DNA nucleases are essential for various biological functions such as replication, recombination, and repair. Restriction endonucleases (REs) are excellent model system for the investigation of DNA recognition and specificity. SgrAI is a type IIF RE that cuts an 8 base pair primary sequence. In addition to its primary cleavage activity it also cleaves secondary sequences, but only appreciably in the presence of the primary sequence. The longer flanking DNA exhibits much greater activated DNA cleavage by SgrAI (>1000 fold activation by secondary site). Interestingly, the asymmetric cleavage seen in one of the two types of secondary site DNA is lost upon activation of SgrAI, suggesting a loss of communication between DNA recognition and activity upon specificity expansion. The structure of SgrAI bound to 22-1HT supports the cryoelectron microscopy structure of activated, oligomeric SgrAI highlighting the significance of the contacts made by the flanking DNA and the role played by N-terminal domain contacts in forming the run-on oligomer. The biological study suggests that the run-on oligomer formation sequesters the host DNA from being cleaved by the activated SgrAI complex. The DNA sequence binding, cleavage preference, and the structure of K96A SgrAI were determined. Unexpectedly, this mutation did not alter the structure of the enzyme, nor did it result in an enzyme lacking sequence preference at the 7ᵗʰ position. Instead, the largest effect of the mutation appears to be in making the enzyme more specific such that it fails to cleave either type of secondary site. It may be that the K96 side chain is required to distort the non YG sequences (specifically GG and TC) of secondary site DNA for proper positioning in the enzyme active site upon activation and specificity expansion. The crystal structure of Mk0566, XPG homologue from M. kandleri, was solved to 2.48 Å resolution and was found to be very similar to that of human FEN-1 and to other archaeal FEN-1/XPG homologues. These results suggest that the main biological role of Mk0566 is in DNA replication; however, they do not preclude involvement in a modified form of nucleotide excision repair.

The P22 c2 repressor protein (P22R) binds to DNA sequence-specifically and helps direct the temperate lambdoid bacteriophage P22 to the lysogenic developmental pathway. To gain insight into its DNA binding mechanism, we solved the 1.6 Å x-ray structure of the N-terminal domain (NTD) of P22R in a complex with a DNA fragment containing the synthetic operator sequence [d(ATTTAAGATATCTTAAAT)]2 This operator has an A-T at position 9L and T-A at position 9R and is termed DNA9T. Van der Waals interactions between protein and DNA appear to confer sequence-specificity. The structure of the P22R NTD – NA9T complex suggests that sequence-specificity arises substantially from interaction of a valine with a complementary binding cleft on the major groove surface of DNA9T. The cleft is formed by four methyl groups on sequential base pairs of 5' TTAA 3'. The valine cleft is intrinsic to the DNA sequence and does not arise from protein-induced DNA conformational change. Protein-DNA hydrogen bonding plays a secondary role in specificity.

Les interactions ADN-protéine sont fondamentales dans de nombreux processus biologiques tels que la régulation des gènes et la réparation de l'ADN. Cette thèse est centrée sur l'analyse des propriétés physiques et dynamiques des interfaces ADN-protéine. À partir de l'étude de quatre complexes ADN-protéine, nous avons montré que l'interface ADN-protéine est dynamique et que les ponts salins et liaisons hydrogène se forment et se rompent dans une échelle de temps de l'ordre de la centaine de picosecondes. L'oscillation des chaînes latérales des résidus est dans certains cas capable de moduler la spécificité d'interaction. Nous avons ensuite développé un modèle de protéine gros grains dans le but de décomposer les interactions ADN-protéine en identifiant les facteurs qui modulent la stabilité et la conformation de l'ADN ainsi que les facteurs responsables de la spécificité de reconnaissance ADN-protéine. Notre modèle est adaptable, allant d'un simple volume mimant une protéine à une représentation plus complexe comportant des charges formelles sur les résidus polaires, ou des chaînes latérales à l'échelle atomique dans le cas de résidus clés ayant des comportements particuliers, tels que les cycles aromatiques qui s'intercalent entre les paires de base de l'acide nucléique / DNA-protein interactions are fundamental in many biological processes such as gene regulation and DNA repair. This thesis is focused on an analysis of the physical and dynamic properties of DNA-protein interfaces. In a study of four DNA-protein complexes, we have shown that DNA-protein interfaces are dynamic and that the salt bridges and hydrogen bonds break and reform over a time scale of hundreds of picoseconds. In certain cases, this oscillation of protein side chains is able to modulate interaction specificity. We have also developed a coarse-grain model of proteins in order to deconvolute the nature of protein-DNA interactions, identifying factors that modulate the stability and conformation of DNA and factors responsible for the protein-DNA recognition specificity. The design of our model can be changed from a simple volume mimicking the protein to a more complicated representation by the addition of formal charges on polar residues, or by adding atomic-scale side chains in the case of key residues with more precise behaviors, such as aromatic rings that intercalate between DNA base pairs