Chez les eucaryotes supérieurs, la progression ordonnée du cycle cellulaire est régie par une dizaine de kinases Cdk-cyclines. Les altérations génétiques ou épigénétiques impliquant des oncogènes ou des gènes codant pour des suppresseurs de tumeurs sont souvent associées à l'expression ou l'activation aberrante des Cdks, favorisant ainsi la prolifération cellulaire incontrôlée et notamment le développement de cancers. Malgré la pertinence oncogénique et thérapeutique de ces protéines, leur détection est restée jusqu'à présent limitée à des méthodes indirectes et invasives. Dans ce contexte, mes travaux de thèse ont permis de développer un biosenseur peptidique fluorescent permettant de reconnaître spécifiquement les Cdk-cyclines. Associé à une stratégie de vectorisation non invasive basée sur l'utilisation de peptides vecteurs pénétrants, le biosenseur a été délivré efficacement dans les cellules. La mise au point d'une quantification ratiométrique du signal a par ailleurs permis d'évaluer l'abondance relative des Cdk-cyclines endogènes. Deux variants plus spécifiques de certains complexes ont pu être développés. Enfin, d'autres versions du biosenseur ont quant à elles permis d'évaluer sa biodistribution in vivo et de mettre au point un essai cellulaire en vue d'un criblage de petites molécules ayant un effet sur l'abondance relative des Cdk-cyclines. / Cdk-cyclins represent key regulators of cell cycle progression among superior eukaryotes. Genetic and epigenetic alterations involving oncogenes or tumor suppressor genes are often associated with aberrant expression or activation of Cdks, leading to the sustained proliferation of cells and by the way to the development of cancers. Despite the oncogenic and therapeutic relevance of these proteins, their detection has so far remained limited to indirect and invasive methods. My Ph.D. thesis work aimed in this context at developing peptidic fluorescent biosensors that specifically recognize Cdk-cyclins. Combined to cell-penetrating peptides, the biosensor was efficiently delivered into cells. Following the development of the signal ratiometric quantification, the relative abundance of endogenous Cdk-cyclins was directly evaluated in living cells. Two other variants, that are more specific towards specific Cdk-cyclin complexes, were also designed. Finally, the development of novel versions of the biosensor allowed us to evaluate its biodistribution in vivo and to set up a cell-based assay to screen small molecules having an effect on Cdk-cyclin relative abundance.

In this thesis, emphasis has been placed on development of new approaches for high-throughput analysis of protein:DNA interactions in vitro and in vivo. In vitro strategies for detection of protein:DNA interaction require isolation of active and soluble protein. However, current methodologies for purification of proteins often fail to provide high yield of pure and tag-free protein mainly because enzymatic cleavage reactions for tag removal do not exhibit stringent sequence specificity. Solving this problem is an important step towards high-throughput in vitro analysis of protein:DNA interactions. As a result, parts of this thesis are devoted to developing new approaches to enhance the specificity of a proteolysis reaction. The first approach was through manipulation of experimental conditions to maximize the yield of the desired protein products from enterokinase proteolysis reactions of two His-tagged proteins. Because it was suspected that accessibility of the EK site was impeded, that is, a structural problem due to multimerization of proteins, focus was based on use of denaturants as a way to open the structure, thereby essentially increasing the stoichiometry of the canonical recognition site over noncanonical, adventitious sites. Promoting accessibility of the canonical EK target site can increase proteolytic specificity and cleavage yield, and general strategies promoting a more open structure should be useful for preparation of proteins requiring endoprotease treatment. One such strategy for efficient EK proteolysis is proposed: by heterodimerizing with a separate leucine zipper, the bZIP basic region and amino-terminus can become more open and potentially more accessible to enterokinase. In vivo strategies have the advantage over their in vitro counterparts of providing a native-like environment for assessing protein:DNA interactions, yet the most frequently used techniques often suffer from high false-positive and false-negative rates. In this thesis, a new bioprobe system for high-throughput detection of protein:DNA interactions in vivo is presented. This system offers higher levels of accuracy and sensitivity as well as accessibility and ease of manipulation in comparison with existing technologies.

In this thesis, emphasis has been placed on development of new approaches for high-throughput analysis of protein:DNA interactions in vitro and in vivo. In vitro strategies for detection of protein:DNA interaction require isolation of active and soluble protein. However, current methodologies for purification of proteins often fail to provide high yield of pure and tag-free protein mainly because enzymatic cleavage reactions for tag removal do not exhibit stringent sequence specificity. Solving this problem is an important step towards high-throughput in vitro analysis of protein:DNA interactions. As a result, parts of this thesis are devoted to developing new approaches to enhance the specificity of a proteolysis reaction. The first approach was through manipulation of experimental conditions to maximize the yield of the desired protein products from enterokinase proteolysis reactions of two His-tagged proteins. Because it was suspected that accessibility of the EK site was impeded, that is, a structural problem due to multimerization of proteins, focus was based on use of denaturants as a way to open the structure, thereby essentially increasing the stoichiometry of the canonical recognition site over noncanonical, adventitious sites. Promoting accessibility of the canonical EK target site can increase proteolytic specificity and cleavage yield, and general strategies promoting a more open structure should be useful for preparation of proteins requiring endoprotease treatment. One such strategy for efficient EK proteolysis is proposed: by heterodimerizing with a separate leucine zipper, the bZIP basic region and amino-terminus can become more open and potentially more accessible to enterokinase. In vivo strategies have the advantage over their in vitro counterparts of providing a native-like environment for assessing protein:DNA interactions, yet the most frequently used techniques often suffer from high false-positive and false-negative rates. In this thesis, a new bioprobe system for high-throughput detection of protein:DNA interactions in vivo is presented. This system offers higher levels of accuracy and sensitivity as well as accessibility and ease of manipulation in comparison with existing technologies.