Selective Inhibition of STAT3 with Respect to STAT1: Insights from Molecular Dynamics and Ensemble Docking Simulations

STAT3 protein, which is known to be involved in cancer development, is a promising target for anticancer therapy. Successful inhibitors of STAT3 should not affect an activity of closely related protein STAT1, which makes their development challenging. The mechanisms of selectivity of several existing STAT3 inhibitors are not clear. In this work, we studied molecular mechanisms of selectivity of 13 experimentally tested STAT3 inhibitors by means of extensive molecular dynamics and ensemble docking simulations. It is shown that all studied inhibitors bind to the large part of the protein surface in an unspecific statistical manner. The binding to the dimerization interface of the SH2 domain, which is usually considered as the main target region, is not energetically preferable. Binding in this region is remarkably similar for STAT1 and STAT3 proteins and cannot explain experimentally observed selectivity toward STAT3. We propose a new mechanism of selectivity called ``selectivity by distraction'' for existing STAT3 inhibitors. This mechanism is based on equilibrium statistical partitioning of inhibitor molecules between protein domains. The unspecific binding of inhibitors to the DNA-binding and the coil-coil domains is stronger in STAT1 in comparison to STAT3 while the energies of their binding to SH2 domains are comparable. This ``distracts'' inhibitor molecules from the SH2 domain of STAT1 and leads to higher effective concentration of inhibitors in the vicinity of the SH2 domain of STAT3.