Molecular Mechanism of Aggregation of the Cataract-Related γD-Crystallin W42R Variant from Multiscale Atomistic Simulations.

The mechanisms leading to aggregation of the crystallin proteins of the eye lens remain largely unknown. We use atomistic multiscale molecular simulations to model the solution-state conformational dynamics of γD-crystallin and its cataract-related W42R variant at both infinite dilution and physiologically relevant concentrations. We find that the W42R variant assumes a distinct conformation in solution that leaves the Greek key domains of the native fold largely unaltered but lacks the hydrophobic interdomain interface that is key to the stability of wild-type γD-crystallin. At physiologically relevant concentrations, exposed hydrophobic regions in this alternative conformation become primary sites for enhanced interprotein interactions leading to large-scale aggregation.

Role of Conformational Flexibility in Monte Carlo Simulations of Many-Protein Systems.

Efficient computational modeling of biological systems characterized by high concentrations of macromolecules often relies on rigid-body Brownian Dynamics or Monte Carlo (MC) simulations. However, the accuracy of rigid-body models is limited by the fixed conformation of the simulated biomolecules. Multi-conformation Monte Carlo (mcMC) simulations of protein solutions incorporate conformational flexibility via a conformational swap trial move within a predetermined library of discrete protein structures, thereby alleviating artifacts arising from the use of a single protein conformation. Here, we investigate the impact of the number of distinct protein structures in the conformational library and the extent of conformational sampling used in its generation on structural observables computed from simulations of hen egg white lysozyme (HEWL), human γD-Crystallin, and bovine γB-Crystallin solutions. We find that the importance of specific protocols for the construction of the protein structure library is strongly dependent on the nature of the simulated system.

Multi-Conformation Monte Carlo: A Method for Introducing Flexibility in Efficient Simulations of Many-Protein Systems.

We present a novel multi-conformation Monte Carlo simulation method that enables the modeling of protein-protein interactions and aggregation in crowded protein solutions. This approach is relevant to a molecular-scale description of realistic biological environments, including the cytoplasm and the extracellular matrix, which are characterized by high concentrations of biomolecular solutes (e.g., 300-400 mg/mL for proteins and nucleic acids in the cytoplasm of Escherichia coli). Simulation of such environments necessitates the inclusion of a large number of protein molecules. Therefore, computationally inexpensive methods, such as rigid-body Brownian dynamics (BD) or Monte Carlo simulations, can be particularly useful. However, as we demonstrate herein, the rigid-body representation typically employed in simulations of many-protein systems gives rise to certain artifacts in protein-protein interactions. Our approach allows us to incorporate molecular flexibility in Monte Carlo simulations at low computational cost, thereby eliminating ambiguities arising from structure selection in rigid-body simulations. We benchmark and validate the methodology using simulations of hen egg white lysozyme in solution, a well-studied system for which extensive experimental data, including osmotic second virial coefficients, small-angle scattering structure factors, and multiple structures determined by X-ray and neutron crystallography and solution NMR, as well as rigid-body BD simulation results, are available for comparison.