Unconstrained enhanced sampling for free energy calculations of biomolecules: a review

Abstract

Free energy calculations are central to understanding the structure, dynamics and function of biomolecules. Yet insufficient sampling of biomolecular configurations is often regarded as one of the main sources of error. Many enhanced sampling techniques have been developed to address this issue. Notably, enhanced sampling methods based on biasing collective variables (CVs), including the widely used umbrella sampling, adaptive biasing force and metadynamics, have been discussed in a recent excellent review (Abrams and Bussi, Entropy, 2014). Here, we aim to review enhanced sampling methods that do not require predefined system-dependent CVs for biomolecular simulations and as such do not suffer from the hidden energy barrier problem as encountered in the CV-biasing methods. These methods include, but are not limited to, replica exchange/parallel tempering, self-guided molecular/Langevin dynamics, essential energy space random walk and accelerated molecular dynamics. While it is overwhelming to describe all details of each method, we provide a summary of the methods along with the applications and offer our perspectives. We conclude with challenges and prospects of the unconstrained enhanced sampling methods for accurate biomolecular free energy calculations.     

Atomistic molecular simulations of protein folding

Theory and experiment have provided answers to many of the fundamental questions of protein folding; a remaining challenge is an accurate, high-resolution picture of folding mechanism. Atomistic molecular simulations with explicit solvent are the most promising method for providing this information, by accounting more directly for the physical interactions that stabilize proteins. Although simulations of folding with such force fields are extremely challenging, they have become feasible as a result of recent advances in computational power, accuracy of the energy functions or 'force fields', and methods for improving sampling of folding events. I review the recent progress in these areas, and highlight future challenges and questions that we may hope to address with these methods. I also attempt to place atomistic models into the context of the energy landscape view of protein folding, and coarse-grained simulations.     

Molecular dynamics simulation of proteins under high pressure: Structure,function and thermodynamics

BackgroundMolecular dynamics (MD) simulation is well-recognized as a powerful tool to investigate protein structure, function, and thermodynamics. MD simulation is also used to investigate high pressure effects on proteins. For conducting better MD simulation under high pressure, the main issues to be addressed are: (i) protein force fields and water models were originally developed to reproduce experimental properties obtained at ambient pressure; and (ii) the timescale to observe the pressure effect is often much longer than that of conventional MD simulations.Scope of reviewFirst, we describe recent developments in MD simulation methodologies for studying the high-pressure structure and dynamics of protein molecules. These developments include force fields for proteins and water molecules, and enhanced simulation techniques. Then, we summarize recent studies of MD simulations of proteins in water under high pressure.Major conclusionsRecent MD simulations of proteins in solution under pressure have reproduced various phenomena identified by experiments using high pressure, such as hydration, water penetration, conformational change, helix stabilization, and molecular stiffening.General significanceMD simulations demonstrate differences in the properties of proteins and water molecules between ambient and high-pressure conditions. Comparing the results obtained by MD calculations with those obtained experimentally could reveal the mechanism by which biological molecular machines work well in collaboration with water molecules.     

Conformational landscapes of membrane proteins delineated by enhanced sampling molecular dynamics simulations

The expansion of computational power, better parameterization of force fields, and the development of novel algorithms to enhance the sampling of the free energy landscapes of proteins have allowed molecular dynamics (MD) simulations to become an indispensable tool to understand the function of biomolecules. The temporal and spatial resolution of MD simulations allows for the study of a vast number of processes of interest. Here, we review the computational efforts to uncover the conformational free energy landscapes of a subset of membrane proteins: ion channels, transporters and G-protein coupled receptors. We focus on the various enhanced sampling techniques used to study these questions, how the conclusions come together to build a coherent picture, and the relationship between simulation outcomes and experimental observables.     

Secondary Structure Propensities in Peptide Folding Simulations: A Systematic Comparison of Molecular Mechanics Interaction Schemes

We present a systematic study directed toward the secondary structure propensity and sampling behavior in peptide folding simulations with eight different molecular dynamics force-field variants in explicit solvent. We report on the combinational result of force field, water model, and electrostatic interaction schemes and compare to available experimental characterization of five studied model peptides in terms of reproduced structure and dynamics. The total simulation time exceeded 18 μs and included simulations that started from both folded and extended conformations. Despite remaining sampling issues, a number of distinct trends in the folding behavior of the peptides emerged. Pronounced differences in the propensity of finding prominent secondary structure motifs in the different applied force fields suggest that problems point in particular to the balance of the relative stabilities of helical and extended conformations.     

Refinement of protein structure homology models via long, all-atom molecular dynamics simulations

Accurate computational prediction of protein structure represents a longstanding challenge in molecular biology and structure-based drug design. Although homology modeling techniques are widely used to produce low-resolution models, refining these models to high resolution has proven difficult. With long enough simulations and sufficiently accurate force fields, molecular dynamics (MD) simulations should in principle allow such refinement, but efforts to refine homology models using MD have for the most part yielded disappointing results. It has thus far been unclear whether MD-based refinement is limited primarily by accessible simulation timescales, force field accuracy, or both. Here, we examine MD as a technique for homology model refinement using all-atom simulations, each at least 100 μs long-more than 100 times longer than previous refinement simulations-and a physics-based force field that was recently shown to successfully fold a structurally diverse set of fast-folding proteins. In MD simulations of 24 proteins chosen from the refinement category of recent Critical Assessment of Structure Prediction (CASP) experiments, we find that in most cases, simulations initiated from homology models drift away from the native structure. Comparison with simulations initiated from the native structure suggests that force field accuracy is the primary factor limiting MD-based refinement. This problem can be mitigated to some extent by restricting sampling to the neighborhood of the initial model, leading to structural improvement that, while limited, is roughly comparable to the leading alternative methods.     

Ten-microsecond molecular dynamics simulation of a fast-folding WW domain

All-atom molecular dynamics (MD) simulations of protein folding allow analysis of the folding process at an unprecedented level of detail. Unfortunately, such simulations have not yet reached their full potential both due to difficulties in sufficiently sampling the microsecond timescales needed for folding, and because the force field used may yield neither the correct dynamical sequence of events nor the folded structure. The ongoing study of protein folding through computational methods thus requires both improvements in the performance of molecular dynamics programs to make longer timescales accessible, and testing of force fields in the context of folding simulations. We report a ten-microsecond simulation of an incipient downhill-folding WW domain mutant along with measurement of a molecular time and activated folding time of 1.5 microseconds and 13.3 microseconds, respectively. The protein simulated in explicit solvent exhibits several metastable states with incorrect topology and does not assume the native state during the present simulations.     

Force field influences in beta-hairpin folding simulations

All-atom force fields are now routinely used for more detailed understanding of protein folding mechanisms. However, it has been pointed out that use of all-atom force fields does not guarantee more accurate representations of proteins; in fact, sometimes it even leads to biased structural distributions. Indeed, several issues remain to be solved in force field developments, such as accurate treatment of implicit solvation for efficient conformational sampling and proper treatment of backbone interactions for secondary structure propensities. In this study, we first investigate the quality of several recently improved backbone interaction schemes in AMBER for folding simulations of a beta-hairpin peptide, and further study their influences on the peptide's folding mechanism. Due to the significant number of simulations needed for a thorough analysis of tested force fields, the implicit Poisson-Boltzmann solvent was used in all simulations. The chosen implicit solvent was found to be reasonable for studies of secondary structures based on a set of simulations of both alpha-helical and beta-hairpin peptides with the TIP3P explicit solvent as benchmark. Replica exchange molecular dynamics was also utilized for further efficient conformational sampling. Among the tested AMBER force fields, ff03 and a revised ff99 force field were found to produce structural and thermodynamic data in comparably good agreement with the experiment. However, detailed folding pathways, such as the order of backbone hydrogen bond zipping and the existence of intermediate states, are different between the two force fields, leading to force field-dependent folding mechanisms.     

Study of the electrostatics treatment in molecular dynamics simulations

This article considers the treatment of long-range interactions in molecular dynamics simulations. We investigate the effects of using different cutoff distances, constant versus distance-dependent dielectric, and different smoothing methods. In contrast to findings of earlier studies, we find that increasing the cutoff over 8 A does not significantly improve the accuracy (Arnold and Ornstein, Proteins 1994;18:19-33), and using a distance-dependent dielectric instead of a constant dielectric also does not improve accuracy (Guenot and Kollman, Protein Sci 1992;1:1185-1205). This might depend on differences in simulation protocols or force fields, or both, because we use the CHARMM22 force field with stochastic boundary conditions, whereas earlier studies used other protocols and energy functions. We also note that the stability of the simulations is highly dependent on the starting structure, showing that accurate molecular simulations not only depend on a realistic simulation protocol but also on correct initial conditions.     

Simulating the free energy of passive membrane permeation for small molecules

Abstract

Computer simulations of passive membrane permeation provide important microscopic insights into the molecular mechanism of this important biological process that are complementary to experimental data. Our review focuses on the main approaches for calculating the free energy, or potential of mean force, for permeation of small molecules through lipid bilayers. The theoretical background for most currently used methods for potential of mean force calculation is described, including particle insertion, thermodynamic integration, umbrella sampling, metadynamics, adaptive biasing force and milestoning. A brief comparison of strengths and weaknesses of the competing approaches is presented. This is followed by a survey of results obtained by the different methods, with special attention to describing the mechanistic insights generated by modelling and illustrating capabilities of the different techniques. We conclude with a discussion of recent advances and future directions in modelling membrane permeation, including latest methodological enhancements, consideration of multiple slow variables and memory effects.     

Headgroup Conformations of Phospholipids from Molecular Dynamics Simulation: Sampling Challenges and Comparison to Experiment

The preferred conformations of the glycerol region of a phospholipid have been explored using replica exchange molecular dynamics (MD) simulations and compared with the results of standard MD approaches and with experiment. We found that due to isomerization rates in key torsions that are slow on the timescale of atomistic MD simulations, standard MD is not able to produce accurate equilibrium conformer distributions from reasonable trajectory lengths (e.g., on the 100 ns) timescale. Replica exchange MD, however, results in quite efficient sampling due to the rapid increase in isomerization rate with temperature. The equilibrium distributions obtained from replica exchange MD have been compared with the results of experimental nuclear magnetic resonance observations. This comparison suggests that the sampling approach demonstrated here is a valuable tool that can be used in evaluating force fields for molecular simulation of lipids.     

Are Current Molecular Dynamics Force Fields too Helical?

Accurate force fields are essential for the success of molecular dynamics simulations. In apparent contrast to the conformational preferences of most force fields, recent NMR experiments suggest that short polyalanine peptides in water populate the polyproline II structure almost exclusively. To investigate this apparent contradiction, with its ramifications for the assessment of molecular force fields and the structure of unfolded proteins, we performed extensive simulations of Ala₅ in water (∼5 μs total time), using twelve different force fields and three different peptide terminal groups. Using either empirical or density-functional-based Karplus relations for the J-couplings, we find that most current force fields do overpopulate the α-region, with quantitative results depending on the choice of Karplus relation and on the peptide termini. Even after reweighting to match experiment, we find that Ala₅ retains significant α- and β-populations. In fact, several force fields match the experimental data well before reweighting and have a significant helical population. We conclude that radical changes to the best current force fields are not necessary, based on the NMR data. Nevertheless, experiments on short peptides open the way toward the systematic improvement of current simulation models.     

Improved side‐chain torsion potentials for the Amber ff99SB protein force field

Recent advances in hardware and software have enabled increasingly long molecular dynamics (MD) simulations of biomolecules, exposing certain limitations in the accuracy of the force fields used for such simulations and spurring efforts to refine these force fields. Recent modifications to the Amber and CHARMM protein force fields, for example, have improved the backbone torsion potentials, remedying deficiencies in earlier versions. Here, we further advance simulation accuracy by improving the amino acid side‐chain torsion potentials of the Amber ff99SB force field. First, we used simulations of model alpha‐helical systems to identify the four residue types whose rotamer distribution differed the most from expectations based on Protein Data Bank statistics. Second, we optimized the side‐chain torsion potentials of these residues to match new, high‐level quantum‐mechanical calculations. Finally, we used microsecond‐timescale MD simulations in explicit solvent to validate the resulting force field against a large set of experimental NMR measurements that directly probe side‐chain conformations. The new force field, which we have termed Amber ff99SB‐ILDN, exhibits considerably better agreement with the NMR data. Proteins 2010. © 2010 Wiley‐Liss, Inc.     

Towards an understanding of complex biological membranes from atomistic molecular dynamics simulations

Computer simulation has emerged as a powerful tool for studying the structural and functional properties of complex biological membranes. In the last few years, the use of recently developed simulation methodologies and current generation force fields has permitted novel applications of molecular dynamics simulations, which have enhanced our understanding of the different physical processes governing biomembrane structure and dynamics. This review focuses on frontier areas of research with important biomedical applications. We have paid special attention to polyunsaturated lipids, membrane proteins and ion channels, surfactant additives in membranes, and lipid–DNA gene transfer complexes.     

Advances in biomolecular simulations: methodology and recent applications

Molecular dynamics simulations are widely used today to tackle problems in biochemistry and molecular biology. In the 25 years since the first simulation of a protein computers have become faster by many orders of magnitude, algorithms and force fields have been improved, and simulations can now be applied to very large systems, such as protein-nucleic acid complexes and multimeric proteins in aqueous solution. In this review we give a general background about molecular dynamics simulations, and then focus on some recent technical advances, with applications to biologically relevant problems.     

A Robust Water Potential Parameterisation

Abstract

We compare molecular dynamics simulation results for the properties of liquid water predicted by four novel water potential models. These models are designed as a combination of parameters taken from the dedicated but brittle TIP3P water potential, and the more flexible but less accurate parameterisations such as the Dreiding and Universal force fields. We find that a hybrid of Dreiding and TIP3P delivers the best results, yielding a density, diffusion coefficient and radial distribution function in good agreement with experiment, performing in some respects even better than the dedicated reference TIP3P model. Another Dreiding based force field predicts semi-quantitative results for the water structure and dynamics while the Universal force field based models are incapable of simulating a condensed phase of water at all, continuing to expand indefinitely. These observations are useful for selecting and designing robust water force field parameterisations that can be used for general simulation purposes.     

Highly sampled tetranucleotide and tetraloop motifs enable evaluation of common RNA force fields

Recent modifications and improvements to standard nucleic acid force fields have attempted to fix problems and issues that have been observed as longer timescale simulations have become routine. Although previous work has shown the ability to fold the UUCG stem–loop structure, until now no group has attempted to quantify the performance of current force fields using highly converged structural populations of the tetraloop conformational ensemble. In this study, we report the use of multiple independent sets of multidimensional replica exchange molecular dynamics (M-REMD) simulations with different initial conditions to generate well-converged conformational ensembles for the tetranucleotides r(GACC) and r(CCCC), as well as the larger UUCG tetraloop motif. By generating what is to our knowledge the most complete RNA structure ensembles reported to date for these systems, we remove the coupling between force field errors and errors due to incomplete sampling, providing a comprehensive comparison between current top-performing MD force fields for RNA. Of the RNA force fields tested in this study, none demonstrate the ability to correctly identify the most thermodynamically stable structure for all three systems. We discuss the deficiencies present in each potential function and suggest areas where improvements can be made. The results imply that although “short” (nsec-μsec timescale) simulations may stay close to their respective experimental structures and may well reproduce experimental observables, inevitably the current force fields will populate alternative incorrect structures that are more stable than those observed via experiment.     

Flexibility of the Bacterial Chaperone Trigger Factor in Microsecond-Timescale Molecular Dynamics Simulations

The bacterial chaperone trigger factor (TF) is the first chaperone to be encountered by a nascent protein chain as it emerges from the ribosome exit tunnel. Experimental results suggest that TF possesses considerable conformational flexibility, and in an attempt to provide an atomic-level view of this flexibility, we have performed independent 1.5-μs molecular dynamics simulations of TF in explicit solvent using two different simulation force fields (OPLS-AA/L and AMBER ff99SB-ILDN). Both simulations indicate that TF possesses tremendous flexibility, with huge excursions from the crystallographic conformation caused by reorientations of the protein’s constituent domains; both simulations also predict the formation of extensive contacts between TF’s PPIase domain and the Arm 1 domain that is involved in nascent-chain binding. In the OPLS simulation, however, TF rapidly settles into a very compact conformation that persists for at least 1 μs, whereas in the AMBER simulation, it remains highly dynamic; additional simulations in which the two force fields were swapped suggest that these differences are at least partly attributable to sampling issues. The simulation results provide potential rationalizations of a number of experimental observations regarding TF’s conformational behavior and have implications for using simulations to model TF’s function on translating ribosomes.     

Comparison of force fields for Alzheimer's A : A case study for intrinsically disordered proteins

Intrinsically disordered proteins are essential for biological processes such as cell signalling, but are also associated to devastating diseases including Alzheimer's disease, Parkinson's disease or type II diabetes. Because of their lack of a stable three‐dimensional structure, molecular dynamics simulations are often used to obtain atomistic details that cannot be observed experimentally. The applicability of molecular dynamics simulations depends on the accuracy of the force field chosen to represent the underlying free energy surface of the system. Here, we use replica exchange molecular dynamics simulations to test five modern force fields, OPLS, AMBER99SB, AMBER99SB*ILDN, AMBER99SBILDN‐NMR and CHARMM22*, in their ability to model Aβ₄₂, an intrinsically disordered peptide associated with Alzheimer's disease, and compare our results to nuclear magnetic resonance (NMR) experimental data. We observe that all force fields except AMBER99SBILDN‐NMR successfully reproduce local NMR observables, with CHARMM22* being slightly better than the other force fields.