Exploring Protein Dynamics and Folding with Constant pH Molecular Dynamics Simulations

Abstract

Solution acidity or pH is a key environmental regulator with profound impact on biological processes such as protein dynamics and functions. Over the past decade, our research group has developed a set of tools to explicitly account for solution pH in molecular dynamics simulations. In this dissertation we demonstrate the new application areas of the constant-pH molecular dynamics tools. First we use constant-pH MD alongside fixed-protonation-state simulations to explore the dynamics of a cytoplasmic heme-binding protein (PhuS) in Pseudomonas aeruginosa, a notorious opportunistic pathogen. Our results provide atomic-level information on how heme binding affects PhuS dynamics which suggests an induced-fit mechanism, in support of the recent hydrogen-deuterium exchange data. Secondly, we apply constant-pH MD to investigate the acid-induced unfolding of BBL, a small component of the ubiquitous pyruvate dehydrogenase multienzyme complex involved in carbohydrate metabolism inside mitochondria. Our data support that BBL is a barrier-limited two-state folder, an issue that has drawn intense debate among experimental groups. More importantly, our simulations reveal that acid-induced unfolding of BBL is triggered by sequential protonation of His166 and Asp162 and thereby offering atomic details unattainable via experimental means. This work is the first time the constant-pH MD in explicit solvent has been applied to protein folding studies. Next, we employ the membrane-enabled constant-pH MD to understand how proton release drives the conformational transition of the transmembrane multi-drug efflux pump AcrB, which is crucial for the intrinsic resistance of E. coli to clinically important antibiotics. Our data address the controversy regarding the proton/drug stoichiometry and reveal the details of how deprotonation of a single residue leads to a global conformational transition in AcrB. This work paves the way for understanding the complete cycle of drug transport in AcrB and validates the membrane-enabled constant-pH MD technique for mechanistic studies of proton-coupled transporters. Lastly, we benchmark the accuracy of the all-atom constant-pH MD with charge-leveling co-ion using titration simulations of five proteins (HP36, BBL, NTL9, HEWL and SNase). The average and maximum absolute errors between the calculated and experimental pKa values based on 10-ns pH-based replica-exchange simulations are 0.7 and 0.9 units, respectively. Detailed analyses indicate that limited sampling is a major source of error. This work demonstrates the all-atom constant-pH MD method a practical tool for accurate prediction of pKa's and atomically detailed studies of proton-dependent conformational dynamics. In summary, my studies offered new mechanistic insights into the various roles of protons in protein dynamics and folding that were previously not well understood. My work further established constant-pH MD as a powerful tool for revealing atomic details of proton-coupled dynamic processes.