The main objective of this work was to design, produce and characterize a new familyof artificial proteins and to introduce new tailored specific binding sites within this structuralframework. Our general goal was to develop method allowing to rapidly generate newprotein binding specifically to any predefined target macromolecule. Binders based onartificial proteins can be viewed as antibody-like molecules but due to their different structurehave more favorable physical properties (expression, solubility, folding efficiency, stability)than antibodies and derivatives.The design and experimental assembly of a first generation artificial protein library isdescribed in part I. Proteins of this library are made by a repetition of a motif idealized from afamily of natural protein repeats (HEAT repeat). These artificial proteins, named αRep, havethe same general fold but the number of the repeated motif vary from protein to protein.Furthermore, a set of positions of each motif is highly variable within the library. Proteinisolated from this first generation library are well expressed, soluble, extremely stable andwere shown to have the designed secondary and tertiary structure.The methods used to improve the diversity and the experimental quality of the protein libraryare described in the second part of this thesis and have allowed us to create a secondgeneration αRep library. This library is based on the same general scheme but its diversitywas optimized by an improved design and experimental procedures known as filtration/shuffling.This highly diverse second generation library (1.7*109 independent clones) was usedto select variants with tailored binding specificities using phage display method. The resultsclearly show that news αReps binding tightly and specifically a range of arbitrarily definedprotein targets can be efficiently selected. The tertiary structure of complexes between αRepand their cognate target molecule were solved and allow to analyze the nature and detailedorganization of this newly engineered molecular recognition capacities.
