A combination of in silico and SAR studies to identify binding hot spots of Bcl-xL inhibitors.

Inhibition of Bcl-2 family protein-protein interactions (PPI) is a very promising direction in cancer chemotherapy. Hence over the last decade, many medicinal chemistry studies endeavoured to discover drug candidates, and a wealth of chemical scaffolds with striking chemical diversity was reported as Bcl-xL inhibitors. This raises the question of whether all these molecules could occupy a unique binding site, or rather discrete pockets of the protein surface. To test if small and chemically diverse Bcl-xL inhibitors are likely to bind a single pocket, and to identify which pocket, we used a battery of computational and modeling approaches. We first checked that the large dataset of Bcl-xL inhibitors we built can actually fit to a universal pharmacophore. Then we defined the probable binding hot spots of interaction through comparison of crystal structures, as well as virtual fragment screening. Finally, new analogues of small polyphenol derivatives were synthesized to precisely probe a hydrogen bond suggested by docking experiments. Bcl-xL inhibition potency of these products confirmed the predicted binding mode. This combination of X-ray structure exploration, molecular modeling studies and medicinal chemistry supports that all these small Bcl-xL inhibitors occupy the same hot spot of interaction. The identification of this binding site should help the design and optimization of small PPI Bcl-xL inhibitors.