BPS2026 – Extensive molecular dynamics simulations of the peptide binding pathway to Abl kinase as a basis for therapeutic peptide design

Peptides represent an important class of therapeutic agents due to their ability to mimic protein interactions while retaining high specificity and affinity. Their potential to modulate protein-protein interactions makes them valuable candidates for drug design and mechanistic studies of protein interactions. However, the molecular details of their binding mechanisms remain less understood compared to protein-small molecule interactions. Molecular dynamics (MD) simulations can provide atomic-level insight into these processes, but the size and flexibility pose significant challenges for sampling. To address this challenge, we used massive two-dimensional replica exchange MD with the gREST/REUS method to study the binding of a peptide substrate to Abl kinase. Enhanced sampling along the binding interface and protein-peptide distance enabled efficient exploration of the recognition process. The analysis showed that the protein-peptide binding landscape was shallower than for protein-small molecule interactions, characterized by extensive areas of the protein surface where the peptide could reside and form stabilizing contacts. Distinct regions were identified along the pathway, each stabilized by conserved interactions. Despite the size and flexibility of the peptide, the bound pose was reached efficiently, suggesting that specific motifs effectively guide recognition. In parallel, the study was extended toward peptide optimization by exploring sequence modifications. Mutation free energy perturbation (FEP) is applied to refine the sequence, followed by strategies for cyclization. These developments expand the applicability of the protocol, offering a route to rational design of optimized peptides and, ultimately, to the study of larger protein-protein interactions.