The size and length scales assessable to molecular dynamics simulation have been greatly increased in recent years by the availability of GPU nodes on supercomputers. However, sampling bottlenecks remain an issue, particularly for large, flexible molecules such as peptides and proteins. Both challenges and successes of work to predict binding modes and affinities of peptide-protein and protein-protein complexes are described. In the case of peptide-protein binding, significant strides have been made by developing a method that combines free energy simulations with sparse, ambiguous external information that restricts sampling to regions of interest. This technique fully accounts for peptide flexibility and is implemented in parallel on large-scale GPU clusters. It is illustrated by studying peptide inhibitors of the P53-regulator-proteins MDM2 and MDMX. Relative binding affinities are evaluated and binding modes are found to be in excellent agreement with experiment.

Describing peptide-protein and protein-protein interactions with molecular dynamics simulation

254th National Meeting (2017)

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