Exploring the Use of Molecular Docking to Identify Bioaccumulative Perfluorinated Alkyl Acids (PFAAs)

Methods to predict the bioaccumulation potential of per- and polyfluorinated alkyl substances (PFAS) are sorely needed, given the proliferation of these substances and lack of data on their properties and behavior. Here, we test whether molecular docking, a technique where interactions between proteins and ligands are simulated to predict both bound conformation and interaction affinity, can be used to predict PFAS binding strength and biological half-life. We show that an easy-to-implement docking program, Autodock Vina, can successfully redock perfluorooctanesulfonate (PFOS) to human serum albumin with deviations smaller than 2 angstrom. Furthermore, predicted binding strengths largely fall within one standard deviation of measured values for perfluorinated alkyl acids (PFAAs). Correlations with half-lives suggest both membrane partitioning and protein interactions are important, and that serum albumin is only one of a number of proteins controlling the fate of these chemicals in organisms. However, few data are available for validation of our approach as a broad screening tool, and available data are highly variable. We therefore call for collection of new data, particularly including proteins other than serum albumin and substances beyond perfluorooctanoic acid (PFOA) and PFOS. The methods we discuss in this work can serve as a framework for guiding such data collection.