The design and synthesis of constrained AVPI peptidomimetics

Abstract

Recent studies have demonstrated that breast cancer cell lines are sensitive to the XIAP (X-linked inhibitor of apoptotic proteins) antagonist Smac (second mitochondrial activator of caspases). Specifically the terminal tetrapeptide, AVPI, inhibits XIAP and thus sensitizes the breast cancer cells to various apoptotic therapeutics. These peptides are poor drug candidates due to their size, hydrophilicity and potential for peptidase cleavage. To address these issues we systematically synthesized peptide mimics that are constrained at various torsion angles along the peptide backbone. The compounds were tested in vivo and in vitro utilizing the laboratories of the Mayo Medical School. One compound, HCl*NH2-(R)-αMe-γ-Lactam-L-Val-L-Pro-L-Ile-OMe (3.1), increased the apoptotic signals above the assay background. Retro-modeling analysis utilizing the Schödinger flexible docking program suite did not provide an easy explanation for why this might be the case. The in silico models suggest that compound 3.1 interacts with only the primary pocket of the AVPI binding site and does not interact, like the native peptide, with both pockets in the active site. The R stereochemistry of the bicyclic compounds is predicted by our models to have a binding efficacy above that of the native peptide presumably due to the interaction with both pockets of the binding site as well as d-orbital interactions of the thiazole ring system with the neighboring tryptophan. The S stereochemistry was also initially proposed as a negative control for the biological assays, however, the instability of the molecule provided several daunting challenges to an already challenging and low yielding synthesis and the negative controls were abandoned when it was clear that the synthesis was futile. Our research shows limited evidence that our AVPI peptidomimetics are a potential scaffold to base other drug-like molecules upon. Future biological data from the bicyclic compounds will hopefully prove our hypothesis and molecular modeling studies correct.