International Journal of Molecular Sciences Review Advances in Molecular Dynamics Simulations and Enhanced Sampling Methods for the Study of Protein Systems Raudah Lazim y, Donghyuk Suh y and Sun Choi * College of Pharmacy and Graduate School of Pharmaceutical Sciences, Ewha Womans University, Seoul 03760, Korea; [email protected] (R.L.); [email protected] (D.S.) * Correspondence: [email protected] These authors contributed equally to this work. y Received: 17 July 2020; Accepted: 11 August 2020; Published: 1 September 2020 Abstract: Molecular dynamics (MD) simulation is a rigorous theoretical tool that when used efficiently could provide reliable answers to questions pertaining to the structure-function relationship of proteins. Data collated from protein dynamics can be translated into useful statistics that can be exploited to sieve thermodynamics and kinetics crucial for the elucidation of mechanisms responsible for the modulation of biological processes such as protein-ligand binding and protein-protein association. Continuous modernization of simulation tools enables accurate prediction and characterization of the aforementioned mechanisms and these qualities are highly beneficial for the expedition of drug development when effectively applied to structure-based drug design (SBDD). In this review, current all-atom MD simulation methods, with focus on enhanced sampling techniques, utilized to examine protein structure, dynamics, and functions are discussed. This review will pivot around computer calculations of protein-ligand and protein-protein systems with applications to SBDD. In addition, we will also be highlighting limitations faced by current simulation tools as well as the improvements that have been made to ameliorate their efficiency. Keywords: molecular dynamics simulation; enhanced sampling; protein-protein interactions; binding free energy; protein-ligand binding affinity 1. Introduction Proteins are vital constituents of living organisms, responsible for myriads of life-sustaining cellular processes such as molecular recognition, signal transduction, protein localization, and enzyme catalysis. These biological events are navigated by protein motions as well as physical interactions established between proteins and partnering molecules such as ligands, peptides, proteins and nucleic acids [1]. Progresses made in structural biology and biophysical characterization methods (X-ray crystallography, nuclear magnetic resonance (NMR) and cryo-electron microscopy) have resulted in the exponential growth in the number of three-dimensional structures available for proteins, protein-ligand and protein-protein complexes. This encourages the study of protein dynamics at the atomic level which is made facile with the advent of molecular dynamics (MD) simulation. The introduction of MD simulation to molecular biology has enabled researchers to access a microscopic visualization of biological processes. While the history of MD simulation of macromolecules began with a simple structural investigation of a small protein, namely bovine pancreatic trypsin inhibitor, the field has matured to a stage that calls for the development of new technology to facilitate the study of larger proteins and protein complexes with more intricate dynamics [2–4]. MD simulations conducted on a variety of protein systems—single proteins, protein-ligand and protein complexes—have Int. J. Mol. Sci. 2020, 21, 6339; doi:10.3390/ijms21176339 www.mdpi.com/journal/ijms Int. J. Mol. Sci. 2020, 21, x FOR PEER REVIEW 2 of 20 Int. J. Mol. Sci. 2020, 21, 6339 2 of 20 ligand and protein complexes—have provided valuable insights on protein stability, protein-ligand binding and protein-protein association, among others. These applications when assimilated into the provideddrug discovery valuable pipeline insights will on proteinempower stability, researchers protein-ligand to identify binding crucial and interactions protein-protein necessary association, for the amongfavorable others. binding These of applicationssmall molecules, when peptides, assimilated and into proteins the drug to discoverybinding pockets pipeline and/or will empower protein- researchersprotein interfaces. to identify The crucialsynergistic interactions use of MD necessary simulations for the and favorable free energy binding calculations of small will molecules, further peptides,aid in harnessing and proteins thermodynamic to binding pockets and kinetic and/or parame protein-proteinters essential interfaces. for the Theprediction synergistic of binding use of MDfree simulationsenergies. The and abilities free energy of these calculations computational will further tools to aid determine in harnessing binding thermodynamic interactions and and binding kinetic parametersaffinities are essential beneficial for for the structure-based prediction of binding drug design free energies. (SBDD) as The such abilities predictions of these could computational be utilized toolsto accelerate to determine drug bindingdevelopment interactions by aiding and bindinghit-to-lead affi optimizationnities are beneficial to afford for drugs structure-based with enhanced drug designspecificity (SBDD) and selectivity. as such predictions could be utilized to accelerate drug development by aiding hit-to-leadThe advancement optimization in to computing afford drugs architecture with enhanced and algorithms specificity andhas selectivity.spurred the extensive growth of MDThe simulation advancement tools inover computing recent decades. architecture Architectural and algorithms developmen has spurredts of computers the extensive resulted growth in the ofinvention MD simulation of graphics tools processing over recent units decades. (GPUs) Architectural and parallel developments computing ofwhile computers the refinement resulted inof thealgorithms invention leads of graphicsto a more processing accurate rendering units (GPUs) of the and potential parallel energy computing surface while (PES) theof the refinement system and of algorithmsbetter conformational leads to a more sampling. accurate The rendering combination of the of potential both facets energy of computing surface (PES) has of enabled the system a more and betterreliable conformational prediction of the sampling. physical The properties combination of pr ofoteins both as facets well ofas computingthe observation has enabled of biologically a more reliablerelevant prediction protein structures of the physical at the properties atomic level, of proteins a feat as that well requires as the observation timescales of beyond biologically milliseconds, relevant proteinexpanded structures length scales, at the atomicand enhanced level, a featsampling that requires techniques timescales that facilitate beyond high milliseconds, free energy expanded barrier lengthcrossing scales, [5]. The and burgeoning enhanced sampling use of MD techniques simulation that as facilitate a main hightool in free the energy computational barrier crossing studies [ 5of]. Theproteins burgeoning as well use as ofa complementary MD simulation asmachinery a main tool in inexperimental the computational studies studies has propelled of proteins actions as well to asimprove a complementary various aspects machinery of MD simu in experimentallations, warranting studies better has propelled predictive actions power toand improve more reliable various in aspectssilico analysis of MD simulations,of protein structures, warranting dynamics better predictive and functions. power and In this more review, reliable we in silicowill provide analysis an of proteinoverview structures, of simulation dynamics strategies and functions., mainly all-atom In thisreview, MD simulation we will provide and enhanced an overview sampling of simulation methods, strategies,used to examine mainly all-atomprotein structures MD simulation and dynamics and enhanced to facilitate sampling the methods, understanding used to examineof mechanisms protein structuresregulating and biological dynamics processes to facilitate namely, the understandingprotein-ligand ofbinding mechanisms and protein-protein regulating biological association. processes The namely,capitalization protein-ligand of the predictiv bindinge and power protein-protein of MD simulation association. and The enhanced capitalization sampling of the methods predictive to powerdetermine of MD binding simulation free andenergies enhanced as well sampling as potentia methodsl binding to determine sites are binding also discussed. free energies Furthermore, as well as potentiallimitations binding of current sites simulation are also discussed. techniques Furthermore, as well as limitationsimprovements of current made simulationto optimize techniques the accuracy as wellof MD as improvementssimulation/enhanced made to sampling optimize methods the accuracy are highlighted. of MD simulation Figure/enhanced 1 reflects samplinga summary methods of the aretopics highlighted. discussed Figurein this1 review. reflects a summary of the topics discussed in this review. Figure 1. Overview of MD simulation and enhanced sampling methods utilized in the study of Figure 1. Overview of MD simulation and enhanced sampling methods utilized in the study of protein-protein (peptide) and protein-ligand complexes. The figure is constructed using the crystal protein-protein (peptide) and protein-ligand complexes. The figure is constructed using the crystal structure of human PIM1 kinase in complexed with imidazopyridazin inhibitor and substrate peptide structure of human PIM1 kinase in complexed