Cyclic peptides are promising modulators of protein-protein interactions: they can bind large protein surfaces with high affinity and specificity and they have enhanced biostability and bioavailability compared to their linear counterparts. Despite several successful demonstrations of the clinical application of cyclic peptides as immunosuppressants, antibiotics, and antifungals, we note that most of these cyclic peptides are actually natural products and their derivatives. The key impediment to fully exploring cyclic peptides as a promising class of therapeutics is that it is currently impossible to accurately predict the three-dimensional conformation that any given cyclic peptide will adopt. Here we present a novel computational method to sample cyclic peptide conformations efficiently to provide fast and reliable structure predictions. By performing bias-exchange metadynamics simulations using multiple dihedral angles as the reaction coordinates, we are able to obtain converged free-energy profiles along these coordinates, regardless of simulation input structures. To extract the equilibrium structural ensemble from the biased simulations, we performed a cluster analysis using the peptide backbone RMSD as the structural similarity metric. After each conformation in the biased simulations is assigned to a structural cluster, its contribution to the equilibrium population of the cluster is properly calculated by reweighting based on the bias history of the conformation. Coupling bias-exchange metadynamics simulations with cluster analysis provides an equilibrium structural ensemble for a cyclic peptide. For a benchmark cyclic peptide, cyclo(-D-(N-Me)Ala1-Ala2-Ala3-Ala4-(N-Me)Ala5-Ala6), the most populated cluster predicted using our method and the amber99sb-ildn force field is in good agreement with the experimental NMR structure, with a backbone RMSD of 0.54 �. We expect that our tool will accelerate the design of novel cyclic peptides, which can be used to target protein-protein interactions for hypothesis testing and as new peptide therapeutics. We will provide open access to the software, with user-friendly tutorial and technical support.

Computational tool for predicting structures of cyclic peptides

247th National Meeting (2014)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.