Exploring Peptide's Antifreeze Activity Using a Semi-Automated Molecular Dynamics-Enabled Screening Framework

The generation of molecular dynamics input files for the study of protein and peptide antifreeze behavior is time-consuming and tedious. This study presents the use of a nonequilibrium molecular dynamics simulations pipeline to infer relative rankings of the ice refreezing inhibition or antifreeze activities of peptides. By leveraging a combination of existing tools, the pipeline developed here allows researchers, using only amino acid sequences and requested ice-water ratios, to quickly generate molecular dynamics-ready input files of proteins and peptides at ice-water interfaces using an amino acid sequence alone. Using this pipeline, this work examines potential relationships between the secondary structure and chain length of plant-derived peptides and their antifreeze activity. Using nine different peptides, in groups of three with different peptide chain lengths, namely, short, intermediate, and long (20-25, 35-40, and 55-60 amino acids, respectively), and distinct secondary structural motifs (α-helix, β-sheet, and random coil), potential relationships between antifreeze activity and peptide structural properties were examined. Our results indicate that peptides with stable and rigid secondary structures, especially those rich in α-helix content, exhibit higher antifreeze activity, regardless of the chain lengths tested. Additional analysis of the simulations also reveals that the peptides demonstrating extensive interactions with water molecules display enhanced antifreeze properties, even those with relatively flexible conformations. The existing findings improve the understanding of structure-function relationships in antifreeze peptides and provide practical insights for designing novel and potentially cost-effective peptides for applications in the cryopreservation of food and biological materials.