The Influence of Chromatin in DNA-RNA Hybrid Metabolism

The influence of chromatin in DNA-RNA

hybrid metabolism

Juan Carlos Martínez Cañas

Tesis doctoral

Universidad de Sevilla

2020

INDEX

Introduction...... 21 1. Sources of DNA damage...... 23 1.1. Replication as a source of genome instability...... 25

1.2. Transcription as a source of genome instability...... 29

1.3. Transcription-replication conflicts as a source of genome instability...... 31

2. R loops ...... 34 2.1. Factors involved in R loop formation...... 36

2.2. The state of the RNA...... 37

2.3. The state of the DNA...... 37

2.4. R loops and genome instability...... 39

2.5. Detection of R loops throughout the genome...... 42

3. Chromatin...... 44 3.1. Histone post-translational modifications...... 45

3.2. Chromatin remodelers...... 48

3.3. Chromatin and DSB repair...... 49

3.4. Chromatin and R loops...... 50

Objectives...... 55 Results…………………………………………………………………………………….57 Chapter I-Cytosine deamination as a tool to determine DNA-RNA hybrid length, frequency and distribution...... 59 1. H3∆1-28, H3K9-23A and H4K31Q histone mutations that lead to DNA-RNA hybrids do not lead to genome instability unless AID is overexpressed...... 61

2. H3K9-23A mutation does not impair replication fork progression and suppresses the replication defects of hpr1∆...... 64

3. R loop formation in histone mutants is not related to an altered pattern of nucleosome positioning...... 67 5

4. R loops are similar in size regardless of whether they induce genome instability or not...... 70

5. Similar R loop length in several R loop-accumulating mutants...... 75

6. R loops are formed at a very low frequency...... 77

7. Bisulfite-qPCR as an alternative method to DRIP ...... 79

8. Exploring the use of AID* deaminase in vivo to detect and map R loops. .. 82

9. AID* overexpression induces mutations genome-wide preferentially in large genes and independently of their GC content and expression levels...... 86

Chapter II- Effect of the loss of chromatin remodelers and histone modifiers in DNA-RNA hybrid metabolism...... 93 1. Chromatin remodeling mutants do not enhance S9.6 signal...... 95

2. Chromatin remodeling mutants do neither lead to genome instability nor to enhanced sensitivity to AID overexpression ...... 96

3. Rtt109 loss leads to the accumulation of DNA-RNA hybrids...... 100

4. Effect of AID* overexpression in rtt109∆ mutant...... 104

5. Rtt109 prevents DNA-RNA hybrid accumulation through its catalytic activity. ……………………………………………………………………………………106

6. DNA-RNA hybrids accumulate in rtt109∆ cells in all cell cycle phases. .... 108

7. Histone H3 deposition during S phase is not decreased in rtt109∆ mutant ……………………………………………………………………………………109

8. DNA-RNA accumulation in rtt109∆ is independent of the H3K56 acetylation state...... 111

9. Mutation of H3K14 and H3K23 (Rtt109 targets) to alanine increases DNA- RNA hybrid accumulation...... 113

10. The genotoxic sensitivity of rtt109∆ mutant is not dependent on DNA-RNA hybrids...... 114

11. The Rad52 foci accumulation phenotype of rtt109∆ mutant is partially caused by DNA-RNA hybrids...... 115

12. H3K9A mutation is able to suppress the genetic instability of hpr1∆. .... 118

13. The SCR defect of rtt109∆ is not caused by DNA-RNA hybrid accumulation...... 118

14. rtt109∆ leads to the spontaneous accumulation of DSBs that are not induced by the presence of DNA-RNA hybrids...... 126

15. Deletion of HPR1 gene in an rtt109∆ background leads to a synergistic effect on genotoxic sensitivity...... 127

Discussion……………………………………………………………………………….129

1. New insights from histone mutants leading to DNA-RNA hybrid accumulation without genetic instability...... 131 2. Setting up the bisulfite deamination assay in yeast reveals that DNA-RNA hybrid length is not a relevant factor for the generation of genetic instability...... 132 3. Detection and mapping DNA-RNA hybrids by newly developed S9.6- independent methodologies based on cytosine deamination in vitro and in vivo. 134 4. A model to explain the different phenotypes of R loop-accumulating histone and hpr1∆ mutants...... 138 5. R loop or related phenotypes not detected in yeast chromatin remodeler mutants...... 140 6. A screening among histone modifiers identifies Rtt109 as a new factor leading to DNA-RNA hybrid accumulation...... 141 7. The lack of acetylation of H3K14 and H3K23 contribute to the DNA-RNA hybrid accumulation in rtt109∆...... 142 8. Defective SCR contributes to the DNA-RNA hybrid accumulation of rtt109∆. ………………………………………………………………………………………143 9. A model to explain the DNA-RNA hybrid accumulation of rtt109∆...... 146 Conclusions ...... 151

Materials and Methods…………………………………………………………………153

Yeast strains, plasmids, primers and antibodies...... 155

Media used in this study...... 164

Techniques ...... 164

Bacterial transformation...... 164 Yeast transformation and gene disruption...... 164 Drop assay...... 164 Plasmid loss assay...... 165 Recombination assay...... 165 Loss of heterozygocity (LOH) at MAT locus...... 165 Cell viability assay...... 166 AID*-induced mutation assay...... 166 Bioinformatic analyses of AID*-induced mutations...... 166 Rad52 foci detection...... 167 Cell culture synchronization and FACS analysis...... 167 Protein extraction and western blot...... 168 Chromosome spreads...... 169 CTAB DNA extraction...... 169 2D-gel electrophoresis...... 170 Bisulfite modification assay...... 170 Lift colony hybridization assay...... 171 DRIP assay...... 172 Chromatin immunoprecipitation (ChIP)...... 172 SCR intermediates analysis...... 173 Bibliography ...... 176

INDEX OF FIGURES

Introduction Figure I1. Schematic representation of DNA damage and their sources...... 24 Figure I2. Replication fork obstacles...... 27 Figure I3. Transcription-associated events that can promote genetic instability. ... 30 Figure I4. Types of transcription-replication conflicts...... 33 Figure I5. Factors that favor DNA-RNA hybridization...... 36 Figure I6. Possible mechanisms through which R loops can contribute to genome instability...... 41 Figure I7. R loop dissolution...... 42 Figure I8. Histone PTMs and chromatin remodelers alter chromatin state...... 48 Results Figure R1. Histone mutants that accumulate DNA-RNA hybrids, do not present genome instability unless AID is overexpressed...... 62 Figure R2. Study of the genotoxic sensitivity of the indicated strains after growth in medium containing hydroxyurea...... 64 Figure R3. H3K9-23A histone mutant suppresses the replication defect of hpr1∆ mutant...... 66 Figure R4. Nucleosome positioning is altered in H3∆1-28 histone mutant...... 67 Figure R5. Around 40% of genes in H3WT and H4WT exhibited loss of nucleosome positioning compared to BY4741...... 68 Figure R6. Examples of nucleosome occupancy...... 69 Figure R7. Possible conformations of DNA affect its sensitivity to sodium bisulfite generating different products that can be detected by PCR...... 71 Figure R8. Detection of R loops by PCR in different regions of the GCN4 gene. .. 72 Figure R9. R loop length in WT, R loop accumulating histone mutants and hpr1∆...... 74 Figure R10. R loop length is similar in several R loop-accumulating mutants...... 76 Figure R11. R loops are formed at a very low frequency...... 78 Figure R12. Scheme of absolute R loop quantification method...... 80

Figure R13. Comparison of bisulfite-qPCR assay with DRIP...... 82 Figure R14. AID* overexpression does not affect viability but induces mutations in CAN1 gene...... 84 Figure R15. Cell viability is not decreased after AID* overexpression...... 85 Figure R16. AID*-induced mutations in CAN1 gene...... 86 Figure R17. AID* overexpression induces mutations genome-wide...... 87 Figure R18. Number of transitions in WT, hpr1∆, H3WT, H3∆1-28, H3K9-23A and H3K31Q strains...... 87 Figure R19. Meta-analysis of AID*-induced mutations at ORFs...... 90 Figure R20. Chromatin remodeling mutants do not enhance S9.6 signal...... 96 Figure R21. Effect of AID overexpression in the levels of Rad52 foci in chromatin remodeling mutants...... 98 Figure R22. Chromatin remodeling mutants do not lead to DNA-RNA hybrid- dependent hyper-recombination...... 99 Figure R23. A screening among yeast null mutants to identify chromatin modifiers involved in DNA-RNA hybrids homeostasis...... 100 Figure R24. Rtt109 loss causes an increase in DNA-RNA hybrid accumulation. 101 Figure R25. Validation of the DNA-RNA hybrid accumulation phenotype in rtt109∆...... 103 Figure R26. Effect of AID* overexpression in rtt109∆ mutant...... 104 Figure R27. AID*-induced mutations in rtt109∆ mutant...... 105 Figure R28. Meta-analysis of AID*-induced mutations at ORFs...... 106 Figure R29. DNA-RNA hybrid accumulation phenotype in rtt109 catalytic mutants...... 107 Figure R30. rtt109∆ mutant accumulates DNA-RNA hybrids independently of the cell cycle phase...... 109 Figure R31. Histone H3 deposition is not decreased in the absence of Rtt109. .. 110 Figure R32. R loops accumulation phenotype of rtt109∆ mutant is not mediated by the acetylation state of H3K56...... 113

Figure R33. Deletion of RTT109 gene in acetylmimetic mutans for lysines 14 and 23 do not accumulate higher levels of S9.6 signal compared to single histone mutants...... 114 Figure R34. Genotoxic sensitivity of rtt109∆ mutant is not suppressed by RNase H overexpression...... 115 Figure R35. Rad52 foci accumulation phenotype of rtt109∆ mutant is partially suppressed after RNase H overexpression...... 116 Figure R36. Study of Rad52 foci accumulation in the non-acetylable H3K9A, H3K14A and H3K23A mutants in HPR1 and hpr1∆ backgrounds...... 117 Figure R37. Repair of a HO-induced DSB by sister chromatid recombination (SCR)...... 120 Figure R38. H3 rtt109∆ and H3K14A mutants share the same SCR defect...... 122 Figure R39. Overexpression of RNase H does not rescue the SCR defect of H3 rtt109∆...... 123 Figure R40. Overexpression of RNase H does not rescue the SCR defect of H3K14A mutant...... 124 Figure R41. H3K23A mutant does not have any SCR defect...... 125 Figure R42. DSB increase in rtt109∆ is not caused by DNA-RNA hybrids...... 126 Figure R43. Deletion of HPR1 gene in a rtt109∆ background leads to a synergistic effect on genotoxic sensitivity...... 128 Discusion Figure D1. Model to explain the different phenotypes associated to R formation in histone and hpr1∆ mutants...... 139 Figure D2. Model of R loop-independent genome instability associated to Rtt109 loss...... 147

RESUMEN

El genoma, el conjunto de genes de un organismo, debe transmitirse de forma fidedigna de una célula a su descendencia. Sin embargo, existen situaciones que comprometen la estabilidad del genoma dando lugar a la aparición de mutaciones, roturas en el ADN, reordenamientos cromosómicos, etc, y que en último lugar pueden desencadenar la muerte celular o incluso la aparición de tumores. La relevancia del mantenimiento de la estabilidad del genoma se hace patente en el hecho de que la evolución ha desarrollado complejos mecanismos de reparación y respuesta al daño en el ADN para prevenir la inestabilidad genética.

Sin embargo, el ADN no está exento de la acción de numerosos factores que pueden condicionar su estabilidad, desde agentes exógenos hasta el propio metabolismo celular. En concreto, se conoce que los procesos de transcripción y replicación del ADN son una fuente importante de inestabilidad genética. Aunque existen diversos motivos por los que estos dos mecanismos pueden generar daño en el ADN, uno de los más estudiados es la formación de bucles R (en inglés R loops). Estos son estructuras formadas por un híbrido de ADN-ARN que desplaza la otra hebra del ADN dejándola en forma de cadena sencilla, más susceptible al ataque de determinados agentes como la enzima modificadora de citosinas AID. Aunque los híbridos de ADN-ARN pueden desempeñar papeles fisiológicos en las células, su acumulación puede potenciar la inestabilidad genómica. Un claro ejemplo de esto lo constituyen los mutantes del complejo THO, un complejo proteico que interviene en la formación de la ribonucleopartícula uniéndose al ARNm en formación y facilitando su transporte al citoplasma. En ausencia de dichos factores, la desprotección del ARNm permite su unión al ADN molde, dando lugar a una acumulación co-transcripcional de híbridos de ADN-ARN que generan altos niveles de inestabilidad genética, principalmente detectada como recombinación entre secuencias repetidas. Asimismo, existen proteínas implicadas en la eliminación de los híbridos una vez formados, como las RNasas H y ciertas helicasas.

En el genoma de los eucariotas, el ADN se encuentra asociado con determinadas proteínas, las histonas, dando lugar a la estructura de la cromatina cuya unidad básica es el nucleosoma. Cada uno de ellos está formado por un octámero proteico que contiene dos copias de cada una de las cuatro histonas H3, H4, H2A y H2B en torno al que se enrollan 140 pb de ADN. Los extremos amino-terminales de las histonas sufren modificaciones post-traduccionales como metilaciones, acetilaciones, fosforilaciones o ADP-ribosilaciones que pueden tener consecuencias tanto para la propia estructura de la cromatina como para procesos básicos del ADN, incluida la formación de estructuras no convencionales, híbridos de ADN-ARN. Estudios previos en el laboratorio identificaron ciertas mutaciones en las colas amino-terminales de las histonas H3 y H4 que dan lugar a una acumulación de híbridos de ADN-ARN. Estos híbridos, a diferencia de los que se acumulan en otros mutantes como los del complejo THO, no dan lugar a un incremento de la inestabilidad genómica, medida por un posible aumento de la recombinación entre secuencias repetidas y la posible acumulación de focos de la recombinasa Rad52.

En esta tesis hemos profundizado en el estudio de dichos mutantes de histonas, analizando los niveles de pérdida de heterozigosidad o pérdida de cromosomas. Hemos conseguido medir la longitud y la frecuencia de los R loops gracias a la puesta a punto de una técnica basada en el uso de bisulfito sódico in vitro desarrollada previamente para el estudio de híbridos en genes de las inmunoglobulinas. Además, hemos utilizado una versión hiper-mutagénica de la AID con el fin de poder mapear los híbridos en todo el genoma in vivo. Hemos observado que los híbridos de ADN-ARN que se acumulan en los mutantes de histonas tienen la misma longitud que aquellos que están presentes en cepas silvestres u otros mutantes que dan lugar a la acumulación de híbridos como los del complejo THO. Los resultados de esta tesis, por tanto, nos han llevado a concluir que la longitud de los híbridos no es un factor determinante en la generación de inestabilidad genética mediada a través de estas estructuras.

Por último, para identificar nuevos factores de la cromatina que tengan un papel en el metabolismo de los R loops, hemos realizado una búsqueda entre mutantes en remodeladores de cromatina o modificadores de histonas. Hemos encontrado que la deleción de la acetil-transferasa Rtt109 provoca un incremento de híbridos que es dependiente de su actividad catalítica. Gracias al análisis de distintas mutaciones en las histonas H3 y H4 hemos asociado dicho incremento con dianas concretas de entre todas las descritas para Rtt109 in vivo o in vitro. Por otra parte, dado que Rtt109 tiene un papel en reparación de roturas de ADN de doble cadena, previamente descrito en nuestro laboratorio, hemos estudiado la relación entre los fenotipos de defecto en la reparación y la acumulación de híbridos de ADN-ARN. Los resultados de esta tesis indican que la acumulación de híbridos de ADN-ARN no es un impedimento para la correcta reparación de roturas de doble cadena, ni tampoco es la causa de todo el daño presente en ausencia de Rtt109. Por lo tanto, concluimos que la persistente acumulación de roturas de doble cadena favorece la formación de híbridos de ADN-ARN. Los resultados de esta tesis ponen de manifiesto la relevancia del contexto cromatínico en el que se encuentra el ADN para la formación co-transcripcional de híbridos de ADN-ARN.

INTRODUCTION

Introduction.

Just as a roller coaster track must be intact in order to safeguard the passenger lives (and their possible future offspring), the genome, which contains all the information needed to form all living organisms from a simple bacteria to a complex human being, must also be preserved. This is highly relevant as reflected by the existence of several complex mechanisms conserved through evolution to maintain the genome stability. In fact, failures to fix the DNA damage can lead to genetic alterations, such as gene loss or chromosome rearrangements, causing ultimately cellular death or even tumors.

Genome instability is the term used to refer to alterations in the DNA sequence. There are many DNA damaging agents, which are classified in exogenous and endogenous factors according to their source. The former is integrated by agents which induce DNA damage from the outside of cells and include exposure to UV and ionizing radiations and environmental carcinogens and chemicals. In contrast endogenous factors arise from the inside, as either a consequence of the own metabolism of cells, generating reactive species of oxygen, as well as mechanisms which involve DNA as a substrate, such as replication and transcription. There are many ways in which these two processes (replication and transcription) can lead to genome instability, but one of the mostly studied is R loop formation. These structures are non-B DNA structures, which appear when a co-transcriptionally synthesized RNA hybridizes with its DNA template leading to a DNA-RNA hybrid and a single stranded DNA (ssDNA) with identical sequence to the RNA forming the hybrid. Despite R loops have their own physiological roles, their unbalanced formation can enhance genome instability, since these non-B DNA structures contain a ssDNA molecule that can be more susceptible to the attack of different genotoxic agents. Furthermore, R loops can block the progression of replication forks, impeding the duplication of the genome previous to mitosis, or generating mutations that can ultimately lead to cell death or cancer.

Cells have evolved several mechanisms to avoid R loop formation and its deleterious effects, such as enzymes that degrade or unwind the RNA of the hybrid

21 within the R loop (RNase H and DNA-RNA helicases, respectively) and proteins that bind to the mRNA preventing its hybridization to the DNA template (THO complex). The lack of any of these mechanisms would enhance R loop formation. Moreover, there are several situations that favor R loop formation, such as the appearance of negative supercoiling behind the replication forks or even the chromatin state.

Inside cells, the DNA is associated to several conserved proteins, called histones, forming the chromatin. Histones associate between them forming nucleosomes around which DNA is wrapped, helping in the compaction of DNA and allowing it to fit inside the nucleus. In addition, these proteins have an active role in the metabolism of the DNA, and can enhance or repress the expression of genes according to the necessity of the cell, through post-translational modifications in the histones or changes in nucleosome positioning. All of this histone-related information code is known as epigenetics and it establishes the state of DNA, splitting it up into euchromatin (open chromatin and transcriptionally active) and heterochromatin (closed chromatin and transcriptionally inactive). These chromatin changes by themselves can affect R loop formation since they can enhance the transcription of certain sequences or create a local alteration of the B DNA structure, favoring the hybridization of the nascent RNA to its DNA template. The more transcriptionally active and open chromatin, the more likely to form R loops. Additionally, the relationship between chromatin and R loops seems to be bidirectional, since R loops can also induce changes in chromatin through the blockage of histone deposition or recruitment of protein factors at R loop- accumulating loci.

Altogether, the study of R loops is essential to understand the variability of genomes and thus, defining the factors that promote or prevent the deleterious effects of DNA-RNA hybrid accumulation is crucial for the field of genome instability. In this thesis, we try to shed light into the mechanisms by which R loop formation is enhanced after chromatin alterations, such as changes in post- translational modification of histones through mutations in histone codifying genes

22 or mutations in enzymes responsible for the modification of histones, as well as mutations in proteins that mobilize whole nucleosomes. Moreover, we also develop S9.6-independent techniques to measure the length of R loops and to map them genome wide, using bisulfite and the Activation-Induced cytidine Deaminase (AID) enzyme, as an alternative to the use of the S9.6 anti-DNA-RNA hybrid antibody (one of the main tools to detect DNA-RNA hybrids).

1. Sources of DNA damage.

As mentioned before, DNA damage can be produced by factors outside of the cell (Figure I1) and among these exogenous factors we can remark the exposure to UV and ionizing radiations, environmental carcinogens and chemicals, such as, alkylating agents, aromatic amines, N-nitrosamines, etc. The exogenous sources of DNA damage can directly affect DNA by causing abasic sites, mismatches leading to mutations, intra- or interstrand crosslinks, single strand breaks (SSBs, the most frequent DNA lesion (Tubbs and Nussenzweig, 2017)), and double strand breaks (DSBs, the most deleterious lesions in the DNA (Rich, Allen and Wyllie, 2000)). Moreover, they can induce the crosslink of proteins to the DNA, blocking its metabolic activity. In addition, IR and UV can also react with oxygen-containing molecules giving rise to the appearance of reactive oxygen species, which can attack the DNA.

Figure I1. Schematic representation of DNA damage and their sources. External factors can contribute to the appearance of DNA damage that can lead to DNA breaks and inter- and intra-strand crosslink (pink lines). The DNA metabolism is also another source of genome instability and includes replication (green) and transcription (red), as it affects the supercoiling of the DNA and can induce the formation of secondary structures such as hairpins, G-quadruplexes or R loops. R loops generated co-transcriptionally generate ssDNA more sensitive to be attacked by chemical and nucleases, and besides, can impede the progression of the replication fork as well as newly coming RNA polymerases. DNA breaks and gaps can also be generated as a consequence of either exogenous damage or different DNA-related processes.

In addition to exogenous sources of DNA damage, the own metabolism of cells threats the stability of the genome (Figure I1). The main endogenous source of DNA damage is reactive oxygen species, which are byproducts of the electron transport chain during cellular respiration, and they can induce oxidative base lesions and SSBs. Nitrogen bases can lose their exocyclic amine giving rise to mismatches and abasic sites. The deamination of nitrogen bases occurs more frequently in ssDNA compared to dsDNA and it is enhanced during active

24 replication, transcription and recombination, processes which also constitute an endogenous source of DNA damage by themselves.

1.1. Replication as a source of genome instability.

The process by which the genome is duplicated prior to mitosis is referred to as replication, and its accuracy is essential to preserve the genomic information to the offspring (reviewed in Bell and Labib (2016)). Replication is initiated bi-directionally from a well-defined and often single origin in bacteria, or from several non-well characterized origins in eukaryotes. At the beginning of S-phase, two hexamers of the MCM2-7 factors are loaded onto the origins where, in a CDK1 and Dbf4- dependent manner, are activated to form the replicative helicase core, which unwinds dsDNA ahead of the replicative polymerases. Once the DNA is unwound, the polymerase Pol α primase synthesizes a short RNA molecule that forms a DNA-RNA hybrid with the template DNA and is used as a primer for the two following polymerases going into action: Pol ε and Pol δ, in charge to replicate the leading and lagging strands, respectively. The replication-factor C, formed by Rfc1- 5 proteins (RFC), binds to 3’ end primer-template junctions and interacts with Proliferating cell nuclear antigen (PCNA) Pol30 in yeast, a clamp that encircles DNA and tethers polymerases firmly to DNA. Thus, replication can carry on from a single primer in the leading strand and discontinuously from the lagging strand which gives rise to Okazaki fragments. Replication goes on until the end of the chromosome or until two replication forks converge.

DNA replication is a very vulnerable process for cells and is very organized and controlled from the beginning to the end. Several pathways coordinate replication with cell cycle progression and with mechanisms involved in the repair of the possible DNA damage generated. As mentioned before, replication itself can be a source of DNA damage. Although replicative polymerases are high fidelity enzymes, provided of 3'-5' exonuclease activity to eliminate misincorporated nucleotides, mutations appear at a steady pace as a consequence of polymerase errors. The mutation rate depends on each polymerase, and has been studied in bacteria, yeast and humans (Nick McElhinny et al., 2010; Clausen et al., 2013; Ide

25 et al., 1993). Moreover, an up-regulation of the synthesis of deoxynucleotide triphosphates (dNTPs) could also increase the mutation rate due to a reduction in the fidelity of replicative polymerases as well as an enhanced action of translesion synthesis polymerases which are error-prone (Pai and Kearsey, 2017). Cells possess mechanisms to counteract the misincorporations such as MisMatch Repair pathway (MMR). The presence of repeated sequences in the DNA used as a template for replication can increase the possibility of mutations seen as expansions or contractions due to the slippage of polymerases, although they can also arise from DNA repair mechanisms such as MMR or Base Excision Repair (BER) and from SSBs or DSBs present in the DNA template.

The unwinding of the DNA by the action of the replicative helicase can generate positive supercoiling ahead of the replication fork, which if not properly resolved, may constitute an obstacle for the progression of replication leading to replication fork stalling or pause and breakage (Figure I2A). However, in normal conditions cells count on enzymes to eliminate the topological stress derived from unwinding the DNA. For instance, Top1 and Top2 topoisomerases relieve the tension by inducing a break on either one or the two strands of the DNA, respectively and religating the ends of the break after relaxing the physical constriction. Nevertheless, the accessibility of topoisomerases is limited in certain situations, such as when two forks encounter, and at other topological barriers such as the nuclear envelope. In these situations, topological stress is believed to be relaxed by coupling helicase action with rotation of the fork relative to the unreplicated DNA (fork rotation). Although this mechanism relaxes topological stress ahead of the forks, it does it at the expense of generating double stranded intertwines behind the fork (DNA precatenates) (Bermejo, Lai and Foiani, 2012), which have to be resolved by Top2 in order not to cause chromosome bridging during mitosis (Baxter, 2015).

Figure I2. Replication fork obstacles. A. Replication fork progression is impeded by positive supercoiling generated ahead of the fork. B. Replication progression is impeded by DNA-binding proteins (Fob1 protein in this case). C. Replication advance is blocked by secondary structures (hairpins, G4-quaruplexes or R loops) present in the DNA. D. DNA damage can stall a moving replication fork. In all panels replication goes from left to right.

The presence of DNA-binding proteins can also impede the unwinding of DNA and hence the progression of the replication fork (Figure I2B). Indeed, proteins bound to DNA can constitute strong barriers to the progression of replication forks, and some of them have even evolved to protect the replisome from collisions with the transcription machinery, such as Fob1 in yeast (Kobayashi, 2003) and Tus protein in E. coli (Hill et al., 1989). However, there are DNA-bound proteins that can lead to non-scheduled replication fork stalling and even collapse. This can be exemplified by camptothecin (CPT) treatment which blocks the Top1 release from

DNA after cleavage (Avemann et al., 1988; Ryan et al., 1991; Strumberg et al., 2000; Tsao et al., 1993). Replication forks encountering such protein adduct on the DNA lead to Top1-induced DSBs (Saleh-Gohari et al., 2005). Additional barriers for the replication advance are non-B DNA structures, such as DNA hairpins, R loops and G4-quadruplexes (Figure I2C).

Replication fork progression is coordinated with the progression of the cell cycle due to the existence of the checkpoints, which operate through transduction cascades mediated by the apical Mec1 kinase, activating ultimately Rad53 and Chk1 effector kinases. The activation of such S phase checkpoint prevents the entrance into mitosis, reduces the speed of the replication forks and increases their stability, inhibits late replication origins and regulates nuclease activity which prevents aberrant replication and recombination intermediates (reviewed in Pardo, Gomez-Gonzalez and Aguilera (2009); Su (2006); Jackson and Bartek (2009); Ciccia and Elledge (2010); Wang et al. (2020)). During replication, the accumulation of ssDNA at stalled replication forks upon encpuntering either DNA damage (Figure I2D) or other sources of obtacles, is the signal that triggers checkpoint activation. RPA is recruited to the ssDNA generated and helps in the recruitment of the Mec1/Ddc2 complex. In addition, the ends of dsDNA surrounding the ssDNA are recognized by Rad24 along with the replication proteins Rfc2-5 forming a RFC-like clamp that loads PCNA-like complex formed by Rad17/Mec3/Ddc1 onto the DNA. This RFC/PCNA-like complex acts as a DNA damage sensor that enhances Mec1 recruitment, amplifying the signal. Thus, cells do not divide until DNA damage is fixed, ensuring that the DNA is correctly and completely replicated before mitosis.

Moreover, upon encountering obstacles, replication forks can even collapse or break, leading to DSBs. In this situation, cells need to restore replication in order to complete genome duplication prior mitosis. Replication-born DSBs can be repaired by Non-Homologous End Joining (NHEJ) or Homologous Recombination (HR) (reviewed in Pardo, Gomez-Gonzalez and Aguilera (2009); Scully et al. (2019); Ceccaldi, Rondinelli and D'Andrea (2016); Marini et al. (2019); Shibata (2017)).

The choice between these two pathways is restricted to the cell cycle stage, being HR the preferred pathway to repair a replication-born DSB. In particular, in order to preserve genome integrity, DSBs are preferentially repaired using the intact sister chromatid as a template (Sister Chromatid Recombination, SCR) (Gonzalez- Barrera et al., 2003; Kadyk and Hartwell, 1992; Johnson and Jasin, 2000). Cells could also repair these DSBs by Break Induced Replication (BIR) or by other types of HR mechanisms using an ectopic homologous sequence or chromosome as a template for repair. However, these processes can compromise the stability of the genome by generating loss of heterozigocity, deletions or insertions, and gross chromosomal rearrangements.

The molecular mechanism of SCR is known to require the product of several genes, most of them belonging to the RAD52 epistatic group, such as RAD51, RAD52, RAD54 and RAD59, and other such as SAE2, SGS1, MUS81 and RRM3. To favor SCR versus other types of HR, cohesins maintain the sister chromatids together (Cortes-Ledesma and Aguilera, 2006). The chromatin context is also involved in specifically regulating SCR, as shown by a role of H3K56 acetylation (Munoz-Galvan et al., 2013) and H3K79 methylation (Conde et al., 2009) in SCR. In addition, a recent study has demonstrated the role of Rpd3 and Hda1 histone deacetylases in SCR through promoting cohesin loading (Ortega, Gomez- Gonzalez and Aguilera, 2019).

1.2. Transcription as a source of genome instability.

Transcription is the process by which RNAs are synthesized by the RNA polymerase using one of the two strands of the DNA as a template. During this process, the transcription machinery unwinds the double helix of DNA creating a bubble in the catalytic pocket of the RNA polymerase, where a short DNA-RNA hybrid of 8-9 bp is formed. Unwinding of DNA leads to the formation of positive and negative supercoiling ahead and behind the RNA polymerase, respectively (Figure I3A). Positive supercoiling can impede the further progression of the RNA polymerase, and the negative supercoiling generated behind provokes the

29 appearance of ssDNA, which is key to understand the genetic instability associated to transcription.

Figure I3. Transcription-associated events that can promote genetic instability. A. Transcription machinery unwinds DNA in order to synthetize RNA, leading to positive supercoiling ahead of the RNA polymerase hampering its further advance as well as negative supercoiling behind the RNA polymerase, allowing DNA-damaging agents to access more easily to DNA. B. Transcription can enhance the formation of DNA secondary structures such as hairpins or R loops, which can be stabilized by G4 quadruplexes.

It is known that the transcription of a DNA sequence is linked to an increase in the mutation rate of that sequence, as observed in bacteria, yeast and mammalian cells (Savic and Kanazir, 1972; Herman and Dworkin, 1971; Datta and Jinks- Robertson, 1995; Green et al., 2003). This phenomenon is termed as transcription- associated mutation (TAM). There is an asymmetry pattern of transcription-induced mutations towards the non-template strand (Green et al., 2003; Beletskii and Bhagwat, 1996; Beletskii and Bhagwat, 1998), possibly related to its enhanced single-stranded character, which makes it more sensitive to the attack of genotoxic agents, as reflected by the higher spontaneous frequency of cytosine deamination in ssDNA respect to dsDNA (Frederico, Kunkel and Shaw, 1990). This TAM asymmetry can be additionally explained by the fact that the transcription-coupled

30 nucleotide excision repair (TC-NER) removes lesions on the template strand preferably (Hanawalt and Spivak, 2008). Alternatively, this TAM asymmetry could also be explained by the presence on the non-template strand of either secondary structures (Figure I3B), such as G4-quadrruplexes, hairpins and R loops, that would leave a few nucleotides unprotected or by a transient exposure of an extended ssDNA region behind the RNA polymerase. In addition, high levels of transcription can also affect the DNA composition by increasing the amount of uracil to the detriment of thymine independently of spontaneous cytosine deamination (Kim and Jinks-Robertson, 2009).

Transcription also induces elevated levels of recombination, a phenomenon termed as transcription-associated recombination (TAR). Recombination can be directly enhanced by DNA breaks occurring as a consequence of the major accessibility of the transcribed DNA, but the current evidence supports that most of TAR events are linked to the capability of the RNA polymerase to stall the replication fork, since in yeast and human cells, transcription only enhances recombination in dividing cells (Prado and Aguilera, 2005; Gottipati et al., 2008).

1.3. Transcription-replication conflicts as a source of genome instability.

Both processes (transcription and replication) use the same linear DNA template and both can take place at the same time, so conflicts between these two machineries are unavoidable.

Depending on the direction of the transcription machinery respect to the replication fork, conflicts can be co-directional, if the two machineries move in the same direction, or head-on, if the two different machineries move in opposite convergent directions. Head-on conflicts have more dramatic consequences than co- directional conflicts. In fact, the formers enhance recombination and can pause replication forks as observed in several cellular types (Liu and Alberts, 1995; Mirkin and Mirkin, 2005; Prado and Aguilera, 2005). This higher toxicity of the head-on conflicts could be explained by the inactivation of the replicative helicases after the clash and/or the blockage of replication and transcription progression caused by

31 the increased positive supercoiling ahead of the two machineries, inhibiting both processes. In contrast, the positive supercoiling generated ahead of a replication fork would be counteracted by the negative supercoiling behind the RNA polymerase in a co-directional conflict. In addition, it has been shown that replisomes can displace an elongating RNA polymerase moving both machineries in the same direction in bacteria (French, 1992; Pomerantz and O'Donnell, 2008). Nevertheless, co-directional conflicts could enhance their toxicity when replication forks encounter an RNA polymerase blocked or backtracked, situation which appears when the RNA polymerase goes back along the DNA template, leading to displacement of the 3' end of the nascent mRNA from the catalytic pocket of the RNA polymerase, stabilizing and inactivating the enzyme.

Several mechanisms have evolved to avoid transcription-replication conflicts, especially in a head-on orientation. In bacteria, most of essential and highly transcribed genes are co-oriented with the replisome (Wang, Berkmen and Grossman, 2007). Yeast possesses proteins to block the progression of the replication machinery through highly transcribed loci such as rDNA and tDNA (Ivessa, Zhou and Zakian, 2000). In addition, it has been demonstrated in yeast and Drosophila that transcription can influence the positioning of the replicative Mcm2-7 helicase, which is restricted to low transcribed regions (Powell et al., 2015; Gros et al., 2015). In mammalian cells, transcription and replication can be spatially and temporally separated, reducing the possibility of both processes to encounter (Smirnov et al., 2014; Gilbert, 2002).

Figure I4. Types of transcription-replication conflicts. A. Schematic head-on conflict between replication fork and RNA polymerase moving in convergent opposite directions. The presence of positive supercoiling ahead of both machineries can aggravate the deleterious effects of the clash. B. Schematic co-directional conflicts between replication and transcription moving in the same direction. DNA-RNA hybrids (not represented) can also potentiate the consequences of the clash in these situations since they can stall the RNA polymerase but the consequences of head on conflicts are more deleterious.

Moreover, there are several mechanisms to release transcriptional blockages. There are proteins that cleave inside the nascent mRNA generating a 3' end to restart transcription, such as GreA and GreB in bacteria (Tehranchi et al., 2010) or TFIIS in yeast (Dutta et al., 2015). Other fators promote the release of the RNA polymerase complex from damaged DNA to repair the damage, such as Mfd and UvrD in bacteria (Baharoglu et al., 2010; Trautinger et al., 2005), as well as the proteasome in eukaryotes (Lafon et al., 2015). Moreover, the promotion of transcription termination can also release transcriptional blockages, such as it is the case for Rho in bacteria (Peters, Vangeloff and Landick, 2011), CPM and Sen1 in yeast and SETX in mammalian cells (Stirling et al., 2012; Mischo et al., 2011; Skourti-Stathaki, Proudfoot and Gromak, 2011). Replication machineries also count on helicases in order to overcome conflicts such as Rep and UvrD in E. coli (Baharoglu et al., 2010) and Rrm3 in yeast (Azvolinsky et al., 2009). The S-phase checkpoint also plays a role in avoiding conflicts by regulating the firing of origins,

33 degradating transcription complexes at highly transcribed genes close to early replication origins or releasing genes from the nuclear envelope, eliminating topological stress (Ciccia and Elledge, 2010; Bermejo et al., 2011; Cimprich and Cortez, 2008; Nguyen et al., 2010; Poli et al., 2016).

2. R loops

There is a DNA-RNA hybrid formed naturally inside the catalytic pocket of the RNA polymerase of about 8-9 bp. This DNA-RNA hybrid is unwound as soon as it exits from the RNA polymerase, but in several circumstances this hybrid can be extended or re-formed behind the transcription machinery, leading to the formation of a secondary structure known as R loop, which is comprised of a DNA-RNA hybrid and a displaced single stranded DNA with identical sequence to the RNA forming the hybrid. It is not exactly known how they are formed but the most extended hypothesis to explain it is the thread-back model, in which the nascent RNA hybridizes with its DNA template as it comes out from the catalytic pocket of the RNA polymerase. Other models argue that R loops can also be formed through DNA-RNA hybrid naturally contained in the transcription bubble. This DNA-RNA hybrid would stay aligned and would be extended by the RNA polymerase. R loops have been extensively studied and their role in genome instability is of growing interest (reviewed in Garcia-Muse and Aguilera (2019); Santos-Pereira and Aguilera (2015); Skourti-Stathaki and Proudfoot (2014)).

The DNA-RNA hybrid forming the R loop is thought to be more stable than dsDNAs, although this depends on its length and composition (Roberts and Crothers, 1992; Shaw and Arya, 2008). It presents an intermediate structure between dsDNA B- and dsRNA A-form, which is essential for its recognition by different enzymes such as DNA-RNA helicases (Sen1) or endonucleases (RNase H) (Shaw and Arya, 2008).

R loop accumulation was inferred in E. coli mutant for Topoisomerase A, in which the overexpression of RNase H enzyme, which degrades specifically the RNA of the DNA-RNA hybrid, reduced the genome instability phenotype observed at the

34 rDNA locus (Drolet et al., 1995). In eukaryotes, R loops were first detected in yeast lacking the Hpr1 protein belonging to the THO complex, involved in messenger ribonucleoparticle (mRNP) formation and export from the nucleus to the cytoplasm (Huertas and Aguilera, 2003). R loops were later on detected in human cells depleted of the splicing factor ASF/SF2 (Li and Manley, 2005), and since then multiple research works support that these secondary structures are formed in all organisms. The frequency at which R loops are formed remains unclear although DNA-RNA immunoprecipitation (DRIP) assays combined with real time qPCR analysis has estimated that they are formed at a frequency of 1-10 % at genic regions and 0.01-0.1 % at intergenic regions. In addition, these secondary structures have a very short half-life of about 10 minutes at promoters (Sanz et al., 2016).

Although these structures are generally associated to genome instability, seen as an increase in the mutation and recombination rates and chromosome rearrangements and loss, it is important not to forget that the discovery of new tools to detect and map R loops across the whole genome has demonstrated that R loops are also present in wild-type cells, revealing that they carry out several functions. Indeed, R loops are not only involved in the replication of the E. coli plasmid or mitochondrial DNA, but they are key elements in the immunoglobulin class switching which takes place in B lymphocytes of vertebrates. Morover, DNA- RNA hybrids are formed during the synthesis of the lagging strand forming Okazaki fragments in all organisms. In addition, it is though that R loops have a role in transcription, promoting efficient transcription termination through a pathway that requires Sen1/SETX protein in collaboration with the endoribonuclease Rat1/XRN2 (Skourti-Stathaki, Proudfoot and Gromak, 2011; Kawauchi et al., 2008). In this regard, R loops would be necessary to pause the RNA polymerase beyond the polyadenylation site of several genes and recruit the transcription termination machinery. In plants, R loops have been described to inhibit the expression of the long-non coding RNA (lncRNA) COOLAIR, which inhibits the bloom in Arabidopsis thaliana (Sun et al., 2013). Apart from their role in replication and transcription, recent works propose that DNA-RNA hybrids could also be important in DNA

35 damage repair, serving as a template for the repair of a DSB lacking a homologous sequence (Keskin et al., 2014) or acting as intermediates during the repair process (Ohle et al., 2016). However, the idea that DNA-RNA hybrids have a positive role in repair is controversial and other models, by contrast, propose that DNA-RNA hybrids could hamper repair (reviewed in Aguilera and Gomez-Gonzalez (2017)).

Figure I5. Factors that favor DNA-RNA hybridization. A. Proteins involved in mRNP biogenesis bind RNA as it exits from the RNA polymerase, reducing the probability of RNA to hybridize with DNA. B. Negative supercoiling relaxes DNA-DNA interactions, enhancing the chance of RNA to reanneal with its DNA template. C. DNA breaks, either DSBs as well as SSBs, release topological stress in the DNA promoting RNA hybridization.

2.1. Factors involved in R loop formation.

The spontaneous unwinding of the double helix and later RNA hybridization with its DNA template is energy costly and therefore unlikely. Although there are proteins capable to promote DNA-RNA hybrid formation such as the bacterial helicase Cas3 in the absence of ATP (Howard et al., 2011), spontaneous R loop formation is promoted by different circumstances regarding the state of the RNA, and the state of DNA.

2.2. The state of the RNA.

The state of the RNA has a key role in R loop formation as deduced by studies in several mutants for proteins belonging to THO/TREX and THSC/TREX-2 complexes or mutations in the splicing factor ASF/SF2. The THO complex is composed of Hpr1, Tho2, Mft1, Thp2 and Tex1 in yeast and THOC1-3 and THOC5-7 in human cells. These proteins interact with Sub2 (human UAP56) and Yra1 (human ALY) to form the TREX complex that works in transcription elongation, mRNA export and genome stability maintenance (Chavez et al., 2000; Strasser et al., 2002). These transcription and mRNA processing factors reduce the ability of the nascent RNA to hybridize with its DNA template, transiently open behind the RNA polymerase by promoting optimal mRNP assembly, thus coating the nascent mRNA and preventing its hybridization with the DNA template when naked (Figure I5A). Moreover, a malfunction of these factors could potentiate R loop formation by increasing the half-life of nascent RNAs; blocking transcription through the stalling of the RNA polymerase and likely stabilization of negative supercoiling behind the transcription machinery; or impeding the 3' end processing or transcription termination which ultimately would affect the RNA release from the transcription site.

2.3. The state of the DNA.

The unwinding of the DNA either by replication or transcription can lead to a transient negative supercoiling accumulated behind both machineries (Figure I5B). The presence of such transiently open DNA favors the hybridization of the nascent RNA with the DNA template. Some studies support this idea since the lack of topoisomerase TopI and TopII increases the formation of R loops at the rDNA (El Hage et al., 2010), and the lack of topoisomerase TopIIIB potentiates R loop formation at highly transcribed genes such as MYC (Yang et al., 2014).

R loop formation could also be favored by the presence of secondary structures in the DNA, such as G4 quadruplexes (Figure I3B), where four guanines bind to each other by G-G Hoogsteen base pair, forming a 4-stranded nucleic acid

37 structure. G4 quadruplexes in one DNA strand impede the realignment with its complementary DNA sequence and favor the encounter of the nascent RNA with its DNA template, especially when G4 quadruplex is formed in a sequence with a high GC skew and most of guanines are present in the non-template strand (Duquette et al., 2004; Duquette et al., 2005). Following the idea that G4 quadruplexes can stabilize R loop formation, the expression of the immunoglobulin S region (composed of G clusters followed by regions with high G density on the non-template strand) from a T7 promoter in bacteria drives R loop formation (Roy and Lieber, 2009). The inversion of the S region respect to the T7 promoter failed to form R loops. In addition, G4 quadruplexes sequencing techniques have revealed that G4 quadruplexes accumulate at promoter regions and genes (Marsico et al., 2019), in a similar pattern as R loops are present in the genome, indicating a likely co-localization between these two secondary structures.

Breaks in the DNA can also induce R loop formation (Figure I5C) as first evidenced in vitro by showing that inducing an SSB in a supercoiled plasmid containing the immunoglobulin S region under the T7 promoter led to the accumulation of DNA-RNA hybrids (Roy et al., 2010). This result suggested that the breakage releases the topological stress in the DNA, allowing its breathing and rotation and thus promoting the hybridization of the mRNA around the cleavage site. In support to this observation, it is also described an accumulation of DNA- RNA hybrids in cells treated with camptothecin (Sordet et al., 2009).

Further analyses have shown that DNA-RNA hybrid formation is not only stimulated by SSBs, but also by the presence of DSBs. In fact, SSBs and DSBs generated by micro-irradiation in human cells lead to an accumulation of DNA-RNA hybrids as detected by immunofluorescence using an inactive RNase H from E. coli (Britton et al., 2014). Moreover, γ-H2AX, a well-known DSB marker (Rogakou et al., 1998), accumulates in human cells exposed to ionizing radiation in a RNase H-sensitive manner (Li, Monckton and Godbout, 2008).

2.4. R loops and genome instability.

R loops and/or DNA-RNA hybrids have physiological roles, as mentioned above, but their unbalanced formation is linked to genome instability. One important key to understand how R loops can lead to genome instability is through the ssDNA generated after RNA annealing with its DNA template. This ssDNA is less stable than dsDNAs, dsRNAs or DNA-RNA hybrids and also it is more sensitive to the attack of DNA damaging agents or enzymes (Figure I1). In this sense, our lab demonstrated that the overexpression of the human Activation-Induced cytidine Deaminase (AID) in yeast enhanced mutations in strains lacking components of the THO complex (Gomez-Gonzalez and Aguilera, 2007). Moreover, DNA-RNA hybrids can stimulate the appearance of mutations by promoting mutagenic replication. RNA forming DNA-RNA hybrids can be used as primers in non- canonical replication as observed at rDNA locus in E. coli cells (Kogoma, 1997) and in yeast (Stuckey et al., 2015) lacking Top1 or RNase H enzymes, both in an origin-independent manner.

In addition to the role of DNA-RNA hybrids in potentiating mutations in DNA, these secondary structures can also enhance recombination and chromosome rearrangements and loss, as it has been reported in THO mutants (Huertas and Aguilera, 2003; Wellinger, Prado and Aguilera, 2006; Gomez-Gonzalez, Felipe- Abrio and Aguilera, 2009; Gomez-Gonzalez et al., 2011). In support of this, the overexpression of RNase H in THO mutants reduces Rad52 foci formation to wild- type levels and partially but significantly diminishes the recombination frequency observed in these strains. The reason why DNA-RNA hybrids are able to cause hyper-recombination lays on the DNA-RNA hybrid capability to stall a replication fork (Figure I6). In this sense, DNA-RNA hybrids accumulated in hpr1∆ mutant from THO complex are able to pause an advancing replication fork leading to an accumulation of replication intermediates as detected by 2D gel electrophoresis (Gomez-Gonzalez, Felipe-Abrio and Aguilera, 2009; Wellinger, Prado and Aguilera, 2006). Along the same line, the use of recombination systems to study replication-transcription conflicts in yeast have revealed that DNA-RNA hybrids

39 potentiate recombination between direct repeats in a RNase H-dependent manner and this genome instability phenotype was sharpened after the overexpression of the Yra1 protein, which stabilizes DNA-RNA hybrids (Garcia-Rubio et al., 2018). Moreover, several replication-associated repair factors have been shown to counteract DNA-RNA hybrids in human cells such as the BRCA1, BRCA2 and other Fanconi anemia proteins (Garcia-Rubio et al., 2015; Bhatia et al., 2014; Hatchi et al., 2015; Schwab et al., 2015). A question coming up from these studies is whether DNA-RNA hybrids themselves can cause replication fork stalling directly or in an indirect manner.

DNA-RNA hybrids can also lead to DSBs through the recruitment of the transcription-coupled nucleotide excision repair (Figure I6). This was proposed after studies in human cells depleted of topoisomerasse I and RNA processing factors (Sollier et al., 2014), in which R loops were processed into DSBs by XPF and XPG endonucleases. These enzymes recognize the flap structures corresponding to the ends of the DNA-RNA hybrids inducing a cleavage leading to a DSB, as demonstrated in vitro (Tian and Alt, 2000) or they could also insert a nick at both ends of the DNA-RNA hybrid creating a region of ssDNA which could turn into a DSB after replication. Moreover, TC-NER machinery can also be recruited at forks stalled due to the presence of DNA-RNA hybrids. Supporting this possibility, ASF/SF2-depleted human cells accumulate R loop-induced DSBs, which depend on replication and are suppressed after XPG depletion suggesting that R loops recruit TC-NER machinery after transcription-replication conflicts (Sollier et al., 2014).

Figure I6. Possible mechanisms through which R loops can contribute to genome instability. A. R loops could pause a moving replisome either directly by themselves, or indirectly through the stalling of the RNA polymerase. B. R loops can be recognized by TC-NER machinery, leading to a ssDNA gap which will be converted into DSB after replication. In addition, R loops provide ssDNA more sensitive to the attack of DNA- damaging agents.

Highlighting the relevance of R loops in genome instability, cells have evolved several mechanisms to overcome the deleterious effects of unbalanced R loop formation. Several proteins are involved in preventing DNA-RNA hybrid formation, such as those participating in RNP biogenesis (Figure I5A). However, the prevention of DNA-RNA hybrids can fail, escaping the first control step leading to their unscheduled accumulation. In these conditions, cells also possess several tools to remove them once formed. Mainly, DNA-RNA hybrid helicases (Figure I7A) are able to displace the RNA from the R loops and RNase H eliminates DNA- RNA hybrids by specifically cleaving the RNA (Figure I7B).

Figure I7. R loop dissolution. A. mRNA unwinding by DNA-RNA helicases such as Sen1/SETX. B. mRNA degradation by RNase H which recognizes specifically DNA-RNA hybrids.

2.5. Detection of R loops throughout the genome.

Given the clear link between R loop and genome instability, several strategies have been developed from the past few years to detect them. The most widely used is the S9.6 antibody, which can bind specifically to DNA-RNA hybrids. The S9.6 antibody was originally isolated by Carrico and colleagues in 1985. Mouse (BALB/c) B-cell was fused with Mouse (BALB/c) Sp2/0–Ag14 myeloma to produce monoclonal antibodies against RNA-DNA hybrids. Animals were immunized with RNA-DNA hybrids formed by using single stranded PhiX174 DNA as template for E. coli DNA dependent RNA polymerase (Boguslawski et al., 1986). Many labs have taken advantage of this antibody and have used it to map DNA-RNA hybrids throughout the genome in yeast (Chan et al., 2014; El Hage et al., 2014; Wahba et al., 2016) and mammals (Ginno et al., 2013; Ginno et al., 2012; Nadel et al., 2015; Sanz et al., 2016; Stork et al., 2016). Most of the methods using the S9.6 antibody are based on the capture of DNA-RNA hybrids from digested DNA followed by the sequencing of either the DNA (DRIP-seq) or the RNA (DRIP-c-Seq) (Chedin, 2016). These studies showed that DNA-RNA hybrids accumulate at highly transcribed genes independently of the RNA polymerase involved, especially at the

42 rDNA locus, and also at promoters, telomeres and other repetitive regions such as transposons or common fragile sites (Chan et al., 2014; Hartono et al., 2018; Wahba et al., 2016; Sanz et al., 2016).

A few authors have revealed that S9.6 antibody also presents high specificity for dsRNAs (Phillips et al., 2013; Hartono et al., 2018), which could reduce the DNA- RNA immunoprecipitation efficiency and have modified the original protocol adding an RNase III in vitro treatment prior to immunoprecipitation with S9.6 (Hartono et al., 2018; Svikovic et al., 2019), since this enzyme degrades dsRNAs. Alternative techniques to S9.6 antibody have been developed such as the use of bisulfite, which converts specifically cytosines into uraciles of the ssDNA displaced by the DNA-RNA hybrid (Ginno et al., 2012; Yu et al., 2003; Huang et al., 2006), the use of the hybrid-binding domain of the RNase H enzyme fused either to GFP (HB- GFP) (Bhatia et al., 2014) or to MNase (Yan et al., 2019) as well as the use of an inactive RNase H in chromatin immunoprecipitation (Chen et al., 2017) which is able to bind DNA-RNA hybrids without resolving them.

Apart from all these techniques, there is an additional tool to indirectly detect R loops based on the heterologous expression of the B-cell specific enzyme AID in yeast (Gomez-Gonzalez and Aguilera, 2007). AID is able to deaminate cytosines of ssDNA turning them into uracils. This enzyme is present in mammals and is involved in triggering somatic hypermutation and class switch recombination in an R loop dependent manner (Muramatsu et al., 2000; Revy et al., 2000). Since first used (Gomez-Gonzalez and Aguilera, 2007), several studies have overexpressed human AID in yeast in order to enhance R loop-dependent genome instability phenotypes (Garcia-Pichardo et al., 2017; Gomez-Gonzalez and Aguilera, 2009; Mischo et al., 2011; Garcia-Benitez, Gaillard and Aguilera, 2017). There is an interesting variant of this enzyme that was named superAID or AID* that is more mutagenic (Wang et al., 2009). AID* has been used in some studies to reproduce mutations found in cancers and to discover regions of the genome more prone to be mutated (Taylor et al., 2013; Taylor, Wu and Rada, 2014), although it has never been used to map R loops.

Independently of the method used to detect R loops, RNase H is a key element in all these techniques that helps to decipher if the signal measured comes from DNA-RNA hybrids or not. RNase H can also be per se a good tool to infer R loops since its overexpression is able to suppress genome instability phenotypes such as growth defects, hyper-recombination or increased DNA damage foci.

3. Chromatin.

Every single linear chromosome is comprised of DNA whose length can be up to 2 meters if disposed in a straight line in eukaryotes. This creates a challenge that is overcome by the binding of the DNA to several proteins called histones, which help to pack the DNA into chromatin, making possible its storage in a very tiny space, such as the nucleus. The basic chromatin unit is the nucleosome. Every single nucleosome is composed of two copies of the four core histone proteins (H2A, H2B, H3 and H4) around which 146 bp of DNA (or 1.65 turns of the helix) is wrapped. Different nucleosomes bind each other through histone H1, facilitating higher order compaction, such as 30 nm fiber and finally, chromosomes, which are visible during meiotic and mitotic divisions.

Chromatin is a very dynamic structure that is continuously changing. This property is given by several factors that modify the composition of the histones and nucleosomes. This fact provides chromatin not only a role for packing DNA inside cells, but also a role in regulating processes that use DNA as a template, including replication, transcription and DNA repair and recombination. The regulation of these processes requires the disruption of DNA-histone interactions in order to allow the access of the proteins involved in these DNA related mechanisms. In this line, cells possess enzymes to covalently modify the histone residues what helps to directly weaken DNA-histone interactions or to indirectly recruit other factors able to remove, slide or replace whole nucleosomes. Histone post-translational modifications (PTMs) and chromatin remodelers are key elements in determining

44 the chromatin state, and they function by opening or closing the chromatin, facilitating or impeding the access of several factors to DNA.

3.1. Histone post-translational modifications.

All the core histones of nucleosomes share a common structure formed by three α- helixes connected by two loops, which facilitates the pairing between H2A with H2B and H3 with H4. The two H3-H4 pairs interact with each other through the histone H3 of each pair forming a (H3-H4)2 tetramer to which two H2A-H2B dimers bind through interactions with the two histones H4 of the tetramer. The C- and N- terminal tails of histones protrude outside the globular domain of histones, which let them interact with histones of different nucleosomes. For this reason, most of histone PTMs take place at the histone tails, which are more accessible, although there are several PTMs affecting some residues of the histone globular domain, which locate at the lateral surface of the nucleosome.

PTMs can directly modify the chromatin structure by relaxing or strengthening the interactions of histones with DNA or they can serve as binding sites for the recruitment of non-histone proteins that ultimately alter chromatin (Figure I8). Several PTMs have been described, including acetylation, methylation, phosphorylation, ubiquitinylation, sumoylation, ADP-ribosylation and deimination. These reactions are catalyzed by specific enzymes which can be classified according to the kind of modification they insert, such as histone acetyltransferases, hystone methyl-transferases, kinases and ubiquitin ligases. In addition, there are enzymes to antagonize these PTMs by erasing them, such as histone de-acetylases, histone de-methylases, phosphatases, etc. PTMs can also crosstalk between them at different levels generating a rich code of epigenetics. In this way, PTMs can establish a global chromatin environment, which divides chromatin into compartments (euchromatin and heterochromatin) and they orchestrate biological tasks based on DNA, modifying chromatin to favor or impede the completion of a specific task.

Histone PTMs have been extensively studied due to the development of mass spectrometry techniques and chromatin immunoprecipitation assays using antibodies that specifically recognizes some of the signatures inserted by histone modifiers. The coupling of ChIP techniques to whole-genome sequencing methodologies have started to shed light into the distribution of these marks throughout the genome and the mechanisms they are involved in. Regarding transcription, high levels of acetylation in histone H3 and H4 tails, trimethylation of H3K36 and H3K79 and ubiquitinylation of H2AB are associated to active genes, whereas methylation of H3K9 and H3K27 (this two last not conserved in budding yeast and the former present in fission yeast) and ubiquitinylation of H2AK119 associate to gene repression (Kouzarides, 2007). However, the histone mark itself is not everything, since the context of the histone modification has also a role in regulating the ultimate function of a specific modification. In this sense, methylation of H3K9 can have an activator or repressor effect depending on whether this mark is present at the coding region or at the promoter, respectively. Apart from the role in transcription of H3K9 as activator (through acetylation) or repressor (through methylation), this residue plays also a role in defining heterocrhomatin at centromeric regions and it has a connection with H3S10 phosphorylation (Fischle et al., 2005), a mark of chromosome condensation during mitosis, also linked to genome instability if accumulated outside mitosis.

Relative to replication, the state of chromatin can also either favor or hinder replication fork progression, as it has been shown that heterochromatin regions are intrinsically more difficult to replicate than euchromatin (Devbhandari et al., 2017; Janssen, Colmenares and Karpen, 2018; Kurat et al., 2017). To overcome nucleosome barriers, replication forks count on several factors that travel with the forks or are recruited when needed. The progression of the replication fork requires, on one hand, the removal of nucleosomes ahead of the fork (nucleosome disassembly) and on the other hand, their assembly behind the replication fork either through the incorporation of newly synthesized histones or through the recycling of the parental nucleosomes. This last process is very important to

46 maintain the epigenetic information and involves, among others factors, chaperones, histone modifiers and ATP-dependent chromatin remodelers.

Replication-coupled nucleosome assembly of newly synthesized histones requires H3K56 acetylation by Rtt109 in yeast (Masumoto et al., 2005). The mechanism also involves the Asf1 chaperone to present H3-H4 dimers to Rtt109 acetyltransferase, which introduces an acetyl group at H3K56 of newly synthesized histones prior to their incorporation into the replicating DNA by CAF1 and Rtt106 factors. H3K56 acetylation is cell cycle regulated appearing only during S phase. However, this mark persists for longer when cells are exposed to DNA damage agents, such as CPT. The persistent presence of acetylated H3K56 under DNA damage conditions correlates with a delay in the disappearance of H2A-P, a DSB mark. Moreover, both H2A-P and H3K56Ac co-immunoprecipitate at CPT-induced DSBs in agreement with a role for H3K56 acetylation not only in replication but also in DNA damage repair (Masumoto et al., 2005).

Figure I8. Histone PTMs and chromatin remodelers alter chromatin state. A. Histone modifiers are recruited onto chromatin and trigger the covalent modification of histone residues which ultimately provoke the completion of a DNA-related task (transcription activation in this case). B. Chromatin remodelers work sliding, evicting or exchanging whole nucleosomes.

3.2. Chromatin remodelers.

Chromatin remodelers can slide, evict or exchange nucleosomes among the genome by using the energy released from the ATP hydrolysis (Figure I8). There are four different families of chromatin remodelers: SWI/SNF, ISWI, CHD and INO80. All of them share an ATPase domain split in two parts: DExx and HELICc but they differ in the adjacent domains, which confer the ability to recognize distinct

48 substrates in order to accomplish their function. SWI/SNF family contains bromodomains which binds to acetylated lysines; ISWI is provided of HAND-SANT- SLIDE domains and recognize unmodified histone tails of nucleosomes; CHD family have chromodomains which bind to methylated lysines; and INO80 family contains a long insertion between the ATPase domain to which actin-related protein (ARP) binds and targets the chromatin remodeler family to histones.

Chromatin remodelers have been implicated in replication, transcription and DNA repair. Swr1 from INO80 family is involved in replacing histone H2A for H2A.Z at promoters, facilitating transcription by weakening DNA histone interactions (Zhang, Roberts and Cairns, 2005), and at zones of DSBs, where RSC complex from SWI/SNF family is recruited through acetylation of H3K14 (Wang et al., 2012) and facilitates the access to damaged DNA. In contrast to Swr1, Ino80 catalyzes the opposite reaction (Brahma et al., 2017), replacing H2A.Z for H2A. CHD1 is recruited to transcriptional active genes through the recognition of H3K4me3 in order to facilitate transcription elongation in human cells (Sims et al., 2007). In yeast, Isw2 is enriched at active replication regions and facilitates the progression of the replication forks (Vincent, Kwong and Tsukiyama, 2008). These are few examples of the chromatin remodeling functions, reflecting the high diverse and specialized roles that they have inside cells.

3.3. Chromatin and DSB repair.

Chromatin modifications play a key role in signaling DNA damage and recruiting DNA repair factors to safeguard the integrity of the genome. The first and best- studied example is the role of H2A phosphorylation. Upon DSBs, the checkpoint kinases Mec1 and Tel1 (ATR and ATM in higher eukaryotes, respectively), catalyze one of the most studied and fast reactions which appears after a DSB, which is the phosphorylation of H2A serine 129 in yeast (H2A-P) and H2AX serine 139 in mammals (Downs, Lowndes and Jackson, 2000; Rogakou et al., 1998; Burma et al., 2001). This phosphorylation mark spreads along the DNA molecule from the break site and constitutes a recognition scaffold for several factors involved in DNA damage repair. In mammalian cells this phosphorylation triggers

49 chromatin poly-ribosylation which helps to recruit NuRD complex causing transcription silencing (Chou et al., 2010; Smeenk et al., 2010) and the subsequent ubiquitination of H2A and H2A.Z by the ubiquitin E3 ligase RNF8 in order to recruit another factors such as BRCA1 (Sobhian et al., 2007).

Chromatin is also modified to allow the access of repair factors onto DNA. In this sense, phosphorylation of H2A.Z also leads to NuA4 and SWR1 recruitment, in both yeast and humans, what provokes an increase in the acetylation levels of histone H4 tail (Downs et al., 2004; Papamichos-Chronakis, Krebs and Peterson, 2006; van Attikum, Fritsch and Gasser, 2007). Moreover, in fission yeast it has been described that H3K14 acetylation increases after HO-induced DSB formation, which suggests that this mark is required for proper activation of DNA damage response by creating a more relaxed and accessible chromatin state (Wang et al., 2012). This could be mediated by the RSC complex, as this SWI/SNF remodeler has been reported to bind to acetylated H3K14 in budding yeast (Kasten et al., 2004). The chromatin remodeler INO80 could also have a role in opening the chromatin by evicting H2A.Z, γH2AX and H3 from DSB sites to favor homologous recombination repair, since INO80-depleted cells fail to complete 5' resection and RPA loading onto ssDNA (van Attikum, Fritsch and Gasser, 2007).

Thus, chromatin is widely modified upon DNA damage by either histone PTMs as well as nucleosome alterations. These chromatin changes act as scaffolds or create a more accessible environment for binding of repair factors to fix the damage.

3.4. Chromatin and R loops.

R loop formation can be modulated by the state of DNA, which in turn is determined by chromatin. So, it is not surprising that changes in chromatin can favor DNA-RNA hybridization. In this sense, defects in nucleosome reassembly during transcription due to the absence of FACT complex give rise to the accumulation of such secondary structures in both budding yeast and human cells (Herrera-Moyano et al., 2014). In addition, de-acetylation inhibition by chemical

50 compounds (TSA and SAGA) or lack of Sin3A in human cells and yeast leads to R loop accumulation (Salas-Armenteros et al., 2017; Wahba et al., 2011). Along the same line, deletion of FFT3, a SNF2-like chromatin remodeler provokes an increased turnover of nucleosomes associated to R loop formation in fission yeast (Taneja et al., 2017) and di-methylation of H3R17 and H4R3 indirectly prevent R loop formation by recruiting TOPIIIB to eliminate negative supercoiling in mammalian cells (Yang et al., 2014).

Several studies have reported a connection between R-loops and certain chromatin marks, such as histone H3 acetylation, mono- and di-methylation of H3K4 and tri-methylation of H3K36 (Sanz et al., 2016; Chen et al., 2015), but the cause-effect relationship between these marks and R loops is not so clear and, in fact, R loops can lead to changes in chromatin. In this regard, in S. pombe R loops modulate gene silencing at centromeres in a mechanism which involves H3K9me2 (Nakama et al., 2012), a histone mark which is reduced in THOC1-depleted human cells and C. elegans (Castellano-Pozo et al., 2013), as well as colocalizes with R loops in Friedreich ataxia and fragile X syndrome patients (Groh et al., 2014; Colak et al., 2014; Loomis et al., 2014).

Furthermore, R loops have also been associated to genome compaction. In particular, our lab observed in yeast, human cells and C. elegans that R loop accumulation in the absence of a component of the THO complex is linked to RNase H-sensitive H3S10 phosphorylation (Castellano-Pozo et al., 2013). In contrast, it is also possible that R loops create a more relaxed chromatin since it has been shown that they co-localize with genome regions more accessible to DNase I in vivo (Sanz et al., 2016).

However, despite all the data supporting the connection of R loops and chromatin, the mechanisms behind are still far from being understood. The discovery of yeast histone mutants that accumulate R loop but not genome instability or H3S10 phosphorylation (Garcia-Pichardo et al., 2017) prompted us to investigate this further.

OBJECTIVES

Objectives

The main goal of this thesis is to study the contribution of histones, chromatin remodelers and histone modifiers in the formation of DNA-RNA hybrids and its associated genome instability in the yeast Saccharomyces cerevisiae. For this purpose, we addressed the following specific objectives:

1. To establish new methodologies for the study of R loop length, frequency and distribution in the yeast Saccharomyces cerevisiae with the aim of determining the relevance of these features in genome instability.

2. To identify new factors involved in DNA-RNA hybrid formation among chromatin remodelers and histone modifiers.

RESULTS

Chapter I-Cytosine deamination as a tool to determine DNA-RNA hybrid length, frequency and distribution.

1. H3∆1-28, H3K9-23A and H4K31Q histone mutations that lead to DNA- RNA hybrids do not lead to genome instability unless AID is overexpressed.

Cytosine deamination by the human enzyme AID has been used as an in vivo tool to target the ssDNA displaced by DNA-RNA hybrids in R loops in both yeast and human cells (Gomez-Gonzalez and Aguilera, 2007; Dominguez-Sanchez et al., 2011). In a screening developed in our lab, several mutations in histones were described to accumulate DNA-RNA hybrids without leading to hyper-recombination unless AID was overexpressed (Garcia-Pichardo et al., 2017). This was in sharp contrast to all previously reported DNA-RNA hybrid accumulating mutants such as deletions in the genes of the THO complex, which lead to a strong DNA-RNA hybrid-dependent hyper-recombination (Huertas and Aguilera, 2003). These histone mutations were located in the N-terminal H3 and H4 histone tails and included a deletion of first 28 residues of histone H3 (H3∆1-28), an histone H3 mutant containing substitutions of the N-terminal tail lysines 9, 14, 18, and 23 to alanines (H3K9-23A) and a substitution of the histone H4 lysine 31 to glutamine (H4K31Q) (Garcia-Pichardo et al., 2017). These mutations were studied in yeast backgrounds (hereafter named as H3 or H4 backgrounds) that contain half the genic dosage of histones since one of the loci encoding for histone H3 and H4 genes (hht1-hhf1) was deleted, and the other one (hht2-hhf2) was replaced with the mutant copy of interest (Dai et al., 2008).

In order to confirm that such histone mutants were actually not leading to increased genome instability despite the accumulation of DNA-RNA hybrids, as they were originally selected for by the use of a recombination system (Garcia-Pichardo et al., 2017), we decided to investigate genetic instability in two different manners: plasmid loss and loss of heterozigocity (LOH). We included the hpr1∆ mutation of the THO complex as a positive control of DNA-RNA hybrid-dependent genetic instability (Figure R1).

Plasmid loss was studied by transforming the yeast strains with the pRS314 empty plasmid for testing its loss plus a combination of an empty plasmid and either

GAL::AID or GAL::RNH1 or both at the same time (Figure R1A). As shown in Figure R1, hpr1∆ caused a high level of plasmid loss as previously reported (Herrera-Moyano et al., 2014; Chavez and Aguilera, 1997). However, none of the histone mutants studied enhanced the levels of plasmid loss with respect to the WT cells in the non-AID overexpression condition. Indeed, the levels of plasmid loss were lower in H3∆1-28 and H3K9-23A mutants. As expected from the presence of DNA-RNA hybrids in these mutants, AID overexpression caused a 4.3- and 1.5-fold increase in H3∆1-28 and H3K9-23A mutants and only a 1.1-fold increase in WT cells. In addition, the overexpression of RNase H together with AID, decreased the percentage of plasmid loss seen in the mutants back to WT levels. These data suggest that the sensitivity to AID-induced damage is enhanced by these histone mutations but that they are not leading to genetic instability by themselves.

A B - AID - RNH + AID - RNH +AID + RNH 80 150 * **

** *** )

60 -7 s

s * 100

y (x10 y

d i

40 n

m as

qu *

l e

P 50 r ***

20 f ***

O L 0 0 T 8 A T Q T 8 A Q W -2 W r1∆ W -2 r1∆ 3 1 -23 4 31 3 1 -23 31 H 9 H K hp H 9 K hp K K H3∆ H4 BY4743 H3∆ H4 H3 H3 H3 H3

Figure R1. Histone mutants that accumulate DNA-RNA hybrids, do not present genome instability unless AID is overexpressed. A. Percentage of pRS314 plasmid loss. Average and SEM of at least 3 independent experiments. * p<0.05, ** p<0.01, *** p<0.001 (Unpaired Student´s t-test). B. Frequency of loss of heterozygosity (LOH) in the MAT locus at chromosome III. Average and SEM of 8-18 independent colonies. * p<0.05, ** p<0.01, *** p<0.001 (Unpaired Student´s t-test).

In order to study LOH, we constructed homozygous diploids strains of the different histone mutants and hpr1∆, again as a positive control, and we measured the LOH at the endogenous MAT locus on chromosome III. Diploid cells (heterozygous in MAT locus) cannot mate due to co-dominant suppression of haploid-specific cell differentiation pathways (Jensen, Sprague and Herskowitz, 1983; Goutte and Johnson, 1988; Li et al., 1995; Johnson et al., 1998). However, loss of either MATa or MATα alleles results in the ability to mate with either MATa or MATα, respectively. So, we crossed the diploid histone mutants with F4 (MATa) and F15 (MATα) tester strains. Mating products were detected by auxotrophy complementation (growth in SD media). The three histone mutants, H3∆1-28, H3K9-23A and H4K31Q, presented lower levels of LOH, as detected by less number of mating products when combined with F4 or F15 strains (0.3-fold in H3∆1-28, 0.2-fold in H3K9-23A and 0.6-fold in H4K31Q with respect to WT) (Figure R1B). We also used the diploid hpr1∆mutant as a positive control given that the loss of Hpr1 is known to increase LOH (Garcia-Rubio et al., 2008), as our data reproduced, with a 4.2 fold increase over the WT levels. This LOH can be due to gene conversion or chromosome loss. To determine the percentage of gene conversion and chromosome loss, we calculated the ratio of the DNA at the PRD1 gene (located at the same chromosome III as the MAT locus) versus the DNA at the GCN4 gene (located at chromosome V) by qPCR with specific primers (see Materials and Methods). A ratio below one indicated that chromosome III was lost. A third PCR reaction at the TRP1 gene (located at chromosome IV), which is in heterozygosis in these diploids (TRP1/trp1∆63), was used as a reference of a gene with only one copy. We observed that 25% (15 of 59 independent clones) of the LOH events at the MAT locus were due to chromosome loss.

Altogether, these data corroborate that H3∆1-28, H3K9-23A and H4K31Q histone mutants do not lead to genome instability by themselves, unless AID is overexpressed (Garcia-Pichardo et al., 2017).

2. H3K9-23A mutation does not impair replication fork progression and suppresses the replication defects of hpr1∆.

The DNA-RNA hybrid accumulation caused by THO null mutants leads to replication fork impairments that make the S-phase checkpoint essential for survival, as it was reported by a synthetic interaction of THO null mutants with the deletion of the RAD24 gene that was exacerbated upon replicative stress (Wellinger, Prado and Aguilera, 2006; Gomez-Gonzalez, Felipe-Abrio and Aguilera, 2009). We thus wondered whether the DNA-RNA hybrid accumulation of the histone mutants would also lead to replication fork impairments and we focused on the H3K9-23A mutant for this purpose. So, we studied the sensitivity to replicative stress of the H3 hpr1∆ mutation alone or in combination with one or two of the following mutations: H3K9-23A and rad24∆ (Figure R2).

YPAD 50 mM HU W303 hpr1Δ rad24Δ hpr1Δ rad24Δ H3WT H3 hpr1Δ H3 rad24Δ H3 hpr1Δ rad24Δ H3K9-23A H3K9-23A hpr1Δ H3K9-23A rad24Δ H3K9-23A hpr1Δ rad24Δ

Figure R2. Study of the genotoxic sensitivity of the indicated strains after growth in medium containing hydroxyurea. Representative images after 3 days of growth are shown.

We observed a slight growth defect of H3K9-23A mutant in the presence of replication stress and this defect was exacerbated in the double mutant H3K9-23A rad24∆ and in the triple mutant H3K9-23A hpr1∆ rad24∆. However, the combination of hpr1∆ and rad24∆ mutations in an H3 background did not lead to the synthetic growth defect described in W303 background strains (Gomez- Gonzalez, Felipe-Abrio and Aguilera, 2009), which were assayed here as positive controls (Figure R2). We speculate that this might be due to the fact that this background contains half the genic dosage of histones what could ameliorate the consequences of the replication defect on the checkpoint.

We thus decided to study replication fork directly by 2D-gel analysis of replication forks. We studied replication intermediates by performing 2D gels at the SPF1 gene, which is close to the ARS508 located at chromosome V and at which replication fork progression has been reported to be impaired in THO null mutants, particularly upon additional replicative stress (Figure R3A) (Gomez-Gonzalez, Felipe-Abrio and Aguilera, 2009). Cells were synchronized in G1 with α-factor during 3 hours and released in fresh YPD medium with 40 mM of HU. DNA from H3WT, H3K9-23A, H3 hpr1∆ and H3K9-23A hpr1∆ was isolated 30 minutes after G1 release to collect cells in S phase.

We analyzed replication intermediates in the 3.9-kb PvuII fragments harboring ARS508 and part of the SPF1 gene (Figure R3B). As shown in Figure R3C, the H3K9-23A mutant did not accumulate replication intermediates compared to WT strain, contrary to what happens with hpr1∆ mutant, which accumulated long Y- shaped molecules as previously described (Gomez-Gonzalez, Felipe-Abrio and Aguilera, 2009). Furthermore, we were not able to see that phenotype in the H3K9- 23A hpr1∆ double mutant, even though it accumulates DNA-RNA hybrids as measured by DRIP at that same region where replication progression was analyzed (Figure R3D). These results indicate that, despite leading to DNA-RNA hybrids, H3K9-23A does not lead to detectable replication impairments and suppressed the fork progression impairment of hpr1∆ cells. This is in agreement

65 with parallel studies from the laboratory that revealed a suppression of the elevated levels of recombination of hpr1∆ cells (Garcia-Pichardo et al., 2017).

A B PvuII

SPF1 ARSH508

PvuII (3.9 kb) C D 40 mM HU H3WT H3 hpr1Δ

SPF1 (3.6Kb) 5

H3K9-23A H3K9-23A hpr1Δ 2

1 S9.6 IP (Relative values) (Relative IP S9.6 0 RNH - + - + - + - + H3WT H3 hpr1∆ H3K9-23A H3K9-23A hpr1∆

Figure R3. H3K9-23A histone mutant suppresses the replication defect of hpr1∆ mutant. A. A diagram of the PvuII restriction sites on the chromosome V region surrounding ARS508 and the SPF1 gene. B. Scheme of the possible replication intermediates. C. 2D gel migration pattern of replication intermediates of the 3.9-kb PvuII DNA fragments from H3WT, H3 hpr1∆, H3K9-23A and H3K9-23A hpr1∆ strains. White head arrow indicates the accumulation of replication intermediates. D. DNA-RNA immunoprecipitation (DRIP) at the SPF1 gene in the indicated strains with or without prior RNase H (RNH) in vitro treatment.

3. R loop formation in histone mutants is not related to an altered pattern of nucleosome positioning.

So far, we have confirmed that DNA-RNA hybrids caused by histone mutations are not affecting replication fork progression nor triggering the subsequent genetic instability phenotypes usually associated to DNA-RNA hybrids and reported for other mutants such as those of the THO complex. Given that all of these mutations affect histones, we wondered whether the reason for DNA-RNA hybrid accumulation in these mutants could be related to changes in nucleosome occupancy and/or positioning. In collaboration with Francisco Antequera´s lab (IBFG-CSIC, Salamanca), we studied the profiles of nucleosome occupancy and positioning in H3∆1-28, H3K9-23A and H4K31Q mutants as compared to the H3WT and H4WT by MNase-seq experiments. We also used H3K9-23R and H4K31A mutations in the same residues that do not lead to R loop accumulation (Garcia-Pichardo et al., 2017). Differences in nucleosome positioning around or below 2% were not considered since this was the level observed between two biological replicates.

A B C D vs BY4741 vs H3WT vs H4WT vs W303 20 20 20 20

15 15 15 15

10 10 10 10 Nucleosome

5 5 5 5 possitioning loss (%) loss possitioning

0 0 0 0 T T 8 A R A 1∆ 2 3 Q 1 1∆ 2∆ W W r - 2 3 1 r h 3 4 p 1 - -2 3 3 p n H H h ∆ 9 9 K K h r 3 K K ∆ H H4 H4 1 H3 H3 rnh

Figure R4. Nucleosome positioning is altered in H3∆1-28 histone mutant. Percentage of nucleosome positioning loss in the indicated strains with respect to: A. BY4741; B. H3WT; C. H4WT; D. W303. Performed in collaboration with F. Antequera’s lab.

It is noteworthy that we observed that the low gene dosage of histone genes had an impact on nucleosome positioning, since there was an 18% of nucleosomes differently positioned in H3WT and H4WT with respect to a canonical BY background (Figure R4A). Moreover, 60% of those changes were observed in both H3WT and H4WT. By studying the average nucleosome positioning at every gene of the genome, we determined that such 18% of nucleosomes that changed their positioning affected to 40% and 44% of all the genes in H3WT and H4WT (2126 and 2296 out of the 5242 genes analyzed, respectively) (Figure R5).

Figure R5. Around 40% of genes in H3WT and H4WT exhibited loss of nucleosome positioning compared to BY4741. A. Average nucleosome occupancy in 2126 of 5242 genes with differential nucleosome positioning from H3WT. B. Average nucleosome occupancy in 2296 of 5242 genes with differential nucleosome positioning from H4WT.

In addition to that, the analysis of the histone mutants revealed that only H3∆1-28 mutation confers significant changes in the nucleosome positioning: 6.5% of nucleosomes lost their positioning as compared with the H3WT strain (Figure R4B). Globally, these differences are very subtle and localize at specific regions of the genome, such as RGI1 locus (Figure R6A). However, we could not find significant changes in the pattern of nucleosome positioning at the GCN4 locus where H3∆1-28 has been reported to accumulate DNA-RNA hybrids (Garcia- Pichardo et al., 2017) (Figure R6B). Moreover, no significant nucleosome positioning defects were observed either in the H3K9-23 mutations (H3K9-23A and H3K9-23R) or in the H4K31 mutations (H4K31Q and H4K31A) (Figure R4B and C). We also studied the nucleosome positioning in another R loop accumulating mutants such as hpr1∆ and rnh1∆ rnh2∆ mutants and compared them to the corresponding WT strains (W303 or BY4741 backgrounds). None of these mutants presented a significant number of misallocated nucleosomes (Figure R4A and D). The lack of any detectable correlation between changes in nucleosome positioning and R loop formation argues against changes in the pattern of nucleosome positioning being neither the cause nor the consequence of DNA-RNA hybrids in any of the mutants.

Figure R6. Examples of nucleosome occupancy. A. Nucleosome occupancy at RGI1 gene. B. Nucleosome occupancy at GCN4 gene. Arrows indicate significant changes in nucleosome positioning.

4. R loops are similar in size regardless of whether they induce genome instability or not.

We next hypothesized that DNA-RNA hybrids leading to genetic instability or not could differ in the length. To test this idea, we performed a technique based on sodium bisulfite that has been originally described to detect and measure DNA- RNA hybrids at the immunoglobulin loci of murine B cells (Yu et al., 2003). Sodium bisulfite converts cytosines (C) into uracils (U) specifically when they are located in ssDNA, as the ones present in R loops. The bisulfite modification assay is based on subsequent steps of DNA amplification (where the polymerase will read uracil (U) as thymine (T), resulting in a C to T conversion) and sequencing of the resulting PCR products to determine the regions of the DNA that have been deaminated. The use of converted primers that anticipate the C to T conversion provoked by bisulfite together with native primers that would bind to a WT sequence of DNA allows the enrichment of deaminated molecules when the frequency of ssDNA molecules is too low (Yu et al., 2003).

DNA is present in cell in many conformations, being the most common the DNA duplex. In this case, bisulfite does not have any effect over the cytosines in the double helix, since they are protected by the complementary strand of DNA (Figure R7A). The DNA can also be melted, generating a region of ssDNA. After bisulfite treatment, all the cytosines of both DNA strands will be converted into uracil so that anytime a converted primer is used we will get PCR amplification (Figure R7B). Last, in the case of an R-loop, the mRNA hybridizes with its DNA template displacing the complementary strand as single strand DNA (Figure R7C). In this case, bisulfite will convert only the cytosines in the ssDNA of the R loop, due to the fact that the mRNA protects the other strand in the form of a DNA-RNA hybrid. Thanks to the use of converted primers, which are strand specific, we expect to get amplification of the non-template strand, while no product is expected if a converted primer for the template strand is used.

Figure R7. Possible conformations of DNA affect its sensitivity to sodium bisulfite generating different products that can be detected by PCR. A. dsDNA. B. ssDNA. C. R loop. Black lines represent DNA; red lines, RNA. Primers are represented as arrows: forward primers are pointing to the right and reverse primers are pointing to the left. Native primers (N) contain the sequence complementary to the original sequence whereas converted primers (C) anticipate the C to T conversion. Blue primers detect conversions on the non-template strand and orange primers detect conversions on the template strand.

We chose the GCN4 gene to measure the length of DNA-RNA hybrids, since it is one of the hybrid-prone DNA regions at which both hpr1∆ and the histone mutants have been shown to accumulate DNA-RNA hybrids (Castellano-Pozo et al., 2013; Garcia-Pichardo et al., 2017). We first amplified this locus using two native primers (see Material and Methods), but after sequencing the resulting PCR product, we observed no C to T conversions, suggesting that R loops must be formed at a very low frequency. Thus, we designed specific converted primers to enrich for DNA-

RNA hybrid containing molecules. As we did not know the length of the R loops or their position along the gene, we decided to screen the GCN4 gene along different regions using a native primer (GCN4 N F, see Material and Methods) upstream the ORF and converted primers that encompassed the ORF (regions 1 to 4) and 200 bp beyond the STOP codon (regions 5 and 6) (GCN4 N or C 1-6 R, see Material and Methods) (Figure R8A).

A 0.8 Template (T) GCN4 1 2 3 4 5 6 Non Template (NT)

Kb 0.3 0.2 0.2 0.2 0.12 0.1 B C 4 - RNH + RNH forward primer: N N N N reverse primer: N C N C location: 1 2 3 4 5 6 H3WT forward primer: N N N N N N N N N C N N N N reverse primer: N C N C N C N C N N N C N C H3∆1-28 H3WT H3K9-23A H3∆1-28 H4K31Q hpr1∆

Figure R8. Detection of R loops by PCR in different regions of the GCN4 gene. A. Scheme of the GCN4 gene regions analysed by PCR. Arrows represent primers, which can be native (N, they align to wild type sequence) or converted (C, they align to bisulfite- mutated sequences). Primers for detecting R loops in either the template (T) or the non- template (NT) strand are plotted in orange and blue, respectively. B. PCRs products from H3WT and H3∆1-28 strains at the indicated regions. C. PCR products from H3WT, H3∆1- 28, H3K9-23A, H4K31Q and hpr1∆ strains at the region 4 with or without prior RNase H (RNH) in vitro treatment.

In a first approach, we performed PCRs in H3WT and H3∆1-28 strains after bisulfite treatment using a native forward primer upstream the ORF and either a native or a converted reverse primer on the non-template strand (blue primers in Figure R8A). We only detected a PCR product using converted primers in the region 4, suggesting that DNA-RNA hybrids accumulate at the 3' end of the GCN4 gene and not beyond it in both H3WT and H3∆1-28 strains (Figure R8B). To 72 distinguish between ssDNA and R loop conformations, we also did a specific PCR for detecting C to T conversions in the template strand using a converted forward primer at region 4 (orange primers in Figure R8B). In this case, we were unable to detect any PCR product, indicating that the region is not opened as ssDNA and in agreement with the existence of an R loop in this region. Next, we extended the analysis to H3K9-23A, H4K31Q and hpr1∆. As shown in Figure R8C, we obtained an amplicon of the expected size for all mutants in the study after performing the PCR for the region 4. However, RNase H in vitro treatment was insufficient to impede this PCR reaction.

We then extracted the PCR product obtained and cloned it into a pGEM-Teasy plasmid. In order to separate the distinct molecules, we transformed E. coli and sequenced the DNA from several clones. Finally, we aligned them to the original GCN4 sequence (see Figure R9A for an example). We observed that most of the sequenced molecules presented long stretches of C to T conversions at the 3' end of the GCN4 gene. However, not all the molecules analyzed had C to T conversions and that such percentage varied slightly between the strains (Figure R9B). Interestingly, although RNase H in vitro treatment prior to bisulfite did not prevent the C to T conversions (Figure R8C), it significantly reduced the number of mutated molecules detected from 82 to 63 % (Chi-Square test with Yates' correction, p<0.05), in agreement with these molecules reflecting real R loops (Figure R9B). We then compared the length of DNA with tracts of consecutive C to T conversions in all the strains (Figure R9C and D). The average values ranged from 125 to 172 bp but we observed no significant differences between H3WT, hpr1∆, H3∆1-28, H3K9-23A, and H4K31Q strains. These results indicate that the length of R loops is similar regardless of whether they produce genome instability or not.

GCN4 H3WT H3∆1-28 H3K9-23A H3K31Q hpr1∆ B

Molecules Total % Molecules with C to T sequenced with C to T changes molecules changes

H3WT 20 24 83

H3∆1-28 53 64 82

H3∆1-28 + RNH 27 43 63

H3K9-23A 18 25 72

H4K31Q 20 28 71

hpr1∆ 22 29 76

C D

600 164 172 151 153 125 153 GCN4 1 2 3 4

H3WT h

t 400

ng e

H3∆1-28 L

H3K9-23A Loop 200

H4K31Q R hpr1∆ 0

H3WT + RNH hpr1∆ H3∆1-28 H3K31Q H3K9-23A

H3∆1-28

Figure R9. R loop length in WT, R loop accumulating histone mutants and hpr1∆. A. Example of sequence alignments showing the C to T conversion obtained after treatment of DNA with bisulfite and amplification of the GCN4 gene by PCR in samples coming from H3WT, H3∆1-28, H3K9-23A, H4K31Q and hpr1∆ strains. B. Number of molecules with C to T changes and its percentage respect to the total number of molecules sequenced from H3WT, H3∆1-28 (with or without RNase H in vitro treatment prior to bisulfite), H3K9-23A, H4K31Q and hpr1∆ strains. C. R loop distribution of all the molecules with C to T changes analysed. Each line represents the length of the conversion track of a particular sequenced molecule. D. R loop length in the indicated strains. Average (red lines) and SEM (blue lines) are plotted.

5. Similar R loop length in several R loop-accumulating mutants.

We next extended the study to other mutants previously described to accumulate R loops. We chose the hpr1∆ and mft1∆ mutants from the THO complex; thp1∆ mutant from the TREX-2 complex; spt16-11 and pob3-7 mutants from the FACT complex; top1∆ mutant of the topoisomerase 1; sen1-1 senataxin mutant and the rnh1∆ rnh2∆ double mutant. We also included the hpr1-101 mutant from the THO complex, which is reported not to accumulate R loops (Huertas and Aguilera, 2003; Gomez-Gonzalez and Aguilera, 2009) (Figure R10).

The DNA from the indicated strains was extracted and treated with sodium bisulfite prior to PCR amplification using the same primers for the region 4 of the GCN4 gene as before. Then, PCR products were purified and cloned into the pGEM- Teasy plasmid to be sequenced. Analysis of the total molecules sequenced revealed that the percentage of molecules with C to T conversions was significantly increased only in sen1-1 mutant with 84% (Chi Square test with Yates' correction, p<0.05) (Figure R10A). The average length of the ssDNA ranged from 72 bp for the top1∆ mutant to 171 bp for rnh1∆ rnh2∆ mutant, compared to 118 bp observed for WT strain (Figure R10B). However, these differences were not significant given that there was a similar variability in all populations, indicating that the R loop length at the GCN4 gene is similar in all the strains analyzed.

Molecules Total % Molecules with C to T sequenced with C to T changes molecules changes

WT 5 14 36 hpr1∆ 10 16 62 hpr1-101 14 19 74 mft1∆ 4 12 33 thp1∆ 10 13 77 spt16-11 5 11 45 pob3-7 4 11 36 top1∆ 24 37 65 sen1-1 16 19 84 rnh1∆ rnh2∆ 9 13 69 B GCN4 400 166 171 160 166

136 h h (bp) 300 t 118

g 148

n 115 e L 200

p 116

o L

- 72

R 100

0 ∆ ∆ ∆ 11 ∆ WT -101 mft1 thp1 top1 hpr1 pob3-7 sen1-1 hpr1 spt16- ∆ rnh2∆ rnh1

THO TREX-2 FACT

Figure R10. R loop length is similar in several R loop-accumulating mutants. A. R loop length (bp) in the indicated strains. Each single circle represents a DNA molecule with C to T conversion. Average (red lines) and SEM (blue lines) are plotted. B. Number of molecules with C to T changes and its percentage respect to total amount of molecules sequenced from the indicated strains.

6. R loops are formed at a very low frequency.

Since the size of R loops cannot explain the different phenotypes observed between histone and THO mutants (leading to genetic instability or not), we wondered whether R loops were formed at a different frequency. For this purpose, we also took advantage of bisulfite to infer frequencies by sequencing amplicons from PCRs with two native primers and detecting the converted molecules by lift colony hybridization assay, as previously reported (Huang et al., 2007; Roy and Lieber, 2009). Thus, after treating the DNA in vitro with sodium bisulfite, we made PCRs using the two native primers (GCN4 N F and GCN4 N6 R, see Material and Methods) surrounding the GCN4 gene. We then cloned the resulting PCR product into the pGEM-Teasy vector and used this construct to transform E. coli cells in order to separate the different molecules present in each sample. Transformants were transferred into a membrane in which the plasmid of each transformant was extracted. This membrane was then hybridized using radioactive-labeled converted primers for the region 4 of the GCN4 gene (referred as C4 in Figure R11). After reading the radioactive signal, they were stripped and hybridized with radioactive- labeled converted primers for the regions 2 and 3 (data not shown, GCN4 C3 R and GCN4 C2 R in Material and Methods). Last, one more stripping and hybridization cycle was performed with a radioactive-labeled native primer to quantify the total number of GCN4 molecules analyzed (referred as N1 in Figure R11). As a positive control of the experiment, we used a plasmid containing the GCN4 gene in which all the cytosines present in the last 500 bp were mutated to thymine (pGTN4 plasmid).

As shown in Figure R11, we checked around 530 clones coming from H3WT and H3∆1-28 DNA and 420 from hpr1∆ DNA, but we were not able to detect signal in any of them, indicating that the frequency of R loops must be below 0.2% at this particular region. These results are in agreement with the fact that we were obliged to use converted primers to be able to detect the C to T conversions (Figure R8 and Figure R9).

A H3WT ∆1-28 hpr1∆

B GCN4 molecules

Sample Analyzed with C to T conversion

H3WT 531 0

∆1-28 530 0

hpr1∆ 422 0

Figure R11. R loops are formed at a very low frequency. A. Representative images of lift colony hybridization assay of GCN4 molecules amplified by PCR (using 2 native primers binding out of the ORF) from H3WT, H3∆1-28 and hpr1∆ strains and pre-treated with sodium bisulfite. Red circles are a positive control of C to T conversion (pGTN4 plasmid). C4 indicates the radioactive-labeled converted primer for the region 4 of the GCN4 gene, and N1, the radioactive-labeled native primer for the region 1. B. Quantification of the GCN4 molecules analyzed in the indicated strains.

7. Bisulfite-qPCR as an alternative method to DRIP

Given the success of the sodium bisulfite technique to detect R-loops and measure their length, we wondered whether we could combine it with the real time PCR in order to quantify DNA-RNA hybrids. So, the bisulfite-treated DNA was used as a template to be amplified by qPCR in two independent reactions that differed only in the pair of primers used. The first reaction (native reaction) was performed with two native primers (forward and reverse) and was used to quantify the total amount of non-mutated DNA per sample. The second reaction (converted reaction) was performed with a native forward primer and a reverse converted primer (GCN4 N F 359 and GCN4 C3 R in Material and Methods) and was used to quantify the total amount of DNA that had been mutated by the bisulfite action. In addition, to determine the absolute amount of DNA in each reaction, we used serial dilutions of known amounts of two different plasmids that contained either the WT (pGCN4) or converted (pGTN4) GCN4 gene for the native and converted reaction respectively (Figure R12A).

The total amount of DNA with C to T conversions was calculated by dividing the amount of DNA obtained with the converted reaction by the total amount of DNA, which corresponds to the sum of DNA with and without C to T conversions (native plus converted reactions) (Figure R12B).

Figure R12. Scheme of absolute R loop quantification method. A. A draw of the plasmids used is plotted on the top. Red part of pGTN4 plasmid represents the C to T mutations contained among 500 bp at the 3’ end of the GCN4 gene. Black lines represent the standard curve of qPCRs whose slope is used to determine the amount of total DNA (native reaction) or converted DNA (converted reaction) present in problem samples represented as orange dots. B. Formula for calculating absolute hybrid frequency.

Since R loops accumulated at the 3' end of the GCN4 gene (region 4), we first tried to detect R loops by using the native forward primer and the converted reverse primer of the region 4 (GCN4 N F and GCN4 C4 R primers in Material and Methods), as previously done for measuring R loop length. We performed the reactions in WT and rnh1∆ rnh201∆ strains but the length of the product (1 Kb) was too large as to give specific amplifications, since they are optimized for 100-200 bp in qPCR protocols. Thus, we designed a new pair of primers, taking into account the R loop length distribution of hybrids obtained in the previous experiment

(Figure R9C). So, we designed the forward primer next to the region where C to T conversions begins and used a reverse (native or converted) primer at the region 3 of the GCN4 gene (GCN4 N F 359 and GCN4 N3 R in Material and Methods), thus leading to an amplicon of 181 bp that was suitable to obtain specific products by qPCR. In parallel, we performed DRIP experiments in the same strains and analyzed the frequency of DNA-RNA hybrids by qPCR (Figure R13).

The analysis of qPCRs revealed that 4% of the molecules analyzed presented C to T conversions in the WT strain, and this value increased up to 7 % in the rnh1∆ rnh2∆ double mutant. Given our previous estimation that DNA-RNA hybrids are formed at a frequency below 0.2 % at the GCN4 gene (Figure R11), we believe that this value is overestimating the real absolute frequency of converted molecules. Moreover, the DRIP experiment performed in parallel showed DNA- RNA hybrid immunoprecipitation in 0.15% of the input arguing that the bisulfite- qPCR technique is not accurately quantifying the absolute levels of DNA-RNA hybrids. However, relative quantification of DNA-RNA hybrids was possible and revealed an increase in DNA-RNA hybrids in rnh1∆ rnh201∆ respect to WT cells (Figure R13). Importantly, in vitro treatment with RNase H prior to bisulfite was able to reduce the signal, indicating that the molecules detected by this bisulfite- qPCR method were actually DNA-RNA hybrids. Thus, the bisulfite-qPCR method could be a valid alternative to the DRIP method to detect variations in DNA-RNA hybrid levels at specific genomic locations without relying on the use of the S9.6 antibody.

GCN4 (0.8Kb) GCN4 (0.8Kb) 6

6 T

4 4

* 2 2

S9.6 (RelativeS9.6 values) Moleculesto Cwith

0 0 conversions(relative values) RNH - + - + RNH - + - + rnh1∆ WT rnh1∆ WT rnh201∆ rnh201∆

Figure R13. Comparison of bisulfite-qPCR assay with DRIP. Relative values of DNA- RNA hybrids detected by bisulfite-qPCR (left) and DRIP (right) at the GCN4 gene in WT, and rnh1∆ rnh201∆ strains. A scheme of the region analyzed is shown above each graph. Average and SEM of at least 3 independent experiments are shown. * p<0.05. (Paired Student´s t-test).

8. Exploring the use of AID* deaminase in vivo to detect and map R loops.

A method based on the bisulfite treatment (bisDRIP-seq) has also been established for the genome-wide mapping of R-loops (Dumelie and Jaffrey, 2017). However, this method still relies on the use of the S9.6 antibody to enrich for R- loops after the bisulfite treatment. We thus decided to explore the use of AID in vivo, as an S9.6-independent alternative tool for R-loop mapping. AID has been demonstrated to be an efficient tool for indirectly detecting R loops in vivo (Gomez- Gonzalez and Aguilera, 2007). However, the low mutation rate and its low processivity made it a disfavored option for the determination of DNA-RNA hybrid length as compared to bisulfite. Since then, an improved mutant version of the AID enzyme was described by the lab of Neuberger (Wang et al., 2009). This version of AID (hereafter referred as super AID or AID*) exhibits an enhanced mutation

82 activity and we wondered whether we could use it to map R loops genome-wide in vivo.

To avoid losing mutations in essential genes, we worked with either WT or hpr1∆ diploid cells. We transformed the diploid strains with a plasmid containing GAL::AID* plus an empty or GAL::RNH1 containing plasmid and we checked their survival upon AID overexpression by plating transformants in glucose or galactose containing medium and incubating them at 30ºC during 3 days (Figure R14A). In agreement with the reported transcription defects of hpr1∆ cells (Chavez et al., 2000), a growth defect was observed in galactose, but this defect was DNA-RNA hybrid-independent as it was not suppressed by RNase H overexpression (Figure R14A). We next tested the effect of AID* overexpression on cell viability after limited times of incubation in galactose. For this purpose, cells were grown in glucose or galactose containing medium during 3, 6, 9, 24 or 48 hours and then plated into glucose containing medium. As shown in Figure R14B, the overexpression of AID* alone or in combination with RNase H had no effect on the cell viability in either WT or hpr1∆ cells in agreement with the fact that chronic AID* overexpression was not lethal in WT or hpr1∆ cells (Figure R14A). We also confirmed the mutator effect of AID* by plating a fraction of the cells in canavanine- containing plates since mutations in the CAN1 gene would make cells resistant to canavanine.

As shown in Figure R14C, canavanine resistant colonies appeared 3 hours after the induction of AID* overexpression and the frequency of resistant colonies was in the order of 10-7 in both WT and hpr1∆ cells. Importantly, we detected no canavanine-resistant colonies after growth in glucose, in agreement with the mutations being induced by the overexpression of AID*. We also observed that the overexpression of RNase H was unable to decrease the frequency of mutations in the CAN1 gene in neither WT nor hpr1∆ cells, suggesting that there might be no R loops at this particular locus. We concluded to use AID overexpression during 24 hours to map R-loops genome-wide as longer incubations did not significantly enhance the mutation frequency.

Figure R14. AID* overexpression does not affect viability but induces mutations in CAN1 gene. A. Drop assay to test the effect of AID* overexpression in WT and hpr1∆ on plates. B. Viability of WT and hpr1∆ diploid strains with and without AID* and RNase H (RNH) at different time points grown in liquid culture. C. Mutation frequency at CAN1 gene in the indicated strains after the overexpression of AID* with or without RNase H (RNH) at different time points. Average and SEM of at least 3 independent experiments are shown.

Figure R15. Cell viability is not decreased after AID* overexpression. Viability of WT, hpr1∆, H3WT, H3∆1-28, H3K9-23A and H4K31Q diploid strains with and without AID* and RNase H (RNH) at 24 hours grown in liquid culture. Average and SEM of at least 3 independent experiments are shown.

We next transformed H3WT, H3∆1-28, H3K9-23A and H4K31Q diploid strains with GAL::AID* containing plasmid plus an empty or GAL::RNH containing plasmid, in order to study the effect of AID* overexpression in such strains. Again, cell viability was not affected by either AID* or RNase H overexpression (Figure R15). We confirmed AID induced effect by measuring the mutation frequency at the CAN1 locus, which was again was in the order of 10-7 after 24 hours of incubation (Figure R16) and extracted the DNA for further genome-wide sequencing. .

10 )

7 *

0 x1

( 1

n e

0.1

f on

i 0.01

t u

M 0.001 RNH - + - + - + - + - + - + WT hpr1∆ H3WT H3∆1-28 H3K9-23A H4K31Q

Figure R16. AID*-induced mutations in CAN1 gene. Mutation frequency in H3WT, H3∆1-28, H3K9-23A, H4K31Q and hpr1∆ strains at CAN1 gene. Average and SEM of 2-4 independent colonies are plotted. * p<0.05 (Unpaired Student´s t-test).

9. AID* overexpression induces mutations genome-wide preferentially in large genes and independently of their GC content and expression levels.

In order to see whether the AID* could be used for mapping R loops genome-wide, we sequenced the whole genome of two to three canavanine resistant colonies from independent experiments with each of the diploid strains and collaborated with Juan Antonio Cordero Varela (IBIS-Sevilla) to perform the alignment with the reference BY4741 genome and search for mutations by SNP calling. Using quite stringent criteria to define the mutations (see Materials and Methods), we detected a total of around a hundred point mutations in each transformant and condition.

As shown in Figure R17, the overexpression of the AID* induced a slight increase in the number of point mutations in all the mutants analyzed when compared to their corresponding WT strains, although RNase H overexpression was only reducing the number of point mutations in hpr1∆. Importantly, different number of mutations was observed between strains when counting cytosines and guanines, as expected from the action of AID* on either DNA strand. By contrast, the number of mutations in adenine and thymine was smaller and more similar between the different strains and conditions.

T A G C

200

s on

i 150

u m

t 100

l a

t 50 To

0 RNH - + - + - + - + - + - + WT hpr1∆ H3WT H3∆1-28 H3K9-23A H4K31Q

Figure R17. AID* overexpression induces mutations genome-wide. Total point mutations for each nucleotide in the indicated strains and conditions. Average and SEM of 1 to 3 independent experiments are plotted in all panels.

Since AID induces C to T transitions (seen as G to A changes in the other strand), we focused specifically in the C to T and G to A mutations (Figure R18). Again, all the mutants displayed an increase in the number of transitions respect to their corresponding WT strains, although this increase was only statistically significant and reduced by RNase H overexpression in the case of hpr1∆.

G to A C to T 100 * 80

C to T & G to A to G & T to C 20 Numbertransitionsof 0 RNH - + - + - + - + - + - + WT hpr1∆ H3WT H3∆1-28 H3K9-23A H4K31Q

Figure R18. Number of transitions in WT, hpr1∆, H3WT, H3∆1-28, H3K9-23A and H3K31Q strains. Total transitions (cytosine to thymine and guanine to adenine changes). Average and SEM of 1 to 3 independent experiments are plotted in all panels. * p<0.05 (Paired Student´s t test).

Next, we analyzed qualitatively these mutations by analyzing the length, GC content and expression levels of the ORFs that were mutated with respect to the average levels of the genome (Figure R19). We observed that mutated ORFs were significantly longer than the average of all the ORFs in the genome in all the strains analyzed with or without RNase H overexpression (Figure R19A). This could be a consequence of the higher probability of being mutated by AID* given their bigger size. However, there was no correlation between the length of ORFs and the number of mutations (Spearman´s test, ρ = 0.15, p>0.05), indicating that AID* does not mutate these genes just because they are larger and hence AID* has more substrate to mutate. By contrast, we detected no big differences in the GC content or the expression levels of the mutated ORFs versus the average of all of the ORFs in the genome (Figure R19B).

Since DNA-RNA hybrids can be favored by the formation of G4-quadruplexes (Duquette et al., 2004; De Magis et al., 2019; Kuznetsov et al., 2018; Wanrooij et al., 2012), we also compared the mutations induced by the AID* with a list of regions of the genome prone to form G4-quadruplexes (Capra et al. (2010). Interestingly, we detected more mutations at predicted G4-quadruplexes forming regions of the genome than those expected by chance (Chi-square, with p<0.05 in all conditions), what could be in agreement with their preference for DNA-RNA hybrid regions.

Next we wondered whether AID*-induced mutations were enriched in DNA-RNA hybrid prone regions and we took advantage of the DRIP-seq data of WT and hpr1∆ strains that has been previously reported (Chan et al. (2014). We observed that for hpr1∆ strains, the number of mutations detected in DNA-RNA hybrid prone ORFs and telomere regions was significantly higher than expected (Chi-square, p<4*10-7), but mutations were not enriched at DNA-RNA hybrid prone formation rDNA locus, which is in agreement with Hpr1 protein being involved in RNA pol II- transcribed genes. WT strain also presented a significant amount of AID*-induced mutations at telomeres and rDNA locus, as expected from the AID* action over the ssDNA of R loops accumulated at these particular regions. However, the

88 overexpression of RNase H only reduced significantly the number of mutations found at telomeres in hpr1∆, whereas it did not have any effect in WT strain.

Finally, since the mRNA forming the hybrid would protect the DNA template strand from the AID* action whereas the non-template strand would be single stranded and thus could be used as substrate for AID*, we hypothesized that AID* action would be enhcanced in the non-template strand. So, we analyzed the number of mutations in cytosines and guanines affecting to template and non-template strands of ORFs. However, the analysis revealed that both strands were mutated indistinctively at the same frequency.

3000 * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *

) 2000

(

t ng

e 1000 L

0 RNH - + - + - + - + - + - + WT hpr1∆ H3WT H3∆1-28 H3K9-23A H4K31Q B

50 * * * ) * * *

* * * * * %

( 40

n e

t 30

on c

C 20 G

0 RNH - + - + - + - + - + - + WT hpr1∆ H3WT H3∆1-28 H3K9-23A H4K31Q C 1.5

1.0

(Exp) 2

Log 0.5

0.0 RNH - + - + - + - + - + - + WT hpr1∆ H3WT H3∆1-28 H3K9-23A H4K31Q

Figure R19. Meta-analysis of AID*-induced mutations at ORFs. A. Average length of mutated ORFs versus the rest of non-mutated ORFs in WT, hpr1∆, H3WT, H3∆1-28, H3K9-23A and H4K31Q strains with and without RNase H (RNH) overexpression. B. Average GC content of mutated ORFs versus the rest of non-mutated ORFs in the indicated strains and conditions. C. Average expression levels of mutated ORFs versus the rest of non-mutated ORFs in the indicated strains and conditions. In all panels, blue lines represent the average values of all ORFs in the genome. Expression data sets used in the analysis come from Beyer et al. (2004). * p<0.05, ** p< 0.01, *** p< 0.001 (Wilcoxon´s test).

Chapter II- Effect of the loss of chromatin remodelers and histone modifiers in DNA-RNA hybrid metabolism.

1. Chromatin remodeling mutants do not enhance S9.6 signal.

Chromatin remodelers have been classified in four main families: SWI/SNF, ISWI, CHD1 and INO80 and their catalytic factors in yeast are Snf2 and Sth1 from the SWI/SNF family, Isw1 and Isw2 from the ISWI family, Chd1 from the CHD1 family and Ino80 and Swr1 from the INO80 family (Clapier and Cairns, 2009). Given that preliminary studies from the lab in human cell cultures suggest that the depletion of the catalytic subunit of chromatin remodelers from SWI/SNF and ISWI complexes leads to R loop-dependent genome instability (A. Bayona, unpublished), we decided to check these complexes in yeast in order to determine whether this mechanism is conserved through evolution. Thus, we explored the effect of the deletion of the catalytic subunit of some of such chromatin remodelers.

We studied the accumulation of DNA-RNA hybrids per cell by immunofluorescence on chromosome spreads using the S9.6 antibody and focused on snf2∆ from the SWI/SNF family, given that STH1 is an essential gene, and on isw1∆ and isw2∆ from the ISWI family. We also included the rnh1∆ rnh201∆ double mutant in all our analyses as a positive control for DNA-RNA hybrid accumulation. In addition, cells were transformed with a GAL::RNH1-containing or an empty plasmid and transformants were grown in galactose to overexpress or not the RNase H. As shown in Figure R20, 76% of the nuclei were positive for S9.6 signal in rnh1∆ rnh201∆ cells compared to the 34% observed in the WT strain. In agreement with the signal observed being caused by DNA-RNA hybrids, the overexpression of RNase H in galactose decreased these values to WT levels. However, neither snf2∆ nor isw1∆ or isw2∆ led to a detectable accumulation ofS9.6 suggesting that these mutants do not accumulate DNA-RNA hybrids in a global scale.

- RNH + RNH

rnh1∆ rnh1∆ WT snf2∆ isw1∆ isw2∆ rnh201∆ WT snf2∆ isw1∆ isw2∆ rnh201∆

DAPI

S9.6

MERGE

100

l *

a gn

i 80

6 .

9 60

t i

w 40

l c

nu % 0 RNH - + - + - + - + - + WT snf2∆ isw1∆ isw2∆ rnh1∆ rnh201∆

Figure R20. Chromatin remodeling mutants do not enhance S9.6 signal. Percentage of nuclei with detectable S9.6 signal in WT, snf2∆, isw1∆, isw2∆ and rnh1∆ rnh201∆ strains. Representative images are shown on the top. Average and SEM of three independent experiments are plotted. *p<0.05. (Paired Student´s t-test).

2. Chromatin remodeling mutants do neither lead to genome instability nor to enhanced sensitivity to AID overexpression.

Although chromosome spreads can be useful to determine global changes in DNA- RNA hybrid accumulation, the differences between samples must be substantial in order to be detected. Therefore, we decided to indirectly infer the putative accumulation of DNA-RNA hybrids by using the AID enzyme and quantifying the number of cells with Rad52 foci, which are centers of DSB repair (Lisby, Rothstein and Mortensen, 2001). For this experiment, cells were transformed with pWJ1344,

96 which contains a RAD52::YFP fusion (Lisby, Rothstein and Mortensen, 2001) together with a combination of two different plasmids containing either GAL::AID or GAL::RNH1 or both at the same time. Cells were grown on galactose to overexpress these proteins and empty plasmids were used in order not to change the carbon source (Figure R21). The levels of cells with Rad52 foci observed in the chromatin remodeling mutants were similar to the WT strain. AID overexpression slightly increased the percentage of cells with Rad52 foci in snf2∆, isw1∆, isw2∆ and rnh1∆ rnh201∆ (1.2-, 1.3-, 1.1- and 1.4-fold respectively compared to the non-AID overexpression condition), whereas we observed a slight 1.2-fold decrease in the WT strain. The overexpression of RNase H together with AID recovered the basal levels of Rad52 foci. However, the differences were not significant and far from what observed in an R-loop accumulating strain such as rnh1∆ rnh201∆ (Figure R21).

Next we wondered whether the lack of each of the chromatin remodeling catalytic subunits could have an impact on recombination levels, a very sensitive indication of genome instability. For this purpose, we checked recombination frequencies in snf2∆, isw1∆ and isw2∆ mutants using the pLZGAID, which contains both the LLacZ recombination system and the GAL1::AID construct (Figure R22A) (Garcia- Pichardo et al., 2017). The LLacZ system consists of two direct repeats of the LEU2 gene separated by the bacterial LacZ gene (Chavez and Aguilera, 1997). Recombination events generate a WT sequence of the LEU2 gene so that recombinants can be selected in a medium lacking leucine. After transforming the strains with the recombination system and the AID enzyme plus an empty or a GAL::RNH1-containing plasmid, cells were grown on glucose or galactose during 3 days to repress or overexpress the AID and the RNase H, respectively. Total and recombinants cells were measured by plating in the appropriate selective media.

rnh1∆ WT rnh201∆ AID - + + - + + RNH - - + - - + BF

DAPI

Rad52-YFP

MERGE

snf2∆ isw1∆ isw2∆

AID - + + - + + - + + RNH - - + - - + - - + BF

DAPI

Rad52-YFP

MERGE

- AID/ - RNH + AID - RNH + AID + RNH

FC:+AID/ -AID x0.6 x1.2 x1.3 x1.1 x1.4

f 52

d 10

i w

s 5

f o

% 0 WT snf2∆ isw1∆ isw2∆ rnh1∆ rnh201∆

Figure R21. Effect of AID overexpression in the levels of Rad52 foci in chromatin remodeling mutants. Percentage of cells with Rad52 foci with and without overexpression of AID and RNase H (RNH) in WT, snf2∆, isw1∆, isw2∆ and rnh1∆ rnh2∆ strains grown in galactose containing media. Representative images are shown on the top. Average and SEM of three independent experiments are plotted (Paired Student´s t-test).

In contrast to rnh1∆ rnh2∆, which led to an RNase H-sensitive hyper-recombination that was further enhanced by AID overexpression as expected, the three chromatin remodeling mutants showed similar recombination levels to the WT strain (Figure R22A). Moreover, the overexpression of the AID and RNase H enzymes did not have any effect on the recombination frequencies. Similar results were observed when we measured pRS315 plasmid loss (Figure R22B).

A B GAL1p AID CYC1t

CEN

LEU2p leu2δ3` lacZ leu2δ5` CYC1t

- AID - RNH + AID - RNH + AID + RNH - AID - RNH + AID - RNH + AID + RNH 40 ** **

100

)

4 -

30 s

(

. i

q 20 m

e 50

l P

ec 10

% R

0 0 WT snf2∆ isw1∆ isw2∆ rnh1∆ WT snf2∆ isw1∆ isw2∆ rnh201∆ Figure R22. Chromatin remodeling mutants do not lead to DNA-RNA hybrid- dependent hyper-recombination. A. Recombination frequency in WT, snf2∆, isw1∆, isw2∆ and rnh1∆ rnh2∆ strains with or without overexpression of AID and RNase H (RNH) in the indicated system. B. Percentage of pRS315 plasmid loss in the indicated strains with and without the overexpression of AID and RNase H (RNH). Average and SEM of three independent experiments are shown ** p<0.01. (Unpaired Student´s t-test).

Altogether, these data indicate that snf2∆, isw1∆ and isw2∆ mutants do not accumulate DNA-RNA hybrids nor exhibit genome instability phenotypes.

3. Rtt109 loss leads to the accumulation of DNA-RNA hybrids.

In order to decipher whether histone modifiers could have a role in DNA-RNA hybrid homeostasis, we performed a screening searching for genes whose deletion increase R loop accumulation by S9.6 immunofluorescence on chromosome spreads. The mutants studied included deletions in genes coding for histone acetyltransferases (HAT) (hat1∆, hat2∆, ada2∆, gcn5∆, spt7∆, spt8∆, rtt109∆, eaf1∆, ahc1∆, sas3∆), histone deacetylases (HDAC) (rpd3∆, sin3∆, sap30∆, rco1∆, hda1∆, hos1∆, set3∆), histone methyltransferases (HMT) (swd3∆, set2∆, dot1∆) and histone demethylases (HDM) (gis1∆, rph1∆, jhd1∆, jhd2∆). As a positive control for the experiment, we also included the rnh1∆ rnh201∆ double mutant, which presented a median of 81 % of cells with S9.6 positive signal, almost 9-fold higher than the WT (Figure R23). The top hit candidate from this screening was rtt109∆, with a median value of 37 % of nuclei with S9.6 signal, compared to the 17 % seen in the WT strain.

100

n g

i 80

. 9

S 60

t i

w 40

l c

u 20

f o

% 0 ∆ ∆ ∆ ∆ ∆ ∆ ∆ ∆ ∆ ∆ ∆ ∆ ∆ ∆ ∆ ∆ ∆ ∆ ∆ ∆ ∆ ∆ ∆ 1∆ 2 2 5 7 8 9 1 1 3 3 3 0 1 1 1 3 3 2 1 1 1 1 2 ∆ WT t t a n t t 0 f c s d n 3 o a s t d t t s h d d a a d c p p a h a p i p c d o e e o i p h h h h g s s t1 e a s r s r h s w s d g r j j rnh1 a rt sa h s rnh201∆ HAT HDAC HMT HDM

Figure R23. A screening among yeast null mutants to identify chromatin modifiers involved in DNA-RNA hybrids homeostasis. Percentage of nuclei with S9.6 signal. Each spot represents one of at least 3 independent experiments. Red line indicates the median values. Blue line is the median value of the WT.

100

In order to confirm the presence of DNA-RNA hybrids in rtt109∆ cells, we performed DRIP in PDC1, GCN4, SPF1 and PDR5 genes, which have been described to accumulate DNA-RNA hybrids in several R loop-accumulating mutants (Castellano-Pozo et al., 2013; Garcia-Pichardo et al., 2017; Herrera- Moyano et al., 2014; Lafuente-Barquero et al., 2017). We also included the 25S region of the rDNA, given that the deletion of RTT109 produces a hyper- amplification of rDNA copies in the yeast genome (Ide, Saka and Kobayashi, 2013). As seen in Figure R24, RTT109 loss caused an increase in DNA-RNA hybrid accumulation, which resulted statistically significant for the SPF1 and PDR5 regions. As expected, the S9.6 signal was reduced after in vitro treatment with RNase H prior to immunoprecipitation, indicating that the signal detected was specific for DNA-RNA hybrids.

PDC1 (1.6Kb) GCN4 (0.8Kb) rDNA 25S (3.4Kb) SPF1 (3.6Kb) PDR5 (4.5Kb)

2.5 )

es ** * * * ** ** u

l 2.0

va ve

i 1.5

t a

l *** *** *** ** e

R 1.0

(

P I

6 0.5

9 S 0.0 RNH - + - + - + - + - + - + - + - + - + - + WT rtt109∆ WT rtt109∆ WT rtt109∆ WT rtt109∆ WT rtt109∆

Figure R24. Rtt109 loss causes an increase in DNA-RNA hybrid accumulation. DNA- RNA Inmunoprecipitation (DRIP) using S9.6 antibody in the indicated regions of the genome, as detected by qPCR in WT and rtt109∆ cells. A scheme of the region analyzed is shown above each graph. Average and SEM of 4 independent experiments are shown. * p<0.05, ** p<0.01, *** p<0.001. (Paired Student´s t-test).

101

Next, we wondered whether the DNA-RNA hybrids accumulated in rtt109∆ could enhance genetic instability. In particular we studied recombination frequencies in WT and rtt109∆ strains using two different systems called co-directional (OUT) and head-on (IN) (Prado and Aguilera, 2005). Both of them consist on a plasmid with an autonomously replicating sequence (ARS) next to two truncated direct repeats of the LEU2 gene under the control of the GAL1 promoter. In the OUT system, transcription of the LEU2 gene goes in the same direction of the replication machinery whereas transcription is opposite to the direction of replication fork progression in the IN system (Figure R25A). Transcription-induction led to an increase in the recombination frequency, specifically in the IN system, as previously reported (Prado and Aguilera, 2005; Garcia-Rubio et al., 2018; Herrera- Moyano et al., 2014). However, recombination levels in rtt109∆ were similar to the WT levels, regardless of the system used and regardless of whether transcription was on or off. Moreover, RNase H overexpression had no significant effect on recombination in any of the conditions tested. These data suggest that the loss of Rtt109 does not boost the deleterious effect of transcription-replication conflicts.

Then, we measured the effect of AID overexpression in rtt109∆ cells as an increase in DNA-RNA hybrid accumulation would potentially enhance the accessibility of this enzyme if these hybrids displace the other strand as ssDNA (R loops). On one hand, we measured the recombination frequency with the pLZGAID plasmid (Garcia-Pichardo et al., 2017) and we observed that AID overexpression significantly enhanced the level of recombination in rtt109∆ cells (Figure R25B). On the other hand, we measured Rad52 foci accumulation with and without overexpression of AID in WT and rtt109∆ strains (Figure R25C). In agreement with what has been described in the literature (Alvaro, Lisby and Rothstein, 2007; Driscoll, Hudson and Jackson, 2007), rtt109∆ mutant led to an accumulation of cells with Rad52 foci of 3-fold over the WT levels. More importantly, we observed that the overexpression of AID further enhanced the levels of Rad52 foci up to 6-fold over the WT levels. All these data suggest that the loss of Rtt109 leads to the accumulation of DNA-RNA hybrids but not to hyper- recombination or transcription-replication conflicts.

102

A CO-DIRECTIONAL (OUT) HEAD-ON (IN)

CEN CEN

ARSH4 GAL1p leu2∆5´ leu2∆3´ ARSH4 leu2∆3´ leu2∆5´ GAL1p

Trx OFF Trx ON x1 x2 x0.8 x1 x1 x1 x1 x1.3 x1 x1.2x1.2 x1 x1 x1.2x1.3x0.9

40 ) 40

)

-4 0

OUT IN 0 OUT IN

( (x1

cy 30 30

requen f

20 fre 20

ion

ion at

10 bin 10

ombina

c Re 0 Rec 0 RNH - + - + - + - + RNH - + - + - + - + WT rtt109∆ WT rtt109∆ WT rtt109∆ WT rtt109∆ B C GAL1p AID CYC1t

CEN WT rtt109∆ LEU2p leu2δ3` lacZ leu2δ5` CYC1t AID - + - +

Glucose Galactose

x1 x1 x1 x2 x1 x2 x3 x6 BF ) ) 30 **

-4 25 15 -4

i *

f (

20 ( DAPI 52

** d 20

10 a

15 R

h t i Rad52-YFP

10 w s

5 ll 10

ce f

5 o

MERGE

ombinationfrequency

ombination frequency ombination ec

ec 0 0 0 R R AID - + - + T T W W WT rtt109∆ rtt109∆ rtt109∆

Figure R25. Validation of the DNA-RNA hybrid accumulation phenotype in rtt109∆. A. Recombination frequency in WT and rtt109∆ strains in the OUT and IN systems overexpressing or not the RNase H (RNH) and when transcription (Trx) is switched on or off. Schemes of the two systems are plotted. B. Recombination frequency with or without overexpression of AID in WT and rtt109∆ strains. C. Percentage of cells with Rad52 foci after the overexpression or not of AID. Representative images are shown. Average and SEM of at least 3 independent experiments are shown in all panels. * p<0.05, ** p<0.01. (Unpaired Student´s t-test).

103

4. Effect of AID* overexpression in rtt109∆ mutant.

We have seen that AID overexpression enhanced either Rad52 foci formation as well as recombination frequency (Figure R25B and C). Next, we proposed to study the effect of AID* (see Chapter I) genome-wide in rtt109∆, just like we did previously with histone and hpr1∆ mutants. Diploid cells were transformed with AID* and were grown in glucose- or galactose-containing media to repress or overexpress the enzyme during 24 hours. In addition to AID*, cells were co- transformed with an empty or GAL::RNH1 containing plasmid to study the dependency on DNA-RNA hybrids. Cell viability was not affected neither by AID* nor by RNase H overexpression (Figure R26A).

Figure R26. Effect of AID* overexpression in rtt109∆ mutant. A. Cell viability after 24 hours of AID* overexpression induction. B. Mutation frequency at CAN1 gene. Average and SEM of 2-3 independent experiments are plotted.

Again, we confirmed the mutator effect of AID* by plating a fraction of the cells in canavanine-containing plates in order to get canavanine-resistant colonies. The mutation frequency of rtt109∆ was similar to that observed for WT cells, independently of RNase H overexpression (Figure R26B).

104

We then extracted the DNA from two to three canavanine-resistant colonies coming from independent experiments and sequenced the whole genome. Surprisingly, we detected less number of mutations induced by AID* in the rtt109∆ colonies (Figure R27A). Rtt109 loss did not affect the percentage of mutations at the different nucelotides nor the number of C to T and G to A transitions, as expected from the AID* action (Figure R27B).

A B C G A T G to A C to T 150 80

* **

s on

i 60

t a

t 100 *

u m

t 40

n i

po 50

a C to T & G to A to G & T to C

t 20

To Numbertransitionsof 0 0 RNH - + - + RNH - + - +

WT rtt109∆ WT rtt109∆

Figure R27. AID*-induced mutations in rtt109∆ mutant. A. Total point mutations in WT and rtt109∆ after AID* overexpression. B. Number of C to T and G to A transitions in the indicated strains. Average and SEM of 2-3 independent experiments are plotted. * p<0.05, ** p< 0.01 (Unpaired Student´s t test).

The analysis of the different genetic features of the the AID-targeted genes features revealed that, as in the WT, mutated ORFs were longer (Figure R28A), but they did not present a higher GC content (Figure R28B) nor were more expressed (Figure R28C) than the mean value of the ORFs in the genome.

Altogether these data suggest that not all the DNA-RNA hybrids accumulated in rtt109∆ is leading to a conconmitant ssDNA accumulation as in R loops.

105

A B C 3000 50 * 1.5 * * * * * * * * * * * * * *

* * ) 40

) %

2000 ( 1.0

t bp

( 30

(Exp)

2 on

ng 20

c e

1000 Log 0.5

L C

G 10

0 0 0.0 RNH - + - + RNH - + - + RNH - + - + WT rtt109∆ WT rtt109∆ WT rtt109∆ Figure R28. Meta-analysis of AID*-induced mutations at ORFs. A. Average length of mutated ORFs in WT and rtt109∆ strains with and without RNase H (RNH) overexpression. B. Average GC content of mutated ORFs in the indicated strains and conditions. C. Average expression levels of mutated ORFs in the indicated strains and conditions. In all panels, blue lines represent the average values of all ORFs in the genome. Expression data sets used in the analysis come from Beyer et al. (2004). ** p< 0.01, *** p< 0.001 (Wilcoxon´s test).

5. Rtt109 prevents DNA-RNA hybrid accumulation through its catalytic activity.

Since Rtt109 is a HAT, we wondered whether the accumulation of DNA-RNA hybrids in rtt109∆ mutant is a consequence of the inability to acetylate its targets. We tested this possibility by using three different plasmids. One of them contains a WT version of the RTT109 gene. The other two plasmids contain a point mutation inside its catalytic domain, changing the same residue of aspartate (D89) to alanine (D89A) or to asparagine (D89N) both impairing its catalytic activity as a HAT (Han et al., 2007). After transforming the rtt109∆ mutant with the indicated plasmids, we checked if the resulting transformants kept accumulating DNA-RNA hybrids by S9.6 immunofluorescence on chromosome spreads and DRIP (Figure R29). By immunofluorescence, we observed that the transformation of rtt109∆ cells with the plasmid containing the WT sequence of the gene reduces the level of S9.6 signal back to WT levels. By contrast, the catalytic mutants (rtt109-D89A and rtt109-D89N) were unable to reduce the S9.6 signal (Figure R29A). By DRIP, we observed similar results (Figure R29B), revealing that Rtt109 prevents DNA-RNA hybrid accumulation through its catalytic activity.

106

A WT rtt109∆

empty empty RTT109 rtt109 rtt109 * D89A D89N 75 vector vector * *

60 DAPI

30 S9.6 15

Nuclei with S9.6 signal (%) signal S9.6 with Nuclei 0 empty empty rtt109 rtt109 RTT109 vector vector D89A D89N MERGE WT rtt109∆

2.0 GCN4 (0.8Kb) 2.0 PDC1 (1.6Kb)

)

) s * s

1.5 * * 1.5 alue

alue *

ive

ive t ** * t * **

1.0 1.0

(Rela

6 .

0.5 . 0.5

S S

0.0 0.0 RNH - + - + - + - + - + RNH - + - + - + - + - + empty empty rtt109 rtt109 empty empty rtt109 rtt109 vector vector RTT109 D89A D89N vector vector RTT109 D89A D89N WT rtt109∆ WT rtt109∆

2.5 SPF1 (3.6Kb) 2.5 PDR5 (4.5Kb)

* *

)

) s * s 2.0 2.0

value

e v

1.5 v 1.5

t t ** *

*** ** (Rela

1.0 (Rela 1.0

. 9

0.5 9 0.5

S S

0.0 0.0 RNH - + - + - + - + - + RNH - + - + - + - + - + empty empty rtt109 rtt109 empty empty rtt109 rtt109 RTT109 RTT109 vector vector D89A D89N vector vector D89A D89N WT rtt109∆ WT rtt109∆

Figure R29. DNA-RNA hybrid accumulation phenotype in rtt109 catalytic mutants. A. Percentage of nuclei with S9.6 foci in the indicated strains, complemented with an empty plasmid, a wild type RTT109 gene or two mutated versions of the gene (rtt109-D89A and rtt109-D89N). Representative images of S9.6 immunofluorescence are shown on the right of the graph. B. DNA-RNA immunoprecipitation (DRIP) at GCN4, PDC1, SPF1 and PDR5 genes in WT and rtt109∆ cells. Average and SEM of at least 3 independent experiments are shown. * p<0.05, ** p<0.01, *** p<0.001 (Paired Student´s t-test).

107

6. DNA-RNA hybrids accumulate in rtt109∆ cells in all cell cycle phases.

Acetylation of lysine K56 of histone H3 (H3K56) is needed to balance the expression of genes that have been replicated during S phase, and the lack of Rtt109, the only one HAT which acetylates lysine K56, leads to an increase in the transcription rate of replicated genes (Han et al., 2007; Driscoll, Hudson and Jackson, 2007; Voichek et al., 2018). We hypothesized that the accumulation of hybrids observed in rtt109∆ mutant could be linked to such increase in the transcription of the genes that have been duplicated during S phase. If that were the case, we would expect the accumulation of DNA-RNA hybrids to be specific of S phase and reproduced by mutations in H3K56.

Thus, we performed chromosome spreads in all the cell cycle phases in the rtt109∆ mutant by synchronizing cells with α-factor during two hours and releasing them in fresh medium containing nocodazole for 30 and 120 minutes to collect S and G2 cells respectively (Figure R30). The analysis of the chromosome spreads in the different cell cycle phases revealed that rtt109∆ mutant accumulates DNA- RNA hybrids independently of the cell cycle phase. This result argues that the DNA-RNA hybrid accumulation phenotype observed in rtt109∆ cells is not just due to an increase in the transcription of replicated genes.

108

α-factor nocodazole

Async. G1 S G2 120’ 30’ 90’ α-factor WT rtt109∆ WT

l 50 rtt109∆ 120’ (G2) 40 * * 30’ (S) 30

20 0’ (G1) 10 Async % of nuclei with S9.6 signa S9.6 with nuclei of % 0 G1 S G2

Figure R30. rtt109∆ mutant accumulates DNA-RNA hybrids independently of the cell cycle phase. Percentage of nuclei with S9.6 signal in WT and rtt109∆ strains in G1, S or G2 phases. A scheme of the experiment is drawn on the top. The DNA content as analyzed by flow cytometry (FACS) of WT and rtt109∆ strains at G1, S and G2 phases is plotted on the left of the graph. Average and SEM of 4 independent experiments are shown. *p<0.05 (Unpaired Student´s t-test).

7. Histone H3 deposition during S phase is not decreased in rtt109∆ mutant

Rtt109 intervenes in replication-coupled nucleosome assembly during S phase by acetylating H3K56 in newly synthesized histone dimers H3-H4 presented by chaperone Asf1 (Asf1-H3-H4 complex) before their incorporation into the replicated DNA (Masumoto et al., 2005). This acetylation mark is recognized by other histone modifiers that ultimately would make CAF1 and Rtt106 transfer the dimer H3K56ac-H4 into the DNA (Su et al., 2012). We thus hypothesized that the accumulation of DNA-RNA hybrids in rtt109∆ cells could be a consequence of the putative defect in the assembly of newly-synthesized histones during replication. To test this possibility, we measured incorporation of histone H3 into the replicating DNA by performing a ChIP in WT and rtt109∆ cells in S phase. We synchronized cells in G1 phase and collected samples at 30 and 45 minutes after α-factor release (Figure R31). qPCRs were made at the PDC1 locus, proximal to the replication origin ARS1211. Contrary to what expected, the analysis of qPCRs

109 revealed an increase in the histone H3 levels into the replicating DNA for rtt109∆ mutant. The observed increase might be due to the reported delay in cell cycle progression of rtt109∆ (Driscoll, Hudson and Jackson, 2007), as we observed in FACS analysis of rtt109∆, especially visible at 45 minutes after α-factor release. Therefore, we believe that the differences of the histone H3 levels observed are due to the slower replication of rtt109∆ respect to the WT and we can rule out a putative defect in H3 levels as a consequence of less deposition as the cause behind DNA-RNA hybrid accumulation in rtt109∆ cells.

In fact, a further MNase-Seq analysis (See Material and Methods) revealed that only the 10 % of nucleosomes lost their positioning in the genome. These differences are very subtle and localize at specific regions of the genome. In addition, and again, we could not find any correlation between changes in nucleosome positioning and R loop accumulation.

A B

WT rtt109∆ ARS1211 PDC1 (1.6Kb)

20 20

45’ WT g

n 1 rtt109Δ i

15 on 15

t i

** y474

input) B

30’ f

po o

vs 10

10 e

)

% (

0’ s o

5 e 5

H3 IP ( H3 IP

l u

Async N 0 0 Time after release 0’ 30’ 45’ from α factor rtt109∆

Figure R31. Histone H3 deposition is not decreased in the absence of Rtt109. A. Chromatin inmunoprecipitation of histone H3 in WT and rtt109∆ mutant during S phase. Above the plot, a scheme of the studied region is shown. DNA content analysis by flow cytometry (FACS) of WT and rtt109∆ strains at different stages of the cell cycle is shown on the left of the graph. Average and SEM of at least 3 independent experiments are shown. ** p<0.01. (Paired Student´s t-test). B. Nucleosome positioning loss of rtt109∆ compared to BY4741 (In collaboration with F. Antequera's lab).

110

8. DNA-RNA accumulation in rtt109∆ is independent of the H3K56 acetylation state.

We have seen so far that rtt109∆ mutant accumulates DNA-RNA hybrids that is not due to an increase in transcription of replicated genes during S phase or a defect in the deposition of histones during replication. Next, we wondered if this phenotype could be due to the lack of H3K56 acetylation (H3K56ac), since the main function of this enzyme is to acetylate this residue during the duplication of the genome, together with Asf1 chaperone (Han et al., 2007; Driscoll, Hudson and Jackson, 2007; Masumoto et al., 2005). After deposition of H3K56ac-H4 dimers into the DNA behind the replication fork, this acetylation mark is removed by the Hst3 and Hst4 sirtuins and their absence is known to cause genetic instability (Celic et al., 2006). So, we checked directly the presence of DNA-RNA hybrids by using S9.6 inmunofluorescence on chromosome spreads in asf1∆, hst3∆, hst4∆, hst3∆ hst4∆ and H3K56A mutants (Figure R32A). From all of such single and double mutants, only asf1∆ presented a significant increase in DNA-RNA hybrids, although minor to that observed in rtt109∆. However, when we performed DRIP experiments, we observed that, in contrast to rtt109∆, neither asf1∆ nor H3K56A mutants accumulated DNA-RNA hybrids at any of the gene locations tested (Figure R32B). These results indicate that the accumulation of DNA-RNA hybrids in rtt109∆ mutant is independent of its role on H3K56 acetylation.

111

T W rtt109∆ asf1∆ hst3∆ hst4∆ hst3∆hst4∆

DAPI

80 l

S9.6 a * *** gn i ***

s 60 ***

. 9

MERGE S

h t

i 40

e l

A A A A c T A

nu 20 f

H3W H3K9 H3K14 H3K23 H3K27 H3K56 o

DAPI % 0 T T A A A A A W asf1∆hst3∆hst4∆ S9.6 rtt109∆ H3WH3K9 H3K14H3K23H3K27H3K56 hst3∆hst4∆ MERGE

PDC1 (1.6Kb) GCN4 (0.8Kb) )

2.0 ) 2.0 * es

** es

l va 1.5 va 1.5

*** *** ***

ve i

*** *** i *** *** t

** t *** **

a l

1.0 l 1.0

(

( P

0.5 P 0.5

9 S 0.0 S 0.0 - + - + - + - + - + - + - + - + - + - + - + - +

T T A A T T A A W W asf1∆ H3W asf1∆ rtt109∆ H3K9 rtt109∆ H3W H3K9 H3K56 H3K56

SPF1 (3.6Kb) PDR5 (4.5Kb) )

2.5 ) 2.5

u l

2.0 * * l 2.0 ** **

ve ve

i *

1.5 i 1.5

t t

a **

a l

** l *** e

*** ** ** e **

1.0 1.0 ** *

(

P I

0.5 I 0.5

9 S 0.0 S 0.0 - + - + - + - + - + - + - + - + - + - + - + - +

T T A A T T A A W W asf1∆ H3W asf1∆ rtt109∆ H3K9 rtt109∆ H3W H3K9 H3K56 H3K56 112

Figure R32. R loops accumulation phenotype of rtt109∆ mutant is not mediated by the acetylation state of H3K56. A. (Left panel). Representative images of chromosome spread in the indicated strains. (Right panel). Percentage of nuclei with S9.6 foci. B. DRIP of the indicated strains at PDC1, GCN4, SPF1 and PDR5 genes. Average and SEM of at least 3 independent experiments are shown. *p<0.05, **p<0.01, ***p<0.001. (Paired Student´s t-test).

9. Mutation of H3K14 and H3K23 (Rtt109 targets) to alanine increases DNA-RNA hybrid accumulation.

Although K3K56 is the best-described substrate of Rtt109, this HAT has also been reported to acetylate K3K9, H3K23 and H3K27 in vivo and H3K14 in vitro (Dahlin et al., 2015). So, we wondered if the accumulation of DNA-RNA hybrids in rtt109∆ mutant was a consequence of a lack or a reduction in the acetylation levels of these lysine residues. For this purpose, we analyzed the presence of DNA-RNA hybrids by S9.6 immunofluorescence on chromosome spreads in H3K9A, H3K14A, H3K23A and H3K27A histone mutants (Figure R32A). The experiment revealed that the H3K14A and H3K23A histone mutants accumulate DNA-RNA hybrids to the same extent as rtt109∆ whereas H3K9A had no effect. This result not only suggests that H3K14 and H3K23 could be the targets of Rtt109 that are responsible for its role preventing DNA-RNA hybrid formation but also nails down the mutations responsible for the previously observed phenotype of H3K9-23A (see Chapter 1 and (Garcia-Pichardo et al., 2017) and allow us to propose that H3K9-23A was leading to DNA-RNA hybrids due to the inability to be acetylated by Rtt109 at the K14 and K23 residues.

We reasoned that if the loss of Rtt109 led to DNA-RNA hybrid accumulation by the inability to acetylate H3K14 and H3K23, the loss of Rtt109 should have no effect in H3K14Q and H3K23Q acetylmimetic mutants. As shown in Figure R33, these mutants led to a slight increase in the percentage of cells with S9.6 positive signal that was significant in the case of H3K14Q. However, we observed no further effect by the deletion of RTT109 arguing that H3K14 and H3K23 are indeed the targets responsible for the R loop accumulating phenotype of rtt109∆.

113

H3-WT H3 K14Q H3 K23Q 60 ** l

RTT109 rtt109∆ RTT109 rtt109∆ RTT109 rtt109∆ a

gn i

DAPI s 6

. 40

h t

S9.6

i wi i e

l 20

nu f

MERGE o

% 0

TT109 rtt109∆ TT109 rtt109∆ TT109 rtt109∆ R R R H3WT H3K14Q H3K23Q

Figure R33. Deletion of RTT109 gene in acetylmimetic mutans for lysines 14 and 23 do not accumulate higher levels of S9.6 signal compared to single histone mutants. Percentage of nuclei with S9.6 signal in acetylmimetic histone H3 mutants for lysines K14 and K23 with or without Rtt109. Average and SEM of 4 independent experiments are shown. Representative images of chromosome spread are shown on the left. * p<0.05, ** p<0.01. (Unpaired Student´s t-test).

10. The genotoxic sensitivity of rtt109∆ mutant is not dependent on DNA- RNA hybrids.

It is described that rtt109∆ mutant is sensitive to agents causing replication stress such as hydroxyurea (HU), camptothecin (CPT) and methyl methanesulfonate (MMS), but not to UV (Fillingham et al., 2008; Han et al., 2007). We wondered if any of these phenotypes was due to the enhanced formation of DNA-RNA hybrids. To test this possibility we transformed a WT strain and the rtt109∆ mutant with an empty plasmid or a GAL::RNH1-containing plasmid and then plated them in galactose-containing minimum media supplemented with the drugs (HU, CPT and MMS). As shown in Figure R34, the sensitivity of rtt109∆ was not suppressed by RNase H overexpression indicating that this phenotype is likely not related to the observed DNA-RNA hybrid accumulation.

114

RNH MOCK 2.5 μg/ml CPT 50 mM HU 0.001 % MMS - WT +

- rtt109Δ +

Figure R34. Genotoxic sensitivity of rtt109∆ mutant is not suppressed by RNase H overexpression. Drop assay in WT and rtt109∆ strains grown in medium containing the replication stress agents hydroxyurea (HU), camptothecin (CPT) or methyl methanesulfonate (MMS) at the indicated strains during 4 days at 30ºC.

11. The Rad52 foci accumulation phenotype of rtt109∆ mutant is partially caused by DNA-RNA hybrids.

Deletion of RTT109 causes an increase in the accumulation of Rad52 foci (Alvaro, Lisby and Rothstein, 2007; Driscoll, Hudson and Jackson, 2007) and we wondered whether this increase in DNA damage resulted from the accumulation of DNA-RNA hybrids. So, we measured Rad52 foci in WT and rtt109∆ strains transformed with pWJ1344 (containing RAD52::YFP) plus an empty or GAL::RNH1-containing plasmid. As shown in Figure R35, we confirmed the reported increase in cells with Rad52 foci in rtt109∆. Strikingly, the overexpression of RNase H in rtt109∆ mutant reduced partially the levels of cells with Rad52 foci from 26% to 16%, implying that part of the DNA damage could be a consequence of the accumulation of DNA-RNA hybrids.

115

WT rtt109∆ RNH - + - +

BF 40

c o f *** *

2 30

5 d

DAPI a R

h 20

l l

e 10

Rad52-YFP c

f o

% 0 RNH - + - + MERGE WT rtt109∆

Figure R35. Rad52 foci accumulation phenotype of rtt109∆ mutant is partially suppressed after RNase H overexpression. Percentage of cells with Rad52 foci. Average and SEM of at least 3 independent experiments are plotted in the graph. Representative images of Rad52-YFP foci in WT and rtt109∆ strains with or without overexpression of RNase H (RNH) are shown on the left. * p<0.05, *** p<0.001 (Paired Student's t test).

Given our previous results regarding H3K14 and H3K23 (Figure R32 and Figure R33), we hypothesized that this partial reduction of cells with Rad52 foci could be related to the DNA-RNA hybrids accumulation coming from a reduction in the acetylation levels of H3K14 and H3K23. So, we measured the accumulation of Rad52 foci in the non-acetylable H3K14A and H3K23A histone mutants, overexpressing or not the RNase H enzyme. We also used the H3K9A strain as a negative control. As shown in Figure R36, only the H3K23A mutant led to a significant accumulation of Rad52 foci compared to the WT strain. Moreover, after the overexpression of RNase H, Rad52 levels were slightly reduced. Although H3K14A mutant did not have higher levels of cells with Rad52 foci respect to the WT, the levels of Rad52 foci also decreased significantly after the overexpression of RNase H. These results suggest that the partial reduction of DNA damage observed in rtt109∆ by RNase H overexpression may be due to the defective

116 acetylation of H3K23 alone or in combination with a reduction in the acetylation levels of H3K14.

HPR1 H3WT H3K9A H3K14A H3K23A RNH - + - + - + - +

Rad52-YFP

MERGE

hpr1 H3WT H3K9A H3K14A H3K23A RNH - + - + - + - +

Rad52-YFP

MERGE

* 30

i * * * * 20

10 % of cells with Rad52 foc Rad52 with cells of % 0 RNH - + - + - + - + - + - + - + - + H3WT H3K9A H3K14A H3K23A H3WT H3K9A H3K14A H3K23A

hpr1 HPR1

Figure R36. Study of Rad52 foci accumulation in the non-acetylable H3K9A, H3K14A and H3K23A mutants in HPR1 and hpr1∆ backgrounds. Representative images of Rad52-YFP foci in the indicated strains with or without overexpression of RNase H are shown. Percentage of cells with Rad52 foci. Average and SEM of at least 3 independent experiments are plotted in the graph. *p<0.05 (Unpaired Student t-test).

117

12. H3K9A mutation is able to suppress the genetic instability of hpr1∆.

H3K9-23A suppressed the replication defects and genome instability observed in hpr1∆ mutant (see Chapter 1 and Garcia-Pichardo et al. (2017)), although which of the mutations (H3K9A, H3K14A, H3K18A or H3K23A) was responsible for this suppression was unknown. The observation that H3K23A leads to elevated number of cells with Rad52 foci (Figure R36) argues that this residue is not responsible for this suppression. We also studied the Rad52 foci accumulation in H3K23A, H3K14A and H3K9A mutants in combination with the deletion of HPR1. As it can be observed in Figure R36, H3K14A hpr1∆ and H3K23A hpr1∆ double mutants presented the same levels of Rad52 foci as hpr1∆ cells, which were reduced after the overexpression of RNase H and only the H3K9A mutation was able to suppress the Rad52 foci accumulation observed in hpr1∆ mutant. Therefore, we can conclude that whereas the inability of Rtt109 to acetylate H3K14 and/or H3K23 leads to DNA-RNA hybrid accumulation and subsequent increase in Rad52 foci, the H3K9A mutation is the responsible for the suppression of the replication defects and genetic instability phenotypes of hpr1∆.

13. The SCR defect of rtt109∆ is not caused by DNA-RNA hybrid accumulation.

From previous studies in our lab, we know that all mutations related to acetylation of H3K56, including rtt109∆ and also hst3∆, hst4∆, hst3∆ hst4∆ as well as changes of H3K56 to alanine (non-acetylable, H3K56A) or glutamine (acetylmimetic, H3K56Q), are defective in the repair of replication-born DSBs by sister-chromatid recombination (SCR) (Munoz-Galvan et al., 2013). Given that DNA-RNA hybrids can accumulate at DNA breaks, we speculated that the DNA-RNA hybrid accumulation could be related to this SCR defect.

In order to study SCR, we took advantage of a plasmid previously developed in our lab (pTHGH) (Ortega, Gomez-Gonzalez and Aguilera, 2019) which contains both, the yeast mating HO endonuclease under the GAL1 promoter and the TINV-HO recombination system. This system is based in two inverted copies of the LEU2

118 gene. One of them has a deletion of the 5´ region and the other one presents a reduced recognition site for the HO endonuclease of 24 bp. The efficiency of the HO endonuclease at this site is very low, which favors that most of the breaks occurs in only one of the single chromatids, while the other one remains unbroken and can be used as a template to repair (Gonzalez-Barrera et al., 2003; Cortes- Ledesma and Aguilera, 2006). Induction of the HO endonuclease generates a single-strand break that replication can convert into a DSB, which will be preferentially repaired by SCR (Kadyk and Hartwell, 1992; Gonzalez-Barrera et al., 2003; Cortes-Ledesma and Aguilera, 2006), although an insignificant amount of DSBs can also be repaired by intra-chromatid recombination. After digestion of the DNA with specific restriction enzymes, we can detect by Southern Blot the appearance of bands corresponding to the DSBs and to intermediates of unequal SCR, which is a good approximation for total SCR (Gonzalez-Barrera et al., 2003) (Figure R37).

119

Figure R37. Repair of a HO-induced DSB by sister chromatid recombination (SCR). A. Scheme of recombination steps during SCR repair of an HO-induced DSB after galactose addition. The HO enzyme creates a nick, which turns into a DSB after replication, and is mainly repaired by SCR, although it can also be repaired by intra- chromatid recombination (ICR, not shown). Digestion with SpeI and XhoI generates different DNA fragments that reflect the different repair interediates and can be detected by Southern blot using a probe against the LEU2 gene. B. Example of Southern blot analysis to detect the SCR and ICR intermediates.

120

We first deleted the RTT109 gene in a histone H3WT strain and confirmed the SCR defect in the H3 background. As shown in Figure R38, although the percentage of DSBs was similar in all mutants, the efficiency of their repair by SCR was severely impaired in the absence of Rtt109 since the percentage of molecules detected corresponding to SCR intermediates 9 hours after the induction of the HO drops from 17% to less than 10%. To check if that phenotype was also observed in the histone mutants related to Rtt109 that accumulate DNA-RNA hybrids, we checked SCR in H3K14A and H3K23A strains and in H3K9A as a control (Figure R38). From all the histone mutants analyzed, only the H3K14A mutant showed a defect in repair by SCR. Moreover, this defect was as strong as the one seen in the rtt109∆ mutant suggesting that, in addition to the reported role for the acetylation of K56 (Munoz-Galvan et al., 2013), the acetylation of H3K14 by Rtt109 might be relevant for SCR.

We wondered whether this SCR defect could be caused by the DNA-RNA hybrid accumulation of H3 rtt109∆ and H3K14A mutants. So, we performed repair kinetics with H3WT, H3 rtt109∆ and H3K14A strains overexpressing or not the RNase H (Figure R39 and Figure R40). In this case, we took two additional samples 1.5 and 4.5 hours after the HO induction to have a better resolution and we performed the experiments in the presence of doxyciclin so that transcription of the leu2 repeat from the tet promoter was not allowed. As a negative control, we used the H3K23A mutant, which was not defective in SCR (Figure R41). We observed again the defect in SCR in both, H3 rtt109∆ and H3K14A mutants, which cannot repair properly the DSB induced by the HO endonuclease, but not in H3K23A. However, the overexpression of RNase H did not have any effect over the efficiency of repair in any of the mutants, arguing that DNA-RNA hybrids are not the cause of the SCR defect observed.

121

H3WT H3K9A H3K14A H3K23A H3 rtt109∆ Time (h) 0 3 6 9 0 3 6 9 0 3 6 9 0 3 6 9 0 3 6 9 SCR

Plasmid

SCR + ICR

DSB1

DSB2

H3WT H3K9A H3K14A H3K23A H3 rtt109∆

40 20

30 15

C S

D 20 10

% % 10 5

0 0 0 3 6 9 0 3 6 9 Time after HO induction (h) Time after HO induction (h)

Figure R38. H3 rtt109∆ and H3K14A mutants share the same SCR defect. On the top, southern blot showing the physical analysis of HO-induced DSB formation and the repair kinetics for H3WT, H3K9A, H3K14A, H3K23A and H3 rtt109∆ strains. Percentage of total DSB corresponding to DSB1 + DSB2 intensities is shown on the left. Percentage of SCR intermediates is shown on the right. All quantifications are related to the total intensity of each lane. Average and SEM of 1 or 2 independent experiments are shown.

122

H3WT H3 rtt109∆

RNH - + - +

Time (h) 0 1.5 3 4.5 6 9 0 1.5 3 4.5 6 9 0 1.5 3 4.5 6 9 01.5 3 4.5 6 9

SCR Plasmid

SCR + ICR

DSB1

DSB2

H3WT H3WT + RNH H3 rtt109∆ H3 rtt109∆ + RNH

60 20

B R

10 % DS % 20 SC % 5

0 0 0 1.5 3 4.5 6 9 0 1.5 3 4.5 6 9 Time after HO induction (h) Time after HO induction (h)

Figure R39. Overexpression of RNase H does not rescue the SCR defect of H3 rtt109∆. On the top, southern blot showing the physical analysis of HO-induced DSB formation and their repair kinetics for H3WT and H3 rtt109∆ strains with or without RNase H overexpression. Percentage of total DSB corresponding to DSB1 + DSB2 intensities is shown on the left. Percentage of SCR intermediates is shown on the right. All quantifications are related to the total intensity of each lane. Average and SEM of 3 independent experiments are shown.

123

H3WT H3K14A RNH - + - + Time (h) 0 1.5 3 4.5 6 9 0 1.5 3 4.5 6 9 0 1.5 3 4.5 6 9 0 1.5 3 4.5 6 9 SCR Plasmid

SCR + ICR

DSB1

DSB2

H3WT H3WT + RNH H3K14A H3K14A + RNH

60 20

S CR

D 10

S % % 20 5

0 0 0 1.5 3 4.5 6 9 0 1.5 3 4.5 6 9 Time after HO induction (h) Time after HO induction (h) Figure R40. Overexpression of RNase H does not rescue the SCR defect of H3K14A mutant. On the top, southern blot showing the physical analysis of HO-induced DSB formation and their repair kinetics for H3WT and H3K14A strains with or without RNase H overexpression. Percentage of total DSB corresponding to DSB1 + DSB2 intensities is shown on the left. Percentage of SCR intermediates is shown on the right. All quantifications are related to the total intensity of each lane. Average and SEM of 3 independent experiments are shown.

124

H3WT H3K23A RNH - + - + Time (h) 0 1.5 3 4.5 6 9 0 1.5 3 4.5 6 9 0 1.5 3 4.5 6 9 0 1.5 3 4.5 6 9 SCR Plasmid

SCR + ICR

DSB1

DSB2

H3W T H3WT + RNH H3K23A H3K23A + RNH

60 20

S C

D 10

S % % 20 5

0 0 0 1.5 3 4.5 6 9 0 1.5 3 4.5 6 9 Time after HO induction (h) Time after HO induction (h)

Figure R41. H3K23A mutant does not have any SCR defect. On the top, southern blot showing the physical analysis of HO-induced DSB formation and their repair kinetics for H3WT and H3K23A strains with or without RNase H overexpression. Percentage of total DSB corresponding to DSB1 + DSB2 intensities is shown on the left. Percentage of SCR intermediates is shown on the right. All quantifications are related to the total intensity of each lane. Average and SEM of 3 independent experiments are shown.

125

14. rtt109∆ leads to the spontaneous accumulation of DSBs that are not induced by the presence of DNA-RNA hybrids.

The fact that rtt109∆ accumulates DNA-RNA hybrids and leads to a defect in SCR that is independent on such hybrids led us to think about the possibility that the accumulation of DNA-RNA hybrids could be in part consequence of its defect in DSB repair. Persistent unrepaired DSBs present in rtt109∆ could favor the RNA hybridization to its DNA template. If that were the case, the removal of DNA-RNA hybrids should not affect the levels of DSBs accumulated in rtt109∆ mutant. To test this hypothesis, we measured the effect of the overexpression of RNase H on the global levels of DSBs indirectly by testing the levels of phosphorylated H2A (H2A- P), a well-known DSB marker (Figure R42). As previously reported (Fillingham et al., 2008), rtt109∆ led to an accumulation of H2A-P. Strikingly, these levels did not decrease after RNase H overexpression, suggesting that rtt109∆, in contrast to other R loop accumulating mutants, accumulates DNA damage independently of the presence of R loops. This result is in agreement with DNA-RNA hybrid formation being favored by the presence of unrepaired DNA breaks in rtt109∆.

4 ** WT rtt109∆

RNH - + - + 3 (ratio)

H2A-P H3 2 P/

H3 1 H2A-

0 RNH - + - + WT rtt109Δ

Figure R42. DSB increase in rtt109∆ is not caused by DNA-RNA hybrids. Western blot of global levels of H2A-P respect to the total amount of histone H3. Representative image of western is shown on the left. ** p<0.01. (Unpaired Student´s t-test).

126

15. Deletion of HPR1 gene in an rtt109∆ background leads to a synergistic effect on genotoxic sensitivity.

So far, our data suggest that the reason why the rtt109∆ mutant accumulates DNA damage is not dependent on DNA-RNA hybrids. Thus, we thought that a genetic interaction could exist between rtt109∆, which leads to persistent unrepaired DSBs that favor DNA-RNA hybridization, and a classical DNA-RNA hybrid-accumulating mutant, such as the THO complex mutant hpr1∆, which reduces the protection of the nascent RNA thus also favoring DNA-RNA hybridization. However, no genetic interaction has been reported between these two factors to our knowledge. Thus, we decided to get this double mutant and study the sensitivity to different genotoxic agents reasoning that further replicative stress would reveal any possible genetic interaction. By S9.6 immunofluorescence on chromosome spreads, we observed similar levels of S9.6 positive cells detected in rtt109∆, hpr1∆, and rtt109∆ hpr1∆ strains (Figure R43A). However, after plating serial dilutions of WT, rtt109∆, hpr1∆, and rtt109∆ hpr1∆ cultures on CPT, HU or MMS-containing media and incubating them during 3 days, we detected that the double mutant was more sensitive to replication stress than any of the single mutants (Figure R43B). These data are in agreement with the source of DNA damage in rtt109∆ mutant being different to hpr1∆.

127

A rtt109∆ WT rtt109∆ hpr1∆ hpr1∆

l 50

a *

gn * i

DAPI s 40

. 9

S 30

i w

i 20 e

S9.6 l c

nu 10

f o

% 0 T MERGE W rtt109∆ hpr1∆

rtt109∆ hpr1∆ B MOCK 50 mM HU 0.005% MMS 2.5 μg/ml CPT

rtt109∆

hpr1∆

rtt109∆ hpr1∆

Figure R43. Deletion of HPR1 gene in a rtt109∆ background leads to a synergistic effect on genotoxic sensitivity. A. Percentage of nuclei with S9.6 signal. Average and SEM of 3 independent experiments are plotted. Representative images are shown. *p<0.05 (unpaired Student´s t-test). B. Representative images of drop assay after 3 days of growth in the indicated strains exposed to different replicative stress drugs.

128

DISCUSSION

129

130

1. New insights from histone mutants leading to DNA-RNA hybrid accumulation without genetic instability.

After the discovery of R loops, many works have related their accumulation with genome instability, promoting the idea that, although they can play a physiological role, unbalanced homeostasis of R loop formation is deleterious (reviewed in Garcia-Muse and Aguilera (2019); Santos-Pereira and Aguilera (2015); Aguilera and Garcia-Muse (2012); Skourti-Stathaki and Proudfoot (2014); Sollier and Cimprich (2015) and Hamperl and Cimprich (2014)). Indeed, R loop accumulation has been associated with several diseases (reviewed in Richard and Manley (2017)). Moreover, many reports have related the R loop associated genetic instability to the impairment of replication fork progression (Aguilera and Gomez- Gonzalez, 2008; Bermejo, Lai and Foiani, 2012). However, not all R loops lead to genome instability, as we reported during the course of this thesis (Garcia- Pichardo et al., 2017). In particular, a parallel work led to the identification of histone H3 and H4 mutants that led to an increase in R loop accumulation but, in contrast to all previously identified R loop accumulating mutants, they were not triggering hyper-recombination or increased formation of Rad52 foci. These mutants were the deletion of the amino-terminal tail of histone H3 (H3∆1-28 mutant) and specific changes of lysines 9, 14, 18 and 23 to alanine in the histone H3 (H3K9-23A) or lysine 31 to glutamine in the histone H4 (H4K31Q).

Although further experiments would be required to address the exact mechanism by which each particular mutation leads to DNA-RNA hybrid accumulation, the results of the second chapter of this thesis point to the loss of histone acetylation by Rtt109 at residues H3K14 and H3K23 as the cause of DNA-RNA hybrid accumulation in some of these mutants (Figure R32 and Figure R33). Moreover, our genome-wide MNase-seq data argue against the possibility that the cause of DNA-RNA hybrid formation is related with changes in the nucleosome distribution among the genome, since only H3∆1-28 resulted in a significant but subtle loss in the positioning of about 6.5 % of nucleosomes (Figure R4B) and these changes in nucleosome positioning did not colocalize with loci at which such histone mutants

131 accumulate DNA-RNA hybrids, such as GCN4 gene (Figure R6B). In addition, our experiments suggest that DNA-RNA hybrids do not lead to significant changes in nucleosome positioning, as demonstrated by the fact that the R-loop accumulating hpr1∆ and rnh1∆ rnh2∆ mutants present a normal nucleosome pattern (Figure R4A and D).

Independently of how DNA-RNA hybrids are generated, in this thesis, we sought to further characterize the phenotypes of these histone mutants regarding both the quality and quantity of R loop accumulation versus other mutant strains as well as the consequences of the R loop accumulation in both genetic instability and replication fork progression. In agreement with the absence of hyper-recombination phenotype previously detected (Garcia-Pichardo et al. (2017), we were not able to see an increase in genome instability in the form of LOH or plasmid loss unless the AID enzyme was overexpressed (Figure R1 and Figure R2). Regarding replication, we observed that the H3K9-23A mutant exhibited a very slight growth defect in the presence of replicative stress (HU) (Figure R2) but this defect was not strong enough to be detected as an impediment in replication fork progression, as observed by 2D-gel electrophoresis (Figure R3C). Thus, our results further supported that these histone mutants do indeed accumulate R loops that are targeted by AID but these R-loops do neither cause replication problems nor genome instability by themselves.

2. Setting up the bisulfite deamination assay in yeast reveals that DNA-RNA hybrid length is not a relevant factor for the generation of genetic instability.

We considered that the length of the R loops might be related to the replication problems or genetic instability phenotypes. For instance, larger R loops could be formed in hpr1∆ that would lead to the formation of an obstacle more difficult to bypass by the replication fork. To determine DNA-RNA hybrid length, we took advantage of sodium bisulfite mutagenesis and subsequent amplification and sequencing of single molecules, a tool that had been previously used at immunoglobulins locus (Huang et al., 2006; Kao et al., 2013; Yu et al., 2003). The 132 comparison of the R loop length between histone and hpr1∆ mutants brought out that the phenotypes observed in these strains could not be explained by R loops of different sizes, since their length was similar among them and respect to wt strain (Figure R9). That observation was extended to other R loop accumulating mutants, such as hpr1∆, mft1∆, thp1∆, spt16-11, pob3-7, top1∆, sen1-1, and rnh1∆ rnh2∆ mutants, which presented similar R loop length average and variation (Figure R10). It is noteworthy that all molecules with C to T conversions coming from top1∆ mutant exhibited a constant R loop length of 72 bp. Further experiments would be required to understand this result but we envision that one possibility is that the positive supercoiling generated ahead of the RNA polymerase in the absence of topoisomerase I, blocks further progression of the transcription machinery. Such block might trigger the accumulation of negative supercoiling behind, which would extend only until to the next nucleosome.

In agreement with other studies (Chan et al., 2014; El Hage et al., 2014; Wahba et al., 2016; Ginno et al., 2013; Ginno et al., 2012; Nadel et al., 2015; Sanz et al., 2016; Stork et al., 2016), we observed that R loops accumulated at the 3' end of the ORF, (Figure R8) and we failed to amplify beyond the stop codon, indicating that R loops accumulated only at the gene body in this particular location. Highlighting that DNA-RNA hybrids accumulate in the form of R loops and not as open DNA, no PCR product was detected when converted primer was used for detecting mutations in the template strand. In contrast to what we expected, RNase H in vitro treatment prior to bisulfite was unable to avoid amplification by PCR (Figure R8C) or to reduce the DNA stretch of bisulfite-induced cytosine deaminations (Figure R9D), but it diminished significantly the number of molecules with C to T conversions detected (Figure R9B) therefore indicating that RNAse H treatment was insufficient but that the action of RNAse H is not partial once it starts in a particular molecule.

Thus, in this thesis we characterized that DNA-RNA hybrid length is not a relevant factor in the induction of genomic instability, as even wild-type cells present DNA- RNA hybrids of the same length as R-loop accumulating mutants that do or do not

133 cause genetic instability. This result implied that additional features must be responsible for the inability of the H3∆1-28, H3K9-23A and H4K31Q histone mutants to impair fork progression and cause hyper-recombination, LOH or plasmid loss. Further experiments performed in parallel in the lab demonstrated that these additional features have to do with the potential of DNA-RNA hybrids to cause H3S10 phosphorylation, a mark of chromosome condensation during mitosis that has also been linked to genome instability if accumulated outside of mitosis (Castellano-Pozo et al., 2013) and that is prevented by H3∆1-28, H3K9-23A and H4K31Q histone mutants (Garcia-Pichardo et al., 2017). Strikingly, H3K9-23A suppressed both the hyper-recombination (Garcia-Pichardo et al., 2017) and the replication fork impairments of hpr1∆ (Figure R3C) arguing that DNA-RNA hybrids are not leading to genetic instability unless H3S10 phosphorylation is allowed. The results of the second chapter of this thesis further nailed down the modification that is responsible for this suppression phenotype to lysine 9 on histone H3, since we observed that rtt109∆ leads to R loop accumulation and that the mutation of H3K9 (a Rtt109 target) to alanine suppressed Rad52 foci accumulation of hpr1∆ mutant, whereas no effect was detected when H3K14 and H3K23 (Rtt109 targets) were changed to alanine (Figure R36). We thus propose that the reported inability to phosphorylate H3S10 in H3K9-23A or H3∆1-28 (Garcia-Pichardo et al., 2017) is due to the lack of H3K9 modifiable residue. This establishes a crosstalk between these consecutive residues (H3K9 and S10), indicating that H3K9 modification could be a pre-requisite for H3S10 phosphorylation. In this support, it is reported a connection between these two residues in human cells (Peng et al., 2018; Fischle et al., 2005).

3. Detection and mapping DNA-RNA hybrids by newly developed S9.6- independent methodologies based on cytosine deamination in vitro and in vivo.

Given the success of the sodium bisulfite technique to detect R loops and measure their length, we wondered whether we could take advantage of this tool in yeast to determine R loop frequencies, as previously developed for murine B cells by lift

134 colony hybridization assay (Huang et al., 2006). We estimated R loop frequency at the GCN4 gene to be below 0.2 % in H3WT, H3∆1-28 and hpr1∆ strains, as we analyzed around 500 molecules from each strain but we were not able to detect a single molecule with C to T conversions in any of them (Figure R11). These results are in agreement with the estimation of R loop frequency at the murine Sγ3 locus in 0.17 % molecules (Huang et al., 2006).

The elevated number of molecules that we had to analyze drove us to try a new methodology to study R loop frequency combining bisulfite with real time PCR. The method seemed to work accurately, since we detected more C to T conversions in the rnh1∆ rnh2∆ double mutant compared to WT strain (7 % vs 4 %, respectively) (Figure R13). However, these percentages of C to T conversions were far from the estimated values below 0.2% obtained by lift colony hybridization assay at GCN4 gene (Figure R11) and Lieber´s group (Huang et al., 2006), specially taking into account that the frequency of R loops by bisulfite-qPCR technique is underestimated, as we are not detecting R loops outside the DNA region delimited by the two primers used. This made us conclude that bisulfite-qPCR assay is not a good tool to infer absolute R loop frequencies, but it can be useful to study relative amounts of R loops between samples (Figure R13) as a good alternative to S9.6 antibody-related technics such as DRIP-qPCR, given its multiple reported lack of specificity-related issues (Phillips et al., 2013; Hartono et al., 2018).

As another S9.6-independent alternative tool for R-loop detection, we also studied the potential of AID* to map R loops genome-wide. This tool, developed by Neuberger´s lab (Wang et al., 2009), had been used previously by Cristina Rada´s group to simulate in yeast the pattern of mutations found in human breast cancer cells, where AID/APOBEC protein family is highly expressed (Taylor et al., 2013). We used diploids cells in the experiment to avoid the reduction in fitness costs caused by accumulated mutations, especially in essential genes (Waters and Parry, 1973; Lada et al., 2013). In agreement with this, AID* overexpression did not affect cell viability (Figure R14, Figure R15 and Figure R26A), but mutations

135 were detected as canavanine-resistant colonies (Figure R14, Figure R16 and Figure R26B).

However, the sequencing analysis revealed a low number of mutations, ranging from 50 to 150 per genome and affecting to all four kinds of nucleotides (Figure R17 and Figure R27). The reduced number of mutations detected is probably indicating that AID* is not acting processively in our experimental conditions. Moreover, the strains used in this experiments possess mechanisms to repair cytosine deaminations provoked by AID*. In fact, overexpression of AID* in ung1∆ mutant (deletion of Ung1, uracil-DNA glycosylase which removes uracil from DNA) increased the total number of mutations detected in wild-type cells but at the expense of a reduction in cytosine mutations (Taylor et al., 2013). This implies that the total number of random mutations increases, but that the enhanced mutations after UNG1 deletion is not mediated by AID* overexpression, but to spontaneous cytosine deamination as a consequence of natural process during cell divisions (Lynch et al., 2008). Furthermore, diploid cells lacking UNG1 were very sensitive to the overexpression of AID* (data not shown).

Nevertheless, supporting the specific action of AID* on DNA-RNA hybrids, we observed that mutations at adenine and thymine nucleotides remained at the same frequency between the different strains and conditions, whereas mutations at cytosines and guanines varied (Figure R17 and Figure R27A). Moreover, specific C to T and G to A transitions resulted significantly enhanced in hpr1∆ (Figure R18) as expected from the action of AID, although this effect could not be suppressed by RNase H overexpression (Figure R17, Figure R18 and Figure R27). In addition, AID*-induced mutations in hpr1∆ occurred more likely than expected at DNA-RNA hybrid accumulating regions described for that mutant (Chan et al., 2014), reinforcing the idea that AID* is targeted to R loops. We also observed that there were more mutations than expected by chance in regions prone to form G4- quadruplexes. This is in agreement with mutations being enriched at R loop forming regions, since G4-quadruplexes favor and stabilize R loops (Duquette et al., 2004; De Magis et al., 2019; Kuznetsov et al., 2018; Wanrooij et al., 2012). In

136 agreement, it has been described that AID may be targeted to immunoglobulin loci through G4-quadruplexes where it triggers class switch recombination (Qiao et al., 2017).

Strikingly, mutations fell into either template and non-template strand at the same frequency, in contrast to what we would expect if AID* was acting on ssDNA of R loops which is the non-transcribed. Nevertheless, genes in the yeast genome are very close one from each other, which could lead to the transient opening of both DNA strands from a unique promoter region, leading only one of these genes to R loop formation. This could explain not only the symmetric distribution of mutations in template/non-template strands, but also the fact that RNase H overexpression failed to avoid AID*-induced mutations.

In contrast to hpr1∆, other DNA-RNA hybrid accumulating mutants such as the histone mutants and rtt109∆ presented the same or less AID*-induced mutations respect to their corresponding wild types. One possible explanation for this result is that such mutations could lead to a modified chromatin which could impede AID* access to ssDNA of R loops. In fact, the lack of acetyl-transferase activity in rtt109∆ mutant might strengthen DNA-histone interactions maybe making the DNA less accessible to AID. Another possibility is that AID* action depends on histone marks which are lost in histone and rtt109∆ mutants. Supporting this idea, it is known that AID action can be affected by histone marks as demonstrated by the fact that its expression in two different cell types induced mutations at different loci associated with histone marks linked to active enhancers (H3K27Ac) and transcription (H3K36me3) (Wang et al., 2014). Notwithstanding, we did not observe that mutations were enriched at high expressed genes or genes with high GC content (Figure R19 and Figure R28). Taking into account that histone mutations can alter the pattern of gene expression, as it happens in hpr1∆ mutant (Gomez- Gonzalez et al., 2011) and that the mRNA profile used as a reference corresponds to a WT strain (Beyer et al., 2004), we believe that RNA-seq analysis from all mutants would be required in order to properly correlate mutations with the expression levels of genes.

137

In summary, we conclude that despite AID* overexpression induces mutations genome-wide, the low number of mutations induced make this tool still insufficient to map R loops genome-wide

4. A model to explain the different phenotypes of R loop-accumulating histone and hpr1∆ mutants.

Altogether, the data from the first chapter of this thesis drive us to propose a model (Figure D1) in which R loops of similar sizes are formed in all strains tested, including WT cells. However, the frequency of such structures is enhanced in H3K14A, H3K23A and hpr1∆ mutants. Genome instability observed in the absence of Hpr1 might be caused firstly by an increase in the frequency of R loop formation and secondly by an increase in the chromatin compaction mediated by H3S10 phosphorylation, what would create a more condensed chromatin difficult to replicate and capable to block further progression of the replication fork. However, the high frequency of R loop formation is not enough to trigger genome instability in when H3K9 is mutated to A. This supports the fact that the previously reported histone mutants prevented H3S10 phosphorylation and were also able to suppress hpr1∆ genome instability phenotype, probably due to the abolishment or a reduction of some unidentified yet modification on H3K9.

138

Figure D1. Model to explain the different phenotypes associated to R formation in histone and hpr1∆ mutants. R loops of similar sizes are formed in WT, H3K14A, H3K23A and hpr1∆ strains, although their frequency is higher in the last three strains. R loop accumulation in hpr1∆ cells triggers H3S10-P, impeding the progression of the replication fork and leading to genome instability. H3K9A prevents H3S10-P.

139

5. R loop or related phenotypes not detected in yeast chromatin remodeler mutants.

None of the chromatin remodeler mutants tested in this thesis led to a detectable accumulation of DNA-RNA hybrids by S9.6 immunoflorescence on chromosome spreads (Figure R20), or to AID susceptibility on Rad52 foci formation, recombination and plasmid loss (Figure R21 and Figure R22). Moreover, it is described in the literature that none of the chromatin remodeler mutants tested displays sensitivity to replicative stress, except for snf2∆ mutant, which presents a growth defect in the presence of hydroxyurea (Minard, Lin and Schultz, 2011). However, this phenotype could arise from an additive effect between HU drug and the repression of the RNR3 gene, involved in dNTPs synthesis, that has been reported in the absence of Snf2 (Minard, Lin and Schultz, 2011). In addition, it has also been described that Snf2 protein is required to complete activation of Mec1 kinase and proper activation of DNA damage response (Kapoor et al., 2015), which may aggravate the consequences of HU presence without implying any replication defect related to DNA-RNA hybrids. In contrast to our results in which recombination levels in isw1∆ occurred at the same frequency than WT strain (Figure R22A), a 3-fold increase in hyper-recombination has been reported for isw1∆ (Babour et al. (2016) that could be explained by differences in the background. It has also been reported that Isw1 is involved in quality control of mRNP along with Rrp6, sequestering malformed mRNP onto chromatin to be eliminated by the nuclear exosome (Babour et al., 2016). Therefore, it is possible that DNA-RNA hybrids are prevented rather that enhanced in isw1∆ cells, where release of the mRNA to the cytoplasm could avoid its hybridization with its DNA template. However, we detected neither higher nor lower levels of DNA-RNA hybrids in isw1∆ compared to WT cells (Figure R20). Our results so far indicate that the deletion of one single chromatin remodeler does not lead to R loop dependent genome instability.

Parallel studies in the lab have shown that the single depletion of the ATPase subunits of human chromatin remodelers, such as hBRG1, hINO80 and hSNF2

140 leads to an accumulation of DNA-RNA hybrids, as well as an increase in RNase H- sensitive γ-H2AX foci formation (Bayona, A. unpublished). Despite the lack of any observable phenotype after the deletion of each of the single genes, we cannot rule out a similar conserved role of SNF2, ISW1 and ISW2 chromatin remodelers in R loop homeostasis in yeast since this might be masked by redundant roles between the different chromatin remodelers in yeast. In this regard, genetic interaction have been observed between Isw1, Isw2 and Chd1 (Tsukiyama et al., 1999) and Snf2 and Isw1 (Erkina et al., 2010).

6. A screening among histone modifiers identifies Rtt109 as a new factor leading to DNA-RNA hybrid accumulation.

In this thesis, we have performed a screening using mutants belonging to four different groups of histone modifiers: HATs, HDACs, HMTs and HDMs. Strikingly, we did not detect the previously reported DNA-RNA hybrid accumulation in sin3∆ mutant in yeast, which was observed also by chromosome spreads (Wahba et al., 2011; Salas-Armenteros et al., 2017), despite this has also been observed in human cells depleted for the SIN3A homolog complex (Salas-Armenteros et al., 2017). In agreement with our observation, DRIP experiments performed in paralel by other members of the lab in the rpd3∆ mutant of the catalytic subunit of the yeast complex containing Sin3 revealed no detectable DNA-RNA hybrid accumulation. We therefore believe that Sin3 or Rpd3 loss are not sufficient to cause DNA-RNA hybrid accumulation in yeast, in contrast to human cells. Furthermore, from all of the mutants analysed in our screening, only Rtt109 loss led to an accumulation of DNA-RNA hybrids as detected by chromosome spreads (Figure R23) and confirmed later by DRIP (Figure R24) and slightly enhaced reactivity to AID (Figure R25B and C).

141

7. The lack of acetylation of H3K14 and H3K23 contribute to the DNA-RNA hybrid accumulation in rtt109∆.

Rtt109 is an acetyl-transferase and the DNA-RNA hybrid accumulation detected is due to a reduction in the acetylation levels of one or more of its substrates as demonstrated by our results with the rtt109∆ strain complemented with versions of RTT109 containing mutations inside its catalytic domain (Figure R29). It is counterintuitive that loss of acetylation leads to DNA-RNA hybrid formation, since previous results from the lab in human cells supported that an open chromatin state, such as that provided by hyper-acetylation which weakens DNA-nucleosome interactions, favors mRNA hybridization with its DNA template (Salas-Armenteros et al., 2017). However, Rtt109 loss might not lead to a closed chromatin state. In this regard, rtt109∆ displays reduced nucleosome occupancy around ARS consensus sequences during replication (Liu et al., 2017) and we observed that 10 % of nucleosomes lost their positioning (Figure R31B). These results may suggest a more open chromatin state, which could favor DNA-RNA hybrid formation. Nevertheless, we failed to observe a histone occupancy decrease in S phase in rtt109∆ mutant in an R-loop accumulating gene (Figure R31A) arguing that nucleosome occupancy is unlikely related to the detected accumulation of hybrids.

The study of DNA-RNA hybrid accumulation by chromosome spreads in different cell cycle phases (Figure R30) also argued against a possible role of the increase in the genic dose after replication that has been reported for rtt109∆ (Voichek, Bar- Ziv and Barkai, 2016). Moreover, a significant reduction in the global transcription rate has been described in rtt109∆, asf1∆, and H3K56R strains with a set of specific genes whose transcription abruptly boost in early or late S phase (Topal et al., 2019) in agreement with H3K56Ac being a mark of transcription activation at promoter regions (Williams, Truong and Tyler, 2008). Such little amount of highly transcribed genes would not be able to explain all the DNA-RNA hybrid signal detected in S phase.

Although the best-known function of Rtt109 is to acetylate H3K56 during replication-coupled nucleosome assembly (Han et al., 2007), neither H3K56A

142 mutant nor hst3∆ and hst4∆ (which remove H3K56 acetylation), accumulate DNA- RNA hybrids as studied by DRIP and chromosome spreads (Figure R32). Further analysis with strains bearing mutations in lysine residues which have been reported to be in vitro and/or in vivo Rtt109 targets, such as H3K9, H3K14, H3K23 and H3K27 (Dahlin et al., 2015; Cote et al., 2019; Fillingham et al., 2008), revealed that only H3K14A and H3K23A mutants accumulated DNA-RNA hybrids at the same extent to rtt109∆ (Figure R32), and that these residues are responsible of DNA-RNA hybrid accumulation in rtt109∆, since mutants for H3K14 and H3K23 which mimic an acetylated state did not enhance the accumulation of such secondary structures after RTT109 deletion (Figure R33).

It is noteworthy that this effect was observed for either H3K14 or H3K23 acetyl- mimimetic mutations what could be explained by a crosstalk between these two residues. Indeed, it has been recently reported that H3K23 acetylation by MORF complex requires H3K14 pre-acetylation in human cells (Klein et al., 2019).

8. Defective SCR contributes to the DNA-RNA hybrid accumulation of rtt109∆.

We thought that DNA-RNA hybrid accumulation derived from reduced H3K14 and H3K23 acetylation might contribute to the increased DNA damage observed in rtt109∆ cells detected by spontaneous accumulation of Rad52 foci (Figure R25C and Figure R35). In this sense, only the H3K23A presented an increase in Rad52 foci. Given the genetic interactions between RTT109 with H3K14 and H3K23 (Figure R33) and the Rad52 foci results from their corresponding mutants (Figure R25, Figure R35 and Figure R36), it is more than likely that part of the DNA damage observed in rtt109∆ mutant arises from a reduction in the acetylation levels of H3K14 and H3K23 residues. Indeed, RNase H overexpression also reduced partially Rad52 foci accumulation in rtt109∆ mutant (Figure R35). By contrast, RNase H overexpression was not sufficient to suppress the rtt109∆ growth defect under replication stress conditions (Figure R34), or the defect in the repair of replication-born DSBs using the sister chromatid as a template (SCR)

143

(Figure R39), indicating that DNA-RNA hybrids are not responsible of these other phenotypes.

The study of the phenotypes of the mutants for the two lysine residues that might be contributing to DNA-RNA hybrid accumulation in rtt109∆ (H3K14 and H3K23) showed that only H3K14A led to an SCR defect (Figure R38 and Figure R40) and this defect was as strong as that observed for rtt109∆ (Figure R38 and Figure R39) or previously reported for H3K56A (Munoz-Galvan et al., 2013). Again, such SCR defect was not caused by DNA-RNA hybrids (Figure R39 and Figure R40). Altogether, we conclude that all Rtt109, H3K14 and H3K56 acetylation are required for proper DSB repair by SCR.

Given the frequent crosstalk between different histone marks, one possibility is that H3K14 acetylation is needed in order to acetylate H3K56. Indeed, supporting this possibility, Asf1-mediated acetylation driven by Rtt109 increases the H3K56 acetylation only in H3-H4 dimers with pre-acetylated H3K14, as suggested by in vitro assays (Cote et al., 2019). Thus, the lack of H3K14 acetylation in H3K14A mutant would explain its strong SCR defect because H3K56 acetylation would not be reduced. Nevertheless, in S. pombe, H3K14 acetylation is required also to activate the DNA damage response through RSC complex recruitment, which ultimately creates a more open chromatin accessible to DNA damage repair factors, facilitating the repair (Wang et al., 2012). Thus, the strong SCR defect and the normal levels of Rad52 foci in H3K14A mutant could be sorted out by a defect in DNA damage checkpoint activation along with reduced levels of H3K56 acetylation.

Thus, Rtt109 loss leads to an increase in DNA-RNA hybrid accumulation, which has little or no effect on DNA damage. We propose that the DNA damage accumulated in rtt109∆ arises from spontaneous DSBs, which cannot be properly repaired due to the defective SCR, and not from DNA-RNA hybrids. In support of this, it is described that Rtt109 loss provokes an increase in H2A phosphorylation and the overexpression of RNase H failed to suppress this phenotype (Figure R42), the SCR defect (Figure R39) or the genotoxic sensitivity (Figure R34). These

144 phenotypes deviate from those observed for other DNA-RNA hybrid accumulating strains such as hpr1∆ mutant, in which DNA damage comes from DNA-RNA hybrids (see Table D1).

H3K14A H3K23A H3K56A rtt109∆ hpr1∆

Genotoxic Strong6 Slight No1 No1 Strong6 sensitivity (no RNH) (RNH)7 Rad52 foci No Yes Yes Yes Yes2 accumulation (RNH) (RNH) (RNH) (RNH)8 Yes Hyper-rec. No No Yes3 Yes3 (RNH)9 Decreased Yes H2A-P - - No11 in HU4 (no RNH) YES Yes SCR defect No YES5 No (no RNH) (no RNH) SL with rad52∆ - - - Yes10 No7 Table D1. Comparison of phenotypes described for H3K14A, H3K23A, H3K56A rtt109∆ and hpr1∆ mutants. RNH: reduced after RNase H overexpression. NO RNH: not reduced after RNase H overexpression. -: not tested. SL: synthetic lethality. 1 Buehl et al. (2018). 2. Clemente-Ruiz, Gonzalez-Prieto and Prado (2011); Munoz-Galvan et al. (2013); Han et al. (2007) 3 Clemente-Ruiz, Gonzalez-Prieto and Prado (2011). 4 Wang et al. (2012). 5 Munoz-Galvan et al. (2013). 6 Fillingham et al. (2008); Han et al. (2007). 7 Gomez-Gonzalez, Felipe-Abrio and Aguilera (2009). 8 Lafuente-Barquero et al. (2017);Wellinger, Prado and Aguilera (2006). 9 Huertas and Aguilera (2003). 10 Jessulat et al. (2008). 11 Gómez-Gonzalez, unpublished).

In fact, RNase H overexpression in hpr1∆ abolishes Rad52 foci formation, and the slight genotoxic sensitivity to replication stress agents, and also reduces the strong hyper-recombination. In addition, hpr1∆ mutant does not show any SCR defect (Ortega-Moreno, P., unpublished) and its viability depends on the S phase checkpoint (Gomez-Gonzalez, Felipe-Abrio and Aguilera, 2009), in contrast to rtt109∆ mutant whose survival depends on the HR machinery, as indicated by a genetic interaction between RTT109 and RAD52 genes (Jessulat et al., 2008). These results suggest that both DNA-RNA hybrid-accumulating mutants (hpr1∆ and rtt109∆) lead to different sources of DNA damage. Consistently, deletion of

145

RTT109 and HPR1 at the same time led to a synergistic growth defect compared with the corresponding single mutants (Figure R43B).

9. A model to explain the DNA-RNA hybrid accumulation of rtt109∆.

In summary, we have discovered rtt109∆ mutant accumulates DNA-RNA hybrids and we propose that this is the consequence of its inability to acetylate several substrates, including H3K14 and H3K23. Given the role of Rtt109 in the repair of spontaneous DNA damage during replication, the persistence of unrepaired DNA breaks in rtt109∆ leads to a further DNA-RNA hybrid accumulation. Altogether, the results from the second chapter of this thesis prompt us to propose a model (Figure D2) in which loss of Rtt109 causes, on one hand an increase in DNA-RNA hybrids mediated by a reduction in the acetylation levels of H3K23 alone or in combination with reduced H3K14 acetylation, and, on the other hand, an accumulation of unrepaired DSBs due to defective SCR. The SCR defect in the absence of Rtt109 would arise from the total abolishment of H3K56 acetylation given that Rtt109 is the only HAT able to acetylate that residue, together with the reduction of H3K14 acetylation, which can be performed by either Rtt109 or by other HATs such as Gcn5. Unrepaired and persistent DSBs would favor DNA-RNA hybrid formation, consistent with previous studies in which it has been demonstrated that DNA nicks an DSBs lead to DNA-RNA hybrid formation (Roy et al., 2010; Sordet et al., 2009; Britton et al., 2014; Li, Monckton and Godbout, 2008; Ohle et al., 2016).

146

Figure D2. Model of R loop-independent genome instability associated to Rtt109 loss. In wt cells, DSBs prolong H3K56 acetylation and along with H3K14 acetylation promote DSB repair through sister chromatid recombination (SCR). In addition, H3K14 acetylation stimulates H3K23 acetylation by Rtt109, which somehow impedes R loop formation. In Rtt109 depleted cells, H3K56 acetylation is abolished, which translates onto DSB repair defect by SCR, whereas H3K14 and H3K23 acetylation levels are reduced. R loop formation is promoted by reduced levels of H3K23Ac and/or H3K14Ac. Consequently, unrepaired and persistent DSBs accumulate in the absence of Rtt109, favoring further R loop formation.

147

148

Conclusions

149

150

Conclusions

1. The DNA-RNA hybrid accumulation in H3∆1-28, H3K9-23A and H4K31Q histone mutations does not lead to LOH or plasmid loss, in contrast to previously reported DNA-RNA hybrid accumulating mutants, such as hpr1∆.

2. The H3K9-23A mutation suppresses the replication fork progression impairment of hpr1∆. We establish that H3K9A, and not H3K14A or H3K23A, is the mutation responsible of the suppression of the genetic instability associated with DNA-RNA hybrids.

3. The lack of major alterations in the nucleosome positioning pattern in R loop-accumulating mutants indicates that changes in nucleosome positioning are neither the cause nor the consequence of DNA-RNA hybrids.

4. The length of R loops, as determined by the bisulfite-modification assay at the GCN4 gene, ranges from 38 bp to 513 bp with an average of 142 bp. Strikingly, R loop length is similar in wild-type and several R-loop accumulating strains, regardless of whether they produce genome instability or not. This indicates that DNA-RNA hybrid length is not relevant for the generation of genetic instability.

5. Bisulfite-modification followed by qPCR can be used as a new tool alternative to S9.6-based methodologies such as DRIP, to detect relative variations in DNA-RNA hybrids at specific genomics locations. Using this method, we estimate that the frequency of DNA-RNA hybrids is low, since we found no positive R loop hits at the GCN4 gene among 500 yeast colonies analyzed.

151

6. Overexpression of a hyper-mutator AID induces mutations genome-wide, preferentially in cytosines of large genes that accumulate R loops and G4- quadruplexes. However, the low number of induced mutations makes this tool insufficient to map R loops genome-wide.

7. The loss of the Snf2, Isw1 or Isw2 catalytic subunits of the SWI/SNF and ISWI family of chromatin remodelers does not enhance R loop accumulation or R loop-dependent genome instability.

8. The loss of the Rtt109 acetyl-transferase leads to an accumulation of DNA- RNA hybrids in all cell cycle phases. Rtt109 prevents DNA-RNA hybrid accumulation through its catalytic activity but not through its reported role of H3K56 acetylation.

9. H3K14 and H3K23 acetylation residues are relevant for Rtt109 role preventing the accumulation of DNA-RNA hybrids, since H3K14A and H3K23A mutations increase DNA-RNA hybrids and H3K14Q and H3K23Q acetyl-mimetic mutants prevent hybrid accumulation in the absence of Rtt109.

10. We favor a model in which the absence of Rtt109 leads to a reduction in H3K14 and H3K23 acetylation levels and whereby to an increase in DNA- RNA hybrid accumulation. In addition, persistent and unrepaired DSBs create an environment favorable to further DNA-RNA hybrid formation.

152

Materials and Methods

153

154

Yeast strains, plasmids, primers and antibodies.

Table M1. Yeast strains used in this thesis.

Yeast Genotype Source strain H3WT MATa his3Δ200 leu2Δ0 lys2Δ0 trp1Δ63 ura3Δ0 (Dai et al., met15Δ0 can1::MFA1pr-HIS3 hht1-hhf1::NatMX4 hht2- 2008) hhf2::[H3]-URA3 H3∆1-28 MATa his3Δ200 leu2Δ0 lys2Δ0 trp1Δ63 ura3Δ0 (Dai et al., met15Δ0 can1::MFA1pr-HIS3 hht1-hhf1::NatMX4 hht2- 2008) hhf2::[∆1-28]-URA3 H3K9-23A MATa his3Δ200 leu2Δ0 lys2Δ0 trp1Δ63 ura3Δ0 (Dai et al., met15Δ0 can1::MFA1pr-HIS3 hht1-hhf1::NatMX4 hht2- 2008) hhf2::[K9-23A]-URA3 H4WT MATa his3Δ200 leu2Δ0 lys2Δ0 trp1Δ63 ura3Δ0 (Dai et al., met15Δ0 can1::MFA1pr-HIS3 hht1-hhf1::NatMX4 hht2- 2008) hhf2::[H4]-URA3 H4K31Q MATa his3Δ200 leu2Δ0 lys2Δ0 trp1Δ63 ura3Δ0 (Dai et al., met15Δ0 can1::MFA1pr-HIS3 hht1-hhf1::NatMX4 hht2- 2008) hhf2::[K31Q]-URA3 H3hpr1Δ MATa his3Δ200 leu2Δ0 lys2Δ0 trp1Δ63 ura3Δ0 This study met15Δ0 can1::MFA1pr-HIS3 hht1-hhf1::NatMX4 hht2- hhf2::[H3]-URA3 hpr1Δkan BY4743 MATa/MATα his3∆/ his3∆ leu2∆0/ leu2∆0 This study met15∆/MET15 lys2∆0/LYS2 ura3∆0/ ura3∆0 YDIH-4B MATa/Matα his3Δ200/his3 leu2Δ0/ leu2Δ0 lys2Δ0/ (Garcia- lys2Δ0 trp1Δ63/ trp1Δ63 ura3Δ0/ura3Δ0 met15Δ0/ Pichardo et met15Δ0 can1::MFA1pr-HIS3/CAN1 hht1- al., 2017) hhf1::NatMX4/ hht1-hhf1::NatMX4 hht2hhf2::[H3]- URA3/ HHT2-HHF2 YDI28-11C MATa/Matα his3Δ200/his3 leu2Δ0/ leu2Δ0 lys2Δ0/ (Garcia- lys2Δ0 trp1Δ63/ TRP1 ura3Δ0/ura3Δ0 met15Δ0/ Pichardo et met15Δ0 can1::MFA1pr-HIS3/ CAN1 hht1- al., 2017) hhf1::NatMX4/ hht1-hhf1::NatMX4 hht2hhf2::[∆1-28]- URA3/hht2-hhf2::[∆1-28]-URA3 YDIK9-2D MATa/Matα his3Δ200/his3 leu2Δ0/ leu2Δ0 lys2Δ0/ (Garcia- lys2Δ0 trp1Δ63/ TRP1 ura3Δ0/ura3Δ0 met15Δ0/ Pichardo et met15Δ0 can1::MFA1pr-HIS3/ can1::MFA1pr-HIS3 al., 2017) hht1-hhf1::NatMX4/ hht1hhf1::NatMX4 hht2-hhf2::[K9- 23A]-URA3/ hht2-hhf2::[K9-23A]-URA3

155

Yeast Genotype Source strain YDIK31- MATa/Matα his3Δ200/his3 leu2Δ0/ leu2Δ0 (Garcia-Pichardo 9A lys2Δ0/ lys2Δ0 trp1Δ63/ TRP1 ura3Δ0/ura3Δ0 et al., 2017) met15Δ0/met15Δ0 can1::MFA1pr-HIS3/ can1::MFA1pr-HIS3 hht1-hhf1::NatMX4/ hht1hhf1::NatMX4 hht2-hhf2::[K31Q]- URA3/hht2-hhf2::[K31Q]-URA3 YDIH-18B MATa/Matα his3/his3 leu2Δ0/ leu2Δ0 lys2Δ0/ (Garcia-Pichardo lys2Δ0 ura3/ura3 met15Δ0/ MET15 CAN1/ et al., 2017) CAN1 hht1-hhf1::NatMX4/ hht1-hhf1::NatMX4 hpr1Δ::TRP1/hpr1Δ::TRP1 W303-1A MATa ade2-1 can1-100 his3-11,15 leu2-3,112 R. Rothstein trp1-1 ura3-1 U678-4C MATa ade2-1 can1-100 his3-11,15 leu2-3,112 Piruat and trp1-1 ura3-1hpr1Δ::HIS3 Aguilera, 1998 DD379 W303-1A rad24Δ::TRP1 Kanellis et al., 2003 WHR24- W303-1A hpr1Δ::HIS3 rad24Δ::TRP1 (Gomez-Gonzalez, 3B Felipe-Abrio and Aguilera, 2009) H3HP1 MATa his3Δ200 leu2Δ0 lys2Δ0 trp1Δ63 ura3Δ0 This study met15Δ0 can1::MFA1pr-HIS3 hht1- hhf1::NatMX4 hht2-hhf2::[H3]-URA3 hpr1Δ::Hyg H3RHP MATa his3Δ200 leu2Δ0 lys2Δ0 trp1Δ63 ura3Δ0 This study met15Δ0 can1::MFA1pr-HIS3 hht1- hhf1::NatMX4 hht2-hhf2::[H3]-URA3 hpr1Δ::Hyg rad24Δ::Hyg K923H MATa his3Δ200 leu2Δ0 lys2Δ0 trp1Δ63 ura3Δ0 This study met15Δ0 can1::MFA1pr-HIS3 hht1- hhf1::NatMX4 hht2-hhf2::[K9-23A]-URA3 hpr1Δ::Hyg K923HR MATa his3Δ200 leu2Δ0 lys2Δ0 trp1Δ63 ura3Δ0 This study met15Δ0 can1::MFA1pr-HIS3 hht1- hhf1::NatMX4 hht2-hhf2::[K9-23A]-URA3 hpr1Δ::Hyg rad24Δ::Kan BY4741 MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 EUROSCARF Ybp249 MATa ade2-1 can1-100 his3-11,15 leu2-3,112 B. Pardo trp1-1 ura3-1 bar1Δ RAD5 WH101R5 MATa ade2-1 can1-100 his3-11,15 leu2-3,112 This study trp1-1 ura3-1 bar1Δ RAD5 hpr1-101 (hpr1- L586P) YMFT1 MATa ade2-1 can1-100 his3-11,15 leu2-3,112 This study trp1-1 ura3-1 bar1Δ RAD5 mft1Δ::Kan

156

Yeast Genotype Source strain THP1 MATa ade2-1 can1-100 his3-11,15 leu2-3,112 trp1- F. Fabre 1 ura3-1 bar1Δ RAD5 thp1Δ::Kan WEIII-36 MATa ade2-1 can1-100 his3-11,15 leu2-3,112 trp1- (Herrera- 1 ura3-1 bar1Δ pob3-7 Moyano et al., 2014) XEI-13 MATa ade2-1 can1-100 his3-11,15 leu2-3,112 lys2 (Herrera- met15 trp1-1 ura3-1 spt16-11 Moyano et al., 2014) RNH-R MATa ade2-1 can1-100 his3-11,15 leu2-3,112 lys2 M. San Martín met15 trp1-1 ura3-1 rnh1Δ::Kan rnh201Δ::Kan YDIH-CAN MATa/Matα his3Δ200/his3 leu2Δ0/ leu2Δ0 lys2Δ0/ This study lys2Δ0 trp1Δ63/ trp1Δ63 ura3Δ0/ura3Δ0 met15Δ0/ met15Δ0 CAN1/CAN1 hht1-hhf1::NatMX4/ hht1- hhf1::NatMX4 hht2hhf2::[H3]-URA3/ HHT2-HHF2 YDI28-CAN MATa/Matα his3Δ200/his3 leu2Δ0/ leu2Δ0 lys2Δ0/ This study lys2Δ0 trp1Δ63/ TRP1 ura3Δ0/ura3Δ0 met15Δ0/ met15Δ0 CAN1/ CAN1 hht1-hhf1::NatMX4/ hht1- hhf1::NatMX4 hht2hhf2::[∆1-28]-URA3/hht2- hhf2::[∆1-28]-URA3 YDIK9-CAN MATa/Matα his3Δ200/his3 leu2Δ0/ leu2Δ0 lys2Δ0/ This study lys2Δ0 trp1Δ63/ TRP1 ura3Δ0/ura3Δ0 met15Δ0/ met15Δ0 CAN1/CAN1 hht1-hhf1::NatMX4/ hht1hhf1::NatMX4 hht2-hhf2::[K9-23A]-URA3/ hht2-hhf2::[K9-23A]-URA3 YDIK31- MATa/Matα his3Δ200/his3 leu2Δ0/ leu2Δ0 lys2Δ0/ This study CAN lys2Δ0 trp1Δ63/ TRP1 ura3Δ0/ura3Δ0 met15Δ0/met15Δ0 CAN1/CAN1 hht1- hhf1::NatMX4/ hht1hhf1::NatMX4 hht2- hhf2::[K31Q]-URA3/hht2-hhf2::[K31Q]-URA3 YSNF2 MATa ade2-1 can1-100 his3-11,15 leu2-3,112 trp1- This study 1 ura3-1 bar1Δ RAD5 snf2Δ::Kan YISW1 MATa ade2-1 can1-100 his3-11,15 leu2-3,112 trp1- This study 1 ura3-1 bar1Δ RAD5 isw1Δ::Kan YISW2 MATa ade2-1 can1-100 his3-11,15 leu2-3,112 trp1- This study 1 ura3-1 bar1Δ RAD5 isw2Δ::Kan YPL001W MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 hat1Δ::Kan EUROSCARF YEL056W MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 hat2Δ::Kan EUROSCARF YDR448W MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 ada2Δ::Kan EUROSCARF YGR252W MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 gcn5Δ::Kan EUROSCARF YBR081C MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 spt7Δ::Kan EUROSCARF YLR055C MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 spt8Δ::Kan EUROSCARF YLL002W MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 EUROSCARF rtt109Δ::Kan

157

Yeast Genotype Source strain YDR359C MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 EUROSCARF eaf1Δ::Kan YOR023C MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 EUROSCARF ahc1Δ::Kan YBL052C MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 EUROSCARF sas3Δ::Kan YNL330C MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 EUROSCARF rpd3Δ::Kan YOL004W MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 EUROSCARF sin3Δ::Kan YMR263W MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 EUROSCARF sap30Δ::Kan YMR075W MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 EUROSCARF rco1Δ::Kan YTOP1 MATa ade2-1 can1-100 his3-11,15 leu2-3,112 This study trp1-1 ura3-1 bar1Δ RAD5 top1Δ::Hyg SEN-R MATa ade2-1 can1-100 his3-11,15 leu2-3,112 M. San Martín lys2 met15 trp1-1 ura3-1 sen1-1 YNL021W MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 EUROSCARF hda1Δ::Kan YPR068C MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 EUROSCARF hos1Δ::Kan YKR029C MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 EUROSCARF set3Δ::Kan YBR175W MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 EUROSCARF swd3Δ::Kan YJL168C MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 EUROSCARF set2Δ::Kan YDR440W MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 EUROSCARF gis1Δ::Kan YDR096W MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 EUROSCARF dot1Δ::Kan YER169W MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 EUROSCARF rph1Δ::Kan YER051W MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 EUROSCARF jhd1Δ::Kan YJR119C MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 EUROSCARF jhd2Δ::Kan YDR138W MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 EUROSCARF hpr1Δ::Kan YOR025W MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 EUROSCARF hst3Δ::Kan

158

Yeast Genotype Source strain YDR191W MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 EUROSCARF hst4Δ::Kan Hst3/4 MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 This study hst3Δ::Kan hst4Δ::Hyg H3Rt MATa his3Δ200 leu2Δ0 lys2Δ0 trp1Δ63 ura3Δ0 This study met15Δ0 can1::MFA1pr-HIS3 hht1-hhf1::NatMX4 hht2-hhf2::[H3]-URA3 rtt109Δ::Kan YJL115W MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 EUROSCARF asf1Δ::Kan BYRtH MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 This study rtt109Δ::Kan hpr1Δ::Hyg H3K9A MATa his3Δ200 leu2Δ0 lys2Δ0 trp1Δ63 ura3Δ0 (Dai et al., 2008) met15Δ0 can1::MFA1pr-HIS3 hht1-hhf1::NatMX4 hht2-hhf2::[H3K9A]-URA3 H3K14A MATa his3Δ200 leu2Δ0 lys2Δ0 trp1Δ63 ura3Δ0 (Dai et al., 2008) met15Δ0 can1::MFA1pr-HIS3 hht1-hhf1::NatMX4 hht2-hhf2::[H3K14A]-URA3 H3K23A MATa his3Δ200 leu2Δ0 lys2Δ0 trp1Δ63 ura3Δ0 (Dai et al., 2008) met15Δ0 can1::MFA1pr-HIS3 hht1-hhf1::NatMX4 hht2-hhf2::[H3K23A]-URA3 H3K27A MATa his3Δ200 leu2Δ0 lys2Δ0 trp1Δ63 ura3Δ0 (Dai et al., 2008) met15Δ0 can1::MFA1pr-HIS3 hht1-hhf1::NatMX4 hht2-hhf2::[H3K27A]-URA3 H3K56A MATa his3Δ200 leu2Δ0 lys2Δ0 trp1Δ63 ura3Δ0 (Dai et al., 2008) met15Δ0 can1::MFA1pr-HIS3 hht1-hhf1::NatMX4 hht2-hhf2::[H3K56A]-URA3 H3K14Q MATa his3Δ200 leu2Δ0 lys2Δ0 trp1Δ63 ura3Δ0 (Dai et al., 2008) met15Δ0 can1::MFA1pr-HIS3 hht1-hhf1::NatMX4 hht2-hhf2::[H3K14Q]-URA3 H3K23Q MATa his3Δ200 leu2Δ0 lys2Δ0 trp1Δ63 ura3Δ0 (Dai et al., 2008) met15Δ0 can1::MFA1pr-HIS3 hht1-hhf1::NatMX4 hht2-hhf2::[H3K23Q]-URA3 H3K14QRt MATa his3Δ200 leu2Δ0 lys2Δ0 trp1Δ63 ura3Δ0 This study met15Δ0 can1::MFA1pr-HIS3 hht1-hhf1::NatMX4 hht2-hhf2::[H3K14Q]-URA3 rtt109Δ::Kan H3K23QRt MATa his3Δ200 leu2Δ0 lys2Δ0 trp1Δ63 ura3Δ0 This study met15Δ0 can1::MFA1pr-HIS3 hht1-hhf1::NatMX4 hht2-hhf2::[H3K23Q]-URA3 rtt109Δ::Kan

159

Table M 2. Plasmids used in this thesis.

Plasmid Description Source pWJ1344 YCp containing RAD52::YFP fusion under (Feng et al., 2007) its own promoter (LEU2 marker) pWJ1314 YCp containing RAD52::YFP fusion under (Alvaro, Lisby and its own promoter (TRP1 marker) Rothstein, 2007) pRS413GAL Ycp pRS413 with the GAL1 promoter and (Gavalda et al., the CYC1-terminator 2016) pRS313- YCp pRS313 containing the GALp::RNH1 (Gavalda et al., GAL::RNH1 fusion 2016) p414GAL Ycp pRS414 with the GAL1 promoter and (Garcia-Rubio et the CYC1-terminator al., 2003) p414GALAID YCp pRS313 containing the GALp::AID (Gomez-Gonzalez fusion and Aguilera, 2007) pLZGAID YCp pRS314 containing LLacZ (Garcia-Pichardo et recombination system and GALp::AID al., 2017) fusion pRS317GAL Ycp pRS317 with the GAL1 promoter and (Garcia-Pichardo et the CYC1-terminator al., 2017) pRS317- YCp pRS317 containing the GALp::RNH1 (Garcia-Pichardo et GAL::RNH1 fusion al., 2017) p317RTT109wt Ycp pRS317 with wild-type RTT109 gene (Han et al., 2007) p317D89A Ycp pRS317 with RTT109 gene mutated (Han et al., 2007) at aspartate 89 to alanine p317D89N Ycp pRS317 with RTT109 gene mutated (Han et al., 2007) at aspartate 89 to asparagine IN YCp pARSGLB-IN in which the LEU2 (Prado and gene was generated in vivo by Aguilera, 2005) homologous recombination between the leu2 repeats OUT YCp pARSGLB-OUT in which the LEU2 (Prado and gene was generated in vivo by Aguilera, 2005) homologous recombination between the leu2 repeats pTINV-HO Ycp pRS414 with TINV and GAL::HO (Ortega, Gomez- Gonzalez and Aguilera, 2019) pRS425AID* YCp pRS425 with the GAL::AID* fusion (Wang et al., 2009) pRS424AID* YCp pRS424 with the GAL::AID* fusion This study

160

Table M 3. Primers used in this thesis.

Primer Sequence 5' to 3' Gene disruption rad24 pFA6a F ATGGATAGTACGAATTTGAACAAACGGCCCTTATTACAATAT AGTCTCAGTTCATTGGGCGGATCCCCGGGTTAATTAAGG rad24 pFA6a R TTAGAGTATTTCCAGATCTGAATCTGAAAGGGACTCACTGAT AACTGGCGCTTTACGCGGATCGATGAATTCGAGCTCG hpr1 pFA6a F ACAATTCAAGAGGCATTAAAACTTGGGCAAAGGAGTAATAC GGATCCCCGGGTTAATTAAGG hpr1 pFA6a R GAATTTCTTATCAGTTTAAAATTTCTATTAAGAGGATAATATC GATGAATTCGAGCTCG top1 TAAAAAAAATCTAAAGGGAGGGCAGAGCTCGAAACTTGAAA CGCGTAAAAATG top1 GCGAACTTGATGCGTGAATGTATTTGCTTCTCCCCTATGCT GCGTTTCTTTGCG mft1 F GGCTGTAAAAAAGGAATCAAAGAACTAAAGCCAAAGGAGAC TAACTCACAATG mft1 R CCTATGTGTCTATATGCCTTTTCTATTTAGTAAGAGCTATGC ATTATACGTGG snf2 F TCGCGACTTTCTGCTATTTTCACGACTTTCGATTAATTATCT GCCCGGATCCCCGGGTTAATTAA snf2 R GTCTACGTATAAACGAATAAGTACTTATATTGCTTTAGGAAG GTAGAATTCGAGCTCGTTTAAAC isw1 F CCATAGCATGATATTACCTGT isw1 R AGTAATGCTGACTCTGTCTAT isw2 F GCTACTCGTCGTTGCTTAGTA isw2 R AATTAGTTAAAGCGGCTCGAC rtt109 F AAGAGAGCATGATAAATCCCCGT rtt109 R ACTAGCCATTCTTTATCCGTCGT

161

Primer Sequence 5' to 3' Gene disruption GCN4 N F TACCAATTGCTATCATGTACCCGT GCN4 N1 R TTGGCAGTAGAAGTGGAAGCA GCN4 C1 R TTAACAATAAAAATAAAAACA GCN4 N2 R GGAGTTGAATCAGTGCTTGACG GCN4 C2 R AAAATTAAATCAATACTTAACA GCN4 N3 R GCATCTTCTAGAACAGGAGTGGG GCN4 C3 R ACATCTTCTAAAACAAAAATAAA GCN4 N4 R AATGAAATCAGCGTTCGCCA GCN4 C4 R AATAAAATCAACATTCACCA GCN4 N5 R GTGTAAAATTCTACTTAAGAA GCN4 C5 R ATATAAAATTCTACTTAAAAA GCN4 N6 R TAATCAGAAGATTATGGGTTC GCN4 C6 R TAATCAAAAAATTATAAATTC GCN4 N4 F TGGCGAACGCTGATTTCATT GCN4 C4 F TAACAAACACTAATTTCATT GCN4 N7 R TATCTAAACCTTAGCGTTTGCATTC qPCR GCN4 F TTGTGCCCGAATCCAGTGA GCN4 R TGGCGGCTTCAGTGTTTCTA TRP F CGGCTTGCAGAGCACAGA TRP R AGCAAGTCAGCATCGGAATCTAG PRD1 F CCGCCTAATTGGTCGTTCAC PRD1 R GTTGTTGATGATTTCGTTGG PDR5 F GTCAGAGGCTATATTTCACTGGAGA PDR5 R TACGTCTTGTTTCGGCCTTAATC SPF1 F CCCGTGGTAAACCTTTAGAAA SPF1 R ATATGAACGGCAAATTGAGAC 25S F TCAACTTAGAACTGGTACGG

162

Primer Sequence 5' to 3' qPCR 25S R GCTTGGTTGAATTTCTTCAC PDC1 F CCTTGATACGAGCGTAACCATCA PDC1 R GAAGGTATGAGATGGGCTGGTAA GCN4 N F 359 AAAACCTAGAAGACAACTCT Primer Sequence 5' to 3' qPCR GCN4 N3 R GCATCTTCTAGAACAGGAGTGGG GCN4 C3 R ACATCTTCTAAAACAAAAATAAA

Table M4. Primary antibodies used in this thesis.

Antibody Source Epitope Reference Use S9.6 Mouse DNA-RNA Hybridoma cell line DRIP: 3 µg hybrids HB-8730 IF: 1:300 Histone H2A Rabbit H2AS129-P Ab15083 (Abcam) ChIP: 5 µg (phospho S129) WB: 1:1000 Histone H3 Rabbit Histone H3 Ab1791 ChIP: 5 µg WB: 1:2000

Table M5. Secondary antibodies used in this thesis.

Specificity Conjugation Reference Use Rabbit IRDye 800CW LI-COR (925-32211) WB: 1:15000 Mouse IRDye LI-COR (92568074 WB: 1:15000 Mouse Cy3 #115-165003 IF: 1:1000 (Jackson laboratories)

163

Media used in this study.

 LB: 0.5% yeast extract, 1% bacto-tryptone, 1% NaC.  YPAD: 1% yeast extract, 2% bacto-peptone, 2% glucose, 20 mg/L adenine).  SD: 2% glucose, 0.17% yeast nitrogen base (YNB) without amino acids, 0.5% ammonium sulfate.  SC: SD containing aminoacids.  SGal: SC without glucose containing 2% filtered-galactose.  SRaf: SC without glucose containing 2% raffinose.  SPO: 1% potassium acid, 0.1% yeast extract, 0.005% glucose). Solid media also contain 2 % agar.

Techniques

Bacterial transformation. Bacteria was transformed with exogenous DNA plasmid according to Sambrook, Fritsch and Maniatis (1989). 100 ng of plasmid were added to competent E.coli cells. Mix was incubated first on ice during 30 minutes and then at 42 ºC during 2 minutes for heat shock. Finally, mix was incubated 45 minutes at 37ºC and then plated on selective media.

Yeast transformation and gene disruption. Yeast strains were transformed using the lithium acetate protocol (Gietz et al., 1995). Mid-log culture was pelleted and resuspended in the transforming mix (0.1 % lithium acetate, PEG-50, 50 µg salmon sperm and 100 ng of plasmid (or 5 µg of cassette for disruption) and incubated during 30 minutes at 30ºC. Heat shot was carried out placing cells at 42ºC during 20 minutes and then were plated in selective media. Additionally, for gene disruption mix was incubated two hours at 30ºC after heat shock step.

Drop assay. Serial dilutions of mid-log culture were plated in medium supplemented with different concentrations of HU, MMS and/r CPT and then incubated at 30ºC during 164

2 and 3 days. Alternatively, media are not supplemented with drugs but with different carbon sources to switch on or switch off transcription of specific genes under the GAL1 promoter introduced in cells by transformation.

Plasmid loss assay. Mid-log transformant culture was grown in non-selective media during 3.5 hours and then plated into YPD or selective media to count total and plasmid-containing cells, respectively.

Recombination assay. Transformants with the recombination system were plated on glocuose- or galactose-containing media and grown at 30ºC during 3 to 4 days. Known dilutions of six independent colonies were plated on YPD or media lacking leucine to count for total or recombinant cells. Recombination frequencies were calculated as the median value of six independent colonies. The average value of three independent transformants was plotted.

Loss of heterozygocity (LOH) at MAT locus. We tested the ability of homozygous diploid strains to mate with either a MATa or a MATα mating tester. Mating products were detected by auxotrophy complementation (growth in SD media). Diploid cells (heterozygous at MAT) cannot mate due to codominant suppression of haploid-specific cell differentiation pathways (Jensen, Sprague and Herskowitz, 1983; Goutte and Johnson, 1988; Li et al., 1995; Johnson et al., 1998). However, loss of either MATa or MATα allele results in the ability to mate with either MATa or MATα, respectively. Diploid colonies were mated with either F4 or F15 (MATa and MATα, respectively) mating testers in YPAD for 5 hours and replica-plated to SD plates. The total mated products (with either F4 or F15) were scored by growth on SD. The LOH frequency values (mated products per total cells) are the median of between 4 to 10 independent colonies per strain. Loss of heterozygosity in this assay might be due to chromosome loss or gene conversion. Chromosome loss was determined in a total of 59 colonies by calculating the ratio of the DNA at the PRD1 gene (located at the same chromosome as the MAT locus, chromosome III) versus the DNA at

165 the GCN4 gene (chromosome V). DNA was quantified by quantitative PCR using the comparative Ct method (Schmittgen and Livak, 2008). A ratio below one indicated that chromosome III was lost. A third PCR reaction at the TRP1 gene, which is in heterozygosis in these diploids, was used as a reference of a gene with only one copy, since the diploid cells were TRP1+/trp1∆63.

Cell viability assay. Cell culture was grown in SRaf lacking specific requirements for plasmid selection until exponential growth at 30ºC in a shaker and then glucose or galactose 2 % (final concentration) was added. Several time points were taken, plated into YPAD and grown at 30ºC. Total cells of each plate were counted and normalized to the number of cells corresponding to time 0 (before glucose or galactose addition). Average of at least three independent experiments was finally calculated and plotted.

AID*-induced mutation assay. Cell culture was grown in SRaf lacking specific requirements for plasmid selection until exponential growth at 30ºC in a shaker and then glucose or galactose 2 % (final concentration) was added. Several time points were taken and plated in YPAD or SC-complete without arginine and supplemented with canavanine in order to count total or mutated cells, respectively. Mutation frequency of three independent experiments was calculated dividing canavanine-resistant colonies between total cells and average was plotted.

Bioinformatic analyses of AID*-induced mutations. DNA from 1-3 independent canavanine-resistant colonies was extracted following CTAB assay and was then prepared for sequencing. Libraries were constructed following Nextera DNA Flex protocol from Illumina. Sequencing was performed in NextSEq500 Illumina machine and analyses were carried out by Juan Antonio Cordero Varela from IBIS.

Adapter and quality trimming was performed using Cutadapt (v1.15), Trimmomatic (v0.36) and FastQC (v0.11.7). Processed reads were mapped to Saccharomyces

166 cerevisiae S288C reference genome (assembly R64-1-1) with BWA (v0.7.16a) and SAMtools (v1.5), filtering only mapped reads properly paired with a minimum Phred Score of 10. After removing PCR duplicates with Picard (v2.7.3), single-nucleotide variants were called using SAMtools (mapQ > 20, baseQ > 10) and filtered (QUAL > 20) with BCFtools (v1.5).

H3WT genome was sequenced as a reference and the variants present in this sample were excluded from all other samples using intersectBed (Bedtools v2.26.0). Replicates from every condition were merged and then unique variants selected using in-house scripts. A Chi-Square test was performed with coverageBed (Bedtools) and in-house R scripts (v3.6.3) to inspect the enrichment of mutations on g-quadruplexes (liftOver from UCSC Genome Browser was used to translate Saccer1 to Saccer3 coordinates in g4 files from Capra et al. (2010) and in telomeres, ORFs and rRNA genes from Chan et al. (2014). Likewise, enrichment of mutations was also tested in template (in contrast to non-template) chain for every condition. Differences in length, expression and %GC were explored with a Wilcoxon signed-rank test in R.

Rad52 foci detection. Yeast strains were transformed wit pWJ1244 or pWJ1213 plasmids (Lisby, Rothstein and Mortensen, 2001). Resulting transformants were grown in SC or SGal selective media (depending on requirements) until exponential growth and then cells were fixed with 2.5 % formaldehyde in 0.1 M KHPO pH 6.4 during 10 minutes and washed twice in 0.1 M KHPO pH 6.6. Afterwards, cells were washed in 0.1 M KHPO pH 7.4 and permeabilized with 80 % ethanol during 10 minutes, followed by resuspension in 1 µg/ml DAPI for staining nuclei. More than 200 nuclei for each experiment were visualized and counted in a fluorescence microscopy Leica DC 350F microscope to obtain the fraction of nuclei with detectable Rad52 foci.

Cell culture synchronization and FACS analysis. Mid-log culture was incubated in fresh media with two doses of 3 µg/ml of α-factor distributed in 3 hours to arrest cells in G1 phase. Cells were then washed 3 times

167 in fresh media and released from G1 in the presence of pronase to eliminate possible residues of α-factor. For FACS analyses, 1 ml of culture was taken at several time points from the release of G1 to see the progression of cells through the cell cycle and then sequentially fixed in ethanol, washed twice with 1 ml of cold 50 mM sodium citrate pH 7 and treated with 250 mg of RNase A during 1 hour at 50ºC followed by 1 mg of proteinase K at 37ºC 1 hour. Propidium iodide was added at a final concentration of 16 µg/ml to stain the DNA. DNA content was then analyzed in a FACScan cytometer (Becton Dickinson). A total of 100,000 cells were counted. Additionally, to arrest cells in G2 phase, nocodazol at 50 ng/ml was added to the media after G1 released.

Protein extraction and western blot. Proteins were extracted from 5 ml of exponential cultures centrifuged and wash in cold water. Pellet was resuspended in 600 µl of 10 % TCA with 200 µl of glass beads and shaked at maximum speed in a multivortex for 10 minutes at 4 ºC. After centrifugation, pellets were resuspended sequentially in 50 µl of water, 50 µl of 1 M Tris and 100 µl of 2X Laemmli buffer (62.5 mM Tris pH 6.8, 2 % SDS, 25 % glycerol, 0.2 % bromophenol blue, 5 % 2-mercaptoethanol, 1 mM PMSF, 66 µg/ml chymostatin). Samples were incubated at 80ºC during 10 minutes and run in SDS- PAGE electrophoresis in a Mini-Protean 3 Cell with running buffer (25 mM Tris- base pH 8.3, 194 mM glycine, 0.1 % SDS) at 100 V. Proteins were transferred into a pre-activated nytrocelulose membrane using a Criterio Blotter (Biorad) for 2 hours at 400 mA in 1X transfer buffer (3g/L Tris basse pH 8.3, 14.32 g/L glycine). Membranes were blocked using commercial Odyssey Blocking Buffer (LI-COR Biosciences) during 1 hour at room temperature. Primary antibody was prepared to the appropriate dilution in blocking buffer with 0.1 % Tween 20 and incubated for 2 hours at room temperature. Three washes of 10 minutes each were performed with 1X TBS and 0.1 % Tween 20 followed by secondary antibody incubation during 45 minutes. Eventually, membranes were washed 3 times in 1X TBS and 0.1 % Tween 20 and scanned. Image acquisition and quantification was performed in an Odyssey CLx Imager (LI-COR Biosciences).

168

Chromosome spreads. This protocol was performed as described in Chan et al. (2014) with some modifications. 10 ml of Mid-log cultures were pelleted and washed in cold spheroplasting buffer (1.2 M Sorbitol, 0.1 M potassium phosphate, 0.5 MgCl2) and then digested by adding at the same buffer 10 mM DTT and 3 mg/ml Zymolyase 20T, and incubated 10 minutes at 37ºC. Digestion was stopped by adding 2 volumes of solution 2 (0.1 M MES, 1 M sorbitol, 1 mM EDTA, 0.5 mM MgCl2 pH 6.4. Spheroplasts were centrifuged and transferred onto slides where they were lysed with 1 % lipsol and fixed with fixative solution (4 % paraformaldehyde, 3.4 % sucrose). The spreading was carried out using a glass rod and the slides were dried overnight in the extraction hood.

Slides were then washed in 1X PBS during 30 minutes and blocked in blocking solution (5 % BSA, 0.2 % milk in 1X PBS) during 10 minutes in humid chambers. Afterwards, slides were incubated with primary antibody S9.6 diluted 1:300 in blocking buffer at 23ºC during 1 hour. Slides were washed 3 times in 1X PBS and incubated with secondary antibody Cy3 diluted 1:1000 in blocking buffer at room temperature during 1 hour. Finally, slides were mounted with 50 µL of Vectashield (Vector laboratories) with 1X DAPI and sealed with nail polish. More than 150 nuclei were visualized and counted in a fluorescence microscopy Leica DC 350F microscope to obtain the fraction of nuclei with detectable DNA-RNA hybrids.

CTAB DNA extraction. Cells were arrested in G1 with α-factor and released into fresh medium containing 40 mM HU for 30 min prior to DNA extraction. DNA extraction was performed with the cetyltrimethylammonium bromide method, and neutral-neutral 2-D gel electrophoresis was performed as described in Wellinger, Schar and Sogo (2003) with some modifications. 100 ml of the cultures were collected, washed with 5 ml of chilled water and carefully resuspended in 1 ml of spheroplasting buffer (1M sorbitol, 10 mMEDTA pH 8, 0.1% b-mercaptoethanol, 2 mg/ml Zimoliase 20T), and then incubated at 30 ºC 1h under soft agitation. The spheroplasts were washed with 500 µl of cold water and resuspended in 400 µl of cold water. Spheroplasts

169 were lysed by adding 500 µl of solution I (1.4 MNaCl, 100mM Tris-Cl pH 7.6, 25mMEDTApH 8, 2%CTAB). RNA was removed by incubating them 30 min at 50 ºC with 400 mg of RNase A. Proteins were removed by incubating them with 800 mg of Proteinase K overnight at 30 ºC under very soft agitation. After centrifugation, pellet and supernatant were treated separately. The supernatant was extracted with 500 µl (24:1) Cloroform:Isoamyl Alcohol. DNAwas precipitated with two volumes of solution II (50mM Tris-Cl pH 7.6, 10mM EDTA pH 8, 1% CTAB) and resuspended in 250 µl of solution III (1.4 M NaCl,1 mM EDTA pH 8, 10 mM Tris-Cl pH 7.6). The original pellet was resuspended in 400 µl of solution III and incubated 1h at 50 ºC. DNA was extracted with 200 µl (24:1) Cloroform:Isoamyl Alcohol and combined with the DNA obtained from the supernatant. The whole sample was precipitated then with 1 volume of isopropanol at room temperature, washed with 70% ethanol and resuspended in 100 µl 10 mM Tris-Cl pH 8. DNA was prepaired to be used for 2D gel electrophoresis or bisulfite modification assay.2D-gel electrophoresis.

DNA isolated by CTAB method was digested with PvuII and then charged into a 0.4 % agarose gel 1 % TBE without ethidium bromide. Electrophoresis was carried out at constant voltage (50 V). From each lane, a single piece of gel containing 3 to 10 kb fragments was cut and used as the well of the second dimension electrophoresis in a 1 % agarose gel 1 % TBE with ethidium bromide. Electrophoresis run overnight at 4 ºC and onced finished, the gel was treated with 0.25 N HCl 10 minutes, denaturation solution (0.5 M NaOH, 1.5 M NaCl) 30 minutes and neutralization solution (1M AcNH4, 0.02 M NaOH).

DNA was then transferred to a Hybond N (GE Healthcare) membranes by Southern blot and analyzed using a 32P-labeled 0.5-kb SPF1 probe.

Bisulfite modification assay. DNA isolated by CTAB method was digested with NdeI, NotI and XhoI. The bisulfite modification assay was performed essentially as described (Yu et al., 2003). Genomic DNA was diluted in 42 µl of distilled water with 17.5 µl of 20 mM hydroquinone and 460.5 µl of 2.5 M sodium bisulfite (pH 5.2). The mixture was

170 sealed with mineral oil in a 500 µl microcentrifuge tube and incubated for 16 hr at 37 ºC in the dark. Bisulfite-treated DNA was purified with the Wizard SV Gel and PCR Clean-Up System (Promega) according to the manufacturer’s instructions. Purified bisulfite-treated DNA was desulfonated with 0.3 M NaOH at 37 ºC for 15 min. Desulfonated DNA was recovered by ethanol precipitation and resuspended in TE (pH 8.0). Bisulfite-modified DNA was used as a template for PCR with either a pair of native primers, or a native primer paired with a ‘converted’ primer, the sequence of which matched the conversions anticipated owing to deamination of C to U in either the transcribed (TS) or nontranscribed (NTS) strand. PCR bands were purified from agarose gels with the Wizard SV Gel and PCR Clean-Up System (Promega) and cloned in pGEMT-easy. Independent clones were sequenced. Only molecules with more than four consecutive expected C to T changes were considered to determine the R-loop length.

Lift colony hybridization assay. This protocol was performed as described in Huang et al. (2006), with some modifications. Bisulfite-treated DNA was used as template for PCR reaction with two native primers outside the GCN4 ORF. PCR product was then run in a electrophoresis and specific-size bands were purified from agarose gels with the Wizard SV Gel and PCR Clean-Up System (Promega) and cloned in pGEMT-easy. Each clone was picked and restreaked as a line on the surface of a new ampicillin agar plate. Each plate contained aproximately 110 different clones. Plates were incubated at 37ºC overnight and clones were then transferred to a nylon membrane by pressing it against the agar plate. Membrane was placed over the surface of denaturing solution (0.5 MNaOH, 1.5 M NaCl) for 15 min with the bacterial side up and then transferred to neutralization solution (0.5 M Tris pH 7.6, 1.5 M NaCl) for 15 min, followed by a rinse with 2X SSC (0.3 M NaCl, 30 mM trisodium citrate pH 7). The DNA was then fixed on the membrane by UV crosslinking. The membrane was then hybridized using first 32P-labeled converted primers and then with 32P-labeled native primers to count for mutated and total GCN4 molecules, respectively. Membranes were exposed overnight and the signal

171 was read in a FLA-5100 Image Fluorescence Analyzer (Fujifilm) and analyzed with Multigauge 2.0 Analysis Software (Science lab).

DRIP assay. Cells coming from exponential groth in YPD or SC depending of the experiment were pelleted by centrifugation and resuspended in 2.4 ml of spheroplasting buffer (1 M sorbitol, 2 mM Tris–HCl pH 8.0,100 mM EDTA pH 8.0, 0.1% v/v beta- mercapto-ethanol, 2mg/ml zymoliase). Samples were incubated at 30 ºC during 30 minutes. After centrifugation, pellet was resuspended in 1.125 ml of solution I (0.8 mM GuHCl, 30 mM Tris–HCl pH 8.0, 30 mMEDTA pH 8.0, 5% Tween 20, 0.5% Triton X-100) together with 40 µl of 10 mg/ml RNase A and incubated at 37 ºC during 30 minutes. Then, 75 µl of 20 mg/ml proteinase K were added and samples stood at 50 ºC for 1 hour. DNA was purified by chloroform-isoamyl alcohol (24:1) and precipitated with 1 volume of isopropanol. With the help of a glass pasteur pipette, dNA was transferred to a new eppendorf where resuspended in 150 µl of 1X TE (1 mM Tris–HCl pH 7.5, 0.5 mM EDTA pH 8.0) and digested overnight with 50 U of HindIII, EcoRI, BsrGI, XbaI and SspI. Half of the DNA was treated with 3 µl of RNase H (New England BioLabs) overnight 37 ºC as RNaseH control. Both samples were incubated with S9.6 antibody- Dynabeads Protein A (Invitrogen) complexes (previously incubated overnight at 4 ºC) during 2.5 hours at 4º C. Samples were then washed 3 times with 1x binding buffer (10 mM NaPO4pH 7.0, 0.14 M NaCl, 0.05% Triton X-100). DNAwas eluted in 100 mL elution buffer (50 mM Tris pH 8.0, 10 mM EDTA, 0.5 % SDS) treated 45 min with 7 µl of 20 mg/ml proteinase K at 55 ºC and purified with Quiagen DNA purification kit. Real-time quantitative PCR was performed using iTaq universal SYBR Green (Biorad) with a 7500 Real-Time PCR machine (Applied Biosystems).

Chromatin immunoprecipitation (ChIP). ChIP was performed as described (Hecht, Strahl-Bolsinger and Grunstein, 1999) with some modifications. For cell extract preparation, pellets were resuspended in 500 µl of lysis buffer (50 mM HEPES-KOH pH 7.5, 150 mM NaCl, 1 mM EDTA pH 8, 1% Triton X-100, 0.1% sodium deoxycholate, 0.1% SDS) supplemented with

172 protease inhibitors (1x Complete Protease Inhibitor Cocktail (Roche) and 1 mM PMSF).The chromatin was sonicated alternating 1 min high intensity and 1 min rest pulses for 15 min in Bioruptor sonication equipment. Samples were centrifuged for 15 min at 13000 rpm to eliminate cell debris. 20 µl of supernatant were processed as Input and 280 µl were immunoprecipitated. The immunoprecipitation was performed overnight at 4 ºC using Dynabeads Protein A (Invitrogen) previously incubated with the antibody for 4 hr rotating at 4 ºC. Beads were washed and chromatin was eluted in 250 µl elution buffer (50 mM Tris-HCl pH 7.4, 10 mM EDTA, 1% SDS) at 65 ºC for 10 min., treated with 6 µl of 50 mg/ml pronase for 1 hr at 42 ºC and decrosslinked for 6 hr at 65 ºC. Quiagen DNA purification kit was used to clean DNA. Real-time quantitative PCR was performed using iTaq universal SYBR Green (Biorad) with a 7500 Real-Time PCR machine (Applied Biosystems).

SCR intermediates analysis. Analysis of the SCR intermediates was performed as described in Ortega, Gomez- Gonzalez and Aguilera (2019). Cells transformed with pTINV-HO were grown in SRaf lacking tryptophan for plasmid selection and in the presence of 5 µg/ml doxycycline in order to reppress LEU2 gene exppression until exponential growth. Galactose at 2 % was added to induced HO endonuclease overexpression and samples were collected after indicated time points. DNA was extracted by phenol- chloroform-isoamyl alcohol (25:24:1), and precipitated in isopropanol. DNA was resuspended in 200 µl of 1X TE (1 mM Tris–HCl pH 7.5, 0.5 mM EDTA pH 8.0) and digested with SpeI and XhoI overnitgh. DNA was precipitated with isopropanol and samples were electrophoresed using a 0.8 % agarose gel. Finaly, DNA was transferred into a Hybond N (GE Healthcare) membrane and hybridized with 32P- labeled 0.22-kb LEU2 probe. Quantification was performed by calculating the signal of the bands corresponding to DSBs, SCR, or SCR + ICR fragments relative to the total DNA in each line.

173

Bibliografy

174

175

Bibliography

Aguilera, A. and Garcia-Muse, T. (2012) 'R loops: from transcription byproducts to threats to genome stability', Mol Cell, 46(2), pp. 115-24.

Aguilera, A. and Gomez-Gonzalez, B. (2008) 'Genome instability: a mechanistic view of its causes and consequences', Nat Rev Genet, 9(3), pp. 204- 17.

Aguilera, A. and Gomez-Gonzalez, B. (2017) 'DNA-RNA hybrids: the risks of DNA breakage during transcription', Nat Struct Mol Biol, 24(5), pp. 439-443.

Alvaro, D., Lisby, M. and Rothstein, R. (2007) 'Genome-wide analysis of Rad52 foci reveals diverse mechanisms impacting recombination', PLoS Genet, 3(12), pp. e228.

Avemann, K., Knippers, R., Koller, T. and Sogo, J. M. (1988) 'Camptothecin, a specific inhibitor of type I DNA topoisomerase, induces DNA breakage at replication forks', Mol Cell Biol, 8(8), pp. 3026-34.

Azvolinsky, A., Giresi, P. G., Lieb, J. D. and Zakian, V. A. (2009) 'Highly transcribed RNA polymerase II genes are impediments to replication fork progression in Saccharomyces cerevisiae', Mol Cell, 34(6), pp. 722-34.

Babour, A., Shen, Q., Dos-Santos, J., Murray, S., Gay, A., Challal, D., Fasken, M., Palancade, B., Corbett, A., Libri, D., Mellor, J. and Dargemont, C. (2016) 'The Chromatin Remodeler ISW1 Is a Quality Control Factor that Surveys Nuclear mRNP Biogenesis', Cell, 167(5), pp. 1201-1214 e15.

Baharoglu, Z., Lestini, R., Duigou, S. and Michel, B. (2010) 'RNA polymerase mutations that facilitate replication progression in the rep uvrD recF mutant lacking two accessory replicative helicases', Mol Microbiol, 77(2), pp. 324-36.

Baxter, J. (2015) '"Breaking up is hard to do": the formation and resolution of sister chromatid intertwines', J Mol Biol, 427(3), pp. 590-607.

Beletskii, A. and Bhagwat, A. S. (1996) 'Transcription-induced mutations: increase in C to T mutations in the nontranscribed strand during transcription in Escherichia coli', Proc Natl Acad Sci U S A, 93(24), pp. 13919-24.

Beletskii, A. and Bhagwat, A. S. (1998) 'Correlation between transcription and C to T mutations in the non-transcribed DNA strand', Biol Chem, 379(4-5), pp. 549- 51.

Bell, S. P. and Labib, K. (2016) 'Chromosome Duplication in Saccharomyces cerevisiae', Genetics, 203(3), pp. 1027-67.

176

Bermejo, R., Capra, T., Jossen, R., Colosio, A., Frattini, C., Carotenuto, W., Cocito, A., Doksani, Y., Klein, H., Gomez-Gonzalez, B., Aguilera, A., Katou, Y., Shirahige, K. and Foiani, M. (2011) 'The replication checkpoint protects fork stability by releasing transcribed genes from nuclear pores', Cell, 146(2), pp. 233- 46.

Bermejo, R., Lai, M. S. and Foiani, M. (2012) 'Preventing replication stress to maintain genome stability: resolving conflicts between replication and transcription', Mol Cell, 45(6), pp. 710-8.

Beyer, A., Hollunder, J., Nasheuer, H. P. and Wilhelm, T. (2004) 'Post- transcriptional expression regulation in the yeast Saccharomyces cerevisiae on a genomic scale', Mol Cell Proteomics, 3(11), pp. 1083-92.

Bhatia, V., Barroso, S. I., Garcia-Rubio, M. L., Tumini, E., Herrera-Moyano, E. and Aguilera, A. (2014) 'BRCA2 prevents R-loop accumulation and associates with TREX-2 mRNA export factor PCID2', Nature, 511(7509), pp. 362-5.

Boguslawski, S. J., Smith, D. E., Michalak, M. A., Mickelson, K. E., Yehle, C. O., Patterson, W. L. and Carrico, R. J. (1986) 'Characterization of monoclonal antibody to DNA.RNA and its application to immunodetection of hybrids', J Immunol Methods, 89(1), pp. 123-30.

Brahma, S., Udugama, M. I., Kim, J., Hada, A., Bhardwaj, S. K., Hailu, S. G., Lee, T. H. and Bartholomew, B. (2017) 'INO80 exchanges H2A.Z for H2A by translocating on DNA proximal to histone dimers', Nat Commun, 8, pp. 15616.

Britton, S., Dernoncourt, E., Delteil, C., Froment, C., Schiltz, O., Salles, B., Frit, P. and Calsou, P. (2014) 'DNA damage triggers SAF-A and RNA biogenesis factors exclusion from chromatin coupled to R-loops removal', Nucleic Acids Res, 42(14), pp. 9047-62.

Buehl, C. J., Deng, X., Luo, J., Buranasudja, V., Hazbun, T. and Kuo, M. H. (2018) 'A Failsafe for Sensing Chromatid Tension in Mitosis with the Histone H3 Tail in Saccharomyces cerevisiae', Genetics, 208(2), pp. 565-578.

Burma, S., Chen, B. P., Murphy, M., Kurimasa, A. and Chen, D. J. (2001) 'ATM phosphorylates histone H2AX in response to DNA double-strand breaks', J Biol Chem, 276(45), pp. 42462-7.

Capra, J. A., Paeschke, K., Singh, M. and Zakian, V. A. (2010) 'G-quadruplex DNA sequences are evolutionarily conserved and associated with distinct genomic features in Saccharomyces cerevisiae', PLoS Comput Biol, 6(7), pp. e1000861.

Castellano-Pozo, M., Santos-Pereira, J. M., Rondon, A. G., Barroso, S., Andujar, E., Perez-Alegre, M., Garcia-Muse, T. and Aguilera, A. (2013) 'R loops are linked to histone H3 S10 phosphorylation and chromatin condensation', Mol Cell, 52(4), pp. 583-90.

177

Ceccaldi, R., Rondinelli, B. and D'Andrea, A. D. (2016) 'Repair Pathway Choices and Consequences at the Double-Strand Break', Trends Cell Biol, 26(1), pp. 52-64.

Celic, I., Masumoto, H., Griffith, W. P., Meluh, P., Cotter, R. J., Boeke, J. D. and Verreault, A. (2006) 'The sirtuins hst3 and Hst4p preserve genome integrity by controlling histone h3 lysine 56 deacetylation', Curr Biol, 16(13), pp. 1280-9.

Chan, Y. A., Aristizabal, M. J., Lu, P. Y., Luo, Z., Hamza, A., Kobor, M. S., Stirling, P. C. and Hieter, P. (2014) 'Genome-wide profiling of yeast DNA:RNA hybrid prone sites with DRIP-chip', PLoS Genet, 10(4), pp. e1004288.

Chavez, S. and Aguilera, A. (1997) 'The yeast HPR1 gene has a functional role in transcriptional elongation that uncovers a novel source of genome instability', Genes Dev, 11(24), pp. 3459-70.

Chavez, S., Beilharz, T., Rondon, A. G., Erdjument-Bromage, H., Tempst, P., Svejstrup, J. Q., Lithgow, T. and Aguilera, A. (2000) 'A protein complex containing Tho2, Hpr1, Mft1 and a novel protein, Thp2, connects transcription elongation with mitotic recombination in Saccharomyces cerevisiae', EMBO J, 19(21), pp. 5824-34.

Chedin, F. (2016) 'Nascent Connections: R-Loops and Chromatin Patterning', Trends Genet, 32(12), pp. 828-838.

Chen, L., Chen, J. Y., Zhang, X., Gu, Y., Xiao, R., Shao, C., Tang, P., Qian, H., Luo, D., Li, H., Zhou, Y., Zhang, D. E. and Fu, X. D. (2017) 'R-ChIP Using Inactive RNase H Reveals Dynamic Coupling of R-loops with Transcriptional Pausing at Gene Promoters', Mol Cell, 68(4), pp. 745-757 e5.

Chen, P. B., Chen, H. V., Acharya, D., Rando, O. J. and Fazzio, T. G. (2015) 'R loops regulate promoter-proximal chromatin architecture and cellular differentiation', Nat Struct Mol Biol, 22(12), pp. 999-1007.

Chou, D. M., Adamson, B., Dephoure, N. E., Tan, X., Nottke, A. C., Hurov, K. E., Gygi, S. P., Colaiacovo, M. P. and Elledge, S. J. (2010) 'A chromatin localization screen reveals poly (ADP ribose)-regulated recruitment of the repressive polycomb and NuRD complexes to sites of DNA damage', Proc Natl Acad Sci U S A, 107(43), pp. 18475-80.

Ciccia, A. and Elledge, S. J. (2010) 'The DNA damage response: making it safe to play with knives', Mol Cell, 40(2), pp. 179-204.

Cimprich, K. A. and Cortez, D. (2008) 'ATR: an essential regulator of genome integrity', Nat Rev Mol Cell Biol, 9(8), pp. 616-27.

Clapier, C. R. and Cairns, B. R. (2009) 'The biology of chromatin remodeling complexes', Annu Rev Biochem, 78, pp. 273-304.

178

Clausen, A. R., Zhang, S., Burgers, P. M., Lee, M. Y. and Kunkel, T. A. (2013) 'Ribonucleotide incorporation, proofreading and bypass by human DNA polymerase delta', DNA Repair (Amst), 12(2), pp. 121-7.

Clemente-Ruiz, M., Gonzalez-Prieto, R. and Prado, F. (2011) 'Histone H3K56 acetylation, CAF1, and Rtt106 coordinate nucleosome assembly and stability of advancing replication forks', PLoS Genet, 7(11), pp. e1002376.

Colak, D., Zaninovic, N., Cohen, M. S., Rosenwaks, Z., Yang, W. Y., Gerhardt, J., Disney, M. D. and Jaffrey, S. R. (2014) 'Promoter-bound trinucleotide repeat mRNA drives epigenetic silencing in fragile X syndrome', Science, 343(6174), pp. 1002-5.

Conde, F., Refolio, E., Cordon-Preciado, V., Cortes-Ledesma, F., Aragon, L., Aguilera, A. and San-Segundo, P. A. (2009) 'The Dot1 histone methyltransferase and the Rad9 checkpoint adaptor contribute to cohesin-dependent double-strand break repair by sister chromatid recombination in Saccharomyces cerevisiae', Genetics, 182(2), pp. 437-46.

Cortes-Ledesma, F. and Aguilera, A. (2006) 'Double-strand breaks arising by replication through a nick are repaired by cohesin-dependent sister-chromatid exchange', EMBO Rep, 7(9), pp. 919-26.

Cote, J. M., Kuo, Y. M., Henry, R. A., Scherman, H., Krzizike, D. D. and Andrews, A. J. (2019) 'Two factor authentication: Asf1 mediates crosstalk between H3 K14 and K56 acetylation', Nucleic Acids Res, 47(14), pp. 7380-7391.

Dahlin, J. L., Chen, X., Walters, M. A. and Zhang, Z. (2015) 'Histone- modifying enzymes, histone modifications and histone chaperones in nucleosome assembly: Lessons learned from Rtt109 histone acetyltransferases', Crit Rev Biochem Mol Biol, 50(1), pp. 31-53.

Dai, J., Hyland, E. M., Yuan, D. S., Huang, H., Bader, J. S. and Boeke, J. D. (2008) 'Probing nucleosome function: a highly versatile library of synthetic histone H3 and H4 mutants', Cell, 134(6), pp. 1066-78.

Datta, A. and Jinks-Robertson, S. (1995) 'Association of increased spontaneous mutation rates with high levels of transcription in yeast', Science, 268(5217), pp. 1616-9.

De Magis, A., Manzo, S. G., Russo, M., Marinello, J., Morigi, R., Sordet, O. and Capranico, G. (2019) 'DNA damage and genome instability by G-quadruplex ligands are mediated by R loops in human cancer cells', Proc Natl Acad Sci U S A, 116(3), pp. 816-825.

Devbhandari, S., Jiang, J., Kumar, C., Whitehouse, I. and Remus, D. (2017) 'Chromatin Constrains the Initiation and Elongation of DNA Replication', Mol Cell, 65(1), pp. 131-141.

179

Dominguez-Sanchez, M. S., Barroso, S., Gomez-Gonzalez, B., Luna, R. and Aguilera, A. (2011) 'Genome instability and transcription elongation impairment in human cells depleted of THO/TREX', PLoS Genet, 7(12), pp. e1002386.

Downs, J. A., Allard, S., Jobin-Robitaille, O., Javaheri, A., Auger, A., Bouchard, N., Kron, S. J., Jackson, S. P. and Cote, J. (2004) 'Binding of chromatin- modifying activities to phosphorylated histone H2A at DNA damage sites', Mol Cell, 16(6), pp. 979-90.

Downs, J. A., Lowndes, N. F. and Jackson, S. P. (2000) 'A role for Saccharomyces cerevisiae histone H2A in DNA repair', Nature, 408(6815), pp. 1001-4.

Driscoll, R., Hudson, A. and Jackson, S. P. (2007) 'Yeast Rtt109 promotes genome stability by acetylating histone H3 on lysine 56', Science, 315(5812), pp. 649-52.

Drolet, M., Phoenix, P., Menzel, R., Masse, E., Liu, L. F. and Crouch, R. J. (1995) 'Overexpression of RNase H partially complements the growth defect of an Escherichia coli delta topA mutant: R-loop formation is a major problem in the absence of DNA topoisomerase I', Proc Natl Acad Sci U S A, 92(8), pp. 3526-30.

Dumelie, J. G. and Jaffrey, S. R. (2017) 'Defining the location of promoter- associated R-loops at near-nucleotide resolution using bisDRIP-seq', Elife, 6.

Duquette, M. L., Handa, P., Vincent, J. A., Taylor, A. F. and Maizels, N. (2004) 'Intracellular transcription of G-rich DNAs induces formation of G-loops, novel structures containing G4 DNA', Genes Dev, 18(13), pp. 1618-29.

Duquette, M. L., Pham, P., Goodman, M. F. and Maizels, N. (2005) 'AID binds to transcription-induced structures in c-MYC that map to regions associated with translocation and hypermutation', Oncogene, 24(38), pp. 5791-8.

Dutta, A., Babbarwal, V., Fu, J., Brunke-Reese, D., Libert, D. M., Willis, J. and Reese, J. C. (2015) 'Ccr4-Not and TFIIS Function Cooperatively To Rescue Arrested RNA Polymerase II', Mol Cell Biol, 35(11), pp. 1915-25.

El Hage, A., French, S. L., Beyer, A. L. and Tollervey, D. (2010) 'Loss of Topoisomerase I leads to R-loop-mediated transcriptional blocks during ribosomal RNA synthesis', Genes Dev, 24(14), pp. 1546-58.

El Hage, A., Webb, S., Kerr, A. and Tollervey, D. (2014) 'Genome-wide distribution of RNA-DNA hybrids identifies RNase H targets in tRNA genes, retrotransposons and mitochondria', PLoS Genet, 10(10), pp. e1004716.

Erkina, T. Y., Zou, Y., Freeling, S., Vorobyev, V. I. and Erkine, A. M. (2010) 'Functional interplay between chromatin remodeling complexes RSC, SWI/SNF

180 and ISWI in regulation of yeast heat shock genes', Nucleic Acids Res, 38(5), pp. 1441-9.

Feng, Q., During, L., de Mayolo, A. A., Lettier, G., Lisby, M., Erdeniz, N., Mortensen, U. H. and Rothstein, R. (2007) 'Rad52 and Rad59 exhibit both overlapping and distinct functions', DNA Repair (Amst), 6(1), pp. 27-37.

Fillingham, J., Recht, J., Silva, A. C., Suter, B., Emili, A., Stagljar, I., Krogan, N. J., Allis, C. D., Keogh, M. C. and Greenblatt, J. F. (2008) 'Chaperone control of the activity and specificity of the histone H3 acetyltransferase Rtt109', Mol Cell Biol, 28(13), pp. 4342-53.

Fischle, W., Tseng, B. S., Dormann, H. L., Ueberheide, B. M., Garcia, B. A., Shabanowitz, J., Hunt, D. F., Funabiki, H. and Allis, C. D. (2005) 'Regulation of HP1-chromatin binding by histone H3 methylation and phosphorylation', Nature, 438(7071), pp. 1116-22.

Frederico, L. A., Kunkel, T. A. and Shaw, B. R. (1990) 'A sensitive genetic assay for the detection of cytosine deamination: determination of rate constants and the activation energy', Biochemistry, 29(10), pp. 2532-7.

French, S. (1992) 'Consequences of replication fork movement through transcription units in vivo', Science, 258(5086), pp. 1362-5.

Garcia-Benitez, F., Gaillard, H. and Aguilera, A. (2017) 'Physical proximity of chromatin to nuclear pores prevents harmful R loop accumulation contributing to maintain genome stability', Proc Natl Acad Sci U S A, 114(41), pp. 10942-10947.

Garcia-Muse, T. and Aguilera, A. (2019) 'R Loops: From Physiological to Pathological Roles', Cell, 179(3), pp. 604-618.

Garcia-Pichardo, D., Canas, J. C., Garcia-Rubio, M. L., Gomez-Gonzalez, B., Rondon, A. G. and Aguilera, A. (2017) 'Histone Mutants Separate R Loop Formation from Genome Instability Induction', Mol Cell, 66(5), pp. 597-609 e5.

Garcia-Rubio, M., Aguilera, P., Lafuente-Barquero, J., Ruiz, J. F., Simon, M. N., Geli, V., Rondon, A. G. and Aguilera, A. (2018) 'Yra1-bound RNA-DNA hybrids cause orientation-independent transcription-replication collisions and telomere instability', Genes Dev, 32(13-14), pp. 965-977.

Garcia-Rubio, M., Chavez, S., Huertas, P., Tous, C., Jimeno, S., Luna, R. and Aguilera, A. (2008) 'Different physiological relevance of yeast THO/TREX subunits in gene expression and genome integrity', Mol Genet Genomics, 279(2), pp. 123- 32.

Garcia-Rubio, M., Huertas, P., Gonzalez-Barrera, S. and Aguilera, A. (2003) 'Recombinogenic effects of DNA-damaging agents are synergistically increased by

181 transcription in Saccharomyces cerevisiae. New insights into transcription- associated recombination', Genetics, 165(2), pp. 457-66.

Garcia-Rubio, M. L., Perez-Calero, C., Barroso, S. I., Tumini, E., Herrera- Moyano, E., Rosado, I. V. and Aguilera, A. (2015) 'The Fanconi Anemia Pathway Protects Genome Integrity from R-loops', PLoS Genet, 11(11), pp. e1005674.

Gavalda, S., Santos-Pereira, J. M., Garcia-Rubio, M. L., Luna, R. and Aguilera, A. (2016) 'Excess of Yra1 RNA-Binding Factor Causes Transcription- Dependent Genome Instability, Replication Impairment and Telomere Shortening', PLoS Genet, 12(4), pp. e1005966.

Gietz, R. D., Schiestl, R. H., Willems, A. R. and Woods, R. A. (1995) 'Studies on the transformation of intact yeast cells by the LiAc/SS-DNA/PEG procedure', Yeast, 11(4), pp. 355-60.

Gilbert, D. M. (2002) 'Replication timing and transcriptional control: beyond cause and effect', Curr Opin Cell Biol, 14(3), pp. 377-83.

Ginno, P. A., Lim, Y. W., Lott, P. L., Korf, I. and Chedin, F. (2013) 'GC skew at the 5' and 3' ends of human genes links R-loop formation to epigenetic regulation and transcription termination', Genome Res, 23(10), pp. 1590-600.

Ginno, P. A., Lott, P. L., Christensen, H. C., Korf, I. and Chedin, F. (2012) 'R- loop formation is a distinctive characteristic of unmethylated human CpG island promoters', Mol Cell, 45(6), pp. 814-25.

Gomez-Gonzalez, B. and Aguilera, A. (2007) 'Activation-induced cytidine deaminase action is strongly stimulated by mutations of the THO complex', Proc Natl Acad Sci U S A, 104(20), pp. 8409-14.

Gomez-Gonzalez, B. and Aguilera, A. (2009) 'R-loops do not accumulate in transcription-defective hpr1-101 mutants: implications for the functional role of THO/TREX', Nucleic Acids Res, 37(13), pp. 4315-21.

Gomez-Gonzalez, B., Felipe-Abrio, I. and Aguilera, A. (2009) 'The S-phase checkpoint is required to respond to R-loops accumulated in THO mutants', Mol Cell Biol, 29(19), pp. 5203-13.

Gomez-Gonzalez, B., Garcia-Rubio, M., Bermejo, R., Gaillard, H., Shirahige, K., Marin, A., Foiani, M. and Aguilera, A. (2011) 'Genome-wide function of THO/TREX in active genes prevents R-loop-dependent replication obstacles', EMBO J, 30(15), pp. 3106-19.

Gonzalez-Barrera, S., Cortes-Ledesma, F., Wellinger, R. E. and Aguilera, A. (2003) 'Equal sister chromatid exchange is a major mechanism of double-strand break repair in yeast', Mol Cell, 11(6), pp. 1661-71.

182

Gottipati, P., Cassel, T. N., Savolainen, L. and Helleday, T. (2008) 'Transcription-associated recombination is dependent on replication in Mammalian cells', Mol Cell Biol, 28(1), pp. 154-64.

Goutte, C. and Johnson, A. D. (1988) 'a1 protein alters the DNA binding specificity of alpha 2 repressor', Cell, 52(6), pp. 875-82.

Green, P., Ewing, B., Miller, W., Thomas, P. J., Program, N. C. S. and Green, E. D. (2003) 'Transcription-associated mutational asymmetry in mammalian evolution', Nat Genet, 33(4), pp. 514-7.

Groh, M., Lufino, M. M., Wade-Martins, R. and Gromak, N. (2014) 'R-loops associated with triplet repeat expansions promote gene silencing in Friedreich ataxia and fragile X syndrome', PLoS Genet, 10(5), pp. e1004318.

Gros, J., Kumar, C., Lynch, G., Yadav, T., Whitehouse, I. and Remus, D. (2015) 'Post-licensing Specification of Eukaryotic Replication Origins by Facilitated Mcm2-7 Sliding along DNA', Mol Cell, 60(5), pp. 797-807.

Hamperl, S. and Cimprich, K. A. (2014) 'The contribution of co-transcriptional RNA:DNA hybrid structures to DNA damage and genome instability', DNA Repair (Amst), 19, pp. 84-94.

Han, J., Zhou, H., Horazdovsky, B., Zhang, K., Xu, R. M. and Zhang, Z. (2007) 'Rtt109 acetylates histone H3 lysine 56 and functions in DNA replication', Science, 315(5812), pp. 653-5.

Hanawalt, P. C. and Spivak, G. (2008) 'Transcription-coupled DNA repair: two decades of progress and surprises', Nat Rev Mol Cell Biol, 9(12), pp. 958-70.

Hartono, S. R., Malapert, A., Legros, P., Bernard, P., Chedin, F. and Vanoosthuyse, V. (2018) 'The Affinity of the S9.6 Antibody for Double-Stranded RNAs Impacts the Accurate Mapping of R-Loops in Fission Yeast', J Mol Biol, 430(3), pp. 272-284.

Hatchi, E., Skourti-Stathaki, K., Ventz, S., Pinello, L., Yen, A., Kamieniarz- Gdula, K., Dimitrov, S., Pathania, S., McKinney, K. M., Eaton, M. L., Kellis, M., Hill, S. J., Parmigiani, G., Proudfoot, N. J. and Livingston, D. M. (2015) 'BRCA1 recruitment to transcriptional pause sites is required for R-loop-driven DNA damage repair', Mol Cell, 57(4), pp. 636-647.

Hecht, A., Strahl-Bolsinger, S. and Grunstein, M. (1999) 'Mapping DNA interaction sites of chromosomal proteins. Crosslinking studies in yeast', Methods Mol Biol, 119, pp. 469-79.

Herman, R. K. and Dworkin, N. B. (1971) 'Effect of gene induction on the rate of mutagenesis by ICR-191 in Escherichia coli', J Bacteriol, 106(2), pp. 543-50.

183

Herrera-Moyano, E., Mergui, X., Garcia-Rubio, M. L., Barroso, S. and Aguilera, A. (2014) 'The yeast and human FACT chromatin-reorganizing complexes solve R-loop-mediated transcription-replication conflicts', Genes Dev, 28(7), pp. 735-48.

Hill, T. M., Tecklenburg, M. L., Pelletier, A. J. and Kuempel, P. L. (1989) 'tus, the trans-acting gene required for termination of DNA replication in Escherichia coli, encodes a DNA-binding protein', Proc Natl Acad Sci U S A, 86(5), pp. 1593-7.

Howard, J. A., Delmas, S., Ivancic-Bace, I. and Bolt, E. L. (2011) 'Helicase dissociation and annealing of RNA-DNA hybrids by Escherichia coli Cas3 protein', Biochem J, 439(1), pp. 85-95.

Huang, F. T., Yu, K., Balter, B. B., Selsing, E., Oruc, Z., Khamlichi, A. A., Hsieh, C. L. and Lieber, M. R. (2007) 'Sequence dependence of chromosomal R- loops at the immunoglobulin heavy-chain Smu class switch region', Mol Cell Biol, 27(16), pp. 5921-32.

Huang, F. T., Yu, K., Hsieh, C. L. and Lieber, M. R. (2006) 'Downstream boundary of chromosomal R-loops at murine switch regions: implications for the mechanism of class switch recombination', Proc Natl Acad Sci U S A, 103(13), pp. 5030-5.

Huertas, P. and Aguilera, A. (2003) 'Cotranscriptionally formed DNA:RNA hybrids mediate transcription elongation impairment and transcription-associated recombination', Mol Cell, 12(3), pp. 711-21.

Ide, H., Yagi, R., Yamaoka, T. and Kimura, Y. (1993) 'Misincorporation of ribonucleotides by DNA polymerase during in vitro DNA replication', Nucleic Acids Symp Ser, (29), pp. 133-4.

Ide, S., Saka, K. and Kobayashi, T. (2013) 'Rtt109 prevents hyper- amplification of ribosomal RNA genes through histone modification in budding yeast', PLoS Genet, 9(4), pp. e1003410.

Ivessa, A. S., Zhou, J. Q. and Zakian, V. A. (2000) 'The Saccharomyces Pif1p DNA helicase and the highly related Rrm3p have opposite effects on replication fork progression in ribosomal DNA', Cell, 100(4), pp. 479-89.

Jackson, S. P. and Bartek, J. (2009) 'The DNA-damage response in human biology and disease', Nature, 461(7267), pp. 1071-8.

Janssen, A., Colmenares, S. U. and Karpen, G. H. (2018) 'Heterochromatin: Guardian of the Genome', Annu Rev Cell Dev Biol, 34, pp. 265-288.

Jensen, R., Sprague, G. F., Jr. and Herskowitz, I. (1983) 'Regulation of yeast mating-type interconversion: feedback control of HO gene expression by the mating-type locus', Proc Natl Acad Sci U S A, 80(10), pp. 3035-9.

184

Jessulat, M., Alamgir, M., Salsali, H., Greenblatt, J., Xu, J. and Golshani, A. (2008) 'Interacting proteins Rtt109 and Vps75 affect the efficiency of non- homologous end-joining in Saccharomyces cerevisiae', Arch Biochem Biophys, 469(2), pp. 157-64.

Johnson, P. R., Swanson, R., Rakhilina, L. and Hochstrasser, M. (1998) 'Degradation signal masking by heterodimerization of MATalpha2 and MATa1 blocks their mutual destruction by the ubiquitin-proteasome pathway', Cell, 94(2), pp. 217-27.

Johnson, R. D. and Jasin, M. (2000) 'Sister chromatid gene conversion is a prominent double-strand break repair pathway in mammalian cells', EMBO J, 19(13), pp. 3398-407.

Kadyk, L. C. and Hartwell, L. H. (1992) 'Sister chromatids are preferred over homologs as substrates for recombinational repair in Saccharomyces cerevisiae', Genetics, 132(2), pp. 387-402.

Kao, Y. P., Hsieh, W. C., Hung, S. T., Huang, C. W., Lieber, M. R. and Huang, F. T. (2013) 'Detection and characterization of R-loops at the murine immunoglobulin Salpha region', Mol Immunol, 54(2), pp. 208-16.

Kapoor, P., Bao, Y., Xiao, J., Luo, J., Shen, J., Persinger, J., Peng, G., Ranish, J., Bartholomew, B. and Shen, X. (2015) 'Regulation of Mec1 kinase activity by the SWI/SNF chromatin remodeling complex', Genes Dev, 29(6), pp. 591-602.

Kasten, M., Szerlong, H., Erdjument-Bromage, H., Tempst, P., Werner, M. and Cairns, B. R. (2004) 'Tandem bromodomains in the chromatin remodeler RSC recognize acetylated histone H3 Lys14', EMBO J, 23(6), pp. 1348-59.

Kawauchi, J., Mischo, H., Braglia, P., Rondon, A. and Proudfoot, N. J. (2008) 'Budding yeast RNA polymerases I and II employ parallel mechanisms of transcriptional termination', Genes Dev, 22(8), pp. 1082-92.

Keskin, H., Shen, Y., Huang, F., Patel, M., Yang, T., Ashley, K., Mazin, A. V. and Storici, F. (2014) 'Transcript-RNA-templated DNA recombination and repair', Nature, 515(7527), pp. 436-9.

Kim, N. and Jinks-Robertson, S. (2009) 'dUTP incorporation into genomic DNA is linked to transcription in yeast', Nature, 459(7250), pp. 1150-3.

Klein, B. J., Jang, S. M., Lachance, C., Mi, W., Lyu, J., Sakuraba, S., Krajewski, K., Wang, W. W., Sidoli, S., Liu, J., Zhang, Y., Wang, X., Warfield, B. M., Kueh, A. J., Voss, A. K., Thomas, T., Garcia, B. A., Liu, W. R., Strahl, B. D., Kono, H., Li, W., Shi, X., Cote, J. and Kutateladze, T. G. (2019) 'Histone H3K23- specific acetylation by MORF is coupled to H3K14 acylation', Nat Commun, 10(1), pp. 4724.

185

Kobayashi, T. (2003) 'The replication fork barrier site forms a unique structure with Fob1p and inhibits the replication fork', Mol Cell Biol, 23(24), pp. 9178-88.

Kogoma, T. (1997) 'Stable DNA replication: interplay between DNA replication, homologous recombination, and transcription', Microbiol Mol Biol Rev, 61(2), pp. 212-38.

Kouzarides, T. (2007) 'Chromatin modifications and their function', Cell, 128(4), pp. 693-705.

Kurat, C. F., Yeeles, J. T. P., Patel, H., Early, A. and Diffley, J. F. X. (2017) 'Chromatin Controls DNA Replication Origin Selection, Lagging-Strand Synthesis, and Replication Fork Rates', Mol Cell, 65(1), pp. 117-130.

Kuznetsov, V. A., Bondarenko, V., Wongsurawat, T., Yenamandra, S. P. and Jenjaroenpun, P. (2018) 'Toward predictive R-loop computational biology: genome- scale prediction of R-loops reveals their association with complex promoter structures, G-quadruplexes and transcriptionally active enhancers', Nucleic Acids Res, 46(15), pp. 7566-7585.

Lada, A. G., Stepchenkova, E. I., Waisertreiger, I. S., Noskov, V. N., Dhar, A., Eudy, J. D., Boissy, R. J., Hirano, M., Rogozin, I. B. and Pavlov, Y. I. (2013) 'Genome-wide mutation avalanches induced in diploid yeast cells by a base analog or an APOBEC deaminase', PLoS Genet, 9(9), pp. e1003736.

Lafon, A., Taranum, S., Pietrocola, F., Dingli, F., Loew, D., Brahma, S., Bartholomew, B. and Papamichos-Chronakis, M. (2015) 'INO80 Chromatin Remodeler Facilitates Release of RNA Polymerase II from Chromatin for Ubiquitin- Mediated Proteasomal Degradation', Mol Cell, 60(5), pp. 784-796.

Lafuente-Barquero, J., Luke-Glaser, S., Graf, M., Silva, S., Gomez-Gonzalez, B., Lockhart, A., Lisby, M., Aguilera, A. and Luke, B. (2017) 'The Smc5/6 complex regulates the yeast Mph1 helicase at RNA-DNA hybrid-mediated DNA damage', PLoS Genet, 13(12), pp. e1007136.

Li, L., Monckton, E. A. and Godbout, R. (2008) 'A role for DEAD box 1 at DNA double-strand breaks', Mol Cell Biol, 28(20), pp. 6413-25.

Li, T., Stark, M. R., Johnson, A. D. and Wolberger, C. (1995) 'Crystal structure of the MATa1/MAT alpha 2 homeodomain heterodimer bound to DNA', Science, 270(5234), pp. 262-9.

Li, X. and Manley, J. L. (2005) 'Inactivation of the SR protein splicing factor ASF/SF2 results in genomic instability', Cell, 122(3), pp. 365-78.

Lisby, M., Rothstein, R. and Mortensen, U. H. (2001) 'Rad52 forms DNA repair and recombination centers during S phase', Proc Natl Acad Sci U S A, 98(15), pp. 8276-82.

186

Liu, B. and Alberts, B. M. (1995) 'Head-on collision between a DNA replication apparatus and RNA polymerase transcription complex', Science, 267(5201), pp. 1131-7.

Liu, S., Xu, Z., Leng, H., Zheng, P., Yang, J., Chen, K., Feng, J. and Li, Q. (2017) 'RPA binds histone H3-H4 and functions in DNA replication-coupled nucleosome assembly', Science, 355(6323), pp. 415-420.

Loomis, E. W., Sanz, L. A., Chedin, F. and Hagerman, P. J. (2014) 'Transcription-associated R-loop formation across the human FMR1 CGG-repeat region', PLoS Genet, 10(4), pp. e1004294.

Lynch, M., Sung, W., Morris, K., Coffey, N., Landry, C. R., Dopman, E. B., Dickinson, W. J., Okamoto, K., Kulkarni, S., Hartl, D. L. and Thomas, W. K. (2008) 'A genome-wide view of the spectrum of spontaneous mutations in yeast', Proc Natl Acad Sci U S A, 105(27), pp. 9272-7.

Marini, F., Rawal, C. C., Liberi, G. and Pellicioli, A. (2019) 'Regulation of DNA Double Strand Breaks Processing: Focus on Barriers', Front Mol Biosci, 6, pp. 55.

Marsico, G., Chambers, V. S., Sahakyan, A. B., McCauley, P., Boutell, J. M., Antonio, M. D. and Balasubramanian, S. (2019) 'Whole genome experimental maps of DNA G-quadruplexes in multiple species', Nucleic Acids Res, 47(8), pp. 3862-3874.

Masumoto, H., Hawke, D., Kobayashi, R. and Verreault, A. (2005) 'A role for cell-cycle-regulated histone H3 lysine 56 acetylation in the DNA damage response', Nature, 436(7048), pp. 294-8.

Minard, L. V., Lin, L. J. and Schultz, M. C. (2011) 'SWI/SNF and Asf1 independently promote derepression of the DNA damage response genes under conditions of replication stress', PLoS One, 6(6), pp. e21633.

Mirkin, E. V. and Mirkin, S. M. (2005) 'Mechanisms of transcription-replication collisions in bacteria', Mol Cell Biol, 25(3), pp. 888-95.

Mischo, H. E., Gomez-Gonzalez, B., Grzechnik, P., Rondon, A. G., Wei, W., Steinmetz, L., Aguilera, A. and Proudfoot, N. J. (2011) 'Yeast Sen1 helicase protects the genome from transcription-associated instability', Mol Cell, 41(1), pp. 21-32.

Munoz-Galvan, S., Jimeno, S., Rothstein, R. and Aguilera, A. (2013) 'Histone H3K56 acetylation, Rad52, and non-DNA repair factors control double-strand break repair choice with the sister chromatid', PLoS Genet, 9(1), pp. e1003237.

Muramatsu, M., Kinoshita, K., Fagarasan, S., Yamada, S., Shinkai, Y. and Honjo, T. (2000) 'Class switch recombination and hypermutation require activation-

187 induced cytidine deaminase (AID), a potential RNA editing enzyme', Cell, 102(5), pp. 553-63.

Nadel, J., Athanasiadou, R., Lemetre, C., Wijetunga, N. A., P, O. B., Sato, H., Zhang, Z., Jeddeloh, J., Montagna, C., Golden, A., Seoighe, C. and Greally, J. M. (2015) 'RNA:DNA hybrids in the human genome have distinctive nucleotide characteristics, chromatin composition, and transcriptional relationships', Epigenetics Chromatin, 8, pp. 46.

Nakama, M., Kawakami, K., Kajitani, T., Urano, T. and Murakami, Y. (2012) 'DNA-RNA hybrid formation mediates RNAi-directed heterochromatin formation', Genes Cells, 17(3), pp. 218-33.

Nguyen, V. C., Clelland, B. W., Hockman, D. J., Kujat-Choy, S. L., Mewhort, H. E. and Schultz, M. C. (2010) 'Replication stress checkpoint signaling controls tRNA gene transcription', Nat Struct Mol Biol, 17(8), pp. 976-81.

Nick McElhinny, S. A., Watts, B. E., Kumar, D., Watt, D. L., Lundstrom, E. B., Burgers, P. M., Johansson, E., Chabes, A. and Kunkel, T. A. (2010) 'Abundant ribonucleotide incorporation into DNA by yeast replicative polymerases', Proc Natl Acad Sci U S A, 107(11), pp. 4949-54.

Ohle, C., Tesorero, R., Schermann, G., Dobrev, N., Sinning, I. and Fischer, T. (2016) 'Transient RNA-DNA Hybrids Are Required for Efficient Double-Strand Break Repair', Cell, 167(4), pp. 1001-1013 e7.

Ortega, P., Gomez-Gonzalez, B. and Aguilera, A. (2019) 'Rpd3L and Hda1 histone deacetylases facilitate repair of broken forks by promoting sister chromatid cohesion', Nat Commun, 10(1), pp. 5178.

Pai, C. C. and Kearsey, S. E. (2017) 'A Critical Balance: dNTPs and the Maintenance of Genome Stability', Genes (Basel), 8(2).

Papamichos-Chronakis, M., Krebs, J. E. and Peterson, C. L. (2006) 'Interplay between Ino80 and Swr1 chromatin remodeling enzymes regulates cell cycle checkpoint adaptation in response to DNA damage', Genes Dev, 20(17), pp. 2437- 49.

Pardo, B., Gomez-Gonzalez, B. and Aguilera, A. (2009) 'DNA repair in mammalian cells: DNA double-strand break repair: how to fix a broken relationship', Cell Mol Life Sci, 66(6), pp. 1039-56.

Peng, Q., Lu, S., Shi, Y., Pan, Y., Limsakul, P., Chernov, A. V., Qiu, J., Chai, X., Shi, Y., Wang, P., Ji, Y., Li, Y. J., Strongin, A. Y., Verkhusha, V. V., Izpisua Belmonte, J. C., Ren, B., Wang, Y., Chien, S. and Wang, Y. (2018) 'Coordinated histone modifications and chromatin reorganization in a single cell revealed by FRET biosensors', Proc Natl Acad Sci U S A, 115(50), pp. E11681-E11690.

188

Peters, J. M., Vangeloff, A. D. and Landick, R. (2011) 'Bacterial transcription terminators: the RNA 3'-end chronicles', J Mol Biol, 412(5), pp. 793-813.

Phillips, D. D., Garboczi, D. N., Singh, K., Hu, Z., Leppla, S. H. and Leysath, C. E. (2013) 'The sub-nanomolar binding of DNA-RNA hybrids by the single-chain Fv fragment of antibody S9.6', J Mol Recognit, 26(8), pp. 376-81.

Poli, J., Gerhold, C. B., Tosi, A., Hustedt, N., Seeber, A., Sack, R., Herzog, F., Pasero, P., Shimada, K., Hopfner, K. P. and Gasser, S. M. (2016) 'Mec1, INO80, and the PAF1 complex cooperate to limit transcription replication conflicts through RNAPII removal during replication stress', Genes Dev, 30(3), pp. 337-54.

Pomerantz, R. T. and O'Donnell, M. (2008) 'The replisome uses mRNA as a primer after colliding with RNA polymerase', Nature, 456(7223), pp. 762-6.

Powell, S. K., MacAlpine, H. K., Prinz, J. A., Li, Y., Belsky, J. A. and MacAlpine, D. M. (2015) 'Dynamic loading and redistribution of the Mcm2-7 helicase complex through the cell cycle', EMBO J, 34(4), pp. 531-43.

Prado, F. and Aguilera, A. (2005) 'Impairment of replication fork progression mediates RNA polII transcription-associated recombination', EMBO J, 24(6), pp. 1267-76.

Qiao, Q., Wang, L., Meng, F. L., Hwang, J. K., Alt, F. W. and Wu, H. (2017) 'AID Recognizes Structured DNA for Class Switch Recombination', Mol Cell, 67(3), pp. 361-373 e4.

Revy, P., Muto, T., Levy, Y., Geissmann, F., Plebani, A., Sanal, O., Catalan, N., Forveille, M., Dufourcq-Labelouse, R., Gennery, A., Tezcan, I., Ersoy, F., Kayserili, H., Ugazio, A. G., Brousse, N., Muramatsu, M., Notarangelo, L. D., Kinoshita, K., Honjo, T., Fischer, A. and Durandy, A. (2000) 'Activation-induced cytidine deaminase (AID) deficiency causes the autosomal recessive form of the Hyper-IgM syndrome (HIGM2)', Cell, 102(5), pp. 565-75.

Rich, T., Allen, R. L. and Wyllie, A. H. (2000) 'Defying death after DNA damage', Nature, 407(6805), pp. 777-83.

Richard, P. and Manley, J. L. (2017) 'R Loops and Links to Human Disease', J Mol Biol, 429(21), pp. 3168-3180.

Roberts, R. W. and Crothers, D. M. (1992) 'Stability and properties of double and triple helices: dramatic effects of RNA or DNA backbone composition', Science, 258(5087), pp. 1463-6.

Rogakou, E. P., Pilch, D. R., Orr, A. H., Ivanova, V. S. and Bonner, W. M. (1998) 'DNA double-stranded breaks induce histone H2AX phosphorylation on serine 139', J Biol Chem, 273(10), pp. 5858-68.

189

Roy, D. and Lieber, M. R. (2009) 'G clustering is important for the initiation of transcription-induced R-loops in vitro, whereas high G density without clustering is sufficient thereafter', Mol Cell Biol, 29(11), pp. 3124-33.

Roy, D., Zhang, Z., Lu, Z., Hsieh, C. L. and Lieber, M. R. (2010) 'Competition between the RNA transcript and the nontemplate DNA strand during R-loop formation in vitro: a nick can serve as a strong R-loop initiation site', Mol Cell Biol, 30(1), pp. 146-59.

Ryan, A. J., Squires, S., Strutt, H. L. and Johnson, R. T. (1991) 'Camptothecin cytotoxicity in mammalian cells is associated with the induction of persistent double strand breaks in replicating DNA', Nucleic Acids Res, 19(12), pp. 3295-300.

Salas-Armenteros, I., Perez-Calero, C., Bayona-Feliu, A., Tumini, E., Luna, R. and Aguilera, A. (2017) 'Human THO-Sin3A interaction reveals new mechanisms to prevent R-loops that cause genome instability', EMBO J, 36(23), pp. 3532-3547.

Saleh-Gohari, N., Bryant, H. E., Schultz, N., Parker, K. M., Cassel, T. N. and Helleday, T. (2005) 'Spontaneous homologous recombination is induced by collapsed replication forks that are caused by endogenous DNA single-strand breaks', Mol Cell Biol, 25(16), pp. 7158-69.

Sambrook, J., Fritsch, E. F. and Maniatis, T. (1989) Molecular Cloning: A Laboratory Manual. Second edn.: Cold Spring Harbor Laboratory Press.

Santos-Pereira, J. M. and Aguilera, A. (2015) 'R loops: new modulators of genome dynamics and function', Nat Rev Genet, 16(10), pp. 583-97.

Sanz, L. A., Hartono, S. R., Lim, Y. W., Steyaert, S., Rajpurkar, A., Ginno, P. A., Xu, X. and Chedin, F. (2016) 'Prevalent, Dynamic, and Conserved R-Loop Structures Associate with Specific Epigenomic Signatures in Mammals', Mol Cell, 63(1), pp. 167-78.

Savic, D. J. and Kanazir, D. T. (1972) 'The effect of a histidine operator- constitutive mutation on UV-induced mutability within the histidine operon of Salmonella typhimurium', Mol Gen Genet, 118(1), pp. 45-50.

Schwab, R. A., Nieminuszczy, J., Shah, F., Langton, J., Lopez Martinez, D., Liang, C. C., Cohn, M. A., Gibbons, R. J., Deans, A. J. and Niedzwiedz, W. (2015) 'The Fanconi Anemia Pathway Maintains Genome Stability by Coordinating Replication and Transcription', Mol Cell, 60(3), pp. 351-61.

Scully, R., Panday, A., Elango, R. and Willis, N. A. (2019) 'DNA double-strand break repair-pathway choice in somatic mammalian cells', Nat Rev Mol Cell Biol, 20(11), pp. 698-714.

Shaw, N. N. and Arya, D. P. (2008) 'Recognition of the unique structure of DNA:RNA hybrids', Biochimie, 90(7), pp. 1026-39.

190

Shibata, A. (2017) 'Regulation of repair pathway choice at two-ended DNA double-strand breaks', Mutat Res, 803-805, pp. 51-55.

Sims, R. J., 3rd, Millhouse, S., Chen, C. F., Lewis, B. A., Erdjument-Bromage, H., Tempst, P., Manley, J. L. and Reinberg, D. (2007) 'Recognition of trimethylated histone H3 lysine 4 facilitates the recruitment of transcription postinitiation factors and pre-mRNA splicing', Mol Cell, 28(4), pp. 665-76.

Skourti-Stathaki, K. and Proudfoot, N. J. (2014) 'A double-edged sword: R loops as threats to genome integrity and powerful regulators of gene expression', Genes Dev, 28(13), pp. 1384-96.

Skourti-Stathaki, K., Proudfoot, N. J. and Gromak, N. (2011) 'Human senataxin resolves RNA/DNA hybrids formed at transcriptional pause sites to promote Xrn2-dependent termination', Mol Cell, 42(6), pp. 794-805.

Smeenk, G., Wiegant, W. W., Vrolijk, H., Solari, A. P., Pastink, A. and van Attikum, H. (2010) 'The NuRD chromatin-remodeling complex regulates signaling and repair of DNA damage', J Cell Biol, 190(5), pp. 741-9.

Smirnov, E., Borkovec, J., Kovacik, L., Svidenska, S., Schrofel, A., Skalnikova, M., Svindrych, Z., Krizek, P., Ovesny, M., Hagen, G. M., Juda, P., Michalova, K., Cardoso, M. C., Cmarko, D. and Raska, I. (2014) 'Separation of replication and transcription domains in nucleoli', J Struct Biol, 188(3), pp. 259-66.

Sobhian, B., Shao, G., Lilli, D. R., Culhane, A. C., Moreau, L. A., Xia, B., Livingston, D. M. and Greenberg, R. A. (2007) 'RAP80 targets BRCA1 to specific ubiquitin structures at DNA damage sites', Science, 316(5828), pp. 1198-202.

Sollier, J. and Cimprich, K. A. (2015) 'Breaking bad: R-loops and genome integrity', Trends Cell Biol, 25(9), pp. 514-22.

Sollier, J., Stork, C. T., Garcia-Rubio, M. L., Paulsen, R. D., Aguilera, A. and Cimprich, K. A. (2014) 'Transcription-coupled nucleotide excision repair factors promote R-loop-induced genome instability', Mol Cell, 56(6), pp. 777-85.

Sordet, O., Redon, C. E., Guirouilh-Barbat, J., Smith, S., Solier, S., Douarre, C., Conti, C., Nakamura, A. J., Das, B. B., Nicolas, E., Kohn, K. W., Bonner, W. M. and Pommier, Y. (2009) 'Ataxia telangiectasia mutated activation by transcription- and topoisomerase I-induced DNA double-strand breaks', EMBO Rep, 10(8), pp. 887-93.

Stirling, P. C., Chan, Y. A., Minaker, S. W., Aristizabal, M. J., Barrett, I., Sipahimalani, P., Kobor, M. S. and Hieter, P. (2012) 'R-loop-mediated genome instability in mRNA cleavage and polyadenylation mutants', Genes Dev, 26(2), pp. 163-75.

191

Stork, C. T., Bocek, M., Crossley, M. P., Sollier, J., Sanz, L. A., Chedin, F., Swigut, T. and Cimprich, K. A. (2016) 'Co-transcriptional R-loops are the main cause of estrogen-induced DNA damage', Elife, 5.

Strasser, K., Masuda, S., Mason, P., Pfannstiel, J., Oppizzi, M., Rodriguez- Navarro, S., Rondon, A. G., Aguilera, A., Struhl, K., Reed, R. and Hurt, E. (2002) 'TREX is a conserved complex coupling transcription with messenger RNA export', Nature, 417(6886), pp. 304-8.

Strumberg, D., Pilon, A. A., Smith, M., Hickey, R., Malkas, L. and Pommier, Y. (2000) 'Conversion of topoisomerase I cleavage complexes on the leading strand of ribosomal DNA into 5'-phosphorylated DNA double-strand breaks by replication runoff', Mol Cell Biol, 20(11), pp. 3977-87.

Stuckey, R., Garcia-Rodriguez, N., Aguilera, A. and Wellinger, R. E. (2015) 'Role for RNA:DNA hybrids in origin-independent replication priming in a eukaryotic system', Proc Natl Acad Sci U S A, 112(18), pp. 5779-84.

Su, D., Hu, Q., Li, Q., Thompson, J. R., Cui, G., Fazly, A., Davies, B. A., Botuyan, M. V., Zhang, Z. and Mer, G. (2012) 'Structural basis for recognition of H3K56-acetylated histone H3-H4 by the chaperone Rtt106', Nature, 483(7387), pp. 104-7.

Su, T. T. (2006) 'Cellular responses to DNA damage: one signal, multiple choices', Annu Rev Genet, 40, pp. 187-208.

Sun, Q., Csorba, T., Skourti-Stathaki, K., Proudfoot, N. J. and Dean, C. (2013) 'R-loop stabilization represses antisense transcription at the Arabidopsis FLC locus', Science, 340(6132), pp. 619-21.

Svikovic, S., Crisp, A., Tan-Wong, S. M., Guilliam, T. A., Doherty, A. J., Proudfoot, N. J., Guilbaud, G. and Sale, J. E. (2019) 'R-loop formation during S phase is restricted by PrimPol-mediated repriming', EMBO J, 38(3).

Taneja, N., Zofall, M., Balachandran, V., Thillainadesan, G., Sugiyama, T., Wheeler, D., Zhou, M. and Grewal, S. I. (2017) 'SNF2 Family Protein Fft3 Suppresses Nucleosome Turnover to Promote Epigenetic Inheritance and Proper Replication', Mol Cell, 66(1), pp. 50-62 e6.

Taylor, B. J., Nik-Zainal, S., Wu, Y. L., Stebbings, L. A., Raine, K., Campbell, P. J., Rada, C., Stratton, M. R. and Neuberger, M. S. (2013) 'DNA deaminases induce break-associated mutation showers with implication of APOBEC3B and 3A in breast cancer kataegis', Elife, 2, pp. e00534.

Taylor, B. J., Wu, Y. L. and Rada, C. (2014) 'Active RNAP pre-initiation sites are highly mutated by cytidine deaminases in yeast, with AID targeting small RNA genes', Elife, 3, pp. e03553.

192

Tehranchi, A. K., Blankschien, M. D., Zhang, Y., Halliday, J. A., Srivatsan, A., Peng, J., Herman, C. and Wang, J. D. (2010) 'The transcription factor DksA prevents conflicts between DNA replication and transcription machinery', Cell, 141(4), pp. 595-605.

Tian, M. and Alt, F. W. (2000) 'Transcription-induced cleavage of immunoglobulin switch regions by nucleotide excision repair nucleases in vitro', J Biol Chem, 275(31), pp. 24163-72.

Topal, S., Vasseur, P., Radman-Livaja, M. and Peterson, C. L. (2019) 'Distinct transcriptional roles for Histone H3-K56 acetylation during the cell cycle in Yeast', Nat Commun, 10(1), pp. 4372.

Trautinger, B. W., Jaktaji, R. P., Rusakova, E. and Lloyd, R. G. (2005) 'RNA polymerase modulators and DNA repair activities resolve conflicts between DNA replication and transcription', Mol Cell, 19(2), pp. 247-58.

Tsao, Y. P., Russo, A., Nyamuswa, G., Silber, R. and Liu, L. F. (1993) 'Interaction between replication forks and topoisomerase I-DNA cleavable complexes: studies in a cell-free SV40 DNA replication system', Cancer Res, 53(24), pp. 5908-14.

Tsukiyama, T., Palmer, J., Landel, C. C., Shiloach, J. and Wu, C. (1999) 'Characterization of the imitation switch subfamily of ATP-dependent chromatin- remodeling factors in Saccharomyces cerevisiae', Genes Dev, 13(6), pp. 686-97.

Tubbs, A. and Nussenzweig, A. (2017) 'Endogenous DNA Damage as a Source of Genomic Instability in Cancer', Cell, 168(4), pp. 644-656.

van Attikum, H., Fritsch, O. and Gasser, S. M. (2007) 'Distinct roles for SWR1 and INO80 chromatin remodeling complexes at chromosomal double-strand breaks', EMBO J, 26(18), pp. 4113-25.

Vincent, J. A., Kwong, T. J. and Tsukiyama, T. (2008) 'ATP-dependent chromatin remodeling shapes the DNA replication landscape', Nat Struct Mol Biol, 15(5), pp. 477-84.

Voichek, Y., Bar-Ziv, R. and Barkai, N. (2016) 'Expression homeostasis during DNA replication', Science, 351(6277), pp. 1087-90.

Voichek, Y., Mittelman, K., Gordon, Y., Bar-Ziv, R., Lifshitz Smit, D., Shenhav, R. and Barkai, N. (2018) 'Epigenetic Control of Expression Homeostasis during Replication Is Stabilized by the Replication Checkpoint', Mol Cell, 70(6), pp. 1121- 1133 e9.

Wahba, L., Amon, J. D., Koshland, D. and Vuica-Ross, M. (2011) 'RNase H and multiple RNA biogenesis factors cooperate to prevent RNA:DNA hybrids from generating genome instability', Mol Cell, 44(6), pp. 978-88.

193

Wahba, L., Costantino, L., Tan, F. J., Zimmer, A. and Koshland, D. (2016) 'S1-DRIP-seq identifies high expression and polyA tracts as major contributors to R-loop formation', Genes Dev, 30(11), pp. 1327-38.

Wang, J. D., Berkmen, M. B. and Grossman, A. D. (2007) 'Genome-wide coorientation of replication and transcription reduces adverse effects on replication in Bacillus subtilis', Proc Natl Acad Sci U S A, 104(13), pp. 5608-13.

Wang, K., Li, L., Zhang, Y. and Gao, D. (2020) 'Crosstalk between signaling pathways and DNA damage response', Genome Instability & Disease, 1(2), pp. 81- 91.

Wang, M., Yang, Z., Rada, C. and Neuberger, M. S. (2009) 'AID upmutants isolated using a high-throughput screen highlight the immunity/cancer balance limiting DNA deaminase activity', Nat Struct Mol Biol, 16(7), pp. 769-76.

Wang, Q., Oliveira, T., Jankovic, M., Silva, I. T., Hakim, O., Yao, K., Gazumyan, A., Mayer, C. T., Pavri, R., Casellas, R., Nussenzweig, M. C. and Robbiani, D. F. (2014) 'Epigenetic targeting of activation-induced cytidine deaminase', Proc Natl Acad Sci U S A, 111(52), pp. 18667-72.

Wang, Y., Kallgren, S. P., Reddy, B. D., Kuntz, K., Lopez-Maury, L., Thompson, J., Watt, S., Ma, C., Hou, H., Shi, Y., Yates, J. R., 3rd, Bahler, J., O'Connell, M. J. and Jia, S. (2012) 'Histone H3 lysine 14 acetylation is required for activation of a DNA damage checkpoint in fission yeast', J Biol Chem, 287(6), pp. 4386-93.

Wanrooij, P. H., Uhler, J. P., Shi, Y., Westerlund, F., Falkenberg, M. and Gustafsson, C. M. (2012) 'A hybrid G-quadruplex structure formed between RNA and DNA explains the extraordinary stability of the mitochondrial R-loop', Nucleic Acids Res, 40(20), pp. 10334-44.

Waters, R. and Parry, J. M. (1973) 'The response to chemical mutagens of the individual haploid and homoallelic diploid UV-sensitive mutants of the rad 3 locus of Saccharomyces cerevisiae', Mol Gen Genet, 124(2), pp. 135-43.

Wellinger, R. E., Prado, F. and Aguilera, A. (2006) 'Replication fork progression is impaired by transcription in hyperrecombinant yeast cells lacking a functional THO complex', Mol Cell Biol, 26(8), pp. 3327-34.

Wellinger, R. E., Schar, P. and Sogo, J. M. (2003) 'Rad52-independent accumulation of joint circular minichromosomes during S phase in Saccharomyces cerevisiae', Mol Cell Biol, 23(18), pp. 6363-72.

Williams, S. K., Truong, D. and Tyler, J. K. (2008) 'Acetylation in the globular core of histone H3 on lysine-56 promotes chromatin disassembly during transcriptional activation', Proc Natl Acad Sci U S A, 105(26), pp. 9000-5.

194

Yan, Q., Shields, E. J., Bonasio, R. and Sarma, K. (2019) 'Mapping Native R- Loops Genome-wide Using a Targeted Nuclease Approach', Cell Rep, 29(5), pp. 1369-1380 e5.

Yang, Y., McBride, K. M., Hensley, S., Lu, Y., Chedin, F. and Bedford, M. T. (2014) 'Arginine methylation facilitates the recruitment of TOP3B to chromatin to prevent R loop accumulation', Mol Cell, 53(3), pp. 484-97.

Yu, K., Chedin, F., Hsieh, C. L., Wilson, T. E. and Lieber, M. R. (2003) 'R- loops at immunoglobulin class switch regions in the chromosomes of stimulated B cells', Nat Immunol, 4(5), pp. 442-51.

Zhang, H., Roberts, D. N. and Cairns, B. R. (2005) 'Genome-wide dynamics of Htz1, a histone H2A variant that poises repressed/basal promoters for activation through histone loss', Cell, 123(2), pp. 219-31.

195