of MII aneuploidy identified thus far. This could also be studied in a more indirect sense by assessing fertility with a Ubr2-/- breeding strategy.

SGO2 and PP2A are required for centromeric cohesion and, as null mutants result in infertility in both males and females (Faridi et al. 2017; Tang et al. 2016), it would be difficult to separate out a kinetochore-regulatory function in MII from that of pericentromeric sister chromatid cohesion throughout MI. For example, oocytes isolated from mice null for Pp2a do not progress beyond the first meiotic division (Tang et al. 2016). Instead, the hypothesis that

UBR2 acts to regulate SGO2/PP2A could be approached from a Ubr2 angle, by identifying if these, or other related proteins, are candidate substrates for UBR2 binding and, potentially, degradation via the proteasome.

3.3.4 Summary

In summary, in this chapter I have shown that Ubr2 is required for maintenance of AcSMC3 between foetal prophase I and metaphase I, similarly to Tex19.1. This is associated with defects in chromosome cohesion – although Ubr2-/- oocytes do not exhibit MII aneuploidy, which has been observed in the Tex19.1-/- oocyte. As the proteins expressed by these genes have a

185 stoichiometric interaction and function together in a variety of tissues, this overlapping phenotype, although not a full phenocopy, implies that the TEX19.1-UBR2 complex acts to promote AcSMC3-containing cohesin complexes along the meiotic chromosome axis, perhaps through regulation of a cohesin regulatory mechanism. Another potential role for

Ubr2 also exists in kinetochore orientation, possibly through direct regulation of the cohesin complex – although this could simply act through a similar mechanism to that which has already been identified. Further work into the role of UBR2 in meiotic sister chromosome cohesion will require identification of UBR2’s N-end Rule ubiquitination substrates in the oocyte if, indeed, this phenotype does act through its canonical pathway.

186

Chapter 4

The Tex19.1-associated Role of Ubr2 in Embryonic

Growth and Placental Trophoblast-Derived Tissues

187

188

4.1 Introduction

In addition to the identified meiotic phenotypes in both male (Y T Kwon et al. 2003; Crichton et al. 2017) and female (see Chapter 3) Ubr2-/- mice, there is a reported female-specific perinatal lethality (Y T Kwon et al. 2003), which varies in penetrance depending on genetic strain background, but the cause of this lethality is not yet clear. Interestingly, a similar sex-specific lethality was observed in Tex19.1-/- mice (Yang et al. 2010; J Reichmann et al. 2013), which is potentially related to defects in placenta function during foetal development having more pronounced consequences for females (J Reichmann et al. 2013). TEX19.1 protein stoichiometrically interacts with UBR2 in a 1:1 ratio (J Reichmann et al. 2017), and in the testis

UBR2 has been shown to stabilize TEX19.1, indicating a functional interaction (Yang et al.

2010). Therefore, it is possible that the female-specific lethality in Ubr2-/- embryos is caused by loss of function of the TEX19.1-UBR2 complex in the placenta and that there may be some overlap between the Tex19.1-/- and Ubr2-/- phenotypes. In order to test this hypothesis, in this chapter I aimed to investigate whether the female-specific lethality reported for Ubr2-/- embryos is associated with defects in placenta development and function, and whether any defects in Ubr2-/- placentas phenocopy those present in Tex19.1-/- mice.

In absence of Tex19.1, embryos exhibit a growth restriction that arises between E12.5 and E14.5.

This is associated with a reduction in placental size, particularly in the trophoectoderm- derived junctional zone (Jz) (Reichmann et al., 2013; Tarabay et al., 2013, and Playfoot et al. 2019, manuscript submitted, MRC HGU). The placenta is a naturally hypomethylated tissue (Oda et al. 2013; Schroeder et al. 2015); global DNA methylation dynamics vary in both embryonic and extraembryonic tissues throughout development. There is a global hypomethylation event in the first few days post-conception, and by E2.5, the majority of DNA methylation is lost, only being maintained at DMRs (differentially methylated regions of the genome), including IAP and L1 retrotransposons, and imprinted loci (Fig 4.1A, and Playfoot et al. 2019, manuscript submitted, MRC HGU). Tex19.1 is highly expressed in the placenta (Celebi et al.

189

2012), and across hypomethylated stages of early development (Fig. 4.1B), before the placental structure itself forms. These include the E2.5 8-cell stage, the E3.5 blastocyst inner cell mass

(ICM) and trophectoderm (TE), and the E6.5 extraembryonic ectoderm (ExE). The TE and

ExE will go on to form components of the junctional zone and placental labyrinth (Rossant and Cross 2001), and remain globally hypomethylated (Oda et al. 2013; Schroeder et al. 2015), whereas the epiblast (Epi) will go on to form the major three layers of early embryonic tissue

(mesoderm, endoderm and ectoderm, Rossant and Cross, 2001) and so regains DNA methylation after implantation (Z. D. Smith et al. 2014; Schroeder et al. 2015) (Fig. 4.1A). The

Epi does not have high expression levels of Tex19.1, as compared to that detected in the other early embryonic tissues in Fig. 4.1B, which correlates with the need for Tex19.1 expression only in a hypomethylated context. Ubr2 is a ubiquitously expressed gene, and early developmental stages are no exception. Ubr2-encoding mRNA can be detected across all cell types in Fig. 4.1B, in both hypomethylated and methylated contexts, and it is most highly expressed alongside the highest expression of Tex19.1 detected, in ExE tissues that will go on to form the trophectoderm-derived placental tissues. Expression data presented here is in RPKM (reads per kilobase million), calculated from mapped raw RNAseq data to the mm10 genome assembly (GRCm38) and corrected for gene length in kb, kilobases (Z. D. Smith et al. 2017)

(GEO GSE84235, analysis carried out by Ian Adams, MRC HGU).

190

A Methylation during post-Fertilisation Development

Embryo

Trophectoderm

-derived

Methylation Placenta

DNA

ExE

TE 8-cell ICM Epi E2.5 E3.5 E6.5

B RNAseq Expression Levels at Developmental Stages

200

180

160

140

120

100 RPKM 80

0 8-cell ICM TE Epi ExE

Ubr2 Tex19.1

191

Figure 4.1 (previous page) Ubr2 and Tex19.1 are coexpressed in hypomethylated developmental contexts A. Schematic representation of global DNA hypomethylation throughout early post-fertilisation development. Both paternal and maternal alleles are globally hypomethylated following fertilisation, and all cells in the 8-cell stage (E2.5) and in the blastocyst (E3.5), including the ICM and TE, are hypomethylated. Exceptions to this include DMRs (differentially methylated regions) such as IAP and L1 retrotransposons and imprinted loci. DNA methylation is restored in embryonic lineages that are derived from the E6.5 Epi post-implantation, but the TE-derived placenta tissues including the ExE remain hypomethylated. B. RNAseq expression data from (Z. D. Smith et al. 2017) for the mouse genes Tex19.1 and Ubr2 in five stages of early embryonic development. RPKM (reads per kilobase million) calculated as described, by Ian Adams, MRC HGU. ICM: inner cell mass. TE: trophectoderm. Epi: epiblast. ExE: extraembryonic ectoderm. Schematic in A adapted from Playfoot et al. 2019, manuscript submitted, MRC HGU.

Tex19.1 has a variety of known roles in the placental trophectoderm-derived tissues, including retrotransposon repression and transcriptional repression of imprinted genes. These are thought to act through TEX19.1 protein’s interaction with the transcriptional corepressor

KAP1, which together recruit SETDB1 (Leung et al. 2014; Jang et al. 2018), a histone methyltransferase that lays down H3K9 trimethylation marks at specific genomic loci, so acting as a genome defence mechanism in a globally hypomethylated context (Oda et al., 2013;

Cheng, Kuo and Ann, 2014, and Playfoot et al. 2019, manuscript submitted, MRC HGU).

The placenta is one of the first organs to develop during embryogenesis, and it is vital that it is able to grow quickly and support development, acting as the conduit for nutrient and oxygen delivery from the maternal blood stream (Rossant and Cross 2001). Intrauterine growth restriction (IUGR) is defined as the smallest 10% of human births, and this has been associated with defects in placental size and limited nutrient transfer (Scifres and Nelson

2009; Woods, Perez-Garcia, and Hemberger 2018). Small placental size can arise as a result of a variety of different disorders in chromatin or genomic stability, including misregulation of imprinted genes (Tunster, Jensen, and John 2013). DNA hypomethylation at DMRs and

192 altered expression of imprinted genes observed in Tex19.1-/- mouse placentas bear some resemblance to those seen in the placentas of human pregnancies that exhibit IUGR

(Takahashi, Kobayashi, and Kanayama 2000; Lefebvre et al. 1998; McMinn et al. 2006;

Koukoura et al. 2011). For example, perturbation of the paternally expressed mouse gene Mest, associated with angiogenesis in placental layers (Mayer et al. 2000), leads to both pre- and post-natal growth restriction (Lefebvre et al. 1998), and this is significantly upregulated in absence of Tex19.1 in the mouse placenta (Playfoot et al. 2019, manuscript submitted, MRC

HGU). The human homolog MEST has also been identified in placental growth defects associated with IUGR, such as Silver Russell Syndrome, or SRS (Hannula et al. 2001).

The human paralog of mouse Tex19.1, TEX19, is not expressed in the human placenta – but the human gene UBR2 is (Fagerberg et al. 2014). Although functionally convergent, there are many key differences in the tissue structure between the human and murine placenta. However, much of the trophectodermal cell lineage is key to nutrient transport and invasion of maternal tissue in both layers, and many of the molecular pathways in this lineage are conserved

(Woods, Perez-Garcia, and Hemberger 2018). Therefore, if the mouse TEX19.1-UBR2 complex is vital for murine placental development and embryonic growth, this could potentially reflect a conserved role for human UBR2 in this tissue.

This chapter aims to investigate a potential role for the gene Ubr2 in murine placental development using a Ubr2-/- mouse, and look for any overlap with the placental phenotype of the Tex19.1-/- mouse (J Reichmann et al. 2013, 2017), which is a null mutant for the known UBR2 interacting protein TEX19.1. I demonstrate a novel role for UBR2 in the placenta, which associates with intrauterine growth restriction (IUGR), and begin to separate the functions of TEX19.1 and UBR2 in the placenta, further elucidating the role of UBR2 in this tissue.

193

194

4.2 Results

4.2.1 Ubr2-/- females on a C57BL/6J background are small and under-represented at weaning

Ubr2-/- mice have been previously reported to exhibit a female-specific embryonic lethality (Y

T Kwon et al. 2003). However, in that study, there was variable severity of the null mutation depending on strain background. Very few male or female Ubr2-/- pups survived in an inbred

129SvJxSvImJ background (1.58% and 1.64% each of all males and females on this background, respectively). Differing strain backgrounds can result in varying survival rates of developmentally challenged pups (Cranston and Fishel 1999; Aihara et al. 2007; J.-Z. Jin and

Ding 2015) and so, interestingly, the Ubr2-/- lethality was reported as female-specific when using either 129SvJxC57BL/6J or 129SvJxCD1 mixed backgrounds, and Kwon et al. were able to obtain male Ubr2-/- mice at an expected rate (23.7% and 24.5% of all males in each background, respectively). The Ubr2-/- mouse generated in our group is on a C57BL/6J inbred background, and so it cannot be assumed that the published embryonic lethality is present in our hands (Crichton et al. 2017). In our group, a premature stop codon was introduced at Cys-

121, which is within the UBR2 UBR box (Uniprot Q6WKZ8-1) and is encoded in exon 3 of the Ubr2 gene; the majority of the UBR box domain is encoded by exons 3 and 4 (T Tasaki et al. 2009). A silent XbaI restriction site was also introduced as a means to genotype the Ubr2-/- allele (Crichton et al. 2017). In contrast, Kwon et al. generated a truncated UBR2 protein in embryonic stem cells (ESCs) by deletion of exons 4 and 5, and partial deletion of exons 3 and

6, by introduction of a lacZ marker gene between the 7th codon of exon 3 and the 16th codon of exon 6. They did not detect any β-galactosidase activity, and UBR2 protein could not be detected in either ESCs, or in the liver or testis of mice produced by chimaeric introduction of these ESCs and backcrossing to produce a Ubr2-/- line (Y T Kwon et al. 2003).

Although the mutations introduced into the Ubr2 gene in these two mouse lines are different, both involve truncation of exon 3, which is known to encode part of the UBR box domain (T

195

Tasaki et al. 2009), and neither Ubr2-/- mouse lines produce any detectable UBR2 protein in the testis (Y T Kwon et al. 2003; Crichton et al. 2017). Therefore, it does not appear that there are any allele-specific differences in how the two mutant proteins are expressed, and I assumed any major differences noted between the two strains was a result of genetic background.

To test if female-biased embryonic lethality is present in the Ubr2-/- mice used here, I examined the breeding data obtained from heterozygous Ubr2+/- x Ubr2+/- crosses used during breeding to generate Ubr2-/- mice for the meiotic phenotypic analyses described in the previous chapter.

These animals were genotyped at 3 weeks post-partum (pp) by PCR against exon 3 (Fig. 4.2A) and the 310bp PCR product digested with the introduced XbaI restriction site, which produces 2 bands at 205bp and 100bp for a Ubr2-/- genotype, or both these and the wild type

310bp band for Ubr2+/- animals (Fig. 4.2B). Crosses between Ubr2+/- mice are expected to generate 25% Ubr2+/+, 50% Ubr2+/- and 25% Ubr2-/- animals, with half of the pups expected to be male and half expected to be female (Fig. 4.2C). However, at weaning, around 3 weeks post-partum, only 15% of pups are Ubr2-/- (53 of 344 mice born from heterozygous Ubr2+/- x

Ubr2+/- crosses); this is a statistically significant deviation from expected Mendelian ratios

(Fig. 4.2D, p=0.004, χ2 test), indicating that there is a lethality in our Ubr2-/- line. As the previously observed lethality is female-specific on some strain backgrounds, I separated the genotypes of the mice produced from heterozygous Ubr2+/- x Ubr2+/- crosses at 3 weeks post- partum by gender. Male Ubr2-/- mice make up 21% of males from heterozygous crosses at weaning. Wild type and heterozygote males make up 23% and 56% of all males. These proportions are not statistically significantly different from expected Mendelian ratios (Fig.

4.2D, p=0.4762, χ2 test). Interestingly, only 9.8% of female mice are Ubr2-/-. Wild type and heterozygote genotypes represent 26.0% and 64.0% of the female pups, respectively.

196

Therefore, female Ubr2-/- mice are present at a statistically significant sub-Mendelian ratio at weaning (Fig. 4.2D, p=0.0008, χ2 test).

Of the 173 females genotyped in our colony, a Mendelian ratio of heterozygote crosses would predict 43 Ubr2-/- females but only 17 are present, which is 9.8% of all females born – this is

39.5% of the expected number, which suggests that any lethality affects 60.5% of Ubr2-/- females. In the previous study on either 129SvJxC57BL/6J or 129SvJxCD1 mixed backgrounds, only 8.2%and 15.2% of expected Ubr2-/- females were weaned, respectively. The first is significantly more penetrant than the Ubr2-/- female penetrance observed from our heterozygous crosses (p=0.0049, χ2 test), and the second is not significantly more penetrant

(p=0.17, χ2 test). This implies a highly variable lethality, and a female-specific perinatal vulnerability

-/- Figure 4.2 (next page) Ubr2 female mice do not survive to adulthood at a Mendelian ratio.

-/- A. The Ubr2 allele has a premature STOP codon introduced into exon 3 at Cys-121, which is within the UBR box domain and will give a truncated protein product. A silent XbaI restriction site was introduced into exon 3 upstream of the STOP codon for genotyping. B. Representative gel -/- electrophoresis of genotyping in the Ubr2 colony. The PCR product (primers indicated in blue in

-/- A) produced by Ubr2 mice have two bands after XbaI cleavage: 210bp and 100bp. * non-specific

+/- +/- +/+ PCR band. C. Heterozygote crosses (Ubr2 x Ubr2 ) are expected to produce 25% Ubr2 , 50%

+/- -/- Ubr2 and 25% Ubr2 offspring. D. Pie charts showing the proportions of mice born from 45

+/- +/- heterozygote (Ubr2 x Ubr2 ) crosses. Absolute numbers shown (total mice: 344, 171 males and 173 females), and as a percentage of the total number of mice. All mice were genotyped at weaning (3 weeks post-partum). Also separated by gender. All statistical differences calculated using χ2 tests: ** p≤0.01. Schematic in A adapted from (T Tasaki et al. 2009).

197

A wild type

UBR2 protein

XbaI

-/- Ubr2 allele 1 2 3 4 5 – – – – – – – – – – – – – – 48 exons total Genotyping PCR primers

truncated

UBR2 protein Cys121-STOP

B C +/- +/-

♂ Ubr2 X ♀ Ubr2

+/+ +/+ +/+ +/+ +/+

+/ +/

- -

bp -

♂ +/- *

500 310 300 210 200 100 ♀ 100 +/-

198

Therefore, the lethality in Ubr2-/- mice on a C57BL/6J inbred background is female-biased. This is an incompletely penetrant phenotype in all strain backgrounds covered here, and the strain appears to affect the level of penetrance; in our C57BL/6J inbred mice, the penetrance of female lethality is significantly reduced as compared to one of the previously published lethalities.

This could reflect either an embryonic or perinatal lethality, which is prevalent before 3 weeks pp, i.e. age of weaning.

In the previous study, surviving adult Ubr2-/- females in the 129SvJxC57BL/6J background exhibited a 20% growth retardation (Y T Kwon et al. 2003). Adult growth retardation of around 8% has also been observed in Tex19.1-/- females (J Reichmann et al. 2013). Adult males do not have a growth retardation in either published phenotype. In order to investigate whether this growth retardation is present on our C57BL/6J inbred Ubr2-/- mice, I weighed surviving adult mice at 8-10 weeks of age. As growth rates differ by gender in the C57BL/6J background (Gall and Kyle 1968), I separated the weights to reflect this. Male Ubr2-/- adult mice are not significantly reduced in size, as compared to age-matched littermate Ubr2+/+ males

(Fig. 4.3A, p=0.58, Student’s T-test), suggesting that males have normal postnatal growth, as has been previously reported. Heterozygote males also do not deviate in size from their Ubr2+/+ littermates (p=0.74, Student’s T-test). However, surviving females are significantly reduced in size by 18% (Fig. 4.3B) compared to both age-matched wild type and heterozygous females

(p=0.006 and p=0.026, Student’s T-Test), which is similar to the previously observed 20% magnitude of growth retardation in Ubr2-/- females (Y T Kwon et al. 2003). Lack of Ubr2 during embryogenesis therefore leads to a female-specific growth restriction at weaning, in all surviving females.

199

A Male Weights at 8-10 weeks pp B Female Weights at 8-10 weeks pp

Ubr2 Genotype Ubr2 Genotype

-/- Figure 4.3 Surviving female Ubr2 mice have a growth retardation A. Beeswarm plots of male mouse body weights (g) at 8-10 weeks post-partum (pp), generated

+/- +/- from Ubr2 x Ubr2 heterozygote crosses. B. Beeswarm plots of female mouse body weights (g)

+/- +/- at 8-10 weeks post-partum (pp), generated from Ubr2 x Ubr2 heterozygote crosses. All statistical differences calculated used Student’s T-Test: * p≤0.05, ** p≤0.01.

4.2.2 Ubr2-/- embryos exhibit intrauterine growth restriction and a placental size defect

Placental growth is important to support the growth of the embryo in utero and placental

defects hinder this growth – this is known as intrauterine growth restriction (IUGR), which

is associated with murine embryonic lethality, smaller size at parturition and poor health

(Beauchamp et al. 2015; G. Li et al. 2017). In human IUGR cases, there is an increase in

neonatal or perinatal morbidity, and this may also be a factor in mouse models of IUGR

(Malhotra et al. 2019) – which could explain the perinatal lethality observed in our Ubr2-/-

females.

200

A E18.5 Embryo Weights B E18.5 Embryo Weights by Gender

* *** *** * **

C E18.5 Placenta Weights D E18.5 Placenta Weights by Gender Female Male

*** *** *** * * * *** *

Female Male

-/- Figure 4.4 – E18.5 Ubr2 mice have an embryonic growth defect associated with a placental defect.

A. Beeswarm plots of E18.5 embryo weights (mg) over 5 litters from Ubr2+/- mating pairs. Littermate wild type (+/+, n=13, grey), heterozygote (+/-, n=21, light grey) and homozygote null (- /-, n=11, green). Means shown as black bars. Due to Home Office regulations, these are combined head and body weights of decapitated embryos. B. Data from A separated into male and female E18.5 embryo weights (mg), means shown as black bars. C. Beeswarm plots of placenta weights (mg) of E18.5 embryos, as in A. Means shown as black bars. D. Data from C separated into male and female placenta weights (mg), means shown as black bars. All statistical differences calculated used Student’s T-Test: * p≤0.05, ** p≤0.01, *** p ≤0.001.

201

As the Ubr2-/- female embryos in (Y T Kwon et al. 2003) were smaller than their Ubr2+/+ littermates at early (E9.5) embryonic stages, an embryonic growth restriction appears be associated with the observed female-specific lethality in absence of UBR2. Female embryonic growth restriction could also potentially be linked to the observed female adult growth restriction phenotype. Around 50% of female E9.5 embryos were deemed to be developmentally arrested at around E7.5-like stages (Y T Kwon et al. 2003). However, the lesser affected E12.5 embryos were histologically analysed and no overt defects were observed in major organs during development. This indicates that UBR2 has a role in embryonic growth, potentially being so severe as to prevent developmental progression, but it is not required for major developmental processes. It should be noted that this variance in penetrance of the

Ubr2-/- phenotype has been observed across differing genetic backgrounds (129SvJxC57BL/6J or 129SvJxCD1 mixed backgrounds, Kwon et al. 2003) and in this study all mice were generated on a C57BL/6J inbred background.

Considering these data, I hypothesised that UBR2 has a role in placental development, and that this is linked to IUGR, which would present as an embryonic growth restriction. To assess this, I first measured the weights of dissected E18.5 embryos. Ubr2-/- embryos are significantly reduced in size; the mean weight of 11 individuals was 794.3±40.1mg (± standard error), 16.50% less than wild type littermate control weights (mean 951.2±46.2mg) and 20.50% less than heterozygous littermates (mean 999.1±33.6mg) (Fig. 4.4A, p=0.0197, p=0.0008 respectively, Student’s T-test). Additionally, there is no significant difference in embryo size between wild type and heterozygote littermates (p=0.3992, Student’s T-test). When separated by gender, this growth restriction can still be observed; female and male Ubr2-/- embryos have mean weights of 774.6±52.6mg and 774.3±42.2mg, respectively, and are significantly smaller than heterozygote littermates (Fig. 4.4B, p=0.0199, p=0.0032 respectively,

Student’s T-test), reduced by 21.1% and 21.7%, respectively. Although Ubr2-/- females weigh

17.0% less than the weights of wild type female littermates (mean weight 933.1±80.3mg), the

202 reduction in embryonic size is not significantly changed between these two groups (p=0.1210,

Student’s T-test). However, males are significantly smaller than wild type littermate males, which have a mean weight of 1035.2±32.4mg (p=0.0002, Student’s T-test), i.e. a 25.2% reduction in size; the scale of reduction in embryonic weight is slightly higher for male

Ubr2-/- embryos than female Ubr2-/- embryos.

Therefore, loss of Ubr2 results in a growth restriction in both male and female embryos at E18.5.

As female adult Ubr2-/- mean weight is 82.1% of mean female wild type weight (Fig. 4.3B), and at E18.5 the mean female Ubr2-/- embryos are 83.0% of wild type littermate weight (Fig. 4.4B), intrauterine growth defect persists into adulthood in Ubr2-/- females. Although they have not caught up with littermates, there is no proportional loss of size and so females do not have a substantial lack of postnatal growth. Males, however, are able to catch up to their Ubr2+/+ littermates (Fig. 4.3A).

As discussed previously, embryonic growth restriction can be associated with placental defects, and a restriction of nutrient flow to the developing placenta. As the Tex19.1-/- embryonic growth restriction is associated with a reduction in placental size, I hypothesised that this may also be the case in the Ubr2-/- embryonic phenotype. In order to ascertain if the embryonic growth restriction is associated with defects in placental function, I measured the weights of placentas dissected from E18.5 placentas. Ubr2-/- placentas have a mean weight of

60.5±2.6mg, and are significantly smaller than the placentas of both wild type (88.9±2.6mg, smaller by 32.03%) and heterozygote (80.9±1.4mg, smaller by 25.30%) littermates (Fig. 4.4C, p=0.0000005, p=0.000001 respectively, Student’s T-test). Interestingly, although there is no difference between the wild type and heterozygote embryo sizes reported above, the heterozygote placentas weigh 9.0% less than their wild type counterparts, which is a statistically significant reduction in weight (p=0.0136, Student’s T-Test) and implies a potentially dose-dependent phenotype.

203

As both female and male embryos are affected at E18.5, I hypothesised that, when separated by gender, this placental defect is also not sex biased (Fig. 4.4D). Firstly, female placentas have a mean weight of 54.4±3.8mg, and are significantly reduced in size by 37.3% compared to the placentas of wild type littermates (mean weight 86.7±2.5mg), and reduced by 25.3% in comparison to the placentas of heterozygote littermates, mean weight 72.8±5.9mg

(p=0.000006, p=0.025 respectively, Student’s T-test). Male Ubr2-/- placentas are also affected; they have a mean weight of 62.6±4.5mg, which is significantly smaller than wild type littermates’ placentas by 29.3% (mean weight 88.6±3.4mg), and heterozygote littermates’ placentas by 20.5%, mean weight 78.8±3.5mg (p=0.0005, p=0.016 respectively, Student’s T-

Test). The dose-dependent effect of partial loss of UBR2 in a Ubr2+/- placenta appears to only affect female heterozygotes, which are significantly smaller than wild type female placentas by 16.0% (p=0.048, Student’s T-Test). These are, however, still significantly larger than Ubr2-

/- female placentas, suggesting that, if placental size is causative of the embryonic growth defect, there is a threshold at which the reduced placental size begins to affect the embryo.

Therefore, there appears to be both a placental and an embryonic growth defect in a Ubr2-/- mouse. Although only Ubr2-/- females are affected by perinatal lethality and post-natal growth restriction, there is no sex bias in placental or embryonic size at E18.5. There also appears to be a female-specific dose-dependent role of Ubr2 in the placenta, as Ubr2+/- female placentas are reduced in size. As female Ubr2+/- embryos are unaffected, this could point to a causative role for placental size in the embryonic growth restriction, and a threshold at which this becomes deleterious.

204

-/- Figure 4.5 (next page) Ubr2 mice exhibit embryonic growth defects earlier in development

A. Beeswarm plots of E12.5 embryo weights (mg) over 12 litters from Ubr2+/- mating pairs. Littermate wild type (+/+, n=33, grey) and heterozygote (+/-, n=40, light grey) and homozygote null (-/-, n=22, green) embryos. Means shown as black bars. B. Beeswarm plots of E12.5 embryo weights (mg) of placentas from, as in A. Means shown as black bars. C. Representative paired E12.5 embryos and placentas. Scale bars are 1mm. D and E. Beeswarm plots of littermate wild type (+/+, n=11, grey) and heterozygote (+/-, n=14, light grey) and homozygote null (-/-, n=9, green) E10.5 embryo and placenta weights (mg), over 4 litters from Ubr2+/- mating pairs, as in A and B. Means shown as black bars. All statistical differences calculated used Student’s T-Test: * p≤0.05, *** p ≤0.001.

205

A E12.5 Embryo Weights B E12.5 Placenta Weights

* *** *

C +/+ +/- -/-

D E10.5 Embryo Weights E E10.5 Placenta Weights

206

4.2.3 The Ubr2-/- and Tex19.1-/-placental defects arise at a similar developmental timepoint

The embryonic growth defect associated with a placental size defect indicates potential

IUGR and, if this is the case, the placental size defect should precede the embryonic growth defect. Kwon et al. (2003) detected a growth restriction at E9.5 in absence of Ubr2, however due to the differences in penetrance of the Ubr2-/- lethality across different strain backgrounds,

I surmised that this may result in a difference in the timing of growth restriction as well. To investigate at what stage the placental defect first appears, I dissected embryos at E12.5, i.e.

12.5 days into embryonic development, and weighed both embryo and placenta. Kwon et al.’s study was limited to embryonic development only, and so measurement of placental weights at E12.5 will also give an insight into the staging of Ubr2-/- placental defects and how this relates to embryonic growth. There is no difference in the embryonic or placental weights between males and females at E18.5, and the phenotype only becomes sexually dimorphic in adulthood, exhibiting a female-specific lethality and growth restriction, and so for these earlier stages I grouped genders together.

Ubr2-/- E12.5 embryos have a mean weight of 78.9±2.5mg, which is subtly but significantly smaller than wild type littermates, which have a mean weight of 86.3±1.6mg (Fig. 4.5A, p=0.0118, Student’s T-test), which is a difference of 7.4mg, or an 8.5% reduction. The placentas of Ubr2-/- E12.5 animals have a mean weight of 69.4±3.4mg, and are significantly smaller than wild type littermates, mean weight 84.1±2.6mg (Fig. 4.5B, p=0.00009, Student’s T-test); this is a 17.5% decrease in placental size. Interestingly, at E18.5 the difference in placental weight was a 37.3% reduction between Ubr2+/+ and Ubr2-/- animals, which is a larger magnitude of size defect and implies a potentially progressive size restriction over time, between E12.5 and E18.5.

Ubr2+/- E12.5 embryos have a mean weight of 84.9±2.1mg, and this is not significantly different from wild type littermate embryos (Fig. 4.5A), which is similar to the observed E18.5 embryos.

Their placentas have a mean weight of 75.7±2.7mg, significantly smaller than wild type

207 littermates (10.1% reduction, Fig. 4.5B, p=0.0284, Student’s T-Test). Again, this is similar to the heterozygous placenta size difference that can be observed in E18.5 Ubr2+/- placentas (Fig.

4.4C), and implies a dose-dependent role of Ubr2 in the placenta, which could indicate a causative role for placental size in the embryonic growth restriction.

In the Tex19.1-/- phenotype, there is no clear embryonic size defect at E12.5, and the placental growth defect does not begin until after E10.5 (Chris Playfoot, unpublished data, MRC HGU).

To ascertain if the Ubr2-/- placental defect has similar timing, I also dissected and weighed placentas and embryos at this earlier stage of development. There is no significant reduction in E10.5 Ubr2-/- embryonic weights (mean 13.5±1.3mg) compared to wild type (mean 13.0±1.1mg) or heterozygote (mean 13.7±1.3mg) littermates (Fig. 4.5D, p=0.762 and p=0.916 respectively,

Student’s T-test). This indicates that the embryonic growth defect begins to present between

E10.5 and E12.5, which is slightly earlier than the size defect recorded in the Tex19.1-/- embryo

(Chris Playfoot, unpublished data, MRC HGU). E10.5 Ubr2-/- mean placental weight is

33.3±2.6mg, which is not a detectable change from either wild type (mean 34.5±1.7mg) or heterozygote (mean 32.0±1.5mg) littermates’ placentas (Fig. 4.5E, p=0.8725 and p=0.2565 respectively, Student’s T-test), implying that the defect in placental growth also presents between E10.5 and E12.5 – this is the same stage as the Tex19.1-/- placental defect (Chris

Playfoot, unpublished data, MRC HGU).

Heterozygote placentas are also not significantly smaller than wild types (p=0.2515, Student’s

T-test), and so the potential dose-dependent phenotype in these placentas presents at the same developmental stage as the null phenotype.

Although around 50% of E9.5 female embryos have been previously reported as developmentally arrested at an E7.5-like stage (Y T Kwon et al. 2003), I did not observe this in any of the E10.5 or E12.5 (Fig. 4.5C) embryos isolated on this strain background, which are

208 also grossly morphologically normal (Fig. 4.5C). This indicates that there may be molecular phenotypes at this stage that are not penetrant in our Ubr2-/- mouse.

Again, this published phenotype is variable and so ack of an observable phenotype here could be a result of changes in penetrance. Taken together, the data presented in this section (Fig.

4.5) indicate that both Ubr2-/- males and females have an embryonic growth defect that presents between E10.5 and E12.5. This growth defect is associated with a smaller placenta, indicating a potential intrauterine growth restriction caused by lack of nutrients during embryogenesis. The Ubr2-/- placental size defect also presents between E10.5 and E12.5, which is similar to the Tex19.1-/-placental phenotype, a related gene whose protein product directly interacts with UBR2. The Tex19.1-/- placental phenotype has many superficial similarities to the timing and IUGR of the Ubr2-/- placental phenotype.

For all further experiments I assumed E10.5 to be prior to the placental phenotype, as there is no detectable size defect, and used E12.5 placentas to study the Ubr2-/- placenta in more detail.

4.2.5 Ubr2 has a role in trophoblast-derived layers of the placenta

The placenta is made up of three tissue layers: the embryonic labyrinth, the central junctional zone and the maternal decidua (Rossant and Cross 2001); see Chapter 1 for more detail on placental structure and development. Tex19.1 is expressed in trophectoderm-derived cell types, which are most abundant in the junctional zone (Jz) and labyrinth (Lb). In a Tex19.1-/- placenta, there is a specific reduction in the trophectoderm-derived cell types of the Jz, leading to a reduction in the size of this tissue layer (J Reichmann et al. 2013). Although the size of the Lb is unaffected in the Tex19.1-/- placenta, there is also a loss of specific cell-type markers in this tissue layer, indicating a role for the gene in both trophectoderm (TE)-derived layers (Chris

Playfoot, unpublished data, MRC HGU). As TEX19.1 and UBR2 interact and have overlapping phenotypes in the testis (Yang et al. 2010; Crichton et al. 2017), I hypothesized that defects in

209 a Ubr2-/- placenta would also overlap the Tex19.1-/- phenotype, i.e. defects in TE-derived tissue layers, primarily the junctional zone (Jz). This is characterised by glycogen-rich cell types and so to distinguish the three layers in histological sections of whole Ubr2-/- E12.5 placenta tissue,

I stained these cell types with PAS-H (periodic acid Schiff-Haematoxylin), which stains glycogen-rich cells that are abundant in the Jz, such as spongiotrophoblast glycogen cells

(SGCs) (P. M. M. Coan et al. 2006).

Despite the smaller size of homozygote null placentas, the tissue sections have grossly normal morphology (Fig. 4.6A). To assess this in more detail, I measured the size of each tissue layer and calculated the proportional contribution of these to the overall placenta size (Fig. 4.6B).

Due to the small size of the placenta and the nature of histological sections, the absolute size of the whole placental tissue measured was variable depending on the section taken for measurement, and so this was carried out as a proportion of whole tissue section size. In the

Ubr2-/- placenta, the Lb is only 29% of the total placenta size, which is a statistically significant

1.27-fold decrease (p=0.042, Student’s T-test) when compared to the wild type E12.5 labyrinth, which makes up 37% of total placenta size. The Jz and the maternal decidua (Md) make up

28% and 43% of the total Ubr2-/- placenta, respectively, which is not a statistically significant change in contribution of either layer (p=0.37 and p=0.57, respectively, Student’s T-test). In the wild type E12.5 placenta, Md is 39% of the total placental size and the Jz makes up 24%.

These are similar to previously reported proportions of placental tissue layers at this stage (P.

M. Coan, Ferguson-Smith, and Burton 2004). The growth defect in E12.5 Ubr2-/- placentas therefore appears to be accompanied by defects in the size of the trophectoderm-derived Lb.

The Tex19.1-/- phenotype has a specific junctional zone reduction and has no change in labyrinth contribution to the total placental cross-sectional area, and so this seems to be a

TEX19.1-independent role of UBR2.

210

-/- Figure 4.6 (next page) Ubr2 placentas are specifically reduced in size in the trophectoderm-derived labyrinth layer. A. Representative 5µm transverse tissue sections of whole fixed E12.5 placentas, stained with PAS to mark glycogen-rich cells of the junctional zone. Md: maternal decidua, Jz: junctional zone, Lb: labyrinth. Scale bars are 1mm. B. Proportional contribution of placental layers to the total tissue size, as measured in histological sections such as A, across 3 biological repeats. All statistical differences calculated using Student’s T-test * p<0.05

211

+/+ -/- A Ubr2 Ubr2

Md Jz

Lb Md

B Placental Tissue Layer Contribution

100%

80% *

60%

40%

ProportionalContribution 20%

0% +/+ -/- Ubr2 Genotype Maternal Decidua Junctional Zone Labyrinth

212

Although the Jz is the only layer whose size is specifically affected in absence of Tex19.1, there are differences in the expression of trophectoderm-specific genes in both the Jz and the Lb. In order to assess whether similar defects are present in Ubr2-/- placentas, I isolated whole placenta RNA at E12.5, generated cDNA and performed qRT-PCR (quantitative real time

PCR) to investigate the expression level of placental cell-type specific marker genes that are known to be perturbed in the Tex19.1-/- placenta (Fig. 4.7). Cdx2 is a homeobox transcription factor that is expressed across the trophectoderm lineage, in cells in both the Lb and Jz

(Sferruzzi-Perri et al. 2009), and so I used this as a general trophectoderm marker gene to look at the cell composition in both tissue layers. Expression is slightly decreased, but not significantly in the absence of Ubr2 across the whole placenta (p=0.06, paired Student’s T- test). As the RNA for these qRT-PCR experiments is isolated from whole placenta at E12.5, the larger junctional zone may obscure any compositional changes in the smaller labyrinth layer, which could explain why there is no detectable decrease in Cdx2 in these placentas.

Pregnancy-specific glycoproteins Psg21 and Psg23, are both highly expressed in both the spongiotrophoblast and in trophoblast giant cells in the junctional zone (Wynne et al. 2006) , and are significantly downregulated in the Tex19.1-/- placenta (J Reichmann et al. 2013). The expression of both of these genes is significantly decreased in the Ubr2-/- placenta (Fig. 4.7, p=0.039, p=0.005 respectively, paired Student’s T-tests). S-TGC (sinusoidal trophoblast giant cell) specific gene Ctsq (Simmons et al. 2008) and trophectoderm-specific gene Prl8a8 are both specifically expressed in these trophectoderm-derived cell types of the junctional zone, and are also both downregulated to varying degrees in the Tex19.1-/- phenotype at E14.5 (Reichmann et al., 2013 and Playfoot et al. 2019, manuscript submitted, MRC HGU). In the E12.5 Ubr2-/- placenta, expression of both of these genes is also significantly decreased (Fig. 4.7, p=0.033 and p=0.045 respectively, paired Student’s T-tests). Therefore, Ubr2 is required for expression of a subset of TE-derived cell-type specific genes that also require Tex19.1 for their normal

213 expression levels in the developing placenta, and these genes are characteristic of both the Lb and Jz, indicating a role for Ubr2 in both tissue layers.

Expression of Cell-Specific Markers in Whole E12.5 Placenta

* ** * *

Cdx2 Psg21 Psg23 Ctsq Prl8a8

Ubr2 Genotype

-/- Figure 4.7 The Ubr2 placental defect can be characterised by cell-specific compositional differences in both the labyrinth and junctional zone tissue layers.

-/- Relative gene expression levels measured by qRT-PCR in the whole E12.5 Ubr2 placenta over 5 repeats, corrected against expression of housekeeping gene β-Actin and relative to gender-matched and paired littermate wild type controls. Black bar is the mean, relative to wild type mean. All statistical differences calculated using a paired Student’s T-test, * p<0.05, ** p<0.01

Taken together, these data suggest that UBR2 has a role in development of the placental labyrinth and, although the junctional zone does not contribute any less to the total placental size than that in Ubr2+/+ placenta (Fig. 4.6B), the expression of some trophectoderm-specific genes expressed in the junctional zone is perturbed and so UBR2 appears to have a role in

214 regulating the development of trophectoderm-derived cells in the placenta. Although TEX19.1 does not have a specific role in labyrinth growth, the compositional differences in the junctional zone that arise in the Ubr2-/- phenotype overlap those observed in the Tex19.1-/- placenta at E14.5, indicating that Ubr2 may have both Tex19.1-dependent and Tex19.1- independent roles in the development of the placenta.

4.2.6 Placental Tex19.1-/- retrotransposon derepression does not act through UBR2

Tex19.1 has previously been shown to regulate expression of retrotransposon RNAs LINE1 and

MMERVK10C in the testes (Ollinger et al. 2008; Judith Reichmann et al. 2012) and in the placenta (J Reichmann et al. 2013; Tarabay et al. 2013). Ubr2-/- testes de-repress the same

MMERVK10C retrotransposon RNAs that are de-repressed in the Tex19.1-/- placenta, and contain no detectable TEX19.1 protein, suggesting that Ubr2 is required for Tex19.1-dependent regulation of MMERVK10C retrotransposon RNA in this tissue. MMERVK10C is unchanged in the Tex19.1-/- placenta, but UBR2 may be required for TEX19.1-dependent regulation of retrotransposons in this tissue, i.e. LINE-1 and VL30. Although the compositional differences identified in both the Jz and Lb (Fig. 4.7) may affect the proportion of hypomethylated cells present in the Ubr2-/- placenta, so potentially obscuring a role in derepression of retrotransposons, these compositional differences can also be observed in the Tex19.1-/- placenta, which do exhibit an upregulation of LINE-1 and VL30. Therefore, I hypothesised that any changes in retrotransposon expression levels detected here, reflecting those that are regulated by Tex19.1 in the placenta, would also be dependent on Ubr2 for their repression in this tissue.

I performed qRT-PCR to quantify the level of RNA present in the tissue for three retrotransposons (LINE-1 Orf2, VL30 and MMERVK10C – Fig. 4.8A). In absence of Ubr2, there is no change in level of either LINE-1 Orf2 or VL30 RNA (p=0.982 and p=0.976

215 respectively, paired Student’s T-tests). The abundance of the control retrotransposon

MMERVK10C does not change in Ubr2-/- placentas either (p=0.26, paired Student’s T-test).

Thus, TEX19.1’s role in regulating retrotransposon RNA levels is independent of UBR2 in the placenta.

Interestingly, E12.5 Tex19.1-/- placentas also exhibit a modest increase in expression of interferon-stimulated genes and activation of innate immune response signaling

(unpublished data, Chris Playfoot, MRC HGU). The activation of this innate immune response is partly dependent on the cytosolic DNA sensor STING (unpublished data, Chris

Playfoot, MRC HGU), but it is not clear whether the increased innate immune signaling in

Tex19.1-/- placentas is triggered by retrotransposon de-repression or through other mechanisms.

Increased expression of retrotransposons has been linked to an elevated interferon (IFN) response in a variety of tissue contexts, including a specific link between LINE-1 derepression and an IFNβ response in the testes (Yu et al. 2015). As Ubr2-/- placentas appear to phenocopy some aspects of the Tex19.1-/- placenta phenotype, but do not de-repress retrotransposons, I tested whether innate immune responses were activated in Ubr2-/- placentas independently of derepression of these retrotransposons. The interferon-stimulated genes activated in Tex19.1-/- placentas are Tmem173, Ifitm2, Ifitm3, Oasl1, Oasl2, Irf7 and Ifi44. However, most of these genes are not detectably upregulated in Ubr2-/- placentas (Fig. 4.8B). These data suggest that, in contrast to the Tex19.1-/- placenta, an innate immune response is not activated in Ubr2-/- placentas. Irf7 is modestly downregulated in the absence of Ubr2 (Fig. 4.8B, p=0.025, paired

Student’s T-Test). This specific effect on only one gene within this group could be attributed to the changes in cell composition across the placenta at this stage of development.

Additionally, the absence of both an immune response and retrotransposon derepression in

Ubr2-/- placentas indicates that both these aspects of the Tex19.1-/- placenta phenotype are

216 independent of Ubr2, and that these may be linked, although there is no direct evidence that one is causal of the other.

-/- Figure 4.8 (next page) The Ubr2 placental defect does not entirely phenocopy the

-/- Tex19.1 placenta

-/- A. Relative retrotransposon expression levels measured by qRT-PCR in the whole E12.5 Ubr2 placenta (green), relative to gender-matched and paired littermate wild type controls (grey), over five gender-matched and littermate paired repeats. B. Relative interferon response gene expression -/- levels measured by qRT-PCR in the whole E12.5 Ubr2 placenta (green), relative to gender- matched and paired littermate wild type controls (grey), over five gender and littermate paired repeats. All expression levels normalised to actin expression and relative to wild type means. * Paired Student’s T-test, p<0.05

217

A Expression of Retrotransposons in the Whole E12.5 Placenta

LINE-1 Orf2 VL30 MMERVK10C

Ubr2 Genotype

B Expression of Interferon-β Response Genes in the Whole E12.5 Placenta

Tmem173 Ifitm2 Ifitm3 Oasl1 Oasl2 Irf7 Ifi44

Ubr2 Genotype

218

4.2.7 Tex19.1-mediated repression of imprinted loci is independent of Ubr2

Our group has recently shown that some of the molecular defects present in Tex19.1-/- placentas are related to the physical interaction of TEX19.1 with an E3 ubiquitin ligase KAP1, which acts as a transcriptional co-repressor to recruit the histone methyltransferase SETDB1 and

H3K9me3 chromatin modifications to specific genomic locations including imprinted genes

(Playfoot et al. 2019, manuscript submitted, MRC HGU). As a result, Tex19.1-/- placentas exhibit reduced amounts of H3K9me3 at specific imprinted genes (Mest, Mcts2 and Impact) and transcriptionally de-repress Mest and Mcts2 around 2-fold in the labyrinth (Playfoot et al. 2019, manuscript submitted, MRC HGU). Impact is not de-repressed in the Tex19.1-/- labyrinth despite depletion of H3K9me3 from this locus, which could potentially reflect an absence of relevant transcription factors for this gene in this tissue. However, we have not been able to determine if the TEX19.1-KAP1 complex also contains UBR2, or whether TEX19.1 exists in separate TEX19.1-KAP1 and TEX19.1-UBR2 complexes, which would confer two separate roles for TEX19.1. In order to address whether Tex19.1-dependent regulation of KAP1 genomic targets depends on UBR2 in the placenta, I therefore investigated whether Tex19.1-dependent

KAP1 genomic targets are dysregulated in Ubr2-/- placentas.

To carry this out I dissected the junctional zone and labyrinth layers from E12.5 placentas, removing the maternal decidua tissue layer and carried out cDNA generation and qRT-PCR as above. In addition to Mest, Mcts2 and Impact, I also analysed Peg3 as a negative control in these layers of the Ubr2-/- placenta, as SETDB1-mediated H3K9me3 histone marks have been implicated in its repression (Skiles et al. 2018), which can be regulated by the TEX19.1-KAP1 complex similarly to Mest, Mcts2 and Impact, but there is no transcriptional upregulation detected in the Tex19.1-/- placenta (unpublished data, Chris Playfoot, MRC HGU) and so its repression appears to be Tex19.1-independent. In absence of Ubr2 there is no gross upregulation of Mest, Mcts2 or Impact in either the placental junctional zone (Fig. 4.9A) or the

219 labyrinth (Fig. 4.9B). Rather, there is a significant decrease in expression of some genes: Mest and Impact in the junctional zone (p=0.0003 and p=0.0242 respectively, paired Student’s T- tests), and Peg3 in the labyrinth (p=0.0212, paired Student’s T-test). These changes could be attributed to the cell-compositional changes observed previously (Fig. 4.6) in the E12.5

Ubr2-/- placenta, as the reduction in proportion of hypomethylated trophectoderm-derived cell types in these layers may result in a proportional change of cells in which these genes are imprinted. Regardless, there is no gross upregulation, whereas the Tex19.1-/- placenta has similar trophectodermal compositional changes and a clear 2x upregulation of Mest and other imprinted gene expression. Therefore, these data indicate that, in contrast to Tex19.1, Ubr2 is not required to maintain repression of imprinted genes in the placenta; this aspect of

TEX19.1’s interaction with KAP1 and the complex’s role in imprinted gene regulation does not appear to involve UBR2.

Figure 4.9 (next page) Ubr2 does not have a role in maintaining placental imprinting

-/- A. Imprinted gene expression levels measured by qRT-PCR in the E12.5 Ubr2 placenta junctional zone (green), over 4 repeats, normalised against expression of housekeeping gene Actin and relative to littermate and gender-matched wild type controls (grey). Mean values shown as a black

-/- bar. B. Imprinted gene expression levels measured by qRT-PCR in the E12.5 Ubr2 placenta labyrinth layer (green), over 4 repeats, normalised against expression of housekeeping gene Actin and relative to littermate and gender-matched wild type controls (grey). Mean values shown as a black bar. Paired Student’s T-test, * p<0.05, *** p<0.001

220

A Expression of Imprinted Genes in the E12.5 Junctional Zone

*** *

Mest Mcts2 Impact Peg3

Ubr2 Genotype

B Expression of Imprinted Genes in the E12.5 Labyrinth *

Mest Mcts2 Impact Peg3

Ubr2 Genotype

221

222

4.3 Discussion

In this chapter, I aimed to investigate a potential role for Ubr2 in development of the murine placenta, and examine a potential link between the functions of Ubr2 and Tex19.1 in this tissue, and in its support of embryonic growth. I identified a Ubr2-/- female-specific lethality that is linked to an embryonic growth restriction and a placental defect primarily affecting trophectoderm-derived cell types. This placental defect also affects males, but there is no discernable lethality, and they recover from the growth defect in adulthood.

Although there are similarities between the Ubr2-/- and Tex19.1-/- placental insufficiencies, including defects in trophectoderm-derived tissues, I showed that many of the currently understood roles for Tex19.1 in the placenta are Ubr2-independent. The Ubr2-/- placenta at E12.5 is grossly normal, but the labyrinth layer is small and there are compositional differences in both trophoblast-derived tissue layers: the junctional zone and the labyrinth. This has similarities to the phenotype of a Tex19.1-/- placenta, which also has an embryonic growth restriction and defects in the trophectoderm-derived layers, although primarily in the junctional zone (Reichmann et al., 2013 and unpublished work by Chris Playfoot, MRC HGU).

Although UBR2 and TEX19.1 have a strong interaction in vitro and in other tissues (Yang et al.

2010; J Reichmann et al. 2017), the previously observed phenotypes in both the Tex19.1-/- junctional zone and labyrinth do not appear to involve UBR2. Therefore, UBR2 and TEX19.1 may be acting independently to regulate placental development in different ways. Conversely, there may be an as-yet undescribed role for TEX19.1 in the placenta that regulates development of the trophoblast lineage, cell composition and, as a result, tissue size – and this role could be acting through UBR2 protein.

223

4.3.1 Strain- and sex-dependent penetrance of the Ubr2-/- phenotype

The previously identified female-specific lethality is also prevalent in our Ubr2-/- mouse, although with variable penetrance over three published strain backgrounds (Y T Kwon et al.

2003), and the C57BL/6J background used in our group. This appears to be least severe on the

C57BL/6J background, in which 39.5% of expected Ubr2-/- females survived to weaning at 3 weeks pp (post-partum). Kwon et al. (2003) found that, on either 129SvJxC57BL/6J or

129SvJxCD1 mixed backgrounds, only 21.7% and 20.0% of expected Ubr2-/- females survived to 3 weeks old. In all three of these strain backgrounds, males survived at expected Mendelian ratios. However, the most severe of these lethalities (129SvJ/SvImJ strain background) affected both male and female Ubr2-/-, and only 1 male and 1 female (6.6% and 6.3%, respectively, of expected Ubr2-/- mice) survived to 3 weeks. A strain-dependent phenotype such as this could be a result of differentially expressed genetic modifiers across different genetic backgrounds. There is evidence to suggest that, even in wild type animals across different strain backgrounds, the size of the placenta and ability of it to support embryonic growth is highly variable, which is linked to growth factor response (Dackor, Caron, and Threadgill

2009). Dackor et al. identified differences between wild type trophectoderm-derived tissues on different strain backgrounds that have some similarity to the trophectoderm-specific defects I identified in our Ubr2-/- placenta: number of spongiotrophoblasts and intensity of glycogen-rich PAS staining was reduced, alongside highly variable expression of trophoblast- derived cell-specific genes such as Ctsq. This appears to indicate a particular vulnerability in placental development to changes in strain background, especially in the trophoblast lineage.

When this is challenged by null mutations such as Ubr2-/-, I would hypothesise that these subtle molecular changes in strain background could exacerbate placental insufficiencies, particularly in the trophoblast lineage.

Another variability in the Ubr2-/- lethalities identified here is sexual dimorphism. Both male and female embryos have a growth restriction at E12.5 and E18.5, associated with placental

224 growth defects and trophectoderm-derived compositional differences in both genders.

However, only female embryos exhibit lethality or growth restriction post-partum in the majority of strain backgrounds, except for the 129SvJ/SvImJ strain background, in which neither males nor females survive to weaning. There is considerable sexual dimorphism in the epigenetic programming of placental tissue (Gabory et al. 2013, 2012; Gong et al. 2018), and these differences in chromatin structure and methylation marks could lead to a difference in vulnerability to mutations that act to regulate epigenetic or hypomethylated genome defence states between the male and female placenta.

The UBR2-interactor TEX19.1 is important for genome defence, acting in complex with KAP1 to lay down H2K9me3 histone methylation via SETDB1 (Playfoot et al. 2019, manuscript submitted, MRC HGU). Aside from genomic imprinting and repression of retrotransposons, which are perturbed in absence of Tex19.1 but I have shown in this chapter do not require

UBR2, one of the key epigenetic functions in a hypomethylated tissue such as the placenta is

X-inactivation. This is a phenomenon only present in female XX cells, which must silence the genes encoded on one X chromosome in order to carry out dosage compensation (Ercan 2014).

Inactivation of this chromosome is mediated by the Xist (X inactive-specific transcript) gene, which encodes cis-acting RNA molecules, specifically expressed from an inactive X, which bind and induce chromatin modifications along their own chromosome (Migeon, Axelman, and Jeppesen 2005). Long-term inactivation depends on CpG dinucleotide DNA methylation

(McGraw et al. 2013); this occurs during early embryonic development (A. Wutz and Jaenisch

2000) and all resulting tissues depend on successful X-inactivation at this specific timepoint.

It has been suggested that X-inactivation can be reversed in specific tissues – for example in the germline, where reprogramming of methylation marks is crucial for epigenetic diversity

(Reik, Dean, and Walter 2001), and in the placenta, the role of which is much less understood.

Cells of the trophoblast lineage undergo de- and remethylation of the X chromosome

(Prudhomme et al. 2015), resulting in highly heterogenous X-inactivation and epigenetic

225 mosaicism across the trophectoderm-derived cell types (Looijenga et al. 1999; Moreira de

Mello et al. 2010; Prudhomme et al. 2015). In addition, various mice exhibiting X-inactivation defects in the placental are deficient in trophectodermal layers (B. Jiang et al. 2012).

Hypomorphic expression of Xist leading to defects in X-inactivation in the trophectodermal lineage have been shown to result in biallelic X-linked gene expression and is associated with placental defects. These include small placentas and an embryonic growth restriction, particularly characterised by trophectodermal tissue reduction such as a significant loss of trophoblast giant cells (TGCs, which are also reduced in the Tex19.1-/- placenta) at E13.5 (Hoki et al. 2011). Although these embryos bear some distinct similarities to both Ubr2-/- and Tex19.1-

/- embryos (Reichmann et al., 2013 and Playfoot et al. 2019, manuscript submitted, MRC HGU), it must be noted that all males progressed through development normally, and did not exhibit any growth or placental defects.

X-inactivation has been associated with regulatory monoubiquitination of histone H2A, leading to repressive chromatin structure and transcriptional silencing (D. E. Cohen and Lee

2002). The canonical E3 ubiquitin ligase associated with this in somatic cells is RING1B (D.

E. Cohen and Lee 2002) but UBR2 has been implicated in mediation of H2A. This appears to be essential for meiotic transcriptional silencing at the XY body of spermatocytes (J Y An et al. 2010; J. M. A. Turner 2007) in presence of TEX19.1, as well as for chromosome stability in somatic cultured fibroblasts (Jee Young An et al. 2012) in absence of TEX19.1, and so could also be important for genomic stability through histone ubiquitination in the X-chromosomes of placental cell types.

Therefore, the trophectodermal defects in absence of Ubr2 indicate a potential role for the protein in X-inactivation in females. This may act through its interaction with TEX19.1, as this protein has been shown to be important for genomic defence and epigenetic silencing

(Playfoot et al. 2019, manuscript submitted, MRC HGU). Investigation of this by study of global methylation marks or expression levels of X-encoded genes such as Xist would give

226 insights into how UBR2 is acting in this tissue. However, it must be noted that although X- inactivation defects may give rise to female-specific lethalities, they do not necessarily explain the male phenotype observed in Ubr2-/- and Tex19.1-/- embryonic stages.

4.3.2 Ubr2’s role in cohesin regulation

Ubr2 has been implicated in cohesin regulation, both alongside and in absence of Tex19.1

(Reichmann et al. 2017, and Chapter 3), and so it is possible that this is the mechanism by which placental growth restriction is taking place in the null phenotype. Study of cohesin protein levels in this tissue is confounded by chromatin preps from E10.5 tissue, preceding the growth restriction phenotype, which have highly variable and inconclusive cohesin levels

(data not shown).

One of the best studied roles for cohesin is in sister chromatid cohesion during the mitotic cell cycle, allowing even division of chromosomes and preventing aneuploidy, which can lead to cell death. Both Ubr2 and Tex19.1 have particularly been implicated in regulation of the acetylated cohesin subpopulation (the AcSMC3 subunit), which is well-studied in mitosis and is required for recruitment of cohesin-maintenance factors such as sororin throughout S and G2 phases of the cell cycle and into mitosis (Nishiyama et al. 2010; Ladurner et al. 2016).

Efficient and timely loading and removal of the cohesin complex during the cell cycle is vital for correct chromosome segregation during cell division, and maintenance of the acetylated cohesin complex during mitosis itself is important. In the placenta, one of the key trophoblast-derived cell types are the trophoblast giant cells (TGCs). These are large polyploid cells located in various tissue layers that are important for remodeling and vascularization of maternal and embryonic tissue to allow highly efficient nutrient transfer

(D. Hu and Cross 2010). They develop from the trophoblast lineage via endoreduplication, in which cells skip mitosis and undergo multiple rounds of DNA synthesis. Not much work has been done to look at the role of the cohesin complex in this specialized cell cycle, but I would

227 hypothesise that cohesin regulation is highly important in these cell types. The cohesinopathy

WABS is associated with defects in the DNA helicase DDX11/ChIR1, which is required for establishment of chromosome cohesion during DNA replication in S phase. This helicase stabilises replication fork structures by interaction with replication fork protectors (Pisani et al. 2018), and it has an in vitro interaction with cohesin complexes (Cortone et al. 2018).

Therefore, cohesin appears to play a role in replication fork stability and loss of chromosome cohesion during a specialised S phase such as in placental TGCs could be a cause of chromosome instability or cell death.

In order to specifically study this cell type in the Ubr2-/- placenta it would first be important to count the number of TGCs and other individual cell types and conduct a more thorough histological examination of the Ubr2-/- placenta. For example, TUNEL stains could be used for quantification of apoptotic cells that may not survive a perturbed cell cycle, leading to the smaller tissue size. In vitro migratory assays could also be used to assess the competency of

TGCs for their normal placental ability to invade the maternal decidua (James, Tun, and

Clark 2016). Finally, there are also some studies into the culture and differentiation of trophoblast stem cells (TSCs) into cell types such as syncytiotrophoblasts (SynTs) which could be a tool to study placental requirements for Ubr2 in cohesins in cell culture in an in vitro setting (Latos and Hemberger 2016; Zhu et al. 2017).

Finally, recent work shows that cohesin is required for much more than simply sister chromatid cohesin and mitotic stability, and the cohesin complex is important for chromatin structure in a broader sense, potentially regulating the structure of TADs and loops within chromatin (G. Wutz et al. 2017; Nuebler et al. 2018; Skibbens 2019). Defects in cohesin complex regulation as a result of either Ubr2-/- or Tex19.1-/- could therefore have wide-reaching implications within placental tissue, affecting interphase chromatin structure or genomic stability, so leading to transcriptional or regulatory changes within individual cell types in the trophoblast lineage. This could be tested by chromosome conformational capture assays

228 such as HiC, in which closely localised regions of the genome are cross-linked and deep sequenced, leading to construction of a spatial chromatin ‘map’.

4.3.3 Roles for UBR proteins in embryonic and placental development

In addition to further investigation of the placental phenotype itself, the molecular mechanism by which Ubr2 is important for placental development is unclear. UBR proteins are a small family of E3 ubiquitin ligases that act to provide substrate specificity in the N-end

Rule pathway of ubiquitination. They have a high level of structural conservation, particularly in their substrate binding domain, and so there is a large potential for redundancy and overlap of function between these proteins (T Tasaki et al. 2009). Tex19.1 has been implicated in the regulation of UBR2’s substrate binding ability in the N-end Rule pathway (J Reichmann et al.

2017) and so it could be that the placental phenotype is a result of the loss of this molecular function.

Some, potentially redundant, roles for UBR1 and UBR2 during development include embryonic neurogenesis and cardiovascular development (Jee Young An et al. 2006). There is also evidence to suggest that UBR4 (also known as p600) has key roles in both embryonic and placental development, presenting a growth restriction phenotype similar to that seen in

Ubr2-/- females (Nakaya et al. 2013). Therefore, whether the role of UBR2 in the placenta is

TEX19.1-dependent or independent, further work into potential N-end Rule functions of the

UBR2 protein in the placenta will be required in order to ascertain the mechanism by which

UBR2 is important for placental development. This could be achieved through selective inactivation of N-end Rule functions of UBR proteins, which is discussed in further detail later in this thesis.

229

4.3.4 Summary

In summary, in this chapter I was able to identify a placental defect in absence of Ubr2 that associates with embryonic growth restriction. This is not gender-specific, affecting both males and females during embryonic stages to a similar level. However, it only persists in females, as a female-specific growth restriction can be observed at 3 weeks post-partum, and a female-specific lethality in 60.5% of expected Ubr2-/- females born from heterozygous matings. UBR2’s major interactor, TEX19.1, has been implicated in genome defence in the hypomethylated trophectoderm lineage, particularly in repression of retrotransposons and imprinted gene expression. However, UBR2 is not required for either function in the developing placenta, and so it is unclear how these placental defects arise in absence of Ubr2.

Further work will be required to elucidate the role of UBR2 in placental development, potentially through other genome defence roles such as X-inactivation or through cohesin regulation as has been identified in other Ubr2-lacking tissues.

230

Chapter 5

Functional mutations in the UBR2 N-end Rule

231

232

5.1 Introduction

As discussed in Chapter 1, UBR2 is one of seven UBR proteins that contain the UBR box; four of which, including UBR2, have been shown to have activity as E3 ubiquitin ligases in the N- end Rule protein recognition system of the UPS (T Tasaki et al. 2009). These proteins are coexpressed in a wide variety of tissues and there is evidence for redundancy between them

(Jee Young An et al. 2006). However, the Ubr2-/- mouse has a specific subset of phenotypes, suggesting that UBR2 has essential molecular functions in specific tissues, whether through the N-end Rule, or through other potential functions of the protein that cannot be compensated by UBR protein redundancy for N-end Rule substrate binding.

UBR2 has a 1:1 stoichiometric interaction with TEX19.1, which is a protein expressed primarily in hypomethylated contexts: the ovary, testis, placenta, and in ES cells (Yang et al.

2010; J Reichmann et al. 2017; Crichton et al. 2017). Data previously published by our lab shows that TEX19.1 is able to regulate turnover of Type II reporters by UBR2 (J Reichmann et al. 2017). The phenotypes of Tex19.1-/- and Ubr2-/- mice have some overlap, appearing to work both agonistically and antagonistically depending on the context of their co-expression, although there are key differences indicating that both genes have roles independent of each other as well (see Chapters 3 and 4 for more details).

One of the best characterised of the Ubr2-/- phenotypes is in male meiosis and spermatogenesis, in the testes. This includes gross spermatocyte apoptosis and male infertility, which can be characterised by gross morphological changes in the seminiferous tubules of the testis, low testis weight and a lack of mature sperm (Y T Kwon et al. 2003), seeming to result from chromosome asynapsis and a zygotene-like arrest (Crichton et al. 2017). The infertile

Tex19.1-/- male also has a zygotene-like asynaptic arrest and reduction in recombination rates

(Crichton et al. 2017). Indeed, TEX19.1 protein is destabilised in the Ubr2-/- testis (Yang et al.

2010), indicating that their interaction is vital for normal spermatogenesis. However, it remains unclear whether the phenotype is a result of lack a functional TEX19.1-UBR2

233 complex, perhaps through TEX19.1’s regulation of Type II N-end Rule activity, or if UBR2 is simply required for TEX19.1 stabilisation in this tissue, producing such a stark phenocopy between these male mice. Aside from UBR2, the UBR protein with the highest Type II N-end

Rule affinity is UBR1 (T Tasaki et al. 2005), and these are both expressed in the testes (T

Tasaki et al. 2005; Y T Kwon et al. 2003). However, there is a difference in localisation of UBR1 and UBR2 in this tissue; UBR1 is highly expressed in spermatogonia, whereas UBR2 is specifically localised to spermatocytes (Y T Kwon et al. 2003). This implies that the Ubr2-/- testis may have lost a specific Type II N-end Rule function in spermatocytes that cannot be covered by protein redundancy, resulting in the observed phenotypes.

Our group have also identified a novel role for UBR2 in regulation of the cohesin complex

(Reichmann et al. 2017 and Chapter 1) – specifically, in the regulation of the chromatin-bound acetylated subpopulation of cohesin, linked to the role of TEX19.1 in acetylated cohesin regulation in the murine oocyte. Both Ubr2 and Tex19.1 are expressed in embryonic stem cells

(ESCs), and presence of TEX19.1 protein is important for repressive chromatin modifications and genomic stability in these contexts (Hackett et al., 2012; Maclennan et al., 2017 and

Playfoot et al. 2019, manuscript submitted, MRC HGU). Another role for both Tex19.1 and Ubr2 in ESCs is in the regulation of the AcSMC3-containing cohesin complex. Tex19.1-/- ES cells have a specific loss of the chromatin-associated acetylated cohesin subunit AcSMC3, as observed in the meiotic oocyte (J Reichmann et al. 2017), but in a Ubr2-/- ES cell line, there is a significant increase in this acetylated subpopulation. This indicates an antagonistic relationship between these proteins in the regulation of acetylated cohesin in this cell type (unpublished data, Karen Dobie, MRC HGU). Ubr2 also has Tex19.1-independent roles in AcSMC3 regulation in mitotic tissues such as the spleen, in which there is again an increase of the acetylated subpopulation of cohesin in absence of Ubr2. As the N-end Rule is the most well characterised molecular function of UBR2, I hypothesised that its role in cohesin regulation

234 across various tissues acts, either directly or indirectly, through this substrate binding pathway.

UBR2’s role in the N-end Rule is as an E3 ubiquitin ligase, which confers substrate specificity for a ubiquitination cascade (Takafumi Tasaki and Kwon 2007). This is regulated by two protein domains (Fig. 5.1A, Tasaki et al., 2009): the UBR box, a highly conserved 70-residue zinc-finger-like structure which is shared across all UBR proteins, and the N-domain, which has high homology with the bacterial ClpS domain and is common to both mouse proteins

UBR1 and UBR2. Alanine-scanning mutagenesis of these domains of mouse UBR1 identified two point mutations: D118A and D233A, which specifically interfere with the ability of UBR1 to bind either Type I or Type II substrates, respectively (T Tasaki et al. 2009).

The Ubr1D118A point mutation lies within the UBR box of UBR1 protein (Fig. 5.1B), which directly binds both Type I and Type II substrates (E Matta-Camacho et al. 2010), although the N-domain is also required to accommodate Type II peptide binding (T Tasaki et al. 2009).

Within the UBR box, the residue D118 specifically conjugates a water molecule with the N- terminal Arginine of Type I peptides (E Matta-Camacho et al. 2010; J Muñoz-Escobar et al.

2017), (Fig. 5.1C) and so mutation of this is expected to leave the whole UBR box intact, maintaining the binding ability towards Type II peptides as normal (T Tasaki et al. 2009).

Alanine screening of other amino acid residues in the UBR box (Fig. 5.1B) resulted in either partial perturbation of the Type I interaction, or perturbation of both Type I and Type II

(Tasaki et al. 2009), presumably destabilising the UBR box structure altogether, and so I surmised that Ubr1D118A would be the best candidate for a Type I mutation in Ubr2.

The Ubr1D233A point mutation lies within the N-domain and is conserved in UBR2, and in the homologous bacterial ClpS protein (Fig. 5.1D). ClpS is a small (12kDa) adaptor protein that acts to provide substrate specificity in the E. coli proteolytic ClpAP machinery, the bacterial equivalent of an N-end Rule (Gottesman et al. 1998) by binding the ClpA catalytic subunit

235

(Zeth et al. 2002). The D233 residue is also essential for ClpS substrate binding (Tan et al.

2016).

To assess the importance of the N-end Rule in UBR2’s major functions and investigate the molecular basis of the null phenotypes, I aimed in this chapter to generate separation-of- function mutations in these two substrate-binding domains: the UBR box and the N-domain, with a view to understand the molecular functions of UBR2 more closely. I generated a

Ubr2D233A mouse model using CRISPR to assess the hypothesis that UBR2 primarily acts through the Type II N-end Rule substrate binding pathway. However, Type II is less important for the essential functions identified in a Ubr2-/- phenotype than has been previously hypothesised. This, paired with introduction of the Ubr2D118A point mutation into ESCs, shows that neither Type I or Type II pathways are apparently important for the recently uncovered novel role of Ubr2 in cohesin regulation, so indicating potential non-canonical roles for Ubr2, outside of the N-end Rule pathway.

236

B >m -UBR1 CGKVF----KSGETTYSCRDCAIDPTCVLCMDCFQSSVHKNHRYKMHTSTGGGFCDCGDTEAWKTGPFC >m-UBR2 CGRVF----KVGEPTYSCRDCAVDPTCVLCMECFLGSIHRDHRYRMTTSGGGGFCDCGDTEAWKEGPYC >m -UBR4 CTFTITQKEFMNQHWYHCHTCKMVDGVGVCTVCAK-VCHKDHEIS-YAKYGSFFCDCGAKEDG----SC >m -UBR5 CSFTWTGAEHINQDIFECRTCGLLESLCCCTECAR-VCHKGHDCKLKRTSPTAYCDCWEK------CKC

D >m -UBR1 YYCVLFNDEHHSYDHVIYSLQRALDCELAEAQLHTTAIDKEGRRAVKAGVYATCQEAKEDIKSHSENVSQHPLHVEVL >m -UBR2 YYCVLFNDEVHTYEQVIYTLQKAVNCTQKEAIGFATTVDRDGRRSVRYGDFQYCDQAKTVIVRNTSRQTK-PLKVQVM >ec -ClpS YKVILVNDDYTPMEFVIDVLQKFFSYDVERATQLMLAVHYQGKAI-CGVFTAEVAETKVAMVNKYARENEHPL-LCTL

Figure 5.1 – The conserved UBR1 and UBR2 UBR box and N-domain. A. Schematic representation of the UBR2 gene, with key domains highlighted. Taken from Tasaki et al. 2009. B. Protein sequence alignment of the UBR box of four UBR proteins, showing homology. Green: cysteine and histidine residues conserved throughout, required for zinc ion binding. Blue: D118 residue, the location of the D118A Asp->Ala point mutation, which is conserved between UBR1 and UBR2. C. Structural representation of chemical interactions within the UBR1 UBR box. Green: a typical bound Type I peptide, in which two N-terminal residues are shown.

Negatively charged D118 (marked by blue box) specifically conjugates an H2O molecule with the N-terminal Arginine (Arg1). Adapted from Matta Camacho et al. 2010. D. Protein sequence alignment of mouse UBR1 and UBR2 N-domains and bacterial ClpS domain. Blue: conserved D233 residue, the location of the D233A Asp->Ala point mutation.

237

238

5.2 Results

5.2.1 In vitro validation of the Ubr2D118A and Ubr2D233A point mutations

In order to specifically interfere with the N-end Rule functions of UBR2, I sought to generate point mutations that interfere with the binding of UBR2 to its N-end Rule substrates.

Analogous mutations D118A and D233A have previously been generated in the Ubr1 gene (T

Tasaki et al. 2009). However, although UBR1 and UBR2 are highly conserved family members

(E Matta-Camacho et al. 2010), and both residues D118 and D233 are conserved in the UBR box and N-domain, there is no direct evidence that these residues will be directly important for UBR2 N-end Rule function. Our lab previously cloned the UBR2 CDS (coding sequence, exons only) into the plasmid p3xFLAG-CMV-10, which encodes an N-terminal triple FLAG tag immediately upstream of a multiple cloning site into which the UBR2 CDS was cloned

(Fig. 5.2A).

In order to introduce either D118A or D233A point mutations into this UBR2 CDS, I used site- directed mutagenesis, in which I designed primers that would introduce the point mutation during a PCR, so generating a linear version of the mutant plasmid that I then circularised. I used Sanger sequencing to verify the introduction of these mutations (Fig. 5.1B and C). This gave me a tool to overexpress the mutant UBR2 proteins and so assess the binding capacity of either Ubr2-T1 (D118A) or Ubr2-T2 (D233A) mutants in an in vitro setting, examining the hypothesis that these are important residues for UBR2 N-end Rule substrate binding.

239

A D118 T1

D233 T2

p3xFLAG-CMV10-UBR2 11574bp

B GAC->GCC D188A

C GAT->GCT

D233A

Figure 5.2 – Site-directed mutagenesis of the Ubr2 CDS. A. Schematic representation of the pCMV10-FLAG plasmid, with the Ubr2 CDS (coding sequence, 5268bp) inserted into the multiple cloning site downstream of the N-terminal 3xFLAG tag sequence. Locations of either the D118 or D233 residues are shown, which are mutated using site- directed mutagenesis in this screen. B. Sanger sequencing data for the UBR2-D118A mutant CDS, the D118A mutation is underlined, GAC->GCC. C. Sanger sequencing data for the UBR2- D233A mutant CDS, the D233A mutation is underlined, GAT->GCT.

240

Firstly, I aimed to investigate the ability of these mutant UBR2 proteins to bind substrates within the N-end Rule. To do so, I transfected HEK293T cells with each of these plasmids, so overexpressing either wild type FLAG-UBR2, FLAG-UBR2-T1D118A or FLAG-UBR2-T2D233A protein and performed a peptide pulldown assay on the cell lysate. To do this I chemically conjugated peptides to resin beads via a C-terminal Cysteine residue, allowing immobilisation of the peptide via a thioether bond. These have a specific amino acid at their

N-terminal end: Type I (arginine, TI), Type II (phenylalanine, TII) or a stabilising negative control (glycine, -ve) (Fig. 5.3A). This will specifically identify proteins able to bind either

Type I, Type II, or both N-end Rule peptides, which I then assayed by Western blotting against the N-terminal FLAG-tag of the overexpressed UBR2 protein.

Wild type UBR2 has very low Type I (TI) binding, as has been shown in previous studies (T

Tasaki et al. 2005) and a very strong Type II (TII) binding ability (Fig. 5.3B and C). The –ve

(stabilising, glycine) peptide negative control pulls down a very low level of FLAG-tagged

UBR2 protein, and so I attributed this to low level non-specific binding of the protein to resin beads. In both the FLAG-UBR2-T1D118A and FLAG-UBR2-T2D233A assays, the –ve control pulls down a very low, or no, detectable level of FLAG-tagged UBR2 protein over all three independent repeats, as observed with the wild type UBR2 protein.

The FLAG-UBR2-T1D118A protein is able to bind TII peptides to a similar level as that observed with wild type UBR2, but there is very minimal detectable interaction with TI peptides over three independent repeats (Fig. 5.3B). Therefore, the D118 residue is important to the ability of the UBR2 UBR box to bind Type I peptides, as has been previously observed in UBR1 (T

Tasaki et al. 2009), but the UBR box itself is intact and able to recognise a Type II substrate.

Finally, the FLAG-UBR2-T2D233A mutant protein has no detectable interaction with TII peptides over three independent repeats, although this is a highly detectable interaction when using the wild type protein (Fig. 5.3C). As discussed above, the TI peptide recognition is low and difficult to detect in in vitro UBR2 studies and so it is difficult to assess whether the

241

FLAG-UBR2-T2D233A mutation affects TI binding. However, as loss of the N-domain in UBR1 has been shown to maintain UBR box functionality in binding to TI substrates (T Tasaki et al. 2009), and the sequences of the UBR box, N domain and surrounding regions of UBR1 and

UBR2 are highly conserved (T Tasaki et al. 2005), I assumed that the D233A mutation does not affect UBR box structure and any potential Type I ability of UBR2 is unaffected.

In vitro peptide pulldown studies for these point mutations in UBR1 show similar results;

Ubr1D118A prevents detectable binding to Type I peptides, and Ubr1D233A prevents Type II binding (T Tasaki et al. 2009). Therefore, the conserved structure between the two proteins appears to translate to conserved functional residues between the two proteins in the UBR box and N-domain, and these residues are important for N-end Rule function as I hypothesised. In contrast to UBR1, UBR2 has a particular affinity, at least in in vitro pulldown assays for Type II peptide recognition, although there are known Type I substrates of UBR2 in vivo, such as the shared UBR1/UBR2 N-end Rule Type I substrate SDE2 (Rageul et al. 2019).

Figure 5.3 – The UBR2-T1 and UBR2-T2 mutations prevent detectable substrate binding. A. Sequences of peptides chemically conjugated to resin beads used in a peptide pulldown assay. N-terminal residues R, F or G are Type I, Type II or negative control peptide sequences, respectively. B. Western blots showing 3 independent repeats of a peptide pulldown after overexpression of UBR2-T1 (D118A) in Hek293T cells. I, input protein. TI, Type I (arginine). TII, Type II (phenylalanine). –ve, negative (glycine, stabilising) control. Probed using anti-FLAG M2 antibody for the N-terminally FLAG tagged UBR2 protein. * indicates a possible isoform of the UBR2 protein. C. Western blots showing 3 independent repeats of a peptide pulldown after overexpression of UBR2-T2 (D233A) plasmid in Hek293T cells, lanes as in B.

242

Type I (TI) RIFSTDTGPGGC Arginine (R)

Type II (TII) FIFSTDTGPGGC Phenylalanine (F)

-ve control GIFSTDTGPGGC Glycine (G)

B kDa FLAG-UBR2-WT FLAG-UBR2-T1 245 I TI TII -ve I TI TII -ve

190 UBR2-FLAG

245

190 UBR2-FLAG

* 245

190 UBR2-FLAG

kDa C FLAG-UBR2-WT FLAG-UBR2-T2 I TI TII -ve I TI TII -ve 245

190 UBR2-FLAG

245

190 UBR2-FLAG

245

190 UBR2-FLAG

243

Although there are detectable changes in protein binding ability when the Ubr2-CDS FLAG point mutant constructs are overexpressed in HEK293T cells, there is no evidence that these mutant proteins are still functional in any pathway aside from their targeted substrate binding. Therefore, I next aimed to validate that, although they have impaired N-end Rule activity, the FLAG-UBR2-T1D118A and FLAG-UBR2-T2D233A mutant proteins are still functional outside their N-end Rule substrate binding. One of the only other validated molecular functions of UBR2 is its stoichiometric interaction with TEX19.1 (J Reichmann et al. 2017), and so I overexpressed the FLAG-tagged plasmid constructs alongside a GFP-tagged

TEX19.1 plasmid (pCAAGS-Tex19.1-GFP), which was previously generated in our lab, in

HEK293T cells and performed co-immunoprecipitation with FLAG-M2 beads to pull down the UBR2 protein, and any interacting proteins.

Wild type UBR2 (Fig. 5.4A and B, FLAG coIP, lane 1) pulls down GFP-TEX19.1, as has been previously observed, but not to GFP alone (Fig. 5.4A and B, FLAG coIP, lane 2). Both FLAG-

UBR2-T1D118A (Fig. 5.4A, FLAG coIP, lane 3) and FLAG-UBR2-T2D233A (Fig. 5.4B, FLAG coIP, lane 3) also co-immunoprecipitate with GFP-TEX19.1. As in the wild type UBR2, GFP expressed alone does not pull down with either T1 or T2 versions of UBR2 (Fig. 5.4A, FLAG coIP, lanes 2 and 4). As a final control, coexpression of the pCMV10 FLAG construct alone with GFP-TEX19.1 (pCAAGS) does not pull GFP-TEX19.1 down (Fig. 5.4A, FLAG coIP, lane

5) and so I was able to conclude that the interactions observed between both wild type and either T1 or T2 mutant UBR2 and TEX19.1 are due to a specific interaction of the proteins themselves.

In this section, using both peptide pulldown assays and coimmunoprecipitation with known interactor TEX19.1, I was able to show that that these specific point mutations within UBR2 are functionally conserved from UBR1 and are successfully able to perturb its canonical substrate binding abilities, via either Type I (Arg) or Type II (Phe) N-end Rule binding. These do not deter UBR2 from other functions, such as its interaction with TEX19.1, which is

244 thought to be important for regulation of UBR2’s Type II N-end Rule binding ability (J

Reichmann et al. 2017). Therefore, these point mutations are ideal tools to separate out the phenotypes of a Ubr2-/- into the protein’s specific molecular functions.

D118A D233A Figure 5.4 (next page) UBR2-T1 and UBR2-T2 proteins are functional outside the N-end Rule A. Representative Western blot of co-immunoprecipitation with FLAG-M2 beads, after coexpression of either wild type FLAG-UBR2 or FLAG-UBR2-T1D118A (pCMV10-Ubr2-FLAG or pCMV10-Ubr2-T1-FLAG) with GFP-TEX19.1 (pCAAGS-Tex19.1-GFP). Controls: FLAG and GFP are the pCMV10-FLAG and pCAAGS-GFP transfected to ensure no crossreaction of the tags with proteins of interest. Probed using anti-FLAG M2 antibody for the N-terminally FLAG tagged UBR2 protein and anti-GFP antibody for the GFP-tagged TEX19.1 protein. * Indicates other isoforms of the UBR2 protein. B. Representative western blot of co-immunoprecipitation of either wild type FLAG-UBR2 or FLAG-UBR2-T1D118A with GFP-TEX19.1. Controls as in A.

245

Input FLAG coIP

UBR2-FLAG + + – – – – + + – – – – UBR2-FLAG

UBR2-T1-FLAG – – + + – – – – + + – – UBR2-T1-FLAG

FLAG – – – – + – – – – – + – FLAG

TEX19.1-GFP + – + – + – + – + – + – Tex19.1-GFP

GFP – + – + – – – + – + – – GFP kDa 245 FLAG (UBR2)

190 *

80 GFP (TEX19.1)

32 25 GFP alone

Input FLAG coIP

UBR2-FLAG + + – – – – + + – – – – UBR2-FLAG

UBR2-T2-FLAG – – + + – – – – + + – – UBR2-T2-FLAG

FLAG – – – – + – – – – – + – FLAG

TEX19.1-GFP + – + – + – + – + – + – Tex19.1-GFP

GFP – + – + – – – + – + – – GFP kDa 245 FLAG (UBR2) 190 *

GFP (TEX19.1) 80

32 25 GFP alone

246

5.2.2 Generation and validation of a Ubr2-T2D233A mouse line using CRISPR-Cas9

As discussed in section 5.1, the major detectable role for UBR2 in the N-end Rule is in the recognition and binding of Type II N-terminal degrons. The protein has a very low detectable binding to Type I degrons, and this is generally assumed to be the role of other coexpressed

UBR proteins. The essential functions of UBR2 are hypothesised to act through the Type II pathway, and so I surmised that the best way to begin investigation of the molecular basis of these phenotypes would be to prevent function of UBR2 in the Type II N-end Rule, by introduction of the D233A point mutation.

In order to generate a Ubr2-T2D233A mouse line, I utilised CRISPR-Cas9 technology. In order to introduce a specific point mutation into this locus, I used wild type Cas9 for generation of double-strand breaks (Cong et al. 2013) and a homology directed repair (HDR) template, which can be introduced into the break site by recombination (Fig. 5.5A). The repair templates were designed as such that either the wild type sequence or the D233A point mutation could be integrated, with a mutant PAM site in both to prevent re-editing after repair template integration (Fig. 5.5B). Ubr2-/- males are infertile and therefore I used both a wild type and mutant repair template to reduce the efficiency of editing with the mutant repair and ensure that heterozygote mice are generated for further breeding and colony establishment. These were injected into the cytoplasm of single-celled mouse zygotes alongside an RNA guide sequence specific to a region within 20bp of the target residue, with a PAM sequence on the 3’ end – this is an NGG motif that aids in binding of the Cas9 machinery. (Fig. 5.5C).

After injection, zygotes were transferred into a surrogate mother in the CD1 background (Fig.

5.5C), and the pups born were screened at weaning (3 weeks pp) for successful integration of the repair templates. 9 individuals were successfully born following one round of zygote injections and genotyped as in Fig. 5.6. Of these individuals, 5 had undergone active genomic editing, but only 3 of these mice were identified as being heterozygote for the mutant allele

247

(30% successful editing). However, these were all heterozygote against mosaic or offtarget mutant alleles, and not wild type Ubr2, and so one round of breeding with wild type C57BL/6J mice was carried out in order to generate 50% true heterozygote offspring, which were then bred together to give a stable heterozygous mouse colony (Fig. 5.5C), expected to produce 25%

Ubr2T2/T2 animals, if at a normal Mendelian breeding ratio.

D233A Figure 5.5 – Generation of a Ubr2 mouse line using CRISPR-Cas9. A. Schematic of CRISPR-wtCas9 homology directed repair. Adapted from www.addgene.org/crispr/guide/. B. CRISPR-HDR guide and repair sequences for the UBR2-T2 CRISPR experiment. PAM sequence is shown in bold and cut site of wtCas9 enzyme marked by ^. Introduced point mutations are marked in repair templates in green. Homology arms are underlined. C. Schematic representation of zygotic injection of CRISPR components and breeding D233A strategy for Ubr2 mouse colony establishment. Components were introduced by injection into the cytoplasm of the single-celled zygote, then cultured and transferred into a CD1 host mother. Pups born were on the C57BL/6J background and were genotyped as being heterozygote for the D233A mutation against indel or mosaic Ubr2 alleles. These were bred with C57BL/6J wild type

+/T2 mice and the 50% heterozygote Ubr2 offspring were then crossed together to produce an

T2/T2 +/+ expected 25% Ubr2 experimental mice, with 25% Ubr2 littermate control animals.

248

targetcleavage (DSB formation) (DSB

Experiment Guide sequence Repair template sequence

Mutant (D233A): TTTTCTCCCCAGAGAGAAGAGTGACACCTACTACTGCATGCTGTTTAA TGCTGAAGTTCACACCTATGAGCAAGTCATTTATACCCTTCAGAAAGC Sense: TGTGAACTG T2 (D233A) CTGCATGCTGTT TAATG^ATGAGG Wild type: TTTTCTCCCCAGAGAGAAGAGTGACACCTACTACTGCATGCTGTTTAA TGATGAAGTTCACACCTATGAGCAAGTCATTTATACCCTTCAGAAAGC TGTGAACTG

Cas9 RNA gRNA dsDNA WT repair ♂ +/T2 dsDNA D233A repair

♀ X +/T2

50% X

CD1 X Ubr2 Alleles Wild type (C57BL/6J) D233A (T2) X Mosaic or offtarget

249

All subsequent colony genotyping was carried out using tetra-primer ARMS-PCR, which depends on the binding ability of a pair of inner primers that overlap at their 3’ end over the point mutant (Fig. 5.6A). The forward and reverse outer primers act to amplify with either one or both inner primers, and also generate a control PCR product that encompasses the mutated region and controls for a successful PCR and any unintentional indels in the

CRISPR-Cas9 site. (S. Ye et al. 2001; Medrano and De Oliveira 2014). Mice generated from heterozygote (Ubr2+/T2 X Ubr2+/T2) crosses were genotyped in this way and then assayed by gel electrophoresis (Fig. 5.6B). The outer control fragment gives a PCR product at 476bp, present in all mice genotyped here. Wild type mice (Ubr2+/+) 2899 and 2900 also have a PCR product at 242bp, and D233A mutants 2895, 2896 and 2899 (Ubr2T2/T2) have a PCR product at 292bp.

Heterozygote mice 2897 and 2901 (Ubr2+/T2) have both smaller bands, as both C and A alleles are present.

These genotypes can be confirmed using Sanger sequencing of the 476bp outer fragment (Fig.

5.6C). Wild type Ubr2+/+ mouse 2899 has the Asp-encoding (D) GAT codon, and a wild type

PAM site (AGG), as the heterozygotes generated from zygotic CRISPR injection were crossed to unedited wild type C57BL/6J mice. Ubr2T2/T2 mouse 2895 has the Ala-encoding (A) GCT codon and a silently mutated PAM site (AAG).

Using these genotyping tools, I was able to generate and maintain a stable mouse line with the Ubr2-T2D233A mutation. For all further experiments in this chapter, Ubr2T2/T2 experimental mice were generated from heterozygote (Ubr2+/T2 X Ubr2+/T2) crosses, and all control animals used were Ubr2+/+ or Ubr2+/T2 generated from these crosses in the Ubr2-T2D233A colony.

250

A C A

Control C allele

A allele

B C GAT->GCT D233A

2895 2896 2897 2898 2899 2900 2901 AGG->AAG silent mutant PAM site

1500

1000

500 2899 (+/+) 400 476

300 292

200 242

100

T2/T2 T2/T2 +/T2 T2/T2 +/+ +/+

+/T2

2895 (T2/T2)

D233A Figure 5.6 Genotyping of Ubr2 mouse colony using tetra ARMS-PCR A. Schematic of tetraARMS-PCR, used for screening and genotyping of mice. Green are outer primer pair, which amplify the outer control band. Blue are inner primers, which bind specifically +/T2 +/T2 to either allele. B. Representative gel electrophoresis of one litter of mice from a Ubr2 x Ubr2 heterozygous cross, showing tetraARMS-PCR products for T2 genotyping. 100kb ladder shown on left. Control (outer primer) band is 476bp. Wild type A allele band is 242bp and UBR2-T2 C allele band is 292bp. 4-digit labels indicate mouse codes. C. Representative Sanger sequencing data generated from earclips of mice at weaning (3 weeks post-partum).

251

My previous work to validate the D233A point mutant in UBR2 (Fig. 5.3C and 5.4B) showed that this is a key residue in the UBR2 N-domain, and the protein is unable to carry out detectable binding to Type II substrates, although the whole protein, particularly the UBR box, appears to be intact and functional, and able to interact with the protein TEX19.1.

However, in an in vivo context, the D233 residue may not be as vital for Type II substrate binding as appears in vitro. Therefore, in order to validate the functional capabilities of the

UBR2-T2 protein, I carried out in vitro peptide pulldown assays as in section 5.2.1 using whole testis in vivo-expressed protein extract to ascertain that the D233A point mutation is also vital for Type II N-end Rule binding in the endogenous UBR2 protein.

For this peptide pulldown, I incubated whole testis protein extract from male mice at 8-10 weeks pp with resin conjugated to either Type I, Type II or negative control peptides as in section 5.2.1 (Fig. 5.3). I then probed this for endogenous UBR2 protein in a Western blot assay (Fig. 5.7). UBR2 protein can be detected in input lanes and in the Type II pulldown lanes for Ubr2+/T2 controls. Importantly, although UBR2-T2 protein is readily detectable in the input samples, and appears to be stable at around 210kDa, similar to the wild type UBR2 protein, there is no detectable UBR2-T2 pulled down by Type II (phenylalanine-containing) peptides from Ubr2T2/T2 testis protein extract. Either UBR2 or UBR2-T2 protein cannot be detected in either Type I pulldown lanes or negative control lanes for either genotype. This demonstrates that endogenous UBR2D233A is a stably expressed protein, as it is readily detectable in input samples, but has lost Type II binding ability and so will not be able to carry out UBR2 Type II N-end Rule functions.

In this section I showed the design and implementation of a CRISPR-Cas9 HDR experiment to generate a successful Ubr2-T2D233A mouse line. Introduction of the point mutation D233A in vivo results in a loss of detectable Type II substrate peptide binding ability, as I observed in section 5.2.1 in vitro (Fig. 5.3C), and so this mouse line can be used to study the Type II N-end

252

Rule functions of E3 ubiquitin ligase UBR2, by comparison to the major phenotypes of a

Ubr2-/- mouse.

+/T2 T2/T2 Ubr2 Ubr2 input TI TII -ve input TI TII -ve kDa 245 UBR2 190

D233A Figure 5.7 – The Ubr2 mutation prevents detectable Type II N-end Rule substrate binding in vivo.

T2/T2 +/T2 Peptide pulldowns with Ubr2 testis protein extract, as compared to Ubr2 control testis, probed with an antibody against endogenous UBR2 protein. Peptides have N-terminal residues R (TI), F (TII) or G (-ve), which confer specificity for the Type I and Type II N-end Rule pathways, or act as a stabilising negative control, respectively.

5.2.3 Phenotyping of the Ubr2-T2D233A mouse line

5.2.3.1 Ubr2’s role in trophoblast-derived placental development is not a Type II

function

One of the most initially obvious phenotypes of the Ubr2-/- mouse is an

underrepresentation of female Ubr2-/- animals born from heterozygote (Ubr2+/- X

Ubr2+/-) crosses, which is described in detail in Chapter 4. In addition to this

observation in our own Ubr2 null colony, which is on a C57BL/6J inbred background,

this female-specific lethality has been documented to varying penetrance, depending

on the genetic background, by Kwon et al. 2003. This lethality is associated with a

female-specific growth restriction in adulthood, and both embryonic and placental

253 size defects in both males and females (Chapter 4). Ubr2-/- embryos are around 80% of Ubr2+/+ littermate size at E18.5, which is 1-2 days before parturition, and the placenta at this stage is around 70% of Ubr2+/+ littermates’ placental weight (Chapter

4).

In order to assess whether Ubr2’s role in female survival acts through the Type II substrate binding ability of UBR2 protein, I assessed the proportions of mice born in the Ubr2-T2D233A mouse colony as a result of heterozygote (Ubr2+/T2 X Ubr2+/T2) crosses, genotyping as described in section 5.2.2, Fig. 5.5C and Fig. 5.6. Crosses between

Ubr2+/T2 mice are expected to generate 25% Ubr2+/+, 50% Ubr2+/T2 and 25% Ubr2T2/T2 animals, with half of the pups expected to be male and half expected to be female (Fig.

5.8A). At weaning (3 weeks post-partum, pp), Ubr2T2/T2 mice represent 22% of pups born (65 of 295 total mice genotyped) over 35 litters, and heterozygote and wild type littermates make up 53% and 25% of pups, respectively (Fig. 5.8B). This does not significantly deviate from the expected Mendelian ratio of pups from these crosses

(p=0.6318, χ2 test). There is no immediately obvious lethality in the Ubr2T2/T2 mice, as was observed in Ubr2-/- mice at 3 weeks pp on the same genetic background

(C57BL6/J), Chapter 4.

The Ubr2-/- lethality has been observed to be female-specific in the majority of genetic backgrounds that a Ubr2-/- has been generated on (Y T Kwon et al. 2003), but of the

295 mice genotypes, there are 152 females and 143 males, which does not deviate from the expected 1:1 ratio of female:male pups (p=0.6810, χ2 test). Female Ubr2T2/T2 mice represent 23% of all female mice genotyped over these 35 litters (34 of 152 females), and heterozygote and wild type females make up 53% and 24%, respectively (Fig.

5.8B). This is, again, not significantly different from the proportion of female mice that were expected (p=0.8209, χ2 test). Finally, male Ubr2T2/T2 mice represent 22% of all

254 male mice in these 35 litters (31 of 143 males genotyped), and heterozygote and wild type males make up 53% and 25%, respectively (Fig. 5.8B) – this is also not a significant change in proportion of genotypes as was expected from Ubr2+/T2 X Ubr2+/T2 crosses (p=0.7622, χ2 test).

Therefore, there is no lethality in Ubr2T2/T2 females, as has been observed in the Ubr2-/- genotype, and so I concluded that UBR2 Type II N-end Rule ability is not required for embryonic survival.

255

A +/T2 +/T2 ♂ Ubr2 X ♀ Ubr2

♂ +/T2

♀ +/T2

All Animals at Weaning Female Male

73 37, 36, 25% 24% 25%

157 81, 76, 53% 53% 53% 31, 65 34, 22% 22% 23%

+/+ +/T2 T2/T2

+/T2 Figure 5.8 Ubr2 crosses produce offspring in Mendelian ratios

+/T2 +/T2 +/+ +/T2 A. Heterozygote crosses (Ubr2 x Ubr2 ) are expected to produce 25% Ubr2 , 50% Ubr2 and

T2/T2 25% Ubr2 offspring. B. Ratios of mice at weaning (3 weeks post-partum) from heterozygous

+/T2 +/T2 (Ubr2 x Ubr2 ) crosses in the T2 colony; all mice, and split into males and females. Statistical testing carried out using χ2 test, no significant changes detected.

256

As there is no definite causal relationship between embryonic growth restriction and lethality detected post-partum, the placental defects or embryonic growth restriction observed in the Ubr2-/- E18.5 embryo (Chapter 4) may still act through UBR2’s Type

II binding ability. It should be noted that the Ubr2-/- phenotype is variably penetrant and so a partial placental defect may not be reflected in a breeding assay alone.

In order to assess the molecular basis of UBR2’s role in placental development, I dissected embryos at E18.5 from heterozygote Ubr2+/T2 X Ubr2+/T2 crosses and weighed the embryos and placentas, as in Chapter 4. The Ubr2-/- phenotype at this stage can be observed by eye; both male and female Ubr2-/- embryos are morphologically normal but are noticeably smaller than littermates. However, Ubr2T2/T2 embryos, while also grossly normal in morphology, do not have an obvious size defect when compared to both littermate Ubr2+/+ and Ubr2+/T2 controls (Fig. 5.9A). The embryo weights reflect this observation (Fig. 5.9B); the mean combined head and body weight of Ubr2T2/T2 embryos is 1082.13±29.9mg (± standard error), and that of Ubr2+/+ littermates is

1096.42±30.1mg, measured over 4 litters of E18.5 animals – this is not a statistically significant change (p=0.7394, Student’s T-test). Ubr2+/T2 embryos have a mean combined weight of 1114.99±38.7mg, which is again not significantly different from either wild type or Ubr2T2/T2 littermates (p=0.7058 and p=0.5083 respectively, Student’s

T-test). Thus, there is no detectable growth restriction in Ubr2T2/T2 embryos at this stage of development.

The mean placental weight of a Ubr2T2/T2 E18.5 embryo is 92.49±4.7mg (± standard error), and that of Ubr2+/+ littermates over 4 litters is 86.38±3.8mg – this is not a significant change in weight (p=0.3196, Student’s T-Test). The placenta of Ubr2+/T2 embryos have a mean weight of 90.13±2.3mg, which is again not significantly different from either wild type or Ubr2T2/T2 littermates’ placental weight (p=0.4181 and p=0.6558

257 respectively, Student’s T-test). This led me to the conclusion that placental development is not regulated by Ubr2’s role in the Type II N-end Rule pathway, but instead through a different molecular function of UBR2.

Taken together, the 3 week pp breeding ratios and E18.5 embryonic and placental weights indicate that neither the female-specific lethality or the placental defect and embryonic growth restriction that I, and others (Y T Kwon et al. 2003), have observed in Ubr2-/- mice act through UBR2’s role in Type II N-end Rule substrate binding.

258

A T2/T2 +/+ +/T2

B E18.5 Embryo Weights C E18.5 Placenta Weights

T2/T2 Figure 5.9 Ubr2 E18.5 embryos do not exhibit a growth or placental defect A. Representative images of littermate embryos from heterozygote pairs before decapitation, showing no gross morphological defects. Scale bar is 500µm B. Weights (mg) of E18.5 embryos

+/T2 over 4 litters from Ubr2 mating pairs. Littermate wild type (+/+, n=13), heterozygote (+/T2, n=12) and homozygotes (T2/T2, n=12). Due to Home Office regulations, these are combined head and body weights of decapitated embryos. Black bar is mean weight. C. Weights (mg) of placentas from E18.5 embryos, as in B. Black bar is mean weight.

259

5.2.3.2 There are no gross infertility phenotypes observed in a Ubr2T2/T2 male mouse

One of the most well characterised phenotypes of a Ubr2-/- mouse, both in our group

(Crichton et al. 2017) and by others (Y T Kwon et al. 2003; Yang et al. 2010; J Y An et

al. 2010), is the role of Ubr2 during spermatogenesis. UBR2 stabilises TEX19.1 protein,

resulting in undetectable TEX19.1 in the Ubr2-/- testes (Yang et al. 2010) and at least a

partial phenocopy of the Tex19.1-/- males, which have a partial meiotic arrest at

prophase I, resulting in infertility (Crichton et al. 2017). This regulation of meiotic

recombination is hypothesised to be a function of the Type II N-end Rule pathway (Y

T Kwon et al. 2003; Yang et al. 2010), and TEX19.1 is thought to act as a cofactor in

Type II N-end Rule activity (J Reichmann et al. 2017).

T2/T2

+/+ +/+

kDa

47 *

TEX19.1

100 Lamin B1 80

T2/T2 Figure 5.10 TEX19.1 protein is stable in the Ubr2 testis Western blotting assay for TEX19.1 (~40kDa) and loading control Lamin B1 (~80kDa) in testis

T2/T2 +/+ protein extract from Ubr2 mice and Ubr2 control animals. * are non-specific bands.

The UBR2-T2 protein is still able to bind ectopically expressed TEX19.1 in an

overexpression in vitro assay in HEK293T cells (Fig. 5.4B), and so in order to

260 investigate the stability of endogenous TEX19.1 in the Ubr2T2/T2 testis, I isolated whole testis protein extract and carried out Western blotting to assess the level of TEX19.1 protein present in the testis of Ubr2T2/T2 and age matched Ubr2+/+ males (Fig. 5.10). The

TEX19.1 protein is expected to be around 40kDa (Uniprot Q99MV2) and is observed in both Ubr2T2/T2 and Ubr2+/+ testis protein extracts, alongside Lamin B1 as a loading control protein. Therefore, endogenous TEX19.1 protein is stable in presence of

UBR2-T2 protein, and this will be a useful mouse model to test whether the UBR2-

TEX19.1 interaction itself is essential for spermatogenesis, or if UBR2 is acting through its Type II ability in prophase I progression.

In order to investigate the role of UBR2 Type II N-end Rule function in male fertility,

I dissected mouse testes from Ubr2+/+ and Ubr2T2/T2 males between 8 and 10 weeks old and assessed the weight. We have previously reported a 67.6% reduction in testis weight between Ubr2+/+ and Ubr2-/- males, and an undetectable level of sperm isolated from the epididymis of Ubr2-/- testes (Crichton et al. 2017). In this study, the average weight of each testis in a Ubr2+/+ male is 85.0±2.8mg (± standard error), and Ubr2T2/T2 testes are 83.7±4.7mg; these are not significantly different in size (Fig. 5.11A, p=0.83,

Student’s T-test). The average testis weight of heterozygous Ubr2+/T2 males is

76.6±2.7mg, which is not significantly smaller than either wild type or T2 testes

(p=0.08 and p=0.39 respectively, Student’s T-tests). This indicates a lack of gross fertility phenotypes in the Ubr2T2/T2 male.

To investigate the Ubr2T2/T2 testes in more detail, I dissected the epididymis of each testis from these mice and performed sperm counts using a haemocytometer. I calculated an estimated mean of 11.9x106±1.1x106 cells/ml mature sperm produced by a Ubr2+/+ testis, and an estimated mean of 10.3x106±1.4x106 cells/ml produced by a

Ubr2T2/T2 testis. The difference between these counts are not statistically significant

261

(Fig. 5.11B, p=0.46, Student’s T-test). Mean sperm counts in the Ubr2+/T2 testis were

9.7x106±2.3x106cells/ml, which again are not statistically significantly deviated from either Ubr2+/+ or Ubr2T2/T2 sperm counts (p=0.40 and p=0.83 respectively, Student’s T- tests). The sperm counted in this assay were generally motile and had a normal morphology, and so the Ubr2T2/T2 males do not have any overt defects in spermatogenesis, which are clearly detectable in the Ubr2-/- testis.

Finally, to assess the Ubr2T2/T2 testis morphology in more detail, I performed tissue sections and stained with H&E to look at the seminiferous tubules (Fig. 5.11C).

Spermatogenesis occurs within each seminiferous tubule and as spermatocytes develop and go through meiosis, they advance towards the lumen, which is the centre of the tubules in these sections. This is therefore a useful assay to look at the different stages of meiosis and ascertain an estimated stage of meiosis in which any cell death or morphological changes may occur. Ubr2-/-seminiferous tubules have an accumulation of zygotene-like arrested cells and absence of round and elongated mature spermatids, indicating a meiotic arrest and cell death, observed as large spaces in the seminiferous tubule lumen (Crichton et al. 2017). There are no clear morphological defects in the Ubr2T2/T2 sections as compared to Ubr2+/+ testes, and both round and elongated spermatids exist in the seminiferous tubules examined here.

Therefore, there is no grossly overt male meiotic arrest in the Ubr2T2/T2 testis.

Taken together, the data presented in this section shows that Ubr2T2/T2 males do not grossly phenocopy the overt spermatogenesis defects seen in Ubr2-/- males, and so the major functions of UBR2 in this tissue do not appear to act through Type II N-end

Rule substrate binding. TEX19.1 protein is stable, indicating that perhaps the interaction between this and UBR2 is required for spermatogenesis to occur.

262

A Testis Weights B Sperm Counts

n=8 n=4 n=11 n=4 n=3 n=7

+/+ T2/T2 C Ubr2 Ubr2

ST L ST

T2/T2 Figure 5.11 Ubr2 males have no overt defects in spermatogenesis

+/+ A. Testis weights, as an average of both testes per mouse in wild type (Ubr2 ), heterozygote

+/T2 T2/T2 (Ubr2 ) and homozygote D233A mutant (Ubr2 ) males at 8 weeks pp. Black bars are means. B. Counts of mature sperm from the same males as in (A), isolated from the epididymis. Black bars +/+ are means. C. Representative H&E stained 5µm-thick testis sections from wild type Ubr2 and

T2/T2 homozygous mutant Ubr2 males. Scale bar is 100µm. ST: seminiferous tubule, L: lumen, arrow indicates direction of meiotic progression, black arrowhead: elongated spermatids, white arrowheads: round spermatids.

263

5.2.3.3 The novel role of Ubr2 in cohesin regulation does not act through Type II N- end Rule

We have previously shown that Ubr2 has a role in splenocytes (J Reichmann et al.

2017) and oocytes (Chapter 3) in the regulation of the AcSMC3-containing subpopulation of cohesin complexes, which is thought to mediate sister chromatid cohesin. In Ubr2-/- splenocytes, chromatin-associated AcSMC3 increases, whereas in Ubr2-/- oocytes it decreases. Importantly, Ubr2 specifically regulates AcSMC3- containing cohesin complexes, which is a relatively minor population of cohesin in cells, and specific effects on this population typically do not cause detectable changes in the abundance of other cohesin subunits, such as total SMC3.

In order to investigate if this role of Ubr2 is perturbed in the Ubr2T2/T2 mutant mouse,

I first looked at the level of chromatin-associated cohesin levels from splenocytes isolated from the Ubr2T2/T2 spleen. Chromatin preparation from the Ubr2-/- spleen exhibits a ~2-fold increase in AcSMC3, and no detectable change in SMC3 or other cohesin subunits (J Reichmann et al. 2017). I assayed the level of total chromatin- associated cohesin in the Ubr2T2/T2 spleen using the SMC3 subunit, which is present in all cohesin complexes, and the AcSMC3 subpopulation, by Western blotting of chromatin prepared from the spleens of 3 independent mice (Fig. 5.12A). Upon quantification of the intensity of Western blot protein bands, normalised to the intensity of chromatin-associated histone H3 protein detected in these samples, I was unable to detect a significant change in either SMC3 or AcSMC3 protein levels (Fig.

5.12B, p=0.7172 and p=0.6515 respectively, Student’s T-Test). This implies that UBR2 acts to regulate mitotic AcSMC3 cohesins through a molecular pathway other than its Type II N-end Rule binding ability.

264

A +/+ T2/T2 +/+ T2/T2 kDa kDa 190 SMC3 190 AcSMC3 135 135

25 Histone H3 25 Histone H3 17 17

B 2

1.5

Associated Associated -

Protein Abundance Protein 0.5

Chromatin Relative 0 +/+ T2/T2 +/+ T2/T2 Ubr2 Genotype Ubr2 Genotype

SMC3 AcSMC3

Figure 5.12 Mitotic cohesin regulation is not perturbed in the Ubr2T2/T2 mouse A. Representative western blots of chromatin-associated proteins SMC3 and AcSMC3 isolated from spleens of wild type (+/+) and homozygous D233A mutant (T2/T2) mice, and chromatin-associated histone H3 protein as a loading control. B. Quantified mean protein levels of SMC3 and AcSMC3 from western blotting in A, corrected against histone H3 protein level, and relative to the wild type mean. Standard error over three pairs of wild type and homozygous mutant mice shown. Differences between wild type and T2/T2 are not significant; SMC3: p=0.72, AcSMC3: p=0.65 (Student’s T-test).

265

As well as a mitotic role for UBR2 in cohesin regulation, in Chapter 3 I outlined a role for UBR2 in regulation of the AcSMC3 cohesin subunit during female meiosis; in the

Ubr2-/- prometaphase I oocyte, AcSMC3 is significantly decreased by around 40% (Fig.

3.1, Chapter 3), implying a protective role of UBR2 for cohesin in this cell type. This contrasts with the decrease in chromatin-associated AcSMC3 in the mitotic spleen or in embryonic stem cells (ESCs) and so it appears to be an oocyte-specific role of

UBR2.

In order to assess whether the meiotic cohesin regulatory function of Ubr2 is perturbed by the Ubr2T2/T2 mutation, I isolated metaphase I oocytes from Ubr2T2/T2 and

Ubr2+/+ age-matched females by superovulation and carried out AcSMC3 and SMC3 immunofluorescence and quantification against ACA (anti-centromeric antibody) signal, as described in Chapter 3. Both AcSMC3 and SMC3 are detectable along the chromosome axis of metaphase I bivalents which are visualised using a DAPI DNA stain, in Ubr2T2/T2 and Ubr2+/+ controls (Fig. 5.13A, 5.14A).

In the Ubr2T2/T2 oocyte, the mean corrected AcSMC3 protein level is 105% of wild type protein level, which is not a significant change in protein level (Fig. 5.13B, p=0.89,

Student’s T-test). Ubr2T2/T2 SMC3 intensity, as an indicator of total chromatin-bound cohesin, is 111% of wild type level, which again is not a significant change (Fig. 5.14B, p=0.51, Student’s T-test). This implies that the AcSMC3-specific cohesin regulatory activity of UBR2 does not require Type II N-end Rule binding.

266

B AcSMC3 Antibody Signal

ive to ive ACA Signal

at AcSMC3Signal Rel

267

Figure 5.13 (previous page) UBR2’s role in meiotic AcSMC3 regulation is not perturbed in Ubr2T2/T2 oocytes A. Representative prometaphase I oocyte chromosome spreads from wild type (+/+) and homozygous D233A mutant (T2/T2) female mice, aged 9-14 weeks, with immunofluorescent staining for centromeric proteins (ACA/CREST antibody) in green and AcSMC3 in red, and DAPI in blue to mark DNA. Scale bars are 10µm. Inset images show an individual bivalent in 3- colour (left) and the AcSMC3 alone (greyscale, right) along chromosome axes. B. Median nuclear intensity of chromosome axial AcSMC3 protein, relative within each bivalent to ACA signal and each nucleus corrected to the wild type mean nuclear intensity within each biological repeat. Mean across oocytes over 3 biological repeats shown as black bar.

Figure 5.14 (next page) UBR2’s role in total meiotic cohesin regulation is not perturbed in Ubr2T2/T2 oocytes A. Representative prometaphase I oocyte chromosome spreads from wild type (+/+) and homozygous D233A mutant (T2/T2) female mice, aged 9-14 weeks, with immunofluorescent staining for centromeric proteins (ACA/CREST antibody) in green and SMC3 in red, and DAPI in blue to mark DNA. Scale bars are 10µm. Inset images show an individual bivalent in 3- colour (left) and the SMC3 alone (greyscale, right) along chromosome axes. B. Median nuclear intensity of chromosome axial SMC3 protein, relative within each bivalent to ACA signal and each nucleus corrected to the wild type mean nuclear intensity within each biological repeat. Mean across oocytes over 3 biological repeats shown as black bar.

268

B SMC3 Antibody Signal

SMC3 Signal SMC3 to Relative ACA Signal

269

As AcSMC3 protein intensity is not directly linked to chromosome cohesion, in order to confirm that the Ubr2T2/T2 allele does not perturb this in post-natal oocytes, I next assessed cohesion along the chromosome arms of metaphase I oocytes by quantifying the number, proportion and position of chiasmata in Ubr2T2/T2 metaphase I oocytes

(Fig. 5.15A), as in Chapter 3.

As quantified in Chapter 3, a mean of 28.2 chiasmata were observed in a Ubr2-/- metaphase I nucleus, as compared to a mean of 27.4 in a control Ubr2+/+ nucleus. In a

Ubr2T2/T2 metaphase I nucleus, the mean number of chiasmata is 26.6±0.6 (± standard error), which is not a significant deviation from the mean of 27.9±1.1 observed in

Ubr2+/+ oocytes isolated from control females from the same colony (Fig. 5.15B, p=0.36,

Student’s T-test). Proportionally, in Chapter 3, a Ubr2-/- metaphase I nucleus contains a mean of 41.7% of bivalents with a single chiasma, and 58.3% with multiple (usually two or three). The Ubr2T2/T2 metaphase I oocytes contained 30.5% single-chiasma and

69.4% multiple-chiasmata bivalents, which is not significantly changed from the wild type proportions observed in oocytes isolated from littermate Ubr2+/+ females. 35.4% of Ubr2+/+ bivalents assessed contained only a single chiasma, and 64.6% contained multiple (Fig. 5.15C, p=0.92, χ2 test). Finally, the Ubr2-/- metaphase I oocyte exhibits a positional terminalisation of single chiasmata towards the telomeric end of the bivalent; 56.1% of Ubr2-/- single-chiasma bivalents are distal to the centromere, with only 5.7% detected in the centromere-proximal third of the chromosome (Chapter 3).

However, 9.8% of single-chiasma Ubr2+/+ bivalents and 14.4% of Ubr2T2/T2 bivalents were observed in the centromeric proximal third of the chromosome. 46.0% and

42.4%, respectively, were observed in the interstitial third, and 44.2% and 43.3% respectively, were observed in the distal, or telomeric, third of the chromosome’s length. This is not a significantly altered proportion of chiasma positions between the

270

Ubr2+/+ and Ubr2T2/T2 oocytes (Fig. 5.15D, p=0.4135, χ2 test). Therefore, the chiasmata

terminalisation observed in the Ubr2-/- metaphase I oocyte is not detectable with the

Ubr2T2/T2 allele, and so UBR2’s role in chromosome arm cohesion between sister

chromatids is not perturbed in absence of Type II N-end Rule functionality.

T2/T2 Figure 5.15 (next page) Ubr2 oocytes do not exhibit defects in chromosome arm cohesion

+/+ A. Representative Metaphase I oocyte chromosome spreads isolated from age-matched Ubr2 and

T2/T2 Ubr2 females. Inset: representative bivalents with a single or multiple chiasmata, indicated by

+/+ white arrows. B. Number of chiasmata present per 20-chromosome oocyte nucleus of Ubr2 (grey)

T2/T2 and Ubr2 (blue) Metaphase I oocytes. Mean values shown by black bar. C. Proportion of bivalents containing single or multiple chiasmata within each 20-chromosome nucleus. D. Position of chiasma relative to the centromeric end of each bivalent with only a single chiasma.

271

Single Chiasma Multiple Chiasmata

+/+ Ubr2 DAPI

T2/T2 Ubr2 DAPI

B Chiasmata Number per Nucleus C Proportions of Chiasmata D Chiasmata Positions

100% 100%

ns 80% 80% ns

60% 60%

ns 40% 40%

20% 20%

n=257 n=219 0% 0% n=166 n=152 +/+ -/- +/+ T2/T2 +/+ T2/T2 n=12 n=11 Ubr2 Genotype Ubr2 Genotype None Single Multiple Proximal Interstitial Distal Ubr2 Genotype

272

Taken together, the data in this section indicates that UBR2’s role in AcSMC3-

containing cohesin regulation, and meiotic chromosome arm cohesin, does not

depend on its ability to bind Type II N-end Rule substrates. Neither mitotic nor

meiotic AcSMC3, although affected in opposite directions in the Ubr2-/- mouse, are

misregulated in the Ubr2T2/T2 mouse.

5.2.3.4 Summary of the Ubr2T2/T2 mouse phenotype

Generation of the Ubr2D233A mouse line was a tool in which I could begin to separate

out the functional capacity of UBR2, in its most well characterised molecular

function: binding to Type II ubiquitination substrates in the N-end Rule pathway.

However, the three major phenotypes that I and others have identified in a Ubr2-/-

mouse were are not present in this mouse model, and so my act through another

molecular function of UBR2.

5.2.4 A stable Ubr2-T1D118A ESC line using CRISPR Cas9

5.2.4.1 Generation of Ubr2-T1D118A ESCs

Of the Ubr2T2/T2 mouse phenotypes, the most interesting to my own research is the

lack of cohesin phenotype in either the spleen or the oocyte. In addition, while

phenotyping the Ubr2T2/T2 mice, it became apparent that the role of UBR2 in regulating

cohesin could be modelled in ESCs in culture, which exhibit a specific upregulation

of AcSMC3 in the Ubr2-/- genotype. Therefore, to further investigate which activities

of UBR2 might be important for cohesin regulation, I designed a CRISPR-Cas9

experiment with homology-directed repair (HDR), with the aim to generate a stable

Ubr2-T1D118A mutant ES cell line, so perturbing the binding of UBR2 to Type I N-end

273

Rule peptides. The D118A point mutation was first identified in an Alanine-scanning mutagenesis screen of Ubr1 (T Tasaki et al. 2009), and acts to specifically perturb detectable binding of UBR2 to Type I N-end Rule peptides. Earlier in this chapter I confirmed that it is both sequence and functionally conserved in the Ubr2 gene, using ectopic expression of a FLAG-UBR2 construct in HEK293T cells (Fig. 5.3B and Fig.

5.4A). In order to carry out HDR, a dsDNA repair oligo with 50bp homology arms encoding the mutant repair template containing both the point mutation and the mutant PAM sequence was cotransfected into E14 ESCs alongside a Cas9 expressing plasmid and pX330 plasmid with the specific guide sequence cloned into it (Fig.

5.16A). These cells were transfected and clonal colonies picked by Fiona Kilanowski

(MRC HGU), and I carried out genetic screening to identify successfully edited clones.

For screening of successfully edited stable Ubr2-T1D118A clones, I designed tetra-ARMS

(amplification-refractory mutation system) PCR genotyping assays, as utilised for genotyping of Ubr2D233A mice and described in section 5.2.2 (Fig. 5.16B). I screened the

PCR reactions by gel electrophoresis (Fig. 5.16C). The control band is at 363bp; the

F10 clone was identified as a heterozygote, and the H10 clone was one of three successfully edited homozygous D118A clones identified from a 96-clone screen (3% success rate); the other two successful clones were designated H5 and E11. I then validated these clones by Sanger sequencing of the 363bp outer primer control fragment (Fig. 5.16D), which shows the GCC alanine-encoding codon in position 118.

In order to confirm that Ubr2-T1D118A is still expressed at a wild type level and the protein is stable, I carried out Western blotting to detect endogenously expressed

UBR2 levels (Fig. 5.16D). The three ESC clones H5, H10 and E11 were probed for

UBR2 protein and Lamin B1 as a loading control, alongside E14 wild type ESCs (+/+).

UBR2 protein is detectable in wild type E14 (+/+) cells, and in the two clones H5 and

274

H10. However, the E11 clone has no detectable UBR2 protein, although Sanger

sequencing data of the 363bp outer tetraARMS fragment shows a normal sequence

immediately around the CRISPR targeted region and successful introduction of the

GAC->GCC point mutation and silent PAM site mutation. I took this to mean that

there is an off-target mutation that destabilises the UBR2 protein, and therefore the

E11 clone was unusable as a Type I N-end Rule mutant cell line.

The H10 clone has a successfully incorporated GCC point mutation at position D118A,

and there is near-wild type levels of UBR2 protein detected and so, for the remainder

of this section, I use this as a stable Ubr2-T1D118A ES cell line to assess the role of UBR2’s

Type I N-end Rule function in cohesin regulation in ES cells.

D118A Figure 5.16 (next page) Ubr2-T1 mutant ES cell line screening A. DNA guide and repair template sequences for CRISPR. Green: point mutations introduced, both for D118A mutation and mutant PAM site. Bold: antisense PAM site used B. Schematic of tetraARMS-PCR, used for screening and genotyping of ESC clones during UBR2-T1 cell line generation. Outer primer and inner primer pairs. C. Representative gel electrophoresis showing tetraARMS-PCR products for T1 clone genotyping. 100kb ladder shown on left. Control (outer primer) band is 363bp. Wild type A allele band is 235bp and UBR2-T1 C allele band is 180bp.

D118A Successful clone H10 shown by *. D. Sanger sequencing data for the successful Ubr2-T1 (H10)

D118A ESC clone (GAC->GCC). E. Protein levels of UBR2 using Western blotting in 3 Ubr2-T1 clones,

-/- wild type E14 and Ubr2 ES cells. Protein product of housekeeping gene Lamin B1 shown as a loading control.

275

A Experiment Guide sequence Repair template sequence

Antisense: Mutant: GAGTGTTTAAAGTGGGGGAACCTACATACTCCTGCAGAGACTGTGCAG GCACTCCATGCA T1 (D118A) TTGCCCCCACATGTG^TTTTATGCATGGAGTGCTTCCTGGGAAGTATC TAAAA^CACAGG CATAGAGACCATCGATATAG

A10 B10 C10 D10 E10 F10 G10 B Tetra ARMS-PCR C H10*

1500 1000 C 500 400 A 300

200 363 100 235 Control 180 C allele A allele

D Clone H10: D118A GAC->GCC

E D118A Ubr2-T1 Clones

H10

E11 +/+

245kDa UBR2

190kDa

Lamin B1 80kDa

276

5.2.4.2 Ubr2-T1D118A ESCs exhibit no cohesin misregulation

In order to investigate whether the D118A mutation, which is predicted to affect binding to Type I substrates, perturbs the ability of UBR2 to regulate cohesin, I measured the level of chromatin-bound cohesin protein present in a G2/M-enriched cell population. I carried out synchronisation of wild type E14 and Ubr2-T1D118A cells to enrich a minimum of 60% of cells into G2/M. To do so, I used a double thymidine block (Fig. 5.17A) – in which the two cell lines were treated with 1.25mM thymidine, a drug that arrests the cell cycle at S phase. A FACS (fluorescence-associated cell sorting) plot of PI-stained (propidium iodide) cycling ES cells has two peaks, one at

G1 and one at G2/M, and a trough that I designated as S-phase. I carried out two rounds of thymidine blocking, resulting an a G1/S-boundary arrest (designated DT+0) before harvesting cells at DT+4, 4 hours after release from the second thymidine block, where cells were enriched in G2/M (Fig. 5.17B). I thresholded this enrichment at a minimum of 60% of harvested cells in G2/M at the DT+4 timepoint for all experiments.

277

2nd 2nd Chromatin Split 1st 1st thymidine release preparation cells thymidine release ST+5 DT+0 DT+4

26h 17h 5h 17h 4h

B G1 S G2/M G1 S G2/M

Cycling

DT+0

DT+4

D118A Wild type E14 Ubr2-T1

Figure 5.17 ES cell synchronisation to a G2/M-enriched population A. Cell treatment timeline for double thymidine block of ES cells. Cells were transferred into GMEM+serum/LIF with 1.25mM thymidine at 26h and 48h after treatment, with a release at 43h and 65h respectively. B. FACS profiles for cycling ES cells, and at timepoints DT+0 and DT+4 for D118A wild type (E14, +/+) and H10 (Ubr2-T1 ) ES cells. ST: single thymidine, DT: double thymidine.

278

In order to isolate chromatin-bound cohesin molecules, I carried out chromatin

preparation on harvested DT+4 cells. Using Western blotting, I looked at the protein

levels of total SMC3 (Fig. 5.18A) and AcSMC3 (Fig. 5.18B). I quantified the level of

each of these proteins, relative to the level of histone H3 present in each sample over

3 independent repeats, and showed that there is a no significant change in the level of

AcSMC3 in the Ubr2-T1D118A mutant cell line (Fig. 5.18C, p=0.23, Student’s T-test). This

contrasts with the 2.27-fold increase in AcSMC3 observed in the Ubr2-/- ES cells

(unpublished data, Karen Dobie, MRC HGU), and the ~2-fold increase observed in

the spleen of Ubr2-/- mice (J Reichmann et al. 2017). Quantification of total SMC3 also

shows no significant change in the level of total cohesin bound to chromatin in these

cells (Fig. 5.18C, p=0.50, Student’s T-test). As cohesin is loaded and unloaded from

chromatin dynamically throughout the cell cycle, although the mESCs were

synchronised to a minimum of 60% G2/M enriched cells, there is still a high level of

variation (standard error) in the cohesin quantified, as has been observed before in

the Ubr2-/- ESCs (unpublished data, Karen Dobie, MRC HGU).

The misregulation of AcSMC3 observed in Ubr2-/- ESCs is not reflected in the

AcSMC3 level in the Ubr2-T1D118A ES cell line. I conclude that UBR2’s role in regulation

of the acetylated subpopulation of cohesin does not act through the Type I N-end

Rule pathway.

Taken together with the data presented in section 5.2.3.3, I have determined that the novel role of Ubr2 in cohesin regulation that our group have recently described (J Reichmann et al.

2017) does not appear to act through either Type I or Type II UBR2 N-end Rule substrate binding, as predicted by the N-end Rule binding abilities of the Ubr2D118A and Ubr2D233A alleles demonstrated in ectopic expression studies in HEK293T cells (Fig. 5.3 and 5.4).

279

A B +/+ T1D118A +/+ T1D118A kDa kDa 190 190

135 SMC3 135 AcSMC3

17 Histone H3 17 Histone H3 11 11

C Chromatin-Associated Cohesin Abundance 1.2

0.8

0.6

Associated Associated Protein Abundance - 0.4

0.2

Relative Chromatin 0 WT T1 WT T1 Ubr2 Genotype SMC3 AcSMC3

Figure 5.18 Chromatin-bound cohesin levels at G2/M are unaffected in Ubr2D118A ESCs A. Representative western blotting for chromatin-associated SMC3 and histone H3 protein as a

D118A loading control, isolated from wild type (WT) E14 and Ubr2-T1 synchronised by a double thymidine block into G2/M. B. Representative western blotting for chromatin-associated AcSMC3 and histone H3 protein, as in A. C. Quantification ± standard error of chromatin- D118A associated SMC3 and AcSMC3 in Ubr2-T1 at G2/M over 3 independent repeats, normalised against H3 protein levels and relative to protein levels in wild type (WT) E14 cells.

280

5.3 Discussion

In this chapter, I carried out validation and implementation of genetic point mutations in the

Ubr2 gene in order to investigate the molecular basis of known essential functions of the E3 ubiquitin ligase. The point mutations D118A and D233A are predicted using in vitro binding assays to perturb UBR2’s ability to bind Type I and Type II N-end Rule substrates, respectively. As has been previously described, both in our group and by others (Kwon et al.

2003; Yang et al. 2010; Crichton et al. 2017; Reichmann et al. 2017 and Chapters 3 and 4), Ubr2 is a vital gene for a subset of specific functions in the mouse, which may be a result of UBR2’s most highly detectable function, the Type II N-end Rule (T Tasaki et al. 2005). These include regulation of trophectoderm-derived placental tissue development associated with perinatal survival, regulation of spermatogenesis during male meiosis, and regulation of the chromatin- associated AcSMC3 component of the cohesin complex across both mitotic and meiotic tissues. By stable introduction of D233A into the Ubr2 gene in a mouse line by CRISPR-Cas9 homology directed repair, I was able to investigate these three major phenotypes of a Ubr2-/- mouse, and showed that none are perturbed in a Ubr2D233A mouse, indicating that they do not act through the Type II N-end Rule ability of UBR2. This indicates a need for further investigation into these phenotypes in order to investigate how UBR2, and potentially other

UBR proteins, factor into these processes.

5.3.1 Lack of redundancy within the UBR family

Within the mouse testis, ovaries and placenta, the genes encoding all four UBR proteins shown to have N-end Rule binding activities (Ubr1, Ubr2, Ubr4 and Ubr5) are highly expressed.

This can be observed by analysis of mapped raw long-RNAseq data from the ENCODE long transcriptome project to the mm9 genome assembly (NCBI 37) and calculation of RPKM

(reads per kilobase million) corrected for gene length in kb, kilobases (Fig. 5.19, analysis by

281

Ian Adams, MRC HGU. Accession number GSE36025, Davis et al., 2018). Ubr2 is particularly highly expressed in the placenta and the ovaries, and the former has a similarly high expression level of Ubr5. In the testis, Ubr2 does not have particularly high expression, compared to the other N-end Rule UBR proteins and, indeed, Ubr5 is the most abundant of the four UBRs in this tissue. The D. melanogaster homolog of UBR5, Hyd, has been implicated in spermatogenesis in the fly (Pertceva et al. 2010), and our lab recently identified a TEX19.1- independent role for UBR5 in mouse spermatogenesis, although the molecular function of this is as yet unknown (unpublished data, Chris Playfoot, MRC HGU).

There is evidence for redundancy between the UBR proteins, particularly the highly homologous family members UBR1 and UBR2. Embryos lacking both proteins have impaired neurogenesis and cardiovascular development leading to embryonic death (Jee Young An et al. 2006), phenotypes which are not present in either single knockout (see Chapter 1, and

Kwon et al. 2001; Yang et al. 2010). Indeed, disturbance of the N-End Rule through single knockouts of UBR proteins only result in a few key phenotypes, which can be attributed to unique functions of each protein. In humans, either full or partial loss of UBR1 protein activity is associated with developmental disorder Johanson-Blizzard Syndrome (JBS), and UBR2 has vital roles in a subset of functions, as characterised in the Ubr2-/- mouse studied in Chapters 3 and 4, and in the following publications (Kwon et al. 2003, Yang et al., 2010; Crichton et al.,

2017). These independent functions could be attributed to differences in cell-specific expression of UBR proteins within a tissue, giving a specific dependence on one UBR protein for N-end Rule functions in certain cell types. For example, Ubr1 and Ubr2 are specifically expressed in spermatogonia and spermatocytes, respectively, in the mouse testis (Y T Kwon et al. 2003). However, in this chapter I showed that perturbation of the UBR2 Type II N-end

Rule in the testis does not result in spermatogenesis defects. Although the role for UBR2 in spermatogenesis may be a Type I N-end Rule function, it may also act through an as-yet uncharacterised molecular role of UBR2. Yeast paralog of the mouse UBR family, Ubr1, has

282 been shown to bind ubiquitination targets such as the CUP9 transcriptional repressor through an internal degron (Xia et al. 2008), and so UBR2 could be acting in a non-N-end

Rule ubiquitination function. Examples of these functions have been identified in mouse

UBR2, such as LINE-1 or histone H2A (Maclennan et al. 2017; Jee Young An et al. 2012). The functional domain through which UBR2 is able to bind and ubiquitinate these substrates is unknown, and so it would be interesting to look at any potential additional binding domains of the UBR2 protein, which is largely unstudied.

As a RING-class E3 ubiquitin ligase, UBR2 depends on binding to E2 ligases such as UBE2A/B and USE1 for ubiquitin transfer to target proteins through its RING domain (Deshaies and

Joazeiro 2009). The point mutation W1170A has been previously identified in the UBR2

RING domain as essential for binding of USE1 and UBE2A/B (C. W. P. Lee et al. 2011).

Therefore, generation of a Ubr2W1170A mutant mouse line would show whether any of the previously identified roles of UBR2 that do not present in an N-end Rule mutant act through a non-N-end Rule ubiquitination pathway, or potentially outside of ubiquitination altogether.

Although this is a highly conserved domain, UBR2 has been confirmed to act through a variety of other E2 ubiquitin conjugating enzymes, including UBE2O and HR6B (H. Xu et al. 2019; J

Y An et al. 2010), and so this point mutation may not necessarily prevent binding of all potential E2 ubiquitin ligases and not generate a fully penetrant catalytic dead mutation.

Coimmunoprecipitation assays could be carried out in order to test the extent of the mutation, as well as the conservation of the UBR2 RING domain structure between different potential

E2 interactors.

Figure 5.19 (next page) Ubiquitous coexpression of UBR proteins across mouse tissues Expression levels of Ubr2 and related UBR protein encoding genes Ubr1, Ubr4 and Ubr5, as measured by long RNAseq, ENCODE GEO (GSE36025, Davis et al., 2018), in RPKM (reads per kilobase per million mapped) across a variety of mouse tissues in the C57BL/6J genetic background. Analysis by Ian Adams (MRC HGU).

283

35 RNAseq Expression Analysis of UBR proteins

20 RPKM 15

Lung

Liver

Heart

Testis

Colon

Ovary

Cortex

Spleen

Kidney

Bladder

Thymus

Adrenal

Placenta

Stomach

SCFat Pad

Duodenum

Cerebellum

FrontalLobe

Small Intestine

LargeIntestine

GenitalFat Pad MammaryGland Ubr1 Ubr2 Ubr4 Ubr5

284

5.3.2 Limitations of this study

It should be noted that the study conducted in this chapter does not have the depth or scope to fully assess the variety of N-end Rule functions of UBR2. Initial investigation into the functions of the UBR N-end Rule E3 ubiquitin ligases led to the general assumption that

UBR2 has a very limited ability within the Type I N-end Rule pathway, instead acting to turn over Type II substrates (T Tasaki et al. 2005). However, recent studies have identified canonical Type I substrates, e.g. encoding an N-terminal arginine, that depend on the presence of one or both of UBR1 and UBR2, the most highly conserved members of the UBR family

(Rageul et al. 2019).

UBR1 is the best studied of the mammalian N-end Rule E3 ubiquitin ligases, and its UBR box has 47% sequence similarity and 68% structural similarity to UBR2. However, UBR1 is able to bind both Type I and Type II peptides to a similar level (T Tasaki et al. 2005). The point mutations utilised in this chapter were first identified in the Ubr1 gene using an alanine screen of the N-end Rule substrate-binding domains. D118 lies within the active site of the UBR box, and specifically conjugates a water molecule to the sidechain of N-terminal amino acid arginine, so giving the UBR box a highly specific binding to Type I substrates (E Matta-

Camacho et al. 2010; J Muñoz-Escobar et al. 2017; T Tasaki et al. 2009).

Regardless, the low detectable in vitro peptide affinity poses issues for the investigation of

Type I functions of UBR2 using a genetic approach, such as in this chapter. As noted in section

5.2.1 (Fig. 5.3), a TI peptide pulldown with wild type UBR2-FLAG protein shows only a low detectable binding affinity to Type I substrates. The UBR2-T1-FLAG construct protein, which contains the D118A mutation, appears to perturb this binding, but the variability of this assay, which is difficult to standardise against loading controls, means that this does not necessarily fully ablate Type I binding ability of the UBR box. Instead, the Ubr2D118A allele, when introduced into an ES cell line as in section 5.2.4 (Fig. 5.15), could be assayed in vivo for

Type I abilities. Ubr2-/- MEFs (mouse embryonic fibroblasts) exhibit a vulnerability to UV

285 irradiation as a result of a lack of Type I N-end Rule degradation of genomic stress regulator

SDE2 (Rageul et al. 2019). If this activity is also required in mESCs, tested by irradiation of the Ubr2-/- ESC line, the Ubr2D118A mESCs could also be used to test the ability of Ubr2D118A to process the SDE2 C-terminal fragment Type I substrate under stress by UV irradiation. This interaction could also be studied in vitro by ectopic expression of the mutant proteins in a cell line such as HEK293T and co-immunoprecipitation for transient protein interactions, although this would require proteasomal inhibition by a treatment such as MG132 to preserve protein substrates destined for ubiquitin-associated proteasomal degradation.

Additionally, although the peptide pulldowns for D233A show a clear reduction in Type II binding ability of UBR2 using both an ectopically overexpressed FLAG construct in

HEK293T cells (section 5.2.1, Fig. 5.3A) and endogenously expressed in the testis (section

5.2.2, Fig. 5.7), this does not necessarily demonstrate a fully ablation of Type II binding ability.

In particular, UBR2-T2 may still be able to bind Type II substrates that do not contain an N- terminal phenylalanine, which is the most common Type II sequence and was used as the TII peptide in this experiment (Fig. 5.3A), or with Type II-like substrates such as non-acetylated

Met-Φ proteins. Further examination of the mutant UBR2 proteins by peptide pulldown with different peptide sequences such as Met-Φ could be used to predict the extent of their binding ability. However, there may be no phenotypes to study, as the UBR proteins are predicted to be redundant in this capacity and UBR1 is able to recognise the Met-Φ N- terminus of known substrate proteins in the Ac/N-end pathway (H. K. Kim et al. 2014;

Nguyen et al. 2018).

5.3.3 Further study of the N-domain and UBR box structures

Interestingly, given the similarities between UBR1 and UBR2, the difference in Type I in vitro binding affinity between the proteins suggests that there is a very subtle change in sequence

286 or structure of the domains within these proteins. The isolated structure of the UBR box of

UBR2 alone is able to bind Type I peptides with a similar affinity to UBR1 in vitro (E Matta-

Camacho et al. 2010), and so there is always potential for Type I substrates, and perhaps another feature of the UBR2 protein structure favours binding to Type II substrates instead.

Type II N-end Rule binding occurs directly at the active site of the UBR box, but requires the

N-domain portion of the protein for Type II binding to occur. The N-domain shares structural similarity with the bacterial ClpS domain, which is a small 12kDa adaptor protein that carries out substrate specificity for the bacterial proteolytic ClpAP machinery (Gottesman et al.

1998). This domain does directly and specifically bind substrates, and so this could indicate a direct interaction between the N-domain and Type II substrates, stabilising the interaction between the UBR box and the N-terminus of the substrate and so favouring Type II degradation in UBR2. The UBR2 residue D233, which was mutated in this study, is conserved in ClpS as D35 and is thought to be important for broad substrate binding and stability of interaction; other residues in this domain are more important for specificity of binding (Tan et al. 2016). Therefore, the D233A mutation introduced could act to simply destabilise Type II interactions and reduce UBR2’s UBR box affinity towards them. As there is no apparent increase in Type I interactions, this implies that the relationship between these two functions of the UBR2 protein cannot be simplified this much, and there may be multiple physical contact points in protein-substrate interactions, as well as possible additional cofactor proteins or domains that may form a larger complex.

5.3.4 Summary

In this chapter, I have demonstrated that the two amino acids D118 and D233 are vital for N- end Rule functioning of the UBR2 protein, and that these canonical functions of UBR2 are not required for some of the major essential roles of the protein in the mouse.

287

Caution must be taken with drawing conclusions using these point mutations, as the full scope of potential N-terminal degron sequences is not covered by the in vitro validation using peptide pulldown assays and there could be some level of N-end Rule function within the mutant UBR2 proteins. However, the canonical phenylalanine Type II or arginine Type I degron binding ability is perturbed, and it appears that the majority of phenotypes identified using the Ubr2-/- mouse line do not act through these two N-end Rule pathways, and instead through a non-canonical function, whether involving ubiquitination or another function of the UBR2 protein altogether.

Further work to generate more genetic models that can be used to separate out the functions of UBR2 will give an insight into the full scope of UBR2’s roles in the mouse, and potentially lead to a greater understanding of the UBR protein family.

288

Chapter 6

Discussion and Conclusions

289

290

The main aim of the work presented in this thesis was to study the role of E3 ubiquitin ligase

UBR2 in the mouse, particularly in meiotic chromosome cohesion and in embryonic survival.

UBR2 is essential in the oocyte for stability of the AcSMC3 cohesin subunit, and loss of the protein results in a loss of chromosome cohesion during meiosis. UBR2 is also required for placental support of embryonic growth, but the molecular basis of its role in placental development is not yet understood.

Beyond the null phenotypes, I identified a gap in the scientific understanding of UBR2’s mechanism of action in the essential roles that have been identified. In order to study this using a genetic approach, I identified functional mutations within the substrate binding domains of UBR2 which perturb its role in the N-end Rule pathway of ubiquitination, and showed that the essential functions of the protein that have been identified in a null mouse are not perturbed in presence of the mutant alleles. This poses the question of non-N-end

Rule activities of UBR2. It also potentially highlights the redundant nature of UBR proteins, as the currently identified N-end Rule substrates of UBR2 can also be regulated by other proteins such as UBR1, whereas UBR2-specific functions may exist outside of the roles encompassed by the UBR protein family.

6.1 Separation of the TEX19.1 and UBR2 Functions

UBR2 is a ubiquitous protein, being expressed across a wide variety of tissues in the mouse, and it is of particular interest that the major phenotypes identified in a Ubr2-/- mouse arise in those tissues expressing TEX19.1, which is able to form a 1:1 stoichiometric complex with the

UBR2 protein (Yang et al. 2010; Maclennan et al. 2017). Indeed, although UBR2’s molecular

N-end Rule substrates span a wide variety of tissues, from the innate immune system to myocardial tissue growth (H. Xu et al. 2019; M. J. Lee et al. 2005), very few Ubr2-/- mouse phenotypes are thus far identified outside of the Tex19.1 expression pattern, for example the

291 misregulation of the acetylated SMC3 cohesin subunit in the splenocyte (J Reichmann et al.

2017). A mouse embryonic fibroblast (MEF) cell line, again a cellular context that does not express Tex19.1, exhibits UBR2 nuclear localisation and chromatin-associated ubiquitination.

This is a result of DNA damage, indicating a regulatory role for UBR2 in the DNA damage response, potentially through regulatory monoubiquitination of histone H2A (Jee Young An et al. 2012). This role has also been identified in silencing of the XY sex body in spermatocyte prophase I (J Y An et al. 2010). It is potentially independent of TEX19.1 in this function, as a

Tex19.1-/- spermatocyte arrests during a zygotene-like stage, as do a subset of Ubr2-/- spermatocytes, but does not appear to exhibit the later pachytene XY sex body silencing defects (Crichton et al. 2017; J Y An et al. 2010). Another possible explanation for increased

DNA damage in somatic cells could be a misregulation of the cohesin complex. Some cohesinopathies, such as WABS, exhibit a DNA breakage phenotype resulting from misregulation of cohesins during S phase and replication fork instability (Cortone et al. 2018;

Pisani et al. 2018). Therefore, large-scale chromatin capture assays or cohesin localisation could be assessed in these Ubr2-/- MEFs to ascertain the cause of genomic instability.

The role of Ubr2 in cohesin regulation is multifaceted. Although downregulation of AcSMC3 in the spleen is Tex19.1-independent, ESCs in culture express both Tex19.1 and Ubr2, and the

Ubr2-/- ESC exhibits a similar chromatin-associated loss of cohesin. Tex19.1-/- ESCs have opposite changes in chromatin-associated AcSMC3 intensity and so we hypothesised that, in this cell type, TEX19.1 acts to regulate UBR2’s control of AcSMC3 levels. However, as discussed in Chapter 3, the Ubr2-/- oocyte has a loss of AcSMC3-containing cohesin complexes associated with the chromosome axis and a reduction in chromosome cohesion which, unexpectedly, at least partially phenocopies the Tex19.1-/- oocyte (J Reichmann et al. 2017). A

Tex19.1-/- female has a pronounced increase in aneuploidy and chromosome missegregation that have not been observed in the Ubr2-/-, indicating that UBR2 and TEX19.1 may also have some

292 independent roles during female meiosis and that they are not directly working together in a

1:1 sense in these cells.

6.1.1 Spermatogenesis

In the testis, UBR2 is required for TEX19.1 protein stability (Yang et al. 2010), presumably due to a requirement for a stable TEX19.1-UBR2 complex. As a result, the mammalian Ubr2-/- spermatocyte, not unexpectedly, has at least a partial phenocopy of the recombination deficits and asynapsis observed in the Tex19.1-/- spermatocyte. However, this is not fully penetrant and surviving Ubr2-/- male mouse also exhibits a later phenotype, progressing into pachytene and exhibiting defects in histone H2A regulatory monoubiquitination required for chromosome stability and transcriptional inhibition in the XY sex body (Jee Young An et al. 2012; J Y An et al. 2010). Defects in these processes have been associated with mosaic expression of X- and Y- encoded genes, sperm heterogeneity and cell death (J. M. A. Turner 2007), and so it could be interesting to assay the chromatin structure of the XY sex body in a Ubr2-/- spermatocyte, using immunofluorescence against monoubiquitinated H2A.

This potentially bimodal functionality of UBR2 in spermatogenesis could be further investigated using genetic systems such as the Ubr2T2/T2 mouse; one, but not both phenotypes may be regulated through the N-end Rule activity of this E3. Therefore, although there are no gross phenotypes in the Ubr2T2/T2 testis from these mice, subtle changes in histone regulation throughout prophase I, or potential low-level cell death, could be assayed in order to separate out these functions of UBR2 upon a molecular basis.

293

6.1.2 Placental development

The most striking differences in phenotype between the Tex19.1-/- and Ubr2-/- mice can be observed in the placenta, outlined in Chapter 4. Both mice exhibit a partial embryonic lethality that presents at a perinatal stage, associated with an IUGR and changes in trophoblast-derived tissue development in the placenta (J Reichmann et al. 2013; Y T Kwon et al. 2003; Jee Young An et al. 2012). However, the Ubr2-/- exhibits no retrotransposition or upregulation of imprinted genes, which are characteristics of the Tex19.1-/- placental deficit

(Woods, Perez-Garcia, and Hemberger 2018; Tunster, Jensen, and John 2013; Garcia-Perez,

Widmann, and Adams 2016; Goodier 2016), posing questions about what role the UBR2 protein plays in placental development and if this involves TEX19.1 protein at all.

TEX19.1 is thought to act in a complex with KAP1 in the placenta, which is a transcriptional corepressor (Jang et al. 2018; Cheng, Kuo, and Ann 2014). This is involved in genome defence or regulation of transcriptional repression in a hypomethylated context (Cheng, Kuo and Ann,

2014 and Playfoot et al. 2019, manuscript submitted, MRC HGU), and their interaction has been hypothesised to be involved in the recruitment of the histone methyltransferase SETDB1, laying down transcriptionally repressive H3K9me3 histone methylation marks . It is not understood whether UBR2 factors into this complex, or if the TEX19.1-UBR2 complex is distinct from TEX19.1-KAP1, or potentially not active in this tissue at all.

The cohesin complex is tightly bound to a wide variety of cellular processes, and this includes genetic or chromosomal stability, mitotic chromosome segregation and regulation of cell cycle progression. All of these processes are vital to normal commencement of tissue growth, particularly in the placenta where the tissue is very fast-growing. For example, a null mutant of the known cohesin regulator ChlR1 helicase has severe cohesin defects in the placenta, leading to abnormal placental growth, perturbation of the cell cycle, aneuploid placental tissue and embryonic lethality at E10.5 (Inoue et al. 2007). Data in the Ubr2-/- splenocyte shows that UBR2 can regulate cohesin independently of TEX19.1, and this function does not depend

294 on the stability of the TEX19.1 protein, as has been observed in the testis (Yang et al. 2010).

However, when TEX19.1 is present it acts to regulates UBR2’s activity towards cohesin.

Therefore, the growth restrictions observed in the Ubr2-/- and the Tex19.1-/- placentas could be a result of cohesin misregulation, leading to aberrant somatic cell cycles and potentially aneuploidy, which could cause apoptosis. Cohesin misregulation could also halt or slow the cell cycle, reducing the rate of proliferation and resulting in a smaller tissue in that sense.

TGCs (trophoblast giant cells) may be particularly vulnerable to this process, as discussed in

Chapter 4.

6.1.3 The interaction between TEX19.1 and UBR2 proteins

As none of the phenotypes identified in the Ubr2 null mouse are grossly overt in a Ubr2T2/T2 mouse, it appears that these phenotypes do not act through the Type II N-end Rule. Further work will be required to separate the functions of TEX19.1 and UBR2, and ascertain the importance of the TEX19.1-UBR2 interaction. This could include perturbation of the TEX19.1-

UBR2 interaction itself.

The molecular interaction site between TEX19.1 and UBR2 has not been mapped; UBR2 is a very large 200kDa (1755 amino acid, aa) protein (T Tasaki et al. 2009) and there is a significant central region of the protein that does not have any known or predicted domain or structural data. It is generally understood that TEX19.1 is not a substrate of UBR2, instead acting as a cofactor or regulatory protein, as the protein is not degraded in presence of UBR2 (Maclennan et al. 2017) – indeed, UBR2 is required for TEX19.1 stability in the testis (Yang et al. 2010).

Other interactors of UBR2 have also been identified in a stable complex with the E3 – for example RECQL4, which is a DNA helicase thought to be responsible for genetic stability through its interaction with the N-end Rule pathway (Yin et al. 2004).

295

Only a small proportion of the UBR2 protein has been mapped or studied in a structural assay, i.e. the UBR box, N-domain, RING domain and the UAIN domain. Bioinformatic analyses could be applied to identify conserved structural motifs within the large central region of the

UBR2 protein, as compared to known mouse proteins, and structural folding could be predicted. This could lead to predictions of vital domains, which could carry out binding to either target proteins outside of the N-end Rule, to ubiquitin or ubiquitin-like structures, or to cofactor proteins such as TEX19.1 and RECQL4.

Upon identification of candidate domains for TEX19.1 interaction, fragments of the UBR2 protein could be isolated for in vitro binding assays, with a view to eventually generate point mutant or deletion mutant UBR2 proteins that are functional in N-end Rule or non N-end

Rule abilities, but not able to detectably interact with TEX19.1. This is a similar approach as to that utilised in this thesis for perturbation of the N-end Rule binding abilities of UBR2, and as discussed in Chapter 5 could also be applied to the RING domain, for generation of a

‘catalytic dead’ UBR2 that is unable to transfer ubiquitin to target proteins. Together, these would be useful tools to ascertain the importance of the TEX19.1-UBR2 complex and separate the functions of these proteins in vivo. While generating multiple mouse lines to incorporate separation-of-function mutants may be costly or time-consuming, the Ubr2D118A mESC line utilised in this thesis demonstrates that we are able to assay at least some of UBR2’s abilities using this approach, particularly cohesin regulation (Chapter 5) or retrotransposon derepression (Maclennan et al. 2017). Interesting candidate point mutations could then be extrapolated into a mouse model for study of essential UBR2 functions in either male or female meiosis, or in the complex placental developmental process.

296

6.2 Redundancy in the UBR Protein Family

As discussed in Chapter 1, there are seven UBR proteins encoded by the mammalian genome, which have arisen through duplication events during evolution (T Tasaki et al. 2005, 2009).

The S. cerevisiae genome encodes only two proteins containing the UBR box domain: UBR1 and

UBR2. UBR1 has documented activity in the N-end Rule, whereas UBR2 is active in non-N- end Rule ubiquitination pathways (Akiyoshi et al. 2013). Mammalian UBR1 and UBR2 are 68% sequence conserved (T Tasaki et al. 2009), and probably arose as a duplication event quite recently during evolution. The only other mammalian UBR proteins that act in the N-end

Rule are UBR4 and UBR5. UBR4 has no known E2 binding domain, and UBR5 carries out ubiquitination through a HECT domain (Edna Matta-Camacho et al. 2012). Finally, UBR3 is known to undergo non-N-end Rule ubiquitination activity (see Chapter 1) and is most closely sequence conserved to S. cerevisiae UBR2.

The large number of E3 ubiquitin ligases encoded in the mammalian genome, the largest class of ubiquitination enzymes and encoded by over 600 genes (Wei Li et al. 2008), implies that there should be little functional redundancy between E3’s. Even within families sharing the same substrate binding domain, variation in the E2 binding mechanism leads to differences in the ubiquitination activity of each E3 (Deshaies and Joazeiro 2009; Metzger et al. 2014). UBR4 and UBR5 are able to bind the same peptide sequences as UBR1 and UBR2 in vitro, but the variation in the rest of their protein structure indicates that substrates could vary between

E3’s (T Tasaki et al. 2009, 2005), allowing the N-end Rule to cover a larger proportion of substrates in the mammalian system as compared to S. cerevisiae.

UBR1 and UBR2, being so highly conserved, share a number of substrates. RGS4 and RGS5 are N-terminally arginylated and levels are controlled by the Type I N-end Rule pathway. This is mediated by both UBR1 and UBR2, implying either a co-operative mechanism involving both E3 ligases, or a redundancy of RGS proteins as substrates for both, which are abundant in myocardial cells and so cannot be regulated by one alone (M. J. Lee et al. 2005).

297

Interestingly, another G-protein regulator RGS2 is regulated by the Ac/N-end Rule pathway, being N-terminally acetylated upon synthesis and then shielded in complex from E3 ubiquitin ligase Teb4 (Park et al. 2015), which demonstrates the level of diversity across N-end Rule regulatory pathways. Another shared in vivo substrate of UBR1 and UBR2 is SDE2, which is a

DNA stress response regulator; upon cleavage, the neo-N-terminus of the C-fragment is a

Type I N-terminal degron that requires both UBR1 and UBR2 for degradation (Rageul et al.

2019).

Finally, a Ubr1-/-Ubr2-/- double null mouse exhibits a fully penetrant embryonic lethality, associated with defects in both cardiovascular development and neurogenesis. Ubr1-/- mice alone survive to adulthood, exhibiting growth defects in the skeletal and adipose tissue (Y. T.

Kwon et al. 2001), and Ubr2-/- mice have a partially penetrant embryonic lethality which exhibit no grossly abnormal embryonic features and is based on placental growth defects

(Kwon et al., 2003 and Chapter 4). This indicates a shared substrate during these developmental events, either their already identified UBR1 or UBR2 substrates, or through an as-yet unidentified shared N-end Rule substrate.

One of the major pathways of the UPS is the misfolded protein response, in which chaperones act to bind and guide unfolded or orphan proteins for degradation via the proteasome. The traditional Arg/N-end Rule relies on sequence-specific N-terminal interactions, which does not lend itself to the versatility of misfolded protein degradation. However, UBR proteins have been implicated in degradation of the non-acetylated Ac/N-end Rule substrates, which are known as Met-φ (J. M. Kim and Hwang 2014; Hwang, Shemorry, and Varshavsky 2010).

This pathway is applicable to a wide variety of misfolded or orphan substrates. Moreover, there has been an N-end Rule like pathway identified in ER chaperone-mediated protein degradation (D. H. Kwon et al. 2018; Cha-Molstad et al. 2015). Through this pathway, the chaperone BiP itself is arginylated to become an N-end Rule like substrate, allowing for binding to a highly variable pool of substrates and degradation of both Arg-BiP and the

298 substrate together via the autophagy pathway. The presence of these peripheral N-end Rule pathways highlight the adaptability of the system beyond simple N-terminal degron binding and degradation in the UPS.

The extent of the redundancy is difficult to study, as we cannot predict whether any other

UBR protein or other E3’s may be involved in particular phenotypes or regulation of certain substrates. Redundancy could take multiple forms: all UBR’s present in specific cell types may be available to access substrates at all times, working together to control protein levels, or different UBR proteins could, either directly or indirectly, control levels of each other, so acting as a “backup plan” and being upregulated in response to loss of another UBR protein or E3 ubiquitin ligase. Therefore, mass spectrometry in order to study the proteome in presence or absence of a protein such as UBR2, or in cells in which UBR2 is inactivated or perturbed using the point mutations discussed in section 6.1.3, could give insights into the balance of E3 ubiquitin ligases in a proteome and whether any crosstalk or redundancy exists between related proteins. It could also give insights into the extent at which point mutation results in perturbation of the UBR2 protein function.

299

300

6.3 Further Applications for UBR2 N-end Rule Point Mutations

Within this thesis, I was able to identify and validate two key point mutations in the Ubr2 gene, and I have only scratched the surface of the potential investigation of UBR2 protein functionality utilising these mutations. Our group and others have been able to identify a novel role for UBR2 in the regulation of the cohesin complex across a variety of tissues, as well as in meiosis and placental growth (Kwon et al., 2003; Yang et al., 2010; Crichton et al., 2017;

Reichmann et al., 2017, and Chapters 3 and 4). This work implies that UBR2 does not necessarily rely on the N-end Rule for all functionality, but there are limits to using this genetic approach and after a phenotype has been identified, it is difficult to link this to a molecular interaction or a particular ubiquitination substrate. Therefore, if these genetic models can be combined with a proteomic approach, this will greatly augment the power of our investigation into the effect of Ubr2-/-.

6.3.1 Identification of UBR2 Substrates and Interactors

Proteomic screening through mass spectrometry is a high-powered way of quickly identifying significant changes in the proteome as a result of genetic changes. By comparing the proteome of Ubr2-/- mouse tissues with either wild type or Ubr2T2/T2 tissues, we could elucidate the molecular changes as a result of loss of UBR2 protein and/or loss of the Type II UBR2 function.

For example, we know that TEX19.1 protein is destabilised in absence of UBR2 in the mouse testis, but the UBR2-T2 protein is able to bind TEX19.1 protein, and TEX19.1 is stable in the

Ubr2T2/T2 testis. These experiments would give insights into the changes in protein dynamics as a result of loss of UBR2 protein, and what is specifically affected by the Type II N-end Rule binding activity of the protein. This would show both direct and indirect effects of the loss of

UBR2 protein, including the extent of UBR protein redundancy as a function of the protein degradation landscape with a lack of UBR2 protein, or a lack of its Type II functionality.

301

Mass spectrometry associated with co-immunoprecipation (coIP) with the UBR2 protein would potentially identify UBR2 interactions of three categories: novel E2 ubiquitin conjugating enzymes, other interacting cofactors such as TEX19.1 or RECQL4, or novel ubiquitination substrates. The TEX19.1-UBR2 interaction was identified in this way in a coIP mass spectrometry assay against TEX19.1 protein (Maclennan et al. 2017). However, transient interactions are common for an E3: both with E2’s, as they are continuously charging and transferring ubiquitin molecules, and with substrates that are short-lived and degraded via the proteasome (Iconomou and Saunders 2016). These can be difficult to detect via a standard co-immunoprecipitation assay, and as a result the majority of direct UBR2 substrates have been identified from the perspective of each individual substrate. In order to systematically and comprehensively compile a list of potential UBR2 substrates, so identifying candidates for the essential UBR2 functions identified in this thesis, I would propose the use of techniques such as BioID (biotinylated protein identification) or TurboID. Both these techniques utilise a proximity-based ligation mechanism, in which a fusion protein is constructed with the protein of interest and expressed in a cell line – for example in ESCs, which are known to require UBR2 for both cohesin and retrotransposon repression – alongside substrate treatment, allowing selective isolation and identification of transiently interacting proteins with affinity capture followed by mass spectrometry (Roux, Kim, and

Burke 2013; Doerr 2018). BioID has been used to success, in combination with MG-132 proteasomal inhibition, on a RING-type E3 complex: the (SCF)β-TrCP1/2 E3 ligases (Coyaud et al. 2015), in which over 50 candidate substrates were identified. Further to this, UBR2-T1 or

UBR2-T2 proteins could be used as a BioID fusion protein in ESCs to assess Type I or Type II

N-end Rule specific interactions, or the assay could be carried out using a wild type UBR2-

BioID fusion protein on a Tex19.1-/- background to assess if presence of TEX19.1 directly affects the complement of substrates able to bind to the UBR2 protein.

302

6.3.2 Further study of the UBR2 N-end Rule Functions

Although the conclusions of this thesis largely focus on the importance of studying non-N- end Rule activities of UBR2, it cannot be ignored that there are many canonical N-end Rule activities of the protein as well. UBR2, along with UBR1, is the canonical N-end Rule E3 substrate interactor, containing both the UBR box and N-domain, and there are many known

N-end Rule substrates of either protein, or both. Introduction of the D118A and D233A point mutations into either the UBR box or the N-domain gives a system in which these N-end Rule functions can be assessed in vivo.

As an example, one such recently identified role of UBR2 within the N-end Rule is turnover of the NLRP1B inflammasome C-terminal fragment in macrophage cells. This is cleaved by anthrax lethal toxin (LT) under infection with B. anthracis bacteria, resulting in degradation of the C-terminal fragment, which contains a Type II neo-N-terminal degron, by UBR1 or

UBR2. This allows the N-terminal fragment to initiate an innate immune response through caspase-1 activation (H. Xu et al. 2019).

This role of UBR2 in the innate immune response was first identified in the mouse macrophage-like RAW 264.7 reporter cell line, but this has not been directly documented in the Ubr2-/- mouse. It would therefore be interesting to induce the B. anthracis immune response in these mice, and mice of the Ubr2T2/T2 genotype to observe the necessity of UBR2 in this pathway in vivo.

Overall, we have only scratched the surface of the potential roles of the N-end Rule in protein degradation and regulation in the eukaryotic system. Tools such as the D118A and D233A point mutations will only further insights into the diversity of the UBR protein family, outlining the importance of protein control and regulation through ubiquitination, a vital and highly adaptable molecular process.

303

304

6.4 Conclusions

UBR2 is an E3 ubiquitin ligase that is primarily implicated in the N-end Rule pathway of the

UPS, particularly through Type II N-terminal degron binding of substrate proteins (T Tasaki et al. 2005, 2009). Thus far, prior to this thesis, UBR2 was identified as an essential regulator of meiotic recombination in the male germline (Yang et al. 2010; Y T Kwon et al. 2003; J Y An et al. 2010). Aside from this function, which appears to depend on the presence of TEX19.1, this is the first study pertaining to UBR2’s role in female meiosis or embryonic survival as a consequence of placental development, the latter of which was identified in 2003 but has not been directly studied (Kwon et al.). At this current time, there is no published work directly addressing the in vivo roles of mouse UBR proteins within the Type I and Type II N-end Rule using a genetic approach.

There is evidence for redundancy between these UBR proteins, particularly between UBR1 and UBR2 (Jee Young An et al. 2006; T Tasaki et al. 2009), and so in depth understanding of the molecular activity of UBR2 in these tissues will lead to a further insight into the wider

UBR family. It is already understood that at least three of the seven mammalian UBR proteins do not primarily use their UBR box as a substrate recognition domain for the ubiquitination pathway, whether they are active within it at all, and although UBR1, UBR2, UBR4 and UBR5 all have in vitro activity in binding to either Type I or Type II substrates (T Tasaki et al. 2005), there is no guarantee that this is their sole mechanism of activity, either within our outside of ubiquitination. Indeed, it appears that the N-end Rule redundancy between these proteins allows a flexibility that can be accounted for in much of the identified N-end Rule UBR2 protein functions and so there may be other, more vital, mechanisms of UBR2 action that have no such compensatory backstop.

One of the more interesting aspects of the UBR2 protein’s activities is its interaction with the genome defence protein TEX19.1, encoded by the gene Tex19.1 and only expressed in the hypomethylated contexts of the placenta, the oocyte, the spermatocyte and, in culture, ESCs.

305

This interaction has been shown to regulate the Type II N-end Rule turnover of the UBR2 ligase (J Reichmann et al. 2017), but this may not necessarily be the primary function of the

TEX19.1-UBR2 complex within these tissues. TEX19.1 itself has been shown to be implicated in a range of UBR2-independent functions, and so future work will be key to ascertain the extent of the TEX19.1-UBR2 complex and what role this complex plays in regulation of either protein’s other functions.

Although the N-end Rule has long been recognised as a primary substrate recognition pathway of the UPS and protein degradation, our current understanding of E3 ubiquitin ligases is shifting towards non-canonical roles, whether within regulatory functions of ubiquitination, or outside of the ubiquitin-proteasome system altogether. The major finding of this thesis is that UBR2 does not appear to act through its presumed activity in the Type II

N-end Rule for a variety of essential functions, and so we must consider the possibilities of different molecular abilities within this protein by combining both genetic and proteomic approaches.

306

References

Adhikary, Santanu, Deepavali Chakravarti, Christopher Terranova, Isha Sengupta, Mayinuer Maitituoheti, Anirban Dasgupta, Dushyant Kumar Srivastava, et al. 2019. “Atypical Plant Homeodomain of UBR7 Functions as an H2BK120Ub Ligase and Breast Tumor Suppressor.” Nature Communications 10 (1): 1398. https://doi.org/10.1038/s41467-019- 08986-5.

Agostinho, Ana, Otto Manneberg, Robin Schendel, Abrahan Hernández‐Hernández, Anna Kouznetsova, Hans Blom, Hjalmar Brismar, and Christer Höög. 2016. “High Density of REC 8 Constrains Sister Chromatid Axes and Prevents Illegitimate Synaptonemal Complex Formation.” EMBO Reports 17 (6): 901–13. https://doi.org/10.15252/embr.201642030.

Aihara, Ken Ichi, Hiroyuki Azuma, Masashi Akaike, Yasumasa Ikeda, Masataka Sata, Nobuyuki Takamori, Shusuke Yagi, et al. 2007. “Strain-Dependent Embryonic Lethality and Exaggerated Vascular Remodeling in Heparin Cofactor II-Deficient Mice.” Journal of Clinical Investigation 117 (6): 1514–26. https://doi.org/10.1172/JCI27095.

Akera, T, L Chmátal, E Trimm, K Yang, C Aonbangkhen, D M Chenoweth, C Janke, R M Schultz, and M A Lampson. 2017. “Spindle Asymmetry Drives Non-Mendelian Chromosome Segregation.” Science 358 (6363): 668–72. https://doi.org/10.1126/science.aan0092.

Akiyoshi, B, C R Nelson, N Duggan, S Ceto, J A Ranish, and S Biggins. 2013. “The Mub1/Ubr2 Ubiquitin Ligase Complex Regulates the Conserved Dsn1 Kinetochore Protein.” PLoS Genetics 9 (2): e1003216. https://doi.org/10.1371/journal.pgen.1003216.

Al-Hakim, Abdallah K., Anna Zagorska, Louise Chapman, Maria Deak, Mark Peggie, and Dario R. Alessi. 2008. “Control of AMPK-Related Kinases by USP9X and Atypical Lys 29/Lys33-Linked Polyubiquitin Chains.” Biochemical Journal 411 (2): 249–60. https://doi.org/10.1042/BJ20080067.

Albertini, D F. 2004. “Micromanagement of the Ovarian Follicle Reserve – Do Stem Cells Play into the Ledger?” Reproduction 127 (5): 513–14. https://doi.org/10.1530/rep.1.00247.

Alkhunaizi, Ebba, Robert M Brosh, Fowzan S Alkuraya, and David Chitayat. 2019. Warsaw Syndrome. GeneReviews. http://www.ncbi.nlm.nih.gov/pubmed/31169992.

307

Allers, Thorsten, and Michael Lichten. 2001. “Differential Timing and Control of Noncrossover and Crossover Recombination during Meiosis.” Cell 106 (1): 47–57. https://doi.org/10.1016/S0092-8674(01)00416-0.

An, J Y, E A Kim, Y Jiang, A Zakrzewska, D E Kim, M J Lee, I Mook-Jung, Y Zhang, and Y T Kwon. 2010. “UBR2 Mediates Transcriptional Silencing during Spermatogenesis via Histone Ubiquitination.” Proceedings of the National Academy of Sciences 107 (5): 1912–17. https://doi.org/10.1073/pnas.0910267107.

An, Jee Young, Euna Kim, Adriana Zakrzewska, Young Dong Yoo, Jun Min Jang, Dong Hoon Han, Min Jae Lee, et al. 2012. “UBR2 of the N-End Rule Pathway Is Required for Chromosome Stability via Histone Ubiquitylation in Spermatocytes and Somatic Cells.” Edited by Joel R. Drevet. PLoS ONE 7 (5): e37414. https://doi.org/10.1371/journal.pone.0037414.

An, Jee Young, Jai Wha Seo, Takafumi Tasaki, Min Jae Lee, Alexander Varshavsky, Yong Tae Kwon, and Alexander Varshavsky. 2006. “Impaired Neurogenesis and Cardiovascular Development in Mice Lacking the E3 Ubiquitin Ligases UBR1 and UBR2 of the N-End Rule Pathway.” Proceedings of the National Academy of Sciences of the United States of America 103 (16): 6212–17. https://doi.org/10.1073/pnas.0601700103.

Anderson, Lorinda K., Aaron Reeves, Lisa M. Webb, and Terry Ashley. 1999. “Distribution of Crossing over on Mouse Synaptonemal Complexes Using Immunofluorescent Localization of MLH1 Protein.” Genetics 151 (4): 1569–79. https://doi.org/10.1371/journal.pgen.1006389.

Bachmair, A, D Finley, and A Varshavsky. 1986. “In Vivo Half-Life of a Protein Is a Function of Its Amino-Terminal Residue.” Science 234 (4773): 179–86. https://doi.org/10.1126/science.3018930.

Baier, Andrea, Manfred Alsheimer, and Ricardo Benavente. 2007. “Synaptonemal Complex Protein SYCP3: Conserved Polymerization Properties among Vertebrates.” Biochimica et Biophysica Acta - Proteins and Proteomics 1774 (5): 595–602. https://doi.org/10.1016/j.bbapap.2007.03.008.

Baker, D J, K B Jeganathan, J D Cameron, M Thompson, S Juneja, A Kopecka, R Kumar, et al. 2004. “BubR1 Insufficiency Causes Early Onset of Aging-Associated Phenotypes and Infertility in Mice.” Nature Genetics 36 (7): 744–49. https://doi.org/10.1038/ng1382.

308

Baker, Sean M., Annemieke W. Plug, Tomas A. Prolla, C. Eric Bronner, Allie C. Harris, Xiang Yao, Donna-Marie M Christie, et al. 1996. “Involvement of Mouse Mlh1 in DNA Mismatch Repair and Meiotic Crossing Over.” Nature Genetics 13 (3): 336–42. https://doi.org/10.1038/ng0796-336.

Banerji, Rajeswari, Robert V. Skibbens, and M. Kathryn Iovine. 2017. “How Many Roads Lead to Cohesinopathies?” Developmental Dynamics. https://doi.org/10.1002/dvdy.24510.

Bartel, B, I Wünning, and A Varshavsky. 1990. “The Recognition Component of the N-End Rule Pathway.” EMBO Journal 9 (10): 3179–89. http://www.ncbi.nlm.nih.gov/pubmed/2209542.

Barton, Sheila C., M. A. H. Surani, and M. L. Norris. 1984. “Role of Paternal and Maternal Genomes in Mouse Development.” Nature 311 (5984): 374–76. https://doi.org/10.1038/311374a0.

Bascom-Slack, C A, L O Ross, and D S Dawson. 1997. “Chiasmata, Crossovers, and Meiotic Chromosome Segregation.” Advances in Genetics 35: 253–84. http://www.ncbi.nlm.nih.gov/pubmed/9348650.

Baudat, Frédéric, and Bernard de Massy. 2007. “Regulating Double-Stranded DNA Break Repair towards Crossover or Non-Crossover during Mammalian Meiosis.” Chromosome Research 15 (5): 565–77. https://doi.org/10.1007/s10577-007-1140-3.

Baumann, Claudia, Xiaotian Wang, Luhan Yang, and Maria M Viveiros. 2017. “Error-Prone Meiotic Division and Subfertility in Mice with Oocyte-Conditional Knockdown of Pericentrin.” Journal of Cell Science 130 (7): 1251–62. https://doi.org/10.1242/jcs.196188.

Beauchamp, B, S Ghosh, M W Dysart, G N Kanaan, A Chu, A Blais, K Rajamanickam, E C Tsai, M. E. Patti, and M. E. Harper. 2015. “Low Birth Weight Is Associated with Adiposity, Impaired Skeletal Muscle Energetics and Weight Loss Resistance in Mice.” International Journal of Obesity 39 (4): 702–11. https://doi.org/10.1038/ijo.2014.120.

Bedford, Lynn, Simon Paine, Paul W. Sheppard, R. John Mayer, and Jeroen Roelofs. 2010. “Assembly, Structure, and Function of the 26S Proteasome.” Trends in Cell Biology. Annual Reviews. https://doi.org/10.1016/j.tcb.2010.03.007.

Ben-Shahar, Tom Rolef, Sebastian Heeger, Chris Lehane, Philip East, Helen Flynn, Mark Skehel, and Frank Uhlmann. 2008. “Eco1-Dependent Cohesin Acetylation during Establishment of Sister Chromatid Cohesion.” Science 321 (5888): 563–66.

309

https://doi.org/10.1126/science.1157774.

Berndsen, Christopher E., and Cynthia Wolberger. 2014. “New Insights into Ubiquitin E3 Ligase Mechanism.” Nature Structural and Molecular Biology 21 (4): 301–7. https://doi.org/10.1038/nsmb.2780.

Bhattacharyya, Tanmoy, Michael Walker, Natalie R. Powers, Catherine Brunton, Alexander D. Fine, Petko M. Petkov, and Mary Ann Handel. 2019. “Prdm9 and Meiotic Cohesin Proteins Cooperatively Promote DNA Double-Strand Break Formation in Mammalian Spermatocytes.” Current Biology 29 (6): 1002-1018.e7. https://doi.org/10.1016/j.cub.2019.02.007.

Bielanska, M. 2002. “Chromosomal Mosaicism throughout Human Preimplantation Development in Vitro: Incidence, Type, and Relevance to Embryo Outcome.” Human Reproduction 17 (2): 413–19. https://doi.org/10.1093/humrep/17.2.413.

Biswas, Uddipta, Michelle Stevense, and Rolf Jessberger. 2018. “SMC1α Substitutes for Many Meiotic Functions of SMC1β but Cannot Protect Telomeres from Damage.” Current Biology, 1–25. https://doi.org/10.1016/j.cub.2017.12.020.

Blengini, Cecilia S., and Karen Schindler. 2018. “Immunofluorescence Technique to Detect Subcellular Structures Critical to Oocyte Maturation.” In Methods in Molecular Biology (Clifton, N.J.), 1818:67–76. https://doi.org/10.1007/978-1-4939-8603-3_8.

Bolcun-Filas, Ewelina, Yael Costa, Robert Speed, Mary Taggart, Ricardo Benavente, Dirk G De Rooij, and Howard J Cooke. 2007. “SYCE2 Is Required for Synaptonemal Complex Assembly, Double Strand Break Repair, and Homologous Recombination.” Journal of Cell Biology 176 (6): 741–47. https://doi.org/10.1083/jcb.200610027.

Borton, Michael T., Maisha S. Rashid, Megan R. Dreier, and William R. Taylor. 2016. “Multiple Levels of Regulation of Sororin by Cdk1 and Aurora B.” Journal of Cellular Biochemistry 117 (2): 351–60. https://doi.org/10.1002/jcb.25277.

Brieño-Enríquez, M A, S L Moak, M Toledo, J J Filter, S Gray, J L Barbero, P E Cohen, and J K Holloway. 2016. “Cohesin Removal along the Chromosome Arms during the First Meiotic Division Depends on a NEK1-PP1γ-WAPL Axis in the Mouse.” Cell Reports 17 (4): 977–86. https://doi.org/10.1016/j.celrep.2016.09.059.

Briley, Shawn M, Susmita Jasti, Jennifer M McCracken, Jessica E Hornick, Barbara Fegley, Michele T Pritchard, and Francesca E Duncan. 2016. “Reproductive Age-Associated

310

Fibrosis in the Stroma of the Mammalian Ovary.” Reproduction, September, 245–60. https://doi.org/10.1530/REP-16-0129.

Broman, Karl W., Lucy B. Rowe, Gary A. Churchill, and Ken Paigen. 2002. “Crossover Interference in the Mouse.” Genetics 160 (3).

Brooker, Amanda S, and Karen M Berkowitz. 2014. “The Roles of Cohesins in Mitosis, Meiosis, and Human Health and Disease.” Methods in Molecular Biology 1170: 229–66. https://doi.org/10.1007/978-1-4939-0888-2_11.

Bukovsky, Antonin, Michael R Caudle, Marta Svetlikova, and Nirmala B Upadhyaya. 2004. “Origin of Germ Cells and Formation of New Primary Follicles in Adult Human Ovaries.” Reproductive Biology and Endocrinology 2 (1): 20. https://doi.org/10.1186/1477-7827-2-20.

Burkhardt, S, M Borsos, A Szydlowska, J Godwin, S A Williams, P E Cohen, T Hirota, M Saitou, and K Tachibana-Konwalski. 2016. “Chromosome Cohesion Established by Rec8-Cohesin in Fetal Oocytes Is Maintained without Detectable Turnover in Oocytes Arrested for Months in Mice.” Current Biology 26 (5): 678–85. https://doi.org/10.1016/j.cub.2015.12.073.

Callaghan, Michelle J, Amanda J Russell, Erica Woollatt, Grant R Sutherland, Robert L Sutherland, and Colin KW Watts. 1998. “Identification of a Human HECT Family Protein with Homology to the Drosophila Tumor Suppressor Gene Hyperplastic Discs.” Oncogene 17 (26): 3479–91. https://doi.org/10.1038/sj.onc.1202249.

Çamdere, Gamze Ö, Kristian K Carlborg, and Douglas Koshland. 2018. “Intermediate Step of Cohesin’s ATPase Cycle Allows Cohesin to Entrap DNA.” Proceedings of the National Academy of Sciences of the United States of America 115 (39): 9732–37. https://doi.org/10.1073/pnas.1807213115.

Canudas, Silvia, and Susan Smith. 2009. “Differential Regulation of Telomere and Centromere Cohesion by the Scc3 Homologues SA1 and SA2, Respectively, in Human Cells.” Journal of Cell Biology 187 (2): 165–73. https://doi.org/10.1083/jcb.200903096.

Carney, Ellen F. 2017. “Epigenetics: Disturbed Imprinting in Pre-Eclampsia.” Nature Reviews Nephrology. https://doi.org/10.1038/nrneph.2017.141.

Carretero, María, Miguel Ruiz-Torres, Miriam Rodríguez-Corsino, Isabel Barthelemy, and Ana Losada. 2013. “Pds5B Is Required for Cohesion Establishment and Aurora B Accumulation at Centromeres.” EMBO Journal 32 (22): 2938–49.

311

https://doi.org/10.1038/emboj.2013.230.

Celebi, Catherine, Aafke van Montfoort, Valérie Skory, Emmanuelle Kieffer, Sandra Kuntz, Manuel Mark, and Stéphane Viville. 2012. “Tex 19 Paralogs Exhibit a Gonad and Placenta-Specific Expression in the Mouse.” Journal of Reproduction and Development 58 (3): 360–65. https://doi.org/10.1262/jrd.11-047K.

Cha-Molstad, Hyunjoo, Ki Sa Sung, Joonsung Hwang, Kyoung A. Kim, Ji Eun Yu, Young Dong Yoo, Jun Min Jang, et al. 2015. “Amino-Terminal Arginylation Targets Endoplasmic Reticulum Chaperone BiP for Autophagy through P62 Binding.” Nature Cell Biology 17 (7): 917–29. https://doi.org/10.1038/ncb3177.

Chan, C. C.W., V. W.S. Liu, E. Y.L. Lau, W. S.B. Yeung, E. H.Y. Ng, and P C Ho. 2006. “Mitochondrial DNA Content and 4977 Bp Deletion in Unfertilized Oocytes.” Molecular Human Reproduction 11 (12): 843–46. https://doi.org/10.1093/molehr/gah243.

Chan, Kok-Lung, Thomas Gligoris, William Upcher, Yuki Kato, Katsuhiko Shirahige, Kim Nasmyth, and Frédéric Beckouët. 2013. “Pds5 Promotes and Protects Cohesin Acetylation.” Proceedings of the National Academy of Sciences of the United States of America 110 (32): 13020–25. https://doi.org/10.1073/pnas.1306900110.

Chao, Wendy, and Patricia A D’Amore. 2008. “IGF2: Epigenetic Regulation and Role in Development and Disease.” Cytokine and Growth Factor Reviews. NIH Public Access. https://doi.org/10.1016/j.cytogfr.2008.01.005.

Chapard, Christophe, Robert Jones, Till van Oepen, Johanna C. Scheinost, and Kim Nasmyth. 2019. “Sister DNA Entrapment between Juxtaposed Smc Heads and Kleisin of the Cohesin Complex.” Molecular Cell 75 (2): 224-237.e5. https://doi.org/10.1016/j.molcel.2019.05.023.

Chen, Guo, and Xingming Deng. 2018. “Cell Synchronization by Double Thymidine Block.” BIO-PROTOCOL 8 (17). https://doi.org/10.21769/bioprotoc.2994.

Cheng, Chun-Ting, Ching-Ying Kuo, and David K Ann. 2014. “KAPtain in Charge of Multiple Missions: Emerging Roles of KAP1.” World Journal of Biological Chemistry 5 (3): 308. https://doi.org/10.4331/wjbc.v5.i3.308.

Chiang, Teresa, Francesca E. Duncan, Karen Schindler, Richard M. Schultz, and Michael A. Lampson. 2010. “Evidence That Weakened Centromere Cohesion Is a Leading Cause of Age-Related Aneuploidy in Oocytes.” Current Biology 20 (17): 1522–28.

312

https://doi.org/10.1016/j.cub.2010.06.069.

Chiang, Teresa, Richard M Schultz, and Michael A Lampson. 2011. “Age-Dependent Susceptibility of Chromosome Cohesion to Premature Separase Activation in Mouse Oocytes.” Biology of Reproduction 85 (6): 1279–83. https://doi.org/10.1095/biolreprod.111.094094.

Christensen, Devin E, Peter S Brzovic, and Rachel E Klevit. 2007. “E2-BRCA1 RING Interactions Dictate Synthesis of Mono- or Specific Polyubiquitin Chain Linkages.” Nature Structural and Molecular Biology 14 (10): 941–48. https://doi.org/10.1038/nsmb1295.

Chui, Ashley J., Marian C. Okondo, Sahana D. Rao, Kuo Gai, Andrew R. Griswold, Darren C. Johnson, Daniel P. Ball, et al. 2019. “N-Terminal Degradation Activates the NLRP1B Inflammasome.” Science 364 (6435): 82–85. https://doi.org/10.1126/science.aau1208.

Chuma, Shinichiro, and Norio Nakatsuji. 2001. “Autonomous Transition into Meiosis of Mouse Fetal Germ Cells in Vitro and Its Inhibition by Gp130-Mediated Signaling.” Developmental Biology 229 (2): 468–79. https://doi.org/10.1006/dbio.2000.9989.

Cimadomo, Danilo, Gemma Fabozzi, Alberto Vaiarelli, Nicolò Ubaldi, Filippo Maria Ubaldi, and Laura Rienzi. 2018. “Impact of Maternal Age on Oocyte and Embryo Competence.” Frontiers in Endocrinology 9 (JUL). https://doi.org/10.3389/fendo.2018.00327.

Clift, Dean, Farid Bizzari, and Adele L Marston. 2009. “Shugoshin Prevents Eohesin Cleavage by PP2ACdc55-Dependent Inhibition of Separase.” Genes and Development 23 (6): 766–90. https://doi.org/10.1101/gad.507509.

Coan, P.M. M, N. Conroy, G.J. J Burton, and A.C. Ferguson-Smith. 2006. “Origin and Characteristics of Glycogen Cells in the Developing Murine Placenta.” Developmental Dynamics 235 (December): 3280–94. https://doi.org/10.1002/dvdy.20981.

Coan, Philip M., Anne C. Ferguson-Smith, and Graham J. Burton. 2004. “Developmental Dynamics of the Definitive Mouse Placenta Assessed by Stereology1.” Biology of Reproduction 70 (6): 1806–13. https://doi.org/10.1095/biolreprod.103.024166.

Cohen, Dena E., and Jeannie T. Lee. 2002. “X-Chromosome Inactivation and the Search for Chromosome-Wide Silencers.” Current Opinion in Genetics and Development. https://doi.org/10.1016/S0959-437X(02)00289-7.

Cole, Francesca, Scott Keeney, and Maria Jasin. 2012. “Preaching about the Converted: How

313

Meiotic Gene Conversion Influences Genomic Diversity.” Annals of the New York Academy of Sciences 1267 (1): 95–102. https://doi.org/10.1111/j.1749-6632.2012.06595.x.

Cong, Le, F Ann Ran, David Cox, Shuailiang Lin, Robert Barretto, Naomi Habib, Patrick D Hsu, et al. 2013. “Multiplex Genome Engineering Using CRISPR/Cas Systems.” Science 339 (6121): 819–23. https://doi.org/10.1126/science.1231143.

Cortone, Giuseppe, Ge Zheng, Pasquale Pensieri, Viviana Chiappetta, Rosarita Tatè, Eva Malacaria, Pietro Pichierri, Hongtao Yu, and Francesca M Pisani. 2018. “Interaction of the Warsaw Breakage Syndrome DNA Helicase DDX11 with the Replication Fork- Protection Factor Timeless Promotes Sister Chromatid Cohesion.” Edited by Lorraine S. Symington. PLoS Genetics 14 (10): e1007622. https://doi.org/10.1371/journal.pgen.1007622.

Costa, Mariana A. 2016. “The Endocrine Function of Human Placenta: An Overview.” Reproductive BioMedicine Online 32 (1): 14–43. https://doi.org/10.1016/j.rbmo.2015.10.005.

Costa, Y, R Speed, R Ollinger, M Alsheimer, C A Semple, P Gautier, K Maratou, et al. 2005. “Two Novel Proteins Recruited by Synaptonemal Complex Protein 1 (SYCP1) Are at the Centre of Meiosis.” Journal of Cell Science 118 (Pt 12): 2755–62. https://doi.org/10.1242/jcs.02402.

Covo, Shay, James W Westmoreland, Dmitry A Gordenin, and Michael A Resnick. 2010. “Cohesin Is Limiting for the Suppression of DNA Damage-Induced Recombination between Homologous Chromosomes.” Edited by James E. Haber. PLoS Genetics 6 (7): 1– 16. https://doi.org/10.1371/journal.pgen.1001006.

Coyaud, Etienne, Monika Mis, Estelle M.N. Laurent, Wade H. Dunham, Amber L. Couzens, Melanie Robitaille, Anne Claude Gingras, Stephane Angers, and Brian Raught. 2015. “BioID-Based Identification of Skp Cullin F-Box (SCF)β-TrCP1/2 E3 Ligase Substrates.” Molecular and Cellular Proteomics 14 (7): 1781–95. https://doi.org/10.1074/mcp.M114.045658.

Cranston, Aaron, and Richard Fishel. 1999. “Female Embryonic Lethality in Msh2-Trp53 Nullizygous Mice Is Strain Dependent.” Mammalian Genome 10 (10): 1020–22. https://doi.org/10.1007/s003359901151.

Crichton, James H., Donncha S. Dunican, Marie MacLennan, Richard R. Meehan, and Ian R. Adams. 2014. “Defending the Genome from the Enemy within: Mechanisms of Retrotransposon Suppression in the Mouse Germline.” Cellular and Molecular Life Sciences. https://doi.org/10.1007/s00018-013-1468-0.

314

Crichton, James H, Christopher J Playfoot, Marie Maclennan, David Read, J Cooke, and Ian R Adams. 2017. “Tex19.1 Promotes Spo11-Dependent Meiotic Recombination in Mouse Spermatocytes.” PLoS Genetics 13 (7): e1006904.

Cross, J. C., M Hemberger, Y Lu, T Nozaki, K Whiteley, M Masutani, and S. L. Adamson. 2002. “Trophoblast Functions, Angiogenesis and Remodeling of the Maternal Vasculature in the Placenta.” In Molecular and Cellular Endocrinology, 187:207–12. Elsevier. https://doi.org/10.1016/S0303-7207(01)00703-1.

Cross, James C. 2000. “Genetic Insights into Trophoblast Differentiation and Placental Morphogenesis.” Seminars in Cell and Developmental Biology. https://doi.org/10.1006/scdb.2000.0156.

———. 2014. “More of a Good Thing or Less of a Bad Thing: Gene Copy Number Variation in Polyploid Cells of the Placenta.” Edited by Marisa S. Bartolomei. PLoS Genetics 10 (5): e1004330. https://doi.org/10.1371/journal.pgen.1004330.

Dackor, Jennifer, Kathleen M. Caron, and David W. Threadgill. 2009. “Placental and Embryonic Growth Restriction in Mice with Reduced Function Epidermal Growth Factor Receptor Alleles.” Genetics 183 (1): 207–18. https://doi.org/10.1534/genetics.109.104372.

Dasgupta, Tanushree, Jisha Antony, Antony W Braithwaite, and Julia A Horsfield. 2016. “HDAC8 Inhibition Blocks SMC3 Deacetylation and Delays Cell Cycle Progression without Affecting Cohesin-Dependent Transcription in MCF7 Cancer Cells.” Journal of Biological Chemistry 291 (24): 12761–70. https://doi.org/10.1074/jbc.M115.704627.

Daughtry, Brittany L, and Shawn L Chavez. 2016. “Chromosomal Instability in Mammalian Pre-Implantation Embryos: Potential Causes, Detection Methods, and Clinical Consequences.” Cell and Tissue Research. NIH Public Access. https://doi.org/10.1007/s00441-015-2305-6.

Davis, Carrie A, Benjamin C Hitz, Cricket A Sloan, Esther T Chan, Jean M Davidson, Idan Gabdank, Jason A Hilton, et al. 2018. “The Encyclopedia of DNA Elements (ENCODE): Data Portal Update.” Nucleic Acids Research 46 (D1): D794–801. https://doi.org/10.1093/nar/gkx1081.

Deardorff, Matthew A., Jonathan J. Wilde, Melanie Albrecht, Emma Dickinson, Stephanie Tennstedt, Diana Braunholz, Maren Mönnich, et al. 2012. “RAD21 Mutations Cause a

315

Human Cohesinopathy.” American Journal of Human Genetics 90 (6): 1014–27. https://doi.org/10.1016/j.ajhg.2012.04.019.

Deshaies, Raymond J., and Claudio A.P. Joazeiro. 2009. “RING Domain E3 Ubiquitin Ligases.” Annual Review of Biochemistry 78 (1): 399–434. https://doi.org/10.1146/annurev.biochem.78.101807.093809.

Dietrich, Axel J. J., and Robert J. P. Mulder. 1983. “A Light- and Electron Microscopic Analysis of Meiotic Prophase in Female Mice.” Chromosoma 88 (5): 377–85. https://doi.org/10.1007/BF00285860.

Diez-Juan, Antonio, Carmen Rubio, Carlos Marin, Sebastian Martinez, Nasser Al-Asmar, Marcia Riboldi, Patricia Díaz-Gimeno, Diana Valbuena, and Carlos Simón. 2015. “Mitochondrial DNA Content as a Viability Score in Human Euploid Embryos: Less Is Better.” Fertility and Sterility 104 (3): 534-541.e1. https://doi.org/10.1016/j.fertnstert.2015.05.022.

Doerr, Allison. 2018. “Proximity Labeling with TurboID.” Nature Methods 15 (10): 764. https://doi.org/10.1038/s41592-018-0158-0.

Du, Fangyong, Federico Navarro-Garcia, Zanxian Xia, Takafumi Tasaki, and Alexander Varshavsky. 2002. “Pairs of Dipeptides Synergistically Activate the Binding of Substrate by Ubiquitin Ligase through Dissociation of Its Autoinhibitory Domain.” Proceedings of the National Academy of Sciences of the United States of America 99 (22): 14110–15. https://doi.org/10.1073/pnas.172527399.

Duncan, F E, T Chiang, R M Schultz, and M A Lampson. 2009. “Evidence That a Defective Spindle Assembly Checkpoint Is Not the Primary Cause of Maternal Age-Associated Aneuploidy in Mouse Eggs.” Biology of Reproduction 81 (4): 768–76. https://doi.org/10.1095/biolreprod.109.077909.

Dunne, Orla M., and Owen R. Davies. 2019. “Molecular Structure of Human Synaptonemal Complex Protein SYCE1.” Chromosoma, January, 1–14. https://doi.org/10.1007/s00412-018- 00688-z.

Dupont, Celine, Martine Bucourt, Fabien Guimiot, Lilia Kraoua, Daniel Smiljkovski, Dominique Le Tessier, Camille Lebugle, et al. 2014. “3D-FISH Analysis Reveals Chromatid Cohesion Defect during Interphase in Roberts Syndrome.” Molecular Cytogenetics 7 (1): 59. https://doi.org/10.1186/s13039-014-0059-6.

316

Eisele, Frederik, and Dieter H Wolf. 2008. “Degradation of Misfolded Protein in the Cytoplasm Is Mediated by the Ubiquitin Ligase Ubr1.” FEBS Letters 582 (30): 4143–46. https://doi.org/10.1016/j.febslet.2008.11.015.

Enguita-Marruedo, Andrea, Wiggert A. Van Cappellen, Jos W. Hoogerbrugge, Fabrizia Carofiglio, Evelyne Wassenaar, Johan A. Slotman, Adriaan Houtsmuller, and Willy M. Baarends. 2018. “Live Cell Analyses of Synaptonemal Complex Dynamics and Chromosome Movements in Cultured Mouse Testis Tubules and Embryonic Ovaries.” Chromosoma 127 (3): 341–59. https://doi.org/10.1007/s00412-018-0668-7.

Eppig, John J., and Karen Wigglesworth. 2000. “Development of Mouse and Rat Oocytes in Chimeric Reaggregated Ovaries after Interspecific Exchange of Somatic and Germ Cell Components1.” Biology of Reproduction 63 (4): 1014–23. https://doi.org/10.1095/biolreprod63.4.1014.

Ercan, Sevinç. 2014. “Mechanisms of X Chromosome Dosage Compensation.” Journal of Genomics 3: 1–19. https://doi.org/10.7150/jgen.10404.

Fagerberg, Linn, Björn M. Hallström, Per Oksvold, Caroline Kampf, Dijana Djureinovic, Jacob Odeberg, Masato Habuka, et al. 2014. “Analysis of the Human Tissue-Specific Expression by Genome-Wide Integration of Transcriptomics and Antibody-Based Proteomics.” Molecular and Cellular Proteomics 13 (2): 397–406. https://doi.org/10.1074/mcp.m113.035600.

Falque, M, R Mercier, C Mézard, D de Vienne, and O C Martin. 2007. “Patterns of Recombination and MLH1 Foci Density along Mouse Chromosomes: Modeling Effects of Interference and Obligate Chiasma.” Genetics 176 (3): 1453–67. https://doi.org/10.1534/genetics.106.070235.

Faridi, R, A U Rehman, R J Morell, P L Friedman, L Demain, S Zahra, A A Khan, et al. 2017. “Mutations of SGO2 and CLDN14 Collectively Cause Coincidental Perrault Syndrome.” Clinical Genetics 91 (2): 328–32. https://doi.org/10.1111/cge.12867.

Farrow, A., M.G.R. Hull, K. Northstone, H. Taylor, W.C.L. Ford, and Jean Golding. 2002. “Prolonged Use of Oral Contraception before a Planned Pregnancy Is Associated with a Decreased Risk of Delayed Conception.” Human Reproduction 17 (10): 2754–61. https://doi.org/10.1093/humrep/17.10.2754.

Feng, Q, J V Moran, H H Kazazian, and J D Boeke. 1996. “Human L1 Retrotransposon Encodes

317

a Conserved Endonuclease Required for Retrotransposition.” Cell 87 (5): 905–16. https://doi.org/10.1016/s0092-8674(00)81997-2.

Finley, Daniel, Helle D Ulrich, Thomas Sommer, and Peter Kaiser. 2012. “The Ubiquitin- Proteasome System of Saccharomyces Cerevisiae.” Genetics 192 (2): 319–60. https://doi.org/10.1534/genetics.112.140467.

Fischer, Andrew H, Kenneth A Jacobson, Jack Rose, and Rolf Zeller. 2008. “Hematoxylin and Eosin Staining of Tissueand Cell Sections.” Cold Spring Harbor Protocols 3 (5): pdb.prot4986. https://doi.org/10.1101/pdb.prot4986.

Ford, J. H., and L. MacCormac. 1995. “Pregnancy and Lifestyle Study: The Long-Term Use of the Contraceptive Pill and the Risk of Age-Related Miscarriage.” Human Reproduction 10 (6): 1397–1402. https://doi.org/10.1093/HUMREP/10.6.1397.

Foss, E, R Lande, F W Stahl, and C M Steinberg. 1993. “Chiasma Interference as a Function of Genetic Distance.” Genetics 133 (3): 681–91. http://www.ncbi.nlm.nih.gov/pubmed/8454209.

Fragouli, E, V Bianchi, P Patrizio, A Obradors, Z Huang, A Borini, J D Delhanty, and D Wells. 2010. “Transcriptomic Profiling of Human Oocytes: Association of Meiotic Aneuploidy and Altered Oocyte Gene Expression.” Molecular Human Reproduction 16 (8): 570–82. https://doi.org/10.1093/molehr/gaq033.

Fragouli, Elpida, Katharina Spath, Samer Alfarawati, Fiona Kaper, Andrew Craig, Claude- Edouard Michel, Felix Kokocinski, Jacques Cohen, Santiago Munne, and Dagan Wells. 2015. “Altered Levels of Mitochondrial DNA Are Associated with Female Age, Aneuploidy, and Provide an Independent Measure of Embryonic Implantation Potential.” Edited by Stuart K. Kim. PLoS Genetics 11 (6): e1005241. https://doi.org/10.1371/journal.pgen.1005241.

Freemont, Paul S. 1993. “The RING Finger: A Novel Protein Sequence Motif Related to the Zinc Finger.” Annals of the New York Academy of Sciences 684 (1): 174–92. https://doi.org/10.1111/j.1749-6632.1993.tb32280.x.

Freemont, Paul S., Isabel M. Hanson, and John Trowsdale. 1991. “A Novel Gysteine-Rich Sequence Motif.” Cell 64 (3): 483–84. https://doi.org/10.1016/0092-8674(91)90229-R.

Frikstad, Kari-Anne M, Elisa Molinari, Marianne Thoresen, Simon A Ramsbottom, Frances Hughes, Stef J.F. Letteboer, Sania Gilani, et al. 2019. “A CEP104-CSPP1 Complex Is

318

Required for Formation of Primary Cilia Competent in Hedgehog Signaling.” Cell Reports 28 (7): 1907-1922.e6. https://doi.org/10.1016/j.celrep.2019.07.025.

Fritzler, M.J., T.D. Kinsella, and E. Garbutt. 1980. “The CREST Syndrome: A Distinct Serologic Entity with Anticentromere Antibodies.” The American Journal of Medicine 69 (4): 520–26. https://doi.org/10.1016/0002-9343(80)90462-3.

Frost, Jennifer M, and Gudrun E Moore. 2010. “The Importance of Imprinting in the Human Placenta.” PLoS Genetics. Public Library of Science. https://doi.org/10.1371/journal.pgen.1001015.

Fujiwara, Hidenobu, Nobukazu Tanaka, Ichiro Yamashita, and Kenji Kitamura. 2013. “Essential Role of Ubr11, but Not Ubr1, as an N-End Rule Ubiquitin Ligase in Schizosaccharomyces Pombe.” Yeast 30 (1): 1–11. https://doi.org/10.1002/yea.2936.

Fukuda, Tomoyuki, Nanaho Fukuda, Ana Agostinho, Abrahan Hernández‐Hernández, Anna Kouznetsova, and Christer Höög. 2014. “STAG 3‐mediated Stabilization of REC 8 Cohesin Complexes Promotes Chromosome Synapsis during Meiosis.” EMBO Journal 33 (11): 1243–55. https://doi.org/10.1002/embj.201387329.

Gabory, Anne, Laure Ferry, Isabelle Fajardy, Luc Jouneau, Jean David Gothié, Alexandre Vigé, Cécile Fleur, et al. 2012. “Maternal Diets Trigger Sex-Specific Divergent Trajectories of Gene Expression and Epigenetic Systems in Mouse Placenta.” PLoS ONE 7 (11). https://doi.org/10.1371/journal.pone.0047986.

Gabory, Anne, Tessa J Roseboom, Tom Moore, Lorna G Moore, and Claudine Junien. 2013. “Placental Contribution to the Origins of Sexual Dimorphism in Health and Diseases: Sex Chromosomes and Epigenetics.” Biology of Sex Differences. https://doi.org/10.1186/2042-6410-4-5.

Gall, G. A. E., and W. H. Kyle. 1968. “Growth of the Laboratory Mouse.” Theoretical and Applied Genetics 38 (7): 304–8. https://doi.org/10.1007/BF01297571.

Garcia-Cruz, R., M. A. Brieño, I. Roig, M. Grossmann, E. Velilla, A. Pujol, L. Cabero, A. Pessarrodona, J. L. Barbero, and M. Garcia Caldés. 2010. “Dynamics of Cohesin Proteins REC8, STAG3, SMC1β and SMC3 Are Consistent with a Role in Sister Chromatid Cohesion during Meiosis in Human Oocytes.” Human Reproduction 25 (9): 2316–27. https://doi.org/10.1093/humrep/deq180.

Garcia-Perez, Jose L, Thomas J Widmann, and Ian R Adams. 2016. “The Impact of

319

Transposable Elements on Mammalian Development.” Development. Oxford University Press for The Company of Biologists Limited. https://doi.org/10.1242/dev.132639.

Gil, P., E Dewey, J Friml, Y Zhao, K C Snowden, J Putterill, K Palme, M Estelle, and J Chory. 2001. “BIG: A Calossin-like Protein Required for Polar Auxin Transport in Arabidopsis.” Genes and Development 15 (15): 1985–97. https://doi.org/10.1101/gad.905201.

Gómez, R, N Felipe-Medina, M Ruiz-Torres, I Berenguer, A Viera, S Pérez, J L Barbero, et al. 2016. “Sororin Loads to the Synaptonemal Complex Central Region Independently of Meiotic Cohesin Complexes.” EMBO Reports 17 (5): 695–707. https://doi.org/10.15252/embr.201541060.

Gong, Sungsam, Michelle D Johnson, Justyna Dopierala, Francesca Gaccioli, Ulla Sovio, Miguel Constância, Gordon Cs Smith, and D Stephen Charnock-Jones. 2018. “Genome- Wide Oxidative Bisulfite Sequencing Identifies Sex-Specific Methylation Differences in the Human Placenta.” Epigenetics 13 (3): 228–39. https://doi.org/10.1080/15592294.2018.1429857.

Goodier, John L. 2016. “Restricting Retrotransposons: A Review.” Mobile DNA. BioMed Central. https://doi.org/10.1186/s13100-016-0070-z.

Gordillo, Miriam, Hugo Vega, Alison H Trainer, Fajian Hou, Norio Sakai, Ricardo Luque, Hülya Kayserili, et al. 2008. “The Molecular Mechanism Underlying Roberts Syndrome Involves Loss of ESCO2 Acetyltransferase Activity.” Human Molecular Genetics 17 (14): 2172–80. https://doi.org/10.1093/hmg/ddn116.

Gottesman, Susan, Eric Roche, Yan Ning Zhou, and Robert T Sauer. 1998. “The ClpXP and ClpAP Proteases Degrade Proteins with Carboxy-Terminal Peptide Tails Added by the SsrA-Tagging System.” Genes and Development 12 (9): 1338–47. https://doi.org/10.1101/gad.12.9.1338.

Groettrup, Marcus, Christiane Pelzer, Gunter Schmidtke, and Kay Hofmann. 2008. “Activating the Ubiquitin Family: UBA6 Challenges the Field.” Trends in Biochemical Sciences 33 (5): 230–37. https://doi.org/10.1016/j.tibs.2008.01.005.

Hackett, Jamie A, James P Reddington, Colm E Nestor, Donncha S Dunican, Miguel R Branco, Judith Reichmann, Wolf Reik, M Azim Surani, Ian R Adams, and Richard R Meehan. 2012. “Promoter DNA Methylation Couples Genome-Defence Mechanisms to Epigenetic Reprogramming in the Mouse Germline.” Development 139 (19): 3623–32.

320

https://doi.org/10.1242/dev.081661.

Haering, Christian H., Jan Löwe, Andreas Hochwagen, and Kim Nasmyth. 2002. “Molecular Architecture of SMC Proteins and the Yeast Cohesin Complex.” Molecular Cell 9 (4): 773– 88. https://doi.org/10.1016/S1097-2765(02)00515-4.

Haglund, Kaisa, Sara Sigismund, Simona Polo, Iwona Szymkiewicz, Pier Paolo Di Fiore, and Ivan Dikic. 2003. “Multiple Monoubiquitination of RTKs Is Sufficient for Their Endocytosis and Degradation.” Nature Cell Biology 5 (5): 461–66. https://doi.org/10.1038/ncb983.

Haig, David. 1999. “Multiple Paternity and Genomic Imprinting.” Genetics 151 (3).

Hamazaki, Jun, Shoshiro Hirayama, and Shigeo Murata. 2015. “Redundant Roles of Rpn10 and Rpn13 in Recognition of Ubiquitinated Proteins and Cellular Homeostasis.” Edited by Alfred L. Goldberg. PLoS Genetics 11 (7): e1005401. https://doi.org/10.1371/journal.pgen.1005401.

Hannula, Katariina, Marita Lipsanen-Nyman, Tero Kontiokari, and Juha Kere. 2001. “A Narrow Segment of Maternal Uniparental Disomy of Chromosome 7q31-Qter in Silver- Russell Syndrome Delimits a Candidate Gene Region.” American Journal of Human Genetics 68 (1): 247–53. https://doi.org/10.1086/316937.

Hansen, Anders S, Iryna Pustova, Claudia Cattoglio, Robert Tjian, and Xavier Darzacq. 2017. “CTCF and Cohesin Regulate Chromatin Loop Stability with Distinct Dynamics.” ELife 6 (May). https://doi.org/10.7554/eLife.25776.

Hara, Kodai, Ge Zheng, Qianhui Qu, Hong Liu, Zhuqing Ouyang, Zhe Chen, Diana R Tomchick, and Hongtao Yu. 2014. “Structure of Cohesin Subcomplex Pinpoints Direct Shugoshin-Wapl Antagonism in Centromeric Cohesion.” Nature Structural and Molecular Biology 21 (10): 864–70. https://doi.org/10.1038/nsmb.2880.

Harris, Lynda K, Ian P Crocker, Philip N Baker, John D Aplin, and Melissa Westwood. 2011. “IGF2 Actions on Trophoblast in Human Placenta Are Regulated by the Insulin-Like Growth Factor 2 Receptor, Which Can Function as Both a Signaling and Clearance Receptor1.” Biology of Reproduction 84 (3): 440–46. https://doi.org/10.1095/biolreprod.110.088195.

Hartl, F. Ulrich, and Manajit Hayer-Hartl. 2002. “Protein Folding. Molecular Chaperones in the Cytosol: From Nascent Chain to Folded Protein.” Science. American Association for

321

the Advancement of Science. https://doi.org/10.1126/science.1068408.

Hassold, T, and P Hunt. 2001. “To Err (Meiotically) Is Human: The Genesis of Human Aneuploidy.” Nature Reviews Genetics 2 (4): 280–91. https://doi.org/10.1038/35066065.

He, Mu, Stephanie Agbu, and Kathryn V. Anderson. 2017. “Microtubule Motors Drive Hedgehog Signaling in Primary Cilia.” Trends in Cell Biology 27 (2): 110–25. https://doi.org/10.1016/j.tcb.2016.09.010.

Hearn, Timothy J, Maria C Marti Ruiz, S M Abdul-Awal, Rinukshi Wimalasekera, Camilla R Stanton, Michael J Haydon, Frederica L Theodoulou, Matthew A Hannah, and Alex A.R. Webb. 2018. “BIG Regulates Dynamic Adjustment of Circadian Period in Arabidopsis Thaliana.” Plant Physiology 178 (1): 358–71. https://doi.org/10.1104/pp.18.00571.

Heck, Jarrod W, Samantha K Cheung, and Randolph Y Hampton. 2010. “Cytoplasmic Protein Quality Control Degradation Mediated by Parallel Actions of the E3 Ubiquitin Ligases Ubr1 and San1.” Proceedings of the National Academy of Sciences of the United States of America 107 (3): 1106–11. https://doi.org/10.1073/pnas.0910591107.

Henderson, Michelle J, Amanda J Russell, Samantha Hird, Marcia Muñoz, Jennifer L Clancy, Gillian M Lehrbach, Sophina T Calanni, David A Jans, Robert L Sutherland, and Colin K.W. Watts. 2002. “EDD, the Human Hyperplastic Discs Protein, Has a Role in Progesterone Receptor Coactivation and Potential Involvement in DNA Damage Response.” Journal of Biological Chemistry 277 (29): 26468–78. https://doi.org/10.1074/jbc.M203527200.

Herbert, Mary, Dimitrios Kalleas, Daniel Cooney, Mahdi Lamb, and Lisa Lister. 2015. “Meiosis and Maternal Aging: Insights from Aneuploid Oocytes and Trisomy Births.” Cold Spring Harb Perspectives on Biology 7 (4): a017970. https://doi.org/10.1101/cshperspect.a017970.

Hernández-Hernández, Abrahan, Sergej Masich, Tomoyuki Fukuda, Anna Kouznetsova, Sara Sandin, Bertil Daneholt, and Christer Höög. 2016. “The Central Element of the Synaptonemal Complex in Mice Is Organized as a Bilayered Junction Structure.” Journal of Cell Science 129 (11): 2239–49. https://doi.org/10.1242/jcs.182477.

Herrán, Yurema, Cristina Gutiérrez-Caballero, Manuel Sánchez-Martín, Teresa Hernández, Alberto Viera, José Luis Barbero, Enrique De Álava, et al. 2011. “The Cohesin Subunit RAD21L Functions in Meiotic Synapsis and Exhibits Sexual Dimorphism in Fertility.” EMBO Journal 30 (15): 3091–3105. https://doi.org/10.1038/emboj.2011.222.

322

Hicke, Linda. 2001. “Protein Regulation by Monoubiquitin.” Nature Reviews Molecular Cell Biology 2 (3): 195–201. https://doi.org/10.1038/35056583.

Hillers, Kenneth J. 2004. “Crossover Interference.” Current Biology 14 (24): R1036–37. https://doi.org/10.1016/J.CUB.2004.11.038.

Hirano, Tatsuya. 2002. “The ABCs of SMC Proteins: Two-Armed ATPases for Chromosome Condensation, Cohesion, and Repair.” Genes and Development. Cold Spring Harbor Laboratory Press. https://doi.org/10.1101/gad.955102.

Hochstrasser, Mark. 2009. “Origin and Function of Ubiquitin-like Proteins.” Nature 458 (7237): 422–29. https://doi.org/10.1038/nature07958.

Hodges, C A, E Revenkova, R Jessberger, T J Hassold, and P A Hunt. 2005. “SMC1beta- Deficient Female Mice Provide Evidence That Cohesins Are a Missing Link in Age- Related Nondisjunction.” Nature Genetics 37 (12): 1351–55. https://doi.org/10.1038/ng1672.

Hoki, Y., R. Ikeda, N. Mise, Y. Sakata, T. Ohhata, H. Sasaki, K. Abe, and T. Sado. 2011. “Incomplete X-Inactivation Initiated by a Hypomorphic Xist Allele in the Mouse.” Development 138 (13): 2649–59. https://doi.org/10.1242/dev.061226.

Holland, Andrew J, Franziska Böttger, Olaf Stemmann, and Stephen S Taylor. 2007. “Protein Phosphatase 2A and Separase Form a Complex Regulated by Separase Autocleavage.” Journal of Biological Chemistry 282 (34): 24623–32. https://doi.org/10.1074/jbc.M702545200.

Honda, Yoshiomi, Masahide Tojo, Kazuhito Matsuzaki, Tadashi Anan, Mitsuhiro Matsumoto, Masayuki Ando, Hideyuki Saya, and Mitsuyoshi Nakao. 2002. “Cooperation of HECT- Domain Ubiquitin Ligase HHYD and DNA Topoisomerase II-Binding Protein for DNA Damage Response.” Journal of Biological Chemistry 277 (5): 3599–3605. https://doi.org/10.1074/jbc.M104347200.

Hons, Michael T., Pim J. Huis Int Veld, Jan Kaesler, Pascaline Rombaut, Alexander Schleiffer, Franz Herzog, Holger Stark, and Jan Michael Peters. 2016. “Topology and Structure of an Engineered Human Cohesin Complex Bound to Pds5B.” Nature Communications 7 (1): 12523. https://doi.org/10.1038/ncomms12523.

Hu, Dong, and James C. Cross. 2010. “Development and Function of Trophoblast Giant Cells in the Rodent Placenta.” International Journal of Developmental Biology 54 (2–3): 341–54. https://doi.org/10.1387/ijdb.082768dh.

323

Hu, Rong-Gui, Jun Sheng, Xin Qi, Zhenming Xu, Terry T. Takahashi, and Alexander Varshavsky. 2005. “The N-End Rule Pathway as a Nitric Oxide Sensor Controlling the Levels of Multiple Regulators.” Nature 437 (7061): 981–86. https://doi.org/10.1038/nature04027.

Huang, C J, Y F Yuan, D Wu, F A Khan, X F Jiao, and L J Huo. 2016. “The Cohesion Stabilizer Sororin Favors DNA Repair and Chromosome Segregation during Mouse Oocyte Meiosis.” In Vitro Cell Developmental Biology of Animals. https://doi.org/10.1007/s11626-016- 0107-0.

Humphryes, Neil, and Andreas Hochwagen. 2014. “A Non-Sister Act: Recombination Template Choice during Meiosis.” Experimental Cell Research. NIH Public Access. https://doi.org/10.1016/j.yexcr.2014.08.024.

Hunt, P A, and T J Hassold. 2002. “Sex Matters in Meiosis.” Science 296 (5576): 2181–83. https://doi.org/10.1126/science.1071907.

Hwang, C.-S., A. Shemorry, and A. Varshavsky. 2010. “N-Terminal Acetylation of Cellular Proteins Creates Specific Degradation Signals.” Science 327 (5968): 973–77. https://doi.org/10.1126/science.1183147.

Iconomou, Mary, and Darren N Saunders. 2016. “Systematic Approaches to Identify E3 Ligase Substrates.” Biochemical Journal. Portland Press Ltd. https://doi.org/10.1042/BCJ20160719.

Inoue, Akira, Tongyuan Li, Sarah K. Roby, Marcus B. Valentine, Madoka Inoue, Kelli Boyd, Vincent J. Kidd, and Jill M. Lahti. 2007. “Loss of ChlR1 Helicase in Mouse Causes Lethality Due to the Accumulation of Aneuploid Cells Generated by Cohesion Defects and Placental Malformation.” Cell Cycle 6 (13): 1646–54. https://doi.org/10.4161/cc.6.13.4411.

Ishiguro, Kei Ichiro, Jihye Kim, Sally Fujiyama-Nakamura, Shigeaki Kato, and Yoshinori Watanabe. 2011. “A New Meiosis-Specific Cohesin Complex Implicated in the Cohesin Code for Homologous Pairing.” EMBO Reports 12 (3): 267–75. https://doi.org/10.1038/embor.2011.2.

Jackson, Matthew P, and Eric W Hewitt. 2016. “Cellular Proteostasis: Degradation of Misfolded Proteins by Lysosomes.” Essays in Biochemistry 60 (2): 173–80. https://doi.org/10.1042/EBC20160005.

Jagiello, G., and J.S. S Fang. 1979. “Analyses of Diplotene Chiasma Frequencies in Mouse

324

Oocytes and Spermatocytes in Relation to Ageing and Sexual Dimorphism.” Cytogenetics and Cell Genetics 23 (1–2): 53–60. https://doi.org/10.1159/000131302.

James, Joanna, Win Tun, and Alys Clark. 2016. “Quantifying Trophoblast Migration: In Vitro Approaches to Address in Vivo Situations.” Cell Adhesion and Migration 10 (1–2): 77–87. https://doi.org/10.1080/19336918.2015.1083667.

Jang, Suk Min, Annamaria Kauzlaric, Jean-Pierre Quivy, Julien Pontis, Benjamin Rauwel, Andrea Coluccio, Sandra Offner, et al. 2018. “KAP1 Facilitates Reinstatement of Heterochromatin after DNA Replication.” Nucleic Acids Research 46 (17): 8788–8802. https://doi.org/10.1093/nar/gky580.

Jessberger, R. 2012. “Age-Related Aneuploidy through Cohesion Exhaustion.” EMBO Reports 13 (6): 539–46. https://doi.org/10.1038/embor.2012.54.

Jiang, Baichun, Wei Zhao, Jupeng Yuan, Yanyan Qian, Wenjie Sun, Yongxin Zou, Chenhong Guo, Bingxi Chen, Changshun Shao, and Yaoqin Gong. 2012. “Lack of Cul4b, an E3 Ubiquitin Ligase Component, Leads to Embryonic Lethality and Abnormal Placental Development.” Edited by Anton Wutz. PLoS ONE 7 (5). https://doi.org/10.1371/journal.pone.0037070.

Jiang, Xiao, Junhao Yan, Yan Sheng, Mei Sun, Linlin Cui, and Zi Jiang Chen. 2018. “Low Anti- Müllerian Hormone Concentration Is Associated with Increased Risk of Embryonic Aneuploidy in Women of Advanced Age.” Reproductive BioMedicine Online 37 (2): 178–83. https://doi.org/10.1016/j.rbmo.2018.04.039.

Jin, Jianping, Xue Li, Steven P. Gygi, and J. Wade Harper. 2007. “Dual E1 Activation Systems for Ubiquitin Differentially Regulate E2 Enzyme Charging.” Nature 447 (7148): 1135–38. https://doi.org/10.1038/nature05902.

Jin, Jiu-Zhen, and Jixiang Ding. 2015. “Strain-Dependent Gene Expression during Mouse Embryonic Palate Development.” Journal of Developmental Biology 3 (1): 2–10. https://doi.org/10.3390/jdb3010002.

Jo, Ukhyun, Winson Cai, Jingming Wang, Yoojin Kwon, Alan D. D’Andrea, and Hyungjin Kim. 2016. “PCNA-Dependent Cleavage and Degradation of SDE2 Regulates Response to Replication Stress.” Edited by Kristijan Ramadan. PLoS Genetics 12 (12): e1006465. https://doi.org/10.1371/journal.pgen.1006465.

Joglekar, Ajit P., and Alexander A. Kukreja. 2017. “How Kinetochore Architecture Shapes

325

the Mechanisms of Its Function.” Current Biology 27 (16): R816–24. https://doi.org/10.1016/j.cub.2017.06.012.

Johnson, Joshua, Jacqueline Canning, Tomoko Kaneko, James K. Pru, and Jonathan L. Tilly. 2004. “Germline Stem Cells and Follicular Renewal in the Postnatal Mammalian Ovary.” Nature 428 (6979): 145–50. https://doi.org/10.1038/nature02316.

Jordan, P W, C Eyster, J Chen, R J Pezza, and S Rankin. 2017. “Sororin Is Enriched at the Central Region of Synapsed Meiotic Chromosomes.” Chromosome Research. https://doi.org/10.1007/s10577-016-9542-8.

Juszkiewicz, Szymon, and Ramanujan S. Hegde. 2018. “Quality Control of Orphaned Proteins.” Molecular Cell 71 (3): 443–57. https://doi.org/10.1016/J.MOLCEL.2018.07.001.

Kaufman, Matthew H., Sheila C. Barton, and M. Azim H. Surani. 1977. “Normal Postimplantation Development of Mouse Parthenogenetic Embryos to the Forelimb Bud Stage.” Nature 265 (5589): 53–55. https://doi.org/10.1038/265053a0.

Kaur, Gurmeet, and Srikrishna Subramanian. 2015. “The UBR-Box and Its Relationship to Binuclear RING-like Treble Clef Zinc Fingers.” Biology Direct 10 (1): 36. https://doi.org/10.1186/s13062-015-0066-5.

Kawai, Ken, Atsuto Moriya, Satoshi Uemura, and Fumiyoshi Abe. 2014. “Functional Implications and Ubiquitin-Dependent Degradation of the Peptide Transporter Ptr2 in Saccharomyces Cerevisiae.” Eukaryotic Cell 13 (11): 1380–92. https://doi.org/10.1128/EC.00094-14.

Kawasumi, Ryotaro, Takuya Abe, Hiroshi Arakawa, Massimiliano Garre, Kouji Hirota, and Dana Branzei. 2017. “ESCO1/2’s Roles in Chromosome Structure and Interphase Chromatin Organization.” Genes and Development 31 (21): 2136–50. https://doi.org/10.1101/gad.306084.117.

Keeney, Scott. 2008. “Spo11 and the Formation of DNA Double-Strand Breaks in Meiosis.” Genome Dynamics and Stability. Berlin, Heidelberg: Springer Berlin Heidelberg. https://doi.org/10.1007/7050_2007_026.

Kerr, J B, R Duckett, M Myers, K L Britt, T Mladenovska, and J K Findlay. 2006. “Quantification of Healthy Follicles in the Neonatal and Adult Mouse Ovary: Evidence for Maintenance of Primordial Follicle Supply.” Reproduction 132 (1): 95–109. https://doi.org/10.1530/rep.1.01128.

326

Kim, Heon Ki, Ryu Ryun Kim, Jang Hyun Oh, Hanna Cho, Alexander Varshavsky, and Cheol Sang Hwang. 2014. “The N-Terminal Methionine of Cellular Proteins as a Degradation Signal.” Cell 156 (1–2): 158–69. https://doi.org/10.1016/j.cell.2013.11.031.

Kim, Hyoung Tae, Kwang Pyo Kim, Fernando Lledias, Alexei F. Kisselev, K. Matthew Scaglione, Dorota Skowyra, Steven P. Gygi, and Alfred L. Goldberg. 2007. “Certain Pairs of Ubiquitin-Conjugating Enzymes (E2s) and Ubiquitin-Protein Ligases (E3s) Synthesize Nondegradable Forked Ubiquitin Chains Containing All Possible Isopeptide Linkages.” Journal of Biological Chemistry 282 (24): 17375–86. https://doi.org/10.1074/jbc.M609659200.

Kim, Jeong Mok, and Cheol Sang Hwang. 2014. “Crosstalk between the Arg/N-End and Ac/N- End Rule.” Cell Cycle, 2014. https://doi.org/10.4161/cc.28751.

Kim, Sung Tae, Yoon Jee Lee, Takafumi Tasaki, Su Ran Mun, Joonsung Hwang, Min Jueng Kang, Srinivasrao Ganipisetti, Eugene C Yi, Bo Yeon Kim, and Yong Tae Kwon. 2018. “The N-Recognin UBR4 of the N-End Rule Pathway Is Targeted to and Required for the Biogenesis of the Early Endosome.” Journal of Cell Science 131 (17): jcs217646. https://doi.org/10.1242/jcs.217646.

Kinsella, Elaine, Natalie Dora, David Mellis, Laura Lettice, Paul Deveney, Robert Hill, and Mark Ditzel. 2016. “Use of a Conditional Ubr5 Mutant Allele to Investigate the Role of an N-End Rule Ubiquitin-Protein Ligase in Hedgehog Signalling and Embryonic Limb Development.” Edited by Michael Schubert. PLoS ONE 11 (6): e0157079. https://doi.org/10.1371/journal.pone.0157079.

Kitajima, T S, S A Kawashima, and Y Watanabe. 2004. “The Conserved Kinetochore Protein Shugoshin Protects Centromeric Cohesion during Meiosis.” Nature 427 (6974): 510–17. https://doi.org/10.1038/nature02312.

Kitajima, T S, T Sakuno, K Ishiguro, S Iemura, T Natsume, S A Kawashima, and Y Watanabe. 2006. “Shugoshin Collaborates with Protein Phosphatase 2A to Protect Cohesin.” Nature 441 (7089): 46–52. https://doi.org/10.1038/nature04663.

Kitamura, Kenji. 2014. “The ClpS-like N-Domain Is Essential for the Functioning of Ubr11, an N-Recognin in Schizosaccharomyces Pombe.” SpringerPlus 3 (1): 257. https://doi.org/10.1186/2193-1801-3-257.

Kitamura, Kenji, and Hidenobu Fujiwara. 2013. “The Type-2 N-End Rule Peptide Recognition Activity of Ubr11 Ubiquitin Ligase Is Required for the Expression of Peptide

327

Transporters.” FEBS Letters 587 (2): 214–19. https://doi.org/10.1016/j.febslet.2012.11.028.

Kline, Antonie D., Joanna F. Moss, Angelo Selicorni, Anne Marie Bisgaard, Matthew A. Deardorff, Peter M. Gillett, Stacey L. Ishman, et al. 2018. “Diagnosis and Management of Cornelia de Lange Syndrome: First International Consensus Statement.” Nature Reviews Genetics. Nature Publishing Group. https://doi.org/10.1038/s41576-018-0031-0.

Koehler, Kara E, Stefanie E Schrump, Jonathan P Cherry, Terry J Hassold, and Patricia A Hunt. 2006. “Near-Human Aneuploidy Levels in Female Mice with Homeologous Chromosomes.” Current Biology. Elsevier. https://doi.org/10.1016/j.cub.2006.07.018.

Koken, Marcel H.M., Paul Reynolds, Iris Jaspers-Dekker, Louise Prakash, Satya Prakash, Dirk Bootsma, and Jan H.J. Hoeijmakers. 1991. “Structural and Functional Conservation of Two Human Homologs of the Yeast DNA Repair Gene RAD6.” Proceedings of the National Academy of Sciences of the United States of America 88 (20): 8865–69. https://doi.org/10.1073/pnas.88.20.8865.

Kolano, Agnieszka, Stéphane Brunet, Alain D. Silk, Don W. Cleveland, and Marie Heĺène Verlhac. 2012. “Error-Prone Mammalian Female Meiosis from Silencing the Spindle Assembly Checkpoint without Normal Interkinetochore Tension.” Proceedings of the National Academy of Sciences of the United States of America 109 (27): E1858–67. https://doi.org/10.1073/pnas.1204686109.

Kong, Xiangduo, Alexander R Ball, Hoang Xuan Pham, Weihua Zeng, H.-Y. Chen, John A Schmiesing, J.-S. Kim, Michael Berns, and Kyoko Yokomori. 2014. “Distinct Functions of Human Cohesin-SA1 and Cohesin-SA2 in Double-Strand Break Repair.” Molecular and Cellular Biology 34 (4): 685–98. https://doi.org/10.1128/mcb.01503-13.

Konstantinidis, Michalis, Samer Alfarawati, Douglas Hurd, Marta Paolucci, John Shovelton, Elpida Fragouli, and Dagan Wells. 2014. “Simultaneous Assessment of Aneuploidy, Polymorphisms, and Mitochondrial DNA Content in Human Polar Bodies and Embryos with the Use of a Novel Microarray Platform.” Fertility and Sterility 102 (5): 1385–92. https://doi.org/10.1016/j.fertnstert.2014.07.1233.

Kostova, Zlatka, Jennifer Mariano, Simone Scholz, Carolin Koenig, and Allan M. Weissman. 2009. “A Ubc7p-Binding Domain in Cue1p Activates ER-Associated Protein Degradation.” Journal of Cell Science 122 (9): 1374–81. https://doi.org/10.1242/jcs.044255.

Kosumi, Keisuke, Yoshifumi Baba, Kazuo Okadome, Taisuke Yagi, Yuki Kiyozumi, Naoya

328

Yoshida, Masayuki Watanabe, and Hideo Baba. 2019. “Tumor Long-Interspersed Nucleotide Element-1 Methylation Level and Immune Response to Esophageal Cancer.” Annals of Surgery, March, 1. https://doi.org/10.1097/sla.0000000000003264.

Koukoura, O., S. Sifakis, A. Zaravinos, S. Apostolidou, A. Jones, J. Hajiioannou, M. Widschwendter, and D.A. Spandidos. 2011. “Hypomethylation along with Increased H19 Expression in Placentas from Pregnancies Complicated with Fetal Growth Restriction.” Placenta 32 (1): 51–57. https://doi.org/10.1016/j.placenta.2010.10.017.

Kragelund, Birthe B, Signe M Schenstrøm, Caio A Rebula, Vikram Govind Panse, and Rasmus Hartmann-Petersen. 2016. “DSS1/Sem1, a Multifunctional and Intrinsically Disordered Protein.” Trends in Biochemical Sciences. Elsevier. https://doi.org/10.1016/j.tibs.2016.02.004.

Kriegenburg, Franziska, Visnja Jakopec, Esben G. Poulsen, Sofie Vincents Nielsen, Assen Roguev, Nevan Krogan, Colin Gordon, Ursula Fleig, and Rasmus Hartmann-Petersen. 2014. “A Chaperone-Assisted Degradation Pathway Targets Kinetochore Proteins to Ensure Genome Stability.” Edited by Jeffrey L. Brodsky. PLoS Genetics 10 (1): e1004140. https://doi.org/10.1371/journal.pgen.1004140.

Kwon, Do Hoon, Ok Hyun Park, Leehyeon Kim, Yang Ouk Jung, Yeonkyoung Park, Hyeongseop Jeong, Jaekyung Hyun, Yoon Ki Kim, and Hyun Kyu Song. 2018. “Insights into Degradation Mechanism of N-End Rule Substrates by P62/SQSTM1 Autophagy Adapter.” Nature Communications 9 (1): 3291. https://doi.org/10.1038/s41467-018-05825-x.

Kwon, Ja Young, Paulomi Aldo, Yuan You, Jiahui Ding, Karen Racicot, Xiaoyan Dong, John Murphy, et al. 2018. “Relevance of Placental Type I Interferon Beta Regulation for Pregnancy Success.” Cellular and Molecular Immunology 15 (12): 1010–26. https://doi.org/10.1038/s41423-018-0050-y.

Kwon, Y. T., Z. Xia, I. V. Davydov, S. H. Lecker, and A. Varshavsky. 2001. “Construction and Analysis of Mouse Strains Lacking the Ubiquitin Ligase UBR1 (E3alpha ) of the N-End Rule Pathway.” Molecular and Cellular Biology 21 (23): 8007–21. https://doi.org/10.1128/MCB.21.23.8007-8021.2001.

Kwon, Y T, Z Xia, J Y An, T Tasaki, I V Davydov, J W Seo, J Sheng, Y Xie, and A Varshavsky. 2003. “Female Lethality and Apoptosis of Spermatocytes in Mice Lacking the UBR2 Ubiquitin Ligase of the N-End Rule Pathway.” Molecular and Cellular Biology 23 (22): 8255–71. https://www.ncbi.nlm.nih.gov/pubmed/14585983.

Kwon, Yong Tae, Anna S Kashina, and Alexander Varshavsky. 1999. “Alternative Splicing

329

Results in Differential Expression, Activity, and Localization of the Two Forms of Arginyl-TRNA-Protein Transferase, a Component of the N-End Rule Pathway.” Molecular and Cellular Biology 19 (1): 182–93. https://doi.org/10.1128/mcb.19.1.182.

La Fuente, R De, M M Viveiros, K Wigglesworth, and J J Eppig. 2004. “ATRX, a Member of the SNF2 Family of Helicase/ATPases, Is Required for Chromosome Alignment and Meiotic Spindle Organization in Metaphase II Stage Mouse Oocytes.” Devevelopmental Biology 272 (1): 1–14. https://doi.org/10.1016/j.ydbio.2003.12.012.

Ladurner, R, E Kreidl, M P Ivanov, H Ekker, M H Idarraga-Amado, G A Busslinger, G Wutz, D A Cisneros, and J M Peters. 2016. “Sororin Actively Maintains Sister Chromatid Cohesion.” EMBO Journal 35 (6): 635–53. https://doi.org/10.15252/embj.201592532.

Lander, E S, L M Linton, B Birren, C Nusbaum, M C Zody, J Baldwin, K Devon, et al. 2001. “Initial Sequencing and Analysis of the Human Genome.” Nature 409 (6822): 860–921. https://doi.org/10.1038/35057062.

Larkins, Brian A., Brian P. Dilkes, Ricardo A. Dante, Cintia M. Coelho, Young‐min Woo, and Yan Liu. 2001. “Investigating the Hows and Whys of DNA Endoreduplication.” Journal of Experimental Botany 52 (355): 183–92. https://doi.org/10.1093/jxb/52.355.183.

Latos, Paulina Anna, and Myriam Hemberger. 2016. “From the Stem of the Placental Tree: Trophoblast Stem Cells and Their Progeny.” Development. https://doi.org/10.1242/dev.133462.

Lee, C. W. P., M. E. Sowa, S. P. Gygi, and J. Wade Harper. 2011. “Alternative Ubiquitin Activation/Conjugation Cascades Interact with N-End Rule Ubiquitin Ligases to Control Degradation of RGS Proteins.” Molecular Cell 43 (3): 392–405. https://doi.org/10.1016/j.molcel.2011.05.034.Alternative.

Lee, Chih-Yung S., Tzu-Lan Yeh, Bridget T. Hughes, and Peter J. Espenshade. 2011. “Regulation of the Sre1 Hypoxic Transcription Factor by Oxygen-Dependent Control of DNA Binding.” Molecular Cell 44 (2): 225–34. https://doi.org/10.1016/j.molcel.2011.08.031.

Lee, J, T Iwai, T Yokota, and M Yamashita. 2003. “Temporally and Spatially Selective Loss of Rec8 Protein from Meiotic Chromosomes during Mammalian Meiosis.” Journal of Cell Science 116 (Pt 13): 2781–90. https://doi.org/10.1242/jcs.00495.

Lee, Min Jae, Takafumi Tasaki, Kayoko Moroi, Jee Young An, Sadao Kimura, Ilia V. Davydov, and Yong Tae Kwon. 2005. “RGS4 and RGS5 Are in Vivo Substrates of the N-End Rule

330

Pathway.” PNAS 102 (42): 15030–35. https://doi.org/10.1073/pnas.0507533102.

Lefebvre, Louis, Stéphane Viville, Sheila C. Barton, Fumitoshi Ishino, Eric B. Keverne, and M. Azim Surani. 1998. “Abnormal Maternal Behaviour and Growth Retardation Associated with Loss of the Imprinted Gene Mest.” Nature Genetics 20 (2): 163–69. https://doi.org/10.1038/2464.

Leung, Danny, Tingting Du, Ulrich Wagner, Wei Xie, Ah Young Lee, Preeti Goyal, Yujing Li, et al. 2014. “Regulation of DNA Methylation Turnover at LTR Retrotransposons and Imprinted Loci by the Histone Methyltransferase Setdb1.” Proceedings of the National Academy of Sciences of the United States of America 111 (18): 6690–95. https://doi.org/10.1073/pnas.1322273111.

Li, Ge, Chentong Xu, Xin Lin, Liujing Qu, Dan Xia, Beiqi Hongdu, Yan Xia, et al. 2017. “Deletion of Pdcd5 in Mice Led to the Deficiency of Placenta Development and Embryonic Lethality.” Cell Death & Disease 8 (5): e2811. https://doi.org/10.1038/cddis.2017.124.

Li, W, and Y Ye. 2008. “Polyubiquitin Chains: Functions, Structures, and Mechanisms.” Cellular and Molecular Life Sciences. NIH Public Access. https://doi.org/10.1007/s00018-008- 8090-6.

Li, Wei, Mario H. MH Bengtson, Axel Ulbrich, Akio Matsuda, VA Venkateshwar A. Reddy, Anthony Orth, SK Sumit K. Chanda, Serge Batalov, and CAP Claudio A. P. Joazeiro. 2008. “Genome-Wide and Functional Annotation of Human E3 Ubiquitin Ligases Identifies MULAN, a Mitochondrial E3 That Regulates the Organelle’s Dynamics and Signaling.” Edited by Hidde Ploegh. PLoS ONE 3 (1): e1487. https://doi.org/10.1371/journal.pone.0001487.

Lianga, Noel, Carole Doré, Erin K. Kennedy, Elaine Yeh, Elizabeth C. Williams, Camille Marie Fortinez, Alick Wang, Kerry S. Bloom, and Adam D. Rudner. 2018. “Cdk1 Phosphorylation of Esp1/Separase Functions with PP2A and Slk19 to Regulate Pericentric Cohesin and Anaphase Onset.” Edited by Dean Dawson. PLoS Genetics 14 (3): e1007029. https://doi.org/10.1371/journal.pgen.1007029.

Linzer, Daniel I H, and Susan J Fisher. 1999. “The Placenta and the Prolactin Family of Hormones: Regulation of the Physiology of Pregnancy.” Molecular Endocrinology 13 (6): 837–40. https://academic.oup.com/mend/article-abstract/13/6/837/2741710.

331

Lister, L M, A Kouznetsova, L A Hyslop, D Kalleas, S L Pace, J C Barel, A Nathan, et al. 2010. “Age-Related Meiotic Segregation Errors in Mammalian Oocytes Are Preceded by Depletion of Cohesin and Sgo2.” Current Biology 20 (17): 1511–21. https://doi.org/10.1016/j.cub.2010.08.023.

Liu, Jin, and Ruth Nussinov. 2010. “Molecular Dynamics Reveal the Essential Role of Linker Motions in the Function of Cullin-RING E3 Ligases.” Journal of Molecular Biology 396 (5): 1508–23. https://doi.org/10.1016/j.jmb.2010.01.022.

Liu, L, and D L Keefe. 2008. “Defective Cohesin Is Associated with Age-Dependent Misaligned Chromosomes in Oocytes.” Reproductive BioMedicine Online 16 (1): 103–12. https://www.ncbi.nlm.nih.gov/pubmed/18252055.

Liu, Y J, C Liu, Z Chang, B Wadas, C S Brower, Z H Song, Z L Xu, et al. 2016. “Degradation of the Separase-Cleaved Rec8, a Meiotic Cohesin Subunit, by the N-End Rule Pathway.” Journal of Biological Chemistry 291 (14): 7426–38. https://doi.org/10.1074/jbc.M116.714964.

Looijenga, Leendert H.J., Ad J.M. Gillis, Annemieke J.M.H. Verkerk, Wim L.J. Van Putten, and J Wolter Oosterhuis. 1999. “Heterogeneous X Inactivation in Trophoblastic Cells of Human Full-Term Female Placentas.” American Journal of Human Genetics 64 (5): 1445–52. https://doi.org/10.1086/302382.

Losada, A. 2014. “Cohesin in Cancer: Chromosome Segregation and Beyond.” Nature Reviews Cancer 14 (6): 389–93. https://doi.org/10.1038/nrc3743.

Luo, Shukun, and Liang Tong. 2018. “Structural Biology of the Separase–Securin Complex with Crucial Roles in Chromosome Segregation.” Current Opinion in Structural Biology 49 (April): 114–22. https://doi.org/10.1016/j.sbi.2018.01.012.

MacLennan, M, J H Crichton, C J Playfoot, and I R Adams. 2015. “Oocyte Development, Meiosis and Aneuploidy.” Seminars in Cell and Developmental Biology. https://doi.org/10.1016/j.semcdb.2015.10.005.

Maclennan, Marie, Marta García-Cañadas, Judith Reichmann, Elena Khazina, Gabriele Wagner, Christopher J Playfoot, Carmen Salvador-Palomeque, et al. 2017. “Mobilization of LINE-1 Retrotransposons Is Restricted by Tex19.1 in Mouse Embryonic Stem Cells.” ELife 6 (August). https://doi.org/10.7554/eLife.26152.

Magli, M.C. 2000. “Chromosome Mosaicism in Day 3 Aneuploid Embryos That Develop to Morphologically Normal Blastocysts in Vitro.” Human Reproduction 15 (8): 1781–86.

332

https://doi.org/10.1093/humrep/15.8.1781.

Malassine, A., J.-L L. Frendo, and D. Evain-Brion. 2003. “A Comparison of Placental Development and Endocrine Functions between the Human and Mouse Model.” Human Reproduction Update 9 (6): 531–39. https://doi.org/10.1093/humupd/dmg043.

Malhotra, Atul, Beth J. Allison, Margie Castillo-Melendez, Graham Jenkin, Graeme R. Polglase, and Suzanne L. Miller. 2019. “Neonatal Morbidities of Fetal Growth Restriction: Pathophysiology and Impact.” Frontiers in Endocrinology 10 (February): 55. https://doi.org/10.3389/fendo.2019.00055.

Mantikou, Eleni, Kai Mee Wong, Sjoerd Repping, and Sebastiaan Mastenbroek. 2012. “Molecular Origin of Mitotic Aneuploidies in Preimplantation Embryos.” Biochimica et Biophysica Acta (BBA) - Molecular Basis of Disease 1822 (12): 1921–30. https://doi.org/10.1016/j.bbadis.2012.06.013.

Marques, Carla, Weimin Guo, Paulo Pereira, Allen Taylor, Cam Patterson, Paul C. Evans, and Fu Shang. 2006. “The Triage of Damaged Proteins: Degradation by the Ubiquitin- Proteasome Pathway or Repair by Molecular Chaperones.” FASEB Journal 20 (6): 741–43. https://doi.org/10.1096/fj.05-5080fje.

Martin, Sandra L. 2006. “The ORF1 Protein Encoded by LINE-1: Structure and Function during L1 Retrotransposition.” Journal of Biomedicine & Biotechnology 2006 (1): 45621. https://doi.org/10.1155/JBB/2006/45621.

Matta-Camacho, E, G Kozlov, F F Li, and K Gehring. 2010. “Structural Basis of Substrate Recognition and Specificity in the N-End Rule Pathway.” Nature Structural and Molecular Biology 17 (10): 1182–87. https://doi.org/10.1038/nsmb.1894.

Matta-Camacho, Edna, Guennadi Kozlov, Marie Menade, and Kalle Gehring. 2012. “Structure of the HECT C-Lobe of the UBR5 E3 Ubiquitin Ligase.” Acta Crystallographica Section F: Structural Biology and Crystallization Communications 68 (10): 1158–63. https://doi.org/10.1107/S1744309112036937.

Matthews, T J, and B E Hamilton. 2009. “Delayed Childbearing: More Women Are Having Their First Child Later in Life.” NCHS Data Brief, no. 21: 1–8. https://www.ncbi.nlm.nih.gov/pubmed/19674536.

Mavragani, Clio P., Irina Sagalovskiy, Qiu Guo, Adrianos Nezos, Efstathia K. Kapsogeorgou, Pin Lu, Jun Liang Zhou, et al. 2016. “Expression of Long Interspersed Nuclear Element 1

333

Retroelements and Induction of Type I Interferon in Patients With Systemic Autoimmune Disease.” Arthritis & Rheumatology 68 (11): 2686–96. https://doi.org/10.1002/art.39795.

Mayer, Wolfgang, Myriam Hemberger, Hans-Georg Frank, Ruth Grummer, Elke Winterhager, Peter Kaufmann, and Reinald Fundele. 2000. “Expression of the Imprinted GenesMEST/Mest in Human and Murine Placenta Suggests a Role in Angiogenesis.” Developmental Dynamics 217 (1): 1–10. https://doi.org/10.1002/(SICI)1097- 0177(200001)217:1<1::AID-DVDY1>3.0.CO;2-4.

McClintock, B. 1950. “The Origin and Behavior of Mutable Loci in Maize.” Proceedings of the National Academy of Sciences of the United States of America 36 (6): 344–55. https://doi.org/10.1073/pnas.36.6.344.

———. 1956. “Controlling Elements and the Gene.” Cold Spring Harbor Symposia on Quantitative Biology 21: 197–216. https://doi.org/10.1101/SQB.1956.021.01.017.

McCoy, Rajiv C, Zachary P Demko, Allison Ryan, Milena Banjevic, Matthew Hill, Styrmir Sigurjonsson, Matthew Rabinowitz, and Dmitri A Petrov. 2015. “Evidence of Selection against Complex Mitotic-Origin Aneuploidy during Preimplantation Development.” PLoS Genetics 11 (10): e1005601. https://doi.org/10.1371/journal.pgen.1005601.

McGraw, Serge, Christopher C Oakes, Josée Martel, M Cecilia Cirio, Pauline de Zeeuw, Winifred Mak, Christoph Plass, Marisa S Bartolomei, J Richard Chaillet, and Jacquetta M Trasler. 2013. “Loss of DNMT1o Disrupts Imprinted X Chromosome Inactivation and Accentuates Placental Defects in Females.” PLoS Genetics 9 (11): e1003873. https://doi.org/10.1371/journal.pgen.1003873.

McMinn, J., M. Wei, N. Schupf, J. Cusmai, E. B. Johnson, A. C. Smith, R. Weksberg, H. M. Thaker, and B. Tycko. 2006. “Unbalanced Placental Expression of Imprinted Genes in Human Intrauterine Growth Restriction.” Placenta 27 (6–7): 540–49. https://doi.org/10.1016/j.placenta.2005.07.004.

Meacham, Geoffrey C., Cam Patterson, Wenyue Zhang, J. Michael Younger, and Douglas M. Cyr. 2001. “The Hsc70 Co-Chaperone CHIP Targets Immature CFTR for Proteasomal Degradation.” Nature Cell Biology 3 (1): 100–105. https://doi.org/10.1038/35050509.

Medrano, Ruan Felipe Vieira, and Camila Andréa De Oliveira. 2014. “Guidelines for the Tetra- Primer ARMS-PCR Technique Development.” Molecular Biotechnology 56 (7): 599–608.

334

https://doi.org/10.1007/s12033-014-9734-4.

Meijering, E., M. Jacob, J.-C.F. Sarria, P. Steiner, H. Hirling, and M. Unser. 2004. “Design and Validation of a Tool for Neurite Tracing and Analysis in Fluorescence Microscopy Images.” Cytometry 58A (2): 167–76. https://doi.org/10.1002/cyto.a.20022.

Meisenberg, Cornelia, Phillip S. Tait, Irina I. Dianova, Katherine Wright, Mariola J. Edelmann, Nicola Ternette, Takafumi Tasaki, et al. 2012. “Ubiquitin Ligase UBR3 Regulates Cellular Levels of the Essential DNA Repair Protein APE1 and Is Required for Genome Stability.” Nucleic Acids Research 40 (2): 701–11. https://doi.org/10.1093/nar/gkr744.

Metzger, Meredith B, Jonathan N Pruneda, Rachel E Klevit, and Allan M Weissman. 2014. “RING-Type E3 Ligases: Master Manipulators of E2 Ubiquitin-Conjugating Enzymes and Ubiquitination.” Biochimica et Biophysica Acta - Molecular Cell Research. NIH Public Access. https://doi.org/10.1016/j.bbamcr.2013.05.026.

Miao, Yilong, Changyin Zhou, Zhaokang Cui, Xiaoxin Dai, Mianqun Zhang, Yajuan Lu, and Bo Xiong. 2017. “Smc1β Is Required for Activation of SAC during Mouse Oocyte Meiosis.” Cell Cycle 16 (6): 536–44. https://doi.org/10.1080/15384101.2017.1282583.

Michels, Karin B., Holly R. Harris, and Ludovic Barault. 2011. “Birthweight, Maternal Weight Trajectories and Global DNA Methylation of LINE-1 Repetitive Elements.” Edited by Sebastian D. Fugmann. PLoS ONE 6 (9): e25254. https://doi.org/10.1371/journal.pone.0025254.

Migeon, Barbara R, Joyce Axelman, and Peter Jeppesen. 2005. “Differential X Reactivation in Human Placental Cells: Implications for Reversal of X Inactivation.” American Journal of Human Genetics 77 (3): 355–64. https://doi.org/10.1086/432815.

Miller, Matthew P, Elçin Unal, Gloria A Brar, and Angelika Amon. 2012. “Meiosis I Chromosome Segregation Is Established through Regulation of Microtubule- Kinetochore Interactions.” ELife 1 (December): e00117. https://doi.org/10.7554/eLife.00117.

Mita, Paolo, Aleksandra Wudzinska, Xiaoji Sun, Joshua Andrade, Shruti Nayak, David J Kahler, Sana Badri, et al. 2018. “LINE-1 Protein Localization and Functional Dynamics during the Cell Cycle.” ELife 7 (January). https://doi.org/10.7554/eLife.30058.

Mogessie, Binyam, Kathleen Scheffler, and Melina Schuh. 2018. “Assembly and Positioning of the Oocyte Meiotic Spindle.” Annual Review of Cell and Developmental Biology 34 (1): 381–403.

335

https://doi.org/10.1146/annurev-cellbio-100616-060553.

Mondal, Gourish, Meredith Stevers, Benjamin Goode, Alan Ashworth, and David A. Solomon. 2019. “A Requirement for STAG2 in Replication Fork Progression Creates a Targetable Synthetic Lethality in Cohesin-Mutant Cancers.” Nature Communications 10 (1): 1686. https://doi.org/10.1038/s41467-019-09659-z.

Monk, D, P Arnaud, S Apostolidou, F A Hills, G Kelsey, P Stanier, R Feil, and G E Moore. 2006. “Limited Evolutionary Conservation of Imprinting in the Human Placenta.” Proceedings of the National Academy of Sciences of the United States of America 103 (17): 6623–28. https://doi.org/10.1073/pnas.0511031103.

Moore, Tom, and Wolf Reik. 1996. “Genetic Conflict in Early Development: Parental Imprinting in Normal and Abnormal Growth.” Reviews of Reproduction 1: 73–77.

Moreira de Mello, Joana Carvalho, Érica Sara Souza de Araújo, Raquel Stabellini, Ana Maria Fraga, Jorge Estefano Santana de Souza, Denilce R Sumita, Anamaria A Camargo, and Lygia V Pereira. 2010. “Random X Inactivation and Extensive Mosaicism in Human Placenta Revealed by Analysis of Allele-Specific Gene Expression along the X Chromosome.” PLoS ONE 5 (6): e10947. https://doi.org/10.1371/journal.pone.0010947.

Morrison, Juliet, Maudry Laurent-rolle, Ana M Maestre, and Ricardo Rajsbaum. 2013. “Dengue Virus Co-Opts UBR4 to Degrade STAT2 and Antagonize Type I Interferon Signaling.” PloS Pathogens 9 (3): e1003265. https://doi.org/10.1371/journal.ppat.1003265.

Muñoz-Escobar, J, E Matta-Camacho, C Cho, G Kozlov, and K Gehring. 2017. “Bound Waters Mediate Binding of Diverse Substrates to a Ubiquitin Ligase.” Structure 25 (5): 719-729.e3. https://doi.org/10.1016/j.str.2017.03.004.

Muñoz-Escobar, Juliana, Guennadi Kozlov, and Kalle Gehring. 2017. “Crystal Structure of the UBR-Box from UBR6/FBXO11 Reveals Domain Swapping Mediated by Zinc Binding.” Protein Science 26 (10): 2092–97. https://doi.org/10.1002/pro.3227.

Musey, Victoria C., Delwood C. Collins, Paul I. Musey, D. Martino-Saltzman, and John R.K. Preedy. 1987. “Age-Related Changes in the Female Hormonal Environment during Reproductive Life.” American Journal of Obstetrics and Gynecology 157 (2): 312–17. https://doi.org/10.1016/S0002-9378(87)80159-X.

Nagaoka, S I, T J Hassold, and P A Hunt. 2012. “Human Aneuploidy: Mechanisms and New Insights into an Age-Old Problem.” Nature Reviews Genetics 13 (7): 493–504.

336

https://doi.org/10.1038/nrg3245.

Nakaya, Takeo, Kei-ichiro Ishiguro, Camille Belzil, Anna M Rietsch, Qunyan Yu, Roderick T Bronson, Yan Geng, et al. 2013. “P600 Plays Essential Roles in Fetal Development.” PLoS ONE 8 (6): e66269. https://doi.org/10.1371/journal.pone.0066269.

Nasmyth, K, and C H Haering. 2009. “Cohesin: Its Roles and Mechanisms.” Annual Review of Genetics 43: 525–58. https://doi.org/10.1146/annurev-genet-102108-134233.

Nasmyth, Kim. 2011. “Cohesin: A Catenase with Separate Entry and Exit Gates?” Nature Cell Biology 13 (10): 1170–77. https://doi.org/10.1038/ncb2349.

Neale, Matthew J., and Scott Keeney. 2006. “Clarifying the Mechanics of DNA Strand Exchange in Meiotic Recombination.” Nature. Nature Publishing Group. https://doi.org/10.1038/nature04885.

Neale, Matthew J., Jing Pan, and Scott Keeney. 2005. “Endonucleolytic Processing of Covalent Protein-Linked DNA Double-Strand Breaks.” Nature 436 (7053): 1053–57. https://doi.org/10.1038/nature03872.

Nelson, S M, E E Telfer, and R A Anderson. 2013. “The Ageing Ovary and Uterus: New Biological Insights.” Human Reproduction Update 19 (1): 67–83. https://doi.org/10.1093/humupd/dms043.

Nerusheva, O. O., S. Galander, J. Fernius, D. Kelly, and A. L. Marston. 2014. “Tension- Dependent Removal of Pericentromeric Shugoshin Is an Indicator of Sister Chromosome Biorientation.” Genes and Development 28 (12): 1291–1309. https://doi.org/10.1101/gad.240291.114.

Nguyen, Kha The, Chang Seok Lee, Sang Hyeon Mun, Nhung Thimy Truong, Sang Ki Park, and Cheol Sang Hwang. 2019. “N-Terminal Acetylation and the N-End Rule Pathway Control Degradation of the Lipid Droplet Protein PLIN2.” Journal of Biological Chemistry 294 (1): 379–88. https://doi.org/10.1074/jbc.RA118.005556.

Nguyen, Kha The, Sang Hyeon Mun, Chang Seok Lee, and Cheol Sang Hwang. 2018. “Control of Protein Degradation by N-Terminal Acetylation and the N-End Rule Pathway.” Experimental and Molecular Medicine. Nature Publishing Group. https://doi.org/10.1038/s12276-018-0097-y.

Nieder, GL., and L Jennes. 1990. “Production of Mouse Placental Lactogen-I by Trophoblast

337

Giant Cells in Utero and in Vitro.” Endocrinology 126 (6): 2809–14. https://doi.org/10.1210/endo-126-6-2809.

Nillegoda, Nadinath B, Maria A Theodoraki, Atin K Mandal, Katie J Mayo, Hong Yu Ren, Rasheda Sultana, Kenneth Wu, Jill Johnson, Douglas M Cyr, and Avrom J Caplan. 2010. “Ubr1 and Ubr2 Function in a Quality Control Pathway for Degradation of Unfolded Cytosolic Proteins.” Edited by Jeffrey L. Brodsky. Molecular Biology of the Cell 21 (13): 2102– 16. https://doi.org/10.1091/mbc.E10-02-0098.

Nishiyama, T., M. M. Sykora, P. J. Huis in ’t Veld, K. Mechtler, and J.-M. Peters. 2013. “Aurora B and Cdk1 Mediate Wapl Activation and Release of Acetylated Cohesin from Chromosomes by Phosphorylating Sororin.” Proceedings of the National Academy of Sciences 110 (33): 13404–9. https://doi.org/10.1073/pnas.1305020110.

Nishiyama, T, R Ladurner, J Schmitz, E Kreidl, A Schleiffer, V Bhaskara, M Bando, et al. 2010. “Sororin Mediates Sister Chromatid Cohesion by Antagonizing Wapl.” Cell 143 (5): 737– 49. https://doi.org/10.1016/j.cell.2010.10.031.

Nuebler, Johannes, Geoffrey Fudenberg, Maxim Imakaev, Nezar Abdennur, and Leonid A Mirny. 2018. “Chromatin Organization by an Interplay of Loop Extrusion and Compartmental Segregation.” Proceedings of the National Academy of Sciences of the United States of America 115 (29): E6697–6706. https://doi.org/10.1073/pnas.1717730115.

Oda, M, D Oxley, W Dean, and W Reik. 2013. “Regulation of Lineage Specific DNA Hypomethylation in Mouse Trophectoderm.” PLoS ONE 8 (6): e68846. https://doi.org/10.1371/journal.pone.0068846.

Ohtake, Fumiaki, Hikaru Tsuchiya, Yasushi Saeki, and Keiji Tanaka. 2018. “K63 Ubiquitylation Triggers Proteasomal Degradation by Seeding Branched Ubiquitin Chains.” Proceedings of the National Academy of Sciences of the United States of America 115 (7): E1401–8. https://doi.org/10.1073/pnas.1716673115.

Ollinger, R, A J Childs, H M Burgess, R M Speed, P R Lundegaard, N Reynolds, N K Gray, H J Cooke, and I R Adams. 2008. “Deletion of the Pluripotency-Associated Tex19.1 Gene Causes Activation of Endogenous Retroviruses and Defective Spermatogenesis in Mice.” PLoS Genetics 4 (9): e1000199. https://doi.org/10.1371/journal.pgen.1000199.

Öllinger, Rupert, Judith Reichmann, and Ian R. Adams. 2010. “Meiosis and Retrotransposon Silencing during Germ Cell Development in Mice.” Differentiation 79 (3): 147–58.

338

https://doi.org/10.1016/j.diff.2009.10.004.

Olsen, Shaun K., and Christopher D. Lima. 2013. “Structure of a Ubiquitin E1-E2 Complex: Insights to E1-E2 Thioester Transfer.” Molecular Cell 49 (5): 884–96. https://doi.org/10.1016/j.molcel.2013.01.013.

Ouyang, Yan, Yong Tae Kwon, Jee Young An, Danny Eller, Shih Chang Tsai, Silvia Diaz-Perez, Joshua J. Troke, Michael A. Teitell, and York Marahrens. 2006. “Loss of Ubr2, an E3 Ubiquitin Ligase, Leads to Chromosome Fragility and Impaired Homologous Recombinational Repair.” Mutation Research - Fundamental and Molecular Mechanisms of Mutagenesis 596 (1-2 SPEC. ISS.): 64–75. https://doi.org/10.1016/j.mrfmmm.2005.12.016.

Pan, H, P Ma, W Zhu, and R M Schultz. 2008. “Age-Associated Increase in Aneuploidy and Changes in Gene Expression in Mouse Eggs.” Developmental Biology 316 (2): 397–407. https://doi.org/10.1016/j.ydbio.2008.01.048.

Pankiv, Serhiy, Terje Høyvarde Clausen, Trond Lamark, Andreas Brech, Jack-Ansgar Bruun, Heidi Outzen, Aud Øvervatn, Geir Bjørkøy, and Terje Johansen. 2007. “P62/SQSTM1 Binds Directly to Atg8/LC3 to Facilitate Degradation of Ubiquitinated Protein Aggregates by Autophagy.” Journal of Biological Chemistry 282 (33): 24131–45. https://doi.org/10.1074/jbc.M702824200.

Paraskevopoulos, Konstantinos, Franziska Kriegenburg, Michael H. Tatham, Heike I. Rösner, Bethan Medina, Ida B. Larsen, Rikke Brandstrup, et al. 2014. “Dss1 Is a 26S Proteasome Ubiquitin Receptor.” Molecular Cell 56 (3): 453–61. https://doi.org/10.1016/j.molcel.2014.09.008.

Park, Sang Eun, Jeong Mok Kim, Ok Hee Seok, Hanna Cho, Brandon Wadas, Seon Young Kim, Alexander Varshavsky, and Cheol Sang Hwang. 2015. “Control of Mammalian g Protein Signaling by N-Terminal Acetylation and the N-End Rule Pathway.” Science 347 (6227): 1249–52. https://doi.org/10.1126/science.aaa3844.

Parte, Seema, Deepa Bhartiya, Hiren Patel, Vinita Daithankar, Anahita Chauhan, Kusum Zaveri, and Indira Hinduja. 2014. “Dynamics Associated with Spontaneous Differentiation of Ovarian Stem Cells in Vitro.” Journal of Ovarian Research 7 (1): 25. https://doi.org/10.1186/1757-2215-7-25.

Patel, Amrita, Gary J. Sibbet, and Danny T. Huang. 2019. “Structural Insights into Non- Covalent Ubiquitin Activation of the CIAP1-UbcH5B∼ubiquitin Complex.” Journal of

339

Biological Chemistry 294 (4): 1240–49. https://doi.org/10.1074/jbc.RA118.006045.

Pellestor, F., Brigitte Andréo, Françioise Arnal, Claude Humeau, and Jacques Demaille. 2002. “Mechanisms of Non-Disjunction in Human Female Meiosis: The Co-Existence of Two Modes of Malsegregation Evidenced by the Karyotyping of 1397 in-Vitro Unfertilized Oocytes.” Human Reproduction 17 (8): 2134–45. https://doi.org/10.1093/humrep/17.8.2134.

Perkins, Adrienne T., Thomas M. Das, Lauren C. Panzera, and Sharon E. Bickel. 2016. “Oxidative Stress in Oocytes during Midprophase Induces Premature Loss of Cohesion and Chromosome Segregation Errors.” Proceedings of the National Academy of Sciences 113 (44): E6823–30. https://doi.org/10.1073/pnas.1612047113.

Perry, Anthony C.F., and Marie Hélène Verlhac. 2008. “Second Meiotic Arrest and Exit in Frogs and Mice.” EMBO Reports. https://doi.org/10.1038/embor.2008.22.

Pertceva, Julia A, Natalia V Dorogova, Elena U Bolobolova, Olga O Nerusheva, Svetlana A Fedorova, and Leonid V Omelyanchuk. 2010. “The Role of Drosophila Hyperplastic Discs Gene in Spermatogenesis.” Cell Biology International 34 (10): 991–96. https://doi.org/10.1042/cbi20100105.

Pickart, Cecile M., and Michael J. Eddins. 2004. “Ubiquitin: Structures, Functions, Mechanisms.” Biochimica et Biophysica Acta - Molecular Cell Research. Elsevier. https://doi.org/10.1016/j.bbamcr.2004.09.019.

Pisani, Francesca M., Ettore Napolitano, Luisa M.R. Napolitano, and Silvia Onesti. 2018. “Molecular and Cellular Functions of the Warsaw Breakage Syndrome DNA Helicase DDX11.” Genes. https://doi.org/10.3390/genes9110564.

Polani, Paul E, and John A Crolla. 1991. “A Test of the Production Line Hypothesis of Mammalian Oogenesis.” Human Genetics 88 (1): 64–70. https://doi.org/10.1007/BF00204931.

Polanski, Z. 2013. “Spindle Assembly Checkpoint Regulation of Chromosome Segregation in Mammalian Oocytes.” Reproduction, Fertility and Development 25 (3): 472–83. https://doi.org/10.1071/RD12145.

Polevoda, Bogdan, Thomas Arnesen, and Fred Sherman. 2009. “A Synopsis of Eukaryotic Nα- Terminal Acetyltransferases: Nomenclature, Subunits and Substrates.” BMC Proceedings 3 (Suppl 6): S2. https://doi.org/10.1186/1753-6561-3-s6-s2.

340

Prieto, Ignacio, Nieves Pezzi, José M Buesa, Leonor Kremer, Isabel Barthelemy, Candelas Carreiro, Fernando Roncal, et al. 2002. “STAG2 and Rad21 Mammalian Mitotic Cohesins Are Implicated in Meiosis.” EMBO Reports 3 (6): 543–50. https://doi.org/10.1093/embo-reports/kvf108.

Prieto, Ignacio, José A. Suja, Nieves Pezzi, Leonor Kremer, Carlos Martínez-A., Julio S. Rufas, and José L. Barbero. 2001. “Mammalian STAG3 Is a Cohesin Specific to Sister Chromatid Arms in Meiosis I.” Nature Cell Biology 3 (8): 761–66. https://doi.org/10.1038/35087082.

Prudhomme, Julie, Agnès Dubois, Pablo Navarro, Danielle Arnaud, Philip Avner, and Céline Morey. 2015. “A Rapid Passage through a Two-Active-X-Chromosome State Accompanies the Switch of Imprinted X-Inactivation Patterns in Mouse Trophoblast Stem Cells.” Epigenetics and Chromatin 8 (1): 52. https://doi.org/10.1186/s13072-015-0044-2.

Puchtler, H., S. N. Meloan, and F. S. Waldrop. 1986. “Application of Current Chemical Concepts to Metal-Hematein and-Brazilein Stains.” Histochemistry 85 (5): 353–64. https://doi.org/10.1007/BF00982665.

Racicot, Karen, Ja Young Kwon, Paulomi Aldo, Vikki Abrahams, Ayman El-Guindy, Roberto Romero, and Gil Mor. 2016. “Type I Interferon Regulates the Placental Inflammatory Response to Bacteria and Is Targeted by Virus: Mechanism of Polymicrobial Infection- Induced Preterm Birth.” American Journal of Reproductive Immunology 75 (4): 451–60. https://doi.org/10.1111/aji.12501.

Rageul, Julie, Jennifer J Park, Ukhyun Jo, Alexandra S Weinheimer, Tri T.M. Vu, and Hyungjin Kim. 2019. “Conditional Degradation of SDE2 by the Arg/N-End Rule Pathway Regulates Stress Response at Replication Forks.” Nucleic Acids Research 47 (8): 3996–4010. https://doi.org/10.1093/nar/gkz054.

Rao, Hai, Frank Uhlmann, Kim Nasmyth, and Alexander Varshavsky. 2001. “Degradation of a Cohesin Subunit by the N-End Rule Pathway Is Essential for Chromosome Stability.” Nature 410 (6831): 955–59. https://doi.org/10.1038/35073627.

Rattani, Ahmed, Magda Wolna, Mickael Ploquin, Wolfgang Helmhart, Seamus Morrone, Bernd Mayer, Jonathan Godwin, et al. 2013. “Sgol2 Provides a Regulatory Platform That Coordinates Essential Cell Cycle Processes during Meiosis I in Oocytes.” ELife 2 (November): e01133. https://doi.org/10.7554/eLife.01133.

Reggiori, Fulvio, and Daniel J Klionsky. 2005. “Autophagosomes: Biogenesis from Scratch?” Current Opinion in Cell Biology 17 (4): 415–22. https://doi.org/10.1016/j.ceb.2005.06.007.

341

Reichman, Rachel, Benjamin Alleva, and Sarit Smolikove. 2017. “Prophase I: Preparing Chromosomes for Segregation in the Developing Oocyte.” In Signaling-Mediated Control of Cell Division, 125–73. Springer, Cham. https://doi.org/10.1007/978-3-319-44820-6_5.

Reichmann, J, K Dobie, L M Lister, D Best, J H Crichton, H J Cooke, W A Bickmore, M Herbert, and I R Adams. 2017. “Tex19 . 1 Regulates Acetylated SMC3 Cohesin and Prevents Aneuploidy in Mouse Oocytes.” BioRxiv (Preprint).

Reichmann, J, J P Reddington, D Best, D Read, R Ollinger, R R Meehan, and I R Adams. 2013. “The Genome-Defence Gene Tex19.1 Suppresses LINE-1 Retrotransposons in the Placenta and Prevents Intra-Uterine Growth Retardation in Mice.” Human Molecular Genetics 22 (9): 1791–1806. https://doi.org/10.1093/hmg/ddt029.

Reichmann, Judith, James H. Crichton, Monika J. Madej, Mary Taggart, Philippe Gautier, Jose Luis Garcia-Perez, Richard R. Meehan, and Ian R. Adams. 2012. “Microarray Analysis of LTR Retrotransposon Silencing Identifies Hdac1 as a Regulator of Retrotransposon Expression in Mouse Embryonic Stem Cells.” Edited by Amos Tanay. PLoS Computational Biology 8 (4): e1002486. https://doi.org/10.1371/journal.pcbi.1002486.

Reik, W, W Dean, and J Walter. 2001. “Epigenetic Reprogramming in Mammalian Development.” Science. American Association for the Advancement of Science. https://doi.org/10.1126/science.1063443.

Revenkova, E, K Herrmann, C Adelfalk, and R Jessberger. 2010. “Oocyte Cohesin Expression Restricted to Predictyate Stages Provides Full Fertility and Prevents Aneuploidy.” Current Biology 20 (17): 1529–33. https://doi.org/10.1016/j.cub.2010.08.024.

Revenkova, Ekaterina, Maureen Eijpe, Christa Heyting, Craig A. Hodges, Patricia A. Hunt, Bodo Liebe, Harry Scherthan, and Rolf Jessberger. 2004. “Cohesin SMC1β Is Required for Meiotic Chromosome Dynamics, Sister Chromatid Cohesion and DNA Recombination.” Nature Cell Biology 6 (6): 555–62. https://doi.org/10.1038/ncb1135.

Rhodes, James, Davide Mazza, Kim Nasmyth, and Stephan Uphoff. 2017. “Scc2/Nipbl Hops between Chromosomal Cohesin Rings after Loading.” ELife 6 (September): 1–20. https://doi.org/10.7554/eLife.30000.

Richards, S, T Hillman, and M Stern. 1996. “Mutations in the Drosophila Pushover Gene Confer Increased Neuronal Excitability and Spontaneous Synaptic Vesicle Fusion.” Genetics 142 (4): 1215–23. http://www.ncbi.nlm.nih.gov/pubmed/8846899.

342

Richardson, Sandra R., Santiago Morell, and Geoffrey J. Faulkner. 2014. “L1 Retrotransposons and Somatic Mosaicism in the Brain.” Annual Review of Genetics 48 (1): 1–27. https://doi.org/10.1146/annurev-genet-120213-092412.

Riris, S, P Webster, and H Homer. 2014. “Digital Multiplexed MRNA Analysis of Functionally Important Genes in Single Human Oocytes and Correlation of Changes in Transcript Levels with Oocyte Protein Expression.” Fertility and Sterility 101 (3): 857–64. https://doi.org/10.1016/j.fertnstert.2013.11.125.

Roest, H. P., W. M. Baarends, J. de Wit, J. W. van Klaveren, E. Wassenaar, J. W. Hoogerbrugge, W. A. van Cappellen, J. H. J. Hoeijmakers, and J. A. Grootegoed. 2004. “The Ubiquitin- Conjugating DNA Repair Enzyme HR6A Is a Maternal Factor Essential for Early Embryonic Development in Mice.” Molecular and Cellular Biology 24 (12): 5485–95. https://doi.org/10.1128/mcb.24.12.5485-5495.2004.

Roest, H P, J. Van Klaveren, J. De Wit, C. G. Van Gurp, M H.M. Koken, M Vermey, J. H. Van Roijen, et al. 1996. “Inactivation of the HR6B Ubiquitin-Conjugating DNA Repair Enzyme in Mice Causes Male Sterility Associated with Chromatin Modification.” Cell 86 (5): 799–810. https://doi.org/10.1016/S0092-8674(00)80154-3.

Rogacheva, Maria V., Carol M. Manhart, Cheng Chen, Alba Guarne, Jennifer Surtees, and Eric Alani. 2014. “Mlh1-Mlh3, a Meiotic Crossover and DNA Mismatch Repair Factor, Is a Msh2-Msh3-Stimulated Endonuclease.” Journal of Biological Chemistry 289 (9): 5664–73. https://doi.org/10.1074/jbc.M113.534644.

Rong, M, A Matsuda, Y Hiraoka, and J Lee. 2016. “Meiotic Cohesin Subunits RAD21L and REC8 Are Positioned at Distinct Regions between Lateral Elements and Transverse Filaments in the Synaptonemal Complex of Mouse Spermatocytes.” Journal of Reproduction and Development 62 (6): 623–30. https://doi.org/10.1262/jrd.2016-127.

Rosenzweig, Rina, Vered Bronner, Daoning Zhang, David Fushman, and Michael H. Glickman. 2012. “Rpn1 and Rpn2 Coordinate Ubiquitin Processing Factors at Proteasome.” Journal of Biological Chemistry 287 (18): 14659–71. https://doi.org/10.1074/jbc.M111.316323.

Rossant, Janet, and James C. Cross. 2001. “Placental Development: Lessons from Mouse Mutants.” Nature Reviews Genetics 2 (7): 538–48. https://doi.org/10.1038/35080570.

Roux, Kyle J., Dae In Kim, and Brian Burke. 2013. “BioID: A Screen for Protein-Protein

343

Interactions.” Current Protocols in Protein Science 74 (November): Unit 19.23. https://doi.org/10.1002/0471140864.ps1923s74.

Rowsey, Ross, Jennifer Gruhn, Karl W Broman, Patricia A Hunt, and Terry Hassold. 2014. “Examining Variation in Recombination Levels in the Human Female: A Test of the Production-Line Hypothesis.” American Journal of Human Genetics 95 (1): 108–12. https://doi.org/10.1016/j.ajhg.2014.06.008.

Sadowski, Martin, and Boris Sarcevic. 2010. “Mechanisms of Mono- and Poly-Ubiquitination: Ubiquitination Specificity Depends on Compatibility between the E2 Catalytic Core and Amino Acid Residues Proximal to the Lysine.” Cell Division 5 (1): 19. https://doi.org/10.1186/1747-1028-5-19.

Sadowski, Martin, Randy Suryadinata, Anthonius Ricardo Tan, Siti Nur Ain Roesley, and Boris Sarcevic. 2012. “Protein Monoubiquitination and Polyubiquitination Generate Structural Diversity to Control Distinct Biological Processes.” IUBMB Life 64 (2): 136–42. https://doi.org/10.1002/iub.589.

Sakai, Hiroshi, Kenoki Ohuchida, Kazuhiro Mizumoto, and Lin Cui. 2011. “Inhibition of P600 Expression Suppresses Both Invasiveness and Anoikis Resistance of Gastric Cancer.” Annals of Surgical Oncology 18: 2057–65. https://doi.org/10.1245/s10434-010-1523-0.

Sakakibara, Y, S Hashimoto, Y Nakaoka, A Kouznetsova, C Höög, and T S Kitajima. 2015. “Bivalent Separation into Univalents Precedes Age-Related Meiosis I Errors in Oocytes.” Nature Communications 6: 7550. https://doi.org/10.1038/ncomms8550.

Sakuno, Takeshi, Kenji Tada, and Yoshinori Watanabe. 2009. “Kinetochore Geometry Defined by Cohesion within the Centromere.” Nature 458 (7240): 852–58. https://doi.org/10.1038/nature07876.

Sandstrom, Andrew, Patrick S Mitchell, Lisa Goers, Edward W Mu, Cammie F Lesser, and Russell E Vance. 2019. “Functional Degradation: A Mechanism of NLRP1 Inflammasome Activation by Diverse Pathogen Enzymes.” Science 364 (6435): eaau1330. https://doi.org/10.1126/science.aau1330.

Schalch, Thomas, and Florian A Steiner. 2017. “Structure of Centromere Chromatin: From Nucleosome to Chromosomal Architecture.” Chromosoma 126 (4): 443–55. https://doi.org/10.1007/s00412-016-0620-7.

Schneider, William M, Meike Dittmann Chevillotte, and Charles M Rice. 2014. “Interferon-

344

Stimulated Genes: A Complex Web of Host Defenses.” Annual Review of Immunology 32 (1): 513–45. https://doi.org/10.1146/annurev-immunol-032713-120231.

Schroeder, Diane I, Kartika Jayashankar, Kory C Douglas, Twanda L Thirkill, Daniel York, Pete J Dickinson, Lawrence E Williams, et al. 2015. “Early Developmental and Evolutionary Origins of Gene Body DNA Methylation Patterns in Mammalian Placentas.” PLoS Genetics 11 (8): e1005442. https://doi.org/10.1371/journal.pgen.1005442.

Schubert, Ulrich, Luis C. Antón, James Gibbs, Christopher C. Norbury, Jonathan W. Yewdell, and Jack R. Bennink. 2000. “Rapid Degradation of a Large Fraction of Newly Synthesized Proteins by Proteasomes.” Nature 404 (6779): 770–74. https://doi.org/10.1038/35008096.

Schulman, Brenda A, and J. Wade Harper. 2009. “Ubiquitin-like Protein Activation by E1 Enzymes: The Apex for Downstream Signalling Pathways.” Nature Reviews Molecular Cell Biology. NIH Public Access. https://doi.org/10.1038/nrm2673.

Schwarzer, W, N Abdennur, A Goloborodko, A Pekowska, G Fudenberg, Y Loe-Mie, N A Fonseca, et al. 2017. “Two Independent Modes of Chromatin Organization Revealed by Cohesin Removal.” Nature 551 (7678): 51–56. https://doi.org/10.1038/nature24281.

Scifres, Christina M., and D. Michael Nelson. 2009. “Intrauterine Growth Restriction, Human Placental Development and Trophoblast Cell Death.” The Journal of Physiology 587 (14): 3453–58. https://doi.org/10.1113/jphysiol.2009.173252.

Seibenhener, M. L., J. R. Babu, T. Geetha, H. C. Wong, N. R. Krishna, and M. W. Wooten. 2004. “Sequestosome 1/P62 Is a Polyubiquitin Chain Binding Protein Involved in Ubiquitin Proteasome Degradation.” Molecular and Cellular Biology 24 (18): 8055–68. https://doi.org/10.1128/mcb.24.18.8055-8068.2004.

Sen, Aritro, and Francesco Caiazza. 2013. “Oocyte Maturation: A Story of Arrest and Release.” Frontiers in Bioscience. http://www.ncbi.nlm.nih.gov/pubmed/23277062.

Sferruzzi-Perri, Amanda N, Anne M Macpherson, Claire T Roberts, and Sarah A Robertson. 2009. “Csf2 Null Mutation Alters Placental Gene Expression and Trophoblast Glycogen Cell and Giant Cell Abundance in Mice1.” Biology of Reproduction 81 (1): 207–21. https://doi.org/10.1095/biolreprod.108.073312.

Shearer, Robert F., Kari-Anne Myrum Frikstad, Jessie McKenna, Rachael A. McCloy, Niantao Deng, Andrew Burgess, Trond Stokke, Sebastian Patzke, and Darren N. Saunders. 2018.

345

“The E3 Ubiquitin Ligase UBR5 Regulates Centriolar Satellite Stability and Primary Cilia.” Edited by Yixian Zheng. Molecular Biology of the Cell 29 (13): 1542–54. https://doi.org/10.1091/mbc.E17-04-0248.

Shemorry, Anna, Cheol Sang Hwang, and Alexander Varshavsky. 2013. “Control of Protein Quality and Stoichiometries by N-Terminal Acetylation and the N-End Rule Pathway.” Molecular Cell 50 (4): 540–51. https://doi.org/10.1016/j.molcel.2013.03.018.

Shen, Qingya, Zhiyu Qiu, Wenping Wu, Jimin Zheng, and Zongchao Jia. 2018. “Characterization of Interaction and Ubiquitination of Phosphoenolpyruvate Carboxykinase by E3 Ligase UBR5.” Biology Open 7: bio037366. https://doi.org/10.1242/bio.037366.

Shi, Yuan, Xiang Chen, Suzanne Elsasser, Bradley B. Stocks, Geng Tian, Byung Hoon Lee, Yanhong Shi, et al. 2016. “Rpn1 Provides Adjacent Receptor Sites for Substrate Binding and Deubiquitination by the Proteasome.” Science 351 (6275): aad9421–aad9421. https://doi.org/10.1126/science.aad9421.

Simmons, David G, David R C Natale, Valerie Begay, Martha Hughes, Achim Leutz, and James C Cross. 2008. “Early Patterning of the Chorion Leads to the Trilaminar Trophoblast Cell Structure in the Placental Labyrinth.” Development 135 (12): 2083–91. https://doi.org/10.1242/dev.020099.

Skibbens, Robert V. 2019. “Condensins and Cohesins - One of These Things Is Not like the Other!” Journal of Cell Science. https://doi.org/10.1242/jcs.220491.

Skiles, William M, Avery Kester, Jane H Pryor, Mark E Westhusin, Michael C Golding, and Charles R Long. 2018. “Oxygen-Induced Alterations in the Expression of Chromatin Modifying Enzymes and the Transcriptional Regulation of Imprinted Genes.” Gene Expression Patterns 28 (June): 1–11. https://doi.org/10.1016/j.gep.2018.01.001.

Slotkin, R. Keith, and Robert Martienssen. 2007. “Transposable Elements and the Epigenetic Regulation of the Genome.” Nature Reviews Genetics. Nature Publishing Group. https://doi.org/10.1038/nrg2072.

Sluimer, Jasper, and Ben Distel. 2018. “Regulating the Human HECT E3 Ligases.” Cellular and Molecular Life Sciences. Springer. https://doi.org/10.1007/s00018-018-2848-2.

Smith, James S, Kienan I Savage, Alexander Thompson, and Ken I Mills. 2016. “Loss of Function Cohesin Complex Gene Mutations Create Neomorphic Cell States

346

Advantageous to Oncogenesis.” Blood 128 (22). http://www.bloodjournal.org/content/128/22/1564?sso-checked=true.

Smith, Zachary D, Michelle M Chan, Kathryn C Humm, Rahul Karnik, Shila Mekhoubad, Aviv Regev, Kevin Eggan, and Alexander Meissner. 2014. “DNA Methylation Dynamics of the Human Preimplantation Embryo.” Nature 511 (7511): 611–15. https://doi.org/10.1038/nature13581.

Smith, Zachary D, Jiantao Shi, Hongcang Gu, Julie Donaghey, Kendell Clement, Davide Cacchiarelli, Andreas Gnirke, Franziska Michor, and Alexander Meissner. 2017. “Epigenetic Restriction of Extraembryonic Lineages Mirrors the Somatic Transition to Cancer.” Nature 549 (7673): 543–47. https://doi.org/10.1038/nature23891.

Soh, Y. Q.Shirleen, Maria M. Mikedis, Mina Kojima, Alexander K. Godfrey, Dirk G. de Rooij, and David C. Page. 2017. “Meioc Maintains an Extended Meiotic Prophase I in Mice.” Edited by Paula E. Cohen. PLoS Genetics 13 (4): e1006704. https://doi.org/10.1371/journal.pgen.1006704.

Speed, R M, and A C Chandley. 1983. “Meiosis in the Foetal Mouse Ovary. II. Oocyte Development and Age-Related Aneuploidy. Does a Production Line Exist?” Chromosoma 88 (3): 184–89. https://doi.org/10.1007/bf00285618.

Spratt, Donald E., Helen Walden, and Gary S. Shaw. 2014. “RBR E3 Ubiquitin Ligases: New Structures, New Insights, New Questions.” Biochemical Journal. https://doi.org/10.1042/BJ20140006.

Stewart, Mikaela D, Tobias Ritterhoff, Rachel E Klevit, and Peter S Brzovic. 2016. “E2 Enzymes: More than Just Middle Men.” Cell Research. Nature Publishing Group. https://doi.org/10.1038/cr.2016.35.

Summers, Daniel W, Katie J Wolfe, Hong Yu Ren, and Douglas M Cyr. 2013. “The Type II Hsp40 Sis1 Cooperates with Hsp70 and the E3 Ligase Ubr1 to Promote Degradation of Terminally Misfolded Cytosolic Protein.” Edited by Jeffrey L. Brodsky. PLoS ONE 8 (1): e52099. https://doi.org/10.1371/journal.pone.0052099.

Sun, Yuxiao, Martin Kucej, Heng Yu Fan, Hong Yu, Qing Yuan Sun, and Hui Zou. 2009. “Separase Is Recruited to Mitotic Chromosomes to Dissolve Sister Chromatid Cohesion in a DNA-Dependent Manner.” Cell 137 (1): 123–32. https://doi.org/10.1016/j.cell.2009.01.040.

347

Surani, M. A.H., S. C. Barton, and M. L. Norris. 1984. “Development of Reconstituted Mouse Eggs Suggests Imprinting of the Genome during Gametogenesis.” Nature 308 (5959): 548–50. https://doi.org/10.1038/308548a0.

Syrjänen, Johanna Liinamaria, Luca Pellegrini, and Owen Richard Davies. 2014. “A Molecular Model for the Role of SYCP3 in Meiotic Chromosome Organisation.” ELife 3 (June). https://doi.org/10.7554/eLife.02963.

Szollosi, D., P. Calarco, and R. P. Donahue. 1972. “Absence of Centrioles in the First and Second Meiotic Spindles of Mouse Oocytes.” Journal of Cell Science 11 (2).

Tachibana-Konwalski, K, J Godwin, L van der Weyden, L Champion, N R Kudo, D J Adams, and K Nasmyth. 2010. “Rec8-Containing Cohesin Maintains Bivalents without Turnover during the Growing Phase of Mouse Oocytes.” Genes and Development 24 (22): 2505–16. https://doi.org/10.1101/gad.605910.

Takahashi, K., T Kobayashi, and N Kanayama. 2000. “P57Kip2 Regulates the Proper Development of Labyrinthine and Spongiotrophoblasts.” Molecular Human Reproduction 6 (11): 1019–25. https://doi.org/10.1093/molehr/6.11.1019.

Takeda, Kojiro, and Mitsuhiro Yanagida. 2005. “Regulation of Nuclear Proteasome by Rhp6/Ubc2 through Ubiquitination and Destruction of the Sensor and Anchor Cut8.” Cell 122 (3): 393–405. https://doi.org/10.1016/j.cell.2005.05.023.

Tan, J L, L Ward, K N Truscott, and D A Dougan. 2016. “The N-End Rule Adaptor Protein ClpS from Plasmodium Falciparum Exhibits Broad Substrate Specificity.” FEBS Letters 590 (19): 3397–3406. https://doi.org/10.1002/1873-3468.12382.

Tanaka, Keiji. 2009. “The Proteasome: Overview of Structure and Functions.” Proceedings of the Japan Academy. Series B, Physical and Biological Sciences 85 (1): 12–36. https://doi.org/10.2183/pjab.85.12.

Tang, An, Peiliang Shi, Anying Song, Dayuan Zou, Yue Zhou, Pengyu Gu, Zan Huang, Qinghua Wang, Zhaoyu Lin, and Xiang Gao. 2016. “PP2A Regulates Kinetochore-Microtubule Attachment during Meiosis I in Oocyte.” Cell Cycle 15 (11): 1450–61. https://doi.org/10.1080/15384101.2016.1175256.

Taniguchi, Tadatsugu, and Akinori Takaoka. 2002. “The Interferon-α/β System in Antiviral Responses: A Multimodal Machinery of Gene Regulation by the IRF Family of Transcription Factors.” Current Opinion in Immunology. Elsevier Current Trends.

348

https://doi.org/10.1016/S0952-7915(01)00305-3.

Tarabay, Y, E Kieffer, M Teletin, C Celebi, A Van Montfoort, N Zamudio, M Achour, et al. 2013. “The Mammalian-Specific Tex19.1 Gene Plays an Essential Role in Spermatogenesis and Placenta-Supported Development.” Human Reproduction 28 (8): 2201–14. https://doi.org/10.1093/humrep/det129.

Tasaki, T, L C Mulder, A Iwamatsu, M J Lee, I V Davydov, A Varshavsky, M Muesing, and Y T Kwon. 2005. “A Family of Mammalian E3 Ubiquitin Ligases That Contain the UBR Box Motif and Recognize N-Degrons.” Molecular and Cellular Biology 25 (16): 7120–36. https://doi.org/10.1128/MCB.25.16.7120-7136.2005.

Tasaki, T, A Zakrzewska, D D Dudgeon, Y Jiang, J S Lazo, and Y T Kwon. 2009. “The Substrate Recognition Domains of the N-End Rule Pathway.” Journal of Biological Chemistry 284 (3): 1884–95. https://doi.org/10.1074/jbc.M803641200.

Tasaki, Takafumi, and Yong Tae Kwon. 2007. “The Mammalian N-End Rule Pathway: New Insights into Its Components and Physiological Roles.” Trends in Biochemical Sciences 32 (11): 520–28. https://doi.org/10.1016/j.tibs.2007.08.010.

Tasaki, Takafumi, Shashikanth M Sriram, Kyong Soo Park, and Yong Tae Kwon. 2012. “The N-End Rule Pathway.” Annual Review of Biochemistry 81: 261–89. https://doi.org/10.1146/annurev-biochem-051710-093308.

Tease, C, and G Fisher. 1986. “Further Examination of the Production-Line Hypothesis in Mouse Foetal Oocytes. I. Inversion Heterozygotes.” Chromosoma 93 (5): 447–52. https://doi.org/10.1007/bf00285827.

Tedeschi, A, G Wutz, S Huet, M Jaritz, A Wuensche, E Schirghuber, I F Davidson, et al. 2013. “Wapl Is an Essential Regulator of Chromatin Structure and Chromosome Segregation.” Nature 501 (7468): 564–68. https://doi.org/10.1038/nature12471.

Telfer, E. E., and R. A. Anderson. 2019. “The Existence and Potential of Germline Stem Cells in the Adult Mammalian Ovary.” Climacteric. Taylor & Francis. https://doi.org/10.1080/13697137.2018.1543264.

Terret, M.Emilie, Katja Wassmann, Irene Waizenegger, Bernard Maro, Jan-Michael Peters, and Marie-Hélène Verlhac. 2003. “The Meiosis I-to-Meiosis II Transition in Mouse Oocytes Requires Separase Activity.” Current Biology 13 (20): 1797–1802. https://doi.org/10.1016/j.cub.2003.09.032.

349

Thomas, Geethu Emily, Marira R. Renjith, and Tapas K. Manna. 2017. “Kinetochore– Microtubule Interactions in Chromosome Segregation: Lessons from Yeast and Mammalian Cells.” Biochemical Journal 474 (21): 3559–77. https://doi.org/10.1042/BCJ20170518.

Thum, Meen-Yau, Hossam I. Abdalla, and Deborah Taylor. 2008. “Relationship between Women’s Age and Basal Follicle-Stimulating Hormone Levels with Aneuploidy Risk in in Vitro Fertilization Treatment.” Fertility and Sterility 90 (2): 315–21. https://doi.org/10.1016/j.fertnstert.2007.06.063.

Toda, Kazuhiro, Kayoko Naito, Satoru Mase, Masaru Ueno, Masahiro Uritani, Ayumu Yamamoto, and Takashi Ushimaru. 2012. “APC/C-Cdh1-Dependent Anaphase and Telophase Progression during Mitotic Slippage.” Cell Division 7 (1): 4. https://doi.org/10.1186/1747-1028-7-4.

Tsuchiya, Hikaru, Fumiaki Ohtake, Naoko Arai, Ai Kaiho, Sayaka Yasuda, Keiji Tanaka, and Yasushi Saeki. 2017. “In Vivo Ubiquitin Linkage-Type Analysis Reveals That the Cdc48- Rad23/Dsk2 Axis Contributes to K48-Linked Chain Specificity of the Proteasome.” Molecular Cell 66 (4): 488-502.e7. https://doi.org/10.1016/j.molcel.2017.04.024.

Tsutsumi, M, R Fujiwara, H Nishizawa, M Ito, H Kogo, H Inagaki, T Ohye, T Kato, T Fujii, and H Kurahashi. 2014. “Age-Related Decrease of Meiotic Cohesins in Human Oocytes.” PLoS One 9 (5): e96710. https://doi.org/10.1371/journal.pone.0096710.

Tunster, S. J., A. B. Jensen, and R. M. John. 2013. “Imprinted Genes in Mouse Placental Development and the Regulation of Fetal Energy Stores.” Reproduction. https://doi.org/10.1530/REP-12-0511.

Turner, Glenn C., Fangyong Du, and Alexander Varshavsky. 2000. “Peptides Accelerate Their Uptake by Activating a Ubiquitin-Dependent Proteolytic Pathway.” Nature 405 (6786): 579–83. https://doi.org/10.1038/35014629.

Turner, James M.A. 2007. “Meiotic Sex Chromosome Inactivation.” Development 134 (10): 1823– 31. https://doi.org/10.1242/dev.000018.

Uhlmann, Frank, Friedrich Lottspelch, and Kim Nasmyth. 1999. “Sister-Chromatid Separation at Anaphase Onset Is Promoted by Cleavage of the Cohesin Subunit Scc1.” Nature 400 (6739): 37–42. https://doi.org/10.1038/21831.

Visnes, T., F. Giordano, A. Kuznetsova, J. A. Suja, A. D. Lander, A. L. Calof, and L. Ström. 2014.

350

“Localisation of the SMC Loading Complex Nipbl/Mau2 during Mammalian Meiotic Prophase I.” Chromosoma 123 (3): 239–52. https://doi.org/10.1007/s00412-013-0444-7.

Walker, William H., Susan L. Fitzpatrick, Grady F. Saunders, Hugo A. Barrera-Saldaña, and Diana Resendez-Perez. 1991. “The Human Placental Lactogen Genes: Structure, Function, Evolution and Transcriptional Regulation.” Endocrine Reviews 12 (4): 316–28. https://doi.org/10.1210/edrv-12-4-316.

Walter, Marius, Aurélie Teissandier, Raquel Pérez-Palacios, and Déborah Bourc’his. 2016. “An Epigenetic Switch Ensures Transposon Repression upon Dynamic Loss of DNA Methylation in Embryonic Stem Cells.” ELife 5 (January). https://doi.org/10.7554/eLife.11418.

Wang, Guolun, Xiaofang Tang, Yujie Chen, Jun Cao, Qinzhu Huang, Xuemei Ling, Wenyan Ren, et al. 2014. “Hyperplastic Discs Differentially Regulates the Transcriptional Outputs of Hedgehog Signaling.” Mechanisms of Development 133 (August): 117–25. https://doi.org/10.1016/j.mod.2014.05.002.

Wang, Li, Xicheng Mao, Donghong Ju, and Youming Xie. 2004. “Rpn4 Is a Physiological Substrate of the Ubr2 Ubiquitin Ligase.” Journal of Biological Chemistry 279 (53): 55218–23. https://doi.org/10.1074/jbc.M410085200.

Wang, Lin Ing, Arunika Das, and Kim S. McKim. 2019. “Sister Centromere Fusion during Meiosis I Depends on Maintaining Cohesins and Destabilizing Microtubule Attachments.” Edited by Giovanni Bosco. PLoS Genetics 15 (5): e1008072. https://doi.org/10.1371/journal.pgen.1008072.

Wang, Min, and Cecile M Pickart. 2005. “Different HECT Domain Ubiquitin Ligases Employ Distinct Mechanisms of Polyubiquitin Chain Synthesis.” EMBO Journal 24 (24): 4324–33. https://doi.org/10.1038/sj.emboj.7600895.

Warburton, D. 2005. “Biological Aging and the Etiology of Aneuploidy.” Cytogenetic and Genome Research 111 (3–4): 266–72. https://doi.org/10.1159/000086899.

Ward, Ayobami, Jessica Hopkins, Matthew Mckay, Steve Murray, and Philip W. Jordan. 2016. “Genetic Interactions Between the Meiosis-Specific Cohesin Components, STAG3, REC8, and RAD21L.” G3 (Bethesda) 6 (6): 1713–24. https://doi.org/10.1534/g3.116.029462.

Watrin, Erwan, Alexander Schleiffer, Koichi Tanaka, Frank Eisenhaber, Kim Nasmyth, and Jan-Michael Peters. 2006. “Human Scc4 Is Required for Cohesin Binding to Chromatin,

351

Sister-Chromatid Cohesion, and Mitotic Progression.” Current Biology 16 (9): 863–74. https://doi.org/10.1016/j.cub.2006.03.049.

Weber, Stewart A, Jennifer L Gerton, Joan E Polancic, Joseph L DeRisi, Douglas Koshland, and Paul C Megee. 2004. “The Kinetochore Is an Enhancer of Pericentric Cohesin Binding.” PLoS Biology 2 (9): E260. https://doi.org/10.1371/journal.pbio.0020260.

Weghofer, Andrea, Santiago Munne, Serena Chen, David Barad, and Norbert Gleicher. 2007. “Lack of Association between Polycystic Ovary Syndrome and Embryonic Aneuploidy.” Fertility and Sterility 88 (4): 900–905. https://doi.org/10.1016/j.fertnstert.2006.12.018.

Wendt, Kerstin S., Keisuke Yoshida, Takehiko Itoh, Masashige Bando, Birgit Koch, Erika Schirghuber, Shuichi Tsutsumi, et al. 2008. “Cohesin Mediates Transcriptional Insulation by CCCTC-Binding Factor.” Nature 451 (7180): 796–801. https://doi.org/10.1038/nature06634.

Whelan, Gabriela, Emanuel Kreidl, Gordana Wutz, Alexander Egner, Jan Michael Peters, and Gregor Eichele. 2012. “Cohesin Acetyltransferase Esco2 Is a Cell Viability Factor and Is Required for Cohesion in Pericentric Heterochromatin.” EMBO Journal 31 (1): 71–82. https://doi.org/10.1038/emboj.2011.381.

Winters, Tristan, Francois McNicoll, and Rolf Jessberger. 2014. “Meiotic Cohesin STAG3 Is Required for Chromosome Axis Formation and Sister Chromatid Cohesion.” EMBO Journal 33 (11): 1256–70. https://doi.org/10.1002/embj.201387330.

Wolf, Peter G., Alexander Cuba Ramos, Julia Kenzel, Brigitte Neumann, and Olaf Stemmann. 2018. “Studying Meiotic Cohesin in Somatic Cells Reveals That Rec8-Containing Cohesin Requires Stag3 to Function and Is Regulated by Wapl and Sororin.” Journal of Cell Science 131 (11): jcs212100. https://doi.org/10.1242/jcs.212100.

Woods, Laura, Vicente Perez-Garcia, and Myriam Hemberger. 2018. “Regulation of Placental Development and Its Impact on Fetal Growth—New Insights From Mouse Models.” Frontiers in Endocrinology 9: 570. https://doi.org/10.3389/fendo.2018.00570.

Wutz, Anton, and Rudolf Jaenisch. 2000. “A Shift from Reversible to Irreversible X Inactivation Is Triggered during ES Cell Differentiation.” Molecular Cell 5 (4): 695–705. https://doi.org/10.1016/S1097-2765(00)80248-8.

Wutz, Gordana, Csilla Várnai, Kota Nagasaka, David A Cisneros, Roman R Stocsits, Wen Tang, Stefan Schoenfelder, et al. 2017. “Topologically Associating Domains and

352

Chromatin Loops Depend on Cohesin and Are Regulated by CTCF, WAPL, and PDS5 Proteins.” EMBO Journal 36 (24): 3573–99. https://doi.org/10.15252/embj.201798004.

Wynne, Freda, Melanie Ball, Andrew S McLellan, Peter Dockery, Wolfgang Zimmermann, and Tom Moore. 2006. “Mouse Pregnancy-Specific Glycoproteins: Tissue-Specific Expression and Evidence of Association with Maternal Vasculature.” Reproduction 131 (4): 721–32. https://doi.org/10.1530/rep.1.00869.

Xia, Zanxian, Ailsa Webster, Fangyong Du, Konstantin Piatkov, Michel Ghislain, and Alexander Varshavsky. 2008. “Substrate-Binding Sites of UBR1, the Ubiquitin Ligase of the N-End Rule Pathway.” Journal of Biological Chemistry 283 (35): 24011–28. https://doi.org/10.1074/jbc.M802583200.

Xing, Zhuo, Yichen Li, Annie Pao, Abigail S Bennett, Benjamin Tycko, William C Mobley, and Y Eugene Yu. 2016. “Mouse-Based Genetic Modeling and Analysis of Down Syndrome.” British Medical Bulletin. Oxford University Press. https://doi.org/10.1093/bmb/ldw040.

Xu, Hao, Jianjin Shi, Hang Gao, Ying Liu, Zhenxiao Yang, and Feng Shao. 2019. “The N-End Rule Ubiquitin Ligase UBR 2 Mediates NLRP 1 B Inflammasome Activation by Anthrax Lethal Toxin.” EMBO Journal, e101996. https://doi.org/10.15252/embj.2019101996.

Xu, Zheng, Bulent Cetin, Martin Anger, Uhn Soo Cho, Wolfgang Helmhart, Kim Nasmyth, and Wenqing Xu. 2009. “Structure and Function of the PP2A-Shugoshin Interaction.” Molecular Cell 35 (4): 426–41. https://doi.org/10.1016/j.molcel.2009.06.031.

Yanagitani, Kota, Szymon Juszkiewicz, and Ramanujan S. Hegde. 2017. “UBE2O Is a Quality Control Factor for Orphans of Multiprotein Complexes.” Science 357 (6350): 472–75. https://doi.org/10.1126/science.aan0178.

Yang, F, Y Cheng, J Y An, Y T Kwon, S Eckardt, N A Leu, K J McLaughlin, and P J Wang. 2010. “The Ubiquitin Ligase Ubr2, a Recognition E3 Component of the N-End Rule Pathway, Stabilizes Tex19.1 during Spermatogenesis.” PLoS ONE 5 (11): e14017. https://doi.org/10.1371/journal.pone.0014017.

Ye, Haifeng, Tuochen Zheng, Wei Li, Xiaoyan Li, Xinxin Fu, Yaoqi Huang, Chuan Hu, et al. 2017. “Ovarian Stem Cell Nests in Reproduction and Ovarian Aging.” Cellular Physiology and Biochemistry 43 (5): 1917–25. https://doi.org/10.1159/000484114.

Ye, S., Sahar Dhillon, Xiayi Ke, Andrew R. Collins, and Ian N. M. Day. 2001. “An Efficient Procedure for Genotyping Single Nucleotide Polymorphisms.” Nucleic Acids Research 29

353

(17): 88e – 88. https://doi.org/10.1093/nar/29.17.e88.

Yin, Jinhu, Yong Tae Kwon, Alexander Varshavsky, and Weidong Wang. 2004. “RECQL4, Mutated in the Rothmund–Thomson and RAPADILINO Syndromes, Interacts with Ubiquitin Ligases UBR1 and UBR2 of the N-End Rule Pathway.” Human Molecular Genetics 13 (20): 2421–30. https://doi.org/10.1093/hmg/ddh269.

Yu, Qiujing, Christopher J. Carbone, Yuliya V. Katlinskaya, Hui Zheng, Ke Zheng, Mengcheng Luo, P. Jeremy Wang, Roger A. Greenberg, and Serge Y. Fuchs. 2015. “Type I Interferon Controls Propagation of Long Interspersed Element-1.” Journal of Biological Chemistry 290 (16): 10191–99. https://doi.org/10.1074/jbc.M114.612374.

Yuan, Jihong, Dongdong Zhang, Lei Wang, Mengyuan Liu, Jian Mao, Yu Yin, Xiaoying Ye, et al. 2013. “No Evidence for Neo-Oogenesis May Link to Ovarian Senescence in Adult Monkey.” Stem Cells 31 (11): 2538–50. https://doi.org/10.1002/stem.1480.

Zadora, Julianna, Manvendra Singh, Florian Herse, Lukasz Przybyl, Nadine Haase, Michaela Golic, Hong Wa Yung, et al. 2017. “Disturbed Placental Imprinting in Preeclampsia Leads to Altered Expression of DLX5, a Human-Specific Early Trophoblast Marker.” Circulation 136 (19): 1824–39. https://doi.org/10.1161/CIRCULATIONAHA.117.028110.

Zenker, Martin, Julia Mayerle, Markus M Lerch, Andreas Tagariello, Klaus Zerres, Peter R Durie, Matthias Beier, et al. 2005. “Deficiency of UBR1, a Ubiquitin Ligase of the N-End Rule Pathway, Causes Pancreatic Dysfunction, Malformations and Mental Retardation (Johanson-Blizzard Syndrome).” Nature Genetics 37 (12): 1345–50. https://doi.org/10.1038/ng1681.

Zeth, Kornelius, Raimond B. Ravelli, Klaus Paal, Stephen Cusack, Bernd Bukau, and David A. Dougan. 2002. “Structural Analysis of the Adaptor Protein ClpS in Complex with the N- Terminal Domain of ClpA.” Nature Structural Biology 9 (12): 906–11. https://doi.org/10.1038/nsb869.

Zhang, Nenggang, Yunyun Jiang, Qilong Mao, Borries Demeler, Yizhi Jane Tao, and Debananda Pati. 2013. “Characterization of the Interaction between the Cohesin Subunits Rad21 and SA1/2.” Edited by Daniela Cimini. PLoS ONE 8 (7): e69458. https://doi.org/10.1371/journal.pone.0069458.

Zhang, Nenggang, Sergey G Kuznetsov, Shyam K Sharan, Kaiyi Li, Pulivarthi H Rao, and Debananda Pati. 2008. “A Handcuff Model for the Cohesin Complex.” Journal of Cell Biology 183 (6): 1019–31. https://doi.org/10.1083/jcb.200801157.

354

Zhao, Xuesen, Jiarui Li, Cheryl A. Winkler, Ping An, and Ju-Tao Guo. 2019. “IFITM Genes, Variants, and Their Roles in the Control and Pathogenesis of Viral Infections.” Frontiers in Microbiology 9 (January): 3228. https://doi.org/10.3389/fmicb.2018.03228.

Zheng, Qiuyang, Timothy Huang, Lishan Zhang, Ying Zhou, Hong Luo, Huaxi Xu, and Xin Wang. 2016. “Dysregulation of Ubiquitin-Proteasome System in Neurodegenerative Diseases.” Frontiers in Aging Neuroscience. Frontiers Media SA. https://doi.org/10.3389/fnagi.2016.00303.

Zhu, Dongmei, Xia Gong, Liyun Miao, Junshun Fang, and Jian Zhang. 2017. “Efficient Induction of Syncytiotrophoblast Layer II Cells from Trophoblast Stem Cells by Canonical Wnt Signaling Activation.” Stem Cell Reports 9 (6): 2034–49. https://doi.org/10.1016/j.stemcr.2017.10.014.

Zickler, D., and N. Kleckner. 1999. “Meiotic Chromosomes: Integrating Structure and Function.” Annual Review of Genetics 33 (1): 603–754. https://doi.org/10.1146/annurev.genet.33.1.603.

Zimmerman, Shawn W., Young-Joo Yi, Miriam Sutovsky, Fred W. van Leeuwen, Gavin Conant, and Peter Sutovsky. 2014. “Identification and Characterization of RING-Finger Ubiquitin Ligase UBR7 in Mammalian Spermatozoa.” Cell and Tissue Research 356 (1): 261– 78. https://doi.org/10.1007/s00441-014-1808-x.

Zuckerman, S. 1951. “The Number of Oocytes in the Mature Ovary.” Recent Progress in Hormone Research 6: 63–109. https://www.scienceopen.com/document?vid=f70f077e-4cab-46ac- 84be-ea88c8a95950.

355