What can molecular modelling bring to the design of artificial inorganic cofactors?

In recent years, the development of synthetic metalloenzymes based on the insertion of inorganic catalysts into biological macromolecules has become a vivid field of investigation. The success of the design of these composites is highly dependent on an atomic understanding of the recognition process between inorganic and biological entities. Despite facing several challenging complexities, molecular modelling techniques could be particularly useful in providing such knowledge. This study aims to discuss how the prediction of the structural and energetic properties of the host-cofactor interactions can be performed by computational means. To do so, we designed a protocol that combines several methodologies like protein-ligand dockings and QM/MM techniques. The overall approach considers fundamental bioinorganic questions like the participation of the amino acids of the receptor to the first coordination sphere of the metal, the impact of the receptor/cofactor flexibility on the structure of the complex, the cost of inserting the inorganic catalyst in place of the natural ligand/substrate into the host and how experimental knowledge can improve or invalidate a theoretical model. As a real case system, we studied an artificial metalloenzyme obtained by the insertion of a Fe(Schiff base) moiety into the heme oxygenase of Corynebacterium diphtheriae. The experimental structure of this species shows a distorted cofactor leading to an unusual octahedral configuration of the iron with two proximal residues chelating the metal and no external ligand. This geometry is far from the conformation adopted by similar cofactors in other hosts and shows that a fine tuning exists between the coordination environment of the metal, the deformability of its organic ligand and the conformational adaptability of the receptor. In a field where very little structural information is yet available, this work should help in building an initial molecular modelling framework for the discovery, design and optimization of inorganic cofactors. Moreover, the approach used in this study also lays the groundwork for the development of computational methods adequate for studying several metal mediated biological processes like the generation of realistic three dimensional models of metalloproteins bound to their natural cofactor or the folding of metal containing peptides.