Many diseases are caused by abnormal protein expression or impaired signal transduction. Therefore, one of the promising targets for treatment of various diseases is to control these mechanisms. Heat shock protein 90 (Hsp90) is an important and evolutionarily conserved chaperone whose expression in eukaryotes increases during cellular stress in order to reduce cellular damage. Hsp90 is a cellular regulator of gene expression, proliferation, cell cycle and is responsible for maturation of over 300 proteins. It can be observed that dysregulation of Hsp90 expression can cause increased expression of proteins, which in turn are inducers of pathological conditions. For this reason, the protein has become an important target for the development of inhibitors that can, by acting on Hsp90, stop the folding of a wide range of pathological proteins. To date, quite a few inhibitors of the Hsp90 N-terminal domain have entered clinical trials, but most have been discontinued due to toxicity and heat shock response activation. In the last twenty years, following the identification of an allosteric binding site in the C-terminal domain of Hsp90, many groups have focused their research on the development of allosteric inhibitors, which unlike N-terminal inhibitors, do not activate the heat shock response and are generally less toxic.
Since there has been no experimentally determined structure of Hsp90 co-crystallized with allosteric C-terminal domain inhibitor, we decided to use computational methods to determine the binding mode, which would improve further design of allosteric inhibitors. Using computational docking, molecular dynamics simulations, and a library of active compounds with known activities, we proposed a binding model of some of the most potent compounds that gives us a deeper insight into the binding pocket and interactions relevant to the inhibitory activity. Because the binding pocket in the C-terminal domain is very voluminous, the research was conducted in two parts, where we initially focused on locating the binding site where the reference active compound TVS-21 and its derivatives bind. The second part of the research was to establish structure-activity relationship of the previously prepared Hsp90 inhibitors. We selected several active and inactive compounds, which we docked in the selected binding pocket. With this, we wanted to explain how an individual modification affects the binding of related active compounds. The lessons learned from the research of the binding of allosteric inhibitors will enable the design of new compounds with improved activity in the future. However, we must be aware that the proposed models are only hypothetical and the results of experimental methods, such as site-directed mutagenesis and/or STD-NMR experiments, would be needed to support our findings. Compound 8 should be used as the reference compound and reference binding model for further research and optimisation.