Trimethylation of histone H3 lysine 9 (H3K9me3) and DNA methylation mark heterochromatin, contributing to gene silencing and normal cellular functions. My research investigated the control of H3K9me3 and DNA methylation in the filamentous fungus Neurospora crassa. The H3K9 methyltransferase complex, DCDC, consists of DIM-5, DIM-7, DIM-9, DDB1, and CUL4. Each component of DCDC is required for H3K9me3. The DIM-9/DDB1/CUL4 subunits are reminiscent of known cullin E3 ubiquitin ligases. I showed that core features of CUL4-based E3 ubiquitin ligases are not required for H3K9me3 and DNA methylation in Neurospora. H3K9me3 is bound by heterochromatin protein 1 (HP1) to recruit the DIM-2 DNA methyltransferase and the HCHC histone deacetylase complex. HCHC consists of HP1, CDP-2, HDA-1, and CHAP. Both HP1 and CDP-2 harbor conserved chromodomains that bind H3K9me3, and CHAP contains two putative AT-hook domains that bind A:T-rich DNA. To test the contributions of these domains to HCHC function, I deleted the chromodomains of HP1 and CDP-2. Deletion of the HP1 chromodomain resulted in a reduction of DNA methylation, which was not exacerbated by deletion of the CDP-2 chromodomain. A strain with deletions of chap and the HP1 chromodomain showed a DNA methylation phenotype comparable to the loss of the HDA-1 catalytic subunit. These findings support a model in which recognition of H3K9me3 and A:T-rich DNA by HP1 and CHAP, respectively, are required for proper HCHC function. To examine the relationships between H3K9me3, DNA methylation, and histone acetylation, I utilized in vivo protein tethering of core heterochromatin components. The requirement of DIM-7 for native heterochromatin, previously implicated in localizing the H3K9 methyltransferase DIM-5, was not bypassed by DIM-5 tethering, indicating that DIM-7 has additional roles within the DCDC. Artificial localization of the HCHC histone deacetylase, by tethering HP1 or HDA-1, resulted in induction of H3K9me3, DNA methylation, and gene silencing, but silencing did not require H3K9me3 or DNA methylation. HCHC-mediated establishment of H3K9me3 was not required for de novo heterochromatin formation at native heterochromatic loci suggesting a role in heterochromatin spreading. Together, this work implicates HDA-1 activity as a key driver of heterochromatin spreading and silencing. This dissertation includes previously published co-authored material.

Les télomères sont des structures critiques situées aux deux extrémités des chromosomes linéaire et empêchent que ces derniers soient dégradés, sujet à des réparations aberrantes ou subissent des recombinaisons homologues. Ces régions particulières sont compactées sous une forme de chromatine très dense que l’on nomme l’hétérochromatine. Une étape clef dans l’établissement et la maintenance de l’hétérochromatine est la tri-méthylation de l’histone H3 en lysine 9 (H3K9me3). Cette marque sert de point d’ancrage pour un certain nombre de protéines qui vont ensuite permettre de propager un environnement répressif. Les télomères étant hétérochromatiques, sont enrichis en H3K9me3 qui est déposée par l’histone methyltransférase SUV39H. Pourtant, la relation entre cette modification H3K9me3 et les télomères a été très peu étudiée. Il a été suggéré que cette marque est impliquée dans l’homéostasie de la taille des télomères, mais la fonction de H3k9me3 reste largement inconnue. Pendant ma thèse, j’ai empêché la catalyse de H3K9me3 aux télomères en supprimant l’histone methyltransférase responsable. Comme modèle d’étude j’ai travaillée sur des cellules embryonnaires de souris (mESC), étant donné que ces cellules de mammifère ont de très longs télomères hétérochromatiques. De façon surprenante, j’ai découvert que c’est l’enzyme SETDB1 qui installe H3K9me3 aux télomères de mESC. J’ai purifiée et comparée les changements de la composition moléculaire de télomères n’ayant plus d’histones trimethylés en lysine 9 (dans des cellules mESC ou SETDB1 est « knocked-out » (KO)) à des télomères sauvages en utilisant la technique de PICh (Proteomics of Isolated Chromatin segments) sur des cellules cultivées en SILAC (Stable Isotope Labelling Amino Aacids). J’ai montré que H3K9me3 contrôle le recrutement de chaperonnes d’histones; et de façon plus surprenante, cette marque semblerait contrôler l’élongation et/ou inhiber la terminaison de la transcription des télomères. En effet les télomères sont transcrits en un long ARN non-codant appelé TERRA. Nos données préliminaires suggèrent que cette voie n’est pas restreinte aux télomères mais aussi à d’autres gènes qui sont sous le contrôle de SETDB1. Il semblerait que la trimethylation de H3K9 dans le corps du gène est nécessaire pour maintenir la processivité de l’ARN polymérase II. Mes données suggèrent que SETDB1 contrôle la trimethylation de H3K9 aux télomères et dans certains corps de gènes ce qui est crucial pour la transcription générale dans les cellules embryonnaires de souris. / Telomeres are critical regions that protect chromosome ends from degradation or aberrant repair. These regions are assembled into heterochromatin. Trimethylation of histone H3 on lysine 3 (H3K9me3) is a biochemical modification found at telomeres and essential in the establishment and maintenance of constitutive heterochromatin. During my thesis, I investigated the function of H3K9 trimethylation. According to my data, this hallmark is deposited by SETDB1 at the telomeres in mouse embryonic stem cells. Using the quantitative PICh method, I showed that this mark controls the recruitment of histone chaperones. Heterochromatin is typically believed to repress gene expression but my data suggests that at telomeres, the H3K9me3 mark instructs transcriptional elongation and/or inhibit transcriptional termination of telomeres into the non-coding RNA TERRA. Preliminary data even suggest this pathway is not only restricted to telomeres but also to other genes under the control of SETDB1. It seems that H3K9 trimethylation in the gene body is necessary to maintain RNA polymerase II processivity. My data suggests SETDB1 controls H3K9 trimethylation at telomeres and gene bodies which is crucial for general transcriptional in mouse embryonic stem cells.

Chromatin diminution is defined as the elimination of DNA during the differentiation of early embryonic cells into pre-somatic cells. While it was first observed in the nematode Parascaris equorum, it also been identified in other parasitic nematodes, hagfish and copepods. In the copepod Mesocyclops edax, up to 90% of genomic DNA is eliminated during chromatin diminution. It was previously shown that the eliminated DNA contained highly repetitive heterochromatic sequences. Here, we digested pre- and post-diminution DNA with BamHI and produced small libraries of clones from each. Analyses revealed no decrease in low copy numbered sequences, such as transposable elements. Rather, both libraries are found to be surprisingly similar in all aspects analysed. Further comparison also demonstrated similarity of our libraries with the DNA sequences eliminated from Cyclops kolensis. Consequently, we suggest that M. edax eliminates portions of euchromatic DNA, in addition to the previously characterized satellite sequences.

Les centromères contribuent à garantir la distribution égale de l'ADN en mitose. Leur identité n'est pas codifiée par la séquence d'ADN, mais de manière épigénétique par le variant de l'histone H3 CENP-A. Dans des lignées cellulaires humaines transformées, CENP-A est incorporé au centromère par son chaperon HJURP, au début de la phase G1. Pendant ma thèse, j'ai utilisé le modèle murin pour étudier les particularités de la chromatine centromérique et son dysfonctionnement dans le cancer. J'ai montré que CENP-A est maintenu sur le génome paternel au cours de la spermatogenèse, contrairement aux autres histones, et peut constituer une marque transgénérationnelle du centromère. Nous avons généré une souris KO pour HJURP pour l'étudier in vivo, et avons détecté son amplification dans de multiples souches de souris. En parallèle, nous avons étudié l'interaction entre la dynamique des variants d'histone et la structure d'ordre supérieur de la chromatine centromèrique. Nous avons découvert que la réorganisation de l'hétérochromatine péricentrique au cours du cycle cellulaire contrôle les deux modes distinctifs d'incorporation des variants d'H2A et la stoechiométrie de CENP A. Pour explorer le lien entre la tumorigenèse et la surexpression de CENP A/HJURP dans des cancers humains, nous avons utilisé un modèle de transformation de fibroblastes murins embryonnaires. Dans le fond génétique nul pour p53 de ces cellules, la surexpression exogène des deux facteurs n'apportait pas un avantage prolifératif mesurable, mais leur accumulation était une conséquence de la transformation. Actuellement, nous analysons si cette surexpression contribue à augmenter la capacité de transformation. / Centromeres are genomic loci ensuring equal distribution of the two sets of chromosomes in mitosis. Their identity is not encoded in the underlying DNA sequence but specified epigenetically by the histone H3 variant CENP-A. In transformed human cell lines, CENP A is deposited at centromeres by the histone chaperone HJURP in a distinct window of the cell cycle. During my PhD I have taken advantage of the mouse model to address cell cycle and developmental features of centromeric chromatin, as well as its dysfunction in cancer.Using an organism-level approach, I could observe that contrary to most histones, CENP-A is retained on the paternal genome during spermatogenesis, acting as a transgenerational mark of the centromere. To study the role of HJURP in vivo, we generated a knockout mouse and discovered that its genomic locus underwent amplification in several mouse subspecies.In parallel, we addressed the crosstalk between histone variant dynamics and higher-order chromatin structure at the centromere, and revealed that the dynamic reorganization of pericentric heterochromatin during the cell cycle controls the distinct incorporation of H2A variants and CENP-A stoichiometry.Finally, to explore the connection between tumorigenesis and CENP-A/HJURP overexpression, recorded in a number of human cancers, we used a mouse embryonic fibroblast model of transformation. We determined that whereas their overexpression did not confer a measurable proliferative advantage in a p53-deficient background, CENP-A/HJURP upregulation was a consequence of transformation. Whether their accumulation has a functional role to enhance tumorigenesis in this system was further investigated.

L'endoribonucléase RNase L est essentiellement connu comme étant un acteur critique de l'immunité innée pour enrayer la progression d'une infection virale en clivant les ARN cellulaires. Son activité est régulée par de nombreux facteurs tels que la 2-5A et son inhibiteur, la RLI. Au cours de cette étude, nous avons démontré une implication de l'activité de la RNase L dans la régulation de la structure du domaine centromérique. Nous présentons dans ce manuscrit, les perturbations majeures engendrées par une augmentation ou une inhibition de l'activité de la RNase L représentées par une délocalisation de HP1-alpha et de CENP-C causant une déstructuration générale des chromosomes. Ces délocalisations de protéines centrales de la structure chromatinienne seraient causées par un défaut de la maturation des transcrits majeures péricentromériques lors d'une modulation de l'activité de la RNase L. Pour terminer, nous avons également identifié un potentiel trafic cyto-nucléaire empreinté par la RNase L. Nous proposons ainsi une fonction nucléaire inattendue de la RNase L par son implication dans la régulation des transcrits péricentromériques assurant l'intégrité structurale de la chromatine. / The endoribonuclease Latente (RNase L) is mostly known as a critical factor in the innate immunity during the cell's defence against a viral infection. The antiviral activity of RNase L which is characterize by it capacity of cleavage of viral RNA, is regulated by several factors like it activator the oligoadénylates 2-5A and his inhibitor RLI. In this manuscript, we have studied the role of the activity of RNase L in the regulation of the structure of centromeric domains. Our results show a general destructuration of chromosomes observed in cells over-expressing RNase L or RLI. These major aberrations are demonstrated by a delocalization of essentials proteins for the structure of chromatin: HP1-alpha and CENP-C. The mislocalization of these proteins could be provoked by a default in the maturation of major transcripts due to a modulation of the activity of RNase L. moreover, in this study, we have identified a mechanism regulating the cyto-nuclear shuttling of RNase L. therefore, we propose that a new nuclear function of RNase L: it's implication in the regulation of pericentromeric transcripts needed to stabilize the integrity of the structure of chromatin.

Pattern formation during development is guided by tightly controlled gene regulatory networks that lead to reproducible cell fate outcomes. However, stochastic choices are often employed to further diversify cell fates. These two mechanisms are closely interlinked in the mouse olfactory system, where stochastic expression of one of one out of >1,000 olfactory receptor (OR) genes is restricted to anatomical segments, or “zones”, organized along the dorsoventral axis of the olfactory epithelium (OE). Despite recent progress in understanding the processes underlying OR choice, the mechanism by which the dorsoventral position of an olfactory sensory neuron (OSN) dictates its OR repertoire has remained elusive and is the focus of this thesis. To gain insight into a possible mechanism I compared the transcriptomes, chromatin landscape, and nuclear architecture of cells isolated from ventral and dorsal zonal segments of the OE. I determined the developmental window in which cells become restricted in their zonal OR repertoire and found this coincided with both the deposition of heterochromatic histone marks H3K9me3 and H3K79me3 on OR genes and their coalescence into a multi-chromosomal compartment. Comparing heterochromatin levels and OR compartment composition in OSNs from different zones, I determined in each case OR genes with more dorsal indexes have higher levels of H3K9me3/H3K79me3 and thus become silenced, while OR genes with more ventral indexes have no heterochromatin and consequently are excluded from OR compartments. Thus, ORs that are “competent” for activation are relatively more accessible, while still being recruited into the OR compartment where they can interact with the proximally positioned enhancer hub. I found that this mechanism is regulated by Nfi family transcription factors that are expressed in a ventral (high) to dorsal (low) gradient in the OE. Deletion of Nfi A, B and X transforms the heterochromatin and OR compartmentalization in ventral OSNs to a more dorsal state, and shifts their preferred OR repertoire towards more dorsal ORs. This result implicates Nfi proteins as key regulators of zonal OR expression. Finally, I probed the nuclear architecture in single cells to look for the source of stochastic choice within zonal segments. I found high variability in inter-chromosomal OR compartment and enhancer hub composition between individual OSNs that stemmed from the unpredictable and variable positioning of chromosomes in the interphase nucleus. Overall, this thesis provides evidence for a mechanism of zonal OR choice that combines deterministic restrictions imposed by a gradient of Nfi with random inter-chromosomal contacts.

Les cellules souches embryonnaires de souris (ESCs) et les cellules souches de l'épiblaste (EpiSCs) représentent, respectivement, les états naïf et amorcé de la pluripotence et sont maintenues in vitro par des voies de signalisation spécifiques. De plus les ESCs cultivées dans un milieu sans sérum avec deux inhibiteurs (2i) sont décrites comme étant les plus naïves. Plusieurs études ont suggéré que chaque type de cellules pluripotentes est caractérisé par une organisation différente de l'épigénome. Nous présentons ici une étude comparative de l'état épigénétique et transcriptionnel des séquences satellites répétées péricentromériques (PCH) entre les ESCs (2i et sérum) et EpiSCs. Nous montrons que H3K27me3 au PCH est très dynamique et peut discriminer les ESCs en 2i des autres cellules souches pluripotentes. Alors que la transcription des séquences satellites est élevée dans les ESCs en sérum, elle est plus faible dans ESCs en 2i et encore plus réprimée dans les EpiSCs. La suppression de la méthylation de l'ADN ou d'H3K9me3 dans les ESCs conduit à un dépôt important de H3K27me3 au PCH, mais peu de changements transcriptionnels de ces séquences. En revanche, l'absence d'H3K9me3 dans les EpiSCs n'empêche pas la méthylation de l'ADN au PCH, mais induit la transcription de ces séquences. La conversion in vitro des ESCs en EpiSCs est plus longue que le passage des cellules de l'ICM en épiblaste in vivo. Cette inefficacité ne peut pas être expliquée par une mise en place retardée du nouveau réseau transcriptionnel. Pour conclure notre étude a révélé que les EpiSCs ont perdu de la plasticité par rapport au ESCs sur l'hétérochromatine ainsi que l’euchromatine, comme le montre la réduction des niveaux d'H3K9ac et des domaines bivalents, étant ainsi plus proche épigénétiquement de cellules somatiques que de la pluripotence naïve. / Mouse embryonic stem cells (ESCs) and epiblast stem cells (EpiSCs) represent naïve and primed pluripotency states, respectively and are maintained in vitro using specific signaling pathways. Furthermore, ESCs cultured in serum-free medium with two inhibitors (2i) are described as being the most naïve. Several studies have suggested that each pluripotent cell type is characterized by a different epigenome organization. Here we present a comparative study of the epigenetic and transcriptional state of centromeric and pericentromeric (CH/PCH) satellite repeats in ESCs (2i and serum ones) and EpiSCs. We show that the pattern of H3K27me3 at PCH is highly dynamic and discriminate 2i-ESCs from the other pluripotent stem cells. Whereas satellites transcription is high in serum-ESCs, it is lower in 2i-ESCs and even more repressed in EpiSCs. Removal of either DNA methylation or H3K9me3 in ESCs leads to enhanced deposition of H3K27me3 but few changes in satellite transcription. By contrast, in EpiSCs removal of H3K9me3 does not prevent DNA methylation at PCH but de-represses the satellite transcription. In vitro conversion from naive to primed pluripotency showed an important delay compared to the in vivo development of ICM cells into post-implantation epiblast. Such inefficiency cannot be explained by a delayed switch to the new transcriptional network. Altogether our study reveals that EpiSCs have lost the chromatin plasticity of ESCs on heterochromatin as well as euchromatin, as shown by the reduction of H3K9ac levels and bivalent domains, thus being closer to somatic cells in terms of epigenetics than naive pluripotency.

Les distorteurs de ségrégation méiotiques sont des éléments génétiques égoïstes qui favorisent leur propre transmission en manipulant la méiose à leur avantage. La diffusion dans les populations d’un distorteur lié au chromosome X (Sex-Ratio) provoque un excès de femelles et cela conduit à un conflit entre le chromosome X et les autres chromosomes. Ces conflits intra-génomiques sont d’importants moteurs de l’évolution des génomes. Mais, peu de choses sont connues sur la nature moléculaire et la fonction des éléments égoïstes Sex-Ratio. Le premier chapitre de cette thèse présente une synthèse actualisée sur les distorteurs de ségrégation méiotiques liés à un chromosome sexuel. Le second chapitre est consacré à l’identification et la caractérisation d’un élément distorteur du système Sex-Ratio « Paris » de Drosophila simulans, dans lequel deux éléments distorteurs liés au chromosome X provoquent ensemble la misségrégation des chromatides sœurs du chromosome Y lors de la méiose II. J’identifie à travers une cartographie génétique par recombinaison un des loci distorteur et je conduis une validation fonctionnelle de son implication dans la distorsion. Il s’agit d’un jeune gène qui évolue rapidement et appartient à une famille de gènes bien connus, impliquée dans la constitution de l’hétérochromatine. Ce gène a émergé par duplication il y a environ 15-22 millions d’années et a connu de façon indépendante de multiples duplications en cis, pseudogenizations, ou bien directement sa perte tout au long de son histoire évolutive. Cela suggère que ce gène pourrait avoir été impliquée dans de multiples conflits génétiques. Le dernier chapitre est consacré à une étude exploratoire de la diversité structurale des chromosomes Y en relation avec la distorsion de ségrégation méiotique du système « Paris ». Les résultats présentés dans ce manuscrit contribuent à augmenter les connaissances sur l’origine moléculaire des conflits génétiques et sur leur impact évolutif. / Segregation distorters are selfish genetic elements that promote their own transmission by subverting the meiotic process to their advantage. The spread of an X-linked distorter (Sex-Ratio) in populations results in an excess of females, which triggers a genetic conflict between the X chromosome and the rest of the genome. Such conflicts are important drivers of genome evolution, but little is known about the molecular nature and the function of the Sex Ratio selfish elements. The first chapter of this manuscript is a review of the current knowledge about X-linked segregation distorters. Then, I present my work on the « Paris » Sex Ratio system of Drosophila simulans, in which two distorter elements on the X chromosome co-operate to prevent Y chromosome sister chromatids segregation during meiosis II. I mapped a gene in one of the distorter loci and achieved the functional validation of its involvement in sex-ratio distortion. It is a young and rapidly evolving gene that belongs to a well-known gene family involved in chromatin state regulation. It emerged through duplication about 15-22 Myrs ago and has experienced multiple independant cis-duplications, loss or pseudogenization throughout its evolutionary history. This suggests that this gene could have been involved in multiple genetic conflicts. Finally, the last chapter is about an opening study of the strucural diversity of Y chromosomes in relation to « Paris » segregation distorter. These findings should help understanding the molecular basis of genetic conflicts and the evolutionary impact of heterochromatin regulation during meiosis.

Embryonic chromatin diminution, the selective excision of large amounts of heterochromatic DNA from presomatic cell lineages, provides an example of an unusually large augmentation of the germline genome and raises questions regarding the source of the increased amount of DNA and its relevance to the biology of the organism. DNA levels in adult germ cell nuclei of the copepod Cyclops kolensis were determined by DNA-Feulgen cytophotometry and compared with those of somatic nuclei of adults and both pre- and postdiminuted embryos from the same mothers. Almost 75 pg DNA/nucleus is excised by diminution, resulting in the return of each generation to the approximately 1 pg DNA/nucleus level found for adult soma. To account for the increase in DNA levels of germ cells observed here, we propose alternative hypotheses to the original model of chromatin diminution: (1) repetitive endocycles or (2) proliferation of genetic elements. Specific tests for these hypotheses using next-generation sequencing and quantitative cytophotometry, as well as the functional significance of germ cell DNA augmentation to the copepod, are discussed.

The segregation of progenitor somatic cells from those of the primordial germ cells that sequester and retain elevated levels of DNA during subsequent developmental events, poses an interesting, alternative pathway of chromosome behavior during the reproductive cycle of certain species of cyclopoid copepods and several other organisms. Separation of maternal and paternal chromosome sets during very early cleavages (gonomery) is often a feature following marked elevations of DNA levels in germ cells for some of these species. Here, we report on the accumulation of large amounts of DNA in germ line nuclei of both female and male juveniles and adults of a freshwater copepod, Mesocyclops edax (Forbes, 1890). We also report the robust uptake of 3H-thymidine by germ cells prior to gametogenesis in this species. By using cytophotometric analysis of the DNA levels in both germ line cells and somatic cells from the same specimens we demonstrate that germ cell nuclei accumulate high levels of DNA prior to the onset of gametogenesis. These elevated amounts coincide with the levels of heterochromatic DNA discarded during chromatin diminution. A new model is proposed of major cytological events accompanying the process of chromatin diminution in M. edax.