Prediction of α-MORFS and folding free energies in intrinsically disordered peptides.

Abstract

Intrinsically disordered proteins (IDPs) and disordered regions (IDRs) of folded proteins play significant roles in cell signaling, regulation, and recognition. However, the lack of well-defined structures challenges our ability to predict the biological role and working mechanisms of IDPs. Computer simulations provide a powerful tool to characterize the conformational ensembles explored by IDPs. Here, we use free energy calculations in atomistic simulations to explore the applicability of binding-induced folding mechanisms for the development of ligands that specifically bind to intrinsically disordered proteins. Molecular recognition features (MoRFs) are functional regions located within longer intrinsically disordered regions that undergo disorder-to-order transitions upon binding their specific interaction partners. A subgroup called α-MoRFs undergoes disordered to α-helix transitions upon binding. In principle, MoRFs can also be utilized as targets for the design of drug molecules that bind to such regions in disease-causing IDPs and render them inactive or modulate their properties. We quantify the free energy cost involved in transitions from intrinsically disordered peptide states into folded secondary structure motifs for variable peptide sequences. We compare α-MoRFs and fully disordered sequences from several known proteins and compare our results to state-of-the-art structure predictions. In contrast to structure prediction tools, our atomistic simulation protocol not only allows us to determine the minimum free energy structure but also quantifies the free energy difference between folded and disordered conformational states. The latter is critical for the development of drugs targeting a specific IDR in a folded conformation because the binding free energy needs to overcompensate the free energy cost of folding. Our general strategy aims to reduce the “undruggability” of IDPs and IDRs, which lack an inherent structure and represent a significant fraction of the proteome.