Summary

Proteins are solvated by water molecules which form water shells around the protein surface. Still water molecules can also penetrate protein cavities such as the binding pocket and are then termed “bound water”. Such bound water molecules can form networks built up by several bound water molecules. The binding of a ligand to a protein cavity or the chemical modification (including post-translational modification) of a protein or its ligands can either displace an individual bound water molecule or the entire network: we regard such a displacement as “defect” of soft matter. We consider this defect in particular a “connectivity defect” since the bound water connects different parts of the protein and the wider hydrogen-bonding network of the surrounding water.

The above-described displacement of bound water molecules is part of the molecular recognition process and affects the thermodynamic properties of the protein and the water molecule(s). We are primarily interested in the microscopic basis of the molecular recognition process, both from an experimental point of view as well from a theoretical standpoint. In our project we will combine experimental physical chemistry, machine-learning approaches and full-atomistic and implicit solvent computer simulations to understand how defects in the water networks in protein cavities are driven by small molecules with the long-term aim of engineering these defects specifically. We will study well-characterized protein-ligand systems to design ligands leading to a specific change in the thermodynamic profile. Ultimately, we aim to rationalize design principles for novel ligands to better understand the molecular recognition process and guide the design of novel ligands. First, we will characterize the experimentally observed water networks in publicly available crystal structures in a data-driven approach. This analysis will be supplemented by atomistic simulations and machine learning approaches for designing defects (by means of, e.g. novel ligands) that perturb a water network. Secondly, we will study if whether and how pKa shifts of titratable amino acids or ligands affect the perturbation of the water network. Experimental validation will complement this step by isothermal titration calorimetry (ITC), protein crystallography and nuclear magnetic resonance (NMR).

The central research question (in combination with the terminology of the CRC) is depicted in Figure 1.