(125d) Computational Prediction and Optimization of Protein Anti-Freeze Activity

Antifreeze proteins (AFPs) are a diverse class of proteins that kinetically depress the observable freezing point of water. We perform molecular dynamics simulation for a collection of AFPs. By analyzing both the dynamic behavior of water near the protein surface and the geometric structure of the protein, we introduce a method that detects the ice binding plane of AFPs and we construct a neural network that is capable of quantitatively predicting experimentally observed thermal hysteresis from three relevant physical variables. We also present a genetic algorithm for the in silico design of AFP mutants with improved antifreeze activity. The algorithm discovers significantly improved mutants for two (rQAE and RiAFP) of three (the above plus wfAFP) structurally diverse AFPs investigated, by encouraging the formation of internal water channels and increasing the total ice-binding area. A subset of residues, mainly nonpolar, is particularly helpful in improving antifreeze activity at the ice-binding surface.