Research Interest


Protein Structure Bioinformatics and Computational Chemistry

Supervisors:  Torsten Schwede 

                     Markus Meuwly



HIV-1 protease


One of the bottlenecks of structure-based drug design is the availability of experimentally determined protein structures. In cases where no protein structure is available, homology models can provide a valuable alternative. Validation of homology models is a crucial aspect in drug development. One important question is to address how errors and inaccuracies of the homology models affect the subsequent molecular modeling of protein-ligand interaction. The results of a recent study estimating the absolute binding free energy of interaction of 16 HIV-1 protease inhibitor complexes correlate well with experimental data [1]. Using this system, we can study the effects of sequence variations and introduce systematic errors in the protein model to simulate the typical errors that occur during homology modeling, e.g. sub-optimal template selection or side chain placement, to quantify the effect on ligand binding affinity and ranking of the 16 inhibitors.

[1] V. Zoete, O. Michielin, M. Karplus, J. Comput. -Aided Mol. Design. 2003, 17, 861-880.


ATP analogue binding in kinases


Protein kinases play an important role in controlling diverse signal transduction pathways in cells, many of which that are disease related. The elucidation of these pathways is therefore essential to understand the molecular mechanics of the disease and to identify viable drug targets, but the identification of the cellular substrates of individual protein kinases remains one of the central challenges in the field. The group of Shokat et al has developed a method to directly tag the substrates of protein kinases. The ATP binding site of a given protein kinase is mutated to accept a radiolabelled ATP analogue (A*TP) as the phosphodonor, allowing the identification of the substrates of the protein kinase [1]. The challenge of this method consists of finding a suitable combination of ATP analogue and mutation site of the enzyme so it retains its activity, but does not alter the substrate specificity. Usually this problem is addressed empirically, and therefore computational approaches offer an attractive alternative to investigate the binding properties of prospective ligands. In this work, six ATP analogues and two different protein kinases, for which experimental data is available, are used to develop a method for investigating ligand binding properties of the ATP analogues to the given engineered protein kinase.

[1] Shah,K.L.Y., Deirmengian, C. and Shokat, K.M. Proc. Natl. Acad. Sci USA, 1997, 94, 3565