Molecular dynamics (MD) simulations are widely used in different fields, particularly in structure-based drug discovery. MD simulations have provided detailed information on the variations and conformational changes of proteins. This method has been used to investigate the structure, dynamics and thermodynamics of biological molecules and their complexes and is an important method for structure prediction and model refinement for proteins.
A recent paper of Bermudez et al. (2016) gives an overview of the applications and impact of MD simulations on structural studies, including protein folding prediction, protein assembly processes and sampling conformational space (Fig.1).
The MD simulations used different molecular parameters, such as the energy function force field (FF) which is used to calculate the potential energy of a system of atoms or coarse-grained particles. Some various commonly used molecular simulation computational programs for molecular dynamics are the Assisted Model Building with Energy Refinement (AMBER), the Chemistry at Harvard Macromolecular Mechanics (CHARMM), the Groningen Molecular Simulation (Gromos) and the Optimized Potentials for Liquid Simulations (OPLS). These software can provide close enough approximation to estimate a wide range of protein conformations that can be studied for a specific biological function.
Protein folding is one of the most important parameters to study in structural biology, therefore there is a need for improved computational methods in protein folding predictions. MD simulations can provide complementary information with high structural resolution. The Blue Gene supercomputer developed by IBM was built for protein folding predictions, as well as Folding@home in which computers all around the world provide computer processing time to perform MD simulations for protein folding and to examine the causes of protein misfolding. In a study of Lindorff-Larsen et al. a supercomputer system for MD simulation called Anton (D.E. Shaw Research) was used to predict the folding of 12 structurally diverse proteins (Chignolin, Trp-cage, BBA, Villin, WW domain, NTL9, BBL, Protein B, Homeodomain, Protein G, α3D and λ-repressor).
To predict protein folding and investigate protein folding pathways MD simulations can also be used with some alternative approaches that are able to give more information on the characteristics of the folding process. Some recent applications of these methods include the coarse-graining (CG) method, implicit solvent models, enhanced sampling algorithms or accelerated molecular dynamics. For example Miao et al., used dual boost accelerated MD to capture the folding of four fast-folding proteins (chignolin, Trp-cage, villin headpiece and WW domain) identifying distinct conformational states (unfolded and intermediate). These examples demonstrated the ability of MD simulations to understand and predict the protein-folding pathways.
MD simulations also are useful for elucidating protein assemblies and relating protein-protein interfaces. One method commonly used for this is course graining (CG), in which a small group of atoms are treated as single particles, to reduce the number of particle-particle interactions to be calculated, however this method might lack detailed molecular information. Dimerization of transmembrane (TM) proteins is usually investigated by CG models. For example Han et al. used coarse-grained and atomistic simulations to model the homodimerization of the synaptobrevin-2 (sybII) transmembrane domain protein and selected TM domain mutants. Another exceptional example is the study by Zhao et al. in which using different techniques combined such as cryo-EM, cryo-electron tomography and MD simulations to reveal the tubular assembly of HIV-1 capsid protein hexamers, resulting in all atom models for the entire HIV-1 capsid. The structural information gained is promising for rational drug design to target the HIV capsid.
Another useful application of MD simulation is for the exploration of the conformational space of macromolecules. MD based sampling is used to obtain functional and dynamic models of pharmacologically targeted proteins to obtain the conformational flexibility and the conformational space of the protein. Perdih et al. used a drug design strategy using multiple protein structures of the MurD for the identification of novel MurD ligase inhibitors using targeted molecular dynamic simulation and the Off-Path simulations (OPS)
Figure 2. Inhibitor design strategy based on an enzyme structural flexibility, bacterial MurD Ligase. Perdih et al. (2014).
In another study performed by Calimet et al. they proposed an atomic resolution mechanism using MD simulation and also a possible model for the gating process of a eukaryotic chloride channel gated by glutamate (GluCL) that can possibly apply for other pentameric ion channels. All these studies showed gainful and effective examples of the application of MD simulation in combination with various other techniques to investigate structural information of pharmacological proteins. It is important to mention that MD simulations cannot completely capture the detailed molecular composition of biological systems. New computational methods and software need to be developed with the improvement of hardware and sampling algorithms.
Bermudez, M. et al. (2016) More than a look into a crystal ball: protein structure elucidation guided by molecular dynamics simulations. Drug Discov Today 21, 1799-1805.
Dill, K.A. and MacCallum, J.L. (2012) The protein-folding problem, 50 years on. Science 338, 1042–1046.
Lindorff-Larsen, K. et al. (2011) How fast-folding proteins fold. Science 334, 517–520.
Miao, Y. et al. (2015) Accelerated molecular dynamics simulations of protein folding. J. Comput. Chem. 36, 1536–1549.
Han, J. et al. (2015) Synaptobrevin transmembrane domain dimerization studied by multiscale molecular dynamics simulations. Biophys. J. 109, 760–771.
Zhao, G. et al. (2013) Mature HIV-1 capsid structure by cryo-electron micetroscopy and all-atom molecular dynamics. Nature 497, 643–646.
Perdih, A. et al. (2014) Inhibitor design strategy based on an enzyme structural flexibility: a case of bacterial MurD ligase. J. Chem. Inform. Model. 54, 1451–1466.
Calimet, N. et al. (2013) A gating mechanism of pentameric ligand-gated ion channels. Proc. Natl. Acad. Sci. U. S. A. 110, E3987–E3996.
Blog written by Thalia Carreno Velazquez