Last updated 06/2019
Antibodies have emerged as a major class of biopharmaceuticals, with indications ranging from autoimmune diseases to cancer. A majority of antibody-related research is currently based on sequence information or stationary structures alone. Here in Innsbruck, we apply a wide range of simulation-based techniques to link structural and dynamic information to pharmaceutically relevant properties such as specificity, stability, hydrophobicity and developability. A key topic of our research is the accurate representation of antibodies as conformational ensembles. Our scope is to provide state-of-the-art tools to develop and optimize novel antibody therapeutics.
It is still entirely unclear what makes a protein an allergen. However, what we do know so far is, that a key step in allergic sensitization is the cleavage of allergen proteins into small peptides. Yet, most protease cleavage sites are found within secondary structure elements, which are not accessible to proteases. The allergen thus has to undergo major conformational rearrangements and local unfolding to become susceptible to proteolysis. This process is strongly linked to a decreasing pH value in the endosome. Using classical and enhanced MD simulations, we study these partial unfolding and refolding events and profile their relation to proteolytic susceptibility. Furthermore, we apply constant pH MD simulations coupled with dynamic NMR experiments to capture shifts in unfolding probability upon changes in pH. Embedded in a strong network of experimental collaborators, we work on elucidating the molecular origins of protein stability, allergenic potency and allergen cross-reactivity.
The human genome comprises more than 560 different proteases, suggesting that nearly 3% of all human genes code for proteases. Their vast variety of biological functions stretch from the degradation of proteins in the digestive tract, over key aspects in the immune system, to involvement in blood coagulation. Depending on its task, substrate recognition of a protease thus needs to range from highly specific to widely promiscuous. We investigate the underlying physical mechanisms on the protein-protein interface, which influence the binding process and determine substrate recognition. We employ and develop MD-based tools to quantify and localize the individual terms contributing to specificity or promiscuity in biomolecular recognition.