Inheritance of germline pathogenic mutations in a single BRCA2 allele predisposes carriers to cancers of the breast, ovaries, pancreas, prostate and other tissues. BRCA2 encodes a complex protein of 3418 residues, with multiple distinct regions implicated in different cellular processes including homologous recombination, replication stress tolerance, maintenance of the G2 and spindle assembly checkpoints, and in completion of cytokinesis. Much evidence suggests that BRCA2 acts as a tumour suppressor by maintaining genome integrity through its functions during DNA replication, repair and chromosome segregation. Neither the mechanisms underlying these functions, nor their significance to tumour suppression, is well understood. In this thesis, I report studies that reveal two novel mechanisms underlying genome maintenance and tumour suppression by BRCA2. The best-studied cellular function of BRCA2 is its role in double strand break repair by homologous DNA recombination, wherein BRCA2 controls the DNA substrate selectivity and activity of the recombinase, RAD51. This control is exerted via evolutionarily conserved BRC repeats motifs, each comprising ~35 residues. BRCA2 homologues in simple organisms such as Caenorhabditis elegans or Ustilago maydis contain only one BRC repeat. By contrast, all known vertebrate BRCA2 orthologues contain 8 BRC repeats whose sequence as well as spacing is conserved. The functional significance of the conserved sequence and spacing of vertebrate BRC repeats in the maintenance of genome integrity is unclear. I report that the BRC1 and BRC4 repeats in chicken BRCA2 are essential for cell viability. Additionally, the BRC6 repeat has an essential, non- redundant role in tolerating replication stress by stabilizing stalled replication forks and preventing the degradation of nascent DNA by MRE11, and the mechanism underlying this function is largely RAD51-independent. Furthermore, my results imply that the BRC repeats in chicken BRCA2 are aligned in the same order as in the human BRCA2, suggesting the BRC6 repeat in chicken BRCA2 may correspond to the BRC6 repeat in human BRCA2. 2 ABSTRACT Although both alleles of BRCA2 must be inactivated to provoke defective homologous DNA recombination, accumulating evidence suggests that BRCA2 heterozygosity suffices for carcinogenesis in murine models and human patients. In a second line of research, I have explored in collaborative studies how heterozygous truncating mutations affecting BRCA2 may promote carcinogenesis. I report that in unchallenged conditions, BRCA2 heterozygous cells appear to be as functionally proficient in key cellular processes as the BRCA2 wildtype cells. However, BRCA2 heterozygosity confers selective sensitivity to replication stress induced by exposure to physiological concentrations of formaldehyde, a known genotoxic agent. Replication fork degradation is observed in BRCA2 heterozygous cells on exposure to formaldehyde, but not hydroxyurea, and this deleterious phenotype has been attributed to the ability of formaldehyde to induce BRCA2 degradation in a proteasome-dependent manner, thereby inducing haploinsufficiency. Overall, my thesis research reveals two novel mechanisms by which BRCA2 maintains genomic integrity. Identification of a novel function for the BRC6 repeat in chicken BRCA2 indicates that individual BRC repeats may have essential functions that are yet to be identified, thereby providing an explanation for the evolutionary conservation of the 8 BRC repeats observed in vertebrate BRCA2. Furthermore, results linking BRCA2 heterozygosity and formaldehyde exposure to increased genomic instability highlights the potential ability of aldehyde exposure in promoting carcinogenesis in heterozygous BRCA2 mutation carriers, which has significant preventive and therapeutic implications.

L’intégrase (IN) du VIH-1 est une enzyme clé du cycle viral du VIH-1 puisque celle-ci catalyse l’insertion stable du génome viral dans celui de la cellule infectée. D’autre part, l’IN participe également à de nombreuses étapes du cycle viral (transcription inverse, import du complexe de préintégration, bourgeonnement…). L’étape d’intégration elle-même fait intervenir de nombreux partenaires cellulaires et viraux interagissant avec l’IN. Certains sont connus et étudiés (LEDGF/P75, TNPO3…), mais il est très probable qu’un très grand nombre de ces partenaires soient encore méconnus malgré leur importance. Depuis quelque année, le rôle des modifications post-traductionnelles de l’IN a commencé à être étudié. En effet plusieurs études montrent que la régulation de l’activité de l’IN pourrait se faire via de telles modifications. Mon travail de thèse s’est orienté sur trois questions autour de ces deux aspects. -Nous avons identifié plusieurs phosphorylations de l’IN par spectrométrie de masse et mis en évidence le rôle essentiel de la phopshorylation de la sérine 24 pour l’infection virale. -Le rôle de la kinase cellulaire GCN2 a été étudié. Nous avons pu montrer un effet restrictif de la protéine sur le cycle viral amenant à un arrêt de la traduction à un temps court après l’infection au VIH-1. L’interaction entre GCN2 et l’IN a été mise en évidence. L’étude du domaine d’interaction entre l’IN et GCN2 a permis la caractérisation d’un résidu essentiel de l’IN, le E85. -L’impact du facteur de réparation RAD51 sur la réplication virale a été étudié. Nous avons montré un effet inhibiteur de cette protéine. Ce travail a permis l’identification d’une molécule chimique RS-1 capable d’inhiber l’intégration dans les cellules infectées via la stimulation de RAD51. / The integrase (IN) of HIV-1 is a key enzyme of the viral cycle of HIV-1 since it catalyzes the stable integration of the viral genome into that of the infected cell. Furthermore, the IN also participates in many steps of the viral cycle (reverse transcription, import of preintegration complex, budding ...). The integration step itself involves many cellular and viral partners interacting with IN. Some of them are studied (LEDGF/p75, TNPO3 ...) but it is very likely that many of these partners are still unknown despite their importance. Recently, the role of post-translational modifications of the IN began to be studied. In fact several studies show that the regulation of the activity of IN could be done through such modifications.My thesis work focused on three issues: -We identified several phosphorylations of IN by mass spectrometry and identified the crucial role of serine 24 to viral infection. -The role of GCN2 cellular kinase was studied. We have shown a restrictive effect of the protein on the viral cycle leading to a translation stop in first hours following infection with HIV-1. The study of the interaction domain between IN and GCN2 allowed the characterization of a critical residue of IN, the E85. -The impact of RAD51 repair factor on viral replication was investigated. We have shown an inhibitory effect of this protein. This work allowed the identification of a chemical molecule RS-1 able to inhibit integration in infected cells through RAD51 stimulation.

Les cellules utilisent deux stratégies principales pour réparer les cassures double-brin (CDB) de l’ADN : la recombinaison homologue (RH) et la ligature d’extrémités non homologues (NHEJ). D’autres voies de réparation plus minoritaires existent qui mènent nécessairement à des altérations génétiques : Single Strand Annealing (SSA) et Alternative End Joining (A-EJ). Nous avons proposé que le choix entre les mécanismes de réparation des CDB nécessite deux étapes : 1) la compétition entre C-NHEJ et la résection, 2) sur les extrémités d’ADN résectées, la compétition entre RH, A-EJ et SSA. Ici, nous avons étudié la régulation de la deuxième étape de ce choix. En outre, la létalité synthétique a été décrite entre RAD52 et BRCA2/PALB2. Étant donné que BRCA2 et PALB2 sont nécessaires pour le chargement de RAD51 sur l’ADN simple brin, cela suggère que la formation d’un nucléofilament RAD51/ADNsb ordonné et RAD52 sont des acteurs essentiels dans le choix de la réparation à la deuxième étape. Nous avons trouvé que l’extinction de RAD51 ou BRCA2 stimule à la fois le SSA et EJ, d’une manière épistatique et que RAD52 contrôle la stimulation de SSA et A-EJ, en absence de RAD51. De plus, par séquençage haut débit, nous montrons que l’inhibition de RAD51 induit une instabilité génomique médiée par la microhomologie au niveau du génome. Cependant l’inhibition de RH n’est pas la réparation directe suffisante vers SSA et A-EJ. En effet, en utilisant des mutants dominants négatifs de RAD51, nous avons trouvé que les mutants du site de fixation/hydrolyse de l’ATP inhibent la RH et stimulent le SSA et que la chimère SMRAD51, qui inhibe la RH, inhibe également le SSA et EJ. Par TEM, nous avons observé que SMRAD51 perturbe spécifiquement la structure de l’ADNsb/SMRAD51. De l’autre côté, deux mutants d’hydrolyse de l’ATP de RAD51 ont montré que la liaison à l’ATP et l’hydrolyse d’ATP sont nécessaires pour une charge efficace de RAD51 sur l’ADN endommagé, dans les cellules vivantes. Ces deux mutants d’ATP ne se fixent pas à l’ADN en opposition à SMRAD51. Enfin, nous montrons que RAD51 n’empêche pas la résection étendue, mais que, in vitro, la protéine RAD51 empêche l’annealing de l’ADNsb complémentaire. Au total, les données montrent que RAD51 joue effectivement un rôle crucial dans la deuxième étape du choix de la voie de réparation des CDB à travers deux mécanismes distincts : 1- il déclenche la RH par son activité catalytique, 2- mail il empêche également les mécanismes non conservateurs dépendants de RAD52, SSA et A-EJ, en altérant l’étape de l’annealing. Par conséquent, le choix en deuxième étape entre la RH et les mécanismes mutagènes, SSA et A-EJ, est orchestré par un antagonisme entre RAD51 et RAD52. / Cells use two primary strategies to repair DNA double-strand break (DSB): Homologous Recombination (HR) and Non-homologous end joining (NHEJ). Beside other mechanisms exist that necessarily lead to genetic alterations: Single Strand Annealing (SSA) and Alternative End Joining (A-EJ). We have proposed that the choice between DSB repair mechanisms requires two steps: 1) competition between C-NHEJ and resection; 2) on resected DNA ends, competition between HR, A-EJ and SSA. Herein we investigated the regulation of the second step of this choice. Furthermore, synthetic lethality has been described between RAD52 and BRCA2/PALB2. Since BRCA2/PALB2 are required for the loading of RAD51 onto the ssDNA, suggesting that both the formation of an ordered RAD51/ssDNA nucleofilament and RAD52 are central players in the choice of repair at the 2nd step.We found that silencing RAD51 or BRCA2 stimulate both SSA and EJ, in an epistatic manner and that silencing RAD51 induced microhomology mediated genomic instability at a genome wide level. Moreover, we show that RAD52 controls the stimulation of SSA and A-EJ, upon RAD51 silencing. However inhibition of HR is not sufficient redirect repair toward SSA and A-EJ. Indeed, using dominant negative mutants of RAD51 we found that the chimera SMRAD51, which inhibits HR, also inhibits SSA and EJ. By TEM we observed that SMRAD51 specifically disrupts the structure of the ssDNA/SMRAD51. On the other side, two ATP hydrolysis mutants of RAD51 showed that ATP binding and hydrolysis is required for efficient loading of RAD51 on damaged DNA, in living cells. These two ATP mutants that do not bind DNA in opposition to SMRAD51, do not inhibit A-EJ and stimulate SSA. Finally we show RAD51 do not prevents extended resection, but that, in vitro, RAD51 protein prevents the annealing of complementary ssDNA.Altogether the data show that RAD51 indeed plays a pivotal role in the second step of DSB repair pathway choice through two separable mechanisms: 1- it triggers HR through its catalytic HR activity 2- but it also prevents RAD52-dependent non-conservative mechanisms SSA and A-EJ, by impairing the annealing step. Therefore, the choice between HR and alternative mutagenic mechanisms A-EJ and SSA (2nd step) is orchestrated by an antagonism between RAD51 and RAD52

O Glioblastoma Multiforme (GBM) é o tipo de tumor cerebral maligno com maior incidência na população. A perda do gene PTEN (fosfatase e tensina homóloga) é uma alteração comum associada ao GBM (até 60%) e esse gene codifica uma enzima que antagoniza a ação de PI3K, inibindo a fosforilação de AKT e, desse modo, regulando vias de sinalização relativas à sobrevivência celular e proliferação. Mutações em PTEN têm sido associadas à instabilidade genômica e ao aumento no número de quebras de fita dupla, além de serem relacionadas também à redução da expressão de RAD51, a qual é uma proteína-chave da via de reparo por recombinação homóloga (HR). Diante disso, o objetivo deste estudo foi avaliar se o status de PTEN interfere na expressão de RAD51 e proteínas parálogas (RAD51C e RAD51B) e, consequentemente, se PTEN é capaz de influenciar a eficiência de HR. Com o objetivo de induzir a formação de quebras de fita duplas (DSBs) no DNA, as células foram tratadas com a droga antitumoral etoposídeo, que produz quebras no DNA, principalmente duplas (DSBs). Duas linhagens de GBM com status diferentes de PTEN foram utilizadas: T98G (PTEN mutado) e LN18 (PTEN tipo selvagem). As células de GBM foram tratadas com etoposídeo em diferentes experimentos ou ensaios: proliferação celular, quantificação da necrose e apoptose, cinética do ciclo celular, imunofluorescência da proteína ?- H2AX, quantificação dos níveis de expressão de RAD51 e parálogas e o silenciamento de PTEN na linhagem LN18. Os resultados mostraram que a linhagem LN18 foi mais sensível à droga nos tempos iniciais (24 e 72 h) (até 61,2% de redução), em comparação com a T98G (até 12,3% de redução); no tempo mais tardio de análise (120 h), ambas as linhagens sofreram redução acentuadana proliferação. Adicionalmente, a LN18 exibiu maior porcentagem de células apoptóticas e necróticas, em comparação com a linhagem T98G, nos tempos de24, 72 e 120 horas após o tratamento. O ensaio de imunofluorescência revelou maior indução de células positivas para ?-H2AX na linhagem LN18 em relação à T98G (p =<0,001), após tratamento com etoposídeo (50 e 75 ?M). Nessas concentrações, a análise da cinética do ciclo celular mostrou um bloqueio na fase G2 em ambas as linhagens (p<0,01) nos tempos analisados (24, 48 e 72h), mas apenas a linhagem LN18 revelou bloqueio na fase S. A expressão de RAD51, RAD51B e C foi mais elevada em LN18 em comparação com a T98G e U87MG, nas células tratados (75?M) e controles. PTEN foi silenciado (siRNA-PTEN) na linhagem LN18 para verificar se a redução da expressão desse gene reduziria também a expressão de RAD51 e parálogas. Após 72 horas de silenciamento, com 69,9% de inibição de PTEN, a expressão de RAD51 e RAD51C também se mostrou reduzida em relação ao grupo controle. Em conjunto, os resultados obtidos no presente estudo indicam que o status de PTEN é crucial para as vias de sobrevivência, controle do ciclo celular e indução de apoptose nas células de GBM, indicando a relação entre PTEN e RAD51 e parálogas nas células de GBM tratadas com um indutor de quebras no DNA. Adicionalmente, outras ferramentas de estudo são requeridas para investigar as vias moleculares e possíveis interações e complexos proteicos envolvendo a participação de PTEN e RAD51 e suas proteínas parálogas / Glioblastoma multiforme (GBM) is the most common malignant brain tumor. Loss of PTEN (Phosphatase and tensin homolog deleted on chromosome 10) gene is the most frequent alteration associated with GBM and encodes a phosphatase enzyme that antagonizes the PI3K, by inhibiting AKT phosphorylation thereby regulating signaling pathways related to cell survival and proliferation. PTEN deficiency has been associated with genomic instability and increased endogenous DSBs, as well as reduced expression of RAD51, which is a key gene with crucial role in HR. In this study, we aimed to evaluate whether the PTEN status in GBM cell lines can affect RAD51 expression and HR efficiency under conditions of treatment with the antineoplastic drug etoposide, which targets the DNA topoisomerase II enzyme, thus leading to the production of DNA breaks. T98G (PTEN mutated) and LN18 (PTEN wild-type) cells were treated with etoposide, and several assays were carried out: cell proliferation, detection and quantification of necrosis and apoptosis, cell cycle kinetics, immunofluorescence staining, RAD51 (and paralogs) protein expression, and PTEN silencing in LN18 cell line, by using the siRNA method. LN18 cells showed a greater reduction in cell proliferation, compared to T98G after treatments (25, 50, 75 e 100 µM) at 24, 72 and 120h. Both cell lines showed a significant increase (p=<0.001) in cell death induction, but LN18 presented a greater percentage of apoptotic and necrotic cells than T98G (24, 72 and 120h). The induction of DSB was analyzed by immunostaining (with ?-H2AX antibody), and for the concentrations (50 and 75 µM) tested, LN18 showed higher levels of ?-H2AX positive cells than that observed for T98G (p=<0.001). The analysis of cell cycle kinetics performed for cells treated with etoposide (50 and 75 µM) and collected at 24, 48 and 72h, LN18 presented a greater G2-blockage, as compared to T98G; only LN18 showed a blockage at the S-phase. The expression of RAD51, RAD51B and C was higher in LN18 compared to T98G and U87MG cells treated with etoposide (75 µM) and controls. When we silenced PTEN in LN18 linage, to check if PTEN silencing may reduce the expression of RAD51 and its paralogs, we found a 69.9% reduction in PTEN protein expressions, and the expression of RAD51 and RAD51C was also found reduced, compared to the control group. Taken together, the results obtained in this study indicate that the status of PTEN is critical for survival pathways, cell cycle control and induction of apoptosis in GBM cells, confirming the relationship between PTEN and RAD51 and its paralogs in GBM cells treated with an inducer of DNA breaks. These results contribute with relevant information for further studies on molecular pathways underlying the interaction between PTEN and RAD51 and its paralogs

O Glioblastoma Multiforme (GBM) é o tipo de tumor cerebral maligno com maior incidência na população. A perda do gene PTEN (fosfatase e tensina homóloga) é uma alteração comum associada ao GBM (até 60%) e esse gene codifica uma enzima que antagoniza a ação de PI3K, inibindo a fosforilação de AKT e, desse modo, regulando vias de sinalização relativas à sobrevivência celular e proliferação. Mutações em PTEN têm sido associadas à instabilidade genômica e ao aumento no número de quebras de fita dupla, além de serem relacionadas também à redução da expressão de RAD51, a qual é uma proteína-chave da via de reparo por recombinação homóloga (HR). Diante disso, o objetivo deste estudo foi avaliar se o status de PTEN interfere na expressão de RAD51 e proteínas parálogas (RAD51C e RAD51B) e, consequentemente, se PTEN é capaz de influenciar a eficiência de HR. Com o objetivo de induzir a formação de quebras de fita duplas (DSBs) no DNA, as células foram tratadas com a droga antitumoral etoposídeo, que produz quebras no DNA, principalmente duplas (DSBs). Duas linhagens de GBM com status diferentes de PTEN foram utilizadas: T98G (PTEN mutado) e LN18 (PTEN tipo selvagem). As células de GBM foram tratadas com etoposídeo em diferentes experimentos ou ensaios: proliferação celular, quantificação da necrose e apoptose, cinética do ciclo celular, imunofluorescência da proteína ?- H2AX, quantificação dos níveis de expressão de RAD51 e parálogas e o silenciamento de PTEN na linhagem LN18. Os resultados mostraram que a linhagem LN18 foi mais sensível à droga nos tempos iniciais (24 e 72 h) (até 61,2% de redução), em comparação com a T98G (até 12,3% de redução); no tempo mais tardio de análise (120 h), ambas as linhagens sofreram redução acentuadana proliferação. Adicionalmente, a LN18 exibiu maior porcentagem de células apoptóticas e necróticas, em comparação com a linhagem T98G, nos tempos de24, 72 e 120 horas após o tratamento. O ensaio de imunofluorescência revelou maior indução de células positivas para ?-H2AX na linhagem LN18 em relação à T98G (p =<0,001), após tratamento com etoposídeo (50 e 75 ?M). Nessas concentrações, a análise da cinética do ciclo celular mostrou um bloqueio na fase G2 em ambas as linhagens (p<0,01) nos tempos analisados (24, 48 e 72h), mas apenas a linhagem LN18 revelou bloqueio na fase S. A expressão de RAD51, RAD51B e C foi mais elevada em LN18 em comparação com a T98G e U87MG, nas células tratados (75?M) e controles. PTEN foi silenciado (siRNA-PTEN) na linhagem LN18 para verificar se a redução da expressão desse gene reduziria também a expressão de RAD51 e parálogas. Após 72 horas de silenciamento, com 69,9% de inibição de PTEN, a expressão de RAD51 e RAD51C também se mostrou reduzida em relação ao grupo controle. Em conjunto, os resultados obtidos no presente estudo indicam que o status de PTEN é crucial para as vias de sobrevivência, controle do ciclo celular e indução de apoptose nas células de GBM, indicando a relação entre PTEN e RAD51 e parálogas nas células de GBM tratadas com um indutor de quebras no DNA. Adicionalmente, outras ferramentas de estudo são requeridas para investigar as vias moleculares e possíveis interações e complexos proteicos envolvendo a participação de PTEN e RAD51 e suas proteínas parálogas / Glioblastoma multiforme (GBM) is the most common malignant brain tumor. Loss of PTEN (Phosphatase and tensin homolog deleted on chromosome 10) gene is the most frequent alteration associated with GBM and encodes a phosphatase enzyme that antagonizes the PI3K, by inhibiting AKT phosphorylation thereby regulating signaling pathways related to cell survival and proliferation. PTEN deficiency has been associated with genomic instability and increased endogenous DSBs, as well as reduced expression of RAD51, which is a key gene with crucial role in HR. In this study, we aimed to evaluate whether the PTEN status in GBM cell lines can affect RAD51 expression and HR efficiency under conditions of treatment with the antineoplastic drug etoposide, which targets the DNA topoisomerase II enzyme, thus leading to the production of DNA breaks. T98G (PTEN mutated) and LN18 (PTEN wild-type) cells were treated with etoposide, and several assays were carried out: cell proliferation, detection and quantification of necrosis and apoptosis, cell cycle kinetics, immunofluorescence staining, RAD51 (and paralogs) protein expression, and PTEN silencing in LN18 cell line, by using the siRNA method. LN18 cells showed a greater reduction in cell proliferation, compared to T98G after treatments (25, 50, 75 e 100 µM) at 24, 72 and 120h. Both cell lines showed a significant increase (p=<0.001) in cell death induction, but LN18 presented a greater percentage of apoptotic and necrotic cells than T98G (24, 72 and 120h). The induction of DSB was analyzed by immunostaining (with ?-H2AX antibody), and for the concentrations (50 and 75 µM) tested, LN18 showed higher levels of ?-H2AX positive cells than that observed for T98G (p=<0.001). The analysis of cell cycle kinetics performed for cells treated with etoposide (50 and 75 µM) and collected at 24, 48 and 72h, LN18 presented a greater G2-blockage, as compared to T98G; only LN18 showed a blockage at the S-phase. The expression of RAD51, RAD51B and C was higher in LN18 compared to T98G and U87MG cells treated with etoposide (75 µM) and controls. When we silenced PTEN in LN18 linage, to check if PTEN silencing may reduce the expression of RAD51 and its paralogs, we found a 69.9% reduction in PTEN protein expressions, and the expression of RAD51 and RAD51C was also found reduced, compared to the control group. Taken together, the results obtained in this study indicate that the status of PTEN is critical for survival pathways, cell cycle control and induction of apoptosis in GBM cells, confirming the relationship between PTEN and RAD51 and its paralogs in GBM cells treated with an inducer of DNA breaks. These results contribute with relevant information for further studies on molecular pathways underlying the interaction between PTEN and RAD51 and its paralogs

Yes / We investigated the meiotic role of Srs2, a multi-functional DNA helicase/translocase that destabilises Rad51-DNA filaments and is thought to regulate strand invasion and prevent hyper-recombination during the mitotic cell cycle. We find that Srs2 activity is required for normal meiotic progression and spore viability. A significant fraction of srs2 mutant cells progress through both meiotic divisions without separating the bulk of their chromatin, although in such cells sister centromeres often separate. Undivided nuclei contain aggregates of Rad51 colocalised with the ssDNA-binding protein RPA, suggesting the presence of persistent single-strand DNA. Rad51 aggregate formation requires Spo11-induced DSBs, Rad51 strand-invasion activity and progression past the pachytene stage of meiosis, but not the DSB end-resection or the bias towards interhomologue strand invasion characteristic of normal meiosis. srs2 mutants also display altered meiotic recombination intermediate metabolism, revealed by defects in the formation of stable joint molecules. We suggest that Srs2, by limiting Rad51 accumulation on DNA, prevents the formation of aberrant recombination intermediates that otherwise would persist and interfere with normal chromosome segregation and nuclear division. / Biotechnology and Biological Sciences Research Council (BB/K009346/1)

Manifestation visible des cross-overs génétiques, les chiasmata lient les paires de chromosomes homologues afin de les orienter correctement sur le fuseau méiotique en Métaphase et Anaphase I. Ils résultent d'un processus complexe et étroitement régulé impliquant l'induction de cassures double-brins et de leur réparation par l'invasion d'un duplex d'ADN homologue faisant office de modèle. La recombinaison est ainsi essentielle pour le synapsis et la ségrégation correcte des chromosomes méiotiques à la première division méiotique, et pour la génération de la variabilité génétique. Bien que les processus permettant à un chromosome de s'apparier seulement à son homologue ne soient pas complètement élucidés, l'appariement des chromosomes homologues est étroitement lié à la recombinaison catalysée par les enzymes d'échange de brins d'ADN RAD51 et DMC1. Ces deux protéines ont des capacités très similaires in vitro, mais sont fonctionnellement distinctes in vivo.La première partie de ma thèse montre l'impact de l'élimination de l'activité d'échange de brins de RAD51 dans la méiose d'Arabidopsis, tout en conservant sa fonction de facteur accessoire pour l'action de DMC1. La recombinaison peut donner lieu à des cross-over (CO) et non-cross-over (NCO) et la recombinase spécifique de la méiose DMC1 a été jugée particulièrement importante dans la production de CO interhomologue. Des résultats récents suggèrent fortement toutefois que DMC1 est la seule recombinase active dans la méiose et doit donc être responsable des résultats de CO et NCO. Etant donné qu'environ 95% de la recombinaison méiotique homologue dans Arabidopsis n'entraîne pas de cross-overs interhomologues, Arabidopsis est un modèle particulièrement sensible pour tester l'importance relative des deux protéines - même des effets mineurs sur la population d'événements non-cross-over devraient produire des effets détectables sur les cross-overs. DMC1 catalyse la réparation de toutes les cassures d'ADN méiotiques en présence d'une protéine RAD51 catalytiquement inactive (fusion RAD51-GFP), et les résultats de mon travail montrent que cela n'a pas d'effet détectable sur les taux relatifs de recombinaison de CO et de NCO : à la fois localement, à l'échelle du chromosome et du génome. Et non plus sur la progression de la division méiotique. Ce travail a abouti à une publication dans le journal PLoS One (Singh G, Da Ines O, Gallego ME & White CI (2017) Analyse de l'impact de l'absence d'activité d'échange de brins de RAD51 dans la méiose d'Arabidopsis PLoS ONE 12: e0183006- 16).Des publications antérieures montrent une synapsis homologue partielle et incomplète en l'absence de rad51 et xrcc3 dans la méiose d'Arabidopsis. Cela s'accompagne de la présence de nombreuses fibres courtes ZYP1 dans ces noyaux, ce qui pourrait indiquer de faibles longueurs de complèxe synaptonémale (SC). Ce synapsis partielle dépend à la fois de SPO11 et de DMC1 et implique des péricentromères, montrant que DMC1 est capable (au moins partiellement) d'entraîner le synapsis dans les péricentromères en l'absence de RAD51. Afin de mieux caractériser ceci et pour tester l'hypothèse que les fibres ZYP1 courtes montrent la présence d'une initiation de SC à ces sites, j'ai méné des expériences d'immunofluorescence et d'imagerie SIM. Utilisant un coloration DAPI et les antiséra ASY1, ZYP1 et CENH3, j'ai conduite des analyses cytogénétiques de le synapsis dans les mutants rad51, xrcc3 et des plantes sauvages. Ces travaux faisaient l'objet de la deuxième partie de mes travaux de thèse. Dans les plantes mutantes, j'observe effectivement des fibres courtes ZYP1 comprenant des centromères, mais elles ne sont pas la règle, ce qui signifie que le synapsis ne commence pas nécessairement à des centromères ou des péricentromères. (...) / The visible manifestation of genetic crossing-over, chiasmata link homologous chromosome pairs to permit them to properly orient on the meiotic Anaphase I spindle. They are the result of an intricate and tightly regulated process involving induction of DNA double- strand breaks and their repair through invasion of a homologous template DNA duplex. Recombination is thus essential for the synapsis and accurate segregation of meiotic chromosomes at the first meiotic division, and in doing so, generates genetic variation. Although the processes permitting a chromosome to pair only with its homologue are not fully understood, successful pairing of homologous chromosomes is tightly linked to recombination catalysed by the DNA strand exchange enzymes RAD51 and DMC1. Both proteins share very similar capabilities in vitro, but are functionally distinct in vivo. The first part of my thesis shows the impact of eliminating the strand exchange activity of RAD51 in Arabidopsis meiosis, while retaining its function as an accessory factor for the action of DMC1. Recombination can give rise to both crossover (CO) and non-crossover (NCO) outcomes and the meiosis-specific recombinase DMC1 has been thought to be of particular importance in the production of inter-homolog CO. Recent results however suggest strongly that that DMC1 is the only active recombinase in wild-type meiosis and thus must be responsible for both CO and NCO outcomes. Approximately 95% of meiotic homologous recombination in Arabidopsis does not result in inter-homologue crossovers and Arabidopsis is thus a particularly sensitive model for testing the relative importance of the two proteins - even minor effects on the non-crossover event population should produce detectable effects on crossing-over. DMC1 catalyses repair of all meiotic DNA breaks in the presence of the catalytically inactive RAD51 (RAD51-GFP fusion) and the results of my work show that this has no detectable effect on the relative rates of CO and NCO recombination, both locally and chromosome- and genome-wide, nor on the progression of the meiotic division. This work has resulted in a publication in the journal PLoS One (Singh G, Da Ines O, Gallego ME & White CI (2017) Analysis of the impact of the absence of RAD51 strand exchange activity in Arabidopsis meiosis. PLoS ONE 12: e0183006–16).Previous publications show partial, incomplete homolog synapsis in the absence of rad51 and xrcc3 in Arabidopsis meiosis. This is accompanied by the presence of many short ZYP1 fibres in these nuclei, possibly indicating short stretches of Synaptonemal Complex (SC). The partial synapsis is both SPO11- and DMC1-dependent and involves peri-centromeres, showing that DMC1 is able to (at least partially) drive synapsis in peri-centromeres in the absence of RAD51. In an effort to better characterize this and to test the hypothesis that the short ZYP1 fibres show the presence of initiation of SC at these sites, immunofluorescence and SIM imaging with DAPI staining and ASY1, ZYP1 and CENH3 antisera were carried out for cytogenetic analyses of synapsis in rad51 and xrcc3 mutants and the WT in the second part of my thesis work. Although I do observe short ZYP1 fibres including centromeres in the mutants, these are not the rule, so synapsis does not necessarily begin at centromeres or peri-centromeres. The superresolution imaging does confirm the presence of stretches of 4-chromatid fibres in xrcc3 plants and this approach will be extended in future work of the group to probe the nature of the RAD51-independent partial meiotic chromosome synapsis.Finally, I have designed and built CRISPR/CAS9 constructs with the aim of creating meiotic DSB hotspots at specific genomic loci. Taking advantage of single nucleotide polymorphism data, these constructs were designed to specifically cleave sites in the Arabidopsis Col-0 ecotype, and not in Ler-0 plants. (...)

No description available.

Repair of DNA double-strand breaks is critical to the maintenance of genomic stability, and failure to repair these DNA lesions can cause loss of chromosome telomeric regions, complex translocations, or cell death. In humans this can lead to severe developmental abnormalities and cancer. A central pathway for double-strand break repair is homologous recombination (HR), a mechanism that operates during the S and G2 phases of the cell cycle and primarily utilizes the replicated sister chromatid as a template for repair. Most knowledge of HR is derived from work carried out in prokaryotic and eukaryotic model organisms. To probe the HR pathway in human cells, we performed a genome-wide siRNA-based screen; and through this screen, we uncovered cellular functions required for HR and identified proteins that localize to sites of DNA damage. Among positive regulators of HR, we identified networks of pre-mRNA-processing factors and canonical DNA damage response effectors. Within the former, we found RBMX, a heterogeneous nuclear ribonucleoprotein (hnRNP) that associates with the spliceosome, binds RNA, and influences alternative splicing. We found that RBMX is required for cellular resistance to genotoxic stress, accumulates at sites of DNA damage in a poly(ADP-ribose) polymerase 1-dependent manner and through multiple domains, and promotes HR by facilitating proper BRCA2 expression. Screen data also revealed that the mammalian recombinase RAD51 is commonly off-targeted by siRNAs, presenting a cautionary note to those studying HR with RNAi and highlighting the vulnerability of RNAi screens to off-target effects in general. Candidate validation through secondary screening with independent reagents successfully circumvented the effects of off-targeting and set a new standard for reagent redundancy in RNAi screens.

Homologous recombination (HR) is a faithful mechanism for the repair of double-stranded DNA breaks (DSBs) and plays a critical role in maintaining the integrity of genomic DNA. The product of the Breast Cancer Susceptibility 2 (BRCA2) gene functions as a recombination mediator in HR-directed repair of DSBs. BRCA2 interacts directly with RAD51, the central recombinase of HR, through highly conserved repetitive motifs of 30-40 amino acids, named BRC repeats, and regulates the formation of the RAD51-ssDNA nucleoprotein filament. There is significant variability in the number of BRC repeats among taxa. However, all mammalian BRCA2 orthologs have eight BRC repeats, which display different characteristics in in vitro studies of RAD51-ssDNA nucleoprotein filament. To test the importance of the number of BRC repeats and to evaluate the role of individual BRC repeats in HR, BRCA2 variants bearing different combinations of BRC repeats were generated using BAC-recombineering, expressed in murine hybridoma cells, and assayed for the ability to stimulate HR using a gene targeting assay. The BRCA2 variant bearing BRC repeats 1 to 4 decreased the efficiency of HR and increased the level of Rad51 protein, whereas the BRCA2 variant bearing BRC repeats 5 to 8 significantly stimulated HR, but had no effect on the level of Rad51. These results supported the hypothesis that BRC repeats are not functionally equivalent, but rather have different, perhaps reinforcing functions in HR. / Canadian Institutes of Health Research