Modeling Protein Conformational Stability On Chromatography Media with a Gō-Like Model

In protein-based drug development, hydrophobic interaction chromatography (HIC) is a vital commercial method used to separate and purify desired proteins from complex mixtures. However, interactions between the protein and the chromatographic surface used for separation may lead to unfolding or misfolding of the protein. We test whether a Gō-like model simulated with Monte Carlo (MC) statistical mechanics can describe folding and unfolding of proteins on chromatographic surfaces as well as the trends connecting surface unfolding to protein structure and the level of surface attraction. The Go-like model implemented in this study uses a residue-level coarse grain representation of the protein as a linear chain of beads and requires the classification of all possible contacts as either native or nonnative. The surface is modeled as a flat plane of point interaction sites, which are treated as native contacts to protein hydrophobic residues. Temperature replica exchange is used to achieve convergence of the unfolded and folded states. The software used to perform the MC simulation is implemented in homegrown Python and C software in order to provide a simpler, more customizable platform. With this model, we explore the tradeoffs between protein stabilization/destabilization and protein-surface interaction strength.