This Thesis Has Been Submitted in Fulfilment of the Requirements for a Postgraduate Degree (E.G

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

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

Identification and characterisation of novel factors involved in the nonsense-mediated mRNA decay (NMD) pathway

Angela Casadio

Thesis presented for the degree of Doctor of Philosophy

The University of Edinburgh

2016

Declaration

I declare that this thesis is of my own composition. The work presented in this thesis is my own, unless otherwise stated, and has not been presented for any other degree or professional qualification.

Angela Casadio

Acknowledgements

First and foremost, I would like to thank my supervisor Javier for making this possible. I have really enjoyed working on a field I knew little about when I started my project, I’ve learnt a lot in the past four years and I will treasure the experience!

A huge thank you goes to Dasa for all the help and good advice she gave me during the course of my PhD. I will always be in your debt for being so patient with my last minute (or even last second) deadlines! Thanks, I really really appreciated it!

Thanks to Nele for being an inexhaustible source of knowledge and for bearing with me when I was asking thousands of questions. You’ve made my life a lot easier and you’ve taught me a lot!

To all past and present people in the lab, Elena, Fiona, Julie, Laurent, Magda, Marianne, Noemi, Ross, Sara, thanks for all the good times and the giggling! You made my time in the lab more enjoyable, and I will always treasure some of our conversations (how could I forget the floppy disc...).

A special thank you to my “lunch pals”, Giulia, Raffaele, Sara and Silvia, for making my breaks enjoyable and for all the laughs. Our lunches were a pleasant escape to the day-to-day routine in the lab.

Thanks to my tissue culture partner, Christine, who shared with me the pain of many hours and the complaints for the state of tissue cultures (and sorry if I complained too much). I was always relieved to come to TC and find a friendly face!

I would also like to thank all the people that helped me one way or another by providing advice and reagents. In particular I’d like to thank Prof. Oliver Mühlemann for the HeLa β-globin cells and Andy Finch for the gift of the siRNAs and for good advice. Thanks to Richard, my second supervisor, Ian and Greg for the usuful suggestions and feedback.

iii

Thanks to Silvia, Bob, Richard M and Ian J for their supervision during the 6-week rotations. With you I have learnt the basic techniques that helped me progress during my PhD project. Thanks for broadening my knowledge in various fields of biology.

I would like to thank all the people in the technical service, in particular Sean, Stephen and Stewart, for making my life easier and helping me when the robot got stuck again and again... Thanks to Matt from the microscopy service for helping to set up the microscope almost every single time! Also thanks to Jimi and Graeme for helping me with the processing and analyses of the mass spec and microarray data.

Grazie a Anne, Amy e Erica, per farmi sentire il calore di una famiglia anche lontano da casa.

Un grazie speciale alla mia famiglia per il costante supporto e l’affetto. Grazie per sopportare la lontananza, ma essere sempre presenti. Questo mio traguardo è stato possibile solo grazie a voi. Vi voglio bene!

Infine, last but not least, grazie di cuore al mio ragazzo Mattias, senza di te non sarebbe stato possibile! Grazie per aver sopportato i miei sbalzi di umore in quest’ultimo periodo e per avermi amata! Grazie grazie grazie! Ti amo.

Abstract

Nonsense mediated mRNA decay (NMD) is a surveillance mechanism that targets transcripts containing premature stop codons (PTCs) for degradation, and that also regulates up to 10% of the whole transcriptome.

During the course of my PhD I set out to identify novel NMD factors by performing a genome-wide RNA interference (RNAi) screen in a transgenic strain of Caenorhabditis elegans carrying an NMD reporter. I identified five novel proteins that are putative NMD factors in worms: NGP-1, NPP-20, AEX-6, PBS-2 and NOAH-2. Knock-down of these proteins led to severe developmental defects: worms were either arrested during various larval stages or died prematurely. The only exception was AEX-6, the knockdown of which led to a milder phenotype.

Homology analysis of the novel C. elegans NMD factors showed that these proteins are conserved in human, with the exception of NOAH-2, which only has a homologue in Drosophila melanogaster, NOMPA.

By performing an NMD assay in human cells, I demonstrated that GNL2 (NGP-1) and SEC13 (NPP-20) are functionally conserved NMD factors in human.

Analysis of the consequences of depletion of GNL2, SEC13, UPF1 or UPF2 on the transcriptome of HeLa cells revealed that these four proteins co-regulate a subset of endogenous NMD targets, whilst also independently regulating the expression of other sets of transcripts.

The findings presented in this thesis further our knowledge of the biology of NMD in both nematodes and humans. They demonstrate the existence of further regulators of this surveillance pathway, and add a layer of complexity to this fine-tuned biological process.

Lay summary

DNA contains the genetic information that cells need to grow and to properly function. To use this information, DNA is transcribed to messenger RNA (mRNA) that is then translated into proteins.

This process is crucial for the survival of the cells. Several mechanisms have therefore been developed by organisms during the course of evolution to check that the genetic information is correctly interpreted. One of these mechanisms is called nonsense-mediated mRNA decay (NMD). Its role is to eliminate mRNAs that would result in proteins shorter than normal ones, which may interfere with the correct functioning of cells. NMD has been associated with several diseases, such as β- thalassemia, muscular dystrophy and intellectual disorders. Functioning NMD is important for proper development of mammals, as the NMD pathway is responsible for regulating up to10% of all mRNAs.

This quality control process has been studied for over twenty years and we already know several of its players. However, not all steps of the NMD pathway are clear, so more effectors could still await discovery.

The aim of my PhD project was to find new factors involved in NMD and to try to understand their role in NMD.

I found five putative NMD effectors in the nematode worm Caenorhabditis elegans (NGP-1, NPP-20, AEX-6, PBS-2, and NOAH-2). Two of the newly identified factors are also functionally conserved in humans (GNL2/NGP-1, and SEC13/NPP-20). I have also shown that GNL2 and SEC13 collaborate with known NMD factors to regulate physiological mRNAs.

This work has increased our understanding of how NMD is regulated, and this deeper understanding will hopefully aid the scientific community in the study of how this process regulates normal development and disease.

Table of contents

Declaration ii

Acknowledgements iii

Abstract v

Lay summary vi

Table of contents vii

List of figures xii

List of tables xv

List of abbreviations xvi

Chapter 1: Introduction 1 1.1 When an mRNA has been translated correctly: mechanisms of normal translation termination 3 1.2 When the life an mRNA is over: cytoplasmic mRNA degradation 5 1.3 When the ribosomes stop elongating: no-go decay (NGD) 7 1.4 When messenger RNAs lack the stop codon: non-stop decay (NSD) 8 1.5 When ribosomes encounter premature stop codons: nonsense-mediated decay (NMD) 8 1.5.1 The NMD factors are conserved throughout evolution in eukaryotes 9 1.5.2 Transcripts bearing a premature stop codon: mechanism of NMD 12 1.5.3 Alternative mechanism of NMD in mammalian cells 21 1.5.4 Physiological importance of NMD 22 1.5.4.1 The role of NMD in development 22

vii

1.5.4.2 The role of NMD in disease 24 1.6 A relative of NMD: staufen-mediated mRNA decay (SMD) 26 1.6.1 Functions of staufen 28 1.7 Project outline 29

Chapter 2: Materials and Methods 31 2.1 Materials 31 2.1.1 Solutions 31 2.1.2 C. elegans strain 33 2.1.3 Cell lines 33 2.1.4 Primers 34 2.1.5 Expression vectors 35 2.1.6 List of materials 36 2.2 Methods 38 2.2.1 Maintenance of C. elegans 38 2.2.2 Synchronisation of population of worms 38 2.2.3 The screen 39 2.2.3.1 Description of the RNAi library 39 2.2.3.2 Description of screen strategy 39 2.2.3.3 Visualisation of GFP-expressing worms 40 2.2.3.4 Determination of clones identity 40 2.2.4 Imaging of worms 41 2.2.4.1 Preparation of slides 41 2.2.4.2 Acquisition of images 43 2.2.5 Maintenance of cell lines 43 2.2.6 siRNA treatment 43 2.2.7 Actinomycin D treatment 44 2.2.8 RNA extraction 44 2.2.8.1 Extraction of RNA from C. elegans 44 2.2.8.2 Extraction of RNA from human cells 44 2.2.9 cDNA synthesis 45 2.2.10 RT-qPCR 46 2.2.10.1 RT-qPCR from C. elegans samples 46

viii

2.2.10.2 RT-qPCR from human samples 46 2.2.11 Visualisation of rRNAs on FA gel 46 2.2.12 Microarray analysis 47 2.2.13 Cloning 47 2.2.14 Bacterial transformation 48 2.2.15 Transfection of expression vectors 48 2.2.15.1 Transfection of expression vectors for mass spectrometry analysis 48 2.2.16 Protein extraction 49 2.2.17 Immunoprecipitations 49 2.2.18 Protein gel electrophoresis 50 2.2.19 Western Blot 50 2.2.20 Mass spectrometry 51 2.2.21 Nuclear-cytoplasmic cells fractionation 51 2.2.22 Polysome profiling 52

Chapter 3: A genome-wide screen in C. elegans to uncover novel NMD factors 54 3.1 Introduction 54 3.1.1 C. elegans 55 3.2 Screen description 56 3.3 Screen results 58 3.3.1 C. elegans phenotypes 61 3.4 Screen validation 69 3.5 Protein conservation 70 3.6 Discussion 78

Chapter 4: Functional conservation of the novel NMD factors in humans 82 4.1 Introduction 82 4.2 Functional conservation of the novel C. elegans NMD factors in human cells 83

4.2.1 Depletion of GNL2 and SEC13 leads to upregulation the PTC-containing NMD reporter transcript 83 4.2.2 Depletion of GNL2 and SEC13 stabilises the NMD reporter transcript 87 4.2.3 Endogenous NMD targets are regulated by GNL2 and SEC13 89 4.2.4 GNL2 and SEC13 take part in the NMD regulatory feedback loop 91 4.3 Characterisation of GNL2 94 4.3.1 Is there a link between the ribosome biogenesis pathway and NMD? 99 4.4 Characterisation of the role of SEC13 in NMD 101 4.5 Discussion 106

Chapter 5: Proteomics and transcriptomics analyses of GNL2 and SEC13 111 5.1 Introduction 111 5.2 Mass spectrometry analysis of GNL2 and SEC13 binding partners 111 5.2.1 Efficient immunoprecipitation of T7-tagged GNL2 and HA-tagged SEC13 111 5.2.2 GNL2 mass spectrometry analysis 112 5.2.3 SEC13 mass spectrometry analysis 117 5.3 Microarray analysis of HeLa NS39 transcriptome 119 5.3.1 Sample preparation 119 5.3.2 Microarray data analysis 120 5.3.2.1 Global effects of NMD factor knockdown on the transcriptome 121 5.3.2.2 Transcripts co-regulated by the NMD factors 121 5.3.2.3 GNL2 and the SMD pathway 125 5.4 Discussion 126 5.4.1 Mass spectrometry analyses 126 5.4.2 Microarray analyses 128

Chapter 6: Discussion 130 6.1 Functional conservation of the novel C. elegans factors in human and fruitfly 132 6.2 GNL2 function 134 6.3 SEC13 function 139 6.4 Concluding remarks 143

References 145

Appendix: Published paper 184

List of figures

Figure Title Page number 1.1 Domain composition of human eRF1 and human eRF3 4 1.2 Domain composition of human PAN2 and human PAN3 6 1.3 Mechanism of NMD 13 1.4 Domain composition of human UPF1, UPF2, UPF3A and UPF3B 15 1.5 Domain composition of human SMG proteins 16 1.6 Degradation of the PTC-containing transcript via SMG6 19 1.7 Degradation of the PTC-containing transcript via the dimer SMG5-SMG7 or SMG5-PNRC2 20 2.1 Structure of the NMD reporter transgene 33 2.2 Preparation of slides for imaging of C. elegans 42 3.1 C. elegans life cycle 56 3.2 Screen strategy 57 3.3 Putative NMD factors identified in the screen 60 3.4 Knock-down of ngp-1 led to worms arrested at L1-L2 stage 62 3.5 Knock-down of npp-20 led to fragile worms 63 3.6 Knock-down of aex-6 led to worms with an egg laying defect 64 3.7 Knock-down of pbs-2 led to pale worms arrested at L2- L3 stage 66 3.8 Knock-down of noah-2 led to sluggish worms arrested at L2-L3 stage 67 3.9 Knock-down of kin-1 led to worms prone to bursting 68 3.10 Expression level of NMD reporter mRNA following knock-down of putative NMD factors 69 3.11 Domain composition of the novel C. elegans NMD proteins and of their human homologues 71

xii

3.12 Protein conservation of GNL2 throughout evolution 73 3.13 Protein conservation of SEC13 throughout evolution 74 3.14 Protein conservation of RAB27A and RAB27B throughout evolution 75 3.15 Protein conservation of PSMB10 and PSMB7 throughout evolution 76 3.16 Protein conservation of NOAH-2 throughout evolution 77 4.1 Knock-down of GNL2 and SEC13 leads to specific up-regulation of the NMD reporter mRNA 85 4.2 Knock-down of PSMB7, alone or in combination with PSMB10, leads toun specific up-regulation of the β-globin mRNA 86 4.3 Depletion of GNL2 or SEC13 leads to a stabilisation of the NMD reporter mRNA 88 4.4 Knock-down of GNL2 and SEC13 leads to up-regulation of a subset of endogenous NMD substrates 90 4.5 Knock-down of a subset of NMD factors up-regulates the mRNA level of other NMD factors 92 4.6 Knock-down of GNL2 and SEC13 leads to an increase of the mRNA levels of a subset of NMD factors 93 4.7 Ribosome biogenesis pathway 94 4.8 Endogenous GNL2 is localised both in the nucleus and in the cytoplasm 95 4.9 Knock-down of GNL2 leads to a drastic reduction of mature 60S ribosomal subunits and of mature ribosomes 96 4.10 Depletion of GNL2 does not affect the incorporation of radio-labelled methionine 97 4.11 Depletion of GNL2 leads to accumulation of 32S ribosomal RNA 98 4.12 Knock-down of LSG1 leads to specific up-regulation of the NMD reporter mRNA 100 4.13 Nuclear pore complex 102

xiii

4.14 Formation of COPII coat vesicles 103 4.15 Knock-down of the NPC components SEH1L, or of the COPII coat vesicles components SEC31A and SEC31B does not lead to specific up-regulation of the NMD reporter mRNA 105 5.1 Immunoprecipitations of T7-GNL2 and HA-SEC13 112 5.2 Heat map of enriched proteins bound to T7-GNL2 and T7 empty vector identified by mass spectrometry analysis of T7 immunoprecipitations 113 5.3 Heat map of enriched proteins bound to HA-SEC13 and HA empty vector identified by mass spectrometry analysis of HA immunoprecipitations 117 5.4 Depletion of UPF1, UPF2, GNL2 and SEC13 120 5.5 UPF1, UPF2, GNL2 and SEC13 co-regulate a subset of transcripts 122 5.6 Heat maps displaying overlap of targets 124 5.7 SMD targets 125 6.1 Hypothetical role of GNL2 in the NMD pathway 137 6.2 Hypothetical role of SEC13 in the NMD pathway 141

xiv

List of tables

Table Title Page number 1.1 S. cerevisiae, C. elegans and H. sapiens NMD factors 11 3.1 List of positive clones identified in the screen 59 5.1 Ribosomal proteins co-immunoprecipitating with T7-GNL2 114 5.2 Proteins co-immunoprecipitating with T7-GNL2 with a putative or a known function in ribosomal biogenesis 115 5.3 Other proteins co-immunoprecipitating with T7-GNL2 116 5.4 Proteins co-immunoprecipitating with HA-SEC13 118 5.5 Summary of mis-regulated transcripts upon knock-down of UPF1, UPF2, GNL2 and SEC13 in HeLa NS39 121

List of abbreviations

ActD Actinomycin D

ADHD Attention deficit hyperactivity disorder

A. thaliana Arabidopsis thaliana

ATP Adenosine triphosphate

BMD Becker muscular dystrophy cDNA Copy DNA

C. elegans Caenorhabditis elegans

CTD C-terminal domain

DECID complex Decay inducing complex

DMD Duchenne muscular dystrophy

D. melanogaster Drosophila melanogaster

DNA Deoxyribonucleic acid dsRNA Double-stranded RNA

D. rerio Danio rerio

E. coli Escherichia coli

EJC Exon junction complex

ER Endoplasmic reticulum eRF Eukaryotic release factor

EV Empty vector

xvi

GFP Green fluorescent protein

GTP Guanosine-5'-triphosphate

HBB β-globin

H. sapiens Homo sapiens

IP Immunoprecipitation

IPTG Isopropyl-β-D-thiogalactopyranoside mRNA messenger RNA

MDCC N-[2-(1-maleimidyl)ethyl]-7-(diethylamino) coumarin-3-carboxamide m7G 7-methylguanosine

NGD No-go decay

NMD Nonsense-mediate decay

NPC Nuclear pore complex

NSCs Neural stem cells

NSD Non-stop decay

ORF Open reading frame

PBP Phosphate binding protein

PCR Polymerase chain reaction

PNS Peripheral nervous system

PTC Premature termination codon

RNA Ribonucleic acid

xvii

RNAi RNA interference rRNA Ribosomal RNA

RT-qPCR Real time – quantitative PCR

S. cerevisiae Saccharomyces cerevisiae siRNA Short interfering RNA

SMD Staufen-mediated decay

SMG Suppressor with morphogenetic effect on genitalia

SMG1C SMG1 complex

SNP Single nucleotide polymorphism

SURF complex SMG1-eRF1-eRF3-UPF1 complex

TAR Thrombocytopenia with absent radii tRNA Transfer RNA uORF Up-stream ORF

UPF Up-frameshift

UTR Untranslated region wt Wild-type

xviii

Chapter 1: Introduction

The genetic information is encoded by genes. In order for cells to access this information, the DNA has to be transcribed into pre-messenger RNA (pre-mRNA) by RNA polymerase II. The C-terminal domain (CTD) domain of RNA polymerase II plays a fundamental role in both transcription and in the subsequent processing of the transcript (McCracken et al., 1997a). The domain is constituted by 52 repeats of

7 residues: Tyrosine1 - Serine2 - Proline3 -Tryptophan4 - Serine5 - Proline6 - Serine7 (Allison et al., 1985, Corden et al., 1985). These residues undergo cycles of phosphorylation and dephosphorylation which are important for RNA polymerase II elongation and for the binding of protein complexes involved in pre-mRNA maturation (Payne et al., 1989, Mortillaro et al., 1996, Yuryev et al., 1996, Cho et al., 1997, McCracken et al., 1997a, McCracken et al., 1997b, Yue et al., 1997).

Three main processes ensure maturation of the pre-mRNA:

1) 5’ capping of the transcript; 2) Splicing of the pre-mRNA; 3) 3’ poly-adenylation of the transcript.

These three events co-operate to produce a mature mRNA (McCracken et al., 1997a).

The capping of the transcript occurs shortly after transcription is initiated. It consists of the addition of a 7-methylguanosine (m7G) cap to the 5’ end of the nascent transcript. The capping enzyme interacts with the CTD of RNA polymerase II, and it has been suggested that elongation happens only after the transcript has been capped. The cap is important to confer stability to the messenger RNA and to induce further maturation of the pre-mRNA (Cho et al., 1997, McCracken et al., 1997b, Yue et al., 1997, Lenasi et al., 2011).

The pre-mRNA is composed by alternating exons and introns. The spliceosome catalyses the process of splicing, which consists of the removal of introns from the pre-mRNA and in the ligation of two neighbouring exons (Berget et al., 1977, Chow et al., 1977, Brody & Abelson, 1985). As a consequence of splicing, exon junction

1 complexes (EJCs) are deposited 20-24 nucleotides up-stream of the boundary between two exons that have been joined together. EJCs are dynamic proteins complexes composed by four core components: eIF4A3, MAGOH, RBM8A/Y14 and CASC3/BTZ (also known as MLN51) (Le Hir et al., 2001a, Le Hir et al., 2001b, Degot et al., 2004, Shibuya et al., 2004, Tange et al., 2005). EJCs associate with additional proteins involved in mRNA export and mRNA surveillance, such as the nonsense mediated decay (NMD) components UPF3A and UPF3B, thus playing a role in the processing of the mRNA both in the nucleus and in the cytoplasm (Le Hir et al., 2000, Le Hir et al., 2001a). Furthermore, EJCs have been shown to enhance the translation of spliced mRNA (Wiegand et al., 2003, Nott et al., 2004). Like the addition of the 5’ cap, splicing occurs co-transcriptionally (Baurén & Wieslander, 1994). In addition to canonical splicing events, there are events that lead to inclusion or exclusion of exons; these events are referred to as alternative splicing. Alternative splicing provides a way to increase complexity of the transcriptome: the majority of all pre-mRNAs can in fact produce multiple isoforms of mature mRNA, leading to the production of multiple protein isoforms (Pan et al., 2008, Wang et al., 2008). Alternative splicing can also be a mean of regulating gene expression. It has been estimated that approximately 30% of all alternative splicing events results in the introduction of premature stop codons (PTCs), making these transcripts NMD substrates (Lewis et al., 2003).

To achieve full maturation of mRNAs, their 3’ ends need to be processed. The transcripts are endonucleolytic cleaved and subsequently a poly(A) polymerase adds a poly(A) tail to the cleaved 3’ ends (Edmonds & Abrams, 1960, Edmonds et al., 1971). Correct poly-adenylation is required for efficient transcription termination (Logan et al., 1987). Furthermore, it has been shown that splicing of the last intron and poly-adenylation are interlinked, as each process is required for the other (Niwa & Berget, 1991, Kyburz et al., 2006).

All post transcriptional steps are tightly regulated and can influence each other. For example, incomplete capping of the transcript inhibits splicing of the introns from the

2 pre-mRNA, and recruits DXO, a protein that removes the defective cap and degrades the transcript by exonuclease activity (Jiao et al., 2013).

Defective splicing events can lead to the introduction of premature stop codons in the transcripts. Multiple quality control mechanisms ensure that aberrantly spliced transcripts do not result in the production of truncated proteins. Firstly, these transcripts are retained on the transcription site through a mechanism that requires the exosome subunit RRP6, and the two NMD factors UPF1 and SMG6 (Custódio et al., 1999, de Almeida et al., 2010, Eberle et al., 2010, de Turris et al., 2011). Unspliced (or defectively spliced) mRNA can then be degraded by the nuclear exosome and the 5’-to-3’ exonuclease XRN2 (West & Proudfoot, 2009, Davidson et al., 2012). In some instances, these aberrant mRNA are exported to the cytoplasm, where they are targeted by nonsense-mediated mRNA decay (Peltz et al., 1993).

Lack of a poly(A) tail can result in the transcript not being released from the transcription site (Custódio et al., 1999). The incompletely poly-adenylated transcripts are subject to degradation by the exosome or by XRN2 (West & Proudfoot, 2009, Davidson et al., 2012).

More quality controls mechanisms are in place in the cytoplasm to ensure that mRNAs with aberrant characteristics are not translated into proteins. Such aberrant characteristics could be the presence of PTCs, the lack of termination codons, or the presence of strong secondary structures in the mRNA.

1.1 When an mRNA has been translated correctly: mechanisms of normal translation termination

Three different codons, UAA, UAG and UGA, signal the end of the coding sequence of an mRNA (Beaudet & Caskey, 1971). Occupation of the A site of the ribosome by one of these codons leads to the recruitment of two release factors: eRF1 and eRF3, which catalyse the release of the protein from the ribosome in a GTP-dependent manner (Beaudet & Caskey, 1971, Zhouravleva et al., 1995). eRF3 is the GTPase, whereas eRF1 is the factor that recognises the stop codon and releases the peptidic

3 chain from the ribosome (Frolova et al., 1996, Song et al., 2000). A highly conserved motif, GGQ, in the domain M (Figure 1.1) of eRF1 is required for translation termination (Frolova et al., 1999). The domain N of eRF1 has been shown to be important for the binding to the ribosomes and for the recognition of the stop codon, and the domain C (Figure 1.1) for binding to eRF3 (Merkulova et al., 1999, Frolova et al., 2000, Song et al., 2000). The domain of eRF3 responsible for binding to eRF1 is the C-terminal domain 2/3 (Figure 1.1) (Merkulova et al., 1999, Frolova et al., 2000). eRF1 forms a complex with GTP-bound eRF3, leading to conformational changes in eRF1 which cause it to resemble a tRNA molecule. This new conformation allows interactions with the stop codon that occupies the A site of the ribosome (Cheng et al., 2009). The GTPase activity of eRF3 is promoted by the binding of the release factors complex to the ribosome, which in turns enhances eRF1-mediated protein release form the ribosome (Frolova et al., 1996, Merkulova et al., 1999, Salas-Marco & Bedwell, 2004).

Figure 1.1 Domain composition of human eRF1 and human eRF3. Human eRF1 is composed by three domains; domain N (in yellow), domain M (in orange) and domain C (in red). Human eRF3 is constituted by two domains: G domain (in pink) and 2/3 domain (in blue).

1.2 When the life of an mRNA is over: cytoplasmic mRNA degradation

Mature mRNA exported to the cytoplasm has a 5’ cap, which is bound by cap- binding proteins (CBPs), and has a 3’ poly(A) tail of approximately 200-250 adenosines, which is bound by several copies of PABP (poly(A)-binding protein). The presence of proteins at both ends of the transcript protects it from the activity of 5’-3’ and 3’-5’ exonucleases.

Degradation of a transcript generally requires two consecutive events: deadenylation followed by decapping of the mRNA (Yamashita et al., 2005).

Deadenylation is a two-step process which involves two distinct enzymatic complexes (Yamashita et al., 2005). The first step consists in a shortening of the poly(A) tail and it is carried out by the PAN2-PAN3 complex. No mRNA decay is observed at this stage. Two PAN3 proteins form a dimer that binds to RNA through three distinct mechanisms: 1. the N-terminal PAM2 motif (Figure 1.2) can interact with PABP; 2. the PK and C-terminal domains can aspecifically bind to RNA (Figure 1.2); 3. the N-terminal domain contains a zinc-finger motif (Figure 1.2) that specifically binds to stretches of adenosines (Uchida et al., 2004, Christie et al., 2013, Jonas et al., 2014, Wolf et al., 2014). The PAN3-homodimer interacts with the WD40 domain and the CS1 region of the nuclease PAN2 (Figure1.2), recruiting it to the substrate mRNA (Boeck et al., 1996, Jonas et al., 2014, Wolf et al., 2014).

Figure 1.2 Domain composition of human PAN2 and human PAN3. Human PAN2 is composed by three domains and an interlinking region (CS1); WD40 domain (in yellow), an inactive USP domain (in green) and a C-terminal nuclease domain (in red). Human PAN3 is constituted by two domains: a pseudo-kinase domain (in orange) and a C-terminal domain (in blue). PAN3 contains two motifs at its N-terminal: a zinc-finger motif (in red) and a PAM2 motif (in purple). USP: ubiquitin-specific protease; ZF: zinc-finger; PAM2: PABP-interacting motif 2; PK: pseudo-kinase.

The second step sees a further shortening of the tail and the degradation of the body of the mRNA and it involves the CCR4-NOT complex (Tucker et al., 2001). This complex is composed by several subunits: CCR4 (there are two paralogues in mammals, CCR4A/CNOT6 and CCR4B/CNOT6L), CNOT1, CNOT2, CNOT3, CNOT9, CNOT10, CNOT11 and CAF1 (there are two paralogues in mammals, CAF1A/CNOT7 and CAF1B/CNOT8/POP2) (Albert et al., 2000, Lau et al., 2009). CCR4 and CAF1 are the catalytic active subunits of the complex and they directly interact. CNOT1 functions as the main scaffolding protein of the complex, directly interacting with the majority of the other subunits (Petit et al., 2012, reviewed by Xu et al., 2014).

Deadenylation is followed by decapping, which consists in the removal of the 5’-cap (Muhlrad et al., 1994). This event is carried out by the DCP1-DCP2 complex. In yeast the complex is composed by Dcp1p and Dcp2p only, whereas in higher eukaryotes additional proteins, such as EDC4/HEDSL, are involved (Beelman et al., 1996, Dunckley & Parker, 1999, Fenger-Grøn et al., 2005). DCP1 and DCP2 interact through their N-terminal domains and this interaction is required for decapping. Binding of DCP1 to DCP2 leads to conformational changes that activate DCP2

6 catalytic activity, resulting in decapping of the complex (van Dijk et al., 2002, Chang et al., 2014).

The 5’-to-3’ exonuclease XRN1 is then recruited to the decapped mRNA and starts to degrade it (Muhlrad et al., 1994). XRN1 contains an N-terminal nuclease domain with a conserved motif, KQRR, which is responsible for substrate recognition (Chang et al., 2011, Jinek et al., 2011).

In some instances the deadenylated mRNA is not decapped and the exosome is recruited to the 3’ end of the transcript, which is subsequently degraded in the 3’-to- 5’ direction. The exosome is a complex constituted by nine core subunits: RRP4, RRP40, RRP41, RRP42, RRP43, RRP45, RRP46, MTR3 and CSL4. Six of the core subunits (RRP41, RRP42, RRP43, PPR45, RRP46 and MTR3) form a ring which accommodates the RNA substrate (Liu et al., 2006). The core subunits lack any catalytic activity, which is carried out by cytoplasmic RRP44 (Wang et al., 2007, Schneider et al., 2009). The exosome is not able to degrade the cap at the 5’ of the mRNA. The scavenger decapping enzyme DCPS acts down-stream of the exosome to release the cap from the mRNA fragment (Wang & Kiledjian, 2001).

1.3 When the ribosomes stop elongating: no-go decay (NGD)

No-go decay is a surveillance pathway that is activated when a translating ribosome stalls on a coding sequence, such as when it encounters a stem loop secondary structure (Doma & Parker, 2006). The Dom34p:Hbs1p complex is recruited to the stalled ribosomes, where it promotes the dissociation of the two ribosomal subunits and the release of peptidyl-tRNA (Chen et al., 2010, Shoemaker et al., 2010). In mammals this dissociation requires an additional protein, ABCE1 (Pisareva et al., 2011). The transcript undergoes an endonucleolytic cleavage up-stream of the stalled ribosome, and the two fragments are subsequently digested by exonuclease activity, carried out by the exosome and by Xrn1p (Doma & Parker, 2006, Passos et al., 2009).

1.4 When messenger RNAs lack the stop codon: non-stop decay (NSD)

Non-stop decay is an evolutionally conserved translation-dependent surveillance mechanism that targets transcripts lacking a termination codon for rapid degradation by the exosome (Frischmeyer et al., 2002, van Hoof et al., 2002). In yeast, this degradation is mediated by the helicase complex ski2p-ski3p-ski8p and by the cofactor ski7p (van Hoof et al., 2002). It has been proposed that translation of the non-stop mRNA is repressed by the ribosome stalling at the 3’ end of the transcript, both in yeast and mammalian cells (Inada & Aiba, 2005, Akimitsu et al., 2007). Ito-Harashima et al. (2007) suggested that the lack of a termination codon could lead to the translation of the poly(A) tail and that the nascent protein is rapidly degraded by the proteasome. Ltn1p is the E3 ubiquitin-ligase that mediates the recognition of the non-stop peptide chain by the proteasome (Wilson et al., 2007, Bengtson & Joazeiro, 2010). The synergic efforts of the exosome and the proteasome ensure that little or no faulty protein is produced from non-stop mRNAs.

1.5 When ribosomes encounter premature stop codons: nonsense-mediated decay (NMD) Nonsense-mediated mRNA decay is a surveillance pathway which targets transcripts containing premature termination codons (PTCs) for degradation. It has also been shown to regulate the expression levels of a subset of endogenous transcripts. Microarray analyses of the transcriptome of yeast strains null for specific NMD factors highlighted that PTC-containing transcripts are not the only NMD substrates: a subset of endogenous mRNAs is also sensitive to NMD regulation (Mulhrad & Parker, 1999, He et al., 2003). Similar results were obtained when the expression of NMD factors was knocked down in human cell lines (Mendell et al., 2004, Wittmann et al., 2006, Yepiskoposyan et al., 2011). It is not clear what features mark these physiological transcripts targets for the NMD pathway, but some of these mRNAs have common characteristics, such as long 3’ UTRs (Mulhrad & Parker, 1999, Yepiskoposyan et al., 2011), alternative up-stream ORFs in the 5’ UTR, termed uORFs, (He et al., 2003, Mendell et al., 2004,

Wittmann et al., 2006, Yepiskoposyan et al., 2011) and the presence of introns in the 3’ UTR (Mendell et al., 2004, Wittmann et al., 2006, Yepiskoposyan et al., 2011). Interestingly, not all the mRNAs with uORFs are substrates for NMD (reviewed by Rehwinkel et al., 2006). It has been proposed that some yeast uORF-containing transcripts harbour elements that stabilise their expression and allow them to escape NMD regulation (Ruiz-Echevarria & Peltz, 2000). In plants, the uORF length determines whether the transcript is going to be degraded by the NMD machinery (Nyikó et al., 2009). In mammals, the mechanism that allows some uORFs- containing transcripts to evade NMD regulation has not yet been characterised (Rehwinkel et al., 2006). As mentioned above, long 3’ UTRs can also trigger NMD. When the ribosome reaches the termination codon, release factors are recruited to allow the nascent peptide to be released and the ribosome to be recycled. Based on current knowledge, a termination codon is recognised as “normal” because of the interactions between the terminating ribosome complex and the proteins that are bound to the 3’ UTR and the poly(A) tail. Specifically, Singh et al. (2008) showed that if the interaction between the cytoplasmic poly(A) binding protein PABPC1 and the release factors eRF3 is impaired, eRF3 is able to interact with the core NMD factor UPF1 thus triggering the NMD response (Singh et al., 2008).

1.5.1 The NMD factors are conserved throughout evolution in eukaryotes

The first insight on the existence of the NMD pathway dates back to 1979 when Chang and Kan discovered a nonsense mutation in the human β-globin chain, which introduced a premature stop codon. Chang and Kan hypothesised that the β-globin transcripts containing the premature termination codon could be more susceptible to degradation (Chang & Kan, 1979). In the same year two other researchers, Losson and Lacroute, studied the effect of nonsense mutations on the level of messenger RNA. They observed that the position of the mutation affected the rate of degradation of the transcript (Losson & Lacroute, 1979).

The factors involved in NMD were only identified a decade later in genetic screens performed in Saccharomyces cerevisiae and Caenorhabditis elegans (Culbertson et al., 1980, Hodgkin et al., 1989, Leeds et al., 1992). The discovery of the conservation of some NMD factors (UPF1/SMG2, UPF2/SMG3, UPF3/SMG4) between yeast and nematodes suggested that they might also be present in other species. Many studies focused on the analysis of sequence homology, and showed that NMD effectors were indeed conserved in other eukaryotes, including humans (Perlick et al., 1996, Applequist et al., 1997, Sun et al., 1998, Lykke-Andersen et al., 2000, Serin et al., 2001, Denning et al., 2001, Chiu et al., 2003, Ohnishi et al., 2003, Gatfield et al., 2003, Arciga-Reyes et al., 2006). The complexity of the system started to become apparent, which resulted in a number of research groups attempting to uncover more factors and to understand the interactions between them, by means of proteomic and genetic analyses (Cali et al., 1999, Longman et al., 2007, Cho et al., 2009, Yamashita et al., 2009, Izumi et al., 2010, Brazão et al., 2012, Casadio et al., 2015). For a summary of the experimental procedures that led to the discovery of NMD effectors refer to Table 1.1.

Table 1.1 S. cerevisiae, C. elegans and H. sapiens NMD factors. The table indicates the experimental approaches that led to the identification of the NMD factors, the species in which they were identified and the relevant publications. The discovery of the NMD factors listed in the last two entries is described in detail in the Results chapters of this thesis.

1.5.2 Transcripts bearing a premature stop codon: mechanism of NMD

In mammalian cells, NMD is a translational-dependent mechanism which often relies on splicing (Kervestin & Jacobson, 2012). Premature stop codons that are approximately 50-55 nucleotides from down-stream exon junction complexes trigger the nonsense-mediated decay of the PTC-harbouring transcripts (Thermann et al., 1998, Zhang et al., 1998a, Zhang et al., 1998b).

During splicing exon junction complexes are deposited 20-24 nucleotides up-stream of the boundaries between two exons (Le Hir et al., 2000). These complexes comprise four core components: MAGOH, RBM8A, CASC3 and eIF4A3, which are exported to the cytoplasm with the mature mRNA (Le Hir et al., 2001a, Le Hir et al., 2001b, Degot et al., 2004, Shibuya et al., 2004, Tange et al., 2005). eIF4A3 is an RNA helicase that binds directly to the spliced mRNA and bridges CASC3 to the RMB8A-MAGOH heterodimer (Palacios et al., 2004, Shibuya et al., 2004). RNPS1 is an additional protein that associates with the EJC and also plays a role in NMD. This was shown by tethering RNPS1 to the 3’ UTR of a β-globin reporter. Binding of RNPS1 down-stream of a stop codon led to down-regulation of the β-globin mRNA (Lykke-Andersen et al., 2001). The NMD factors UPF2 and UPF3A/B are found associated with the EJC (Figure 1.3, panel a). UPF3A and UPF3B are paralogue proteins and they both localise predominantly in the nucleus, where they associate with the exon junction complex (Lykke-Andersen et al., 2000, Serin et al., 2001). Both paralogues interact with RBM8A, and this interaction is important in order to activate the NMD pathway (Kim et al., 2001, Gehring et al., 2003). Thanks to a nuclear export signal present in their sequence, UPF3A and UPF3B remain associated with the mRNA during the export to the cytoplasm. UPF3A is less active in the NMD pathway than its paralogue UPF3B (Kunz et al., 2006), and the levels of UPF3A are tightly regulated by the levels of UPF3B (Chan et al., 2009). Once in the cytoplasm, UPF3 recruits UPF2 (Lykke-Andersen et al., 2000, Serin et al., 2001).

Figure 1.3 Mechanism of NMD. (a) UPF3 and UPF2 are bound to the EJC. The SURF complex (SMG1-UPF1-eRF1-eRF3) assembles on the ribosome stalling at the PTC. (b) Interactions between the SURF complex and the EJC lead to the conformational changes in the two protein complexes and in the formation of the DECID complex. DHX34 is recruited to facilitate the transition from SURF to DECID complex. (c) As a result of the conformational changes UPF1 is phosphorylated. The release factors, SMG8, SMG9 and DHX34, are released and additional proteins (SMG6, SMG5-7) are recruited. The targeted transcript is rapidly degraded.

When the ribosome encounters a premature stop codon it stalls, as shown in toeprinting analysis of β-globin reporters (a wild-type and two PTC-containing transcripts) carried out in rabbit reticulocyte lysates (Peixeiro et al., 2012). The SURF complex (SMG1-UPF1-eRF1-eRF3) is recruited to the stalling ribosome (Figure 1.3, panel a). This protein complex is constituted by the two release factors eRF1 and eRF3, the core NMD factor UPF1, and by the SMG1C complex, SMG1- SMG8-SMG9 (Kashima et al., 2006). SMG8 and SMG9 are two phosphoproteins that interact with the N-terminus of SMG1 (Figure 1.5). SMG9 is required for the interaction of SMG8 with SMG1: a complex composed of SMG1 and SMG8 cannot be detected when SMG9 is depleted in human cells, whilst SMG9 and SMG1 interact even upon SMG8 knock-down (Arias-Palomo et al., 2011). Binding of SMG8 and SMG9 to SMG1 inhibits SMG1 kinase activity (Yamashita et al., 2009, Arias- Palomo et al., 2011). There is evidence that contrasts the idea that UPF1 is recruited to NMD substrates exclusively when the ribosome pauses at the PTC. Hogg and Goff (2010) performed affinity purification of tagged mRNA in HEK cells and found that UPF1 binds to the 3’ UTR of most transcripts. They showed that the binding is not sequence specific, and they proposed that UPF1 binding to 3’ UTR is directly proportional to the length of the UTR, i.e. the longer the 3’ UTR sequence is, the more UPF1 proteins bind to it (Hogg & Goff, 2010). The binding of UPF1 to mRNAs was further analysed by CLIP in mouse embryonic stem cells treated with cycloheximide, where translation was impaired. UPF1 was found to bind to coding sequences (Hurt et al., 2013). It would be interesting to know whether the PTC-independent binding of UPF1 to mRNA is a conserved feature in different organisms and in different cell types. If so, this would suggest that UPF1 is already localised on most mRNAs, poised to act upon interaction with other NMD components and with release factors. This does not exclude the possibility that additional UPF1 may be recruited specifically upon ribosome stalling at the PTC. Interaction of the SURF complex with the down-stream EJC-UPF3-UPF2 complex leads to conformational changes and the formation of a new protein complex, the DECID complex (decay inducing) (Kashima et al., 2006) (Figure 1.3, panel b).

Figure 1.4 Domain composition of human UPF1, UPF2, UPF3A and UPF3B. UPF1 is composed of three domains: CH domain (in red), helicase domain (in green), and the C-terminal SQ domain (in blue). Human UPF2 is constituted by three MIF4G1 domains (in yellow) and by a C-terminal domain (in orange). Both UPF3A and UPF3B contain a RRM-like domain (in light blue and dark blue, respectively) and a C-terminal EBM motif (in black). The regions required for interaction with other NMD factors, EJC components and degradation-inducing factors are highlighted underneath each protein. RRM: RNA recognition motif; EBM: EJC-binding motif.

Figure 1.5 Domain composition of human SMG proteins. SMG1 is composed of three domains: HEAT domain (in yellow), FAT domain (in blue), and the PIKK kinase domain (in orange). SMG5 is constituted by three domains: the N-terminal 14-3-3 domain (in pink) an α-helical domain (in green) and an inactive PIN domain (in dark purple). SMG6 contains two N-terminal EBM motifs (in black) and four domains: USR domain (in brown), 14-3-3 domain (in pink) an α-helical domain (in green) and an active PIN domain (in purple). SMG7 is composed of three domains: 14-3-3 domain (in pink) an α- helical domain (in green) and PC domain (in yellow). SMG8 contains a conserved region (CR, in grey). SMG9 contains an NTPase domain (in red). The regions required for interaction with other NMD factors, EJC components and degradation-inducing factors are highlighted underneath each protein. EBM: EJC-binding motif.

Two main events lead to DECID complex formation. It is unclear whether the two events happen concurrently or one happens before the other: 1) UPF1 harbours a regulatory domain, termed CH, which binds in cis to the helicase domain and inhibits the ATP-dependent helicase activity of UPF1 itself. An additional inhibition is performed by the N-terminal SQ domain of UPF1 which binds in cis to the helicase domain blocking its function (Fiorini et al., 2013). The N- terminal domain of UPF2 binds to the CH domain of UPF1, thus leading to activation of UPF1 helicase activity (Chamieh et al., 2008, Chakrabarti et al., 2011). DHX34 associates with the SURF complex, and facilitates the interaction of UPF2 with UPF1, thus promoting the remodelling of the SURF complex into the DECID complex (Hug & Cáceres, 2014).

2) SMG1 interacts with UPF2, and this interaction promotes SMG-1 kinase activity, leading to the phosphorylation of UPF1 at N- and C-terminal residues. At least four UPF1 amino acids are phosphorylated in vivo by SMG1: T28, S1078, S1096 and S1116 (Yamashita et al., 2001, Ohnishi, et al.,2003, Kashima et al., 2006, Okada- Katsuhata et al., 2012). It is not clear whether this interaction dislodges SMG8 and SMG9 from SMG1, thus activating the kinase activity of SMG1.

Until recently not much was known about the helicase activity of UPF1 and its relevance for NMD. Croquette and colleagues showed that UPF1 can unwind double-stranded RNA molecules and translocate along single strands of RNA. Furthermore, they showed that binding of UPF2 to the CH domain of UPF1 releases the CH inhibition, activating the helicase activity, and increases UPF1 processivity (Fiorini et al., 2015). During the translocation onto single-stranded nucleic acids, UPF1 is able to dislodge bound proteins (Fiorini et al., 2015). This observation is in

17 line with a previous study that showed that the helicase activity of UPF1 was required for recycling of the protein complex bound to NMD target transcripts (Franks et al., 2010).

Phosphorylation of UPF1 is a key event for the activation of the NMD response as it leads to a remodelling of the SURF complex (Figure 1.3) (Yamashita et al., 2001). Firstly, it promotes the dissociation of the release factor eRF3 from UPF1 (Kashima et al., 2006). Secondly, phosphorylation of the residues T28 and S1096 at the N- terminus and C-terminus of UPF1 creates a platform for the binding of the decay- inducing proteins SMG6 and the dimer SMG5-SMG7 (Okada-Katsuhata et al., 2012) (Figure 1.3, panel c). Binding of SGM5-7 results in the dissociation of eRF1 from UPF1 and the release of the ribosome from the PTC (Okada-Katsuhata et al., 2012). The dimer SMG5-7 interacts with PP2A, that dephosphorylates UPF1 (Ohnishi et al., 2003). SMG6 also interacts with UPF1 independently of the phosphorylated T28, via interactions with the helicase domain of UPF1 (Chakrabarti et al., 2014, Nicholson et al., 2014).

Figure 1.6 Degradation of the PTC-containing transcript via SMG6. SMG6 cleaves the targeted transcripts in proximity of the PTC. The two mRNA fragments are then digested by the exosome and XRN1.

SMG6 is the endonuclease that is responsible for the cleavage of the faulty transcript in the proximity of the PTC (Figure 1.6), and for this activity it requires SMG1 and UPF1 (Huntzinger et al., 2008, Eberle et al., 2009, Nicholson et al., 2014). SMG6 cleaves the faulty transcript in two positions: 3 nucleotides and ~ 20-40 nucleotides down-stream of the PTC (Boehm et al., 2014). After this endonucleolytic cleavage event, the disassembly of the protein complex from the 3’ fragment of the faulty transcript requires UPF1 ATPase-dependent helicase activity (Franks et al., 2010). The two fragments thus obtained are rapidly degraded by exonucleases: XRN1 degrades the 3’ fragment and the exosome degrades the 5’ fragment (Gatfield & Izaurralde, 2004, Eberle et al., 2009).

SMG5 is important for the recruitment of SMG7 to the DECID complex, but does not have any degradation activity (Jonas et al., 2013); SMG7, on the other hand, promotes degradation of the faulty transcripts by interacting with one of the catalytic subunits of the CCR4-NOT deadenylase complex, POP2 (also known as CNOT8) (Figure 1.7, panel a). This interaction with POP2 is required for SMG7-mediated decay (Loh et al., 2013). Deadenylation of the faulty transcript is followed by decapping (SMG7 interacts with the decapping component DCP2) and then degradation of the transcript via XRN1 and the exosome (Chen & Shyu, 2003, Lejeune et al., 2003, Unterholzner & Izaurralde, 2004, Loh et al., 2013). Alternatively, SMG5 can dimerise with PNRC2 instead of SMG7. This interaction has similar effects on the fate of the faulty transcript (Cho et al., 2013a). PNRC2 interacts with the decapping component DCP1A, thus stimulating the decapping of the transcript and its 5’-to-3’ degradation (Cho et al., 2009) (Figure 1.7, panel b).

SMG6- and SMG5-SMG7-mediated degradation of the target mRNA are likely to be complementary pathways, since depletion of both SMG5 and SMG6 has a stronger stabilisation effect on an NMD reporter transcript than either knock-down individually (Jonas et al., 2013).

Figure 1.7 Degradation of the PTC-containing transcript via the dimer SMG5-SMG7 or SMG5-PNRC2. (a) the dimer SMG5-SMG7 binds to phosphorylated UPF1. SMG7 then recruits POP2, a component of the deadenylase complex, and the decapping component DCP2. The faulty transcript is then subject to degradation by two exonucleases: XRN1 and the exosome. (b) the dimer SMG5-PNRC2 binds to phosphorylated UPF1. PNRC2 then recruits the decapping component DCP1A. The targeted transcript is decapped and degraded by the exonuclease XRN1.

1.5.3 Alternative mechanism of NMD in mammalian cells

The definition of exon junctions may not be as clear as initially thought. CLIP experiments, mapping the binding sites of the core EJC factor eIF4A3 to mRNAs in vivo, showed that eIF4A3 binds around 20-24 nucleotides from the exon boundaries (as described by Le Hir et al., 2000), but also up to 200 nucleotides from the boundaries. Furthermore, eIF4A3 is not present in all exons, showing differential exon-binding in the same transcript (Saulière et al., 2012). A separate study made use of a different technique (RIPIT) and also showed differential binding of the EJC to exons and transcripts (Singh et al., 2012). The mechanism of NMD described in the previous paragraph relies on interaction of the NMD machinery with the EJC, implying that differential binding of the EJC to exons and transcripts would have an impact on the NMD response. This suggests that NMD may recognise target mRNAs independently of the presence of a down-stream exon junction complex. Further evidence exists in support of this notion. For EJC-independent NMD, a critical event is the interaction of PABPC1, the cytoplasmic poly(A)-binding protein, with the ribosome at the termination codon: impaired interaction between PABPC1 and the ribosome triggers degradation of the transcript via NMD (Bühler et al., 2006, Eberle et al., 2008). It has been shown that interaction of PABPC1 with the release factor eRF3 inhibits NMD (Singh et al., 2008). Expression of an eRF3 mutant lacking the domain required for interaction with PABPC1 (but still able to bind to both eRF1 and UPF1) leads to degradation of a PTC-containing β-globin reporter that would normally escape NMD-regulation (Peixeiro et al., 2012). This study further highlights the importance of the interaction between the terminating ribosome, eRF1, eRF3 and PABPC1 to “mark” a stop codon as normal, thus protecting the transcript from NMD-mediated degradation.

EJC-independent NMD requires UPF1 and SMG1. The requirement for UPF2, UPF3, SMG6, and SMG5-7 is transcript-specific, but importantly at least one of the two degradation pathways (mediated by SMG6 or by SMG5-7) is required for EJC- independent NMD (Metze et al., 2013). There is evidence to suggest the existence of a UPF3A- and UPF3B-independent branch of NMD, which is still dependent on the EJC and other NMD factors, but does not require either UPF3 protein. Wilkinson and colleagues demonstrated that introduction of a PTC-containing TCRβ transgene in cells depleted of UPF3 did not lead to up-regulation of the transcript mRNA, whereas the mRNA of a PTC- containing HBB transgene was up-regulated in the same cells. They also showed over 300 endogenous transcripts appear to be regulated by this branch of NMD (Chan et al., 2007).

A UPF2-independent branch of NMD has also been reported. A UPF1 mutant that does not bind to UPF2 is still able to activate the NMD response when overexpressed in cells depleted of wild-type UPF1 (Ivanov et al., 2008). Furthermore, UPF3B and the EJC components RBM8A, MAGOH and eIF4A3 can still activate the NMD pathway when tethered to a reporter mRNA, even when UPF2 is knocked down (Gehring et al., 2005). It is not clear how the inhibitory intra-molecular binding of the CH domain to the helicase domain of UPF1 is released in the absence of UPF2. It is possible that additional proteins participate in the activation of UPF1, but if so they still await to be discovered.

1.5.4 Physiological importance of NMD

1.5.4.1 The role of NMD in development

Nonsense mediate decay is a non-essential pathway for development in yeast and C. elegans, but it is required in Drosophila, zebrafish and mammals (Culbertson et al., 1980, Hodgkin et al., 1989, Medghalchi et al., 2001, Metzstein & Krasnow, 2006, Wittkopp et al., 2009, Anastasaki et al., 2011). NMD factors are required for embryogenesis in mammals. Crossing of mice heterozygous for individual NMD factors (Upf1, Upf2, Smg1, and Smg6) results in no progeny when both alleles of the

22 same gene are knocked out (Medghalchi et al., 2001, Weischenfeldt et al., 2008, McIlwain et al., 2010, Li et al., 2015). Furthermore, knockout of either Upf1, Upf2 or Smg6 is lethal at early embryonic stages, before E7.5 (Medghalchi et al., 2001, Weischenfeldt et al., 2008, Li et al., 2015).

In keeping with this observation, analysis of Smg6 knockout embryonic stem cells shows that the NMD function of Smg6 is required to exit pluripotency (Li et al., 2015).

Knockout of Upf2 has been studied in two systems: in the bone marrow and in Sertoli cells. In both instances it leads to severe developmental defects. Sertoli cell- specific conditional knockout of Upf2 renders the Sertoli cells unable to differentiate and to support normal spermatogenesis (Bao et al., 2014). Conditional knockout of Upf2 in the entire haematopoietic system leads to severe anaemia resulting in death. Analysis of the bone marrow shows that the pool of haematopoietic progenitors is strongly reduced, and that the ability of B- and T-cells to respond to external stimuli was severely affected (Weischenfeldt et al., 2008).

Mouse neural stem cells (NSCs) down-regulate Upf1, Upf2, Upf3b, Smg1 and Smg6 when they undergo neural differentiation and maturation. A decrease of the NMD response is an essential event for proper neural differentiation, as NMD targets transcripts encoding neural differentiation factors for degradation (Lou et al., 2014). Although Upf3b mRNA and protein levels are down-regulated during differentiation of NSCs, the remaining levels are sufficient to maintain an active NMD pathway. Missense mutations in Upf3b, which result in impaired NMD, lead to the development of defective neurites which are less competent in forming complex branching (Alrahbeni et al., 2015). In vivo and in vitro analyses of mouse embryonic and adult brains revealed that Rbm8a is down-regulated when the neural progenitor cells initiate differentiation, and that over-expression of Rbm8a in neural progenitor cells favours proliferation over differentiation (Zou et al., 2015). This suggests that an active NMD pathway is important for normal neural development, and that titration of NMD activity may regulate cell fate decisions.

The study of NMD in mammalian development has lagged behind the study of the molecular mechanisms regulating the pathway. It is likely that the homeostasis of many systems and cell types is controlled by NMD, and that the mis-regulation of NMD may affect a multiplicity of choices with regards to cell survival, self-renewal and differentiation.

1.5.4.2 The role of NMD in disease

The NMD pathway and its factors are important for normal development of organisms and for the maintenance of the homeostasis of the transcriptome, but they also impact the outcome of many genetic diseases. As many as 25%-30% of all genetic diseases are caused by the arising of premature stop codons (Frischmeyer & Dietz, 1999). The ability of the NMD pathway to recognise and degrade the faulty transcripts plays a role in the determination of the disease phenotype.

In some cases NMD ameliorates the disease phenotype by targeting a transcript encoding a truncated protein which would have dominant negative effects. The best studied example is β-thalassemia, a type of disorder caused by mutations in the β-globin gene (HBB). Some mutations in the β-globin gene result in PTCs that are recognised by the NMD machinery. The faulty transcripts encoding these truncated versions of HBB are therefore degraded. An example of this is a 22 nucleotide duplication that introduces a new splice site which is preferentially used by the spliceosome. This leads to a shift in the coding sequence and results in a PTC at codon 37 (Svasti et al., 2008). The transcript encoding this aberrant form of β-globin is degraded by the NMD machinery, and patients that are heterozygous for this kind of mutations are generally asymptomatic, although still carrying the mutated β-globin allele. Other mutations in the β-globin gene (such as the mutation at codon 121) result in premature stop codons that are not recognised by the NMD machinery. The faulty transcripts are then translated into truncated forms of the β-globin protein which act in a dominant negative fashion. This has deleterious effects on the patients, such as

24 the formation of inclusion bodies, and severe anaemia requiring transfusions and in some cases bone marrow transplants (Thein et al., 1990).

In other cases NMD worsens the outcome of the disease by targeting for degradation a transcript encoding a truncated protein that is partially functional. The truncated form of the protein could retain some of the function of the wild-type protein, and would thus lead to a milder phenotype. For example, in patients affected by Duchenne muscular dystrophy (DMD), deletions of portions of the coding region of the dystrophin gene lead to a frameshift. This results in the introduction of a PTC that is recognised by the NMD machinery, and the transcript is targeted for degradation (Monaco et al., 1988, Kerr et al., 2001). In Becker muscular dystrophy (BMD), instead, deletions of portions of the coding region of the dystrophin gene do not alter the coding frame, but give rise to a truncated form of dystrophin. This truncated dystrophin is not targeted by the NMD machinery and partially retains the function of the wild-type protein. Due to expression of this hypomorphic dystrophin variant, BMD patients suffer from a milder version of dystrophy compared to DMD patients (Monaco et al., 1988, Kerr et al., 2001).

Mutations in genes encoding NMD factors can also lead to the development of pathologies. Missense mutations in the UPF3B gene have been associated with a variety of mental disorders, from schizophrenia to attention deficit hyperactivity disorder (ADHD) to mental retardation (Tarpey et al., 2007, Laumonnier et al., 2010, Addington et al., 2011, Szyszka et al., 2012, Jolly et al., 2013, Xu et al., 2013). Alrahbeni et al. (2015) showed that the NMD function of the UPF3B missense mutants is impaired by making use of a tethering assay. Tethering of wild-type UPF3B to an mRNA reporter led to the down-regulation of the transcript. Tethering of mutants UPF3B proteins to the same mRNA reporter led to a failure in the down- regulation of the transcript, indicating that these mutants are NMD-defective (Alrahbeni et al., 2015). Patients lacking a portion of chromosome I (q21.1, which contains the RBM8A gene), and carrying rare SNPs in the regulatory regions of the remaining RBM8A allele, are

25 affected by an autosomal recessive disorder called thrombocytopenia with absent radii (TAR). Patients affected by this pathology have low amount of platelets, due to a reduction of the precursor cells in the bone marrow, and have skeletal abnormalities that can vary from having no radii to lack of upper limbs (Albers et al., 2012).

In a recent study, Nguyen et al. (2013) analysed the link between NMD factors and intellectual disability. They found that deletions of UPF2, UPF3A, and RBM8A and duplications of UPF2, SMG6, RBM8A, EIF4A3, and RNPS1 significantly correlated with patients with disorders, suggesting that variations to the physiological copy number of these genes cause improper neuronal development. Patients heterozygous for UPF2 deletion show a variety of disorders, from hearing impairment to language disorder. Sequence analysis showed no mutations in the intact UPF2 allele, suggesting that loss of a single allele is sufficient to impair neural development. Analysis of the transcriptome showed that ~1,000 genes are mis-regulated in patient- derived lymphoblastoid cell lines. 38% of the transcripts mis-regulated in patients with UPF2 deletion were found to be mis-regulated in patients carrying UPF3B mutations (Nguyen et al., 2013). This suggests a co-operation of NMD factors to maintain proper neural development. This process seems to require just the right amount of NMD activity.

1.6 A relative of NMD: staufen-mediated mRNA decay (SMD)

Staufen-mediated mRNA decay is a translation-dependent surveillance mechanism that requires staufen1 (STAU1) and UPF1 (Kim et al., 2005). Tethering of STAU1 to the 3’ UTR of a reporter mRNA led to a specific down-regulation of the transcript; when siRNAs against UPF1 were used in conjunction with STAU1 tethering, the reporter mRNA was stabilised. This led to the conclusion that STAU1 is recruiting UPF1 to a subset of transcripts (Kim et al., 2005). In order for a transcript to be recognised as an SMD substrate, it has to harbour staufen-binding sites in its 3’ UTR. These sites have two main features: 1) they need to be down-stream of a termination codon; 2) they are composed by stretches of

26 double-stranded RNA (dsRNA). The dsRNA can derive by either of two events: base pairing between two Alu elements (one present in the 3’ UTR of the transcript and another one in a long non-coding RNA), or due to base pairing between two sequences in the same transcript (Gong & Maquat, 2011). The human paralogue of STAU1, staufen2 (STAU2), is also active in the SMD pathway (Park et al., 2013). Binding of either paralogue to the 3’ UTR of target mRNAs promotes degradation (Kim et al., 2005, Kim et al., 2007, Park et al., 2013). UPF1 is an essential factor for both SMD and NMD (Sun et al., 1998). Maquat and colleagues showed that the canonical NMD factor UPF2 competes for UPF1 binding with STAU1. Both STAU paralogues and UPF2 bind UPF1 in the same N-terminal region, hence leading Maquat and colleagues to propose that NMD and SMD are competitive pathways (Gong et al., 2009, Miki et al., 2011). It has been shown that two additional NMD factors, PNRC2 and SMG1, also play a role in SMD (Kim et al., 2005, Cho et al., 2012). SMG1 interacts with STAU1; SMG1 kinase activity is required for functional SMD (Cho et al., 2013b). UPF1 binds both STAU1 and PNRC2, bridging the two proteins together. Like in the NMD pathway, PNRC2 recruits DCP1A and XRN1 for degradation of the SMD targets (Cho et al., 2012). STAU1 and STAU2 have been shown to bind to each other forming either homo- or hetero-dimers; this dimerisation is important for efficient UPF1 binding (Gleghorn et al., 2013, Park et al., 2013). Overexpressed STAU1 and STAU2 interact with partially overlapping subsets of mRNAs (Furic et al. 2008); it is thus possible that each of the three staufen dimers may bind different targets.

Maquat and colleagues based the definition of SMD as a separate pathway from NMD on the observations that: 1- Depletion of STAU1 from cells did not affect the transcript levels of two well characterised NMD substrates. 2- When STAU1 was tethered to the reporter mRNA, depletion of either UPF2 or UPF3 could not restore the expression level of the transcript. 3- Depletion of STAU1, when either UPF2 or UPF3 were tethered to the same reporter transcript, had no effect on the transcript level (Kim et al., 2005).

It is not clear whether SMD is a pathway in contraposition to NMD, or whether it merely constitutes a sub-branch of NMD which requires a specific set of NMD factors only. This is an etymological issue which is likely to be debated in the field for years to come. What is clear, however, is that staufen does mediate mRNA decay.

1.6.1 Functions of staufen

Staufen is an RNA binding protein evolutionarily conserved from C. elegans to H. sapiens. It was initially identified in D. melanogaster, where it was shown to be important for transport of maternal mRNAs and translation activation (St. Johnston et al., 1991, Micklem et al., 2000). The two human paralogues, staufen1 and staufen2, differ in their expression pattern: STAU1 is ubiquitous whilst STAU2 is preferentially expressed in the brain (Marión et al., 1999, Duchaine et al., 2002).

Human staufen has three inter-linked roles: 1) mRNA transport in different tissues, such as polarised epithelial cells, neurons and oocytes (Villacé et al., 2004, Gautrey et al., 2005, De Santis et al., 2015). 2) Translation regulation/activation. This function of staufen is dependent on the correct localisation of its mRNA targets, in line with its previously listed function (Villacé et al., 2004, De Santis et al., 2015). An example of the function of staufen in translation activation is observed during retroviral infection, when staufen binds viral mRNAs and targets them to sites of active translation, thus helping in the formation of new viral particles (Hanke et al., 2013, Banerijee et al., 2014). 3) mRNA degradation: both STAU1 and STAU2 are involved in SMD and are implicated in the degradation of a subset of transcripts (Kim et al., 2005).

1.7 Project outline

Our understanding of the nonsense-mediated mRNA decay pathway has greatly improved since the first observations that transcripts containing premature termination codons were less stable than their wild-type counterparts (Chang & Kan, 1979, Losson & Lacroute, 1979). Several studies have investigated the mechanism through which nonsense-mediated mRNA decay operates, but there are still many open questions in the field, and it is likely that novel NMD factors still await to be discovered.

In order to address whether more proteins take part in the NMD pathway, a Post- Doctoral fellow in our laboratory, Dr. Dasa Longman, developed a transgenic strain of C. elegans (named PTCxi) with a GFP NMD reporter integrated in the genome. Using a commercially available library (created in the laboratory of Dr. Julie Ahringer), Dr. Longman silenced the expression of ~86% of the nematode genome and looked for effects on the NMD reporter. This approach proved to be successful, as she identified two novel factors: SMGL-1/NBAS and SMGL-2/DHX34. Both factors were found to be conserved in humans and zebrafish, where they retained their function in NMD (Longman et al., 2007, Anastasaki et al., 2011). The role of DHX34 in NMD has been extensively characterised in another publication from our laboratory (Hug & Cáceres, 2014). The C. elegans screen performed by Dr. Dasa Longman, and described in the previous paragraph, proved to be a valuable tool for identifying novel NMD factors (Longman et al., 2007).

Prior to the genome-wide screen, Longman et al. (2007) performed a genetic screen in which mutations were randomly induced in the genome of the PTCxi nematodes. As a result they obtained several non-viable worms that were expressing the GFP NMD reporter which were not characterised further (Longman et al., 2007). When performing the RNAi-based genome-wide screen, they only identified two novel NMD factors; it is therefore possible that the screen performed by Cáceres and colleagues did not discover all the uncharacterised nematode NMD factors (Longman et al., 2007).

In light of this, and following the development of a new RNAi library, targeting over 1700 genes that were not silenced by the previous library, we decided to follow the same approach as Longman et al. (2007), making use of the new library in an attempt to uncover novel NMD factors in C. elegans, and to subsequently test their role in NMD in humans.

These considerations led us to determine the aims of my PhD project:

1 – To attempt to discover novel NMD factors, by performing a genome-wide RNAi screen in C. elegans and subsequently assessing whether the role of any identified factor in NMD is conserved in humans.

2 – To characterise the role of any newly identified factor in the NMD pathway, attempting to uncover the steps at which it may be acting.

Chapter 2: Materials and Methods

2.1 Materials

2.1.1 Solutions

1x M9 Implemented 1x M9

3 g KH2PO4 To 50 ml of M9 add:

6 g Na2HPO4 100 µl cholesterol (5 mg/ml in ethanol) 5 g NaCl 10 µl ampicillin (250 mg/ml)

1 ml 1 M MgSO4 60 µl tetracycline (10 mg/ml)

dH2O up to 1 l 40 µl IPTG (250 mg/ml) Sterilise by autoclaving 20 µl fungizone (250 µl/ml)

NGM plates 1% Agarose gel 3 g NaCl 1 g of agarose 17 g Agar 100 ml 1X TBE 2.5 g BD BactoTM Peptone

975 ml dH2O Heat up for 2 min in the microwave. Autoclave 50 min the cool down for Bring the solution to RT and then add 15 min in a 55°C water bath. Then 10 µl of SYBR® Safe DNA gel stain. add:

1 ml 1 M CaCl2 Pour the gel. 1 ml cholesterol (5 mg/ml in ethanol)

1 ml 1 M MgSO4

25 ml 1 M KPO4 Dispense using a peristaltic pump.

Leave the plates to dry before use.

10X FA gel buffer 1X FA gel buffer 200 mM MOPS 100 ml 10X FA gel buffer

50 mM C2H3NaO2 20 ml 37% (12.3 M) formaldehyde

10 mM EDTA 880 ml dH2O pH to 7 with NaOH

FA gel (1% agarose) 5X RNA loading buffer 1 g agarose 16 µl saturated aqueous bromophenol 10 ml 10X FA gel buffer blue solution

Up to 100 ml with dH2O 80 µl 500 mM EDTA pH 8 After the agarose has melt, cool the 720 µl 37% (12.3 M) formaldehyde gel mixture to 65°C and add: 2 ml 100% glycerol 1.8 ml 37% (12.3 M) formaldehyde 3.084 ml formamide 1 µl of 10 mg/ml ethidium bromide 4 ml 10X FA gel buffer

Up to 10 ml with dH2O

IP buffer Blocking solution (for Western Blot) 10 mM TRIS pH 8 500 ml 1X PBS (autoclaved) 150 mM NaCl 5 g BSA 1 mM EDTA pH 8 250 µl Tween® 20 1% IGEPAL® CA-630 0.2% Sodium Deoxycholate dH2O

Filter the solution.

Before use, add: Protease Inhibitor 1 mM DTT

10X RSB buffer Implemented 1X RSB buffer 200 mM TRIS pH 7.4 1 ml 10X RSB buffer 1M KCl 20 µl 1M DTT

100 mM MgCl2 20 µl RNasin (used at 20 u/µl) dH2O 50 µl 100% IGEPAL® CA-630 10 µl 100 µg/ µl cycloheximide Filter the solution. 125 µl 2 u/µl Turbo DNase 1 tablet of complete mini EDTA-free

Up to 10 ml with dH2O

10% sucrose 45% sucrose 5 g sucrose 22.5 g sucrose 5 ml 10X RSB buffer 5 ml 10X RSB buffer

Up to 50 ml with dH2O Up to 50 ml with dH2O

60% sucrose 60 g sucrose

Up to 100 ml with dH2O

2.1.2 C. elegans strain The transgenic PTCxi C. elegans strain used in this study was described in Longman et al (2007). This nematode strain carries an NMD reporter (depicted in Figure 2.1) called PTCx and integrated in the genome. The reporter consists of a fusion between GFP and LacZ. It features a premature stop signal in the first exon of LacZ, which makes it a substrate of NMD. The levels of the transgenic transcript are very low and no GFP expression can be detected (Longman et al., 2007).

Figure 2.1 Structure of the NMD reporter transgene. The PTCx transgene consists in a fusion between GFP (in green) and LacZ (in blue). The premature stop codon is situated in the first exon of LacZ and it is shown in the schematic of the transgene. The star (*) indicates the stop codon at the end of LacZ.

2.1.3 Cell lines For this study three human cell lines were used: HeLa, a HeLa cell line stably expressing a wild-type β-globin transgene and a HeLa cell line stably expressing an NMD-sensitive β-globin transgene, which contains a PTC in codon 39 (Thermann et al., 1998). The two HeLa cell lines carrying the β-globin transgenes were a gift from Prof. Oliver Mühlemann (University of Berne).

2.1.4 Primers

Primers for qRT-PCR Probe Roche Gene Primer Sequence Number Assay Forward TGAAGGTGATGCAACATACG n.a. n.a. GFP Reverse CGGGCATGGCACTCTTGAAA Forward TCTGTCACATCACCGTACAA ama-1 n.a. n.a. Reverse CAGAGAGTATCCTGGACGAT Forward AGAGCTACGAGCTGCCTGAC ACTB 9 n.a. Reverse CGTGGATGCCACAGGACT Forward CTGTGAGCCCCGTTCCTAC POLR2J 1 n.a. Reverse GTCGGTGTCAGGGTGAGG Forward TGCAGGCTGCCTATCAGAA HBB 83 n.a. Reverse GCGAGCTTAGTGATACTTGTGG UPF1 n.a. n.a. n.a. 115402

UPF2 n.a. n.a. n.a. 137425

GNL2 n.a. n.a n.a. 144594

SEC13 n.a. n.a. n.a. 138390

RAB27A n.a. n.a. n.a. 114763

RAB27B n.a. n.a. n.a. 115465 Forward TGATAAGTTGCCTTATGTCACCA PSMB7 36 n.a. Reverse TCCTCCATGTCTGGCCTAAA Forward CATCGCCCCCAAAATCTAC PSMB10 19 n.a. Reverse CCATCCGTGTGGTCATCTC Forward ATCAGGGCGCTTGAAAGATA ARHGEF18 18 n.a. Reverse TTCTTGTAGCAGCAAAAGTACGTC Forward GACTGCGGTCTCCTAAAGGTC BMP2 49 n.a. Reverse GGAAGCAGCAACGCTAGAAG TMC7 n.a. n.a. n.a. 142251

SMG1 n.a. n.a. n.a 106402

SMG5 n.a. n.a n.a. 128927 Forward TGGATGCATCATTGTGGTAGA EFL1 12 n.a. Reverse TTGTCGCAGAACTGCCTGT

Forward TTGCAGGAACAGAGTTTGTAGC LSG1 40 n.a. Reverse TCTCTGGCTCTCCTCGAAAG Forward GGACATCCACTATTCGGTGAA SBDS 40 n.a. Reverse CTCTTTTAACTGCTTTATCACTTCCA Forward ATCCAGATGTGGGATCTTCG SEC31A 55 n.a. Reverse CTCCAAGCAATTGCCAAAAT Forward CTGCCATGAGCTTTCTACCC SEC31B 83 n.a. Reverse GCATGAAAAAGCCGATCTCT Forward TGGGAAGAAATAGTAGGAGAATCAA SEH1L 21 n.a. Reverse TCCACCAGAGTTGTCCTTTTAAC

Primers used for sequencing Name Primer Sequence Forward GTTTTCCCAGTCACGACGTT L4440 Reverse TGGATAACCGTATTACCGCC HA Forward CCTATGATGTGCCTGACTACG SV40pA Reverse GAAATTTGTGATGCTATTGC SEC13 Forward AGATCAACAACGCTCACACC Forward GGGCAAGGATCGTGATTTGG Forward CCCTCACATTGAAACTTACCTG GNL2 Forward CTTGGGAGAATGCTGAGGAC Forward GGCACAAAGGGAAGAGGAAC

Primers used for cloning Name Primer Sequence Forward CTGCTAGCATGGTGAAGCCCAAGTACAA GNL2 Reverse CGACTGTGATCATTACTGCTTTTGTCTGAATTTTTTGCG

2.1.5 Expression vectors The expression vectors used are: HA-SEC13 pRK5 construct, obtained from Addgene (Plasmid number 46332) (Bar- Peled et al., 2013). The HA tag is fused to the N-terminal of SEC13. The sequence of SEC13 was confirmed by sequencing. HA pRK5 empty plasmid was obtained by cutting out SEC13 from the Addgene original vector, using the restriction enzymes SalI and NotI (NEB) and by inserting a peptide with overhangs compatible with SalI and NotI restriction sites.

Sequence of the peptide:

5’ TCGACACCGGTGGGCCCTGATCAGATATCCAATTGGC 3’

3’ GTGGCCACCCGGGACTAGTCTATAGGTTAACCGCCGG 5' pCG T7 empty vector. The multiple cloning site contains consensus sequences for two restriction enzymes: XbaI and BamHI. pCG T7 GNL2 was obtained by cloning GNL2 into the empty vector. T7 tag is fused to the N-terminal of GNL2. The sequence of GNL2 was confirmed by sequencing. pCDNA 3xFLAG empty vector (Life Technologies).

2.1.6 List of materials

Product Name Manufacturer Catalogue No. Agarose Life Technologies 16500500 Amphotericin β from Streptomyces spores Sigma-Aldrich A2411 (fungizone) Ampicillin Sigma-Aldrich A9393 anti-GNL2 antibody (rabbit) Sigma-Aldrich HPA027163 anti-HA magnetic beads Pierce 88836 anti-mouse IgG HRP-conjugated (horse) NEB 7076 anti-rabbit IgG HRP-conjugated (goat) NEB 7074 LifeSpan anti-SEC13 antibody (rabbit) LS-C185388 Biosciences anti-β-tubulin (mouse) Sigma-Aldrich T4026 anti-T7 tag antibody agarose beads Novagen 69026 anti-XRN2 (rabbit) Bethyl A301-103A BamHI-HF NEB R3136S BD BactoTM Peptone BD Biosciences 211820 BenchMark™ Pre-Stained Protein Ladder Life Technologies 10748-010

Calcium chloride (CaCl2) Sigma-Aldrich 746495 ChemiGlow West Chemiluminescence ProteinSimple 60-12596-00 Substrate Kit Cholesterol Sigma-Aldrich C8667 Complete Mini EDTA-free Tablets Roche 05892791001 Cycloheximide Sigma-Aldrich C7698 DMEM Life Technologies 41966029 DTT Sigma-Aldrich D0632 Foetal calf serum (FCS) Sigma-Aldrich F-7524

ThermoFisher iBlot® 2 Mini transfer stacks IB23002 Scientific IGEPAL® CA-630 Sigma-Aldrich I8896 QIAquick Gel Extraction Kit Qiagen 28704 QIAprep Spin Miniprep Kit Qiagen 27104 Glycerol Sigma-Aldrich G5516 Library Efficiency® DH5α™ Competent Life Technologies 18263-012 Cells Lightcycler® 480 Probes Master Roche 04887301001 Lightcycler® 480 SYBR Green I Master Roche 04887352001 Lipofectamine® 2000 Life Technologies 11668019

Magnesium chloride (MgCl2) Sigma-Aldrich M8266 Magnesium sulphate (MgSO4) Sigma-Aldrich 208094 NE-PER™ Nuclear and cytoplasmic ThermoFisher 78833 extraction reagents Scientific NuPAGE® LDS Sample Buffer 4X Life Technologies NP0007 NuPAGE® Novex® 4-12% Bis-Tris Protein Life Technologies NP0321BOX Gels NuPAGE® Novex® 3-8% Tris-Acetate Life Technologies EA0375BOX Protein Gels NuPAGE® MOPS SDS Running Buffer 20X Life Technologies NP0001 NuPAGE® Sample Reducing Agent 10X Life Technologies NP0009 NuPAGE® Tris-Acetate SDS Running Life Technologies LA0041 Buffer 20X ON-TARGET plus GNL2 siRNA Fisher Scientific L-020392-01 ON-TARGET plus NBAS siRNA Fisher Scientific L-020986-00 ON-TARGET plus non-targeting siRNA Fisher Scientific D-001810-02 ON-TARGET plus PSMB7 siRNA Fisher Scientific L-006021-00 ON-TARGET plus PSMB10 siRNA Fisher Scientific L-006019-00 ON-TARGET plus RAB27A siRNA Fisher Scientific L-004667-00 ON-TARGET plus RAB27B siRNA Fisher Scientific L-004228-00 ON-TARGET plus SEC13 siRNA Fisher Scientific L-012351-00 ON-TARGET plus SEC31A siRNA Fisher Scientific L-014166-01 ON-TARGET plus SEC31B siRNA Fisher Scientific L-012991-00 ON-TARGET plus SEH1L siRNA Fisher Scientific L-013475-01 ON-TARGET plus UPF2 siRNA Fisher Scientific L-012993-01 OPTI-mem I Life Technologies 31985062 Pierce™ anti-HA magnetic beads Life Technologies 88836 Phusion® High-Fidelity DNA polymerase NEB M0530S Ponceau S Solution Sigma-Aldrich P7170 Potassium chloride (KCl) Sigma-Aldrich P9333 Potassium dihydrogen phosphate (KH2PO4) Sigma-Aldrich P0662 QIAshredder Qiagen 79654 RNasin® Ribonuclease Inhibitor Promega N2515 RNAzol® RT Sigma-Aldrich R4533

SalI-HF NEB R3138T Sepharose Protein G beads GE Healthcare 17-0756-01 SilverQuestTM silver staining kit NovexTM LC6070 DharmaFECT 1 Fisher Scientific T-2001

Sodium azide (NaN3) Sigma-Aldrich S2002 Sodium chloride (NaCl) Sigma-Aldrich 31434

Sodium hydrogen phosphate (Na2HPO4) Sigma-Aldrich 255793 Sodium hydroxide (NaOH) Sigma-Aldrich S8045 Sodium hypochlorite (NaOCl) Sigma-Aldrich 425044 ssRNA ladder NEB N0362S Sucrose Sigma-Aldrich 84097 SYBR® Safe DNA gel stain Life technologies S33102 Tetracycline Sigma-Aldrich T7660 TOPO Zero Blunt PCR Cloning Kit Life technologies 450245 TRIS base Sigma-Aldrich T1503 TRIzol Reagent Life technologies 15596-026 TURBO™ DNase Ambion AM2238 TrypLE™ express enzyme (1X), phenol red Life Technologies 12605010 Tween® 20 Sigma-Aldrich P9416 XbaI NEB R0145T

2.2 Methods

2.2.1 Maintenance of C. elegans PTCxi strain of C. elegans was cultured on NGM plates, kept in an incubator at 20°C in the dark.

2.2.2 Synchronisation of population of worms Adult worms were washed from the NGM plates with 2 ml of 1x M9 solution and transferred to a 50 ml falcon tube (Corning Life Sciences DL, Cat. No. 14-432-22). The volume was adjusted to 50 ml with 1x M9 solution. The tube was centrifuged at 2500 rpm for 10 min at 4°C). The 1x M9 solution was aspirated, leaving 4.5 ml in the tube. 500 µl of NaOCl solution (Sigma-Aldrich, Cat. No. 425044) and 500 µl of 5M NaOH were added to the 4.5 ml of the worms and 1x M9 solution mix. Tubes were shaken a couple of times to allow the chemicals to act on every worm and after

5 min, the M9-NaOCl-NaOH mix was diluted with dH2O up to 50 ml. Tubes were centrifuged for 10 min at 2500 rpm at 4°C and the supernatant completely removed.

Two more washes were performed with dH2O. The pelleted embryos were

38 resuspended in 20 ml of 1x M9 solution and incubated at 20°C O/N in a shaking incubator, set to 150 rpm.

2.2.3 The screen

2.2.3.1 Description of the RNAi library The ORF-RNAi library v.1.1 was developed in Dr. Marc Vidal’s laboratory (Reboul et al., 2003, Rual et al., 2004) and is available from Source Bioscience (www.lifesciences.sourcebioscience.com). The library consists of a strain of Escherichia coli bacteria (the HT115 (DE3) strain) expressing dsRNAs under the control of IPTG-inducible T7 RNA polymerase promoters. The library contains dsRNAs against 11,511 clones, and each one of them is expected to specifically target a single gene. It covers 10,953 out of 19,920 genes (roughly 55% of the nematode genome); of these 10,953 genes, 1736 genes were not previously targeted by the RNAi library developed in Dr. Julie Ahringer’s laboratory. 485 genes are targeted by at least two different RNAi clones (Rual et al., 2004). The library consists in glycerol stock of the bacterial clones; each clone is in a different well, for a total of 120 96-well plates. Each plate has an ID number which allow correct identification of the clones.

2.2.3.2 Description of screen strategy Liquid cultures of bacterial clones were started on day one: 500 ml of LB with 1 ml of ampicillin (250 mg/ml) were dispensed in 96-deep well plates and inoculated with 2 µl of glycerol stock of the bacterial clones (96-well format). The 96-deep well plates were incubated at 37°C overnight in a shaking incubator (set on 200 rpm). This procedure was performed with the aid of the BioMek FXP liquid handling robot (Beckman Coulter). Embryos were collected from adult PTCxi worms on day one and let to hatch O/N in a shaking incubator at 20°C at 150 rpm. On day 2, an aliquot of the bacterial cultures was used to make a new glycerol stock. 10 µl of 12.5 mg/ml IPTG were added to the bacterial cultures (these steps were

39 performed with the aid of the BioMek FXP liquid handling robot). The cultures were put back in the shaking incubator for 4 hours. The synchronised population of L1 worms were washed once in 50 ml of 1x M9 solution and resuspended in 50 ml of 1x M9 solution. The number of worms in 50 ml of 1x M9 was counted at the microscope (Leica) and the worms were diluted to the appropriate volume to obtain 25 L1 larvae in 50 µl of 1x M9 solution. The mixture of worms and 1x M9 was implemented as indicated in the Materials section. 50 µl of the implemented mixture of worms and 1x M9 were aliquoted in 96-well plates and, with the aid of the BioMek FXP liquid handling robot, 100 µl of the induced bacteria suspensions were added to the worms. The 96-well plates were put in the shaking incubator at 18°C at 150 rpm for 2 days.

2.2.3.3 Visualisation of GFP-expressing worms Search of GFP-positive worms in 96-well plate was carried out manually. GFP expression was visualised using a Leica MZFL III Stereo microscope.

2.2.3.4 Determination of clones identity To determine the identity of the silenced gene, positive clones were noted indicating the ID number of the plate and the coordinates of the well. To demonstrate that the clone was corresponding to the library annotations, bacterial clones expressing the dsRNAs were grown in LB with ampicillin at 37°C in a shaking incubator O/N. The plasmid DNA was purified the day after using the QIAprep Spin Miniprep Kit (Qiagen), following the manufacturer’s instructions. The plasmid DNA was then sequenced with 3130/3730 Genetic Analyser (Life Technology) using the primers L4440 Forward and Reverse. To determine which gene was silenced by the dsRNAs, the obtained sequences were blasted against C. elegans genome (www.wormbase.org/tools/blast_blat).

2.2.4 Imaging of worms

2.2.4.1 Preparation of slides A 5% agar solution in water was prepared and it was left to cool down until roughly at RT. Subsequently 0.1% sodium azide (NaN3) was added to the agar solution. A drop of agar was then placed onto the centre of the slide. Another slide was positioned on top of the drop of agar and it was pressed down gently avoiding the formation of bubbles. The agar drop was flattened to a circle about 0.4 mm thick. After the agar solidified, the slides were gently separated so that the agar pad would adhere on one of them. The slides were left on the bench top with the agar side up. Worms were collected in an Eppendorf tube and centrifuged at 1,200 rpm for 10 min. The supernatant was discarded and the pelleted worms were resuspended in 10 µl of 1x M9. A drop of the M9 + worms mixture was put on the agar pad. A cover slip was gently put over the worms and Vaseline was added around the edges of the coverslip to seal it.

Figure 2.2 Preparation of slides for imaging of C. elegans On a slide put a drop of a solution of 5% agar and 0.1% NaN3. Flatten the agar drop with the help of a second slide, to obtain a 0.4 mm-thick pad. When it has solidified, separate the two slide and add a drop of a mixture of C. elegans and 1x M9. Lay a cover slip over the worms and seal with Vaseline.

2.2.4.2 Acquisition of images Brightfield and GFP images of worms were acquired using a Zeiss Axioplan 2 microscope.

2.2.5 Maintenance of cell lines HeLa, HeLa β-globin wt and HeLa β-globin NS39 were kept in an incubator set at

37°C, 5% CO2 and 21% O2. Cells were cultured in DMEM + 10% FCS.

2.2.6 siRNA treatment HeLa β-globin wt and HeLa β-globin NS39 were transfected with siRNAs pools. On day 0, cells were plated in 6-well plates at 15% confluency. On day 1, cells were transfected as follow: 10 µl of one of the siRNA pools (5 µM) was mixed with 190 µl of OPTI-mem in an Eppendorf tube. 5 µl of DharmaFECT I were mixed with 195 µl of OPTI-mem in a different Eppendorf tube. After an incubation of 5 min the content of the two tubes was mixed together and incubated for 20 min. In the meantime, the old medium was removed from the cells and 1.5 ml of fresh medium were added in each well. The transfection mix was added to the cells at the end of the incubation. On day 2 the medium containing the transfection mix was removed from the cells and 2 ml of fresh medium were added in each well. On day 3 cells were split and plated again in a 6-well plate to reach the right confluency for a second round of transfection. On day 4 cells were transfected in the same way as it was done at day 1. On day 5, the medium containing the transfection mix was removed from the cells and fresh medium was added in each well; if the cells were reaching confluency they were propagated in 10 cm plates. Cells were collected on day 6 to extract RNA. Cells depleted of PSMB7 and PSMB10, alone or in combination, were collected on day 5. Depletion of UPF1 was achieved by using a combination of two individual siRNAs in equal molarity that were previously reported (Paillusson et al. 2005, Azzalin & Lingner 2006). Sequence of the two siRNAs: hUPF1-I GAGAATCGCCTACTTCACT hUPF1-II GATGCAGTTCCGCTCCATT

2.2.7 Actinomycin D treatment Cells were treated with siRNA pools against either GNL2 or SEC13, or mock depleted as described in the previous paragraph. On day 6 the cells were collected and freshly seeded in 6 6-well plates. Actinomycin D (2 µg/ml) was added in each well and cells were collected at six different time points: 0.5, 1, 2, 4, 6 and 8 hours after the addition of Actinomycin D.

2.2.8 RNA extraction

2.2.8.1 Extraction of RNA from C. elegans Collect worms in an Eppendorf tube and centrifuge them at 1,200 rpm for 10 min. Remove almost completely the supernatant. Add 300 µl of TRIzol and with a 10 ml syringe, fitted with a 1 ml needle, aspirate up and down for 10 times. Add 200 µl of TRIzol and let the tube stand at RT for 5 min. Add 100 µl of chloroform and vortex for 15 s and the incubate at RT for 15 min. Centrifuge at 12,000 g for 15 min at 4°C. Take the top phase (clear), being careful to avoid both the interphase (white) and the bottom phase (pink), and transfer it to a new Eppendorf tube. Resuspend it 1:1 with 70% ethanol and proceed using the RNeasy Mini Kit (Qiagen, Cat. No. 74104), following the manufacturer’s instructions from step 3 of the protocol (Part 1 of the Quick-start protocol). The RNA concentration and purity was assessed at the NanoDrop 8000 (Thermo Scientific).

2.2.8.2 Extraction of RNA from human cells Two different methods were used to extract RNA from cells. 1) Cells were washed once with PBS and then they were collected in an Eppendorf tube by adding 350 µl of RTL buffer in a well of a 6-well plate or by adding 600 µl of RTL buffer to a 10 cm plate. Cells resuspended with RTL buffer were collected in an Eppendorf tube. The mixture cells-RTL buffer was transferred on a shredding column (QIAshredder, Qiagen Cat. No. 79654). The sample was centrifuged at 13,200 rpm for 2 min at RT. The flow- through was resuspended in an equal volume of 70% ethanol. The sample was processed for the extraction of RNA using the RNeasy Mini Kit (Qiagen, Cat. No.

74104), following the manufacturer’s instructions from step 2 of the protocol (Part 1 of the Quick-start protocol). The RNA concentration and purity was assessed at the NanoDrop 8000 (Thermo Scientific). 2) Cells were washed once with ice-cold PBS. To cells growing in 6-well plates, 500 µl of RNAzol® RT were added in each wells. To cells growing in 10 cm plates, 1 ml of RNAzol® RT was added. The mixture of cells and RNAzol® RT was collected in an eppendorf and 200 µl of water (per 500 µl of RNAzol® RT) were added. The tubes were shaken vigorously for 15 s and subsequently incubated for 15 min at RT. The tubes were then centrifuged at 12,000 g for 15 min at 4ºC. The supernatant was transferred to a new tube and an equal volume of 100% isopropanol was added. The mixture was incubated for 10 min at RT and was then centrifuged for 10 min at 12,000 g at RT. The precipitated RNA formed a pellet at the bottom of the tube and the supernatant could be discarded. The RNA pellets were washed twice with 500 µl of 75% ethanol, centrifuged at 8,000 g for 3 min at RT. After the second wash, the tubes were centrifuged for an additional minute to completely get rid of the ethanol. The pellets were resuspended in RNase-free water, vortexed for 20 s and centrifuged to collect the total RNA to the bottom of the tube. RNA concentration and purity was assessed using the NanoDrop 8000 (Thermo Scientific) and the Agilent 2100 Bioanalyzer (Agilent).

2.2.9 cDNA synthesis The cDNA was synthesised starting from 500 ng of RNA using Transcriptor Universal cDNA master (Roche, Cat. No. 05893151001), following the manufacturer’s instructions. Cycling conditions: Step Temperature (°C) Time (min) Primer Annealing 25 5 Reverse Transcription 55 20 Denaturation 85 5 Hold 4 10

2.2.10 RT-qPCR

2.2.10.1 RT-qPCR from C. elegans samples

The cDNA was diluted 1:10 with dH2O. RT-qPCR was performed using the SyBr Green system (Roche) according to the manufacturer’s instructions. Cycling conditions:

Step Sub step Temperature (°C) Time (hh:mm:ss) Pre-incubation 95 00:05:00 Denaturation 95 00:00:10 Amplification Primer Annealing 60 00:00:10 Extension 72 00:00:10 95 00:00:05 Melting curve 65 00:01:00 97 Cooling 40 00:00:30

The expression of GFP was normalised to the reference gene ama-1.

2.2.10.2 RT-qPCR from human samples

The cDNA was diluted 1:10 with dH2O. RT-qPCR was performed using the Roche Universal Probe Library (UPL) system (Roche) according to the manufacturer’s instructions. The expression of the various gene analysed was normalised to the geometric mean of two reference genes: ACTB and POLR2J.

2.2.11 Visualisation of rRNAs on FA gel The RNA was resuspended with 5X RNA loading buffer, incubated at 65°C for 5 min and cooled down on ice. The samples were then loaded on pre-equilibrated formaldehyde (FA) gel. The gel was run at 5 V for 1.5 h. The bands were visualised using the Gel Doc XR system (BioRad).

2.2.12 Microarray analysis Total RNA was extracted using RNAzol® RT as described in section 2.2.8.2; concentration and purity were analysed using the NanoDrop 8000 (Thermo Scientific) and the Agilent 2100 Bioanalyzer (Agilent). The depletion levels of each gene of interest were analysed by RT-qPCR. The samples were then shipped on dry ice to OGT - Oxford Gene Technology (Oxford, UK) for hybridisation to SurePrint G3 Human Gene Expression 8x60k microarray chips (Agilent) and subsequent scanning. Bioinformatics analyses were performed by Graeme Grimes of the in-house Bioinformatics service. Analysis of the microarray data was performed with R using the Limma package. Expressed genes were filtered by calculating the 95% percentile of the negative control probes, then keeping probes that were at least 10% brighter on 3 arrays.

2.2.13 Cloning Phusion® High-Fidelity DNA polymerase (NEB) was used to amplify GNL2 cDNA following the manufacturer’s instructions. The forward and reverse primers contained overhangs corresponding to the consensus sequence for NheI and BclI respectively, to facilitate cloning into the final vector pCG-T7. NheI and BclI produce overhangs compatible with XbaI and BamHI respectively. Cycling conditions:

Step Temperature (°C) Time Initial Denaturation 98 30 s 35 cycles of: Denaturation 98 10 s Annealing 62 30 s Extension 72 90 s Final extension 72 10 min Hold 4 10 min

The PCR product was run on a 1% agarose gel. The band corresponding to GNL2 was excised from the gel and purified with a QIAquick Gel Extraction Kit (Qiagen) following the manufacturer’s instructions.

The purified DNA was ligated into the pCR™Blunt II-TOPO® vector (Life Technologies) following the manufacturer’s instructions. The ligated product was then transformed in DH5α competent cells (Life Technologies). The pCR™Blunt II-TOPO®-GNL2 vector was mini-prepped using the QIAprep Spin Miniprep Kit (Qiagen), digested with NheI 37°C for 1h, and then with BclI at 50°C for 1h. After the digestion, the reaction was run on a 1% agarose gel, and the band corresponding to GNL2 with overhangs was purified. The pCG-T7 vector was digested with XbaI and BamHI at 37°C for 1h, run on a 1% agarose gel, and the band corresponding to the digested vector was purified. pCG-T7 and GNL2 were ligated overnight at 4°C. The vector was then sequenced to confirm the correctness of the insert.

2.2.14 Bacterial transformation 50 µl of bacteria were transformed with 1-10 µl of DNA plasmid. The mixture bacteria/DNA plasmid was incubated for 30 min on ice. Bacteria were then heat- shocked by incubating the tubes at 42°C for 20 s, and then put back on ice for 2 min. 950 µl of LB were added to the bacteria and the tubes incubated at 37°C for 1h in a gently shaking incubator. Bacteria were then seeded on LB-agar plates with the appropriate antibiotic.

2.2.15 Transfection of expression vectors HeLa cells were seeded on day 0 in 6-well plates or in 10 cm plates so that they would be roughly 80% confluent the day after. On day 1 cells were transfected with expression vectors using Lipofectamine® 2000 according to the manufacturer’s instructions. Cells were then collected after 48 hours.

2.2.15.1 Transfection of expression vectors for mass spectrometry analysis Cells were plated in 10 cm plates transfected with 250 ng of one of the following expression vectors: - pRK5 HA empty vector;

- pRK5 HA-SEC13; - pCG T7 empty vector; - pCG T7-GNL2; and 4 µg of the carrier vector 3x FLAG empty vector. 30 µl of Lipofectamine® 2000 in a final transfection volume of 2 ml. 24 hours post-transfection cells were transferred to 14 cm plates and collected the day after.

2.2.16 Protein extraction Cells were washed once with 1X PBS. 1X PBS was added to the cells in order to scrape them off the plate. Cells were then transferred to eppendorf tubes and centrifuged at 2500 rpm for 5 min. After removing the supernatant, the pelleted cells were resuspended in the appropriate volume of IP buffer (1 ml for 107 cells) and incubated on ice for 20 min. The tubes were then centrifuged at 13200 rpm for 15 min at 4ºC. The supernatant was then transferred to a new tube.

2.2.17 Immunoprecipitations Beads preparation: 20 µl of beads (for each sample) were washed 3 times with 1X PBS. The protein extracts were pre-cleared by incubating them with sepharose G beads for 1h at 4°C on a rotating mixer. The tubes were subsequently centrifuged for 2 min to collect the beads at the bottom of the tube. An aliquot of the protein extracts was collected as “inputs”. The protein extracts were incubated with the relevant anti-tag beads overnight at 4°C on a rotating mixer. The beads were then washed 4 times to remove any residual unbound proteins. Protein extracts obtained from cells transfected with T7-GNL2 and T7 empty vector were incubated with anti-T7 agarose beads. Protein extracts obtained from cells transfected with HA-SEC13 and HA empty vector were incubated with anti-HA magnetic beads (Pierce).

2.2.18 Protein gel electrophoresis Prior to loading inputs and beads on a gel, they were resuspended as follow:

Inputs: 3.5 µl 4X LDS loading dye: 2.5 µl 10X Reducing agent: 1 µl dH2O: 3 µl

Beads: 20 µl 4X LDS loading dye: 5 µl 10X Reducing agent: 2 µl dH2O: 13 µl

Beads and inputs were incubated at 95°C for 5 min, briefly centrifuged and loaded on a protein gel. T7 inputs and immunoprecipitations were loaded on a 3-8% Tris-Acetate protein gel (Life Technologies) and the gel was run in 1X Tris-Acetate SDS running buffer (Life Technologies). HA inputs and immunoprecipitations were loaded on a 4-12% Bis-Tris protein gel (Life Technologies) and the gel was run in 1X MOPS SDS running buffer (Life Technologies).

2.2.19 Western Blot The gels were then transferred onto nitrocellulose membranes (ThermoFisher Scientific) using the iBlot® 2 gel transfer device, program 3 (20V) for 7 minutes. The membranes were then incubated for 2 min with Ponceau S (Sigma-Aldrich) to determine whether proteins had been transferred from the gel to the membrane. The membranes were then washed three times with dH2O and incubated for 30 min at RT with the blocking solution. The membrane was washed once with 1X PBS + 0.02% Tween® 20. Membrane were then incubated over-night at 4°C on a shaker with the primary antibodies.

Primary antibodies were diluted in the blocking solution at the following concentration:

Rabbit anti-GNL2 (Sigma-Aldrich) 1:1,000 Rabbit anti- SEC13 (LS Biosciences) 1:1,000 Mouse anti-TUBB (anti-tubulin β, clone 2.1) (Sigma-Aldrich) 1:5,000 Rabbit anti-XRN2 (Bethyl) 1:2,000 The day after the membranes were washed five times with 1X PBS + 0.02% Tween® 20. And incubated for an hour with the appropriate secondary antibodies. Secondary antibodies were diluted in the blocking solution at the following concentration:

Horse anti-mouse 1:20,000 Goat anti-rabbit 1:10,000 Membranes were washed five times with 1X PBS + 0.02% Tween® 20. Proteins were visualised using ChemiGlow West Chemiluminescence substrate kit (ProteinSimple).

2.2.20 Mass spectrometry Proteins were extracted as described in section 2.2.15 and immunoprecipitated using beads against the tag proteins, as described in section 2.2.16. The samples were then processed by our collaborator Dr. Jim Willis (Cancer Research UK – IGMM).

2.2.21 Nuclear-cytoplasmic cells fractionation Nuclear-cytoplasmic fractionation of HeLa cells was performed with the NE-PER™ Nuclear and cytoplasmic extraction reagents (ThermoFisher Scientific), following the manufacturer’s instructions with a minor modification (highlighted in bold).

1) Cells were washed and harvested with trypsin (Life Technologies) and centrifuged at 500 g for 5 minutes. Washed and resuspended in 1X PBS. Cells were transferred to a clean eppendorf tube and centrifuged again. The supernatant was carefully removed.

2) Ice-cold CERI + protease inhibitor was added to the pellet cells (100 µl for 106 cells). The tube was vortexed vigorously for 15 s, then incubated on ice for 10 min. 3) Ice-cold CERII was added to the tube (5.5 µl for 106 cells). The tube was vortexed vigorously for 5 s, then incubated on ice for 1 min. The tube was vortexed for 5 s and centrifuged at 16,000 g for 5 min at 4°C. 4) The supernatant, corresponding to the cytoplasmic fraction, was transferred to a clean pre-chilled tube. 5) The pellet was then washed twice with the CERI buffer. 6) The pellet was then resuspended in ice cold NER + protease inhibitor (50 µl for 106 cells). The tube was vortexed vigorously for 15 s. the tube was incubated on ice for 40 min (vortexed for 15 s every 10 min). 7) The tube was centrifuged at 16,000 g for 10 min at 4°C. 8) The supernatant, corresponding to the nuclear fraction, was transferred to a clean pre-chilled tube.

2.2.22 Polysomes profiling

HeLa cells were ~ 60%-70% confluent when they were harvested. 1) The culturing medium was removed from the cells. The plate was positioned on ice and the cells were washed with ice-cold 1X PBS. 2) 500 µl of implemented 1X RSB buffer was added to the cells (~ 4*106 cells), being careful to cover the entire surface of the dish. Collect cells in eppendorf tubes and incubate them on ice for 10 min. 3) The cells were then transferred onto QIAshredders and centrifuged for 2 min at maximum speed to homogenise the samples. 4) The flow-through were transferred to clean eppendorf tubes and centrifuged for 10 min at maximum speed at 4°C. The supernatant from each tube was then transferred to a clean eppendorf tube and. 5) 10%-45% sucrose gradients were prepared using a BioComp gradient station instrument (model 153) and were then allowed to set and cooled down at 4°C for 30 min.

6) The samples were loaded on top of the gradients and then ultra-centrifuged at 41,000 g for 2 h at 4°C 7) The absorbance was measured at 254 nm.

Chapter 3: A genome-wide screen in C. elegans to uncover novel NMD factors

3.1 Introduction

Genome-wide screens have proven to be valuable tools to analyse a variety of processes in C. elegans, such as the formation of dauer larvae at high temperature, moulting and spindle assembly (Ailion & Thomas, 2003, Frand et al., 2005, Maia et al., 2015). In order to interrogate the nematode’s genome for additional NMD factors, Dr. Dasa Longman, a post-doctoral fellow in our laboratory, developed a transgenic strain of C. elegans, named PTCxi, carrying a GFP-based NMD reporter (PTCx) (Longman et al. 2007). This reporter consists of a fusion between GFP and LacZ, with a premature termination codon (PTC) positioned in the first exon of LacZ (Figure 3.2, panel a). The PTC arises from a point mutation in position 2,193 which changes a cytosine to thymidine, resulting in a change of codon from CAA to TAA, and therefore introducing a stop codon. This feature makes the reporter sensitive to NMD regulation. Worms carrying this reporter have considerably reduced expression of GFP, since this transcript is now recognised by the NMD machinery and targeted for degradation. When the NMD pathway is defective, for instance when one of the factors that are required for NMD is silenced, the GFP mRNA is no longer degraded, resulting in GFP expression. The PTCxi transgenic strain of C. elegans carries the PTCx reporter integrated in the genome. The PTCx reporter was injected in the gonads of young adults which were subsequently irradiated with gamma rays. This caused a multi-copy integration of the reporter. Mapping of the integration revealed that the reporter was integrated in chromosome III in proximity of the unc-32 locus (this analysis was carried out by the Tim Schedl’s laboratory, Dasa Longman personal communication). In their study, Longman et al. made use of an RNAi library developed in the laboratory of Dr. Julie Ahringer. This library targets 86% of the predicted genes in C. elegans. In this screen two novel NMD factors were discovered, and named smgl-1 and smgl-2 (smg-lethal-1 and 2), since knock-down of either gene is lethal in nematodes (Longman et al., 2007).

Taking advantage of the PTCxi transgenic strain, I performed a similar genome-wide screen, but using a different RNAi library that included many previously untested genes. This library, termed the ORF-RNAi library v.1.1, was developed in the laboratory of Dr. Marc Vidal (see Materials and Methods for a description of the library). It consists of 11,511 clones, targeting 55% of the C. elegans genome (10,953 genes; as some of the genes are targeted by more than one clone). Importantly, 1,736 genes were not previously silenced by the Ahringer library (Reboul et al., 2003, Rual et al., 2004).

In order to fully understand the screening methods and strategy, it is important to briefly introduce the life cycle of C. elegans.

3.1.1 C. elegans

C. elegans worms are free-living diploid nematodes present in the soil in temperate climates. They are of small dimensions (adult worms are ~1 mm long) and can be of two genders: hermaphrodite or male. C. elegans has been used for genetics studies for over forty years, since Sydney Brenner set out to find a suitable organism system to study many processes in vivo (Brenner, 1974). Hermaphrodites worms lay eggs which hatch into larvae. There are four larval stages of development prior to the maturation of the worms into adults: L1-L4 (Figure 3.1). Moulting determines the transition from one larval stage to the following stage (Hope (ed.), 1999). Two of the most important parameters for the life cycle of C. elegans are temperature and the presence of food. C. elegans display optimal growth between 16°C and 25°C, with the best growth conditions at 20°C (Hope (ed.), 1999). In adverse conditions, for instance during starvation, L2 larvae develop into dauer larvae (Figure 3.1). These worms undergo several changes: amongst other things, they are less active from a metabolic point of view, they do not feed and are thinner than L3 larvae. C. elegans can survive for several months at this stage. When normal growth conditions

55 are restored, dauer larvae moult into L4 larvae and the normal development of the worms is reinstated (Cassada & Russell, 1975, Klass & Hirsh, 1976).

Figure 3.1 C. elegans life cycle Embryos laid by adult worms hatch into L1 larvae. Four subsequent larval stages (L1-L4) lead to the development of mature adults. Under stress conditions, L2 larvae can moult into dauer larvae which will moult into L4 larvae when normal growth conditions are restored.

3.2 Screen description

The screen of C. elegans for the presence of novel NMD factors was carried out by feeding, as described in Kamath et al. (2001). A diagram depicting the screen is represented in Figure 3.2, panel b. On day one bacterial clones from the library were grown in liquid culture at 37°C. C. elegans embryos were collected from adult worms, resuspended in M9 physiological solution, without food, and allowed to hatch over-night at 20°C to obtain arrested populations of L1 larvae. Next day, the expression of dsRNAs was induced by adding IPTG to the bacterial cultures. After four hours the library clones were aliquoted in 96-well plates and 25 L1 larvae were added per well. The worms were allowed to feed on the bacteria expressing gene- specific dsRNAs for two days. As expected, this resulted in the down-regulation of the corresponding gene product. The plates containing the nematodes were then visually scored for the presence of GFP fluorescence, indicating that the inactivated gene is a potential NMD factor in C. elegans.

Figure 3.2 - Screen strategy. (a) Structure of the PTCx NMD reporter. The reporter consists of a fusion between GFP (in green) and LacZ (in blue). The PTC is positioned in the first exon of LacZ. (b) Embryos collected from PTCxi worms were allowed to hatch over-night at 20°C without food. The bacterial clones from the library were grown over-night at 37°C. The expression of the dsRNAs was induced by adding IPTG to the bacterial culture for 4 hours. Synchronised populations of L1 PTCxi larvae were aliquoted in 96-well plates and supplemented with bacteria expressing dsRNAs. Worms were allowed to feed on the bacteria for 60-72 hours. The presence of GFP-positive worms indicated inactivation of putative novel NMD factors.

3.3 Screen results

Clones that scored positive for the appearance of GFP were tested multiple times to determine whether the up-regulation of GFP was reproducible. I identified seven positive clones, recognised by unique identifiers: Y48G8AL.6, T19A6.2, Y77E11A.13, Y87G2A.4, C47B2.4, F52B11.3 and ZK909.2 (listed in Table 3.1). To determine the identity of the positive clones and confirm it matched the library annotation, I cultured the bacteria corresponding to the positive clones and sequenced the vector. The obtained sequences were blasted using the online tool at wormbase.org, and resulted in 100% nucleotide identity between the blasted sequences and the targeted genes.

Y48G8AL.6 corresponds to the core NMD factor smg-2 (suppressor with morphological effect on genitalia). Inactivation of smg-2 led to high expression of GFP (Figure 3.3, panel II), confirming that the screen was working successfully. The other six positive clones (Table 3.1) correspond to genes that were not previously linked to NMD, and had been proposed to take part in a variety of processes, from GTP-binding to trafficking of vesicles. All of the positive clones express GFP at higher level than the worms treated with empty vector (Figure 3.3, panel III-VIII), where GFP cannot be visualised (Figure 3.3, panel I). The identity of the positive clones is as follows: T19A6.2 (ngp-1), Y77E11A.13 (npp-20), Y87G2.4 (aex-6), C47B2.4 (pbs-2), F52B11.3 (noah-2) and ZK909.2 (kin- 1).

Table 3.1 - List of positive clones identified in the screen. The table indicates the clone identifier and the corresponding gene name for the six novel putative NMD factors. Also shown are the previously proposed functions of these factors, and the previously characterised phenotypes resulting from knock down of these genes. The human homologues (if present) are also indicated.

Figure 3.3 Putative NMD factors identified in the screen. Panels i-viii show brightfield images of the worms, whilst Panels I-VIII show GFP expression in the worms. Worms treated with an empty vector were used as negative controls as they do not express any GFP (panel I) and worms treated with dsRNAs against smg-2 were used as positive controls as they express high level of GFP (panel II). Scale bar: 100 µm.

3.3.1 C. elegans phenotype

Next, we investigated whether these newly identified NMD genes were required for the development of the worms.

Transgenic nematodes treated with an empty vector (negative control) did not display any aberrant phenotype and reach the stage of fertile adulthood in the time frame of the experiment (Figure 3.3, panel i). I confirmed that worms feeding on RNAi against smg-2 (Figure 3.3, panel ii) exhibited morphological defects of the genitalia, in accordance with a previous report (Hodgkin et al., 1989). These worms could produce progeny (Figure 3.3, panel ii). In contrast, depletion of the putative NMD factors identified in this screen, with the exception of AEX-6, led to impairments in the normal development of nematodes. Worms depleted of NGP-1 were previously reported to arrest their growth during larval development (Simmer et al., 2003). In agreement with this study, treatment of the PTCxi worms with dsRNAs targeting ngp-1 led to severe developmental defects, with the majority of the worms arrested at the L1-L2 stage of larval development. The treated worms never reached adulthood and were thin and pale (Figure 3.3, panel iii, Figure 3.4, panel B) compared to the wild-type worms (Figure 3.3, panel i, Figure 3.4, panel A). These observations suggest that ngp-1 is important for the correct development of C. elegans.

Figure 3.4 Knock-down of ngp-1 led to worms arrested at L1-L2 stage. Panel A shows a brightfield image of a wild-type worm, whilst Panel B shows a brightfield image of a worm depleted of ngp-1. Scale bar: 100 µm.

Worms treated with dsRNAs against npp-20 developed further than the worms depleted of NGP-1, but not to the extent of the wild-type worms (Figure 3.3, panel i, Figure 3.5, panel A). They were very fragile, and died by bursting at the L3-L4 larval stage (Figure 3.5, panel B) or when they reached adulthood (as indicated by the

62 presence of early-stage embryos) (Figure 3.3, panel iv). npp-20 is therefore required for proper development of the worms, as previously reported (Kamath et al., 2003).

Figure 3.5 Knock-down of npp-20 led to fragile worms. Panel A shows a brightfield image of a wild-type worm, whilst Panel B shows a brightfield image of a worm depleted of npp-20. The red arrow indicates the remaining of the burst worm. Scale bar: 100 µm.

C. elegans depleted of aex-6 were previously reported to be constipated due to abdominal muscles not working properly (Thomas, 1990) and we found in our screen

63 that these worms depleted of aex-6 displayed an egg laying defect (Figure 3.6, panel B). In healthy worms, oocytes and embryos of progressive developmental stages can be observed in the gonad. These are aligned and oriented in the same direction (Figure 3.6, panel A). In worms lacking aex-6, the embryos could not be actively expelled from the gonad due to the inability of the abdominal muscles to work properly, resulting in misalignment of the embryos. Embryos were eventually expelled due to pressure. This phenotype is in accordance with what was previously reported as the muscles responsible for expulsion of the embryos are also involved in defecation.

Figure 3.6 Knock-down of aex-6 led to worms with an egg-laying defect. Panel A shows a brightfield image of a wild-type worm, whilst Panel B shows a brightfield image of a worm depleted of aex-6. The red arrows indicate the alignment (A) or misalignment (B) of the embryos in the gonad on both side of the vulva (indicated by the blue arrowhead). Scale bar: 100 µm.

Depletion of pbs-2 was reported to lead to larval arrest and sterility in C. elegans (Simmer et al., 2003). Worms depleted of pbs-2 in our screen arrested during larval development at the L2-L3 stage. They also displayed a swollen intestine and were extremely pale (Figure 3.3, panel vi, Figure 3.7, panel B) in comparison with wild- type worms (Figure 3.3, panel i, Figure 3.7, panel A). The swollen intestine phenotype was not previously reported by other studies and it would be interesting to develop null mutants for pbs-2 to validate my observations.

Figure 3.7 Knock-down of pbs-2 led to pale worms arrested at L2-L3 stage. Panel A shows a brightfield image of a wild-type worm, whilst Panel B shows a brightfield image of a worm depleted of pbs-2. The orange arrowheads indicate the width of the intestine. Scale bar: 100 µm. noah-2 was identified by Frand et al. (2005) and reported to be critical for moulting, an essential process for nematodes in order to progress through the different developmental stages. In accordance with this report, we found that worms depleted of noah-2 were arrested at the L2-L3 stage. Worms were pale and sluggish (Figure 3.3, panel vii, Figure 3.8, panel B) in comparison with wild-type worms (Figure 3.3, panel i, Figure 3.8, panel A).

Figure 3.8 Knock-down of noah-2 led to sluggish worms arrested at L2-L3 stage. Panel A shows a brightfield image of a wild-type worm, whilst Panel B shows a brightfield image of a worm depleted of noah-2. Scale bar: 100 µm.

The final positive clone identified in the screen corresponds to kin-1. It is an essential gene, and its depletion led to slower growth and ultimately to the death of the worms, which burst through the vulva (Figure 3.9, panel B), as previously reported (Rual et al., 2004) and in contrast to the wild-type worms (Figure 3.9, panel A).

Figure 3.9 Knock-down of kin-1 led to worms prone to bursting. Panel A shows a brightfield image of a wild-type worm, whilst Panel B shows a brightfield image of a worm depleted of kin-1. The yellow arrow indicates the bursting of the worm through the vulva (indicated by the blue arrowhead). Scale bar: 100 µm.

3.4 Screen validation

GFP up-regulation at the protein level (Figure 3.3, panels III-VIII) is a powerful tool to detect positive clones during the screen, however NMD acts at the transcript level. In order to assess whether a higher expression of GFP protein was due to a stabilisation of the PTC-containing transcript, RNA levels of the NMD reporter were analysed by RT-qPCR. In normal conditions the PTCx transcript is an NMD substrate and therefore it is recognised by the NMD machinery and targeted for degradation. PTCxi worms treated with an empty vector have low levels of the PTCx transcript, whilst worms depleted of smg-2 strongly up-regulate the GFP transcript. Out of the six positive clones identified in the screen, depletion of five of them resulted in up-regulation of the GFP mRNA, confirming that NGP-1, NPP-20, AEX- 6, PBS-2 and NOAH-2 are putative NMD factors in C. elegans (Figure 3.10).

Figure 3.10 – Expression level of NMD reporter mRNA following knockdown of putative NMD factors. RT-qPCR for the GFP-based NMD reporter in PTCxi C. elegans fed on bacteria containing an empty vector control plasmid (ev), or plasmids encoding dsRNA against a positive control (smg-2), and the six novel putative NMD factors identified in the screen. Gene expression levels were normalised to the housekeeping gene ama-1 and are displayed relative to the empty vector control. Error bars represent the standard error of the mean of at least 3 biological replicates. * = p-value <0.05, Mann-Whitney test.

The knock-down efficiency could not be evaluated by RT-qPCR because of the nature of the RNAi vectors used in the screen. The vectors encode dsRNAs which correspond to the whole ORF of the targeted genes. This implies that gene-specific

69 primers are unable to distinguish between mRNA and dsRNAs, therefore making RT-qPCR not suitable to analyse the knock-down efficiency. Antibodies against the protein encoded by the novel putative NMD factors are not available. I therefore assumed the genes were efficiently knocked down when GFP expression and phenotypic effects were detectable by microscopy.

3.5 Protein conservation

NGP-1 is a putative GTP-binding protein. The protein comprises two domains: an N- terminal domain, NGP1NT, and a GTP-binding domain, MMR_HSR1 (Figure 3.11, panel a). The NGP1NT domain (in light green) is highly conserved, but no function has yet been associated with it. The MMR_HSR1 domain (in green) is typical of this subfamily of GTPases. This GTP-binding domain is characterised by five G-motifs (indicated by black bars), which are permutated with respect to other GTPases, and are ordered G5-G4-G1-G2-G3. It is conserved from Saccharomyces cerevisiae to higher vertebrates (Figure 3.12, panel a), including Homo sapiens, where the protein is called GNL2. The percentage of amino acid identity across the species varies between 39% and 86%, indicating a high conservation of the amino acid sequence (Figure 3.12, panel b).

Figure 3.11 - Domain composition of the novel C. elegans NMD proteins and of their human homologues. (a) Protein composition of NGP-1 and of GNL2. Both proteins have two conserved domains: the N-terminal NGP1NT domain (in light green), and a GTP-binding domain, MMR_HSR1 (in green). The five G-motifs typical of the GTP-binding domain are indicated by black bars. (b) Protein composition of NPP-20 and of SEC13. Both proteins are characterised by six WD40 domains (in orange). (c) Protein composition of AEX-6 and of its two human homologues RAB27A and RAB27B. The three proteins comprise the characteristic RAB domain (in purple). (d) Protein composition of PBS-2 and of its two human homologues PSMB7 and PSMB10. All three proteins contain a proteasome domain (light blue). The two human proteins also contain a C-terminal domain, called Pr_beta_C (in blue). (e) Protein composition of NOAH-2 and Drosophila NOMPA. The two proteins are composed of four PAN_AP domains (in red) and one ZP domain (in orange). They also have an N-terminal signal peptide (SP, in yellow) and a C-terminal trans-membrane signal (TM, in black) Ce: Caenorhabditis elegans; Hs: Homo sapiens; Dm: Drosophila melanogaster.

NPP-20 was identified as a nucleoporin member in C. elegans (Galy et al., 2003). It is highly conserved between eukaryotes: there is 40% identity in the amino acid sequence between human and C. elegans, which reaches 97% between human and mouse (Figure 3.13). Both the human and the worm proteins are composed of six WD40 repeats (Figure 3.11, panel b). WD40 domains (in orange) are associated with protein-protein interactions.

C. elegans AEX-6 is a member of the Rab small GTPases family. Like other members of this family, AEX-6 is involved in the trafficking of vesicles. It acts in the docking of synaptic vesicles and of dense core vesicles (Mahoney et al., 2006, Feng et al., 2012). AEX-6 is conserved through evolution in eukaryotes and there are two paralogues proteins in human, mouse and zebrafish. Both AEX-6 and the two human homologues, RAB27A and RAB27B, are characterised by a highly conserved RAB domain (Figure 3.11, panel c and Figure 3.14, panel a). As the three proteins are mainly composed by the RAB domain (in purple), they are characterised by a high percentage of identity in their amino acid sequence (58% between the C. elegans protein and its human homologues RAB27A and RAB27B) (Figure 3.14, panel b).

C. elegans PBS-2 was identified as a proteasome subunit (Davy et al., 2001), and, in accordance with this function, the protein is characterised by the proteasome domain (Figure 3.11, panel d). The same domain (in light blue) is present in the two human homologues, PSMB7 and PSMB10, which, in addition, contain a C-terminal domain typical of the beta-subunits of the proteasome (Pr_beta_C, in blue). PBS-2 homologues can be found in higher eukaryotes and in yeast, with a percentage of identity in the amino acid sequences that varies between 35% (human and C. elegans) to almost 90% in mammals (Figure 3.15).

NOAH-2 only has a homologue in D. melanogaster, called NOMPA (Figure 3.16). The two proteins display a low percentage of amino acid identity, 10%. Nevertheless, both proteins are characterised by the same domains (Figure 3.11, panel e). NOAH-2 and NOMPA contain four apple-like domains (PAN_AP, in red), that are predicted to be involved in the binding of proteins or carbohydrates. Both proteins also comprise a domain called zona pellucida (ZP, in orange), which is usually present in secreted glycoproteins involved in development. This is in accordance with NOAH-2

72 being required for moulting (Frand et al., 2005). An N-terminal signal peptide (SP, in yellow) and a C-terminal trans-membrane motif (TM, in black) are present in both proteins.

Figure 3.12 - Protein conservation of GNL2 throughout evolution. (a) Amino acid alignment of human GNL2 with a subset of its homologues, including C. elegans NGP-1. Conservation of a specific amino acid in all species considered is indicated in black. Intermediate shades of gray are used to display varying levels of conservation. (b) The table shows the percentage of amino acid identity. Hs: Homo sapiens; Mm: Mus musculus; Dr: Danio rerio; Dm: Drosophila melanogaster; Sc: Saccharomyces cerevisiae; Ce: Caenorhabditis elegans.

Figure 3.13 - Protein conservation of SEC13 through evolution. (a) Amino acid alignment of human SEC13 with a subset of its homologues, including C. elegans NPP-20. Conservation of a specific amino acid in all species considered is indicated in black. Intermediate shades of gray are used to display varying levels of conservation in

74 subsets of species. (b) The table shows the percentage of amino acid identity. Hs: Homo sapiens; Mm: Mus musculus; Dr: Danio rerio; Dm: Drosophila melanogaster; Sc: Saccharomyces cerevisiae; Ce: Caenorhabditis elegans.

Figure 3.14 - Protein conservation of RAB27A and RAB27B through evolution. (a) Amino acid alignment of the human paralogues RAB27A and RAB27B with a subset of their homologues, including C. elegans AEX-6. Conservation of a specific amino acid in all species considered is indicated in black. Intermediate shades of gray are used to display varying levels of conservation in subsets of species. (b) The table shows the percentage of amino acid identity. Hs: Homo sapiens; Mm: Mus musculus; Dr: Danio rerio; Dm: Drosophila melanogaster; Sc: Saccharomyces cerevisiae; Ce: Caenorhabditis elegans.

Figure 3.15 - Protein conservation of PSMB10 and PSMB7 through evolution. (a) Amino acid alignment of the human paralogues PSMB10 and PSMB7 with a subset of their homologues, including C. elegans PBS-2. Conservation of a specific amino acid in all species considered is indicated in black. Intermediate shades of gray are used to display varying levels of conservation in subsets of species. (b) The table shows the percentage of amino acid identity. Hs: Homo sapiens; Mm: Mus musculus; Dr: Danio rerio; Sc: Saccharomyces cerevisiae; Ce: Caenorhabditis elegans.

Figure 3.16 - Protein conservation of NOAH-2 through evolution. Amino acid alignment of C. elegans NOAH-2 with its D. melanogaster homologue NOMPA. The conservation of specific amino acids between the two proteins is indicated in black. The percentage of amino acid identity between the two proteins is indicated at the bottom of the alignment. Ce: Caenorhabditis elegans; Dm: Drosophila melanogaster.

3.6 Discussion

Forward genetics screens are based on the introduction of random mutations in the genome of the organisms of interest (for example C. elegans). Mutant worms with the desired phenotype are then analysed in an attempt to identify the gene that is linked to that specific phenotype. One obvious limitation of forward genetic screens is that all lethal phenotypes are over-looked. An example of a forward genetic screen is the one performed by Hodgkin et al. (1989). They carried out a screen on a strain of C. elegans carrying a mutation in unc-54 (named r293), to look for mutations that would suppress the uncoordinated phenotype (a paralysis of the worms). After exposing the unc-54(r293) worms to mutagenesis, they scored the progeny for dominant suppressors (the phenotypic effect of which was observed on the F1 progeny) and for recessive suppressors (the phenotypic effect of which was detectable in F2). They identify six genes belonging to the second class of suppressors, named smg-1 to smg-6. The development of RNAi libraries presented itself as a solution to this issue, because it allowed for reverse genetic screens. In RNAi reverse genetic screens specific genes are silenced and the effects of such down-regulation on the worms (in terms of development and behaviour) are studied, allowing the analysis of genes that are essential. Another advantage of RNAi screens, that made them a powerful tool to use in genetic screens, is that dsRNAs can cross cell boundaries (Fire et al., 1998). This resulted in the development of three methods of delivery: injections, soaking and feeding. Feeding worms with double stranded RNA (produced by bacteria) is the easiest and cheapest way of performing a screen on a wide number of animals (Kamath et al., 2001).

Screening of the PTCxi transgenic strain of C. elegans, carrying an NMD reporter, with the ORF-RNAi library v.1.1 led to the discovery of five novel putative NMD factors: NGP-1, NPP-20, AEX-6, PBS-2 and NOAH-2. It also identified SMG-2 amongst the positive clones, but failed to identify four known NMD factors that were targeted by the library: SMG-6, SMG-7, SMGL-1 and SMGL-2. It is possible that the dsRNAs targeting these four genes were not efficient in down-regulating these gene products.

The previous screen carried out in our laboratory made use of a library developed in Julie Ahringer’s laboratory, which targets 86% of the nematode genome (Fraser et al., 2000, Kamath et al., 2003, Longman et al., 2007). This resulted in the discovery of two novel NMD factors, SMGL-1 and SMGL-2. Analysis of the list of genes targeted by the Ahringer Library reveals that ngp-1 and pbs-2 were amongst them. Simmers et al. (2003) reported that there is a variability of 10% to 30% between screens, even if carried out in parallel in the same laboratory, which could explain why those two genes were not identified as NMD factor in the original screen by Longman et al. (2007). Furthermore, the studies that first described the Ahringer and Vidal libraries had already indicated the inability of the screening strategy to identify all known phenotypes (Fraser et al., 2000, Kamath et al., 2003, Rual et al., 2004). For instance, Ahringer and colleagues could not detect 10% of the known embryonic lethal genes in their screen of chromosome I (Fraser et al., 2000). It is thus not surprising that the screen performed by Longman et al. (2007) and the screen I performed did not identify all known NMD factors. It is probable that a repeat of the screen with the same or a different library could lead to the discovery of additional factors involved in the regulation of NMD.

Knock-down of four out of five of the gene encoding the novel factors (ngp-1, npp- 20, pbs-2 and noah-2) led to strong developmental defects and, in the case of npp-20, to the death of the worms. Similar to the findings of this screen, the screen performed by Dasa Longman identified two novel NMD factors, SMGL-1 and SMGL-2, which are required for the development of C. elegans; since silencing of either gene is lethal in nematodes (Longman et al., 2007). Both these findings are in contrast with what was observed for the core NMD factors, the silencing of which led to morphological defects but not lethality (Hodgkin et al., 1989). This suggests that NMD is dispensable for the development of C. elegans. It is therefore possible that the severe phenotypes observed upon silencing of the new putative NMD factors are due to the functions of the factors in pathways unrelated to NMD.

NGP-1 has not been previously associated with any functions in C. elegans. Its yeast homologue, Nog2p, has been characterised as a GTPase involved in the biogenesis of the pre-60S subunit of the ribosome (Saveanu et al., 2001, Matsuo et al., 2014).

Nog2p is required for the cleavage of ITS2 (internal transcribed spacer 2) from the 27SB pre-rRNA (Saveanu et al., 2003, Talkish et al., 2012), and it has been suggested to be involved in the export of the pre-60S ribosomal subunit into the cytoplasm (Zhang et al., 2007). If this function is conserved in C. elegans, it could be the reason why worms depleted of ngp-1 are displaying severe defects in development. The D. rerio and the Arabidopsis thaliana homologues of NGP-1 are required for the biogenesis of the pre-60S ribosomal subunit (Im et al., 2011, Essers et al., 2014), and so is the human homologue GNL2 (Tafforeau et al., 2013), strongly suggesting that the role of this GTPase in ribosome biogenesis is conserved throughout evolution.

NPP-20 was identified as a nucleoporin member in a screen looking for components of the nuclear pore complex (Galy et al., 2003); its specific function in such complex was not analysed. Sec13p, the yeast homologue of NPP-20, is a component of the nuclear pore complex (Lutzmann et al., 2002, Enninga et al., 2003), and a component of the COPII coat vesicles, which are involved in the trafficking of cargo from the endoplasmic reticulum (ER) to the Golgi (Barlowe et al., 1994). These functions are conserved in human and they will be discussed in details in Chapter 4.

AEX-6 has been implicated in the exocytosis of both synaptic and dense core vesicles. In particular, it has been proposed that AEX-6 and its interactors RBF-1 (homologue of rabphilin) take part in the docking of both vesicle types (Mahoney et al., 2006, Feng et al., 2012). In human, as in C. elegans, RAB27 has a role in the trafficking of dense-core vesicles (Fukuda, 2006). Moreover, RAB27 has been shown to be involved in the trafficking of vesicles in several cell-types, such as in natural-killer cells, neutrophils and melanosomes (Liu et al., 2010, Herrero-Turrion et al., 2008, Klomp et al., 2007).

PBS-2 was identified as a component of the 26S proteasome complex (Davy et al., 2001) but the specific role of PBS-2 was not further characterised in C. elegans. There are two human homologues of PBS-2: PSMB7 and PSMB10. PSMB7 is a core component of the 20S proteasome. The 20S proteasome is formed by four protein complexes disposed as four concentric rings; each ring is composed of seven subunits. The inner rings are composed by β subunits, and the outer rings are

80 composed by α subunits. Three of the β subunits (PSMB5, PSMB6, and PSMB7) have proteolytic activity. The α subunits form a narrow ring that plays a role in regulating the function of the 20S proteasome (for a recent review, see Bhattacharyya et al., 2014). The proteasome is responsible for the degradation of proteins in the cell and it is therefore not surprising that down-regulation of one of its core subunits leads to cell death. PSMB10 replaces PSMB7 in a specialised form of the proteasome, the immunoproteasome. The best characterised function of the immunoproteasome is to produce the antigens to activate the immune system (for a recent review of the immunoproteasome, see Ferrington & Gregerson, 2012). noah-2 has been identified in a screen for genes involved in moulting (Frand et al., 2005), but there are no insights on the role it may play in the process. Worms depleted of noah-2 expression are severely affected and cannot progress through the larval stages, which fits with the idea of the worms being unable to moult and therefore develop into the next larval stage. It also fits with the observation that NOAH-2 only has a homologue in D. melanogaster, an animal that develops through moulting stages. The D. melanogaster homologue, nompA, was also identified in a genetic screen. The fly genome was screened to identify genes that were responsible for the animals not responding to touch stimuli, and in particular for mutants that generated abnormal mechanoreceptor potentials (Kernan et al., 1994). This led to the identification of nompA, which was then found to be expressed in the peripheral nervous system (PNS) of the embryos, specifically in type I sensory organs (Chung et al., 2001).

Chapter 4: Functional conservation of the novel NMD factors in humans

4.1 Introduction

The majority of NMD factors was identified in genetic or RNAi screens performed in S. cerevisiae (Culbertson et al., 1980, Leeds et al., 1992) or in C. elegans (Hodgkin et al., 1989, Cali et al., 1999, and Longman et al., 2007). Thanks to homology conservation, these factors were subsequently found and functionally characterised in humans (Perlick et al., 1996, Applequist et al., 1997, Sun et al¸ 1998, Denning et al., 2001, Serin et al., 2001, Chiu et al., 2003, and Ohnishi et al., 2003), as well as in other species, including D. melanogaster (Alonso & Akam, 2003, Gatfield et al., 2003) and in plants (Hori & Watanabe, 2005, Arciga-Reyes et al., 2006, Yoine et al., 2006, Kerenyi et al., 2007, Wu et al., 2007, Lloyd & Davies, 2013).

In the screen I performed I identified NGP-1, NPP-20, AEX-6, PBS-2, and NOAH-2 as a putative novel NMD factor in worms (Figure 3.4). Homology analysis revealed that NGP-1, NPP-20, AEX-6 and PBS-2 have homologues proteins in human. NOAH-2, instead, is not conserved in vertebrates, but has a homologue in the fruit fly, termed NOMPA. In an experiment performed by our collaborators Raúl Vallejo Baier and Claudio Alonso (University of Sussex), nompA knock-down within the embryonic peripheral nervous system (PNS) led to a significant up-regulation of an NMD reporter; this up-regulation was comparable to the effect observed following UPF1 depletion. This demonstrates that the Drosophila homologue of NOAH-2 behaves as a tissue-specific NMD factor in fruit fly embryos (Casadio et al., 2015).

As a follow up from the RNAi screen I wanted to investigate whether the human homologues of NGP-1, NPP-20, AEX-6 and PBS-2 are functionally conserved in human NMD using an NMD assay in human cells. To understand the mechanisms underlying NMD well characterised NMD substrates, such as genes coding for immunoglobulin, for the T-cell receptor β and for β-globin, have been used over the years (Mühlemann et al., 2001, Thermann et al., 1998). One of the first NMD targets to be identified was the β-globin transcript carrying a point mutation in codon 39, which gives rise to a premature stop codon (referred to

82 as NS39). This mutation naturally occurs in a subset of β-thalassemia patients and leads to reduced levels of the mRNA encoding β-globin (Chang & Kan, 1979, Losson & Lacroute, 1979). It was subsequently discovered that the degradation of the transcript was due to NMD (Peltz et al., 1993). HeLa cells carrying the wild-type β-globin allele or its mutated version with a premature stop codon in position 39 (NS39 allele), developed by Kulozik and colleagues (Thermann et al., 1998), were previously used in our laboratory and in others to assay for the function of putative NMD factors (the stable cell lines were a kind gift from Prof. Oliver Mühlemann, University of Berne). When NMD is impaired the transcript levels of the wild-type β-globin transgene remain unchanged. However, depletion of known NMD factors leads to increased expression of the mRNA of the β-globin transgene with the PTC (NS39). Thus, these two cell lines can be used as tools to assess whether a specific protein plays a role in the NMD response.

Taking advantage of these two cell lines I investigated whether the putative novel C. elegans NMD factors conserved their function in human NMD.

4.2 Functional conservation of the novel C. elegans NMD factors in human cells

I set out to investigate the role of the human homologues of the newly discovered putative C. elegans NMD factors in two human cell lines: in HeLa cells stably expressing a wild-type β-globin transgene, and in HeLa stably expressing a β-globin transgene containing a premature stop codon in position 39 (named β-globin NS39). The β-globin NS39 is a well characterised NMD substrate, and as such this transcript is rapidly degraded by the NMD machinery.

4.2.1 Depletion of GNL2 and SEC13 leads to up-regulation the PTC-containing NMD reporter transcript

To deplete HeLa cells carrying either the wild-type β-globin transgene or the β- globin NS39 I used siRNA pools targeting the putative NMD factors: GNL2 (ngp-1), SEC13 (npp-20), RAB27A and RAB27B (aex-6), individually or in combination, or

PSMB7 and PSMB10 (pbs-2), alone or in combination (the treatment of cells with siRNA pools is described in Materials and Methods, paragraph 2.2.6). Cells were also treated with control siRNA pools: a non-targeting pool, referred to as “scrambled” hereafter; a pool against UPF2 (a well characterised NMD factor), and a treatment lacking siRNA but including all other reagents, referred to as “mock” hereafter. Depletions of GNL2, SEC13 and both RAB27 paralogues were done for five days. The expression of the factors was reduced to 10%-30%, as judged by the level of residual transcript measured by RT-qPCR (Figure 4.1, panel c). Depletions of PSMB7 and PSMB10 were done for four days, as depletion of PSMB7 caused cell death at day 5. This four days RNAi treatment resulted in a robust depletion of PSMB7 and PSMB10 transcripts (Figure 4.2, panel c). The transcript level of the two β-globin reporters was analysed by RT-qPCR. Of the tested factors, only depletion of GNL2 and SEC13 specifically up-regulated the NS39 transcript, but not wild-type β-globin (Figure 4.1, panel b and a respectively). Depletion of PSMB7 alone or in combination with PSMB10 up-regulated both the wild-type and NS39 transcripts, indicating that this up-regulation could be due to an NMD-independent mechanism (Figure 4.2, panels a and b). These findings strongly suggested that GNL2 and SEC13 are both NMD factors in human.

Figure 4.1 - Knock-down of GNL2 and SEC13 leads to specific up-regulation of the NMD reporter mRNA. RT-qPCR for β-globin in HeLa cells carrying a wild-type (wt) β-globin reporter (a) or a β- globin transgene with a PTC at the codon 39 (NS39) (b). Cells were depleted of the putative NMD factors for 5 days. Cells were also mock-treated or treated with a non-targeting siRNA pool (scrambled) as negative controls. Cells were depleted of UPF2 as a positive control. (c) RT-qPCR for the depleted genes. Gene expression levels were normalised to the geometric mean of two housekeeping genes (POLR2J and ACTB), and are displayed relative to the

85 mock control. Error bars represent the standard error of the mean of at least 3 biological replicates. * = p-value <0.05, Mann-Whitney test.

Figure 4.2 - Knock-down of PSMB7, alone or in combination with PSMB10, leads to unspecific up-regulation of the β-globin mRNA. RT-qPCR for β-globin in HeLa cells carrying a wild-type (wt) β-globin reporter (a) or a β- globin transgene with a PTC at the codon 39 (NS39) (b). Cells were depleted of PSMB7 or PSMB10 (alone or in combination) for 4 days. Cells were also mock-treated or treated with a

86 non-targeting siRNA pool (scrambled) as negative controls. Cells were depleted of UPF2 as a positive control. (c) RT-qPCR for the depleted genes. Gene expression levels were normalised to the geometric mean of two housekeeping genes (POLR2J and ACTB), and are displayed relative to the mock control. Error bars represent the standard error of the mean of at least 3 biological replicates. * = p-value <0.05, Mann-Whitney test.

4.2.2 Depletion of GNL2 and SEC13 stabilises the NMD reporter transcript

One of the hallmarks of NMD is the rapid degradation of the mRNA substrates, and consequently, when NMD is defective, there is stabilisation of the mRNA substrates, leading to enhanced half-life of these transcripts (Mendell et al., 2004). The stability of the wild-type β-globin transcript should not be affected by an impaired NMD. To obtain a conclusive proof that the mechanism leading to the up-regulation of the NS39 β-globin transcript was indeed NMD, I decided to measure the increase in mRNA stability when NMD is abrogated.

Cells were depleted for five days of GNL2 or SEC13, or mock-treated. On the fifth day of depletion the cells were treated with Actinomycin D to inhibit transcription, RNA was harvested at different time points and analysed by RT-qPCR. As expected, depletion of GNL2 and SEC13 did not affect the half-life of the wild-type β-globin: the wild-type transcript is very stable in all three conditions analysed (Figure 4.3, panel a). In mock-treated cells, the transcript encoding β-globin NS39 was unstable and rapidly degraded (with a half-life of ~ 2 hours). Its half-life was increased to 15 hours or 9 hours when either GNL2 or SEC13 were depleted, respectively (Figure 4.3, panel b). This indicates that GNL2 and SEC13 are NMD factors.

Figure 4.3 – Depletion of GNL2 or SEC13 leads to a stabilisation of the NMD reporter mRNA. RT-qPCR for β-globin in HeLa cells carrying a wild-type (wt) β-globin reporter (a) or a β- globin transgene with a PTC at the codon 39 (NS39) (b). Cells were mock-treated (as negative control. Shown in blue) or were depleted of either GNL2 (in green) or SEC13 (in purple) for 5 days. On the fifth day cells were treated with ActD and collected after 0.5, 1. 2. 4. 6. 8 hours. Gene expression levels were normalised to the geometric mean of two housekeeping genes (POLR2J and ACTB), and are displayed relative to the first time point (0.5 h). Error bars represent the standard error of the mean of at least 3 biological replicates.

4.2.3 Endogenous NMD targets are regulated by GNL2 and SEC13

To validate these findings independently of the reporter system, I analysed the effect of GNL2 and SEC13 knock-down on a panel of previously described NMD substrates (Mendell et al., 2004, Longman et al., 2013): ARHGEF18, BMP2, PAXIP1 and TMC7. These endogenous NMD substrates were selected to be analysed for convenience as RT-qPCR primers were already available in the laboratory. Of these four endogenous NMD substrates only ARHGEF18 and PAXIP1 have annotated NMD features: a uORF (up-stream open reading frame) and the presence of an intron in the 3’ UTR respectively. Cells were depleted of either GNL2 or SEC13 for five days; using depletion of two known NMD factors, UPF1 and UPF2, as positive controls. Basal expression of the selected substrates was identified in mock-treated cells. Depletion of either core NMD factor, UPF1 or UPF2, or depletion of SEC13 led to up-regulation of the messenger RNA of the four endogenous NMD substrates analysed: ARHGEF18, BMP2, PAXIP1 and TMC7. Depletion of GNL2 led to the up-regulation of ARHGEF18, BMP2 and TMC7 transcripts (Figure 4.4). This confirms that GNL2 and SEC13 are bona fide NMD factors. PAXIP1 mRNA was not up-regulated upon depletion of GNL2 (Figure 4.4, panel d), which suggests that this target may be regulated by a subset of NMD factors only.

Figure 4.4 - Knock-down of GNL2 and SEC13 leads to up-regulation of a subset of endogenous NMD substrates. RT-qPCR for ARHGEF18 (a), BMP2 (b), PAXIP1 (c), and TMC7 (d). Cells were depleted of GNL2 or SEC13 for 5 days. Cells were mock-treated as a negative control, or depleted of either UPF1 or UPF2 as positive controls. Gene expression levels were normalised to the geometric mean of two housekeeping genes (POLR2J and ACTB), and are displayed relative to the mock control. Error bars represent the standard error of the mean of at least 3 biological replicates. * = p-value <0.05, ** = p<0.01, *** = p<0.001, Mann-Whitney test. The RNAs analysed in this experiment were also analysed for the NMD assay reported in Figure 4.1.

4.2.4 GNL2 and SEC13 take part in the NMD regulatory feedback loop

It has been shown that transcripts encoding NMD factors are themselves regulated by other effectors of the pathway, thus forming a feed-back auto-regulatory loop (Huang et al., 2011, and Yepiskoposyan et al., 2011). When analysing transcriptome changes, specifically mRNAs that were up-regulated upon knock-down of a subset of NMD factors, Mühlemann and colleagues noticed an enrichment of the NMD pathway in the GO term analysis. RT-qPCR analyses confirmed that there was an up- regulation of mRNAs encoding UPF2, SMG1, SMG5, SMG6 and SMG7 upon UPF1 knock-down. When UPF2, UPF3B, SMG1, SMG6 and SMG7 were knocked down both the mRNA and the protein levels of UPF1 were increased. Sequence analyses showed that NMD-inducing features are present in the transcripts coding for NMD factors. Cloning of the long 3’ UTR of UPF1, SMG5 and SMG7 downstream of a reporter construct was able to induce NMD-mediated degradation the reporter. To further validate their findings Yepiskoposyan et al. (2011) performed RNA IP of FLAG-UPF1 and identified the mRNA coding for many NMD factors amongst the enriched UPF1-bound mRNAs (Yepiskoposyan et al., 2011). Wilkinson and colleagues used a similar approach and came to similar conclusions although with some differences, for example whilst the level of UPF3B were unaffected by UPF1 knock-down in the study of Mühlemann and colleagues, Wilkinson and colleagues showed a two-fold up-regulation of UPF3B mRNA upon UPF1 knock-down (Huang et al., 2011, and Yepiskoposyan et al., 2011).

Figure 4.5 - Knock-down of a subset of NMD factors up-regulates the mRNA level of other NMD factors Summary of three studies (Huang et al., 2011, Yepiskoposyan et al., 2011, and Longman et al., 2013) analysing the response of transcripts encoding NMD factors following knock- down of a subset of NMD factors. The transcripts up-regulated in all studies are displayed in green. The transcripts for which contrasting data have been reported are displayed in yellow. k.d. = knock-down.

To understand whether GNL2 and SEC13 take part in this regulatory feedback loop affecting the level of other NMD factors mRNA, the transcript levels of a selected number of known NMD factors were analysed by RT-qPCR. Cells were depleted of either GNL2 or SEC13, as described previously, and the transcript levels of UPF1, UPF2, SMG1 and SMG5 were assessed. Depletion of GNL2 led to a significant up- regulation of SMG5 mRNA and to increased levels of UPF1, SMG1 and to some extent of UPF2. When SEC13 was depleted the transcripts of UPF1 and UPF2 were significantly up-regulated and SMG5 levels were also increased (Figure 4.6). Taken together all these findings strongly suggested a role for GNL2 and SEC13 in the

NMD regulatory circuit. To further validate the observation that GNL2 and SEC13 may take part in the regulation of other NMD factors, it would be important to repeat the experiment and analyse the effect of depletion of GNL2 and SEC13 on the protein levels of UPF1, UPF2, SMG1 and SMG5, since antibodies against these factors are commercially available. It would be interesting to show whether a correlation between the changes at the mRNA level and changes at the protein level exists.

Figure 4.6 - Knock-down of GNL2 and SEC13 leads to an increase of the mRNA level of a subset of NMD factors. RT-qPCR for UPF1, UPF2, SMG1 and SMG5. Cells were depleted of either GNL2 (in green) or SEC13 (in purple) for 5 days. Cells were also mock-treated as a negative control. Gene expression levels from 8 independent experiments were normalised to the geometric mean of two housekeeping genes (POLR2J and ACTB), and are displayed as fold changes compared to the mock control. * = p<0.05, Student’s t test.

4.3 Characterisation of GNL2

Human GNL2 is a member of the MMR1-HSR1 family of GTPases. Its homologues in yeast, zebrafish and Arabidopsis thaliana have been shown to be involved in the biogenesis of the 60S ribosome subunit (Im et al., 2011, Essers et al., 2014, Matsuo et al., 2014). Human GNL2 has been shown to be required for ribosomal biogenesis in HeLa cells, but its specific role has not been characterised in details (Tafforeau et al., 2013).

Figure 4.7 Ribosome biogenesis pathway. Ribosomal RNA is synthesised in the nucleolus as polycistronic RNA. It undergoes sequential endonucleolytic and exonucleolytic cleavages in the nucleolus, nucleus and cytoplasm leading to the formation of mature ribosomal subunits. During the maturation of the rRNA molecules several proteins associate with the pre-rRNA. Association of GNL2 (in orange) with the pre-60S rRNA happens in the nucleus; GNL2 is thought to be released from the immature ribosomal subunit prior to the export to the cytoplasm. The GTPases LSG1 (in blue) and EFL1 (in green) associate with the pre-60S subunit in the cytoplasm and are released from it before it final maturation stages.

A series of human GNL2 deletion constructs, expressed as GFP fusions, was used to determine its subcellular localisation, as well as to assess its GTPase activity (Chennupati et al., 2011). Wild-type GNL2 localised predominantly in the nucleus and in the nucleolus. In order to assess the localisation of endogenous GNL2 in HeLa cells, I therefore elected to perform a nuclear/cytoplasmic fractionation, followed by western blot analysis. To ensure correct fractionation, the western blot membrane was also probed with antibodies against XRN2, used as a nuclear marker, and tubulin, used as a cytoplasmic marker. GNL2 could be identified in both the nucleus and the cytoplasm (Figure 4.8).

Figure 4.8 Endogenous GNL2 is localised both in the nucleus and in the cytoplasm. Nuclear and cytoplasmic fractionation of HeLa cells reveals that GNL2 is localised in both cellular compartments. Anti-XRN2 was used as a nuclear marker, whilst anti-tubulin was used as cytoplasmic marker. N: nucleus; C: cytoplasm.

To confirm the role of GNL2 in ribosome biogenesis, I analysed the polysome profiling of cells that were depleted of GNL2 (or treated with a scrambled siRNA pool) for four or five days. In the control it was possible to identify the peaks corresponding to the small ribosomal subunit (40S), the large ribosomal subunit (60S), the mature ribosome (80S), and a series of peaks corresponding to the

95 polysomes (Figure 4.9, panels a and c). When GNL2 was depleted, either for four or five days, there was a drastic change in the profile. The peak corresponding to the 40S subunit was still visible, and looked mainly unaffected, but the peak corresponding to the 60S subunit was almost completely abolished. The peak corresponding to the mature ribosome was drastically reduced but still present. No peaks could be detected for the polysomes (Figure 4.9, panels b and d). This is consistent with defects in 60S maturation.

Figure 4.9 Knock-down of GNL2 leads to a drastic reduction of mature 60S ribosomal subunits and of mature ribosomes. HeLa cells were treated with either a scrambled siRNA pool or a siRNA pool against GNL2 for 4 days (a and b, respectively) and for 5 days (c and d, respectively). Cell extracts were loaded on a sucrose gradient and the fractionation profiles were analysed at 254 nm. The peaks shown correspond to the small (40S) and large (60S) subunits of the ribosome, the mature ribosomes (80S) and polysomes.

As NMD is a translation-dependent process, it could be argued that the up-regulation of NMD targets observed upon GNL2 depletion (described in Chapter 4.2) is due to translation inhibition rather than being a direct effect of GNL2 knock-down. However, the incorporation of radio-labelled methionine was unaffected in GNL2 depleted cells compared to cells treated with a scrambled siRNA pool (Figure 4.10; this figure was adapted from Casadio et al., 2015, and the experiment was performed by Dr. Nele Hug), suggesting that translation is still active in GNL2 depleted cells. To compare the samples equal amounts of proteins from cell lysates were loaded on a gel. Before lysis the cells were treated with cycloheximide to inhibit translation and then deprived of methionine for 20 minutes so that no de-novo protein production could occur. Cells were then provided with radio-labelled methionine and cysteine for 1 hour, prior to protein extraction. Were GNL2-depleted cells unable to translate, no incorporation of radio-labelled amino acids would have been recorded, and only unlabelled proteins would have been loaded on the gel. It is therefore likely that GNL2 knock-down does not critically impair translation.

Figure 4.10 Depletion of GNL2 does not affect the incorporation of radio-labelled methionine. HeLa cells were depleted of GNL2, treated with a scrambled siRNA pool or mock-treated. Cells were treated with cycloheximide to inhibit translation, then released from inhibition and supplied with radio-labelled methionine-containing medium. Cells were lysed and equal amount of protein were loaded on a gel. The incorporation of methionine was then measured using autoradiography. The experiment was carried out by Dr. Nele Hug,

In order to investigate whether the depletion of GNL2 affects the processing of rRNAs, I depleted HeLa cells of GNL2 and compared their total RNA profile to that of a scrambled control by running the samples on a denaturing gel. The scrambled control sample showed two bands corresponding to the 28S and 18S rRNAs, whereas the GNL2-depleted sample was also enriched for a precursor of the 60S ribosomal subunit, the 32S rRNA (Figure 4.11). This accumulation could be indicative of ribosomal biogenesis defects.

Figure 4.11 Depletion of GNL2 leads to accumulation of 32S ribosomal RNA. RNA extracted from cells treated with a scrambled siRNA pool or a siRNA pool against GNL2 was run on a denaturing gel. The red arrows indicate the accumulation of one of the pre-60S ribosomal RNA intermediates, 32S rRNA.

These observations confirmed that human GNL2, like its homologues in yeast, zebrafish and A. thaliana, is required for ribosomal biogenesis, and that GNL2 is important for maturation of the 60S subunit of the ribosome.

The accumulation of the precursor 32S rRNA does not prevent complete maturation of the pre-60S ribosomal subunit, as indicated by the presence of a band corresponding to the 28S rRNA (Figure 4.11), suggesting that some level of translation is still on-going in the cells. The result of this experiment is in line with the observation that methionine incorporation is not affected in GNL2-depleted cells (Figure 4.10). The polysomes profiling data are in contrast with the above considerations showing an extreme phenotype with translation likely to be severely impaired when cells are depleted of GNL2 (Figure 4.9, panels b and c). It is not clear why the polysomes profiling phenotype is so severe compared to the other experiment, but since the outcomes of the other two experiments are milder and in line with each other, we are more inclined to believe that the polysomes profiling is an outlier. Furthermore, the GFP-visual screen in C. elegans in which GNL2 was identified requires translation of the GFP reporter upon its stabilisation by NMD inhibition. Altogether, this suggests that the effect of NMD inhibition upon GNL2 depletion is not due to a general effect on translation inhibition.

4.3.1 Is there a link between the ribosome biogenesis pathway and NMD?

We next sought to answer whether the two functions of GNL2 are separated, or whether there is a link between the ribosome biogenesis pathway and NMD. In order to test whether such link exists, we made use of the two HeLa cell lines carrying either the wild-type β-globin reporter or the NS39 β-globin reporter described above. Knock-down of three genes encoding proteins involved in the ribosomal pathway was performed in these cell lines. Two of these proteins are small GTPases, EFL1 (also known as EFTUD1) and LSG1, and they are both required for ribosomal biogenesis in HeLa cells (Figure 4.7; Tafforeau et al., 2013). Their yeast

99 homologues, Efl1p and Lsg1p, take part in the maturation of the pre-60S ribosome downstream of GNL2 (Senger et al., 2001, Kallstrom et al., 2003). LSG1, like GNL2, is a member of the MMR1-HSR1 GTPases family. The third protein, SBDB, is a cofactor for EFL1 (Gijsbers et al., 2013) (the siRNA pools for the silencing of the genes coding for EFL1, LSG1 and SBSD were provided by Andrew Finch, IGMM – Cancer Research UK).

100

Figure 4.12 - Knock-down of LSG1 leads to specific up-regulation of the NMD reporter mRNA. RT-qPCR for β-globin in HeLa cells carrying a wild-type (wt) β-globin reporter (a) or a β- globin transgene with a PTC at codon 39 (NS39) (b). Cells were depleted of EFL1, LSG1 or SDBS for 5 days. Cells were also mock-treated or treated with a non-targeting siRNA pool (scrambled) as negative controls. Cells were depleted of UPF2 or GNL2 as positive controls. (c) RT-qPCR for the depleted genes. Gene expression levels were normalised to the geometric mean of two housekeeping genes (POLR2J and ACTB), and are displayed relative to the mock control. Error bars represent the standard error of the mean of at least 3 biological replicates. * = p-value <0.05, Mann-Whitney test.

Cells were individually depleted of each of these proteins for five days. As negative controls, the cells were also mock-depleted or treated with a scrambled siRNA pool; as positive controls, cells were depleted of either UPF2 or GNL2. The expression of the factors was reduced of ~20%, as judged by the level of residual transcript measured by RT-qPCR (Figure 4.12, panel c). Depletion of LSG1 specifically up-regulated the NS39 transcript, but not the wild- type β-globin, whilst depletion of either EFL1 or SBDS had no effect on both the wild-type and NS39 reporters (Figure 4.12, panel b and a respectively). This suggests that LSG1 is a putative NMD factor, and that there may be a link between the ribosome biogenesis pathway and NMD.

4.4 Characterisation of the role of SEC13 in NMD SEC13 is a component of the nuclear pore complex and also a component of the outer cage of the COPII-coat vesicles, which transport cargos from the endoplasmic reticulum to the Golgi apparatus. In the nuclear pore complex, SEC13 is part of a sub-complex forming the outer rings. In vertebrates, nine other proteins are components of this sub-complex (Figure 4.13); one in particular, SEH1L, is conserved throughout evolution and shares the same structure as SEC13 (reviewed by Grossman et al., 2012).

101

Figure 4.13 Nuclear pore complex. The nuclear pore complex is formed by sub-complexes which contain different proteins. In yeast the outer ring Nups is a sub-complex containing seven proteins, including Sec13p and Seh1p. The outer ring Nups face both the cytoplasm and the nucleoplasm.

In the COPII-coat vesicles, SEC13 forms an heterotetramer with SEC31, thus forming the outer cage of the vesicles (Figure 4.14; Stagg et al., 2006). SEC13 also binds to SEC16 at the budding sites (Figure 4.14; Whittle & Schwartz, 2009). In human, there are two paralogues of SEC31: SEC31A and SEC31B. SEC31A is the

102 most abundant form and it is most widely expressed (data obtained from genecards.org).

103

Figure 4.14 Formation of COPII coat vesicles. SEC16 and SEC13 form a heterotetramer which binds to the transmembrane protein SEC12. GTP-bound SAR1 also binds to SEC12. SAR1-GTP interacts with the SEC23-24 dimer which is responsible for cargo selection and loading into the nascent vesicle. Several cargo molecules assemble in proximity and the increased protein concentration leads to the formation of a vesicle bud. SEC13-31 heterotetramers are recruited to form the outer cage of the vesicle and the vesicle is fully formed. COPII coat vesicles transport cargo molecules from ER to Golgi.

To dissect whether these two functions of SEC13 are required for NMD, I depleted cells of either SEH1L or SEC31A and SEC31B (alone or in combination). As negative controls, cells were mock-depleted or treated with a scrambled siRNA pool. As positive controls, cells were depleted of either UPF2 or SEC13. Depletions were performed over five days; expression of UPF2 and SEH1L was reduced of 80-90% compared to mock, whilst depletions of SEC13, SEC31A and SEC31B were not very efficient, their expression being reduced of 20-50%, as judged by the level of residual transcript measured by RT-qPCR (Figure 4.15, panel c). Depletions of SEC13 partners in both the nuclear pore complex and in the COPII coat vesicle complex did not lead to the up-regulation of the wild-type β-globin transcript or of the PTC-containing reporter mRNA (Figure 4.15, panel a and b). Depletions of SEC31A and SEC31B were not efficient. It is therefore possible that cells were still able to produce an amount of these proteins sufficient to mask any loss-of-function phenotype.

104

Figure 4.15 - Knock-down of the NPC component SEH1L or of the COPII coat vesicle components SEC31A and SEC31B does not lead to specific up-regulation of the NMD reporter mRNA. RT-qPCR for β-globin in HeLa cells carrying a wild-type (wt) β-globin reporter (a) or a β- globin transgene with a PTC at codon 39 (NS39) (b). Cells were depleted of SEH1L,

105

SEC31A or SEC31B (alone or in combination) for 5 days. Cells were also mock-treated or treated with a non-targeting siRNA pool (scrambled) as negative controls. Cells were depleted of UPF2 or SEC13 as positive controls. (c) RT-qPCR for the depleted genes. Gene expression levels were normalised to the geometric mean of two housekeeping genes (POLR2J and ACTB), and are displayed relative to the mock control. Error bars represent the standard error of the mean of at least 3 biological replicates. * = p-value <0.05, Mann- Whitney test.

4.5 Discussion

The C. elegans RNAi screen described in this thesis led to the identification of five novel putative NMD factors: NGP-1, NPP-20, AEX-6, PBS-2 and NOAH-2. NOAH- 2 is absent from vertebrates but is conserved in D. melanogaster, and our collaborators have shown that its homologue NOMPA is involved in NMD (Casadio et al., 2015). Of the remaining four factors, only two (NGP-1/GNL2 and NPP- 20/SEC13) showed functional conservation in human NMD.

Depletion of one of the human homologues of PBS-2, PSMB7 (alone or in combination with PSMB10), led to up-regulation of both the wild-type and NS39 β- globin transcripts in our reporter cell lines. Cells depleted of PSMB7, but not PSMB10, died after five days of depletion, whilst they were still viable after four days. It is possible that cells depleted of PSMB10, a core component of the immunoproteasome, are viable because the expression levels of this protein are low in HeLa cells (data obtained from the Human Protein Atlas, Uhlén et al., 2015). Depletion of PSMB10 in a cell type with an active immunoproteasome could lead to the same unspecific up-regulation of the two β-globin transcripts observed in HeLa cells depleted of PSMB7. Depletion of the human homologues of AEX-6, RAB27A and RAB27B (alone or in combination) had no detectable effect on the two β-globin transcripts. A possible explanation for the lack of any detectable effects could be that incomplete knock-down resulted in residual RAB27A and RAB27B protein levels sufficient to maintain the NMD response. Unfortunately the anti-RAB27A and anti-RAB27B antibodies I tested did not work (data not shown) and I was therefore unable to determine the protein levels after five days of siRNA treatment. Furthermore, there are 65 members of the RAB family in the human genome (reviewed by Rojas et al.,

106

2012), thus it is possible that I could not observe an effect in the NMD assay due to redundancy. RAB27 is specifically involved in the exocytosis of dense-core vesicles (Fukuda et al., 2006) and it is also involved in the docking of vesicles in specific cell types, such as melanosomes (Klomp et al., 2007). It is therefore possible that RAB27 could be involved in NMD in a cell type-specific manner.

GNL2 is a GTPase belonging to the MMR1-HSR1 subfamily and comprises the permutated G-motifs (G5-G4-G1-G2-G3) typical of this subfamily. Chennupati et al. (2011) created a series of G-motifs mutants to analyse whether the GTPase activity observed in other species is retained by human GNL2. They found that both the G5 and G4 motifs are important for the binding of GTP. In a study conducted in plants, Im et al. (2011) showed that the Oryza sativa and the Arabidopsis thaliana homologues of GNL2 can hydrolyse GTP in vitro, and that this activity increases in the presence of the 60S ribosomal subunit. The ability of human GNL2 to bind and (potentially) hydrolyse GTP could be important for its function in ribosome biogenesis, but could also be important for NMD. It would therefore be interesting to create a series of mutant GNL2 proteins with point mutations in the G-motifs to investigate whether the mutations may affect NMD.

Depletion of another GTPase of the MMR1_HSR1 family, LSG1, led to up- regulation of the transcript of the PTC-containing reporter, suggesting that LSG1 itself, or ribosome biogenesis more generally, plays a role in NMD. LSG1 has been reported to be involved in the late steps of the 60S maturation and to be prevalently localised in the cytoplasm (Reynaud et al., 2005). This finding suggests that there could be a functional link between the ribosome maturation process and NMD, that GTPases of the MMR1_HSR1 family could play an energy provision role in NMD through the hydrolysis of GTP, and/or that GNL2 and LSG1 could have partial redundancy. Additional controls need to be performed to further validate a link between LSG1, ribosomal biogenesis and NMD. First of all, the protein levels should be assessed after depletion to determine the efficiency of the siRNA treatment. The effect of

107

LSG1 knock-down on the maturation of the pre-60S ribosomal subunit could also be tested by performing polysomes profiling and by analysing the potential accumulation of pre-60S intermediate on a denaturing gel. Should pre-60S accumulation occur without NMD being affected, it would imply that knock-down is being achieved to sufficient levels to impair LSG1 function, and that LSG1 is unlikely to play a role in NMD. Should both processes be affected, further characterisation of LSG1 would be required. Should only NMD be affected, it would suggest that different levels of LSG1 are required for the two processes.

The mechanism through which GNL2 is active in the NMD pathway needs to be investigated. Finch et al. (2011) proposed a model for the release of eIF6 from the pre-60S ribosomal subunit. In their model, hydrolysis of GTP by the action of the GTPase EFL1 led to a conformational change in SBDS, that ultimately led to the dissociation of eIF6 from the ribosomal subunit (Finch et al., 2011). It is possible that GNL2 could act in a similar manner to facilitate the release of stalled ribosomes from premature termination codons. This event is less efficient that the release from normal termination codons (reviewed in Celik et al., 2015); the GTP-hydrolysing activity of GNL2 could provide the required additional energy.

SEC13 is a small protein of 37 kDa which comprises six WD40 protein interaction domains, and is a component of the nuclear pore complex (NPC) (Siniossoglou et al., 1996). In vertebrates SEC13, together with eight other proteins (the outer ring Nups), can be found on both sides of the nuclear membrane (Lutzmann et al., 2002). In order to investigate whether this function of SEC13 is important for NMD, I depleted cells of another Nup protein found in the outer ring, SEH1L. Like SEC13, SEH1L is highly conserved throughout evolution and folds as a β-propeller. I could achieve 90% knock-down of SEH1L, but observed no effects on the NMD reporter transcript. This suggests that SEH1L is not involved in NMD. This in turns signifies that the nuclear pore complex function of SEC13 may not be required for its role in NMD. Alternatively, it is possible that other components of the complex are involved

108 in NMD. Knock-down of several nuclear pore complex components would allow to discriminate between these two possibilities.

SEC13 is also a component of the COPII coat vesicles. These vesicles transport cargos (secreted proteins and lipids) from the endoplasmic reticulum to the Golgi apparatus. At the budding sites of the vesicles, SEC23 and SEC24 form a complex. This complex is important for the interaction with the two proteins that form the outer cage of the vesicles: SEC13 and SEC31 (Stagg et al., 2006, Stagg et al., 2008). There are four SEC24 proteins (A-D) that have different binding affinity for different cargos (Tang et al., 1999, Mancias & Goldberg, 2008). The energy that allows the vesicles to form at the budding sites and to fuse with the membrane at the acceptor sites is provided by hydrolysis of GTP by SAR1, with the help of SEC23 (Bi et al., 2002). Yeast only has one homologue of each protein involved in the process, whilst in human there are multiple paralogues of each protein (with the exception of SAR1 and SEC13).

To test whether the COPII coat function of SEC13 is important for NMD, the other component of the outer coat of the vesicles, SEC31, was depleted from HeLa cells. The depletion level achieved for SEC31A/B was poor: a minimum of 50% of the mRNA coding for SEC31 was still present in the cells. It is possible that expression of SEC31A or SEC31B is necessary for cell viability, or it may be that the siRNA pools targeting these two genes are less efficient than siRNA pools targeting other genes. It is therefore difficult to assess whether SEC31A and SEC31B have an effect on NMD.

Additional controls should have included assaying cells depleted of SEHIL or SEC31 for functionality of the nuclear pore complex and/or of the vesicular transport to evaluate whether knock-down of the factors affected those functions. Should NPC/COPII coat function, but not NMD, be affected, it would imply that knock- down is being achieved to sufficient levels to disrupt the roles of SEH1L and SEC31 paralogues, and that these factors are unlikely to be involved in NMD. COPII coat assembly and transport of the cargo from ER to Golgi could be assayed making use of confocal microscopy. To determine whether the nuclear pore is still functional, nuclear and cytoplasmic fractions could have been separated and the mRNA

109 population in both compartments could have been analysed by RTq-PCR upon knock-down of NPC components and in controls.

The specific role of SEC13 in the NMD pathway is still uncharacterised. In a recent study, Izaurralde and colleagues performed mass-spectrometry analysis on a C-terminal domain, termed PC domain, of the core NMD factor SMG7. They identified SEC13 in the list of proteins co-purifying with the PC domain (Loh et al., 2013). SMG7 forms a complex with SMG5, and these two proteins are required in the last step of NMD to trigger decapping of the mRNA, thus leading to the degradation of the substrate transcript (Unterholzner & Izaurralde, 2004). The interaction between SEC13 and the PC domain of SMG7 suggests that SEC13 may be involved in the last step of the NMD pathway. To assess whether this is the case, it would be informative to perform immunoprecipitation of SEC13 and analyse whether it interacts with SMG5, SMG6, SMG7, and/or the hyperphosphorylated form of UPF1 which is found at this stage of NMD.

110

Chapter 5: Proteomics and transcriptomics analyses of GNL2 and SEC13

5.1 Introduction

The role of the novel NMD factors GNL2 and SEC13 in the surveillance pathway is still unclear. In order to shed light on the function of these proteins in NMD, I decided to make use of two unbiased approaches: a proteomics and transcriptomics approach. Firstly, I performed a mass spectrometry analysis of the proteins co- immunoprecipitating with GNL2 and SEC13 to determine their interacting partners. The mass spectrometry data presented below are preliminary, as the experiment was only carried out once without any technical or biological replicates. Secondly, I analysed what proportion of the transcriptome is regulated by GNL2 and SEC13 by means of microarrays. I compared the effects of the depletion of GNL2 and SEC13 with the effects of the depletion of two core NMD factors, UPF1 and UPF2, in order to assess which transcripts are co-regulated by these NMD factors.

5.2 Mass spectrometry analysis of GNL2 and SEC13 binding partners

Antibodies against human GNL2 and SEC13 are commercially available. Although these antibodies can detect the endogenous proteins in Western Blots, they are not able to efficiently purify GNL2 or SEC13 from cell lysates (data not shown). Due to this technical issue, I elected to transfect HeLa cells with tagged versions of GNL2 and SEC13 to perform immunoprecipitations and mass spectrometry analyses.

5.2.1 Efficient immunoprecipitation of T7-tagged GNL2 and HA-tagged SEC13

HeLa cells were transfected with T7-GNL2 or HA-SEC13 and the appropriate empty vector controls. Cells were harvested 48 hours post transfection and the proteins were extracted as described in the Materials and Methods chapter (paragraph 2.2.15). The

111 protein extracts were either incubated with anti-T7 coupled beads or with anti-HA coupled beads. The proteins bound to the beads were then analysed on a western blot gel to determine whether efficient immunoprecipitation had been achieved. Both T7-GNL2 and HA-SEC13 could be immunoprecipitated efficiently (Figure 5.1, panels a and b, respectively). The western blot membranes were also probed with an anti-tubulin antibody to ensure that the gel was loaded with equal amount of protein.

Figure 5.1 Immunoprecipitation of T7-GNL2 and HA-SEC13. Western blot showing (a) T7 immunoprecipitation and (b) HA immunoprecipitation. The membranes were probed with anti-GNL2 and anti-SEC13 antibodies, respectively. To ensure that equal amount of sample and control were loaded, the membranes were also probed with anti-tubulin antibody. EV: empty vector; IP: immunoprecipitation.

The samples for mass spectrometry were prepared in the same manner. After incubation of the protein extracts with the relevant beads, the beads were washed to remove residual unbound proteins. The samples were then processed and analysed by Dr. Jimi Wills (IGMM – Cancer Research UK).

5.2.2 GNL2 mass spectrometry analysis

67 proteins were enriched in the T7-GNL2 immunoprecipitated sample compared to the control (Figure 5.2). GNL2 (Q5T0F3) was identified as one of the proteins with the highest amino acidic sequence coverage (65%), indicating efficient GNL2 immunoprecipitation.

112

Figure 5.2 Heat map of enriched proteins bound to T7-GNL2 and T7 empty vector identified by mass spectrometry analysis of T7 immunoprecipitations. EV: empty vector

113

35 of the enriched proteins are ribosomal proteins, belonging to three different categories (40S, 60S and mitochondrial ribosomal proteins), accounting for ~52% of all enriched proteins (Table 5.1). These results are in line with a role for GNL2 in ribosomal biogenesis.

Table 5.1 Ribosomal proteins co-immunoprecipitating with T7-GNL2. The ribosomal proteins belong to three different categories; for each category the number of proteins identified and their names are reported.

9 proteins co-immunoprecipitating with GNL2 (listed in Table 5.2), as well as GNL2 itself, were previously assayed for a role in ribosomal biogenesis. Five (GNL2, NPM1, RSL1D1, RPS11 and FBL) were found to be required for correct processing of ribosomal intermediates (Tafforeau et al., 2013), and a sixth protein, DDX21, was subsequently reported to play a role in transcription and in ribosomal biogenesis by binding to the pre-rRNA (Calo et al., 2015). It is therefore possible that the four GNL2-co-immunoprecipitating proteins assayed by Tafforeau et al. (2013) (GNL3, DHX9, RBM28 and SRP14) also play a role in ribosomal biogenesis, but are not essential proteins for this process. These findings are consistent with a role for GNL2 in ribosomal biogenesis.

114

Table 5.2 Proteins co-immunoprecipitating with T7-GNL2 with a putative or a known function is ribosomal biogenesis. GNL2-co-immunoprecipitating nucleolar proteins assayed by Tafforeau et al. (2013) for a role in ribosomal biogenesis.

Amongst the other GNL2 interactors (listed in Table 5.3), there are two proteins involved in splicing, SRSF3 and SRSF7, and an export factor, ALYREF. SRSF3, SRSF7 and ALYREF are all RNA-binding proteins (Baltz et al., 2012, Castello et al., 2012); it is therefore possible that these interactions are mediated by mRNA. To verify whether this is the case, protein lysates should be incubated with or without RNase A. Following immunoprecipitation of GNL2, should the proteins of interest be undetectable in the RNase A-treated samples, this would signify that the interactions were RNA-mediated. Should they be detectable, on the other hand, this would imply that the interactions are either direct or mediated by other proteins.

115

Table 5.3 Other proteins co-immunoprecipitating with T7-GNL2. The table reports the UniProt accession identification, the gene symbol, the percentage of coverage of the protein sequence found in the mass spectrometry and a brief description.

116

5.2.3 SEC13 mass spectrometry analysis

20 proteins were enriched in the HA-SEC13 immunoprecipitated sample compared to the control (Figure 5.3).

Figure 5.3 Heat map of enriched proteins bound to HA-SEC13 and HA empty vector identified by mass spectrometry analysis of HA immunoprecipitations. EV: empty vector

117

The protein with the highest coverage (96%) of the amino acidic sequence is SEC13 (Table 5.4). The COPII coat vesicle component SEC16A was amongst the proteins co- immunoprecipitating with SEC13. This observation is in accordance with previously published data, which reported an interaction between the two proteins (Hughes et al., 2009, Whittle & Schwartz, 2010).

Several ribosomal proteins were also identified. Ribosomal proteins are common mass spectrometry contaminants, and could have been identified for this reason. In alternative, it is possible that these specific factors could associate with the nuclear pore, of which SEC13 is a component.

Table 5.4 Proteins co-immunoprecipitating with HA-SEC13. The table reports the UniProt accession identification, the gene symbol, the percentage of coverage of the protein sequence found in the mass spectrometry and a brief description.

118

5.3 Microarray analysis of HeLa NS39 transcriptome

Nonsense-mediate mRNA decay was first identified as a mechanism to degrade transcripts harbouring a premature stop codon, thus avoiding the production of truncated proteins that could have deleterious effects (Leeds et al., 1991). Since then, NMD has been associated with the regulation of a larger proportion of the transcriptome, between 3 and 10%. This additional function of NMD is conserved in all the species analysed: in S. cerevisiae (Lelivelt & Culbertson, 1999, He et al., 2003), in C. elegans (Longman et al., 2013), in D. melanogaster (Rehwinkel et al., 2005, Chapin et al., 2014), in D. rerio (Longman et al., 2013), in A. thaliana (Drechsel et al., 2013), and in H. sapiens (Mendell et al., 2004, Wittmann et al., 2006, Yepiskoposyan et al., 2011, Longman et al., 2013).

5.3.1 Sample preparation

In order to assess what proportion of transcripts is affected by the depletion of either GNL2 or SEC13, I analysed the transcriptome of the NS39 HeLa cells upon depletions of those two factors. Cells were also treated with a scrambled siRNA pool or mock depleted as negative controls. To elucidate whether the changes observed were indeed due to NMD, cells were also depleted of two well characterised NMD factors: UPF1 and UPF2. For each condition three biological replicates were prepared.

UPF1 mRNA was knocked down to ~50% of its expression compared to mock, whilst UPF2, GNL2 and SEC13 mRNAs were knocked down to ~10%-20% of their expression compared to mock (Figure 5.4).

119

Figure 5.4 Depletion of UPF1, UPF2, GNL2 and SEC13. RT-qPCR for UPF1 (a), UPF2 (b), GNL2 (c), and SEC13 (d). The mRNA expression level of each gene was analysed in mock depleted cells or in cells treated with a scrambled siRNA pool, as well as in cells depleted of the relevant factor. Gene expression levels were normalised to the geometric mean of two housekeeping genes (POLR2J and ACTB), and are displayed relative to the mock control. Error bars represent the standard error of the mean of 3 biological replicates.

The transcriptome of HeLa NS39 was analysed using the Agilent SurePrint G3 Human Gene Expression 8x60k microarrays. Total RNA extracted from cells treated with siRNA pools against UPF1, UPF2, GNL2, and SEC13, as well as with a scrambled siRNA pool, was then processed by OGT – Oxford Gene Technology.

5.3.2 Microarray data analysis

The microarray data for UPF1, UPF2, GNL2 and SEC13 were normalised to the scrambled control and a gene list containing information about fold change and statistical significance was produced by Graeme Grimes from the in-house bioinformatics services.

120

For each samples, I analysed changes in gene expression that were statistically significant (with adjusted p-value < 0.05). All genes which fell into this category had a minimum fold-change value of 1.5 relative to the scrambled siRNA control. The total number of gene expressed was 29,106.

5.3.2.1 Global effects of NMD factor knock-down on the transcriptome

Knock-down of UPF1 led to changes in the expression levels of 3,242 genes, with ~63% of the transcripts being up-regulated. Depletion of UPF2 had a smaller effect on the transcriptome, with 129 up-regulated mRNAs out of the 136 mis-regulated transcripts (~95% of the transcripts were up-regulated). Knock-down of GNL2 led to similar changes to UPF2 with regards to the number of transcript with altered expression, but most of the mRNAs were down-regulated, with only 43 transcripts being up-regulated (~27%). Depletion of SEC13 had the smallest effect on the transcriptome, with only 65 mis-regulated mRNAs, but ~69% of the transcripts were up-regulated (Table 5.5). The amplitude of the changes observed could be due to differential knock-down efficiency at the protein level, with residual proteins being sufficient to partially maintain their functions.

Table 5.5 Summary of mis-regulated transcripts upon knock-down of UPF1, UPF2, GNL2 and SEC13 in HeLa NS39.

5.3.2.2 Transcripts co-regulated by the NMD factors

I then proceeded to ask whether a subset of the transcripts was co-regulated by the four NMD factors. I compared the list of significantly up-regulated transcripts upon knock-down of each NMD factor, in order to determine the extent of the overlap between the targets of each protein (Figure 5.5).

121

Figure 5.5 UPF1, UPF2, GNL2, and SEC13 co-regulate a subset of transcripts. (a) Venn’s diagram showing the overlap between the up-regulated targets upon knock-down of the four NMD targets. Diagram generated with Venny (http://bioinfogp.cnb.csic.es/tools/venny/). (b) Table showing the percentage of overlap between UPF2 and UPF1, GNL2 and UPF1, and SEC13 and UPF1. * = p-value <10-16, Fisher’s exact test.

A proportion of transcripts were up-regulated only when individual factors were depleted, suggesting that the regulation of these targets may be NMD-independent.

All targets that were not exclusively regulated by UPF2, GNL2 or SEC13 were co- regulated by UPF1. This suggests that these genes are regulated by NMD.

Interestingly, GNL2, SEC13 and UPF1 co-regulated the expression of 5 transcripts independently of UPF2. This suggests that the regulation of these five transcripts could be via the UPF2-independent branch of NMD.

Three genes were regulated by all four factors: HBD, DDIT3, and GDF15.

122

DDIT3 had been previously reported to be up-regulated upon knock-out of Upf2 in mice bone marrow-derived microphages and thymocytes (Weischenfeldt et al., 2008). The transcriptome analysis was performed in the HeLa NS39 cell line, which carries the β-globin NS39 transgene, a well-characterised NMD target. It was therefore surprising that the β-globin gene (HBB) was not amongst the up-regulated transcripts, but that a closely related gene, δ-globin (HBD), was. The microarray chip contains two probes that bind HBB and one that binds HBD. Blast of the sequence of the first HBB probe identified the 3’ UTR of the β-globin gene as a hit. As the NS39 transgene does not include this region, the probe would not detect upregulation of this reporter transcript. Blast of the sequence of the second probe reported 100% identity between the probe sequence and an exon that is only present in the β-globin transcript HBB-004. As the NS39 transgene is based on the sequence of the β-globin transcript HBB-001, the probe would not detect upregulation of this reporter transcript. Both HBB probes present in the array are therefore unable to bind the NS39 transcript, which explains why HBB was not amongst the up-regulated targets.

Blast of the sequence of the probe corresponding to HBD gave two hits: the first corresponded to HBD (with 100% identity between the probe and the gene sequence) and the second to HBB (with 90% identity). HeLa cells do not express HBD (data obtained from the Human Protein Atlas, Uhlén et al., 2015). It is therefore likely that the mRNA binding to the HBD probe corresponds to the up-regulated β-globin NS39 transgene.

UPF2 and UPF1 had the highest overlap of up-regulated transcripts: ~91% of the UPF2 targets were also UPF1 targets. SEC13 and UPF1 co-regulated ~58% of SEC13 targets (26 out of 45), 65% of which was exclusively regulated by SEC13 and UPF1. Of these 26 targets, 4 were also regulated by UPF2 and 8 by GNL2. 23 GNL2 targets, corresponding to 53% of the up-regulated mRNAs, were also up- regulated upon UPF1 knock-down. Of these 23 transcripts, 11 were also co-regulated by UPF2 and 8 by SEC13 (Figure 5.5, panel b).

123

The percentages reported above are referred to the overlap of targets that are significantly up-regulated (p-value < 0.05), and may underestimate the number of common targets. Analysis of the behaviour of genes significantly up-regulated in SEC13-depleted cells within the UPF1 knock-down sample revealed that the overlap of co-regulated mRNAs is greater than 75%, if significance is not taken into account for the UPF1-depleted sample (Figure 5.6, panel a). Similarly, comparison between genes that are significantly up-regulated upon GNL2 knock-down and the same genes in the UPF1-depleted sample shows that the overlap is higher than 60% (Figure 5.6, panel b). Comparison of UPF2 targets with the same transcripts in the UPF1 sample showed an almost complete overlap (over 94%) (Figure 5.6, panel c).

Figure 5.6 Heatmaps displaying overlap of targets. (a) Heatmap comparing the expression levels of targets that are significantly up-regulated in SEC13-depleted cells with the levels of the same targets in UPF1-depleted cells. (b) Heatmap comparing the expression levels of targets that are significantly up-regulated in GNL2-depleted cells with the levels of the same targets in UPF1-depleted cells. (c) Heatmap comparing the expression levels of targets that are significantly up-regulated in UPF2- depleted cells with the levels of the same targets in UPF1-depleted cells. Expression levels are expressed as the base 2 logarithm of the fold change compared to the scrambled control. Blue indicates transcripts that are down-regulated, and yellow transcripts that are up-regulated. FC: fold change.

124

5.3.2.3 GNL2 and the SMD pathway

Amongst the 7 genes that were exclusively co-regulated by GNL2 and UPF1, there were two genes that have been reported to be targets of the SMD pathway: AMIGO2 and c-JUN (Gong et al., 2009, Kim et al., 2007).

NMD and SMD (Staufen-mediated mRNA decay) are two regulatory pathways that share three effectors: UPF1, SMG1, and PNRC2. To test whether GNL2 could also be shared between the two pathways, I analysed by RT-qPCR the effects of GNL2 depletion on three SMD substrates: AMIGO2, c-JUN and FLJ21870. As SMD is UPF1-dependent and UPF2-independent (Gong et al., 2009), I included knock-down of these two factors as positive and negative controls.

Depletion of UPF1 led to up-regulation of all three SMD targets. Knock-down of GNL2 led to the up-regulation of the transcripts encoding AMIGO2 and c-JUN, whilst the transcript of FLJ21870 was unaffected by GNL2 depletion (Figure 5.7). The probe for this transcript was absent from the microarray chip. Depletion of UPF2 should have no effect on the expression of SMD targets, as UPF2 is not involved in the pathway. As expected, the mRNA level of c-JUN was not altered following UPF2 knock-down. However, the transcript coding for AMIGO2 was ~3 fold up-regulated following UPF2 knock-down.

125

Figure 5.7 SMD targets. RT-qPCR for the SMD targets c-JUN, AMIGO2 and FLJ21870 in HeLa NS39 cells. Cells were depleted of UPF1, UPF2, and GNL2 for 5 days. Cells were also mock-treated or treated with a non-targeting siRNA pool (scrambled) as negative controls. Gene expression levels were normalised to the geometric mean of two housekeeping genes (POLR2J and ACTB), and are displayed relative to the mock control. Error bars represent the standard error of the mean of at 4 biological replicates.

Based on these observations it is not possible to determine whether GNL2 is one of the factors shared between NMD and SMD.

5.4 Discussion

5.4.1 Mass spectrometry analyses

The mass spectrometry data presented in this chapter are based on a single experiment carried out without any technical or biological replicates. In order to confirm its results, the experiment needs to be repeated, and the experimental design could be improved. Increasing the number of technical and biological replicates should reduce the “noise” caused by aspecific protein binding to the beads. Making use of tagged versions of GFP or other inert proteins may provide a more suitable negative control. Positive controls should also be included: immunoprecipitation of a known NMD factor should clarify whether it is possible to detect the formation of NMD complexes in our system; immunoprecipitation of components of the other protein complexes GNL2 and SEC13 are known to be involved in should provide the same readout for these other complexes. Several commercially available antibodies against either GNL2 or SEC13 were tested and found unable to efficiently precipitate the proteins of interest, leading to the decision to perform immunoprecipitations of transiently transfected tagged- GNL2 and SEC13. Transfection of tagged versions of GNL2 and SEC13 led to over- expression of these proteins (Figure 5.1). This could affect the stoichiometry of the protein complexes these factors are components of, and thus potentially result in the identification of spurious interactors at the expense of physiological interactors.

126

It would beneficial for the study of GNL2 and SEC13 interactors to generate stable cell lines expressing endogenous levels of the tagged version of the proteins. A simple approach would be to create a tetracycline-inducible cell line in which titration of tetracycline levels could be optimised to obtain near-endogenous levels of expression of the tagged proteins. A more elegant approach would be to make use of CRISPR/Cas technology to introduce a tag on one or both of the endogenous GNL2/SEC13 alleles. This should result in physiological levels of expression of tagged versions of the proteins, and therefore represent the ideal system to identify bona fide interactors. The mass spectrometry experiment, although very preliminary, revealed a number of potential interactors for GNL2 and SEC13, but no NMD factor was identified. It is possible that these interactions are transient and therefore difficult to capture. It would be interesting to repeat the mass spectrometry analysis in the presence of a highly transcribed NMD substrate, such as β-globin NS39, to try and capture the NMD complexes. Alternatively, it is possible that other functions of GNL2 and SEC13 involve higher affinity protein interactions, and that therefore only a small proportion of the two proteins is in complex with other NMD factors. To attempt to obviate to this, it would be interesting to perform nuclear/cytoplasmic fractionation of the cells, and to analyse the protein complexes formed by GNL2 and SEC13 in the two cellular compartments. This could allow an easier identification of potential interactions between GNL2 or SEC13 and known NMD factors.

The interactions between GNL2 or SEC13 and the proteins identified in the mass spectrometry analysis will need to be confirmed by western blot analysis. In the set of experiments we performed, we decided not to treat the protein extracts with RNase A, in order to identify the maximum number of interactors. To determine whether the interactions are mediated by RNA, protein samples will need to be incubated with or without RNase A prior to the binding with either anti-T7 or anti- HA beads. To verify whether the interactions are direct or mediated by other proteins, an in vitro binding assay could be performed. As a follow up, mutants and truncated versions of GNL2, SEC13 and any interactors of interest could then be developed to map the protein regions/domains important for binding. In addition to

127 this, interacting proteins could also be tested in the NMD assay to assess whether they are important for the NMD response. SEC13 could be immunoprecipitated, but, surprisingly, of all the components of the COPII coat vesicles or the NPC, only SEC16A could be detected. This, in conjunction with the high number of ribosomal proteins that co-immunoprecipitated with HA-SEC13, suggests low quality of the sample.

5.4.2 Microarray analyses

The microarray data show that GNL2 and SEC13, like other NMD factors, can regulate physiological transcripts as well as PTC-containing transcripts. Furthermore, the data show that the overlap of targets that are significantly up-regulated both in UPF2 and in UPF1 knock-down is higher than 90%, whereas only 50%-60% of GNL2 and SEC13 targets are shared with UPF1. This suggests that targets that are not co-regulated by UPF1 are regulated independently of NMD. The data will need to be validated in independent samples by RT-qPCR. It would also be interesting to repeat these experiments in different cell lines, to determine how conserved the targets of the various genes are in cells with different transcriptional profiles.

Upon knock-down of NMD factors, transcripts that are subject to NMD regulation are expected to be up-regulated. When UPF1, GNL2, SEC13 and UPF2 were silenced, many transcripts were also down-regulated. For the purpose of this thesis, these transcripts were over-looked. There are two possible explanations for these transcripts being down-regulated: the first is that changes in these mRNAs are due to alteration of the other functions of the four NMD factors tested. A second possibility is that the changes we see in these transcripts are due to indirect effects: for example, a transcript important for the maintenance of the stability of other mRNAs may be mis-regulated, consequently leading to the down-regulation of its targets. It is possible that the down-regulated transcripts arise from a combination of the two options.

128

The transcript levels of two SMD targets, AMIGO2 and c-JUN, were up-regulated when both UPF1 and GNL2 were depleted. It is possible that GNL2 could also play a role in the SMD pathway, like UPF1, SMG1 and PNRC2. The duplicitous effects of GNL2 knock-down on three SMD substrates (Figure 5.6) make it necessary to analyse a wider range of SMD targets to obtain a clear answer. It would be interesting to perform microarray analysis of cells treated with siRNA against STAU1 and STAU2 and to compare the transcriptome changes upon depletion of these two SMD factors and of GNL2.

In addition to transcriptome-wide analyses, tethering assays could be performed. Tethering of STAU1 down-stream of a termination codon was shown to induce decay of the transcript, and depletion of UPF1 restored the transcript levels (Kim et al., 2005). A similar approach could be used: GNL2 and STAU1 could be tethered to a transcript and the level of the transcript could be measured. Subsequently, cells could be treated with siRNAs against GNL2, when STAU1 is tethered to the transcript, and with siRNAs against STAU1, when GNL2 is tethered to the mRNA. RT-qPCR analysis of the transcript level would show whether GNL2 and STAU1 act together in the SMD pathway, and which protein binds first to the transcript to induce decay.

Five mRNAs were up-regulated upon depletion of GNL2, SEC13 and UPF1, but not when UPF2 was knocked-down, suggesting that the regulation of these substrates could be due to the UPF2-independent and UPF3-dependent branch of NMD. When validating these targets it would be interesting to include samples depleted of UPF3A and UPF3B, to confirm whether these NMD factors participate in the co-regulation of the five identified targets.

129

Chapter 6: Discussion

Genetic and RNAi screens have been valuable tools for the identification of novel NMD factors (Culbertson et al., 1980, Hodgkin et al., 1989, Leeds et al., 1992, Cali et al., 1999, Longman et al., 2007). In the genome-wide screen performed by Longman et al. (2007), they took advantage of an RNAi library, developed in the Ahringer laboratory, targeting 86% of the predicted genes in C. elegans. Longman et al. (2007) screened a transgenic strain of C. elegans, the PTCxi strain, which carries an NMD reporter (termed PTCx) integrated in the genome. Two novel NMD factors, SMGL-1 and SMGL-2, were identified.

A second C. elegans RNAi library, the ORF-RNAi library v.1.1 developed in the Vidal laboratory, targets 1,736 genes that were not silenced by the Ahringer library. In order to uncover novel NMD factors, I set out to perform a genome-wide RNAi screen using the same transgenic strain of C. elegans, the PTCxi strain, described in Longman et al. (2007) and the ORF-RNAi library v.1.1. I identified five factors that are putative NMD factors in nematodes: NGP-1, NPP-20, AEX-6, PBS-2 and NOAH-2. Knock-down of these factors led to stabilisation of the PTCx mRNA (Figure 3.4).

The smg genes were discovered in a genetic screen in C. elegans performed by Hodgkin et al. (1989). In this screen Hodgkin et al. (1989) identified mutations that suppress nonsense alleles of several C. elegans genes. These mutations also led to morphological defects in worm genitalia, hence why they were termed smg (suppressor with morphological effect on the genitalia). No lethality was associated with mutations in smg genes, suggesting that the NMD pathway is not essential for viability in C. elegans. In contrast with these observations, yet similarly to the phenotypes described by Longman et al. (2007) for smgl-1 and smgl-2 knock-down, ngp-1, npp-20, pbs-2 and noah-2 were found to be required for proper development of the worms (Figure 3.3). The severity of the phenotypes observed is likely due to functions of these genes other than their roles in NMD. This could well be the case, as these genes encode factors involved in several different processes, listed below.

130

Three of the factors identified in the screen as putative NMD factors were also identified as component of other processes in different screens. Their function, however, has not been characterised further. NPP-20 was identified as a component of the nuclear pore, whilst PBS-2 was found to be a component of the proteasome (Galy et al., 2003, Davy et al., 2001). NOAH-2 is required for moulting, but the molecular mechanism through which it acts has not been identified (Frand et al., 2005). No function has been described for NGP-1, other than its putative role in NMD presented in this thesis (Chapter 3). By sequence homology, NGP-1 is a putative GTPase involved in ribosomal biogenesis. Its homologues in yeast, zebrafish, plant and human have all been reported to play a role in the maturation of the large subunit of the ribosome, therefore it is likely that this function is conserved in nematodes as well (Saveanu et al., 2001, Im et al., 2011, Tafforeau et al., 2013, Essers et al., 2014, Matsuo et al., 2014). Knock-down of aex-6 led to a milder phenotype than those described above for ngp- 1, npp-20, pbs-2 and noah-2. The worms had deficient abdominal muscles, which resulted in constipation and in the inability of actively expel the embryos from the gonad. AEX-6 is a member of the RAB family, which, in C. elegans, comprises 31 members. AEX-6 has been implicated in the docking of dense core and synaptic vesicles (Mahoney et al., 2006, Feng et al., 2012).

Although the screen was only a tool towards the mean of identifying novel NMD factors in human, it would be interesting to definitively prove whether the factors identified in the screen are bona fide NMD factors in C. elegans. When developing the PTCxi transgenic strain, Longman et al. (2007) also generated a “wild-type” C. elegans transgenic strain. This strain carries a wild-type reporter that consists of the same fusion between GFP and LacZ as the PTCx reporter, but lacks the termination codon down-stream of the GFP coding sequence, thus becoming insensitive to NMD regulation. The wild-type worms express GFP (Longman et al., 2007). To confirm that NGP-1, NPP-20, AEX-6, PBS-2 and NOAH-2 are NMD factors in C. elegans, the effect of their depletion in “wild-type” worms should be evaluated by RT-qPCR. Furthermore, the effect of the depletion of individual factors on

131 endogenous NMD targets, such as rpl-7a and rpl-12 transcripts, should be assessed by RT-qPCR.

6.1 Functional conservation of the novel putative C. elegans factors in human and fruitfly

In agreement with a function in moulting, NOAH-2 is not conserved in vertebrates, but it has a homologue in D. melanogaster, NOMPA. Our collaborators, Raúl Vallejo Baier and Claudio Alonso (University of Sussex), showed that NOMPA is an NMD factor in Drosophila (Casadio et al., 2015).

In order to test whether the human homologues of the putative C. elegans NMD factors were functionally conserved in human NMD, I tested them in an NMD assay. The assay is performed in two stable HeLa cell lines carrying either a wild-type β- globin transgene or a β-globin transgene (NS39 β-globin) with a premature stop codon in position 39, a well characterised NMD substrate.

When tested in an NMD assay in HeLa cells, the human homologues of AEX-6 (RAB27A and RAB27B) failed to stabilise the mRNA of the NMD reporter. The human homologues of PBS-2, PSMB7 and PSMB10, also failed in the NMD assay. PSMB10 did not stabilise the NMD reporter mRNA, whilst PSMB7 up- regulated the transcript levels of both the wild-type and PTC-containing β-globin genes. As wild-type β-globin is not an NMD substrate, this upregulation is likely due to a mechanism alternative to NMD (Figure 4.1 and Figure 4.2). A possible explanation for this observation lies in the known function of PSMB7. PSMB7 is one of the catalytic subunit of the 20S proteasome (for a recent review, see Bhattacharyya et al., 2014). It is likely that knock-down of PSMB7 is sufficient to impair the function of the proteasome leading to accumulation of proteins, which in turn may lead to translational pausing and transcript accumulation. This would result in higher levels of both wild-type and NS39 β-globin mRNA. Knock-down of PSMB10 did not lead to the same unspecific up-regulation of both transcripts. PSMB10 is a component of the immunoproteasome (reviewed by Ferrington & Gregerson, 2012), and its expression level is low in HeLa cells (data

132 obtained from the Human Protein Atlas, Uhlén et al., 2015). This could explain why, unlike PSMB7, its depletion did not affect the expression of either β-globin transcript.

Both GNL2 and SEC13, the human homologues of NGP-1 and NPP-20 respectively, specifically stabilised the transcript levels of the PTC-containing β-globin transgene (Figure 4.1) and of endogenous NMD substrates (Figure 4.4), thus showing functional conservation in the NMD pathway. It would have also been interesting to test GNL2, SEC13, as well as both RAB27 paralogues, in an NMD assay using a different reporter for which the “50-55 nucleotide” rule (see paragraph 1.5.2) does not apply, such as the TCRB reporter (Wang et al., 2002). Whilst this would be a further confirmation of the functional conservation of GNL2 and SEC13 in NMD, already observed with the β-globin reporter and endogenous substrates, it would be a particularly useful experiment to assess the potential role in NMD of the RAB27 paralogues, for which the effect on endogenous substrates was not tested.

Another assay to confirm the involvement of GNL2 and SEC13 in the NMD pathway would be the tethering of the two proteins of interest to a reporter mRNA, to determine whether their binding leads to degradation of the reporter. Coupling of this experiment with depletions of known NMD factors would also allow to elucidate the hierarchical recruitment of NMD effectors to RNA substrates.

In an attempt to further characterise the role of GNL2 and SEC13 in the NMD pathway, I analysed the consequences of their depletion on the transcriptome, and compared them to the effects of depletions of UPF1 and UPF2. Knock-down of GNL2 and SEC13 led to up-regulation of 43 and 45 mRNAs respectively. The majority of these transcripts were also up-regulated when UPF1 was knocked-down. ~53% of the transcripts were common targets between GNL2 and UPF1. ~58% of the transcripts up-regulated upon SEC13 knock-down were also UPF1 targets (Figure 5.5). These observations are in line with a role for GNL2 and SEC13 in the NMD pathway.

133

Transcripts up-regulated upon GNL2 knock-down are either exclusively regulated by GNL2 or co-regulated by UPF1: all transcripts co-regulated by GNL2 and SEC13 or by GNL2 and UPF2 are also regulated by UPF1. The same observation is true for SEC13 (Figure 5.5). This suggests that the up-regulated transcripts are either regulated by NMD (the ones shared with UPF1) or regulated by alternative processes in which GNL2 and SEC13 are involved. Strikingly, not all the transcripts that are NMD targets are co-regulated by all four NMD factors. Whereas UPF1 seems to be required for regulation of all these targets, the requirement for GNL2, SEC13 and UPF2 seems to be transcript-specific. For example, 8 transcripts are regulated by UPF1, UPF2 and GNL2, but not by SEC13. This suggests that GNL2, SEC13 and UPF2 could be active in branches of the NMD pathway that do not require the other factors. It would be interesting to analyse the consequences of the depletion of other NMD factors on the transcriptome to determine the extent of the overlap with GNL2 and SEC13.

6.2 GNL2 function

In this thesis I have shown that GNL2 is required for a fully functional NMD pathway in C. elegans and in H. sapiens. In a previous study, human GNL2 was shown to also play a role in ribosomal biogenesis: its depletion leads to the accumulation of one of the precursor rRNAs (32S) of the 60S ribosomal subunit (Tafforeau et al., 2013). When I depleted cells of GNL2 and run total RNA on a denaturing gel, I could detect the accumulation of the 32S rRNA, recapitulating the result reported by Tafforeau et al. (2013) (Figure 4.11). Translation is required for functional NMD (Kervestin & Jacobson, 2012). If GNL2 knock-down leads to a reduction in the number of active ribosomes, it is likely this will impact NMD. It is possible that the levels of NMD factors may be reduced, which in turn might result in aberrant stoichiometry in NMD protein complexes, and in a decreased ability to recognise NMD target transcripts. It is also possible that following target recognition, reduced levels of deadenylation and decapping enzymes may fail to lead to the degradation of target mRNAs. Alternatively, it is possible that the levels of RNA Polymerase II are reduced so much that the mRNA content of the cells is decreased

134 to a further extent than the protein content. In this case, it is possible that NMD complexes may find it easier to bind and induce degradation of their targets than in normal cells. Further characterisation of the effects of impaired ribosome biogenesis is required to discriminate between these possibilities.

NMD and ribosomal biogenesis are processes with different cellular localisations, the former being a cytoplasmic surveillance pathway and the latter being mainly nucleolar/nuclear. Analysis of the localisation of endogenous GNL2 showed that it is found in both the nucleus and the cytoplasm (Figure 4.8). This raises two possibilities: GNL2 has two independent functions and, depending on its cellular localisation, it is involved in either ribosomal biogenesis or in NMD. A more intriguing alternative would be represented by the possibility that the two pathways may be interlinked. This hypothesis is supported by the observation that another GTPase involved in ribosomal biogenesis, LSG1, can specifically up-regulate the mRNA level of the β-globin NS39 transgene when tested in NMD reporter cells (Figure 4.9). The putative role of LSG1 in the NMD pathway has not been investigated further, but as it belongs to the same GTPase protein family as GNL2, it is possible that it carries out a similar function. It would be interesting to determine whether LSG1 is required for NMD in human cells. To do so, two approaches could be used. Firstly, the effects of LSG1 depletion on endogenous NMD targets and on the mRNAs encoding NMD factors could be analysed by RT-qPCR. With these analyses it would be possible to determine whether depletion of LSG1 can lead to increased stability of target mRNAs, and whether LSG1 takes part in the NMD regulatory feedback loop. Secondly, interactions between LSG1 and known NMD factors could be explored by mean of immunoprecipitations followed by western blot analyses.

What specific role GNL2 may play in the NMD pathway remains an open question. Interestingly, GNL2 was identified in two distinct studies as an RNA-binding protein, specifically poly(A)-containing RNA (Baltz et al., 2012, Castello et al., 2012). This is very interesting because it implies that GNL2 can interact with two different RNA species: ribosomal and messenger RNA. It is difficult to reconcile a role in NMD for a protein that is involved in rRNA maturation and which

135 predominantly localises in the nucleus. The findings that GNL2 is present both in the nucleus and in cytoplasm (Figure 4.8), and that it can bind mRNA (Baltz et al., 2012, Castello et al., 2012) are encouraging and make it easier to envisage a role for GNL2 in NMD; nevertheless, it is still premature to pinpoint the activity of GNL2 to a specific step of the NMD pathway. GNL2 could bind to mRNA in the proximity of a stalling ribosome and could either interact with the release factors and UPF1, or it could interact with the ribosome directly and bridge it to the mRNA. By doing so, and by interacting with UPF1, GNL2 could provide a platform for the assembly of the NMD complexes, thus leading to degradation of the transcript. It would be informative to perform CLIP experiment to determine which mRNAs are directly bound by GNL2 and whether GNL2 has a preferred consensus motif. This kind of analysis could give an insight on the transcripts that are regulated by GNL2 and could help identify features that might trigger the NMD response. Moreover, it would be interesting to perform CLIP of UPF1 in parallel to determine whether the two proteins bind in close proximity to the mRNA. Based on the data available for UPF1 CLIP in mouse embryonic stem cells, that position UPF1 in the coding sequence of most mRNAs (Hurt et al., 2013), an intriguing idea would be that GNL2 and UPF1 are present together on mRNAs and that, when the right conditions arise, they induce mRNA decay.

GNL2 can co-immunoprecipitate with UPF1 (Casadio et al., 2015). It is possible that upon recruitment of GNL2 to the stalled ribosome on the PTC it may interact with UPF1, the two release factors and the ribosome. By hydrolysing GTP, GNL2 may provide energy for the dissociation of the ribosomal subunits from the PTC and contribute to the recycling of the ribosome (Figure 6.1). This hypothesis stems from the observation that the release of the ribosome from a PTC is less efficient than its release from a normal termination codon (reviewed in Celik et al., 2015). The GTPase activity of GNL2 may therefore generate energy to facilitate this process.

136

Figure 6.1 Hypothetical role of GNL2 in the NMD pathway. UPF3 and UPF2 are bound to the EJC. The SURF complex (SMG1-UPF1-eRF1-eRF3) assembles on the ribosome stalling at the PTC. GNL2 is recruited to the ribosome. Interactions between the SURF complex and the EJC lead to the conformational changes in the two protein complexes and in the formation of the DECID complex. DHX34 is recruited to facilitate the transition from SURF to DECID complex. As a result of the conformational changes UPF1 is phosphorylated. By hydrolysing GTP, GNL2 could facilitate the release of the ribosomal subunits from the PTC. The release factors, SMG8, SMG9, GNL2 and DHX34, are released and additional proteins (SMG6, SMG5-7) are recruited. The targeted transcript is rapidly degraded.

Many underlying assumptions of this hypothesis are yet to be tested. Firstly, it has not been unequivocally demonstrated that human GNL2 can hydrolyse GTP. Chennupati et al. (2011) showed that two motifs of the GTP-binding domain are important for the interaction with GTP: G5 and G4 (for the domain composition

137 of GNL2 refer to Figure 3.11), but did not demonstrate GTPase activity. This activity has been demonstrated for the A. thaliana and O. sativa homologues of GNL2, but not for the human protein (Im et al., 2011). In order to verify the GTPase activity of GNL2, an in vitro GTPase assay could be performed. One such assay is based on the Escherichia coli phosphate-binding protein (PBP) fused to a fluorescent probe (MDCC). PBP can bind to the inorganic phosphate released after the hydrolysis of GTP to GDP. By binding inorganic phosphate, PBP induces a conformational change in MDCC, which results in an increased emission of fluorescence (for a detailed protocol see Shutes & Der, 2005). Secondly, it has not been demonstrated that GNL2 can bind to the SURF complex components eRF1 and eRF3 and to the ribosome. This hypothesis could by tested by immunoprecipitating GNL2 and testing for interactions with eRF1, eRF3 and ribosome components by western blot. To determine whether these interactions could be mediated by RNA, the protein extracts could be subject to RNase treatments prior to immunoprecipitation.

It would be important to determine the full range of GNL2 interactors involved in the NMD pathway. These could be revealed by means of immunoprecipitation of GNL2, followed by western blot analysis, probing for all known NMD factors. Although we detected interaction of GNL2 with UPF1 (Casadio et al. 2015, experiment performed by Dr. Dasa Longman), it is not known whether GNL2 has preferential binding affinity for a specific UPF1 phosphorylation status. SMG1-mediated phosphorylation of UPF1 is induced upon interaction of the SURF complex with the EJC complex (Kashima et al., 2006). Subsequent binding of SMG6 and SMG5-SMG7 to phosphorylated UPF1 results in the recruitment of PP2A, which then dephosphorylates UPF1 (Ohnishi et al., 2003). Antibodies that specifically bind to the phosphorylated isoform of UPF1 would allow to understand whether phosphorylated UPF1 interacts with GNL2. As an alternative approach, the interaction of GNL2 with UPF1 mutants could be tested. A number of UPF1 mutants which lack various protein domains exist, and could be used to pinpoint the regions important for GNL2 interaction. We have previously used this approach successfully in our laboratory to determine the nature of the interaction between DHX34 and UPF1 (Hug & Cáceres, 2014).

138

UPF1 was not detected in the mass spectrometry analysis of GNL2 interactors, but was identified by immunoprecipitation followed by western blot. This suggests that further NMD factors, which were not identified by mass spectrometry, may be identified as GNL2 interactors in immunoprecipitation followed by western blot analysis. The inability to detect UPF1 by mass spectrometry could be due to the potential transient nature of the interaction, which would make it hard to capture.

Chennupati et al. (2011) created a series of mutant and truncated GNL2 alleles to assess their cellular localisation and their GTP-binding abilities (Chennupati et al., 2011). The same mutants could be used to further characterise the role of GNL2 in NMD and in the ribosome biogenesis pathway. Cells could be depleted of endogenous GNL2 and subsequently transfected with siRNA-resistant plasmids coding for GNL2 mutants to determine whether they can rescue the NMD phenotype and/or the defects in the maturation of the pre-60S subunits. The same mutants could also be used to map the interaction of GNL2 with other components of the NMD pathway and of the ribosome biogenesis pathway. Furthermore, when performing the tethering assay, some of the mutants could be used to determine whether they are still able to induce decay of the reporter. This set of experiments would clarify which domains of GNL2 are important for its function in both NMD and ribosome biogenesis, and which domains are important for binding to other effectors involved in these pathways.

6.3 SEC13 function

SEC13 is a component of the nuclear pore complex and a component of COPII coat vesicles (Siniossoglou et al., 1996, Stagg et al., 2006). To determine whether any of these two functions is important for the role of SEC13 in the NMD pathway, I depleted components of the NPC (SEH1L) and of COPII coat vesicles (SEC31A and SEC31B) from HeLa cells. Depletion of these proteins did not lead to an up- regulation of the PTC-containing NMD reporter, suggesting that neither of these previously characterised functions of SEC13 appears to be required for its role in NMD (Figure 4.10).

139

SEC13 comprises six WD40 protein-protein interaction domains. It is therefore possible that SEC13 acts as a binding platform for other NMD factors. In a recent study, Loh et al. (2013) found SEC13 amongst the proteins co- immunoprecipitating with the PC domain of SMG7. This interaction would position SEC13 in the last step of the NMD pathway. Preliminary interaction analyses carried out by a post-doctoral fellow in our laboratory, Dr. Nele Hug, and a former technician, Marianne Keith, show that SEC13 can co-immunoprecipitate with full length SMG7, confirming the observation of Loh et al. (2013). In the same set of experiments, SEC13 was also found to co-immunoprecipitate with UPF1 and DHX34, a protein that is recruited to the SURF complex to facilitate the transition to the DECID complex (Hug & Cáceres, 2014).

These interactions raise the possibility that SEC13 could be recruited to the SURF complex, prior to or at the same time as DHX34, where it binds to UPF1 before it interacts with the EJC complex. The SEC13 WD40 domains could then mediate the interaction of UPF1 with other NMD factors, such as DHX34 and SMG7 (Figure 6.2).

140

Figure 6.2 Hypothetical role of SEC13 in the NMD pathway. UPF3 and UPF2 are bound to the EJC. The SURF complex (SMG1-UPF1-eRF1-eRF3) assembles on the ribosome stalling at the PTC. Interactions between the SURF complex and the EJC lead to the conformational changes in the two protein complexes and in the formation of the DECID complex. DHX34 (and potentially SEC13) is recruited to the SURF complex to facilitate the transition to DECID complex. As a result of the conformational changes, UPF1 is phosphorylated. The release factors, SMG8, SMG9, and DHX34, are released and additional proteins (SMG6, SMG5-7) are recruited. The targeted transcript is rapidly degraded.

141

In order to understand at which stage SEC13 acts in the NMD pathway, and when it is recruited to the target mRNA, it would be important to determine which of the known NMD factors interact with SEC13. These could be revealed by means of immunoprecipitation of SEC13, followed by western blot analysis, probing for all known NMD factors. The interaction of SEC13 with UPF1, DHX34 and SMG7 which were preliminarily characterised in our laboratory will need to be confirmed with independent experiments. To determine whether the interactions are mediated by RNA, immunoprecipitated samples could be treated with RNase. Furthermore, the analysis of UPF1 mutants could be used to pinpoint the regions important for SEC13 interaction.

The mass spectrometry analysis of the proteins co-immunoprecipitating with SEC13 were of poor quality, as indicated by the presence of only one component of COPII coat vesicles, and the absence of NPC components. Repeating this experiment could lead to the identification of a wider range of SEC13 interactors, including those involved in NMD.

Mutations in the WD40 domain of another protein have been reported to disrupt the interaction with a known binding partner (Denisenko et al., 1998). It would be interesting to introduce point mutations in SEC13 WD40 domains and assess whether they disrupt the NMD function of SEC13. It would also be interesting to create SEC13 truncations to map the interaction of SEC13 with other components of NMD pathway, starting with DHX34 and SMG7, for which a putative interaction has already been described (Loh et al., 2013, unpublished observations by Dr. Nele Hug), and to test whether the truncated mutants are still able to induce decay of a reporter mRNA in tethering assays.

Although the results presented in Figure 4.10 suggest that neither the NPC nor the COPII coat vesicles transport system are involved in NMD, they do not rule it out completely. Depletion of a higher number of components of these two complexes in our NMD reporter system, accompanied by assessment of the impairment of NPC/COPII coat function in the depleted cells, would allow to determine whether the NMD function of SEC13 is linked to one of its previously described functions.

142

Both functions could be important for NMD. If the role of SEC13 in the COPII coat were to be essential for its function in NMD, it would point toward a compartmentalisation of NMD. One of the factors identified in the screen performed by Dr. Dasa Longman, NBAS, has been found to be a components of the syntaxin 18 complex, which is involved in the transport of vesicles from the Golgi apparatus to the endoplasmic reticulum (Longman et al., 2007, Aoki et al., 2009). COPII coat vesicles are involved in the transportation of cargo from ER to Golgi. It is therefore possible that SEC13 and NBAS could facilitate the correct localisation and assembly of NMD factors involved in the degradation of transcripts that localise to the Golgi or ER.

An alternative scenario is that the nuclear pore function of SEC13 is required for NMD. The nuclear pore complex could transport faulty transcripts that need to be degraded by the NMD machinery to prevent the production of truncated proteins, from the nucleus to the cytoplasm. It is not clear how and why SEC13 would then interact with DHX34 and SMG7. Although both functions could be linked to NMD, the preliminary interaction data available are not easily reconcilable with either SEC13 function. Only in depth analysis of SEC13 interactors can shed light on which (if any) previously characterised role of SEC13 is required for NMD. In any case, SEC13 is likely to act as a scaffold protein promoting the formation of protein complexes by means of its WD40 domains.

6.4 Concluding remarks

During the course of my PhD I identified five novel putative NMD factors in C. elegans. The function of one of these new effectors, NOAH-2, was shown to be conserved in Drosophila by our collaborators. Furthermore, I demonstrated that the homologues of two of the novel putative C. elegans NMD factors, GNL2 and SEC13, are functionally conserved in the NMD pathway in H. sapiens.

The work presented in this thesis led to uncover novel NMD factors in three different organisms, indicating that many more factors could still be discovered.

143

Characterising the exact role played by the two novel human NMD factors in the surveillance pathway will push forward our understanding of NMD. The increasing evidence supporting the existence of alternative pathways of NMD leads to the idea that additional uncharacterised proteins could be involved in the NMD regulation of transcripts, acting on specific subset of mRNAs. Their function could potentially be required for NMD in specific cell types or at different stages of development, when the expression of particular subsets of genes is fundamental for proper development of an organism. Performing a similar screen to the one described in this thesis in higher eukaryotes may allow to answer this question. Whilst the delivery of shRNA/siRNA/CRISPR- Cas constructs to human cell lines for loss-of-function experiments may prove to be more problematic and costly than allowing nematodes to feed on a bacterial library, such a screen could allow the identification of factors that are not conserved in other eukaryotic species. Another disadvantage may be represented by the cell line chosen for the screen, which would only express a specific subset of genes, compared to a developing nematode which will express its full complement of coding genes over the course of its life. This could however prove to be an advantage should the loss- of-function of a particular gene be lethal at early developmental stages but not in the cell type of choice.

The importance of the NMD pathway is underlined by its requirement for embryogenesis and proper development in mice, and emphasised by the observation that mutations that impair the functions of NMD factors, such as mutations in the UPF3B gene, have been associated with several mental disorders in humans (Medghalchi et al., 2001, Weischenfeldt et al., 2008, McIlwain et al., 2010, Alrahbeni et al., 2015, Li et al., 2015).

Approximately 30% of all genetic diseases are caused by mutations leading to PTCs and, in order to develop treatments to cure them, it is of fundamental importance for us to achieve a comprehensive understanding of the molecular mechanisms underlying the NMD pathway. The uncovering of novel NMD effectors in this work will hopefully aid in the achievement of this goal.

144

References

Addington A. M., Gauthier J., Piton A., Hamdan F. F., Raymond A., Gogtay N., Miller R., Tossell J., Bakalar J., Inoff-Germain G., Gochman P., Long R., Rapoport J. L., Rouleau G. A. A novel frameshift mutation in UPF3B identified in brothers affected with childhood onset schizophrenia and autism spectrum disorders. Mol Psychiatry. 2011 Mar;16(3):238-9

Ailion M., Thomas J. H. Isolation and characterization of high-temperature-induced Dauer formation mutants in Caenorhabditis elegans. Genetics. 2003 Sep;165(1):127-44

Akimitsu N., Tanaka J., Pelletier J. Translation of nonSTOP mRNA is repressed post- initiation in mammalian cells. EMBO J. 2007 May 2;26(9):2327-38

Albers C. A., Paul D. S., Schulze H., Freson K., Stephens J. C., Smethurst P. A., Jolley J. D., Cvejic A., Kostadima M., Bertone P., Breuning M. H., Debili N., Deloukas P., Favier R., Fiedler J., Hobbs C. M., Huang N., Hurles M. E., Kiddle G., Krapels I., Nurden P., Ruivenkamp C. A., Sambrook J. G., Smith K., Stemple D. L., Strauss G., Thys C., van Geet C., Newbury-Ecob R., Ouwehand W. H., Ghevaert C. Compound inheritance of a low-frequency regulatory SNP and a rare null mutation in exon-junction complex subunit RBM8A causes TAR syndrome. Nat Genet. 2012 Feb 26;44(4):435-9, S1-2

Albert T. K., Lemaire M., van Berkum N. L., Gentz R., Collart M. A., Timmers H. T. Isolation and characterization of human orthologs of yeast CCR4-NOT complex subunits. Nucleic Acids Res. 2000 Feb 1;28(3):809-17

Allison L. A., Moyle M., Shales M., Ingles C. J. Extensive homology among the largest subunits of eukaryotic and prokaryotic RNA polymerases. Cell. 1985 Sep;42(2):599-610

145

Alonso C. R., Akam M. A Hox gene mutation that triggers nonsense-mediated RNA decay and affects alternative splicing during Drosophila development. Nucleic Acids Res. 2003 Jul 15;31(14):3873-80

Alrahbeni T., Sartor F., Anderson J., Miedzybrodzka Z., McCaig C., Müller B. Full UPF3B function is critical for neuronal differentiation of neural stem cells. Mol Brain. 2015 May 27;8:33

Anastasaki C., Longman D., Capper A., Patton E. E., Cáceres J. F. Dhx34 and Nbas function in the NMD pathway and are required for embryonic development in zebrafish. Nucleic Acids Res. 2011 May;39(9):3686-94

Aoki T., Ichimura S., Itoh A., Kuramoto M., Shinkawa T., Isobe T., Tagaya M. Identification of the neuroblastoma-amplified gene product as a component of the syntaxin 18 complex implicated in Golgi-to-endoplasmic reticulum retrograde transport. Mol Biol Cell. 2009 Jun;20(11):2639-49

Applequist S. E., Selg M., Raman C., Jäck H. M. Cloning and characterization of hUPF1, a human homolog of the Saccharomyces cerevisiae nonsense mRNA- reducing UPF1 protein. Nucleic Acids Res. 1997 Feb 15;25(4):814-21

Arciga-Reyes L., Wootton L., Kieffer M., Davies B. UPF1 is required for nonsense- mediated mRNA decay (NMD) and RNAi in Arabidopsis. Plant J. 2006 Aug;47(3):480-9. Epub 2006 Jun 30

Arias-Palomo E., Yamashita A., Fernández I. S., Núñez-Ramírez R., Bamba Y., Izumi N., Ohno S., Llorca O. The nonsense-mediated mRNA decay SMG-1 kinase is regulated by large-scale conformational changes controlled by SMG-8. Genes Dev. 2011 Jan 15;25(2):153-64

Azzalin C.M., Lingner J. The human RNA surveillance factor UPF1 is required for S phase progression and genome stability. Curr Biol. 2006 Feb 21;16(4):433–9.

146

Baltz A. G., Munschauer M., Schwanhäusser B., Vasile A., Murakawa Y., Schueler M., Youngs N., Penfold-Brown D., Drew K., Milek M., Wyler E., Bonneau R., Selbach M., Dieterich C., Landthaler M. The mRNA-bound proteome and its global occupancy profile on protein-coding transcripts. Mol Cell. 2012 Jun 8;46(5):674-90

Banerjee A., Benjamin R., Balakrishnan K., Ghosh P., Banerjee S. Human protein Staufen-2 promotes HIV-1 proliferation by positively regulating RNA export activity of viral protein Rev. Retrovirology. 2014 Feb 13;11:18

Bao J., Tang C., Yuan S., Porse B. T., Yan W. UPF2, a nonsense-mediated mRNA decay factor, is required for prepubertal Sertoli cell development and male fertility by ensuring fidelity of the transcriptome. Development. 2015 Jan 15;142(2):352-62

Barlowe C., Orci L., Yeung T., Hosobuchi M., Hamamoto S., Salama N., Rexach M. F., Ravazzola M., Amherdt M., Schekman R. COPII: a membrane coat formed by Sec proteins that drive vesicle budding from the endoplasmic reticulum. Cell. 1994 Jun 17;77(6):895-907

Bar-Peled L., Chantranupong L., Cherniack A.D., Chen W.W., Ottina K.A., Grabiner B.C., Spear E.D., Carter S.L., Meyerson M., Sabatini D.M. A tumor suppressor complex with GAP activity for the Rag GTPases that signal amino acid sufficiency to mTORC1 Science. 2013 May 31;340(6136):1100-6

Baurén G., Wieslander L. Splicing of Balbiani ring 1 gene pre-mRNA occurs simultaneously with transcription. Cell. 1994 Jan 14;76(1):183-92

Beaudet A. L, Caskey C. T. Mammalian peptide chain termination. II. Codon specificity and GTPase activity of release factor. Proc Natl Acad Sci U S A. 1971 Mar;68(3):619-24

147

Beelman C. A., Stevens A., Caponigro G., LaGrandeur T. E., Hatfield L., Fortner D. M., Parker R. An essential component of the decapping enzyme required for normal rates of mRNA turnover. Nature. 1996 Aug 15;382(6592):642-6

Bengtson M. H., Joazeiro C. A. Role of a ribosome-associated E3 ubiquitin ligase in protein quality control. Nature. 2010 Sep 23;467(7314):470-3

Berget S. M., Moore C., Sharp P. A. Spliced segments at the 5' terminus of adenovirus 2 late mRNA. Proc Natl Acad Sci U S A. 1977 Aug;74(8):3171-5

Bhattacharyya S., Yu H., Mim C., Matouschek A. Regulated protein turnover: snapshots of the proteasome in action. Nat Rev Mol Cell Biol. 2014 Feb;15(2):122- 33

Bi X., Corpina R. A., Goldberg J. Structure of the Sec23/24-Sar1 pre-budding complex of the COPII vesicle coat. Nature. 2002 Sep 19;419(6904):271-7

Boeck R., Tarun S. Jr, Rieger M., Deardorff J. A., Müller-Auer S., Sachs A. B. The yeast Pan2 protein is required for poly(A)-binding protein-stimulated poly(A)- nuclease activity. J Biol Chem. 1996 Jan 5;271(1):432-8

Boehm V., Haberman N., Ottens F., Ule J., Gehring N. H. 3' UTR length and messenger ribonucleoprotein composition determine endocleavage efficiencies at termination codons. Cell Rep. 2014 Oct 23;9(2):555-68

Brazão T. F., Demmers J., van IJcken W., Strouboulis J., Fornerod M., Romão L., Grosveld F. G. A new function of ROD1 in nonsense-mediated mRNA decay. FEBS Lett. 2012 Apr 24;586(8):1101-10

Brody E., Abelson J. The "spliceosome": yeast pre-messenger RNA associates with a 40S complex in a splicing-dependent reaction. Science. 1985 May 24;228(4702):963-7

148

Brenner S. The genetics of Caenorhabditis elegans. Genetics. 1974 May;77(1):71-94

Bühler M., Steiner S., Mohn F., Paillusson A., Mühlemann O. EJC-independent degradation of nonsense immunoglobulin-mu mRNA depends on 3' UTR length. Nat Struct Mol Biol. 2006 May;13(5):462-4

Cali B. M., Kuchma S. L., Latham J., Anderson P. smg-7 is required for mRNA surveillance in Caenorhabditis elegans. Genetics. 1999 Feb;151(2):605-16

Calo E., Flynn R. A., Martin L., Spitale R. C., Chang H. Y., Wysocka J. RNA helicase DDX21 coordinates transcription and ribosomal RNA processing. Nature. 2015 Feb 12;518(7538):249-53

Casadio A., Longman D., Hug N., Delavaine L., Vallejos Baier R., Alonso C. R., Cáceres J. F. Identification and characterization of novel factors that act in the nonsense-mediated mRNA decay pathway in nematodes, flies and mammals. EMBO Rep. 2015 Jan;16(1):71-8

Cassada R. C., Russell R. L. The dauerlarva, a post-embryonic developmental variant of the nematode Caenorhabditis elegans. Dev Biol. 1975 Oct;46(2):326-42

Castello A., Fischer B., Eichelbaum K., Horos R., Beckmann B. M., Strein C., Davey N. E., Humphreys D. T., Preiss T., Steinmetz L. M., Krijgsveld J., Hentze M. W. Insights into RNA biology from an atlas of mammalian mRNA-binding proteins. Cell. 2012 Jun 8;149(6):1393-406

Celik A., Kervestin S., Jacobson A. NMD: At the crossroads between translation termination and ribosome recycling. Biochimie. 2015 Jul;114:2-9

Chakrabarti S., Jayachandran U., Bonneau F., Fiorini F., Basquin C., Domcke S., Le Hir H., Conti E. Molecular mechanisms for the RNA-dependent ATPase activity of Upf1 and its regulation by Upf2. Mol Cell. 2011 Mar 18;41(6):693-703

149

Chakrabarti S., Bonneau F., Schüssler S., Eppinger E., Conti E. Phospho-dependent and phospho-independent interactions of the helicase UPF1 with the NMD factors SMG5-SMG7 and SMG6. Nucleic Acids Res. 2014 Aug;42(14):9447-60

Chamieh H., Ballut L., Bonneau F., Le Hir H. NMD factors UPF2 and UPF3 bridge UPF1 to the exon junction complex and stimulate its RNA helicase activity. Nat Struct Mol Biol. 2008 Jan;15(1):85-93

Chan W. K., Huang L., Gudikote J. P., Chang Y. F., Imam J. S., MacLean J. A. 2nd, Wilkinson M. F. An alternative branch of the nonsense-mediated decay pathway. EMBO J. 2007 Apr 4;26(7):1820-30

Chan W. K., Bhalla A. D., Le Hir H., Nguyen L. S., Huang L., Gécz J., Wilkinson M. F. A UPF3-mediated regulatory switch that maintains RNA surveillance. Nat Struct Mol Biol. 2009 Jul;16(7):747-53

Chang J. C., Kan Y. W. beta 0 thalassemia, a nonsense mutation in man. Proc Natl Acad Sci USA. 1979 Jun;76(6):2886-9

Chang J. H., Xiang S., Xiang K., Manley J. L., Tong L. Structural and biochemical studies of the 5'→3' exoribonuclease Xrn1. Nat Struct Mol Biol. 2011 Mar;18(3):270-6

Chang C. T., Bercovich N., Loh B., Jonas S., Izaurralde E. The activation of the decapping enzyme DCP2 by DCP1 occurs on the EDC4 scaffold and involves a conserved loop in DCP1. Nucleic Acids Res. 2014 Apr;42(8):5217-33

Chapin A., Hu H., Rynearson S. G., Hollien J., Yandell M., Metzstein M. M. In vivo determination of direct targets of the nonsense-mediated decay pathway in Drosophila. G3 (Bethesda). 2014 Mar 20;4(3):485-96

150

Chen C. Y., Shyu A. B. Rapid deadenylation triggered by a nonsense codon precedes decay of the RNA body in a mammalian cytoplasmic nonsense-mediated decay pathway. Mol Cell Biol. 2003 Jul;23(14):4805-13

Chen L., Muhlrad D., Hauryliuk V., Cheng Z., Lim M. K., Shyp V., Parker R., Song H. Structure of the Dom34-Hbs1 complex and implications for no-go decay. Nat Struct Mol Biol. 2010 Oct;17(10):1233-40

Cheng Z., Saito K., Pisarev A. V., Wada M., Pisareva V. P., Pestova T. V., Gajda M., Round A., Kong C., Lim M., Nakamura Y., Svergun D. I., Ito K., Song H. Structural insights into eRF3 and stop codon recognition by eRF1. Genes Dev. 2009 May 1;23(9):1106-18

Chennupati V., Datta D., Rao M. R., Boddapati N., Kayasani M., Sankaranarayanan R., Mishra M., Seth P., Mani C., Mahalingam S. Signals and pathways regulating nucleolar retention of novel putative nucleolar GTPase NGP-1(GNL-2). Biochemistry. 2011 May 31;50(21):4521-36

Chiu S. Y., Serin G., Ohara O., Maquat L. E. Characterization of human Smg5/7a: a protein with similarities to Caenorhabditis elegans SMG5 and SMG7 that functions in the dephosphorylation of Upf1. RNA. 2003 Jan;9(1):77-87

Cho E. J., Takagi T., Moore C. R., Buratowski S. mRNA capping enzyme is recruited to the transcription complex by phosphorylation of the RNA polymerase II carboxy- terminal domain. Genes Dev. 1997 Dec 15;11(24):3319-26

Cho H., Kim K. M., Kim Y. K. Human proline-rich nuclear receptor coregulatory protein 2 mediates an interaction between mRNA surveillance machinery and decapping complex. Mol Cell. 2009 Jan 16;33(1):75-86

151

Cho H., Kim K. M., Han S., Choe J., Park S. G., Choi S. S., Kim Y. K. Staufen1- mediated mRNA decay functions in adipogenesis. Mol Cell. 2012 May 25;46(4):495- 506

Cho H., Han S., Choe J., Park S. G., Choi S. S., Kim Y. K. SMG5-PNRC2 is functionally dominant compared with SMG5-SMG7 in mammalian nonsense- mediated mRNA decay. Nucleic Acids Res. 2013a Jan;41(2):1319-28

Cho H., Han S., Park O. H., Kim Y. K. SMG1 regulates adipogenesis via targeting of staufen1-mediated mRNA decay. Biochim Biophys Acta. 2013b Dec;1829(12):1276-87

Chow L. T., Gelinas R. E., Broker T. R., Roberts R. J. An amazing sequence arrangement at the 5' ends of adenovirus 2 messenger RNA. Cell. 1977 Sep;12(1):1- 8

Christie M., Boland A., Huntzinger E., Weichenrieder O., Izaurralde E. Structure of the PAN3 pseudokinase reveals the basis for interactions with the PAN2 deadenylase and the GW182 proteins. Mol Cell. 2013 Aug 8;51(3):360-73

Chung Y. D., Zhu J., Han Y., Kernan M. J. nompA encodes a PNS-specific, ZP domain protein required to connect mechanosensory dendrites to sensory structures. Neuron. 2001 Feb;29(2):415-28

Corden J. L., Cadena D. L., Ahearn J. M. Jr., Dahmus M. E. A unique structure at the carboxyl terminus of the largest subunit of eukaryotic RNA polymerase II. Proc Natl Acad Sci U S A. 1985 Dec;82(23):7934-8

Culbertson M. R., Underbrink K. M., Fink G. R. Frameshift suppression Saccharomyces cerevisiae. II. Genetic properties of group II suppressors. Genetics. 1980 Aug;95(4):833-53

152

Custódio N., Carmo-Fonseca M., Geraghty F., Pereira H. S., Grosveld F., Antoniou M. Inefficient processing impairs release of RNA from the site of transcription. EMBO J. 1999 May 17;18(10):2855-66

Davidson L., Kerr A., West S. Co-transcriptional degradation of aberrant pre- mRNA by Xrn2. EMBO J. 2012 May 30;31(11):2566-78

Davy A., Bello P., Thierry-Mieg N., Vaglio P., Hitti J., Doucette-Stamm L., Thierry- Mieg D., Reboul J., Boulton S., Walhout A. J., Coux O., Vidal M. A protein-protein interaction map of Caenorhabditis elegans 26S proteasome. EMBO Rep. 2001 Sep;2(9):821-8 de Almeida S. F., García-Sacristán A., Custódio N., Carmo-Fonseca M. A link between nuclear RNA surveillance, the human exosome and RNA polymerase II transcriptional termination. Nucleic Acids Res. 2010 Dec;38(22):8015-26

Denisenko O., Shnyreva M., Suzuki H., Bomsztyk K. Point mutations in the WD40 domain of Eed block its interaction with Ezh2. Mol Cell Biol. 1998 Oct;18(10):5634- 42

De Santis L., Gandolfi F., Pennarossa G., Maffei S., Gismano E., Intra G., Candiani M., Brevini T. A. Expression and intracytoplasmic distribution of staufen and calreticulin in maturing human oocytes. J Assist Reprod Genet. 2015 Apr;32(4):645- 52 de Turris V., Nicholson P., Orozco R. Z., Singer R. H., Mühlemann O. Cotranscriptional effect of a premature termination codon revealed by live-cell imaging. RNA. 2011 Dec;17(12):2094-107

153

Degot S., Le Hir H., Alpy F., Kedinger V., Stoll I., Wendling C., Seraphin B., Rio M. C., Tomasetto C. Association of the breast cancer protein MLN51 with the exon junction complex via its speckle localizer and RNA binding module. J Biol Chem. 2004 Aug 6;279(32):33702-15

Denning G., Jamieson L., Maquat L. E., Thompson E. A., Fields A. P. Cloning of a novel phosphatidylinositol kinase-related kinase: characterization of the human SMG-1 RNA surveillance protein. J Biol Chem. 2001 Jun 22;276(25):22709-14

Doma M. K., Parker R. Endonucleolytic cleavage of eukaryotic mRNAs with stalls in translation elongation. Nature. 2006 Mar 23;440(7083):561-4

Drechsel G., Kahles A., Kesarwani A. K., Stauffer E., Behr J., Drewe P., Rätsch G., Wachter A. Nonsense-mediated decay of alternative precursor mRNA splicing variants is a major determinant of the Arabidopsis steady state transcriptome. Plant Cell. 2013 Oct;25(10):3726-42

Duchaîne T. F., Hemraj I., Furic L., Deitinghoff A., Kiebler M. A., DesGroseillers L. Staufen2 isoforms localize to the somatodendritic domain of neurons and interact with different organelles. J Cell Sci. 2002 Aug 15;115(Pt 16):3285-95

Dunckley T., Parker R. The DCP2 protein is required for mRNA decapping in Saccharomyces cerevisiae and contains a functional MutT motif. EMBO J. 1999 Oct 1;18(19):5411-22

Eberle A. B., Stalder L., Mathys H., Orozco R. Z., Mühlemann O. Posttranscriptional gene regulation by spatial rearrangement of the 3' untranslated region. PLoS Biol. 2008 Apr 29;6(4):e92

Eberle A. B., Lykke-Andersen S., Mühlemann O., Jensen T. H. SMG6 promotes endonucleolytic cleavage of nonsense mRNA in human cells. Nat Struct Mol Biol. 2009 Jan;16(1):49-55

154

Eberle A. B., Hessle V., Helbig R., Dantoft W., Gimber N., Visa N. Splice-site mutations cause Rrp6-mediated nuclear retention of the unspliced RNAs and transcriptional down-regulation of the splicing-defective genes. PLoS One. 2010 Jul 12;5(7):e11540

Edmonds M., Abrams R. Polynucleotide biosynthesis: formation of a sequence of adenylate units from adenosine triphosphate by an enzyme from thymus nuclei. J Biol Chem. 1960 Apr;235:1142-9

Edmonds M., Vaughan M. H. Jr., Nakazato H. Polyadenylic acid sequences in the heterogeneous nuclear RNA and rapidly-labeled polyribosomal RNA of HeLa cells: possible evidence for a precursor relationship. Proc Natl Acad Sci U S A. 1971 Jun;68(6):1336-40

Enninga J., Levay A., Fontoura B. M. Sec13 shuttles between the nucleus and the cytoplasm and stably interacts with Nup96 at the nuclear pore complex. Mol Cell Biol. 2003 Oct;23(20):7271-84

Essers P. B., Pereboom T. C., Goos Y. J., Paridaen J. T., Macinnes A. W. A comparative study of nucleostemin family members in zebrafish reveals specific roles in ribosome biogenesis. Dev Biol. 2014 Jan 15;385(2):304-15

Feng W., Liang T., Yu J., Zhou W., Zhang Y., Wu Z., Xu T. RAB-27 and its effector RBF-1 regulate the tethering and docking steps of DCV exocytosis in C. elegans. Sci China Life Sci. 2012 Mar; 55(3):228-35

Fenger-Grøn M., Fillman C., Norrild B., Lykke-Andersen J. Multiple processing body factors and the ARE binding protein TTP activate mRNA decapping. Mol Cell. 2005 Dec 22;20(6):905-15

Ferrington D. A., Gregerson D. S. Immunoproteasomes: structure, function, and antigen presentation. Prog Mol Biol Transl Sci. 2012;109:75-112

155

Finch A. J., Hilcenko C., Basse N., Drynan L. F., Goyenechea B., Menne T. F., González Fernández A., Simpson P., D'Santos C. S., Arends M. J., Donadieu J., Bellanné-Chantelot C., Costanzo M., Boone C., McKenzie A. N., Freund S. M., Warren A. J. Uncoupling of GTP hydrolysis from eIF6 release on the ribosome causes Shwachman-Diamond syndrome. Genes Dev. 2011 May 1;25(9):917-29

Fiorini F., Boudvillain M., Le Hir H. Tight intramolecular regulation of the human Upf1 helicase by its N- and C-terminal domains. Nucleic Acids Res. 2013 Feb 1;41(4):2404-15

Fiorini F., Bagchi D., Le Hir H., Croquette V. Human Upf1 is a highly processive RNA helicase and translocase with RNP remodelling activities. Nat Commun. 2015 Jul 3;6:7581

Fire A., Xu S., Montgomery M. K., Kostas S. A., Driver S. E., Mello C. C. Potent and specific genetic interference by double-stranded RNA in Caenorhabditis elegans. Nature. 1998 Feb; 19;391(6669):806-11.

Frand A. R., Russel S., Ruvkun G. Functional genomic analysis of C. elegans molting. PLoS Biol. 2005 Oct;3(10):e312

Franks T. M., Singh G., Lykke-Andersen J. Upf1 ATPase-dependent mRNP disassembly is required for completion of nonsense- mediated mRNA decay. Cell. 2010 Dec 10;143(6):938-50

Fraser A. G., Kamath R. S., Zipperlen P., Martinez-Campos M., Sohrmann M., Ahringer J. Functional genomic analysis of C. elegans chromosome I by systematic RNA interference. Nature. 2000 Nov 16;408(6810):325-30

Frischmeyer P. A., Dietz H. C. Nonsense-mediated mRNA decay in health and disease. Hum Mol Genet. 1999;8(10):1893-900

156

Frischmeyer P. A., van Hoof A., O'Donnell K., Guerrerio A. L., Parker R., Dietz H. C. An mRNA surveillance mechanism that eliminates transcripts lacking termination codons. Science. 2002 Mar 22;295(5563):2258-61

Frolova L., Le Goff X., Zhouravleva G., Davydova E., Philippe M., Kisselev L. Eukaryotic polypeptide chain release factor eRF3 is an eRF1- and ribosome- dependent guanosine triphosphatase. RNA. 1996 Apr;2(4):334-41

Frolova L. Y., Tsivkovskii R. Y., Sivolobova G. F., Oparina N. Y., Serpinsky O. I., Blinov V. M., Tatkov S. I., Kisselev L. L. Mutations in the highly conserved GGQ motif of class 1 polypeptide release factors abolish ability of human eRF1 to trigger peptidyl-tRNA hydrolysis. RNA. 1999 Aug;5(8):1014-20

Frolova L. Y., Merkulova T. I., Kisselev L. L. Translation termination in eukaryotes: polypeptide release factor eRF1 is composed of functionally and structurally distinct domains. RNA. 2000 Mar;6(3):381-90

Fukuda M. Rab27 and its effectors in secretory granule exocytosis: a novel docking machinery composed of a Rab27 · effector complex. Biochem Soc Trans. 2006 Nov;34(Pt 5):691-5

Furic L., Maher-Laporte M., DesGroseillers L. A genome-wide approach identifies distinct but overlapping subsets of cellular mRNAs associated with Staufen1- and Staufen2-containing ribonucleoprotein complexes. RNA. 2008 Feb;14(2):324-35

157

Gallegos M. E., Balakrishnan S., Chandramouli P., Arora S., Azameera A., Babushekar A., Bargoma E., Bokhari A., Chava S. K., Das P., Desai M., Decena D., Saramma S. D., Dey B., Doss A. L., Gor N., Gudiputi L., Guo C., Hande S., Jensen M., Jones S., Jones N., Jorgens D., Karamchedu P., Kamrani K., Kolora L. D., Kristensen L., Kwan K., Lau H., Maharaj P., Mander N., Mangipudi K., Menakuru H., Mody V., Mohanty S., Mukkamala S., Mundra S. A., Nagaraju S., Narayanaswamy R., Ndungu-Case C., Noorbakhsh M., Patel J., Patel P., Pendem S. V., Ponakala A., Rath M., Robles M. C., Rokkam D., Roth C., Sasidharan P., Shah S., Tandon S., Suprai J., Truong T. Q., Uthayaruban R., Varma A., Ved U., Wang Z., Yu Z. The C. elegans rab family: identification, classification and toolkit construction. PLoS One. 2012;7(11):e49387

Galy V., Mattaj I. W., Askjaer P. Caenorhabditis elegans nucleoporins Nup93 and Nup205 determine the limit of nuclear pore complex size exclusion in vivo. Mol. Biol. Cell. 2003 Dec; 14(12):5104-15

Gatfield D., Unterholzner L., Ciccarelli F. D., Bork P., Izaurralde E. Nonsense- mediated mRNA decay in Drosophila: at the intersection of the yeast and mammalian pathways. EMBO J. 2003 Aug 1;22(15):3960-70

Gatfield D., Izaurralde E. Nonsense-mediated messenger RNA decay is initiated by endonucleolytic cleavage in Drosophila. Nature. 2004 Jun 3;429(6991):575-8

Gautrey H., McConnell J., Hall J., Hesketh J. Polarised distribution of the RNA- binding protein Staufen in differentiated intestinal epithelial cells. FEBS Lett. 2005 Apr 11;579(10):2226-30

Gehring N. H., Neu-Yilik G., Schell T., Hentze M. W., Kulozik A. E. Y14 and hUpf3b form an NMD-activating complex. Mol Cell. 2003 Apr;11(4):939-49

158

Gehring N. H., Kunz J. B., Neu-Yilik G., Breit S., Viegas M. H., Hentze M. W., Kulozik A. E. Exon-junction complex components specify distinct routes of nonsense-mediated mRNA decay with differential cofactor requirements. Mol Cell. 2005 Oct 7;20(1):65-75

Gijsbers A., García-Márquez A., Luviano A., Sánchez-Puig N. Guanine nucleotide exchange in the ribosomal GTPase EFL1 is modulated by the protein mutated in the Shwachman-Diamond syndrome. Biochem Biophys Res Commun. 2013 Aug 2;437(3):349-54

Gleghorn M. L., Gong C., Kielkopf C. L., Maquat L. E. Staufen1 dimerizes through a conserved motif and a degenerate dsRNA-binding domain to promote mRNA decay. Nat Struct Mol Biol. 2013 Apr;20(4):515-24

Gong C., Kim Y. K., Woeller C. F., Tang Y., Maquat L. E. SMD and NMD are competitive pathways that contribute to myogenesis: effects on PAX3 and myogenin mRNAs. Genes Dev. 2009 Jan 1;23(1):54-66

Gong C., Maquat L. E. lncRNAs transactivate STAU1-mediated mRNA decay by duplexing with 3' UTRs via Alu elements. Nature. 2011 Feb 10;470(7333):284-8

Grossman E., Medalia O., Zwerger M. Functional architecture of the nuclear pore complex. Annu Rev Biophys. 2012;41:557-84

Hanke K., Hohn O., Liedgens L., Fiddeke K., Wamara J., Kurth R., Bannert N. Staufen-1 interacts with the human endogenous retrovirus family HERV-K(HML-2) rec and gag proteins and increases virion production. J Virol. 2013 Oct;87(20):11019-30

He F., Li X., Spatrick P., Casillo R., Dong S., Jacobson A. Genome-wide analysis of mRNAs regulated by the nonsense-mediated and 5' to 3' mRNA decay pathways in yeast. Mol Cell. 2003 Dec;12(6):1439-52

159

Herrero-Turrión M.J., Calafat J., Janssen H., Fukuda M., Mollinedo F. Rab27a regulates exocytosis of tertiary and specific granules in human neutrophils. J Immunol. 2008 Sep 15;181(6):3793-803

Hodgkin J., Papp A., Pulak R., Ambros V., Anderson P. A new kind of informational suppression in the nematode Caenorhabditis elegans. Genetics. 1989 Oct;123(2):301-13

Hogg J. R., Goff S. P. Upf1 senses 3'UTR length to potentiate mRNA decay. Cell. 2010 Oct 29;143(3):379-89

Hope I. A. (ed.) C. elegans: A Practical Approach. Oxford University Press, Oxford. 1999. Pp. 281

Hori K., Watanabe Y. UPF3 suppresses aberrant spliced mRNA in Arabidopsis. Plant J. 2005 Aug;43(4):530-40

Huang L., Lou C. H., Chan W., Shum E. Y., Shao A., Stone E., Karam R., Song H. W., Wilkinson M. F. RNA homeostasis governed by cell type-specific and branched feedback loops acting on NMD. Mol Cell. 2011 Sep 16;43(6):950-61

Hug N., Cáceres J. F. The RNA helicase DHX34 activates NMD by promoting a transition from the surveillance to the decay-inducing complex. Cell Rep. 2014 Sep 25;8(6):1845-56

Hughes H., Budnik A., Schmidt K., Palmer K. J., Mantell J., Noakes C., Johnson A., Carter D. A., Verkade P., Watson P., Stephens D. J. Organisation of human ER-exit sites: requirements for the localisation of Sec16 to transitional ER. J Cell Sci. 2009 Aug 15;122(Pt 16):2924-34

160

Huntzinger E., Kashima I., Fauser M., Saulière J., Izaurralde E. SMG6 is the catalytic endonuclease that cleaves mRNAs containing nonsense codons in metazoan. RNA. 2008 Dec;14(12):2609-17

Hurt J. A., Robertson A. D., Burge C. B. Global analyses of UPF1 binding and function reveal expanded scope of nonsense-mediated mRNA decay. Genome Res. 2013 Oct;23(10):1636-50

Im C. H., Hwang S. M., Son Y. S., Heo J. B., Bang W. Y., Suwastika I. N., Shiina T., Bahk J. D. Nuclear/nucleolar GTPase 2 proteins as a subfamily of YlqF/YawG GTPases function in pre-60S ribosomal subunit maturation of mono- and dicotyledonous plants. J Biol Chem. 2011 Mar 11;286(10):8620-32

Inada T., Aiba H. Translation of aberrant mRNAs lacking a termination codon or with a shortened 3'-UTR is repressed after initiation in yeast. EMBO J. 2005 Apr 20;24(8):1584-95

Ito-Harashima S., Kuroha K., Tatematsu T., Inada T. Translation of the poly(A) tail plays crucial roles in nonstop mRNA surveillance via translation repression and protein destabilization by proteasome in yeast. Genes Dev. 2007 Mar 1;21(5):519-24

Ivanov P. V., Gehring N. H., Kunz J. B., Hentze M. W., Kulozik A. E. Interactions between UPF1, eRFs, PABP and the exon junction complex suggest an integrated model for mammalian NMD pathways. EMBO J. 2008 Mar 5;27(5):736-47

Izumi N., Yamashita A., Iwamatsu A., Kurata R., Nakamura H., Saari B., Hirano H., Anderson P., Ohno S. AAA+ proteins RUVBL1 and RUVBL2 coordinate PIKK activity and function in nonsense-mediated mRNA decay. Sci Signal. 2010 Apr 6;3(116):ra27

161

Jiao X., Chang J. H., Kilic T., Tong L., Kiledjian M. A mammalian pre-mRNA 5' end capping quality control mechanism and an unexpected link of capping to pre-mRNA processing. Mol Cell. 2013 Apr 11;50(1):104-15

Jinek M., Coyle S. M., Doudna J. A. Coupled 5' nucleotide recognition and processivity in Xrn1-mediated mRNA decay. Mol Cell. 2011 Mar 4;41(5):600-8

Jolly L. A., Homan C. C., Jacob R., Barry S., Gecz J. The UPF3B gene, implicated in intellectual disability, autism, ADHD and childhood onset schizophrenia regulates neural progenitor cell behaviour and neuronal outgrowth. Hum Mol Genet. 2013 Dec 1;22(23):4673-87

Jonas S., Weichenrieder O., Izaurralde E. An unusual arrangement of two 14-3-3-like domains in the SMG5-SMG7 heterodimer is required for efficient nonsense-mediated mRNA decay. Genes Dev. 2013 Jan 15;27(2):211-25

Jonas S., Christie M., Peter D., Bhandari D., Loh B., Huntzinger E., Weichenrieder O., Izaurralde E. An asymmetric PAN3 dimer recruits a single PAN2 exonuclease to mediate mRNA deadenylation and decay. Nat Struct Mol Biol. 2014 Jul;21(7):599- 608

Kallstrom G., Hedges J., Johnson A. The putative GTPases Nog1p and Lsg1p are required for 60S ribosomal subunit biogenesis and are localized to the nucleus and cytoplasm, respectively. Mol Cell Biol. 2003 Jun;23(12):4344-55

Kamath R. S., Martinez-Campos M., Zipperlen P., Fraser A. G., Ahringer J. Effectiveness of specific RNA-mediated interference through ingested double- stranded RNA in Caenorhabditis elegans. Genome Biol. 2001;2(1):RESEARCH0002

162

Kamath R. S., Fraser A. G., Dong Y., Poulin G., Durbin R., Gotta M., Kanapin A., Le Bot N., Moreno S., Sohrmann M., Welchman D. P., Zipperlen P., Ahringer J. Systematic functional analysis of the Caenorhabditis elegans genome using RNAi. Nature 2003 Jan 16;421(6920):231-7

Kashima I., Yamashita A., Izumi N., Kataoka N., Morishita R., Hoshino S., Ohno M., Dreyfuss G., Ohno S. Binding of a novel SMG-1-Upf1-eRF1-eRF3 complex (SURF) to the exon junction complex triggers Upf1 phosphorylation and nonsense- mediated mRNA decay. Genes Dev. 2006 Feb 1;20(3):355-67

Kerényi Z., Mérai Z., Hiripi L., Benkovics A., Gyula P., Lacomme C., Barta E., Nagy F., Silhavy D. Inter-kingdom conservation of mechanism of nonsense-mediated mRNA decay. EMBO J. 2008 Jun 4;27(11):1585-95

Kernan M., Cowan D., Zuker C. Genetic dissection of mechanosensory transduction: mechanoreception-defective mutations of Drosophila. Neuron. 1994 Jun;12(6):1195- 206

Kerr T. P., Sewry C. A., Robb S. A., Roberts R. G. Long mutant dystrophins and variable phenotypes: evasion of nonsense-mediated decay? Hum Genet. 2001 Oct;109(4):402-7

Kervestin S., Jacobson A. NMD: a multifaceted response to premature translational termination. Nat Rev Mol Cell Biol. 2012 Nov;13(11):700-12

Kim V. N., Kataoka N., Dreyfuss G. Role of the nonsense-mediated decay factor hUpf3 in the splicing-dependent exon-exon junction complex. Science. 2001 Sep 7;293(5536):1832-6

Kim Y. K., Furic L., Desgroseillers L., Maquat L. E. Mammalian Staufen1 recruits Upf1 to specific mRNA 3'UTRs so as to elicit mRNA decay. Cell. 2005 Jan 28;120(2):195-208

163

Kim Y. K., Furic L., Parisien M., Major F., DesGroseillers L., Maquat L. E. Staufen1 regulates diverse classes of mammalian transcripts. EMBO J. 2007 Jun 6;26(11):2670-81

Klass M., Hirsh D. Non-ageing developmental variant of Caenorhabditis elegans. Nature. 1976 Apr 8;260(5551):523-5

Klomp A. E., Teofilo K., Legacki E., Williams D. S. Analysis of the linkage of MYRIP and MYO7A to melanosomes by RAB27A in retinal pigment epithelial cells. Cell Motil Cytoskeleton. 2007 Jun;64(6):474-87

Kunz J. B., Neu-Yilik G., Hentze M. W., Kulozik A. E., Gehring N. H. Functions of hUpf3a and hUpf3b in nonsense-mediated mRNA decay and translation. RNA. 2006 Jun;12(6):1015-22

Kyburz A., Friedlein A., Langen H., Keller W. Direct interactions between subunits of CPSF and the U2 snRNP contribute to the coupling of pre-mRNA 3' end processing and splicing. Mol Cell. 2006 Jul 21;23(2):195-205

Lau N. C., Kolkman A., van Schaik F. M., Mulder K. W., Pijnappel W. W., Heck A. J., Timmers H. T. Human Ccr4-Not complexes contain variable deadenylase subunits. Biochem J. 2009 Aug 27;422(3):443-53

Laumonnier F., Shoubridge C., Antar C., Nguyen L. S., Van Esch H., Kleefstra T., Briault S., Fryns J. P., Hamel B., Chelly J., Ropers H. H., Ronce N., Blesson S., Moraine C., Gécz J., Raynaud M. Mutations of the UPF3B gene, which encodes a protein widely expressed in neurons, are associated with nonspecific mental retardation with or without autism. Mol Psychiatry. 2010 Jul;15(7):767-76

Le Hir H., Izaurralde E., Maquat L. E., Moore M. J. The spliceosome deposits multiple proteins 20-24 nucleotides upstream of mRNA exon-exon junctions. EMBO J. 2000 Dec 15;19(24):6860-9

164

Le Hir H., Gatfield D., Izaurralde E., Moore M. J. The exon-exon junction complex provides a binding platform for factors involved in mRNA export and nonsense- mediated mRNA decay. EMBO J. 2001a Sep 3;20(17):4987-97

Le Hir H., Gatfield D., Braun I. C., Forler D., Izaurralde E. The protein Mago provides a link between splicing and mRNA localization. EMBO Rep. 2001b Dec;2(12):1119-24

Leeds P., Peltz S. W., Jacobson A., Culbertson M. R. The product of the yeast UPF1 gene is required for rapid turnover of mRNAs containing a premature translational termination codon. Genes Dev. 1991 Dec;5(12A):2303-14

Leeds P., Wood J. M., Lee B. S., Culbertson M. R. Gene products that promote mRNA turnover in Saccharomyces cerevisiae. Mol Cell Biol. 1992 May;12(5):2165- 77

Lejeune F., Li X., Maquat L. E. Nonsense-mediated mRNA decay in mammalian cells involves decapping, deadenylating, and exonucleolytic activities. Mol Cell. 2003 Sep;12(3):675-87

Lelivelt M. J., Culbertson M. R. Yeast Upf proteins required for RNA surveillance affect global expression of the yeast transcriptome. Mol Cell Biol. 1999 Oct;19(10):6710-9

Lenasi T., Peterlin B. M., Barboric M. Cap-binding protein complex links pre-mRNA capping to transcription elongation and alternative splicing through positive transcription elongation factor b (P-TEFb). J Biol Chem. 2011 Jul 1;286(26):22758- 68

Lewis B. P., Green R. E., Brenner S. E. Evidence for the widespread coupling of alternative splicing and nonsense-mediated mRNA decay in humans. Proc Natl Acad Sci U S A. 2003 Jan 7;100(1):189-92

165

Li T., Shi Y., Wang P., Guachalla L. M., Sun B., Joerss T., Chen Y. S., Groth M., Krueger A., Platzer M., Yang Y. G., Rudolph K. L., Wang Z. Q. Smg6/Est1 licenses embryonic stem cell differentiation via nonsense-mediated mRNA decay. EMBO J. 2015 Jun 12;34(12):1630-47

Liu Q., Greimann J. C., Lima C. D. Reconstitution, activities, and structure of the eukaryotic RNA exosome. Cell. 2006 Dec 15;127(6):1223-37

Liu D., Meckel T., Long E. O. Distinct role of rab27a in granule movement at the plasma membrane and in the cytosol of NK cells. PLoS One. 2010 Sep 21;5(9):e12870

Lloyd J. P., Davies B. SMG1 is an ancient nonsense-mediated mRNA decay effector. Plant J. 2013 Dec;76(5):800-10

Logan J., Falck-Pedersen E., Darnell J. E. Jr., Shenk T. A poly(A) addition site and a downstream termination region are required for efficient cessation of transcription by RNA polymerase II in the mouse beta maj-globin gene. Proc Natl Acad Sci U S A. 1987 Dec;84(23):8306-10

Loh B., Jonas S., Izaurralde E. The SMG5-SMG7 heterodimer directly recruits the CCR4-NOT deadenylase complex to mRNAs containing nonsense codons via interaction with POP2. Genes Dev. 2013 Oct 1;27(19):2125-38

Longman D., Plasterk R. H. A., Johnstone I. L., Cáceres J. F. Mechanistic insights and identification of two novel factors in the C. elegans NMD pathway. Genes Dev. 2007 May 1;21(9):1075-85

Longman D., Hug N., Keith M., Anastasaki C., Patton E. E., Grimes G., Cáceres J. F. DHX34 and NBAS form part of an autoregulatory NMD circuit that regulates endogenous RNA targets in human cells, zebrafish and Caenorhabditis elegans. Nucleic Acids Res. 2013 Sep;41(17):8319-31

166

Losson R., Lacroute F. Interference of nonsense mutations with eukaryotic messenger RNA stability. Proc Natl Acad Sci USA. 1979 Oct;76(10):5134-7

Lou C. H., Shao A., Shum E. Y., Espinoza J. L., Huang L., Karam R., Wilkinson M. F. Posttranscriptional control of the stem cell and neurogenic programs by the nonsense-mediated RNA decay pathway. Cell Rep. 2014 Feb 27;6(4):748-64

Lutzmann M., Kunze R., Buerer A., Aebi U., Hurt E. Modular self-assembly of a Y- shaped multiprotein complex from seven nucleoporins. EMBO J. 2002 Feb 1;21(3):387-97

Lykke-Andersen J., Shu M. D., Steitz J. A. Human Upf proteins target an mRNA for nonsense-mediated decay when bound downstream of a termination codon. Cell. 2000 Dec 22;103(7):1121-31

Lykke-Andersen J., Shu M. D., Steitz J. A. Communication of the position of exon- exon junctions to the mRNA surveillance machinery by the protein RNPS1. Science. 2001 Sep 7;293(5536):1836-9

Mahoney T. R., Liu Q., Itoh T., Luo S., Hadwiger G., Vincent R., Wang Z. W., Fukuda M., Nonet M. L. Regulation of synaptic transmission by RAB-3 and RAB-27 in Caenorhabditis elegans. Mol Biol Cell. 2006 Jun;17(6).2617-25

Maia A. F., Tanenbaum M. E., Galli M., Lelieveld D., Egan D. A., Gassmann R., Sunkel C. E., van den Heuvel S., Medema R. H. Genome-wide RNAi screen for synthetic lethal interactions with the C. elegans kinesin-5 homolog BMK-1. Sci Data. 2015 May 12;2:150020

Mancias J. D., Goldberg J. Structural basis of cargo membrane protein discrimination by the human COPII coat machinery. EMBO J. 2008 Nov 5;27(21):2918-28

167

Marión R. M., Fortes P., Beloso A., Dotti C., Ortín J. A human sequence homologue of Staufen is an RNA-binding protein that is associated with polysomes and localizes to the rough endoplasmic reticulum. Mol Cell Biol. 1999 Mar;19(3):2212-9

Matsuo Y., Granneman S., Thoms M., Manikas R. G., Tollervey D., Hurt E. Coupled GTPase and remodelling ATPase activities form a checkpoint for ribosome export. Nature. 2014 Jan; 2;505(7481):112-6

McCracken S., Fong N., Yankulov K., Ballantyne S., Pan G., Greenblatt J., Patterson S. D., Wickens M., Bentley D. L. The C-terminal domain of RNA polymerase II couples mRNA processing to transcription. Nature. 1997a Jan 23;385(6614):357-61

McCracken S., Fong N., Rosonina E., Yankulov K., Brothers G., Siderovski D., Hessel A., Foster S., Shuman S., Bentley D. L. 5'-Capping enzymes are targeted to pre-mRNA by binding to the phosphorylated carboxy-terminal domain of RNA polymerase II. Genes Dev. 1997 Dec 15;11(24):3306-18

McIlwain D. R., Pan Q., Reilly P. T., Elia A. J., McCracken S., Wakeham A. C., Itie-Youten A., Blencowe B. J., Mak T. W. Smg1 is required for embryogenesis and regulates diverse genes via alternative splicing coupled to nonsense-mediated mRNA decay. Proc Natl Acad Sci U S A. 2010 Jul 6;107(27):12186-91

Medghalchi S. M., Frischmeyer P. A., Mendell J. T., Kelly A. G., Lawler A. M., Dietz H. C. Rent1, a trans-effector of nonsense-mediated mRNA decay, is essential for mammalian embryonic viability. Hum Mol Genet. 2001 Jan 15;10(2):99-105

Mendell J. T., Sharifi N. A., Meyers J. L., Martinez-Murillo F., Dietz H. C. Nonsense surveillance regulates expression of diverse classes of mammalian transcripts and mutes genomic noise. Nat Genet. 2004 Oct;36(10):1073-8

168

Merkulova T. I., Frolova L. Y., Lazar M., Camonis J., Kisselev L. L. C-terminal domains of human translation termination factors eRF1 and eRF3 mediate their in vivo interaction. FEBS Lett. 1999 Jan 22;443(1):41-7

Metze S., Herzog V. A., Ruepp M. D., Mühlemann O. Comparison of EJC-enhanced and EJC-independent NMD in human cells reveals two partially redundant degradation pathways. RNA. 2013 Oct;19(10):1432-48

Metzstein M. M., Krasnow M. A. Functions of the nonsense-mediated mRNA decay pathway in Drosophila development. PLoS Genet. 2006 Dec 29;2(12):e180

Micklem D. R., Adams J., Grünert S., St. Johnston D. Distinct roles of two conserved Staufen domains in oskar mRNA localization and translation. EMBO J. 2000 Mar 15;19(6):1366-77

Miki T., Kamikawa Y., Kurono S., Kaneko Y., Katahira J., Yoneda Y. Cell type- dependent gene regulation by Staufen2 in conjunction with Upf1. BMC Mol Biol. 2011 Nov 16;12:48

Monaco A. P., Bertelson C. J., Liechti-Gallati S., Moser H., Kunkel L. M. An explanation for the phenotypic differences between patients bearing partial deletions of the DMD locus. Genomics. 1988 Jan;2(1):90-5

Mortillaro M. J., Blencowe B. J., Wei X., Nakayasu H., Du L., Warren S. L., Sharp P. A., Berezney R. A hyperphosphorylated form of the large subunit of RNA polymerase II is associated with splicing complexes and the nuclear matrix. Proc Natl Acad Sci U S A. 1996 Aug 6;93(16):8253-7

Mühlemann O., Mock-Casagrande C. S., Wang J., Li S., Custódio N., Carmo- Fonseca M., Wilkinson M. F., Moore M. J. Precursor RNAs harboring nonsense codons accumulate near the site of transcription. Mol Cell. 2001 Jul;8(1):33-43

169

Muhlrad D., Decker C. J., Parker R. Deadenylation of the unstable mRNA encoded by the yeast MFA2 gene leads to decapping followed by 5'3' digestion of the transcript. Genes Dev. 1994 Apr 1;8(7):855-66

Muhlrad D., Parker R. Aberrant mRNAs with extended 3' UTRs are substrates for rapid degradation by mRNA surveillance. RNA. 1999 Oct;5(10):1299-307

Nguyen L. S., Kim H. G., Rosenfeld J. A., Shen Y., Gusella J. F., Lacassie Y., Layman L. C., Shaffer L. G., Gécz J. Contribution of copy number variants involving nonsense-mediated mRNA decay pathway genes to neuro-developmental disorders. Hum Mol Genet. 2013 May 1;22(9):1816-25

Nicholson P., Josi C., Kurosawa H., Yamashita A., Mühlemann O. A novel phosphorylation-independent interaction between SMG6 and UPF1 is essential for human NMD. Nucleic Acids Res. 2014 Aug;42(14):9217-35

Niu X., Hong J., Zheng X., Melville D. B., Knapik E. W., Meng A., Peng J. The nuclear pore complex function of Sec13 protein is required for cell survival during retinal development. J Biol Chem. 2014 Apr 25;289(17):11971-85

Niwa M., Berget S. M. Mutation of the AAUAAA polyadenylation signal depresses in vitro splicing of proximal but not distal introns. Genes Dev. 1991 Nov;5(11):2086-95

Nott A., Le Hir H., Moore M. J. Splicing enhances translation in mammalian cells: an additional function of the exon junction complex. Genes Dev. 2004 Jan 15;18(2):210-22

Nyikó T., Sonkoly B., Mérai Z., Benkovics A. H., Silhavy D. Plant upstream ORFs can trigger nonsense-mediated mRNA decay in a size-dependent manner. Plant Mol Biol. 2009 Nov;71(4-5):367-78

170

O'Donohue M. F., Choesmel V., Faubladier M., Fichant G., Gleizes P. E. Functional dichotomy of ribosomal proteins during the synthesis of mammalian 40S ribosomal subunits. J Cell Biol. 2010 Sep 6;190(5):853-66

Ohnishi T., Yamashita A., Kashima I., Schell T., Anders K. R., Grimson A., Hachiya T., Hentze M. W., Anderson P., Ohno S. Phosphorylation of hUPF1 induces formation of mRNA surveillance complexes containing hSMG-5 and hSMG-7. Mol Cell. 2003 Nov;12(5):1187-200

Okada-Katsuhata Y., Yamashita A., Kutsuzawa K., Izumi N., Hirahara F., Ohno S. N- and C-terminal Upf1 phosphorylations create binding platforms for SMG-6 and SMG-5:SMG-7 during NMD. Nucleic Acids Res. 2012 Feb;40(3):1251-66

Paillusson A., Hirschi N., Vallan C., Azzalin C.M., Muhlemann O. A GFP-based reporter system to monitor nonsense-mediated mRNA decay. Nucleic Acids Res. 2005 Mar 30;33(6):e54

Palacios I. M., Gatfield D., St Johnston D., Izaurralde E. An eIF4AIII-containing complex required for mRNA localization and nonsense-mediated mRNA decay. Nature. 2004 Feb 19;427(6976):753-7

Pan Q., Shai O., Lee L. J., Frey B. J., Blencowe B. J. Deep surveying of alternative splicing complexity in the human transcriptome by high-throughput sequencing. Nat Genet. 2008 Dec;40(12):1413-5

Park E., Gleghorn M. L., Maquat L. E. Staufen2 functions in Staufen1-mediated mRNA decay by binding to itself and its paralog and promoting UPF1 helicase but not ATPase activity. Proc Natl Acad Sci U S A. 2013 Jan 8;110(2):405-12

Passos D. O., Doma M. K., Shoemaker C. J., Muhlrad D., Green R., Weissman J., Hollien J., Parker R. Analysis of Dom34 and its function in no-go decay. Mol Biol Cell. 2009 Jul;20(13):3025-32

171

Payne J. M., Laybourn P. J., Dahmus M. E. The transition of RNA polymerase II from initiation to elongation is associated with phosphorylation of the carboxyl- terminal domain of subunit IIa. J Biol Chem. 1989 Nov 25;264(33):19621-9

Peixeiro I., Inácio Â., Barbosa C., Silva A. L., Liebhaber S. A., Romão L. Interaction of PABPC1 with the translation initiation complex is critical to the NMD resistance of AUG-proximal nonsense mutations. Nucleic Acids Res. 2012 Feb;40(3):1160-73

Peltz S. W., Brown A. H., Jacobson A. mRNA destabilization triggered by premature translational termination depends on at least three cis-acting sequence elements and one trans-acting factor. Genes Dev. 1993 Sep;7(9):1737-54

Perlick H. A., Medghalchi S. M., Spencer F. A., Kendzior R. J. Jr., Dietz H. C. Mammalian orthologues of a yeast regulator of nonsense transcript stability. Proc Natl Acad Sci USA. 1996 Oct 1;93(20):10928-32

Petit A. P., Wohlbold L., Bawankar P., Huntzinger E., Schmidt S., Izaurralde E., Weichenrieder O. The structural basis for the interaction between the CAF1 nuclease and the NOT1 scaffold of the human CCR4-NOT deadenylase complex. Nucleic Acids Res. 2012 Nov;40(21):11058-72

Pisareva V. P., Skabkin M. A., Hellen C. U., Pestova T. V., Pisarev A. V. Dissociation by Pelota, Hbs1 and ABCE1 of mammalian vacant 80S ribosomes and stalled elongation complexes. EMBO J. 2011 May 4;30(9):1804-17

Reboul J., Vaglio P., Rual J. F., Lameschl P., Martinez M., Armstrong C. M., Li S., Jacotot L., Bertin N., Janky R., Moore T., Hudson J. R. Jr., Hartley J. L., Brasch M. A., Vandenhaute J., Boulton S., Endress G. A., Jenna S., Chevet E., Papasotiropoulos V., Tolias P. P., Ptacek J., Snyder M., Huang R., Chance M. R., Lee H., Doucette- Stamm L., Hill D. E., Vidal M. C. elegans ORFeome version 1.1: experimental verification of the genome annotation and resource for proteome-scale protein expression. Nat. Genet. 2003 May;34(1):35-41

172

Rehwinkel J., Letunic I., Raes J., Bork P., Izaurralde E. Nonsense-mediated mRNA decay factors act in concert to regulate common mRNA targets. RNA. 2005 Oct;11(10):1530-44

Rehwinkel J., Raes J., Izaurralde E. Nonsense-mediated mRNA decay: Target genes and functional diversification of effectors. Trends Biochem Sci. 2006 Nov;31(11):639-46

Reynaud E. G., Andrade M. A., Bonneau F., Ly T. B., Knop M., Scheffzek K., Pepperkok R. Human Lsg1 defines a family of essential GTPases that correlates with the evolution of compartmentalization. BMC Biol. 2005 Oct 7;3:21

Rojas A. M., Fuentes G., Rausell A., Valencia A. The Ras protein superfamily: evolutionary tree and role of conserved amino acids. J Cell Biol. 2012 Jan 23;196(2):189-201

Rual J. F., Ceron J., Koreth J., Hao T., Nicot A. S., Hirozane-Kishikawa T., Vamdenhaute J., Orkin S. H., Hill D. E., van den Heuvel S., Vidal M. Toward improving Caenorhabditis elegans phenome mapping with an ORFeome-based RNAi library. Genome Res. 2004 Oct;14(10B):2162-8

Ruiz-Echevarría M. J., Peltz S. W. The RNA binding protein Pub1 modulates the stability of transcripts containing upstream open reading frames. Cell. 2000 Jun 23;101(7):741-51

Salas-Marco J., Bedwell D. M. GTP hydrolysis by eRF3 facilitates stop codon decoding during eukaryotic translation termination. Mol Cell Biol. 2004 Sep;24(17):7769-78

173

Saulière J., Murigneux V., Wang Z., Marquenet E., Barbosa I., Le Tonquèze O., Audic Y., Paillard L., Roest Crollius H., Le Hir H. CLIP-seq of eIF4AIII reveals transcriptome-wide mapping of the human exon junction complex. Nat Struct Mol Biol. 2012 Nov;19(11):1124-31

Saveanu C., Bienvenu D., Namane A., Gleizes P.E., Gas N., Jacquier A., Fromont- Racine M. Nog2p, a putative GTPase associated with pre-60S subunits and required for late 60S maturation steps. EMBO J. 2001 Nov 15;20(22):6475-84

Saveanu C., Namane A., Gleizes P. E., Lebreton A., Rousselle J. C., Noaillac- Depeyre J., Gas N., Jacquier A., Fromont-Racine M. Sequential protein association with nascent 60s ribosomal particles. Mol. Cell. Biol. 2003 Jul;23(13):4449-60

Schneider C., Leung E., Brown J., Tollervey D. The N-terminal PIN domain of the exosome subunit Rrp44 harbors endonuclease activity and tethers Rrp44 to the yeast core exosome. Nucleic Acids Res. 2009 Mar;37(4):1127-40

Senger B., Lafontaine D. L., Graindorge J. S., Gadal O., Camasses A., Sanni A., Garnier J. M., Breitenbach M., Hurt E., Fasiolo F. The nucle(ol)ar Tif6p and Efl1p are required for a late cytoplasmic step of ribosome synthesis. Mol Cell. 2001 Dec;8(6):1363-7

Serin G., Gersappe A., Black J. D., Aronoff R., Maquat L. E. Identification and characterization of human orthologues to Saccharomyces cerevisiae Upf2 protein and Upf3 protein (Caenorhabditis elegans SMG-4). Mol Cell Biol. 2001 Jan;21(1):209-23

Shibuya T., Tange T. Ø., Sonenberg N., Moore M. J. eIF4AIII binds spliced mRNA in the exon junction complex and is essential for nonsense-mediated decay. Nat Struct Mol Biol. 2004 Apr;11(4):346-51

174

Shoemaker C. J., Eyler D. E., Green R. Dom34:Hbs1 promotes subunit dissociation and peptidyl-tRNA drop-off to initiate no-go decay. Science. 2010 Oct 15;330(6002):369-72

Shutes A., Der C. J. Real-time in vitro measurement of GTP hydrolysis. Methods. 2005 Oct;37(2):183-89

Simmer F., Moorman C., van der Linden A. M., Kuijk E., van der Berghe P. V., Kamath R. S., Fraser A. G., Ahringer J., Plasterk R. H. Genome-wide RNAi of C. elegans using the hypersensitive rrf-3 strain reveals novel gene functions. PLoS Biol. 2003 Oct;1(1):E12

Singh G., Rebbapragada I., Lykke-Andersen J. A competition between stimulators and antagonists of Upf complex recruitment governs human nonsense-mediated mRNA decay. PLoS Biol. 2008 Apr 29;6(4):e111

Singh G., Kucukural A., Cenik C., Leszyk J. D., Shaffer S. A., Weng Z., Moore M. J. The cellular EJC interactome reveals higher-order mRNP structure and an EJC-SR protein nexus. Cell. 2012 Nov 9;151(4):750-64

Siniossoglou S., Wimmer C., Rieger M., Doye V., Tekotte H., Weise C., Emig S., Segref A., Hurt E. C. A novel complex of nucleoporins, which includes Sec13p and a Sec13p homolog, is essential for normal nuclear pores. Cell. 1996 Jan 26;84(2):265- 75

Sloan K. E., Mattijssen S., Lebaron S., Tollervey D., Pruijn G. J., Watkins N. J. Both endonucleolytic and exonucleolytic cleavage mediate ITS1 removal during human ribosomal RNA processing. J Cell Biol. 2013 Mar 4;200(5):577-88

175

Song H., Mugnier P., Das A. K., Webb H. M., Evans D. R., Tuite M. F., Hemmings B. A., Barford D. The crystal structure of human eukaryotic release factor eRF1-- mechanism of stop codon recognition and peptidyl-tRNA hydrolysis. Cell. 2000 Feb 4;100(3):311-21

St. Johnston D., Beuchle D., Nüsslein-Volhard C. Staufen, a gene required to localize maternal RNAs in the Drosophila egg. Cell. 1991 Jul 12;66(1):51-63

Stagg S. M., Gürkan C., Fowler D. M., LaPointe P., Foss T. R., Potter C. S., Carragher B., Balch W. E. Structure of the Sec13/31 COPII coat cage. Nature. 2006 Jan 12;439(7073):234-8

Stagg S. M., LaPointe P., Razvi A., Gürkan C., Potter C. S., Carragher B., Balch W. E. Structural basis for cargo regulation of COPII coat assembly. Cell. 2008 Aug 8;134(3):474-84

Sun X., Perlick H. A., Dietz H. C., Maquat L. E. A mutated human homologue to yeast Upf1 protein has a dominant-negative effect on the decay of nonsense- containing mRNAs in mammalian cells. Proc Natl Acad Sci USA. 1998 Aug 18;95(17):10009-14

Svasti S., Boonchoy C., Vanichsetakul P., Winichagoon P., Fucharoen S. Molecular mechanism of beta-thalassaemia caused by 22-bp duplication. Ann Hematol. 2008 Aug;87(8):633-7

Szyszka P., Sharp S. I., Dedman A., Gurling H. M., McQuillin A. A nonconservative amino acid change in the UPF3B gene in a patient with schizophrenia. Psychiatr Genet. 2012 Jun;22(3):150-1

176

Tafforeau L., Zorbas C., Langhendries J. L., Mullineux S. T., Stamatopoulou V., Mullier R., Wacheul L., Lafontaine D. L. The complexity of human ribosome biogenesis revealed by systematic nucleolar screening of Pre-rRNA processing factors. Mol Cell. 2013 Aug 22;51(4):539-51

Talkish J., Zhang J., Jakovljevic J., Horsey E. W., Woolford J. L. Jr. Hierarchical recruitment into nascent ribosomes of assembly factors required for 27SB pre-rRNA processing in Saccharomyces cerevisiae. Nucleic Acids Res. 2012 Sep 1;40(17):8646-61

Tang B. L., Kausalya J., Low D. Y., Lock M. L., Hong W. A family of mammalian proteins homologous to yeast Sec24p. Biochem Biophys Res Commun. 1999 May 19;258(3):679-84

Tange T. Ø., Shibuya T., Jurica M. S., Moore M. J. Biochemical analysis of the EJC reveals two new factors and a stable tetrameric protein core. RNA. 2005 Dec;11(12):1869-83

Tarpey P. S., Raymond F. L., Nguyen L. S., Rodriguez J., Hackett A., Vandeleur L., Smith R., Shoubridge C., Edkins S., Stevens C., O'Meara S., Tofts C., Barthorpe S., Buck G., Cole J., Halliday K., Hills K., Jones D., Mironenko T., Perry J., Varian J., West S., Widaa S., Teague J., Dicks E., Butler A., Menzies A., Richardson D., Jenkinson A,. Shepherd R., Raine K., Moon J., Luo Y., Parnau J., Bhat S. S., Gardner A., Corbett M., Brooks D., Thomas P., Parkinson-Lawrence E., Porteous M. E., Warner J. P., Sanderson T., Pearson P., Simensen R. J., Skinner C., Hoganson G., Superneau D., Wooster R., Bobrow M., Turner G., Stevenson R. E., Schwartz C. E., Futreal P. A., Srivastava A. K., Stratton M. R., Gécz J. Mutations in UPF3B, a member of the nonsense-mediated mRNA decay complex, cause syndromic and nonsyndromic mental retardation. Nat Genet. 2007 Sep;39(9):1127-33

177

Thein S. L., Hesketh C., Taylor P., Temperley I. J., Hutchinson R. M., Old J. M., Wood W. G., Clegg J. B., Weatherall D. J. Molecular basis for dominantly inherited inclusion body beta-thalassemia. Proc Natl Acad Sci U S A. 1990 May;87(10):3924- 8

Thermann R., Neu-Yilik G., Deters A., Frede U., Wehr K., Hagemeier C., Hentze M. W., Kulozik A. E. Binary specification of nonsense codons by splicing and cytoplasmic translation. EMBO J. 1998 Jun 15;17(12):3484-94

Thomas J. H. Genetic analysis of defecation in Caenorhabditis elegans. Genetics. 1990 Apr;124(4):885-72

Tucker M., Valencia-Sanchez M. A., Staples R. R., Chen J., Denis C. L., Parker R. The transcription factor associated Ccr4 and Caf1 proteins are components of the major cytoplasmic mRNA deadenylase in Saccharomyces cerevisiae. Cell. 2001 Feb 9;104(3):377-86

Uchida N., Hoshino S., Katada T. Identification of a human cytoplasmic poly(A) nuclease complex stimulated by poly(A)-binding protein. J Biol Chem. 2004 Jan 9;279(2):1383-91

Uhlén M., Fagerberg L., Hallström B M., Lindskog C., Oksvold P., Mardinoglu A., Sivertsson Å., Kampf C., Sjöstedt E., Asplund A., Olsson I., Edlund K., Lundberg E., Navani S., Szigyarto C. A., Odeberg J., Djureinovic D., Takanen J. O., Hober S., Alm T., Edqvist P. H., Berling H., Tegel H., Mulder J., Rockberg J., Nilsson P., Schwenk J. M., Hamsten M., von Feilitzen K., Forsberg M., Persson L., Johansson F., Zwahlen M., von Heijne G., Nielsen J., Pontén F. Proteomics. Tissue-based map of the human proteome. Science. 2015 Jan 23;347(6220):1260419

Unterholzner L., Izaurralde E. SMG7 acts as a molecular link between mRNA surveillance and mRNA decay. Mol Cell. 2004 Nov 19;16(4):587-96

178 van Dijk E., Cougot N., Meyer S., Babajko S., Wahle E., Séraphin B. Human Dcp2: a catalytically active mRNA decapping enzyme located in specific cytoplasmic structures. EMBO J. 2002 Dec 16;21(24):6915-24 van Hoof A., Frischmeyer P. A., Dietz H. C., Parker R. Exosome-mediated recognition and degradation of mRNAs lacking a termination codon. Science. 2002 Mar 22;295(5563):2262-4

Villacé P., Marión R. M., Ortín J. The composition of Staufen-containing RNA granules from human cells indicates their role in the regulated transport and translation of messenger RNAs. Nucleic Acids Res. 2004 Apr 30;32(8):2411-20

Wang Z., Kiledjian M. Functional link between the mammalian exosome and mRNA decapping. Cell. 2001 Dec 14;107(6):751-62

Wang J., Gudikote J. P., Olivas O. R., Wilkinson M. F. Boundary-independent polar nonsense-mediated decay. EMBO Rep. 2002 Mar;3(3):274-9

Wang H. W., Wang J., Ding F., Callahan K., Bratkowski M. A., Butler J. S., Nogales E., Ke A. Architecture of the yeast Rrp44 exosome complex suggests routes of RNA recruitment for 3' end processing. Proc Natl Acad Sci U S A. 2007 Oct 23;104(43):16844-9

Wang E. T., Sandberg R., Luo S., Khrebtukova I., Zhang L., Mayr C., Kingsmore S. F., Schroth G. P., Burge C. B. Alternative isoform regulation in human tissue transcriptomes. Nature. 2008 Nov 27;456(7221):470-6

Weischenfeldt J., Damgaard I., Bryder D., Theilgaard-Mönch K., Thoren L. A., Nielsen F. C., Jacobsen S. E., Nerlov C., Porse B. T. NMD is essential for hematopoietic stem and progenitor cells and for eliminating by-products of programmed DNA rearrangements. Genes Dev. 2008 May 15;22(10):1381-96

179

West S., Proudfoot N. J. Transcriptional termination enhances protein expression in human cells. Mol Cell. 2009 Feb 13;33(3):354-64

Whittle J. R., Schwartz T. U. Structure of the Sec13-Sec16 edge element, a template for assembly of the COPII vesicle coat. J Cell Biol. 2010 Aug 9;190(3):347-61

Wiegand H. L., Lu S., Cullen B. R. Exon junction complexes mediate the enhancing effect of splicing on mRNA expression. Proc Natl Acad Sci U S A. 2003 Sep 30

Wilson M. A., Meaux S., van Hoof A. A genomic screen in yeast reveals novel aspects of nonstop mRNA metabolism. Genetics. 2007 Oct;177(2):773-84

Wittkopp N., Huntzinger E., Weiler C., Saulière J., Schmidt S., Sonawane M., Izaurralde E. Nonsense-mediated mRNA decay effectors are essential for zebrafish embryonic development and survival. Mol Cell Biol. 2009 Jul;29(13):3517-28

Wittmann J., Hol E. M., Jäck H. M. hUPF2 silencing identifies physiologic substrates of mammalian nonsense-mediated mRNA decay. Mol Cell Biol. 2006 Feb;26(4):1272-87.

Wolf J., Valkov E., Allen M. D., Meineke B., Gordiyenko Y., McLaughlin S. H., Olsen T. M., Robinson C. V., Bycroft M., Stewart M., Passmore L. A. Structural basis for Pan3 binding to Pan2 and its function in mRNA recruitment and deadenylation. EMBO J. 2014 Jul 17;33(14):1514-26

Wu J., Kang J. H., Hettenhausen C., Baldwin I. T. Nonsense-mediated mRNA decay (NMD) silences the accumulation of aberrant trypsin proteinase inhibitor mRNA in Nicotiana attenuata. Plant J. 2007 Aug;51(4):693-706

Xu K., Bai Y., Zhang A., Zhang Q., Bartlam M. G. Insights into the structure and architecture of the CCR4-NOT complex. Front Genet. 2014 May 16;5:137

180

Xu X., Zhang L., Tong P., Xun G., Su W., Xiong Z., Zhu T., Zheng Y., Luo S., Pan Y., Xia K., Hu Z. Exome sequencing identifies UPF3B as the causative gene for a Chinese non-syndrome mental retardation pedigree. Clin Genet. 2013 Jun;83(6):560-4

Yamashita A., Ohnishi T., Kashima I., Taya Y., Ohno S. Human SMG-1, a novel phosphatidylinositol 3-kinase-related protein kinase, associates with components of the mRNA surveillance complex and is involved in the regulation of nonsense- mediated mRNA decay. Genes Dev. 2001 Sep 1;15(17):2215-28

Yamashita A., Chang T. C., Yamashita Y., Zhu W., Zhong Z., Chen C. Y., Shyu A. B. Concerted action of poly(A) nucleases and decapping enzyme in mammalian mRNA turnover. Nat Struct Mol Biol. 2005 Dec;12(12):1054-63

Yamashita A., Izumi N., Kashima I., Ohnishi T., Saari B., Katsuhata Y., Muramatsu R., Morita T., Iwamatsu A., Hachiya T., Kurata R., Hirano H., Anderson P., Ohno S. SMG-8 and SMG-9, two novel subunits of the SMG-1 complex, regulate remodeling of the mRNA surveillance complex during nonsense-mediated mRNA decay. Genes Dev. 2009 May 1;23(9):1091-105

Yepiskoposyan H., Aeschimann F., Nilsson D., Okoniewski M., Mühlemann O. Autoregulation of the nonsense-mediated mRNA decay pathway in human cells. RNA. 2011 Dec;17(12):2108-18

Yoine M., Ohto M. A., Onai K., Mita S., Nakamura K. The lba1 mutation of UPF1 RNA helicase involved in nonsense-mediated mRNA decay causes pleiotropic phenotypic changes and altered sugar signalling in Arabidopsis. Plant J. 2006 Jul;47(1):49-62

181

Yue Z., Maldonado E., Pillutla R., Cho H., Reinberg D., Shatkin A. J. Mammalian capping enzyme complements mutant Saccharomyces cerevisiae lacking mRNA guanylyltransferase and selectively binds the elongating form of RNA polymerase II. Proc Natl Acad Sci U S A. 1997 Nov 25;94(24):12898-903

Yuryev A., Patturajan M., Litingtung Y., Joshi R. V., Gentile C., Gebara M., Corden J. L. The C-terminal domain of the largest subunit of RNA polymerase II interacts with a novel set of serine/arginine-rich proteins. Proc Natl Acad Sci U S A. 1996 Jul 9;93(14):6975-80

Zhang J., Sun X., Qian Y., Maquat L. E. Intron function in the nonsense-mediated decay of beta-globin mRNA: indications that pre-mRNA splicing in the nucleus can influence mRNA translation in the cytoplasm. RNA. 1998a Jul;4(7):801-15

Zhang J., Sun X., Qian Y., LaDuca J. P., Maquat L. E. At least one intron is required for the nonsense-mediated decay of triosephosphate isomerase mRNA: a possible link between nuclear splicing and cytoplasmic translation. Mol Cell Biol. 1998b Sep;18(9):5272-83

Zhang J., Harnpicharnchai P., Jakovljevic J., Tang L., Guo Y., Oeffinger M., Rout M. P., Hiley S. L., Hughes T., Woolford J. L. Jr. Assembly factors Rpf2 and Rrs1 recruit 5S rRNA and ribosomal proteins rpL5 and rpL11 into nascent ribosomes. Genes Dev. 2007 Oct 15;21(20):2580-92

Zhouravleva G., Frolova L., Le Goff X., Le Guellec R., Inge-Vechtomov S., Kisselev L., Philippe M. Termination of translation in eukaryotes is governed by two interacting polypeptide chain release factors, eRF1 and eRF3. EMBO J. 1995 Aug 15;14(16):4065-72

182

Zou D., McSweeney C., Sebastian A., Reynolds D. J., Dong F., Zhou Y., Deng D., Wang Y., Liu L., Zhu J., Zou J., Shi Y., Albert I., Mao Y. A critical role of RBM8a in proliferation and differentiation of embryonic neural progenitors. Neural Dev. 2015 Jun 21;10:18

183

Appendix: Published paper

Part of the work presented in this thesis has been published in the following article, which is attached as an appendix:

146

Published online: December 1, 2014

Scientific Report

Identification and characterization of novel factors that act in the nonsense-mediated mRNA decay pathway in nematodes, flies and mammals

Angela Casadio1, Dasa Longman1,*, Nele Hug1, Laurent Delavaine1, Raúl Vallejos Baier2, Claudio R Alonso2 & Javier F Cáceres1,**

Abstract insects, plants and mammals [reviewed by 4,5]. A genome-wide RNAi screen in C. elegans identified two additional NMD factors that Nonsense-mediated mRNA decay (NMD) is a surveillance mecha- are conserved throughout evolution and, unlike the core smg-1-7 nism that degrades mRNAs harboring premature termination genes, are essential for embryonic development [6]. Accordingly, codons (PTCs). We have conducted a genome-wide RNAi screen in they were termed smgl-1 and smgl-2 (for smg-lethal-1 and 2, respec- Caenorhabditis elegans that resulted in the identification of five tively). Their human homologs, NBAS (for neuroblastoma amplified novel NMD genes that are conserved throughout evolution. Two of sequence) and DHX34, act in concert with core NMD factors to co- their human homologs, GNL2 (ngp-1) and SEC13 (npp-20), are also regulate a large number of endogenous RNA targets [7]. required for NMD in human cells. We also show that the C. elegans The ATP-dependent RNA helicase, UPF1/SMG2, is a central gene noah-2, which is present in Drosophila melanogaster but NMD factor and undergoes cycles of phosphorylation and dephos- absent in humans, is an NMD factor in fruit flies. Altogether, these phorylation that are essential for its activity. UPF1 is phosphory- data identify novel NMD factors that are conserved throughout lated at multiple [S/T]Q motifs at its C- and N-terminus by the evolution, highlighting the complexity of the NMD pathway and SMG1 complex, which contains the protein kinase SMG1 and the suggesting that yet uncovered novel factors may act to regulate SMG8,9 subunits [3]. NMD is initiated by the assembly of the SURF this process. complex, comprising SMG1, UPF1 and the translation release factors eRF1 and eRF3, in the vicinity of a PTC. Subsequently, an Keywords C. elegans; nonsense-mediated decay; RNAi screen; smg genes interaction of this complex with an exon junction complex (EJC), Subject Categories RNA Biology deposited downstream as a consequence of the splicing process, DOI 10.15252/embr.201439183 | Received 17 June 2014 | Revised 30 October leads to the formation of the decay-inducing complex (DECID) that 2014 | Accepted 3 November 2014 | Published online 1 December 2014 results in mRNA degradation [8]. The interaction of the SURF EMBO Reports (2015) 16: 71–78 complex with the EJC allows the binding of UPF2 to the N-terminal domain of UPF1 resulting in a large conformational change that activates the UPF1 helicase activity [9, 10]. UPF1 dephosphoryla- Introduction tion is carried out at a later stage and requires the activity of SMG5-7 together with protein phosphatase 2A (PP2A) [reviewed The NMD pathway targets mRNAs harboring premature termination by 11]. codons (PTCs) for degradation, but also regulates the stability of a In order to establish whether there are more factors that could wide array of endogenous transcripts [reviewed by 1–3]. Genetic regulate the NMD pathway within the context of a multicellular screens in the nematode Caenorhabditis elegans resulted in the iden- organism, we carried out a genome-wide RNAi screen in tification of seven genes required for NMD, termed smg-1-7 (for C. elegans that builds on the success of our previous effort [6], but suppressor with morphological effect on genitalia). Likewise, a simi- using a different RNAi library that included many previously lar approach in Saccharomyces cerevisiae identified three NMD untested genes. We identified five novel NMD genes that are genes, UPF1-3 (for up-frameshift), that are orthologs of C. elegans highly conserved throughout evolution and demonstrate that they smg-2, smg-3 and smg-4, respectively. Subsequently, orthologs for participate in the NMD pathway in human cells and Drosophila all the smg genes were identified in several species including embryos.

1 MRC Human Genetics Unit, Institute of Genetics and Molecular Medicine, Western General Hospital, University of Edinburgh, Edinburgh, UK 2 School of Life Sciences, University of Sussex, Brighton, UK *Corresponding author. Tel: +44 131 332 2471; Fax: +44 131 467 8456; E-mail: [email protected] **Corresponding author. Tel: +44 131467 8426; Fax: +44 131 467 8456; E-mail: [email protected]

ª 2014 The Authors. Published under the terms of the CC BY 4.0 license EMBO reports Vol 16 |No1 | 2015 71 Published online: December 1, 2014 EMBO reports Novel NMD factors in nematodes, flies and mammals Angela Casadio et al

Results and Discussion that their silencing by RNAi restores GFP expression. Accordingly, we searched for the appearance of green worms, dead or alive, A genome-wide RNAi screen to identify new genes required for NMD following inactivation of individual genes by RNAi. As a negative control, we fed PTCxi animals empty RNAi vector, which had no Our previous RNAi screen in C. elegans led to the identification of effect on the level of GFP expression (Fig 2A, panel I), whereas inac- smgl-1/NBAS and smgl-2/DHX34 that act in NMD in nematodes and tivation of the core NMD factor, SMG-2/UPF1, which induced strong vertebrates [6,12]. We revisited this approach with the use of a GFP expression, was used as a positive control (Fig 2A, panel II). different RNAi library: the C. elegans ORF-RNAi library v.1.1 that Screening of the entire library resulted in the identification of five contains 11,511 clones targeting 55% of the nematode genome [13]. RNAi clones that scored positive by increased GFP expression This library includes dsRNAs against 1,736 genes that were not (Fig 2A, panels III–VII). The clones identified in this screen are: targeted previously [14]. As earlier, we used the C. elegans PTCxi T19A6.2, Y77E11A.13, Y87G2A.4, C47B2.4 and F52B11.3 (Table 1). strain that expresses a GFP-based reporter harboring a PTC that is Confirming that these newly identified C. elegans genes act as NMD integrated in the genome [6] (Fig 1). This PTCx reporter has reduced factors, quantitative RT–PCR analysis showed that downregulation GFP expression, since its transcript is subject to NMD-mediated of each of these genes led to an increase in the GFP reporter mRNA degradation. Thus, novel NMD genes were identified by the criterion level, as was seen with depletion of smg-2 (Fig 2B).

ATG PTC * PTCx G1 G2 G3 G4 L1 L2 L9 L10 L11

Bacterial ORF-RNAi library v1.1 PTCxi

ORF

Collect embryos and hatch O/N at 20°C AmpR Grow O/N at 37°C

Induce dsRNAs expression (IPTG, 4 h) Aliquot 25 synchronized L1 larvae per well

20°C for 60-72 h

Score GFP expression

NMD factor

Figure 1. A genome-wide RNAi screen designed to identify novel NMD factors in C. elegans. The PTCx NMD reporter consists of a fusion between GFP (in green) and LacZ (in blue) genes; a PTC is present downstream of the GFP coding region, making the transcript an NMD substrate.

72 EMBO reports Vol 16 |No1 | 2015 ª 2014 The Authors Published online: December 1, 2014 Angela Casadio et al Novel NMD factors in nematodes, flies and mammals EMBO reports

A Brightfield GFP Brightfield GFP

i I iv IV

empty vector npp-20

ii II v V

smg-2 aex-6

iii III vi VI

ngp-1 pbs-2 vii VII

noah-2

B 6

5 * * 4 * * 3 2 * * 1

0 Relative GFP mRNA expression mRNA Relative GFP pbs-2 aex-6 ngp-1 smg-2 noah-2 npp-20 empty vector

Figure 2. Newly identified NMD factors. A RNAi was induced with an empty vector as a negative control (panel I), whereas a smg-2 clone (panel II) was used as a positive control. Panels i and ii show brightfield images of the PTCxi strain treated with the negative and positive controls, respectively. Depletion of five genes (panels III to VII) resulted in increased GFP expression. Panels iii to vii show brightfield images of the phenotypes of the affected worms. The scale bars correspond to 100 lm. B Depletion of the novel NMD genes in C. elegans leads to upregulation of the PTCx NMD reporter mRNA, which was monitored by quantitative RT–PCR relative to the expression of ama-1 reference gene. The values shown are the average fold-change (mean Æ SEM) from at least three independent experiments relative to empty vector-depleted worms. Statistical analysis was performed using the Mann–Whitney U-test for non-parametric distributions. *P < 0.05.

Novel NMD genes in C. elegans gene and encodes a putative GTPase that comprises a GTP-binding domain formed by five G-motifs, which is typical of the HSR1_MMR1 All of the newly identified genes, with the exception of noah-2, are GTP-binding protein subfamily. It also contains a conserved N- conserved throughout evolution and have clear orthologs in human, terminal domain (NGP1NT) (Supplementary Fig S1A). Its yeast mouse, zebrafish and yeast (Supplementary Figs S1 and S2). The homolog, Nog2p, is involved in ribosomal biogenesis playing a role C. elegans gene ngp-1 (T19A6.2) corresponds to the human GNL2 in the processing of the pre-60S particles [15]. The npp-20 gene

ª 2014 The Authors EMBO reports Vol 16 |No1 | 2015 73 Published online: December 1, 2014 EMBO reports Novel NMD factors in nematodes, flies and mammals Angela Casadio et al

Table 1. List of novel putative NMD factors identified in this study. Gene name Predicted function in C. elegans Gene name Clone ID (C. elegans) C. elegans phenotype (H. sapiens) T19A6.2 ngp-1 Nuclear/nucleolar GTP-binding protein family Embryonic lethal GNL2 Larval arrest Maternal sterile Y77E11A.13 npp-20 Nuclear pore complex protein Embryonic lethal SEC13 Larval arrest Y87G2A.4 aex-6 Rab protein involved in trafficking of vesicles Aboc expulsion missing RAB27A Constipated RAB27B C47B2.4 pbs-2 Proteasome p subunit Embryonic lethal PSMB7 Larval arrest PSMB10 F52B11.3 noah-2 PAN and ZP domain-containing protein Embryonic lethal Not conserved Larval arrest Conserved in Drosophila (nompA)

(Y77E11A.13) corresponds to human SEC13, which encodes a protein the affected worms (Fig 2A, panel vi). Finally, depletion of noah-2 that comprises six WD-40 domains (Supplementary Fig S1B) and is a led to an early larval arrest at L2–L3 stages and subsequent larval constituent of the endoplasmic reticulum and the nuclear pore lethality (Fig 2A, panel vii). complex (NPC) [16]. The aex-6 gene (Y87G2A.4) is a member of the Rab small GTPase superfamily. It has two homologs in humans, Drosophila nompA gene is required for NMD RAB27A and RAB27B (Supplementary Fig S1C), with RAB27B functioning in the trafficking of dense-core vesicles [17]. The pbs-2 gene There is no ortholog of noah-2 in mammalian genomes; nonethe- (C47B2.4) is a member of the proteasome B-type family and is a 20S less, the gene is clearly present in Drosophila melanogaster core beta subunit of the proteasome (Supplementary Fig S1D), with (Supplementary Fig S2), suggesting that it most likely emerged at two human homologs, PSMB7 and PSMB10 [18]. Finally, the noah-2 some point during the early evolution of the Ecdyzozoa before the gene (F52B11.3) encodes a PAN and ZP domain-containing protein split between arthropods and nematodes. We assessed its potential that is required for embryonic and larval development, reproduction, role in NMD in Drosophila embryos by means of a previously coordinated locomotion and molting (Supplementary Fig S1E) [19]. It described NMD fluorescent GFP reporter [22,23]. Unlike the previ- is related to the Drosophila extracellular matrix component nompA ously characterized Drosophila NMD genes upf1, upf2 and upf3 (no-mechanoreceptor-potential A) [20]. There are no homologs of that are ubiquitously expressed [22,24], expression of nompA in noah-2 in vertebrates (Supplementary Fig S2). Drosophila embryos is confined to type I sense organs of the peripheral nervous system (PNS) [20]. Thus, we used the UAS/ The newly identified NMD genes are required for proper Gal4 system [25] to drive expression of a nompA RNAi construct development in C. elegans within the PNS using a NompA-Gal4 line [26]. We found that nompA knockdown within its expression domain in the embryonic Depletion of all these novel genes resulted in developmental defects, PNS led to a significant upregulation of the NMD reporter (Fig 3C, in contrast to smg-2 depletion that did not compromise development F and G), which was comparable to the effect observed following (Fig 2A, compare panels ii with iii–vii). Thus, these novel NMD UPF1 depletion (Fig 3B, E and G), as compared to no RNAi treat- factors are different from core smg-1-7 genes and display similar ment (Fig 3A, B and D), which showed no expression of the behavior to smgl-1, 2 that are essential for viability [6]. In reporter. This demonstrates that the Drosophila ortholog of noah-2 C. elegans, ngp-1 is an essential gene. Its knockdown led to a vari- behaves as a tissue-specific NMD factor in fruit fly embryos. able larval arrest, where the majority of the affected worms were arrested at L1–L2 stages, compared to control worms that invariably GNL2 and SEC13 act in the NMD pathway in human cells reached adulthood within the time limit of the experiment (Fig 2A, panels i and iii, respectively). Those worms that escaped early arrest Next, we investigated a potential role for the human homologs of the failed to reach adulthood and produced no embryos. Depletion of factors identified in this screen in NMD in human cells. HeLa cells npp-20 resulted in worms arrested at L2–L3 larval stages. The stably expressing an integrated human b-globin (HBB) gene, either in majority of the worms were very fragile and died by bursting awild-typeversionorcarryinganNMD-inducingmutation(NS39) (Fig 2A, panel iv). By contrast, depletion of aex-6 resulted in a mild [27], were individually depleted of each of these genes. The level of but highly consistent phenotype with worms able to progress depletion of these factors is shown in Supplementary Fig S3C. As through the developmental stages normally; however, adult worms expected, depletion of the human homologs of the novel NMD factors were constipated, as previously reported [21], and also exhibited an did not significantly affect the levels of the wild-type b-globin mRNA egg-laying defect and reduced brood size (Fig 2A, panel v). Deple- (Fig 4A). By contrast, depletion of UPF2 (positive control) or of GNL2 tion of pbs-2 resulted in very sick and pale larvae that were arrested and SEC13, but not of RAB27A-B (depleted individually or in combi- around L2 stage, displaying a swollen intestine in the majority of nation), resulted in a significantly increased level of the b-globin

74 EMBO reports Vol 16 |No1 | 2015 ª 2014 The Authors Published online: December 1, 2014 Angela Casadio et al Novel NMD factors in nematodes, flies and mammals EMBO reports

Wild-type (no RNAi) Upf1 RNAi nompA RNAi A B C G

400

D A4 A5 E A4 A5 F A4 A5 300 200

100

0 WT Upf1 nompA DAPI DAPI DAPI Relative NMD sensor activity (%) RNAi RNAi 22C10 22C10 22C10 GFP-NMD sensor GFP-NMD sensor GFP-NMD sensor Quantification

Figure 3. Drosophila nompA is required for NMD in Drosophila embryos. A–C Embryos expressing a GFP-NMD sensor (green) under the control of nompA regulatory sequences show signal in support cells linked to the embryonic peripheral nervous system (PNS) [labeled by 22C10 signal (magenta)]. DAPI signal is shown in blue. Expression of UAS-RNAi constructs against Upf1 (B) or nompA (C) genes using a nompA-Gal4 driver leads to upregulation of the GFP-NMD sensor when compared to wild-type (no RNAi) (A), revealing a reduction in NMD activity in the knockdown conditions. Scale bars represent 100 lm. D–F Higher magnification (40×) of the areas marked by a rectangle in (A–C) further illustrates the upregulation of the GFP-NMD sensor in Upf1- (E) and NompA- depleted cells (F). Optical fields include embryonic abdominal segments A4–A5. Scale bars represent 10 lm. G Quantification of GFP signal in cells marked by an arrow in panels (D–F) shows a significant upregulation of NMD sensor expression upon downregulation of Upf1 (dark gray) and NompA (black) compared to wild-type (no RNAi) (light gray). Results represent the average of five biological replicates (mean Æ SEM). Pair-wise comparisons were performed using a one-tailed t-test (non-parametric) between treatments and wild-type. ***P < 0.001.

NMD reporter (NS39) mRNA when compared to mock-depleted cells UPF1. We immunopurified Flag-tagged UPF1 expressed at physio- (Fig 4B). Whereas individual depletion of PSMB10 clearly showed no logical levels from transiently transfected HEK 293T cells that also effect on the levels of the NMD reporter, knockdown of PSMB7 (either co-expressed T7-tagged GNL2 in the presence of RNase A. We individually or in combination with PSMB10) led to an upregulation used transiently expressed Flag-empty vector (F-EV) co-expressed of both the wild-type and NMD reporters, making it difficult to with T7-tagged GNL2, as a negative control. We observed that conclude whether PSMB7 had a specific role in NMD (Supplementary UPF1 specifically co-immunoprecipitated with T7-tagged GNL2 in Fig S3A and B). Altogether, these experiments show that GNL2 and an RNA-independent manner (Supplementary Fig S4). Future stud- SEC13 have a clear effect in the NMD response in human cells, ies will aim to test the interaction of GNL2 and SEC13 with whereas it still remains possible that the remaining tested genes may components of the NMD machinery. have an NMD effect that is substrate or tissue specific. To rule out indirect effects of GNL2 and SEC13, we first inves- GNL2 and SEC13 participate in an autoregulatory feedback loop tigated whether these factors have a general role in mRNA translation, which would impact on NMD. This is unlikely, since the Transcripts encoding NMD factors are sensitive to depletion of very nature of the RNAi screen in C. elegans requires that the different NMD factors as part of a negative feedback regulatory loop NMD reporter is indeed translated. In agreement, knockdown of that acts to tightly control NMD homeostasis [29,30]. We investi- GNL2 or SEC13 did not result in a general inhibition of gated whether depletion of GNL2 or SEC13 would have an impact translation, as measured by metabolic labeling of HeLa cells on the levels of transcripts encoding NMD factors in human cells. (Supplementary Fig S3E and F). Next, we examined the half-life of Interestingly, we found that depletion of GNL2 in HeLa cells resulted the wild-type or NMD-sensitive NS39 b-globin reporter mRNAs in a significant upregulation of the levels of SMG5 mRNA, as well as upon depletion of GNL2 and SEC13. The stability of wild-type increased mRNA levels for UPF1, UPF2 and SMG1 (Fig 4F). Simi- b-globin mRNA was unaffected by GNL2 or SEC13 depletion larly, SEC13 depletion resulted in a significant upregulation in the (Fig 4C). By contrast, depletion of GNL2 or SEC13 led to a marked levels of mRNAs encoding UPF1 and UPF2, and to a lesser extent of stabilization of a PTC-containing b-globin mRNA, confirming that SMG5 mRNA (Fig 4F). Thus, the novel NMD factors, GNL2 and both GNL2 and SEC13 act in the NMD pathway (Fig 4D). Further- SEC13, participate in a negative regulatory feedback loop controlling more, depletion of GNL2 or SEC13 in HeLa cells led to a marked the expression of NMD factor mRNAs. upregulation of three endogenous transcripts that were previously reported to be sensitive to NMD regulation [7, 28] (Fig 4E). In agreement, knockdown of GNL2 or SEC13 also resulted in an Conclusions increased half-life of one of those NMD substrates (ARHGEF18) mRNA (Supplementary Fig S3D). As further proof of the role of Even though the NMD pathway is a highly conserved process, the novel factors identified in this screen in the NMD pathway, several mechanisms have evolved to define a PTC across different we probed for the interaction of GNL2 with the core NMD factor species [3]. Whereas in mammalian cells, NMD is linked to

ª 2014 The Authors EMBO reports Vol 16 |No1 | 2015 75 76 Published online:December1,2014 Figure reports EMBO MOreports EMBO 4 . Vol A C EF

16 Relative expression % HBB mRNA remaining Relative HBB mRNA Expression 10 10 12 |No 0 1 2 3 4 5 6 7 8 9 0 2 4 6 8 wt β-globin 1 |

RGF8TC BMP2 TMC7 ARHGEF18 Mock 2015 ***

ActD treatment (h) Scramble siRNA ** mock si wt β-globin GNL2 ** siUPF2 ** siGNL2

siSEC13 *** mRNA si si

** siRAB27A SEC13 UPF1

* siRAB27B ** siSEC13 siGNL2 ~53h t½ Mock siRAB27A + siRAB27B *** si ½~n.d. t½ UPF2 ½~34h t½ * ** *** oe M atr nnmtds le n mammals and flies nematodes, in factors NMD Novel B D Relative fold-change % HBB mRNA remaining Relative HBB mRNA Expression 0 1 2 3 10 12 0 2 4 6 8 NS39 β-globin

P1UF M1SMG5 SMG1 UPF2 UPF1 Mock

* Scramble siRNA ActD treatment (h) si GNL2 NS39 β-globin * PTC siUPF2

siGNL2 * mRNA * siSEC13 * si

SEC13 siRAB27A siRAB27B siGNL2 ~2h t½ Mock siSEC13 siRAB27A + siRAB27B * ½~15h t½ ½~9h t½ neaCsdoe al et Casadio Angela ª 2014 h Authors The Published online: December 1, 2014 Angela Casadio et al Novel NMD factors in nematodes, flies and mammals EMBO reports

◀ Figure 4. GNL2 and SEC13 are required for NMD in human cells. A, B HeLa cells stably expressing a wild-type b-globin reporter (A) or a b-globin NS39 NMD reporter (B) were mock-depleted or depleted of UPF2, GNL2, SEC13, RAB27A, RAB27B or both RAB27A and RAB27B. The level of the b-globin mRNA was monitored by quantitative RT–PCR relative to two reference genes (POLR2J and ACTB). The values shown are the average fold-change (mean Æ SEM) from four independent experiments relative to mock-depleted cells (control). Statistical analysis was performed using the Mann–Whitney U-test for non-parametric distributions. *P < 0.05. The level of depletion of NMD factors is shown in Supplementary Fig S3C. C, D Analysis of the half-life of b-globin reporters. Samples were collected at the indicated time points, and the mRNA levels of the HBB reporters were monitored by qRT–PCR and normalized to POLR2J and ACTB reference genes. The values shown are the average fold-change (mean Æ SEM) from three independent experiments relative to the first time point. E Depletion of GNL2 and SEC13 leads to a significant upregulation in the mRNA levels of endogenous NMD substrates. Samples were analyzed as described in (A, B) *P < 0.05; ***P < 0.01, ***P < 0.001. F GNL2 and SEC13 contribute to the negative NMD feedback loop, regulating the levels of transcripts encoding NMD factors. RT– qPCR analysis of total cellular RNA from HeLa cells depleted of GNL2 (in green) and SEC13 (in purple) is shown. The graph shows distribution of relative fold-change from eight independent experiments relative to mock-depleted cells (control). Statistical analysis was performed using Student’s t-test. *P < 0.05.

pre-mRNA splicing, exon boundaries are not used to define PTCs in scored for the appearance of GFP expression, indicating that the other organisms, including S. cerevisiae [31], S. pombe [32,33], depleted protein is required for NMD in C. elegans. Drosophila [34] and C. elegans [6]. RNAi screens have been widely used in C. elegans to identify Fly stocks genes involved in many different cellular pathways, and we had used this approach in the past to identify novel NMD factors [6]. We used the following fly stocks all obtained from the Bloomington Here, we revisited this approach with the use of a different RNAi Stock Center (Indiana, USA): w1118;P{GMR29A10-GAL4}attP2, y1 w*; library that includes dsRNAs against 1,736 genes that were not P{UAS-mCD8::GFP.L}LL5 (BM 5137), y1 v1;P{TRiP.JF02919}attP2 targeted in our previous screen and have identified five novel and y1 v1; P{TRiP.GL01485}attP2. Animals were reared at 25°Con NMD genes that are required for proper development in nema- cornmeal, molasses and yeast medium. todes. Due to the high degree of evolutionary conservation, we could analyze the role of these newly identified NMD factors in For more detailed Materials and Methods see the Supplementary mammalian cells and we chose HeLa cells as our experimental Information. system. The only exception was the noah-2 gene that does not have a human counterpart. For this, we studied its functional Supplementary information for this article is available online: homolog in Drosophila and found that it acts in the NMD pathway http://embor.embopress.org in insects. Importantly, we show that two human homologs, GNL2 (ngp-1) and SEC13 (npp-20), are also required for NMD in human Acknowledgements cells. Only recently, we uncovered the mechanism by which the We thank Oliver Mühlemann for the gift of the HeLa b-globin cell lines. We RNA helicase DHX34, which was identified in our first RNAi thank Pedro Patraquim and Sofia Pinho for technical support. A.C. is the reci- screen, promotes mRNP remodeling and triggers the conversion pient of an MRC studentship. R.V.B. was supported by Becas Chile. This work from the SURF complex to the DECID complex resulting in NMD was supported by MRC core funding (JFC) and by grants from the Wellcome activation [35]. Further studies will help to delineate the mecha- Trust to J.F.C (095518/Z/11/Z) and to C.A. (098410/Z/12/Z). nism by which SEC13 and GNL2 activate NMD in human cells, as well as their involvement in the described alternative NMD Author contributions branches [36–38]. In summary, our work has led to the identifica- AC, DL and JFC conceived, designed and interpreted the experiments in nema- tion of novel NMD factors in nematodes, flies and mammals, todes and human cells. AC, DL, NH and LD performed the experiments and revealing that the machinery underlying NMD is more complex data analysis. RVB and CRA designed, performed and interpreted the experi- than previously thought. ments in Drosophila. The manuscript was co-written by all authors.

Conflict of interest Materials and Methods The authors declare that they have no conflict of interest.

Genome-wide RNAi screen References The NMD reporter is based on the GFP::lacZ vector pDP96.04 and is driven by the ubiquitous sec-23 promoter [6]. The PTCxi transgenic 1. Karam R, Wengrod J, Gardner LB, Wilkinson MF (2013) Regulation of strain carrying this GFP-based NMD reporter integrated in the nonsense-mediated mRNA decay: implications for physiology and genome was described previously [6]. PTCxi transgenic worms were disease. Biochim Biophys Acta 1829: 624 – 626 grown on standard NGM plates seeded with OP50 E. coli bacteria at 2. Popp MW-L, Maquat LE (2013) Organizing principles of mammalian 20°C. The RNAi library used for the screen was created in the labo- nonsense-mediated mRNA decay. Annu Rev Genet 47: 139 – 165 ratory of Marc Vidal and is commercially available [13]. RNAi was 3. Schweingruber C, Rufener SC, Zünd D, Yamashita A, Mühlemann O performed in liquid format by feeding synchronized population of (2013) Nonsense-mediated mRNA decay—mechanisms of substrate PTCxi L1 larvae with bacterial clones expressing dsRNA correspond- mRNA recognition and degradation in mammalian cells. Biochim Biophys ing to individual genes in 96-well plates [6]. Worms were then Acta 1829: 612 – 623

ª 2014 The Authors EMBO reports Vol 16 |No1 | 2015 77 Published online: December 1, 2014 EMBO reports Novel NMD factors in nematodes, flies and mammals Angela Casadio et al

4. Behm-Ansmant I, Kashima I, Rehwinkel J, Saulière J, Wittkopp N, 22. Avery P, Vicente-Crespo M, Francis D, Nashchekina O, Alonso CR, Izaurralde E (2007) mRNA quality control: an ancient machinery Palacios IM (2011) Drosophila Upf1 and Upf2 loss of function inhibits recognizes and degrades mRNAs with nonsense codons. FEBS Lett 581: cell growth and causes animal death in a Upf3-independent manner. 2845 – 2853 RNA 17: 624 – 638 5. Chang Y-F, Imam JS, Wilkinson MF (2007) The nonsense-mediated decay 23. Metzstein MM, Krasnow MA (2006) Functions of the nonsense-mediated RNA surveillance pathway. Annu Rev Biochem 76: 51 – 74 mRNA decay pathway in Drosophila development. PLoS Genet 2:e180 6. Longman D, Plasterk RHA, Johnstone IL, Cáceres JF (2007) Mechanistic 24. Alonso CR, Akam M (2003)AHox gene mutation that triggers insights and identification of two novel factors in the C. elegans NMD nonsense-mediated RNA decay and affects alternative splicing during pathway. Genes Dev 21: 1075 – 1085 Drosophila development. Nucleic Acids Res 31: 3873 – 3880 7. Longman D, Hug N, Keith M, Anastasaki C, Patton EE, Grimes G, Cáceres 25. Brand AH, Perrimon N (1993) Targeted gene expression as a means of JF (2013) DHX34 and NBAS form part of an autoregulatory NMD circuit altering cell fates and generating dominant phenotypes. Development that regulates endogenous RNA targets in human cells, zebrafish and 118: 401 – 415 Caenorhabditis elegans. Nucleic Acids Res 41: 8319 – 8331 26. Pfeiffer BD, Jenett A, Hammonds AS, Ngo T-TB, Misra S, Murphy C, Scully 8. Kashima I, Yamashita A, Izumi N, Kataoka N, Morishita R, Hoshino S, A, Carlson JW, Wan KH, Laverty TR, et al (2008) Tools for neuroanatomy Ohno M, Dreyfuss G, Ohno S (2006) Binding of a novel SMG-1-Upf1-eR- and neurogenetics in Drosophila. Proc Natl Acad Sci USA 105: 9715 – 9720 F1-eRF3 complex (SURF) to the exon junction complex triggers Upf1 27. Lykke-Andersen J, Shu MD, Steitz JA (2000) Human Upf proteins target phosphorylation and nonsense-mediated mRNA decay. Genes Dev 20: an mRNA for nonsense-mediated decay when bound downstream of a 355 – 367 termination codon. Cell 103: 1121 – 1131 9. Chamieh H, Ballut L, Bonneau F, Le Hir H (2008) NMD factors UPF2 and 28. Mendell JT, Sharifi NA, Meyers JL, Martinez-Murillo F, Dietz HC (2004) UPF3 bridge UPF1 to the exon junction complex and stimulate its RNA Nonsense surveillance regulates expression of diverse classes of mamma- helicase activity. Nat Struct Mol Biol 15: 85 – 93 lian transcripts and mutes genomic noise. Nat Genet 36: 1073 – 1078 10. Chakrabarti S, Jayachandran U, Bonneau F, Fiorini F, Basquin C, Domcke S, 29. Huang L, Lou C-H, Chan W, Shum EY, Shao A, Stone E, Karam R, Song H-W, Le Hir H, Conti E (2011) Molecular mechanisms for the RNA-dependent Wilkinson MF (2011) RNA homeostasis governed by cell type-specific and ATPase activity of Upf1 and its regulation by Upf2. Mol Cell 41: 693 – 703 branched feedback loops acting on NMD. Mol Cell 43: 950 – 961 11. Nicholson P, Mühlemann O (2010) Cutting the nonsense: the degrada- 30. Yepiskoposyan H, Aeschimann F, Nilsson D, Okoniewski M, Mühlemann O tion of PTC-containing mRNAs. Biochem Soc Trans 38: 1615 – 1620 (2011) Autoregulation of the nonsense-mediated mRNA decay pathway 12. Anastasaki C, Longman D, Capper A, Patton EE, Cáceres JF (2011) Dhx34 in human cells. RNA 17: 2108 – 2118 and Nbas function in the NMD pathway and are required for embryonic 31. Amrani N, Ganesan R, Kervestin S, Mangus DA, Ghosh S, Jacobson A development in zebrafish. Nucleic Acids Res 39: 3686 – 3694 (2004) A faux 3’-UTR promotes aberrant termination and triggers 13. Rual J-F, Ceron J, Koreth J, Hao T, Nicot A-S, Hirozane-Kishikawa T, nonsense-mediated mRNA decay. Nature 432: 112 – 118 Vandenhaute J, Orkin SH, Hill DE, van den Heuvel S, et al (2004)Toward 32. Brogna S, Wen J (2009) Nonsense-mediated mRNA decay (NMD) mecha- improving Caenorhabditis elegans phenome mapping with an ORFeome-based nisms. Nat Struct Mol Biol 16: 107 – 113 RNAi library. Genome Res 14: 2162 – 2168 33. Wen J, Brogna S (2010) Splicing-dependent NMD does not require the 14. Kamath RS, Fraser AG, Dong Y, Poulin G, Durbin R, Gotta M, Kanapin A, Le EJC in Schizosaccharomyces pombe. EMBO J 29: 1537 – 1551 Bot N, Moreno S, Sohrmann M, et al (2003) Systematic functional analysis 34. Gatfield D, Unterholzner L, Ciccarelli FD, Bork P, Izaurralde E (2003) of the Caenorhabditis elegans genome using RNAi. Nature 421: 231 – 237 Nonsense-mediated mRNA decay in Drosophila: at the intersection of 15. Saveanu C, Bienvenu D, Namane A, Gleizes PE, Gas N, Jacquier A, the yeast and mammalian pathways. EMBO J 22: 3960 – 3970 Fromont-Racine M (2001) Nog2p, a putative GTPase associated with 35. Hug N, Cáceres JF (2014) The RNA helicase DHX34 activates NMD by pre-60S subunits and required for late 60S maturation steps. EMBO J 20: promoting a transition from the surveillance to the decay-inducing 6475 – 6484 complex. Cell Rep 8: 1845 – 1856 16. Shaywitz DA, Orci L, Ravazzola M, Swaroop A, Kaiser CA (1995) Human 36. Gehring NH, Kunz JB, Neu-Yilik G, Breit S, Viegas MH, Hentze MW, SEC13Rp functions in yeast and is located on transport vesicles budding Kulozik AE (2005) Exon-junction complex components specify distinct from the endoplasmic reticulum. J Cell Biol 128: 769 – 777 routes of nonsense-mediated mRNA decay with differential cofactor 17. Fukuda M (2006) Rab27 and its effectors in secretory granule exocytosis: requirements. Mol Cell 20: 65 – 75 a novel docking machinery composed of a Rab27.effector complex. 37. Chan W-K, Huang L, Gudikote JP, Chang Y-F, Imam JS, MacLean JA, Biochem Soc Trans 34: 691 – 695 Wilkinson MF (2007) An alternative branch of the nonsense-mediated 18. Nandi D, Woodward E, Ginsburg DB, Monaco JJ (1997) Intermediates in decay pathway. EMBO J 26: 1820 – 1830 the formation of mouse 20S proteasomes: implications for the assembly 38. Ivanov PV, Gehring NH, Kunz JB, Hentze MW, Kulozik AE (2008) Interac- of precursor beta subunits. EMBO J 16: 5363 – 5375 tions between UPF1, eRFs, PABP and the exon junction complex suggest 19. Frand AR, Russel S, Ruvkun G (2005) Functional genomic analysis of C. an integrated model for mammalian NMD pathways. EMBO J 27: 736 – 747 elegans molting. PLoS Biol 3:e312 20. Chung YD, Zhu J, Han Y, Kernan MJ, Brook S, York N (2001) nompA License: This is an open access article under the encodes a PNS-specific, ZP domain protein required to connect terms of the Creative Commons Attribution 4.0 mechanosensory dendrites to sensory structures. Neuron 29: 415 – 428 License, which permits use, distribution and reproduc- 21. Ailion M, Thomas JH (2003) Isolation and characterization of tion in any medium, provided the original work is high-temperature-induced Dauer formation mutants in Caenorhabditis properly cited. elegans. Genetics 165: 127 – 144

78 EMBO reports Vol 16 |No1 | 2015 ª 2014 The Authors Figure S1 A Ce NGP-1 NGP1NT MMR_HSR1

1 60 190 337 411 651

Hs GNL2 NGP1NT MMR_HSR1 39% 1 43 174 322 411 731

B Ce NPP-20 WD40

1 313

Hs SEC13 WD40 40% 1 322

C Ce AEX-6 RAB

1 215

Hs RAB27A Hs RAB27B RAB RAB 47% 45% 1 221 1 218

Ce PBS-2 proteasome

1 43 224 277

Hs PSMB7 Hs PSMB10 proteasome proteasome 40% 35% 1 40 221 277 1 36 217 273

E Ce NOAH-2 SP PAN_AP ZP TM

1 741

Dm NOMPA SP PAN_AP ZP TM 10% 1 1557

* 20 * 40 * 60 * 80 * 100 Hs GNL2 : MVKPKYK-GRS------TINPSKASTNPDRVQG--AGGQNMRDRATIRRLNMYR-QKERRNSRGKIIKPLQYQSTVASGTVARVEPNIKW : 80 Mm Gnl2 : MVKPKYK-GRS------TINRSAASTNPDRVQG--AGGQNMRDRGTIRRLNMYR-QKERRNSRGKVIKPLQYQSTVASGTVARVEPNIKW : 80 Dr gnl2 : MVKAKFK-GKS------SINTSNSSSNPDRVKG--AGGNNMRDRATIKRLNMYR-QKQRCNSRGKVIKPLQYQNTVAPGTVARVEPNIKW : 80 Dm ns2 : MPKVRSTPGKPRTQ------GFNHSNHSMNPERPKSGLKGVAHPRTKGTIKRLQMYRNFKAKRDRTGKILTPAPFQGRLPAGTMARVEPTPKW : 87 Sc NOG2 : MGTGKKEKSRR------IREGDTKDGNLRVKG----ENFYRDSKRVKFLNMYTSGKEIRNKKGNLIRAASFQDSTIP--DARVQPDRRW : 77 Ce ngp-1 : MVKARKEGKVSKKLKQKPDYGPVKNRHTFRGSNHSMNPDRKPD--AKDKSQRSKATINRLRMYKSFKPIRDSKGKILKAAPFQDTLASGTQARIEPNRKW : 98

* 120 * 140 * 160 * 180 * 200 Hs GNL2 : FGNTRVIKQSSLQKFQEEMDTVMKDPYKVVMKQSKLPMSLLHDRIRPHNLKVHILDTESFETTFGPKSQRKRPNLFASDMQSLIENAEMSTESYDQGKDR : 180 Mm Gnl2 : FGNTRVIKQASLQKFQEEMDKVMKDPYKVVMKQSKLPMSLLHDRIQPHNAKVHILDTESFESTFGPKSQRKRPNLFASDMQSLLENAEMSTESYDQGKDR : 180 Dr gnl2 : FANTKVIKQSSLQKFQEEMNAVKKDPYRVVMRQSKLPMSLLHDRIKAHNSKVHILDTETFETTFGPKAQRKRPNLSVGELKEFAEQAEVSAQSYSAEKDR : 180 Dm ns2 : FSNSRVISQTALQKFQDEIGKAVKDPYQVIMKPSQLPVTLLNEAAKYK--RVHLLDTESFDSTFGPKKQRKRVSLKVRDLEDLSKAADDQADKYDSAKDL : 185 Sc NOG2 : FGNTRVISQDALQHFRSALGETQKDTYQVLLRRNKLPMSLLEEKDADESPKARILDTESYADAFGPKAQRKRPRLAASNLEDLVKATNEDITKYEEKQVL : 177 Ce ngp-1 : FGNTRIIGQEQLQKFQANLGKVLSDPFQVVMKQTKLPISLLQEKAKTQ--RVHVTETESFEYTFGKKSLRKKAKLTEASLEEMSKSAEDRDVKYKPELDL : 196

* 220 * 240 * 260 * 280 * 300 Hs GNL2 : DLVT------EDTGVRNEAQEEIYKKGQSKRIWGELYKVIDSSDVVVQVLDARDPMGTRSPHIETYLKKEKPWKHLIFVLNKCDLVPTWATKRWVAVL : 272 Mm Gnl2 : DLVM------EDTGVRNEAQEEIYKKGQSKRIWGELYKVIDSSDVVVQVLDARDPMGTRSPHIEAYLKKEKPWKHLIFVLNKCDLVPTWATKRWVAVL : 272 Dr gnl2 : DLVS------EDSGVKEEAREEIFKKGQSKRIWGELYKVIDSSDVIIQVLDARDPMGTRSQSIETYLKKEKPWKHLIFVLNKCDLIPTWVTKHWVAVL : 272 Dm ns2 : DLIR------EDTGEKKAVRDWVFGAGQSKRIWNELHKVVDASDVLLQVLDARDPMGTRSKYIEEFLRKEKPHKHLFFILNKVDLVPVWVTQRWVAIL : 277 Sc NOG2 : DATLGLMGNQEDKENGWTSAAKEAIFSKGQSKRIWNELYKVIDSSDVVIHVLDARDPLGTRCKSVEEYMKKETPHKHLIYVLNKCDLVPTWVAAAWVKHL : 277 Ce ngp-1 : SREK------ELGDRPENMNPLFRAGQSNRVWGELYKVIDSSDVVVQVVDARDPMGTRCRHVEEFLRKEKPHKHLVTVINKVDLVPTWVTRKWIGEL : 287

* 320 * 340 * 360 * 380 * 400 Hs GNL2 : SQDYPTLAFHASLTNPFGKGAFIQLLRQFGKLHTDKKQISVGFIGYPNVGKSSVINTLRSKKVCNVAPIAGETKVWQYITLMRRIFLIDCPGVVYPSE-D : 371 Mm Gnl2 : SQDYPTLAFHASLTNPFGKGAFIQLLRQFGKLHTDKKQISVGFIGYPNVGKSSVINTLRSKKVCNVAPIAGETKVWQYITLMRRIFLIDCPGVVYPSE-D : 371 Dr gnl2 : SQEYPTLAFHASLTNSFGKGSLIQLLRQFGKLHSDKKQISVGFIGYPNVGKSSIINTLRSKKVCNVAPLAGETKVWQYITLMRRIFLIDCPGVVYPSD-D : 371 Dm ns2 : SAEYPTIAFHASLQHPFGKGALINLFRQLGKLHLDKKQISVGFIGYPNVGKSSVINALRSKKVCKVAPIAGETKVWQYITLMKRIFLIDCPGVVYPTA-E : 376 Sc NOG2 : SKERPTLAFHASITNSFGKGSLIQLLRQFSQLHTDRKQISVGFIGYPNTGKSSIINTLRKKKVCQVAPIPGETKVWQYITLMKRIFLIDCPGIVPPSSKD : 377 Ce ngp-1 : SKEMPTIAFHASINNSFGKGAVINLLRQFAKLHPDRPQISVGFIGYPNVGKSSLVNTLRKKKVCKTAPIAGETKVWQYVMLMRRIYLIDSPGVVYPQG-D : 386

* 420 * 440 * 460 * 480 * 500 Hs GNL2 : SETDIVLKGVVQVEKIKSPEDHIGAVLERAKPEYISKTYKIDSWENAEDFLEKLAFRTGKLLKGGEPDLQTVGKMVLNDWQRGRIPFFVKPPNAEPLVAP : 471 Mm Gnl2 : SETDIVLKGVVQVEKIKAPQDHIGAVLERAKPEYISKTYKIESWENAEDFLEKLALRTGKLLKGGEPDMLTVSKMVLNDWQRGRIPFFVKPPNAELPTDS : 471 Dr gnl2 : SETDIVLKGVVQVEKIRNPEDHIGAVLERAKAEYIQKTYRIPSWSSAEDFLEKLAFRTGKLLKGGEPDLPTVSKMVLNDWQRGRIPFFVKPPGVET--DE : 469 Dm ns2 : TDTEKVLKGVVRVELVTNPEDYVDSLLKRVRPEYISKNYKIEHWNTSTHFLEQLAQKTGKLLKGGEPDVTVTARMVLNDWQRGKLPFYVPPEGFAVPKSQ : 476 Sc NOG2 : SEEDILFRGVVRVEHVTHPEQYIPGVLKRCQVKHLERTYEISGWKDATEFIEILARKQGRLLKGGEPDESGVSKQILNDFNRGKIPWFVLPP------: 469 Ce ngp-1 : SETQIILKGVVRVENVKDPENHVQGVLDRCKPEHLRRQYGIPEFTDVDDFLTKIAIKQGRLLKGGDPDIVAVSKVVLNEFQRGKLPYFVPPPGCEER--- : 483

* 520 * 540 * 560 * 580 * 600 Hs GNL2 : QLLPSSSLEVVPEAAQNNPGEEVTETAGEGSESIIKEETEENSHCDANTEMQQILTRVRQNFGKINVVPQFSGDDLVPVEVSDLEEELESFSDEEEEEQE : 571 Mm Gnl2 : QLPPSSPLEVPTETTQNNPEEETTETEVERSDSITEKEPEGDCSQDRNSEMQQILARVRQNFGKINVGPQFSADDLVPVEMSDLED-LESSGEEEEQEQE : 570 Dr gnl2 : ENKAQAMLEMPEMESEMDELENRLEAQDVEEASSSQHE-EQKLLKKKKEEVQKLFNDVKQNFGKINVAPEFSEEDLVPVEMPDMMDASGSEDEENEEEEE : 568 Dm ns2 : EGKEE------EVVAEDPNEDAKSEAPTFVSES------VKKAREFKQI-----QDFRKIRVGLEYEQEDVKDLDHIDLELLEQQKAERAAKKKA : 554 Sc NOG2 : ------EKEGEEKPKKKE : 481 Ce ngp-1 : ------AKKDFSQAPINEMCADDDEQLPLDAQLELDDVARNGGPDEDDDEAIDDEEQEELTE : 539

* 620 * 640 * 660 * 680 * 700 Hs GNL2 : QQRDDAEES-SSEPEEE------NVG-NDTKAVIKALDEKIAKYQKFLD-KAKAKKFSAVRISKGLSEKIFAKPEEQRKTLEEDVDDRAPSKKGKKR : 659 Mm Gnl2 : QPGEDAEEERSPDTQEE------PVG-NDTKAVLRALDEKIAKYQRFLN-KAKAKKFSAVRISKDLSEKVFAKYKEEKKTSAEDSD-AAPTKKARKW : 658 Dr gnl2 : EEDEEKAEEETGEQQEEESSVSDSTPKAGKKSTQEVNKEMDEKIAKLKHFLD-RAKSKHFSAIRIPKGLSERVFTETVAKQNAAKNQQQ--KAVKQGKRR : 665 Dm ns2 : RIHNLGEEDEES------SDGADEFYSEDEYNEDLQRVVHKKAKSKKPQQLAITSSGKFRVAKIQPGDSDGAGSSDDDGPSTSKAP-- : 634 Sc NOG2 : VEKTA------: 486 Ce ngp-1 : IGSTCSGLTDLS------GISDLAGDLSDADDVADDDEKPSSSDSKTKKSKTDVASVRTRGKRGGKKANKAKLSGKTIVQAASEKKDK-- : 621

* 720 * 740 * 760 * Hs GNL2 : KAQREEEQEHSNKAPRALTSKERRRAVRQQRPKKVGVRYYETHNVKNRNRNKKKTNDSEGQKHKRKKFRQKQ : 731 Mm Gnl2 : DAQMEEEP--SNKTQRMLTCKERRRAARQQQSKKVGVRYYETHNVKNRNRNKKKTSDSEGQKHRRNKFRQKQ : 728 Dr gnl2 : RA-VEDEP----AQPAKLTSKEKRRIERAQKVKKVGVRYYETHNVKNKNKNRKIPSD--GKKANRAKR---- : 726 Dm ns2 : ------RLTAKQKRSLERSQKRKKIGSNFYETTNVKNRNRNKKKDA------: 674 Sc NOG2 : ------: - Ce ngp-1 : ------SVVEFAKPGATKTNMWRKKLEKRRKVKHQK------: 651

Mm Gnl2 Dr gnl2 Dm ns2 Sc NOG2 Ce ngp-1 Hs GNL2 86% 66% 45% 34% 39% Mm Gnl2 66% 46% 34% 39% Dr gnl2 44% 35% 37% Dm ns2 34% 41% Sc NOG2 36%

* 20 * 40 * 60 * 80 Hs SEC13 : MVSVINTVDTSHEDMIHDAQMDYYGTRLATCSSDRSVKIFDV-RNGGQILIADLRGHEGPVWQVAWAHPMYGNILASCSY : 79 Mm Sec13 : MVSVMNTVDTSHEDMIHDAQMDYYGTRLATCSSDRSVKIFDV-RNGGQILIADLRGHEGPVWQVAWAHPMYGNILASCSY : 79 Dr sec13 : MVSVINTVDTSHEDMIHDAQMDYYGTRLATCSSDRSVKIFDV-KNGGQILVADLRGHEGPVWQVAWAHPMYGNILASCSY : 79 Dm sec13 : MVSLLQEIDTEHEDMVHHAALDFYGLLLATCSSDGSVRIFHS-RKN-NKALAELKGHQGPVWQVAWAHPKFGNILASCSY : 78 Sc SEC13 : MVVIAN----AHNELIHDAVLDYYGKRLATCSSDKTIKIFEV-EGETHKLIDTLTGHEGPVWRVDWAHPKFGTILASCSY : 75 Ce npp-20: MTTVRQRIDTQHRDAIHDAQLNIYGSRLATCGSDRLVKIFEVRPNGQSYPMAELVGHSGPVWKVSWAHPKYGGLLASASY : 80

* 100 * 120 * 140 * 160 Hs SEC13 : DRKVIIWREEN-GTWEKSHEHAGHDSSVNSVCWAPHDYGLILACGSSDGAISLLTYTGE-GQWEVKKINNAHTIGCNAVS : 157 Mm Sec13 : DRKVIIWKEEN-GTWEKTHEHSGHDSSVNSVCWAPHDYGLILACGSSDGAISLLTYTGE-GQWEVKKINNAHTIGCNAVS : 157 Dr sec13 : DRKVIIWKEEN-STWDKMYEYTGHDSSVNSVCWGPYDFGLILACGSSDGAISVLTCSGD-GHWDIKKINNAHTIGCNAVS : 157 Dm sec13 : DRKVIVWKSTTPRDWTKLYEYSNHDSSVNSVDFAPSEYGLVLACASSDGSVSVLTCNTEYGVWDAKKIPNAHTIGVNAIS : 158 Sc SEC13 : DGKVLIWKEEN-GRWSQIAVHAVHSASVNSVQWAPHEYGPLLLVASSDGKVSVVEFKEN--GTTSPIIIDAHAIGVNSAS : 152 Ce npp-20: DKKVIIWNEQQ-GRWQKAYEWAAHEASTTCVAFAPHQYGLMLASASADGDIGILRYDNSSNEWISSKIQKCHEQGVNSVC : 159

* 180 * 200 * 220 * 240 Hs SEC13 : WAPAVVPGSLIDHPSGQKPNYIKRFASGGCDNLIKLWKEEEDGQ-WKEEQKLEAHSDWVRDVAWAPSIGLPTSTIASCSQ : 236 Mm Sec13 : WAPAVVPGSLIDQPSGQKPNYIKKFASGGCDNLIKLWREEEDGQ-WKEEQKLEAHSDWVRDVAWAPSIGLPTSTIASCSQ : 236 Dr sec13 : WAPAVVPGSLIEQPTGQKPNYIKRFVSGGCDNLVKLWKEE-DGQ-WKEDQKLEAHSDWVRDVGWAPSIGLPTSTIASCSQ : 235 Dm sec13 : WCPAQAPDPAFDQRVTSRSAAVKRLVSGGCDNLVKIWRED-NDR-WVEEHRLEAHSDWVRDVAWAPSIGLPRSQIATASQ : 236 Sc SEC13 : WAPATIE----EDGEHNGTKESRKFVTGGADNLVKIWKYNSDAQTYVLESTLEGHSDWVRDVAWSPTV-LLRSYLASVSQ : 227 Ce npp-20: WAPGSAD------PAAKKRLVSAGNDKNVKIWAFDDATNEWILEKTLAGHTDFVREAAWCPVTNNGQHTIVSCGM : 228

* 260 * 280 * 300 * 320 Hs SEC13 : DGRVFIWTCDDASSNTWSPKLLH--KFNDVVWHVSWSITANILAVSGGDNKVTLWKESVDGQWVCISDVN------KGQG : 308 Mm Sec13 : DGRVFIWTCDDASGNMWSPKLLH--KFNDVVWHVSWSITANILAVSGGDNKVTLWKESVDGQWVCISDVN------KGQG : 308 Dr sec13 : DGRVFIWTCDDPAGNTWTAKLLH--KFNDVVWHVSWSITGNILAVSGGDNKVTLWKESVDGQWACISDVN------KGQG : 307 Dm sec13 : DRHVIVWSSN-ADLSEWTSTVLH--TFDDAVWSISWSTTGNILAVTGGDNNVTLWKENTEGQWIRINYESGTAIQSKQPS : 313 Sc SEC13 : DRTCIIWTQD-NEQGPWKKTLLKEEKFPDVLWRASWSLSGNVLALSGGDNKVTLWKENLEGKWEPAGEVH------: 296 Ce npp-20: EGNLVLFRTSNIETEEWKAKLLE--TAPCALYHSSFSPCGSFLSVAGDDNVITIWRENLQGQWIKVPRDN------KE : 298

* 340 * 360 Hs SEC13 : SVSASVTEGQQNEQ------: 322 Mm Sec13 : SVSASITEGQQNEQ------: 322 Dr sec13 : AVS-SITDSQQSEQ------: 320 Dm sec13 : HLPHSHSQQQQALQQHQQQAPSHPGPSSDSEHSSNLSNSQLSN : 356 Sc SEC13 : ------Q------: 297 Ce npp-20: REGMSQAVGAPGAQR------: 313

Mm Dr Dm Sc Ce Sec13 sec13 sec13 SEC13 npp-20 Hs SEC13 96% 87% 50% 47% 40% Mm Sec13 87% 51% 47% 39% Dr sec13 51% 46% 39% Dm sec13 39% 34% Sc SEC13 35%

* 20 * 40 * 60 * 80 Hs RAB27A: MSDGDYD------YLIKFLALGDSGVGKTSVLYQYTDGKFNSKFITTVGIDFREKRVVYRASGPDGATGRGQRIHLQL : 72 Mm Rab27a: MSDGDYD------YLIKFLALGDSGVGKTSVLYQYTDGKFNSKFITTVGIDFREKRVVYRANGPDGAVGRGQRIHLQL : 72 Dr rab27a: MSDGDYD------YLIKFLALGDSGVGKTSFLYQYTDSKFNSKFITTVGIDFREKRVVYKSSGPDGTAGRGQRIHIQL : 72 Hs RAB27B: MTDGDYD------YLIKLLALGDSGVGKTTFLYRYTDNKFNPKFITTVGIDFREKRVVYNAQGPNGSSGKAFKVHLQL : 72 Mm Rab27b: MTDGDYD------YLIKLLALGDSGVGKTTFLYRYTDNKFNPKFITTVGIDFREKRVVYDTQGADGASGKAFKVHLQL : 72 Dr rab27b: MTDGDYD------YLIKLLALGDSGVGKTTFLYRYTDNKFNPKFITTVGIDFREKRVVYTTNSPNGTTGKTFKVHLQL : 72 Ce aex-6 : --MGDYD------YLIKFLALGDSGVGKTSFLHRYTDNTFTGQFISTVGIDFKEKKVVYKSSR-GGFGGRGQRVLLQL : 69 Dm Rab27 : MRAAPPEPEPLQLAGSGEQFLVLGDSGVGKTCLLYQYTDGRFHTQFISTVGIDFREKRLLYNSRG------RRHRIHLQI : 74

* 100 * 120 * 140 * 160 Hs RAB27A: WDTAGQERFRSLTTAFFRDAMGFLLLFDLTNEQSFLNVRNWISQLQMHAYCENPDIVLCGNKSDLEDQRVVKEEEAIALA : 152 Mm Rab27a: WDTAGQERFRSLTTAFFRDAMGFLLLFDLTNEQSFLNVRNWISQLQMHAYCENPDIVLCGNKSDLEDQRAVKEEEARELA : 152 Dr rab27a: WDTAGQERFRSLTTAFFRDAMGFLLLFDLTNEQSFLNVRNWMSQLQTHAYCENPDIVLCGNKSDLEEQRAVKAENARELA : 152 Hs RAB27B: WDTAGQERFRSLTTAFFRDAMGFLLMFDLTSQQSFLNVRNWMSQLQANAYCENPDIVLIGNKADLPDQREVNERQARELA : 152 Mm Rab27b: WDTAGQERFRSLTTAFFRDAMGFLLMFDLTSQQSFLNVRNWMSQLQANAYCENPDIVLIGNKADLPDQREVNERQARELA : 152 Dr rab27b: WDTAGQERFRSLTTAFFRDAMGFLLMFDLTSQQSFLNVRNWMSQLQANAYCENPDIVLVGNKADLADQREVQEKQAKELA : 152 Ce aex-6 : WDTAGQERFRSLTTAFFRDAMGFILIFDITNEQSFLNIRDWLSQLKVHAYCEQPDIIICGNKADLENRRQVSTARAKQLA : 149 Dm Rab27 : WDTAGQERFRSLTTAFYRDAMGFLLIFDLTSEKSFLETANWLSQLRTHAYSEDPDVVLCGNKCDLLQLRVVSRDQVAALC : 154

* 180 * 200 * 220 * 240 Hs RAB27A: EKYGIPYFETSAANGTNISQAIEMLLDLIMKRMERCVD------KSWIPEGVVRSNGHASTDQLS-EEKEKGACGC- : 221 Mm Rab27a: EKYGIPYFETSAANGTNISHAIEMLLDLIMKRMERCVD------KSWIPEGVVRSNGHTSADQLS-EEKEKGLCGC- : 221 Dr rab27a: EKYGVPYFETSAANGENVSRAVEVLLDLIMKRMERCVD------KSWIPDGTVRSNGHSTADLTEPHDPEQSKCAC- : 222 Hs RAB27B: DKYGIPYFETSAATGQNVEKAVETLLDLIMKRMEQCVE------KTQIPDTVN----GGNSGNLDGEKPPEKKCIC- : 218 Mm Rab27b: EKYGIPYFETSAATGQNVEKSVETLLDLIMKRMEKCVE------KTQVPDTVN----GGNSGKLDGEKPAEKKCAC- : 218 Dr rab27b: DKYGIPYFETSAATGSEVDKAVVTLLDLVMKRMEQCVE------KPSADANTS----DGSS-KLSAAPAQQKKCAC- : 217 Ce aex-6 : DQLGLPYFETSACTSTNVEKSVDCLLDLVMQRIQQSVE------TSSLPLSECR----GVSLDGDPSAASSYCANC- : 215 Dm Rab27 : RRYRLPYIETSACTGANVKEAVELLVGRVMERIENAACNREFSLLLTQSRCLPNIAYGQPEDLVRLHDRREEPCSRRNCR : 234

Hs RAB27A: -- : - Mm Rab27a: -- : - Dr rab27a: -- : - Hs RAB27B: -- : - Mm Rab27b: -- : - Dr rab27b: -- : - Ce aex-6 : -- : - Dm Rab27 : NC : 236

Mm Dr Hs Mm Dr Ce Dm Rab27a rab27a RAB27B Rab27b rab27b aex-6 Rab27 Hs RAB27A 95% 83% 70% 69% 67% 58% 47% Mm Rab27a 84% 71% 70% 68% 57% 46% Dr rab27a 70% 70% 68% 59% 47% Hs RAB27B 94% 83% 59% 45% Mm Rab27b 82% 58% 43% Dr rab27b 59% 43% Ce aex-6 44%

* 20 * 40 * 60 * 80 Hs PSMB10 : ------MLKPALE-PRGGFSFENCQRNASLERVLP--GLKVPHARKTGTTIAGLVFQDGVILGADTRATNDSVVADKS : 69 Mm Psmb10 : ------MLKQAVE-PTGGFSFENCQRNASLEHVLP--GLRVPHARKTGTTIAGLVFRDGVILGADTRATNDSVVADKS : 69 Dr psmb10 : ------MLNTSTKTLTGGFSFENTRRNAVLEANLSEKGYSAPNARKTGTTIAGLVFKDGVILGADTRATDDMVVADKN : 72 Dr psmb10.2 : ------MALTSHVLEPSLCGFNFENATRNIVLENGAEEGKIKPPKALKTGTTIAGVVFKDGVVLGADTRATSDEVVADKM : 74 Hs PSMB7 : ------MAAVSVYAPPVGGFSFDNCRRNAVLEADFAKRGYKLPKVRKTGTTIAGVVYKDGIVLGADTRATEGMVVADKN : 73 Mm Psmb7 : ------MAAVSVFQPPVGGFSFDNCRRNAVLEADFAKKGFKLPKARKTGTTIAGVVYKDGIVLGADTRATEGMVVADKN : 73 Dr psmb7 : ------MATVSVCQYQPGGFSFENCRRNALLEADITKLGFSSPAARKTGTTICGIVYKDGVVLGADTRATEGMIVADKN : 73 Sc PUP1 : ------MAGLSFDNYQRNNFLAENSH----TQPKATSTGTTIVGVKFNNGVVIAADTRSTQGPIVADKN : 59 Ce pbs-2 : MATADVKSTVPHMDMDRGGAFDFSNCIRNQAMCKMGG----KAPKLTSTGTTIVAVAFKGGLVMGADSRATAGNIIADKH : 76

* 100 * 120 * 140 * 160 Hs PSMB10 : CEKIHFIAPKIYCCGAGVAADAEMTTRMVASKMELHALSTGREPRVATVTRILRQTLFRYQGHVGASLIVGGVDLTGPQL : 149 Mm Psmb10 : CEKIHFIAPKIYCCGAGVAADTEMTTRMAASKMELHALSTGREPRVATVTRILRQTLFRYQGHVGASLVVGGVDLNGPQL : 149 Dr psmb10 : CMKIHYIAPNIYCCGAGVAADAEVTTQMMSSNVELHSLSTGRPPLVAMVTRQLKQMLFRYQGHIGSSLIVGGVDVNGAQL : 152 Dr psmb10.2 : CAKIHYIAPNIYCCGAGTAADTEKTTDMLSSNLTIFSMNSGRNPRVVMAVNIIQDMLFRYHGMIGANLILGGVDCTGSHL : 154 Hs PSMB7 : CSKIHFISPNIYCCGAGTAADTDMTTQLISSNLELHSLSTGRLPRVVTANRMLKQMLFRYQGYIGAALVLGGVDVTGPHL : 153 Mm Psmb7 : CSKIHFISPNIYCCGAGTAADTDMTTQLISSNLELHSLTTGRLPRVVTANRMLKQMLFRYQGYIGAALVLGGVDVTGPHL : 153 Dr psmb7 : CSKIHYISPNIYCCGAGTAADTEMTTQIISSNLELHSLSTGRLPRVATANRMLKQMLFRYQGYIGAALVLGGVDCTGPHL : 153 Sc PUP1 : CAKLHRISPKIWCAGAGTAADTEAVTQLIGSNIELHSLYTSREPRVVSALQMLKQHLFKYQGHIGAYLIVAGVDPTGSHL : 139 Ce pbs-2 : CEKVHKLTESIYACGAGTAADLDQVTKMLSGNLRLLELNTGRKARVITALRQAKQHLFNYQGYIGAYLLIGGVDPTGPHL : 156

* 180 * 200 * 220 * 240 Hs PSMB10 : YGVHPHGSYSRLPFTALGSGQDAALAVLEDRFQPNMTLEAAQGLLVEAVTAGILGDLGSGGNVDACVITK-TGAKLLRTL : 228 Mm Psmb10 : YEVHPHGSYSRLPFTALGSGQGAAVALLEDRFQPNMTLEAAQELLVEAITAGILSDLGSGGNVDACVITA-GGAKLQRAL : 228 Dr psmb10 : YSVYPHGSYDKLPFLTMGSGAASAISVFEDRYKPNMELEEAKQLVRDAITAGIFCDLGSGSNVDLCVITD-KKVDYLRTY : 231 Dr psmb10.2 : YTVGPYGSMDKVPYLAMGSGDLAAMGILEDRFKVNMDLEQAKALVSDAIQAGIMCDLGSGNNIDLCVITK-EGVDYIRPH : 233 Hs PSMB7 : YSIYPHGSTDKLPYVTMGSGSLAAMAVFEDKFRPDMEEEEAKNLVSEAIAAGIFNDLGSGSNIDLCVISK-NKLDFLRPY : 232 Mm Psmb7 : YSIYPHGSTDKLPYVTMGSGSLAAMAVFEDKFRPDMEEEEAKKLVSEAIAAGIFNDLGSGSNIDLCVISK-SKLDFLRPF : 232 Dr psmb7 : YSIYPHGSTDKLPYVTMGSGSLAAMAVFEDRYRPDMEEEDAKSLVRDAIAAGIFNDLGSGSNIDVCVITK-GKVDYLRPH : 232 Sc PUP1 : FSIHAHGSTDVGYYLSLGSGSLAAMAVLESHWKQDLTKEEAIKLASDAIQAGIWNDLGSGSNVDVCVMEIGKDAEYLRNY : 219 Ce pbs-2 : YMCSANGTTMAFPFTAQGSGSYAAITILERDFKVDMTKDEAEKLVQRALEAGMHGDNASGNSLNLVIIEP-SETVFKGPI : 235

* 260 * 280 * Hs PSMB10 : SSPTEPVKR------SGRYHFVPGTTAVLTQTVKPLTLELVEETVQAMEVE--- : 273 Mm Psmb10 : STPTEPVQR------AGRYRFAPGTTPVLTREVRPLTLELLEETVQAMEVE--- : 273 Dr psmb10 : DQPVHKNQR------GGTYRYKPGTTAVLSKTVTPLTLDVVDESVHVMDTE--- : 276 Dr psmb10.2 : KESPYNYKR------QAKYKYKSGTTPILTKTVNKLELDLVQETVQMMETSASS : 281 Hs PSMB7 : TVPNKKGTR------LGRYRCEKGTTAVLTEKITPLEIEVLEETVQTMDTS--- : 277 Mm Psmb7 : SVPNKKGTR------LGRYRCEKGTTAVLTEKVTPLEIEVLEETVQTMDTS--- : 277 Dr psmb7 : DIANKKGVREDKLDGRPVTGSYRYKHGTTGVLSKAVTPLNLDMVEESVQTMDTS--- : 286 Sc PUP1 : LTPNVREEK------QKSYKFPRGTTAVLKESIVN-ICDIQEEQVDITA----- : 261 Ce pbs-2 : VPEFCKRPE------PNDLVYKFQAGATKVLKHKTYK------YDVVESMDITH-- : 277

Mm Dr Dr Hs Mm Dr Sc Ce Psmb10 psmb10 psmb10.2 PSMB7 Psmb7 psmb7 PUP1 pbs -2 Hs PSMB10 88% 63% 51% 55% 56% 53% 44% 35% Mm Psmb10 61% 50% 55% 56% 53% 43% 34% Dr psmb10 55% 62% 62% 65% 45% 34% Dr psmb10.2 55% 56% 55% 42% 36% Hs PSMB7 96% 76% 51% 40% Mm Psmb7 77% 51% 41% Dr psmb7 48% 37% Sc PUP1 37%

* 20 * 40 * 60 * 80 * 100 Ce noah-2: ------MWGVIFLLLSIVPAAQSVFECSSHETTAFVRIPRARLDG--TPVVISTAG---HDLTCAQYCRNNIEPTTGAQRVCASFNFD------: 77 Dm nompA : MRPRKGIHVLLTTLVVSLSLSKINGQTTCKNGLGRVLYERLPNQQLQGYDDDVVRDTAPPFRVLEKCQDLCLRDRSGSNNLVRTCTSFDFQPGSRITSFG : 100

* 120 * 140 * 160 * 180 * 200 Ce noah-2: -----GRETCYFFDDAATPAGTSQLTANPSANNFYYEKTCIPNVSAHEACTYRSFSFERARNTQLEG--FVKKSVTVENREHCLSACLKEKEFVCKSVNF : 170 Dm nompA : GNSEYEESLCYLTSEQAGPEGIGSLMLVP--NSVHFNEICLTSSRPERECPSRRYVFERHPRKKLKLPISDIKEITAANRSDCEDKCLNEFSFVCRSANF : 198

* 220 * 240 * 260 * 280 * 300 Ce noah-2: HYDTSLCELSVEDKRSKPTHVRMSEKIDYYDNNCLSRQNRCGPSGGNLVFVKTTNFEIR-YYDHTQSVEAQESYCLQKCLDSLNTFCRSVEFNPKEKNCI : 269 Dm nompA : DSTMRSCTLSRFTRRTHPELMEDDPNSDYLENTCLNAERRCD---GLAVFVKEENKRLGGPFEVDIFNNMTLEECQTMCLRAEKYFCRSVEFDDQSKQCI : 295

* 320 * 340 * 360 * 380 * 400 Ce noah-2: VSDEDTFSRADQQGQVVG--KDYYEPICVAADLS------SSTCRQQAAFERFIGSSIEGEVVASAQGVTISDCISLCFQN--LNCKSINYDR : 352 Dm nompA : LSEEDSISQKDDISISSSPTHHFYDLVCLDNQRANDYPDNSVTSHLFSSGRRPDTAFQRYRNSRLGGEFHSEITGRSLSECLDECLRQTSFQCRSAVYSD : 395

* 420 * 440 * 460 * 480 * 500 Ce noah-2: TASSCFIYAVGRQD------: 366 Dm nompA : RFRTCRLSRYNQKDGMRIIYDADYDYYENLMLNVVGGGADGDGGGHGGSSDGKRPGDQSGSNWRQPNKHDDRYGSGGSSVGGGSHGGTGSGGSRLPPGEG : 495

* 520 * 540 * 560 * 580 * 600 Ce noah-2: ------ANIKAN------: 372 Dm nompA : VDYGRPYDRYPDIAGNEYDRNPYGGDRDRDRYPPDRYGSRYPTGGDGIGYNRPYDRFPDDYDRYPAGAGVNGDRDRERDRDRYPVVGDRDRYPGAVDRDR : 595

* 620 * 640 * 660 * 680 * 700 Ce noah-2: ------: - Dm nompA : YPGVRDRDRYPEPYPPERYADRYGDRRYPERERDRDRDRLPYRPLPYPGINDNSLPSDLPHTRPYPTDDDAPFRPYGYGGGRYGENRYEGRYPPRFPPSR : 695

* 720 * 740 * 760 * 780 * 800 Ce noah-2: ------: - Dm nompA : ERDPVGGYTGRDAPDSIFPDRRYRPSSMDSPRYPYLPDSRGPPGRYDDIVHSARRPEPDSAKRYPPAPIAPTGSSSKYASTPNRFPVGNDRYPIDIYKYG : 795

* 820 * 840 * 860 * 880 * 900 Ce noah-2: ------PSMDYYEFNCESQFGG------: 388 Dm nompA : NRPGPNDLGRRPGLERPPPPFYDYDYEERYGDRYGPYDREYDGPPGRRPPSGGPYGRYDSPFNRPYGGNGLDDRPLPLPGLGLSHPPTPYGGGGAGHVGV : 895

* 920 * 940 * 960 * 980 * 1000 Ce noah-2: ------: - Dm nompA : GVGVNSGPPRPPITRCEESDNFKQIAARHKMRRHFVRRALIVPSLIQCERECIESRDFVCRSFNYRDSAASGYEDRDRDRDRDSPNCELSDRDSRELDIH : 995

* 1020 * 1040 * 1060 * 1080 * 1100 Ce noah-2: ------MALCTNEGIRFIVNTKEPYTGAIYAAERFSTCSQVVENAKQISITFPPPTVSSDCGTVIRDGKMEALVVVSLD : 461 Dm nompA : DPGTFDASNYDFYERSIGRSDGECMDVTQTCNEEGMEFTIRTPEGFLGRIYTYGFYDRCFFRGNGGTVNVLRLSGPQGYPDCGTQRYGDTLTNIVVVQFS : 1095

* 1120 * 1140 * 1160 * 1180 * 1200 Ce noah-2: GVLPHQVTTEWDRFYRVSCDVSMDKMVKEGSVVVTTIYEASSQNTTVLDVATPPPVSAELQILNQLEEPLHKASIGDPLLLVITSEQAGPHNMMVTECTA : 561 Dm nompA : ----DNVQTSRDKRYNLTCIFRGPGEAVVSSGYIG-AGSGSPIPIEYLPAENTLSSKVRLSILYQ-GRPTTTIAVGDPLTFRLEAQDGYNHVTDIFATNV : 1189

* 1220 * 1240 * 1260 * 1280 * 1300 Ce noah-2: TRVGGFGDTVPFTLIENGCPRYPALVGPVEQDFDKNRLKSDLRAFRLDGS------YDVQIVCSIMFCAGPNG------: 628 Dm nompA : VARDPYSGRSIQLIDRFGCPVDPFVFPELDKLRDGDTLEARFNAFKIPESNFLVFEATVRSCREGCQPAYCPGPAGRQEPSFGRRRRSLNTTEIPEPEAL : 1289

* 1320 * 1340 * 1360 * 1380 * 1400 Ce noah-2: ------CPVSNCLDSGTNELFMSHGRKKRSADLEAGETEEKLSAIIRVFAKGEDEEEMEMANNTMMTSMSDSTELLCIAEPFFVSSVVSLS : 713 Dm nompA : ALEGSSQLEASTLDEVTVVNSTTVSATLGQVPLNETQLGEKTKETEEPEQVREMIEVFETREEIEKESYPR--KLVAPVETVCMTPAEYHGLITAIILLM : 1387

* 1420 * 1440 * 1460 * 1480 * 1500 Ce noah-2: VLCFALSAIIAIWGCHSLHSKPVKQVAA------: 741 Dm nompA : ILLFSITLVAGLGYRRYWKSISKNRLVDRHSPIHSLGHSHSSIRTHERFTEIGHMPNLNGGGGGAAAGTGGGANQSASNRASNAFRTNMSMFGGSLHKTF : 1487

* 1520 * 1540 * 1560 * Ce noah-2: ------: - Dm nompA : ATGNLARMCQLPVINPMRSTNQSSHQFEDPSEPIYTDPSLFERSRQVADHSVTHPQNEFPTTVRNYRCEV : 1557

Percentage Identity: 10% Figure S3

A B

12 12 * 10 10 *

Expression 8 Expression 8 A 6 * * 6 * 4 4

2 2

0 0 Relative HBB mRN Relative HBB mRN A Mock Mock si UPF2 si UPF2 si PSMB7 si PSMB7 si PSMB10 si PSMB10 Scramble siRNA Scramble siRNA si PSMB7 + PSMB10 si PSMB7 + PSMB10 PTC

wt β-globin NS39 β-globin

C D ARHGEF18 Mock t½ ~ 8h siGNL2 t½ ~ 22h siSEC13 t½ ~ 58h 1.2 g n i 100 1 n a i m e 0.8 50 0.6

0.4

Level of the depleted genes 25 A ARHGEF mRNA r 0.2 %

0 13 0 2 4 6 8 Relative mRN

Mock ActD treatment (h) si UPF2 si GNL2 si SEC13 si PSMB7 si RAB27A si RAB27B si PSMB10 + si RAB27B si PSMB7 + PSMB10 si RAB27 A

E F siRNA siRNA

1.4 si UPF2 si GNL2 si SEC13 si UPF2 si GNL2 si SEC13 Mock Scramble Mock Scramble 1.2 kDa 1 175- 0.8 83- 0.6 0.4 62- 0.2

48- S] Met/Cys Incorporation 0 35 24- [ Mock si UPF2 17- si GNL2 si SEC13

7- Scramble siRNA Coomassie Blue Autoradiograph Figure S4

input anti-Flag IP F-EV F-EV F-EV F-UPF1 F-UPF1 T7- GNL2 - + + + + T7- EV + - - - - * WB: α-GNL2

WB: α-PABP1

WB: α-UPF1 Identification and characterization of novel factors that act in the nonsense-mediated mRNA decay pathway

in nematodes, flies and mammals

Angela Casadio1, Dasa Longman1*, Nele Hug1, Laurent Delavaine1, Raúl Vallejos Baier2, Claudio R. Alonso2 and Javier F. Cáceres1*

Supplementary Information

Supplementary Methods

Cell culture and siRNA-mediated depletion

HeLa cells were grown in Dulbecco's modified Eagle's medium (Life Technologies)

supplemented with 10% fetal calf serum and incubated at 37°C in the presence of 5% CO2. For siRNA-mediated depletions, HeLa cells were transfected using a pool of four siRNAs (OnTarget

Plus siRNA, Thermo Scientific) against each gene tested and were transfected using

DharmaFECT 1 (Thermo Scientific) following manufacturer’s instructions. Depletion of UPF1 was achieved by using a combination of two individual siRNAs previously described in [1,2].

Catalogues numbers of OnTarget Plus siRNA pools: GNL2 (L-020392-010005), SEC13 (L-

012351-00-0005), RAB27A (L-004667-00-0005), RAB27B (L004228-00-0005), PSMB7 (L-

006021-00-0005), PSMB10 (L-006019-00-0005), UPF2 (L-012993-01-0005), scramble (D-

001810-02-20). UPF1 siRNAs sequence: hUPF1-I GAGAATCGCCTACTTCACT and hUPF1-II

GATGCAGTTCCGCTCCATT. Transfections were performed in six well plates in four biological replicates; cells were seeded at 15% confluency at day 0 and transfected on day 1 with

5 µM siRNA. On day 3, cells were passaged in order to reach the right confluency level for a second round of transfection on day 4. Cells were then collected 2 days after the second round of transfection. Cells treated with siRNAs against PSMB7 and PSMB10 were collected 1 day after the second round of transfection. Total RNA was isolated using the Qiagen RNeasy kit following manufacturer’s instruction. The quality of RNA was assessed by Agilent 2100 Bioanalyzer using Casadio et al.

RNA 6000 Nano Kit, and NanoDrop 8000 spectrophotometer. The efficiency of the knockdown was assessed by RT-qPCR.

Half-life mRNA analysis

HeLa cells stably expressing HBB gene, wild type or NS39, were mock-depleted or depleted of

GNL2 or SEC13 for 5 days, as described above. On day 5, Actinomycin D (2 µg/ml) was added to the depleted cells. Total RNA was isolated from cells treated with Actinomycin D for 0.5, 1, 2,

4, 6, and 8 hours. The quality of RNA was assessed by Agilent 2100 Bioanalyzer using RNA

6000 Nano Kit, and NanoDrop 8000 spectrophotometer. For each time point, the mRNA levels of the HBB reporters were monitored by qRT-PCR and normalized to POLR2J and ACTB reference genes. The values shown are the average fold-change (mean ± SEM) from three independent experiments relative to the first time point. The decay curve was calculated using

GraphPad Prism software (exponential one phase decay curve) and displayed on log2 y-axis. For the half-life analysis of ARHGEF18 mRNA, untransfected Hela cells were treated as above.

cDNA synthesis and RT-PCR (C. elegans) cDNA was synthesised starting from 500 ng of RNA using Transcriptor Universal cDNA master

(Roche) according to manufacturer’s instructions. Quantitative RT-PCR was performed using the

SYBR Green system (Roche) following manufacturer’s instructions. The expression of GFP was normalized to that of the reference gene ama-1. The sequence of the primers used is listed in

Supplementary Table S1. cDNA synthesis and quantitative RT-PCR (human cells) cDNA was synthesized from 1 µg of total RNA using Transcriptor Universal cDNA master (Roche) according to manufacturer’s instructions. Quantitative RT-PCR was performed using the probe system (Roche) following

2 Casadio et al. manufacturer’s instructions. The expression of β-globin was normalized to the geometric mean of two reference genes (ACTB and POLR2J). To check depletion levels of the putative factors and of the positive control UPF2, their mRNA level was normalized to the geometric mean of the two reference genes. The expression of ARHGEF18,

BMP2 and TMC7 was normalized to the geometric mean of two reference genes (ACTB and POLR2J). To analyze the half-life of the two β-globin reporters and of ARHGEF18 mRNAs, the expression of their RNAs was measured at the indicated time points and normalized to the geometric mean of two reference genes (ACTB and POLR2J). To analyze the feedback loop, mRNAs corresponding to known NMD factors were normalized to the geometric mean of two reference genes (ACTB and POLR2J). The sequences of the primers used for RT-qPCR and catalogue numbers for the Roche assays are listed in the Supplementary Table S1.

Translation inhibition assay

HeLa cells were depleted with siRNA for 5 days, as described previously. After siRNA treatment overall translation activity was measured as described by [3]. Cells were labelled for 1h with

[35S]-Met/Cys. Afterwards, labeled protein extract concentration was determined using Bradford assay. Equal amount of protein was separated on 4-12% Bis-Tris gels (Life Technologies). After gel drying, the gels were stained using Coomassie Colloidal Blue (Life Technologies).

Radioactivity in each sample was determined using autoradiography, radioactive signals and

Coomassie staining signals were quantified with ImageQuantTL software (GE Healthcare).

Immunoprecipitation and Western Blotting

HEK293T cells were transiently transfected with Flag-tagged UPF1 or Flag empty vector (F-EV) together with either T7-tagged GNL2 or T7 empty vector (T7-EV). GNL2 cDNA was cloned by PCR amplification (Forward

Primer: CTGCTAGCATGGTGAAGCCCAAGTACAA and Reverse Primer:

CGACTGTGATCATTACTGCTTTTGTCTGAATTTTTTGCG) and ligated into pCG-T7 vector. Flag-UPF1 was previously described [4]. In total, 10ng of each plasmid together with 2 µg of carrier DNA per well was transfected in 6-well plates, expanded into 10 cm dishes and harvested 48hrs post transfection. All subsequent steps were carried out at 4ºC. Cells were washed 2x in PBS and 107 cells were lysed in 1 ml of IP buffer (10mM Tris HCl pH8, 150mM

NaCl, 1% NP-40, 0.2% Sodium Deoxycholate, 1mM EDTA, 1mM DTT, protease inhibitors (Roche)) for 20 min.

3 Casadio et al.

For RNase treatment half of protein extract was incubated with 20 µg/ml RNase A (ThermoScientific) for 20 min.

Cell extracts were spun 20 min at full speed and supernatants were incubated with Anti-Flag M2 magnetic beads

(Sigma) o/n at 4ºC. Beads were washed 5x in IP buffer and resuspended in LDS loading buffer supplemented with reducing agent (Life Technologies). Samples were incubated at 94ºC for 5 min, and subjected to

SDS/polyacrylamide gel electrophoresis followed by Western blotting using rabbit anti-GNL2 antibody (1:500)

(Sigma), rabbit anti-PABP1 antibody (1: 1000) (Cell Signaling), or goat anti-RENT1 (UPF1) antibody (1:3000)

(Bethyl). Proteins were visualized by Fluorchem (Protein Simple).

Immunostaining experiments in Drosophila embryos

Late stage 16 embryos were collected, dechorionated in bleach 50% and fixed in 1x PBS formaldehyde 4% according to standard protocols. Nuclei were stained with DAPI, neurons of the PNS were labelled with a mouse monoclonal anti-futsch antibody (22C10, Hybridoma bank) and GFP was detected in fixed tissue using a rabbit polyclonal anti-GFP antibody (Life Technologies). Secondary antibodies were Alexa Fluor 555 Donkey Anti-Mouse

IgG (H+L) and Alexa Fluor 488 Donkey Anti-Rabbit IgG (H+L), respectively (both from Life Technologies).

Embryos were mounted in Vectashield reagent (Vector Labs) and unsaturated images of GFP staining of late stage 16 embryos were taken using a DM6000 microscope and LAS software (Leica Microsystems), using identical settings between experimental and control samples. Images were imported into ImageJ (NIH), converted into jpeg format, exported to Adobe Photoshop PS6 (Adobe Systems Incorporated) and average levels of pixel intensity extracted from the green channel were measured in abdominal PNS fields comparing identical areas (250 µm2) in Upf1-RNAi, nompA-RNAi and wild-type (no RNAi) control embryos. After subtraction of background, ratios between GFP expression in control and genetically induced RNAi treatments were determined and plotted using GraphPad Prism (GraphPad Software).

Supplementary References 1. Paillusson A, Hirschi N, Vallan C, Azzalin CM, Mühlemann O (2005) A GFP-based reporter system to monitor nonsense-mediated mRNA decay. Nucleic Acids Res 33: e54. 2. Azzalin CM, Lingner J (2006) The human RNA surveillance factor UPF1 is required for S phase progression and genome stability. Curr Biol 16: 433–439. 3. Chazal P-E, Daguenet E, Wendling C, Ulryck N, Tomasetto C, Sargueil B, Le Hir H (2013) EJC core component MLN51 interacts with eIF3 and activates translation. Proc Natl Acad Sci U S A 110: 5903–5908.

4 Casadio et al.

4. Hug N, Cáceres JF (2014) The RNA Helicase DHX34 Activates NMD by Promoting a Transition from the Surveillance to the Decay-Inducing Complex. Cell Rep 8: 1845–1856.

Legends for Supplementary Figures

Figure S1. Schematic representation of the novel NMD proteins. (A) The C. elegans NGP-1 protein and its human homolog GNL2 are characterized by a conserved N-terminal NGP1NT domain (yellow) and by a GTPase domain (green). The five G-motifs (black) that are typical of this class of GTPases are indicated. (B) The C. elegans

NPP-20 protein and its human homolog SEC13 are characterized by the presence of six WD40 domains (orange).

(C) The C. elegans AEX-6 protein and its two human homologs RAB27A and RAB27B are small GTPase members of the RAB subfamily and are characterized by a highly conserved RAB domain (purple). (D) The C. elegans PBS-2 protein and its two human homologs PSMB7 and PSMB10 are characterized by the presence of a proteasome domain (light blue). The two human proteins also contain a proteasome beta subunits C-terminal domain (dark blue). (E) The C. elegans NOAH-2 protein and its D. melanogaster homolog nompA are characterized by 4

PAN_AP-domains (red) and by a ZP-domain (dark orange). Both proteins contain an N-terminal signal peptide sequence (yellow) and a C-terminal transmembrane (TM) signal sequence (black). Predicted conserved protein domains were generated using SMART (http://smart.embl-heidelberg.de). For each conserved protein % of identity is indicated.

Figure S2. Alignment of orthologous proteins identified in the RNAi screen. (a) ngp-1 (b) npp-20 (c) aex-6 (d) pbs-2 and (e) noah-2. Sequences were aligned using CLUSTAL 2.1, and output was produced using Genedoc (version 2.7.00). Residues highlighted in black: 100% homology, dark grey: 80% homology, pale grey: 60% homology. Similarity is defined by a scoring matrix (PAM-65), which reflects the degree to which residues are interchangeable or equivalent during evolution. The table below each alignment indicates % identity of proteins across species. Hs: human; Mm: mouse; Dr: zebrafish; Dm; fruitfly; Ce: C.elegans; Sc; yeast

Figure S3. Role of the newly identified human homologs in NMD.

5 Casadio et al.

(A, B) HeLa cells stably expressing wild-type β-globin reporter (A) or β-globin NS39 NMD- reporter (B) were mock-depleted, or depleted of the indicated factors for four days. The level of the β-globin mRNA was monitored by quantitative RT-PCR and normalized to two reference genes (POLR2J and ACTB). The values shown are the average fold-change (mean ± SEM) from four independent experiments relative to mock-depleted cells (control). Statistical analysis was performed using the Mann-Whitney test for non-parametric distributions. * p< 0.05. (C) The level of mRNA of the indicated factors following siRNA-mediated depletion in HeLa cells were measured by qRT-PCR. These samples were used in further analysis shown in Figures 4 and S3.

All values are the average of at least three biological replicas showing +/- S.E.M. (D) GNL2 and

SEC13 regulate the half-life of ARHGEF18 mRNA. HeLa cells were mock-depleted or depleted of GNL2 or SEC13 for 5 days and treated with Actinomycin D as indicated. Samples were collected after the indicated times and the mRNA level of ARHGEF18 was monitored by qRT-

PCR and normalized to POLR2J and ACTB reference genes. The values shown are the average fold-change (mean ± SEM) from three independent experiments relative to the first time point.

(E) Depletion of new NMD factors does not affect global translation. HeLa cells were depleted of UPF2, GNL2 and SEC13 and metabolic labeling was carried out by incubation with [35S]-

Met/Cys for 1h. Total protein extracts were analyzed by Coomassie staining. The incorporation of [35S]-Met/Cys was determined by autoradiography. (F) Bar charts showing quantification of the [35S]-Met/Cys incorporation normalized to the total protein concentration. Autoradiograph signals and Coomassie staining were analyzed using the ImageQuant TL software (GE

Healthcare). All values represent the average fold-change (mean ± SD) from at least two independent experiments relative to mock-depleted cells. siRNA mediated depletion of NMD factors were measured as described before by RT-qPCR (depletion (mean ± SD) was 0.22±0.03 for UPF2; 0.12±0.01 for GNL2 and 0.32±0.02 for SEC13).

6 Casadio et al.

Figure S4. GNL2 co-immunoprecipitates with the core NMD factor UPF1. Immunoprecipitation of Flag-tagged

UPF1 co-purified T7-tagged GNL2 in an RNA-independent manner. As a control of RNase A treatment, UPF1 interaction with PABP1 protein was lost in the absence of RNA. All tagged proteins were expressed at near endogenous levels. F-EV: Flag-empty vector; T7-EV: T7-empty vector. The asterisk indicates an unspecific band.

7 Casadio et al.

Supplementary Table S1. List of primers and PCR assays used UPL Probe Number Assay ID Gene Primers Sequence (Roche) (Roche) Forward TCTGTCACATCACCGTACAA Ce ama-1 n.a. n.a. Reverse CAGAGAGTATCCTGGACGAT Forward TGAAGGTGATGCAACATACG GFP n.a. n.a. Reverse CGGGCATGGCACTCTTGAAA Forward AGAGCTACGAGCTGCCTGAC Hs ACTB 9 n.a. Reverse CGTGGATGCCACAGGACT Forward CTGTGAGCCCCGTTCCTAC Hs POLR2J 1 n.a. Reverse GTCGGTGTCAGGGTGAGG Forward Hs HBB n.a. n.a. 102141 Reverse Forward Hs UPF2 n.a. n.a. 137425 Reverse Forward Hs GNL2 n.a n.a. 144594 Reverse Forward Hs SEC13 n.a. n.a. 138390 Reverse Forward Hs RAB27A n.a. n.a. 114763 Reverse Forward Hs RAB27B n.a. n.a. 115465 Reverse Forward TGATAAGTTGCCTTATGTCACCA Hs PSMB7 36 n.a. Reverse TCCTCCATGTCTGGCCTAAA Forward CATCGCCCCCAAAATCTAC Hs PSMB10 19 n.a. Reverse CCATCCGTGTGGTCATCTC Forward Hs UPF1 n.a. n.a 115402 Reverse Forward Hs SMG1 n.a. n.a 106402 Reverse Forward Hs SMG5 n.a n.a. 128927 Reverse Forward Hs TMC7 n.a. n.a 142251 Reverse Forward ATCAGGGCGCTTGAAAGATA Hs ARHGEF18 18 n.a. Reverse TTCTTGTAGCAGCAAAAGTACGTC Forward GACTGCGGTCTCCTAAAGGTC Hs BMP2 49 n.a. Reverse GGAAGCAGCAACGCTAGAAG