Item

Abstract

Protein design can learn from mechanism observed in natural evolution. We have applied the concept of modular assembly and fragment recruitment to construct new well-folded proteins. By combination of parts from two differently folded proteins, a flavodoxin-like and a (ba) 8 -barrel protein, we were able to build new ba-barrel. The crystal structure of one of the chimeric pro- teins showed that the fragments retain their structures within the new context. It further revealed an unexpected b-strand that is important for the integrity of the fold. In order to construct a stable protein without this strand we used computational design to optimize the interfaces of the two fragments and to stabilize the fold. Our approach illustrates how different protein folds could have evolved from each other and developed anew. It fur- ther shows the potential of this concept for the design of new proteins from folding fragments where each fragment contributes its own properties, e.g. ligand binding sites. Furthermore, we want to use computational tools to design enzyme function and ligand specificity into our proteins. Therefore, we performed a structural analysis of a number of published receptors in order to learn in molecular detail how precise the predictions of available programs are. Our in-depth analysis did not confirm ligand bind- ing, various methods showed no indication for binding as had been previously described. Thus, although the computation of side-chain conformations appears to be working, the design pro- tocol for ligand-protein interactions has to be revisited and improved.