Mechanisms of Centriole Duplication and Spindle Assembly in Human Cells

Home , CDK5RAP2, Centrin 2, CEP135, CEP152, CEP350, CEP72

Mechanisms of centriole duplication and spindle assembly in human cells

Qianzhu Wu

A thesis submitted in conformity with the requirements for the degree of Doctorate of Philosophy Molecular Genetics University of Toronto

Mechanisms of centriole duplication and spindle assembly in human cells

Qianzhu Wu

Doctorate of Philosophy

Molecular Genetics

University of Toronto

2019

Abstract

The centrosome is the major microtubule-organizing center in animal cells. Centrosomes duplicate once and only once per cell cycle thereby ensuring that two centrosomes are present in mitosis, each of them organizing one of the two spindle poles. Numerical and structural centrosomal abnormities lead to a number of devastating human diseases, including ciliopathies, microcephaly and cancer. Despite extensive effort at elucidating the molecular mechanisms underpinning centriole assembly, our understanding of this process remains incomplete and many novel factors remain to be identified. Here, using a high-throughput, semi-automated centriole duplication screen, I identified TRIM36 as a novel regulator of centriole duplication and mitotic spindle assembly. TRIM36 is a microtubule-binding RING-type E3 ligase. In centriole duplication, TRIM36 regulates daughter centriole assembly in a manner that is dependent on the

RING domain. Using super-resolution microscopy, I discovered that TRIM36 plays a role in the early steps of centriole formation, where it mediates the recruitment of SAS6. TRIM36 depletion

ii also caused severe defects in mitotic spindle assembly and chromosome congression. My results indicated that the dual role of TRIM36 in centriole duplication and spindle formation involves, at least in part, the recruitment of γ-tubulin to centrosomes. Co-depletion of RBM14, a negative regulator of centriole duplication, compensated for the loss of γ-tubulin recruitment and rescued both the centriole duplication and spindle assembly defects. These results provide insights into the critical role of the E3 ligase TRIM36 in centriole duplication, mitotic spindle assembly and cell cycle progression.

iii Acknowledgments

I would like to begin by thanking my supervisor, Dr. Laurence Pelletier. It was truly a wonderful experience working in your lab, with your enthusiasm and guidance, and with a team of talented scientists. I would not go this far without your support and guidance. Thank you!

A big thank you to my committee members, Dr. Brian Raught & Dr. John Brumell. You always ask great and thoughtful questions, providing inspiring and insightful. Your comments have been very much appreciated.

Thank you to all of our collaborators - Brian, Étienne, Anne-Claude, Frank, Nero and Dan - This work would not have been possible without your contributions.

A big shout out to all members of the Pelletier lab, past and present and my friends at the LTRI. You made every day a fun and memorable experience. You have been amazing colleagues, friends and bench-mate (yes talking about you, Suzy). Thank you Sally & Johnny for going through my paper manuscript, which took a huge portion in this thesis. Thank you Ladan for being there in difficult times. Thank you Dave, Bahareh and João for great suggestions. A special thank you to Yi 1, Yi 2 and Gagan for all their support and help in the past years. You always kept me motivated and set me back on track. I will miss the baking days and the lab lunches. Thank you!

Thank you to all my friends, Yiwang & Huayun, Kai & Huijuan, who have been supportive and generous with your time during my Ph.D. The entire experience would be so different without you.

To my family, THANK YOU! Mom, Dad & Jianjin – I don’t think I can ever express how grateful I am for your love and support. I am incredibly fortunate to have you in my life. You all make me feel deeply blessed. Table of Contents

Abstract ...... ii

Acknowledgments ...... iv

Table of Contents ...... v

List of Tables ...... ix

List of Figures ...... x

List of Abbreviations ...... xiv

Chapter 1. Introduction ...... 1

Introduction ...... 2

1.1 The Centrosomes ...... 2

1.1.1 Organization of the human centrosome ...... 2

1.1.2 Centrosome biogenesis ...... 8

1.1.3 Spindle assembly ...... 17

1.1.4 Centrosome structure/numerical abnormalities and diseases ...... 22

1.2 The Ubiquitination System ...... 26

1.2.1 The covalent conjugation of ubiquitin with other cellular proteins regulates diverse eukaryotic cell functions...... 26

1.2.2 Ubiquitination and de-ubiquitination in centriole duplication and spindle assembly ...... 30

1.3 The TRIM E3 ligases ...... 34

1.3.1 Sub-classes of TRIM family ...... 34

1.3.2 TRIM36 ...... 36

1.4 Rationale of the Thesis ...... 37

Chapter 2. Identification of novel centriole duplication regulators ...... 39

Identification of novel centriole duplication regulators ...... 40

2.1 Statement of Contributions ...... 40

2.2 Summary ...... 41

2.3 Introduction ...... 41

2.3.1 BioID ...... 41

2.4 Results ...... 44

2.4.1 Generation of the centriole duplication interactome ...... 44

2.4.2 Functional characterization of centriole duplication interactome and identification of novel centriole duplication regulators ...... 54

2.5 Discussion ...... 66

2.6 Material and Methods ...... 68

2.6.1 Cell lines and Tissue culture ...... 68

2.6.2 Sample preparation for BioID ...... 69

2.6.3 BioID followed by mass spectrometry ...... 69

2.6.4 SAINT analysis ...... 70

2.6.5 U-2 OS S-phase arrested centriole over-duplication assay ...... 70

2.6.6 Statistical methods ...... 71

Chapter 3. TRIM36 is a novel regulator of centriole duplication and microtubule organization factor ...... 72

TRIM36 is a novel regulator of centriole duplication and a microtubule organization factor ...... 73

3.1 Statement of Contributions ...... 73

3.2 Summary ...... 73

3.3 Introduction ...... 74 vi

3.3.1 TRIM36 participates in centrosome biological events ...... 74

3.3.2 Using PLK4-Induced Centriole Over-duplication in combination with Super- Resolution Imaging to Delineate the Centriole Assembly Pathway...... 75

3.4 Results ...... 76

3.4.1 Generation and characterization of TRIM36 antibodies reveal it’s a centrosomal protein...... 76

3.4.2 Proximity-dependent BioID combined mass spectrometry approach maps TRIM36 proximity interactors...... 81

3.4.3 TRIM36 is a novel regulator of centriole duplication...... 85

3.4.4 Depletion of TRIM36 causes microtubule-organization defects...... 105

3.4.5 TRIM36 cooperates with γ-tubulin to regulate centriole assembly and microtubule formation...... 114

3.5 Discussion ...... 128

3.5.1 TRIM36 is a novel positive centriole duplication regulator that cooperates with γ-tubulin to regulate centriole assembly ...... 128

3.5.2 TRIM36 depletion impairs γ-tubulin recruitment to the centrosomes, disrupts spindle formation and causes mitotic defects ...... 130

3.6 Materials and Methods ...... 132

3.6.1 Cell lines and tissue culture ...... 132

3.6.2 Cloning ...... 132

3.6.3 RNAi ...... 133

3.6.4 RNAi rescue of centriole duplication defects ...... 133

3.6.5 Purification of GST-TRIM36 from Sf9 cells ...... 134

vii

3.6.6 PLK4 Induced Centriole Over-Duplication Assays ...... 134

3.6.7 Immunofluorescence microscopy ...... 134

3.6.8 3-D SIM Imaging ...... 135

3.6.9 Live-Imaging ...... 135

3.6.10 Ubiquitin-Activated Interaction Traps (UBAITs)...... 135

3.6.11 Statistical methods ...... 136

3.6.12 Western Blots ...... 136

3.6.13 Co-Immunoprecipitation ...... 136

3.6.14 Antibodies and primers used in the study ...... 137

Chapter 4 Conclusion and Future Directions ...... 140

Conclusion and Future Directions ...... 141

4.1 Identification of novel centriolar components using proximity based BioID profiling .. 141

4.2 TRIM36 in centriole duplication and spindle formation ...... 142

4.3 TRIM36 as an E3 ligase ...... 144

4.4 TRIM36 in human diseases ...... 145

Reference ...... 147

Appendix...... 181

viii

List of Tables

Table 1-1. A brief summary of nomenclature of key centrosome proteins...... 10

Table 1-2. A summary of identified microcephaly proteins, their localization and functions. .... 25

Table 2-1. A list of centriole duplication factors used as baits in this study...... 46

Table 2-2. A representative subset of previously validated centrosomal proteins and PPIs identified in the proximity interaction network...... 50

Table 2-3. A representative subset of previously validated microtubule associated proteins identified in the proximity interaction network...... 53

Table 2-4. Prey candidates selected for the esiRNA centriole duplication screen...... 56

Table 2-5 A subset of MTOC and MT cytoskeleton proteins identified in the proximity interaction network of USP54...... 61

Table 3-1. TRIM36 Proximity Interactors Detected by BioID...... 85

Table 3-2. TRIM36 Interactors Detected by UBAIT...... 117

Table 3-3. Primary Antibodies Used in this Work...... 137

Table 3-4. Secondary Antibodies Used in this Work...... 138

Table 3-5. Primers Used in this Work...... 139

List of Figures

Figure 1-1. An illustration of the centriole cartwheel structure...... 3

Figure 1-2. An illustration of the centrosome structure...... 7

Figure 1-3. The link between centrosome cycle and cell cycle...... 9

Figure 1-4. A schematic illustration of centrosome biogenesis events...... 17

Figure 1-5. An illustration of mitotic spindles...... 22

Figure 1-6. A Schematic illustration of the ubiquitin proteasome system...... 29

Figure 2-1. A schematic overview of BioID method...... 43

Figure 2-2. Examples of expression and biotinylation validation of FLAG-BirA* HEK293 T- REx cells by Western blot and immunofluorescence imaging...... 45

Figure 2-3. The protein proximity interaction map of 13 centriole duplication regulators...... 47

Figure 2-4. U-2 OS S-phase arrested centriole over-duplication assay...... 57

Figure 2-5. Z-score distribution of the esiRNA centriole over-duplication pilot screen...... 59

Figure 2-6 USP54 regulates centriole duplication in a moderate manner...... 60

Figure 2-7 Protein proximity interaction network of FLAG-BirA*-USP54...... 62

Figure 2-8 U-2 OS S-phase arrested centriole over-duplication assay...... 64

Figure 2-9 Z-score distribution of centriole over-duplication screen...... 66

Figure 3-1. Purified TRIM36 was analyzed using Fast protein liquid chromatography (FPLC). 78

Figure 3-2. Purified TRIM36 confirmation by mass spectrometry...... 79

Figure 3-3. Endogenous TRIM36 localizes to centrosomes...... 80

Figure 3-4. BioID of FLAG-BirA*-TRIM36 identifies proximity interactors enriched in centriole duplication pathway and microtubule organization...... 83

Figure 3-5. TRIM36 is required for centriole duplication in HeLa cells...... 86

Figure 3-6. TRIM36 is required for centriole duplication in HeLa cells/ ...... 88

Figure 3-7. TRIM36 is required for centriole over-duplication in U-2 OS S-phase arrested cells...... 89

Figure 3-8. TRIM36 is required for PLK4-induced centriole over-duplication in U-2 OS S-phase arrested cells...... 90

Figure 3-9. Endogenous rescue experiments in the U-2 OS over-duplication assay...... 92

Figure 3-10. N-terminus (Tripartite/RBCC motif with COS box domain) is crucial to centriole duplication regulation, especially the RING domain...... 94

Figure 3-11. TRIM36 catalytic activity is required for centriole duplication in HeLa cells...... 95

Figure 3-12. RING domain from another TRIM family member, TRIM37, is not sufficient in recuing centriole duplication defect caused by TRIM36 depletion...... 97

Figure 3-13. Intact RING domain is crucial to TRIM36’s function in centriole duplication regulation...... 98

Figure 3-14. RING domain comparison among TRIM family proteins...... 99

Sequence alignment of the RING domain of TRIM36 compared with RING domains from other TRIM proteins using Jalview (http://www.jalview.org/). The conservative residues are denoted by highlights (purple). The sequences corresponding to the loop regions are shown in the alignment...... 99

Figure 3-15. TRIM36 is upstream of cartwheel protein SAS6 in centriole assembly pathway. 101

Figure 3-16. The total cellular levels of SAS6 decreased upon TRIM36 depletion in S-phase arrested U-2 OS cells...... 102

Figure 3-17. The total cellular levels of CEP135 remained the same upon TRIM36 depletion in S-phase arrested U-2 OS cells...... 103

Figure 3-18. A schematic representation showing the role of TRIM36 in the centriole assembly pathway...... 104

Figure 3-19. TRIM36 is associated with and required for microtubule organization...... 107

Figure 3-20. TRIM36 is required for mitotic spindle assembly and chromosome congression. 108

Figure 3-21. TRIM36 depletion leads to mitotic arrest in HeLa cells...... 110

Figure 3-22. TRIM36 RING domain is required for microtubule localization...... 111

Figure 3-23. Over-expressed TRIM36 localizes on microtubules in HeLa cells...... 112

Figure 3-24. TRIM36 RBCC domain is crucial for its microtubule localization...... 113

Figure 3-25. Using UBAIT to identify TRIM36 interactors...... 115

Figure 3-26. TUBG1 associates with TRIM36...... 119

Figure 3-27. TRIM36 is required for γ-Tubulin recruitment to the centrosome region in HeLa mitotic cells...... 120

Figure 3-28. TRIM36 is required for γ-Tubulin/TUBG1 recruitment to the centrosome region in HeLa mitotic cells...... 121

Figure 3-29. γ-Tubulin is essential for SAS6 recruitment to the centrosome region...... 122

Figure 3-30. RBM14 depletion restores γ-Tubulin level at the centrosome region after TRIM36 depletion...... 124

Figure 3-31. Co-depletion of RBM14 and TRIM36 restores the γ-Tubulin level and spindle assembly...... 126

Figure 3-32. Co-depletion of RBM14 and TRIM36 restores the γ-Tubulin level and centriole duplication...... 127

xii

Figure 3-33. A schematic representation model for how TRIM36 affects centriole duplication and spindle formation pathway...... 132

xiii

List of Abbreviations

3D-SIM Three-dimensional structured illumination microscopy

APC/C Anaphase-promoting complex/cyclosome

AP-MS Affinity purification followed by mass spectrometry

BioID Biotin identification

Co-IP Co-immunoprecipitation

CDK Cyclin-Dependent-kinase

CRISPR Clustered regularly interspaced short palindromic repeats

DAPI 4', 6-diamidino-2-phenylindole

EM Electron microscopy

GCP γ-Tubulin complex proteins

HU Hydroxyurea

MAP Microtubule-associated protein

MCPH Microcephaly

MT Microtubule

MTOC Microtubule-organizing center

OMX Optical microscopy experiment

PCM Pericentriolar material

PPI Protein protein interaction

RING Really interesting new gene

SAC Spindle assembly checkpoint

SCF Skp1, Cullin, F-box E3 ubiquitin ligase complex

SPB Spindle pole body

TEM Transmission electron microscopy

TRIM Tripartite motif family xiv

UBAIT Ubiquitin-Activated Interaction Traps

UBD Ubiquitin binding domain

γ-TuRC Gamma-tubulin Ring Complex

Chapter 1. Introduction

Introduction 1.1 The Centrosomes

The centrosome is a major microtubule-organizing center (MTOC) in mammalian cells. It plays an important role in a number of fundamental cellular functions such as cell division, cell signaling and motility, intracellular trafficking and cell polarity by organizing an astral array of microtubules during the cell cycle (Pelletier et al., 2006; Pelletier, 2007; Nigg and Raff, 2009; Arquint, Gabryjonczyk and Nigg, 2014). Structurally, the centrosome is composed of two centrioles (a ‘mother centriole’ referring to the centriole formed in the previous duplication cycle, and a ‘daughter centriole’ referring to the centriole formed in the new duplication cycle), surrounded by pericentriolar material (PCM) (Woodruff, Wueseke and Hyman, 2014). Centrioles are also required for the formation of cilia and flagella in quiescent cells where they migrate to the plasma membrane and transform into basal bodies (Gönczy, 2012). They also participate in mitotic spindle assembly to ensure equal segregation of sister chromatids during mitosis (Ault and Rieder, 1994; Zou et al., 1999). Here, I provide more detailed descriptions of centrosome structure, duplication cycle, and contribution to human diseases, as basis for the rationale to gain a better understanding of the biology of centrosomes.

1.1.1 Organization of the human centrosome

The term ‘centrosome’ was first introduced by Theodor Boveri in 1887 (Boveri, 1887). In later studies on living sea urchin, he described the morphology and physiological roles of centrosomes, identifying them as permanent structures during the cell cycle that give rise to the dicentric spindle apparatus (Boveri, 1888). In most organisms, the centrosome is a non- membrane-bound organelle (Woodruff et al., 2017). One exception is the multilayered spindle pole body (SPB) in budding yeast (Saccharomyces cerevisiae) that is embedded in the nuclear envelope (NE) at fusion sites between the inner (INM) and outer (ONM) nuclear membranes (Byers and Goetsch, 1975; Kupke, Malsam and Schiebel, 2017). The SPB is the budding yeast equivalent of a human centrosome. A set of four proteins (Mps2, Bbp1, Ndc1, and Nbp1), termed the SPB insertion network (SPIN), collectively contribute to the insertion of the SPB into the NE (Kupke, Malsam and Schiebel, 2017). In human cells, the non-membrane-bound 2

centrosome contains microtubule-based barrel-shaped centrioles surrounded by pericentriolar material (PCM) (Woodruff, Wueseke and Hyman, 2014). The mature human centriole is a ~450 nm long and ~250 nm wide cylindrical structure (Gönczy, 2012), primarily composed of 9-fold symmetric microtubule triplets referred to as “A”, “B” and “C” (Gönczy, 2012; Nigg and Holland, 2018). The centriole diameter and symmetry are defined by a scaffold structure located at the proximal end of the procentriole known as ‘the cartwheel’ (van Breugel et al., 2011). The size of the cartwheel dictates the diameter of the centriole barrel (Marshall, 2016). In vitro experiments involving protein structures have confirmed that Spindle assembly abnormal protein 6 (SAS6) forms 9 homodimers through its N-terminal domain and self-assembles into the cartwheel hub (Kitagawa et al., 2011a; van Breugel et al., 2011), from which radiates 9 spokes connecting to the “A” tubule. In fully mature centrioles, the “C” microtubule only reaches half way up the symmetry (Nigg and Holland, 2018) (Figure 1-1).

A B

C B A SAS6 dimer

Pinhead

Figure 1-1. An illustration of the centriole cartwheel structure. (A) Electron micrographs of the proximal end of the centriole taken from (Winey and O’Toole, 2014a), showing the microtubule triplets having an anticlockwise twist. Scale bar: 100 nm. (B) Schematic cross-sectional view of the cartwheel structure, showing the pinheads (red) linking SAS6 homodimers (purple) and microtubule triplet (green) together. The microtubule triplets are termed “A”, “B” and “C” (Gönczy, 2012; Hirono, 2014; van Breugel et al., 2014; Winey and O’Toole, 2014a).

Unlike the cartwheels of the proximal lumen, the structure and functions of the distal lumen are much less understood. Cryoelectron tomography revealed that the distal lumen harbors a 45 nm scaffold formed by periodic stack of rings (Ibrahim et al., 2009). The lumen contains the centriole duplication regulator Centrin (Salisbury et al., 2002) and its interacting protein POC5, which has a critical role in centriole elongation and will be discussed in section 1.1.2 (Azimzadeh et al., 2009a). Upon reaching maturation, the centriole harbors 9 centriolar appendages at its distal and subdistal ends that are required for anchoring cytoskeletal microtubules and membrane docking during ciliogenesis (Tanos et al., 2013; Winey and O’Toole, 2014b). NIN, CEP164, CEP83 and FOP are among the identified centriolar appendage proteins that have key roles in regulating ciliogenesis (Mojarad et al., 2017). A previous work showed that the CEP19/CEP350/FOP module co-localizes to the basal bodies at the distal end of appendages, while a study by Wang et al. reported that the CEP350/FOP complex is localized to the sub-distal appendage (Mojarad et al., 2017; Wang et al., 2018). These studies provided important insights into the structure and functions of the distal ends of centrioles.

In human cells, a tether structure containing CEP250 (C-NAP1), LRRC45 and Rootletin (encoded by Crocc) links the two centrioles together, and is dissolved at the onset of mitosis to enable centrosome separation (Fry et al., 1998; He et al., 2013; Prosser et al., 2015; Nigg and Holland, 2018). The separation is facilitated by NEK2 kinase mediated phosphorylation of C- NAP1, Rootletin and LRRC45 (Fry et al., 1998; Bahe et al., 2005; He et al., 2013; Prosser et al., 2015; Nigg and Holland, 2018). Suzanna L. Prosser and colleagues identified NEK5, a protein in the NIMA-related protein kinases family, as a component of the proximal end of centrioles (Prosser et al., 2015). They demonstrated that NEK5 is essential for centrosome linker organization and proper centrosome separation (Prosser et al., 2015). In summary, both NEK2 and NEK5 are important for the timely separation of centrosomes before mitosis to ensure genomic stability (Fry et al., 1998; Prosser et al., 2015).

The highly structured centrioles are embedded in an electron-dense, protein rich PCM, which is recruited and organized by the mother centriole and is crucial for centrosome’s function as a MTOC (Woodruff, Wueseke and Hyman, 2014). Sub-diffraction microscopy provided important insights into the dynamic and hierarchical structure of PCM. These studies showed that the PCM protein components occupy separate domains and exhibit layered organizations in 4

a ring-shaped manner around the proximal end of the mother centriole during interphase (Lawo et al., 2012a; Mennella et al., 2012; Sonnen et al., 2012). (Lawo et al., 2012b). Super-resolution imaging was made possible by three dimensional structured illumination microscopy (3D SIM), a method that generates a 3 dimensional reconstruction with doubled spatial resolution compared to conventional wide-field microscopy (±120 nm in the x-,y-direction and ±300 nm in the z- direction) (Gustafsson et al., 2008). In this method, three mutually coherent light beams are generated by a diffraction grating from a slightly spatially incoherent beam that interfere with the sample with each creating an intensity interference pattern with both axial and lateral structures (Gustafsson et al., 2008). Through spatial frequency mixing, high-resolution information are encoded and recorded into the observed images based on spatially structured excitation intensity (Gustafsson et al., 2008). The method has demonstrated to have high efficacy in the imaging of centrosome structures (Lawo et al., 2012a; Mennella et al., 2012; Sonnen et al., 2012). Several core PCM components, such as PCNT (Pericentrin), CDK5RAP2, CEP192 and CEP152, were found to be involved in the construction of the inner scaffold for recruiting and loading of γ- Tubulin, resulting in PCM expansion and maturation at mitosis. Dr. Lawo et al. observed that the C-terminal PACT domain of PCNT facilitates the targeting of PCNT to the centrioles, whereas PCNT still labels the outermost layer of centrosomes during interphase (Lawo et al., 2012a). By measuring the toroid diameters detected using antibodies specific to different PCNT epitopes, the team found that the C-terminal of PCNT localizes near centrioles while its N-terminal extends into the centrosome periphery, adopting an elongated conformation that spans ~100 nm wide (Lawo et al., 2012a). By phosphorylating NEDD1, PLK1 controls the recruitment of PCNT, which acts as a scaffold for the recruitment of CDK5RAP2 (also known as CEP215) and other PCM proteins including CEP192 (Haren, Stearns and Lüders, 2009; Lee and Rhee, 2011; Lawo et al., 2012a). Subsequently, γ-TuRC is recruited to the PCM (Haren, Stearns and Lüders, 2009).

In earlier studies, electron microscopy identified electron-dense spherical granules around the centrosomes that have diameters ranging from 70-100 nm in multiple vertebrate cells (Kubo et al., 1999a). The granules were termed ‘centriolar satellites’ (Kubo et al., 1999a). The recruitment of PCM proteins is highly dependent on microtubules. PCM1, a 228-kD protein that localizes to small granules around centrioles, was the first PCM component to be identified (Kubo et al., 1999a). In follow up studies, PCM1 is commonly used as a localization marker to

identify, define, and evaluate new satellite proteins (Hori and Toda, 2017). Satellite proteins are defined as proteins that co-localize with PCM1, or proteins de-localized from the centrosomes upon PCM1 depletion (Lee and Stearns, 2013; Firat-Karalar et al., 2014; Klinger et al., 2014; Hori and Toda, 2017). To date, more than 100 satellite proteins have been identified, with many associating with centrosomes (Kubo et al., 1999a; Klinger et al., 2014; Hori and Toda, 2017). The centriolar satellite mainly participates in four cellular functions: centriole duplication, centrosome maturation, ciliogenesis, and microtubule organization (Kubo et al., 1999b; Dammermann and Merdes, 2002). While PCM1 serves as a scaffold for satellites, proteins including CEP290 (Valente et al., 2006; Klinger et al., 2014), BBS4 (Lee and Stearns, 2013), OFD1 (Tang et al., 2013), and CEP63 (Firat-Karalar et al., 2014) all contribute to the maintenance of the pericentriolar localization patterns of satellites (Hori and Toda, 2017). Satellites largely depend on microtubule structures to move around the centrosome. Depletion of microtubule machinery proteins, such as ANK2, MAPT, and MAP7D3, disrupts satellite localization (Gupta, Coyaud, et al., 2015) (Figure 1-2).

PCM Mother Centriole Distal Appendage Sub-distal Appendage

γ-TuRC Tether Satellite

Daughter Centriole

Figure 1-2. An illustration of the centrosome structure. (A) Electron micrographs of centrioles showing longitudinal section view of the mother and daughter centrioles with cross sections highlighting the distal and subdistal appendages. Reprinted from (Winey and O’Toole, 2014a) based on original images in (Paintrand et al., 1992). (B)The mother centriole and the daughter centriole are linked by a tether structure (Nigg and

Holland, 2018), surrounded by the PCM (yellow) (Woodruff, Wueseke and Hyman, 2014). Distal and subdistal appendages decorate the distal end of the mother centriole (Tanos et al., 2013; Winey and O’Toole, 2014b). The γ-TuRC (orange) is recruited to the PCM during centrosome maturation and conducts microtubule organizing functions (Zhu et al., 2008; Haren, Stearns and Lüders, 2009). Satellites are shown along the microtubules around the centrioles (Kubo et al., 1999b; Hori and Toda, 2017).

In addition to vertebrates, the centrosome also functions as the main MTOC in many organisms, including higher fungi, amoebozoa, and several other eukaryotes (Bornens and Azimzadeh, 2007; Azimzadeh, 2014). While the overall structure of the centrosome is conserved, there are variations in size, microtubule arrangements, subcellular localizations and functions. The centrosomes of Caenorhabditis elegans and Drosophila are composed of 9-fold symmetric doublet and singlet microtubules rather than the triplet structures found in mammalian centrosomes (Pelletier et al., 2006). Another key difference is that centrioles of Caenorhabditis elegans and Drosophila lack subdistal appendages (Azimzadeh, 2014). In Saccharomyces cerevisiae, the functionally equivalent organelles to mammalian centrosomes are called spindle pole bodies (SPB), a multi-layered cylindrical shaped organelle embedded in the nuclear envelope (Marshall, 2009; Azimzadeh, 2014). In some lineages of Amoebozoa, for example D. discoideum, the centrosome-like structure is referred to as a nuclear-associated body (NAB), in which a three-layered core is embedded in a cloud of amorphous material similar to pericentriolar material in mammalian centrosomes (Azimzadeh, 2014).

1.1.2 Centrosome biogenesis

Centrosome biogenesis is a tightly regulated process during proliferation to coordinate the different cellular functions involving this organelle (Pelletier, 2007). The centrosome duplicates only once during the cell cycle in a template based manner (Tsou and Stearns, 2006; Godinho and Pellman, 2014; Marshall, 2016) (Figure 1-3).

Centriole Separation M Centrosome G1 Maturation

Procentriole S Formation

Centriole Elongation

Figure 1-3. The link between centrosome cycle and cell cycle. The centrosome is duplicated once and only once during the cell cycle. The process includes two key steps: centriole duplication and centrosome maturation. Centriole duplication is initiated at G1/S transition, and the centrosome is matured at the onset of mitosis for the subsequent spindle formation.

The process of centriole duplication is relatively conserved across species. In Caenorhabditis elegans, this was elucidated by studying the events of the first mitotic division of embryos. SPD-2 initiates centriole duplication by recruiting ZYG-1 to the mother centrioles (Kemp et al., 2004; Pelletier et al., 2006). Subsequently, the two protiens promote the recruitment of centriole components SAS-5 and SAS-6, followed by the recruitment of SAS-4 (Delattre et al., 2004; Pelletier et al., 2006). In Drosophila, Asterless (Asl) plays an essential role in both centriole duplication initiation and mitotic PCM recruitment (Bonaccorsi, Giansanti and Gatti, 1998; Varmark et al., 2007; Blachon et al., 2008; Dzhindzhev et al., 2010). At the

beginning of centriole duplication, Asl recruits SAK to the mother centrioles (Dzhindzhev et al., 2010; Novak et al., 2014). SAK drives centriole duplication initiation by phosphorylating Ana2, leading to the loading of SAS6 (Kim, Fong and Bryan Tsou, 2014). At the onset of mitosis, newly formed daughter centrioles are converted and gain the ability to recruit PCM and duplicate (Fu et al., 2016). This process involves the recruitment of Asl to the daughter centrioles prior to mitosis by SAS-4 and Ana1, which itself is recruited by Bld10, a key protein important in maintaining centriole structural integrity (Roque et al., 2012; Hirono, 2014; Fu et al., 2016).

Human Drosophila C.elegans CEP192 Spd2 SPD-2 CEP152 Asterless - PLK4 SAK ZYG-1 SAS6 SAS-6 SAS-6 STIL Ana-2 SAS-5 CENPJ Sas-4 SAS-4 CEP135 Bld10 - CEP295 Ana1 - CDK5RAP2 Cnn SPD-5 PCNT D-PLP -

Table 1-1. A brief summary of nomenclature of key centrosome proteins.

In mammalian cells, the initiation of centriole duplication occurs at G1/S transition and is facilitated by Polo-like kinase 4 (PLK4), (SAK in Drosophila, ZYG-1 in Caenorhabditis elegans) (Habedanck et al., 2005). Unlike other PLK family proteins that typically contains 2 polo-box (PB) motifs, PLK4 has 3 PB motifs with the first two having a crucial role in dimerization and subsequent recruitment to the centrioles (Habedanck et al., 2005). Following dimerization, PLK4 is auto-phosphorylated at its kinase activation loop and destruction motif at its Linker 1 (L1) region. Phosphorylation of the kinase activation loop resulted in activation of kinase activity (Habedanck et al., 2005). Phosphorylation of the destruction motif directs SCF β- 10

TrCP targeting to PLK4 to facilitate proteasome dependent protein degradation (Habedanck et al., 2005; Arquint and Nigg, 2016).

In a species-dependent manner, PLK4 is recruited to the centrioles by either or both of scaffold proteins CEP152 (Asl in Drosophila) and CEP192 (Spd2 in Drosophila, SPD-2 in Caenorhabditis elegans) (Sonnen et al., 2013; T.-S. Kim et al., 2013). In Caenorhabditis elegans, the recruitment of ZYG-1 solely depends on SPD-2, whereas in Drosophila, Asl also participates in PLK4 recruitment (Pelletier et al., 2006; Dzhindzhev et al., 2010). Curiously, Klebba et al. described opposing roles for PLK4 kinase activity at two PLK4 binding sites for Asl (Asl-A and Asl-C) (Klebba et al., 2015). The N-terminal domain Asl-A promotes PLK4 homo-dimerization and auto-phosphorylation in interphase cells, resulting in its activation (Klebba et al., 2015). In contrast, the C-terminal Asl-C region stabilizes PLK4 during mitosis at its peak and thereby targeting PLK4 to centrioles (Klebba et al., 2015). As a result, both over- expression or depletion of Asl can stabilize PLK4 in Drosophila (Klebba et al., 2015). In a recent study, Boese et al. showed that PLK4 phosphorylates Asl at a number of sites in Drosophila (Boese et al., 2018). They specifically studied 13 phospho-sites within the Asl-A region, and identified a phosphomimetic mutant, Asl-A-13A, which inhibited centriole duplication through suppressing PLK4 catalytic activity (Boese et al., 2018). However, the duplication defect can be masked by Asl-C, which promotes centriole amplification (Boese et al., 2018). The dual function of Asl shows how key centriole proteins are strictly regulated during the cell cycle. Similarly in human cells, CEP152 localization to the centrioles is dependent on CEP192, and the two bind to the PB motifs of PLK4 in a competitive fashion (T.- S. Kim et al., 2013). Altogether, the interactions with PLK4 facilitate PLK4 recruitment (Sonnen et al., 2013).

Another important substrate of PLK4 is the SCL/TAL1 interrupting locus (STIL) protein, which is the human homolog of Caenorhabditis elegans SAS-5 and Drosophila Ana-2. STIL has been reported to be mutated in primary microcephaly (Arquint et al., 2015; Arquint and Nigg, 2016). The coiled-coil region of STIL interacts with the PB3 and L1 domains of PLK4, and this is pivotal for STIL localization to the centrioles from late G1 phase to metaphase (Ohta et al., 2014a; Arquint et al., 2015; Moyer et al., 2015). Dzhindzhev et al. discovered that in Drosophila, Ana-2 is phosphorylated by PLK4 on S38 located in a conserved region of Ana-2, 11

and that this is necessary for Ana-2 recruitment to the pro-centrioles (Dzhindzhev et al., 2017). Subsequently, PLK4 phosphorylates Ana-2 at 4 serine residues (S318, S365, S370, S373) within the STAN domain, a motif that defines Ana-2 orthologs (Dzhindzhev et al., 2014). This step results in the recruitment of SAS6 to the pro-centrioles (Dzhindzhev et al., 2014). Similarly in mammalian cells, PLK4 phosphorylates STIL at S1108 and S1116 within its STAN domain, promoting SAS6 binding to STIL and enabling cartwheel formation, a crucial step in centriole biogenesis (Arquint et al., 2015).

Interestingly, PLK4 starts off as a uniform ring around the mother centriole, and concentrates into a single spot where the pro-centriole generates (Kim et al., 2013; Ohta et al., 2014b). However, PLK4 localization does not affect the recruitment of STIL, which is loaded at a single site in a manner that is dependent on active PLK4 (Dzhindzhev et al., 2017). Recently, Leda et al. proposed a mechanistic biophysical model in silico that predicts PLK4 re-positioning from a ring to a spot is driven by both auto-phosphorylation and phosphorylation of STIL, with the latter functioning to retain PLK4 at the centriole surface (Leda, Holland and Goryachev, 2018). The resulting PLK4-STIL complex at that specific spot is thus crucial for centriole biogenesis (Leda, Holland and Goryachev, 2018). This is in line with the discovery made by Yi.L from the Pelletier lab, showing that STIL promotes PLK4 activity through its direct binding with CEP85 (Liu et al., 2018). CEP85 is a novel centriole duplication regulator discovered by Yi that generate a positive feedback loop that ensures full activation of PLK4 (Liu et al., 2018).

After recruitment to the pro-centrioles, SAS6 oligomerizes into the 9-fold symmetrical cartwheel structure (Kitagawa et al., 2011b). Although it is widely accepted that SAS6 is recruited by phosphorylated STIL, the underlying mechanisms remain unknown. Fong et al. proposed a model in which SAS6 is first recruited to the proximal lumen of the mother centrioles, where the cartwheel structure is formed (Fong et al., 2014). Next, the established cartwheel dissociates from the lumen and is transferred to the outside wall in a PLK4 and STIL dependent manner to initiate pro-centriole generation (Fong et al., 2014). The interactions among PLK4, STIL and SAS6 set the stage for a tightly regulated step-wise biogenesis of centrioles.

Following the formation of the cartwheel, STIL interacts with Centrosome Protein J (CENPJ, also known as CPAP, Centrosomal P4.1-Associated Protein, homolog of SAS-4 in

Caenorhabditis elegans), which promotes centriole elongation at the proximal end (Tang et al., 2011a). Similar to STIL, CENPJ is encoded by a microcephaly gene and is localized to the centrosome (Tang et al., 2011b). A homozygous missense mutation (E1235V) in CENPJ was found in patients with microcephaly that impaired CENPJ binding to STIL and localization to the centrioles (Tang et al., 2011a). It has been reported that over-expression of CENPJ produces elongated centrioles (Schmidt et al., 2009; Y.-N. Lin et al., 2013). Next, Dr. David Comartin from the Pelletier lab showed that CENPJ controls centriole length through interaction with CEP120 and SPICE1 (Spindle and Centriole-associated protein 1, previously known as CCDC52) (Archinti et al., 2010; Comartin, Gagan D. Gupta, et al., 2013). Similar to CENPJ, CEP120 over-expression also leads to centriole elongation while depletion of CENPJ/CEP120/SPICE1 reduced centriole length. BioID, a proximity-dependent biotin identification followed by mass spectrometry, identified potential interactors with CEP120 (Roux et al., 2012; Comartin, Gagan D. Gupta, et al., 2013; Gupta, Coyaud, et al., 2015), including Ankyrin 2 (ANK2), which interacts exclusively with CEP120 among known centrosome proteins (Gupta, Coyaud, et al., 2015). ANK2 depletion inhibited centriole duplication and over-duplication induced by PLK4 over-expression, and was shown to be critical for CEP120 localization to the centrosomes (Gupta, Coyaud, et al., 2015).

Using super-resolution microscopy, it was confirmed that the centriole protein CEP135 is dependent on CENPJ/CEP120/SPICE1 for localization to the procentriole assembly region (PCAR), a PLK4-rich area where new centrioles begin to form (Comartin, Gagan D. Gupta, et al., 2013). CEP135 is encoded by another human microcephaly gene, and is the homolog of Bld10 discovered in Chlamydomonas reinhardtii, where it was shown to be indispensible for cartwheel formation (Y.-C. Lin et al., 2013). Lin and colleagues also demonstrated an interaction between CEP135 and SAS6, suggesting a role for CEP135 in linking the cartwheel structure to the microtubules (Y.-C. Lin et al., 2013). At the later stage of centriole elongation, the conserved WD40 domain protein POC1 (Proteome of Centriole protein 1), POC5 and CEP295 (also known as KIAA1731, homolog of Drosophila Ana 1) have been identified as centriole length modulators at the distal end (Azimzadeh et al., 2009b; Keller et al., 2009; Chang et al., 2016). On binding to CEP135 and CEP152, CEP295 localizes to the microtubule wall outside of the cartwheel structure at the proximal end of the centriole, and this is essential for POC1 and POC5

recruitment (Chang et al., 2016). POC5 is recruited to the daughter centrioles during G2/M before centrosome maturation, and is involved in the building of the distal ends of daughter centrioles (Azimzadeh et al., 2009b).

Centrioles have consistent diameters and lengths, a result of careful centriole length control. In contrast to the previously described positive centriole elongation regulators, CP110 was found to localize to the distal ends of centrioles, from where it is removed prior to axoneme formation during ciliogenesis (Schmidt et al., 2009). Consistent with this observation, CP110 depletion promotes centriolar microtubule elongation (Schmidt et al., 2009). CEP97 is an interacting partner with CP110 (Spektor et al., 2007) and is required for recruiting CP110 to the centrioles, and its depletion leads to abnormal mitotic spindles and disrupted cilia assembly (Spektor et al., 2007). Additionally, the centriolar kinesin KIF24 interacts with both CEP97 and CP110, and removal of KIF24 resulted in the loss of CP110 decorating mother centrioles (Kobayashi et al., 2011), providing further insights into the mechanism of centriole length control. In summary, a number of proteins including CEP120, CENPJ, CEP135, POC1, POC5, CEP97 and CP110 are involved in maintaining the length of pro-centrioles during centriole assembly.

Prior to entering mitosis, PCM accumulates around the mother centriole and the newly formed daughter centrioles. This is pivotal for centrosome maturation and spindle assembly, where the γ-Tubulin ring complex (γ-TuRC) promotes microtubule nucleation. Polo-like Kinase 1 (PLK1) initiates centrosome maturation by phosphorylating Pericentrin (PCNT) at S1235 and S1241 residues (Lee and Rhee, 2011). The phosphorylation of PCNT enables centrosomal recruitment of key PCM proteins such as CEP192 and Aurora A (Lee and Rhee, 2011). CEP192 is a key substrate for both PLK1 and Aurora A kinases, and its phosphorylation provides binding sites for the γ-Tubulin ring complex (γ-TuRC) (Joukov, Walter and De Nicolo, 2014). Additionally, Neural Precursor Cell Expressed, Developmentally Down-Regulated 1 (NEDD1) is also required for γ-TuRC targeting to the centrosomes (Haren et al., 2006). They observed mitotic spindle defects in NEDD1 depleted cells. The depletion of NEDD1 and γ-Tubulin, a major component of γ-TuRC, resulted in centriole duplication failure (Haren et al., 2006). In addition, LATS2, encoded by the human homolog of tumor suppressor protein warts/lats in Drosophila, localizes to the centrosomes and together with its interacting partner AJUBA, 14

regulate γ-tubulin recruitment to centrosomes and spindle organization during mitosis (Abe et al., 2006). In summary, during centrosome maturation, the centrosomal recruitment of γ-Tubulin ring complex (γ-TuRC) depends on a large number of proteins, including PCNT, CEP192, CDK5RAP2, NEDD1, LATS2 and Ajuba, in preparation for subsequent mitotic spindle assembly (Figure 1-4).

NEK2 Centriole PLK1 M Separation

CEP152 CEP192 PLK4 P Cartwheel G1/S STIL PLK4 SAS6 Formation

CENPJ CEP120 CP110 SPICE1

CEP135 S POC1A/1B/5 Centriole γ-TuRC α/β-Tubulin Elongation

Pericentrin NEDD1 CEP192 CDK5RAP2 G2/M Centrosome γ-TuRC Maturation

Centrosome M Separation

Figure 1-4. A schematic illustration of centrosome biogenesis events. Centriole biogenesis is spatial-temporally aligned with the cell cycle (Pelletier, 2007; Nigg and Holland, 2018), and the stages of the cell cycle are shown on the left. The fibrous linker between mother and daughter centrioles is dissolved during centriole separation, and the pro-centriole biogenesis is initiated during G1/S transition. In late G2, the mother and grandmother centrioles lose their tether linker, leading to centrosome separation to form spindle poles individually (Bahe et al., 2005; Gönczy, 2012).

1.1.3 Spindle assembly

Following its full maturation, the centrosome locates to the mitotic spindle pole and serves as the major microtubule organizing center (MTOC) during mitotic spindle assembly (Prosser and Pelletier, 2017). First described by Walther Flemming over 150 years ago, the mitotic spindle has a main structure composed of microtubules (MTs) with diameters of 250 Å (Paweletz, 2001; Kollman et al., 2011). MTs are formed by 13 polar filaments of α- and β- tubulin heterodimers, with β-Tubulin oriented towards the plus end and α-Tubulin towards the minus end (Tilney et al., 1973; Nogales et al., 1999). MT assembly is dependent on GTPs bound by α- and β-tubulins (Cassimeris et al., 1987; Erickson and O’Brien, 1992).

MTs nucleate from the MTOCs with their minus ends embedded in the organelle. Components of the MTOCs from different organisms include centrosome, spindle microtubules, kinetochores, nuclear envelope, and the Golgi apparatus (Godinho and Pellman, 2014; Wu and Akhmanova, 2017). MTs serve as a structural support for cell shape determination, and road tracks for motor-driven intracellular transport (Moritz et al., 1995). The MT nucleation function of the centrosome is mainly driven by centrosomal recruitment of γ-Tubulin ring complex (γ- TuRC). The TuRC contains γ-Tubulin and γ-Tubulin complex proteins (GCP) that serve as an assembly template for microtubule formation in eukaryotic cells (Wise, Krahe and Oakley, 2000; Kollman et al., 2011).

Based on extending directions and binding destinations, MTs in the bipolar spindle array extending from the centrosomes are categorized into three subsets: Kinetochore MTs (K-MTs),

Non-kinetochore MTs (nK-MTs), and Astral MTs (A-MTs) (Dumont and Mitchison, 2009). The nK-MTs are also referred to as interpolar microtubules, and they are potentially involved in keeping the spindle poles apart and ensuring spindle integrity (Dumont and Mitchison, 2009). The nK-MTs drive spindle bipolarity through a process termed MT sliding where nk-MTs are cross-linked by Kinesin 11 (Eg5) and Kinesin 14 at the overlapping zone of two half spindles (Slangy et al., 1995; Kapitein et al., 2005; Lüdecke et al., 2018). The A-MTs have their plus ends extending to the cellular cortex, and they help to secure spindle placement by balancing the forces applied on the MTs (Grill et al., 2003). The plus ends of K-MTs are embedded in the kinetochores of sister chromatids, and the embedded K-MTs are referred to as K-fibers (Dumont and Mitchison, 2009). Embedding is a pivotal process that connects chromosomes to the spindle structure to mediate the bipolar pulling of chromosomes (Dumont and Mitchison, 2009) (Figure 1-5A).

The appropriate embedment of plus-end MTs with kinetochores at centromeres of paired sister chromatids is essential for proper chromosome segregation (Heald and Khodjakov, 2015). The searching-binding-breaking embedment process, referred to as “search and capture”, continues until all attachments are stabilized and chromosomes are bi-oriented (Kirschner and Mitchison, 1986; Cassimeris et al., 1987; Heald and Khodjakov, 2015). Unattached kinetochores trigger the spindle assembly checkpoint (SAC), which becomes silenced once proper attachments of all kinetochores have been achieved (Sacristan and Kops, 2014). The ubiquitin ligase anaphase-promoting complex/cyclosome (APC/C) is activated as the mitotic process progresses. APC/C triggers mitotic exit by inducing poly-ubiquitination and proteasome-dependent degradation of Cyclin B and Securin, rendering it a critical target for the mitotic checkpoint complex (MCC) at the SAC (Primorac and Musacchio, 2013; Sacristan and Kops, 2014).

The “search and capture” is mediated by the dynamics of the tubulin dimers, which are added or removed to allow MTs to shrink or grow (Heald and Khodjakov, 2015). The stochastic switching is driven by GTP hydrolysis at the β-tubulin side of the heterodimer. The MT end capped by β-tubulin is defined as the “plus-end” (Erickson and O’Brien, 1992; Alushin et al., 2014). The GDP-tubulin heterodimer is more prone to depolymerization, a process often referred to as “catastrophe” (Desai and Mitchison, 1997).

The dynamic instability of MTs is essential for controlling MT length, and the precise regulation of MT length is critical for determining how spindles are oriented and positioned in cells (Pearson and Bloom, 2004; Siller and Doe, 2009). In the budding yeast Saccharomyces cerevisiae, the mitotic spindle is aligned with the cell polarity axis, and this is determined by activation of the Rho GTPase CDC42 (David Pruyne and Anthony Bretscher, 2000). As an initial step, the plus-ends of A-MTs are transported along the actin cables, positioning one SPB at the bud neck division (David Pruyne and Anthony Bretscher, 2000; Siller and Doe, 2009). The bud neck is defined by accumulation of Septin proteins and will later serve as separation plane during division (David Pruyne and Anthony Bretscher, 2000; Siller and Doe, 2009). This process is dependent on two microtubule plus-end proteins, Kar9 and Bim1 (Siller and Doe, 2009). In the next step, A-MTs shorten and attach to the cortex in a dynein- and dynactin- dependent manner, thereby precisely positioning the pre-anaphase spindle (Adames and Cooper, 2000).

Positioning accuracy is highly dependent on SPB asymmetry as such A-MT only pulls one of the two SPBs to the bud neck. This is achieved by the specific localization to daughter SPBs of Kar9 and dynein (Liakopoulos et al., 2003; Pearson and Bloom, 2004; Grava et al., 2006). The spindle positioning determines the ingression site during cytokinesis and subsequently determines the fates of the two daughter cells, and this is especially important in neurons and stem cells (Siller and Doe, 2009; Pelletier and Yamashita, 2012).

A large number of proteins take on different roles to ensure proper spindle formation. Goshima et al. showed that the Dgt complex, which they named “augmin”, is essential for γ- Tubulin recruitment to the spindle microtubules in Drosophila (Goshima et al., 2008). In the following year, an 8-subunit human augmin complex (HAUS) was identified to be essential for spindle and centrosome integrity (Lawo et al., 2009). The HAUS subunits localize to spindle poles and microtubules, and regulate mitotic spindle assembly by opposing NUMA (Nuclear mitotic apparatus protein). Depletion of the HAUS complex resulted in centrosome fragmentation (Lawo et al., 2009). Motor proteins, including dyneins and kinesins, also participate in spindle assembly by driving centrosome separation or trans-locating the minus ends generated within the spindles to the spindle poles (Grill et al., 2003; Brugués et al., 2012; Prosser and Pelletier, 2017). SPICE1 also participates in spindle assembly, and was first characterized by Marco Archinti et al. from the Luders lab as a protein that localizes to centrioles 19

and mitotic spindle microtubules (Archinti et al., 2010). SPICE1 depletion leads to severe spindle assembly defects, resulting in monopolar spindles and multipolar spindles (Archinti et al., 2010) (Figure 1-5B).

In the absence of centrosomes, other MTOCs take over to maintain the structure and function of mitotic spindles through various mechanisms (Khodjakov et al., 2000; Karsenti and Vernos, 2001; Gadde and Heald, 2004; Mahoney et al., 2006). For example, chromatin is able to create a local environment enabling the nucleation and organization of spindle MTs (Kalab, Weis and Heald, 2002; Gruss and Vernos, 2004). The GTPase Ran binds to importin-β and induces the release of cargos including the spindle assembly factors (SAFs), promoting MT nucleation around the chromatins (Kalab, Weis and Heald, 2002; Gruss and Vernos, 2004). MTs can also mediate microtubule nucleation (Mahoney et al., 2006). The aforementioned Augmin complex in Drosophila can recruit γ-TuRC to the proximity of an existing spindle MT. Time-lapse imaging demonstrated that with Dgt depletion, monopolar spindles cannot convert to bipolar spindles, indicating Dgt depletion impairs chromosome-mediated MT nucleation (Goshima et al., 2008).

A-MT

nK-MT

K-MT

Figure 1-5. An illustration of mitotic spindles. (A). There are three types of microtubules in the spindle array: kinetochore microtubules (K- MTs), attaching the chromosomes to the spindle poles and pulling the sister chromatids apart (Grill et al., 2003; Dumont and Mitchison, 2009); non-kinetochore microtubules (nK-MTs), extending to the opposing poles, providing stability to the spindle (Grill et al., 2003; Dumont and Mitchison, 2009); and astral microtubules (A-MTs), with their plus ends extended to the cellular cortex, securing spindle placement through balancing the forces applied on the microtubules (Grill et al., 2003; Dumont and Mitchison, 2009). (B). Supernumerary centrosomes tend to form multipolar spindles, leading to the failure of K-MT attachment to the kinetochores (Chan, 2011; Nigg and Holland, 2018)

1.1.4 Centrosome structure/numerical abnormalities and diseases

The potential association between centrosome/spindle abnormalities and human diseases has been discussed throughout the years. In 1902, Dr. Theodor Boveri first proposed that centriole over-amplification and aneuploidy may contribute to human tumorigenesis (Boveri, 2008), and this is commonly describeded as the “Boveri hypothesis” (Godinho and Pellman, 2014). Gino Galeotti and David von Hansemann, Boveri’s contemporaries, observed that abnormal mitoses are common features of cancer cells (Godinho and Pellman, 2014). In more recent years, a series of clinical studies have focused on the underlying mechanisms of centrosome amplification in human solid and haematological tumors (Godinho and Pellman, 2014). In p53 deficient epidermis, centrosome amplification induced by PLK4 over-expression accelerated tumorigenesis (Serçin et al., 2016). Centrosome defects in human cancers are categorized into structural or numerical abnormalities (Godinho and Pellman, 2014; Nigg and Holland, 2018). Structural defects, including changes in centrosome shape, size, position and composition (Godinho and Pellman, 2014), are related to deregulation of protein levels or activation of centrosomal proteins (Schnerch and Nigg, 2016; Nigg and Holland, 2018). Numerical defects reflect changes in centrosome copy numbers in human cells, a potential result of deregulation of the centriole duplication cycle (Duensing et al., 2003).

Aneuploidy is usually a consequence of abnormal centrosome amplification that can lead to the formation of multipolar or mono-polar mitotic spindles (Nigg and Holland, 2018). As discussed in previous sections, the SACs are triggered upon sensing of unattached kinetochores to arrest cells in mitosis, potentially leading to cell death (Sacristan and Kops, 2014). When multipolar spindles are formed, merotelic K-MT attachments arise (Godinho and Pellman, 2014; Sacristan and Kops, 2014). The defective attachments can escape censorship posed by SACs, resulting in progression of aberrant mitosis (Godinho and Pellman, 2014). Consistent with the theory, a study conducted in Drosophila showed that excess centrosomes in larva brain tissue significantly increased tumorigenic risk (Nano and Basto, 2017). Studies using a doxycycline- inducible mouse model also showed that centrosome amplification induced by elevated PLK4 levels promoted genome instability and spontaneous tumorigenesis, which was partially rescued by p53 (Holland et al., 2012; Levine et al., 2017).

An interesting early observation described that in human tumors, multi-polar anaphases occur less frequently than multi-polar metaphases (Timonen and Therman, 1950). This led to the discovery of an alternative outcome for cells with supernumerary centrosomes where centrosomes are clustered to form pseudo-bipolar spindles (Brinkley, 2001; Godinho and Pellman, 2014). Centrosome clustering is a process where cells find a way to produce viable progeny by shaping the multi-polar spindle intermediates into pseudo-bipolar spindles that appear to function normally during mitosis (Milunović-Jevtić et al., 2016). Genome wide screens have identified proteins involved in centrosome clustering, including chromosomal passenger complex (CPC) components (Aurora-B, INCENP, Survivin, and Borealin) and Ndc80 complex (HEC1, SPC24, and SPC25), which localizes to the contact points for microtubule attachment on kinetochores (Leber et al., 2010). Centrosome clustering could also be induced as a result of an elongated anaphase by SAC activation (Yang et al., 2008). However, depletion of the HAUS complex inhibited centrosome clustering (Leber et al., 2010). Knock-down of components of the HAUS complex prevented γ-Tubulin recruitment resulted in reduced K-MT attachment to kinetochores and loss of spindle tension, preventing centrosome clustering (Leber et al., 2010). In summary, the mechanisms of centrosome clustering remain largely unknown. Since cancer cells with centrosome amplification require centrosome clustering to proliferate, exploring ways to prevent centrosome clustering may lead to new anticancer strategies.

Centrosome and centriole proteins have also been implicated with diseases in neurogenesis. For example, mutations in the gene encoding PCNT lead to microcephalic osteodysplastic primordial dwarfism type II (MOPD II) (Rauch et al., 2008), in which patients are characterized with low body weight and small head size (Rauch et al., 2008). Mutations in PCNT are also the cause of Seckel syndrome, also characterized with having markedly reduced brain and body size (Griffith et al., 2008). In addition, autosomal recessive primary microcephaly (MCPH), a severe developmental disorder characterized by reduced neural cell proliferation and small brain size during embryonic development (Barbelanne and Tsang, 2014), is caused by genetic mutations in genes encoding centrosome-localizing proteins (Barbelanne and Tsang, 2014; Faheem et al., 2015; Nano and Basto, 2017). To date, 25 microcephaly genes have been identified (Table 1-2) (Vulprecht et al., no date; Bond and Woods, 2006; Y.-C. Lin et al., 2013; Barbelanne and Tsang, 2014; Faheem et al., 2015; Martin et al., 2016; DiStasio et al., 2017; Moawia et al., 2017; Nano and Basto, 2017; Perez et al., 2019; Reilly et al., 2019). Apart from the known key centriole duplication proteins like STIL and CEP135, a recently identified microcephaly gene, KIF14, localizes to spindle MTs and has been shown to have a role in centriole duplication (Gupta, Coyaud, et al., 2015).

Many hypotheses and models have been proposed to explain the mechanism linking abnormalities in centrosome structure and function to developmental diseases. It has been proposed that since the inheritance of centrosomes is asymmetric, the need to divide asymmetrically during normal brain development requires neuronal progenitors to have increased sensitivity to centrosome defects (Bond and Woods, 2006; Wang et al., 2009; Pelletier and Yamashita, 2012). This is consistent with the fact that extra centrioles tend to cause developmental microcephaly (Marthiens et al., 2013).

Locus Microcephaly Proteins Cellular Localization Function MCPH1 Microcephalin Nucleus DNA Damage Checkpoint MCPH2 WDR62 Centrosomes/Neurons Mitotic Spindle Orientation MCPH3 CDK5RAP2 Centrosomes Centriole Maturation MCPH4 CASC5 Kinetochores Spindle Checkpoint MCPH5 ASPM Spindle Poles Microtubule minus-end Regulation MCPH6 CENPJ Centrioles Centriole Duplication MCPH7 STIL Centrioles Centriole Duplication MCPH8 CEP135 Centrioles Centriole Duplication MCPH9 CEP152 Centrioles Centriole Duplication MCPH10 ZNF335 Nucleus Cell Division MCPH11 PHC1 Nucleus Cell Cycle Control MCPH12 CDK6 Cytoplasm Cytoskeletal Stability MCPH13 CENPE Centromeres Centromere Integrity MCPH14 SAS6 Centrioles Cartwheel Formation MCPH15 NLS1 Nucleus Nuclear Import MCPH16 ANKLE2 Cytoplasm Nuclear Envelope Reassembly MCPH17 CITK Cytokinesis MCPH18 WDFY3 Cytoplasm Autophagy MCPH19 COPB2 Golgi apparatus Retrograde Trafficking MCPH20 KIF14 Spindle MTs Cytokinesis MCPH21 NCAPD2 Nucleus Chromosome Condensation MCPH22 NCAPD3 Nucleus Chromosome Condensation MCPH23 NCAPH Nucleus Chromosome Condensation MCPH24 NUP37 Nucleus Nuclear Pore Complex MCPH25 MAP11 Mitotic Spindles Cytokinesis

Table 1-2. A summary of identified microcephaly proteins, their localization and functions.

Centrioles form the basal bodies that dock to the cell membrane and generate cilia or Flagella (Marshall, 2009; Nigg and Raff, 2009). Dysfunctions and defects in cilia lead to ciliopathies, including Kartagener’s Syndrome, Bardet-Biedl syndrome (BBS) and Meckel- Gruber syndrome (MKS) (Afzelius, 1976; Priya et al., 2016; Hartill et al., 2017). Many MKS proteins are localized to the transition zone at the base of the cilia that functions in protein transport (Hartill et al., 2017). Disruptions at the transition zone result in Meckel syndrome, which has distinct features including occipital encephalocele, polycystic kidneys, and polydactyly (Hartill et al., 2017). The etiology of Bardet-Biedl syndrome has been linked to approximately 20 genes encoding for proteins that localize to the centrosome, with some (e.g.

ARL6/BBS3 and IFT27/BBS19) having a role in the control of intraciliary trafficking (Priya et al., 2016). BBS is characterized with obesity, diabetes, and renal problems, and has an autosomal recessive/oligogenic mode of inheritance (Priya et al., 2016). As the role for cilia in signaling becomes increasingly recognized, investigations into their functions could potentially lead to the development of new therapeutic approaches.

1.2 The Ubiquitination System

Ubiquitination is a post-translational modification process that involves conjugation of the 76-amino acid ubiquitin to target proteins to mediate downstream processes, including altering metabolic stability or other non-proteolytic functions. Ubiquitination plays a pivotal role in controlling centrosome number in cells and ensuring effective spindle formation (Zhang and Galardy, 2016). The importance of the ubiquitination system in the centrosome cycle has been widely addressed in a large number of studies. Ubiquitination is closely linked to centrosome biology through regulation of centrosomal proteins degradation (Sacristan and Kops, 2014; Zhang and Galardy, 2016). Targeted ubiquitination and deubiquitination tightly control the level and activity of several key players of centriole duplication that in turn facilitate the orderly assembly of the mitotic spindle (Zhang and Galardy, 2016). Here, I discuss the growing evidence showing the importance of the ubiquitination system in regulating the centrosome system, and provide a general picture of how these systems are tied together.

1.2.1 The covalent conjugation of ubiquitin with other cellular proteins regulates diverse eukaryotic cell functions.

Ubiquitin is a 76-amino acid polypeptide that contains a compact β-fold and a flexible C- terminus (Komander and Rape, 2012). It marks many cellular substrates to regulate a broad set of cellular events, including protein degradation and non-proteolytic cell signaling (Hershko and Ciechanover, 1998; Komander and Rape, 2012). Ubiquitination is a crucial post-translational modification achieved through mono/multi-ubiquitin chain attachment to specific sites on 26

substrate proteins. The format of the linkage between ubiquitin molecules contains information that decides the fate of the substrate (Hershko and Ciechanover, 1998; Komander and Rape, 2012). The linkage forms between the C-terminal Glycine of one ubiquitin and the amino group of seven Lysine sites (Lys6, Lys11, Lys27, Lys29, Lys33, Lys48, or Lys63) or the N-terminal methionine (Met1) of another ubiquitin (Kwon et al., 2008; Komander and Rape, 2012).

Ubiquitination is highly regulated and mediated by a hierarchy of enzymes (Figure 1-6). The ubiquitin activating enzyme (E1) activates the C-terminal Gly residue of ubiquitin in the presence of ATPs (Schulman and Harper, 2009). The activated ubiquitins are subsequently transferred to Ub-conjugating enzymes (E2s) (Ye and Rape, 2009). Finally, the ubiquitin is transferred to a protein target by ubiquitin ligases (E3), which interacts with E2 and the substrate specifically (Hershko and Ciechanover, 1998; Komander and Rape, 2012; Zhang and Galardy, 2016). In this step, the C-terminus of ubiquitin is linked in an amide isopeptide linkage to a ε- amino group of one of the Lys residues of the protein substrate. There are 2 major types of E1s, more than 50 E2s, and over 800 E3s that participate in the post-translational modification process (Zhang and Galardy, 2016; Kwon and Ciechanover, 2017). In most cases, E3s are responsible for selecting substrates and deciding the fates of the proteins.

There are three major types of E3s: RING (really interesting new gene) domain family, HECT (homologous to E6-AP C terminus) domain family, and the later discovered U-box domain proteins (Hershko and Ciechanover, 1998; Ciechanover, 2015). In ubiquitination process mediated by RING E3 ligases, the ubiquitin is transferred from E2 to the substrate directly, with RING E2 functioning as a platform (Hershko and Ciechanover, 1998). On the other hand, HECT E3 ligases and the ubiquitins from E2 form a thiolester intermediate before the ubiquitins are transferred to the substrates (Hershko and Ciechanover, 1998). Most of the E3s discussed in this chapter belong to RING family.

One of the reasons why ubiquitination influences a large variety of biological processes has to do with the flexible topology of the ubiquitin chains (Komander and Rape, 2012). The ubiquitin chains can be categorized by the number of lysine residues ubiquitinated on the substrate. Typically, ubiquitination can be divided into mono-ubiquitination and multi- ubiquitination, with the latter further broken down into homogenous ubiquitin chain, mixed

ubiquitin chain, or branched ubiquitin chain (Shih, Sloper-Mould and Hicke, 2000; Ye and Rape, 2009; Akutsu, Dikic and Bremm, 2016). Mono-ubiquitination largely occurs at multiple lysine sites, and has been traditionally thought to have no proteolytic roles (Shih, Sloper-Mould and Hicke, 2000; Livneh et al., 2017). However, recent studies found evidence that numerous proteins with specific structural disorders are targeted by the proteasome following monoubiquitination or multi-monoubibquitination (Braten et al., 2016).

Poly-ubiquitin chains, on the other hand, are more “canonical”. As mentioned above, there are 8 linkage sites, and proteomics analysis reveals that all of these sites are actively used for different functions (Yau and Rape, 2016). Among them, the most well-studied is Lys48, which generates proteasomal degron and leads to protein degradation (Tenno et al., 2004; Jacobson et al., 2009; Kwon and Ciechanover, 2017). Lys63 is widely known for its role in autophagic degradation of protein substrates and non-degradative processes including protein transport, protein kinase activation and DNA repair (Pickart and Fushman, 2004; Tenno et al., 2004; Lim and Yue, 2015). Typically, Lys11 is found forming mixed or branched chains with Lys48 or Lys63, and hence also plays a part in protein degradation (Xu et al., 2009). Lys33 and Lys29 are largely assembled by HECT E3s, and are involved in multiple cellular processes such as DNA repair, Wnt signaling and kinase activity regulation (Kazlauskaite et al., 2014; Kristariyanto et al., 2015; Yu et al., 2016). Misfolded proteins are mainly tagged with Lys48 chains that are linked with proteasome-associated receptors such as RPN10 and RPN13 through their ubiquitin binding domains (UBD) (Ciechanover and Kwon, 2017). If the misfolded proteins escape proteasome, they tend to form aggregates that are typically addressed by the autophagic proteolysis. HDAC6 (histone deacetylase 6) bind to these cargos and store them into aggresomes by recognizing both Lys48 and Lys63 linkages (Olzmann et al., 2007; S. T. Kim et al., 2013).

When it comes to the editing of ubiquitin chains, deubiquitinase (DUB) works to cut the linkage and erase the modifications (Mevissen and Komander, 2017). Six distinct types of DUBs have been described, including the ubiquitin-specific proteases (USPs), the ubiquitin C-terminal hydrolases (UCHs), the ovarian tumor proteases (OTUs), the Josephin family, the motif interacting with ubiquitin - containing novel DUB family (MINDYs), and the JAB1/MPN/Mov34 family (JAMMs) (Komander, Clague and Urbé, 2009; Reyes-Turcu, Ventii and Wilkinson, 2009; Clague et al., 2013; de Poot, Tian and Finley, 2017; Mevissen and 28

Komander, 2017). By cleaving ubiquitin or ubiquitin-like proteins from targets, DUBs serve multiple purposes in the ubiquitination pathway, including ubiquitin pro-protein activation, ubiquitin recycling, mono-ubiquitination generation and ubiquitin modification removal from target proteins (Nijman et al., 2005; Reyes-Turcu, Ventii and Wilkinson, 2009).

A E2 ATP E1 E2 E1 Substrate

RING HECT

Substrate Substrate

E3 E3

E2 E2

Substrate Substrate Substrate Substrate Substrate

Mono Linear Non-linear Multi-chain Branched

Figure 1-6. A Schematic illustration of the ubiquitin proteasome system. (A). The ubiquitin is activate by an E1 ubiquitin-activating enzyme in an ATP-dependent manner (Hershko and Ciechanover, 1998). It is then transferred to an E2 ubiquitin-conjugating enzyme (Ye and Rape, 2009). HECT E3 ligases thiolester intermediate before the ubiquitins are 29

transferred to the substrates, while the RING E3 ligases directly transfer the ubiquitin from E2 enzyme to the substrate (Komander and Rape, 2012; Zhang and Galardy, 2016). (B). The different topologies of ubiquitin chains. There are different types of ubiquitin modification, including mono-ubiquitylation, linear ubiquitin chain, non-linear ubiquitin chain, multi-ubiquitin chain and branched ubiquitin chain (Komander and Rape, 2012).

1.2.2 Ubiquitination and de-ubiquitination in centriole duplication and spindle assembly

Two major categories of ubiquitin ligases involved in the centrosome cycle are the aforementioned anaphase-promoting complex/cyclosome (APC/C) and Skp-cullin-F-box (SCF) with F-box protein complexes, both having key roles in cell cycle regulation (Zhang and Galardy, 2016). Both E3s contain the RING finger as their catalytic module, and Cullin sub-units that recruit E2 enzymes (Vodermaier, 2004; Zhang and Galardy, 2016). APC/C and SCF are evolutionarily related complexes composing of multiple sub-units that enable them to be involved in many cellular functions (Vodermaier, 2004). For example, SCF complex associates with F-box containing proteins that recognize different degrons in target proteins, and as a result, the number of potential substrate is very large (Vodermaier, 2004). Similarly, APC is fully functional only when it binds to activators, including Cdc20 and Cdh1, that bind to APC substrates (Peters, 2002; Primorac and Musacchio, 2013). During the critical transition between metaphase and anaphase, Cdc20 is responsible for linking SAC components and APC complex, whereas during G1 phase, Cdh1 takes over and bind to APC substrates such as Securin to promote cell cycle progression (Kapanidou, Curtis and Bolanos-Garcia, 2017). There is growing evidence that APC/C and SCF have functions beyond cell cycle regulation, including degradation of Polo-like kinases and Aurora kinases (Peters, 2002; Genschik, Sumara and Lechner, 2013).

As indicated in the previous section, centriole-guided and de novo biogenesis are initiated by PLK4 alike. Therefore, the abundance of PLK4 during the cell cycle is carefully and tightly

controlled. Ines Cunha-Ferreira and colleagues reported that the degradation of PLK4 is mediated by the SCF E3 with its F-box moiety Slimb (β-TrCP ortholog in Drosophila) complex (Cunha-Ferreira et al., 2009). They observed reduced PLK4 degradation in the presence of MG132, a 26S proteasome inhibitor. Following the discovery, the same group demonstrated that the trans-autophosphorylation of PLK4 at Ser293 and Thr297 sites within the β-TrCP targeting region is critical for its proteasome-mediated degradation (Cunha-Ferreira et al., 2013). Interestingly, while the autophosphorylation of PLK4 is triggered by STIL binding, the PLK4 pool that coexists with STIL stays intact during the entire procentriole generation process. It is likely that apart from activating PLK4 kinase activity, STIL also protects it from β-TrCP targeting through direct binding (Nigg and Holland, 2018). By regulating PLK4 levels, the SCF/Slimb complex controls centriole numbers in different tissues (Cunha-Ferreira et al., 2009, 2013).

The ubiquitination of PLK4 is also triggered by the E3 ligase Mind bomb (MIB1) (Cajanek, Glatter and Nigg, 2015). MIB1 associates with centriolar satellites, co-localizing with PCM-1, and relocates to centrioles at the beginning of procentriole biogenesis in response to increased regional levels of PLK4 (Cajanek, Glatter and Nigg, 2015). Cajanek and colleagues found that the catalytic dead mutant form of MIB1 (C997S) failed to bind PLK4, and polyubiquitination of PLK4 was significantly increased upon MIB over-expression (Cajanek, Glatter and Nigg, 2015). MIB1-induced ubiquitination led to PLK4 protein degradation, and additionally reduced the binding of PLK4 to CEP152 and CEP192 (Cajanek, Glatter and Nigg, 2015). In summary, SCF/β-TrCP and MIB1 regulate PLK4 stability and abundance at the beginning of centriole duplication.

PLK4 is not the only regulator of centriole duplication that is under control of E3 ligases. The cartwheel proteins SAS6 and STIL are also closely controlled by the ubiquitination system (Puklowski et al., 2011; Arquint and Nigg, 2016). SAS6 is a cell cycle regulated protein that is degraded between anaphase and early G1 (Puklowski et al., 2011). Interestingly, SAS6 is a substrate for both cell cycle regulating E3 ligase complexes APC/C and SCF/F-box (Strnad et al., 2007; Puklowski et al., 2011). Mediated by Cdh1, the APC complex drives the ubiquitination of SAS6 by targeting the KEN domain at its C-terminus (Strnad et al., 2007). On the other hand, SAS6 interacts with the F-box protein FBXW5 as demonstrated by co-immunoprecipitation 31

studies (Puklowski et al., 2011). Expression of excessive FBXW5 increased SAS6 stability (Puklowski et al., 2011). In S-phase arrested U2-OS cells, FBXW5 depletion significantly promoted centriole over-duplication (Puklowski et al., 2011), suggesting that FBXW5 has a role in centriole duplication regulation through regulating SAS6 levels. STIL is not detectable or remains at a very low levels during G1 phase, but its levels slowly increases towards G1/S transition (Arquint et al., 2012a). Level of STIL protein continues to increase as cells approach mitosis, peaking at metaphase (Arquint et al., 2012b). STIL protein levels drop at the same time as Cyclin B1, and co-depletion of Cdh1 and Cdc20 can significantly stabilize STIL (Arquint et al., 2012b). These data point support the hypothesis that STIL is likely regulated by the APC/Cdc20 complex, but this remains to be validated.

FBXW5 is phosphorylated by PLK4 at Ser151 residue to suppress ubiquitination mediated degradation of SAS6 (Puklowski et al., 2011). FBXW5 is itself a ubiquitination target and substrate of APC/Cdh1 (Puklowski et al., 2011). Level of FBXW5 peaks at G1/S transition, and thereafter it is targeted for degradation by APC/Cdc20 by targeting its D-box domain (Puklowski et al., 2011). Another target of APC/Cdh1 that localizes at the proximal ends of centrioles is CENPJ, the level of which is significantly reduced in late mitosis (Tang et al., 2009).

CP110 is targeted by the SCF/F-box complex during the G2 phase through the F-box protein Cyclin F (D’Angiolella et al., 2010). Depleting Cyclin F caused centrosomal abnormalities, specifically multipolar spindles lagging chromosomes (D’Angiolella et al., 2010). Cyclin F protein levels on centrosomes intensified from S phase to G2 phase, whereas CP110 protein levels peaked at S phase and decreased when cells entered G2 (D’Angiolella et al., 2010). More recently, T. Nagai and colleagues provided evidence showing that the CP110 interacting partner CEP97 degradation is mediated by an E3 ligase cullin-3 (CUL3)–RBX1– KCTD10 complex (Nagai et al., 2018). The complex belongs to the family of CULLIN-RING ubiquitin ligases, and is composed of the scaffold CULLIN3, a RBX RING domain, and a BTB domain serving as the adapter between the complex and its substrates (Genschik, Sumara and Lechner, 2013). As the BTB domain linker in the (CUL3)–RBX1–KCTD10 complex, KCTD10 is localized on the centrosomes in close proximity to CEP170 (Nagai et al., 2018). Every subunit of this complex is indispensable in promoting CEP97 degradation and subsequent control of 32

ciliogenesis (Nagai et al., 2018). The role of (CUL3)–RBX1–KCTD10 in centriole elongation regulation has not been investigated, but is an important question that should be addressed.

Ubiquitination also ensures mitotic spindle integrity. Nek2A, a cell-cycle related kinase of the never in mitosis A (NIMA) family, is required for centrosome separation and bipolar spindle formation (Faragher and Fry, 2003). The degradation of Nek2A in early mitosis is mediated by APC/C-Cdc20 through the recognition by Cdc20 of the KEN-box (Hames et al., 2001). Another example is the regulation of Eg5, a member of the kinesin-5 family of plus-end- directed microtubule-based motor proteins (Lawrence et al., 2004; Kapitein et al., 2005). Eg5 crosslinks MTs in the spindle and coordinates their sliding during spindle formation (Kapitein et al., 2005). In late mitosis, Eg5 is targeted by APC-Cdh1 (Eguren et al., 2014). When Cdh1 is over-expressed, levels of Eg5 are reduced, resulting in chromosome congression defects (Eguren et al., 2014).

Deubiquitinating enzymes (DUBs) also play an important role in regulating the levels of centriole duplication proteins. CP110 interacts with the DUB USP33 at centrosomes (Li et al., 2013). They observed that CP110 levels closely correlated with USP33 levels throughout the cell cycle. USP33 depletion significantly reduced CP110 levels, suggesting that CP110 is stabilized by USP33 through a process of deubiquitination (Li et al., 2013). Functionally, removal of USP33 inhibited hydroxyurea (HU)-induced centrosome amplification in U-2 OS cells (Li et al., 2013). These results indicate that USP33 positively regulates centriole duplication by stabilizing CP110.

In addition, Dr. Gomez-Ferreria from the Pelletier lab identified microtubule binding K63-DUB CYLD, of which the crystal structure of the USP domain was solved by the group of David Barford, in the array of CEP192 interactors using mass spectrometry (Komander et al., 2008; Gomez-Ferreria et al., 2012). Co-depletion of CYLD alleviated the severe spindle assembly defects in CEP192-depleted cells, indicating that CEP192 may regulate microtubule nucleation and stability through antagonizing CYLD (Gomez-Ferreria et al., 2012).

In summary, tight control of centriole duplication and spindle formation cannot be achieved without the ubiquitin-proteasome system. The APC/C and SCF complexes have prominent roles in regulating the levels and activities of key hub proteins such as PLK4. 33

1.3 The TRIM E3 ligases

The Tripartite Motif (TRIM) family proteins are RING type E3 ligases that have diverse biological roles, including intracellular signaling, development, apoptosis, protein quality control, innate immunity, autophagy, and carcinogenesis (Hatakeyama, 2017). Some TRIM family members have been associated with centriole biogenesis and spindle assembly. TRIM28, for example, mediates SUMOylation of Nucleophosmin (NPM), a multi-functional protein whose inactivation leads to centrosome over-duplication and genomic instability (Grisendi et al., 2005; Neo et al., 2015). TRIM18 (Midline-1, MID1) binds to MTs and targets the catalytic subunit of protein phosphatase 2A (PP2A) for degradation (Han, Du and Massiah, 2011). Its ortholog in Xenopus, Xnf7 (Xenopus nuclear factor 7), was identified as a MT-binding protein that localized to metaphase spindles both in Xenopus egg extracts and cultured cells. When Xnf7 is depleted, spindles were hypersensitive to MT-depolymerizing agents such as Nocodazole (Maresca et al., 2005). Here, I describe the definition, classification and cellular functions of TRIM family proteins to provide a rationale for my decision to focus on a TRIM family member in this thesis.

1.3.1 Sub-classes of TRIM family

The mammalian TRIM family was first characterized using a functional genomic approach, based on systematic data dating back to 2001 (Reymond et al., 2001). Membership in the TRIM family has expanded from 37 proteins to 75 proteins throughout the years (Reymond et al., 2001). Genes encoding TRIM proteins are found mutated in several human diseases, including familial Mediterranean fever, X-linked Opitz/GBBB syndrome and mulibrey nanism (Reymond et al., 2001; Han, Du and Massiah, 2011). TRIM family is defined by a specific N- terminal RING–B-box–coiled-coil (RBCC) E3 ligase: RING finger domain, one or two characteristic B-box domains and a coiled-coil motif (Napolitano and Meroni, 2012). This N- terminal arrangement is conserved throughout evolution, as previous studies have identified

genes encoding proteins belonging to a large family sharing the RBCC motif in Xenopus and C. elegans (Shou et al., 1996; Frank and Roth, 1998; Slack et al., 2000).

The RING motif is a specific type of Zinc Finger domain, and is defined by the consensus sequence CX2CX(9–39)CX(1–3)HX(2–3)C/HX2CX(4–48)CX2C (Stone et al., 2005). The particular arrangement of Cysteines and Histidines enables eight metal ligand residues to coordinate the zinc ions in a cross-brace structure (Meroni and Diez-Roux, 2005; Stone et al., 2005). The B-box domain, which belongs exclusively to the TRIM family, is structurally similar to the RING domain (Reymond et al., 2001; Napolitano and Meroni, 2012). B-box domains can be divided into types 1 and 2, each has specific Cysteine and Histidine arrangements (Napolitano and Meroni, 2012). If a TRIM protein contains 2 B-boxes, type 1 is always located N-terminal to type 2 (Reymond et al., 2001; Napolitano and Meroni, 2012). While the functions of B-box domains remain to be elucidated, a study of MID1 (TRIM18) suggested that type 1 B-box domain amplifies the E3 ligase activity of RING domain (Han, Du and Massiah, 2011). The coiled-coil domain facilitates cross-interaction among TRIM proteins, especially homo- interactions, suggesting formation of higher order structures (Sanchez et al., 2014).

Subfamilies of TRIM proteins are defined based on their C-terminal domain arrangements using bioinformatics approaches (Short and Cox, 2006). Common C-terminal domains of TRIM family proteins include B30.2-like/RFP (Ret finger protein)/SPRY (SplA and ryanodine receptor), AR (acid rich) region, ARF (ADP-ribosylation factor), PHD (plant homeodomain finger), Bromo domain, NHL (NCL-1/HT2A/LIN-41 repeat), immunoglobulin, and MATH (meprin and tumor necrosis factor receptor-associated factor homology) domains (Short and Cox, 2006). TRIM proteins are classified into 9 subfamilies, C-I to C-IX. C-I subfamily of proteins contains an arrangement of COS/FN3/B30.2-like domains; C-II proteins contain COS/AR region; C-III proteins contain COS/FN3 domains; C-IV proteins contain B30.2- like domains; C-V proteins have no clearly identified C-terminal domains; C-VI proteins contain PHD/Bromo domains; C-VII proteins contain immunoglobulin/NHL repeats; C-VIII proteins contain MATH domain; and C-IX proteins contain ADP-ribosylation factor (ARF) domains (Short and Cox, 2006).

1.3.2 TRIM36

TRIM36, originally known as RBCC728 or HAPRIN (haploid germ cell-specific RBCC protein), is encoded by a cDNA from 5q22.3 chromosomal region, where allelic changes have been associated with higher grades of conventional renal cell carcinoma (RCC), urothelial carcinoma (UC) and prostate carcinoma (PC) (Balint et al., 2004). Quantitative RT-PCR analysis of human tissue samples revealed that TRIM36 is highly expressed in testis, prostate and brain, whereas it is lowly expressed in kidney, lung and heart (Balint et al., 2004). Intriguingly, TRIM36 is significantly up regulated in prostate cancer samples (4-33 folds compared to normal tissues), suggestive of a role in tumorigenesis (Balint et al., 2004; Liang et al., 2018). Remarkably, another group also observed that with MAPK/ERK inhibition, TRIM36 expression significantly delayed prostate cancer cell cycle progression and inhibited cell proliferation in vitro and in vivo (Liang et al., 2018).

As a member of the TRIM family, the 728 amino acid TRIM36 protein contains a RING finger C3HC4 structure, two B-box domains, a coiled-coil domain, a COS box domain, a fibronectin type III domain, and a C-terminal domain of Spla kinase and ryanodine receptor (SPRY) (Balint et al., 2004; Short and Cox, 2006). It has been reported that TRIM36 is associated with developmental functions including fertilization, acrosome reaction and cortical rotation in embryogenesis (Cuykendall and Houston, 2009b; He et al., 2017). Using a yeast two- hybrid screen, Naoto Miyajima and colleagues identified CENP-H (Centromere protein-H) as an interacting partner with TRIM36, and showed that TRIM36 has ubiquitin ligase activity with the E2 enzyme Ubc4 (Miyajima et al., 2009). These findings suggest a role for TRIM36 in maintaining chromosomal stability. A study conducted by Emi Yoshigai and colleagues demonstrated that TRIM36 is essential for normal somite formation during early development of Xenopus embryos (Yoshigai et al., 2009). In addition, Tawny N. Cuykendall and Douglas W. Houston showed that depletion of maternal TRIM36 disrupted vegetal MT array formation in Xenopus embryos (Cuykendall and Houston, 2009b). Kieran M. Short and Timothy C. Cox identified an association between TRIM36 and MTs in human cells, and observed that over- expressed TRIM36 localized to MTs in COS-1 cells, similar to other C-I subfamily proteins (Short and Cox, 2006).

The C-I subfamily contains 6 proteins: TRIM36, TRIM1, TRIM18, TRIM9, TRIM46 and TRIM67 (Short and Cox, 2006). They share similar domain alignment, characterized by COS/FN3/B30.2-like domains at their C-termini (Short and Cox, 2006). There is evidence that this subfamily are highly involved in human neuronal cell growth and development (Berti et al., 2002; Han, Du and Massiah, 2011; Singh et al., 2017; Boyer et al., 2018; Do et al., 2018). A point mutation (P508T) in the B30.2-like domain of TRIM36 was found in patients with Anencephaly (APH), a neural tube defects (NTDs) disease characterized with missing brain tissues and cranium (Singh et al., 2017). Intriguingly, TRIM9 and TRIM67 were identified as targets in Paraneoplastic Cerebellar Degeneration (PCD), a rare neurological syndrome with a high cancer risk (Do et al., 2018). As described in the previous section, TRIM18 (MID1) associates with PP2A and binds to MTs (Han, Du and Massiah, 2011). Similar to TRIM36, MID1 was also observed to be upregulated in lung adenocarcinoma, and functionally affected cell cycle progression, proliferation and apoptosis (Liang et al., 2018; Zhang et al., 2018). Although the underlying mechanisms remain unclear, the one possible link between TRIM family members and neuronal diseases and cancer involves their roles in MT organization and cell cycle progression.

1.4 Rationale of the Thesis

When I joined the lab in 2012, my first project was to perform a high-throughput siRNA screen focusing on regulators of centriole duplication. TRIM36 emerged as a prominent candidate. TRIM36 was among a small cohort (30 versus 500 total) of proteins that scored as hits in all three screens (ciliogenesis screen, satellite screen, and centriole duplication screen) that our lab had completed at that time. The ciliogenesis screen was conducted in serum-starved RPE-1 cells to induce cilia formation in order to identify regulators of ciliogenesis (Gupta, Coyaud, et al., 2015). The satellite screen selected candidates that affected PCM1 and CEP290 intensity and distribution (Gupta, Coyaud, et al., 2015). The centriole duplication screen identified proteins that suppressed centriole amplification (Gupta, Coyaud, et al., 2015).

Apart from its potential association with MT structures (Short and Cox, 2006), little was known about TRIM36 at that time,. I was interested in investigating the role of TRIM36 in centriole duplication as an E3 ligase as well as the underlying mechanisms. The first chapter in the Results describes the discovery of novel centriole duplication regulators using a proximity based protein-protein interaction landscape of known centriole proteins using BioID. Functional RNAi screens identified TRIM36 as a potential candidate for further characterization. The overarching aim of the second chapter is to investigate the role of TRIM36 in centriole duplication and spindle formation using domain analysis and dependency assays to elucidate its position and function in the spatiotemporally regulated centrosome biogenesis pathway.

Chapter 2. Identification of novel centriole duplication regulators

Identification of novel centriole duplication regulators

2.1 Statement of Contributions

My PhD project started with exploring novel centriole duplication regulators through expanding protein proximity interaction map followed by identification of novel centriole duplication regulators that was a part of a broader scale project aiming to generate proximity profiles for 58 components of the centrosome-cilium interface and centriolar satellites in our laboratory that was published in Cell:

A Dynamic Protein Interaction Landscape of the Human Centrosome-Cilium Interface. Cell 163, 1484-1499. Gupta, Gagan D., Coyaud, É., Gonçalves, J., Mojarad, Bahareh A., Liu, Y., Wu, Q., Gheiratmand, L., Comartin, D., Tkach, Johnny M., Cheung, Sally W.T., et al. (2015)

My contribution in this collaborative project included charactering some of the BioID cell lines, preparing some of the BioID cell lines for mass spectrometry, helping Gagan with performing the centriole over-duplication assay, manually collecting and quantifying the centriole over-duplication assay data for subsequent data analysis. Some of the data presented in this chapter are included in this paper.

Sally W.T. Cheung generated cell lines stably expressing FLAG-BirA* fusion centriole duplication proteins for BioID including: CEP192, CEP152, STIL, SAS6, CEP135, CEP120, CENPJ, CEP120, CNTROB, POC5, POC1A, POC1B, CP110 and USP54.

Dr. Étienne Coyaud and Dr. Brian Raught (Princess Margaret Cancer Centre, University Health Network) performed all BioID mass spectrometry experiments presented in this study Dr. Anne-Claude Gingras analyzed the data and uploaded it to the Prohits website.

Dr. Gagan D. Gupta from Pelletier lab performed the centriole over-duplication assay, conducted IF image processing and statistical analysis shown in this chapter.

2.2 Summary

Centriolar proteins play a pivotal role in centriole biogenesis as well as in centriolar functions including cell division, cell signaling and motility, intracellular trafficking and cell polarity. Identification of novel centriolar proteins and their characterization have frequently lead to more clarity in the centriole duplication pathway and regulation. In order to identify novel centriole duplication regulators, I employed BioID followed by mass spectrometry to map protein proximity interactions of centriolar proteins. I generated a protein proximity interaction landscape of 13 known centriole duplication regulators (Table 2.1), providing a comprehensive view of the protein interaction relationships among these regulators. Next, I asked whether the preys were associated with centriole duplication regulation function by performing a centriole overduplication screen using S-phase arrested U-2 OS cells in combination with high-throughput imaging and quantification of centrioles. As part of this screen, I identified TRIM36 and UPS54 as novel regulators of centriole duplication that I further characterized together other potential candidates.

2.3 Introduction

2.3.1 BioID

The combination of affinity purification followed by mass spectrometry (AP-MS) is a powerful tool to study protein-protein interactions (PPI). In the AP-MS method, target proteins (‘baits’) are fused with epitope tags to serve as affinity capture probes. Commonly used epitope tags include FLAG, Myc, Strep, GFP, His-6 and protein A (Morris et al., 2014). The epitope tags eliminate the reliance on specific antibodies that need to be generated for each ‘bait’ protein. If the antibody binding site overlaps with the PPI interface, the antibody will disrupt the PPI (Morris et al., 2014). However, the utility of AP-MS in identifying PPIs between centrosome and cilium proteins remains limited due to the insoluble nature of these organelles (Gupta, Coyaud, et al., 2015). Harsh lysis conditions often destroy PPIs, resulting in inaccurate and partial views of the whole interaction landscape.

The BioID (proximity-dependent biotin identification) method was optimized and applied to this project in order to circumvent the limitations of AP-MS in identifying PPIs between centrosome and cilium proteins (Roux et al., 2012) (Figure 2-1). First discovered in Escherichia coli, BirA is a 35 kD DNA-binding protein ligase that facilitates the biotinylation of an acetyl- CoA carboxylase subunit (Roux et al., 2012). The original BirA ligase is highly selective for endogenous substrates and could not be widely applied as a technology. In this study, I used a BirA mutant, referred to as BirA*, that contains a R118G mutation that renders the enzyme more promiscuous towards target proteins(Roux et al., 2012). Proteins of interest (‘baits’) are fused with BirA* and expressed in cells under condition of excess biotin (Roux et al., 2012; Morriswood et al., 2013; Firat-Karalar et al., 2014). In the presence of ATP, BirA* activates biotin and synthesizes biotinyl-5’-AMP (bioAMP) (Lane et al., 1964). This activated form of biotin interacts with lysine residues, generating biotinylated polypeptides(Roux et al., 2012), which can be captured by high-affinity avidin/streptavidin- mediated purification (Green, 1963).

Unfiltered datasets from AP-MS and BioID contain non-specific binding proteins (contaminants) that contribute to false positive protein interactions (Choi et al., 2011). Statistical models are used to distinguish true interactions from contaminants, which include proteins that bind to affinity matrix alone and frequent flyers that bind to a large number of bait proteins (Skarra et al., 2011). An advanced computational approach termed Significance Analysis of INTeractome (SAINT) is used to identify a bona fide PPI (Choi et al., 2011, 2012; Skarra et al., 2011). By estimating specific mixture of distributions of each bait-prey pair that represent true or false interactions, SAINT converts label-free quantification metrics, such as spectral count, into probabilities of true interactions after normalizing spectral counts to the length of the proteins as well as the total number of peptides counts in the total purification pool. For accuracy, SAINT leverages negative controls to model spectral count distributions to identify potential false PPIs. Compared to other statistical tools such as PP-NSAF and CompPASS (Sowa et al., 2009), SAINT is able to estimate Bayesian false discovery rate (BFDR) based on probability calculations (Choi et al., 2011; Skarra et al., 2011). In combination, BioID and SAINT represent a powerful tool to study PPIs that are important for centriole duplication.

Non-specific Proteins

Biotin

BirA* Bait

Prey Lysis and Solubilization

Streptavidin Affinity Purification

Mass Spectromatry

Figure 2-1. A schematic overview of BioID method. (Adapted from Roux et al., 2012) The biotin–ligase fusion protein (bait) is expressed in Flp-In T- REx HEK 293 cells. Harsh cell lysis conditions help to break-down centrosomes, releasing and denaturing of centrosomal proteins. Affinity purification of biotinylated proteins using streptavidin followed by mass spectrometry identified proteins that interact with the bait.

2.4 Results

2.4.1 Generation of the centriole duplication interactome

I generated an interaction map based on BioID data of 13 known centriole duplication regulators (Table 2-1). The interaction map is part of a published paper from the Pelletier lab, in which 58 “bait” polypeptides localized to the centriole, centriolar appendages, centriolar satellites or the ciliary transition zone defining the “centrosome-cilium interface” were included in a systematic BioID analysis (Gupta, Coyaud, et al., 2015). First, each protein is tagged with FLAG-BirA* at its N-terminal, and ligated into the pcDNA5 FRT/Tet-On vector. The tetracycline-inducible fusion constructs are transfected and stably expressed in Flp-In T-REx HEK 293 cells (Roux et al., 2012; Gupta, Gingras, et al., 2015). Subsequently, the cell lines were characterized by Western blotting and immunofluorescence imaging to examine the expression and localization of the proteins of interest, along with its biotinylation of proximately interacting partners (Roux et al., 2012; Gupta, Coyaud, et al., 2015) (Figure 2-2). The cell lines were seeded into 6-well plates and treated with 1 μg/mL tetracycline or 1 μg/mL tetracycline plus 50 μM biotin for 24 hours. Due to its strong affinity with biotin, horseradish peroxidase (HRP) conjugated Streptavidin was used for detecting the biotinylation by Western blotting, and Alexa Fluor™ 647 Streptavidin conjugate was used in IF imaging (Fig 2-3) (Roux et al., 2012; Gupta, Coyaud, et al., 2015).

Figure 2-2. Examples of expression and biotinylation validation of FLAG-BirA* HEK293 T-REx cells by Western blot and immunofluorescence imaging.

(A) Immunofluorescence staining of HEK293 T-Rex expressing Tet-inducible FLAG-BirA*- POC5 revealed the fusion protein localized on centrosomes (± Tet (1 µg/mL, 24 h) and ± Biotin 45

(50 µM, 24 h)). Cells were co-stained with anti-FLAG, anti-Streptavidin and anti-Pericentrin antibodies. Scale bar: 15 μm. (B) Western blot using anti-FLAG (upper panel) and anti- Streptavidin (lower panel) HEK293 T-Rex expressing Tet-inducible FLAG-BirA*-POC5 under following conditions: ± Tet (1 µg/mL, 24 h) and ± Biotin (50 µM, 24 hours).

List of Centriole Function Reference Duplication Regulator CEP192 Initiation, maturation (Zhu et al., 2008) CEP152 Initiation (Sonnen et al., 2013) SAS6 Cartwheel formation (Keller et al., 2014) STIL Cartwheel formation (Arquint et al., 2012a) CNTROB Elongation (Zou et al., 2005) CENPJ Elongation (Tang et al., 2009) CEP135 Elongation (Y.-C. Lin et al., 2013) (Comartin, Gagan D. Gupta, et al., CEP120 Elongation 2013) SPICE1 Elongation (Archinti et al., 2010) POC5 Elongation (Azimzadeh et al., 2009b) POC1A Elongation (Keller et al., 2009) POC1B Elongation (Keller et al., 2009) CP110 Elongation (Schmidt et al., 2009)

Table 2-1. A list of centriole duplication factors used as baits in this study.

The characterized cell lines expressing moderate amounts of aforementioned centriole duplication regulators were harvested for BioID and SAINT analyses. For each regulator, at least four replicates (two biological replicates, each having two technical replicates) were analyzed. Dr. Etienne Coyaud performed mass spectrometry experiments. The final probability scores were computed for each pair of bait and prey, and a BFDR cutoff of 0.02 was applied (Gupta, Coyaud, et al., 2015). Only interactions with an average SAINT score (AvgP) > 0.8 and maximum SAINT score (MaxP) >0.85 were chosen to build the interactome. After manually removing common background proteins (Gupta, Coyaud, et al., 2015), 2283 high-confidence protein interactions were identified, including 1644 unique preys. The interaction map was visualized

using Cytoscape 3.6.1 to generate a self-organized, prefuse force-directed topology map with spring embedded (Shannon et al., 2003), as shown in Figure 2-3.

Figure 2-3. The protein proximity interaction map of 13 centriole duplication regulators. An entire of 2012 proximity interactions were identified with 13 centriole duplication regulators after applying cutoff of BFDR < 0.02, AvgP > 0.8 and MaxP > 0.85. Large red nodes represent bait proteins used in the study, while gray nodes represent interacting/proximity preys. interactors represented by small gray nodes. Gray arrows represent bait-prey interactions, and their corresponding edge thickness is aligned to total peptide spectral counts.

A reference list termed centrosome and cilium database (CCDB) were used as benchmark for validating the defected interactions. The CCDB was constructed with proteins with existing

evidence of association with the centrosome and cilium interphase (Gupta, Coyaud, et al., 2015). The CCDB is composed of 1554 proteins merged from Syscilia Gold Standard ciliary protein (van Dam et al., 2013; Christie and Blake, 2018) and the second generation of the human CentrosomeDB (Alves-Cruzeiro, Nogales-Cadenas and Pascual-Montano, 2014). The CentrosomeDB contains 1357 centrosome genes from human and Drosophila with corresponding compiled information (Alves-Cruzeiro, Nogales-Cadenas and Pascual-Montano, 2014). In the aforementioned centriole duplication dataset, there are 302 proteins belonging to CCDB. The high overlap ratio provided rationale for our approach to identify novel centriole duplication regulators.

A large amount of centrosomal proteins were identified (Table 2-2). Previously validated PPIs were also present in the interaction map, including CCP110-CEP97, CENPJ-CEP135, CEP152-CEP192, SPICE1-CEP120, and STIL-CEP85, and these highlighted in blue in Table 2- 2 (Liu et al., no date; Spektor et al., 2007; Comartin, Gagan D. Gupta, et al., 2013; Sonnen et al., 2013; Y.-C. Lin et al., 2013).

Some centrosomal proteins used as ‘baits’ localize on the microtubules during certain stages of the cell cycle, including SPICE1, which localizes on the spindles during mitosis (Archinti et al., 2010). Thus not surprisingly, the dataset was also enriched for proteins with pivotal roles in MT regulation and spindle assembly (Table 2-3). The entire Augmin/HAUS complex was clustered in the dataset, along with MAPs, LUZP1, DCTN1, NAP1L1 (Lawo et al., 2009), providing potential links and cross-overs among centriole duplication regulators and spindle assembly regulators.

Bait Prey Total Spec Count Saint Score CEP76 139 1 CCP110 CEP97 632 1 CEP120 373 1 CEP152 67 1 CEP170 1322 1 CEP192 331 1 CEP290 45 1 CEP350 51 1 CEP44 63 1 CENPJ CEP55 224 1 CEP63 26 1 CEP72 91 1 CEP76 133 1 CEP85 188 1 CEP97 1267 1 CEP104 18 0.99 CEP135 11 0.94 CEP192 119 1 CEP350 132 1 CEP72 45 1 CEP95 193 1 CEP97 70 1 CEP55 44 0.95 CEP170 443 1 CEP192 475 1 CEP120 CEP350 106 1 CEP44 42 1 CEP55 183 1 CEP72 57 1 CEP85 32 1 CEP95 46 1 CEP97 77 1 CEP128 23 0.99 CEP192 256 1 CEP152 CEP85 26 1 CEP63 24 0.97 CEP120 70 1 CNTROB CEP192 679 1 49

CEP85 20 1 POC1A CEP44 89 1 POC1B CEP44 61 1 CEP192 294 1 POC5 CEP72 45 1 CEP128 18 0.97 CEP152 106 1 CEP192 325 1 SAS6 CEP350 119 1 CEP72 51 1 CEP97 117 1 CEP120 690 1 CEP135 432 1 CEP350 173 1 CEP44 44 1 SPICE1 CEP55 136 1 CEP72 137 1 CEP95 33 1 CEP85 28 0.99 CEP192 15 0.97 CEP85 207 1 STIL CEP250 14 0.96

Table 2-2. A representative subset of previously validated centrosomal proteins and PPIs identified in the proximity interaction network. BioID results were collected from at least four replicates (two biological replicates, each having two technical replicates) of Tet-inducible FLAG-BirA*-bait protein expression in the presence of tetracycline and biotin followed by Streptavidin affinity-purification combined with mass- spectrometry to identify biotinylated peptides. Spectral counts reflect the abundance of a prey polypeptide in the SAINT analysis. (From left to right) Prey: Gene Symbol for identified “prey” proteins; Total Spec Count: total amount of spectral counts for all replicates; Saint Score: overall SAINT score for 4 replicates (Choi et al., 2011, 2012; Skarra et al., 2011). MS and was performed by Dr. Etienne Coyaud. Spectral counts reflect the abundance of a prey polypeptide in the SAINT analysis.

Bait Prey Total Spec Count Saint Score CP110 HAUS6 110 1 CKAP4 255 1 CKAP5 3387 1 DCTN1 228 1 DCTN2 272 1 HAUS1 393 1 HAUS2 137 1 HAUS3 822 1 HAUS4 393 1 HAUS5 650 1 HAUS6 1170 1 HAUS7 273 1 HAUS8 577 1 KIF14 168 1 LUZP1 439 1 MAP3K7 60 1 CENPJ MAP4K4 88 1 MAP4K5 56 1 MAP7D3 386 1 TUBG1 164 1 TUBGCP3 55 1 KATNB1 98 0.99 MAP2K2 16 0.99 TUBE1 24 0.99 MAP1S 13 0.97 HDAC6 16 0.94 TUBGCP4 14 0.94 MAP7 13 0.92 CKAP2 138 0.9 DCTN5 18 0.89 TUBGCP2 129 0.89 HAUS3 73 1 HAUS6 157 1 HAUS7 19 1 CEP120 LUZP1 134 1 MAP7D3 148 1 NAP1L1 791 1 NAP1L4 288 1 51

HAUS5 62 0.99 KIF14 20 0.97 HAUS2 30 0.93 HAUS3 84 1 HAUS6 121 1 HAUS7 26 1 KIF14 52 1 CEP135 LUZP1 471 1 MAP7D3 143 1 HAUS4 37 0.99 HAUS8 46 0.88 CKAP5 503 1 HAUS3 48 1 KIF14 48 1 MAP4K4 14 0.97 CEP152 TUBGCP2 60 0.96 HAUS7 15 0.95 HAUS6 57 0.9 CKAP5 196 1 CNTROB LUZP1 99 1 CKAP2 161 1 POC1A CKAP5 393 1 MAP7D3 140 1 HAUS1 224 1 HAUS2 224 1 HAUS3 673 1 HAUS4 305 1 POC5 HAUS5 641 1 HAUS6 1101 1 HAUS7 200 1 HAUS8 443 1 HAUS3 229 1 HAUS4 83 1 HAUS5 149 1 HAUS6 271 1 SAS6 HAUS7 40 1 HAUS2 68 0.99 HAUS8 128 0.99 HAUS1 60 0.96 52

KIF14 12 0.94 CKAP2 580 1 CKAP2L 48 1 CKAP5 582 1 HAUS1 83 1 HAUS2 64 1 HAUS3 219 1 HAUS4 79 1 HAUS5 186 1 HAUS6 378 1 HAUS7 84 1 HAUS8 122 1 SPICE KIF14 91 1 LUZP1 600 1 MAP1S 60 1 MAP7 138 1 MAP7D3 892 1 MAP9 110 1 NAP1L1 1208 1 NAP1L4 486 1 MAPRE2 186 0.97 MAP7D2 14 0.95 MAP7D1 88 0.92 STIL HAUS7 10 0.89

Table 2-3. A representative subset of previously validated microtubule associated proteins identified in the proximity interaction network. BioID results were collected from at least four replicates (two biological replicates, each having two technical replicates) of Tet-inducible FLAG-BirA*-bait protein expression in the presence of tetracycline and biotin followed by Streptavidin affinity-purification combined with mass- spectrometry to identify biotinylated peptides. Spectral counts reflect the abundance of a prey polypeptide in the SAINT analysis. (From left to right) Prey: Gene Symbol for identified “prey” proteins; Total Spec Count: total amount of spectral counts for all replicates; Saint Score: overall SAINT score for 4 replicates (Choi et al., 2011, 2012; Skarra et al., 2011). MS and was performed by Dr. Etienne Coyaud. Spectral counts reflect the abundance of a prey polypeptide in the SAINT analysis.

2.4.2 Functional characterization of centriole duplication interactome and identification of novel centriole duplication regulators

To identify novel factors from the preys that regulate centriole assembly, I first performed a pilot RNAi screen using Endoribonuclease-prepared siRNA (esiRNAs) on 72 candidate preys picked from the CEP120/SPICE1 interactome, since CEP120 and SPICE1 are two known centriole duplication regulators that interact with each other (Table 2-4) (Comartin, Gagan D. Gupta, et al., 2013). The esiRNAs were generated by Sally Cheung as previously described (Kittler et al., 2005). The screen employed a centriole over-duplication assay in U-2 OS cells because this assay is more sensitive compared to assays using cycling cells since it provides a higher dynamic range due to the increase of centriole numbers (Balczon et al., 1995a). The human osteosarcoma U-2 OS cell line was derived from a moderately differentiated sarcoma of 15-year-old girl in 1964 (Niforou et al., 2008). When arrested in S-phase using HU or the DNA polymerase inhibitor Aphidicolin, the U-2 OS cells undergo multiple rounds of centriole replication, potentially due to a deficient p53 pathway (Prosser, Straatman and Fry, 2009).

Candidate Z-score TRIM36 -16.143 Centrobin -13.571 Poc1B -12.429 PFDN4 -12.143 CEP120 -11.571 CEP72 -10.286 PIBF1 -10 SPICE -9 Poc1A -9 SKA2 -8.8571 MAP7D3 -8.8571 CPAP -8.7143 CEP63 -8.7143 CCDC77 -8.5714 USP54 -8.4286 MAP1S -8 NAP1L1 -7.8571 PFDN2 -7.7143 MAP7D3si5 -7.2857 KIF14 -7.1429 KIAA1218 -6.7143 AZI -6.5714 SKA3 -6.5714 KIF7 -6.5714 CAMSAP1 -6.5714 ANK2 -6.4286 CCDC93 -6.4286 MAP126 -6.2857 PELO -6.1429 CCDC92 -6.1429 CEP135 -6 MTUS -5.8571 KIAA1731 -5.8571 CCDC112 -5.8571 C10orf88 -5.7143 CCDC85C -5.7143 KIAA00586 -5.5714 KIAA0586 -5.4286 CCDC22 -5.1429 SAS6 -5 MAP7 -5 TRIM36 -5

MAPT -5 MAP7D1 -4.8571 MAP9 -4.7143 CNTROB -4.7143 CSPP1 -4.5714 MTUS1 -4.4286 KIAA1525 -4.4286 KIAA1672 -4.2857 MAP7D3si6 -4 FAM91A1 -4 CCDC138dq2 -3.4286 PFDN5 -3.1429 PFDN6 -3.1429 NCKAP5L -2.7143 CCDC14 -2.4286 C2orf44 -1.8571 ALMS -1.7143 IFT20 -1.4286 CC2D1A -1.4286 KIAA00753 -1.2857 C15orf23 -1 PFDN1 -1 ZC2HCA1 -0.8571 MIIP -0.8571 SKA1 -0.7143 TMEM209 -0.4286 NT 0 CEP55 0 CCDC138 0.14286 TEX9 0.42857 FAM161A 0.71429 KIAA1217 1.28571 CP110 1.85714

Table 2-4. Prey candidates selected for the esiRNA centriole duplication screen. The % of cells with centriole over-duplication was determined for each RNAi condition. Z- c−cμ scores are calculated using the following formula: 푧 = , where c-cμ represents the differences cSD between over-duplication level for any given test conditions versus the mean over-duplication % of cells transfected with non-targeting siRNA (Gupta, Coyaud, et al., 2015). Csd represents the standard deviation of over-duplication among cells transfected with non-targeting esiRNA.

Positive controls, including SPICE1 and CEP120, are highlighted in yellow, whereas positive control is labeled in yellow. TRIM36 is labeled in red, and USP54 is labeled in orange.

In the screen, U-2 OS cells were first seeded onto 12-well coverslips coated with hydrophobic Teflon except for the well areas, reverse transfected with esiRNA pools (3,000 cells/well; 20ng/μl esiRNA) for 24 hours, and then flooded with McCoy's 5A (Modified) media containing 1.6μg/ml aphidicolin (Sigma-Aldrich) for 48 hours to arrest the cells in S-phase and induce over-duplication (Figure 2-4). Using Centrin foci as a centriole marker and PCNT as a centrosome marker, cells were counted based on centriole over-duplication (more than 4 Centrin foci in one cell), and the number of cells harboring over-duplicated centrioles were used to calculate Z-scores (the differences between over-duplication level for any given test conditions and mean over-duplication level of cells transfected with negative control esiRNA (Luciferase), divided by standard deviation of over-duplication among cells transfected with negative control esiRNA) (Figure 2-4, 2-5).

HU - + + CEP120 RNAi - - +

Figure 2-4. U-2 OS S-phase arrested centriole over-duplication assay.

(A) A demonstration of 12-well Teflon plate used in the study. The coverslips are covered with hydrophobic Teflon except for the well areas, keeping the droplets containing cells and esiRNAs separate with each other. (B) A schematic overview of the centriole over-duplication assay in S- phase arrested U2-OS cells. When arrested in S-phase, the U-2 OS cells undergo multiple rounds of centriole replication. When cells are treated with RNAi targeting centriole duplication factors (as shown here for CEP120), the number of centrioles indicated by Centrin foci will be less compared to negative controls.

The screen identified several promising candidate hits that strongly inhibited centriole over-duplication in centriole assembly. These hits included TRIM36, which I further pursued in follow up studies. I will elaborate on the results and discussions of TRIM36 characterization in the following chapter. The other hits included USP54, PFDN4, PFDN2, PIBF1, SKA2, KIF14, MAP1S and CCDC77. Prefoldins (PFDN) have been associated with MT stability (Delgehyr et al., 2012), while PIBF1 plays a critical part in ciliogenesis, with mutations in this gene occuring in patients with Joubert Syndrome, a disease classified under ciliopathies (Ott et al., 2019). SKA2, KIF14 and MAP1S closely associate with MTs and spindle dynamics (Tegha-Dunghu et al., 2014; Kevenaar et al., 2016; Sivakumar and Gorbsky, 2017). The function of CCDC77 is not known.

2.5 Z-score 0.0

-2.5

-5.0

e r

o -7.5 TRIM36 sc

- USP54

Z -10.0 Positive controls: SPICE1 and CEP120 -12.5 Negative control: Non-target

-15.0

-17.5 58

Figure 2-5. Z-score distribution of the esiRNA centriole over-duplication pilot screen. The % of cells with centriole over-duplication was determined for each RNAi condition. Z- c−cμ scores are calculated using the following formula: 푧 = , where c-cμ represents the differences cSD between over-duplication level for any given test conditions versus the mean over-duplication % of cells transfected with non-targeting siRNA (Gupta, Coyaud, et al., 2015). Csd represents the standard deviation of over-duplication among cells transfected with non-targeting esiRNA. Positive controls, including SPICE1 and CEP120, are labeled in green, whereas positive control is labeled in yellow. TRIM36 is labeled in red, and USP54 is labeled in orange.

USP54 is a deubiquitinating enzyme of the family Ubiquitin-Specific Proteases (USPs) that are frequently deregulated in cancer (Quesada et al., 2004; Fraile et al., 2016; Miguela and Lujambio, 2017). Recently, a study has shown that down-regulation of USP54 in colorectal carcinoma cells inhibited tumorigenesis through a yet unknown mechanism (Fraile et al., 2016). An earlier study had previously associated USP54 with a putative centrosome function based on the identification of a proximity interaction between USP54 and PLK4 (Firat-Karalar et al., 2014). In our interaction map, USP54 was identified as a prey for CEP135, further supporting its potential role in centriole duplication regulation.

To validate the centriole duplication defect observed after USP54 esiRNA treatment, I used two independent Dharmacon ON-TARGETplus siRNAs against USP54. One of the siRNA didn’t yield a centriole duplication phenotype, suggesting potential off-target event (Sigoillot and King, 2011). The other treatment resulted in a 70% to 20-40% decrease in number of cells with over-duplicated centrioles in the U-2 OS S-Phase arrest centriole over-duplication assay. These results showed that USP54 has a moderate effect on centriole duplication (Figure 2-6). To exclude off target effects, in the future, I will perform rescue experiments to further validate the role of USP54 in centriole duplication.

ll *** ns

ce >4 c

i 100

t <=4

)

% o

( 50

t n

ce 0 r

e t i s -1 -2

p i i rge IL s s a T 4 4 -T S 5 5 n P P o S S N U U

Figure 2-6 USP54 regulates centriole duplication in a moderate manner. Bar graph showing the percentage of S-phase arrested U-2 OS cells processed in which either >4 or 4 centrioles were detectable (***p<0.001, **p<0.01, p<0.05 by Student’s t-test, n=50, four replicates per condition, mean ± SEM).

With the link between USP54 and centriole duplication established, I sought to perform BioID analysis of USP54 to map its proximity interactors and further expand the proximity interaction map (Roux et al., 2012). This work was performed in collaboration with Dr. Étienne Coyaud in the lab of Dr. Brian Raught. Sally Cheung from the Pelletier lab created a HEK293 T- REx cell line expressing Tet-inducible FLAG-BirA*-USP54 protein. GO analysis revealed that the acquired dataset contained multiple known centrosomal proteins as indicated in Figure 2-7, including 6 known components of basal body (CYLD, CSNK1A1, RPGRIP1L, DLG5, SDCCAG3 and OFD1), 25 known components of MTOC and 42 components of MT cytoskeleton (Panther 14.1, Table 2-5, Figure 2-7). Together with its deubiquitinase CYLD, CEP192 was also identified in the interactome (Komander et al., 2008; Gomez-Ferreria et al., 2012). Dr. Gomez-Ferreria CYLD in the array of CEP192 interactors using mass spectrometry (Komander et al., 2008; Gomez-Ferreria et al., 2012). She found that co-depletion of CYLD alleviated the severe spindle assembly defects in CEP192-depleted cells, indicating that CEP192 may regulate microtubule

nucleation and stability through antagonizing CYLD (Gomez-Ferreria et al., 2012). The functions of newly discovered proteins in the dataset remain to be characterized in future studies.

MTOC MT CC2D1A SDCCAG3 CEP192 CEP192 CEP55 YTHDF2 CEP85 RASSF7 CSNK1A1 PLEKHA7 CYLD LZTS2 DLG5 TPGS1 FLII MPHOSPH9 HAUS3 FLII HAUS8 DLG5 KRT18 DVL1 LATS2 HAUS3 LZTS2 LRRC49 MPHOSPH9 TUBB6 NDEL1 CSNK1A1 NME7 BAG2 OFD1 RPGRIP1L PLEKHA7 DNAJA1 RASSF7 TUBAL3 RPGRIP1L KRT18 SDCCAG3 CYLD SORBS1 KIF14 TCHP RAE1 TPGS1 TCHP YTHDF2 YWHAE DYRK1A TTK NDEL1 OFD1 PKP4

Table 2-5 A subset of MTOC and MT cytoskeleton proteins identified in the proximity interaction network of USP54.

PPP1CA NAT2 HIST1H2AJ RPGRIP1L TBK1 KRT19 RPL18A CNOT7 POLR2H

CYLD LPP ACTB ERBB2IP KRT8 CAD KLHL15 CEP192 SNX9 CNOT10 SPATA2 RPF2 CSE1L DNAJA3 JUP SCRIB KANK2 PMFBP1 RPS27L CTNND1 CCDC85C TRIP6 YTHDC2 ADSL TNRC6B

KIAA1462 50.17 PDLIM5 NUP155 RASSF8 LOC100505818 EIF2C2 4.33 GPATCH1 CNOT1 42.33 4.0 USMG5 10.17 3.83 32.17 27.67 TNRC6A CCDC138 ECD 1.83 AP3M1 25.17 YWHAB RPS11 69.83 7.5 72.17 124.17 41.33 8.83 3.67SIPA1L1 LIMD1 MKL1 PKP4 3.0 HSPH1 6.0 15.5 3.5 GFAP KDELR2 NDEL1 6.83 4.33 2.83 CNIH4 13.0 14.33 15.17 WDR6 7.67 9.83 PDLIM7 3.67XIAP30.0 LATS2 33.0 100.0 8.5RPL36A RPL37A SORBS1 ABLIM1 4.17 34.17 RPS15A 50.67 6.33 18.83 AP2B110.67 KRT18 RASSF7 57.5 19.83 3.0 48.5 3.33 3.17 41.17 1.83 15.17 7.5 4.0 16.5 PRKAG1 OFD1 11.67 5.0 8.67 2.5 2.67 RQCD1 WDR77 HAUS3 MSH5 STUB1 EXOG 4.17 1.67 MAGED1 5.67 5.17 5.5 6.17 2.67 5.67 2.83 WDR83 SIPA1L237.5 6.17 38.67 9.83 PARD3 SKA3 ANKHD1 7.67 2.83FLII 8.33 2.5 2.33 5.33 4.33 3.5 5.67 CSNK1A1 3.5 3.67 9.67 7.33 1.83 4.83 DCAF7 SEC24B XPO59.67 LZTS2 2.67 3.83 13.33 14.33 IKBKAP YWHAE 123.83 4.0 CC2D1A9.83 18.33 FlagBirA-USP541.67 3.67 199.33 5.67 1.83 34.67 TUBB2A DNAJA1 2.83 DIP2A MPP5 16.5 DNMBP 4.67 IRAK1 44.5 4.0 HAUS8 12.0 49.0 3.67 SKA2 164.33 24.0 11.0 2.83 TUBAL3 2.33 4.33 TK1 USP9X 1.83 18.83 30.0 YWHAZ49.5 136.17 6.33 17.5 CDIPT 2.83 TUBB62.0 5.83 22.33 25.67 YWHAQ DEPDC1B 510.17 KIF14 2.33 4.33 16.83 6.17 35.67 2.83TCHP18.33 TRIM37 PDZD11 10.33 DYRK1A IPO5 4.33 8.5 PPP1R13B RPL23 5.5 22.5 33.83 5.5 2.83 CCT7 334.67 2.0 4.67 5.5 44.0 4.83 23.33 ATXN10 TP53BP2 31.33 6.0 2.83 3.83 ELP3 5.83 BAG2 46.17 6.33 RPL21 NEBL PAK4 7.5RPL10 TJP1 8.33 60.67 2.83NUP2056.17 RPS23 371.0 5.33 DCP1B CEP85 EIF3L 9.5 33.17 EDC33.17 RPS24 EIF2C1 3.17 6.17 26.83 13.0 2.33 RAE1 37.67 YWHAH 2.83 DLG5 DSP YTHDF2 TAT MGLL 14.83 16.0 2.83 PLEKHA7 TJP2 3.17 KPNA2 FAM83H NUP133 DVL3 YWHAG DVL2 PLEKHA5 NOL6 PALLD SEC23A FARSA KIAA1671 IRS4 CXADR EPHA2 RPL36 DVL1 MAP4K4 MLLT4

XRN1 SYNM CEP55 TTK RAP1BL TPGS1 FBRSL1 NME7 Bait

SDCCAG3 CTNNA1 LRRC49 DNAJA2 EPPK1 MPHOSPH9 MTOC components

Bait-Prey interaction

Figure 2-7 Protein proximity interaction network of FLAG-BirA*-USP54. Proximity interactions of USP54 are identified with 13 centriole duplication regulators after applying cutoff of BFDR < 0.02, AvgP > 0.8 and MaxP > 0.85. Large blue nodes represent bait proteins used in the study, while gray nodes represent interacting/proximity preys. interactors represented by small gray nodes. Gray arrows represent bait-prey interactions, and their corresponding edge thickness is aligned to total peptide spectral counts. Pink nodes indicate MTOC proteins identified in the interactome.

The esiRNA screen on 72 candidates the value of the BioID method in combination with functional screens in identifying novel centriole duplication regulators. Dr. Gagan Gupta and I performed a large-scale automated centriole duplication screen to assess 500 proteins (see complete list in appendix 1.1) in U-2 OS S-phase arrested cells using commercial siRNAs (Figure 2-8). The 500 proteins comprise ~30% of the entire centrosome-cilium protein interaction network described in the Cell paper (Gupta, Coyaud, et al., 2015). Multiple negative controls were used, including CEP120, SPICE1, STIL, SAS6 (labeled in green in Figure 2-9).

Cell drops containing corresponding siRNAs in each hydrophobic well

B Non-Target CEP120 si

/Pericentrin

Centrin-2 DAPI/

Figure 2-8 U-2 OS S-phase arrested centriole over-duplication assay. (A) A demonstration of 96-well Teflon plate used in the study. The coverslips are covered with hydrophobic Teflon except for the well areas, keeping the droplets containing cells and esiRNAs separate with each other. (B) Centrioles in U-2 OS cells over-duplicates when the cells are arrested in S-phase. When cells are treated with RNAi targeting centriole duplication factors (as shown here for CEP120), the number of centrioles indicated by Centrin foci will be less compared to non-target. Cells were co-stained with anti-Centrin serum (green) and anti- 64

Pericentrin (red) antibodies. DAPI was used to stain DNA (blue). Centrosome areas are magnified 2.5X for visibility. Scale bar: 15 μm.

The result analysis was conducted in a double-blind manner (see Section 2.6.5 for details). For each well that represented each RNAi condition, I counted the percentage of cells with more than 4 centrioles. Dr. Gupta calculated Z-scores for each RNAi condition using the c−cμ following formula: 푧 = , where c-cμ represents the differences between over-duplication cSD level for any given test conditions and mean over-duplication level of cells transfected with non- targeting siRNA (Gupta, Coyaud, et al., 2015). Out of 500 proteins assessed, there were 122 candidates suppressed centriole amplification (with a Z-score < -2), and 55 of these hits were in CCDB (Gupta, Coyaud, et al., 2015) (Figure 2-9). Of the proteins previously identified in our pilot screen, USP54 and TRIM36 emerged as “hits” in this large-scale screen, indicative of the robustness of both screens. Other interesting “hits” include CKAP proteins (Cytoskeleton- Associated Protein), for example CKAP2 and CKAP4. Previous work has shown that CKAP2 maintains integrity of centrosome and MTOCs after being phosphorylated by CDK1 (Tsuchihara et al., 2005; Seki and Fang, 2007; Case et al., 2013).

Notably, TRIM36 yielded a phenotype in all three functional screens (centriole duplication, ciliogenesis, and satellite dynamics) as described in our publication (Gupta, Coyaud, et al., 2015). There were in total 30 proteins out of 500 yielded a phenotype in all three screens, suggesting their core functions in centrosome biological processes (Gupta, Coyaud, et al., 2015). The evidences supported the rationale for performing follow up studies on TRIM36 to be described in the next chapter. The screen also provided a large pool of candidates that potentially involved in centriole amplification process for follow up studies.

1 Z-score 0

-1 e

r -2 o

sc -3

- Z

-4 TRIM36 USP54 -5 Positive controls: SPICE1 and CEP120 Negative control: Non-target -6

Figure 2-9 Z-score distribution of centriole over-duplication screen. The number of cells with centriole over-duplication was shown as a percentage of total cells counted for each RNAi condition. Z-scores are calculated using the following formula: 푧 = c−cμ, cSD where c-cμ represents the differences between over-duplication level for any given test conditions and mean over-duplication level of cells transfected with non-targeting siRNA

(Gupta, Coyaud, et al., 2015). Csd represents the standard deviation of over-duplication among cells transfected with non-targeting siRNA. Negative controls, including CEP135, STIL, SPICE1 and CEP120, are labeled in green, whereas positive control is labeled in yellow.

2.5 Discussion

Of the 1644 unique preys identified in our centriole duplication BioID interactome, 302 proteins belong to CCDB, our benchmark database, highlighting the confidence and robustness of our proximity interaction map. This served as a great starting pool for subsequent screening and characterization experiments to identify novel centriole duplication regulators. Together with Dr. Gupta, I conducted a semi-automated high-throughput centriole duplication screen using S- phase arrested U2-OS cells that have been sensitized for duplication detection (Balczon et al.,

1995b). The main readout was centriole over-duplication defects, leading to identification of proteins that positively regulate the centriole duplication process.

It is important to recognize that there are two caveats associated with the methodology. First, the centriole duplication screen was performed under a sensitized environment. Whether the positive centriole duplication regulators identified also impact de novo centriole assembly needs to be further characterized. The next steps include performing similar screens using HeLa cycling cells and PLK4 assay (detailed in Chapter 3). CRISPR knockout screens focusing on the most promising candidates can be used for validation (F Ann Ran et al., 2013; F. Ann Ran et al., 2013). In addition, a systematic localization screen can be performed on the potential centriole duplication factor candidates.

The second caveat is that the centriole duplication screen does not provide as much information on negative centriole duplication regulators. The number of studies of negative regulators are outnumbered by those of positive regulators. More detailed knowledge on the roles of negative regulators is important to fully understand the centrosome amplification process.

I have identified USP54 as a positive centriole duplication regulator, as its depletion led to centriole duplication defects. The DUB is overexpressed in intestinal cancer stem cells (CSCs) and its depletion reduced the tumorigenesis of colon cancer cells (Fraile et al., 2016). It has also been associated with human leukemia (Xiao et al., 2016). Although the underlying mechanism remains to be determined, the finding is consistent with the hypothesis that centriole amplification contributes to tumorigenesis. USP54 contains an inactive ubiquitin carboxyl- terminal hydrolase domain and a microtubule interacting and trans-port (MIT) domain (Rigden et al., 2009). The importance of these domains in relation to the regulation of centriole duplication remains to be determined.

In BioID, the biotinylation reaction is performed over a 24 hours period because of the extremely slow labeling kinetics (Roux et al., 2012; Morriswood et al., 2013). Recently, more enzyme-catalyzed proximity labeling methods with shorter labeling time have been explored to study the spatial and temporal PPIs of endogenous proteins in living cells. An engineered ascorbate peroxidase (APEX) was developed for proteomic labeling (Rhee et al., 2013). APEX is a monomeric 28 kDa peroxidase that oxidizes numerous phenol derivatives to short-lived 67

phenoxyl radicals in the presence of H2O2 (Martell et al., 2012; Rhee et al., 2013). With a time frame as short as 1 min, endogenous proteins can be biotinylated by APEX in reactions containing biotin-phenol and H2O2 (Rhee et al., 2013). TurboID and miniTurbo are also novel enzyme-catalyzed proximity labeling approaches with a short labeling time (~10 min) that have the benefit of offering significantly higher efficiencies compared to BioID (Branon et al., 2018). In TurboID, the original BirA enzyme is mutated at 15 sites; and in miniTurbo, the N-terminal of the original BirA enzyme is deleted in addition to mutations at 13 sites (Branon et al., 2018). These methods combined with MS are powerful tools for the identification of dynamic protein interactions during centriole duplication (Rhee et al., 2013; Branon et al., 2018).

2.6 Material and Methods

2.6.1 Cell lines and Tissue culture Flp-In T-REx HEK 293 cell lines were cultured at 37ºC in Dulbecco’s modified Eagle’s medium containing 4.5 g/L D-glucose, 110 mg/L sodium pyruvate and L-glutamine, and supplemented with 10% tetracycline-free fetal bovine serum and 1% GlutaMAX (all tissue culture reagents were from Invitrogen). U-2 OS cells were grown in McCoy 5A medium supplemented with 10% FBS and 1% GlutaMAX. To generate pcDNA5-FLAG-BirA* fusion plasmid, coding sequences of proteins of interest were amplified by PCR using primers containing appropriate restriction sites. PCR products were digested and cloned into the pcDNA5-FLAG-BirA* backbone. The plasmids were used to co-transfect Flp-In T-REx HEK 293 cells with pOG44 (Flp-recombinase expression vector) under the presence of Lipofectamine® 2000 (Invitrogen). Transfected cells were selected with the Puromicin. Cell lines were characterized through immunofluorescence and Western blotting. Cells were incubated in1μg/ml tetracycline (Sigma-Aldrich) and 50μM biotin (BioShop, Burlington, ON, Canada), 1μg/ml tetracycline only or no tetracycline nor biotin for 24 hours. Biotinylated proteins were detected using HRP-conjugated streptavidin (1:50000; Invitrogen) or and Alexa Fluor™ 647 Streptavidin conjugate (Invitrogen).

2.6.2 Sample preparation for BioID For each technical replicate, cells were grown on 5 × 150 cm2 plates until sub-confluent (80%). Cells were incubated in complete media with 1 μg/ml tetracycline (Sigma) and 50 μm biotin for 24h (BioShop, Burlington, ON, Canada). Cells were collected and pelleted at 2000 rpm for 5 min at 4°C. Pellets were washed twice with PBS, and dried pellets were frozen at -80°C. Cell pellets were resuspended and lysed in 10ml modified RIPA lysis buffer (50 mm Tris- HCl pH 7.5, 150 mm NaCl, 1 mm EDTA, 1 mm EGTA, 1% Triton X-100, 0.1% SDS, 1:500 protease inhibitor mixture (Sigma-Aldrich, Saint-Louis, MO), 250U Turbonuclease (Accelagen, San Diego, CA)) and rotated at 4°C for 1 hour, followed by brief sonication (30 s at 35% power, Sonic Dismembrator 500; Fisher Scientific) to disrupt any visible aggregates, then centrifuged at 16,000 rpm for 30 minutes at 4°C. Supernatant was incubated a fresh 15ml conical tube with 30μl of packed, pre-equilibrated Streptavidin sepharose beads (GE) for 3 hours at 4°C with end-over- end rotation. Beads were pelleted by centrifugation at 2000 rpm for 2 minutes and transferred with 1ml lysis buffer to a fresh Eppendorf tubewashed six times with 50 mm ammonium bicarbonate pH 8.3, and treated with TPCK-trypsin for tryptic digestion overnight (Promega, Madison, WI, 16 h at 37 °C). The supernatant containing the tryptic peptides was collected and lyophilized. Peptides were resuspended in 0.1% formic acid and 1/5th of the sample was analyzed per MS run.

2.6.3 BioID followed by mass spectrometry All mass-spectrometry was performed by Dr. Etienne Coyaud and Dr. Brian Raught as previously described (Gupta, Coyaud, et al., 2015). LC analytical columns (75μm inner diameter) and pre-columns (100μm inner diameter) were made in-house from fused silica capillary tubing from InnovaQuartz (Phoenix, AZ) and packed with 100Å C18-coated silica particles (Magic, Michrom Bioresources, Auburn, CA). Lyophilized samples from the previous step were analyzed by mass spectrometry. Digested peptides were subjected to nanoflow liquid chromatography - electrospray ionization - tandem mass spectrometry (nLC-ESI-MS/MS) with a Proxeon EASY-nLC pump in-line with a hybrid linear quadrupole ion trap (Velos LTQ) Orbitrap mass spectrometer (Thermo Fisher Scientific, Waltham, MA) using a 90 minute reversed phase (10-40% acetonitrile, 0.1% formic acid) buffer gradient running at 250nL/min. Thermo.RAW files were converted to the .mzXML

format using Proteowizard for protein identification (Kessner et al., 2008) then searched against Human RefSeq Version 45 (appended with a reversed decoy database based on RefSeq v45) using the X!Tandem search engine (Nesvizhskii et al., 2003; Craig and Beavis, 2004).

2.6.4 SAINT analysis For proteins identification, Protein Prophet cut-off score of 0.85 was chosen, and proteins identified were analyzed with SAINT v. 3.3 with the following settings: total peptide >2, nburn 2,000, niter 5,000, lowMode 0, minFold 1, normalize 0 (Choi et al., 2011, 2012; Gupta, Coyaud, et al., 2015). For each regulator in the BioID, at least four replicates (two biological replicates, each having two technical replicates) were analyzed against 10 control runs (293 T-REx FLAG- BirA* only (Gupta, Coyaud, et al., 2015). Control runs were collapsed to the highest 4 spectral counts for each protein ID. The final probability scores were computed for each pair of bait and prey, and a BFDR cutoff of 0.02 was applied (Gupta, Coyaud, et al., 2015). Only interactions with an average SAINT score (AvgP) > 0.8 and maximum SAINT score (MaxP) >0.85 were chosen to build the interactome. Common background proteins (e.g. histones and ribosomal subunits) were removed manually (Gupta, Coyaud, et al., 2015) before export to ProHitsWeb (http://prohits-web.lunenfeld.ca/). The interaction map was visualized using Cytoscape 3.6.1 to generate a self-organized, prefuse force-directed topology map with spring embedded (Shannon et al., 2003).

2.6.5 U-2 OS S-phase arrested centriole over-duplication assay U-2 OS cells were seeded on 96-well coverslips (G-slides) coated with hydrophobic teflon mentioned above, then reverse transfected with Dharmacon ON-TARGETplus siRNA SMART pools (3,000 cells/well; 60nM siRNA) for 24 hours. The cells were then flooded with McCoy's 5A (Modified) media with 1.6μg/ml aphidicolin (Sigma-Aldrich) for 48 hours. The coverslips were fixed in ice-cold methanol for 30 min, then incubated with antibodies targeting Centrin-2 and PCNT. To achieve unbiased samples, automated imaging on Deltavision Elite DV system (GE Healthcare) is conducted first using just PCNT signal (20x 0.85 NA) for obtaining coordinates of centrosomes with MATLAB. The coordinates were used to generate a point list and automated imaged under 100x 1.4NA lenses, taking 30 Z-stacks of 0.2µm in all channels

(DAPI, Centrin, Pericentrin). Images were deconvolved and projected, then centrioles were counted manually in a double-blind manner with three replicates.

2.6.6 Statistical methods Z-scores for each RNAi condition in the U-2 OS S-phase arrested screen was calculated c−cμ using the following formula: 푧 = , where c-cμ represents the differences between over- cSD duplication level for any given test conditions and mean over-duplication level of cells transfected with non-targeting siRNA (Gupta, Coyaud, et al., 2015).

Chapter 3. TRIM36 is a novel regulator of centriole duplication and microtubule organization factor

TRIM36 is a novel regulator of centriole duplication and a microtubule organization factor

3.1 Statement of Contributions

My contribution to the work detailed in this section includes conducting all the experiments, acquiring and analyzing all the data with the following exceptions:

Dr. Yi Luo conducted CRISPR knock-in experiment to build an endogenously RNAi- resistant cell line for TRIM36.

Dr. Gagan D. Gupta from Pelletier lab performed the centriole over-duplication assay, conducted IF image processing and statistical analysis shown in this chapter.

Zhiqin Li and Dr. Nero Thevakumaran from Frank Sicheri lab (The Lunenfeld- Tanenbaum Research Institute) performed virus transfection in insect cells to express proteins for purification.

. Dr. Johnny Tkach constructed FLAG-AAAGGSG-Ubiquitin fusion protein and expressed it in HEK293 Flp-In T-REx cells, which were used as control cells in the following mass spectrometry experiment performed by Dr. Etienne Coyaud in Brian Raught Lab.

3.2 Summary

In mammalian cells, two centrosomes are present and each organizes one of the spindle poles during mitosis. Structural centrosome defects and aberrant centriole numbers can cause monopolar and multipolar spindles, which can cause aneuploidy and contribute to cancer initiation / progression (Prosser and Pelletier, 2017; Nigg and Holland, 2018). The microtubule nucleation 73

ability of the centrosome is majorly due to the centrosomal recruitment of the γ-tubulin ring complex (γ-TuRC). Y-TuRC contains γ-tubulin and γ-tubulin complex proteins (GCP) that serve as an assembly template for microtubule formation by anchoring and stabilizing the minus end of microtubules in eukaryotic cells (Moritz et al., 1995).

Although the molecular mechanisms that drive the recruitment of γ-tubulin to the centrosome and centriole assembly have been intensively studied, these processes still remain incompletely understood. The ways centriole proteins are regulated by other factors such as post- translational modifications also remain to be fully characterized. As mentioned in the previous chapter, we have previously established a protein-protein proximity interaction network at the centrosome-cilia interface using BioID (Roux et al., 2012) in order to identify novel regulators of centrosome function. This large scale proteomic study included three functional screens that revealed new factors that control centriole duplication, cilia assembly, and centriolar satellite distribution (Gupta, Coyaud, et al., 2015). We identified 30 genes out of 500 that affect all screen parameters, including TRIM36 (Tripartite Motif-Containing protein 36), a RING-type E3 ligase that was reported to regulate microtubule plus-end growth and vegetal microtubule formation in Xenopus (Cuykendall and Houston, 2009b; Olson, Oh and Houston, 2015).

In this chapter, I characterize the characterization of TRIM36 as a novel centriole duplication and microtubule organization factor that facilitates γ-tubulin recruitment to the centrosome in detail. TRIM36 depletion also leads to chromosome congression defects. Using affinity-based purification coupled with mass spectrometry and co-Immunoprecipitation assays, I characterized the interaction between TRIM36 and γ-tubulin, and show that they both mediate SAS6 recruitment to the centrosome in a STIL-independent manner.

3.3 Introduction

3.3.1 TRIM36 participates in centrosome biological events

TRIM36 was initially identified as a proximity interactor of SPICE1 by BioID. It belongs to a functional cluster identified in the proximity interactome described in Chapter 2 (Gupta,

Coyaud, et al., 2015). Proteins in the cluster segregate around CEP120 and SPICE1, two major centriole duplication and elongation proteins previous described by Dr. Comartin, a former student from the Pelletier laboratory, (Comartin, Gagan D. Gupta, et al., 2013). Dr. Comartin showed that CEP120 and SPICE1 controlled centriole length through interaction with CENPJ (Archinti et al., 2010; Comartin, Gagan D. Gupta, et al., 2013). CEP120 over-expression leads to centriole elongation while depletion of CENPJ/CEP120/SPICE1 reduced centriole length (Comartin, Gagan D. Gupta, et al., 2013). He also confirmed that the localization to the procentriole assembly region (PCAR) of CEP135 is dependent on CENPJ/CEP120/SPICE1, a PLK4-rich area where new centrioles begin to form (Comartin, Gagan D. Gupta, et al., 2013). Several proteins identified in the CEP120/SPICE1 interactor module have been validated as centriole duplication regulators in our RNAi U-2 OS S-phase arrested centriole duplication screen described in Chapter 2. This includes Ankyrin 2 (ANK2) and TRIM36 (Gupta, Coyaud, et al., 2015). ANK2 has been shown to interact exclusively with CEP120 among known centrosome proteins, and its depletion inhibited centriole duplication whereas it’s over- duplication induced by PLK4 over-expression (Gupta, Coyaud, et al., 2015). In this chapter, I investigated the other protein, TRIM36, and demonstrated that it is also a crucial centriole duplication factor.

3.3.2 Using PLK4-Induced Centriole Over-duplication in combination with Super-Resolution Imaging to Delineate the Centriole Assembly Pathway.

Centrosomes duplicate once and only once per cell cycle thereby ensuring that two centrosomes are present in mitosis to organize the two spindle poles, and their numbers are strictly regulated by a set of proteins in a canonical duplication cycle that can be divided into ordered steps (Nigg and Holland, 2018). Various systems have been developed to delineate this complex process, including the centriole duplication assay described in the last chapter. Another system using PLK4 over-expression which induces centriole over-duplication in U-2 OS cells (Kleylein-Sohn et al., 2007) can effectively amplify centrosome components recruitment to the centrosome region, thus raising the number of procentriole per parent from one to multiple 75

(Kleylein-Sohn et al., 2007). The additional procentrioles form a flower-like pattern around, reducing the possibility of detection failures of procentrioles (hereafter the ‘PLK4 assay’).

Three dimensional structured illumination microscopy (3D-SIM) based super-resolution microscopy (Gustafsson et al., 2008) can remove the limitation of traditional deconvolution imaging and provide deeper layers and higher orders of structural information of centrosomes and centrosome protein organization (Lawo et al., 2012a). Compared to the conventional wide- field microscopy, 3D-SIM generates a 3 dimensional reconstruction with doubled spatial resolution (±120 nm in the x-,y-direction and ±300 nm in the z-direction) allowing higher organizational properties orders of cellular structures to be defined (Gustafsson et al., 2008; Lawo et al., 2012a; Mennella et al., 2012; Sonnen et al., 2012). In this chapter, I describe the recruitment dependency data based on the automated image analysis approach combining 3D- SIM and PLK4 assay developed by Dr. David Comartin and Dr. Gagan Gupta from L. Pelletier Lab, to quantify the fluorescence intensity of centriole protein signal in the region surrounding the mother centriole overlapping the Myc-PLK4 signal (Comartin, Gagan D. Gupta, et al., 2013). The region was coined the ProCentriole Assembly Region (PCAR). The Myc-PLK4 signal was used as a ‘mask’ to define the area in which the fluorescence intensity of the centriole protein was quantified. The high-resolution of 3D-SIM enables differentiation between the centriole protein localization within or outside the PCAR. In summary, 3D-SIM combined with PLK4 assay offers insights in details of centriole protein recruitment to the pro-centriole generation sites.

3.4 Results

3.4.1 Generation and characterization of TRIM36 antibodies reveal it’s a centrosomal protein.

Due to unavailability of a good commercial TRIM36 antibody at the time when this study began, I generated a custom anti-TRIM36 serum using purified TRIM36 full-length protein. The expression of a recombinant GST-TRIM36 in Sf9 insect cells was performed by Zhiqin Li from Frank Sicheri lab. Insect cells expressing GST-TRIM36 were harvested and lysed in lysis buffer

(detailed in Section 3.6.5) followed by homogenization and sonication. After centrifugation at 19000 RPM for 30 min at 4°C. The supernatant was incubated with glutathion-sepharose beads for 30 min at 4°C. GST tag was cleaved overnight at 4°C by the 27kD TEV protease (Sigma- Aldrich). Eluted purified protein from the beads was fractionated using Fast protein liquid chromatography (FPLC) on an equilibrated Superdex 200 10/300 GL size exclusion column which separate proteins based on their molecular weight (Figure 3-1.).

Two main peaks indicating proteins with the highest concentration appeared at Fraction 10 (800mAu (280nm)) and 12 (200mAu (280nm)) (Figure 3-1). The two peaks correspond to protein sizes of approximately 85kD and 400kD respectively for fraction 12 and fraction 10 (Healthcare, 2002). The molecular weight of TRIM36 is 83kD, thus the FPLC results suggested the existence of self-multimerization apart from oligomers. To confirm that the purified protein corresponds to TRIM36, I have run the purified protein (fraction 12) on an 8% SDS- Polyacrylamide gels subjected to electrophoresis (SDS-PAGE), which confirmed the presence of a band at the expected size of ~83kD. An additional band with a slightly lower molecular weight of ~75kD was also obtained (Figure 3-2A). I collected the bands separately, and sent the samples in the SDS-PAGE gel to the Sickkids Mass Spectrometry Facility for identification with mass spectrometry. The results were analyzed and visualized using Scaffold application (Proteasome Software). The identified proteins, their molecular weights, and spectral counts are reported (Figure 3-2B). TRIM36 protein was enriched in both band samples with the highest spectral counts, indicating that the top band corresponds to purified TRIM36, while the bottom band might contain fragmented TRIM36 and additional contaminants such as HSP72 (Figure 3-2B).

Figure 3-1. Purified TRIM36 was analyzed using Fast protein liquid chromatography (FPLC). Chromatogram showing two rising from a baseline was drawn on sequence of fractions axis. Peaks with 800mAu (280nm) and 200mAu (280nm) appeared at fraction 10 and 12, indicating the existence of oligomers, dimers and multimers. Fractions were also analyzed on SDS-page gels, and the ones with right size proteins were collected and concentrated to 500mL volume. Before analyzed by FPLC, GST tag was cleaved from GST-TRIM36 fusion protein overnight at 4°C by the 27kD TEV protease (the lower band).

Figure 3-2. Purified TRIM36 confirmation by mass spectrometry. (A) Purified protein (before and after TEV cleavage of GST tag) was collected and applied on an 8% SDS-PAGE gel. The size of uncleaved GST-TRIM36 was ~110-120kD (83kD+28kD), and there was an obvious shift after cleavage. Two separate bands (indicated by red arrows) at the right size (~83KD) was analyzed using mass spectrometry in (B). (B)The data was visualized using Scaffold application. The identified proteins, their molecular weights, and spectral counts are shown in the figure.

The purified TRIM36 protein was used as an antigen for rabbit immunization to generate custom polyclonal antibody by Covance Research Products, Inc. I tested the resulting immune sera to confirm presence of anti-TRIM36 antibodies using Western blotting and immunofluorescence imaging.

Western blotting results showed that the serum was able to detect endogenous TRIM36 in HeLa cells (Figure 3-4A). Immunofluorescence imaging demonstrated that TRIM36 signal is

localized to the centrosome following TRIM36 serum staining, and the signal went away (~80% decrease) following TRIM36 depletion using a deconvolved siRNA (Invitrogen) (Figure 3-4B, C). Pericentrin signal was used as a ‘mask’ to define the centrosome region, in which TRIM36 signal was quantified using ImageJ (https://imagej.nih.gov/ij/).

arget

Non- TRIM36 si 95

72 *

B DAPI TRIM36 Pericentrin Merge C

20000

Non-Target

y at centrosome

t i

s 10000 n

e ***

si M36

I 0

R i TRIM36 T s 6 3 IM R T Non-Target

Figure 3-3. Endogenous TRIM36 localizes to centrosomes. (A) Western blot of whole cell lysates of HeLa cells after 72 hours RNAi treatment with non- target siRNA or siRNAs against TRIM36. The top band represents TRIM36 (~83 kD), which goes away with TRIM36 depletion. Asterisk represents unspecific band. (B) Representative images (A) and (B) quantification of centrosomal signal intensity of TRIM36 of fixed HeLa cells co-stained with anti-TRIM36 serum (green) and anti-Pericentrin (red) antibodies. DAPI was used to stain DNA (blue). Boxed insets show 3x magnified centrioles as labeled by TRIM36 and Pericentrin. Scale bar: 15 μm. (***p<0.001, **p<0.01, p<0.05 by Student’s t-test, n=50, mean ± SEM)

3.4.2 Proximity-dependent BioID combined mass spectrometry approach maps TRIM36 proximity interactors.

Since we observed that TRIM36 is localized in the centrosomes, we performed BioID analysis of TRIM36 to map its proximity interactors and to get a better understanding in its role at the centrosome (Roux et al., 2012). This work was done in collaboration with Dr. Étienne Coyaud in Dr. Brian Raught lab at the Princess Margaret Cancer Centre in Toronto. I generated a HEK293 T-REx cell pools expressing Tet-inducible FLAG-BirA*-TRIM36 protein, and characterized the cell line by assessing protein expression and biotinylation using Western blotting and immunofluorescence imaging (Figure 3-5A, B). The characterized cell line expressing moderate amounts of FLAG-BirA*-TRIM36 were harvested for BioID and SAINT analyses. Four replicates (two biological replicates, each having two technical replicates) were analyzed. Dr. Etienne Coyaud performed mass spectrometry experiments. Probability scores were computed for each pair of TRIM36 and its prey, then a BFDR cutoff of 0.02 was applied (Gupta, Coyaud, et al., 2015). Only interactions with an average SAINT score (AvgP) > 0.9 and maximum SAINT score (MaxP) >0.95 were chosen to build the interactome.

By comparing BioID results for cells expressing FLAG-BirA*-TRIM36 to cells expressing the FLAG-BirA* tag alone using SAINT analysis as described in Chapter 2, we identified multiple known centriole duplication factors and microtubule-associated proteins. Identified proteins included MAP7D3, KIF14, and NEDD1 (Haren et al., 2006; Yadav, Verma and Panda, 2014; Reilly et al., 2019). This suggests potential role for TRIM36 in centriole duplication and microtubule regulation (Figure 3-5C, Table 3-1).

A B Flag-BirA*-TRIM36 Tet - + + Flag-BirA*-TRIM36 Biotin - - + Tet - + + Biotin - - +

130 DAPI WB: FLAG 95

250 Flag

130 WB: 95 Streptavidin

72 Streptavidin 55

Pericentrin

Merge

C SEPT11 FBXW11 CAMSAP1

HAUS4 MAP7D3 CDK18 DLGAP5 TTK

HAUS5 MAPT CCDC34 HAUS3 MAP7 HAUS7 SEPT6 46.25 LUZP1 MAP1S 4.5 75.0

CCDC1242.75 128.25 KIF14 4.0 22.0 8.5 GTSE1 8.5 CKAP5 CEP44 MAP9 12.25 7.0 10.25 KIF2A 20.5 30.25 4.5 KIF1B NEDD1 14.0 180.5SKA1 MAP1B 13.0 53.5 87.0 MAP7D1 SEPT9 164.0 33.25 8.5 10.0 CDK2 19.25 4.25 7.0 53.25 34.75 76.25 3.25 FlagBirA- KIF2C10.25 CDC27 CDC37L1 SEPT2 50.5 MRPL27 7.5 14.5 8.75 7.75 23.5 MTUS1 PLK1 5.25 C15orf23 TRIM36 44.75 1.0 23.5 3.0 3.25 264.75 8.25 24.75 22.75 43.75 2.0 CEP413.75 TOP2A KIAA0284 MARK2 10.75PAK4 CKAP2 38.5 244.75 4.25 36.25 52.0 NOTCH2 2.75 11.0MAPRE2 NRL 82.25 11.75 263.5 CDC16 MAPRE3 20.5 24.25 HAUS8 18.25 SEPT7 27.25 18.5 SKA2 CEP170 HAUS1 CAMSAP2 SKA3 HAUS2 NCL

CKAP2L MAP7D2 MAP4 CLASP1 Identified centriole duplication regulators SLC25A4

CCNB1 Mircotubule related proteins

HAUS6 SEPT8 CKB BioID Bait to Prey 31 Average Spec

Figure 3-4. BioID of FLAG-BirA*-TRIM36 identifies proximity interactors enriched in centriole duplication pathway and microtubule organization. Validating expression of FLAG-BirA*-TRIM36 and biotinylation in HEK293 T-REx cells by Western blot and immunofluorescence imaging. (A) Western blot using anti-FLAG (upper panel) and anti-Streptavidin (lower panel) of total extracts of HEK293 T-Rex expressing Tet-inducible FLAG-BirA*-TRIM36 under following conditions: ± Tet (1 µg/mL, 24 h) and ± Biotin (50 µM, 24 h). Note the biotinylation of TRIM36 only exists in the presence of tetracycline and biotin. (B) Immunofluorescence staining of cells described in (A) revealed FLAG-BirA*-TRIM36 localizes on microtubules in HEK293 T-REx cells. HEK 293 T-REx cells stably expressing Tet- inducible FLAG-BirA*-TRIM36 fusion protein were labeled with anti-FLAG, anti-Pericentrin and Alexa Fluor 647-coupled Streptavidin in cycling cells. (C) Proximity interactions of TRIM36 are identified with 13 centriole duplication regulators after applying cutoff of BFDR < 0.02, AvgP > 0.9 and MaxP > 0.95. Edge thickness is proportional to the total peptide counts.

Prey Ctrl Spec Counts SpecSum AvgSpec AvgP MaxP SaintScore Gene Counts CKAP2 249|271|287|252 1059 264.75 0|0|0|0 1 1 1 MAP4 231|227|297|299 1054 263.5 46|47|38|50 1 1 1 CEP170 208|230|275|266 979 244.75 12|18|13|9 1 1 1 LUZP1 172|175|191|184 722 180.5 2|2|0|0 1 1 1 CKAP5 168|171|158|159 656 164 10|9|10|8 1 1 1 MAP7D3 106|98|152|157 513 128.25 4|4|2|3 1 1 1 GTSE1 84|94|88|82 348 87 0|0|0|0 1 1 1 CKAP2L 88|94|71|76 329 82.25 0|0|0|0 1 1 1 SEPT9 69|78|81|77 305 76.25 8|3|6|5 1 1 1 CAMSAP1 53|54|96|97 300 75 0|0|0|0 1 1 1 KIF14 38|34|73|69 214 53.5 0|0|0|0 1 1 1 MAP1B 61|59|42|51 213 53.25 13|14|8|6 1 1 1 SKA3 43|43|64|58 208 52 4|2|0|1 1 1 1 SEPT2 46|45|55|56 202 50.5 10|4|11|10 1 1 1 SEPT11 39|42|55|49 185 46.25 4|3|4|5 1 1 1 SEPT7 35|44|51|49 179 44.75 7|4|5|8 1 1 1 NRL 60|84|13|18 175 43.75 0|0|0|0 1 1 1 SKA2 43|60|24|27 154 38.5 1|1|0|0 1 1 1 CAMSAP2 22|26|49|48 145 36.25 0|0|0|0 1 1 1 83

MAP7D1 23|27|47|42 139 34.75 0|0|0|0 1 1 1 KIF2A 33|30|38|32 133 33.25 0|0|0|1 1 1 1 MAP7 25|26|34|36 121 30.25 0|0|0|0 1 1 1 SEPT8 21|21|33|34 109 27.25 0|0|0|0 1 1 1 NOTCH2 25|28|26|20 99 24.75 0|0|6|0 1 1 1 SLC25A4 0|0|52|45 97 24.25 0|11|0|0 1 1 1 MTUS1 13|17|31|33 94 23.5 0|0|0|0 1 1 1 TOP2A 46|47|0|1 94 23.5 0|0|0|0 1 1 1 MAPRE2 27|24|17|23 91 22.75 4|6|2|3 1 1 1 DLGAP5 23|24|19|22 88 22 0|0|0|1 1 1 1 CLASP1 18|14|25|25 82 20.5 0|0|0|0 1 1 1 SEPT6 24|24|0|34 82 20.5 0|0|0|0 1 1 1 SKA1 11|11|30|25 77 19.25 1|0|1|0 1 1 1 HAUS6 10|14|23|27 74 18.5 0|0|1|6 0.99 1 0.99 CCNB1 16|15|20|22 73 18.25 0|0|0|0 1 1 1 KIF2C 15|15|18|10 58 14.5 0|0|0|0 1 1 1 MAP1S 11|10|16|19 56 14 0|0|0|0 1 1 1 CCDC124 12|17|10|13 52 13 4|2|4|2 1 1 1 MAPT 11|7|17|14 49 12.25 0|0|0|0 1 1 1 MAP7D2 6|7|16|18 47 11.75 0|0|0|0 1 1 1 NCL 20|20|2|2 44 11 3|3|4|5 1 1 1 HAUS8 6|6|14|17 43 10.75 0|0|0|0 1 1 1 CDC27 10|10|9|12 41 10.25 0|0|0|0 1 1 1 HAUS3 5|4|15|17 41 10.25 0|0|0|0 1 1 1 MAP9 10|8|12|10 40 10 0|0|0|0 1 1 1 SKP1 6|8|12|10 36 9 1|2|1|1 1 1 1 CDC37L1 0|8|11|16 35 8.75 0|0|0|0 1 1 1 CEP44 6|8|11|9 34 8.5 0|0|0|0 1 1 1 HAUS5 7|6|8|13 34 8.5 1|1|0|0 1 1 1 TTK 6|7|10|11 34 8.5 2|1|0|0 1 1 1 PAK4 9|10|7|7 33 8.25 0|0|0|0 1 1 1 PLK1 7|8|6|10 31 7.75 1|0|0|0 1 1 1 MRPL27 15|13|0|2 30 7.5 0|0|0|0 1 1 1 CCDC34 0|0|13|15 28 7 0|0|0|0 1 1 1 NEDD1 3|1|13|11 28 7 0|0|0|0 1 1 1 CKB 15|8|0|0 23 5.75 0|0|1|1 1 1 1 C15orf23 8|7|3|3 21 5.25 0|0|0|0 1 1 1 FBXW11 0|0|9|9 18 4.5 0|0|0|0 1 1 1 HAUS7 2|0|8|8 18 4.5 1|1|0|0 1 1 1 HAUS1 0|0|7|10 17 4.25 0|0|0|0 1 1 1 KIF1B 0|2|11|4 17 4.25 0|0|0|0 1 1 1 CDK18 0|7|5|4 16 4 0|0|0|0 1 1 1

MAPRE3 6|6|3|0 15 3.75 0|0|0|0 1 1 1 CDK2 0|0|6|7 13 3.25 0|0|0|0 1 1 1 MARK2 0|0|6|7 13 3.25 0|0|0|0 1 1 1 KIAA0284 0|0|7|5 12 3 0|0|0|0 1 1 1 HAUS2 1|1|4|5 11 2.75 0|0|0|0 1 1 1 HAUS4 1|2|5|3 11 2.75 0|0|0|0 1 1 1 CDC16 4|4|0|0 8 2 0|0|0|0 1 1 1 CEP41 2|2|0|0 4 1 0|0|0|0 0.95 0.95 0.95

Table 3-1. TRIM36 Proximity Interactors Detected by BioID. BioID results were collected from two biological replicates/four technical replicates of Tet- inducible FLAG-BirA*-TRIM36 expression in the presence of tetracycline and biotin followed by Streptavidin affinity-purification combined with mass-spectrometry to identify biotinylated peptides. Spectral counts reflect the abundance of a prey polypeptide in the SAINT analysis. (From left to right) Prey Gene: Gene Symbol for identified “prey” proteins; Spec Counts: spectral counts for each “prey” protein in each of the technical replicates; SpecSum: total amount of spectral counts for 4 replicates; AveSpec: the average spectral counts for 4 replicates; Ctrl Counts: the spectral counts for each protein in the 2 control replicates; AvgP: the average probability of the 4 replicates; MaxP: the maximum probability between the 4 replicates; SaintScore: overall SAINT score for 4 replicates (Choi et al., 2011, 2012; Skarra et al., 2011). MS and was performed by Dr. Etienne Coyaud. Spectral counts reflect the abundance of a prey polypeptide in the SAINT analysis.

3.4.3 TRIM36 is a novel regulator of centriole duplication.

In the last chapter, I described two centriole duplication screens, an esiRNAi based pilot centriole duplication screen and an RNAi based centriole duplication screen in U-2 OS cells arrested in S-phase. Since TRIM36 was identified to be a strong positive centriole duplication factor by both screens, I first confirmed the role of TRIM36 in centriole duplication by depleting endogenous TRIM36 in HeLa cells with two deconvolved siRNAs. Both siRNAs led to dramatically reduced TRIM36 protein levels (Figure 3-6A, C). Using Centrin foci as a centriole marker and PCNT as a centrosome marker, I observed significant loss of centrioles in mitotic HeLa

cells transfected with TRIM36 siRNAs (40 to 50% reduction of mitotic cells with four centrioles) (Figure 3-6B).

A B *** *** Non-Target TRIM36 si <4

s 100 =4

ll ce

PCNT

o t

i 50 m

Centrin-2 f

o %

DAPI 0 i s -1 -2 5 i i 3 s s 1 6 6 C P 3 3 E IM IM Non-Target C R R

T T

arget

TRIM36si-1 Non- TRIM36 si-2 95 TRIM36 72 *

α-Tubulin 55

Figure 3-5. TRIM36 is required for centriole duplication in HeLa cells. (A) Representative IF microscopy images of fixed HeLa cells at 72 hours post transfection with non-target siRNA or siRNAs against TRIM36. DAPI was used to label DNA (blue). Cells were labeled for Centrin-2 (green) and Pericentrin (red). Boxed insets show 4x magnified centrioles on each spindle pole. Scale bar: 15 μm. Asterisk represents unspecific band. (B) Bar graph showing the percentage of mitotic HeLa cells processed as in (A) in which either <4 or 4 centrioles were detectable (***p<0.001, **p<0.01, p<0.05 by Student’s t-test, n=50, four replicates per condition, mean ± SEM). (C) Western blot of whole cell lysates of HeLa cells collected from the same experiment showing depletion of TRIM36 protein levels under RNAi conditions.

To rule out potential off-target effects, I expressed a Tet-inducible RNAi-resistant TRIM36 transgene in HeLa T-Rex cells. HeLa cells were treated with control siRNA or TRIM36 siRNA,

followed by transient transfection by either FLAG-BirA* or FLAG-BirA*-TRIM36-resistant form, and labeled with antibodies against Centrin-2 and FLAG. Centrioles were counted by examining Centrin-2 signal number and the percentage of mitotic cells. Analysis monitors centrioles number under each condition and data are reported as % of cells with less than 4 centrioles (Figure 3-7A, B). In the control siRNA treated sample, around 16% of mitotic HeLa cells contain less than 4 centrioles, while in the TRIM36 siRNA treated sample, ~60% mitotic cells contain less than 4 centrioles, demonstrating significant inhibition of centriole duplication under TRIM36 depletion. Over-expression of the siRNA-resistant construct resulted in 25% of mitotic cells containing less than 4 centrioles. This data demonstrated that the rescue experiment in HeLa cycling cells when using siRNA-resistant construct was successful and the inhibition of centriole duplication didn’t occur, confirming that the results obtained with TRIM36 siRNA activity was not an off-target effect (Figure 3-7A, B).

A Non-Target TRIM36 si

Tet (-) Tet (+) Tet (-) Tet (+)

Flag

Centrin-2 DAPI

Non-Target TRIM36 si B Tet (-) Tet (+) Tet (-) Tet (+) ** *** 100 =4

s <4

ll >4

i t

o 50

o % 0

130 WB: Flag

Figure 3-6. TRIM36 is required for centriole duplication in HeLa cells/ (A) Representative IF microscopy images of fixed HeLa cells stably expression of RNAi-resistant FLAG-BirA*-tagged TRIM36 induced by Tet (1 µg/mL, 24 h). HeLa cells were under RNAi treatment for 72 hours with non-target siRNA or siRNAs against TRIM36. DAPI was used to label DNA (blue). Cells were labeled for Centrin-2 (green) and Pericentrin (red). Boxed insets show 5x magnified centrioles on each spindle pole. Scale bar: 15 μm. (B) Bar graph showing the percentage of mitotic HeLa cells processed as in (A) in which either <4, 4, or >4 centrioles were detectable (***p<0.001, **p<0.01, p<0.05 by Student’s t-test, n=50, three replicates per condition, mean ± SEM). (C) Western blot of whole cell lysates of HeLa cells collected from the same experiment showing stable expression of RNAi-resistant FLAG-BirA*-tagged TRIM36 induced by Tet.

Next, I depleted endogenous TRIM36 in S-phase arrested U-2 OS cells with the same two deconvolved siRNAs. Results indicated significant loss of over-duplicated centrioles in U-2 OS cells transfected with TRIM36 siRNAs (~40 to 50% reduction) (Figure 3-7). Lastly, I depleted endogenous TRIM36 in PLK4-induced over-duplication assay in U-2 OS cells, and similarly I obtained significant loss of over-duplicated procentrioles around the mother centriole in U-2 OS cells transfected with TRIM36 siRNAs (~50% reduction) (Figure 3-8). These results confirmed that TRIM36 was requited for centriole duplication.

Figure 3-7. TRIM36 is required for centriole over-duplication in U-2 OS S-phase arrested cells. (A) Representative IF microscopy images of fixed U-2 OS cells at 72 hours post transfection with non-target siRNA or siRNAs against TRIM36 and treated with Hydroxyurea (HU) for 48 hours prior to fixation for the induction of S phase arrest and centriole over-duplication. Cells were stained for DNA (blue), Centrin-2 (green) as a marker of centriole number, and Pericentrin (red) as a centrosome marker. Boxed insets show 4x magnified centrioles on each spindle pole. Scale bar: 15 μm. Asterisk represents unspecific band. (B) Bar graph showing the percentage of S-phase arrested U-2 OS cells processed as in (A) in which either >4 or <=4 centrioles were detectable 89

(***p<0.001, **p<0.01, p<0.05 by Student’s t-test, n=50, four replicates per condition, mean ± SEM). (C) Western blot of whole cell lysates of S-phase arrested U-2 OS cells collected from the same experiment showing depletion of TRIM36 protein levels under RNAi conditions.

Figure 3-8. TRIM36 is required for PLK4-induced centriole over-duplication in U-2 OS S- phase arrested cells. (A) Representative IF microscopy images of fixed U-2 OS cells at 72 hours post transfection with non-target siRNA or siRNAs against TRIM36 and treated with Hydroxyurea (HU) for 24 hours prior to fixation with centrosomes induced to over-duplicate by PLK4 overexpression. Cells were stained for DNA (blue), Centrin-2 (green) as a marker of centriole number, and Pericentrin (red) as a centrosome marker. Boxed insets show 4x magnified centrioles on each spindle pole. Scale bar: 15 μm. Asterisk represents unspecific band. (B) Bar graph showing the percentage of U-2 OS cells over-expressing PLK4 induced by Tet (1 µg/mL, 16h) processed as in (A) in which either >4 or <=4 centrioles were detectable (***p<0.001, **p<0.01, p<0.05 by Student’s t-test, n=50, four replicates per condition, mean ± SEM). (C) Western blot of whole cell lysates of U-2 OS cells 90

over-expressing PLK4 collected from the same experiment showing depletion of TRIM36 protein levels under RNAi conditions.

To rescue the phenotype at the endogenous level of TRIM36, Dr. Yi Luo, a postdoctoral fellow in our lab and I generated a modified U-2 OS cell-line in which endogenous TRIM36 was edited by CRISPR to be resistant to TRIM36 siRNA-1. As previously demonstrated, the human codon–optimized Cas9 nucleases can be re-directed toward any genomic locus followed by a 5’- NGG protospacer-adjacent motif (PAM) sequence by altering the 20-nt guide sequence within the single-guided RNA (sgRNA) (F Ann Ran et al., 2013; F. Ann Ran et al., 2013). The system is a robust and versatile tool for stimulating targeted double-stranded DNA breaks (DSBs) in eukaryotic cells. It mediates genome editing via non-homologous end joining (NHEJ) or homology-directed repair (HDR) in the presence of an exogenously introduced repair template (F Ann Ran et al., 2013).

To maximize HDR efficiency, we used a custom single-stranded DNA oligonucleotide (ssODN) as the repair template for modification within siRNA targeting region (Figure 3-9A) (Paquet et al., 2016). The customized 200-nt antisense ssODN synthesized by IDT contains 90-nt flanking sequences on each side of the 20-nt repair template region including TRIM36 siRNA-1 targeting site. The sequence was mutated to be siRNA-resistant and an XbaI digestion site was introduced for genotyping analysis (Figure 3-9B). The ssODN was co-transfected with sgRNAs cloned into pX458 (a plasmid expressing Cas9) in U-2 OS cells and transfected cells were selected using 1μg/ml Puromycin. The targeted genomic region from single clones was PCR-amplified using primers that anneal outside of the region of homology. PCR product digestion with XbaI confirmed the occurrence of HDR events. Two single clones were selected and one of them was sequenced, demonstrating that it was modified as RNAi resistant. Generated cell-line was then tested for the rescue experiment in U-2 OS over-duplication assay. TRIM36 depletion by siRNA- 1 significantly reduced the percentage of cells with over-duplicated centrioles from ~80% to 30%, but the RNAi had little effect on the mutated cell line, leaving ~70% cells with over-duplicated centrioles after siRNA-1 treatment (Figure 3-9C). The endogenous rescue experiments further

excluded possible off-target effects of the RNAi treatment and validates the specificity of centriole duplication phenotype caused by TRIM36 depletion.

Exon Silent Mutation Positions

siRNA-1 Targeting Region PAM Checking Primer XbaI Cutting Site B C 100 ** *

U-2 OS U-2 OS TRIM36 siRNA-1 Resistant

l e

c 50

f o

Non-TargetTRIM36 si-1TRIM36 si-2Non-TargetTRIM36 si-1TRIM36 si-2 % 95 TRIM36 72 0

α-Tubulin 55

Non-TargetTRIM36 si-1TRIM36 si-2Non-TargetTRIM36 si-1TRIM36 si-2 U-2 OS U-2 OS TRIM36 siRNA-1 Resistant

Figure 3-9. Endogenous rescue experiments in the U-2 OS over-duplication assay. (A) A demonstration of ssODN design. The siRNA-1 targeting sequence was mutated to be siRNA-1-resistant and an XbaI digestion site was introduced for following genotyping analysis. (B) The single clone that contains mutated siRNA-1 targeting site all was sequenced then tested for rescue experiment in U-2 OS over-duplication assay. (C) TRIM36 depletion by siRNA-1 significantly reduced percentage of cells with over-duplicated centrioles to ~80% to 30%, but the RNAi hardly affected the mutated cell line (shown as single-clone #8), leaving ~70% cells with over-duplicated centrioles after siRNA-1 treatment. 92

As a member of TRIM C-I subfamily, TRIM36 contains a conserved N-terminal RBCC motif (Meroni and Diez-Roux, 2005; Short and Cox, 2006), a COS box domain reported to bind microtubule (Short and Cox, 2006), a fibronectin type III (FN3) and a SPRY/B30.2/B30.2-like domain (Balint et al., 2004) (Figure 3-10A). The RBCC motif is conserved across all TRIM family proteins, while the arrangement of COS/FN3/B30.2-like domains are specific to C-I subfamily of proteins (Short and Cox, 2006; Esposito, Koliopoulos and Rittinger, 2017). To identify which domains of TRIM36 are functionally indispensable in the control of centriole duplication, I generated different Tet-inducible FLAG-BirA* tagged TRIM36 fragments resistant to RNAi (Figure 3-10A). The fragments were transiently transfected into HeLa cells treated with control siRNA or TRIM36 siRNA, and their expressions were examined by Western blot (Figure 3-11B). In the control siRNA treated sample, ~10% of mitotic HeLa cells contain less than 4 centrioles, while in the TRIM36 siRNA treated sample, ~60% mitotic cells contain less than 4 centrioles. Over-expression of the siRNA-resistant TRIM36 full length (1-718), RING domain only (1-132), or RBCC domain (1-418) constructs reduced the percentage of mitotic cells containing less than 4 centrioles to ~20%, ~30% and ~30%, respectively, while the C-terminal domains failed to do so (Figure 3-11C). This result showed that the RING finger domain (1-132) and the N-terminal fragment (tripartite/RBCC motif plus COS box domain, 1-418) were sufficient to rescue the centriole duplication defect caused by TRIM36 depletion, but fragments lacking the RING domain could not (Figure 3-10B, C).

A 1 33 118 157 196 212 244 256 356 413 419 510 720 728 FL RING B1 B2 CC COS FN3 SPRY

1-132 RING

1-418 RING B1 B2 CC COS 133-728 B1 B2 CC COS FN3 SPRY 256-728 CC COS FN3 SPRY

ns B C ns

** )

) **

)

728

3 -

- ***

132

418

256

1 1

133 =4

(

( (

( 100

* *

* <4

B B

B >4

t o

t 50 i

250 m f

130 o 95 % 72 0 t ) ) ) ) 2 8 55 ge N.A. FL 28 28 r 13 -41 -7 -7 a - 1 3 6 -T (1 ( WB: Flag n 13 25 o ( ( N TRIM36 si

Figure 3-10. N-terminus (Tripartite/RBCC motif with COS box domain) is crucial to centriole duplication regulation, especially the RING domain. (A) A schematic representation of FLAG-BirA* tagged full-length and four truncation fragments generated for TRIM36. (From top to down) Full-Length (1-728), RING domain (1-132), RBCC domain (1-418), Full-length with RING truncated (133-728), C-terminal with RBB domains truncated (256-728). (B) Western blot of whole cell lysates against FLAG tag showing transient expressions of indicated RNAi-resistant FLAG-BirA*-TRIM36 fragments under 72 hours RNAi treatment with non-target siRNA or siRNAs against TRIM36. (C) Quantification of the percentage of mitotic HeLa cells processed as in (B) in which either <4, 4, or >4 centrioles were detectable in the presence of indicated TRIM36 fragments (***p<0.001, **p<0.01, p<0.05 by Student’s t-test, n=50, three replicates per condition).

Furthermore, I sought to determine whether the catalytic activity of TRIM36 was required for its role in centriole duplication. To do this, I mutated a key Histidine residue to Alanine (H50A) that disrupts the entire RING structure (Meroni and Diez-Roux, 2005; Stone et al., 2005) (Figure 3-11A), and expressed the Tet-inducible FLAG-BirA*-TRIM36 (H50A)-RNAi resistant form in HeLa T-Rex cells (Figure 3-12B). The results indicated that the H50A mutation compromised TRIM36 ability to rescue the centriole duplication defect, since the percentage of mitotic cells containing less than 4 centrioles remained at ~70% after over-expression of the mutated siRNA- resistant construct, similar to the TRIM36 RNAi treated sample (Figure 3-11B, C).

1 33 50 118 157 196212 244256 356 413 419 510 720 728 FL RING B1 B2 CC COS FN3 SPRY

H50A

B Non-Target TRIM36 si

Tet (-) Tet (+) Tet (-) Tet (+)

Flag

Centrin-2 DAPI

C Non-Target TRIM36 si

Tet (-) Tet (+) Tet (-) Tet (+) ns ***

100 >4 s

ll <4

ce =4

o t

i 50

o % 0

130 WB: Flag

Figure 3-11. TRIM36 catalytic activity is required for centriole duplication in HeLa cells.

(A) Schematic sequences of TRIM36 indicating catalytic dead mutation site (red dot). (B) Representative IF microscopy images of fixed HeLa cells stably expression of RNAi-resistant FLAG-BirA*-tagged TRIM36 (H50A) induced by Tet (1 µg/mL, 24 h). HeLa cells were under RNAi treatment for 72 hours with non-target siRNA or siRNAs against TRIM36. DAPI was used to label DNA (blue). Cells were labeled for Centrin-2 (green) and Pericentrin (red). Boxed insets show 5x magnified centrioles on each spindle pole. Scale bar: 15 μm. (C) Bar graph showing the percentage of mitotic HeLa cells processed as in (A) in which either <4, 4, or >4 centrioles were detectable (***p<0.001, **p<0.01, p<0.05 by Student’s t-test, n=50, three replicates per condition, mean ± SEM). (C) Western blot of whole cell lysates of HeLa cells collected from the same experiment showing stable expression of RNAi-resistant FLAG-BirA*-tagged TRIM36 induced by Tet.

As the N-terminal region is a conserved domain of the TRIM family, to verify this regulation is due to TRIM36 RING domain, I also tested TRIM37 RING domain upon TRIM36 depletion and confirmed that the centriole duplication function is specific to TRIM36 (Figure 3- 12A, B). When comparing the RING structure of TRIM36 with other TRIM family E3 ligases, I noticed that the sequence between the 5th and 6th conserved cysteine residues in TRIM36 is significantly longer and potentially forms a distinct loop with unknown functions (Jalview) (Figure 3-12C). Indeed, TRIM36 RING fragments lacking this loop could no longer rescue the centriole duplication defect caused by TRIM36 depletion, suggesting the intact RING finger including this unique loop is crucial to the regulation of centriole duplication (Figure 3-13A, B). However, it should be noted that the RING structure alone would be heavily affected by the loop deletion, thus in the future, I will carry out the rescue experiment using full-length TRIM36 lacking this loop. I then compared the RING sequence among the entire TRIM family, and I identified similar loop structure in only a selected few TRIM proteins, including TRIM9 and TRIM67 and both belonging to the C-1 subfamily as TRIM36 based on domain arrangement similarities (Short and Cox, 2006) (Figure 3-14). The function of the extended loop structure is currently unknown and warrants further investigation.

A B ns

** =4

100 <4

t Centrin-2

i 50

DAPI %

t e 0 g r N.A. t a e -T g n r N.A. o a N Flag-BirA* -T n TRIM36 RING TRIM37 RING o N Flag-BirA* TRIM36 RINGTRIM37 RING TRIM36 si

C TRIM36 si

TRIM36 RING TRIM37 RING Conservation

Quality

FigureConsencus 3-12. RING domain from another TRIM family member, TRIM37, is not sufficient in recuing centriole duplication defect caused by TRIM36 depletion. (A) Representative IF microscopy images of fixed HeLa cells transiently expressing FLAG-BirA*- tagged TRIM36/TRIM37 RING domain respectively induced by Tet (1 µg/mL, 24 h). HeLa cells were under RNAi treatment for 72 hours with non-target siRNA or siRNAs against TRIM36. DAPI was used to label DNA (blue). Cells were labeled for Centrin-2 (green) and Pericentrin (red). Boxed insets show 5x magnified centrioles on each spindle pole. Scale bar: 15 μm. (B) Bar graph showing the percentage of mitotic HeLa cells processed as in (A) in which either <4 or 4 centrioles were detectable (***p<0.001, **p<0.01, p<0.05 by Student’s t-test, n=50, three replicates per condition, mean ± SEM). (C) Schematic peptide sequence conservation comparison between TRIM36 RING and TRIM37 RING domain using Jalview (http://www.jalview.org/), showing the characteristic “Loop” sequence and critical Cysteine and Histidine sites of the RING domain of the TRIM family.

1 33 118 1-132 RING

61 112

B >4 100 <4

s =4

o t

i 50

o %

0 t rge N.A. a -T n o N TRIM36 RING Loopless RING TRIM36 si

Figure 3-13. Intact RING domain is crucial to TRIM36’s function in centriole duplication regulation. (A) A schematic representation of FLAG-BirA* tagged TRIM36 RING domain lacking the loop region. (B) Bar graph showing the percentage of mitotic HeLa cells processed as in (A) in which either <4 or 4 centrioles were detectable. The loopless RING domain fails to rescue the defect in centriole duplication in HeLa mitotic cells. (n=50, two replicates per condition).

A Loop

Figure 3-14. RING domain comparison among TRIM family proteins. Sequence alignment of the RING domain of TRIM36 compared with RING domains from other TRIM proteins using Jalview (http://www.jalview.org/). The conservative residues are denoted by highlights (purple). The sequences corresponding to the loop regions are shown in the alignment.

To determine the stage at which TRIM36 functions during centriole assembly, we investigated the effect of TRIM36 depletion on the recruitment of key centriole duplication factors (CEP152, CEP192, STIL, SAS6, CEP120, SPICE1, CENPJ, CEP135, CP110) (Nigg and Holland, 2018) to the site of daughter centriole assembly. This was carried out using three-dimensional structured-illumination microscopy (3D-SIM) in U-2 OS cells as previously described by our lab 99

(Comartin, Gagan D Gupta, et al., 2013). To achieve this, Dr. Gagan Gupta and I quantified the fluorescence intensity of centriole protein signal in the PCAR (Comartin, Gagan D. Gupta, et al., 2013). The Myc-PLK4 signal was used as a ‘mask’ in which the target centrosome protein signal within that region was quantified. The image analysis, including generation of signal masks and quantifications of signals were performed by Dr. Gupta using MatLab scripts he has previously developed for this assay. The results indicated that the recruitment of CEP152 to the PCAR was not significantly affected by TRIM36 depletion, the PLK4 ring remained intact after TRIM36 depletion, suggesting that TRIM36 functions downstream of PLK4 (Figure 3-15A). CEP192 enrichment, however, was affected upon TRIM36 RNAi treatment, potentially due to its adverse effect on PCM from loss of centrioles (Figure 3-15B).

I observed ~40% increase in STIL intensity at centrosomes (Figure 3-15C), while the centriolar intensity of SAS6 was significantly reduced by ~40% (Figure 3-15D). In addition, the Western blot analysis revealed a marked decrease in SAS6 levels along with TRIM36 depletion, suggesting that SAS6 intensity decrease around PCAR may be a result of its instability after TRIM36 depletion (Figure 3-16). Since the U-2 OS cells were arrested in S-phase in this assay, the instability should not be a result of cell cycle regulation. Previous studies showed SAS6 oligomerized into the 9-fold symmetrical cartwheel structure and was recruited to the proximal lumen of the mother centrioles in a PLK4 and STIL dependent manner to initiate pro-centriole generation (Kitagawa et al., 2011b; Fong et al., 2014). However, The discrepancy of STIL and SAS6 results in my experiments could be indicative of an alternative SAS6 recruitment mechanism that occurs in a STIL independent manner as suggested before (Ito et al., 2018).

100

)

) 4

A B C 4

(

( 192

152 ns P

P 2.0

E E

2.0 STIL 4.0 C

C *** y at Centrosome

Centrin-2 Centrin-2

Centrin-2 y at Centrosome

t 1.0

y at Centrosome

i i

1.0 *** t s

i 2.0

n I

Myc-PLK4 Myc-PLK4

Myc-PLK4 0 0 0 et g et et

Merge Merge Merge r g a 36 si r rg T M a 36 si a 36 si - I T M STIL mean - I -T IM et n R et n R et g mean CEP152 o T g g n R r 36 si N r 36 si mean CEP192 o T r 36 si o T a a N a N -T IM -T IM -T IM

n R n R n R )

o T o T o T 4

N N ) N

) (

D 4 E F

(

120

( P

2.0 ns E

C SAS6 *** CENPJ 3.0 1.0

Centrin-2 y at Centrosome

CP110 t

Centrin-2

2.0 i

y at Centrosome s

t 1.0

y at Centrosome

t e

n ***

e t

1.0 n

I n

Myc-PLK4 I

Myc-PLK4

Myc-PLK4 0 0 0 et

et Merge g Merge r et Merge rg a 36 si rg a 36 si T M a 36 si T M et - I et T M - I g n R g - I et n R r mean CEP120 o T SAS6 mean SAS6 g o 36 si r n R mean CENPJ T a N a 36 si o T r 36 si N T M T M a - I - I N -T IM n R n R n R o T o T o T N

N N

)

4 3

G H 10 I

(

CP110 CEP135 SPICE1 2.0 1.0

1.0

Centrin-2 y at Centrosome

Centrin-2

y at Centrosome

i i

s 1.0 *** s

*** s

e t

t ***

I I

Myc-PLK4

Myc-PLK4 0 0 0 et et et g g Merge r Merge g Merge r a 36 si r 36 si a 36 si M a T M -T I T IM - I n R - n R et T et n R et mean CP110 o T

SPICE1 mean SPICE1 o

g meanCEP135 T rg N r o rg N a 36 si a 36 si N a 36 si -T IM -T IM -T IM n R n R n R o T o T o T N N N

Figure 3-15. TRIM36 is upstream of cartwheel protein SAS6 in centriole assembly pathway. Representative 3D-SIM pseudocolor-merged (left, all bottom panels) and individual gray-scale images showing centrosomes induced to over-duplicate by PLK4 overexpression and depleted of TRIM36 by RNAi in U-2 OS cells. Centrosomes are labeled for Centrin-2 (blue), Myc-PLK4 (red), and CEP152 (green; A), CEP192 (green; B), STIL (green; C), SAS6 (green; D), CEP120 (green; E), SPICE1 (green; F), CENPJ (green; G), CEP135 (green; H), CP110 (green; I). (Right)

101

Quantification of corresponding protein level at centrosomes using Myc-PLK4 as mask. Scale bar: 1 μm. (***p<0.001, **p<0.01, p<0.05 by Student’s t-test, n=20, whiskers represent min to max).

Non-target TRIM36 si SAS6 si SAS6 si

72 SAS6

α-Tubulin 55

Figure 3-16. The total cellular levels of SAS6 decreased upon TRIM36 depletion in S-phase arrested U-2 OS cells. Western blotting analysis of whole cell lysates of U-2 OS cells at 72 hours post transfection with non-target siRNA or siRNAs against TRIM36 and treated with Hydroxyurea (HU) for 24 hours prior to fixation with centrosomes induced to over-duplicate by PLK4 overexpression. The samples were probed with anti-SAS6 antibody, showing SAS6 protein level decrease under depletion of TRIM36.

I have also observed a significant decrease in centrosomal recruitment of centriole elongation factor CEP120 (Figure 3-15F) and SPICE1 in cells treated with TRIM36 siRNA (Figure 3-15G). CENPJ also associates with CEP120 and SPICE1 for centriole length control (Comartin, Gagan D. Gupta, et al., 2013). However, protein levels of CENPJ remained the same at centrosomes (Figure 3-15E). This indicates potential independent recruitment/degradation mechanisms for CENPJ compared to CEP120/SPICE1, possibly through its direct interaction with STIL. CEP135 is recruited by CEP120/SPICE1/CENPJ to the mother centriole and PCAR (Comartin, Gagan D. Gupta, et al., 2013). With TRIM36 depletion, the CEP135 intensity decreased around PCAR (Figure 3-15H). The Western blotting analysis indicated only slight 102

decrease in CEP135 levels along with TRIM36 depletion, suggesting that the recruitment of CEP135 to the PCAR is blocked (Figure 3-17). The intensity of the capping protein CP110 (Schmidt et al., 2009) is also significantly reduced upon TRIM36 RNAi treatment (Figure 3-15I).

Western blotting analysis will be used on all the examined centriole proteins in the PLK4 assay to confirm that SAS6 protein stability is affected by TRIM36 depletion. Overall, the results demonstrated that TRIM36 acts downstream of STIL but upstream of SAS6 in the control of centriole duplication. This implies a role of TRIM36 at early stages of centriole duplication, potentially during cartwheel formation, which warrants further investigation (Figure 3-18).

Non-target TRIM36 si

CEP135 135

α-Tubulin 55

Figure 3-17. The total cellular levels of CEP135 remained the same upon TRIM36 depletion in S-phase arrested U-2 OS cells. Western blotting analysis of whole cell lysates of U-2 OS cells at 72 hours post transfection with non-target siRNA or siRNAs against TRIM36 and treated with Hydroxyurea (HU) for 24 hours prior to fixation with centrosomes induced to over-duplicate by PLK4 overexpression. The samples were probed with anti-CEP135 antibody, showing CEP135 protein level unchanged under depletion of TRIM36 compared to control RNAi conditions.

103

S G2 M

CENPJ CEP152 CEP120 CEP192 CP110 Pericentrin PLK4 SPICE1 NEDD1 P CEP192 STIL CEP135 CDK5RAP2 PLK4 SAS6 POC1A/1B/5 γ-TuRC α/β-Tubulin γ-TuRC

TRIM36

daughter centriole side view:

Initiation Cartwheel formation Centriole elongation Maturation

Figure 3-18. A schematic representation showing the role of TRIM36 in the centriole assembly pathway.

104

3.4.4 Depletion of TRIM36 causes microtubule-organization defects.

A number of microtubule-associated proteins such as HAUS complex, SEPT family, and KIFs (Figure 3-5C) were identified as TRIM36 proximity interactors. The aforementioned HAUS complex (Section 1.1.3) is essential for spindle and centrosome integrity (Lawo et al., 2009). The HAUS subunits localize to spindle poles and microtubules, and regulates mitotic spindle assembly by opposing NUMA (Lawo et al., 2009). Septins (SEPT) are a family of GTP-binding proteins that participates in a spectrum of cellular processes including cytokinesis and ciliogenesis by assembling into cytoskeletal filaments (Fung, Dai and Trimble, 2014). The kinesin superfamily proteins (KIFs) serve as sources of force for intracellular transportation of membranous organelles and protein complexes in a microtubule-dependent manner (Miki et al., 2001; Niwa, 2015). Previous studies revealed that the depletion of maternal TRIM36 disrupted vegetal MT array formation in Xenopus embryos, and that over-expressed TRIM36 is localized to MTs in COS-1 cells, similar to other C-I subfamily proteins (Meroni and Diez-Roux, 2005; Short and Cox, 2006; Cuykendall and Houston, 2009a; Olson, Oh and Houston, 2015). These findings suggest that the TRIM36 ortholog could carry out related functions in human cells. To determine if this was indeed the case, I first tested the impact of TRIM36 depletion using RNAi on microtubule plus-end dynamics in HeLa cells expressing the plus-end marker EB1. In sharp contrast to control- transfected cells where EB1 is found concentrated around the centrosomes and at the tip of microtubules, in TRIM36 depleted cells, EB1 appeared dispersed along the microtubules lattice (Figure 3-19A), accompanied by marked increased in acetylated microtubules (Figure 3-19B, C). MTs acquire various post-translational modifications (PTM) such as polyglutamylation, polyglycylation, detyrosination, acetylation and phosphorylation (Verhey and Gaertig, 2007). Acetylation is enriched on microtubules that show slow subunit turnover and are more resistant to depolymerization drugs including Nocodazole, thus defined as “stable” (Verhey and Gaertig, 2007). This result suggests that TRIM36 depletion disrupts microtubules dynamics, followed by their stabilization and increased acetylation.

During mitosis, I also observed a high incidence of aberrant mitotic spindles and severe chromosome congression defects in HeLa cells after TRIM36 depletion (Figure 3-20A, B). In addition, anti-kinetochore staining against Calcium-responsive trans-activator (CREST) revealed disruptions in chromosome alignment upon TRIM36 RNAi treatment with the aforementioned two 105

deconvolved siRNAs, along with pronounced spindle assembly defects (Figure 3-20A). Analysis monitors spindle morphology under each condition and data are reported as % of cells with multipolar spindles/fragmented spindles/chromosome congression defects. In the control siRNA treated sample, around 20% of mitotic HeLa cells contain spindle assembly defects, while in the TRIM36 siRNA treated sample, ~70% mitotic cells contain spindle defects, demonstrating significant inhibition of spindle assembly under TRIM36 depletion (Figure 3-20B).

106

DAPI/ EB1/ Pericentrin Merge

arget

Non- TRIM36si

B Acetylated

α-Tubulin/ α-Tubulin Pericentrin Merge

arget

Non- TRIM36 si

Non-target TRIM36si

Acetylated 55 α-Tubulin

α-Tubulin 55 normalized

Ac-α-Tubulin

1.12 1.00 intensity

Figure 3-19. TRIM36 is associated with and required for microtubule organization. (A) Immunofluorescence microscopy images of fixed HeLa cells stained for DNA (DAPI, blue), EB1 (green) and Pericentrin (red) at 72 hours post transfection with Non-target siRNA or siRNAs against TRIM36. Scale bar: 15 μm. (B) Immunofluorescence microscopy images of fixed HeLa cells stained for α-Tubulin (green), acetylated - α-Tubulin (red) and Pericentrin (blue) at 72 hours 107

post transfection with Non-target siRNA or siRNAs against TRIM36. Scale bar: 15 μm. (C) HeLa cells were transfected as described in (B), and whole cell extracts were analyzed by Western blotting, probed for acetylated-α-tubulin and α-tubulin. Band intensity was measured using ImageJ.

DAPI/ α-Tubulin/ CREST Merge

arget

Non- TRIM36 si

B *** *** normal spindle

s 100 spindle formation defect

t o

t 50

f o

% 0 t 1 2 i- i- rge s s a 6 6 -T 3 3 n IM IM o R R N T T

Figure 3-20. TRIM36 is required for mitotic spindle assembly and chromosome congression. (A Immunofluorescence microscopy images of fixed HeLa cells stained for DNA (DAPI, blue), α-Tubulin (green) and CREST (red) at 72 hours post transfection with Non-target siRNA or siRNAs against TRIM36. Scale bar: 15 μm. (B) Quantification of spindle formation defects in cells processed as in (A). Spindle morphology under each condition was examined and data are

108

reported as % of cells with multipolar spindles/fragmented spindles/chromosome congression defects. (***p<0.001, **p<0.01, *p<0.05 by Student’s t-test, n=50, three replicates per condition).

Live imaging using Nikon A1R-HD confocal system and analyzed using NIS-Elements software revealed that the down-regulation of TRIM36 resulted in significantly prolonged mitosis when aberrant spindles are repetitively formed and disassembled (Figure 3-21A), leading to chromosome congression and segregation failure. In the control siRNA treated sample, HeLa cells stay in mitosis for average ~2 hours, while in the TRIM36 siRNA treated sample, HeLa cells stay in mitosis for more than 4 hours (Figure 3-21B). This result was consistent with our Western blot analysis revealing a sharp increase in total cellular levels of phospho (pSer10)-histone H3 (p-H3), a marker of mitotic cells, upon TRIM36 depletion (Figure 3-21C) (Crosio et al., 2009). To further demonstrate that the down-regulation prolongs mitosis progression, in the future, I will quantify Mitotic Index (MI = number of p-H3positive cells/total cell number) for each knockdown condition using immunofluorescence imaging.

A B SiR DNA α-Tubulin/

0:00 0:10 0:20 0:30 0:40 0:50 1:00 900 ***

)

(

arget

i t T 500

1:10 1:20 1:30 1:40 1:50 2:00 2:10 a

D Non-

0 t i s 0:00 0:10 0:20 0:30 0:40 0:50 1:00 rge 6 a 3 -T IM n R o T N

1:10 1:20 1:30 1:40 1:50 2:00 2:10 C

2:20 2:30 2:40 2:50 3:00 3:10 3:20

Non-targetTRIM36 si TRIM36 si TRIM36 Phospho 17 Histone H3 3:30 3:40 3:50 4:00 4:10 4:20 4:30 α-Tubulin 55

109

Figure 3-21. TRIM36 depletion leads to mitotic arrest in HeLa cells. (A) Time-lapse fluorescence montages of spindle microtubules (green) and SiR-DNA (Maroon) during the mitosis of HeLa cells at 72 hours post transfection with Non-target siRNA or siRNAs against TRIM36. Cells were imaged using Nikon A1R-HD confocal at 20X magnification using Resonance scanner. Images were processed and analyzed using NIS-Elements software. Scale bar: 15 μm. Numbers on the panels indicating progression time in minutes. (B) Quantification of (A) showing duration of HeLa cells staying in mitosis post transfection with Non-target siRNA or siRNA against TRIM36. (***p<0.001, **p<0.01, *p<0.05 by Student’s t-test, n=50, whiskers represent min to max). (C) Whole cell extracts of HeLa cells treated as described in (B) were analyzed by Western blotting, probed for Phospho-Histone H3.

Furthermore, I observed that over-expressed FLAG-BirA*-TRIM36 was localized on microtubules throughout the cell cycle (Figure 3-22A). The H50A mutation fusion protein can not localize on the microtubules (Figure 3-22B), suggesting the intact of the RING is required for microtubule localization. Hyper-acetylated microtubules and mitotic spindle defects were also observed after TRIM36 overexpression (Figure 3-23A, B), suggesting that maintaining normal levels of TRIM36 was critical for proper microtubules dynamic. Therefore, one can speculate that the physiological levels of TRIM36 are tightly controlled. Hence, either up-regulation or down-regulation of TRIM36 protein levels has the potential to disrupt MTs organization.

110

Flag-BirA*- TRIM36 TRIM36 (H50A) interphase mitosis interphase mitosis

DAPI

FLAG

Merge

Figure 3-22. TRIM36 RING domain is required for microtubule localization. (A) Representative IF images of HeLa cells transiently over-expressing FLAG-BirA*-tagged TRIM36 and its catalytic mutated form FLAG-BirA*-tagged TRIM36 (H50A) fixed 24 hours after transfection and stained with antibodies against the FLAG tag and Pericentrin. DAPI was used to stain DNA. Scale bar: 15μm.

111

A B Flag-BirA*-TRIM36 interphase mitosis

DAPI

Flag-BirA*-TRIM36 Non-transfected Acetylated α-Tubulin 55

FLAG α-Tubulin 55

normalized

Ac-α-Tubulin 1.00 intensity 1.72

α-Tubulin

Merge

Figure 3-23. Over-expressed TRIM36 localizes on microtubules in HeLa cells. (A) Representative HeLa cells transiently over-expressing FLAG-BirA*-tagged TRIM36 fixed 24 hours after transfection and stained with antibodies against the FLAG tag and α-Tubulin. DAPI was used to stain DNA. Scale bar: 15μm. (B) HeLa cells were transfected as described in (B), and whole cell extracts were analyzed by Western blotting, probed for acetylated-α-tubulin and α- tubulin. Band intensity was measured using ImageJ.

To identify which domain of TRIM36 interacts with microtubules, we transiently over- expressed our panel of TRIM36 fragments in HeLa cells (Figure 3-10A). In both interphase and mitosis, the N-terminal fragment (tripartite motif plus COS box domain, 1-418) was robustly associated with microtubules, while fragments containing the RING domain only or lacking the RING domain failed to localize on microtubules (Figure 3-24). In section 3.4.3, I have demonstrated that the RING finger domain (1-132) and the N-terminal fragment (tripartite/RBCC motif plus COS box domain, 1-418) were sufficient to rescue the centriole duplication defect caused by TRIM36 depletion, while other fragments lacking the RING domain could not (Figure 3-10B, C). The combined results demonstrate that TRIM36

112

microtubule binding is independent of its role in centriole duplication regulation, since as per our findings its RING domain is essential for controlling centriole number, but is not required for microtubule binding.

A Flag-BirA*- 1-132 1-418 133-728 256-728 interphase mitosis interphase mitosis interphase mitosis interphase mitosis

DAPI

FLAG

Pericentrin

Merge

B Fragment MT binding ability FL + RING B1 B2 CC COS FN3 SPRY 1-132 - RING

1-418 + RING B1 B2 CC COS

133-728 - B1 B2 CC COS FN3 SPRY 256-728 - CC COS FN3 SPRY

Figure 3-24. TRIM36 RBCC domain is crucial for its microtubule localization. (A) Representative IF images of HeLa cells in interphase or mitosis transiently (24 hours) over- expressing TRIM36 truncation fragments (RING domain (1-132), RBCC domain (1-418), Full- length with RING truncated (133-728), C-terminal with RBB domains truncated (256-728)) were labeled for DNA (DAPI, blue), FLAG tag (green) and Pericentrin (red). (B) Table showing relative microtubule binding abilities of the aforementioned TRIM36 fragments.

113

3.4.5 TRIM36 cooperates with γ-tubulin to regulate centriole assembly and microtubule formation.

To further investigate the functional partners of TRIM36, we applied the Ubiquitin- Activated Interaction Traps (UBAITs) method (O’Connor et al., 2015, 2018). UBAIT is an E3- ubiquitin fusion protein containing an affinity purification tag, the target E3 ligase, a peptide linker (AAAGGSG) and a C-terminal ubiquitin (Figure 3-25A). The C-terminal ubiquitin moiety forms an amide linkage to proteins that interacts with the E3 in an E1- and E2-dependent manner, enabling covalent co-purification of the E3 with its partner proteins (O’Connor et al., 2015, 2018). It has been demonstrated that RING E3 UBAIT is able to capture putative substrates and the resulting conjugates can be purified from transfected HEK293 cells and detected by AP-MS (O’Connor et al., 2018). In comparison with traditional AP-MS or BioID, UBAIT captures better weak and/or transient interactions between the E3 ligase and targets since the substrates are covalently trapped to E3 ligase. On the down side, it allows in parallel to capture other proteins that interact with the putative substrates.

I constructed Tet-inducible FLAG-tagged TRIM36 (full length)-AAAGGSG-Ubiquitin fusion and expressed it in HEK293 Flp-In T-REx cells. Over-expressed FLAG-tagged TRIM36 (full length)-AAAGGSG-Ubiquitin fusion localizes on the microtubules. Dr. Johnny Tkach helped me to construct FLAG-AAAGGSG-Ubiquitin fusion protein and expressed it in HEK293 Flp-In T-REx cells, which were used as control cells in the following mass spectrometry experiment performed by Dr. Etienne Coyaud in Brian Raught Lab. Considering the down side of UBAIT method, other negative controls will be used in the future analysis, including a “ΔGG” mutant UBAIT, where the two terminal glycine residues of ubiquitin are deleted; and a control in which the affinity tagged protein of interest is expressed without any ubiquitin sequences (O’Connor et al., 2018). Here, proteins with an average SAINT score > 0.98 were chosen to be listed in the interactome. UBAIT results were collected from two biological replicates/four technical replicates of samples in the presence of tetracycline.

114

The MS results indicated that proteins identified by this method were enriched for cell cycle, microtubule stabilization and centriole assembly functions. The BioID dataset is enriched with more microtubule-associated proteins possibly as a result of the localization of TRIM36 on microtubules (for example KIF14, CKAP2, CEP170, MTUS1, MAP9, MAP7 and LUZP1), while the UBAIT dataset is enriched with ubiquitination-related proteins (for example CDK1, CDK4, UBE2N and UBE2E2) (Table 3-2). The identifications of CDK1 and CDK4 are particularly interesting, since the presence of CDK1 promotes centriole duplication by binding to STIL and inhibiting STIL phosphorylation by PLK4 (Novak et al., 2016), while the absence of CDK4 impairs centriole duplication in p53-null cells (Adon et al., 2010).

A B

RBM14 O 31 UBAIT-TRIM36 Flag TRIM36 AAGGSG Ub C 1 27 2 Interacting Protein TUBG1

BioID Bait to Prey

31 Average Spec

Figure 3-25. Using UBAIT to identify TRIM36 interactors. (A) A schematic representation showing UBAIT structure. The structure contains an affinity purification tag (FLAG), the target E3 ligase, a peptide linker (AAAGGSG) and a C-terminal ubiquitin. (B) UBAIT results identified TUBG1 and RBM14 as TRIM36 interactors. Data was analyzed using SAINT. Edge thickness is proportional to total peptide counts.

PreyGene Spec SpecSum AvgSpec ctrlCounts AvgP MaxP SaintScore ACTG1 350|374|625|813 2162 540.5 26|62 1 1 1 PTBP1 220|225|168|237 850 212.5 41|32 1 1 1

115

ACTN4 178|134|161|166 639 159.75 20|31 1 1 1 TCP1 128|120|134|114 496 124 19|19 1 1 1 CDK1 117|110|78|129 434 108.5 3|7 1 1 1 TMOD3 47|67|119|159 392 98 0|0 1 1 1 RBM10 118|118|28|125 389 97.25 38|16 1 1 1 CAPZA1 77|71|105|93 346 86.5 0|2 1 1 1 KIF5B 40|48|59|75 222 55.5 12|7 1 1 1 UBA2 57|47|46|68 218 54.5 3|1 1 1 1 PSMC5 51|46|54|65 216 54 1|2 1 1 1 KIF11 63|40|57|47 207 51.75 0|0 1 1 1 CDC5L 57|55|30|42 184 46 0|0 1 1 1 HDAC2 66|43|42|24 175 43.75 14|9 1 1 1 CAPZB 33|34|52|52 171 42.75 0|1 1 1 1 PSMC2 32|33|40|57 162 40.5 2|0 1 1 1 UBE2N 38|35|34|44 151 37.75 5|8 1 1 1 RBM39 41|43|26|36 146 36.5 10|7 1 1 1 PSMC3 23|23|37|58 141 35.25 0|0 1 1 1 NAP1L1 44|41|28|27 140 35 7|8 1 1 1 ACTN1 39|32|35|28 134 33.5 0|0 1 1 1 CAPZA2 31|23|35|44 133 33.25 0|0 1 1 1 NUDC 32|23|35|35 125 31.25 0|3 1 1 1 MAPRE1 35|27|30|31 123 30.75 0|0 1 1 1 HDAC1 23|45|12|39 119 29.75 1|1 1 1 1 DCTN1 21|27|21|46 115 28.75 0|0 1 1 1 USP39 32|47|14|22 115 28.75 4|3 1 1 1 CDC37 22|21|33|35 111 27.75 3|3 1 1 1 NOP2 41|31|19|14 105 26.25 4|10 1 1 1 USP14 30|29|20|23 102 25.5 3|1 1 1 1 CDK4 29|34|17|19 99 24.75 3|3 1 1 1 RAB5C 22|28|22|25 97 24.25 3|7 1 1 1 CUL1 25|24|16|29 94 23.5 3|2 1 1 1 CUL3 29|26|21|17 93 23.25 2|8 0.99 1 1 MAP2K1 15|16|26|34 91 22.75 0|0 1 1 1 TNPO1 33|26|15|13 87 21.75 3|6 1 1 1 ARCN1 23|22|19|22 86 21.5 3|3 1 1 1 RBM14 21|31|21|12 85 21.25 0|2 1 1 1 SEPT2 24|25|14|22 85 21.25 4|4 1 1 1 TYMS 33|22|13|15 83 20.75 0|1 1 1 1 TRIM28 9|6|27|35 77 19.25 6|9 1 1 1 SRP14 15|16|21|24 76 19 4|6 1 1 1 TUBG1 27|25|8|14 74 18.5 0|0 1 1 1 SRP72 17|10|14|31 72 18 0|1 1 1 1

116

SAE1 17|13|15|26 71 17.75 2|3 1 1 1 SEPT7 20|20|6|19 65 16.25 2|1 1 1 1 CDK2 14|23|10|13 60 15 0|0 1 1 1 RBM25 14|22|8|12 56 14 1|3 1 1 1 UBE2K 16|10|10|15 51 12.75 1|0 1 1 1 UBE2M 11|14|14|9 48 12 0|0 1 1 1 CKAP5 10|8|12|15 45 11.25 2|4 1 1 1 TXNL1 21|14|3|3 41 10.25 0|1 1 1 1 RBM28 13|9|5|11 38 9.5 0|0 1 1 1 SEPT9 7|11|6|14 38 9.5 2|3 1 1 1 KIF4A 4|6|3|18 31 7.75 0|0 1 1 1 UBE2S 5|11|11|4 31 7.75 0|0 1 1 1 TRIM25 4|14|3|9 30 7.5 1|3 1 1 1 KIF2A 12|2|8|5 27 6.75 0|0 1 1 1 KIF1A 3|5|7|7 22 5.5 1|1 1 1 1 RBM4 13|5|3|0 21 5.25 0|0 1 1 1 PLK1 5|8|2|4 19 4.75 0|0 1 1 1 TUBGCP2 9|6|2|2 19 4.75 1|0 1 1 1 MAD2L1 7|6|5|0 18 4.5 0|0 1 1 1 MAPK3 4|3|4|7 18 4.5 0|0 1 1 1 MARK2 10|4|2|2 18 4.5 0|0 1 1 1 RBM34 5|12|0|0 17 4.25 0|0 1 1 1 MAP3K7 5|6|0|5 16 4 0|0 1 1 1 UBE2E2 7|5|0|4 16 4 0|0 1 1 1 RBM12 4|3|7|0 14 3.5 0|0 1 1 1 CDK5 0|6|4|2 12 3 0|0 1 1 1 CDC23 0|4|5|2 11 2.75 0|0 1 1 1 RBM27 4|7|0|0 11 2.75 0|0 1 1 1 RBM17 4|2|0|4 10 2.5 0|0 1 1 1 UBA3 0|5|4|0 9 2.25 0|0 1 1 1 AURKB 3|4|0|0 7 1.75 0|0 1 1 1 MAP1S 0|2|2|0 4 1 0|0 0.98 0.98 0.98 MAPK14 2|0|0|2 4 1 0|0 0.98 0.98 0.98

Table 3-2. TRIM36 Interactors Detected by UBAIT. UBAIT results were collected from two biological replicates/four technical replicates of Tet- inducible FLAG-TRIM36 expression in the presence of tetracycline followed by affinity- purification combined with mass-spectrometry to identify biotinylated peptides. HEK293 Flp-In T-REx cells expressing FLAG-AAAGGSG-Ubiquitin fusion protein were used as the control. Spectral counts reflect the abundance of a prey polypeptide in the SAINT analysis. (From left to

117

right) Prey Gene: Gene Symbol for identified “prey” proteins; Spec Counts: spectral counts for each “prey” protein in each of the technical replicates; SpecSum: total amount of spectral counts for 4 replicates; AveSpec: the average spectral counts for 4 replicates; Ctrl Counts: the spectral counts for each protein in the 2 control replicates; AvgP: the average probability of the 4 replicates; MaxP: the maximum probability between the 4 replicates; SaintScore: overall SAINT score for 4 replicates (Choi et al., 2011, 2012; Skarra et al., 2011). MS and was performed by Dr. Etienne Coyaud. Spectral counts reflect the abundance of a prey polypeptide in the SAINT analysis.

Notably, TUBG1 (tubulin gamma-1 chain) is enriched in UBAIT-TRIM36 dataset (Figure 3-26B, Table 3-2). It has an average 18.5 spectral counts in UBAIT-TRIM36 samples, compared to 0 spectral counts in control samples. TUBG1 is a universally expressed isoform of γ-tubulin and localizes to the centrosomes and mitotic spindle (Yuba-Kubo et al., 2005). Previous studies showed that silencing of γ-tubulin inhibits centriole duplication and spindle formation in HeLa cells (Haren et al., 2006). A more recent work carried out by Rosa Ramírez Cota et al. showed that down-regulation of GCP6 and GCP3, which are other components of γTuRC, led to centriole duplication and spindle formation defects (Teixidó-Travesa et al., 2016). These results demonstrated that the integrity of γTuRC is critical for centriole duplication. Since TRIM36 depletion also leads to centriole duplication defects and mitotic defects, I next sought to confirm the association of TRIM36 and TUBG1 using co-IP in HeLa cells (Figure 3-26). FLAG-BirA* tagged full length TRIM36, FLAG-BirA* tagged TRIM36 RING domain were transfected into HeLa cells respectively with Myc tagged TUBG. Immuno-precipitated FLAG-tagged TRIM36 (full length/RING domain) were immunoblotted using anti-FLAG and anti-Myc to confirm the expression of the transgenes and potential interactions with the Myc-TUBG1, respectively. As shown in Figure 3-25, Myc-TUBG1 interacts with FLAG-BirA* tagged full length TRIM36 and FLAG-BirA* tagged TRIM36 RING domain, but not with the FLAG-BirA* only tag in HeLa cells, suggesting that TUBG1 interacts with TRIM36, especially its RING domain.

118

Input FLAG IP

FLAG-BirA*

FLAG-BirA*-TRIM36

FLAG-BirA*-TRIM36 RING

FLAG-BirA*

FLAG-BirA*-TRIM36 FLAG-BirA*-TRIM36 RING

130 95 72 WB: FLAG 55

55 IP: Myc-TUBG1

Figure 3-26. TUBG1 associates with TRIM36. Co-immunoprecipitation of TUBG1/γ-tubulin with TRIM36 full-length protein and its RING domain. HeLa cells were co-transfected with FLAG-BirA*-TRIM36 or FLAG-BirA*-TRIM36 RING constructs (or FLAG-BirA* vector which was used as the control), and Myc-TUBG1. Cells were harvested and lysed as described in 3.6.11. Western blotting of anti-FLAG immunoprecipitates and inputs (cell lysates) were performed using antibodies against FLAG and Myc.

Therefore, I sought to examine γ-tubulin localization at the spindle poles in the absence of TRIM36. I used PCNT IF signal as a ‘mask’ to define the centrosome area, in which the fluorescence intensity of γ-tubulin was quantified (Figure 3-27A). Along with defects in spindle morphology, I observed a ~40% decrease of centrosomal total intensity of γ-tubulin in mitotic HeLa cells depleted of TRIM36 (Figure 3-27B), indicating that TRIM36 depletion results in decreased γ-tubulin levels at the centrosomes. The Western blotting analysis revealed that γ- tubulin levels remained unchanged with TRIM36 depletion, suggesting that the loss of γ-tubulin at the centrosomes was a result of its inhibited recruitment to the centrosomes (Figure 3-27C). Same results were observed in the PLK4 assay, where the cells were treated with non-target siRNA

119

or siRNAs against TRIM36 in S-phase arrested U-2 OS cells over-expressing PLK4. 3D-SIM showed significant reduction (~35%) of γ-tubulin levels at PCAR (Figure 3-28A, B). Since γ- tubulin is essential for centriole duplication and spindle formation (Haren et al., 2006), giving rise to my hypothesis that the centriole duplication and spindle defects I observed after TRIM36 depletion is a result of loss of γ-tubulin at the centrosomes/spindle poles.

A B C

Non-target TRIM36 si

)

10 (

DAPI 10.0 Non-target ** TRIM36 si

γ-Tubulin 55 γ-Tubulin

y at Centrosome 5.0

n e

t α-Tubulin 55

n I

α-Tubulin total 0 t i s rge 6 a 3

γ-Tubulin -T M n I o R N T

Pericentrin

Merge

Figure 3-27. TRIM36 is required for γ-Tubulin recruitment to the centrosome region in HeLa mitotic cells. (A) Representative IF microscopy images of fixed HeLa cells at 72 hours post transfection with Non-target siRNA or siRNAs against TRIM36. Cells were labeled for DNA (DAPI, blue), α- tubulin (green) and Pericentrin (red). Scale bar: 15μm. (B) γ-Tubulin level was quantified using Pericentrin as mask. (***p<0.001, **p<0.01, p<0.05 by Student’s t-test, n=50, whiskers represent min to max). (C) Western blotting analysis of whole cell lysates of HeLa cells at 72 hours post

120

transfection with non-target siRNA or siRNAs against TRIM36 The samples were probed with anti-γ-Tubulin antibody, showing γ-Tubulin protein level unchanged under depletion of TRIM36.

A B γ-Tubulin

y at Centrosome

i s

CP110

n I

Myc-PLK4 Relative γ-Tubulin

Merge 0 t i s rge 6 t i a 3 ge s -T IM r 6 n R a 3 o T -T M N n I o R N T

Figure 3-28. TRIM36 is required for γ-Tubulin/TUBG1 recruitment to the centrosome region in HeLa mitotic cells. (A) Representative 3D-SIM pseudocolor-merged (all bottom panels) and individual gray-scale images showing over-duplicated centrioles in S-phase arrested U-2 OS cells over-expressing PLK4. Cells were with non-target siRNA or siRNAs against TRIM36. Centrosomes are labeled for Centrin-2 (blue), Myc-PLK4 (red), and γ-Tubulin (green). (B) Quantification of γ-Tubulin level at centrosome using Myc-PLK4 as mask (***p<0.001, **p<0.01, p<0.05 by Student’s t-test, n=30, whiskers represent min to max).

121

In 3.4.3, I discussed that in the PLK4 assay, TRIM6 depletion led to instability of SAS6 and reduced its intensity at the centrosome, while STIL intensity at the centrosome remained unchanged. Here, I examined the recruitment of key cartwheel proteins (SAS6 and STIL) in the absence of TUBG1 in S-phase arrested U-2 OS cells expressing Tet-inducible PLK4. I observed a marked decrease of SAS6 at PCAR upon TUBG1 depletion, while its impact on STIL recruitment is not significant (Figure 3-29A, B). The fact that TUBG1 depletion phenocopies the loss of SAS6 after TRIM36 depletion suggests a potential mechanism of TRIM36 in regulating centriole duplication and microtubule organization process through the recruitment of γ-tubulin to centrosomes.

A B

) 3.0

4 )

4 3.0

(

SAS6

STIL

10 ( ns

2.0 2.0

CP110

Centrin-2

y at Centrosome

t i

y at Centrosome ***

e I

t 1.0 n

I 1.0

Myc-PLK4

Merge

SAS6 mean SAS6 STIL mean

Merge 0 0 t i t i t i t i s s s s rge rge rge rge a a a a -T -T -T -T n n n n o o o o N N N N γ-Tubulin γ-Tubulin γ-Tubulin γ-Tubulin

Figure 3-29. γ-Tubulin is essential for SAS6 recruitment to the centrosome region. Representative 3D-SIM pseudocolor-merged (all bottom panels) and individual gray-scale images showing STIL (A, left) and SAS6 (B, left) at centrosomes induced to over-duplicate by PLK4 overexpression in S-phase arrested U-2 OS cells. Cells were with non-target siRNA or siRNAs against γ-Tubulin. Centrosomes are labeled for Centrin-2 (blue), Myc-PLK4 (red), and STIL/SAS6 (green). Scale bar: 15μm. (A, B, right) Quantification of STIL/SAS6 level at centrosome using Myc-PLK4 as a mask (***p<0.001, **p<0.01, p<0.05 by Student’s t-test, n=20, whiskers represent min to max).

122

Intriguingly, RBM14 is also enriched in the UBAIT-TRIM36 dataset (Figure 3-25B, Table 3-2). It has an average 21.25 spectral counts in UBAIT-TRIM36 samples, compared to 1 spectral counts in control samples (Table 3-2). RBM14 is reported to be a negative regulator of centriole duplication that binds to γ-tubulin (Shiratsuchi et al., 2015). Its depletion leads to the formation of ectopic centriolar protein complexes that harbor normal centrosome functions; the complexes accumulate γ-tubulin and serve as sites for microtubule nucleation (Shiratsuchi et al., 2015).

Since the depletion of RBM14 or TRIM36 have antagonizing effects on γ-tubulin to the centrosomes, I further investigated the functional relationship between these proteins. To do this, I co-depleted RBM14 and TRIM36 in HeLa cells, and measured the fluorescence intensity of γ- tubulin under each condition (Non-target RNAi, TRIM36 RNAi, RBM14 RNAi and TRIM36/RBM14 RNAi) (Figure 3-30). I found that the co-depletion of RBM14 and TRIM36 restored centrosomal γ-tubulin levels compared to TRIM36 depletion alone (Figure 3-30B). Previous work showed that the depletion of RBM14 leads to the formation of ectopic centriolar protein complexes harboring STIL and CENPJ that are able to accumulate γ-tubulin (Shiratsuchi et al., 2015), which may explain the restoration of γ-tubulin at the spindle poles under co- depletion condition.

123

Non-TargetNon-Target/TRIM36Non-Target/RBM14 siTRIM36 si/RBM14 si si Non-TargetNon-Target/TRIM36Non-Target/RBM14 siTRIM36 si/RBM14 si si 95 RBM14 TRIM36 72 72

α-Tubulin 55 α-Tubulin 55

RNAi: Non-Target Non-Target/TRIM36 Non-Target/RBM14 TRIM36 / RBM14

-Tubulin

DAPI

) 6

10 *

( ***

2.0

yat Centrosome

n I

0 γ-Tubulin

RNAi: Non-Target

TRIM36 / RBM14 Non-Target/TRIM36Non-Target/RBM14

Figure 3-30. RBM14 depletion restores γ-Tubulin level at the centrosome region after TRIM36 depletion. (A) Western blot of whole cell lysates collected from RNAi treated HeLa cells showing depletion and co-depletion of TRIM36 and RBM14 protein levels under RNAi conditions. (B) 124

Immunofluorescence microscopy images of fixed mitotic HeLa cells processed as in (A) and stained for DNA (DAPI, blue) and γ-tubulin (green). Scale bar: 15 μm. (C) Quantification of γ- tubulin level was quantified at the centrosome region in cells processed as in (A). (***p<0.01 by Student’s t-test, n=50, three replicates per condition).

I also examined the % of cells with spindle defects under each condition (Non-target RNAi, TRIM36 RNAi, RBM14 RNAi and TRIM36/RBM14 RNAi). Indeed, co-depletion of RBM14 and TRIM36 rescued spindle disorganization after TRIM36 depletion (Figure 3-31A). The cells containing spindle defects after TRIM36 depletion was ~65%, and the number was reduced to ~15% with co-depletion of RBM14 and TRIM36 (Figure 3-31B). Next, I examined the% of cells with centriole duplication defects under each condition (Non-target RNAi, TRIM36 RNAi, RBM14 RNAi and TRIM36/RBM14 RNAi). The significant loss of centrioles in mitotic HeLa cells transfected with TRIM36 RNAi was rescued by co-depletion of RBM14 and TRIM36 (~40% reduction of mitotic cells with four centrioles under TRIM36 depletion compared with ~20% under co-depletion) (Figure 3-32). These results suggest that the levels of γ-tubulin at centrosomes are tightly controlled, which is crucial for the regulation of centriole duplication and cell cycle progression. The antagonizing effects of TRIM36 and RBM14 on centrosomal γ-tubulin recruitment warrants further investigation. First step is to investigate whether TRIM36 over-expression leads to γ-tubulin accumulation at centrosomes by inducing formation of pseudo/ectopic centrioles, similar to the effects of RBM14 depletion. Next, I will double-knock down RBM14 and another factor that affects centriole duplication and spindle formation, including SPICE1 and CEP192. The experiment is to examine whether the antagonizing role of RBM14 and TRIM36 is specific.

125

RNAi: Non-Target Non-Target/TRIM36 Non-Target/RBM14 TRIM36 / RBM14

α-Tubulin

DAPI

B *** normal spindle spindle formation defect

100

t o

t 50

o % 0

RNAi:

Non-Target TRIM36/RBM14 Non-Target/TRIM36Non-Target/RBM14

Figure 3-31. Co-depletion of RBM14 and TRIM36 restores the γ-Tubulin level and spindle assembly. (A) Immunofluorescence microscopy images and (B) quantification of spindle morphology of fixed HeLa cells at 72 hours post transfection with non-target siRNA or siRNAs against TRIM36 or siRNAs against RBM14. DAPI was used to label DNA (blue). Cells were labeled for α- tubulin (green). Scale bar: 15 μm. (***p<0.001, **p<0.01, p<0.05 by Student’s t-test, n=50, three replicates per condition, mean ± SEM).

126

RNAi: Non-Target Non-Target/TRIM36 Non-Target/RBM14 TRIM36 / RBM14

PCNT

Centrin-2

DAPI

B ** =4

100 <4 s

ll >4

t o

t 50

o % 0 RNAi:

Non-Target

TRIM36 / RBM14 Non-Target/TRIM36Non-Target/RBM14

Figure 3-32. Co-depletion of RBM14 and TRIM36 restores the γ-Tubulin level and centriole duplication. (A) Immunofluorescence microscopy images and (B) bar graph showing the percentage of mitotic HeLa cells processed as in which either <4 or 4 centrioles were detectable. DAPI was used to label DNA (blue). Cells were labeled for Centrin-2 (green) and Pericentrin (red). Boxed insets show 4x magnified centrioles on each spindle pole. Scale bar: 15 μm. (***p<0.001, **p<0.01, p<0.05 by Student’s t-test, n=50, four replicates per condition, mean ± SEM).

127

3.5 Discussion

3.5.1 TRIM36 is a novel positive centriole duplication regulator that cooperates with γ-tubulin to regulate centriole assembly

As the major MT organizing centers in mammalian cells, centrosomes are involved in the control of spindle formation during mitosis and play a pivotal role in chromosome congression and segregation (Prosser and Pelletier, 2017; Nigg and Holland, 2018). In this work, I discussed the dual role of TRIM36 in centriole duplication and spindle formation, where TRIM36 potentially acts through a common mechanism involving the recruitment of γ-tubulin to the centrosome region.

I proved that TRIM36 locates at early stages of centriole duplication pathway using three-dimensional structured-illumination microscopy (3D-SIM) in U-2 OS cells as our lab described previously (Comartin, Gagan D Gupta, et al., 2013). In addition, I have studied multiple domains of TRIM36, and provided evidences that as an RING-type E3 ligase from TRIM family, the RING domain at its N-terminus is crucial for its role in centriole duplication. From N-terminal to C-terminal, TRIM36 contains a RING finger C3HC4 structure, two B-box domains, a coiled-coil domain, a COS box domain, a FN III domain and a SPRY domain (Balint et al., 2004; Short and Cox, 2006). Intriguingly, my analysis of domain sequence comparison reveals that the RING of C-I TRIMs contain a loop structure of various lengths between the 5th and 6th conserved cysteine residues. The feature is so far specific to the C-I subfamily, and it will be interesting to follow up on the potential function of the loop structure. It may act as potential binding domain for E2 or putative substrates. As the catalytic domain, RING serves as binding sites for interacting E2 and facilitates Ub transferring from E2 to the substrates (Brown et al., 2015). I will be interesting to delete the loop structure from full length TRIM36, and examine whether the deletion affects centriole duplication.

In pursuing functional partners and putative substrates of TRIM36, I exploited the Ubiquitin-Activated Interaction Traps (UBAITs) method (O’Connor et al., 2015, 2018). γ-tubulin and RBM14 were identified in the interaction network. γ-tubulin is a crucial component of γTuRC, and has been recognized as a critical regulator of centriole duplication both in mice and human

128

(Yuba-Kubo et al., 2005; Haren et al., 2006; Kollman et al., 2011). Another work showed that down-regulation of other components of γTuRC, including GCP6 and GCP3, led to centriole duplication defects (Teixidó-Travesa et al., 2016). These studies demonstrated the integrity of γTuRC was essential for normal centriole duplication. I demonstrated that TRIM36 interacts with γ-tubulin using co-IP, and it regulates γ-tubulin level at the centrosome. The results provided a potential explanation of the centriole duplication defects observed after TRIM36 depletion. Since the recruitment of γ-tubulin to the centrosome was affected by TRIM36 depletion, the formation of γTuRC was disrupted, leading to centriole duplication defects (Teixidó-Travesa et al., 2016).

I also found that both γ-tubulin recruitment deficiency caused by TRIM36 depletion and RNAi mediated depletion of γ-tubulin decreased SAS6 intensity at the centrosome region. A lack of γ-tubulin could be the cause of subsequent centriole duplication and spindle formation defects that I observed with TRIM36 depletion. One hypothesis is that depletion of TRIM36 impedes centrosomal γ-tubulin recruitment, thus inhibits cartwheel formation process, providing another possible explanation of centriole duplication defects caused by TRIM36 depletion. In the absence of centrosomal γ-tubulin, SAS6 is unstable. In the canonical centriole assembly pathway, STIL recruits SAS6 to mother centrioles to form cartwheels (Arquint et al., 2015; Arquint and Nigg, 2016). Interestingly, the cartwheel dissolvent event doesn’t have an effect on STIL, which theoretically forms the linker between the cartwheel and the centriole microtubule walls with CENPJ (CPAP) (Tang et al., 2011b; Hirono, 2014). It should be noted that the degradation of SAS6 was observed in S-phase arrested U-2 OS cells with PLK4 overexpression. In the future, it is important to examine SAS6 protein level at different cell cycle stages after TRIM36 depletion.

RBM14 is reported to be a negative regulator of centriole duplication that binds to γ- tubulin (Shiratsuchi et al., 2015). Shiratsuchi et al. observed a significant increase of centrin foci after siRNA-mediated depletion of RBM14 (Shiratsuchi et al., 2015). Its depletion leads to the formation of ectopic centriolar protein complexes containing STIL and CENPJ that harbor normal centrosome functions; the complexes accumulate γ-tubulin and serve as platforms to incorporate SAS6 and as sites for microtubule nucleation (Shiratsuchi et al., 2015). By co- depleting TRIM36 and RBM14, the centrosomal level of γ-tubulin is restored and the centriole duplication defect is rescued. One possible explanation is that while TRIM36 depletion impairs SAS6 stability and γ-tubulin to the centrosomes, the ectopic centriolar protein complexes formed 129

after RBM14 depletion cluster around the spindle poles, recruiting SAS6 and γ-tubulin, restoring centrioles in the end.

3.5.2 TRIM36 depletion impairs γ-tubulin recruitment to the centrosomes, disrupts spindle formation and causes mitotic defects

As centriole duplication and spindle formation are functionally closely associated events, there are a number of proteins involved in both processes, and I have demonstrated that TRIM36 is one of them. TRIM36 is identified in a cohort of microtubule associating proteins that are in CEP120/SPICE1 interaction network, as identified and characterized in previous studies (Gupta, Coyaud, et al., 2015). The cluster of proteins includes the entire Augmin/HAUS complex is clustered in the dataset, along with MAPs, LUZP1, DCTN1, NAP1L1 and ANK2, and a number of them cause similar phenotypes after depletion as CEP120, SPICE1 and CEP135 (Gupta, Coyaud, et al., 2015). I have presented that both TRIM36 depletion and over-expression disrupt spindle assembly. This could be a result of both directly binding to the microtubules and binding to other proteins that facilitate spindle assembly. I have observed that the RBCC domain of TRIM36 has a high affinity with MTs, whilst a catalytic dead mutation at its RING domain will disrupt the association. RING domain only does not have microtubule binding abilities. Note that the expression of RING domain is sufficient in rescuing centriole duplication defects after TRIM36 depletion in HeLa mitotic cells, demonstrating its function in maintaining normal centriole duplication is independent from its MT binding ability.

I have demonstrated that TRIM36 depletion led to prolonged mitosis. Aberrant spindles are repetitively formed and disassembled, resulting in chromosome congression defects. In line with my observation, study conducted by Emi Yoshigai et al. and Houston et al. demonstrated that TRIM36 is essential for vegetal microtubule array formation and normal somite formation during early development of Xenopus embryos (Cuykendall and Houston, 2009b; Yoshigai et al., 2009; Olson, Oh and Houston, 2015).

Our work showed that TRIM36 also has a dual role in both processes, potentially through mediating centrosomal levels of γ-tubulin. As a fundamental component of γ-TuRC, γ-tubulin has 130

an essential role in basal body duplication in Paramecium (Ruiz et al., 1999). Moreover, TUBG1 deficient mouse embryos were presented with serious mitotic arrest and disorganized spindles (Yuba-Kubo et al., 2005). In mitotic HeLa cells, γ-Tubulin depletion reduced centriole numbers on spindle poles, resulting in spindle pole separation defects (Haren et al., 2006).

The defects in spindle disorganization after TRIM36 depletion was also rescued by co- depletion of RBM14 in HeLa mitotic cells. Shiratuschi et al. reported observing extra MTOC sites, and they tended to be clustered into pseudo-bipolar spindles after cold treatment (Shiratsuchi et al., 2015). My results showed that the pseudo-bipolar spindles These results suggest that the levels of γ-tubulin at centrosomes are tightly controlled, which is crucial for the regulation of centriole duplication and cell cycle progression. It would be interesting to capture the dynamics of spindle formation using live-imaging under co-depletion of RBM14 and TRIM36.

In summary, our data indicate that TRIM36 has a pivotal role in centriole duplication, mitotic progression and spindle formation regulation (Figure 3-33). Interestingly, the results correlated with previous clinical findings that TRIM36 expression is significantly up-regulated in human primary cancer cell lines (Liang et al., 2018) and down-regulated in several non-small cell lung cancer (NSCLC) cell lines (Zhan et al., 2015), with the effects being a delay in cell cycle in both cases (Miyajima et al., 2009; Liang et al., 2018). Altogether, the data suggest that TRIM36 regulation may play a promising role in cancer treatment.

131

Spindle Pole

PLK4 Wild Type P -TuRC STIL γ SAS6

γ -TuRC

Spindle Pole

TRIM36 Depletion

Degradation

Figure 3-33. A schematic representation model for how TRIM36 affects centriole duplication and spindle formation pathway.

3.6 Materials and Methods

3.6.1 Cell lines and tissue culture HeLa cells were grown at 37ºC in Dulbecco’s modified Eagle’s medium containing 4.5 g/L D-glucose, 110 mg/L sodium pyruvate and L-glutamine, and supplemented with 10% fetal bovine serum and 1% GlutaMAX (all tissue culture reagents were from Invitrogen). Myc-tagged PLK4 U-2 OS cells line was a kind gift from Dr. E. Nigg. The cell line was maintained in McCoy 5A media supplemented with 10% Tet-free FBS, 2mM L-glutamine and 0.5mg/ml G418. All cells were cultured in a 5% CO2 humidified atmosphere at 37°C.

3.6.2 Cloning cDNAs were amplified using KOD DNA Polymerase (Toyobio EMD Millipore) by PCR, digested with restriction enzymes (NEB), and ligated with digested vectors using the T4 DNA Ligase (Rapid Ligation Kit from Fermentas). TRIM36 mutated sequences were generated by QuikChange (Agilent Technologies) Site-Directed Mutagenesis protocol to and cloned into pCDNA5-TO/FRT-FLAG-BirA* vector. For TRIM36 siRNA-2 resistant construct, TRIM36 was 132

mutated at wobble-codon positions at its siRNA-2 target sequence (primer: 5’- GCCCACTGTCTTGCTCATACCTGGGGCTCACGGCGTCGCGGGGAGAACGGAGCCATT -3’). For TRIM36 H50A mutant, TRIM36 was mutated at its RING domain (primer: 5’- TACACATTTATGACAGATACTGGCTTGGCAAGGGAGAATCAATGGGTGG-3’).

3.6.3 RNAi All siRNAs were transfected using Lipofectamine RNAiMax (Invitrogen) according to the manufacturer’s protocol. For TRIM36, deconvolved Stealth siRNAs were acquired from Thermo Fisher (HSS183096, Cat #1299001, target sequence: gagttgcttctagcgataaactaca; HSS183097, Cat #1299001, target sequence: ccgggatgcagttagtccaagatat). For RBM14, we acquired Silencer® Select siRNA from Thermo Fisher (Cat # 4392420). For TUBG1, we acquired deconvolved ON-TARGETplus siRNAs from Dharmacon (Cat # J-005160-08). We used Dharmacon Luciferase GL2 Duplex non-targeting siRNA as the negative control.

3.6.4 RNAi rescue of centriole duplication defects

TRIM36 siRNA-2 resistant constructs generated in 3.6.2 were transfected into HeLa FRT/Flp-In cells to generated stable cell lines (Invitrogen). Transgene expression was induced with tetracycline (Bioshop) at 0.1 µg/mL for 24 h for rescue experiments, and at 1 µg/mL for 24 h for all other experiments.

For endogenous rescue experiments in the U-2 OS assay, a modified U-2 OS cell-line in which endogenous TRIM36 was edited by CRISPR to be resistant to TRIM36 siRNA-1. A custom single-stranded DNA oligonucleotide (ssODN) was used as the repair template for modification within siRNA targeting region (Paquet et al., 2016). The customized 200-nt antisense ssODN synthesized by IDT contains 90-nt flanking sequences on each side of the 20-nt repair template region including TRIM36 siRNA-1 targeting site. The sequence was mutated to be siRNA- resistant and an XbaI digestion site was introduced for genotyping analysis. The ssODN was co- transfected with sgRNAs cloned into pX458 (a plasmid expressing Cas9) in U-2 OS cells and transfected cells were selected using 1μg/ml Puromycin. The targeted genomic region from single clones was PCR-amplified using primers that anneal outside of the region of homology.

133

3.6.5 Purification of GST-TRIM36 from Sf9 cells TRIM36 was cloned into pFastBac-GST (Life Technologies) and transformed into BL21 E. coli (Life Technologies) cells. The recombinant bacmid was transfected into Sf9 insect cells by Z.Li from Sicheri Lab. Next, insect cells were harvested and lysed in lysis buffer (50 mM HEPES pH7.5, 200 mM NaCl, 5% glycerol, 5mM β-ME, 0.8 mM PMSF, 30mL lysozyme and protease inhibitors (Roche)) followed by homogenization and sonication. The supernatant after lysis was incubated with glutathion-sepharose beads for 30 min at 4°C. Lysis buffer described above was used to wash the beads and GST tag was cleaved by TEV overnight at 4°C. Eluted sample was applied on Superdex 200 10/300 GL size exclusion to separate proteins based on their apparent molecular weight and analyzed on SDS page gels.

3.6.6 PLK4 Induced Centriole Over-Duplication Assays U-2 OS T-REx cells with inducible Myc-PLK4 were seeded in 6-well plates and treated with RNAi as described above. At 48 hours post-RNAi, cells were arrested in S-phase by addition of hydroxyurea to a final concentration of 8 mM, and induced to over-express PLK4 by addition of tetracycline (2mg/mL) for 20 hours before fixation.

3.6.7 Immunofluorescence microscopy Glass coverslips with cells were first washed with 4°C PBS, then fixed with ice-cold methanol at −20°C for 15 min. blocked in 1% BSA Fraction V (OmniPur) and 0.05% Tween-20 solution in PBS at room temperature for one hour. Cells were incubated with the primary antibodies in blocking solution for 1 hour, washed 3 times for 5min each in blocking solution and then incubated with fluorophore-conjugated secondary antibodies (Molecular Probes, 1/500 dilution) and DAPI (0.1 μg/ml) in blocking solution for 1 hour. The cells were then washed for 3 times (5min each) in blocking solution. The coverslips were inverted and mounted on glass slides with standard mounting solution (ProLong Gold antifade, Invitrogen™ Molecular Probes™). A list of antibodies used is provided in Table 3-3. All imaging was performed on Delta Vision microscopes (GE Healthcare-Applied Precision) using 60X (NA 1.42) plan apochromat oil objective (Olympus). Images were deconstructed, projected and exported as tiff files by SoftWoRx (GE-Healthcare-Applied Precision), and subsequently analyzed using

134

IMAGEJ. All image quantifications were performed by Dr. Gagan Gupta using MATLAB software.

3.6.8 3-D SIM Imaging 3D-SIM images was acquired using a Delta Vision OMX microscope, reconstructed and aligned in SoftWoRx (GE-Healthcare-Applied Precision) as previously described (Lawo et al., 2012a; Comartin, Gagan D. Gupta, et al., 2013). 3D-SIM image quantifications were performed by Dr. Gagan Gupta using MATLAB software.

3.6.9 Live-Imaging HeLa cells stably expressing GFP-Tubulin were transfected with control or TRIM36 siRNA in 4-well Nunc™ Lab-Tek™ II Chamber Slide™ (Thermo Fisher) for 72 hours. 3 hours before imaging, the cells were washed with PBS for 3 times, refed with phenol red-free DMEM supplemented with 10% FBS and 1X Glutamax (Life Technologies), and stained with Sir-DNA dye (Cytoskeleton). Cells were imaged using Nikon A1R-HD confocal system at 37°C in 5% CO2 at 20X magnification using Resonance scanner. Image stacks of 25μm were taken every 10 minutes overnight. Images were processed and analyzed using NIS-Elements software.

3.6.10 Ubiquitin-Activated Interaction Traps (UBAITs)

For TRIM36 UBAIT constructs, Tet-inducible FLAG-tagged TRIM36 (full length)- AAAGGSG-Ubiquitin fusion was expressed in HEK293 Flp-In T-REx cells. HEK293 Flp-In T- REx cells expressing FLAG-AAAGGSG-Ubiquitin fusion protein were used as control cells in the following mass spectrometry experiment performed by Dr. Etienne Coyaud in Brian Raught Lab. For each replicate in AP-MS, cells on fivr 15-cm dishes were induced with tetracycline (1 µg/mL) for 24h and grown to 80-90% confluency before harvested in PBS. For affinity purification, cells were lysed in four volumes of ice-cold lysis buffer (50 mM HEPES, pH 8.0, 100 mM KCl, 2 mM EDTA, 10% glycerol, 0.1% NP-40, 1 mM PMSF, 1 mM DTT, and protease inhibitor cocktail (Sigma)). Lysates were freeze/thawed and centrifuged at 14000 rpm for 20 min at 4ºC. Soluble lysate (inputs) was added to mouse anti-FLAG M2 magnetic beads (Sigma) and allowed to bind for 2 h at 4ºC by nutation. Beads were washed in aforementioned lysis buffer for three times and two times in rinsing buffer (20 mM Tris-HCl, pH 8.0, 2 mM CaCl2). For trypsin digestion, beads were resuspended in 20 mM Tris-HCl (pH 8.0), 500 ng of trypsin (Sigma) was 135

added, and beads were allowed to rotate at 37ºC or 4 h. Beads were then magnetized, an additional 500 ng of trypsin added, and incubated overnight at 37ºC without agitation. After the second trypsin digestion, formic acid was added to 2%. The sample was analyzed by mass spectrometry and SAINT as previously described in chapter 2. Proteins with an average SAINT score > 0.98 were chosen to be listed in the interactome. UBAIT results were collected from two biological replicates/four technical replicates of samples in the presence of tetracycline.

3.6.11 Statistical methods p-values were calculated from performing two-tailed unpaired student t-tests using GraphPad Prism 4 software. Experiment sample numbers and replicates are as indicated in figure legends. Unless otherwise specified, all the error bars represent S.E.M, and following shorthand is used throughout: *** p<0.001, **p<0.01, *p<0.05.

3.6.12 Western Blots Laemmli buffer was added directly to cells were pelleted at 4°C for 5 mins at 2500 rpm, and the lysate was boiled at 95°C for 5 mins. Proteins were resolved using 8% SDS-PAGE (SDS- Polyacrylamide gels subjected to electrophoresis), then transferred to PVDF membranes (Immobilon-P, Milipore) at 4°C. The membranes were blocked in Tris buffer containing 0.1% Tween-20 (TBST) with 5% skim-milk powder (BioShop) and incubated with primary antibodies diluted with blocking buffer overnight at 4°C. Membranes were then rinsed 3 times for 20 minutes in TBST, then incubated with secondary HRP conjugated antibodies (1/5000) for 1 hour at room temperature followed by a repeated set of 3 rinses in TBST. Western blots were developed using SuperSignal Pico or Dura reagents (Pierce/Thermo Scientific) on HCL films.

3.6.13 Co-Immunoprecipitation For co-immunoprecipitation experiments, NP-40 lysis buffer (50 mM HEPES, pH 7.4, 100 mM KCl, 0.3% NP-40, 10% glycerol, 1 mM EGTA, 1 mM MGCl2, protease inhibitors (Roche)) was used to resuspend cell pellets. The resuspended cells were incubated on ice for 30 min and freeze-thawed on dry ice to aid lysis for 5 min. Lysate was centrifuged at 20,000 g for 20 min at 4ºC, a portion of cleared soluble fraction was set aside as input. For anti-FLAG immunoprecipitation, lysate was nutated with anti-FLAG M2 Affinity Gel (Sigma-Aldrich)

136

overnight at 4ºC, followed by three washes in ten volumes of lysis buffer the next day. Samples were eluted in Laemmli sample buffer.

3.6.14 Antibodies and primers used in the study

Primary antibodies Source Mouse Anti-Acetylated Tubulin Sigma (T6793) Rat Anti-Alpha Tubulin Cedarlane/BioRad (MCA77G) Mouse Anti-Alpha Tubulin Sigma (T6199) Human Nuclear ANA-Centromere Cedarlane/Fitzgerald (CS1058) CREST Goat Anti-C-Myc Polyclonal Abcam (ab19234) Rabbit Anti-CENPJ (CPAP) Cedarlane/Proteintech (11517-1-AP) Mouse Anti-Centrin, clone 20H5 Cedarlane/Millipore (04-1624) Rabbit Anti-CEP120 (Xie et al., 2007) Rabbit Anti-CEP135 (Bird and Hyman, 2008) Rabbit Anti-CEP152 (asterless) Cedarlane/Bethyl (A302-480A) Rabbit Anti-CEP192 (27 hSPD2) (Zhu et al., 2008) Rabbit Anti-CP110 Polyclonal Cedarlane/Proteintech (12780-1-AP) Mouse Anti-Cyclin A (BF683) NEB/Cell Signaling (CST) (4656) Mouse Anti-EB1 (1A11/4) Santa Cruz (sc-47704) Monoclonal Rabbit Anti-FLAG Sigma (F7425) Mouse Anti-FLAG M2 Magnetic Sigma (M8823) Beads Mouse Anti-FLAG M2 Monoclonal Sigma (F3165) Mouse Anti-Gamma-Tubulin Sigma (T6557) Mouse Anti-Histone H3 (Phospho Abcam (ab14955) S10) Goat Anti-Pericentrin 2 (C-16) Santa Cruz (SC-28145) Goat Anti-Pericentrin 2 (N-20) Santa Cruz (SC-28143) Rabbit Anti-RBM14 Polyclonal Abcam (ab12325) Mouse Anti-SAS6 (91.390.21) Santa Cruz (SC-81431) Rabbit Anti-SPICE1 (Archinti et al., 2010) Rabbit Anti-STIL (Liu et al., 2018) Rabbit Anti-TRIM36 self-made

Table 3-3. Primary Antibodies Used in this Work.

137

Secondary antibodies Source Alexa Fluor 488 donkey anti-mouse Invitrogen Alexa Fluor 594 donkey anti-mouse Invitrogen Alexa Fluor 647 donkey anti-mouse Invitrogen Alexa Fluor 488 donkey anti-rabbit Invitrogen Alexa Fluor 594 donkey anti-rabbit Invitrogen Alexa Fluor 647 donkey anti-rabbit Invitrogen Alexa Fluor 488 donkey anti-goat Invitrogen Alexa Fluor 594 donkey anti-goat Invitrogen Alexa Fluor 647 donkey anti-goat Invitrogen Alexa Fluor 647 donkey anti-rat Invitrogen goat anti-mouse IgG (H+L)-HRP BioRad goat anti-rabbit IgG (H+L)-HRP BioRad Bovine anti-goat IgG-HRP Cedarlane/Jackson

Table 3-4. Secondary Antibodies Used in this Work.

Primers used Sequence (5'-3') TRIM36_siRNA01_QC_F cggggagaacggagccattcttgcagcttgtcactggaggc cactcccacttttaccaggtatga TRIM36_siRNA01_QC_R tcatacctggtaaaagtgggagtggcctccagtgacaagct gcaagaatggctccgttctccccg TRIM36_siRNA02_QC_F gcccactgtcttgctcatacctggggctcacggcgtcgcggg gagaacggagccatt TRIM36_siRNA02_QC_R aatggctccgttctccccgcgacgccgtgagccccaggtatg agcaagacagtgggc AscI-TRIM36 attggcgcgccatcggagtctggggagatgag TRIM36-PacI gtcattaattaagctacatgtcctcttggtatt Trim36 (1-132) PacI_R gtcattaattaagcagaccattgattcctcgtt Trim36 (1-418) PacI_R gtcattaattaaggtctatgccactagagaaaa Trim36 (133-728) AscI_F attggcgcgccatttcgaaacttcactttgga Trim36 (256-728) AscI_F attggcgcgccattaaaggaaaagctttcaaa TRIM36 H50A F tacacatttatgacagatactggcttggcaagggagaatcaa tgggtgg TRIM36 H50A R ccacccattgattctcccttgccaagccagtatctgtcataaat gtgta TRIM36-PacI stop-codon free ctggttaattaacatgtcctcttggtatt version PacI-AAAGGSG-Ub(F) ctggttaattaaggcggccgcaggcggatctggtatgcagat tttcgtga XbaI-Ub(R) gcactctagactattaacgaagtctcaacac TRIM36 RING loop deletion for ttcccttgccctggctgt 138

TRIM36 RING loop deletion agttcttttacacatttatgac rev TRIM36 RNAi_Resistant KI agttgtccggggtagtgtga check F TRIM36 RNAi_Resistant KI aactggaagtctttgctgcgg check R

Table 3-5. Primers Used in this Work.

139

Chapter 4 Conclusion and Future Directions

140

Conclusion and Future Directions

4.1 Identification of novel centriolar components using proximity based BioID profiling

Proteomics facilitated by mass spectrometry has been a powerful method to determine the composition of purified centrosomes (Andersen et al., 2003). Here I describe an approach exploiting BioID followed by mass spectrometry to map protein proximity interactions of centriolar proteins. It allows for capturing transient interactions along centriole duplication pathway (Firat-Karalar et al., 2014; Gupta, Gingras, et al., 2015). BirA* fused proteins of interest (‘baits’) are expressed in HEK293 cells with excess biotin (Roux et al., 2012; Morriswood et al., 2013; Firat-Karalar et al., 2014). This activated biotin by BirA* interacts with lysine residues of ‘preys’ at certain proximity, and the ‘preys’ can then be captured by high- affinity avidin/streptavidin- mediated purification (Green, 1963).

However, BioID has its limitations. Due to slow labeling kinetics, the biotinylation reaction is performed over a 24 hours period, which is longer than one cycle of centriole duplication (Roux et al., 2012; Morriswood et al., 2013). As a result, BioID would not be a good tool in examining PPIs in each stage of the centriole duplication pathway. Proximity labeling methods with shorter labeling time are thus sought to study the spatial and temporal PPIs. As discussed in Chapter 2, an engineered ascorbate peroxidase (APEX) was developed for short proteomic labeling (1-2 hours) (Rhee et al., 2013). In the presence of H2O2 APEX is a monomeric 28 kDa peroxidase that oxidizes numerous phenol derivatives to short-lived phenoxyl radicals (Martell et al., 2012; Rhee et al., 2013). With a time frame as short as 1 min, endogenous proteins can be biotinylated by APEX in reactions containing biotin-phenol and

H2O2 (Rhee et al., 2013). TurboID and miniTurbo are another two novel enzyme-catalyzed proximity labeling approaches with a short labeling time (~10 min) that generated from modifications of BioID (Branon et al., 2018). It will be interesting to explore their efficacy in the identification of dynamic protein interactions during centriole duplication (Rhee et al., 2013; Branon et al., 2018).

141

In order to identify potential centriole duplication factors from the proximity interactome, I conducted a following semi-automated high-throughput centriole duplication RNAi screen in S-phase arrested U-2 OS cells. Compared to assays using cycling cells, this assay has a higher sensitivity in detecting centriole duplication factors since it provides a higher dynamic range due to the centriole over-duplication in S-phase arrested U-2 OS (Balczon et al., 1995a). Centriole over-duplication defects was the main read-out of the assay, and 122 candidates that suppressed centriole duplication were identified from 500 proteins assessed that were chosen from the aforementioned interactome. The screen provided an intensive pool of potential centriole duplication factors to follow up on. CCDC53 is one of them. CCDC53, a subunit of the WASH complex, was identified in the screen that strongly suppress centriole over-duplication when down-regulated (Visweshwaran et al., 2018). The WASH complex were identified as actin nucleation promoting factors by activating the Arp2/3 complex (Derivery et al., 2009; Visweshwaran et al., 2018). Recent work has shown that centrosomes also serve as Actin nucleation sites, and the actin-filament nucleation is dependent on Apr2/3 complexes, a process promoted by WASH complex (Farina et al., 2016). Dr. Yi from Pelletier lab further tightened the bond by demonstrating that centriole/centrosomal proteins CEP192-PLK4/AURKB complex can be recruited to cell cortex drive protrusive activity and cell motility through formin-dependent actin remodeling (Luo et al., 2019). Investigating the role of CCDC53 will further reveal the correlations between centrioles functions and actin networks.

4.2 TRIM36 in centriole duplication and spindle formation

Lying at the core of the protein cluster discovered using BioID, TRIM36 is a crucial candidate that regulates centriole duplication pathway. I proved that TRIM36 locates at early stages of centriole duplication pathway and regulates γ-tubulin level at centrosome. γ-tubulin is critical for centriole duplication (Yuba-Kubo et al., 2005; Haren et al., 2006). SAS6 is unstable in the absence of centrosomal γ-tubulin. Depletion of TRIM36 impedes centrosomal γ-tubulin recruitment, thus inhibits cartwheel formation process. By co-depleting TRIM36 and RBM14, a negative regulator of centriole duplication and γ-tubulin recruitment, the centrosomal level of γ- tubulin is restored and the centriole duplication defect is rescued. Interestingly, the SAS6 142

degradation event doesn’t have an effect on STIL, rather led to a slight increase of STIL level at the centrosomes.

The observation raises questions revolving around SAS6 recruitment mechanisms that multiple theories have been trying to explain. In a model that Fong et al. proposed in 2014, SAS6 is initially recruited to the proximal lumen of the mother centrioles, where the cartwheel structure is formed, and in the next step, the established cartwheel dissociates from the lumen and transfers to the outside wall in a PLK4 and STIL dependent manner (Fong et al., 2014). Recent work by Ito et al. revealed two complementing pathways to recruit SAS6 to the spindle pole body of fission yeast (Ito et al., 2018). The fission yeast Pericentrin, Pcp1, can be pulled down together with Bld10 (CEP135 ortholog), SAS6 and SAS4 (CENPJ ortholog) by co- Immunoprecipitation (Ito et al., 2018). They also demonstrated that SAS-6 is recruited by two complementary pathways by Ana2 and PLP respectively in Drosophila through performing single and double depletion of the two proteins (Ito et al., 2018). This is contrary to the canonical view of PCNT coming into the pathway at a relatively late stage (Lee and Rhee, 2011; Lawo et al., 2012b). Ito et al. showed evidence of Pcp1 recruiting SAS6 for centriole biogenesis independent of STIL (Ito et al., 2018), as neither PLK4 nor STIL exists in fission yeast’s genome (Ito et al., 2018). The newly discovered mechanism suggests centriole duplication pathway is not a simple linear process. Proteins with original identified functions at different stages are potentially intertwined with each other.

The hypothesis that the centriole duplication pathway is not a simple linear process is in agreement with my observation that centrosomal γ-tubulin recruitment determines SAS6 stability. Traditionally, similar as Pericentrin, γ-tubulin is considered to function at the maturation stage of centriole duplication (Haren et al., 2006; Haren, Stearns and Lüders, 2009). γ-tubulin may serve as a template for the assembly of centriole microtubule walls, positioning it at the same timing as cartwheel formation during the process. The presence TRIM36, on the other hand, is necessary for γ-tubulin accumulation at the centrosome. Whether γ-tubulin is a direct substrate of TRIM36 remains unknown and needs future exploration. If it is, how ubiquitination impairs centrosomeal γ-tubulin localization will be an interesting topic to investigate. On the other side, if TRIM36 affects γ-tubulin indirectly through forming a complex with other proteins, finding out what could be the bridge will be another project worth pursuing. 143

In this thesis, I have demonstrated that over-expressed TRIM36 localized on the MTs, and both its depletion and over-expression disrupted spindle assembly. This could be a result of both directly binding to the microtubules and binding to other proteins that facilitate spindle assembly. I have observed that the RBCC domain of TRIM36 has a high affinity with microtubules, whilst a catalytic dead mutation at its RING domain will disrupt the association. In line with my observation, study conducted by Emi Yoshigai et al. and Houston et al. demonstrated that TRIM36 is essential for vegetal microtubule array formation and normal somite formation during early development of Xenopus embryos (Cuykendall and Houston, 2009b; Yoshigai et al., 2009; Olson, Oh and Houston, 2015).

I have provided evidence that one potential mechanism behind TRIM36 regulating spindle formation is also through γ-tubulin recruitment. The hypothesis is best supported by the observation that co-depletion of TRIM36 and RBM14 diminishes spindle defects caused by TRIM36 depletion. RBM14 is reported as a negative centriole duplication regulator, and its depletion results in accumulation of γ-tubulin at the spindle poles (Shiratsuchi et al., 2015). In the future, it will be intriguing to investigate if the antagonism is specific between the pair, and to discover more antagonising pairs in spindle assembly regulation.

4.3 TRIM36 as an E3 ligase

In section 3.4.3, I explained that the TRIM36 RIMG domain only is sufficient in rescuing centriole duplication defects caused by TRIM36 depletion, while the TRIM36 truncations that lack RING domain failed to do so. I also demonstrated a catalytically dead form of TRIM36 with a key catalytic site Histidine residue mutation to Alanine (H50A) compromised its ability to rescue the centriole duplication defect (Meroni and Diez-Roux, 2005; Stone et al., 2005). These results revealed that the integrity of RING domain is crucial for TRIM36 as a centriole duplication factor.

Previous work has shown that TRIM36 has E3 ubiquitin ligase activity with the E2 enzyme Ubc4 (Miyajima et al., 2009). The next step is to follow up with TRIM36 E3 ligase activity in human cells, exploring the type of Ub chains it will generate, which is crucial in determining the fate of its substrates. It will be critical to search for substrates that can further 144

map TRIM36 into centriole duplication process. In Chapter 3, I have demonstrated that the interaction network of TRIM36 identified using UBAIT provided a valuable resource for identification of potential TRIM36 substrates. Besides RBM14 and γ-tubulin, CDK1 is also enriched in the UBAIT-TRIM36 interactome (average spectral counts of 108.5 in UBAIT- TRIM36 vs. 5 in control). In mitosis, CDK1 binds to STIL, preventing the formation of PLK4- STIL complex, and inhibiting STIL phosphorylation by PLK4 (Zitouni et al., 2016). The phosphorylation event where PLK4 phosphorylates STIL at S1108 and S1116 within its STAN domain is critical for SAS6 binding to STIL and enabling cartwheel formation (Arquint et al., 2015). If TRIM36 triggers ubiquitination of CDK1 that leads to its proteasome-mediated degradation, CDK1 protein level will increase after TRIM36 depletion, enforcing its inhibition of STIL phosphorylation, leading to cartwheel formation failure and centriole duplication defects. It will be interesting to verify the interaction between CDK1 and TRIM36, and to examine CDK1 protein level after TRIM36 down-regulation. It will also be important to examine its ubiquitination status with TRIM36 up- or down-regulation to determine whether it is ubiquitinated by TRIM36.

4.4 TRIM36 in human diseases

In light of the pivotal roles of TRIM36 in both centriole duplication and spindle formation, it will be interesting to further investigate its role in tumorigenesis considering the link between centrosome amplification and cancer (Nigg, 2006; Nigg and Holland, 2018). Intriguingly, quantitative RT-PCR analysis on human tissue samples reveals that TRIM36 is significantly up regulated in prostate cancer samples (4-33 folds compared to normal tissues) (Balint et al., 2004; Liang et al., 2018). Remarkably, another group also observed that via inhibition of the MAPK/ERK phosphorylation pathway, TRIM36 expression significantly delayed prostate cancer cell cycle progression and inhibited cell proliferation in vitro and in vivo (Liang et al., 2018). Similarly, another C-I subfamily member, TRIM18 (MID1), associates with PP2A and binds to microtubules (Han, Du and Massiah, 2011). MID1 is also found to be upregulated in lung adenocarcinoma, and it is demonstrated to affect cell cycle progression, proliferation and apoptosis (Liang et al., 2018; Zhang et al., 2018). The evidences suggest a 145

promising cancer therapy targeting TRIM36 and TRIM family proteins and controlling their protein level in human cells.In this thesis, I have described a novel role for the E3 ligase TRIM36 in regulating centriole duplication and spindle formation. TRIM36 interacts with and promotes the centrosomal recruitment of the key γ-TuRC component protein γ-tubulin. RBM14, a negative factor of centriole duplication (Shiratsuchi et al., 2015), have antagonizing effect on γ-tubulin recruitment with TRIM36. Given the critical role of γ-tubulin in centriole duplication and spindle formation, future work investigating the antagonizing effect of TRIM36 and RBM14 in γ-tubulin -dependent processes will be improtant.

146

Reference

Abe, Y. et al. (2006) ‘LATS2-Ajuba complex regulates γ-tubulin recruitment to centrosomes and spindle organization during mitosis’, FEBS Letters. John Wiley & Sons, Ltd, 580(3), pp. 782– 788. doi: 10.1016/j.febslet.2005.12.096.

Adames, N. R. and Cooper, J. A. (2000) ‘Microtubule interactions with the cell cortex causing nuclear movements in Saccharomyces cerevisiae.’, The Journal of cell biology. The Rockefeller University Press, 149(4), pp. 863–74. Available at: http://www.ncbi.nlm.nih.gov/pubmed/10811827 (Accessed: 30 March 2019).

Adon, A. M. et al. (2010) ‘Cdk2 and Cdk4 Regulate the Centrosome Cycle and Are Critical Mediators of Centrosome Amplification in p53-Null Cells’, Molecular and Cellular Biology. doi: 10.1128/MCB.00253-09.

Afzelius, B. A. (1976) ‘A human syndrome caused by immotile cilia.’, Science (New York, N.Y.). American Association for the Advancement of Science, 193(4250), pp. 317–9. doi: 10.1126/SCIENCE.1084576.

Akutsu, M., Dikic, I. and Bremm, A. (2016) ‘Ubiquitin chain diversity at a glance’, J Cell Sci. The Company of Biologists Ltd, 129(5), pp. 875–880. doi: 10.1242/JCS.183954.

Alushin, G. M. et al. (2014) ‘High-resolution microtubule structures reveal the structural transitions in αβ-tubulin upon GTP hydrolysis.’, Cell. NIH Public Access, 157(5), pp. 1117–29. doi: 10.1016/j.cell.2014.03.053.

Alves-Cruzeiro, J. M. da C., Nogales-Cadenas, R. and Pascual-Montano, A. D. (2014) ‘CentrosomeDB: a new generation of the centrosomal proteins database for Human and Drosophila melanogaster.’, Nucleic acids research. Oxford University Press, 42(Database issue), pp. D430-6. doi: 10.1093/nar/gkt1126.

Andersen, J. S. et al. (2003) ‘Proteomic characterization of the human centrosome by protein correlation profiling.’, Nature. doi: 10.1038/nature02166.

147

Archinti, M. et al. (2010) ‘SPICE - a previously uncharacterized protein required for centriole duplication and mitotic chromosome congression’, Journal of Cell Science. doi: 10.1242/jcs.069963.

Arquint, C. et al. (2012a) ‘Cell-cycle-regulated expression of STIL controls centriole number in human cells.’, Journal of cell science. The Company of Biologists Ltd, 125(Pt 5), pp. 1342–52. doi: 10.1242/jcs.099887.

Arquint, C. et al. (2012b) ‘Cell-cycle-regulated expression of STIL controls centriole number in human cells.’, Journal of cell science. The Company of Biologists Ltd, 125(Pt 5), pp. 1342–52. doi: 10.1242/jcs.099887.

Arquint, C. et al. (2015) ‘STIL binding to Polo-box 3 of PLK4 regulates centriole duplication.’, eLife. eLife Sciences Publications, Ltd, 4. doi: 10.7554/eLife.07888.

Arquint, C., Gabryjonczyk, A.-M. and Nigg, E. A. (2014) ‘Centrosomes as signalling centres.’, Philosophical transactions of the Royal Society of London. Series B, Biological sciences. The Royal Society, 369(1650). doi: 10.1098/rstb.2013.0464.

Arquint, C. and Nigg, E. A. (2016) ‘The PLK4-STIL-SAS-6 module at the core of centriole duplication’. doi: 10.1042/BST20160116.

Ault, J. G. and Rieder, C. L. (1994) ‘Centrosome and kinetochore movement during mitosis’, Current Opinion in Cell Biology. Elsevier Current Trends, 6(1), pp. 41–49. doi: 10.1016/0955- 0674(94)90114-7.

Azimzadeh, J. et al. (2009a) ‘hPOC5 is a centrin-binding protein required for assembly of full- length centrioles.’, The Journal of cell biology. The Rockefeller University Press, 185(1), pp. 101–14. doi: 10.1083/jcb.200808082.

Azimzadeh, J. et al. (2009b) ‘hPOC5 is a centrin-binding protein required for assembly of full- length centrioles’, The Journal of Cell Biology, 185(1), pp. 101–114. doi: 10.1083/jcb.200808082.

Azimzadeh, J. (2014) ‘Exploring the evolutionary history of centrosomes.’, Philosophical 148

transactions of the Royal Society of London. Series B, Biological sciences. The Royal Society, 369(1650). doi: 10.1098/rstb.2013.0453.

Bahe, S. et al. (2005) ‘Rootletin forms centriole-associated filaments and functions in centrosome cohesion.’, The Journal of cell biology. The Rockefeller University Press, 171(1), pp. 27–33. doi: 10.1083/jcb.200504107.

Balczon, R. et al. (1995a) ‘Dissociation of centrosome replication events from cycles of DNA synthesis and mitotic division in hydroxyurea-arrested Chinese hamster ovary cells.’, The Journal of cell biology. The Rockefeller University Press, 130(1), pp. 105–15. doi: 10.1083/jcb.130.1.105.

Balczon, R. et al. (1995b) ‘Dissociation of centrosome replication events from cycles of DNA synthesis and mitotic division in hydroxyurea-arrested Chinese hamster ovary cells.’, The Journal of cell biology. The Rockefeller University Press, 130(1), pp. 105–15. Available at: http://www.ncbi.nlm.nih.gov/pubmed/7790366 (Accessed: 14 March 2019).

Balint, I. et al. (2004) ‘Cloning and characterisation of the RBCC728/TRIM36 zinc-binding protein from the tumor suppressor gene region at chromosome 5q22.3’. doi: 10.1016/j.gene.2004.02.045.

Barbelanne, M. and Tsang, W. Y. (2014) ‘Molecular and cellular basis of autosomal recessive primary microcephaly.’, BioMed research international. Hindawi Limited, 2014, p. 547986. doi: 10.1155/2014/547986.

Berti, C. et al. (2002) ‘TRIM9 is specifically expressed in the embryonic and adult nervous system’, Mechanisms of Development. Elsevier, 113(2), pp. 159–162. doi: 10.1016/S0925- 4773(02)00013-8.

Blachon, S. et al. (2008) ‘Drosophila asterless and Vertebrate Cep152 Are Orthologs Essential for Centriole Duplication’, Genetics. Genetics Society of America, 180(4), p. 2081. doi: 10.1534/GENETICS.108.095141.

Boese, C. J. et al. (2018) ‘Asterless is a Polo-like kinase 4 substrate that both activates and

149

inhibits kinase activity depending on its phosphorylation state.’, Molecular biology of the cell. American Society for Cell Biology, 29(23), pp. 2874–2886. doi: 10.1091/mbc.E18-07-0445.

Bonaccorsi, S., Giansanti, M. G. and Gatti, M. (1998) ‘Spindle Self-organization and Cytokinesis During Male Meiosis in asterless Mutants of Drosophila melanogaster’, The Journal of Cell Biology. The Rockefeller University Press, 142(3), p. 751. Available at: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2148166/?report=reader (Accessed: 6 April 2019).

Bond, J. and Woods, C. G. (2006) ‘Cytoskeletal genes regulating brain size’, Current Opinion in Cell Biology. Elsevier Current Trends, 18(1), pp. 95–101. doi: 10.1016/J.CEB.2005.11.004.

Bornens, M. and Azimzadeh, J. (2007) ‘Origin and Evolution of the Centrosome’, in Eukaryotic Membranes and Cytoskeleton. New York, NY: Springer New York, pp. 119–129. doi: 10.1007/978-0-387-74021-8_10.

Boveri, T. (1887) ‘Ueber den Antheil des Spermatozoon an der Theilung des Eies’. Available at: https://scholar.google.com/scholar?cluster=3848878939670590021&hl=en&as_sdt=2005&sciod t=0,5 (Accessed: 5 May 2019).

Boveri, T. (1888) ‘Über partielle Befruchtung’. Available at: https://scholar.google.com/scholar_lookup?journal=Sitz.- Ber.+Ges.+Morph.+Physiol.+München&title=Ueber+partielle+Befruchtung&author=T+Boveri &volume=4&publication_year=1888&pages=64-72& (Accessed: 5 May 2019).

Boveri, T. (2008) ‘Concerning the origin of malignant tumours by Theodor Boveri. Translated and annotated by Henry Harris.’, Journal of cell science. The Company of Biologists Ltd, 121 Suppl 1(Supplement 1), pp. 1–84. doi: 10.1242/jcs.025742.

Boyer, N. P. et al. (2018) ‘Mammalian TRIM67 Functions in Brain Development and Behavior.’, eNeuro. Society for Neuroscience, 5(3). doi: 10.1523/ENEURO.0186-18.2018.

Branon, T. C. et al. (2018) ‘Efficient proximity labeling in living cells and organisms with TurboID’, Nature Biotechnology. Nature Publishing Group, 36(9), pp. 880–887. doi:

150

10.1038/nbt.4201.

Braten, O. et al. (2016) ‘Numerous proteins with unique characteristics are degraded by the 26S proteasome following monoubiquitination.’, Proceedings of the National Academy of Sciences of the United States of America. National Academy of Sciences, 113(32), pp. E4639-47. doi: 10.1073/pnas.1608644113. van Breugel, M. et al. (2011) ‘Structures of SAS-6 suggest its organization in centrioles.’, Science (New York, N.Y.). American Association for the Advancement of Science, 331(6021), pp. 1196–9. doi: 10.1126/science.1199325. van Breugel, M. et al. (2014) ‘Structure of the SAS-6 cartwheel hub from Leishmania major.’, eLife. eLife Sciences Publications, Ltd, 3, p. e01812. doi: 10.7554/eLife.01812.

Brinkley, B. R. (2001) ‘Managing the centrosome numbers game: from chaos to stability in cancer cell division’, Trends in Cell Biology. Elsevier Current Trends, 11(1), pp. 18–21. doi: 10.1016/S0962-8924(00)01872-9.

Brown, N. G. et al. (2015) ‘RING E3 mechanism for ubiquitin ligation to a disordered substrate visualized for human anaphase-promoting complex’, Proceedings of the National Academy of Sciences. doi: 10.1073/pnas.1504161112.

Brugués, J. et al. (2012) ‘Nucleation and transport organize microtubules in metaphase spindles.’, Cell. Elsevier, 149(3), pp. 554–64. doi: 10.1016/j.cell.2012.03.027.

Byers, B. and Goetsch, L. (1975) ‘Behavior of spindles and spindle plaques in the cell cycle and conjugation of Saccharomyces cerevisiae.’, Journal of bacteriology. American Society for Microbiology (ASM), 124(1), pp. 511–23. Available at: http://www.ncbi.nlm.nih.gov/pubmed/1100612 (Accessed: 6 May 2019).

Cajanek, L., Glatter, T. and Nigg, E. (2015) ‘The E3 ubiquitin ligase Mib1 regulates Plk4 and centriole biogenesis’, Journal of Cell Science, (128), pp. 1674–1682. doi: 10.1242/jcs.166496.

Case, C. M. et al. (2013) ‘CKAP2 Ensures Chromosomal Stability by Maintaining the Integrity of Microtubule Nucleation Sites’, PLoS ONE. doi: 10.1371/journal.pone.0064575. 151

Cassimeris, L. U. et al. (1987) ‘Dynamic instability of microtubules’, BioEssays. Nature Publishing Group, pp. 149–154. doi: 10.1002/bies.950070403.

Chan, J. Y. (2011) ‘A clinical overview of centrosome amplification in human cancers.’, International journal of biological sciences. Ivyspring International Publisher, 7(8), pp. 1122– 44. Available at: http://www.ncbi.nlm.nih.gov/pubmed/22043171 (Accessed: 10 February 2019).

Chang, C.-W. et al. (2016) ‘CEP295 interacts with microtubules and is required for centriole elongation.’, Journal of cell science. Company of Biologists, 129(13), pp. 2501–13. doi: 10.1242/jcs.186338.

Choi, H. et al. (2011) ‘SAINT: probabilistic scoring of affinity purification-mass spectrometry data.’, Nature methods. NIH Public Access, 8(1), pp. 70–3. doi: 10.1038/nmeth.1541.

Choi, H. et al. (2012) ‘Analyzing protein-protein interactions from affinity purification-mass spectrometry data with SAINT.’, Current protocols in bioinformatics. NIH Public Access, Chapter 8, p. Unit8.15. doi: 10.1002/0471250953.bi0815s39.

Christie, K. R. and Blake, J. A. (2018) ‘Sensing the cilium, digital capture of ciliary data for comparative genomics investigations’, Cilia. BioMed Central, 7(1), p. 3. doi: 10.1186/s13630- 018-0057-0.

Ciechanover, A. (2015) ‘The unravelling of the ubiquitin system’, Nature Reviews Molecular Cell Biology. Nature Publishing Group, 16(5), pp. 322–324. doi: 10.1038/nrm3982.

Ciechanover, A. and Kwon, Y. T. (2017) ‘Protein Quality Control by Molecular Chaperones in Neurodegeneration’, Frontiers in Neuroscience. Frontiers, 11, p. 185. doi: 10.3389/fnins.2017.00185.

Clague, M. J. et al. (2013) ‘Deubiquitylases From Genes to Organism’, Physiological Reviews. American Physiological Society Bethesda, MD, 93(3), pp. 1289–1315. doi: 10.1152/physrev.00002.2013.

Comartin, D., Gupta, G. D., et al. (2013) ‘CEP120 and SPICE1 cooperate with CPAP in centriole elongation’, Current Biology, 23(14), pp. 1360–1366. doi: 10.1016/j.cub.2013.06.002. 152

Comartin, D., Gupta, G. D., et al. (2013) ‘Report CEP120 and SPICE1 Cooperate with CPAP in Centriole Elongation’, Current Biology, 23, pp. 1360–1366. doi: 10.1016/j.cub.2013.06.002.

Craig, R. and Beavis, R. C. (2004) ‘TANDEM: Matching proteins with tandem mass spectra’, Bioinformatics. doi: 10.1093/bioinformatics/bth092.

Crosio, C. et al. (2009) ‘Mitotic Phosphorylation of Histone H3: Spatio-Temporal Regulation by Mammalian Aurora Kinases’, Molecular and Cellular Biology. doi: 10.1128/mcb.22.3.874- 885.2002.

Cunha-Ferreira, I. et al. (2009) ‘The SCF/Slimb Ubiquitin Ligase Limits Centrosome Amplification through Degradation of SAK/PLK4’, Current Biology. Cell Press, 19(1), pp. 43– 49. doi: 10.1016/J.CUB.2008.11.037.

Cunha-Ferreira, I. et al. (2013) ‘Regulation of autophosphorylation controls PLK4 self- destruction and centriole number.’, Current biology : CB. Elsevier, 23(22), pp. 2245–2254. doi: 10.1016/j.cub.2013.09.037.

Cuykendall, T. N. and Houston, D. W. (2009a) ‘c’, Development (Cambridge, England). Company of Biologists, 136(18), pp. 3057–65. doi: 10.1242/dev.036855.

Cuykendall, T. N. and Houston, D. W. (2009b) ‘Vegetally localized Xenopus trim36 regulates cortical rotation and dorsal axis formation.’, Development (Cambridge, England). Company of Biologists, 136(18), pp. 3057–65. doi: 10.1242/dev.036855.

D’Angiolella, V. et al. (2010) ‘SCF(Cyclin F) controls centrosome homeostasis and mitotic fidelity through CP110 degradation.’, Nature. NIH Public Access, 466(7302), pp. 138–42. doi: 10.1038/nature09140. van Dam, T. J. et al. (2013) ‘The SYSCILIA gold standard (SCGSv1) of known ciliary components and its applications within a systems biology consortium.’, Cilia. BioMed Central, 2(1), p. 7. doi: 10.1186/2046-2530-2-7.

Dammermann, A. and Merdes, A. (2002) ‘Assembly of centrosomal proteins and microtubule organization depends on PCM-1.’, The Journal of cell biology. The Rockefeller University Press, 153

159(2), pp. 255–66. doi: 10.1083/jcb.200204023.

David Pruyne and Anthony Bretscher (2000) Polarization of cell growth in yeast I. Establishment and maintenance of polarity states. Available at: http://jcs.biologists.org/content/joces/113/3/365.full.pdf (Accessed: 30 March 2019).

Delattre, M. et al. (2004) ‘Centriolar SAS-5 is required for centrosome duplication in C. elegans’, Nature Cell Biology. Nature Publishing Group, 6(7), pp. 656–664. doi: 10.1038/ncb1146.

Delgehyr, N. et al. (2012) ‘Drosophila Mgr, a Prefoldin subunit cooperating with von Hippel Lindau to regulate tubulin stability.’, Proceedings of the National Academy of Sciences of the United States of America. National Academy of Sciences, 109(15), pp. 5729–34. doi: 10.1073/pnas.1108537109.

Derivery, E. et al. (2009) ‘The Arp2/3 Activator WASH Controls the Fission of Endosomes through a Large Multiprotein Complex’, Developmental Cell. doi: 10.1016/j.devcel.2009.09.010.

Desai, A. and Mitchison, T. J. (1997) ‘MICROTUBULE POLYMERIZATION DYNAMICS’, Annual Review of Cell and Developmental Biology. Annual Reviews 4139 El Camino Way, P.O. Box 10139, Palo Alto, CA 94303-0139, USA , 13(1), pp. 83–117. doi: 10.1146/annurev.cellbio.13.1.83.

DiStasio, A. et al. (2017) ‘Copb2 is essential for embryogenesis and hypomorphic mutations cause human microcephaly’, Human Molecular Genetics. Narnia, 26(24), pp. 4836–4848. doi: 10.1093/hmg/ddx362.

Do, L. D. et al. (2018) ‘TRIM9 and TRIM67 Are New Targets in Paraneoplastic Cerebellar Degeneration’, The Cerebellum. Springer US, pp. 1–10. doi: 10.1007/s12311-018-0987-5.

Duensing, S. et al. (2003) ‘Excessive centrosome abnormalities without ongoing numerical chromosome instability in a Burkitt’s lymphoma’, Molecular Cancer. doi: 10.1186/1476-4598-2- 30.

Dumont, S. and Mitchison, T. J. (2009) ‘Force and length in the mitotic spindle.’, Current 154

biology : CB. Elsevier, 19(17), pp. R749-61. doi: 10.1016/j.cub.2009.07.028.

Dzhindzhev, N. S. et al. (2010) ‘Asterless is a scaffold for the onset of centriole assembly’, Nature, 467(7316), pp. 714–718. doi: 10.1038/nature09445.

Dzhindzhev, N. S. et al. (2014) ‘Plk4 phosphorylates Ana2 to trigger Sas6 recruitment and procentriole formation.’, Current biology : CB. Elsevier, 24(21), pp. 2526–32. doi: 10.1016/j.cub.2014.08.061.

Dzhindzhev, N. S. et al. (2017) ‘Two-step phosphorylation of Ana2 by Plk4 is required for the sequential loading of Ana2 and Sas6 to initiate procentriole formation.’, Open biology. The Royal Society, 7(12). doi: 10.1098/rsob.170247.

Eguren, M. et al. (2014) ‘A Synthetic Lethal Interaction between APC/C and Topoisomerase Poisons Uncovered by Proteomic Screens’, Cell Reports. Cell Press, 6(4), pp. 670–683. doi: 10.1016/J.CELREP.2014.01.017.

Erickson, H. P. and O’Brien, E. T. (1992) ‘Microtubule Dynamic Instability and GTP Hydrolysis’, Annual Review of Biophysics and Biomolecular Structure. Annual Reviews 4139 El Camino Way, P.O. Box 10139, Palo Alto, CA 94303-0139, USA , 21(1), pp. 145–166. doi: 10.1146/annurev.bb.21.060192.001045.

Esposito, D., Koliopoulos, M. G. and Rittinger, K. (2017) ‘Structural determinants of TRIM protein function’. doi: 10.1042/BST20160325.

Faheem, M. et al. (2015) ‘Molecular genetics of human primary microcephaly: an overview’, BMC Medical Genomics. BioMed Central, 8(S1), p. S4. doi: 10.1186/1755-8794-8-S1-S4.

Faragher, A. J. and Fry, A. M. (2003) ‘Nek2A kinase stimulates centrosome disjunction and is required for formation of bipolar mitotic spindles.’, Molecular biology of the cell. American Society for Cell Biology, 14(7), pp. 2876–89. doi: 10.1091/mbc.e03-02-0108.

Farina, F. et al. (2016) ‘The centrosome is an actin-organizing centre’, Nature Cell Biology. doi: 10.1038/ncb3285.

155

Firat-Karalar, E. N. et al. (2014) ‘Proximity Interactions among Centrosome Components Identify Regulators of Centriole Duplication’, Current Biology. Cell Press, 24(6), pp. 664–670. doi: 10.1016/J.CUB.2014.01.067.

Fong, C. S. et al. (2014) ‘SAS-6 assembly templated by the lumen of cartwheel-less centrioles precedes centriole duplication.’, Developmental cell. NIH Public Access, 30(2), pp. 238–45. doi: 10.1016/j.devcel.2014.05.008.

Fraile, J. M. et al. (2016) ‘The deubiquitinase USP54 is overexpressed in colorectal cancer stem cells and promotes intestinal tumorigenesis.’, Oncotarget. Impact Journals, LLC, 7(46), pp. 74427–74434. doi: 10.18632/oncotarget.12769.

Frank, D. J. and Roth, M. B. (1998) ‘ncl-1 is required for the regulation of cell size and ribosomal RNA synthesis in Caenorhabditis elegans.’, The Journal of cell biology. The Rockefeller University Press, 140(6), pp. 1321–9. Available at: http://www.ncbi.nlm.nih.gov/pubmed/9508766 (Accessed: 12 February 2019).

Fry, A. M. et al. (1998) ‘C-Nap1, a Novel Centrosomal Coiled-Coil Protein and Candidate Substrate of the Cell Cycle–regulated Protein Kinase Nek2’, The Journal of Cell Biology. Rockefeller University Press, 141(7), pp. 1563–1574. doi: 10.1083/JCB.141.7.1563.

Fu, J. et al. (2016) ‘Conserved molecular interactions in centriole-to-centrosome conversion.’, Nature cell biology. Europe PMC Funders, 18(1), pp. 87–99. doi: 10.1038/ncb3274.

Fung, K. Y. Y., Dai, L. and Trimble, W. S. (2014) ‘Cell and Molecular Biology of Septins Equal contribution 2’, International Review of Cell and Molecular Biology, 310, pp. 289–339. doi: 10.1016/B978-0-12-800180-6.00007-4.

Gadde, S. and Heald, R. (2004) ‘Mechanisms and Molecules of the Mitotic Spindle’, Current Biology. Cell Press, 14(18), pp. R797–R805. doi: 10.1016/J.CUB.2004.09.021.

Genschik, P., Sumara, I. and Lechner, E. (2013) ‘The emerging family of CULLIN3-RING ubiquitin ligases (CRL3s): cellular functions and disease implications.’, The EMBO journal. European Molecular Biology Organization, 32(17), pp. 2307–20. doi: 10.1038/emboj.2013.173.

156

Godinho, S. A. and Pellman, D. (2014) ‘Causes and consequences of centrosome abnormalities in cancer.’, Philosophical transactions of the Royal Society of London. Series B, Biological sciences. The Royal Society, 369(1650). doi: 10.1098/rstb.2013.0467.

Gomez-Ferreria, M. A. et al. (2012) ‘CEP192 interacts physically and functionally with the K63- deubiquitinase CYLD to promote mitotic spindle assembly.’, Cell cycle (Georgetown, Tex.). Taylor & Francis, 11(19), pp. 3555–8. doi: 10.4161/cc.21574.

Gönczy, P. (2012) ‘Towards a molecular architecture of centriole assembly’, Nature Reviews Molecular Cell Biology. Nature Publishing Group, 13(7), pp. 425–435. doi: 10.1038/nrm3373.

Goshima, G. et al. (2008) ‘Augmin: a protein complex required for centrosome-independent microtubule generation within the spindle.’, The Journal of cell biology. The Rockefeller University Press, 181(3), pp. 421–9. doi: 10.1083/jcb.200711053.

Grava, S. et al. (2006) ‘Asymmetric Recruitment of Dynein to Spindle Poles and Microtubules Promotes Proper Spindle Orientation in Yeast’, Developmental Cell. Cell Press, 10(4), pp. 425– 439. doi: 10.1016/J.DEVCEL.2006.02.018.

GREEN, N. M. (1963) ‘AVIDIN. 1. THE USE OF (14-C)BIOTIN FOR KINETIC STUDIES AND FOR ASSAY.’, The Biochemical journal. Portland Press Ltd, 89(3), pp. 585–91. doi: 10.1042/bj0890585.

Griffith, E. et al. (2008) ‘Mutations in pericentrin cause Seckel syndrome with defective ATR- dependent DNA damage signaling.’, Nature genetics. Europe PMC Funders, 40(2), pp. 232–6. doi: 10.1038/ng.2007.80.

Grill, S. W. et al. (2003) The Distribution of Active Force Generators Controls Mitotic Spindle Position Downloaded from. Available at: http://science.sciencemag.org/ (Accessed: 13 February 2019).

Grisendi, S. et al. (2005) ‘Role of nucleophosmin in embryonic development and tumorigenesis’. doi: 10.1038/nature03915.

Gruss, O. J. and Vernos, I. (2004) ‘The mechanism of spindle assembly: functions of Ran and its 157

target TPX2.’, The Journal of cell biology. The Rockefeller University Press, 166(7), pp. 949– 55. doi: 10.1083/jcb.200312112.

Gupta, G. D., Coyaud, É., et al. (2015) ‘A Dynamic Protein Interaction Landscape of the Human Centrosome-Cilium Interface.’, Cell. NIH Public Access, 163(6), pp. 1484–99. doi: 10.1016/j.cell.2015.10.065.

Gupta, G. D., Gingras, A.-C., et al. (2015) ‘A Dynamic Protein Interaction Landscape of the Human Centrosome-Cilium Interface’, CELL, 163, pp. 1484–1499. doi: 10.1016/j.cell.2015.10.065.

Gustafsson, M. G. L. et al. (2008) ‘Three-dimensional resolution doubling in wide-field fluorescence microscopy by structured illumination.’, Biophysical journal. The Biophysical Society, 94(12), pp. 4957–70. doi: 10.1529/biophysj.107.120345.

Habedanck, R. et al. (2005) ‘The Polo kinase Plk4 functions in centriole duplication’, Nature Cell Biology. doi: 10.1038/ncb1320.

Hames, R. S. et al. (2001) ‘APC/C-mediated destruction of the centrosomal kinase Nek2A occurs in early mitosis and depends upon a cyclin A-type D-box.’, The EMBO journal. European Molecular Biology Organization, 20(24), pp. 7117–27. doi: 10.1093/emboj/20.24.7117.

Han, X., Du, H. and Massiah, M. A. (2011) ‘Detection and Characterization of the In Vitro E3 Ligase Activity of the Human MID1 Protein’, Journal of Molecular Biology. Academic Press, 407(4), pp. 505–520. doi: 10.1016/J.JMB.2011.01.048.

Haren, L. et al. (2006) ‘NEDD1-dependent recruitment of the gamma-tubulin ring complex to the centrosome is necessary for centriole duplication and spindle assembly.’, The Journal of cell biology. The Rockefeller University Press, 172(4), pp. 505–15. doi: 10.1083/jcb.200510028.

Haren, L., Stearns, T. and Lüders, J. (2009) ‘Plk1-dependent recruitment of gamma-tubulin complexes to mitotic centrosomes involves multiple PCM components.’, PloS one. Public Library of Science, 4(6), p. e5976. doi: 10.1371/journal.pone.0005976.

Hartill, V. et al. (2017) ‘Meckel-Gruber Syndrome: An Update on Diagnosis, Clinical 158

Management, and Research Advances.’, Frontiers in pediatrics. Frontiers Media SA, 5, p. 244. doi: 10.3389/fped.2017.00244.

Hatakeyama, S. (2017) ‘TRIM Family Proteins: Roles in Autophagy, Immunity, and Carcinogenesis.’, Trends in biochemical sciences. Elsevier, 42(4), pp. 297–311. doi: 10.1016/j.tibs.2017.01.002.

He, R. et al. (2013) ‘LRRC45 Is a Centrosome Linker Component Required for Centrosome Cohesion’, Cell Reports. Cell Press, 4(6), pp. 1100–1107. doi: 10.1016/J.CELREP.2013.08.005.

He, Z. et al. (2017) ‘TRIM36 hypermethylation is involved in polycyclic aromatic hydrocarbons- induced cell transformation*’. doi: 10.1016/j.envpol.2017.03.001.

Heald, R. and Khodjakov, A. (2015) ‘Thirty years of search and capture: The complex simplicity of mitotic spindle assembly’, The Journal of Cell Biology. The Rockefeller University Press, 211(6), p. 1103. doi: 10.1083/JCB.201510015.

Healthcare, G. (2002) Superdex 75 10/300 GL and Superdex 200 10/300 GL. Available at: www.gelifesciences.com/protein-purification (Accessed: 15 January 2019).

Hershko, A. and Ciechanover, A. (1998) THE UBIQUITIN SYSTEM, Annu. Rev. Biochem. Available at: www.annualreviews.org (Accessed: 13 February 2019).

Hirono, M. (2014) ‘Cartwheel assembly.’, Philosophical transactions of the Royal Society of London. Series B, Biological sciences. The Royal Society, 369(1650). doi: 10.1098/rstb.2013.0458.

Holland, A. J. et al. (2012) ‘The autoregulated instability of Polo-like kinase 4 limits centrosome duplication to once per cell cycle.’, Genes & development. Cold Spring Harbor Laboratory Press, 26(24), pp. 2684–9. doi: 10.1101/gad.207027.112.

Hori, A. and Toda, T. (2017) ‘Regulation of centriolar satellite integrity and its physiology.’, Cellular and molecular life sciences : CMLS. Springer, 74(2), pp. 213–229. doi: 10.1007/s00018-016-2315-x.

159

Ibrahim, R. et al. (2009) ‘Electron tomography study of isolated human centrioles’, Microscopy Research and Technique. John Wiley & Sons, Ltd, 72(1), pp. 42–48. doi: 10.1002/jemt.20637.

Ito, D. et al. (2018) ‘An ancestral role of pericentrin in centriole formation through SAS-6 recruitment’, bioRxiv. Cold Spring Harbor Laboratory, p. 313494. doi: 10.1101/313494.

Jacobson, A. D. et al. (2009) ‘The lysine 48 and lysine 63 ubiquitin conjugates are processed differently by the 26 s proteasome.’, The Journal of biological chemistry. American Society for Biochemistry and Molecular Biology, 284(51), pp. 35485–94. doi: 10.1074/jbc.M109.052928.

Joukov, V., Walter, J. C. and De Nicolo, A. (2014) ‘The Cep192-organized aurora A-Plk1 cascade is essential for centrosome cycle and bipolar spindle assembly.’, Molecular cell. NIH Public Access, 55(4), pp. 578–91. doi: 10.1016/j.molcel.2014.06.016.

Kalab, P., Weis, K. and Heald, R. (2002) ‘Visualization of a Ran-GTP gradient in interphase and mitotic Xenopus egg extracts.’, Science (New York, N.Y.). American Association for the Advancement of Science, 295(5564), pp. 2452–6. doi: 10.1126/science.1068798.

Kapanidou, M., Curtis, N. L. and Bolanos-Garcia, V. M. (2017) ‘Cdc20: At the Crossroads between Chromosome Segregation and Mitotic Exit’, Trends in Biochemical Sciences. Elsevier Current Trends, 42(3), pp. 193–205. doi: 10.1016/J.TIBS.2016.12.001.

Kapitein, L. C. et al. (2005) ‘The bipolar mitotic kinesin Eg5 moves on both microtubules that it crosslinks’, Nature. Nature Publishing Group, 435(7038), pp. 114–118. doi: 10.1038/nature03503.

Karsenti, E. and Vernos, I. (2001) ‘The mitotic spindle: a self-made machine.’, Science (New York, N.Y.). American Association for the Advancement of Science, 294(5542), pp. 543–7. doi: 10.1126/science.1063488.

Kazlauskaite, A. et al. (2014) ‘Parkin is activated by PINK1-dependent phosphorylation of ubiquitin at Ser65’, Biochemical Journal. Portland Press Limited, 460(1), pp. 127–141. doi: 10.1042/BJ20140334.

Keller, D. et al. (2014) ‘Mechanisms of HsSAS-6 assembly promoting centriole formation in 160

human cells.’, The Journal of cell biology. The Rockefeller University Press, 204(5), pp. 697– 712. doi: 10.1083/jcb.201307049.

Keller, L. C. et al. (2009) ‘Molecular architecture of the centriole proteome: the conserved WD40 domain protein POC1 is required for centriole duplication and length control.’, Molecular biology of the cell. American Society for Cell Biology, 20(4), pp. 1150–66. doi: 10.1091/mbc.e08-06-0619.

Kemp, C. A. et al. (2004) ‘Centrosome Maturation and Duplication in C. elegans Require the Coiled-Coil Protein SPD-2’, Developmental Cell. Cell Press, 6(4), pp. 511–523. doi: 10.1016/S1534-5807(04)00066-8.

Kessner, D. et al. (2008) ‘ProteoWizard: Open source software for rapid proteomics tools development’, Bioinformatics. doi: 10.1093/bioinformatics/btn323.

Kevenaar, J. T. et al. (2016) ‘Kinesin-Binding Protein Controls Microtubule Dynamics and Cargo Trafficking by Regulating Kinesin Motor Activity’, Current Biology. Cell Press, 26(7), pp. 849–861. doi: 10.1016/J.CUB.2016.01.048.

Khodjakov, A. et al. (2000) ‘Centrosome-independent mitotic spindle formation in vertebrates’, Current Biology. Cell Press, 10(2), pp. 59–67. doi: 10.1016/S0960-9822(99)00276-6.

Kim, M., Fong, C. S. and Bryan Tsou, M.-F. (2014) ‘Centriole Duplication: When PLK4 Meets Ana2/STIL’, Current Biology. Cell Press, 24(21), pp. R1046–R1048. doi: 10.1016/J.CUB.2014.09.079.

Kim, S. T. et al. (2013) ‘The N-end rule proteolytic system in autophagy’, Autophagy. Taylor & Francis, 9(7), pp. 1100–1103. doi: 10.4161/auto.24643.

Kim, T.-S. et al. (2013) ‘Hierarchical recruitment of Plk4 and regulation of centriole biogenesis by two centrosomal scaffolds, Cep192 and Cep152.’, Proceedings of the National Academy of Sciences of the United States of America. National Academy of Sciences, 110(50), pp. E4849-57. doi: 10.1073/pnas.1319656110.

Kirschner, M. and Mitchison, T. (1986) ‘Beyond self-assembly: From microtubules to 161

morphogenesis’, Cell. Cell Press, 45(3), pp. 329–342. doi: 10.1016/0092-8674(86)90318-1.

Kitagawa, D. et al. (2011a) ‘Structural basis of the 9-fold symmetry of centrioles.’, Cell. Elsevier, 144(3), pp. 364–75. doi: 10.1016/j.cell.2011.01.008.

Kitagawa, D. et al. (2011b) ‘Structural basis of the 9-fold symmetry of centrioles.’, Cell. Elsevier, 144(3), pp. 364–75. doi: 10.1016/j.cell.2011.01.008.

Kittler, R. et al. (2005) Production of endoribonuclease-prepared short interfering RNAs for gene silencing in mammalian cells, NATURE METHODS |. Available at: http://deqor.mpi-cbg.de (Accessed: 16 March 2019).

Klebba, J. E. et al. (2015) ‘Two Polo-like kinase 4 binding domains in Asterless perform distinct roles in regulating kinase stability.’, The Journal of cell biology. The Rockefeller University Press, 208(4), pp. 401–14. doi: 10.1083/jcb.201410105.

Kleylein-Sohn, J. et al. (2007) ‘Plk4-Induced Centriole Biogenesis in Human Cells’, Developmental Cell. Cell Press, 13(2), pp. 190–202. doi: 10.1016/J.DEVCEL.2007.07.002.

Klinger, M. et al. (2014) ‘The novel centriolar satellite protein SSX2IP targets Cep290 to the ciliary transition zone.’, Molecular biology of the cell. American Society for Cell Biology, 25(4), pp. 495–507. doi: 10.1091/mbc.E13-09-0526.

Kobayashi, T. et al. (2011) ‘Centriolar kinesin Kif24 interacts with CP110 to remodel microtubules and regulate ciliogenesis.’, Cell. Elsevier, 145(6), pp. 914–25. doi: 10.1016/j.cell.2011.04.028.

Kollman, J. M. et al. (2011) ‘Microtubule nucleation by γ-tubulin complexes’, Nature Reviews Molecular Cell Biology. Nature Publishing Group, 12(11), pp. 709–721. doi: 10.1038/nrm3209.

Komander, D. et al. (2008) ‘The Structure of the CYLD USP Domain Explains Its Specificity for Lys63-Linked Polyubiquitin and Reveals a B Box Module’, Molecular Cell. Cell Press, 29(4), pp. 451–464. doi: 10.1016/J.MOLCEL.2007.12.018.

Komander, D., Clague, M. J. and Urbé, S. (2009) ‘Breaking the chains: structure and function of

162

the deubiquitinases’, Nature Reviews Molecular Cell Biology. Nature Publishing Group, 10(8), pp. 550–563. doi: 10.1038/nrm2731.

Komander, D. and Rape, M. (2012) ‘The Ubiquitin Code’. doi: 10.1146/annurev-biochem- 060310-170328.

Kristariyanto, Y. A. et al. (2015) ‘Assembly and structure of Lys33-linked polyubiquitin reveals distinct conformations.’, The Biochemical journal. Portland Press Limited, 467(2), pp. 345–52. doi: 10.1042/BJ20141502.

Kubo, A. et al. (1999a) ‘Centriolar satellites: molecular characterization, ATP-dependent movement toward centrioles and possible involvement in ciliogenesis.’, The Journal of cell biology. The Rockefeller University Press, 147(5), pp. 969–80. Available at: http://www.ncbi.nlm.nih.gov/pubmed/10579718 (Accessed: 27 January 2019).

Kubo, A. et al. (1999b) ‘Centriolar satellites: molecular characterization, ATP-dependent movement toward centrioles and possible involvement in ciliogenesis.’, The Journal of cell biology. The Rockefeller University Press, 147(5), pp. 969–80. Available at: http://www.ncbi.nlm.nih.gov/pubmed/10579718 (Accessed: 14 February 2019).

Kupke, T., Malsam, J. and Schiebel, E. (2017) ‘A ternary membrane protein complex anchors the spindle pole body in the nuclear envelope in budding yeast.’, The Journal of biological chemistry. American Society for Biochemistry and Molecular Biology, 292(20), pp. 8447–8458. doi: 10.1074/jbc.M117.780601.

Kwon, M. et al. (2008) ‘Mechanisms to suppress multipolar divisions in cancer cells with extra centrosomes.’, Genes & development. Cold Spring Harbor Laboratory Press, 22(16), pp. 2189– 203. doi: 10.1101/gad.1700908.

Kwon, Y. T. and Ciechanover, A. (2017) ‘The Ubiquitin Code in the Ubiquitin-Proteasome System and Autophagy’, Trends in Biochemical Sciences. Elsevier Current Trends, 42(11), pp. 873–886. doi: 10.1016/J.TIBS.2017.09.002.

Lane, J. D. et al. (1964) The Enzymatic Synthesis of Holotranscarboxylase from

163

Apotranscarboxylase and ( +)-Biotin II. INVESTIGATION OF THE REACTION MECHANISM, THE JOURNAL ~~~BIOLO~ICAL CHEMISTRY. Available at: http://www.jbc.org/ (Accessed: 16 May 2019).

Lawo, S. et al. (2009) ‘HAUS, the 8-Subunit Human Augmin Complex, Regulates Centrosome and Spindle Integrity’, Current Biology. Cell Press, 19(10), pp. 816–826. doi: 10.1016/J.CUB.2009.04.033.

Lawo, S. et al. (2012a) ‘Subdiffraction imaging of centrosomes reveals higher-order organizational features of pericentriolar material’, Nature Cell Biology. Nature Publishing Group, 14(11), pp. 1148–1158. doi: 10.1038/ncb2591.

Lawo, S. et al. (2012b) ‘Subdiffraction imaging of centrosomes reveals higher-order organizational features of pericentriolar material’, Nature Cell Biology. doi: 10.1038/ncb2591.

Lawrence, C. J. et al. (2004) ‘A standardized kinesin nomenclature.’, The Journal of cell biology. The Rockefeller University Press, 167(1), pp. 19–22. doi: 10.1083/jcb.200408113.

Leber, B. et al. (2010) ‘Proteins required for centrosome clustering in cancer cells.’, Science translational medicine. American Association for the Advancement of Science, 2(33), p. 33ra38. doi: 10.1126/scitranslmed.3000915.

Leda, M., Holland, A. J. and Goryachev, A. B. (2018) ‘Autoamplification and Competition Drive Symmetry Breaking: Initiation of Centriole Duplication by the PLK4-STIL Network.’, iScience. Elsevier, 8, pp. 222–235. doi: 10.1016/j.isci.2018.10.003.

Lee, J. Y. and Stearns, T. (2013) ‘FOP is a centriolar satellite protein involved in ciliogenesis.’, PloS one. Public Library of Science, 8(3), p. e58589. doi: 10.1371/journal.pone.0058589.

Lee, K. and Rhee, K. (2011) ‘PLK1 phosphorylation of pericentrin initiates centrosome maturation at the onset of mitosis’, J Cell Biol. Rockefeller University Press, 195(7), pp. 1093– 1101. doi: 10.1083/JCB.201106093.

Levine, M. S. et al. (2017) ‘Centrosome Amplification Is Sufficient to Promote Spontaneous Tumorigenesis in Mammals.’, Developmental cell. NIH Public Access, 40(3), p. 313–322.e5. 164

doi: 10.1016/j.devcel.2016.12.022.

Li, J. et al. (2013) ‘USP33 regulates centrosome biogenesis via deubiquitination of the centriolar protein CP110.’, Nature. NIH Public Access, 495(7440), pp. 255–9. doi: 10.1038/nature11941.

Liakopoulos, D. et al. (2003) ‘Asymmetric Loading of Kar9 onto Spindle Poles and Microtubules Ensures Proper Spindle Alignment’, Cell. Cell Press, 112(4), pp. 561–574. doi: 10.1016/S0092-8674(03)00119-3.

Liang, C. et al. (2018) ‘TRIM36, a novel androgen-responsive gene, enhances anti-androgen efficacy against prostate cancer by inhibiting MAPK/ERK signaling pathways article’, Cell Death and Disease. doi: 10.1038/s41419-017-0197-y.

Lim, J. and Yue, Z. (2015) ‘Neuronal Aggregates: Formation, Clearance, and Spreading’, Developmental Cell. Cell Press, 32(4), pp. 491–501. doi: 10.1016/J.DEVCEL.2015.02.002.

Lin, Y.-C. et al. (2013) ‘Human microcephaly protein CEP135 binds to hSAS-6 and CPAP, and is required for centriole assembly’, The EMBO Journal, 32, pp. 1141–1154. doi: 10.1038/emboj.2013.56.

Lin, Y.-N. et al. (2013) ‘CEP120 interacts with CPAP and positively regulates centriole elongation.’, The Journal of cell biology. The Rockefeller University Press, 202(2), pp. 211–9. doi: 10.1083/jcb.201212060.

Liu, Y. et al. (2018) ‘Direct binding of CEP85 to STIL ensures robust PLK4 activation and efficient centriole assembly’, Nature Communications, 9(1). doi: 10.1038/s41467-018-04122-x.

Liu, Y. et al. (no date) ‘Direct binding of CEP85 to STIL ensures robust PLK4 activation and efficient centriole assembly’. doi: 10.1038/s41467-018-04122-x.

Livneh, I. et al. (2017) ‘Monoubiquitination joins polyubiquitination as an esteemed proteasomal targeting signal’, BioEssays. John Wiley & Sons, Ltd, 39(6), p. 1700027. doi: 10.1002/bies.201700027.

Lüdecke, A. et al. (2018) ‘Diffusive tail anchorage determines velocity and force produced by

165

kinesin-14 between crosslinked microtubules.’, Nature communications. Nature Publishing Group, 9(1), p. 2214. doi: 10.1038/s41467-018-04656-0.

Luo, Y. et al. (2019) ‘Atypical function of a centrosomal module in WNT signalling drives contextual cancer cell motility’, Nature Communications. Springer US, 10(1), p. 2356. doi: 10.1038/s41467-019-10241-w.

Mahoney, N. M. et al. (2006) ‘Making Microtubules and Mitotic Spindles in Cells without Functional Centrosomes’, Current Biology. Cell Press, 16(6), pp. 564–569. doi: 10.1016/J.CUB.2006.01.053.

Maresca, T. J. et al. (2005) ‘Xnf7 Contributes to Spindle Integrity through Its Microtubule- Bundling Activity’, Current Biology. Cell Press, 15(19), pp. 1755–1761. doi: 10.1016/J.CUB.2005.08.049.

Marshall, W. F. (2009) ‘Centriole evolution.’, Current opinion in cell biology. NIH Public Access, 21(1), pp. 14–9. doi: 10.1016/j.ceb.2009.01.008.

Marshall, W. F. (2016) ‘Cell Geometry: How Cells Count and Measure Size’. doi: 10.1146/annurev-biophys-062215-010905.

Martell, J. D. et al. (2012) ‘Engineered ascorbate peroxidase as a genetically encoded reporter for electron microscopy.’, Nature biotechnology. NIH Public Access, 30(11), pp. 1143–8. doi: 10.1038/nbt.2375.

Marthiens, V. et al. (2013) ‘Centrosome amplification causes microcephaly’, Nature Cell Biology. Nature Publishing Group, 15(7), pp. 731–740. doi: 10.1038/ncb2746.

Martin, C.-A. et al. (2016) ‘Mutations in genes encoding condensin complex proteins cause microcephaly through decatenation failure at mitosis.’, Genes & development. Cold Spring Harbor Laboratory Press, 30(19), pp. 2158–2172. doi: 10.1101/gad.286351.116.

Mennella, V. et al. (2012) ‘Subdiffraction-resolution fluorescence microscopy reveals a domain of the centrosome critical for pericentriolar material organization’, Nature Cell Biology, 14(11), pp. 1159–1168. doi: 10.1038/ncb2597. 166

Meroni, G. and Diez-Roux, G. (2005) ‘TRIM/RBCC, a novel class of “single protein RING finger” E3 ubiquitin ligases’, BioEssays, 27, pp. 1147–1157. doi: 10.1002/bies.20304.

Mevissen, T. E. T. and Komander, D. (2017) ‘Mechanisms of Deubiquitinase Specificity and Regulation’, Annual Review of Biochemistry. Annual Reviews , 86(1), pp. 159–192. doi: 10.1146/annurev-biochem-061516-044916.

Miguela, V. and Lujambio, A. (2017) ‘Uncovering the role of USP54 in cancer.’, Oncotarget. Impact Journals, LLC, 8(7), pp. 10765–10766. doi: 10.18632/oncotarget.14820.

Miki, H. et al. (2001) ‘All kinesin superfamily protein, KIF, genes in mouse and human.’, Proceedings of the National Academy of Sciences of the United States of America. National Academy of Sciences, 98(13), pp. 7004–11. doi: 10.1073/pnas.111145398.

Milunović-Jevtić, A. et al. (2016) ‘Centrosomal clustering contributes to chromosomal instability and cancer.’, Current opinion in biotechnology. NIH Public Access, 40, pp. 113–118. doi: 10.1016/j.copbio.2016.03.011.

Miyajima, N. et al. (2009) ‘TRIM36 interacts with the kinetochore protein CENP-H and delays cell cycle progression’, Biochemical and Biophysical Research Communications. Academic Press, 381(3), pp. 383–387. doi: 10.1016/J.BBRC.2009.02.059.

Moawia, A. et al. (2017) ‘Mutations of KIF14 cause primary microcephaly by impairing cytokinesis’, Annals of Neurology. John Wiley & Sons, Ltd, 82(4), pp. 562–577. doi: 10.1002/ana.25044.

Mojarad, B. A. et al. (2017) ‘CEP19 cooperates with FOP and CEP350 to drive early steps in the ciliogenesis programme.’, Open biology. The Royal Society, 7(6). doi: 10.1098/rsob.170114.

Moritz, M. et al. (1995) ‘Microtubule nucleation by γ-tubulin-containing rings in the centrosome’, Nature. Nature Publishing Group, 378(6557), pp. 638–640. doi: 10.1038/378638a0.

Morris, J. H. et al. (2014) ‘Affinity purification–mass spectrometry and network analysis to understand protein-protein interactions’, Nature Protocols. Nature Publishing Group, 9(11), pp. 2539–2554. doi: 10.1038/nprot.2014.164. 167

Morriswood, B. et al. (2013) ‘Novel bilobe components in Trypanosoma brucei identified using proximity-dependent biotinylation.’, Eukaryotic cell. American Society for Microbiology (ASM), 12(2), pp. 356–67. doi: 10.1128/EC.00326-12.

Moyer, T. C. et al. (2015) ‘Binding of STIL to Plk4 activates kinase activity to promote centriole assembly.’, The Journal of cell biology, 209(6), pp. 863–78. doi: 10.1083/jcb.201502088.

Nagai, T. et al. (2018) ‘Cullin-3-KCTD10-mediated CEP97 degradation promotes primary cilium formation.’, Journal of cell science. The Company of Biologists Ltd, 131(24), p. jcs219527. doi: 10.1242/jcs.219527.

Nano, M. and Basto, R. (2017) ‘Consequences of Centrosome Dysfunction During Brain Development’, in Advances in experimental medicine and biology, pp. 19–45. doi: 10.1007/978- 3-319-57127-0_2.

Napolitano, L. M. and Meroni, G. (2012) ‘TRIM family: Pleiotropy and diversification through homomultimer and heteromultimer formation’, IUBMB Life. doi: 10.1002/iub.580.

Neo, S. H. et al. (2015) ‘TRIM28 Is an E3 Ligase for ARF-Mediated NPM1/B23 SUMOylation That Represses Centrosome Amplification’. doi: 10.1128/MCB.01064-14.

Nesvizhskii, A. I. et al. (2003) ‘A statistical model for identifying proteins by tandem mass spectrometry’, Analytical Chemistry. doi: 10.1021/ac0341261.

Niforou, K. N. et al. (2008) ‘The proteome profile of the human osteosarcoma U2OS cell line’, Cancer Genomics and Proteomics. International Institute of Anticancer Research, 5(1), pp. 63– 77.

Nigg, E. A. (2006) ‘Origins and consequences of centrosome aberrations in human cancers’, International Journal of Cancer. doi: 10.1002/ijc.22245.

Nigg, E. A. and Holland, A. J. (2018) ‘Once and only once: mechanisms of centriole duplication and their deregulation in disease’, Nature Publishing Group. doi: 10.1038/nrm.2017.127.

Nigg, E. A. and Raff, J. W. (2009) ‘Centrioles, Centrosomes, and Cilia in Health and Disease’,

168

Cell, 139(4), pp. 663–678. doi: 10.1016/j.cell.2009.10.036.

Nijman, S. M. B. et al. (2005) ‘A Genomic and Functional Inventory of Deubiquitinating Enzymes’, Cell. Cell Press, 123(5), pp. 773–786. doi: 10.1016/J.CELL.2005.11.007.

Niwa, S. (2015) ‘Kinesin superfamily proteins and the regulation of microtubule dynamics in morphogenesis’, Anatomical Science International. Springer Japan, 90(1), pp. 1–6. doi: 10.1007/s12565-014-0259-5.

Nogales, E. et al. (1999) ‘High-resolution model of the microtubule.’, Cell. Elsevier, 96(1), pp. 79–88. doi: 10.1016/S0092-8674(00)80961-7.

Novak, Z. A. et al. (2014) ‘Asterless licenses daughter centrioles to duplicate for the first time in Drosophila embryos.’, Current biology : CB. Elsevier, 24(11), pp. 1276–82. doi: 10.1016/j.cub.2014.04.023.

Novak, Z. A. et al. (2016) ‘Cdk1 Phosphorylates Drosophila Sas-4 to Recruit Polo to Daughter Centrioles and Convert Them to Centrosomes.’, Developmental cell. Elsevier, 37(6), pp. 545–57. doi: 10.1016/j.devcel.2016.05.022.

O’Connor, H. F. et al. (2015) ‘Ubiquitin-Activated Interaction Traps (UBAITs) identify E3 ligase binding partners.’, EMBO reports. European Molecular Biology Organization, 16(12), pp. 1699–712. doi: 10.15252/embr.201540620.

O’Connor, H. F. et al. (2018) ‘Ubiquitin-Activated Interaction Traps (UBAITs): Tools for Capturing Protein-Protein Interactions’, in. Humana Press, New York, NY, pp. 85–100. doi: 10.1007/978-1-4939-8706-1_7.

Ohta, M. et al. (2014a) ‘Direct interaction of Plk4 with STIL ensures formation of a single procentriole per parental centriole.’, Nature communications. Nature Publishing Group, 5, p. 5267. doi: 10.1038/ncomms6267.

Ohta, M. et al. (2014b) ‘Direct interaction of Plk4 with STIL ensures formation of a single procentriole per parental centriole.’, Nature communications. Nature Publishing Group, 5, p. 5267. doi: 10.1038/ncomms6267. 169

Olson, D. J., Oh, D. and Houston, D. W. (2015) ‘The dynamics of plus end polarization and microtubule assembly during Xenopus cortical rotation’. doi: 10.1016/j.ydbio.2015.01.028.

Olzmann, J. A. et al. (2007) ‘Parkin-mediated K63-linked polyubiquitination targets misfolded DJ-1 to aggresomes via binding to HDAC6.’, The Journal of cell biology. Rockefeller University Press, 178(6), pp. 1025–38. doi: 10.1083/jcb.200611128.

Ott, T. et al. (2019) ‘The Frog Xenopus as a Model to Study Joubert Syndrome: The Case of a Human Patient With Compound Heterozygous Variants in PIBF1.’, Frontiers in physiology. Frontiers Media SA, 10, p. 134. doi: 10.3389/fphys.2019.00134.

Paintrand, M. et al. (1992) ‘Centrosome organization and centriole architecture: Their sensitivity to divalent cations’, Journal of Structural Biology. doi: 10.1016/1047-8477(92)90011-X.

Paquet, D. et al. (2016) ‘Efficient introduction of specific homozygous and heterozygous mutations using CRISPR/Cas9’, Nature. Nature Publishing Group, 533(7601), pp. 125–129. doi: 10.1038/nature17664.

Paweletz, N. (2001) ‘Walther Flemming: pioneer of mitosis research’, Nature Reviews Molecular Cell Biology. Nature Publishing Group, 2(1), pp. 72–75. doi: 10.1038/35048077.

Pearson, C. G. and Bloom, K. (2004) ‘Dynamic Microtubules Lead the Way for Spindle Positioning’, Nature Reviews Molecular Cell Biology. Nature Publishing Group, 5(6), pp. 481– 492. doi: 10.1038/nrm1402.

Pelletier, L. et al. (2006) ‘Centriole assembly in Caenorhabditis elegans’, Nature, 444(7119), pp. 619–623. doi: 10.1038/nature05318.

Pelletier, L. (2007) ‘Centrioles: duplicating precariously.’, Current biology : CB. Elsevier, 17(17), pp. R770-3. doi: 10.1016/j.cub.2007.07.025.

Pelletier, L. and Yamashita, Y. M. (2012) ‘Centrosome asymmetry and inheritance during animal development.’, Current opinion in cell biology. NIH Public Access, 24(4), pp. 541–6. doi: 10.1016/j.ceb.2012.05.005.

170

Perez, Y. et al. (2019) ‘Mutations in the microtubule-associated protein MAP11 (C7orf43) cause microcephaly in humans and zebrafish’, Brain. Narnia, 142(3), pp. 574–585. doi: 10.1093/brain/awz004.

Peters, J.-M. (2002) ‘The Anaphase-Promoting Complex: Proteolysis in Mitosis and Beyond’, Molecular Cell. Cell Press, 9(5), pp. 931–943. doi: 10.1016/S1097-2765(02)00540-3.

Pickart, C. M. and Fushman, D. (2004) ‘Polyubiquitin chains: polymeric protein signals’, Current Opinion in Chemical Biology. Elsevier Current Trends, 8(6), pp. 610–616. doi: 10.1016/J.CBPA.2004.09.009. de Poot, S. A. H., Tian, G. and Finley, D. (2017) ‘Meddling with Fate: The Proteasomal Deubiquitinating Enzymes.’, Journal of molecular biology. NIH Public Access, 429(22), pp. 3525–3545. doi: 10.1016/j.jmb.2017.09.015.

Primorac, I. and Musacchio, A. (2013) ‘Panta rhei: The APC/C at steady state’, Journal of Cell Biology. doi: 10.1083/jcb.201301130.

Priya, S. et al. (2016) ‘Bardet-Biedl syndrome: Genetics, molecular pathophysiology, and disease management.’, Indian journal of ophthalmology. Wolters Kluwer -- Medknow Publications, 64(9), pp. 620–627. doi: 10.4103/0301-4738.194328.

Prosser, S. L. et al. (2015) ‘Nek5 promotes centrosome integrity in interphase and loss of centrosome cohesion in mitosis.’, The Journal of cell biology. Rockefeller University Press, 209(3), pp. 339–48. doi: 10.1083/jcb.201412099.

Prosser, S. L. and Pelletier, L. (2017) ‘Mitotic spindle assembly in animal cells: A fine balancing act’, Nature Reviews Molecular Cell Biology. Nature Publishing Group, 18(3), pp. 187–201. doi: 10.1038/nrm.2016.162.

Prosser, S. L., Straatman, K. R. and Fry, A. M. (2009) ‘Molecular dissection of the centrosome overduplication pathway in S-phase-arrested cells.’, Molecular and cellular biology. American Society for Microbiology (ASM), 29(7), pp. 1760–73. doi: 10.1128/MCB.01124-08.

Puklowski, A. et al. (2011) ‘The SCF-FBXW5 E3-ubiquitin ligase is regulated by PLK4 and 171

targets HsSAS-6 to control centrosome duplication’, Nature Cell Biology. doi: 10.1038/ncb2282.

Quesada, V. et al. (2004) ‘Cloning and enzymatic analysis of 22 novel human ubiquitin-specific proteases’, Biochemical and Biophysical Research Communications. Academic Press, 314(1), pp. 54–62. doi: 10.1016/J.BBRC.2003.12.050.

Ran, F. A. et al. (2013) ‘Double Nicking by RNA-Guided CRISPR Cas9 for Enhanced Genome Editing Specificity’, Cell. Cell Press, 154(6), pp. 1380–1389. doi: 10.1016/J.CELL.2013.08.021.

Ran, F. A. et al. (2013) ‘Genome engineering using the CRISPR-Cas9 system’, Nature Protocols. Nature Publishing Group, 8(11), pp. 2281–2308. doi: 10.1038/nprot.2013.143.

Rauch, A. et al. (2008) ‘Mutations in the pericentrin (PCNT) gene cause primordial dwarfism.’, Science (New York, N.Y.). American Association for the Advancement of Science, 319(5864), pp. 816–9. doi: 10.1126/science.1151174.

Reilly, M. L. et al. (2019) ‘Loss-of-function mutations in KIF14 cause severe microcephaly and kidney development defects in humans and zebrafish’, Human Molecular Genetics. Narnia, 28(5), pp. 778–795. doi: 10.1093/hmg/ddy381.

Reyes-Turcu, F. E., Ventii, K. H. and Wilkinson, K. D. (2009) ‘Regulation and cellular roles of ubiquitin-specific deubiquitinating enzymes.’, Annual review of biochemistry. NIH Public Access, 78, pp. 363–97. doi: 10.1146/annurev.biochem.78.082307.091526.

Reymond, A. et al. (2001) ‘The tripartite motif family identifies cell compartments.’, The EMBO journal. European Molecular Biology Organization, 20(9), pp. 2140–51. doi: 10.1093/emboj/20.9.2140.

Rhee, H.-W. et al. (2013) ‘Proteomic mapping of mitochondria in living cells via spatially restricted enzymatic tagging.’, Science (New York, N.Y.). NIH Public Access, 339(6125), pp. 1328–1331. doi: 10.1126/science.1230593.

Rigden, D. J. et al. (2009) ‘Ab initio protein modelling reveals novel human MIT domains.’, FEBS letters, 583(5), pp. 872–8. doi: 10.1016/j.febslet.2009.02.012.

172

Roque, H. et al. (2012) ‘Drosophila Cep135/Bld10 maintains proper centriole structure but is dispensable for cartwheel formation’, Journal of Cell Science, 125(23), pp. 5881–5886. doi: 10.1242/jcs.113506.

Roux, K. J. et al. (2012) ‘A promiscuous biotin ligase fusion protein identifies proximal and interacting proteins in mammalian cells.’, The Journal of cell biology. The Rockefeller University Press, 196(6), pp. 801–10. doi: 10.1083/jcb.201112098.

Ruiz, F. et al. (1999) ‘Basal body duplication in Paramecium requires γ-tubulin’, Current Biology. Cell Press, 9(1), pp. 43–46. doi: 10.1016/S0960-9822(99)80045-1.

Sacristan, C. and Kops, G. J. P. L. (2014) ‘Joined at the hip: kinetochores, microtubules, and spindle assembly checkpoint signaling’, Trends in Cell Biology, 25(1). doi: 10.1016/j.tcb.2014.08.006.

Salisbury, J. L. et al. (2002) ‘Centrin-2 Is Required for Centriole Duplication in Mammalian Cells’, Current Biology. Cell Press, 12(15), pp. 1287–1292. doi: 10.1016/S0960-9822(02)01019- 9.

Sanchez, J. G. et al. (2014) ‘The tripartite motif coiled-coil is an elongated antiparallel hairpin dimer’, Proceedings of the National Academy of Sciences. doi: 10.1073/pnas.1318962111.

Schmidt, T. I. et al. (2009) ‘Control of Centriole Length by CPAP and CP110’, Current Biology. Cell Press, 19(12), pp. 1005–1011. doi: 10.1016/J.CUB.2009.05.016.

Schnerch, D. and Nigg, E. A. (2016) ‘Structural centrosome aberrations favor proliferation by abrogating microtubule-dependent tissue integrity of breast epithelial mammospheres.’, Oncogene. Nature Publishing Group, 35(21), pp. 2711–22. doi: 10.1038/onc.2015.332.

Schulman, B. A. and Harper, J. W. (2009) ‘Ubiquitin-like protein activation by E1 enzymes: the apex for downstream signalling pathways.’, Nature reviews. Molecular cell biology. NIH Public Access, 10(5), pp. 319–31. doi: 10.1038/nrm2673.

Seki, A. and Fang, G. (2007) ‘CKAP2 is a spindle-associated protein degraded by APC/C-Cdh1 during mitotic exit’, Journal of Biological Chemistry. doi: 10.1074/jbc.M701688200. 173

Serçin, Ö. et al. (2016) ‘Transient PLK4 overexpression accelerates tumorigenesis in p53- deficient epidermis’, Nature Cell Biology. Nature Publishing Group, 18(1), pp. 100–110. doi: 10.1038/ncb3270.

Shannon, P. et al. (2003) ‘Cytoscape: a software environment for integrated models of biomolecular interaction networks.’, Genome research. Cold Spring Harbor Laboratory Press, 13(11), pp. 2498–504. doi: 10.1101/gr.1239303.

Shih, S. C., Sloper-Mould, K. E. and Hicke, L. (2000) ‘Monoubiquitin carries a novel internalization signal that is appended to activated receptors.’, The EMBO journal. European Molecular Biology Organization, 19(2), pp. 187–98. doi: 10.1093/emboj/19.2.187.

Shiratsuchi, G. et al. (2015) ‘RBM14 prevents assembly of centriolar protein complexes and maintains mitotic spindle integrity.’, The EMBO journal. European Molecular Biology Organization, 34(1), pp. 97–114. doi: 10.15252/embj.201488979.

Short, K. M. and Cox, T. C. (2006) ‘Subclassification of the RBCC/TRIM Superfamily Reveals a Novel Motif Necessary for Microtubule Binding * □ S’. doi: 10.1074/jbc.M512755200.

Shou, W. et al. (1996) Finely Tuned Regulation of Cytoplasmic Retention of Xenopus Nuclear Factor 7 by Phosphorylation of Individual Threonine Residues, MOLECULAR AND CELLULAR BIOLOGY. Available at: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC231081/pdf/160990.pdf (Accessed: 12 February 2019).

Sigoillot, F. D. and King, R. W. (2011) ‘Vigilance and validation: Keys to success in RNAi screening.’, ACS chemical biology. NIH Public Access, 6(1), pp. 47–60. doi: 10.1021/cb100358f.

Siller, K. H. and Doe, C. Q. (2009) ‘Spindle orientation during asymmetric cell division’, Nature Cell Biology. Nature Publishing Group, 11(4), pp. 365–374. doi: 10.1038/ncb0409-365.

Singh, N. et al. (2017) ‘A homozygous mutation in TRIM36 causes autosomal recessive anencephaly in an Indian family’, Human Molecular Genetics. Oxford University Press, 26(6), p. ddx020. doi: 10.1093/hmg/ddx020.

174

Sivakumar, S. and Gorbsky, G. J. (2017) ‘Phosphatase-regulated recruitment of the spindle- and kinetochore-associated (Ska) complex to kinetochores.’, Biology open. Company of Biologists, 6(11), pp. 1672–1679. doi: 10.1242/bio.026930.

Skarra, D. V et al. (2011) ‘Label-free quantitative proteomics and SAINT analysis enable interactome mapping for the human Ser/Thr protein phosphatase 5.’, Proteomics. NIH Public Access, 11(8), pp. 1508–16. doi: 10.1002/pmic.201000770.

Slack, F. J. et al. (2000) ‘The lin-41 RBCC gene acts in the C. elegans heterochronic pathway between the let-7 regulatory RNA and the LIN-29 transcription factor.’, Molecular cell. Elsevier, 5(4), pp. 659–69. doi: 10.1016/S1097-2765(00)80245-2.

Slangy, A. et al. (1995) ‘Phosphorylation by p34cdc2 regulates spindle association of human Eg5, a kinesin-related motor essential for bipolar spindle formation in vivo’, Cell. Cell Press, 83(7), pp. 1159–1169. doi: 10.1016/0092-8674(95)90142-6.

Sonnen, K. F. et al. (2012) ‘3D-structured illumination microscopy provides novel insight into architecture of human centrosomes.’, Biology open. Company of Biologists, 1(10), pp. 965–76. doi: 10.1242/bio.20122337.

Sonnen, K. F. et al. (2013) ‘Human Cep192 and Cep152 cooperate in Plk4 recruitment and centriole duplication’, Journal of Cell Science, 126(14), pp. 3223–3233. doi: 10.1242/jcs.129502.

Sowa, M. E. et al. (2009) ‘Defining the human deubiquitinating enzyme interaction landscape.’, Cell. NIH Public Access, 138(2), pp. 389–403. doi: 10.1016/j.cell.2009.04.042.

Spektor, A. et al. (2007) ‘Cep97 and CP110 Suppress a Cilia Assembly Program’, Cell. Cell Press, 130(4), pp. 678–690. doi: 10.1016/J.CELL.2007.06.027.

Stone, S. L. et al. (2005) ‘Functional Analysis of the RING-Type Ubiquitin Ligase Family of Arabidopsis 1[w]’, Plant Physiology, 137, pp. 13–30. doi: 10.1104/pp.104.052423.

Strnad, P. et al. (2007) ‘Regulated HsSAS-6 levels ensure formation of a single procentriole per centriole during the centrosome duplication cycle.’, Developmental cell. NIH Public Access, 175

13(2), pp. 203–13. doi: 10.1016/j.devcel.2007.07.004.

Tang, C.-J. C. et al. (2009) ‘CPAP is a cell-cycle regulated protein that controls centriole length’. doi: 10.1038/ncb1889.

Tang, C.-J. C. et al. (2011a) ‘The human microcephaly protein STIL interacts with CPAP and is required for procentriole formation’, The EMBO Journal, 30, pp. 4790–4804. doi: 10.1038/emboj.2011.378.

Tang, C.-J. C. et al. (2011b) ‘The human microcephaly protein STIL interacts with CPAP and is required for procentriole formation’, The EMBO Journal, 30, pp. 4790–4804. doi: 10.1038/emboj.2011.378.

Tang, Z. et al. (2013) ‘Autophagy promotes primary ciliogenesis by removing OFD1 from centriolar satellites.’, Nature. NIH Public Access, 502(7470), pp. 254–7. doi: 10.1038/nature12606.

Tanos, B. E. et al. (2013) ‘Centriole distal appendages promote membrane docking, leading to cilia initiation.’, Genes & development. Cold Spring Harbor Laboratory Press, 27(2), pp. 163–8. doi: 10.1101/gad.207043.112.

Tegha-Dunghu, J. et al. (2014) ‘MAP1S controls microtubule stability throughout the cell cycle in human cells.’, Journal of cell science. The Company of Biologists Ltd, 127(Pt 23), pp. 5007– 13. doi: 10.1242/jcs.136457.

Teixidó-Travesa, N. et al. (2016) ‘MZT1 regulates microtubule nucleation by linking γTuRC assembly to adapter-mediated targeting and activation’, Journal of Cell Science. doi: 10.1242/jcs.195321.

Tenno, T. et al. (2004) ‘Structural basis for distinct roles of Lys63- and Lys48-linked polyubiquitin chains’, Genes to Cells. John Wiley & Sons, Ltd (10.1111), 9(10), pp. 865–875. doi: 10.1111/j.1365-2443.2004.00780.x.

Tilney, L. G. et al. (1973) ‘Microtubules: evidence for 13 protofilaments.’, The Journal of cell biology. Rockefeller University Press, 59(2 Pt 1), pp. 267–75. doi: 10.1083/JCB.59.2.267. 176

Timonen, S. and Therman, E. (1950) The Changes in the Mitotic Mechanism of Human Cancer Cells. Available at: http://cancerres.aacrjournals.org/content/canres/10/7/431.full.pdf (Accessed: 6 April 2019).

Tsou, M. F. B. and Stearns, T. (2006) ‘Controlling centrosome number: Licenses and blocks’, Current Opinion in Cell Biology, 18(1), pp. 74–78. doi: 10.1016/j.ceb.2005.12.008.

Tsuchihara, K. et al. (2005) ‘Ckap2 regulates aneuploidy, cell cycling, and cell death in a p53- dependent manner’, Cancer Research. doi: 10.1158/0008-5472.CAN-04-4223.

Valente, E. M. et al. (2006) ‘Mutations in CEP290, which encodes a centrosomal protein, cause pleiotropic forms of Joubert syndrome’, Nature Genetics. Nature Publishing Group, 38(6), pp. 623–625. doi: 10.1038/ng1805.

Varmark, H. et al. (2007) ‘Asterless Is a Centriolar Protein Required for Centrosome Function and Embryo Development in Drosophila’, Current Biology. Cell Press, 17(20), pp. 1735–1745. doi: 10.1016/J.CUB.2007.09.031.

Verhey, K. J. and Gaertig, J. (2007) ‘The tubulin code’, Cell Cycle. doi: 10.4161/cc.6.17.4633.

Visweshwaran, S. P. et al. (2018) ‘The trimeric coiled‐coil HSBP1 protein promotes WASH complex assembly at centrosomes’, The EMBO Journal. doi: 10.15252/embj.201797706.

Vodermaier, H. C. (2004) ‘APC/C and SCF: Controlling Each Other and the Cell Cycle’, Current Biology. Cell Press, 14(18), pp. R787–R796. doi: 10.1016/J.CUB.2004.09.020.

Vulprecht, J. et al. (no date) ‘STIL is required for centriole duplication in human cells’. doi: 10.1242/jcs.104109.

Wang, L. et al. (2018) ‘A distal centriolar protein network controls organelle maturation and asymmetry.’, Nature communications. Nature Publishing Group, 9(1), p. 3938. doi: 10.1038/s41467-018-06286-y.

Wang, X. et al. (2009) ‘Asymmetric centrosome inheritance maintains neural progenitors in the neocortex’, Nature. Nature Publishing Group, 461(7266), pp. 947–955. doi:

177

10.1038/nature08435.

Winey, M. and O’Toole, E. (2014a) ‘Centriole structure.’, Philosophical transactions of the Royal Society of London. Series B, Biological sciences. The Royal Society, 369(1650). doi: 10.1098/rstb.2013.0457.

Winey, M. and O’Toole, E. (2014b) ‘Centriole structure.’, Philosophical transactions of the Royal Society of London. Series B, Biological sciences. The Royal Society, 369(1650). doi: 10.1098/rstb.2013.0457.

Wise, D. O., Krahe, R. and Oakley, B. R. (2000) ‘The γ-Tubulin Gene Family in Humans’, Genomics. Academic Press, 67(2), pp. 164–170. doi: 10.1006/GENO.2000.6247.

Woodruff, J. B. et al. (2017) ‘The Centrosome Is a Selective Condensate that Nucleates Microtubules by Concentrating Tubulin’, Cell, 169(6), p. 1066–1077.e10. doi: 10.1016/j.cell.2017.05.028.

Woodruff, J. B., Wueseke, O. and Hyman, A. A. (2014) ‘Pericentriolar material structure and dynamics.’, Philosophical transactions of the Royal Society of London. Series B, Biological sciences. The Royal Society, 369(1650). doi: 10.1098/rstb.2013.0459.

Wu, J. and Akhmanova, A. (2017) ‘Microtubule-Organizing Centers’, Annual Review of Cell and Developmental Biology. Annual Reviews , 33(1), pp. 51–75. doi: 10.1146/annurev-cellbio- 100616-060615.

Xiao, H. et al. (2016) ‘Mutations in epigenetic regulators are involved in acute lymphoblastic leukemia relapse following allogeneic hematopoietic stem cell transplantation’, Oncotarget. Impact Journals, LLC, 7(3), p. 2696. doi: 10.18632/ONCOTARGET.6259.

Xu, P. et al. (2009) ‘Quantitative Proteomics Reveals the Function of Unconventional Ubiquitin Chains in Proteasomal Degradation’, Cell. Cell Press, 137(1), pp. 133–145. doi: 10.1016/J.CELL.2009.01.041.

Yadav, S., Verma, P. J. and Panda, D. (2014) ‘C-terminal region of MAP7 domain containing protein 3 (MAP7D3) promotes microtubule polymerization by binding at the C-terminal tail of 178

tubulin’, PLoS ONE. doi: 10.1371/journal.pone.0099539.

Yang, Z. et al. (2008) ‘Extra centrosomes and/or chromosomes prolong mitosis in human cells.’, Nature cell biology. NIH Public Access, 10(6), pp. 748–51. doi: 10.1038/ncb1738.

Yau, R. and Rape, M. (2016) ‘The increasing complexity of the ubiquitin code’, Nature Cell Biology. Nature Publishing Group, 18(6), pp. 579–586. doi: 10.1038/ncb3358.

Ye, Y. and Rape, M. (2009) ‘Building ubiquitin chains: E2 enzymes at work.’, Nature reviews. Molecular cell biology. NIH Public Access, 10(11), pp. 755–64. doi: 10.1038/nrm2780.

Yoshigai, E. et al. (2009) ‘Trim36/Haprin plays a critical role in the arrangement of somites during Xenopus embryogenesis’, Biochemical and Biophysical Research Communications, 378, pp. 428–432. doi: 10.1016/j.bbrc.2008.11.069.

Yu, Z. et al. (2016) ‘Lys29-linkage of ASK1 by Skp1ÀCullin 1ÀFbxo21 ubiquitin ligase complex is required for antiviral innate response’, eLife, 5(e14087). doi: 10.7554/eLife.14087.001.

Yuba-Kubo, A. et al. (2005) ‘Gene knockout analysis of two γ-tubulin isoforms in mice’, Developmental Biology. Academic Press, 282(2), pp. 361–373. doi: 10.1016/J.YDBIO.2005.03.031.

Zhan, W. et al. (2015) ‘TRIM59 Promotes the Proliferation and Migration of Non-Small Cell Lung Cancer Cells by Upregulating Cell Cycle Related Proteins’, PLOS ONE. Edited by A. Ahmad, 10(11), p. e0142596. doi: 10.1371/journal.pone.0142596.

Zhang, L. et al. (2018) ‘MID1–PP2A complex functions as new insights in human lung adenocarcinoma’, Journal of Cancer Research and Clinical Oncology. Springer Berlin Heidelberg, 144(5), pp. 855–864. doi: 10.1007/s00432-018-2601-0.

Zhang, Y. and Galardy, P. J. (2016) ‘Ubiquitin, the centrosome, and chromosome segregation’, Chromosome Res, 24, pp. 77–91. doi: 10.1007/s10577-015-9511-7.

Zhu, F. et al. (2008) ‘The Mammalian SPD-2 Ortholog Cep192 Regulates Centrosome

179

Biogenesis’, Current Biology. Cell Press, 18(2), pp. 136–141. doi: 10.1016/J.CUB.2007.12.055.

Zitouni, S. et al. (2016) ‘CDK1 Prevents Unscheduled PLK4-STIL Complex Assembly in Centriole Biogenesis’, Current Biology. doi: 10.1016/j.cub.2016.03.055.

Zou, C. et al. (2005) ‘Centrobin: a novel daughter centriole-associated protein that is required for centriole duplication.’, The Journal of cell biology. The Rockefeller University Press, 171(3), pp. 437–45. doi: 10.1083/jcb.200506185.

Zou, H. et al. (1999) ‘Identification of a Vertebrate Sister-Chromatid Separation Inhibitor Involved in Transformation and Tumorigenesis’, Science. American Association for the Advancement of Science, 285(5426), pp. 418–422. doi: 10.1126/science.285.5426.418.

180

Appendix

List of proteins screened in the automated U-2 OS S-phase arrested centriole over- duplication screen

siRNA plate 1

1 2 3 4 5 6 7 8 9 10 11 12 CCDC85 FAM17 FAM19 RNF21 MPHOS FAM21 A NT S100A9 TTK USP54 FAM21A NT C 9B 9X 9 PH9 C FAM21 SIPA1L PLK1S YTHDF CCDC B PLK4 RCOR1 USP1 UBR4 GID8 NME7 PLK4 B 2 1 1 138 CEP35 CCDC1 CCDC C SAS6 WDR83 MIB1 NDEL1 NAV1 FLII SRP72 DIAPH3 SAS6 0 8 14 CEP1 ZNF51 ARPC CCDC1 ZNF59 CCDC2 SSX2I CEP1 D C3orf14 FHDC1 NUP54 KIAA1211 35 2 3 02A 8 2 P 35 CEP1 CCSE PMFBP PLEKH CEP1 E KRI1 ZC3H15 PCNT NEDD1 WDR62 GLMN SEC13 20 R2 1 A7 20 SEC24 ZSWIM C21orf C15orf3 FAM83 F STIL KCTD12 KIF23 LZTS2 PDIA4 ACTR1A STIL B 8 2 8 B CEP1 KIAA10 FAM83 DYNC1 ANKR CEP1 G ACTR1B ACVR1 SIK2 IPO9 PALLD CRKL 52 33 H H1 D26 52 CNTR COMM COMM KIAA12 WASH KIAA01 OR51F CNTR H CCDC53 IFT81 RBM33 NCKAP5L OB D3 D2 17 1 96 2 OB

siRNA plate 2

1 2 3 4 5 6 7 8 9 10 11 12 SEC16 CSNK1 KIAA12 CENP A NT MIF PHIP AP3M1 AP4E1 ANXA1 ISOC1 NT A A1 44 E C19orf5 CCDC1 PHF21 PHLDB CAMS B PLK4 FSD1 RAB35 SVIL KIF20B CEP85L PLK4 5 16 A 2 AP3 WRAP7 CCDC6 CCDC7 SLC25 MAP7 C SAS6 TXLNG CEP44 LUZP1 PIBF1 TXLNA SAS6 3 1 7 A31 D3 CEP1 CCDC6 SPATA PLEKH NUDC CEP1 D MAP7D2 TOX3 HSPA6 RPAP3 BAG2 NUDCD3 35 6 24 G1 D2 35 CEP1 NUP10 ZNF31 NUP16 YTHDF WDR5 CEP1 E TTC4 YLPM1 CPSF7 WDR33 TRIM26 20 7 8 0 2 5 20 STAMBP MB21D QRICH MAGO ANKH PIK3C F STIL UTP15 IBTK TRIM37 SDCCAG3 STIL L1 2 2 H D1 2B CEP1 TBC1D3 UBE2 C17orf6 CEP1 G CEP170B MTUS1 MKL1 GTSE1 SRRD TTC16 TRIM4 52 1 O 7 52 CNTR TMEM CCDC4 TMEM CNTR H EMC2 RRBP1 UBXN4 EMC1 EMC8 EMC3 EMC7 OB 177 7 57 OB

siRNA plate 3

1 2 3 4 5 6 7 8 9 10 11 12 CTTNB WDR6 FAM20 TBC1D KIAA03 A NT UXT SPCS2 RAB2A FRYL SPCS3 NT P2NL 0 8A 2B 55 PDXDC SEMA6 FERM CCDC1 ARHG SLC35 B PLK4 CACHD1 CDCA3 RABL3 USP6NL PLK4 1 A T2 15 AP1 F2 ANKLE WDR4 KIAA03 C8orf4 RAB11 C SAS6 GPR107 VPS51 PALD1 NBEA SLC5A3 SAS6 2 1 19L 7 FIP1 CEP1 FAM171 FAM91 NAP1L KIAA07 CEP1 D SEC31A WDR6 AZI1 OFD1 IFT20 PCM1 35 A1 A1 1 53 35 CEP1 CEP57 CCDC9 CCDC1 COMM C10orf8 RAPH CEP1 E ODF2L CEP85 CEP95 TPGS1 20 L1 2 12 D4 8 1 20 KIAA058 CCNB LRGU F STIL CEP72 IFT57 CEP55 PIDD FAT1 TTC38 MAPKBP1 STIL 6 2 K CEP1 HIST1H4 SERPI FAM18 CEP1 G DCDC2 RBM48 PREX2 USP35 PIAS1 INADL IFIT1B 52 G NA3 6B 52 CNTR ALKBH CCDC ANKR KIAA0 CNTR H IFIT1B ALS2CL TTC26 CHIA BICC1 TUBAL3 OB 3 41 D17 368 OB

siRNA plate 4

1 2 3 4 5 6 7 8 9 10 11 12

181

FAM12 OTOP STARD TPD52L LAMA PRR5- A NT EFHC2 DAW1 DIRC2 GJD2 NT 4B 1 9 2 5 ARHGAP8 CEP29 SSX2I B PLK4 STPG2 INCENP EHD1 GAS8 HAUS6 PCM1 NT CEP135 PLK4 0 P CENTRO ANK2S TRIM3 NAP1L C SAS6 ANK2S5 KIF2A CEP63 MAP9 MAP7 CPAP SAS6 BIN 4 6 1S1 CEP1 CEP12 CEP12 MAP7D KIA173 CSPP1 MTUS CEP1 D CSPP1 STIL SPICE SAS6 35 8 0 3S6 1 S4 1 35 CEP1 NAP1L1S KIA005 CEP15 PERICENT CEP1 E TRIM36 USP54 PLK4 RLUC PLK1 PLK1 20 17 86 2 RIN 20 PERICEN F STIL CEP135 STIL TRIN CEP1 CEP1 G 52 52 CNTR H CNTROB OB

siRNA set2 plate1 1 2 3 4 5 6 7 8 9 10 11 12 ABRAC ACTR1 A NT AAK1 ABCF1 ACSL3 ADSL AGO1 AGO2 AGO3 AGPAT6 NT L 0 AGTPB ALDH1 ANKH AP2M B PLK4 AGPAT9 AHI1 ALG9 ALMS1 ANO6 AP3D1 PLK4 P1 B1 D1 1 ARFG ARHG ARL6IP ATP13 ATP1B C NT APC APMAP APOL2 ASPH ATP2A2 NT AP2 AP21 5 A1 3 CEP1 ATP6AP ATP6V ATP6V BCAP CEP1 D ATP6AP1 AUP1 B9D1 B9D2 BAIAP2 BET1 35 2 0A1 1F 31 35 CEP1 C9orf7 CACYB CAMS CEP1 E BICD2 BRI3BP BTRC C2CD3 C2CD5 CAMLG CAMSAP2 20 8 P AP1 20 CC2D1 CC2D2 CCHC CCP11 F STIL CAPZA1 CAPZB CBX1 CCNB1 CCT8 CDC23 STIL A A R1 0 SASS CDK5R CDK5R CENP CEP10 CEP15 CEP16 CEP17 SASS G CDC27 CEP19 CEP192 6 AP2 AP3 H 4 2 4 0 6 CNTR CKAP2 CKAP CNTR H CEP89 CEP97 CETN1 CFL1 CHID1 CISD2 CKAP2 CKAP5 OB L 4 OB

siRNA set2 plate2 1 2 3 4 5 6 7 8 9 10 11 12 CNOT1 CNOT A NT CLASP1 CLASP2 CLCC1 CLCN7 CLGN CLIP1 CNOT1 CNOT7 NT 0 11 CSNK1 CTNN B PLK4 CNPY2 CNPY3 CNTRL BMI1 COPA COPS3 CRK CTSB PLK4 D D1 CYB5R DCTN C NT CTSV CYLD DAD1 DAPK3 DCLK1 DCTN1 DCTN4 DCTN6 NT 3 5 CEP1 CEP1 D DDRGK1 DDX24 DDX54 DEK DGKE DHRS7 DHX29 DHX35 DLG5 DNAJC16 35 35 CEP1 EIF2AK EIF4E CEP1 E DPM1 DSP DTNB DVL2 DVL3 EI24 EIF4E2 EIF5B 20 3 NIF1 20 F STIL EIF6 EMC4 EMC6 EMD EML4 EPB41 ERAP1 ERAP2 ERC1 ERLEC1 STIL SASS EXOC FAM18 FBXO2 FBXW1 FGFR1 SASS G ERLIN1 ERO1L EVC2 FBF1 FKBP10 6 4 4A 8 1 OP 6 CNTR GIGYF CNTR H FKBP2 FKBP8 FOPNL GARS GDI2 GFPT1 GLUD2 GNAI2 GNL3 OB 1 OB

siRNAset2 plate3

1 2 3 4 5 6 7 8 9 10 11 12 GPATC GRAM HADH HAUS A NT GOSR2 GPN1 GPX1 GSK3A HAUS1 HAUS8 NT H1 D1A A 4 HIST1H2 HMOX HSD17 HSPA1 IGFBP B PLK4 HM13 HMMR HYLS1 IFT74 IK PLK4 AE 2 B12 3 2 KDELC KDM1 C NT INPPL1 INVS IPMK IPO4 ITPR1 ITSN2 KANK2 KEAP1 NT 2 A CEP1 KIAA100 KIAA14 KIAA16 CEP1 D KIF14 KIF1A KIF1B KIF2C KIF3A KIF7 KNSTRN 35 9 30 71 35 CEP1 LEPRE LMBR CEP1 E KRAS LAMC1 LAS1L LATS1 LBR LCA5 LMAN1 LMF2 20 1 1L 20 MAP1L F STIL LMO7 LRRC49 LTV1 LZTS3 MAGT1 MANF MAP1S MAP2 MAP4 STIL C3B 182

SASS MAPR MAPR MCM1 MESD SASS G MAP7D1 MAPK9 MAPT MAST4 MED4 MIA3 6 E2 E3 0 C2 6 CNTR MMGT MPDU MRE11 MRPS CNTR H MINPP1 MKS1 MLF2 MLH1 MOGS MTDH OB 1 1 A 35 OB

siRNAset2plate4

1 2 3 4 5 6 7 8 9 10 11 12 NCAP NDUFA NDUFS NDUF A NT NARG2 NEK2 NBR1 NCOR1 NDC1 NDUFS8 NT D2 5 1 S3 NOMO NPHP B PLK4 NDUFV2 NEK8 NELFB NENF NIN NINL NPHP1 NSDHL PLK4 2 4 OSBPL C NT NUP214 NUP62 NUP88 OCRL ODF2 OS9 PARK7 PCNA PDE3B NT 8 CEP1 PGAM PGRM PGRM PICAL PLEK CEP1 D PDIA5 PDZD11 PHB PIGU PLOD2 35 5 C1 C2 M HA5 35 CEP1 POFUT PPP2R PRKAB PRPF1 PSMC CEP1 E POC1A POC1B PPIL4 PREB PTPN1 20 1 2A 1 9 1 20 RBM1 RCOR F STIL PYGL RABL5 RABL6 RCC2 REEP4 RER1 RFC4 RGPD5 STIL 4 3 SASS RNF21 RPGRI RPGRI SASS G RHBDD2 RIOK1 RPAP2 RPGR RPL27 RPN1 RPP30 6 3 P1 P1L 6 CNTR SCAM SEC11 SEC22 CNTR H RQCD1 RRM1 RRP1 SCD SCFD1 SCLT1 SEC23IP OB P2 A B OB

siRNAset2plate5

1 2 3 4 5 6 7 8 9 10 11 12 SECISB SERPI SH3D1 SIPA1L SIPA1L A NT SEC63 SEL1L SIL1 SKA2 SLC1A3 NT P2 NH1 9 1 3 SLC30A SLC33 SLC9A SORBS SPAG B PLK4 SLC30A5 SMG7 SMPD4 SOAT1 SPATA2 PLK4 6 A1 1 1 5 SQSTM C NT SPICE1 SRP9 SRPR SRPRB SRSF4 SSR3 STIM1 STT3B STX5 NT 1 CEP1 TBL1X TDRD CEP1 D TANC1 TBK1 TCHP TCTE1 TCTN1 TCTN2 TCTN3 TEX264 35 R1 3 35 CEP1 TMED1 TMEM TMEM1 TMEM CEP1 E TEX9 TFCP2 TJP1 TJP2 TMED5 TMEM216 20 0 17 99 209 20 TMEM23 TMEM2 TMEM3 TMEM TMEM6 TMEM TNRC F STIL TNIP1 TNPO3 TNRC6A STIL 1 37 3 43 7 9 18 SASS TOR1 TP53B TP53I1 TRIM5 SASS G TNRC6B TOP3B TOR1A TP53 TRIM13 TRIP6 6 AIP1 P2 1 6 6 CNTR TUBGC TUBGC TXNDC TXNDC UBE3 CNTR H TTC13 TWF1 UBA5 UBE2J1 UBIAD1 OB P2 P3 12 16 D OB

siRNAset2plate6

1 2 3 4 5 6 7 8 9 10 11 12 UNC93 A NT UFC1 UFL1 UFM1 UFSP1 UFSP2 UGGT2 UPF3B VAPB VEZT NT B1 WDR1 YTHDC YWHA YWHA B PLK4 VMA21 VRK2 WFS1 XRN1 YWHAE YWHAZ PLK4 8 2 B G ZMPSTE C NT ZMYM6NB NT 24 CEP1 CEP1 D 35 35 CEP1 CEP1 E 20 20 F STIL STIL

SASS SASS G 6 6 CNTR H CNTROB OB

183