Discovering Putative Protein Targets of Small Molecules: A Study of the p53 Activator Nutlin.

Small molecule drugs bind to a pocket in disease causing target proteins based on complementarity in shape and physicochemical properties. There is a likelihood that other proteins could have binding sites that are structurally similar to the target protein. Binding to these other proteins could alter their activities leading to off target effects of the drug. One such small molecule drug Nutlin binds the protein MDM2, which is upregulated in several types of cancer and is a negative regulator of the tumor suppressor protein p53. To investigate the off target effects of Nutlin, we present here a shape-based data mining effort. We extracted the binding pocket of Nutlin from the crystal structure of Nutlin bound MDM2. We next mined the protein structural database (PDB) for putative binding pockets in other human protein structures that were similar in shape to the Nutlin pocket in MDM2 using our topology-independent structural superimposition tool CLICK. We detected 49 proteins which have binding pockets that were structurally similar to the Nutlin binding site of MDM2. All of the potential complexes were evaluated using molecular mechanics and AutoDock Vina docking scores. Further, molecular dynamics simulations were carried out on four of the predicted Nutlin-protein complexes. The binding of Nutlin to one of these proteins, gamma glutamyl hydrolase, was also experimentally validated by a thermal shift assay. These findings provide a platform for identifying potential off-target effects of existing/new drugs and also opens the possibilities for repurposing drugs/ligands.

Stapled peptides in the p53 pathway: computer simulations reveal novel interactions of the staples with the target protein.

Atomistic simulations of a set of stapled peptides derived from the transactivation domain of p53 (designed by Verdine & colleagues, JACS 2007 129:2456) and complexed to MDM2 reveal that the good binders are uniquely characterized by higher helicity and by extensive interactions between the hydrocarbon staples and the MDM2 surface; in contrast the poor binders have reduced helicity and their staples are mostly solvent exposed. Our studies also find that the best binders can also potentially inhibit MDMX with similar affinities, suggesting that such stapled peptides can be evolved for dual inhibition with therapeutic potential.

Differential binding of p53 and nutlin to MDM2 and MDMX: computational studies.

Half of human tumours have mutated p53 while in the other half, defective signalling pathways block its function. One such defect is the overexpression of the MDM2 and MDMX proteins. This has led to an intense effort to develop inhibitors of p53-MDM2/MDMX interactions. Nutlin is the first such compound described to block p53-MDM2 interactions. Molecular dynamics simulations have been used to explore the differences in binding of p53 and nutlin to MDM2/MDMX. Simulations reveal that p53 has a higher affinity for MDM2 than MDMX, driven by stronger electrostatic interactions. p53 is displaced from MDM2 by nutlin because it is more flexible, thus paying a larger entropic penalty upon sequestration by MDM2. The inherent plasticity of MDM2 is higher than that of MDMX, enabling it to bind both p53 and nutlin. The less flexible MDMX interacts with the more mobile p53 because the peptide can adapt conformationally to dock into MDMX, albeit with a reduced affinity; nutlin, however is rigid and hence can only interact with MDMX with low affinity. Evolutionarily, the higher affinity of MDM2 for p53 may enable MDM2 to bind p53 for longer periods as it shuttles it out of the nucleus; in contrast, MDMX only needs to mask the p53 TA domain. This study enables us to hypothesize gain of function mutations or those that have decreased affinity for nutlin. These conclusions provide insight into future drug design for dual inhibitors of MDM2 and MDMX, both of which are oncoproteins found overexpressed in many cancers.