Natural and synthetic heterocyclic cations that bind to the DNA minor groove have demonstrated effectiveness as therapeutic agents for cancer, parasitic and viral diseases, as well as powerful probes for use to extend our fundamental understanding of DNA molecular recognition. Crystal and NMR structures with a variety of minor groove binding compounds have shed light on the structural varieties of these systems, the important solvent molecules in the complexes, and the induced fit effects for binding of both DNA and the bound small molecule. Topics of specific importance in DNA recognition are the development of a greater variety of cell-permeable minor groove agents that have increased DNA binding sequence selectivity. In this dissertation, the structural and energetic basis of the interaction between DNA and minor groove binders has been systematically investigated. A set of powerful and complementary biophysical methods have been used: gel electrophoresis with ligation ladder assay, circular dichroism, mass spectrometry, surface plasmon resonance and isothermal titration calorimetry have been applied to determine the binding stoichiometry, binding affinity, kinetics and thermodynamics, and also the structural influence that minor groove binders can have on DNA. The results of several minor groove complexes clearly show that based on DNA sequences, minor groove binders can have multiple binding modes and consequently affect the geometry of DNA minor groove and the overall DNA curvature in distinct manners. In addition, the binding enthalpy of a minor groove binder is essentially salt concentration and binding mode independent. Besides the investigation of DNA-minor groove binder complex, the binding and inhibition of transcription factor PU.1 has also been studied. The highly positive charged PU.1 targets DNA by inserting an α-helix in the major groove of the 5’-GGAA-3’ site, and displays a strong salt concentration dependency. A set of minor groove binders have been rationally designed based on the high-affinity DNA sequence for PU.1 to target the flanking sequences of the 5’-GGAA-3’ site. They display a structure-related PU.1 inhibition efficacy. This work demonstrates that minor groove binders are capable of modulating PU.1 by targeting the opposite groove and supports future efforts to develop agents for other transcription factors.

Les facteurs de transcription FLI-1 et SPI-1/PU.1 appartiennent à la famille ETS et reconnaissent le même motif sur l’ADN GGAA. Leur activation est observée de manière récurrente dans les érythroleucémies murines induites par le virus de Friend. Ces observations suggèrent un rôle crucial de ces deux facteurs dans la transformation de la lignée érythrocytaire potentiellement par la dérégulation de gènes cibles communs. Mon travail de thèse a consisté à tester la contribution de ces deux facteurs au phénotype des cellules érythroleucémiques et à rechercher les gènes cibles directs communs.Nous avons pu montrer que FLI-1 et SPI-1/PU.1 ont des contributions additives au phénotype des cellules érythroleucémiques surexprimant les deux facteurs. Par une approche transcriptomique, nous avons identifié une grande proportion de gènes cibles directs communs à FLI-1 et à SPI-1/PU.1 impliqués dans différentes étapes de la biogenèse des ribosomes. La déplétion de ces facteurs induit une réduction de la biogenèse des ribosomes qui n’induit pas de stress ribosomique stabilisant p53. Néanmoins, nous avons mis en évidence une contribution spécifique de RPL11, un médiateur essentiel du stress ribosomique, à la différenciation des cellules érythroleucémiques induites par l’absence de ces facteurs.Nous avons mené en parallèle l’inventaire par ChIP-Seq des sites de recrutement de FLI-1 et de SPI-1/PU.1 sur le génome entier de 3 lignées érythroleucémiques indépendantes. Cette stratégie nous a permis de montrer que les régions de recrutement communes sont la conséquence de la proximité de consensus spécifiques et distincts et du recrutement de FLI-1 et de SPI-1/PU.1 sur leur propre consensus. / The transcription factors FLI-1 and SPI-1/PU.1 belong to the ETS family and recognize the same DNA motif GGAA. Their activation is recurrently observed in murine erythroleukemia induced by Friend virus. These observations suggest a crucial role of these two factors in erythroid lineage transformation potentielly by deregulating common target genes. My thesis work consisted of testing both factors contribution to the phenotype of erythroleukemia cells and of searching for common direct target genes.We showed that FLI-1 and SPI-1/PU.1 have additive contributions to the phenotype of erythroleukemia cells overexpressing both factors. By a transcriptomic approach, we identified a high proportion of common direct target genes of FLI-1 and SPI-1/PU.1 involved in ribosome biogenesis at different levels. The déplétion of these factors induced a decrease of ribosome number which doesn’t induce a ribosomal stress leading to the p53 stabilization. However, we highlighted a specific contribution of RPL11, an essential ribosomal stress médiator, in erythroleukemia cell differentiation induced by depletion of both factors. In parallel, we mapped at whole génome scale by ChIP-Seq the recruitment site of FLI-1 and SPI-1/PU.1 in 3 independent erythroleukemia cell lines. This strategy allowed us to show that the common recruitment régions are the conséquence of a very close association of clearly distinct and specific consensus binding sites for FLI-1 and SPI-1/PU.1 and that each of those factor sis recruited to its own consensus.