Chromatin remodelers utilize the energy derived from ATP hydrolysis to mobilize nucleosomes. ISWI remodelers mobilize and evenly space nucleosomes to regulate gene expression. ISW2, an ISWI remodeler in yeast, has been shown to reposition nucleosome near promoter regions and represses both mRNA and antisense non coding RNA transcription. ISW2 is composed of four subunits and the catalytic Isw2 subunit consists of several conserved domains. The highly conserved ATPase domain is present at the N-terminus whereas the conserved HAND, SANT and SLIDE domain are towards the carboxyl terminal end of Isw2. Nucleosome mobilization by ISW2 requires both extranucleosomal DNA and the N-terminal tail of histone H4. DNA crosslinking and peptide mapping revealed that the ATPase domain contacts nucleosome two helical turns away (SHL2) from dyad to a site close to the H4 tail, whereas the HAND, SANT and SLIDE domain contact a 30bp stretch of DNA comprising the edge of nucleosome and ~20bp of extranucleosomal DNA. The ATPase domain and the C-terminal domains were investigated for their role in regulation of ISW2 activity both in-vitro and in-vivo. It appears that there are distinct modes of ISW2 regulation by these domains. Mutation of a patch of five acidic amino acids on the region of ATPase domain that contact SHL2 was found to be crucial for both ISW2 remodeling and nucleosome stimulated ATPase activity. Acidic patch mutant ISW2 was unable to mobilize nucleosome or hydrolyze ATP in absence of H4 tail. This indicates that the region of ATPase domain contacting nucleosome at SHL2 and H4 tail act in two separate and independent pathways to regulate ISW2 remodeling. Both HAND and SLIDE domain were shown to crosslink entry/exit site and linker DNA respectively. The roles of C-terminal domains were investigated either by deletion of the individual domain or mutation of conserved basic residues on the surface of these domains that are suspected to interact extranucleosomal with DNA. Deletion of HAND domain had minimal effect on in vitro ISW2 activity, however whole genome transcription analysis revealed one key role of this domain in ISW2 regulation. In absence of HAND domain, ISW2 had minimal role on repression of genes that were RPD3 (co-factor) dependent, however significantly derepressed genes that were RPD3 independent. At these loci, nucleosome positions were altered and ISW2 recruitment was reduced in absence of a functional HAND domain. Thus the HAND domain regulates recruitment and remodeling of ISW2 at those genes where ISW2 acts independent of other cofactors. The SANT domain, C-terminal to HAND domain, appears to control the "step size" of nucleosome remodeling and was found to be required for processive nucleosome remodeling by ISW2. Both H4 tail and SANT domain appear to control two distinct stages of ISW2 remodeling. A long alpha helical spacer separates SANT domain from SLIDE domain. SLIDE domain was found to be the protein-protein interaction domain that interacts with accessory Itc1 subunit to maintain ISW2 complex integrity. The two ways by which SLIDE domain regulate ISW2 is by binding or recruitment of ISW2 to promoter regions and additionally by binding independent regulation of both ATPase and remodeling activity. The remodeling mechanism of ISW2 was further compared with another ISWI type remodeler in yeast, Isw1a; using time resolved nucleosome remodeling combined with high resolution site specific histone DNA crosslinking at six different nucleosomal positions to track the movement of the nucleosomes. Nucleosome remodeled by the same remodeler showed discontinuous nucleosome movement between two tracking points indicating formation of small "bulges". One key difference in remodeling mechanism was that although both ISW2 and Isw1a moved nucleosomes towards longer linker DNA, only Isw1a remodeled nucleosomes "backtracked" ~11bp during remodeling. Backtracking of remodeling was prominently observed at nucleosomal regions in close proximity to translocase binding sites suggesting the potentially different mechanisms shared by similar remodeling complexes.

ATAD2 est un facteur très conservé mais peu caractérisé qui possède différents domaines fonctionnels : un domaine AAA ATPase et un bromodomaine (BRD). Normalement, ATAD2 est exprimé fortement dans les cellules germinales males ainsi que dans les cellules souches embryonnaires (cellules ES). De plus, la surexpression de cette protéine a été détectée dans de nombreux cancers. Il a été montré qu'ATAD2 agit comme co-activateur des récepteurs aux androgènes et aux œstrogènes. Cette protéine semble aussi agir comme co-facteur de l’oncogène Myc et joue un rôle dans la voie pRb/E2F. La surexpression d’ATAD2 prédit un mauvais prognostic dans les cancers du poumon et du sein. Toutes ces caractéristiques font d'ATAD2 un candidat de choix comme biomarqueur pronostic et une cible potentielle pour des agents thérapeutiques dans le cadre de cancers agressifs.Dans ce projet de thèse, nous montrons que hATAD2 interagit avec l'histone H4 acétylée via son bromodomaine, et que le domaine ATPase est responsable de la multimérisation d’ATAD2 et permet au BRD d’interagir avec les lysines acétylées dans les cellules. Des investigations complémentaires, comprenant notamment des études structurales, montrent que le BRD d'ATAD2 est responsable de son interaction spécifique avec la forme acétylée de la lysine 5 de l'histone H4. Nous avons aussi analysé le domaine AAA ATPase et découvert des éléments qui contrôlent son rôle dans la multimérisation des protéines. De plus, nous avons étudié ATAD2 dans la lignée de cellules cancéreuses pulmonaires, H1299, ainsi que dans les cellules ES et démontré que ce facteur est essentiel pour la prolifération des cellules en l'absence des facteurs de croissance. En combinant des approches ChIP-seq, ChIP-protéomics et RNA-seq dans les cellules ES, nous avons montré qu'ATAD2 est très enrichi dans les régions à haute activité transcriptionnelle et maintient la chromatine accessible pour les facteurs impliqués dans les activités de la chromatine. Ces données indiquent qu'ATAD2, dans son contexte physiologique, assure un rôle essentiel dans les activités générales de la chromatine, telles que la transcription, en maintenant l'accessibilité de la chromatine pour les facteurs de transcription.Enfin, afin de caractériser la structure d’ATAD2 et celle de son homologue dans Schizosaccharomyces pombe, ABOI, différents fragments contenants le domaine AAA ATPase ont été produits dans des bactéries ainsi que dans des cellules d'insectes en utilisant des vecteurs d’expression de baculovirus. Les conditions de production de fragments solubles ont été établies et certains de ces fragments ont été purifiés. Néanmoins, l’obtention de la structure cristalline de l'ATAD2 nécessite des travaux supplémentaires. / ATAD2 is an evolutionarily conserved but poorly characterized factor that bears different types of func¬tional domains: an AAA ATPase domain and a bromodomain (BRD). ATAD2 is normally highly ex¬pressed in male germ cells and in embryonic stem cells (ESC), however the overexpression of this protein has been detected in a large variety of independent cancers. ATAD2 is proposed to act as a co-activator of androgen and estrogen receptors and in addition, this protein also seems to act as a co-factor for Myc oncogene and plays a role in the pRb/E2F pathway. Moreover, the overexpression of ATAD2 predicts poor prognosis in lung and breast cancers. All of these characteristics make ATAD2 a valuable prognosis biomarker and a promising therapeutic target in aggressive cancers.Herein, we show that hATAD2 binds to acetylated H4 tail through its BRD, and that its ATPase domain enables ATAD2 multimerization, affecting the ability of the BRD to bind acetylated lysine in cells. Additional investigations, including structural studies, show that ATAD2’s BRD is responsible for its specific interaction with acetylated lysine 5 of histone H4. We have also functionally analyzed the AAA ATPase domain and discovered elements that control its role in protein multimerization. In addition, we studied ATAD2 in ESC and in the H1299 lung cancer cell line, and demonstrated that this factor has crucial roles in cell proliferation in the absence of growth factors. Moreover, by using a combination of ChIP-seq, ChIP-proteomics and RNA-seq experiments in ESC, we found that ATAD2 is highly enriched in regions with high transcriptional activity and that it keeps chromatin accessible for chromatin templated factors. These data indicate that ATAD2 in its physiological context ensures a critical role in general chromatin-templated activities, such as transcription, by maintaining the accessibility of chromatin for transcription factors. Finally, in order to structurally characterize either ATAD2 or its homologue in Schizosaccharomyces pombe, ABOI, different fragments containing the AAA ATPase domain were produced in bacteria as well as in insect cells using baculovirus expression vectors. Conditions to produce soluble fragments were established and some of these fragments were purified. Nonetheless, solving the crystal structure of ATAD2 still requires further investigation.