In this thesis we have attempted to study four basic aspects of DNA-protein interactions: Affinity, specificity, accessibility and kinetics. With NF-kB as our model transcription factor, we wanted to investigate how a particular dimer recognizes a specific binding sequence? How fast are these interactions? And finally, how does the NF-kB interact with it binding site in the chromatin context? Specificity of NF-kB-DNA interactions has recently come into focus after it was shown that these dimers can bind to the sequences which do not fall into the NF-kB general consensus motif. We studied seven such sequences for their specificity for four NF-kB dimers. Our results show that p50 homodimers are least discriminative and can bind specifically to all these sequences. While as, RelA homodimers were highly discriminative and did not bind to most of these nontraditional sequences. We used two different methods to measure binding affinities: traditional gel mobility shift assay (EMSA) and a novel technique called as UV laser footprinting. Our results show that UV laser footprinting is the better method to determine the binding constants.For studying the dynamics of NF-kB-DNA binding, we combined UV laser footprinting with stopped flow device. This combination, not only give us one base pair resolution but also milli-second time resolution. Using p50 homodimers as a model transcription factor, we showed that the binding of this factor follows a two-step mechanism. First step involves the fast recognition of the sequence and second step follows a slower kinetics most likely for the stabilization of the complex. Our experiments suggest that flanking sequences play a role in the recognition and stabilization process of the complex formation.Finally, we also studied the accessibility of nucleosomes to NF-kB. Our in vitro data sheds light on the in vivo requirements for the alterations in chromatin structure necessary for the productive binding of NF-kB. These include either a removal of H2A-H2B dimers from the nucleosome and/or chromatin remodeler induced relocation of the histone octamer.Our data sheds light on the in vivo requirements for the alterations in chromatin structure necessary for the productive binding of NF-kB. We hypothesize that some factors like PU.1 might be able to target the chromatin remodeling/dimer eviction machinery to particular nucleosomes and lead to productive binding of NF-kB.

Dans cette thèse nous avons étudié quatre aspects fondamentaux de l’interaction ADN-protéine, notamment : l’affinité, la spécificité, l’accessibilité et la cinétique. En particulier, nous avons adressé les questions suivantes : comment un dimer du facteur de transcription NF-kB reconnait spécifiquement sa séquence d’ADN-cible, quelle est la rapidité de ces interactions, comment NF-kB interagit avec son site de fixation dans le contexte de la chromatine? Récemment, la spécificité de l’interaction NF-kB – ADN a reçu une attention particulière après l’observation que NF-kB peut se lier à des séquences qui n’entrent pas dans la classification de ses motifs « consensus ». Nous avons étudié la spécificité d’interaction de sept de ces motifs avec quatre dimers de NF-kB. Nos résultats montrent que le homo-dimer p50 sont les moins discriminatives et peuvent s’associer spécifiquement avec ces sept séquences. Par contre, les homo-dimers RelA se sont révélés hautement discriminatives ne pouvant pas s’associer spécifiquement avec ces séquences. Pour mesurer l’affinité de l’interaction nous avons utilisés deux méthodes distinctes : le traditionnel gel de retard (EMSA) et une nouvelle technique – « l’empreinte » au laser UV. Nos résultats montrent que le deuxième approche est plus approprié pour la mesure des constantes spécifiques de dissociation.Pour étudier la dynamique de l’interaction NF-kB – ADN, nous avons couplé l’empreinte au laser UV avec un appareil de mélange-rapide à façon. Cette combinaison nous a permis d’atteindre une résolution spatiale d’un nucléotide et temporaire de quelques millisecondes. Nous avons montré que l’homo-dimer p50 s’associe avec sa séquence-cible (MHC) H2 en suivant une cinétique à 2 pas. Le premier, de durée ~100 ms, reflète une reconnaissance initiale rapide, tandis que le deuxième, de durée ~1s, reflète une stabilisation lente du complexe. Nos expériences suggèrent aussi que les séquences voisines du site de reconnaissance jouent aussi un rôle dans la stabilisation du complexe.Finalement, nous avons étudié aussi l’accessibilité du nucléosome pour le NF-kB. Nos données in vitro montre que l’invasion spécifique de l’ADN à l’intérieure du nucléosome par NF-kB nécessite une perturbation majeure de la structure du nucléosome telle que l’éviction d’au moins un dimer d’histones H2A-H2B. / In this thesis we have attempted to study four basic aspects of DNA-protein interactions: Affinity, specificity, accessibility and kinetics. With NF-kB as our model transcription factor, we wanted to investigate how a particular dimer recognizes a specific binding sequence? How fast are these interactions? And finally, how does the NF-kB interact with it binding site in the chromatin context? Specificity of NF-kB-DNA interactions has recently come into focus after it was shown that these dimers can bind to the sequences which do not fall into the NF-kB general consensus motif. We studied seven such sequences for their specificity for four NF-kB dimers. Our results show that p50 homodimers are least discriminative and can bind specifically to all these sequences. While as, RelA homodimers were highly discriminative and did not bind to most of these nontraditional sequences. We used two different methods to measure binding affinities: traditional gel mobility shift assay (EMSA) and a novel technique called as UV laser footprinting. Our results show that UV laser footprinting is the better method to determine the binding constants.For studying the dynamics of NF-kB-DNA binding, we combined UV laser footprinting with stopped flow device. This combination, not only give us one base pair resolution but also milli-second time resolution. Using p50 homodimers as a model transcription factor, we showed that the binding of this factor follows a two-step mechanism. First step involves the fast recognition of the sequence and second step follows a slower kinetics most likely for the stabilization of the complex. Our experiments suggest that flanking sequences play a role in the recognition and stabilization process of the complex formation.Finally, we also studied the accessibility of nucleosomes to NF-kB. Our in vitro data sheds light on the in vivo requirements for the alterations in chromatin structure necessary for the productive binding of NF-kB. These include either a removal of H2A-H2B dimers from the nucleosome and/or chromatin remodeler induced relocation of the histone octamer.Our data sheds light on the in vivo requirements for the alterations in chromatin structure necessary for the productive binding of NF-kB. We hypothesize that some factors like PU.1 might be able to target the chromatin remodeling/dimer eviction machinery to particular nucleosomes and lead to productive binding of NF-kB.