Computational Drug Discovery

Proteins form different conformations in combination with other proteins—as multimers or heterocomplexes—than when they are present as monomers. Static descriptions of protein structures are therefore often not sufficient to describe their dynamic properties. Nevertheless, current simulations of protein-protein interactions are usually based on monomeric structures, mostly due to limited computational resources. Since small molecules of interest for drug discovery can bind differently in different conformations, we often cannot base our scientific understanding on a single structure. Our group focusses on building biologically relevant models to explain protein-ligand interactions and to develop novel inhibitors.

Our approach allows to investigate whether multimerisation contributes to major conformational changes and thus influences protein-drug interactions. By simulating multiple crystal structures and models of molecules with and without binding partners, we systematically compare these structures in terms of movement and changes at the interface of the components. The aim is to find out which components of the molecules are relevant for dimerisation or multimerisation. Based on this analysis, possibilities open up for site-specific mutagenesis with the goal of target validation as well as for the development of novel inhibitors of protein-protein interactions at binding sites that are not present in monomeric structures.