THE RNA EXOSOME AND THE MAINTENANCE OF GENOME INTEGRITY
Judit Domingo Prim
The RNA exosome and the maintenance of genome integrity
Judit Domingo Prim ©Judit Domingo Prim, Stockholm University 2018
ISBN print 978-91-7797-298-3 ISBN PDF 978-91-7797-299-0
Printed in Sweden by Universitetsservice US-AB, Stockholm 2018 Distributor: Department of Molecular Biosciences, The Wenner-Gren Institute Cover: Adrià Verdaguer Abuin and Judit Domingo Prim Aprendre per saber despendre, heus aquí el vell secret.
– Lluís Llach
SAMMANFATTNING
DNA innehåller den nödvändiga genetiska informationen för att våra celler genom budbärar-RNA ska kunna syntetisera alla proteiner som behövs för cellens funktion. Cellerna utsätts dagligen för strålning, kemiska ämnen eller felaktigheter i interna metaboliska processer som kan orsaka olika typer av DNA-skador. De allvarligaste skadorna är dubbelsträngsbrott i DNA (DSB) som, om de inte repareras, kan orsaka mutationer, kromosomala deletioner och sjukdomar såsom cancer.
Cellen kan använda två huvudmekanismer för att reparera DSB: homolog rekombination (HR) och icke-homolog sammanfogning (NHEJ). HR medför inga förändringar i DNA-t och är därför den reparationsmekanism som cellen väljer att använda i första hand. HR-mekanismens noggrannhet beror på att den är aktiv när DNA-t replikeras och kan använda systerkromatiden som mall i reparationsprocessen. I grova drag, när DSB inträffar avlägsnas en av DNA-kedjorna medan den andra skyddas av proteinet RPA. RPA kommer att locka till sig andra proteiner såsom RAD51 och RAD54, som kommer att hitta och binda till den närliggande systerkromatiden vars sekvens kommer att kopieras under reparationen.
Under de senaste åren har forskningen upptäckt en ny komponent i DNA- reparationsmekanismen: RNA. Denna avhandling beskriver funktionen av RNA och av exosomen (ett proteinkomplex som påverkar RNA-metabolism) vid reparationen av DNA-skador. I artikel I har vi med hjälp av immunfluorescens och immunoprecipitering av kromatin visat att en av exosomens två katalytiska subenheter rekryteras till DSB. I frånvaro av detta protein, som heter EXOSC10 i människa och i bananflugor RRP6, blir cellerna mer känsliga för yttre skador, t.ex. strålning, och visar defekter i rekryteringen av nödvändiga HR- faktorer såsom RAD51 till DSB. Detta händer även i celler som överuttrycker en katalytiskt inaktiv mutant av EXOSC10 eller RRP6. Dessa resultat tyder på att RNA-nedbrytning av exosomen är nödvändig för DNA-reparation.
5 I artikel II har vi med hjälp av massiv sekvensering av RNA, så kallad next- generation sequencing eller NGS, bekräftat att icke-kodande RNA produceras i DSB-s närhet. RNA-polymeras II är ansvarigt för att syntetisera dessa icke-kodande RNA, vars syntes induceras av DSB och som därför kallas för dilncRNAs (damage-induced, long non-coding RNAs). DilncRNA är förstadiemolekyler till kortare diRNA, som andra forskargrupper har upptäckt i tidigare studier och som har föreslagits ha en roll i DNA- reparation. Artikel II visar också att det finns två typer av diRNA-molekyler: de som är 21-22 nukleotider långa och produceras av Dicer och de som är resultatet av nedbrytning av andra ribonukleaser såsom kanske exosomen. Slutligen har vi i artikel III visat att exosomen, och EXSOC10 i synnerhet, bryter ned de dilncRNA-molekyler som produceras i DSB. Celler med sänkta EXSOC10-nivåer visar högre nivåer av dilncRNA och diRNA. Dessa RNA hindrar rekryteringen av RPA till DNA-skadorna vilket i sin tur medför en uppreglering av DNA-resektionsprocessen som är ett avgörande steg i HR- reparationen.
I denna avhandling kan vi således konstatera att exosomen, särskilt EXOSC10, är nödvändig för nedbrytningen av de RNA-molekyler som produceras vid dubbelsträngsbrott i DNA och att denna nedbrytningsprocess är en förutsättning för en korrekt homolog rekombination.
Català
L’ADN conté la informació genètica necessària per les nostres cèl·lules per sintetitzar, a través de l’ARN, totes les proteïnes necessàries per al seu funcionament. Malauradament, les cèl·lules es troben diàriament sotmeses a fonts tan endògenes com exògenes que poden danyar l’ADN. El tipus més sever d’aquests danys són els talls de doble cadena (DSBs), que a falta de ser reparats, poden causar mutacions, delecions cromosòmiques i malalties tant rellevants com el càncer.
La cèl·lula té dos mecanismes principals per reparar aquest dany, anomenats reparació homòloga (HR) i unió d’extrems no homòlegs (NHEJ),
6 sent HR el mecanisme d’elecció sempre que sigui possible ja que amb aquest tipus de reparació no és produeixen errors. Això és degut a que aquest mecanisme és actiu quan l’ADN s’està replicant i la cromàtida germana es troba disponible. En termes generals, quan es produeix el dany una de les cadenes d’ADN serà reseccionada mentre l’altre, ràpidament protegida per RPA, atraurà proteïnes com RAD51 i RAD54 que envairan la cromàtida germana per poder copiar íntegrament la seqüència d’ADN. Durant els darrers anys de recerca, s’ha descobert un curiós interactuant en la reparació homòloga, l’ARN. En aquesta tesis s’ha investigat aquesta implicació de l’ARN conjuntament amb la funció en la reparació de l’ADN d’un nou complex proteic, l’exosoma, conegut fins ara com un complex amb moltes funcions importants dins el metabolisme de l’ARN.
A l’Article I hem demostrat a través de tècniques de microscopia de fluorescència i immunoprecipitació de cromatina que la subunitat catalítica principal de l’exosoma, anomenada RRP6 en mosques o EXOSC10 en humans, és reclutada als DSBs. En absència d’aquesta proteïna, o amb la sobre-expressió d’un mutant catalíticament inactiu, les cèl·lules es tornen més sensibles a fonts externes de dany com la radiació, mostrant defectes en el reclutament de proteïnes necessàries per HR com RAD51. Aquests resultats suggereixen que l’activitat ribonucleolítica, o de degradació de l’ARN, de l’exosoma és necessària per la reparació de l’ADN.
A l’Article II hem confirmat, gràcies a la seqüenciació massiva d’ARN, la síntesis d’ARN prop dels DSBs. L’ARN polimerasa II és l’encarregada de sintetitzar aquests ARNs no codificants induïts pel dany, anomenats dilncRNA. Els dilncRNAs són processats a una forma més curta, els diRNAs, els quals s’han descrit en previs articles com a funcional en la reparació de l’ADN. En aquest article hem demostrat que aquests diRNAs poden ser formats o bé per Dicer, formant ARNs curts d’uns 21-22 nucleòtids que són incorporats a Argonaute, o bé, per altres ribonucleases com podria ser l’exosoma. Finalment, a l’Article III hem demostrat que l’exosoma és una d’aquestes ribonucleases encarregades de degradar els ARNs produïts als DSBs. En absència d’EXOSC10 els nivells tant de dilncRNAs com diRNAs augmenten, fent que la cèl·lula no pugui reclutar RPA als llocs de dany i, com a 7 conseqüència, provocant una desregulació del procés de resecció, bàsic per la reparació homòloga de l’ADN. Així doncs, finalitzada aquesta tesis podem concloure que l’exosoma, i concretament EXOSC10, és necessari per degradar els ARNs produïts als llocs de tall per tal de permetre una correcte reparació de l’ADN per recombinació homòloga.
8
LIST OF PUBLICATIONS
Paper I. Consuelo Marín-Vicente*, Judit Domingo-Prim*, Andrea E. Eberle and Neus Visa. RRP6/EXOSC10 is required for the repair of double-strand breaks by homologous recombination. Journal of Cell Science, 2015. 128, 1097-1107. http://jcs.biologists.org/content/128/6/1097 * These authors contributed equally to this work.
Paper II. Franziska Bonath, Judit Domingo-Prim, Marcel Tarbier, Marc Friedländer and Neus Visa. Next generation sequencing reveals two populations of damage induced small RNAs at endogenous DNA double-strand breaks. Manuscript submitted, under peer-review
Paper III. Judit Domingo-Prim, Martín Endara-Coll, Franziska Bonath, Sonia Jimeno, Marc Friedländer, Pablo Huertas and Neus Visa. EXOSC10 is required for RPA recruitment and controlled DNA end resection at double-strand breaks. Manuscript.
Other publications not included in this thesis:
Jaclyn Quin, Stefanie Böhm, Judit Domingo-Prim, Anna Vintermist, Neus Visa and Ann-Kristin Östlund Farrants. Non-coding RNAs from the rDNA intergenic repeat are transcribed by RNA polymerase I and II. Manuscript.
9
CONTENTS
SAMMANFATTNING 5
LIST OF PUBLICATIONS 9
CONTENTS 10
LIST OF ABBREVIATIONS 12
INTRODUCTION 15
I. DNA damage and repair of DNA double-strand breaks 15 I.1 Homologous recombination 17 I.2 Non-homologous end joining 18 I.3 Alternative-EJ repair mechanisms 19
II. Chromatin remodeling in DNA repair 21
III. DNA:RNA hybrids in genome instability 23
IV. DNA damage-response RNAs 25
V. Ribonucleases in DNA repair 27
VI. The exosome 29 VI.1 The exosome in RNA processing and degradation 30 VI.2 The exosome in transcription termination 31 VI.3 The exosome in chromatin dynamics 32 VI.4 Core-independent functions of the exosome 33 VI.5 The exosome in DNA repair 34
METHODS 35
I. How to induce DNA double-strand breaks 35 I.1 Ionizing radiation 35
10 I.2 UV laser micro-irradiation 36 I.3 Site-specific DSBs 37
II. How to measure resection 40 II.1 QAOS 40 II.2 SMART 41
III. RNA analysis at DSBs 42 III.1 ssRT-qPCR 42 III.2 RNA-seq 43
AIMS 45
RESULTS AND DISCUSSION 47 Paper I 47 Paper II 50 Paper III 53
CONCLUSIONS AND FUTURE PERSPECTIVES 55
ACKNOWLEDGEMENTS 59
REFERENCES 61
11
LIST OF ABBREVIATIONS
AGO2 Argonaute RISC DNA-PKcs DNA-dependent catalytic component 2 protein kinase, alt-EJ Alternative end-joining catalytic subunit ATM Ataxia talangectasia DRIP DNA:RNA mutated immunoprecipitation ATR Ataxia talangectasia DSB Double-strand break Rad3-related protein dsDNA Double-strand DNA bp Base pair endoRNase Endoribonuclease BRCA1/2 Breast cancer 1/2 EXO1 Exonuclease 1 BrdU Bromodeoxyuridine exoRNase Exoribonuclease CDK Cyclin-dependent EXOSC10 Exosome component 10 kinase GFP Green fluorescent cDNA Complementary DNA protein ChIP Chromatin Gy Gray immunoprecipitation HeLa Henrietta Lacks cells CtIP CTBP-interacting HO Homothallic protein endonuclease CUT Cryptic unstable HP1a Heterochromatin transcript protein 1a DDR DNA damage response HR Homologous DDRNA DNA damage-response recombination RNA KAP1 KRAB-domain DDX1 DEAD-box helicase 1 associated protein 1 diRNA Damage-induced RNA kb Kilobase pair dilncRNA DSB-induced long non- lncRNA Long non-coding RNA coding RNA miRNA MicroRNA DIvA Damage-induced via MMEJ Microhomology end- AsiSI cell line joining DNA Deoxyribonucleic acid MMSET Multiple myeloma SET DNA2 DNA replication domain helicase/nuclease 2 MRN Mre11-Rad50-Nbs1 complex
12 mRNA Messenger RNA SMART Single molecule NRD Nrd1-Nab3-Sen1 analysis of resected complex tracks NEXT Nuclear exosome snRNA Small nuclear RNA target complex snoRNA Small nucleolar RNAs ncRNA Non-coding RNA SSA Single-strand annealing NHEJ Non-homologous end SSB Single-strand break joining ssDNA Single-strand DNA nm Nanometers SSR See-saw reporter nt Nucleotide ssRT Strand-specific PAS Poly-adenylation signal retrotranscription PARP1 Poly(ADP-Ribose) SUMO Small ubiquitin-like polymerase 1 modifier Pre-mRNA Precursor messenger SU(VAR)3-9 Suppressor of RNA variegation 3-9 QAOS Quantitative SWI/SNF SWItch/Sucrose non- amplification of ssDNA fermentable qPCR Quantitative TRAMP Trf4/5, Air1/2, Mtr4 polymerase chain polyadenylation reaction complex RNA Ribonucleic acid tRNA Transfer RNA RNAi RNA interference UV Ultraviolet RNAPII RNA polymerase II U2OS Human bone RNase Ribonuclease osteosarcoma RNP Ribonucleoprotein epithelial cells complex XRCC4 X-ray repair cross ROS Reactive oxygen complementing species 4-OHT 4-hydrotamoxyfen RPA Replication protein A 53BP1 P53 binding protein 1 rRNA Ribosomal RNA X Serine 139 SETX Senataxin phosphorylated siRNA Small interfering RNA histone variant H2Ax Ionizing radiation
13
14
INTRODUCTION
The exosome is responsible for the processing and degradation of many different types of RNAs, including non-coding RNAs (ncRNAs). Some reports previous to this thesis suggested that these ncRNAs are required for the assembly of DNA damage response (DDR) foci and for DNA repair by homologous recombination (HR). Immunoprecipitation and high- throughput mass spectrometry methods revealed links between the exosome and proteins involved in DNA repair. These protein-protein interactions, and the fact that short ncRNAs might be involved in the DDR led us to investigate whether the exosome plays a role in DNA repair.
I. DNA damage and repair of DNA double-strand breaks
“Our cells contain common molecules, such as water or oxygen, that can damage DNA”, Tomas Lindahl 1.
DNA breaks can be p ultraviolet (UV) radiation, crosslinking agents or topoisomerase poisons, but also by endogenous factors, which are the main source of damage in vivo. Endogenous damage is often caused by derivative products from cell metabolism, such as reactive oxygen species (ROS), regulation of DNA folding and remodeling or by defects in RNA processing 2,3. The resulting DNA lesions can be of different types: modifications of DNA bases, such as pyrimidine dimers, creation of abasic sites, single-strand DNA breaks (SSBs), where only one DNA strand is interrupted, or double-strand breaks (DSBs)4. DSBs are the most cytotoxic lesions that threaten genomic integrity and cause replication fork collapse. If DSBs are not properly repaired, they result in loss of physical continuity of the genome, loss of chromosomal fragments or chromosomal rearrangements. It is estimated that in mammalian cells around ten DSBs per cell are formed daily. Failure to
15 repair a DSB has deleterious consequences and can lead to cell death or deregulated growth and cancer development. The DDR to DSBs consists of checkpoint pathways to stop the progression of the cell cycle until the structure of the broken chromosome has been restored. The MNR complex, composed of three proteins that are evolutionarily conserved: Mre11, Rad50 and Nbs1, is one of the first factors to bind DSBs. MRN recruits ATM, the major checkpoint kinase, ATR and DNA-PK, which are the kinases responsible for the phosphorylation of the H2AX histone variant at Ser139 ). This phosphorylation in the vicinity of the DSBs generates binding sites for adaptor proteins and promotes chromatin remodeling to increase the accessibility of the DDR effectors to the damaged site 5–7.
Figure 1. DNA double-strand break repair mechanisms. The choice of DNA repair mechanism relies primarily on whether DNA end resection occurs or not. If resection occurs, there are three pathways that can compete with each other: HR, SSA and MHEJ. NHEJ competes before resection has occurred.
The two main mechanisms for DSB repair are HR and non-homologous end joining (NHEJ). Alternatively, there are two other error-prone mechanisms named micro-homology end joining (MMEJ) and single-strand annealing (SSA) (Figure 1). NHEJ is the principal DSB repair mechanism in mammalian
16 cells and it is active through the entire cell cycle, but it is considered error- prone as it is not guided by a template nucleic acid. During the mid-S and G2 phases of the cell cycle, when the amount of replicated DNA in the cell is highest and the sister chromatids are available, HR becomes the major pathway. HR promotes faithful repair and is typically error-free 6.
I.1 Homologous recombination
Homologous recombination is the pathway of choice, if possible, for the repair of DSBs and rescue of collapsed or stalled replication forks. Indeed, failure of the HR pathway results in genome instability and can cause serious deleterious effects that have been associated with different human diseases, including cancer 8. The choice between HR and NHEJ is dependent on the cell cycle phase, as mentioned above, and the type of DNA ends. For instance, one-ended DSBs, mainly found in replication forks, tend to be repaired by HR. For the repair of two-ended DSBs, the repair pathway choice is dictated by the chromatin structure, the DNA end complexity and the transcriptional status9. The HR pathway is finally established when the DNA ends are resected into long single-stranded DNA (ssDNA) by MRE11-CtIP and EXO1-DNA2 in the 5’- 3’ direction (see below). DNA end resection prevents NHEJ and promotes HR in cells that have already replicated their DNA 5. Once the DNA is resected, RPA binds the ssDNA protecting DNA ends from degradation and preventing spontaneous annealing with complementary sequences. RPA is replaced by RAD51 and its paralogs BRCA2 and RAD52. RAD51, stimulated by RAD54, is responsible for the strand invasion of the sister chromatid 6. It is the same protein, RAD54, an ATP-dependent double-strand DNA (dsDNA) translocase, which promotes the disassembly of RAD51 from the heteroduplex DNA. This translocation is necessary for successful homologous DNA pairing and strand exchange 10. Specialized DNA polymerases, , are next required to synthesize DNA from the intact DNA template and restore the DNA duplex through a complex series of events 11,12. The synthesis forms Holliday junctions, which need to be resolved by either nucleolytic cleavage or by the action of helicases and
17 topoisomerases to convergently migrate the junctions 13. The branch migration of the Holliday junctions and the final reposition of the nucleosomes is also promoted by RAD54 10.
I.1.1 DNA end resection and RPA
The cell cycle plays a critical role in regulating DNA end resection because it takes place only when cyclin-dependent kinases (CDKs) are active and can phosphorylate CtIP 14. DNA end resection proceeds at both sides of the DSB and is a two-step process of nucleolytic degradation of a single DNA strand at both sides of the DSB 6,15. The first step is catalyzed by the MRN complex together with CtIP. This process is also referred to as short-range DNA end resection, as generally affects up to 20 nucleotides (nt) in mammals and 300 nt in yeast. To initiate short-range resection, CtIP, after being phosphorylated, acts as a co-factor of MRE11 and promotes an endonucleolytic cleavage of the 5’- terminated DNA strand 16. Next, it continues with the degradation in the 3’ to 5’ direction generating a 3’ ssDNA overhang. This initial step is required for the subsequent resection in the 5’-3’ direction. The long-range resection factors are EXO1, a dsDNA specific exonuclease, and DNA2, which also needs the BLM helicase to process the dsDNA 4. The ssDNA that results from the DNA end resection becomes rapidly coated by the ssDNA-binding protein RPA. The key function of RPA is to protect ssDNA from degradation, prevent the formation of secondary structures and promote the recruitment of downstream factors 4,8. RPA binding can also regulate EXO1 processivity. It has been described that multiple cycles of binding and release are required for extensive long-range resection until EXO1 is stripped from the DNA by RPA, which attenuates resection 8,17,18.
I.2 Non-homologous end joining
The choice of NHEJ is determined by the interplay between BRCA1 and 53BP1. Whereas BRCA1 stimulates DNA end resection and HR, 53BP1 inhibits resection, together with the MRN complex by protecting the DNA ends. In this way, 53BP1 promotes NHEJ.
18 In the NHEJ pathway, the DSBs are repaired by blunt end ligation through a process that is initiated by the Ku heterodimer, Ku70 and Ku80, which rapidly binds and protects the DNA ends from resection. The formed Ku heterodimer creates a scaffold to recruit DNA-PKcs, which promotes synapsis of the two DNA ends and limits the DNA end processing by Artemis nuclease. DNA-PKcs also facilitates the recruitment of the ligase complex, which is formed by DNA ligase IV, XRCC4 and XLF. XRCC4 stabilizes the binding of the DNA ligase IV and stimulates its activity, while XLF maintains the stability of the broken DNA ends during repair 5,19. The ligation is independent of sequence homology, which explains why NHEJ is error- prone. This repair mechanism can occur through all the cell cycle, although it is predominant in G0/G1 when the cells cannot perform HR 6.
The advantage of the NHEJ pathway is its fast kinetics, having a clear role in protecting genome integrity by suppressing chromosomal translocations 5,6.
I.3 Alternative-EJ repair mechanisms
Alternative-EJ pathways make a minor contribution to DSB repair in non- malignant cells. They are recombinational repair pathways that repair during S and G2 phases and are initiated by DNA end resection as HR does. Short resection by MRN11-CtIP might be sufficient for MMEJ, whereas extensive resection is likely to be needed to expose the homologous repeats required for SSA 20. SSA is responsible for the repair of DSBs that occur in S-phase in unreplicated DNA, as it does not require a donor sequence. In this mechanism, long regions with sequence homology in the RPA-covered ssDNA anneal with each other. Rad52 promotes the joining between the interspersed nucleotide repeats, but the sequence between the repeats is deleted in the repair product 5,6,20. The other alternative-EJ pathway, MMEJ, can serve as a back-up pathway in cells that are deficient in either NHEJ or HR. However, MMEJ also seems to be necessary for the repair of a subset of DSBs in cells with full NHEJ and HR capacity. MMEJ involves micro-homology regions and is promoted by PARP1, which competes with Ku and have end-bridging activity 6,19. This
19 pathway joins together DNA ends on different chromosomes, which generates chromosomal translocations and mutagenic rearrangements 6. For this reason, MMEJ often has harmful consequences for genomic integrity.
20
II. Chromatin remodeling in DNA repair
DSBs can arise in condensed and decondensed chromatin, independently of the degree of chromatin packaging. However, the chromatin architecture in the region that surrounds the DSB and the nucleosome packing may limit the access of the DNA-repair proteins to the break, which results into higher mutation rates in compact heterochromatin domains. Multiple mechanisms of ATP-dependent chromatin remodeling and post- translational modification of histones cooperate to provide accessibility to the damaged site during the repair 21. However, much remains to be understood about the complex and dynamic interplay among diverse histone regulatory mechanisms. This complexity is enhanced by the fact that different chromatin environments will undergo different types of rearrangements. The discovery of the importance of these reversible chromatin changes led to the “access, repair, restore” model for DNA repair, which was initially proposed in the context of nucleotide excision repair 22 and thereafter extended to other types of DNA damage 23.
One of the initial modifications in the chromatin involves the repressive histone mark H3K9me3. H3K9 methyltransferases such as SUV3-9 relocate to DSBs and favor chromatin compaction 24. The compact chromatin domain may be beneficial to stabilize the damaged ends and enhance DDR signaling, as proposed by Burgess et al. 25. Afterwards, chromatin relaxation is necessary for amplification and activation of the DDR effector pathway. The methylated H3K9 histone facilitates the recruitment of Tip60 26, which acetylates histones H2A and H4. Tip60, associated with the NuA4 chromatin remodeling complex, can also be recruited to DSBs by MDC1 27. Histone acetylation extends for hundreds of kilobases (kb) away from the break and promotes chromatin unpacking, giving relaxed chromatin structures at the site of the break to allow the recruitment of the DNA-repair machinery. These changes in chromatin structure actively contribute to activation of the upstream ATM/ATR-dependent signaling of the DDR 26. As a result of ] 5 I = heterochromatin domains regardless of whether the damage was produced in euchromatin or heterochromatin 5,28.
21 In heterochromatic regions, immediately after a DSB occurs, chromatin is decondensed by ATP-dependent chromatin remodelers, as for example KAP1, the SWI/SNF complex or INO80, which contribute to the establishment of a chromatin environment that allows for downstream signaling and repair events 21,28. Besides chromatin decompaction, another feature of the DDR at heterochromatic sites is the relocation of the DNA breaks to the periphery of heterochromatic domains. This relocation relies on the activation of DNA damage checkpoint kinases and requires DNA end resection 29,30. The pre-existing chromatin context at the site of the DSB influences the decision to use HR or NHEJ. DSBs are repaired with slower kinetics in heterochromatin than in euchromatin. The histone mark H3K36me3, guided by the H3K36 methyltransferase MMSET, is enriched in transcribed genes and exonic sequences and seems to be essential for the recruitment of RAD51 to DSBs 31,32. Therefore, it has been described that a locus that is being transcribed is prone to engage HR because it has higher RAD51 levels than the chromatin that shows repressive marks 33.
22
III. DNA:RNA hybrids in genome instability
Accurate DNA replication and DNA repair are crucial for the maintenance of the genome. Moreover, correct RNA synthesis and processing reactions are also important for the stability of the genome. Genomic loci with high transcriptional activity are characterized by high mutation and recombination frequencies and are associated with an increased formation of DNA:RNA hybrids and R-loops 34. These structures can initiate premature transcription termination and can be a source of genomic instability if they are not correctly processed 35–37. In a R-loop structure, an RNA molecule invades a complementary DNA duplex and hybridizes with the complementary DNA strand whereas the other strand is displaced forming a ssDNA loop (Figure 2a).
Figure 2. R-loop potential instability outcomes. a) R-loop structure. b) RNAPII stalling and collision between transcription and replication forks is a source of genome instability.
The prevention and resolution of R-loops are often coupled to RNA binding and processing factors that assemble with the nascent RNA and prevent R- loop formation. Depletion or mutation of RNA binding proteins has deleterious effects on the stability of the genome because wrongly assembled RNPs favor DNA-RNA hybridization 38,39. Defects in RNA processing can result in the production of defective transcripts, which can inhibit RNA release and export, and increase the frequency of DNA:RNA hybrid-mediated R-loop formation. Moreover, RNA processing defects can
23 induce genome instability indirectly by impairing the expression of genes that are required for chromatin regulation and DNA repair 40. Regardless of their origin, R-loops may stall RNA polymerases and lead to collisions between the transcription and the DNA replication machineries, which can result in DNA breaks (Figure 2b) 41.
Once the R-loop is formed, several pathways can contribute to its resolution. In one of them, the RNA in the DNA:RNA hybrid is degraded by ribonuclease H (RNase H)-like enzymes. Alternatively, the R-loops can be resolved by DNA:RNA helicases such as Senataxin (SETX) that unwind the DNA:RNA hybrid and in this way contribute to restore the DNA duplex 42,43. Studies about divergently transcribed enhancer RNAs have shown that the RNA exosome may also mediate the degradation of this RNAs at the DNA:RNA hybrids or promote earlier transcription termination to avoid the R-loop formation and subsequent damage. It has been demonstrated that exosome-depleted cells have a greater propensity to accumulate DNA:RNA hybrids that lead to genomic instability, which can be visualized by a high accumulation of H2AX foci 44. DNA:RNA hybrids can also be formed as a consequence of a DSB in a transcribed loci 45. For instance, the DNA supercoiling produced as a result of RNA polymerase II (RNAPII) arrest at the DNA lesions facilitates DNA strand openings that favors nascent RNA hybridization 46–48. Also, the 5’-end resection might facilitate the RNA binding to ssDNA 47. Regardless of their origin, it has been discussed whether R-loops are necessary to promote faithful repair or, contrarily, if they have to be removed in case they are formed 41. It has been recently proposed that Drosha drives RNA invasion at DSB forming DNA:RNA hybrids in order to facilitate the DDR 49. A more specific function for the DNA:RNA hybrids in DSB repair is the regulation of DNA end resection and RPA recruitment, as suggested by experiments in Schizosaccharomyces pombe (S. pombe) 47. However, there is strong experimental evidence suggesting that the persistence of R-loops compromises DNA repair 41,47,48,50. It has been shown that SETX 45 and the HR proteins BRCA1 and BRCA2 35 help to resolve R-loops, which supports the view that the DDR contributes to their removal. In all cases, RNA processing and turnover plays an important role in both the management of R-loops and the maintenance of the genomic stability 40.
24
IV. DNA damage-response RNAs
The vast majority of the genome is transcribed and many of the ncRNAs that are produced are biologically functional. Some of these ncRNAs, termed either DNA damage-response RNAs (DDRNAs) or DSB-induced RNAs (diRNAs), have been related to the assembly of DDR foci and for DNA repair by HR 49,51–56.
Transcription takes place in the vicinity of the DSBs without the need of a specialized promoter sequence. It has been proposed that the resulting transcripts are processed by Dicer and Drosha, producing DDRNAs/diRNAs. DDRNAs/diRNAs have been reported in plants 52,57, insects 58, mouse and human cells 51,53–55. The function of DDRNAs/diRNAs is still unclear, but they may function as guides and promote the recruitment of protein complexes to DSBs. Once processed, DDRNAs/diRNAs associate with AGO2 to mediate DSB repair 53,56. There are observations suggesting that DDRNAs/diRNAs act in ] 1 in the coordination of DDR focus formation 54, and that a DDRNAs/diRNAs-AGO2 complex promotes RAD51 recruitment to DSBs 53. Furthermore, this DDRNA/diRNA-AGO2 interaction is required for the recruitment of the chromatin remodeling proteins MMSET and TIP60 to the DSBs to enhance local histone H4 methylation and acetylation, respectively 56. In both cases, it seems that the RNA sequence is sufficient to mediate the locus-targeting activity of the complex. However, it remains unclear if the sequence complementarity of DDRNAs/diRNAs mediates their specific localization or which partners they recruit 59. How DNA transcription is regulated around DSBs and how these transcripts are produced need to be further investigated. It has been shown that damaged loci have repressive chromatin modifications that inhibit transcription elongation at the DSBs. It is well established that ATM activation after DNA damage causes chromatin compaction, which suggest that transcription is not physically impeded by the DNA lesions itself, but is instead actively controlled by the chromatin changes that are triggered by DDR-activation 60. Other studies suggest that RNAPII might be only paused at damaged chromatin, rather than being excluded from it 61. Related to these observations, very recent research has established that there is de
25 novo transcription from the break site and that long damage-induced RNAs (dilncRNAs) are synthesized 55 (see Paper II in this thesis). However, the diRNA production has only been shown at repetitive sequences or in reporter constructs and its production at endogenous DSBs is still questioned 49,57.
26
V. Ribonucleases in DNA repair
The interconnection and cooperation between RNA processing and DNA repair are not fully characterized yet. In the past years, it has been clear that deficiencies in RNA processing can lead to DNA damage. However, more recently, many RNA processing proteins have been shown to have a dual role, being involved in the maintenance of genome stability and DNA repair3. As explained above, when a DSB occurs, dilncRNAs are synthetized from the break and need to be processed. The type III RNases Dicer and Drosha have been characterized as being involved in the processing of the dilncRNAs into its 22 nt-long small diRNA through a pathway that is similar to the microRNA (miRNA) biogenesis pathway 49,51,54,55 (see Paper II in this thesis). According to recently proposed models, diRNAs serve as guides and bind to AGO2, another type III-like RNase, that forms a scaffold to recruit RAD51 to the DSBs53. Apart from the miRNA biogenesis-like pathway, there are other endo and exonucleases that degrade RNAs at the surroundings of DSBs. Usually, these RNases have both a role in preventing DNA damage and contribute to DNA repair. For example, nascent RNAs, either from canonical or damage- induced transcription, can form DNA:RNA hybrids at the surroundings of DSBs. RNase H, which specifically degrades RNA bound to DNA, is the main RNase involved in preventing DNA damage caused by the R-loops 35,62,63. On the other hand, regulating DNA:RNA hybrid formation and degradation is necessary for the DNA end resection process 47.
Three other RNases, as XRN1 64, XRN2 50 and RRP6/ EXOSC10 44,64 have been reported to act by preventing genomic instability, minimizing the risk of DNA damage and at the same time contributing to the DDR in different ways. In yeast, XRN1, together with RRP6 and TRF4 have been shown to activate the checkpoint kinase Mec1 by promoting RPA coating of the ssDNA 64. XRN2, similar to XRN1, is a 5’-3’ exonuclease mainly implicated in transcription termination that has also been involved in genomic instability by preventing R-loop formation. In absence of XRN2, the excess of R-loops
27 inhibits DNA repair 50. RRP6/EXOSC10, a 3’-5’ exoribonuclease (exoRNase), is also recruited to DSBs and is necessary for RNA degradation in order to allow RPA incorporation, as shown in Paper I and III in this thesis. The RNases mentioned above can act together with helicases such as SETX45,65 or DDX1 48 that help unwinding the DNA:RNA hybrids or the RNA, respectively, as well as other cofactors, like the C1D family proteins 66, the nuclear exosome targeting (NEXT) complex 67 or TREX2 35, that have also been implicated, directly or indirectly, in DNA repair.
Overall, even if diRNAs have a function in regulating DDR foci formation, which is still highly controversial, RNases and RNA degradation at DSBs are really important to maintain the RNA homeostasis in the surroundings of the break.
28
VI. The exosome
The RNA exosome is a multiprotein RNase complex that participates in a multitude of cellular RNA processing and degradation events. It was discovered more than two decades ago as a complex required for the 3’- end processing of ribosomal RNAs (rRNAs) 68. This multisubunit assembly consists of a catalytically inactive nine-subunit core that associates with two additional subunits, RRP6, or its ortholog in human cells EXOSC10, and DIS3 (Figure 3) 69. More recently, a third catalytic subunit, DIS3L, has been discovered 70. RRP6/EXOSC10, DIS3 and DIS3L have 3’-5’ exoRNase activity, and DIS3 also has endoribonuclease (endoRNase) activity. RRP6/EXOSC10 and DIS3 are associated with the exosome core in the cell nucleus, whereas DIS3L is mostly cytoplasmic 71.
Figure 3. The exosome core and its catalytic subunits. The exosome core has a ring structure formed by nine subunits. This core interacts either to both or each of the two catalytic subunits, Rrp6/EXOSC10 and Rrp44/DIS, separately, to carry out the ribonuclease function of the exosome.
Structural and functional studies have revealed that the exosome core guides the RNA to either RRP6/EXOSC10 or DIS3. Moreover, there are alternatives paths for the RNA to access the DIS3 active sites: either directly72 or through an extended path that requires RRP6 to thread the RNA into the central channel of the core and stimulate the RNase activities associated with DIS3 73,74. In the nucleus, the exosome takes part in the processing and/or degradation of the nuclear-encoded RNAs and plays a crucial role in the
29 maturation of rRNAs, transfer RNAs (tRNAs), ncRNAs and small nucleolar RNAs (snoRNAs). It is aslo involved in the removal of RNA fragments and cryptic unstable transcripts (CUTs) and long ncRNAs (lncRNAs). In the cytoplasm, it is responsible for the turnover of many RNAs, including messenger RNAs (mRNAs), and is part of the quality control system 75,76.
Moreover, RRP6 and DIS3 have other interaction partners than the core exosome, and have core-independent functions that are still poorly understood (see below, section VI.4) 77.
VI.1 The exosome in RNA processing and degradation
The RNA exosome is essential for transcriptome-wide regulation of gene expression and RNA surveillance, and because of its many diverse substrates the exosome regulates diverse cellular processes across eukaryotes 78.
The exosome functions at sites of transcription and associates with several factors for efficient precursor-mRNA (pre-mRNA) processing and elongation79,80. These interactions with transcription elongation factors and RNA-binding proteins at active loci indicates that the machinery for transcription elongation and the machinery for pre-mRNA surveillance act together 81. The exosome is recruited to newly synthetized pre-mRNAs through interactions that are highly dynamic and depending on transcription (Figure 4). It is present along the complete length of protein-coding genes, with the highest concentration at the 5’-end, suggesting that the quality control mechanisms start as soon as transcription is initiated. Nevertheless, this association is independent of the pre-mRNA quality 76. The mechanisms of quality control can recognize at least three types of defect in the recently transcribed transcripts: the presence of unspliced introns, polyadenylation defects and inefficient ribonucleoprotein (RNP) assembly. In all cases, nuclear surveillance is based on the fact that correct pre-mRNP packaging and efficient processing will protect the pre-mRNAs from exoribonucleolytic attack 82.
30
Figure 4. Multiple interactions support the recruitment of the exosome to nascent pre-mRNPs. Protein-protein interactions with transcription elongation factors connect the exosome to Pol-II transcripts. Interactions with mRNA-binding proteins and cap-binding factors connect the exosome to the nascent pre-mRNA. Figure adapted from Eberle and Visa, 2015 76.
VI.2 The exosome in transcription termination
Transcription terminates when the RNA polymerase disengages the DNA and releases the RNA transcript. In coding genes, transcription termination of RNAPII occurs downstream the 3’ end processing site of the transcribed RNA, which is located approximately 20 nt downstream of the poly(A) signal (PAS). There is also a variety of non-coding transcripts that are terminated similarly to the coding genes with a PAS-dependent termination, but for most snoRNAs, small nuclear RNAs (snRNAs) and CUTs, transcription termination depends on specialized components such as the Nrd1-Nab3-Sen1 (NRD) complex 83. The NRD complex has been characterized in yeast and is known to co- precipitate with both exosome subunits and the Trf4/Air2/Mtr4 polyadenylation (TRAMP) complex. TRAMP guides the RNA transcript to the exosome, and these interactions led to the proposal that the exosome participates in RNA processing once transcription is terminated 78,84. However, experiments done in cells lacking RRP6 revealed transcription termination defects in the absence of this protein 85. Following these experiments, a reverse-torpedo model involving the 3’-5’ exonuclease
31 activity of the exosome, more specifically of DIS3, have been described in yeast. This model proposes the existence of a fail-safe mechanism that is going to be active when the 5’-torpedo mechanism is not possible due to non-productive pre-mRNA cleavage or when there is a free 3’-ended RNA86. For this reason, RNAPII backtracking events, which provide a 3’ end as an entry point for DIS3, favor exosome-dependent transcription termination87. The NRD complex does not exist in metazoans, where transcription termination is highly dependent on PASs through the torpedo model and is modulated by chromatin 88. Furthermore, in human cells, it has been proposed that EXOSC10 facilitates transcription termination by recruiting 3’ end processing factors to the PAS 89.
VI.3 The exosome in chromatin dynamics
Gene expression is regulated by the packaging of DNA in the chromatin. In many organisms, short ncRNAs are involved in the establishment of heterochromatin states, the silent form of the chromatin 90. In fission yeast, this silencing is mediated by an RNA processing pathway that involves RNA interference (RNAi) 91. Animals cells instead use the Piwi-interacting RNA (piRNA) pathway 92. In both cases, these ncRNAs act as guides in the recruitment of histone methyltransferases to the histones in the chromatin. The methylated histones act as binding sites for other factors such as HP1a that keep the chromatin condensed 93. An important regulator of ncRNAs is the exosome. In S. pombe, the 3’-5’ exoribonuclease activity of the exosome is needed for heterochromatin- associated gene silencing 94. In Drosophila melanogaster (D. melanogaster) it has been shown that the exosome is tethered to heterochromatic loci through an interaction with the histone metyltransferase SU(VAR)3-9, and this interaction mediates the association of the exosome to a subset of heterochromatic regions 95. Under normal conditions RRP6 degrades pervasive heterochromatin-associated transcripts derived from repetitive sequences from subtelomeric and pericentromeric regions to maintain the packaging of the heterochromatin95. The degradation of these pervasive transcripts is necessary to avoid interactions between the transcripts and HP1a 96, an interaction that would destabilize the packaging of the 32 heterochromatin. Instead, transcripts formed at protein coding genes are assembled into protective ribonucleoprotein complexes and are efficiently exported away from the transcription locus 97. At these regions, the exosome has an antagonistic function against the histone mark H3K56Ac, which promotes RNAPII occupancy, controlling mRNA and ncRNA expression 98.
VI.4 Core-independent functions of the exosome
As mentioned above, the eukaryotic exosome is composed by several catalytic RNase subunits – RRP6/EXOSC10, DIS3 and DIS3L – that usually are associated with the exosome core. However, it has been shown that RRP6 and DIS3 have other interaction partners and can act independently of the exosome 99. These partners may regulate autonomous functions of RRP6 and DIS3, but the mechanisms remain unknown.
One of the independent functions of RRP6 is the degradation of some specific transcripts, as shown by studies based on the use of mutant catalytic subunits that cannot interact with the exosome core 100. In Saccharomyces cerevisiae (S. cerevisiae) RRP6-targeted transcripts have been identified, as for example the 5.8S rRNA and some snRNAs, that are not stabilized after RRP6 disruption from the core exosome 78,100. In D. melanogaster, RRP6 is required for cell proliferation and error-free mitosis, while depletion of core subunits does not have the same effect 77. It has been proposed that this function is related to the degradation of cell cycle and mitosis-related transcripts by RRP6 77. Related to these observations, we have recently shown that RRP6 is involved in HR and, at least in D. melanogaster, this function does not require the exosome core (see Paper I in this thesis). Regarding RNA surveillance, RRP6 also appears to have a specialized role in transcripts that are targeted by NMD 100 but, in this case, the core- independent functions are not restricted to RRP6. In has been shown that DIS3 and the TRAMP complex are sufficient for the RNA turnover of hypomodified tRNAs 101.
33 VI.5 The exosome in DNA repair
Short ncRNAs, named as DDRNA/diRNAs, have been proposed to be required for the assembly of DDR foci and for DNA repair by HR 51–54. As a key regulator of most types of RNA in the cell, the exosome has been linked to the regulation of this type of ncRNAs. Studies in human cells have revealed that the exosome, through a SUMO- dependent interaction is complexed with SETX and is targeted to sites of transcription-induced DNA damage 65. Moreover, a genome-wide siRNA- based screen identified EXOSC10 among RNA binding proteins that contribute to HR repair 102. In S. cerevisiae, RRP6 together with TRF4 are responsible for the recruitment of RPA to promote the formation of RPA- coated ssDNA at the DSB ends 64. All these observations support that the exosome may contribute to the DDR, but a more direct role for the exosome at DSBs remains to be investigated (see Paper I and Paper III in this thesis).
34
METHODS
I. How to induce DNA double-strand breaks
DNA is, in normal conditions, subject to endogenous and exogenous damaging agents. In addition to these constant sources of damage, in order to study DNA repair, we expose the DNA to a specific sources of exogenous damage that are efficient and can be used in a reproducible and modulated manner. As explained in the introduction, damage comes in several varieties, and the focus on this thesis is the study of DSB repair. Many of the well-known exogenous damaging agents that are used as research tools are chemotherapeutic drugs such as DNA-alkylating agents, crosslinking agents, topoisomerase inhibitors like etoposide or replication inhibitors such as hydroxyurea 5,103. Another well-established but more variable source, as it produced different types of DNA damage, is the use of radiation. There are different types of radiation, being the most used the UV and 103.
Finally, in order to study site-specific DSBs some methods similar to the ones used for gene editing, have been developed. These methods involve the expression of homing endonucleases such as I-PpoI or I-SceI, restriction enzymes as AsiSI, zinc finger nucleases, TALE nucleases and, more recently, CRISPR-Cas9 104. In the present thesis, different methods have been used to produce DSBs, as described below.
I.1 Ionizing radiation
Among all exogenous agents, is one of the most effective methods and it has been widely used. This kind of radiation leads to extensive base damage that creates both SSBs and DSBs. DNA damage in irradiated cells is
35 generated either directly, due to the high amount of energy that is delivered to a localized site in the DNA, or indirectly, through the generation of free radicals as ROS 103. The intensity of the energy that is given to the cells is measured in Grays (Gy), which are defined as the absorption of one joule of radiation energy per kilogram of matter. It has been estimated that one Gy induces around 1000 SSB and 20-40 DSBs per cell in mammalian cells 105. The principal advantage of using this method is its fast effects, which allow analyzing the damage and repair pathways on a short time period after the damage is produced. Ionizing radiation produces heterogeneous damage, not only because it gives a mixture of types of damage, but affecting different loci and different cells. Such heterogeneity can be either and advantage, as analyzing the random damage gives a general overview of the DDR, or a disadvantage when the analysis of a specific site is needed.
I.2 UV laser micro-irradiation
Fluorescence microscopy has always been a powerful tool to visualize and to study DNA damage. For this reason, together with the current accessibility of microscopes capable of delivering highly focused UV light, micro-irradiation techniques have evolved into a powerful tool to study the recruitment of DNA repair proteins and the pathways these proteins are involved in.
The generation of DNA damage through UV light is basically dependent on the laser wavelength. In general, UV-A is the wavelength used for this kind of experiments, oscillating between 350 and 405 nanometers (nm), which produce a mixture of damage types (base lesion, SSBs and DSBs). In order to enhance the production of DSBs, exogenous photosensitizers such as bromodeoxyuridine (BrdU) are given to the cells and incorporated into the DNA before micro-irradiation 106. After BrdU incorporation, usually for 24h, live cells are placed in the microscope where a small field or stripe per cell is selected and micro-irradiated. The micro-irradiated field constitutes a highly localized pattern, which gives the major advantage of this method, facilitating to establish the recruitment of specific molecules of interest to
36 the damaged area. This recruitment can be easily monitored with live cell imaging as the damage is already produced with a fluorescence microscope.
I.3 Site-specific DSBs
The need of studying the DDR in specific sites of the genome led to the development of site-specific DSB induction methods. The sequence specificity of this methods allow for detailed molecular studies of the DSB- flanking regions by chromatin immunoprecipitation (ChIP), DNA:RNA hybrid immunoprecipitation (DRIP), RT-PCR and RNA-seq. These methods produce only DSBs and avoid other deleterious consequences of radiation that could cause indirect effects. The principal limitations are related to timing of expression of the enzymes and the efficiency of the cleavage.
I.3.1 I-PpoI endonuclease
The generation of sequence-specific DSBs was established with the yeast homothallic (HO) endonuclease and translated to human cells overexpressing the homing endonucleases I-PpoI and I-SceI 7. I-PpoI, from Physaurum, cleaves the sequence CTCTCTTAANGGTAGC within the 28S rDNA sequence in human cells and has been used to study the DDR inside the nucleolus 107. A special feature of the rDNA locus is that it is mainly composed of repetitive sequences. Moreover, there are also partial copies of the rDNA that have spread through the genome and remain inserted in long introns of genes such as RYR2 or DAB1. Therefore, this system provides the possibility of studying and comparing DSBs produced both at high-copy and single-copy loci.
I.3.2 AsiSI restriction enzyme
The restriction enzyme AsiSI has the recognition sequence of 8 basepairs (bps) GCGATNCGC that cuts rarely the human genome. In fact, it has been experimentaly shown that AsiSI produces approximately about 80 DSBs at
37 the human genome when expressed in U2OS cells 45. This restriction enzyme was expressed fused to a modified estrogen receptor ligand binding domain (AsiSI-ER) and stably transfected to U2OS cells, which forms the so-called “DSBs inducible via AsiSI (DIvA) system”. In this system, AsiSI is mainly found in the cytoplasm and it is translocated to the nucleus upon exposure of the cells to 4-hydroxytamoxifen (4-OHT) 108. The main advantage for this system is the time control of inducible site- specific DSBs. This system has been widely used for high-throughput analysis of DNA repair proteins by ChIP-seq 33,45,109, transcription 33,110 and DNA:RNA hybrid formation at DSBs by DRIP and DRIP-seq 45,49.
I.3.3 GFP reporters
The GFP reporters to study the DSB repair arose from the need to investigate the repair pathway choices. These reporters are based on engineered GFP genes containing a recognition site for I-SceI, a yeast mitochondrial intron-coded homing endonuclease. There is a different reporter cassette for each repair pathway (Figure 5). The assay is based in that, initially, all of them are GFP negative and successful repair of the I- SceI-induced DSBs restores a functional GFP gene that is expressed and produces a fluorescent signal that is measured by flow cytometry 111. The HR reporter cassette, named DR-GFP, contains two mutated copies of the GFP gene with a 22 bp deletion that has been replaced by two I-SceI sites with inverted orientation. Repair by HR reconstitutes the GFP gene by intramolecular o intermolecular gene conversion 111,112.
The NHEJ cassette, also called EJ5-GFP, is composed of the GFP gene with a long 3 kb intron flanked by recognition sequences for the I-SceI endonuclease. If repaired by NHEJ the intron is removed and the GFP function is restored. As this reporter measures total-NHEJ, a novel reporter, EJ2-GFP, was developed to measure MMEJ. In this cassette, the GFP coding sequence is disrupted by three stop codons flanked by I-SceI recognition sites that have 8 bp sequence of microhomology, which, if annealed by MMEJ, will restore the GFP coding frame 111,113.
38 To study the SSA pathway, the SA-GFP reporter cassette was developed consisting of two GFP fragments 5’GFP and SceI-GFP3’ which have 266 bp of homology. The I-SceI-induced DSB at GFP3’ when repaired by SSA anneals to the homolog sequence at 5’GFP, restoring the GFP function 114. More recently, a dual reporter has been developed to study the balance between HR and NHEJ pathways. This construct is called See-Saw reporter (SSR) and measures the balance between both pathways. The SSR cassette consists of the GFP gene flanked by two truncated portions of the RFP gene, sharing 302 bp of homology with each other. The I-SceI target was inserted at the end of the GFP gene. GFP is not functional after being cut by I-SceI and is restored after being repaired by NHEJ. Repair by HR uses the 302 bp of homology to restore the RFP gene 115.
Figure 5. Schematic drawing of GFP reporters. The drawing shows the structure of each construct, the location of the I-SceI cleavage sites, and how the GFP signal is restored after resolution of the I-SceI-induced cleavage.
39
II. How to measure resection
DNA end resection is an obligatory step for the HR repair pathway. There are several ways to visualize and measure resected DNA depending on whether the DSBs are created randomly or site-specific. To monitor global changes in resection, the most commonly used technique has been immunofluorescence. By this technique, it is possible to measure the ssDNA either indirectly, quantifying RPA-coated ssDNA, or directly, measuring the exposure of ssDNA-containing BrdU. However, this is a low-resolution technique that cannot monitor changes in resection length or speed. Recently, high-resolution techniques have been developed based on ChIP or quantitative PCR (qPCR) in site-specific double strand breaks. One such method is the quantitative analysis of single-stranded DNA method (QAOS). Another recent development is the measurement of individual DNA fibers after random damage as is done in the single molecule analysis of resected tracks (SMART) 116.
II.1 QAOS
This quantitative analysis measures the amount of ssDNA that is produced at specific position in the vicinity of the DSB by qPCR. The principle of this method is the use of restriction enzymes that recognize and cut only double-stranded DNA. QAOS requires the design of PCR primers flanking the restriction enzyme cut and at different distances from the DSB that is being analyzed, produced by a site-specific DSB method. If DNA end resection occurs, the restriction enzymes are not able to cut the ssDNA, therefore the primers can amplify that region. Then, the quantified ssDNA is calculated relative to the ssDNA amount on a reference sample. This method was firstly developed in yeast 117, and adapted to human at the DIvA cell line 118.
In DIvA cells, genomic DNA is isolated after AsiSI induction and digested or mock digested with specific restriction enzymes that cut at the proximity of the AsiSI target and a control that cuts in a non-damaged region, allowing to measure the percentage of ssDNA generated after resection 118. With 40 this method is possible to measure the percentage of cells that are resecting and how far from the DSB resection is occurring. Moreover, by using QAOS, it is possible to differentiate the resection patterns in a sequence-specific manner and discern, for example, between transcribed and non-transcribed regions.
II.2 SMART
The single molecule analysis of resected tracks (SMART) is a novel assay to study the extent of DNA end resection in random DSBs that can be produced by any kind of DNA-damaging agent. It is an immunofluorescence-based technique that relies on a BrdU treatment to label the DNA during replication. The genomic DNA is gently isolated and stretched by DNA-combing onto silanized coverslips to obtain individual DNA fibers. This procedure of isolation and DNA-combing is standardized to ensure reproducibility.
The ssDNA is detected by immunofluorescence using an anti-BrdU antibody, and the method relies on the fact that the BrdU is only exposed in single-stranded DNA. The resected ssDNA can then be visualized and the length of the stretched fibers can be measured by fluorescence microscopy15,116. It is considered a quantitative method as the analysis involves a large number of fibers, which gives an overview of the resection patterns in the whole cell culture. However, only the DSB that have been resected can be visualized and it is not possible to differentiate the type of sequences that are being resected.
41
III. RNA analysis at DSBs
Controversial results from RNA studies in the DNA repair field 49,57 brought us to search for a system in which to investigate transcription and RNA processing at DSBs. The RNA studies presented in this thesis are based on site-directed DSBs formed by expression of the I-PpoI endonuclease. The investigated loci have been the DSB produced at the 28S rDNA, a repetitive and highly expressed locus, and a DSB produced by the same endonuclease in a small insertion of the 28S that is located in an intron of the RYR2 gene (Figure 6). Although the DNA sequences in the vicinity of the RYR2 DSB share homology with the 28S rDNA, they are divergent enough to allow for the design of specific probes to analyze these two loci independently from each other.
Given the high degree of evolutionary conservation of the rDNA locus, analyses of I-PpoI-induced DSBs can be carried out in samples from different organisms. In this thesis, the generation of I-PpoI-induced DSBs both in human HeLa cells and in mouse embryonic stem cells has been particularly useful (see Paper II).
Figure 6. I-PpoI induced DSBs. The drawing illustrates the structure of the rDNA and RYR2 loci. The arrowheads show the location of the I-PpoI target sequence at the 28S rDNA (left) and the inserted fragment in an intron of the RYR2 gene (right).
III.1 ssRT-qPCR
In strand-specific reverse transcription (ssRT), the RT reaction is primed with specific oligonucleotides for the selective synthesis of one or a few cDNAs of interest. In this way, strand-specific information can be obtained, which is particularly relevant for the quantification of RNAs that are produced in loci where both strands are transcribed.
42 The analysis of diRNAs is difficult to detect in loci that are already transcribed before the DSB is produced. For this reason, and hypothesizing that if there is transcription after a DSB it will start from the break and it will proceed in both upstream and downstream directions 58, strand- specific analysis of the RNAs was the method of choice for quantitative and accurate analyses of changes in RNA levels before and after DSB induction. The strand-specific primers were designed to reverse-transcribe either the RNA synthesized upstream the break from the negative DNA strand, or the RNA synthesized downstream the break from the positive strand (Figure 7). The latter will also reverse-transcribe the RNA coming from the canonical- RNA transcription from the promoters of the 28S or RYR2 genes. For each reverse transcription reaction, strand-specific primers from the region of interest and for a reference gene were used and amplified by quantitative PCR to measure the RNA levels (see Paper II and Paper III).
Figure 7. The figure shows RNA transcription and polarities at both sites of the DSBs. Arrows define the ssRT oligos (1) and the qPCR primers used for amplification of the strand-specific transcript (2).
III.2 RNA-seq
Next-generation sequencing of RNA is, nowadays, widely used to study the presence and the quality of RNAs in a sample. There exist different platforms and approaches in order to sequence the RNA depending on the type of RNA and the RNA length. For small RNA-seq, as diRNAs, the sequencing platform that has been mostly used is Illumina. This method has low sequencing error rate, particularly important for short RNAs, and offers a greater sequencing depth, allowing to detect low expressed transcripts119.
43 RNA-seq analyses have confirmed the existence of diRNAs in previous reports 49,51,55,57,58. However, sequencing at unique genomic loci in a natural chromatin context has been unsuccessful or with low read quality. I-PpoI endonuclease has been the system of choice in order to address this adversity, allowing to reach high sequencing depth due to its target sites both at the genomic repetitive region of rDNA and unique intronic sites, as explained below (see Paper II). The principal advantage of this method is that small RNA-seq reveals unknown sequences that cannot be analyzed by PCR-based methods.
44
AIMS
The general aim of the project was to understand the role of the exosome in DNA repair. This work will provide new knowledge about the function of the exosome and will contribute to a better understanding of the fundamental mechanisms that act to preserve the stability of the genome.
The exosome is responsible for the processing and degradation of many different types of RNAs, including ncRNAs. The fact that short ncRNAs might be involved in the DDR and that previous studies from the Visa lab identified exosome-interacting factors involved in DNA repair 95, led us to investigate whether the exosome plays a role in DNA repair.
The exosome consists of eleven subunits, as described above, and the focus of this thesis has been on the catalytically active ones, in particular EXOSC10 and its ortholog in D. melanogaster RRP6.
More specifically, the aims for this thesis were:
I. To determine whether the exosome was involved in the DDR and whether this role was conserved between D. melanogaster and humans. This question was studied in Paper I. II. To clarify the existence of de novo transcription at DSBs and to study diRNA production and processing. This question was studied in Paper II. III. To characterize the specific function for the exosome in DNA repair and to establish whether this function was linked to the degradation of RNAs synthesized at DSBs. This question was studied in Paper III.
45
46
RESULTS AND DISCUSSION
Paper I
The involvement of ncRNAs in the repair of DSBs was one of the major findings in the DNA repair field reported previously to this thesis (see Section IV). In Paper I, we asked whether the exosome was involved in DNA repair. Using mass-spectrometry and co-immunoprecipitation, we showed that RRP6 co-immunoprecipitates with the histone variant H2Av upon DNA damage in S2 cells of D. melanogaster, which directly links the exosome with a potential role in DNA repair. Moreover, in fly and human cells, induces a rapid relocalization of RRP6 and EXOSC10, respectively, to sites of DNA damage, and this recruitment is dependent on the ATM/ATR kinases from the DDR cascade.
RNAi was used for depletion experiments, and depletion of RRP6/EXOSC10 I] ] -induced damage, as inferred from inhibited dephosphorylation of H2Av (or H2AX in human cells). Defects in DNA repair were predicted to increase radiation sensitivity. Indeed, depletion of RRP6/EXOSC10 renders the cells more sensitive to radiation.
Immunofluorescence experiments were also carried out to study the effect of RRP6/EXOSC10 depletion on the recruitment of DNA repair factors to DDR foci. These experiments showed that the depletion impairs the recruitment of the homologous recombination factor RAD51 to the damaged sites, without affecting RAD51 levels. Moreover, RRP6/EXOSC10 interact with RAD51, as shown by co-immunoprecipitation (in S2 cells) and proximity ligation assay (in HeLa cells). The interaction could be direct or indirect. The defect in RAD51 recruitment was also observed upon overexpression of RRP6-Y361A–V5, a catalytically inactive RRP6 mutant, without affecting the interaction with RAD51. By contrast, the analysis of the 53BP1 protein by immunofluorescence revealed that its recruitment is not affected in EXOSC10-depleted cells.
47 The results presented in Paper I point to a hypothetical mechanism where the 3’-5’ ribonucleolytic activity of RRP6/EXOSC10 is required for DNA repair, degrading the RNA molecules that could inhibit the recruitment of HR factors.
Major findings:
RRP6/EXOSC10 is recruited to DSBs and its recruitment is dependent, directly or indirectly, on ATM/ATR kinases. Cells depleted of RRP6/EXOSC10 are more sensitive to radiation than control cells. RRP6/EXOSC10 interacts with RAD51. The catalytic activity of RRP6 is required for the recruitment of the homologous factor RAD51 to DSBs.
The results presented here are also consistent with data from a genome- wide siRNA screen in which depletion of EXOSC10 inhibited homologous recombination in a human cell line 102. Moreover, another study showing that, in yeast, Rrp6 and its cofactor Trf4 are necessary for the loading of RPA to ssDNA 64, supported the involvement of the exosome catalytic subunit in the assembly of the DNA repair machinery.
The analysis in Paper I has been focused on RRP6/EXOSC10, and the multisubunit structure of the exosome makes it difficult to establish whether the entire exosome has a role in DNA repair. In S2 cells, depletion of the core subunit RRP4 did not affect RAD51 recruitment, which suggests that RRP6 alone, not the entire exosome, is required for DSB repair. In human cells the depletion of exosome core subunits also reduces the levels of EXOSC10 120, and we have not been able to study the contribution of the core. However, Richard et al. 65 showed that the human EXOSC9 is recruited to DDR foci, which suggests that the entire exosome is involved in DNA repair. The most interesting issue that arises from the results presented here is the molecular mechanism by which RRP6/EXOSC10 contributes to the recruitment of RAD51 to DSBs.
48 At that moment, as shown by others, DNA damage induces the formation of short RNAs in the vicinity of the damaged sites. These RNAs are referred to as DDR RNAs (DDRNAs) or diRNAs 51,52, and have been described as Drosha- and Dicer-dependent RNA products that act together with AGO2 to facilitate the recruitment of RAD51 to DSBs 53. The biogenesis of diRNAs is investigated in Paper II and their specific roles in DNA repair need to be further investigated. Furthermore, the prevalent roles of RRP6/EXOSC10 in the processing of many types of RNAs suggest that RRP6 and EXOSC10 might participate in diRNA processing and degradation, which would be the link between RRP6/ EXOSC10 and DSB repair. This possibility is investigated in Paper III.
http://jcs.biologists.org/content/128/6/1097
49
Paper II
In Paper II, we aimed at solving the controversy generated during recent years about diRNA biogenesis, processing and function at endogenous DSBs. Previous studies suggest that transcription takes place at DNA DSBs47,55, that the transcripts that originate at DSBs are processed by the ribonucleases Drosha and Dicer into diRNAs, and that diRNAs are involved in DNA repair by HR 51–55. However, the existence of diRNAs at endogenous loci has been questioned 49, as most of the abovementioned conclusions are based on the use of reporter constructs or on the analysis of telomeric repeats. Using the homing endonuclease I-PpoI, we have investigated diRNA production in genetically unperturbed mammalian cells at the repetitive 28S locus and at the unique RYR2 gene. In this paper, we confirmed that there is de novo synthesis of dilncRNAs by RNAPII at both the 28S and the RYR2 loci. Moreover, we showed by extensive RNA-seq that diRNAs are produced at the repetitive rDNA locus, which provides a direct evidence for diRNA production at endogenous loci in mammalian cells. Instead, these short RNAs are undetectable at unique genic and intergenic loci. Taken together, these observations imply that de novo transcription near the DSB is not sufficient for diRNA production.
In Paper II, we also revealed the existence of two subtypes of diRNAs that emerge from the 28S locus and result from the processing of dilncRNAs at this site. According to our model, one of the subpopulations arises from the dilncRNAs being able to anneal with already existing transcripts made at the same locus that harbors the DSB. The resulting dsRNA is processed by Dicer and gives rise to 21-22 nt diRNAs with a typical nucleotide-sequence bias at the 5’ end. The other diRNA population is Dicer-independent and shows a more heterogeneous size distribution. Our model proposes that the Dicer-independent diRNAs are products of dilncRNAs degradation, probably by different endoribonucleases and/or exoribonucleases that remain to be identified.
50 Surprisingly, the use of genetic knock-out strains in combination with I-PpoI cleavage and RNA-seq analysis showed that Drosha and its partner DGCR8 are dispensable for diRNA production. This has been an unexpected observation because studies from different laboratories have provided convincing evidence in support for the role of Drosha in DSB repair, and the assumption has been that Drosha was needed to produce diRNAs. Our results clearly rule out this possibility. Interestingly, in a very recent publication, Drosha has been shown to promote DNA:RNA hybrid formation 49, which provides an alternative explanation for the observed DNA repair defects in Drosha-depleted cells. In Paper II, we also showed that diRNAs from the 28S locus are incorporated into Argonaute and this observation agrees with previously published work 53,56.
Major findings:
There is de novo transcription at DSBs both at unique and repetitive sequences of the genome giving rise to dilncRNAs. DiRNAs are produced at repetitive sequences. There is a subpopulation of Dicer-dependent diRNAs of 21-22 nt. DiRNA biogenesis does not require Drosha.
In accordance with our results, other groups have also shown that de novo transcription by RNAPII, which is a key point in our model, is a general phenomenon that occurs at DSBs 47,55 and that RNAPII can initiate transcription at free DNA ends in the absence of a promoter in insect cells58. In Paper II we could characterize diRNA production at repetitive loci such as the 28S rDNA. The rDNA has been shown to be mainly repaired by HR thanks to its repetitive sequence, independently of the cell cycle stage 107. This could be the reason why diRNAs, if they are needed for HR, are easier to detect on a locus that is basically repairing by this pathway. For the same reason, whether or not diRNAs are produced at unique sites remain to be further investigated. We have estimated that less than one diRNA would be
51 produced per DSB, and therefore, it cannot be relevant. However, the main limitation on the I-PpoI system is that it does not allow time-course studies of early events. If diRNAs are produced during a short-time peak, using the I-PpoI system we might not be able to detect them.
52
Paper III
In the previous papers we have investigated the involvement of the exosome in DNA repair and the production of diRNAs at DNA DSBs. In Paper III we have further demonstrated that the exosome complex, not only EXOSC10 but also DIS3, is recruited to DSBs and is required for HR. However, EXOSC10 is the main subunit of the exosome involved in the repair of DSBs. In this paper, we intended to resolve which are the mechanisms by which EXOSC10 acts in DNA repair. By micro-irradiation and immunofluorescence experiments we showed that, in the absence of EXOSC10, cells fail to recruit the HR factors RAD51 and RPA, but not the upstream factor CtIP. In order to elucidate whether EXOSC10 acts upstream RPA, we have analyzed DNA end resection by both SMART, after inducing random damage to the cells, and QAOS, at site-specific DSBs in DIvA cells. With both systems, we showed that DNA-resected tracks get longer upon EXOSC10 depletion.
Furthermore, and based on results by Ohle et al. 47 showing that in yeast, transcription is needed for resection, we have analyzed whether the failure to recruit RPA and the deregulation of DNA end resection was due to the fact that the exosome could be involved in diRNA processing and degradation. By ssRT-qPCR and RNA-seq, we have shown that EXOSC10- depleted cells have increased levels of both dilncRNAs and diRNAs after damage. Moreover, DRIP-qPCR experiments revealed the persistence of DNA:RNA hybrids in EXOSC10-depleted cells. Surprisingly, in absence of DIS3, less RNA is synthetized upon damage and RNAPII-ChIP experiments revealed an increased accumulation of RNAPII at the surroundings of the break. With these results in mind, our working hypothesis was that EXOSC10 is required to degrade the dilncRNAs in order to allow RPA incorporation. In order to prove this hypothesis, we treated the EXOSC10-depleted cells with RNase A and showed that RNA degradation by RNase A restored the RPA recruitment. Moreover, inhibition of transcription abolished the resection defect, which confirms that transcription is needed for proper DNA end
53 resection, even though RNAs have to be removed to allow the downstream events of the HR. Taken together, we propose that the primary defect in the absence of EXOSC10 is the RPA recruitment. If cells fail to recruit RPA to ssDNA DNA end resection may be deregulated 18. Our results suggest that the excess of dilncRNAs and diRNAs that are accumulated in EXOSC10-depleted conditions compete with RPA for binding to the ssDNA. These results suggest a model where the catalytic activity of EXOSC10 is required for degradation of the newly synthetized transcripts in order to facilitate the recruitment of RPA and the proper assembly of the homologous recombination machinery.
Major findings:
EXOSC10 and DIS3 are recruited to sites of DNA damage. EXOSC10 is required to degrade the dilncRNAs at DSBs in order to allow the recruitment of RPA. De novo transcription after DNA damage is required for DNA end resection. EXOSC10 is required to limit DNA end resection.
The current findings are supported by previous results reporting that, in yeast, transcription is required for DNA end resection 47. Moreover, as seen in Paper II and by other groups, there is de novo transcription occurring at DSBs 47,55. It seems that de novo transcription is a specific feature for the HR pathway, and much remains to be investigated about the biogenesis and functional significance of the damage-induced transcription. In any case, as we show in the present Paper, the damage-induced transcripts have to be removed in order to avoid DNA:RNA hybrid formation and allow the proper recruitment of HR repair proteins 47,48.
54
CONCLUSIONS AND FUTURE PERSPECTIVES
In this thesis, we have solved the main questions that were aimed to be investigated and we have clarified many important questions about RNA at DSBs. RNA has been considered a source of genome instability 37. RNA binding proteins and ribonucleases had been related to DNA repair because they contribute to preserve the genome stability 35,39,48,50. However, in recent years, many reports, including this thesis, have revealed that RNA transcription and degradation might be a new feature of the DSB repair pathways 47,48,50,55. Moreover, as demonstrated here, EXOSC10 should be regarded as one of the actors involved in the HR pathway.
The model presented in Figure 8 summarizes the findings in this thesis. RNAPII initiates transcription of the DSB-flanking sequences producing de novo damage-induced RNA transcripts. These transcripts will be processed by either Dicer or other RNases into diRNAs. DiRNAs are incorporated into Argonaute complexes and may therefore function as guides, as proposed by others. Moreover, the transcription process is necessary to promote DNA end resection. On the other hand, the dilncRNAs, and perhaps also de diRNAs, have to be degraded because otherwise they form DNA:RNA hybrids and compete with RPA for the binding to the free ssDNA tracks produced by the long-range resection enzymes. The exosome catalytic subunit EXOSC10 degrades dilncRNAs and leaves the ssDNA free to assemble with RPA, which in turn is required for the proper regulation of the DNA end resection.
55
Figure 8. The model summarizes the findings from this thesis on the initial steps of the HR pathway that involve transcription, dilncRNA degradation, resection and RPA coating of ssDNA.
Previous studies have convincingly established that when a DSB occurs, ongoing transcription is shut down 121–123, while results demonstrate that transcription of the DSB-flanking regions takes place. However, the transcription of the sequences flanking a DSB might be intrinsically different from the transcription of the genes located near the DSB. Indeed, the neighbouring genes are controlled by bona fide promoters that respond to regulatory networks, which might not be the case at DSBs. Therefore, we argue that the onset of promoterless transcription at a DSB is not incompatible with the general inhibition of promoter-driven transcription revealed by previous studies. The role of the DIS3 subunit of the exosome needs further investigation. The results presented in Paper III suggest that DIS3 might contribute to the termination of stalled RNAPII molecules, and that in the absence of such termination, the synthesis of dilncRNAs is inhibited. 56 The main open question that is nowadays under investigation is the function of these RNAs in the repair of DSBs. Performing RNA analysis with a system that allows for better timing control of the DNA cleavage should help to elucidate this question. Another technical aspect of relevance for studies of RNA at DSBs would be to better control the inactivation of EXOSC10. Depletion of proteins by RNA interference, as siRNA, takes long time and can cause indirect effects to the cells, mostly during the lack of the targeted protein. The use of an inhibitor of EXOSC10 activity, together with a tight timing control of DSB production, would allow us to further study the kinetics of RNA synthesis and degradation and the correlation with other HR events, and would give more information about the functionality of RNA in DNA repair.
On the other hand, an EXOSC10 inhibitor could also be helpful as a new therapeutic for treating cancer. One possible use for such an inhibitor would be to sensitize the cancer cells to other sources of damage such as radiation in the context of a radiotherapy scheme.
Another open question that arises from this thesis is the endogenous damage that the depletion of EXOSC10 per se could cause, either directly or indirectly. In experiments of EXOSC10 depletion, we have observed that the depleted cells show signs of genomic instability, for example slightly elevated levels of H2AX phosphorylation. This damage could be caused directly, by the lack of EXOSC10, as it has been shown for other ribonucleases 50,62, or indirectly, due to its role in DNA repair. Taking into account that most of the DSBs formed in the cells are caused by endogenous sources, it is reasonable to infer that cells lacking EXOSC10 fail to repair endogenous damage, which results in increased genomic instability.
57
58
ACKNOWLEDGEMENTS
First and foremost, I would like to express my gratitude to my supervisor, Neus Visa. Your constant and structured guidance on the project but, at the same time, believing in me and my experiments, allowed me to grow as a scientist. Thanks for all your support both inside and outside the lab. My co-supervisor, Anki Östlund-Farrants, always with good advice and constructive discussions. After the last five years, I just feel the fortune I had working with inspiring people and disconnecting with incredible friends. The former members in the group: thank you for all the time we spent together, all our “panesitos” full of laughs and extended talks, and all your support and encouragement both in the lab and in real life. Just for being here, a huge part of this thesis is also yours. The present members in the group: after teaching you how to take care of THE tip, I can depart peacefully knowing that this lab is going to be as ideal as it has always been. Thank you for the last months, the stress is always less with some beers. The coffee-mates: thanks for all your help every time I needed it since the first day. Everyone in MBW: it is a pleasure working with all of you. And especial thanks to the innebandy team for all the fun we had. The EMBO-STFs: first, to Gemma, thanks for allowing me to travel back to my home lab. You have always been my mentor. And thanks to the Lab-4. Our Lab-4. I still remember that recommendation letter that made a group of friends the perfect lab team. That’s what we were. Second, to the people at CABIMER, it has been much easier to finish this thesis with your help. And then my afterwork: the bodycombat people. Thanks for that way to stop our minds with a “kiai” while push-kicking the tension week after week. This thesis wouldn’t have been possible without min svensk familj, which have been much more than friends. THE vikings, feeding us with all possible bbqs and helping me to grow first as a person and then as a scientist. And the real survivors: the other PhD, sorry, now you are just the other; the tall,
59 that reach the sambo level; the independent, always here since the first day; the exiled, time to time living on my sofa; and the feminist, or how apps are useful to meet people. Impossible to forget the elderly viking, who traveled from the past to teach us. And the grumpy dog, always happy to see me again every time I cross the door. Thank you for all the trips. The castles we have visited together. All the fun we had. The food, and more food, and even more. The beers. Your #team. Mil gracias imbéciles! All the friends who didn’t feel like vikings and went back to the warm. And the always-warmer friends: thank you all to make the word distance mean “feel closer”. Cambrils, none of us lives there anymore, but every time I go you are there. Every time I need you, you are here. And Barcelona, these friends that after all, it seems that the exams we did together really linked us. And especially to one, definitely you are the start of the entire story. Família, això també és gràcies a valtros, sobretot gràcies per animar-me a emprendre aquest camí i per donar-me suport dia a dia amb les vostres trucades riguroses de cada setmana i els kilos de pernil de la iaia que he acabat important durant aquests cinc anys. And then you. I can only say a huge THANK YOU to believe in me. To believe in us.
60
REFERENCES
1. Lindahl, T. The Intrinsic Fragility of DNA (Nobel Lecture). Angew. Chemie - Int. Ed. 55, 8528–8534 (2016). 2. Turgeon, M.-O., Perry, N. J. S. & Poulogiannis, G. DNA Damage, Repair, and Cancer Metabolism. Front. Oncol. 8, (2018). 3. Mikolaskova, B. et al. Maintenance of genome stability: the unifying role of interconnections between the DNA damage response and RNA-processing pathways. Curr. Genet. (2018). doi:10.1007/s00294- 018-0819-7 4. Ranjha, L., Howard, S. M. & Cejka, P. Main steps in DNA double- strand break repair: an introduction to homologous recombination and related processes. Chromosoma 1–28 (2018). doi:10.1007/s00412-017-0658-1 5. Gospodinov, A. & Herceg, Z. Chromatin structure in double strand break repair. DNA Repair (Amst). 12, 800–10 (2013). 6. Ceccaldi, R., Rondinelli, B. & D’Andrea, A. D. Repair Pathway Choices and Consequences at the Double-Strand Break. Trends Cell Biol. xx, 1–13 (2015). 7. Polo, S. & Jackson, S. Dynamics of DNA damage response proteins at DNA breaks: a focus on protein modifications. Genes Dev. 25, 409– 33 (2011). 8. Kaniecki, K., De Tullio, L. & Greene, E. C. A change of view: homologous recombination at single-molecule resolution. Nat. Rev. Genet. (2017). doi:10.1038/nrg.2017.92 9. Shibata, A. Regulation of repair pathway choice at two-ended DNA double-strand breaks. Mutat. Res. - Fundam. Mol. Mech. Mutagen. 803–805, 51–55 (2017). 10. Mazin, A. V., Mazina, O. M., Bugreev, D. V. & Rossi, M. J. Rad54, the motor of homologous recombination. DNA Repair (Amst). 9, 286– 302 (2010). 11. Mcvey, M. et al. Eukaryotic DNA polymerases in homologous recombination. Annu. Rev. Genet. 393–421 (2016). doi:10.1146/annurev-genet-120215-035243.Eukaryotic 12. Tumini, E., Barroso, S., -Calero, C. P. & Aguilera, A. Roles of human POLD1 and POLD3 in genome stability. Sci. Rep. 6, 38873 (2016). 13. Forget, A. L. & Kowalczykowski, S. C. Single-molecule imaging brings Rad51 nucleoprotein filaments into focus. Trends in Cell Biology 20, 269–276 (2010). 14. 5 8 >? @ Ƕ7 1
61 break. 17, 11–16 (2010). 15. Huertas, P. & Cruz-García, A. Speed matters: How subtle changes in DNA end resection rate affect repair. Mol. Cell. Oncol. 2, e982964 (2016). 16. Anand, R., Ranjha, L., Cannavo, E. & Cejka, P. Phosphorylated CtIP Functions as a Co-factor of the MRE11-RAD50-NBS1 Endonuclease in DNA End Resection. Mol. Cell 64, 940–950 (2016). 17. Symington, L. S. Mechanism and Regulation of DNA End Resection in Eukaryotes. 51, 195–212 (2016). 18. Myler, L. R. et al. Single-molecule imaging reveals the mechanism of Exo1 regulation by single-stranded DNA binding proteins. Proc. Natl. Acad. Sci. 113, E1170–E1179 (2016). 19. Deriano, L. & Roth, D. B. Modernizing the nonhomologous end- joining repertoire: alternative and classical NHEJ share the stage. Annu. Rev. Genet. 47, 433–55 (2013). 20. Sallmyr, A. & Tomkinson, A. E. Repair of DNA double-strand breaks by mammalian alternative end-joining pathways. J. Biol. Chem. jbc.TM117.000375 (2018). doi:10.1074/jbc.TM117.000375 21. Fortuny, A. & Polo, S. E. The response to DNA damage in heterochromatin domains. (2018). doi:10.1007/s00412-018-0669-6 22. Smerdon, M. J. DNA repair and the role of chromatin structure. Curr. Opin. Cell Biol. 3, 422–8 (1991). 23. Green, C. M. & Almouzni, G. When repair meets chromatin. First in series on chromatin dynamics. EMBO Rep. 3, 28–33 (2002). 24. Ayrapetov, M. K., Gursoy-Yuzugullu, O., Xu, C., Xu, Y. & Price, B. D. DNA double-strand breaks promote methylation of histone H3 on lysine 9 and transient formation of repressive chromatin. Proc. Natl. Acad. Sci. 111, 9169–9174 (2014). 25. Burgess, R. C., Burman, B., Kruhlak, M. J. & Misteli, T. Activation of DNA Damage Response Signaling by Condensed Chromatin. Cell Rep. 9, 1703–1718 (2014). 26. Sun, Y. et al. Histone H3 methylation links DNA damage detection to activation of the tumour suppressor Tip60. Nat. Cell Biol. 11, 1376– 1382 (2009). 27. Stadler, J. & Richly, H. Regulation of DNA repair mechanisms: How the chromatin environment regulates the DNA damage response. Int. J. Mol. Sci. 18, (2017). 28. Price, B. D. & D’Andrea, A. D. Chromatin remodeling at DNA double- strand breaks. Cell 152, (2013). 29. Chiolo, I. et al. Double-strand breaks in heterochromatin move outside of a dynamic HP1a domain to complete recombinational repair. Cell 144, 732–744 (2011). 30. Tsouroula, K. et al. Temporal and Spatial Uncoupling of DNA Double
62 Strand Break Repair Pathways within Mammalian Heterochromatin. Mol. Cell 63, 293–305 (2016). 31. Bannister, A. J. et al. Spatial distribution of di- and tri-methyl lysine 36 of histone H3 at active genes. J. Biol. Chem. 280, 17732–17736 (2005). 32. Barski, A. et al. High-Resolution Profiling of Histone Methylations in the Human Genome. Cell 129, 823–837 (2007). 33. Aymard, F. et al. Transcriptionally active chromatin recruits homologous recombination at DNA double-strand breaks. Nat. Struct. Mol. Biol. 21, (2014). 34. Aguilera, A. The connection between transcription and genomic instability. EMBO J. 21, 195–201 (2002). 35. Bhatia, V. et al. BRCA2 prevents R-loop accumulation and associates with TREX-2 mRNA export factor PCID2. Nature 511, 362–365 (2014). 36. Santos-Pereira, J. M. & Aguilera, A. R loops: New modulators of genome dynamics and function. Nat. Rev. Genet. 16, 583–597 (2015). 37. Huertas, P. & Aguilera, A. Cotranscriptionally formed DNA:RNA hybrids mediate transcription elongation impairment and transcription-associated recombination. Mol. Cell 12, 711–721 (2003). 38. Li, X. & Manley, J. L. Inactivation of the SR protein splicing factor ASF/SF2 results in genomic instability. Cell 122, 365–378 (2005). 39. Salas-Armenteros, I. et al. Human THO–Sin3A interaction reveals new mechanisms to prevent R-loops that cause genome instability. EMBO J. e201797208 (2017). doi:10.15252/embj.201797208 40. Chan, Y. a., Hieter, P. & Stirling, P. C. Mechanisms of genome instability induced by RNA-processing defects. Trends Genet. 30, 245–253 (2014). 41. Aguilera, A. & Gómez-González, B. DNA-RNA hybrids: The risks of DNA breakage during transcription. Nat. Struct. Mol. Biol. 24, 439– 443 (2017). 42. Hamperl, S. & Cimprich, K. a. The contribution of co-transcriptional RNA:DNA hybrid structures to DNA damage and genome instability. DNA Repair (Amst). 19, 84–94 (2014). 43. Santos-Pereira, J. M. & Aguilera, A. R loops: new modulators of genome dynamics and function. Nat. Rev. Genet. 16, 583–597 (2015). 44. Pefanis, E. et al. RNA Exosome-Regulated Long Non-Coding RNA Transcription Controls Super-Enhancer Activity. Cell 161, 774–789 (2015). 45. Cohen, S. et al. Senataxin resolves RNA:DNA hybrids forming at DNA
63 double-strand breaks to prevent translocations. Nat. Commun. 9, 533 (2018). 46. Giono, L. E. et al. The RNA response to DNA damage. J. Mol. Biol. 1– 16 (2016). doi:10.1016/j.jmb.2016.03.004 47. Ohle, C. et al. Transient RNA-DNA Hybrids are Required for Efficient Double-Strand Break Repair. Cell 1–13 (2016). doi:10.1016/j.cell.2016.10.001 48. Li, L. et al. DEAD Box 1 Facilitates Removal of RNA and Homologous Recombination at DNA Double Strand Breaks. Mol. Cell. Biol. MCB.00415-16 (2016). doi:10.1128/MCB.00415-16 49. Lu, W.-T. et al. Drosha drives the formation of DNA:RNA hybrids around DNA break sites to facilitate DNA repair. Nat. Commun. 9, 532 (2018). 50. Morales, J. C. et al. XRN2 Links Transcription Termination to DNA Damage and Replication Stress. PLoS Genet. 12, e1006107 (2016). 51. Francia, S. et al. Site-specific DICER and DROSHA RNA products control the DNA-damage response. Nature 488, 231–235 (2012). 52. Wei, W. et al. A role for small RNAs in DNA double-strand break repair. Cell 149, 101–112 (2012). 53. Gao, M. et al. Ago2 facilitates Rad51 recruitment and DNA double- strand break repair by homologous recombination. Cell Res. 1–10 (2014). doi:10.1038/cr.2014.36 54. Francia, S., Cabrini, M., Matti, V. & Oldani, A. DICER , DROSHA and DNA damage-response RNAs are necessary for the secondary recruitment of DNA damage response factors. (2016). 55. Michelini, F. et al. Damage-induced lncRNAs control the DNA damage response through interaction with DDRNAs at individual double-strand breaks. Nat. Cell Biol. 19, (2017). 56. Wang, Q. & Goldstein, M. Small RNAs recruit chromatin modifying enzymes MMSET and Tip60 to reconfigure damaged DNA upon double-strain break and facilitate repair. Cancer Research (2016). doi:10.1158/0008-5472.CAN-15-2334 57. Miki, D. et al. Efficient generation of diRNAs requires components in the posttranscriptional gene silencing pathway. Sci. Rep. 7, 1–11 (2017). 58. Michalik, K. M., Böttcher, R. & Förstemann, K. A small RNA response at DNA ends in Drosophila. Nucleic Acids Res. 40, 9596–9603 (2012). 59. D’Adda di Fagagna, F. A direct role for small non-coding RNAs in DNA damage response. Trends Cell Biol. 24, 171–178 (2014). 60. Francia, S. Non-Coding RNA: Sequence-Specific Guide for Chromatin Modification and DNA Damage Signaling. Front. Genet. 6, 320 (2015). 61. Britton, S. et al. DNA damage triggers SAF-A and RNA biogenesis
64 factors exclusion from chromatin coupled to R-loops removal. Nucleic Acids Res. 42, 9047–9062 (2014). 62. Wahba, L., Amon, J. D., Koshland, D. & Vuica-Ross, M. RNase H and Multiple RNA Biogenesis Factors Cooperate to Prevent RNA:DNA Hybrids from Generating Genome Instability. Mol. Cell 44, 978–988 (2011). 63. Zimmer, A. D. & Koshland, D. Differential roles of the RNases H in preventing chromosome instability. Proc. Natl. Acad. Sci. 113, 12220–12225 (2016). 64. Manfrini, N. et al. RNA-processing proteins regulate Mec1/ATR activation by promoting generation of RPA-coated ssDNA. EMBO Rep. 16, 221–31 (2015). 65. Richard, P., Feng, S. & Manley, J. L. A SUMO-dependent interaction between Senataxin and the exosome, disrupted in the neurodegenerative disease AOA2, targets the exosome to sites of transcription-induced DNA damage. Genes Dev. 27, 2227–2232 (2013). 66. Jackson, R. A., Wu, J. S. & Chen, E. S. C1D family proteins in coordinating RNA processing, chromosome condensation and DNA damage response. Cell Div. 11, 2 (2016). 67. Blasius, M., Wagner, S. a., Choudhary, C., Bartek, J. & Jackson, S. P. A quantitative 14-3-3 interaction screen connects the nuclear exosome targeting complex to the DNA damage response. Genes Dev. 28, 1977–1982 (2014). 68. Mitchell, P., Petfalski, E., Shevchenko, A., Mann, M. & Tollervey, D. The exosome: A conserved eukaryotic RNA processing complex I ] ;o ; 8 Cell 91, 457–466 (1997). 69. Schmid, M. & Jensen, T. H. The exosome: a multipurpose RNA-decay machine. Trends Biochem. Sci. 33, 501–510 (2008). 70. Tomecki, R. et al. The human core exosome interacts with differentially localized processive RNases: hDIS3 and hDIS3L. EMBO J. 29, 2342–2357 (2010). 71. Lebreton, A., Tomecki, R., Dziembowski, A. & Séraphin, B. Endonucleolytic RNA cleavage by a eukaryotic exosome. Nature 456, 993–996 (2008). 72. Han, J. & van Hoof, A. The RNA Exosome Channeling and Direct Access Conformations Have Distinct In Vivo Functions. Cell Rep. 16, 3348–3358 (2016). 73. Zinder, J. C., Wasmuth, E. V. & Lima, C. D. Nuclear RNA Exosome at 3.1 Å Reveals Substrate Specificities, RNA Paths, and Allosteric Inhibition of Rrp44/Dis3. Mol. Cell 1–12 (2016). doi:10.1016/j.molcel.2016.09.038
65 74. Wasmuth, E. V & Lima, C. D. The Rrp6 C-terminal domain binds RNA and activates the nuclear RNA exosome. 1–15 (2016). doi:10.1093/nar/gkw1152 75. Tomecki, R., Drazkowska, K. & Dziembowski, A. Mechanisms of RNA degradation by the eukaryotic exosome. ChemBioChem 11, 938–945 (2010). 76. Eberle, A. B. & Visa, N. Quality control of mRNP biogenesis: Networking at the transcription site. Semin. Cell Dev. Biol. 32, 37–46 (2014). 77. Graham, A. C., Kiss, D. L. & Andrulis, E. D. Core exosome- independent roles for Rrp6 in cell cycle progression. Mol. Biol. Cell 20, 2242–53 (2009). 78. Fox, M. J. & Mosley, A. L. Rrp6: Integrated roles in nuclear RNA metabolism and transcription termination. Wiley Interdiscip. Rev. RNA 7, n/a-n/a (2015). 79. Hessle, V. et al. The Exosome Associates Cotranscriptionally with the Nascent Pre-mRNP through Interactions with Heterogeneous Nuclear Ribonucleoproteins. Mol. Biol. Cell 20, 3459–3470 (2009). 80. Andersen, P. R. et al. The human cap-binding complex is functionally connected to the nuclear RNA exosome. 20, 1367–1376 (2013). 81. Andrulis, E. D. et al. The RNA processing exosome is linked to elongating RNA polymerase II in Drosophila. 100, 837–841 (2002). 82. Jensen, T. H., Dower, K., Libri, D. & Rosbash, M. Early formation of mRNP: License for export or quality control? Molecular Cell 11, 1129–1138 (2003). 83. Mischo, H. E. & Proudfoot, N. J. Disengaging polymerase: Terminating RNA polymerase II transcription in budding yeast. Biochim. Biophys. Acta - Gene Regul. Mech. 1829, 174–185 (2013). 84. Kim, K., Heo, D. H., Kim, I., Suh, J. Y. & Kim, M. Exosome cofactors connect transcription termination to RNA processing by guiding terminated transcripts to the appropriate exonuclease within the nuclear exosome. J. Biol. Chem. 291, 13229–13242 (2016). 85. Fox, M. J., Gao, H., Smith-Kinnaman, W. R., Liu, Y. & Mosley, A. L. The exosome component Rrp6 is required for RNA polymerase II termination at specific targets of the Nrd1-Nab3 pathway. PLoS Genet. 11, e1004999 (2015). 86. Lemay, J. F. & Bachand, F. Fail-safe transcription termination: Because one is never enough. RNA Biol 12, 927–932 (2015). 87. Lemay, J.-F. et al. The RNA exosome promotes transcription termination of backtracked RNA polymerase II. Nat. Struct. Mol. Biol. 21, 919–926 (2014). 88. Proudfoot, N. J. Transcriptional translation in mammals: Stopping de RNA polymerase II juggernaut. Science (80-. ). 352, (2016).
66 89. 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. 38, 8015–8026 (2010). 90. Castel, S. E. & Martienssen, R. a. RNA interference in the nucleus: roles for small RNAs in transcription, epigenetics and beyond. Nature reviews. Genetics 14, 100–12 (2013). 91. Grewal, S. S. & Elgin, S. C. R. Transcription and RNAi in the formation of heterochromatin. Nature 447, 399–406 (2007). 92. Lin, H. & Yin, H. A Novel Epigenetic Mechanism in Drosophila Somatic Cells Mediated by PIWI and piRNAs. Cold Spring Harb Symp Quant Biol. 73, 273–281 (2008). 93. Elgin, S. C. R. & Reuter, G. Position-Effect Variegation, Heterochromatin Formation, andGeneSilencing in Drosophila. Cold Spring Harb. Perspect Biol. 1–26 (2013). doi:10.1101/cshperspect.a017780 94. Reyes-Turcu, F. E., Zhang, K., Zofall, M., Chen, E. & Grewal, S. I. S. Defects in RNA quality control factors reveal RNAi-independent nucleation of heterochromatin. 18, 1132–1138 (2012). 95. Eberle, A. B., Jordán-pla, A., Ga, A. & Hessle, V. An Interaction between RRP6 and SU ( VAR ) 3- 9 Targets RRP6 to Heterochromatin and Contributes to Heterochromatin Maintenance in Drosophila melanogaster. 3–9 (2015). doi:10.1371/journal.pgen.1005523 96. Keller, C. et al. HP1 Swi6 Mediates the Recognition and Destruction of Heterochromatic RNA Transcripts. Mol. Cell 47, 215–227 (2012). 97. Giacometti, S. et al. Mutually Exclusive CBC-Containing Complexes Contribute to RNA Fate. Cell Rep. 18, 2635–2650 (2017). 98. Rege, M. et al. Chromatin Dynamics and the RNA Exosome Function in Concert to Regulate Transcriptional Homeostasis. Cell Rep. 13, 1610–1622 (2015). 99. | 5 >8 }8 5 8 >8 I I Ƕ7 I functionally distinct complexes. 1–13 (2011). doi:10.1261/rna.2364811.forms 100. Callahan, K. P. & Butler, J. S. Evidence for core exosome independent function of the nuclear exoribonuclease Rrp6p. Nucleic Acids Res. 36, 6645–6655 (2008). 101. Kadaba, S., Wang, X. & Anderson, J. T. Nuclear RNA surveillance in Saccharomyces cerevisiae: Trf4p-dependent polyadenylation of nascent hypomethylated tRNA and an aberrant form of 5S rRNA. RNA 12, 508–521 (2006). 102. Adamson, B., Smogorzewska, A., Sigoillot, F. D., King, R. W. & Elledge, S. J. A genome-wide homologous recombination screen identifies the RNA-binding protein RBMX as a component of the
67 DNA-damage response. Nat. Cell Biol. 14, 318–328 (2012). 103. Mehta, A. & Haber, J. E. Sources of DNA Double-Strand Breaks and Models of Rec. Cold Spring Harb. Perspect. Biol. 6, 1–19 (2014). 104. Gasiunas, G. & Siksnys, V. RNA-dependent DNA endonuclease Cas9 of the CRISPR system: Holy Grail of genome editing? Trends Microbiol. 21, 562–567 (2013). 105. Lomax, M. E., Folkes, L. K. & O’Neill, P. Biological consequences of radiation-induced DNA damage: Relevance to radiotherapy. Clin. Oncol. 25, 578–585 (2013). 106. Holton, N. W., Andrews, J. F. & Gassman, N. R. Application of Laser Micro-irradiation for Examination of Single and Double Strand Break Repair in Mammalian Cells. J. Vis. Exp. 1–12 (2017). doi:10.3791/56265 107. Sluis, M. Van & Mcstay, B. A localized nucleolar DNA damage response facilitates recruitment of the homology-directed repair machinery independent of cell cycle stage. 1–13 (2015). doi:10.1101/gad.260703.115. 108. Massip, L., Caron, P., Iacovoni, J. S., Trouche, D. & Legube, G. Deciphering the chromatin landscape induced around DNA double strand breaks. Cell Cycle 9, 2963–2972 (2010). 109. Iacovoni, J. S. et al. High- ] >? double strand breaks in the mammalian genome. EMBO J. 29, 1446– 1457 (2010). 110. Marnef, A., Cohen, S. & Legube, G. Transcription-Coupled DNA Double-Strand Break Repair: Active Genes Need Special Care. J. Mol. Biol. 429, 1277–1288 (2017). 111. Seluanov, A., Mao, Z. & Gorbunova, V. Analysis of DNA Double- strand Break (DSB) Repair in Mammalian Cells. J. Vis. Exp. 1–6 (2010). doi:10.3791/2002 112. Pierce, A. J., Johnson, R. D., Thompson, L. H. & Jasin, M. XRCC3 promotes homology-directed repair of DNA damage in mammalian cells. Genes Dev. 13, 2633–2638 (1999). 113. Bennardo, N., Cheng, A., Huang, N. & Stark, J. M. Alternative-NHEJ Is a Mechanistically Distinct Pathway of Mammalian Chromosome Break Repair. PLoS Genet. 4, e1000110 (2008). 114. Stark, J. M., Pierce, A. J., Oh, J., Pastink, A. & Jasin, M. Genetic Steps of Mammalian Homologous Repair with Distinct Mutagenic Consequences. Mol. Cell. Biol. 24, 9305–9316 (2004). 115. Gomez-Cabello, D., Jimeno, S., Fernández-Ávila, M. J. & Huertas, P. New Tools to Study DNA Double-Strand Break Repair Pathway Choice. PLoS One 8, e77206 (2013). 116. Huertas P., C.-G. A. Single Molecule Analysis of Resection Tracks. Genome Instab. Methods Mol. Biol. Humana Press. New York, NY vol
68 1672, (2018). 117. Holstein, E. & Lydall, D. Quantitative Amplification of Single- Stranded DNA. DNA Repair Protoc. Chapter 23, 323–339 (2012). 118. Zhou, Y., Caron, P., Legube, G. & Paull, T. T. Quantitation of DNA double-strand break resection intermediates in human cells. Nucleic Acids Res. 42, 1–11 (2014). 119. Chu, Y. & Corey, D. R. RNA sequencing: platform selection, experimental design, and data interpretation. Nucleic Acid Ther. 22, 271–4 (2012). 120. Kammler, S., Lykke-Andersen, S. & Jensen, T. H. The RNA Exosome Component hRrp6 Is a Target for 5-Fluorouracil in Human Cells. Mol. Cancer Res. 6, 990–995 (2008). 121. Iannelli, F. et al. A damaged genome’s transcriptional landscape through multilayered expression profiling around in situ-mapped DNA double-strand breaks. Nat. Commun. 8, 1–7 (2017). 122. Shanbhag, N. M., Rafalska-Metcalf, I. U., Balane-Bolivar, C., Janicki, S. M. & Greenberg, R. A. An ATM-Dependent Transcriptional Silencing Program is Transmitted Through Chromatin in Cis to DNA Double Strand Breaks. 141, 970–981 (2011). 123. Kakarougkas, A. et al. Requirement for PBAF in Transcriptional Repression and Repair at DNA Breaks in Actively Transcribed Regions of Chromatin. Mol. Cell 55, 723–732 (2014).
69